Contaminated Sediments PDF

STP 1442
Contaminated Sediments:
Characterization, Evaluation,
Mitigation~Restoration, and
Management Strategy Performance
Jacques Locat, Rosa Galvez Cloutier, Ronald Chaney,

and Kenneth Demars, editors
ASTM Stock Number: STPI442
INTERNATIONAL
ASTM International
100 Barr Harbor Drive
PO Box C700
West Conshohocken, PA 19428-2959
Printed in the U.S.A.
Library of Congress Cataloging-in-Publication Data

ISBN: 0-8031-3466-5
Contaminated sediments : characterization, evaluation,
mitigation/restoration, and management strategy performance / Jacques
Locat...[et al.].
p. cm. - - (STP ; 1442)
"ASTM stock number: STP 1442."
"Second International Symposium on Contaminated Sediments ... in Quebec
City, Canada on May 26-28 May 2003"--Foreword.
Includes bibliographical references and index.
ISBN 0-8031-3466-5
1. Contaminated sediments-Management-Congresses.2. Soil
remediation-Congresses. I. Locat, Jacques. I1. International
Symposium on Contaminated Sediments (2nd : 2003 : Quebec, Quebec) III. Series:
ASTM special technical publication ; 1442.
TD878.C663 2003
628.5'5-dc21
2003049606
Copyright 9 2003 ASTM Intemational, West Conshohocken, PA. All rights reserved. This material
may not be reproduced or copied, in whole or in part, in any printed, mechanical, electronic, film, or
other distribution and storage media, without the written consent of the publisher.
Photocopy Rights
Authorization to photocopy items for internal, personal, or educational classroom use,
or the internal, personal, or educational classroom use of specific clients, is granted by
ASTM International (ASTM) provided that the appropriate fee is paid to the Copyright
Clearance Center, 222 Rosewood Drive, Danvers, MA 01923; Tel: 978-750-8400; online:
http://www.copyright.com/.
Peer Review Policy

Each paper published in this volume was evaluated by two peer reviewers and at least one editor.
The authors addressed all of the reviewers' comments to the satisfaction of both the technical
editor(s) and the ASTM Intemational Committee on Publications.
To make technical information available as quickly as possible, the peer-reviewed papers in this
publication were prepared =camera-ready" as submitted by the authors.
The quality of the papers in this publication reflects not only the obvious efforts of the authors and
the technical editor(s), but also the work of the peer reviewers. In keeping with long-standing
publication practices, ASTM International maintains the anonymity of the peer reviewers. The ASTM
International Committee on Publications acknowledges with appreciation their dedication and
contribution of time and effort on behalf of ASTM International.
Printed in Ann Arbor, MI

2003
Foreword
The Second International Symposium on Contaminated Sediments: Characterization, Evaluation,
Mitigation/Restoration, and Management Strategy Performance in Quebec City, Canada on 26-28
May 2003 is sponsored by ASTM International Committee D18 on Soil and Rock. The symposium
chairs and co-chairs of this publication are Jacques Locat, Laval University (CGS) and Rosa GalvezCloutier, Laval University (CSCE, ASTM); and Ronald C. Chaney, Humboldt State University
(ASTM) and Kenneth Demars, University of Connecticut (ASTM).
iii
Contents
vii
OVERVIEW
SECTIONI: SEDIMENT CHARACTERIZATION

The Origin and Behavior of a Flood Capping Layer Deposited on Contaminated
Sediments of the Sagnenay Fjord (Quebec)---EMILmNPELLETmR,GASTONDESROSmRS,
JACQUES LOCAT, ALFUNSO MUCCI, AND HI~L~NE TREMBLAY (KEYNOTE PAPER)
The Weathering Behavior of Contaminated Industrial Sediments after Their Exposure

to Atmospheric O x y g e n ~ M I C H A E L SCHUBERT, PETER MORGENSTERN,
RAINER WENNRICH, KLAUS FREYER, ALBRECHT PASCHKE, AND HOLGER WEISS
Deep-Freeze Sampling Methods for Soft Sediments--MATHIASRICKING

ANDTOBIASSCHULZE
19
28
Quality Evaluation of Eutrophic Sediments at St. Augustin Lake, Quebec, Canada-ROSA GALVEZ-CLOUTIER, MARIE-EVE BRIN, GERARDO DOM1NGUEZ, SERGE LEROUEIL,
AND SYLVAIN ARSENAULT
35
Trace Metal Levels in Nearshore Sediments Close to Industrial Discharges off

Cuddalore (Bay of Bengal)---THRESIAMMA JOSEPH, K. K. BALACHANDRAN.
MAHESWARI NAIR, V. KESAVADAS, K. K. C. NAIR, AND JOSEPH SEBASTIAN PAIMPILLIL
53
Randomization Tests: A Statistical Tool to Assess Heavy Metal Pollution in Car River
Basin Sediments (RS, Brazil)----MARIALUOAK. RODRIGUES,
MARIA TERESA RAYA-RODRJGUEZ~ AND VALI~RIO D. PILLAR
62
Study of the Geochemical Distribution of Heavy Metals in Sediments in Areas

Impacted by Coal Mining--ELBA C. TEIXEIRA,MARIALUCIAK. RODRIGUES,
MARTA F. C. ALVES, AND JANE R. BARBOSA
72
Characterization of a Catastrophic Flood Sediment Layer: Geological, Geotechnical,

Biological, and Geochemical Signatures----Ht~L~NETREMBLAY,GASTONDESROSERS,
JAQUES LOCAT, ALFONSO MUCCI, AND i~MILIENPELLETIER
Characterization of Contaminated Sediments in Hamilton Harbour, Lake Ontario-ALEXJ. ZEMAN AND TIMOTHY S. PATTERSON
87
102
vi
CONTENTS
In Situ Flume Measurements of Sediment Erodability in Saguenay Fjord

( Q u e b e c , C a n a d a ) - - - A N N - L A U R E MOREAU, JACQUES LOCAT, PHILIP HILL,
BERNARD LONG, AND YVON OUELLET
119
SECTIONII: MITIGATION AND RESTORATION METHODS

Reclamation Using Waste Sediment by Sand Capping Technique--THIAM SOONTAN,
GODAKURU P. KARUNARATNE, VICTOR CHOA, AND MYINT WIN BO (KEYNOTE PAPER)
]41
Durability Study for Geotextile Tube Use in Talimu River Sediment Control-DAVE TA-TEH CHANG, CHAO-PING SUNG, BOR-LING CHEN, AND NUAN-HSUAN HO
156
Factors Controlling Contaminant Transport Through the Flood Sediments of the

Saguenay Fjord: Numerical Sensitivity Analysis---SIBYLLEDUERi
167
AND RENI~ THERRIEN
Steps for Restoration of a Polluted Egyptian Closed Lagoon on the Alexandria

Mediterranean Coast---OSMANA. EL-RAYIS
183
Rehabilitation of Brine-Saturated Sediments--MOiR D. HAUG,S. LEEBARBOUR,

194
AND CHRIS A. JOHNS
Use of Selective Sequential Extraction for the Remediation of Contaminated

Sediments--CATHEP,]NEN. MULLIGAN AND BEHNAZ DAHR AZMA
208
Effects of Overburden Stresses on Soil Reclamation by ElectroosmosiS--ANSM KABIR

AND MIGUEL PICORNELL
224
A Reactive Geocomposite to Remediate Contaminated, Subaqueous

S e d i m e n t s - - - T H O M A S C. SHEAHAN, AKRAM ALSHAWABKEH, LORETTA A. FERNANDEZ,
AND KAREN S. HENRY
236
SECTIONIll: MONITORING AND PERFORMANCE

Sediment Transport and Deposition Processes Near Ocean Outfalls in Southern
C a l i f o r n i a - - H O M A J. LEE, MARLENE A. NOBLE, AND JINGPING XU (KEYNOTE PAPER)
Numerical Model for Contaminant Transport in Consolidating Sediments-PATRICKJ. FOX
253
266
Assessment of the Lead Release from Cables Buried in Sediments into the Water
C o I u m n - - A N T O N I N A DEGTIAREVA, MARIA ELEKTOROWICZ, AND TAGH! EBAD!
282
Numerical Modeling of Hydrodynamic Circulation and Cohesive Sediment Transport

in Hartwell Lake, South Carolina/Georgia, USA--~EBNEMELClANDPAULA. WORK 296
Retention of Heavy Metals in the Post '96 Flood Sediment Layer Deposited in the
Saguenay River, Quebec, Canada--ROSA GALVEZ-CLOUTIER,MYRIAMMURIS,
JACQUES LOCAT, AND CYRIL BOURG
Index
310
323
Overview
Recent advances in our understanding of contaminated sediments have been assembled in this
Special Technical Publication, which is one of the major scientific contributions to the Second
International Symposium on Contaminated Sediments held in Qurbec City from May 26 to 28, 2003.
This volume is part of the overall technical program of ASTM Committee D 18 on soil and rocks.
For many decades, waterways have been exposed to a wide variety of contaminants. Even if regulations and a better control of contaminants have been established to reduce their emission, many contaminants are still present in bottom sediments. In fact, some of them are persistent and continue to
pose a potential risk to the environment with direct and cumulative toxic impacts on aquatic life, organisms, and eventually on human health.
In recent years, major advances have been made in the study and understanding of contaminated
sediments, particularly via major projects in areas such as Los Angeles, California, Saguenay Fjord
in Quebec, and Singapore.
The symposium covers the areas of sediment characterization, contaminant evaluation, mitigation/restoration methods, and management strategy performance from the geological, geotechnicai,
biological, and geophysical perspectives. It reviews recent advances in contaminated sediments-management-related research and focuses on engineering aspects of contaminant transport, erosion, stability, monitoring, and modeling. The main goal of the symposium is to identify both established and
innovative physico-cbemical and biological tests and methods used to characterize and evaluate properties and behavior of contaminated sediments, as well as the potential for contaminant transfer.
The papers gathered in this publication cover the primary goal of the symposium and reflect research activities in many parts of the world. Keynote papers, selected for this volume, reflect recent
work carried out on large coastal investigations (e.g., in the Los Angeles area), and on natural and artificial capping of contaminated sediments. Other papers in this volume have been assembled into
three groups: (!) sediment characterization, (2) mitigation and restoration methods, and (3) monitoring and performance. Each of these sections begins with the corresponding keynote paper.
Sediment characterization of contaminated sediments has become more and more complex. It involves ex situ techniques from standard tests (e.g., physical properties) to biological analyses in
addition to all the chemical analyses, but also in situ ones like erodability tests. Mitigation and
restoration methods assembled herein are diversified and touch on many different environments from
river sediments and harbor lagoons to land reclamation. It involves techniques ranging from the use
of geotextiles and geocomposites to selective sequential extraction methods. The monitoring and performance aspects of contaminated sediments are largely supported by extensive site investigations,
like the Southern California project, but also by the development of modeling tools.
A few papers included in this volume summarize a five-year research effort aimed at evaluating
the performance of a catastrophic capping layer resulting from the major 1996 Saguenay flood disaster that proved to be very beneficial to the Saguenay Fjord environment and ecosystem by covering most of the ancient contaminated sediments!
Finally, the editors would like to thank all contributing authors for their effort and timely response.
This book represents the achievements of a process strongly supported by various learning societies
or agencies, including ASTM International (Committee D-18), the Canadian Geotechnical Society,
viii
OVERVIEW
the Canadian Society of Civil Engineering, the Society for Environmental Toxicology and Chemistry
(St. Lawrence Chapter), and the National Science and Engineering Research Council of Canada. The
Editors are very grateful to Mrs. H61~ne Tremblay, Secretary of the Symposium, and to Mrs. Crystal
Kemp for their dedication towards ensuring the completion of this Special Technical Publication.
Jacques Locat
Laval University(CGS)
Rosa Galvez-Cloutier
Laval University(CSCE, ASTM)
Section I: Sediment Characterization
Emilien Pelletier, ~Gaston Desrosiers, 1 Jacques Locat, 2 Alfonso Mucci,3 and H61~ne
Tremblay2
The Origin and Behavior of a Flood Capping Layer Deposited on Contaminated
Sediments of the Saguenay Fjord (Quebec)
Reference: Pelletier, E., Desrosiers, G., Locat, J., Mucci, A., and Tremblay, H., "The
Origin and Behavior of a Flood Capping Layer Deposited on Contaminated
Sediments of the Saguenay Fjord (Quebec)," Contaminated Sediments:
Characterization, Evaluation, Mitigation~Restoration, and Management Strategy

Performance, ASTMSTP 1442, J. Locat, R. Galvez-Cloutier, R. C. Chaney, and K. R.
Demars, Eds., ASTM International, West Conshohocken, PA, 2003.
Abstract: The upper section of the Saguenay Fjord was impacted by a catastrophic flood
in July 1996. Contaminated sediments were capped by a layer of clean silty post-glacial
sediments with background levels of trace metals and polycyclic aromatic hydrocarbons
(PAHs). The capping layer was characterized by geotechnical and geochemical methods
and its biological recolonization was monitored by annual sampling of the macrofauna.
The strong dominance of surface deposit feeders Cirratulidae and Ampharetidae was
observed at most stations in the first 2-3 years followed by carnivorous annelids such as
Lumbrineridae and Nephtidae species indicating a well recolonized benthic habitat in the
Baie des Ha!Ha!. The presence of benthic fauna was a major factor in modifying the
density of sediments by physical mixing and irrigation, and in changing the surface
roughness. The slope stability of the capping layer is considered as very good except in
limited deltaic sectors at the head of the Baie des Ha!Ha! The new layer showed a good
efficiency to isolate contaminated sediments from the sediment/water interface. Although
manganese and iron were remobilized as the new layer became anoxic, mercury, arsenic
and PAHs showed a very limited mobility through the flood layer which allowed
geochemists to calculate the present fluxes of toxicants to the Baie des Ha!Ha! without
interferences from older contaminated sediments trapped below the flood layer. The
present flux of PAHs to bay (0.9 ng.cm'2.g q) is about 300 times lower than the 1974 flux
and 8 times lower than the 1986 flux.
Keywords: Saguenay Fjord, flood layer, contaminated sediments, extreme
meteorological event, biogeochernical barrier, metal remobilization, bioturbation, layer
instability
'Professors, Institut des sciences de lamer (ISMER), Universit6 du Qu6bec
Rimouski, 310 all6e des Ursulines, Rimouski (Que) Canada G5L 3AI.
2Professor and post-doctoral fellow, respectively, D6partement de g6ologie et g6nie
g6ologique, Universit6 Laval, Qu6bec (Que), Canada G1K 7P4.
3Professor, Department of Earth & Planetary Sciences, McGill University,
3450,University St., Montreal (Que) Canada H3A 2A7.
3
Copyright 9 2003 by ASTM International
www.astm.org
CONTAMINATED SEDIMENTS
Introduction
The Saguenay Fjord is the largest fjord in eastern Canada. Its long, narrow glaciallyscoured submerged valley joins the St. Lawrence Estuary at Tadoussac, about 200 km
downstream of Quebec city. Typical of classical fjords, it has a U-shaped cross-section
and two shallow sills that subdivide it into two distinct basins. The upstream basin
reaches a maximum depth of 280 m near Baie EternitY. Intrusions of cold and dense
seawater from the intermediate layer of the St. Lawrence Estuary are responsible for the
periodical renewal of the fjord deep waters (Seibert et al. 1979). The upper section of the
fjord opens in two branches, giving birth to the North Arm toward the Saguenay River
and the Baie des Ha!Ha!. The Saguenay River is the main freshwater tributary of the
fjord with an average annual flow rate of 1600 mJs-I and spring flood flow rates rarely
exceed 3000 m3s"l due to the presence of a number of dams and dikes regulating the
outflow of the Saguenay drainage basin.
Figure 1 - Map of the Upstream Section of the Saguenay Fjord Showing in Dark Gray
the BenthicArea Directly Affected by the 1996 Flood.

On July 18th, 1996, heavy rain started in the Saguenay region and continued for 3
more days, resulting in precipitation of about 200 mm with peaks at 280 mm in some
areas (Nicolet et al 1997). Because of a wet early summer and high moisture conditions,
this precipitation event caused widespread flooding in the steep catchments along the
Saguenay River and Fjord. The resulting floods caused over $1 billion in damages as the
Chicoutimi and Ha! Ha! rivers made their way through the city of Chicoutimi and the
PELLETIER ET AL. ON FLOOD CAPPING LAYER
industrial town and port terminal of Ville La Bale. The incision of new river channels and
widening of existing channels through late-glacial deltaic fill sediments and the transport
of an estimated 15 million m 3 of solids to the Saguenay Fjord resulted in a massive
sedimentation event in the North Arm of the Fjord and the Baie des Ha!Ha!. Shortly after
this catastrophic event, a multidisciplinary team of marine scientists (physical
oceanographers, geological engineers, geochemists, chemists and biologists) from
universities and governmental agencies was assembled to study the effects and behavior
of the capping layer deposited on contaminated sediments of the Saguenay Fjord. This
huge research effort mainly supported by NSERC (Natural Sciences and Engineering
Research Council of Canada) and ALCAN International Limitre was motivated by the
uniqueness of the event and the challenge of integrating the knowledge from many
disciples into a conceptual model which could serve as a unique source of information for
those involved in the management of further catastrophic events occurring in coastal
environments.
The objective of this paper is to provide a first summary of the main findings on the
performance of the capping layer either for geological, geotechnical and geochemical
parameters, and chemical and biological indicators.
Sediment Deposition and Capping Process

Water samples collected in the Baie des Ha!Ha! on September 1996 have clearly
shown the influence of the flood on the abundance, distribution and chemical
composition of suspended particle matter (SPM). The abundance o f SPM in surface
water (0-5 m) was still 2-3 times higher than values recorded before 1990 (Pelletier et al
1999a). The deep water layer of the Baie des Ha!Ha! was still loaded with about 5 mg.L-l
of terregeneous particles (low organic carbon content), a concentration about 15 to 20
times higher than values recorded 20 years before the flood (Sundby and Loring 1978).
Sediment box cores, collected in the North Ann and the Baie des Ha!Ha! three
weeks after the flood revealed that 10 to 50 cm of a fluidic sediment (with a shear
strength <0.5 kPa) were deposited in both sections o f the Fjord (Fig. 1). The flood
deposit was still clearly visible in cores collected in Bale des Ha!Ha! in 1998 and years
after as a lighter brown-gray layer on the gray-black indigenous sediment. Grain size
analysis of 13 samples of surface sediment (0-5 mm) collected in September 1996 in the
Baie des Ha!Ha! showed a constant composition of the surface layer with an average high
water content o f 57 % in weight, and a composition of 18% clay, 72% silt and 10% sand
(Pelletier et al 1999b). Chemical analysis revealed a relatively high contain in inorganic
carbon content (up to 0.5%) but very low concentrations in mercury, lead, and arsenic.
Higher levels of inorganic carbon were attributed to detrital, calcareous materials from
the post-Winsconsin marine clays transported by the flood event.
Subaqueous capping is known as an efficient means for isolating contaminated
sediments and residues from the aquatic environment (Zeman 1994). The natural capping
process which occurred in the Baie des Ha!Ha! and North Arm in July 1996 covered an
area of about 60 km 2 at water depth varying from 50 to 180 rn, The thickness of the new
sediment layer, as estimated a few weeks after the flood and before the consolidation
process, ranged from about 60 cm at station 2 near the delta of the small Rivirre des
Ha!Ha! close to the city o f L a Baie to a few mm in the deepest basin of the Oord near
Baie Eternit6 (Fig. 1). The North Arm also received a large volume of post-glacial
sediments and the capping layer was estimated to 40 cm near Saint-Fulgence to 5-10 cm
at the junction with the Bale des Ha!Ha!
During the flood, the sediments were most probably transported into the Fjord by
hyperpycnal flow meaning that the mass of sediments was heavy enough to sink to the
seafloor, increasing the turbidity, current regime, and erosion of the seafloor. The
modification of the current regime has resulted in the formation of many dunes on the
main flowing channels, typical of strong currents (Tremblay et al. 2001). The bedmaterial load was principally transported by traction and suspension and by turbidity
currents (Cremer et al. 2002; Tremblay et al. 2003). The coarser particles were
transported by a traction process and were mostly accumulated at the river deltas. The
finer particles were held temporarily in suspension by fluid turbulence and were
deposited gradually in a fining downstream sequence. Following the quick accumulation
and progradation of the deltaic structures, gravitational movements on the unstable delta
f~ont created successive sequences of deposition in the basin (Cr6mer et al. 2002). Each
episodic deposition was constituted of fine sandy laminations at its base, topped with
clayey silt.
To evaluate the extent of the flood layer and its morphology, a series of box core
samples, geophysical and multibeam sonar surveys (SIMRAD EM1000) were carried out
and compared with pre-flood surveys. The extent of the covered area is shown on
backscatter images (Fig. 2).
Figure 2 - Temporal Evolution of the Backscatter Data Over Multibeam Sonar
Surveys (modified from Urgeles et al. 2002)

On these images, a difference in backscatter of sediments before (in 1993) and after
the flood (in 1997) can be observed. The darker zones in 1997 correspond to the limits of
the capping layer that seems to end just downstream to the eordluence of the Bale des
Ha!Ha! and the North Arm (Urgeles et al. 2002). Box cores revealed a thickness varying
from almost nothing to 60 cm in most of the covered area, generally decreasing in the
downstream direction. The accumulation of sediment reached 7 m at the river mouths of
the Bale des Ha!Ha! (Kammerer et at 1998).
Anthropogenic Contamination of the Fjord
The rapid expansion oftbe metallurgic industry and urbanization in the Saguenay
- Lac Saint-Jean region in years 1940-1970 contributed to the degradation o f the aquatic
environment, and particularly the Saguenay Fjord. The presence of high concentrations
of mercury (Hg) in the sediments of the Fjord was first pointed out by Loring (1975) who
observed concentrations reaching 6 to 12 mg.kg"t in the North Arm and 3 to 6 mg.kg-~ in
the Baie des Ha!Ha! for sediment sampled in 1964. The flux of rig to sediments to the
fjord reached 60 000 ng.cm2.yr~ in early 1970s and then was progressively reduced with
the cutbacks of liquid effluents from a chlor-alkali plant in 1970-1974 (Smith and Loring
1981). An estimated 60 tons of mercury still reside in the sediments at sediment depths
varying between 15 cm and over one meter. In addition to Hg, other trace metals such as
cadmium (Cd), chromium (Cr), copper (Cu), lead (Pb), and zinc (Zn) have been found in
elevated concentrations in the sediments of the fjord, and their presence has been
attributed to the industrial and urban activities of Chicoutimi and other upstream human
communities (Gagnon et al. 1993).
The flux ofpolycyclic aromatic hydrocarbons (PAHs) to the fjord has been
estimated at 11 000 ng.cm'2.yr~ in 1966 in the North Arm (Martel et al. 1987). This flux
was reduced by one order of magnitude between 1970 and 1981 with the enforcement of
new environmental regulations in Canada.
Hg (mg.kg "~)
0
0
3 0
JO
Cd (mg.kg "1)
1
2
Total PAIls (mg.kg "I)

3
10
10
~v~20
20
2O
4O
40
40
5O
50
50
I0
5 10 1 5 2 O 2 5
Figure 3: Hg, Cd and Total PAHs Profiles From a Core Sampled in 1991 in the North
Arm near St. Fulgence. The Local Sedimentation Rate is Approximately 3 cm.yr "1.
CONTAMINATEDSEDIMENTS
Changes in Hg, Cd and PAHs concentrations in a sediment core collected in the

North Arm in 1991 are shown in Fig. 3. Total PAHs decreased from about 20 mg.kg"~ to
less than 5 mg.kg-I within 17 years whereas Cd concentrations also decreased by a factor
of 4-5 during the same period. The behavior of mercury was more puTgling as the
concentration in the core seems to increase in the early1980s and then decreased again in
the early 1990s to a value of 0.4 mg.kg-I (two times the pre-industrial background level)
The presence of PAHs in the Baie des Ha!Ha! and elsewhere in the Fjord was first
reported by Martel et al. (1987) who found total PAHs concentrations reaching 3.5
mg.kg1 at a sediment depth corresponding to the early 1970s. The flux of PAHs to the
bay was estimated to be 279 ng.cm2.yr"~ in 1974, a value about 25 times lower than the
flux to North Arm at the same period. The explanation for such a difference hold in the
fact that most of the PAH-loaded sediments came from the Saguenay River, and the Baie
des Ha!Ha! was only indirectly touched by the contamination. A new core was sampled
in the bay in 1988 and PAHs were analyzed (Ouellet 1990). The top 40 mm of the core
showed an average total PAHs concentration of 1.83 mg.kg"1 confirming the important
reduction already observed in the North Arm although the sedimentation rate in the bay
was <0.3 cm.yr-~.
In the last few years, a new environmental problem surfaced with the use of
antifouling paints on large commercial and navy ships. The powerful biocide tributyltin
(TBT) contained in these paints is released into the water column on the passage of the
vessels and accumulated in suspended particulate matter and sediments along waterways
and near important harbors. The Saguenay Fjord acts as a sink for TBT because of the
slow renewal process of its deep water mass. Concentrations over 100 ng.g"~ of total
organotins (expressed as Sn) have been observed in sediments and organisms collected in
the Baie des Ha!Ha! and elsewhere in the Fjord from 1999 to 2001.
The Performance of the Capping Layer

Geotechnical Results
The flood layer is a turbidite, except at the fiver mouths, usually characterized by a
thin sandy layer at the bottom. The grain size distribution varies from sand to clay. Near
the fiver mouths, the sediments were coarser and mainly constituted of sand and gravel.
Compared to sediments existing before the flood, they have higher water content, a
lower consistency and a lower plasticity index (Maurice et al. 2000). In the North Arm,
the new layer is harder to identify because the sandy layer at its base is very thin or
absent, and an important bioturbation process has quickly erased the signature of the
turbidite through mixing of sediments.
The temporal evolution of some sediment properties was monitored using
backscatter data (Fig. 2). The intensity of the hackscattering depends on the water
content of the sediment, its density and surface roughness (Urgeles et al. 2002).
Sediment with a low water content shows high hackscattering which increases with the
increasing density. On the 2001 image the extent of the dark zones is quite similar to the
1993 image, meaning that the properties of the flood layer have returned to the pre-flood
conditions.
In their recent work, Maufice et al. (2000) estimated that the consolidation of the
layer was completed within 3 months after the flood. Authors also noted that
bioturbation, easily observed in cores, clearly influenced the consolidation process of the
newly deposited sediments. On the non bioturbated profile (Fig. 4a) one can easily
determine the thickness of the flood layer by quick changes of the water content, intact
(C,) and remolded (Cur) undrained shear strength and liquidity index (IL). In the flood
layer, the water content (w %) decreases regularly with depth and the shear strength
increases following the consolidation process. The peaks between 25-30 cm correspond
to the sandy layer. The liquidity index is high, around 4, meaning that the consistency is
very low, like sludge. For samples where bioturbation is well established (Fig. 4b), an
reverse situation is observed. Water content increases with depth, the resistance
decreases and the sediment is more consistent. These changes are directly attributed to
the action of organisms that drain and mix the sediment, causing a decrease in the water
content and an accelerated densification. Without bioturbation, the value OflL is around
4 at the sediment surface with a slight decrease with depth, whereas it rapidly decreases
to a value of 2 when bioturbation is present.
w%
100
Cu (kPa)
200
0
0.0
10
IL
2.5
w%
5.0
0
0
' I '
100 200
0
cu (kPa)
IL
0.0 2.5 5.0

10
' II
.w
o~
'li
3o
40
50
40 ~
(b) bioturbated
,
,*P
~
(a) r~n-bioturbat~l
Figure 4 - Geotechnical Profiles o f the 1996 Turbidite

a) Non Bioturbated Sample Taken in the Baie des H a / H a / i n 1997 and b) Bioturbated
Sample Taken in the North Arm in 1998 (Modified from Maurice et al, 2000)
The profile in Fig. 4b shows that the organism activities contribute to an increase in
the shear strength of sediments at the surface. For bioturbated sediments, Cu is around 5
kPa while it is less than 1 kPa for non-bioturbated sediments. Consequently, this increase
in Cu at the sediment surface contributes to increase the resistance to erosion. To
evaluate if the bottom current in the Fjord could erode the sediment surface, a study on
the erodability has been conducted by measuring the in situ critical shear stress and
erosion rates using a circular benthic flume (Miniflume) (Moreau et al. 2003). The
results reflect the spatial variability of sediment properties observed for the bioturbated
flood layer (Tremblay et al. 2003), with critical shear stress (re) varying between 0.06 and
0.44 Pa (corresponding to current velocity of 5.4 and 14.7 cm.s'l), and the mean erosion
10
rate for total erosion phase (Eme=.)between 3.42 xl0 4 and 1.25 xl0 -5 kg.m2.s l. These
results reveal that the surface layer could be resuspended even by low currents.
Geochemical Results
The distribution of four metals (Fe, Mn, Hg, and As) was monitored in the sediments
and their response to diagenetic processes following the deposition of the flood material.
The geochemical cycles of iron and manganese in sediments dictate the diagenetic
behavior o f many elements since their oxides are strong adsorbents and play a pivotal
role in determining the distribution and partitioning of other elements in the sedimentary
column (Gagnon et al. 1997; Mucci et al. 2000)
Under stead-state conditions, such as those encountered at the deep inner basin near
Baie l~temit6 (Fig. 1, St 30), which received only an ineffectual, discontinuous, thin film
of the flood material, reactive Fe and Mn (i.e., soluble in a 1M HCI solution) are
concentrated in the oxic sediments (Fig. 5). The concentration of the reactive Fe and Mn
oxides near the sediment-water interface results from the reductive dissolution of
authigenic and detrital oxides as they are buried below the oxygen penetration depth
(-5ram) and dissolved Fe(II) and Mn(II) diffuse back to the oxic layer where they are
oxidized and precipitated (Mucci and Edenborn 1992). The highest arsenic
concentrations were also found near the sediment/water interface where this element is
actively concentrated by adsorption onto the iron oxides (Fig. 5). The distribution of rig
in these sediments reflects, mostly, the history of metal discharge to the Fjord. Similar
metal accumulation profiles in the sediments were observed in 1991 in the North Arm
(Fig. 1). Sulfate reducing conditions are encountered within the first two cm below the
sediment/water interface as evidenced by the porewater sulfate depletion and/or the
appearance of acid volatile sulfides (AVS) in the solid sediments (Fig. 6; Gagnon et al.
1995; Mucci et al. 2000a).
Solid Mn(HCI) (ppm)
0 .. 1000 2000 3000.
-Ill-- Solid Fc(HCI) (wt%)

3
4
~- 2o
~ 30
,o i
50
- ' .................................
zx Porewater Mn (ppm)
o PorewaterFe (ppm)
Solid As(tot) (ppm)

0
10
9.,- - ~
20
, .
--r
30
.'i.~ '
o Porewater As (ppm)
Solid Hg(tot) (ppb)

1000
2000
/
I
\
.,.,,,
. . . . . . . . . . .
,..
- - AVS gmol g ' )
Figure 5 - Steady State Diagenetic Conditions in the Fjord as Represented by the

Geochemistry of Sediments at Station 30, a Site Unaffected by the Flood Deposit.
Manganese remobilization occurred almost unimpeded following the emplacement

of the flood deposit. As indicated in the abridged time-series from station 9 in the Bale
11
des Ha! Ha! (Fig. 6 a and b), authigenic manganese oxides (i.e., Mnrtc0 present at the
original sediment-water interface and delivered with the flood material were reduced and
most of the Mn(II) diffused freely to the newly established interface where it was
oxidized and re-precipitated. Dissolved Mn (II) profiles show that it originates both from
the former sediment-water interface and the dissolution of detrital oxides delivered with
the flood material (Fig. 6 a and b). The shape of the dissolved manganese profiles is
dictated by the strength of the sources (i.e., reactivity of reducible phases at the former
sediment/water interface and within the flood deposit) and sinks (i.e, precipitation of
authigenic oxides close to the new sediment/water interface, the formation of a mixed
carbonate and adsorption to AVS at depth (Mucei and Edenborn 1992; Saulnier and
Mucci 2000).
(a)
--a- Solid Mn(HCl) (ppm)
00
400
YS~
-a-- Solid Fe(HCI) (wl%)
-4-- Solid As(to0 (ppm)
10
'
Porewater Mn (oom)
~ pmewater Fe (ohm)
(b)
---a- Solid Mn(HCI)(ppm)
+
Q
1000 2000 3000
Solid Fr
1
2
(wt%)
_3 4
20
30 0
Solid Hg(to0 (ppb)
100 200 300 400
t~16
c~ Potewater As (m)m~
~
0
Solid As(tot) (pOre)

10
20
30
~
0
Solid Hg(tot) (0pb)

100 200 300 400
10
g 2o
4O
50
o: i
A~
'-'~ ....
Porewater Mn (ppm)
"t
D PoerewatcrFe(ppm)
o Porewa~r As(ppm)
--
AVS (ixrnoLg-j
Figure 6 - Evolution of Sediment and Porewater Geochemistry at Station 9 in the Bale

des Ha!Ha!: (a) Less than one Month after the Flood and (b) Nearly Three Years after
the Flood Event. The Dashed Line Marks the Lower Boundary of the Flood Layer.
In contrast to Mn, reactive iron associated with the authigenic oxides at the original
interface displayed limited mobility. Following the reductive dissolution of these oxides,
Fe(II) was trapped as sulfides (i.e., mostly AVS with minor authigenic pyrite) under the
sulfate-reducing conditions that were rapidly established. At station 9, the sulfides (see
12
AVS profile in Fig. 6b) are found immediately above and below the solid, Fe(HCI) peak
and the lower boundary of the flood deposit whereas a single and broader sulfide peak is
only observed below the former sediment-water interface at station 5 (data not shown).
The difference between the two behaviors may be explained by the higher sedimentation
rate (<0.2cm.yr"~ in the Baie des Ha!Ha! vs -lcm.yr "l in the North Arm) and a greater
reduction potential of the indigenous sediments in the North Arm. Mucci and Edenborn
(1992) proposed a conceptual model to explain the different, diagenetic bebaviour of iron
and manganese in the Saguenay Fjord sediments following the Saint Jean Vianney
landslide of May 1971. A revision of this model as well as a numerical model describing
the remob'dization of reactive manganese can be found in Mucci et al. (2003).
Arsenic is reduced from As(V) to As(Ill) under suboxie conditions and forms a
stable sulfide mineral under sulfidic conditions. In marine sediments, however, it usually
co-precipitates with authigenic iron sulfides (Belzile and Lebel 1986). Likewise, we
observe that most of the arsenic, which was also concentrated at the original interface by
adsorption to the authigenic iron oxides (Fig. 5) appears to be trapped by co-precipitation
with the iron sulfides (i.e., AVS; Fig. 6b), thus limiting its diffusion through the flood
deposit. The porewater arsenic profile (Fig. 6b) attests to its limited mobility through the
flood layer.
Very little mercury was delivered to the tjord with the flood material (Fig 6a). The
vertical distributions of rig in the sediments recovered since the flood event indicate that
most of the Hg is still found within the indigenous, contaminated sediments buried under
the flood deposit but a fraction was remobilized and is associated with the AVS within
the deposit (Fig. 6b). Based on the time-series data, the peak concentration at the new
sediment-water interface would originate from inputs of rig-laden particles under the
normal sedimentation regime (Fig. 6b).
Recolonization of the New Layer in the Baie des Ha!Ha!
The benthic recolonization process of the new sediment layer deposited in the Baie
des Ha!Ha! was monitored by a series of stations located on the longitudinal axis of the
bay (Fig. 1). Five stations were sampled annually using a Van Veen grab of 1/8 m 2. A
particular attention was given to stations 2 and 13 which were located at the head of the
bay near La Baie and at the junction between the North Arm and the Baie des Ha!Ha!,
respectively. For station 2, no macrobenthic organisms were found in fall 1996, but the
recolonisation process started .q2uicklyafter our sampling and the average density peaked
in 1999 with 3101 + 868 ind.m (Fig. 7). This average density suddenly dropped to 802
q- 269 ind.m"2 in 2000, and then returned to values previously observed in 2001 with 3133
314 ind.rn"2. Station 13 showed an increasing average density of organisms from 1996
to 1998 (421 82 ind.m'2), but again an unexpected progressive reduction of the biota
was observed until 2001 (128 11 ind.m'2). For stations 5, 7, 9 located a few km away
from direct sources of sediment, the densities were higher in 1996 than those observed
for the same year at stations 2 and 13. Their average density seems to strongly decrease
in 1997 and dominate in 1998 for these three stations. We again observed a decrease in
the following years at these stations, except for station 9 in 2001(176 8 ind.m-2) (Fig.
7). The Kruskal-Wallis nonparametric test showed significant differences between the
five annually sampled stations from 1996 to 2001 (Chi-square = 16,66, df = 4, p = 0.22).
13
The strong predominance of the polychaete annelids was observed for all stations in
the first year. In this particular zoological group, the dominant species responsible for the
recolonization process at station 2 is the surface deposit feeders Chaetozone setosa and
Ampharete cfarctica, and the subsurface deposit feeder Cossura longocirrata. In the
following years, even if the deposit feeders are still well represented, carnivorous
polychaetes became dominant with the strong presence of the Lumbrinerisfragilis
species and Nephtydae Aglaophamus neotenus for the other stations.
4000
. . . . . . . . . . . . . . . . . .
= 3500
I 3000
2500
2000
1500
1000
~1996
I-I1997
I-I1998
11111999
PA2000
112001
',ooo
MF'N~
2
13
Station
Figure 7- Distribution of Macrobenthic Organisms Found in Sediments of Five Stations
of the Baie des Ha!Ha!from 1996 to 2001.

Station 2 was catastrophically affected by the 1996 flood as no benthic organisms
were found at this station in September 1996 (Pelletier et al. 1999b). The Spionidae
group was supplanted in 1997 by Lumbrineridae (I,umbrinereisfragilis) and Cirratulidae
(Chaetozone setosa) in most stations. For example, station 13 was occupied by
Spionidae only in 1996, and we found at this same station in 1997 a much higher
diversity with six annelid families among which Ampharetidae (Ampharete cfarctica)
and Capitellidae (Capitella capitata) were present. In all stations after the flood, the
polychaete annelids represented the dominant benthic fauna by their density and
diversity, thus indicating a rapid reeolonisation process of this new sediment layer with
the presence of pioneer species such as Spionidae (Pearson and Rosenberg 1978). From
1997, Spionidae were gradualy replaced by species being more characteristic of deep
muddy circalittoral biotopes and species indicating poss~le organic contamination
(Pearson and Rosenberg 1978; Peletier et al. 1999b). While polychaetes were
recolonizing this new sedimentary layer they created a network of tubes and burrows
down to the ancient layer. This new bioturbation of the top layer of sediment has been
directly observed by axial tomodensitometry (de Montety et al. 2000). The high mean
density observed at station 2 is due to the high abundance of surface deposit feeders,
particularly Cirratulidae (Chaetozone setosa) and Ampharetidae (.4mpharete cf arctica).
Other stations showed a lower mean density and a gradual reduction of deposit feeders
compared to carnivores like Lumbrinereisfragilis of the family of Lumbrineridae and
Nephtydae (Aglaophamus neotenus). Station 2 received terrigenous sediment with a
relatively high organic content when compared to the other stations along the bay.
14
However, the recolonization process which quickly started at station 2 was unstable in
time and clearly related to the capping instability in that area. Indeed, the significant fall
of average densities observed at this station in 2000 was caused by the local erosion of
the sedimentary surface layer in 2000 (Michaud et al. 2001, 2002).
Occurrence of PAHs in the New Layer

In sediment sampled in the Baie des Ha!Ha! in 1988 and analyzed by Ouellet (1990)
the concentration of a major PAl-/contributor, benzo(a)pyrene (BaP), was about 0.25
mg.kg~ in the top 10 mm and increased with depth to a maximum of 1.963 mg.kg"z in
the 10-12 cm layer. Since 1998, we analyzed each year the evolution of the distribution
profile of high molecular weight PAHs in the flood deposit layer at station 9 in the Baie
des Ha!Ha!. The post-glacial sediments eroded by the flood and transported to the bay
were very poor in PAHs with a total concentration of about 0.1-0.2 mg.kg'~ which is well
below the background level (0.5 mg.kg"l) found by Martel et al (1987) in the preindustrial sediment layers of the Saguenay Fjord. The lowest concentration of BaP
(0.003 mg.kg1) was measured at 4-12 cm depth in a core collected in 1998. Interestingly,
the top 2 cm layer of the same core showed BaP concentrations ranging from 0.02 to
0.035 mg.kg-l. This thin enriched surface layer is attributed to the late deposition (within
the year after the flood) of very fine clay and black carbon particles which were enriched
in hydrophobic high molecular weight PAHs. These particles were mixed with
underlying sediments in the following years by the bioturhation ofmacrobenthic
organisms recolonizing the area. The BaP profile in 2001 was characterized by a
maximum value of 0.047 mg.kg"~ in the 5-6 cm layer with a steady decrease toward the
surface reaching 0.021 mg.kg~ in the top 0-1 cm layer. The lower values appearing near
the surface seem to indicate a very low input of new PAHs ~om the natural
sedimentation process of the bay in the last three years. Using the average concentration
of BaP in the top 5 cm (0.03 mg.kg"l) and a sedimentation rate of 30 mg.cm'2.yr"1
(Leclerc et al. 1986) the present flux of BaP in the Baie des Ha!Ha! can be estimated to
be about 0.9 ng.cm-2.yr~. This value is about 300 times lower than the 1974 flux (Martel
et al. 1987) and 8 times lower than the 1988 flux (Ouellet 1990).
Slope Stability Analysis

The nature of the capping layer in the Baie des Ha!Ha! and the North Arm varies
greatly near major slopes found along active deltas. The capping layer in nearshore areas
consists mostly of sand and gravel which have been rapidly deposited over pre-existing
organic rich fine-grained sediments. The multibeam image of the upper part of the Bale
des Ha!Ha! showed that the rapid loading of the underlaying sediments has produced
some failures visible as submarine slides (Tremblay et al. 2001). No measurements of
pore pressures are available in these sectors but as time goes on, excess pore pressure
related to the rapid loading process shall vanish. Elsewhere, Maurice et al. (2000) have
shown that all excess pore pressures due to rapid sedimentation and consolidation have
been dissipated so that for most of the area covered by the capping layer, the stability
analysis can be carried out assuming hydrostatic conditions.
The analysis of the stability of the capping layer included considerations for both
drained and undrained conditions and also seismic acceleration. On such low angle
15
slopes with such a thin layer (assume here at an average of 0.5 m) we can use infinite
slope stability for evaluating the factor of safety against failure. In such a situation, it was
hypothesized that the failure could take place under undrained conditions. The following
equation for the infinite slope stability (Lee et al. 2000) was used:
=[
kr, z cos fl
1
FS Lr~ cos a-~ -Y~-dhr~sin fl J
[1]
where k is the strength ratio (C/ty'~o, where C, is the undrained shear strength and o%0
the effective stress at a given depth), z is the depth to the failure plane, y and y' the total
and buoyant unit weight, respectively, r the slope angle and ah the equivalent seismic
acceleration. From the geotechnical investigation carded by Maurice et al. (2000) the
value ofk varies around 0.5, which is often the case for recently deposited sediments of
the Saguenay Fjord (Perret et al. 1995). Using Eq. [1], the parametric analysis has
indicated that in absence of an earthquake, slopes as high as 10 degrees are stable, even
with a k value of 0.2. Considering a seismic event similar to the one which occurred in
1988, a slope at 6 degrees is stable for an acceleration of 0.3. Even with a k value at 0.5
and the same seismic loading, slopes at 10 degrees are still stable. Therefore, it appears
that under seismic loading conditions, only limited sectors would be affected near actual
deltas of the Bale des Ha! Ha!. The actual slope stability of these sectors could also he
very dependent on remaining high pore pressures resulting from the rapid loading
generated by the capping layer. More work remains to be done in these areas to ascertain
actual stability conditions.
Conclusion
In summary, the flood capping layer was clearly identified and characterized in the
Baie des HafHal using geotechnical and geochemical tools. The presence of benthic
fauna was monitored over 5 years and was a major factor in modifying the density of
sediments by physical mixing and irrigation, and in changing the surface roughness. In
all stations after the flood, the polychaete annelids represented the dominant benthic
fauna by their density and diversity indicating a rapid recolonization process with the
presence o f pioneer species such as Spionidae. Then, Spionidae were rapidly replaced by
carnivorous species of the family of Lumbrineridae and Nephtydae. The slope stability
of the capping layer is considered as very good, except in some limited deltaic sectors at
the head of the Bale des Ha!Ha! Although manganese and iron were remobilized as the
new layer became anoxie, mercury, arsenic and PAHs showed a very limited mobility
through the flood layer. The impermeability of the flood layer trapped the old aromatic
hydrocarbons far below the sediment surface and allowed the geoehemists to calculate
the present flux of PAH to the sediment floor o f the Baie des HatHa! The
multidisciplinary approach, adopted from the beginning, was proved to be an efficient
and high productive path to reach a clear understanding of a very complex geological,
chemical and biological process. Team leaders and students are grateful to NSERC,
ALCAN International Lt~e, FCAR (Qu6bec), and Pare Matin Saguenay-Saint-Laurent for
financial support.
16
References
Cr6mer, J.-F., Long, B., Desrosiers, G., de Montety, L., and Locat, J., 2002, "Application
de la scanographie ~i l'6tude de la densit6 des s6diments et/t h earact6risation des
structures s6dimemaires: exemple des s6diments d6pos6s darts h rivi&e Saguenay
(Qu6bec, Canada) apr6s la crue de juillet 1996", Canadian Geotechnical Journal,
Vol. 39, pp. 440-450.
de Montety, L., Long, B., Desrosiers, G., Cr6mer, J.-F., and Locat, J., 2000,
"Quantification des structures biog6nes en fonction d'un gradient de perturbation
dans la baie des Ha! Ha!/t l'aide de la tomodensitom6trie axiale", Proceedings of the
53rd Canadian Geotechnical Conference, Montreal, pp. 131-135.
Gagnon, C., Pelletier, E., and Maheu, S., 1993, "Distribution of Trace Metals and some
Major Constituents in Sediments of the Saguenay Fjord, Canada," Marine Pollution
Bulletin, Vol. 26, pp.107-110.
Gagnon C., PeUetier E., and Mucci A., 1997, "Behaviour of Anthropogenic Mercury in
Coastal Marine Sediments", Marine Chemistry, No 59, pp. 159-176.
Kammerer, E., Hughes Clarke, J.E., Locat, J., Doucet, N. and Godin, A., 1998,
"Monitoring Temporal Changes in Seabed Morphology and Composition using
multibeam Sonars: A Case Study of the 1996 Saguenay River Flood", Proceedings
of the Canadian Hydrographic Conference, Victoria, Canada, pp. 450-461.
Leclerc, A., Gagnon, M.J., C6t6, R., and Rami, A., 1986, "Compaction and Movement of
Intertidal Water in Bottom Sediments of the Saguenay Fjord, Quebec, Canada ",
Sedimentary Geology, Vol. 46, pp. 213-230.
Lee, H.J., Locat, J., Dartnell, P., Minasian, D., and Wong, F., 2000, "A GIS-Based
Regional Analysis of the Potential for Shallow-Seated Submarine Slope Failures" In
Proceedings of the 8th International Symposium on Landslides, Cardiff, U.K., pp.
917-922
Loring, D.H., 1975, "Mercury in the Sediments of the Gulf of St Lawrence", Canadian
Journal of Earth Sciences, Vol. 12, pp. 1219-1225.
Martel, L., Gagnon, M., Mass6, R and Leclerc A., 1987, "The Spatio-temporal Variations
and Fluxes of Polycyclic Aromatic Hydrocarbons in the Sediments of the Saguenay
Fjord, Qu6bec, Canada", WaterResearch, Vol. 21, pp. 699-707.
Maurice, F., Locat, J., and Leroueil, S., 2000, "Caract6ristiques g6otechniques et
6volution de la touche de s6diment d6pos6e lors du d61uge de 1996 dans la Baie des
Hal Ha! (Fjord du Saguenay, Qu6bee)". Proceedings of the 53rd Canadian
Geotechnical Conference, Montreal, pp. 123-130.
17
Michaud, E., Long, B., Cr6mer, J. F., Desrosiers, G., and de Montety, L., 2001, "Les
phases de construction delta'fques, influence d'un apport catastrophique ",
Proceedings of the Canadian Coastal Conference, Qudbec, pp. 533-547.
Michaud, E., Desrosiers, G., Long, B., de Montety, L., Cr6mer, J.-F., PeUetier, E., Locat,
J., Gilbert, F., and Stora, G., 2002, "Use of axial tomography to follow temporal
changes of benthic communities in an unstable sedimentary environment : (Ha! Ha!
Bay, Saguenay Fjord), Journal of Experimental Marine Biology and Ecology, Vol.
4060, pp. 1-8.
Moreau, A-L., Loeat, J., Long, B., Hill P., and Ouellet Y., 2003, "In situ Flume
Measurements of Sediment Erodability in Saguenay Fjord (Quebec, Canada)"
Contaminated Sediments: Characterization, Evaluation, Mitigation~Restoration, and
Management Strategy Performance, ASTM STP 1442, J. Locat, R. Galvez-Cloutier,
R.C. Chaney, and K. Demars, Eds., ASTM International, West Conshohoeken, PA.
Mucci A. and Edenborn H. M., 1992, "Influence of an organic-Poor Landslide Deposit on
the Early Diagenesis of Iron and Manganese in a Coastal Marine Sediment",
Geochimica. Cosmochimica. Acta, Vo156, pp. 3909-3921
Mucci A., Richard L.-F., and Lucotte M., 2000, "The Differential Geochemical Behavior
of Arsenic and Phosphorus in the Water Column and Sediments of the Saguenay
Fjord Estuary, Canada", Aquatic Geochemistry, No 6, pp. 293-324.
Mucci A., Boudreau B. and Guignard C., 2003, "Diagenetic Mobility of Trace Elements
in Sediments Covered by a Flash Flood Deposit: Mn, Fe and As", Applied
Geochemistry, (in press)
Nicolet, R., Roy, L., Ar~s, IL Dufour, J., Marinier, G., and Morin, G., 1997, "Rapport de
la Commission scientifique et technique sur la gestion des barrages ", Bibliothbque
nationale du Qudbec, Qu6bec, Canada.
Ouellet, S., 1990, "Etude des hydrocarbures aliphatiques et aromatiques polycycliques
(HAP) dans les s~diments du Fjord du Saguenay", MSc thesis, Universit6 du
Qu6bec h Rimouski, 148 pp.
Pearson, T. H. and Rosenberg, R., 1978, "Maerobenthic Succession in Relation to
Organic Enrichment and Pollution in Marine Environment", Oceanography Marine
Annual Review, No 16, pp 229-311.
Pelletier, E., Mostajir, B., Roy, S., Gosselin, M., Gratton, Y., Chanut, J.-P., Belzile, C.,
Demers, S., and Thibault, D., 1999a, "Crue 6clair de juillet 1996 darts la r6gion du
Saguenay (Qu6bec). 1. Impacts sur la colorme d'eau de la baie des Ha!Ha! et du t]ord
du Saguenay", Canadian Journal of Fisheries and Aquatic Sciences, No 56, pp.
2120-2135.
18
Pelletier, E., Deflandre, B., Nozais, C., Tita, G., Desrosiers, G., Gagn6, J.-P., and Mucci,
A., 1999b, "Crue 6clair de juliet 1996 dans la r6gion du Saguenay (Qu6bec). 2.
Impacts stir les s~diments et le biote de la baie des Ha! Ha! Et du Fjord du
Saguenay", Canadian Journal of Aquatic Sciences, No 56, pp 2136-2147.
Perret, D., Locat, J., and Leroueil, S., 1995, "Strength Development with Burial in FineGrained Sediments from the Saguenay Fjord, Quebec", Canadian Geotechnical
Journal, Vol. 32, No. 2, pp. 247-262.
Saulnier I., and Mucci A., 2000, "Trace Metal Remobilization Following the
Resuspension of Estuarine Sediments: Saguenay Fjord, Canada", Applied
Geochemistry, No 15, pp. 203-222.
Seibert, G.H., Trites, R.W., and Reid, J., 1979, "Deepwater Exchange Processes in the
Saguenay Fjord", Journal of Fisheries Research Board of Canada, No 36, pp. 42-53.
Smith, J.N. and Loring, D.H., 1981, "Geochronology of Mercury Pollution in the
Sediments of the Saguenay Fjord, Quebec", Environmental Science & Technology,
Vol. 15, pp. 944-951.
Sundby, B. and Loring, D.H., 1978, "Geochemistry of Suspended Particulate Matter in
the Saguenay Fjord", Canadian Journal of Earth Sciences, Vol. 15, pp., 1002-1011.
Tremblay, H., Locat, J. and Urgeles, R., 2001, "G6odynamique littorale de la couche
diluvienne de 1996, Baie des Ha!Ha!, Fjord du Saguenay, Qu6bec, Canada",
Proceedings. of the 2001 Canadian Coastal Conference, Qu6bee. pp. 507-522.
Tremblay, H., Desrosiers, G., Locat, J., Mucci, A., and Pelletier, 15,, 2003,
"Characterization of a Catastrophic Sediment Layer : Geological, Geotechnical,
Biological and Geochemical Signatures", Contaminated Sediments:
Performance, ASTMSTP 1442, J. Locat, R. Glavez-Cloutier, R.C. Chaney, and K.
Demars, Eds. ASTM International, West Conshohocken, PA.
Urgeles, R., Locat, J., Schmitt, T. and Hughes Clarke J., 2002, "The July 1996 Flood
Deposit in the Saguenay Fjord, Quebec, Canada: Implications for Source of Spatial
and Temporal Backscatter Variations", Marine Geology, No 184, pp. 41-60.
Zeman, A.J., 1994, "Subaqueous Capping of Very Soft Contaminated Sediments",
Canadian Geotechnical Journal, Vol. 31, pp. 570-577.
Michael Schubert 1, Peter Morgenstem I, Rainer Wennrich I, Klaus Freyer j, Albrecht

Paschke t, and Holger Weiss I
The Weathering Behaviour of Contaminated Industrial Sediments

After Their Exposure to Atmospheric Oxygen
Reference: Schubert, M., Morgenstern, P., Wennrich, R., Freyer, K., Paschke, A., and
Weiss, H., "The Weathering Behaviour of Contaminated Industrial Sediments After
Their Exposure to Atmospheric Oxygen," Contaminated Sediments: Characterization,
Evaluation, Mitigation~Restoration, and Management Strategy Performance, ASTM STP
1442, J. Locat, R. Galvez-Cloutier, R. C. Chaney, and K. R. Demars, Eds., ASTM
International, West Conshohocken, PA, 2003.
Abstract: Mining and ore processing of"copper shale" in the Mansfeld region,
Germany, were stopped in 1990 as a consequence of the obsolete level of the applied
technology. One of the residues of the former pyrometallurgical activities is a flue dust,
which was scrubbed out of the process gases. The scrubbing resulted in an extremely
fine-grained sludge, which was stored in "ponds" where it settled down as sediment. The
need to assess the risk of heavy metal emissions posed by such sludge deposits when
they dry up and become exposed to weathering prompted the question over the depth to
which heavy metals can possibly be mobilized. A drill hole was bored into the dried-up
sediment of such a "pond" and the sediment material was analysed. The results show a
substantial influence of weathering in the upper 0.5 m of the material. In particular the
oxidation of the heavy metal sulphides abundant in the sediment and the subsequent
dissolution of the sulphates is nowadays a major problem for the quality of the nearby
ground- and surface waters.
Keywords: heavy metals, sediments, weathering, mobilization
Introduction
For more than 800 years the Mansfeld region of Saxony-Anhalt / Germany was
known for the mining of a low-grade metalliferrous Permian black shale - the
Kupferschiefer or "copper shale." However, for economical and ecological reasons
mining and ore processing stopped in 1990. Whereas the shutdown of the works had an
immediate effect on the local air quality the ground- and the surface waters even now,
about twelve years later, still suffer a continuing contamination closely linked to the
former mining activities. Now that the actual smelters are closed and demolished the
sources of that contamination are the various by-products left behind by the former
mining and ore processing.
~Research scientist, Centre for Environmental Research Leipzig-Halle, Permoserstrasse

15, D-04318 Leipzig, Germany, [email protected]
19
Copyright9 2003by ASTM International
www.astm.org
20
Two major by-products can still be found abundantly in the area. These two major
by-products are waste rocks and slag, which were piled up in huge heaps during the last
century. These dumps are today a characteristic feature of the local landscape (Wege et
al. 2000). However, a third by-product with a less striking impact on the topography but
with a noticeable negative effect on the local ground- and surface water quality is a flue
dust, which has been scrubbed out of the process gases which left the blast-furnaces at
temperatures of about 400~ The scrubbing was basically done by spraying a veil of
water into the gas stream and thereby binding the dust particles in a slurry of about 3 g/1
of solids. The slurry was transferred into conic tanks where the solids were allowed to
bed down. The excess water was finally decanted and the water content of the remaining
sludge could be reduced to about 30% w/w.
The scrubbing of the process gasses, that went on for almost a century until 1990
consequently resulted in an accumulation of several hundred thousands of tons of the
sludge. By the end of the pyrometallurgical activities in 1990 a total of about 225 000
tons of sludge had been piled up. Aside from its still substantial water content the main
components of the sludge (by dry weight) were Zn ( - 20%), Pb ( - 13%), SiO2 (~ 18%),
Stotal(~ 18%) and Ctotal (~ 13%).
Due to its high heavy metal concentrations the sludge had to be considered as an
industrial waste product and to be put in safekeeping. For that purpose it was washed into
huge "ponds" that sat atop of the slag and waste rock heaps. To enable a further
dewatering of the material the basis of these ponds had not been sealed. As a
consequence the water and considerable amounts of the extremely fine grained sludge
oozed away into the slag and waste rock heaps. Finally the ponds dried up and the
sediment became exposed to the atmospheric oxygen. The weathering of these dried up
sediments are nowadays one of the main sources for the inorganic and organic
contaminants that are found in the groundwater and the surface waters of the region.
As it could be shown by other authors who work on sites comparable to the
sediment deposit described in this paper, a detailed mineralogical analysis of tailings
solids conducted in combination with adequate pore water and pore gas analysis can
provide a comprehensive insight into the future geochemical evolution of the materials
(Blowes et al. 1990). However, the dense, clay-like, and virtually dry tailings material we
focus on here does neither allow a pore gas nor a promising pore water analysis. Due to
the extremely fine grained particles (about 1 #m) a mineralogical analysis of the material
has its limitations, too. Weiss et al. (1997) who carded out a mineralogical analysis of
the unweathered sludge pointed out that all inorganic compounds of the sludge should be
referred to as "chemical compounds" rather than "minerals" because they formed rapidly
in the smelting process and not in a "natural" manner. Apart from this limitation the
predominant crystalline phases in the unaltered sludge have been identified as wurzite /
sphalerite (ZnS, 37%), galena (PbS, 6.9%), anglesite (PbSO4, 6.3%), and quartz (SiO2,
2.5%). Traces of pyrite / marcasite (FeS2) have also been detected. However, about 38%
of the inorganic material was found to be amorphous. For a more detailed discussion of
the physical and chemical properties of the sediment see Weiss et al. (1997) and
Morency et al. (1998).
The major local disposal site today is the so called "Pond 10," which sits in a basin
atop of a waste rock heap. Pond 10 contains about 225 000 m a of the described industrial
sediment. Most of the sludge has been brought to this pond between 1982 and 1990
coming directly from the production site, i.e. from the scrubbers. It was washed into the
SCHUBERT ET AL. ON EXPOSURE TO ATMOSPHERIC OXYGEN
basin using huge amounts of water (1000 - 1200 m3/d). During these years the sediment
in the pond was permanently covered with water and thus virtually sealed from the
atmospheric oxygen. After the close down of the copper works in 1990 Pond 10 dried up
and remained untouched until midsummer 1993 when a transfer of some additional
10 000 tons of flue dust and alike materials to this site began. These additional materials,
which had before been stored at other less suitable sites, cover the whole 25 000 m 2 site
with a layer of about 0.5 m in thickness.
Today a central problem for the quality of the groundwater and surface waters in
the vicinity of Pond 10 is the weathering of the dried-up sediment. Especially the
oxidation of heavy metal sulphides, which are abundant in the sediment and the
subsequent dissolving and leaching of the suiphates gives rise to considerable heavy
metal emissions out of the sediment disposal site into the surrounding environment.
A theoretical prediction of these leaching processes is difficult due to the many
different potential effects, such as the complexation and co-precipitation of heavy metals
and their adsorption onto the solid matrix particles such as organic components or iron /
manganese oxy-hydroxides. With the aim to assess the leaching behaviour of the material
some basic information on the leachability of the heavy metals in the unaltered sludge
have been obtained previously to the experiments discussed in this paper (Paschke et al.
2001). Besides a standardized leaching test which aimed to investigate the initial contact
of the material with distilled water over a 24 hours period (German Standard Methods for
the Examination of Water, Waste Water, and Sludge; DIN 38 414-$4), a five-step
shaking cascade test was performed. In each step of the cascade test 200 g solids and
1000 ml water were shaken in an overhead shaker for 24 h. After each step the leached
material was removed by centfifugation and the eluate was shaken again with a fresh
solid sample at the same liquid/solid ratio. As a third test Paschke et al. (2001) applied
the standardized pHstat shaking test developed by Obermann and Cremer (1992). For the
test the solid sample was eluted with an aqueous solution over periods of 24 h and 336 h,
while the pH was constantly maintained at 4 (at a liquid/solid ratio of 1:10).
Table 1 - Leachate pH-Value and Heavy Metal Concentration of Leaching Tests [mg/1];
Element Concentrations in the Unaltered Sludge by dry weight (ECsl,dm)[g/kg]

pH
Pb
Cu
DIN $4 (24h)
6.8
4.03
0.025
639
pHstat (24h)
4.0
30.43
0.028
2056
1.02
pHstat (336h)
4.0
23.40
0.056
2710
3.92
Cascade (step 1)
6.7
3.78
0.028
1206
0.44
Cascade (step 5)
6.5
8.89
0.127
4923
1.68
ECsludge
--
130.6
15.9
Zn
208.8
Cd
0.29
0.54
Some of the results achieved by Pasehke et al. (2001) which are relevant for the
presented paper are compared to the respective element concentrations in the unaltered
sludge in Table 1 (for the element concentrations see also Table 2). The results show that
the pHstat test (pH 4) resulted in much higher eluate concentrations than the DIN $4
21
22
(pH 6.8) for most of the investigated elements. The shaking cascade concentrations
successively approached or even exceeded the pHstat test results.
The comparison of the amounts of heavy metals eluted from the unaltered sediment
shows clearly the difference in the behaviour between the easily mobilisable elements Cd
and Zn on the one hand, and Cu and Pb on the other.
With the aim to assess the risk potential of the sediment deposited in Pond 10 the
question was raised down to which depth the dried up sediment can be chemically altered
by weathering. Since the sediment is a very dense, dry, and clay-like material it was
expected that oxidation and solution processes do only occur within a relatively thin top
layer. Provided that a significant chemical alteration of the sediment in this layer does
take place, the question which metals are mainly mobilized by oxidation and solution
processes becomes the problem of central interest.
Experimental
The goal of the experiments was to investigate the weathering behaviour of the
dried-up sediment in-situ under the given conditions. Due to the history of the site the
following facts could be presupposed.
- The industrial sediment stored in Pond 10 has a thickness of about 9 m. It consists
virtually entirely of dried-up, clay-like flue dust.
- The sludge that has been washed into Pond 10 continuously between 1982 and
summer 1990 was brought here directly from the production site and can thus be
considered as "fresh" flue dust. In this period the sediment body was permanently
covered with water.
- Following September 1990 the pond remained practically untouched for about
three years. During that time it dried up and the surface of the sediment became exposed
to atmospheric oxygen.
- Starting in summer 1993 additional materials have been brought to Pond 10 from
other sites less suitable for the storage. That material represents the top 0.5 m of the
sediment body and can not be considered as "fresh" sludge. The conditions under which
the material had been stored before being transferred to Pond 10 are not reproducible.
With the aim to investigate the influence of oxidation and weathering on the
chemical composition of the sediment and to examine its chemical alteration as a
function of depth a drill hole was bored at the centre of Pond 10. The hole penetrated the
whole 9 m of the sediment body and reached the basis of the pond. Thus the drill core
material represents the sludge that has been dumped in the basin from the very beginning
in 1982 to the last deposits washed into the basin in 2001.
The core material was stored under cool and damp conditions. Initially the drill
core was examined visually for inhomogeneities in its physical characteristics such as
grain size distribution, colour and moisture content. Subsequently the core material was
investigated for alterations in its chemical composition using XRF-analysis.
The samples for the XRF-analysis were taken from the axis of the core which had a
diameter of 15 cm. Thus they can be expected to be representative for the sediment under
the actual conditions on site. For the quantitative analysis of heavy metal concentrations
that are out of the range of the heavy metal concentrations in available reference
materials, the original material was diluted with SiO2 powder (Riedel-de-Haen). Dilution
factors of 5 - 10 reduced the concentrations of the elements of interest to the desired level
23
and yielded sample compositions, which matched the working range of calibrations
performed by the EDXRF spectrometer (XLAB 2000, Spectro A.I.). The determination
of the matrix constituents A1203 and SiO2 was provided by wavelength dispersive X-ray
fluorescence-measurements of the undiluted material (SRS-3000, Siemens AG). In any
case the prepared sample material was mixed with wax, 20% w/w, (Hoechst wax for
XRF-analysis) as a binder and compacted in a hydraulic press at a pressure of 100 MPa.
Results and Discussion
The visual examination of the drill core material confirmed the expected physical
homogeneity of the sediment body. Virtually the whole drill core could be described as a
clay-like, black, slightly moist material.
Figures 1 and 2 illustrate concentration profiles determined in the drill core
material. The concentrations of the samples taken from a depth of < 50 cm have not been
referred to since that layer of material was, as mentioned previously, added to the
original "fresh" sediment after 1993.
To enable better comparability the concentrations of the elements discussed here
were normalized to a mean composition of the fresh, i.e. unaltered sediment. That mean
composition was derived from the analytical results of the samples taken from the drill
core section between 2.5 and 9.0 m which is assumed to be not affected by weathering
and should therefore represent the chemical composition of the sludge as it came from
the scrubbers. The mean composition was defined as summarized in Table 2.
Table 2 - Mean Element Concentrations (MC) in the Sediment and the Relative Standard
Deviations (RSD)for the Discussed Elements (n = 8)
MC [g/kg]
RSD [%]
Si
AI
Fe
Zn
Pb
Mn
Cu
Cd
85.3
15.3
24.1
208.8
130.6
0.78
15.9
0.54
4.5
10.4
12.9
9.9
15.9
12.50
15.5
8.10
Figure 1 shows the normalized concentrations of silica, aluminium, and iron in the
drill core samples. As it can be seen the concentrations of these elements vary in a range
of less than + 20%.
Figure 2 shows the normalized concentrations of the heavy metals of concern. It
can be seen that the heavy metal concentrations do also vary within a + 20% range of the
mean concentrations in the zone below 2.5 m. That implies the chemically homogeneous
composition of the fresh sludge. However, the samples taken in 50, 75 and 100 cm show
that in these upper layers the concentrations of cadmium, manganese and zinc are
significantly lower than the respective mean concentrations. In contrast to this the
concentrations of lead and copper remain within the + 20% range.
24
1,6
* Si
~"
1,4
1,2
AI
-o- Fe
1,0
l/'o"
"',
."
0,8
0,6
g~
0,4
I00
200
300
400
500
600
700
800
depth Icml
Figure 1 - Normalized Concentrations of Si, Al, and Fe vs. Depth in the Sediment
1,6
.~1,4
- o~ Pb
- - e - Zn
- x - Cu
- ~x- M n
-.o-- Cd
1,2
& 1,0
O
~o,g
0,4
100
200
300
400
500
600
700
800
depth [em]
Figure 2 - Normalized Concentrations of Heavy Metals vs. Depth in the Sediment
The obtained chemical data indicate that the fresh sludge was chemically fairly
homogeneous, The concentrations of silicon, aluminium, and iron change in a range of
only about + 20%. Silicon, which appears as SiO2 in the whole sludge body is not
considerably affected by solution processes. Aluminium appears as A1203 which also
shows a very poor solubility and is thus not leached out of the material. Iron has been
identified as FeS in the unaltered sediment. Since the sulphide is not stable in the upper,
oxidized layers it is subject to alteration. However, FeO(OH) and FeSO4 show only poor
solubilities too and are for that reason not leached out of the sediment. Consequently the
iron concentration in the altered zone does not change significantly and stays in the
mentioned 20% range.
In the lower part of the depth profile the heavy metal concentrations do also stay in
a range of just about + 20%. However, in the upper 50 cm of the discussed section the
concentrations of zinc, manganese and cadmium are considerably lower than in the mean
unaltered sediment. The cadmium concentration determined in the most upper sample is
only about 40% of the cadmium concentration in the unaltered sludge. Zinc and
manganese show concentrations of about 50 and 60%, respectively, compared to the
unaltered sludge. On the other hand it can be seen that the concentrations of lead and
copper do not show such a general decrease in the upper section. The concentrations
remain in the + 20% range.
The mobilization of zinc, manganese and cadmium and the more or less stable
behaviour of lead and copper can be explained with the solubility of the respective
sulphates. In the fresh sludge the discussed metals appear mainly as metal-II-sulphides.
The respective solubility data are summarized in Table 3. As it can be seen the sulphides
of the discussed metals are virtually insoluble.
As long as the sediment was covered with water the reducing conditions in the
actual sludge body were stable. After the pond had dried up oxidation processes started at
the surface of the sediment and the sulphides were transformed into sulphates. That
oxidation front penetrated into the sludge deposit. However, due to the clay-like
structure of the material, the hydraulic conductivity of the tailings material is < 10-9 naJs,
most of the rain water runs off the sediment surface towards the surrounding coarse
grained waste rock dams where it oozes away immediately.
Since the sediment was exposed to rainfall the sulphates that show high solubilities
were mobilized much more easily than the sulphates that show only poor solubilities. As
can be seen in Table 3 the solubilities of the discussed sulphates decrease in the order
ZnSO4 > CdSO4 > MnSO4 > CuSO4 > FeSO4 > PbSO4. That is in good correspondence
with the data illustrated in the Figures 1 and 2. The concentrations of Zn, Cd, and Mn,
i.e., of the metals that have easily soluble sulphates, are considerably lower in the upper
sludge layer due to oxidation and mobilization. On the other hand the concentrations of
the metals that have sulphates with only low or very low solubilities (Fe, Cu, and Pb) do
exhibit more or less stable concentrations even in the top layer of the sediment.
Table 3 - Solubilities o f the Sulphides and Sulphates [g/l] o f the Discussed Metals in
"Cold Water" (Lide 1993); s = Highly Soluble, sl s = Slightly Soluble, * = FeS04 x 1-120
Zn
Cd
Mn
Cu
Fe
Pb
M-II-Sulphate
755
520
143
sl s *
0.0425
M-II-Sulphide
0.00067
0.0013
0.0047
0.00033
0.0062
0.00086
Besides FeSO4 iron is, to a certain degree, likely to be bound in oxy-hydroxides,

which are also poorly soluble. An adsorption of heavy metals onto these oxy-hydroxides
is a possible way of heavy metal fixation, too, but due to the low iron concentrations in
the sediment, compared to the extreme heavy metal concentrations even in the altered
zone, that mechanism does not play an important role in the case discussed here.
Besides the different solubilities of the sulphates and oxides of the discussed metals
their different affinity to organic matter might be another effect responsible for the
25
26
observed weathering behaviour. Copper shows a much higher tendency to form metalorganic complexes than zinc, manganese, cadmium and lead (Homburg et al. 1993,
Asche and Beese 1986, Gruhn et al. 1985). That means that in the presence of organic
matter copper becomes adsorbed or chemically bound and thus fixed in complex
insoluble metal-organic structures, whereas zinc, manganese and cadmium are much
more mobile. Since the sludge consists of up to 20% of organic compounds, such as
PAHs, the formation of metal-organic complexes is likely to be of some importance for
the fixation of copper in the oxidized zone of the sediment.
Conclusions
The weathering of the dried-up sediment, i.e. the oxidation of heavy metal
sulphides and the subsequent dissolution of the sulphate salts, gives rise to considerable
heavy metal emissions in the local ground- and surface waters.
The elements Si, Ai, Fe, Pb, and Cu show concentration changes versus depth in a
range of only about +_20% of the respective mean concentrations of the "fresh" sludge.
On the other hand the concentrations of Cd, Mn, and Zn display considerably lower
concentrations in the upper 50 cm of the original sediment body.
After the sediment had dried up an oxidation process started at the surface and
transformed the suiphides into sulphates and oxides. That gave rise to a mobilization of
Zn, Mn and Cd whereas Fe, Pb, and Cu show more or less stable behaviour due to the
poor solubilities of their sulphates or oxides/hydroxides. Another reason for the different
weathering behaviour of the metals might be their dissimilar affinity to organic matter.
References
Asche, N. and Beese, F., 1986, "Untersuchungen zur Schwermetalladsorption in einem

sauren Waldboden," Zeitschrift fuer Pflanzenernaehrung und Bodenkunde, Vo1.149,
pp.172-180.
Blowes, D. W. and Jambor, J. L., 1990, "The Pore Water Geochemistry and the
Mineralogy of the Vadose Zone of Sulphide Tailings, Waite Amulet, Quebec, Canada,"
Applied Geochemistry, Vol. 5, pp. 327-346.
Gruhn, A.. Matthess, G., Pekdeger, A., and Scholtis, A., 1985, "Die Rolle der geloesten
organischen Substanz beim Transport von Schwermetallen in der ungesaettigten
Bodenzone," Zeitschrift der deutschen geologischen Gesellschaft, Vol. 136, pp. 417-427.
Homburg, V. and Bruemmer, G., 1993, "Verhalten von Schwermetallen in Boeden Untersuchungen zur Schwermetallmobilitaet," Zeitschrift fuer Pflanzenemiihrung und
Bodenkunde, Vol. 156, pp. 467-477.
Lide, D.R., 1993, CRC Handbook of Chemistry and Physics, 73 ed., CRC Press, Boca
Raton, USA.
Morency, M., Weiss, H., Freyer, K., Bourne, J., Fontaine, D., Mineau, R., Moeder, M.,
Morgenstern, P., Popp, P., Preda, M., Treutler, H.-C., and Wennrich, R., 1998,
"Oxidation treatment of a sulphide-bearing scrubber dust from the Mansfeld Region Germany," Science of Total Environment, Vol. 223, No. 2-3, pp. 87-97.
Obermann, P., Cremer, S., 1992, ,,Mobilisierung von Schwermetallen in Porenwassern
von belasteten B6den und Deponien," in Materialien zur Ermittlung und Sanierung yon
Altlasten, Vol. 6, Landesamt fur Wasser und Abfall Nordrhein-Westfalen, Germany.
Paschke, A., Freyer, K., Treutler, H. C., Wennrich, R., Popp, P., Moeder, M., Weiss, H.,
and Schueuermann, G., 2001, "Assessment of Theisenschlamm Leaching Behaviour", in
Fine-rained residues from copper smelting and their environmental impacts, A case
study from the Mansfeld District, Germany, UFZ-Report Nr. 22/2001, UFZ Verlag,
Leipzig, Germany, pp. 71-82.
Wege. F.-W., 2000, "Das Haldenkonzept der Landkreise Mansfelder Land und
Sangerhausen," in: Bergbau- und Rueckstandshalden des Mansfelder
Kupferschieferbergbaus, Schriftenreihe des Mansfeld-Museums, Neue Reihe, 5 Eigenverlag des Foerdervereins Mansfeld-Museum und des Mansfeld-Museums,
Hettstedt, Germany, pp. 2-13.
Weiss, H., Morency, M., Freyer, K., Bourne, J., Fontaine, D., Ghaleb, B., Mineau, R.,
Moeder, M., Morgenstern, P., Popp, P., Preda, M., Treutler, H.-C., and Wennrich R.,
1997, "Physical and chemical characterization of a complexly contaminated scrubber
dust - a by-product of copper smelting in Sachsen-Anhalt, Germany," Science of Total
Environment, Vol. 203, No. 1, pp. 65-78.
27
Mathias Ricking, i and Tobias Schulze i
Deep-Freeze Sampling Methods for Soft Sediments

Reference: Ricking, M., and Schulze, T., "Deep-Freeze Sampling Methods for Soft
Sediments", Contaminated Sediments: Characterization, Evaluation,
Mitigation/Restoration, and Management Strategy Performance, ASTM STP 1442, J.

Locat, R. Galvez-Cloutier, R. C. Chaney, and K. R. Demars, Eds., ASTM International,
West Conshohocken, PA, 2003.
Abstract: Freeze coring systems were applied for sampling high-water-content

sediments or very coarse sediments for a couple of years. At the Free University within
the last 20 years different systems were developed and applied for sampling the upper
1.5-2.5 m in an undisturbed stratified way. The difficulty in controlling the depth
penetration has been overcome by applying a small echo sounding system onboard. The
LINDE-LANCE device runs on liquid nitrogen pressurized through a copper pipe, where
the sediment freezes in concentric layers onto the lance at -193~ A low weight system
with a closed aluminium lance (CO2-LANCE) running on a eutectic mixture of dry ice
and alcohol that freezes the concentric sediment layers onto the lance at -90~ was
developed for a much easier sampling by means of a Zodiac. Advantages and limits are
addressed.
Keywords: freeze-core techniques, sediment, high-resolution analysis, environmental

chemistry
Free University of Berlin, Department of Earth Sciences, Environmental Geochemistry

and Hydrogeology, Malteserstr. 74-100, House B, 12249 Berlin (Germany); e-mail:
ricking@zedat, fu-berlin.de.
28
Copyright9 2003by ASTM lntcrnational
www.astm.org
RICKING AND SCHULZE ON DEEP-FREEZESAMPLING METHODS
29
Introduction
Freeze coring systems were applied for sampling high-water-content sediments

(Renberg and Hansson 1993, Lotter et al. 1997 Anderson et al. 2000, Quinlan et al. 2002)
or very coarse sediments (Hill 1999) for a couple of year
The main scientific interest in freeze coring systems is the high resolution analysis of
the sediment record in paleoecological studies (GFZ-Potsdam-UWITEC corer; Quinlan et
al. 2002), and pigment and other chemical analysis in high-water-content surface
sediments ifa fine resolution is required (Lotter et al. 1997; Anderson et al. 2000). In
very coarse systems only the few top layers are sampled for biofilm analysis (Hill 1999).
Within the urban area of Berlin the eutrophic to hypertrophic riverine system of the
Havel and Spree Rivers is enlarged by several canals, connecting the Havel and Spree
River upstream of their confluence, and numerous lake-like broadenings, with very slow
flowing conditions. In these regions the sedimentation rates are very high, up to 3
cm/year.
The muds are characterized by high water contents of more than 90% in the upper
decimetres, accompanied by total organic carbon (TOC) contents up to 15%. In highly
organic sediments methane ebullition often occurs when the core is ascended to the water
phase. Such gas ebullition may destroy or homogenize the laminations.
Therefore, freeze coring systems were developed and optimized for a high-resolution
analysis of heavy metals and organic compounds within the top sediment layers (e.g.,
Koch et al. 1999, Schwarzbauer et al. 2001).
At the Free University (Department of Earth Sciences) within the last 20 years 3
different systems were applied and developed for sampling the upper 1.5-2.5 m in an
undisturbed stratified way. These layers represent an accumulation time from 1960 to
now, as determined by gamma spectrometric measurements of Pb-210 and Cs-137 of
sediment core slices. The difficulty in controlling the depth penetration has been
overcome by applying a small echo sounding system onboard.
Methods
The first system developed was a hollow core system, later on published by Pachur et
al. (1984), which was used for a few cores only (Ballschmiter et al. 1980). The core
length is restricted to 1 m. The main disadvantage was that it was too laborious and could
be applied only by means of a drilling platform. Based on the experiences with the
hollow core sediment coring system, during the mid to late 1980s a system used for the
subway construction was adapted to obtain aquatic sediments in the (West-) Berlin area.
The LINDE-LANCE system runs on liquid nitrogen pressurized through a copper
pipe, where the sediment freezes in concentric layers onto the lance at -193~ The
system is 2.50 m in length and has a diameter of 10 cm.
During the early 1990s a low weight system with a closed aluminum lance (CO2LANCE) running on an eutectic mixture of dry ice and alcohol, which freezes the
concentric sediment layers onto the lance at -90~ was developed for a much easier
30
sampling by means of a Zodiac. The system is 1.78 m in length and has a diameter of 7
cm.
The LINDE-LANCE system is used from a drilling platform where the system is
lowered very slowly down to the sediment surface by means of a wire to the pre-selected
depth via an echo sounding system (Figure 1). To protect the lance from penetrating
water an Armaflex| tape was used, characterized by a watertight protection at very low
temperatures. Through a copper lance the liquid nitrogen is pressed out o f the DEWAR
container by applying a nitrogen gas pressure of 0.5-0.8 MPa.
Nitrogen Gas Dewar with Liquid Nitrogen
ipe
Stainless Steel Lance
Sediment
Frozen Sedimen
Figure 1: The liquid nitrogen system (LINDE-LANCE)
Liquid nitrogen volatilises in the lance and the sediment freezes onto the lance due to
the cold advection nitrogen gas. About 100 L liquid nitrogen are required for each core.
The copper pipe was adjusted at different depths to sample especially the sediment/waterinterface. The maximum core length is in the range of 2.5 m with a diameter of 15-30
cm.
It takes at least 2-2.5 hours to obtain one core. The application is limited to a water
depth o f 10 m. After the lance has warmed to ambient temperature, it is filled with hot
water and the frozen sediment crust is detached. The core is cut onboard into pieces of 50
RICKING AND SCHULZE ON DEEP-FREEZESAMPLING METHODS

cm by means of a stainless steel saw or can be obtained as one piece. To remove
contamination the non-frozen sediment is washed away by pure water, surfaces are
cleaned with a metal scraper and the pieces are wrapped in pre-cleaned aluminium foil.
The sediment core is transported on dry ice to the laboratory, where it is stored at
-20~ in the dark. The core is subsampled into slices using a glowing silver wire, heated
by using a current of 12 V, or using a stainless steel saw under a laminar flow bank. The
freeze-dried samples can be used for geochemical, isotopical, and biostratigraphical
analyses.
The CO2-LANCE (Figure 2) is applied from a ship/zodiac by use of an anchor rope or
a thin coring linkage. Ten kg crushed dry ice and 0.3 L alcohol are filled into the lance
where an eutectic mixture develops
CO2-Bubbles
luminiumaRCS
Sediment
Frozen Sediment
Dry lee &
Figure 2: The dry-ice system (CO:-LANCE)
The closed lance is lowered down to the pre-selected depth immediately. The sediment
freezes undisturbed in layers of up to 20-25 cm thickness on the outside of the lance. The
sampling takes at least 20-30 minutes and is finished when no gas bubbles are appearing
on the water surface. Then the sediment core is lifted to the surface, the lance is filled
with hot water, the core is detached, and the contamination of the sediment core is
washed away by pure water. The sediment pieces or the whole core is wrapped in
aluminium foil and transported on dry ice to the laboratory, where the storage was
31
32
realized at -20~ in the dark. The preparation for analysis was realized under the same
conditions as for the LINDE-LANCE.
Discussion
Besides a small smearing of the inner few mm, which are withdrawn, preliminary
results indicate no modification of the samples and compounds, as deduced by
microscopic analysis of laminated sediments.
Using now a stainless steel lance eliminates problems with heavy metals like Zn by
applying the zinced LINDE-LANCE. Earlier experiments revealed no contamination via
Zn (78% of the outer LINDE-LANCE mantle). Tests by members of the working group
in 1991, where 6 lance parallel sections were analysed in 1 cm intervals revealed no
contamination by the steel lance at a concentration level of>_ 500 lag/g Zn. Pre-industrial
sediments in Lake Bogen (Brandenburg, Germany) or Lake Nikolas (Berlin, Germany)
displayed no contamination via the lance at a concentration level of 15-70 ~tg/g Zn.
There are no pore-water processes registered during the freezing-process, as the
sediment freezes immediately, revealed by the water content analysis with a measured
uncertainty area of 1.1% for the water content analysis.
The main advantages for these methods are the definitive control on the depth of
penetration, also controllable by video systems, no hydraulic systems, the easier
handling, and the gaining of more sediment material and the relatively low costs per core.
Limiting factors for the LINDE-LANCE are the necessity to protect the lance from
penetration of water by sealing de connections between the stainless steel parts with a
special tape for refrigerator techniques.
Limiting factors for the CO2-LANCE are water depths of maximum 40 m, in
maximum 4 cores per day and less sediment material than with the LINDE-LANCE
method.
Both methods will be used for sampling surface sediments and sediment cores for the
German Environmental Specimen Bank (ESB). Warved sediments of Lake Belau
(Northern Germany) with high water contents illustrate the feasibility of the freeze coring
system (Figure 3).
First results relating to possible modifications during the sampling and storage will be
presented at the symposia, initial test revealed no modification.
Conclusions
Both methods are suitable for sampling high-water-content sediments without

disturbing the stratigraphy. The LINDE-LANCE system gains more material, but is
limited to water depths of less than 10 m. The CO2-LANCE provides less material but is
applicable to water depths of up to 40 m. The frozen sediment cores have to be prepared
for transport within 30 min after archiving them.
RICKING AND SCHULZE ON DEEP-FREEZE SAMPLING METHODS
Figure 3: C02-LANCE sediment core of Lake Belau (Northern Germany)
REFERENCES
Anderson, N. J., Clarke, A., Juhler, R. K., McGowan, S., and Renberg, I,, 2000, "Coring
of laminated lake sediments for pigment and mineral magnetic analyses,
Sondre Stromfjord, southern West Greenland", Geology of Greenland Survey
Bulletin Vol. 186, pp. 83-87.
Hill, M. T. R. 1999, "A freeze-corer for simultaneous sampling of benthic
macroinvertebrates and bed sediment from shallow streams", Hydrobiologia
Vol. 412, No. 10, pp. 213-215.
Koch, M., Ricking, M., Rotard, W., Mailahn, W., Knoth, W., Pribyll, J., and Pachur, H.J., 1999, "PCDD/Fs, PCBs and PAHs in sediments of North-East Germany
and their source identification", Posterpresentation on Dioxin "99 in Venice,
Organohalogen Compounds, Vol. 43, pp. 359-363.
Pachur, H.-J., Denner, H. D., and Walther, M., 1984, "A freezing device for sampling the
sediment-water interface in lakes", Catena, Vol. 11, pp. 65-70.
33
34
Quinlan, R., Leavitt, P. R., Dixit, A. S., Hall, R. I., and Smol, J. P., 2002, "Landscape
effects of climate, agriculture, and urbanization on benthic invertebrate
communities of Canadian prairie lakes", Limnology and Oceanography, Vol.
47, No. 2, pp. 378-391.
Schwarzbauer, J., Ricking, M., Franke, S., and Francke, W., 2001, "Organic pollutants in
sediments of Havel and Spree River. Part 5: Halogenated compounds",
Environmental Science and Technology, Vol. 35, No. 2, pp. 4015-4025.
Rosa Galvez-Cloutier t, Marie-Eve Brin 2, Gerardo Dominguez 3, Serge Leroueil4 and

Sylvain Arsenault ~.
Quality Evaluation of Eutrophic Sediments at St. Augustin Lake, Quebec, Canada
Reference: Galvez-Cioutier, R., Brin, M.E., Dominguez, G., Leroueil, S. and Arsenault,
S., "Quality Evaluation of Eutrophic Sediments at St. Augustin Lake, Quebec,
Canada,"
Contaminated
Sediments:
Characterization, Evaluation,
Mitigation~Restoration, and Management Strategy Performance, ASTM STP 1442, J.
Abstract: St. Augustin Lake is located 20 kilometers west from Quebec City in the
territory of Quebec's Provincial Capital. The whole basin land is largely urbanized with
some agricultural and wetland zones, some forest spots and surrounded by an important
network of highways. The lake neighbors an industrial park that includes airport
installations and storage facilities for chemical products. In 1993, a first quality study
revealed the presence of some toxic metals in lake sediments. Although, lower levels
were found, these analyses were performed at a time where neither standards methods
nor quality guidelines were available in Canada. Recently launched, a major City
project aims to open the lake to public recreational activities and involves the cleaning
of water. St. Augustin Lake is in a eutrophic state most of the year, and all usages have
been limited. However, a recent paleo-limnologic study showed that the lake used to be
in the past (60 years ago) rather mesotrophic. The water clean-up program cannot
succeed without the evaluation and restoration (if needed) of the bottom sediments.
Indeed, bottom sediments may constitute important reservoirs for contaminants such as
phosphorus, nitrogen, heavy metals and hydrocarbons. These contaminants may be
strongly attached to organic colloids, clay and carbonates that are usually present in
sediments. This study shows metal profiles and their distribution and correlations
between metals and some sediment constituents.
Keywords: sediment, heavy metals, eutrophication, nutrients.
i Associate Professor, Department of Civil Engineering, Faculty of Science and Engineering,
Laval University, Quebec, Quebec GIK 7P4.
2,3 Graduate students, Department of Civil Engineering, Faculty of Sciences and Engineering,
Laval University, Quebec, Quebec G 1K 7P4.
4 Professor, Department of Civil Engineering, Faculty of Science and Engineering,Laval
University, Quebec, Quebec G1K 7P4.
5Biologist, EXXEP consultants, 3470 Chemin de Quatre Bourgeois, Ste. Foy, Quebec, Canada,
GIW 2K4.
35
www.astm.org
36
1.0 Introduction
This study is included in the project "La grande Corvd"' launched by the ancient city
council of St. Augustin City whose final objective is the opening of the lake to its
citizens for recreational activities that aim to include bathing and fishing. However, the
Saint-Augustin Lake is in a eutrophic state almost all year round, and all usages have
been limited. The water clean-up program cannot succeed without the evaluation and
restoration (if needed) of bottom sediments. Indeed, bottom sediments may represent
important reservoirs for nutrients and contaminants such as phosphorus, nitrogen, heavy
metals and hydrocarbons. These contaminants may be attached to organic colloids, clay,
carbonates or be co-precipited with iron and manganese in hydroxides present in the
sediments constituents. The objective of the present study was to firstly characterize
physically and chemically the sediments and secondly to evaluate environmental quality
of the sediments.
2.0 Study Area and Sampling Schedule
2.1 Description
St. Augustin Lake is located 20 kilometers west from Quebec City on the territory of
Quebec's National Capital between latitudes N 46 ~ 42' and N 48041 , and longitudes W
71~ and W 71~ '. The Lake is 2225 m long and 380 m width. Its whole basin has
been largely urbanized but still presents agricultural zones, wetlands, and forest spots all
surrounded by an important network of highways. It also includes hydroplanes port
installations. The lake neighbors an industrial park and landfill. In Figure 1, a general
diagram shows a summary of the main limnological and geochemical characteristics of
the St. Augustin Lake. As it can be seen, St. Augustin Lake is a shallow lake with the
maximum depth of 6 m. The sediment occurs in a 30 cm - lm thick layer above the
bedrock. The average annual precipitation around the area is 800mm, apart form the
precipitation, the lake is fed with groundwater and it counts with one tributary and one
discharge. In 1993, a first quality study (Landry 1993) revealed the presence of toxic
metals in the sediments. Although, low levels were found: cadmium (0.5 - 1 ppm),
chrome (15 - 28 ppm), copper (10 - 25 ppm), nickel (20 - 30 ppm), lead (25 - 30 ppm)
and arsenic (1 - 4 ppm), these analyses were performed at a time where no standard
methods, no quality guidelines were available in Canada. Despite the fact that the Lake
can reach extremes eutrophic conditions, a recent paleo-limnologic study (Roberge et al.
2002) has shown that the lake used to be mesotrophic about 60 years ago. The method
alpha was employed using the Pb 210 datation. The constant rate of supply (CRS) has
been used in the Roberge et al.,'s study to change the data for a period of time.
GALVEZ-CLOUTIER ET AL. ON EUTROPICSEDIMENTS
37
Figure 1 -Schematic Conception of the Actual situation at St. Augustin Lake
2.2 Sampling Schedule
During fall 2001, a systematic sampling program used a stochastic approach (Baudo
1990) in order to obtain 42 sediment samples shown in Figure 2that were representative
of the sediments at the bottom of the Lake. The sampling used 8 transepts that defined
triangular areas from which samples were obtained, covering the whole lake surface.
The coordinates at each sampling station were recorder by using a Global positioning
system (GPS) equipment. In addition, 5 extra samples were taken at hot spot areas such
as the discharging points of: surface runoff from the highway system, agricultural
runoff, hydroplane port, Keno Camping Station. As shown in Figure 3, the sediment
samples were taken with an Ekrnan grab containing a volume of 5.3 L and with the
following dimensions 150 X 150 X 225mm. The grab took this volume of the surface
sediments to different depths inside the water column varying between 1 to 6 m. At
each site, pH, redox, temperature and conductivity were measured in-situ using a WTW
TETRA COM 325 multimeter. The sediments were stored in a cold room (4~ C) in
polypropylene containers.
38
Figure 2- The Samples Site and the Transects at the St. Augustin Lake
39
Figure 3 - Typical Sediments (left) and Ekrnan Grab (right)
2.3 Sediments characteristics
Sediments presented high organic matter and in some instances strong sulphur and
decomposition odours. The characterization program covered two aspects: 1)
mineralogy, and grain size evaluation and, 2) Contamination: concentration of priority
contaminants such as heavy metals (Cd, Cr, Ni, Pb, Hg etc.), metalloids (As), anions
(CO3, NO3, PO4 and SO4), organochlorinated pesticides known to be used in the area
and polycycling aromatics hydrocarbons (PAHs).
3.0 Experimental methods

3.1 Grain size analysis and mineralogy
After destruction of organic matter with oxygen peroxide, grain size analysis was run
separately on fine and coarse fractions. Fine grain size analysis was done using a laser
counter (Mastersizer of Malvern Instuments). This equipment measures sizes from 0,05
pan to 556 pan. Results are summarized by a representative graphic (Figure 4). The fine
fraction consisted mainly of silty loam and sand. Coarse grain size (sieve # 30 > 600
Jxm) was done gravimetrically using sieve analysis according to BNQ 2560-040 (BNQ
1982). Two methods of classification are shown for the coarse fraction. The first one
tames from the Unified Soil Classification System (ASTM D2487) and the second one
from the United State Department of Agriculture. Only seven samples required this
method. Mineralogy was performed using X-ray diffraction and scanning electron
microscopy. X-ray diffraction was run on a D 5000 Siemens insmarnent. Operating
conditions were the following: 40 kv and 30 ma, Cu lamp and angle span from 10~ to
70 ~ and shooting speed of l~
and aperture of 0.02 ~ / 1.2 seconds. Samples were
40
1 cm x 2 cm. Scanning Electron Microscopy was performed in a JEOL 840A instrument

using a retrodiffusion method of radiation that allows for the study of both morphology
and elemental analysis. For X-ray diffraction and SEM analysis, samples F2 and H2
were chosen due to their high content on Pb and Zn.
3.2 Anions and metals analysis

The interstitial water was taken as the supernatant of a sediment sample after 10
minutes of centrifugation. The filtration of the supernatant and the dilution were required
for measuring the ions concentration with an ion chromatograph (Dionex). Heavy
metals were measured by AAS following the procedure described by the Saint.
Lawrence Centre (SLC 1993). Sediments were digested using 5ml HNO3 and 2 ml of
peroxide (30 %) for 1 gram (dry weight). Then, the mix was heating until the volume
reached 1 at 2 ml. After, 10 ml of regal water (HNO3 20% and HC1 5%) was adding in
the solution, during 60 minutes the heating continued. Mercury and arsenic were
measure out by and external certified laboratory (Bodycote). Quality control included
the usage of reference materials like the PACS-2 from the National Research Council of
Canada
3.3 Organic compounds analyses
Pesticides were measured on HPLC and PAHs by GC-MS according to EPA standard
methods by an external certified laboratory (Bodycote).
4.0 Results
4.1 Geochemistry of water and sediments

The summary of in-situ results is presented in Table 1. The pH and conductivity
values of water present in the grab were homogeneous and stable for most of the
stations. On the contrary, redox values presented a large variability from 174 mV to -114
mV. Among the 42 samples, 13 stations (about the third) presented negative value of
redox potential. Otherwise, there is no correlation between the pE and the depth of the
sample (R 2= 0.02). It seems that the depth of water does not have an influence on the
potential redox. In eutrophic lakes dissolved oxygen can rapidly be consumed due to
oxidation process. At this point anaerobic conditions may reach various stratification
layers (Wetzel 1983).
GALVEZ-CLOUTIER El" AL. ON EUTROPICSEDIMENTS
41
Table 1 - In-sire Characteristics
Parameters
Average and Standard Deviation
pH
Conductivity(ItS)
Temperature (~
pE (mV)
Alkalinity (mg/L CaCO3)
Water content (%)
7.70
796.63 +_ 12.44
8.71 + 0.60
36.98 + 68.96
118.14 + 4.48
347.30 + 115.44
Organic matter (measured according to Centre d'Expertise en Analyse

Environnementale du Quebec, 2001, M.A. IO0-S.T. 10,) ranged from 6 to 19%, with the
average value of 11%. In fact, two kinds of organic layers were visually detectable as
shown in the scheme of Figure 1.
Volume(%)
Aect~m~ative
'/i
P~r~cleD ~ t e r O~m)
Figure 4 - Grain Size Distribution of the Representative Sample F2.
42
From the result of grain size analysis a general pattern arises. This typical pattern is
represented in a Figure 4. This one illustrates the general distribution o f the fine grain
size. The coarse grain results are shown in Table 2
Table 2- Grain Size Analysis of Coarse Samples
Sample
HI
CD
Composition
(%)
Sand : 50
(WL) J
(%)
USCS
Classification
USDA
Clay : 5
Silt : 40
Fragments : 5
Sand : 30
Clay : <1
Silt : 60
38.8
Silty sand (SM)
Sandy loam
58.8
Organic silt (OH)
Silty loam
Clay : 5
Silt : 53
Fragments : 2
Sand : 45
Clay : 2
Silt : 53
Fragments : <1
Sand : 45
58.8
Organic silt (OH)
Silty loam
58.8
Organic silt (OH)
Silty loam
Clay
:0
Silt : 25
38.8
Silty sand (SM)
Loamy sand
Clay : 5
Silt : 53
Fragments : < 1
Sand : 92
58.8
Organic silt (OH)
Silty loam
Clay : 0
Silt : 8
Fragments : < 1
N/A
Sand (SP)
Sand
Fragments : 9
Sand : 40
F4
D1
HI extra
Fragments : 30
Sand : 42
EF
II
I Water limit
Theses results confirm the similarity between both classifications. The sediments
were characterized by a grain size sample S curve and can be classified as silty loam.
The diffraction results show the strong presence of quartz, feldspars (albite) and mica
(biotite) as primary minerals. As secondary minerals, calcite, and dolomite were found
in trace concentrations, few clays (mostly illite and chlorite) were also found in low
GALVEZ-CLOUTIER El" AL. ON EUTROPICSEDIMENTS
43
concentrations. Figure 5 shows various grains of sediments. In particular, sphalerite

indicated as zinc sulphate was identified. In the same figure, the second structure at the
left is a grain of clay characterized by its multiple lattice structure. Organic matter was
identified abundantly in all samples. In Figure 6 from sample F2, white spots can be
seen, which were identified as pyrite (FeS2).
Figure 5 - SEM Photography of sample H2.
Figure 6 -SEMPhotography of sample F2
4.2 Contamination
St. Augustin Lake is also reservoir of highway runoffs. These highways are
maintained in winter by de-icing products. For this reasons chloride and sulphate were
measured in interstitial water sediments. As an example, the longitudinal profile of C1
concentrations is shown in Figure 7. Chloride concentrations were generally high
44
CONTAMINATED
SEDIMENTS
compared to other lakes in Quebec. In particular, site D2 and H3 reached 200 ppm. Site
H3 is closer to the discharging point of highway runoff. Sulphate was also high,
concentrations ranged from 100 to 350 ppm with a strong peak at the Keno Camping site
of almost 500 ppm. Unfortunately, in Quebec there are no quality guidelines for C1 or
SO4. It seems that anions phosphorus and nitrogen were not present in soluble state
(ortho-phosphates and nitrates) in the interstitial water. Further investigations must be
done to quantify the concentration of these anions responsible of eutrophisation
Chloride concentrations
250
ilOO
otEEttttI tt Ht
O
. . . . . .
......
= . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
i ........
i .....
i,,,,
i i i r ] . . . . . . . . . . . . . . . . . . . . . . . .
Sampling Sites
Figure 7- Chloride Concentrations on Interstitial Water Samples of the Sediments of St.

Augustin lake
The evaluation of contaminants included the concentrations of heavy metals (Cu, Cd,
Cr, Ni, Zn, Pb and Hg), transition metal (As), organochlorinated pesticides and PAHs.
Heavy metals were dosed in the 42 samples while Hg, As, pesticides and PAHs were
measured on 5 samples representatives of each transect (A3, D3, F1, G3, H2).
Presently, quality criteria in Canada are mostly based on the comparison of total
concentrations against guidelines that defines levels of toxicity (Galvez-Cloutier et al.
1995). In Canada three guidelines are frequently used: that one of the St. Lawrence
Centre (1996), that one of the Ontario's Ministry of Environment and that one of the
Canadian Council of Ministries of Environment (OME 1990, EC 1992, CCEM 1999,
SCL 1996). The table in Figure 1 shows the average values of contaminants found in the
sediments, all metals were found in various concentrations being Zn the most abundant.
Concentrations were higher than those reported in the Landry (1993) study. Table 3
GALVEZ-CLOUTIER ET AL. ON EUTROPIC SEDIMENTS
45
summarizes the results of the heavy metal analysis. The second column shows the
average concentration in 2001. These results could be compared with the results from
1993 (third columns) and the background values (fourth column). At the fifth column
the SEM (minor effect level) values of the EC (1992) guidelines are shown. The order of
decreasing concentration was Zn > Pb > Cu > Cr > Ni > As > Hg. Cadmium, zinc and
lead profiles are shown, respectively, in figures 8-10. The majority of the results
obtained in 2002 exceed the background value and the SEM criteria. Moreover, the
present results are higher than those obtained in 1993.
Table 3 -Summary of Metals Concentrations

Metals and
Metalloids
St. Augustin
Landry
Background
SEM
Sediments
Data (1993)
values
Criteria
Concentration
From (CCEM) I
SLAP2
(mg/kg)
(mg/kg)
(mg/kg)
(mg/kg)
Arsenie (As)
6.68 1.13
3.20
2.5
7
Cadmium (Cd)
2.38 + 0.65
0.63 + 0.10
0.32
0.9
Chrome (Cr)
25.87 + 7.24
22.72 4.36
47
55
Copper (Cu)
40.87 + 9.47
18.38 + 4.00
31
28
Lead (Pb)
77.82 + 26.47
24.27 + 4.50
6
42
Mercury (Hg)
0.12 4- 0.02
0.05 +0.01
0.074
0.2
Nickel (Ni)
42.18 + 9.54
28.77 + 5.27
n/a
35
Zinc (Zn)
280.83 74.91 121.33+ 27.22
104
150
Canadian Council of Environment Ministers; ~St-Lawrence Action Plan
Figure 8 - Concentration Profile of Cadmium at St. Augustin Lake
46
Figure 9 - Concentration Profile of Zinc at St. Augustin Lake

As it can be seen from Table 3 and Figures 8, 9 and 10, most heavy metals exceeded
the SEM level of toxicity. In the ease of Cd and Zn, over 90% of samples exceeded the
SEM level. Zinc is an essential element vital for growth while cadmium may be highly
toxic.
Figure 10- Concentration Profile of Lead at St. Augustin Lake

Compared to geogenie background values (approximately 100 ppm for Zn, 20 ppm
for Pb, 30 ppm for Cr and 0.2 ppm for Cd) and compared to SEM criteria it is evident
that sediments at the St. Augustin lake have been contaminated most probably by the
anthropogenic activities that had taken place in the past. Comparing values from the
47
Landry study (1993) concentrations have increased for all metals measured. Spatially,
contamination was not homogenous. For instance, figure 11 shows the spatial
distribution of Cd at the sediment's surface. Samples fi'om the darker area exceeded 3
ppm of Cd. As it can be seen, the central-north and extremes zones o f the lake were the
most contaminated.
Figure 11 - Spatial Distribution of Cd.
48
Figurel 2 -Concentration of the Mercury of 5 Samples of the St. Augustin lake.
Results for Mercury and Arsenic are presented in Figures 12 and 13. While Hg values
did not exceed SEM criteria, arsenic was found in greater concentrations. Higher
concentrations of Arsenic were obtained also at the longitudinal extremes of the lake.
The CCEM (1999) has indicated a background value for As of 2.5 ppm. Due to the high
toxicity of arsenic it has to be considered for further evaluation.
Figure 13 - Arsenic Concentrations offive Samples of the St-Augustin lake
49
Organochlorinated pesticides and PAH were not found in any of the 5 samples
analyzed. Results fall under the detection limits, for pesticides from 0.02 to 0.10 ppm
and for PAH 0.25 ppm.
5.0 Discussion
5.1 The clay and the organic matter content
Sediments at the St. Augustin Lake consist of silt and sand with very high
concentrations of organic matter (OM). By colour, this organic matter can be divided in
two layers of about 30 cm each one. One highly oxidized and deeper and one superficial
and less oxidized. Together with organic matter, clay and carbonates are other adsorbing
materials with affinities for heavy metals. Correlation analysis was performed for all
heavy metals against 3 reactive materials: organic matter, clay content and carbonates.
As an example, Figure 14 shows the positive correlations between Zinc concentration
and OM and clay. Indeed, to sites presenting higher contents on OM or Clay
corresponded higher concentrations of Zn. The relationship was positive with OM
(logarithmic relation) and less pronounced with clay (linear relationship). These results
confirm the affinity of heavy metals for OM and clay colloids.
161 .
,,~_
y= 3 , 5 9 5 2 L n ~
9,0186
0L-- I~ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
0
5O
100
150
2OO
25O
3O0
350
40O
45O
Zinc Concentration (mg/Kg)
i"_6r anic M "e(

Figure 14 - Correlation between Zn Concentration and Clay and Organic Matter
50
5.2 Eutrophication
The high organic matter found in the sediments is a result of the eutrophication status
of the lake. The analyses of nitrate and phosphate did not reveal the presence of these
anions in the interstitial water. Further investigations will determine the forms of
phosphorus in the sediments and their implication in eutrophisation. Otherwise, the
water contains phosphorus at an average concentration of 70 ~tg/L. This level of P may
characterize hypereutrophic lakes according to Ryding and Rast (1994). According to
the Landry (1993) study, nutrients were already present in high concentrations in the
water column. Given the fact that most nutrients inputs (domestic effluents, agricultural
farm discharges) have been closed or reduced, results seem to indicate that these
nutrients, particularly phosphorus, may come from what was accumulated at the bottom
sediments over the last 20 years. The second assumption is that the P and N are
transported by groundwater coming, for example, from the agricultural subsoil
basements.
5.3 The second part
A second phase will study these two possibilities. In addition, phosphorus speciation
into organic and inorganic forms will be determined in order to complete a global mass
balance for nutrients. Although some correlations can be drawn between contaminants
and sediments components, toxicity, mobility or availability of heavy metals are not
evaluated by the method of total concentrations, speciation and distribution information
is essential to evaluate final fate of contaminants or clean-up alternatives. The second
phase of this study will carry heavy metal speciation and bioassays for further evaluation
of toxicity. These analyses will complete the evaluation and determination of the role of
sediments in the process of retention and leaching of heavy metals. The fact is that
heavy metals and arsenic could become available for organisms when a change of
conditions like pH, redox potential, oxygen concentration or bioturbation occurs.
6.0 Conclusions
From the above results and discussions, it is concluded that St. Augustin Lake
presents with respect to sediments two geoenvironmental problems: I) contamination by
metals and 2) contamination by nutrients. Indeed, the five heavy metals and a metalloid
measured were found in concentrations exceeding the background values found in the
literature. In the case of Cd, Pb and in a lesser extent Cr, concentrations exceeded, for
over 50% of the samples, the level of minimum toxicity. These metals were not
distributed homogenously along the surface of the lake; on the contrary, concentrations
were higher at the central north zone and longitudinal extremes of the lake, sometimes
corresponding to points close to the remaining discharges (e.g. highway runoff,
51
recreational activities centre). Total concentrations do not give insight with respect to
toxicity, thus further metal analysis needs to include metal speciation and bioassays. A
high concentration of phosphorus has been found in the water column. The sediments
displayed a high content of organic matter. This is the cause of the eutrophic state of the
lake. Further studies need to be carried in order to identify the sources of P and N and its
distribution in water and sediment components. Finally, the problems at St. Augustin
Lake are representative of a more global and widely spread environmental problem not
only particular to Quebec or the rest of Canada but all over the world. Any attempt to
clean up the water can be jeopardized by the presence of remaining contaminants hold
within the sediments. Thus, this aspect needs to be settled before. Today it exist several
methods for capping, isolating or treating sediments; these alternatives need to be
explored for the sediments in St. Augustin Lake.
Acknowledgments
The authors wish to thank EXXEP Consultants and the FCAR funds for making this
study possible.
References
American Public Health Association, American Water Work Association,

WaterPollution Control Federation (APHA, AWWA, WPCF), 1989, Standards
Methods for the Examination of Water and Wastewater, Port City press, Baltimore.
Baudo, R., 1990, .Sediment Sampling, Mapping and Data Analysis, In: Sediments
Chemistry and Toxicity of in Place Pollutants, Edited by Baudi, R., Giesy, J.P., and
Muntau, H., Lewis Publishers, pp. 15-50.
Bureau de Normalisation du Qu6bec., 1982, Granulats-Analyse Granulom~triquepar
Tamisage, Minist&e de l'Industrie, du Commerce et du Tourisme, BNQ 260-0401982-08-11.
Centre d'Expertise en Analyse Environnementale du Qu6bec, 2001, M~thode d'Analyse
M.A. 100-S.T. 10, D~termination des Solides Totaux et des Solides Totaux Volatils;
M~thode Gravim~trique.
Conseil Canadien des Ministres de l'Environnement (CCEM), 1999, Recommandations
Canadiennes pour la Qualit~ des SOdiments: Protection de la Vie Aquatique,
Winnipeg.
52
Environment Canada (EC), 1992, Review and Recommendations for Canadian lnterim
Environmental Quality Criteria for Contaminated Sites, Inland Water Directorate,
Water Quality Branch, Scientific Series No 197.
Ministry of Environment Ontario (MOE), 1990, Provincial Sediment Quality Guidelines
(draft project),, p. 21.
Landry, P., 1993, Expertise et Am~nagements R~alis~s au lac St. Augustin et
l'Int~rieur de son Bassin de Drainage, 48 p.
Galvez-Cloutier, R., Yong, R.N., Chan, J., and Bahout, E., 1995, CriticalAnalysis on
Sediment Quality Criteria and Hazard-Risk Assessment Tools,. Dredging,
Remediation and Containment Contaminated Sediments ASTM STP No 1293, pp.
306-318.
Roberge, K., Pienitz, R., and Arsenanlt, S., 2002, Eutrophisation Rapide du lac SaintAugustin, Quebec,: Etude Pal$olimnologique pour une Reconstitution de la Qualit~ de
l'Eau, Le Naturaliste Canadien, Vol 126, No 2, pp68-82..
Ryding, S.O., and Rast, W., 1994, Le Contr61e de l'Eutrophisation desLacs et des
R~servoirs, Masson, Unesco, 294p.
St. Lawrence Centre (SLC), 1993, Methods Manual for Sediments Characterization,
Action Plan St.Lawrence.
St. Lawrence Centre (SCL), 1996, Rapport Synth~se sur l'Etat du St-Laurent. Volume
1 :L 'Ecosyst~me du St-Laurent, Environnement Canada- R6gion du Qu6bec,
Conservation de l'Environnement et l~ditions MultiMondes, Montr6al.
Yong, R.N., 1995, The Fate of Toxic Pollutants in Contaminated Sediments. Dredging,
Remediation and Containment Contaminated Sediments ASTM STP No 1293, pp.
306-318.
Wetzel, R.G., 1983, Limnology, 2ndedition, Saunders College Publishing, 767 p.
Thresiamma Joseph/K. K. Balachandran, I Maheswari Nair/V. Kesavadas/K. K. C.

Nair/and Joseph Sebastian Paimpillil2
Trace Metal Levels in Nearshore Sediments Close to Industrial

Discharges off Cuddalore (Bay of Bengal)
Reference: Joseph, T., Balachandran, K. K., Nair, M., Kesavadas, V., Nair, K. K. C.,
and Paimpillil., J. S., "Trace Metal Levels in Nearshore Sediments Close to
Industrial Discharges off Cuddalore (Bay of Bengal)," Contaminated Sediments:
Performance, ASTM STP 1442. J. Locat, R. Galvez-Cloutier, R. C. Charley, and K. R.
Abstract: Trace metals in the sediment and in the overlying water column along with
texture characteristics in the vicinity of industrial discharges at Cuddallore were
analyzed, covering the seasonal changes to identify probable anthropogenic influence
on this coast. During monsoon, cadmium and zinc had insignificant correlation with
other trace metals. Other metals, except Mn, had a positive significant correlation. The
spatial distribution of metals and the correlation were drastically changed during postmonsoon, loosing significant correlation for Pb, Ni, and Mn with other metals and the
observed correlations being mostly negative. Cobalt and copper had strong positive
correlation with all metals. Among the trace metals in water column Cu and Ni had
strong positive correlation (0.88). Most of the correlation was positive except for Fe
and Ni. During post-monsoon, levels of Cd, Co, Cu, Pb, Zn, and Mn had dropped down
to 33-94%. The maximum amount of lowering was noticed for Pb and Cu. Ni (148%)
and iron (110%) levels had an increasing trend. The trace metals in the water column
had a similar decreasing trend (41-76%) for Cu, Ni, Fe and Zn. Cluster analysis had
shown three distinct clusters of sampling points and was classified as least (4 stations),
moderate (5 stations) and high (5 stations) trace metal contaminated regions. The least
contaminated stations were all in the offshore area. Wilks lambda (0.027) denoted the
good discriminatory power of the current model with the three groups. Partial lambdas
indicated that Cr contributed the most, Cd the second most, Ni the third most and Mn
contributed the least for the overall discrimination. The other trace metals also
contribute to this function, but with negative sign. The forcing mechanisms behind the
rapid changes in trace metal levels; their spatial distribution patterns and correlation
strengths were indicated.
Keywords: Trace metals, anthropogenic influence, discriminatory power, monsoon
effect
Scientist, Regional Center, National Institute of Oceanography, Cochin, India.
2 Director, Enviro Solutions, Ferric Lane, Aluva, Kerala, India.
53
www.astm.org
54
Introduction
Marine sediments fringing the west coast of India exhibit well-defined patterns in
their texture, chemical and mineral composition (Murthy et a1.1981). Present
knowledge of the distribution patterns of heavy metals in shelf sediments off the west
coast of India is based on the studies of Murthy et a1.(1981) and Balachandran (2001).
On the east coast of India, the available information on the concentration of heavy
metals in the sediments are limited to the work of Padma and Periakali (1999) and
Senthilnathan and Balasubramanin (1999) and seriously lack a geochemical data from
Cuddallore region.
The Cuddalore Power Company (CPC) situated in the city of Chennai, had set up a
thermal power plant south of Cuddalore Port (South India) with two units with 660 MW
capacities. The site of the plant is about 5 km from Cuddalore Port and 15 km south of
Cuddalore Town. On the eastern side of the plant site is the Bay of Bengal and the
Uppanar estuary borders the western side. The SIPCOT industrial area (mainly
chemical industries) is situated on the western shore of the Uppanar estuary (Fig. 1).
The cooling water is taken from the sea and the heated cooling water is discharged into
sea. In addition to this plant a desalination plant and a demineralization plant were also
set up in the vicinity. The quantity of cooling water intake was of the order of 302, 400
ma/hr. A portion (950 m3/hr) of this water is required for the desalination plant and the
desalinated water output is 328 m3/hr. The remaining water is discharged at the rate of
622 m3/hr. This is mixed with the heated cooling water and is discharged (302,000
m3/hr) into the sea.
The rivers that drain into Bay of Bengal in this region are the Pannaiyar River,
Gadilam River and Uppanar River in the south near Cuddalore Port and Vellar River in
the south near Portonovo. Uppanar River runs somewhat parallel to the coast in this
area and drains into the sea at Cuddalore Port. The Gadilam River also joins the sea at
Cuddalore and a narrow canal connects the estuarine portion of these two rivers.
The continental shelf in this region is narrow (20 kin), taking the 50 m depth as the
edge of the shelf. Near the shore up to 10 m depth, the slope of the sea bottom is
greater and beyond the 50-m the slope is very steep. The slop between 10 and 20 m
contours appears to be gentle. The 20-m-depth contour is at a distance of 9-10 km from
the shore. The nearshore bathymetry in this area appears to undergo seasonal changes.
The coastal region from Nellore to Cuddalore is reported to be strong upwelling zone
during southwest monsoon. As information on geochemistry associated with textural
characteristics of this coastal region can throw some light on the processes by which
they have been incorporated in the sediments, the present studies have been carried out
in this direction.
Materials and Methods

The shelf-sediment samples were taken from 24 stations with three transects
perpendicular to the coast covering the anticipated discharge point. The first sampling
was during the end of southwest monsoon season and the second one was during the
middle of the post-monsoon season (1997-98), thus enabling to decipher the seasonal
JOSEPH ET AL. ON LEVELS IN NEARSHORE SEDIMENTS
Gadilam River
N
D
I
ire
11o45'
Uppanar River - ~
11o40 '
000
17
O0
OOO
16
9 O O 9
12
9
.ql,- ~
tOO
1
O 9
5
4-
79o45,
Figure ! . ~ / i r ~ g
O0
21
9
24
O 9
$
~-
79o~,
~ m ~ o ~ oHCmta//o~
TI
55
56
changes in sediment geochemistry. The sampling stations are shown in Fig. 1 and the
sediment sampling technique is discussed in Anon (1998). Water samples were
collected using Hydrobios-Niskinsamplers and sediments using a Van Veen grab.
Water samples were filtered through 0.45 lain Miillipore filter papers and the filtrate
acidified to pH 4.5 for extraction. The analytical method used for the extraction of
dissolved heavy metals was chelation with ammonium pyrrollidine dithiocarbamate
(APDC) and extraction with methyl isobutyl ketone (MIBK). The organic layer was
then stripped off the metals with concentrated nitric acid and the aqueous extract
analyzed in an air-acetylene flame of AAS (PEAAanalyst100). Textural
characteristics were determined following pipette analysis (Krumbein and Petijohn.
1938) and organic carbon using wet digestion followed by back titration with ferrous
ammonium sulphate (El Wakeel and Riley. 1957). Sediments, dried and finely
powdered, were digested in Teflon TM (trifluoroethylene) bomb using aquaregiahydrofluric acid mixture at 90~ for 2 hours, cooled and made up to 50 mL in a boric
acid medium (Loring and Rantala. 1992). Analytical accuracy was checked by running
a procedural and a standard reference material (BCSS-1) as a control and the precision
was found to be better than 15%.
The sediments in the nearshore area move along the coast due to the action of waves,
which generate longshore currents and cause sediment transport. The littoral transport
in the region appears to be northerly from April to October and southerly from
November to March. The annual net transport is northerly between Pondichery and
Allur. The longshore transport in the Cuddalore area is about 6.98 x 102 m3/yr when the
direction is northerly (Anon. 1998). In general, the annual net transport in this region
appears to be low as compared to other places. The net direction of transport is towards
southwesterly as compared to other places.
Many trace metals such as copper, chromium, zinc and lead get rapidly fixed to
particulate matter and carried to the bottom because of which the sediments are
considered to provide reliable estimate of metal pollution levels. It was seen that all the
trace elements were more enriched in the nearshore sediments as compared to offshore
region. The variation in the levels were in narrow range and comparable to the reported
values for the East Coast of India. Organic carbon in these sediments were very low
(Tables 1 and 2) and hence, these concentrations could be taken as lithogenic in origin.
The highest heavy metal concentrations in the shelf sediments (Muraleedharan Nair
et al. 1992) at the southern tip of West Coast of India were of the order of Cr 250, Cu
38, Ni 80, Pb 37 and Zn 11 ppm. The concentrations of copper during the study period
had ranged from 0.77 ppm to 16 ppm (Table. 3a, 3b). The mean concentration of nickel
was comparable to that reported for unpolluted areas. The high values for certain trace
metals may be due to the domestic and industrial pollutants. The concentrations of
various trace elements follow closely the texture of sediments (Table. 2). Higher
concentrations of elements are associated with the silt-clay fractions of sediments. In
Table 4, the concentrations of heavy metals reported from some of the unpolluted
coastal regions are given. A perusal of these values with the average values for
Cuddalore shelf sediments revealed that the sediments of the coastal region do not
JOSEPH ET AL. ON LEVELS IN NEARSHORESEDIMENTS
Table 1 - Sediment characteristics offCuddallore

Stations
Sand %
Clay %
Silt %
1
2
4
5
6
7
8
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
54.32
93.21
92.53
94.45
95.80
95.6t
98.73
98.00
97.15
97.07
94.96
91.06
16.66
95.50
92.99
32.07
19.99
90.94
85.87
52.97
94.87
97.23
8.00
2.95
6.90
4.15
0.30
4.30
0.35
1.50
2.25
2.50
3.70
8.50
30.45
2.35
4.50
39.55
43.80
9.00
13.55
33.55
4.35
2.75
37.68
3.84
0.57
1.40
3.90
0.09
0.02
0.50
0.60
0.43
1.34
0.44
52.89
2.15
2.51
28.38
36.21
0.06
0.58
13.48
0.78
0.02
Organic Carbon (mg/g)

14.27
3.57
3.57
2.98
1.19
5.95
6.55
9.52
2.98
8.33
5.95
3.57
11.90
0.60
12.50
9.52
0.60
3.57
5.36
4.76
0.60
0.60
Table 2 - Sediment Characteristics along different transects (Ranges and Averages)

Transect
T1
T2
T3
Sand %
54.32-98.73
(89.24)
16.66-98.00
(84.34)
19.99-97.23
(70.87)
Clay %
Silt %
0.30- 8.00
(3.85)
1.50-30.45
(7.32)
2.75-43.80
(18.88)
0.02-37.68
(6.90)
0.43-52.89
(8.34)
0.02-36.21
(10.25)
( ) Values in paranthesis are the mean for each transect.
Organic carbon (rag/g)

2.58-14.27
(5.44)
0.60-11.90
(6.12)
0.60-12.50
(4.68)
57
58
Table 3a - Heavy Metals in Sedtments , September, 1997
Stn.
Cd
Co
Cr
Cu
Fe
Mn
1
2
4
6
7
8
10
11
13
15
16
17
18
20
22
23
24
2.1
2.5
1.9
1.5
1.6
2.4
1.2
1.7
1.9
2.4
2.1
1.6
1.6
2.1
1.2
0.8
1.2
25.1
24.5
16.1
9.4
14.7
15.7
22.9
16.5
15.3
30.6
28.2
26.8
24.7
25.1
10.6
11.3
16.1
59.2
45.3
29.9
29.1
36.9
36.3
37.0
33.7
84.8
82.9
83.6
79.2
65.7
77.4
38.8
35.5
46.6
13.5
10.6
4.4
4.5
5.4
6.1
8.0
6.4
5.2
16.3
7.3
5.3
11.7
15.1
3.7
4.0
7.8
5.5
212
14.8 388
10.1 306
6.7
330
8.7 577
9.2 523
10.2 286
10.0 525
9.4 332
21.3
460
1 4 . 8 386
16.0 320
17.6 372
17.1
377
9.7
354
7.7 424
12.1 535
Ni
Pb
Zn
33.0
27.1
17.3
8.2
17.2
14.9
23.4
20.0
16.5
41.9
27.0
25.6
29.3
34.1
14.4
15.4
22.8
55.4
37.9
25.6
25.8
36.1
35.6
46.6
28.1
34.6
63.0
47.2
35.8
63.6
64.5
25.9
27.7
27.3
30.5
111.0
14.0
9.7
12.5
16.6
25.5
20.5
15.8
50.7
34.1
32.4
47.9
42.0
15.4
13.1
39.8
* concentration in ppm dry weight except for Fe, mg/g

Table 3b - Heavy Metals in Sedtments , January, 1998
Stn.
Cd
Co
Cu
Fe
Mn
Ni
1
3
5
6
7
13
14
15
17
18
19
20
21
22
23
24
1.7
1.2
1.7
0.4
0.7
1.3
1.7
0.9
1.7
2.1
1.3
1.7
1.7
0.4
0.4
1.7
24.5
13.5
19.6
12.2
12.5
21.4
29.5
26.8
26.6
23.6
9.7
21.7
19.2
11.5
10.9
17.0
8.9
0.8
3.6
0.9
0.7
3.5
10.0
3.6
8.1
6.1
0.9
6.3
4.4
0.9
0.8
0.9
21.4
5.4
9.4
7.5
7.7
11.6
21.8
17.4
21.0
19.2
6.7
17.7
15.3
9.5
7.3
8.9
454
30.8
360
15.4
143 25.2
497
12.5
363
14.2
380 22.4
357
31.5
425
29.1
562
26.9
445
28.2
423
29.0
105 39.2
383
26.6
402
55.1
285 104.0
424
35.3
Pb
Zn
20.9
9.8
21.4
10.6
9.1
10.4
21.4
10.6
21.0
10.3
21.1
21.0
0.0
10.0
9.5
10.6
34.6
16.1
16.4
8.6
6.6
16.3
48.5
33.0
35.9
30.5
11.8
27.4
13.3
8.8
6.9
9.5
JOSEPH ET AL. ON LEVELS IN NEARSHORESEDIMENTS
Table 4 - A v e r a g e concentration of heavy metals in coastal sediments in the study
area and for other regions (cited from Balachandran, 2001)

No.
Region
Fe %
Zn
Mn
Ni
Co
Cu
Cd
1.
2.
3.
4.
Present study
SW Coast of
India (Cochin)
Cochin estuary,
India
Pulikkat estuary
(east coast of India)
Pondicherry harbour
(east coast of India)
Mumbai, India
Thana creek, Mumbai
8.9
35
385
27
19
--
3.3
99
171
72
22
32
--
5.5
1266
337
67
25
53
--
495
. . . . . .
5.
6.
7.
8.
2.8
--
. . . . . .
-138
1059
-- 206
980
.
89
132
51
50
8.6
42
105
132
2.5

Table 5 - Trace metal concentration in sediment and water column during
Southwest monsoon and post-monsoon period.
Water
SW Monsoon
Post-monsoon
Ratio
Sediment
SWMonsoon
Post-monsoon
Ratio
Cu
Pb
Cd
0.7
1.0
1.4
. . . . . .
. . . . . .
. . . . . .
8.0
3.4
0.4
40.1
13.5
0.3
1.7
1.2
0.7
Co
19.6
18.6
0.9
Ni
Zn
Fe
Mn
1.2
1.6
1.4
3.0
7.2
2.4
31
67
2.1
----
31.3
19.3
0.3
11.6
12.8
1.1
395
343
0.9
22.8
33.7
1.5

indicate any sign of pollution. The effects of nearshore circulation in the trace metal
levels in coastal sediments was clear during the period of study with the high
concentration levels during the withdrawal of southwest monsoon season. Enrichment
during the post-monsoon period was noticed only for nickel and the levels of iron
remained the same. The significant reduction in concentration levels was noticed for
copper and lead. During the same period, the overlying water column had experienced
enrichment in heavy metals as shown in Table 5.
A cluster analysis using the heavy metal concentrations had shown two cluster
formations. The stations on the northern transect and nearshore stations in middle
transect formed the first cluster. The offshore stations in all the three transect form the
next cluster. Most of the heavy metals had very high levels at stations in cluster one.
59
60
This indicates that the sources of trace metals are coastal points north of transect one
and it hints towards the estuarine and industrial contributions.
The trace metals cadmium and zinc had no significant correlation with other metals.
Other metals had a positive significant correlation except for Mn. After a time gap of
six months, the correlation had changed drastically with lack of significant correlation
for Pb, Ni and Mn with other metals and with negative correlation. Cobalt and copper
had strong positive (> 0.70) correlation with other metals. Cadmium and zinc had
strong positive correlation with Cu, Co and Fe. Among the trace metals in water
column Cu and Ni had strong positive correlation (0.88). Most of the correlation was
positive except for Fe and Ni. After a six-month gap, no correlation among the metals
was statistically significant and most of them were negative.
The spatial distribution of copper in the coastal sediments had shown similar patterns
during the two seasons with the maximum concentration centered at coastal stations of
transect 2. There was a noticeable reduction (42%) in copper concentration at the
coastal stations during the post-monsoon season. The cadmium distribution was not
having a well-defined spatial pattern and with no nearshore enrichment. During the
post-monsoon season, a slight increase in cadmium levels was noticed at coastal
stations in the northern transect. The distribution pattern of lead remained to be similar
to the copper distribution with the coastal stations at mid transect showing the
maximum levels and with an offshore decrease of concentration. During the postmonsoon season, the lead concentration had reduced by 50% and the well-marked high
levels at mid transect had disappeared with almost same concentration at coastal
stations. The pattern of cobalt distribution remained nearly the same during the two
seasons with a pronounced peak at coastal station in mid transect. There was no
significant reduction in concentration levels of cobalt during the two seasons. The
spatial distribution pattern for zinc had shown a pocket with very high zinc levels (>100
ppm) in the coastal waters on the southern transect at the end of the monsoon season.
During the post-monsoon sampling period, the zinc levels in coastal sediments had
reduced by one third of its monsoon level. This spatial distribution patterns indicated a
flux of zinc in the northward direction or a local enriching contribution from a land
based source.
In a canonical correlation analysis using trace metals in the three transect, the main
root had accounted for 92% of the variability of the canonical variate. The main
contributions to the canonical root were from Cd, Cu, Ni and Co in the order of
magnitude. The sampling stations were grouped into a contaminated coastal (station
no. 1,15, l 6,17,18), moderately contaminated nearshore (st.nos. 2,3,13,14,19) and an
uncontaminated offshore (st.nos. 8,9,10,24) zones, based on the distance from shore
line. A discriminant analysis was conducted to see which trace metal discriminate
between the three groups. Wilks lambda (0.27) denoted the good discriminatory power
of the current model with the three groups. Partial lambdas for overall discrimination
showed the magnitude of contribution in the order Cr > Cd > Ni > Mn. The first root
accounts for 81% of the explained variance, inferring that a single function could
explain 81% discriminatory power, thus establishing the role of Cr, Cd and Ni in
classifying the contamination levels of sediment. The other trace metals also contribute
to this function, however with negative sign.
JOSEPH ET AL. ON LEVELS IN NEARSHORE SEDIMENTS
Conclusion
The industrial development along the coastline had shown some impact on the
environment of the shelf region. Heavy metal distribution in the coastal region showed
an alongshelf pattern. Metal distribution in this shelf is controlled by the fine sediment
dynamics which are mainly regulated by the low-energy alongshore currents. As the
shelf currents do not effectively disperse the anthropogenic discharges, the metal
enrichment is restricted to the sediments of shallower regions only. The present study,
in general, indicates that trace metals do not seem to constitute a potential threat to
marine environment of this region and the release of metals into the coastal zone
through anthropogenic activities are adequately dealt with by the dominant coastal
processes of this region resulting in minimal biological impact. Taking the data
presented here as the baseline, it may be worthwhile to resurvey the area to understand
the impact of the recent developmental activities and to what level the area has been
polluted.
Reference
Anon, 1998. Rapid environmental impact assessment of proposed jetty, cooling water
intake and disposal of heated cooling water off Cuddalore coast for CPC thermal
power plant. Technical report of National Institute of Oceanography, Goa, India.
pp.93.
Balachandran, K.K, 2001. "Chemical oceanographic studies of the coastal waters of
Cochin", Ph.D Thesis, Cochin University of Science and Technology, Cochin,
pp. 148.
El Wakeel, S.K. and Riley, J.P., 1957. "The determination of organic carbon in marine
muds". Journal of Du. Conseil International Pour. Explor. Mer., Vol.22, pp. 180.
Krumbein, W.C. and Pettijohn, F.J., 1938. "Manual of sedimentary petrography".
Appletion Century Crofts Inc., New York, pp.549.
Loring, D.H. and Rantala, R.T.T., 1992. "Manual for Geochemical Analsis", Earth
Science Reviews, Vol.32, pp.235-283.
Muraleedharan Nair, M.N. and Ouseph, P.P., 1992. "Elemental distribution in the
nearshore sediments along the Kerala Coast". Proceedings, Fourth Kerala Science
Congress, pp.37-38.
Murthy, P.S.N. and Veerayya, M., 1981. "Studies on the sediments of Vembanad Lake,
Kerala State: Part IV- Distribution of trace elements", Indian Journal of Marine
Science, Vol. 10, pp. 165-172.
Padma, S. and Periakali, P., 1999. "Physico-chemical and geochemical studies in
Pulikkat Lake, east coast of India", Indian Journal of Marine Science, Vol.28,
pp.434-437.
Senthilnathan, S. and Balasubramanian, T., 1999. "Heavy metal distribution in
Pondicherry Harbour, southeast coast of India", Indian Journal of Marine Science,
Vol.28, pp.380-382.
61
Mafia Lucia K. Rodrigues, I Mafia Teresa Raya-Rodriguez, 2 and Val6rio D. Pillar 2

Randomization Tests: A Statistical Tool to Assess Heavy Metal Pollution in Cai
River Basin Sediments (RS, Brazil)
Reference: Rodrigues, M. L. K., Raya-Rodriguez, M. T., and Pillar, V. D.,

"Randomization Tests: A Statistical Tool to Assess Heavy Metal Pollution in Cai
River Basin Sediments (RS, Brazil)," Contaminated Sediments: Characterization,
Evaluation, Mitigation~Restoration, and Management Strategy Performance, ASTM STP

1442, J. Locat, R. Galvez-Cloutier, R. C. Chancy, and K. R. Demars. Eds., ASTM
Abstract: This study evaluated the use of statistical randomization tests in data analysis
regarding heavy metals content in sediments from the Cai River basin (Brazil). During
one year, four samplings took place at eleven sites in the Cai River and tributaries. The
results showed significant differences between some locations, enabling spatial
segregation of sites affected by natural and anthropic sources. Considering a significance
level (a) of 5%, the control station, above polluted areas, showed the best quality; Cai
River station after the mouth of Pinhal/Belo stream indicated influence from the
industrial complex of Caxias do Sul region, as well as copper compounds use in
vineyards; and the Cadeia and Forromeco streams showed effects of natural contribution
from basalts and surface runoffon naked soils. For a =10%, it was possible to identify
chromium pollution by tanneries in the Cadeia stream. Long-term studies should be
performed, since understanding temporal aspects such as alteration processes in
sediments involves a much longer time scale.
Keywords: randomization tests, multivariate statistics, sediment, heavy metals
i Chemical Engineer, Divisao de Quimica, FEPAM, Rua Aur61io Porto 45, 90620-010
Porto Alegre, RS, Brazil.
2 Professor, Departamento de Ecologia, UFRGS, Av. Bento Gonqalves 9500, 91540-000
Porto Alegre, RS, Brazil.
62
www.astm.org
RODRIGUES ET AL. ON RANDOMIZATION TESTS
63
Introduction
Rivers act as carriers of dissolved heavy metals or those associated with suspended
solid particles. They are also potential reservoirs of these elements, by deposition in the
bottom sediment and assimilation by the biota. In aquatic ecosystems, sediment is the
main compartment of accumulation, reprocessing and transference of heavy metals
(FOrstner and Wittman 1981). The potentially mobile metallic elements, with greater
bioavailability, can be released from sediments by microbial activity and changes in the
various physicochemical factors that characterize the aquatic environment.
In order to analyze data resulting from monitoring multiple variables in
environmental samples, statistical methodologies must be sought in order to show the
significance of differences, both spatial and temporal, and enable the preparation of
consistent diagnoses.
Among the techniques that involve multivariate statistics, randomization tests have
demonstrated their usefulness in comparing groups of sampling units described by a
given number of variables (Pillar and Orlrci 1996). The groups of sampling units may be
defined by one or more external factors, such as location, time of data collection, or
experimental treatments, arranged in different designs. The advantage of these tests is
that they do not require the assumptions involved when classic statistical tests are used
(Manly 1991). A randomization test generates a probability P which is used to judge
whether the groups differ significantly. P is generated under the null hypothesis (Ho) that
the difference between groups, measured by an appropriate test criterion, is the result of
chance alone. In other words, under 11othe observations for the variables in a given
sampling unit are independent from the group to which the sampling unit belongs. I f P is
sufficiently low, compared to a defined level of significance a, it is rare to find by chance
a difference between groups as large as the observed one. Therefore, the null hypothesis
is rejected. On the contrary, i f P is larger than a, Ho is accepted, and it is concluded that
there is no significant difference between the groups.
In classic statistics, P is obtained assuming that the distribution of the test criterion
(e.g., F, t) agrees with the theoretical distribution, usually the normal one. In
randomization tests, P is generated from the data themselves, considering the conditions
established in tlo. The observed set of data is viewed as one of the many possible
outcomes were Ho true. That is, under 11o any permutation of the data vectors among the
groups is equally possible. A data vector contains observations on one or more variables
describing a sampling unit. Based on this, an iterative algorithm is used for the random
generation of permutations. Sampling units may be stratified in blocks, in which case
random permutations are restricted to within blocks. The method described in Pillar and
Orl6ci (1996) uses as test criterion a sum of squares between groups (Qb) defined as
(1)
Qb = Qt - Qw
where
n--1
/1
2
r l h=l i=h+l
(2)
64
is the total sum of squares involving n(n-1)/2 pair-wise squared distances ( d~i) between n
sampling units and
k
Qw = ~.,Qw~
(3)
Cffil
is the sum of squares within k groups, such that

n--1
tl
2
Qwc = - - ~ ' . ~-',d;,,le
(4)
nr h=l i=h+l
is the sum of squares involving no(no-1)/2 pair-wise squared distances between nc

sampling units belonging to group c. At each iteration a new Qbois computed and
compared to the Qb observed in the data. After many iterations, usually not less than
1000, the proportion of iterations in which Qoo > Qb is taken as the probability P(Qbo>_
Qb) sought by the test. By convention, the observed data set is taken as the first iteration
and thus the minimum value for P is
P = (1 / total number of iterations)
(5)
The greater the number of iterations, the more reliable will be the test result, since
the value of P will be closer to that which would be obtained if we had worked with all
the possible permutations.
In this paper we evaluate the usefulness of a randomization test to analyze the
association of heavy metals content in sediments with location and time of sample
collection, in the Cai River basin, south Brazil.
Methodology
Study Area
The Cai Basin, located in the state of Rio Grande do Sul in the extreme south of
Brazil, has a surface of 5000 km 2 (Figure 1). The Cai River is a tributary of Guaiba Lake
and extends about 260 km from the headwaters in the Plateau (1000 m altitude) until the
mouth almost at the sea level. On the banks of the Cai River significant urban and
industrial concentrations occur, especially from petrochemical, metallurgical, tannery,
food and textile branches, that are potential sources of heavy metals. The region also has
concentrated agricultural activities, such as viticulture, vegetable and fruit production, in
which fungicides and inputs containing these elements are applied. The local geology
consists mainly of Serra Geral Formation rocks, in which basalts that are naturally rich in
metallic elements predominate (FEPAM/GTZ 1997).
65
COMI~LEX
Cmd~ do
Ysrro~llha ~
/"
_
Mont~
Sampling
.....
sites
1 - CA 210
2-3- CA
Pinha//Belo
136
t'trom
,lt~qt
,
I~"
4 - Forremeco stream
PETROCHEM/CAL~'~
COMPLEX
6 - Ctdoig stremn
7 - CA O~O
8 - CA 24.1
9 - Born Jm-dimstremn
I0- CA 1 0 . 6
11 - CA 001
~0"30'
3o"
0
lOkm
51 30'
9
Figure
I -
~t'
,,
i~
. . . . . .
Location of Sampling Stations in the Cai River Basin, Rio Grande do Sul State
(RS), Brazil
66
Sampling and Analysis of Sediment Samples

As described by Rodrigues (1997), from July 1994 to July 1995 two sampling
expeditions were performed during the flood season and two during the dry season, at
seven stations along the Cai River. In addition, four stations near the mouths of some
important tributaries were also sampled. Sampling stations located in the Cai River were
identified by the letters CA followed by the distance in kilometers from the mouth
(Figure 1). As a control point, sediment was collected in the Cai River, at a site located
above the areas impacted by domestic and industrial discharges.
The elements cadmium (Cd), copper (Cu), lead (Pb), nickel (Ni), chromium (Cr),
vanadium (V), zinc (Zn), aluminum (A1), iron (Fe), manganese (Mn) and titanium (Ti)
were analyzed in the silt-clay fraction (dry weight), by inductively coupled plasma
optical emission spectrometry (ICP/OES). In the extraction of total metals content,
0.25000 g of sediment samples were digested with 2 mL of hydrofluoric acid (HF) and3
mL of nitric acid (HNO3) in PTFE closed vessels, heated in an oven at 150~ for 4 hours.
In order to eliminate HF residues, the extracts were dried on a hot plate and then diluted
to an appropriate volume. For potentially mobile metals, a partial determination (Malo
1977) was used, involving digestion with hydrogen peroxide (H202) and hot extraction
with 0.3 M hydrochloric acid (HC1).
Grade p.a. reagents that had previously been tested as to metal content and ultrapurified water were used to process the samples duplicates. The certified reference
material CANMET STSD-4 was analyzed at the same time in order to assess the
accuracy of results for total metals content. The mean recovery was between 95 and
105%, except for Cr (75%) and for A1 (35%), which was further analyzed by X-ray
fluorescence.
Additionally, organic matter was determined by calcinating the samples at 450~
until constant weight. Grain size analysis was performed using sieving and sedimentation
techniques.
Statistical Analysis of the Data Using Randomization Tests

The initial analysis aimed at assessing whether the sampling sites differed
significantly in terms of metal concentration in the sediments by following the
randomization test described by Pillar and Orl6ci (1996). In data processing, the sampling
dates were fixed as blocks in order to evaluate the existence of a significant difference
between the sampling sites. The null hypothesis in this case was that the metal content in
the sediments was independent from the sampling site. In a second analysis the sampling
locations were taken as blocks to test whether the sampling dates differed. In this case the
null hypothesis was that the metal content in the sediments was independent from the
sampling date.
The database, with 22 columns corresponding to the variables, and 44 rows
designating the sample units, was processed via the MULTIV computer program (Pillar
2000). Total AI, analyzed only in the third sampling campaign (n=l 1), was ignored, in
order to avoid gaps in the data matrix. The data set was submitted to standardization by
centering and normalization within variables. The test criterion Qb was computed from
pair-wise Euclidean distances between sampling units. The probability P(Qbo> Qb) was
obtained by I 000 iterations of random permutation. Specific contrasts between groups,
67
pair-wise or multiple, were also evaluated, for which probabilities P(Qbo> Qb) were
generated and interpreted in a w a y similar to a Scheff6 test (Pillar and Orl6ci 1996).
A significance level a = 0.05 was used to interpret the results, which means that the
probability o f error in rejecting Ho is 0.05. Comparing the probability P obtained in the
randomization test to the chosen level o f significance, the null hypothesis established for
each case was rejected or not. In the final assessment, in order to provide a better
demonstration o f the existence o f tendencies, the results were also discussed as regards an
(z = 0.10, since a strict criterion often do not apply to environmental studies involving
extremely c o m p l e x phenomena, with a high degree o f uncertainty.

The analytical results o f the sediment samples are summarized in Table 1. In the
first statistical analysis o f data, comparing the different sampling locations, the use o f
randomization test generated a probability P=-0.001, which indicated the sites differed
significantly. Table 2 displays the results o f the contrasts between sites, which enabled
the following observations.
Table 1- Heavy Metals and Organic Matter Contents (Dry Weight, Fraction < 63 pm) in
the Sedimentsfrom Cal River Basin (RS, Brazil), Average o f l l Sampling Locations in
the Periodfrom July 1994 to July 1995
Variable
Mean
P/T
(%)
44.3
Cd T (mg/kg)
44
1.10
Cd P (mg/kg)
44
0.487
Cu T (mg/kg)
44
102
47,3
Cu P (mg/kg)
44
48.2
Pb T (mg/kg)
44
34.8
58.3
Pb P (mg/kg)
44
20.3
Fe T (mg/kg)
44
53 457
30.6
Fe P (mg/kg)
44
16 333
Ni T (mg/kg)
44
50.2
27.3
Ni P (mg/kg)
44
13.7
Zn T (mg/kg)
44
136
61.5
Zn P (mg/kg)
44
83.7
Mn T (mg/kg)
44
1 080
88.0
Mn P (mg/kg)
44
950
Cr T (mg/kg)
44
117
46.8
Cr P (mg/kg)
44
54.8
AI T (mg/kg)
11
89 155
8.3
AlP (mg/kg)
44
7415
Ti T (mg/kg)
44
10 429
0.8
TiP (mg/kg)
44
78.9
V T (mg/kg)
44
230
21.6
V P (mg/kg)
44
49.6
Orsanic Matter (%)
44
14.8
Note: T = total content; P = potentially mobile content
Minimum
Maximum
0.445
< 0.220
48.1
16.0
21.2
11.9
24 659
9 946
7.68
1.28
67.4
24.2
432
389
14.1
2,31
84 300
4966
8 196
32.1
101
19.2
10.7
1.73
0.811
158
82.5
49.2
36.0
72 276
23 355
71.2
26.8
352
321
1 704
1 479
305
223
106 800
9633
14 071
142
356
78.9
18.5
Standard
deviation
0.293
0,145
24.8
15.5
7.81
5.25
10 101
3 368
14.9
5,79
57.4
54.1
299
297
56.7
50.5
7 508
1277
1 501
23.6
54.3
15.0
1.47
68
Results of Randomization Tests by Analyzing Contrasts Between Sediment

Sampling Stations in the Cai River Basin (RS, Brazil), from July 1994 to July 1995
Table 2 -
Contrasts
Sites
Decision about
Decision
Ho(ct =0.O5)
I~1000000000
10-100000000
100-10000000
1000-1000000
10000-100000
100000-10000
1000000-1000
10000000-100
100000000-10
1000000000-1
01-100000000
010-10000000
0100-1000000
01000-100000
010000-10000
0100000-1000
01000000-100
010000000-10
0100000000-1
001-10000000
0010-1000000
00100-100000
001000-10000
0010000-1000
00100000-100
001000000-10
0010000000-1
0001-1000000
00010-100000
000100-I0000
0001000-1000
00010000-100
000100000-10
0001000000-1
00001-100000
000010-10000
0000100-1000
00001000-100
000010000-10
0000100000-1
000001-10000
0000010-1000
00000100-100
000001000-10
0000010000-1
0000001-1000
00000010-100
000000100-10
0000001000-1
00000001-100
000000010-10
0000000100-1
000000001-10
0000000010-1
0000000001-1
CA 210 - Pinhal/Beio
CA 210 - CA 136
CA 210 - Forromeco
CA 210 - Cadeia
CA 210 - CA 075
CA 210 - CA 050
CA 210 - CA 24,1
CA 210 - Born Jardim
CA 210 - CA 10,6
CA 210 - CA 001
Pinhal/Belo - CA 136
Pinhal/Belo - Forromeco
Pinhal/Belo - Cadeia
Pinhal/Belo - CA075
Pinhal/Belo - CA 24,1
Pinhal/Belo - Born Jardim
Pinhal/Belo - CA 10,6
CA 136 - Forromeco
CA 136 - Cadeia
CA 136 - CA075
CA 136 - CA 050
CA 136 - CA 24,1
CA 136 - CA 10,6
CA 136 - CA 001
Forromeco - Cadeia
Forromeco - CA075
Forromeco - CA 050
Forromeco - CA 24,1
Forromeco - Bom Jardim
Forromeco - CA 10,6
Forromeco - CA 001
Cadeia - CA 075
Cadeia - CA 050
Cadeia - CA 24,1
Cadeia- Bom Jardim
Cadeia- CA 10,6
Cadeia - CA 001
CA075 - CA 050
CA075 - CA 24, I
CA075 - Born Jardim
CA075 - CA 10,6
CA075 - CA 001
CA 050 - CA 24,1
CA 050 - CA 10,6
CA 050 - CA 24,1
CA 24,1 - Born Jardim
CA 24,1 - CA 10,6
CA 24,1 - CA 001
Born Jardim - CA 10,6
Born Jardim - CA 001
CA 10,6 - CA 001
0.001
0.001
0.001
0.001
0.001
0.002
0.002
0.003
0.001
0.004
0.455
0.016
0.001
0.308
0.309
0.408
0.232
0.245
0.174
0.001
0.001
0.018
0.016
0.027
0.012
0.013
0.007
0.074
0.523
0.473
0.133
0.054
0.165
0.153
0.021
0.063
0.010
0.006
0.024
0.026
0.917
0.726
0.138
0.477
0.448
0.932
0.361
0.969
0.983
0.391
0.931
0.889
0.291
0.227
0.981
Rejected
Rejected
Rejected
Rejected
Rejected
Rejected
Rejected
Rejected
Rejected
Rejected
Accepted
Rejected
Rejected
Accepted
Accepted
Accepted
Accepted
Accepted
Accepted
Rejected
Rejected
Rejected
Rejected
Rejected
Rejected
Rejected
Rejected
Accepted
Accepted
Accepted
Accepted
Accepted
Accepted
Accepted
Rejected
Accepted
Rejected
Rejected
Rejected
Rejected
Accepted
Accepted
Accepted
Accepted
Accepted
Accepted
Accepted
Accepted
Accepted
Accepted
Accepted
Accepted
Accepted
Accepted
Accepted
Different
Different
Different
Different
Different
Different
Different
Different
Different
Different
Not different
Different
Different
Not different
Not different
Not different
Not different
Not different
Not different
Different
Different
Different
Different
Different
Different
Different
Different
Not different
Not different
Not different
Not different
Not different
Not different
Not different
Different
Not different
Different
Different
Different
Different
Not different
Not different
Not different
Not different
Not different
Not different
Not different
Not different
Not different
Not different
Not different
Not different
Not different
Not different
Not different
69
The control point (CA 210), which presented the lowest concentrations of the
evaluated elements and consequently the best environmental quality, differed
significantly from all the other sampling sites at a 5% level o f significance (a).
For a = 0.05, the sampling site CA 136 was only similar to the Pinhal/Belo stream
station, that receives part o f industrial and domestic runoff waters from Caxias do Sul
city. When compared to the control point, these two locations (especially CA136) showed
increased concentrations o f the trace-elements Cd, Cu, Cr, Pb, Ni and Zn, which could be
associated to metallurgical, electroplating and textile industries concentrated in this
region. The results also suggested the influence o f the use o f copper-based fungicides in
viticulture.
Considering ~ = 0.05, the Cadeia stream was different from the other stations,
except for point CA 050 in the Cal River, which suffers its direct influence, and for
Forromeco stream that equally drains major agricultural areas in the basin, and is also
subject to erosion. In the sediment o f these streams, comparatively high contents o f Fe.
Ti, Mn, A1, and V were observed. The relief o f these regions, located on steep slopes with
a predominance o f basalts, together with the action o f rainfall on naked soils, could favor
the bank erosion and the entrainment o f material to the stream beds.
If an ~ = 0.10 were chosen, all the sampling points would be considered
significantly different in relation to the Cadeia stream. For ~ = 0.05 the power of the test
was not sufficient to segregate the Forromeco and Cadeia streams, but it is known that a
large input o f pollution by chromium has an impact on the latter due to the contribution
o f tannery effluents (FEPAM/GTZ, 1997).
In the final stretch o f the Cal River, under influence o f a petrochemical complex, all
the sampling stations were similar for ~ = 0.05, including the upstream points CA 050
and CA 24.1. Therefore, in this study, no influence of the petrochemical complex was
detected on the contents o f the elements analyzed in the sediment samples. However, the
grain size distribution in the sediments showed it is a preferential site for fine particles
settlement and may represent a reservoir for trace elements. Due to possible changes in
aquatic environmental properties and dredging activities, which can modify redox
conditions, this would be a potentially fragile site concerning trace elements
remobilization to the water column.
Table 3 - Results of Randomization Tests by Analyzing Contrasts Between Sediment
Sampling Dates in the Cai River Basin (RS), from July 1994 to July 1995
Contrasts
Dates
1 -1 0 0
1 0 -1 0
I 0 0 -1
0 1 -1 0
0 1 0 -1
0 0 1 -1
July/94 - November/94
July/94 - MarclV95
July/94 - July/95
November/94- March/95
November/94- July/95
March/95 - July/95
P
0.251
0.031
0.016
0.337
0.082
0.203
Decision about
1to( 0~=0.05)
Accepted
Rejected
Rejected
Accepted
Accepted
Accepted
Decision
Not different
Different
Different
Not different
Not different
Not different
In the second statistical analysis, related to the temporal aspect, the randomization
test indicated that the sampling dates differed significantly (P=0.025). Table 3 contains
the results obtained in the contrasts, showing similar tendencies between the subsequent
70
collection dates, to a 5 % level of significance. For intercalated collection dates, there

was a significant difference when one considered at = 0.05, except for the pair of dates of
November/94-July/95, where the difference was significant with a = 0.10.
Summary and Conclusions
The randomization tests proved to be an appropriate tool for the spatial assessment
of the proposed database, showing consistency in the segregation of areas that suffer the
impact of different sources of heavy metals contribution.
Considering at =0.05, the results showed a better environmental quality for the
control point and contamination of the area affected by the proximity of Caxias do Sul
industrial complex with metals Cd, Cu, Pb, Ni and Zn. Forromeco and Cadeia streams,
with comparatively high contents ofAl, Fe, Mn, Ti and V, did not differ significantly,
showing the effect of natural contribution by basaltic rocks, associated with surface
runoff through naked soils on steep slopes. Considering at =0.10, Cadeia stream showed a
tendency to differ from all other sampling units, indicating the influence of Cr emission
by tanneries. In the stretch located between the Cai River mouth and kilometer 50,
including the Bom Jardim stream that drains a petrochemical complex area, all collection
stations did not differ significantly, with a = 0.05. These results indicated that, during the
monitoring period, no detectable influence of the petrochemical complex on the contents
of the elements analyzed in the sediment was observed. However, the final stretch of the
Cai River, under influence of the complex, is a preferential site for fine particles
settlement and may represent a reservoir for trace elements, which can be subjected to
remobilization to the water column. Therefore, this section of the river also deserves
great attention in terms of monitoring and preservation of environmental quality.
As to the temporal aspect, the randomization tests showed similarity between the
subsequent collection dates, to a 5% level of significance. Meanwhile, long-term studies
should be performed, since understanding processes of change in the sediments involves
a larger time scale.
Be that as it may, the possibility of applying this statistical tool, to data
interpretation and to preparing spatial and temporal environmental diagnoses, was clearly
demonstrated.
References
1.FEPAM-GTZ, 1997, "Levantamento dos Principals Usos do Solo eda ./~gua na Bacia
Hidrogr~ifica do Rio Cai," Vol.1, Projeto de Cooperag~o Trcnica Brasil /
Alemanha, Porto Alegre, Brasil, 138 p.
2.F6rstner, U., Wittman, G.T.W., 1981, "Metal Pollution in the Aquatic Environment,"
Springer-Verlag, Berlin, 486 p.
3.Malo, B.A., 1977, "Partial Extration of Metals from Aquatic Sediments,"
Environmental Science & Technology, Vol. l 1, pp. 277-282.
RODRIGUES ET AL. ON RANDOMIZATIONTESTS
71
4.Manly, B. F. J., 1991, "Randomization and Monte Carlo Methods in Biology,"

Chapman and Hall, London, 281 p.
5.Pillar, V.D.P., Orl6ci, L., 1996, "On Randomization Testing in Phytocoenology:
Multifaetor Comparisons of Relev~ Groups," Journal of Vegetation Science, Vol.
7, pp. 585-592.
6.Pillar, V.D.P., 2000, "MULTIV. Software para Anhlise Multivariada e Testes de
Aleatorizag]o," Departamento de Botfinica da Universidade Federal do Rio
Grande do Sul, Porto Alegre, Brasil.
7.Rodrigues, M.L.K., 1997, "Diagn6stico da Poluig~o por Elementos-Tra~o no
Sedimento da Bacia Hidrogr~ifiea do Rio Cai, RS," dissertation, Instituto de
Bioci6ncias da Universidade Federal do Rio Grande do Sul, Porto Alegre, Brasil,
124 p.
Elba C. Teixeira, i Maria Lucia K. Rodrigues, 2 Marta F. C. Alves, ~and Jane R.

Barbosa 2
Study of the Geochemical Distribution of Heavy Metals in Sediments in Areas

Impacted by Coal Mining
Reference: Teixeira, E. C., Rodrigues, M. L. K., Alves, M. F. C., and Barbosa, J. R.,
"Study of the Geochemical Distribution of Heavy Metals in Sediments in Areas
Impacted by Coal Mining," Contaminated Sediments: Characterization, Evaluation,

Locat, R. Galvez-Cloutier, R. C. Chancy, and K. R. Demars, Eds., ASTM International,
Abstract: This study evaluated the geochemical distribution and potential mobility of heavy
metals in sediments impacted by coal extraction in the Candiota Region of Brazil. Sampling was
performed in Candiota stream and tributaries, at eleven sites, in February and June 2000. The
silt/clay fraction of sediments was analyzed by total and sequential extraction schemes
established by the EPA and European Commission, respectively. The results indicated higher
mobility for manganese, intermediary for nickel, zinc, copper, and lower for iron, chromium,
and aluminum. The mouth of Poach Stream, most affected by mining activities, presented an
increase of iron in the oxidizable fraction, as well as zinc, nickel, and copper, suggesting
association with sulfides and presence of pyrite in the sediments. The same behavior was
observed in the Candiota Stream, after the entrance of the Poac~. Due to a lower pH (4.3) caused
by mining acid drainage, at the mouth of Poacfi Stream, manganese content decreased in
sediments and increased ten times in the water column.
Keywords: sequential extraction, geochemical distribution, sediments, heavy metals,
coal
' Research Scientist and Chemist, respectively, Programa de Pesquisas, FEPAM, Rua
Carlos Chagas, 55, 90030-020 Porto Alegre, RS, Brazil.
2 Research Scientist and Chemist, respectively, Divisgo de Quimica, FEPAM, Rua
Aurflio Porto 45, 90620-010 Porto Alegre, RS, Brazil.
72
www.astm.org
TEIXEIRA ET AL. ON AREAS IMPACTED BY COAL MINING
Introduction
The load of heavy metals discharged into aquatic systems is probably reflected by
their accumulation in the sediments. These pollutants are preferentially adsorbed on the
finest particles and in different phases of the sediments, becoming available to the biota
through physicochemical changes of the system. Therefore, it is important to estimate
which fraction of the metallic elements could be potentially released into the water
column or directly transferred to the biota, leading to the knowledge of the different
types and degrees of pollution in the sediments.
Among the techniques used to identify anthropogenic contributions, several authors
have mentioned the sequential extraction as an appropriate method to determine the
chemical distribution of heavy metals in contaminated sediments (Kersten and Forstner
1995). Many of the proposed sequential schemes were derived from the work of Tessier
et al. (1979). However, the majority of the employed techniques presented significant
differences, impairing the comparison of the results and the evaluation of heavy metals
mobility. Although the recommended extractants were selective, few of them were
specific enough to isolate well defined fractions. Aiming at the standardisation of
procedures, the European Community (Ure et al. 1993, Fiedler et al. 1994, Quevauviller
et al. 1994) proposed a Certified Reference Material (CRM) - BCR 601 for sequential
schemes, involving a three-step extraction (Lopez-Sanchez et al. 1998).
Concerning the geochemical distribution of heavy metals in aquatic sediments, in Rio
Grande Sul, Brazil, some studies are available, which applied sequential extraction
procedures in regions contaminated by coal mining (Teixeira et al. 1997, Teixeira et al.
1994). In the southwest of the State, in the Candiota Region, is located the biggest
Brazilian coal reserve (Great Candiota Deposit), estimated 12 billion tons and
accounting for 38% of the coal reserves of Brazil (Neves and Chaves 2000). The total
production of coal in Rio Grande do SUl State is approximately 3 000 000 t/year,
representing about 60% of the nation's overall production, which is mainly used to
generate electrical power. The main economic activities in this region are: coal and
limestone mining, coal power plant, cement industries, agriculture and cattle-raising.
Such activities are favored by the easy access to the site, the topography, geology and
availability of water resources. All these activities are located inside the limits of the
Candiota Stream Basin, approximately 1430 kmL Regarding the anthropogenic input,
the most affected tributaries of Candiota Stream are the small-sized watercourses Poacfi
and Candiotinha.
The purpose of this work was to study the geochemical distribution of the metallic
elements copper (Cu), nickel (Ni), zinc (Zn), chromium (Cr), aluminum (A1), iron (Fe),
and manganese (Mn) in the stream sediments impacted by coal mining along the
Candiota Stream Basin, Rio Grande do Sul State, Brazil, in order to assess their
potential mobility and bioavailability in the aquatic environment.
73
74
Study Area
The Candiota Stream Basin (Figure 1) is included in the basin of the upper course of
Jaguar~o River, the main tributary of the Mirim Lagoon. This river delimits the border
between Brazil and Uruguay in the Southeast of Rio Grande do Sul State.
The climate of this region is classified as Cfa (subtropical climate, according to Von
Koepen classification system), with well-distributed precipitation during all the year.
The minimum temperatures occur in winter (June and July) and the maximum ones in
summer (January and February), and the annual average temperature is approximately
18~ (DNM 1992).
The geological structure of the Candiota Region is formed mainly by the rocks of the
Paran~i Basin, which are characterized regionally as an extensive deposition depression
located in the Center-East part of the South American Continent (Schneider et al. 1974).
The Pre-Cambrian basement of Candiota Region is represented, in the Northeast part of
the area, by metamorphic rocks of the Porongos Group and early Palaeozoic
sedimentary units of the Bom Jardim Group (Arroio dos Nobres Formation) and
Camaqu~ (Guaritas Formation). The crystalline basement occurs in the far Northwest
and Southeast parts of the region and is represented, respectively, by gneiss of the
Cambai Complex and granite rocks of the associations CGgPM (Pirtheiro Machado
Granite-Gnaiss Complex) and CGDF (Dom Feliciano Granite Complex). The other
litho-stratigraphic units observed at Candiota Region belong to the Permian age
sedimentary packages of the Paran~i Basin.
The lithology of the region is constituted by fine whitish-gray sandstone, locally
coarse, intercalated with dark-gray silt, carbon silt and coal layers.
Experimental Method
Sediments samples were collected at eleven sites along the Candiota Stream Basin, in
February and June 2000, taking into consideration the proximity of potential pollutant
sources and the access to sampling spots (Figure 2). Three sites were considered as
backgrounds, due to the absence of any human activity. According to their geological
diversity they were classified as follows: CA01 - background site of Candiota Stream,
with soils formed over the arenite rocks of the top of Rio Bonito Formation; PO01 - site
located over sedimentary rocks of Rio Bonito Formation, which is under the natural
influence of the coal layers that emerge along the fiver bed in the headwaters of Poac~i
Stream; and CC01 - site situated over schists intercalated with marbles of Porongos
Group.
Corers were used in the sample collection, avoiding disturbance of fine sediments. In
order to obtain representative samples and enough material for analysis, at each
sampling station, at least four plastic pipes were employed. In this study, only the first 5cm layer was taken for the sample composition, which corresponded to a recent
deposition. The samples were transported to the laboratory, under refrigeration at 4~
and wet-screened through a 63 pm nylon mesh, to obtain the silt/clay fraction
(Salomons and F6rstner 1980). The fraction < 63 rtrn was oven dried at a temperature <
40~ and disaggregated in an agate mortar for chemical analyses.
Figure 1 -Location Map of Candiota Region
75
76
Figure 2 -
Schematic Drawing of Candiota Stream Basin, RS
TEIXEIRA ET AL. ON AREAS IMPACTEDBY COAL MINING
The sequential extraction scheme, applied to the samples and simultaneously to the
CRM BCR-601 comprised three phases: carbonatic (first step), reducible (second step)
and oxidizable (third step) (Lopez-Sanchez et al. 1998). In the present study, a fourth
step was added - residual fraction, to obtain a mass balance for each studied element.
Since the BCR-601 certification was given only for certain metals (Ni and Cd for the
three phases, Cr for the first phase, and Zn for the first two phases), it was necessary to
determine the analytical accuracy for the other investigated elements (Mn, Cu, Fe, Cr,
and At) by comparing the sum of concentrations obtained in the four steps of the
sequential extraction to the total metallic content.
The total metals and the fourth phase extractions were conducted in accordance with
USEPA analytical method for Microwave Assisted Acid Digestion of Siliceous and
Organically Based Matrices (EPA 3052), with addition of fluoric, nitric and
hydrochloric acids. Boric acid was employed to eliminate fluoric acid residues. The
metal determinations in all the extracts was accomplished by inductively coupled
plasma optical emission spectrometry (ICP/OES). In the quality control of total content
determinations, the CRM CANMET STSD-3 was also analyzed. Total sulfur (St)
determination followed the procedures of Test Method for Sulfur in the Analysis Sample
of Coal and Coke Using High Temperature Tube Furnace Combustion Methods (D4239), and the analysis of organic carbon (OC) content was performed through the
method proposed by Walkley-Black (1965).
Grade p.a. reagents that had previously been tested with respect to metal content and
ultra-purified water were used to process the sample duplicates and CRM triplicates.
Statistical Analysis of the Data Base

Initially, for each metallic element, linear regression and Student's t test were applied
to verify the existence of significant differences between the sampling dates. In the latter
test, the concentration differences between dates were previously checked as to normal
distribution. Considering a 5% confidence level, the results showed similar contents for
the investigated elements in the two sampling dates, except for Mn, that did not show a
recognizable pattern of variability. Meanwhile, concerning the whole data set, the dates
did not differ significantly, allowing the presentation of the average values calculated
for February and June 2000.
In an attempt to synthesize all the information obtained in the sequential extraction
scheme and to recognize patterns of distribution for the metallic elements in the
different phases, a metric ordination of the sampling units (spatial and temporal
measures resulting from data collection), by principal components analysis, was
additionally performed. This multivariate statistical procedure was carried out with the
software Statisticafor Windows 4.3TM, and also included the variables St and OC.
Certified Reference Materials

Table 1 displays the results obtained for the sequential extraction scheme and for the
total extraction applied to the CRM BCR-601, as well as the relative recovery between
the two procedures.
77
78
For the elements that presented certification, the accuracy of the sequential extraction
scheme was satisfactory, except for the element Ni in the reducible fraction, where an
overestimation (approx. 80%) has been verified. This fact is probably explained by a
partial attack of oxidizable and/or residual phases. Further significant differences were
observed in the second phase especially for Cu and Cr concentrations (Table 1), when
compared with corresponding indicative values mentioned by Rauret et al. (1999) for
the BCR-601. A considerable error for Ni in the reducible fraction was also reported by
the European Commission (Ure et al. 1999). In order to reduce uncertainties, LopezSanchez et al. (1998) recommended modifications of the BCR sequential extraction
scheme, among them the increase in the extractant concentration and a better control of
the final pH in the second phase.
Table 1 - Sequential Extraction Applied to the Certified Reference Material (CRM) -
BCR 601
Ni
lag/g
Ni
~tg/g
Zn
bt~g
Zn
pg/g
Cr
/,tg/g
Cr
p,g/g
Cu
lag/g
Mn
mg/g
Fe
mg/g
AI
mg/g
First step
Second step
Third step
7.44
10.9
6.96
8.011
6.051
8.551
215
202
120
2641
1821
1372
0.3
7.5
24.2
0.361
1.432
18 .32
11.4
47.2
99.6
0.255
0.161
0.064
0.041
3.05
2.92
0.093
2.95
1.58
Four step
T
57.8
83.1
-----
195
733
-----
102
134
-----
56.1
215
0.644
1.12
41.6
47.6
59.7
64.3
Total extraction
71.1
--785
Recovery, %
116
--93.3
BCR-601 certified concentration
2 BCR-601 indicative value
-----
139
96.4
-----
210
101
1.13
99.8
49.9
95.3
66.4
96.8
Concerning the accuracy of total metals determination, evaluated in this work by the
analysis of the CRM CANMET STSD-3, the mean recovery for the studied elements
was between 95 and I00%, except for Cr (78%).
The reproducibility of the analyses was largely satisfactory, with a coefficient of
variation lower than 2% for all the determinations, including samples.
Total and Sequential Extraction of the Samples

As mentioned before, with the application of linear regression and Student's t test, no
significant difference was observed, at a 5% confidence level, between the data sets
obtained for February and June 2000. Among the metallic elements, Mn was an
exception, but no distribution pattern was identified. Therefore, in an attempt to
facilitate the data interpretation and the disclosure of results, the average geochemical
distribution of the studied elements was determined. The same procedure was adopted
for total contents of metals, sulfur and organic carbon. The mean values are presented in
Table 2, including a mass balance for each studied element, which evaluated the
efficiency of the extractions by comparing the sum of the four phases and the
concentration obtained in the total extraction.
Although total extraction enables the quantification of metallic elements in
sediments, it didn't allow a quantitative assessment of the anthropogenic contribution
and the potential risk of an element release. Generally, the results of total metal contents
TEIXEIRA ET AL. ON AREAS IMPACTEDBY COAL MINING
in the sediments of Candiota, Candiotinha and Poac~ streams, when compared to the
corresponding background sites, did not show a recognisable trend. On the other hand,
the sequential extraction scheme results pointed out the phases in which the elements
occurred in the sediments, thus allowing the evaluation of their mobility and availability
to the aquatic environment.
The results of geochemical distribution obtained with the sequential extraction
scheme are illustrated in Figures 3 to 5. The potential mobility of the metallic elements
in the Candiota region, indicated by the sum of the first three extraction phases, was
higher for Mn. For Fe, Cr and AI, between 70 and 90% of the total concentration was
found in the residual fraction, indicating a lower mobility for these elements.
Cu presented a higher percentage of extraction in the oxidizable fraction of Candiota
stream (CA03 and CA06) and Poac~i Stream (PO02), in relation to the points considered
as backgrounds (CA01 and PO01, respectively). These results, especially for PO02, can
be related to the contents of St and OC (Table 2), indicating some contamination of
sediments by coal. The preferential association of Cu with the oxidizable fraction has
been verified in some works, that pointed out this metal correlation to organic matter
(Salomons, 1995) and to sulfides in sediments (Teixeira et al. 2001, Teixeira et al.
1997).
Although more than 70% of Fe concentration was found in the residual fraction, this
element also appeared associated to the oxidizable fraction in sites CA06 and PO02. The
presence of Fe in the oxidizable fraction (PO02), associated with the higher S content in
the sediment, also indicates the presence of sulfide in the sediments.
Zn showed association with the carbonatic fraction for site CA04 (30%), in response
to the sediments contamination by waste coal combustion (Figure 2). This association
could be attributed to the substitution of calcium (Ca) by Zn in calcite (Davidson and
Clarke 1996). Otherwise, in sites CA06 and PO02, Zn presented preferential association
to the oxidizable fi'action, with percentage contents varying between 25 and 31%, what
could be attributed to an association with sulfides, probably spharelite (Finkelman 1995,
Davidson and Clarke 1996). Accordingly, in this study, the highest concentration of St
was found at the sampling station PO02. The presence of Zn in the sediments was more
evident at sites contaminated by coal (Figure 3). The contribution of Zn to CA06
sampling point showed influence of coal mining drainage from Seival Mine, CRMMesh IV, and inputs from the basin of the Poac~ Stream.
As to the Ni results, a high percentage in the oxidizable fraction was found at the
sites CA04, CA06 and PO02 (between 37 and 49%), in response to local anthropogenic
sources (coal mining and ashes). The association of this metal to sulfides in sediments
and coal has been reported in several works (Teixeira et al. 1994, Pires et al. 1997,
Finkelman 1995).
79
80
Table
Cu
p_g/g
Zn
ttg/g
Ni
Itg/g
Cr
I~g/g
Mn
mg/g
Fe
mg/g
AI
mg/g
St%
0(2 %
2 -
Sequential and Total Extraction Applied to Sediment Samplesfrom the

Candiota Region, RS, Brazil
First step
Second step
Third step
Four step
TE
Total exa'.
Recov., %
First step
Second step
Third step
Four step
I;
Total extr.
Recov., %
First step
Second step
Third step
Four step
,r
Total extr.
Recov., %
First step
Second step
Third step
Four step
~Total extr.
Recov., %
First step
Second step
Third step
Four step
t;
Total extr.
Recov., %
First step
Second step
Third step
Four step
r
Total extr.
Recov., %
First step
Second step
Third step
Four step
,r
Total r
Recov., %
CA01
0.25
2.15
1.95
9.50
13.8
14.0
98.6
10.8
12.3
8.55
42.3
73.8
79.6
93
2.15
2.35
3.30
7.67
15.5
13.7
113
0.049
0.496
4.64
29.9
35.0
38.3
91.5
0.761
0.357
0.110
0.105
1.33
1.26
106
0.077
6.32
2.52
24.9
33.8
31.5
107
0.062
0.888
2.01
76.5
79.5
88.1
90
0.08
3.39
CA02
0.60
2.25
1.80
7.75
12.4
11.4
109
10.0
10.4
6.8
31.0
58.1
56.9
102
1.93
1.90
3.40
5.61
12.8
15.0
86
0.035
0.330
3.24
30.7
34.3
32. I
107
0.539
0.200
0.035
0.055
0.828
0.750
I10
0.188
4.90
1.44
22.2
28.7
28.5
100
0.062
0.805
1.69
76.0
78.6
81.9
96
0.06
3.06
CA03
0.85
2.83
4.05
14.9
22.6
21.2
107
19.5
18.5
11.7
59.6
109
100
109
1.59
1.80
3.02
I 1.3
17.7
19.4
91
0.075
0.836
7.41
42.5
50.8
46.7
109
0.220
0.102
0.263
0.118
0.466
0.416
112
0.321
5.26
2.67
23.7
31.9
36.6
87
0.048
1.01
2.00
73.9
76.9
81.9
94
0.11
3.7
CA04
0.20
2.10
3.05
13.8
19.1
18.6
103
54.8
18.0
13.4
56.4
143
163
87
3.10
3.45
4.70
I 1.8
23.0
19.2
120
0.106
0.604
5.93
36.4
43.0
42.g
101
0.375
0.136
0.906
0.136
0.738
0.745
99
0.070
2.62
1.36
30.0
34.0
35.5
96
0.186
1.23
1.91
6.82
71.6
84.8
84
0.22
3.04
CA05
0.45
2.35
1.55
11.4
15.7
13.8
114
7.36
9.65
7.85
45.7
70.5
64.5
109
1.65
2.30
4.06
5.65
13.7
15.3
90
0.041
0.313
4.59
38.7
43.6
37.3
117
0.348
0.192
0.40
0.089
0.668
0.655
102
0.159
2..99
1.34
21.1
25.6
27.1
94
0.035
0.577
1.98
80.2
82.8
79.2
105
0.08
3.71
CA06
0.30
0.173
3.20
13.4
17.0
16.0
106
11.1
11.5
19.9
36.5
78.9
75.1
105
3.00
2.60
10.1
6.30
22.0
22.5
98
0.065
0.098
6.95
36.5
43.6
39.2
III
0.368
0.097
0.061
0.054
0.580
0.577
101
0.122
2.47
5.08
19.7
27.3
28.7
96
0.194
0.758
1.39
75.7
78.0
83.7
93
0.15
2.76
CA07
0.29
1.55
0.81
12.3
15.0
16.2
92.6
5.70
11.0
3.80
61.8
82.3
80.8
102
1.40
1.95
1.95
12.3
17.6
17.9
99
0.055
0.200
2.35
36.2
38.8
38.4
101
0.362
0.323
0.063
0.186
0.934
0.981
95
0.039
2.08
1.08
29.9
33.1
32.9
1 O0
0.044
0.875
2.62
67.0
70.6
71.8
98
0.07
1.79
CC01
0.50
2.38
2.15
11.4
16.4
17.4
94.2
5.80
9.50
6.50
47.8
69.6
73.3
95
5.05
3.91
6.60
42.4
58.0
52.8
110
0.150
0.780
16.5
76.2
93.5
81.2
115
0.540
0.195
0.040
0.111
0.884
0.930
95
0.149
4.43
2.07
18.2
24.8
26.5
94
0.038
0.552
2.13
55.9
58.7
59.3
99
0.06
3.34
CC02
0.25
2.10
1.53
7.92
11.8
12.6
93.6
2.95
9.20
6.4
57.9
76.4
72.5
105
0.700
1.55
1.94
9.20
13.4
11.8
113
0.048
0.413
4.78
33.8
39.0
34.2
114
0.701
0.372
0.072
0.162
1.31
1.34
98
0.736
6.16
1.70
33.6
41.6
38.7
107
0.038
1.22
2.08
78.6
81.9
67.6
121
0.05
2.00
PO01
0.45
2.05
2.48
8.75
13.7
13.9
98.6
7.33
4.70
5.36
34.5
51.9
56.4
92
1.79
1.40
1.47
6.70
11.4
10.8
105
0.070
1.30
4.19
37.7
43.3
35.9
121
0.031
0.099
0.038
0.083
0.535
0.533
100
0.122
6.20
1.71
33.0
41.0
31.7
129
0.052
0.644
1.34
78.3
80.3
57.3
140
0.10
2.44
PO02
0.08
1.45
3.33
9.16
14.0
14.3
98.0
2.92
1.20
8.56
16.4
29.1
33.2
88
1.39
1.10
2.91
3.02
8.4
7.2
117
0.046
1.00
5.56
32.0
38.6
43.4
89
0.029
0.012
0.013
0.027
0.082
0.093
88
0.040
1.41
5.37
17.0
23.8
28.2
84
0.091
0.671
0.151
69.6
71.8
82.2
87
0.37
7.02
Figure 3 - Geochemical Distribution of Copper and Zinc in Sediments from Candiota

Region, RS, Brazil
High percentages of Al and Cr were found in the residual fraction of the sediment
samples (Figures 4 and 5). In other sequential extraction studies, the labile Cr has been
shown to be associated to clay minerals (Teixeira et al 2001, Teixeira et al 1997).
The increased content of Mn in the available fractions and specially in the carbonatic
phase (Figure 4), confirmed its high potential mobility, as mentioned above. The
geochemical distribution of this element, concentrated on the carbonatic phase, followed
by the reducible one, has been verified in other studies where the sequential extractions
in sediments was applied (Teixeira et al. 1997, Kersten and Frrstner 1995).
However in the Poac~ Stream, in comparison to the site PO01, the sampling station
PO02 showed a decrease in Mn mobility and, consequently, an increase in the residual
fraction. This fact was due to a lower pH (4.3) caused by mining acid drainage that
brought about an increase of the metal availability (Streck 2001) and a reduction in its
mobile fraction in sediments.
81
82
Figure 4 -
Geochemical Distribution of Nickel, Chromium and Manganese in Sediments

from Candiota Region, RS, Brazil
Figure 5 - Geochemical Distribution of Iron and Aluminum in Sedimentsfrom Candiota
Region, RS, Brazil
Principal Components Analysis

Following this method, the data base was synthesized by a set of orthogonal axes.
The first, considered the best summary of the relationship exhibited by the data, and
accounting for 21.9% of the variance, was retained for interpretation. Attempting to a
better visualization of tendencies, the second axis (18.9% of the data variability) was
also used to prepare a diagram (Figure 6), representing the dispersion of sampling units.
In the first component, the most relevant variables, with positive correlations were:
Fe (0.73), Cu (0.67), Ni (0.57) and Zn (0.57), all in the oxidizable phase; St (0.68); and
OC (0.67). The good correlation between these parameters in the oxidizable fraction
suggested the presence of pyrite in the sediments collected in sites CA06 and, mainly,
PO02. The results were in agreement with those presented above about the geochemical
83
84
distribution, that indicated contamination in these sediments by coal and ashes,

originating from coal mining.
oCC01J
2
1
CA07J CAOTFo
CC01F
CA04J 9
CAO3F
/9
-2
ComponentI
Figure 6 - Dispersion of the Sampling Units, According to Ordination Scores in
Principal Components I and 2 (F=February, d=dune)
Conclusions
The results obtained with total metal extractions in sediment samples from the
Candiota Region didn't allow a quantitative assessment of the anthropogenic
contribution. On the other hand, the application of sequential extraction displayed the
phases in which the elements occurred, allowing the prediction of their potential
mobility.
The results indicated higher potential of release from sediments for manganese,
intermediary for nickel, zinc, copper, and lower for iron, chromium, and aluminum.
Metal geochemical distributions showed that sediment contamination occurred
particularly in sites CA06 and PO02, which represented the area strongly affected by coal
mining activities. This contamination was demonstrated by the higher percentages of Fe,
Cu, Zn, and Ni in the oxidizable phase, and by the significant sulfur content in
sediments, notably in PO02. These results revealed the sulfide contamination of
sediments produced by coal mining.
In the Poac~i Stream sediments, due to a lower water pH (4.3) caused by mining acid
drainage, the site PO02 presented a decrease in Mn concentration in the three mobile
phases and an increase in the residual fraction content, in comparison to the control
station PO01. At the same time, an increase of the Mn availability was observed in the
water column, where the concentration was ten times higher in comparison with the
control site. Eventhough, the Mn potential mobility at the site PO02, represented by the
sum of the three mobile phases, was still superior to the other metals.
Applying the metric ordination by principal components analysis, it was possible to
identify associations between variables in the different phases of the sequential
extraction scheme. Additionally, it allowed evaluation of the degree of relevance of each
variable in the diagnosis of the area, and detection of trends over space or time. The first
component summarized the most relevant information, showing the good correlation
between St, OC and the metals Fe, Cu, Ni, and Zn present in the third phase. The
association between these variables in the oxidizable fraction confirmed the presence of
pyrite in the sediments at site CA06 and especially PO02. The results were in agreement
with those presented above about the geochemical distribution that indicated
contamination in these sediments by coal and ashes originating from coal processing
activities. As concluded by linear regression and Student's t test application, no relevant
temporal trend was detected for the whole data set with this statistical tool.
Aknowledgements
Acknowledgements to PADCT/CIAMB, FAPERGS, and CNPq for the financial
support; and to Cledion Peixoto - technician of FEPAM, for the accomplishment of the
sanlpling.
References
ASTM, 1995, "Standard Test Methods for Sulfur in the Analysis Sample of Coal and
Coke Using High Temperature Tube Furnace Combustion Methods," ASTM-D
4239-94. 5 (5): 363-375.
Davidson, R.M., Clarke, L.B., 1996, "Trace Elements in Coal," United Kingdom, IEA
Coal Research, London, IEAPER/21.
DNM, Departamento Nacional de Meteorologia (Brasil), 1992, "Normas Climatolrgicas
(1961 - 1990)," Brasilia.
EPA, 1996, "Microwave Assisted Acid Digestion of Siliceous and Organically Based
Matrices," Method 3052, pp 1-20.
Fiedler, H.D., L6pez-Sfinchez, J.F., Rubio, R., Rauret, G., Quevauviller, Ph.,Ure, A.M.
and Muntau, H., 1994, "Study of the Stability of Extractable Trace Metal
Contents in a River Sediment Using Sequential Extraction,"Analyst, Vol. 119,
pp. 1109-1114.
Finkelman, R.B., 1995, "Modes of Occurrence of Environmentally-Sensitive Trace
Elements in Coal." ln: Swaine, D.J. and Goodarzi, F., Environmental Aspects of
Trace Elements in Coal, Kluwer Academic Publishers, Netherlands, pp. 24-50.
Kersten, M., FSrstner, U., 1995, "Speciation of Trace Metals in Sediments and
Combustion Waste," In: A.M. Ure and C.M. Davidson, Chemical Speciation in
the Environment, pp. 234-275.
Lopez-Sanchez, J. F., Sahuquillo, A., Fiedler, H.D., Rubio, R., Rauret, G., Manteau, H.
and Quevauviller, P., 1998, "CRM 601, a Stable Material for its Extractable
Content of Heavy Metals," Analyst, Vol. 123, pp. 1675-1677.
Neves, G. R., Chaves, S. H. A., 2000, "A Regi~o Carbonifera "Tradicional" do Rio
Grande do Sul," Carv~o e Meio Ambiente, Centro de Ecologia/UFRGS, Porto
Alegre, RS, pp. 108-129.
85
86
Pires, M., Fiedler, H., Teixeira, E.C., 1997, "Geochemical Distribution of Trace
Elements in Coal: Modelling and Environmental Aspects," Fuel, Vol. 76, No.
14/15, pp. 1425-1437.
Quevauviller, Ph., Rauret, G., Muntau, H., Ure, A. M., Rubio, R., Lopez-Sanchez, J.F.,
Fiedler, H.D., Griepink, B., 1994, "Evaluation of a Sequential Extraction
Procedure for the Determination of Extractable Trace Metal Contents in
Sediments," Fresenius Journal of Analytical Chemistry, Vol. 349, pp. 808-814.
Salomons, W., 1995, "Environmental Impact of Metals Derived from Mining Activities:
Processes, Predictions, Prevention," Journal of Geochemical Exploration, Vol.
52, pp. 5-23.
Salomons, W., F0rstner, U., 1980, "Trace Metal Analysis on Polluted Sediments, Part I:
Assessment of Sources and Intensities," Environmental Technology Letters,
Vol. 1, pp. 494-505.
Schneider, R. L., M0.hlmann, H., Tommasi, E., Medeiros, R. A., Daemon, R. F. and
Nogueira, A. A., 1974. Proceeding ... XXVIII Congresso Brasileiro de Geologia,
1974, Porto Alegre, pp. 41-67.
Streck, C., 2001, "Estudo da Qualidade das ,~guas Superflciais e Estimativa dos Huxos
de Elementos-Traqo na Regi~o de Candiota-RS," M.Sc. Dissertation, PPGEE/
PUCRS, 159 p.
Teixeira, E.C., Ortiz, L.S., Sanchez, J.C.D., Alves, M.F.C.C., 2001, "Distribution of
Selected Heavy Metals in Fluvial Sediments of the Coal Mining Region of
Baixo Jacui, RS, Brasil," Environmental Geology, Vol. 41, pp. 128-144.
Teixeira, E.C., Sanchez, J.D., Femandes, I., Formoso, M.L., Pegorini, J., Pestana,
M.H.D., 1997, "A Preliminary Study of Metals in Sediments from Areas
Influenced by Coal Processing and Steel Industry Activities in Baixo Jacui
Region, RS-Brazil," Environmental Technology, Vol. 18, pp. 581-592.
Teixeira, E.C., Pestana, M.H.D., Sanchez, J.D., Fernandes, I., 1994, "Geochemical
Distribution of Metallic Elements in the Mineral Matter of Brazilian Coals,"
Environmental Technology, No. 15, pp. 98%996.
Tessier, A., Campbell, P.G.C., Bisson, M., 1979, "Sequential Extraction Procedure for
the Speciation of Particulate Trace Metals," Analytical Chemistry, Vol. 51, No. 7,
pp. 844-851.
Ure, A., Quevauviller, Ph., Muntan, H., Griepink, B., 1993, "Improvements in the
Determination of Extractable Contents of Trace Metals in Soil and Sediment
Prior to Certification, BCR Information-Chemical Analysis," BrusselsLuxembourg: ECSC-EEC-EAEC, 85 p.
Walkley-Black, C.A, 1965, "Chemical and Microbiological Properties," Methods of Soil
Analysis, Madison, Wisc : Soil Science Society of America, Vol. 2, pp. 1372-1378
(Agronomy, n.9).
Hrlrne Tremblay, l Gaston Desrosiers, 2 Jacques Locat, 3 Alfonso Mucci, 4 and l~milien
PeUetier2
Characterization of a Catastrophic Flood Sediment Layer: Geological,

Geoteehnieai, Biological, and Geochemical Signatures
Reference: Tremblay, H., Desrosiers, G., Locat, J., Mucci, A., and Pelletier, I~.
"Characterization of a Catastrophic Flood Sediment Layer: Geological,
Geotechnical, Biological, and Geochemical Signatures," Contaminated Sediments:
Performance, ASTMSTP 1442, J. Locat, R. Galvez-Cloutier, R. C. Chancy, and K. R.
Abstract: In i 996, an important rainstorm took place in the Saguenay region, Canada,
that caused severe flooding and erosion of a large amount of post-glacial sedimentary
material. Consequently, about 20 million tons of sediment were deposited in the upstream
part of the Saguenay Fjord, covering the recent, contaminated sediments with a l 0 to 60
cm thick layer of relatively clean material. This flood layer was distinguished from the
pre-flood sediments by various properties: low consistency and low resistance, high water
content, the absence of benthic organisms, or the presence of inherited geochemical
components. In this study, we evaluated the criteria used by the various investigators to
identify the 1996 flood deposit, then finally compared its signature. Subsequently, we
assessed the spatial variability of the deposit and its effect on the interpretation of
temporal studies.
Keywords: sediments, flood layer, geological signature, geotechnical signature,
biological signature, geochemical signature, spatial variability
IPostdoctorate Fellow, Dept. of Geology and Engineering Geology, Laval University,
Quebec, Quebec, Canada, G 1K 7P4.
2professor, Institut des Sciences de la Mer de Rimouski (ISMER), 310 allre des
Ursulines, Rimouski, Quebec, Canada, G5L 3AI.
3professor, Department of Geology and Engineering Geology, Laval University, Quebec,
Quebec, Canada G 1K 7P4.
4professor, Department of Earth and Planetary Sciences, McGill University, 3450
University St., Montreal, Quebec, Canada, H3A 2A7.
87
www.astm.org
88
Introduction
Background on Signature of Catastrophic Sediment Layer
Sediments can record different events through time, such as earthquakes, tsunamis,
floods, forest fires, and sea-level fluctuations. These events are often used as absolute
markers when the event date is known. The signature of these events is also used to
determine the extent of the affected area. The main problem is to be able to identify and
recognize the layer corresponding to one particular event.
Studies on sediments from a recent flood event (in 1995) have been conducted on
the Northern California shelf, U.S. (Drake 1999, Leithold and Hope 1999). The authors
indicate that the flood layer is easily recognized by its lighter color, higher water content,
the high content of fine particles, a sharp basal contact and common occurrences of
"fresh" terrigenous plant material.
In the Saguenay Fjord, Canada, many large slides have occurred on land and under
water. These disasters were identified in cores and distinguished from the "normal"
sediments by a lighter color, the absence of bioturbation, a higher water content, a
coarsening with depth within the layer, and a higher inorganic carbon content (Perret et
al. 1995, St-Onge et al. 1999, St-Onge and Hillaire-Marcel 2000). In their study, Schafer
et al. (1983) identified different events in the sedimentary history of the Saguenay Fjord
such as floods and landslides. The landslides, especially the 1971 St-Jean-Vianney
landslide, which occurred in the marine clay from the Laflamme post-glacial sea, are
distinguished from recent sediments by the presence of marine clay pellets. St-Onge and
Hillaire-Marcel (2000) identified up to 10 accidental sediment layers in a 37.6 m long
piston core taken in 1999 in the inner basin, including a 15m thick deposit triggered by an
estimated magnitude 7 earthquake in 1663. The other nine are older than the 1663 event.
Saguenay Fjord Context
Saguenay Fjord characteristics - The Saguenay Fjord is located in the Province of

Quebec, Canada, about 200 km northeast from Quebec City. It is a long (93 km) and
narrow (1 to 6 km) glacial valley that joins the St. Lawrence estuary (Fig. 1). It is
composed of 3 main basins from west to east, 275, 180 and 250 m deep, respectively that
are separated by shallow sills 130, 65 and 20 m deep, in the same order (Schafer et al.
1990). The geological, geotechnical, hydrographic and biological characteristics of the
fjord are described by Schafer et al. (1990), Perret et al. (1995), and Boss6 et al. (1996).
Recent sediment accumulation rates vary from 7 crrdyr, at the head of the fjord (in the
Bras Nord) to less than 0.1 cm/yr in the deepest basin (Smith and Walton 1980).
TREMBLAY ET AL, ON CATASTROPHIC FLOOD SEDIMENT LAYER
1'
7005n,
_~.~_..
- ~ ~.~_f,- / ~-~.~
4e~2~,oOo~-_
- -
/.
'-znO~n,
Saguenay Fjord-
~stations Vl, V2
Riv. des
. ../~
,,r~/~ i
%~,~
'/~ Atlantic
Oceanj
r~, ~ 1_~7"~,
,-~
200k~
~~ . . . .
~-~
_.
\-"f-"-h
b Km
i.
[ QUI~BEC
r~
station
~
-~-~---j.~-J~
I 48020'/'~. -f
7nOAn,
\ ~
I
Ba'edes
Ha. I Ha ! ~
|Ville de HaIHat
La Baie
70o50 ,
89
i 70040,
70030,
~" I
to Saint Lawrence R i v e r ' ~

.
Figure 1 - Location map.

The 1 9 9 6 f l o o d layer - The most recent catastrophic input of sediment in the
Saguenay Fjord resulted from the 1996 flood. From July 18th to 20th, 1996, exceptional
rainfalls occurred in the upper part of the fjord, mainly in the Bras Nord and the Baie des
Ha! Ha! (Fig. 1) (Yu et al. 1997). This extreme precipitation event caused flooding of the
main tributary rivers, where about 15 to 20 million tons of post-glacial sediments were
eroded and transported down to the fjord. The sediments were deposited as deltaic
deposits close to river mouths, and, in much of the covered area, as turbidites, i.e with a
sharp basal contact and normally graded beds. This catastrophic sediment deposit
covered the pre-existing contaminated sediments in this part of the fjord, with a clean
layer of sediment of about 10 to 60 cm thick in most of the covered area. The newly
deposited sediments act as a natural capping layer.
The 1996 event has created a major perturbation on the seafloor of the fjord. During
the first investigation, barely 3 weeks after the flood, the benthic community had almost
been eliminated. The biodiversity decreased from 40 to 4 species and the density of the
most abundant species, nematodes, was reduced by ten-fold (Pelletier et al. 1999). The
new deposit has also modified the geochemical steady state at the sediment-water
interface, causing the upward migration of the redox limit and the remobilization of some
metals such as iron and manganese (Mucci et al. 2000a).
This catastrophic flood event has been studied by many researchers who have
worked in the Saguenay Fjord environment in the past two decades. Their interests are
varied, and combine in one project different disciplines such as, geology, geotechnics,
geochemistry, biology and hydraulics. Although they work on different aspects of the
project, they have two common needs: the identification of the flood layer and the
evaluation of the spatial variability of its thickness for temporal evolution of given
parameters. As the investigators used different parameters to evaluate the thickness of the
1996 layer, it is of interest to compare the way each discipline defines the new layer. The
next sections will present the different signatures of the flood layer and a study on the
variability of its thickness. One station was selected for the comparative study, station 9
in the Baie des Ha! Ha!, where the 1996 layer is easily recognizable (Fig. 1), and stations
V1 and V2 for the variability study.
90
Methodology
To evaluate the extent and thickness of the flood layer as well as the evolution of its
properties with time, a large number of cores were taken using a 0.06 m 2 (22 x 30 cm)
Ocean Instruments Mark II box corer to a maximum sub-bottom depth of 60 cm. Up to 4
sub-samples were taken in each box with 10.16 cm diameter PVC tube with a beveled
end to facilitate its penetration and a fixed piston at the top to avoid sediment
compaction. Samples were extruded, on board whenever possible, or in the laboratory a
few days following the end of cruises. A visual description was noted and a variety of
properties were measured. The results were compiled as a function of depth and the
position of the interface between the pre-flood and the 1996 flood sediments was noted.
Positioning of sampling stations was provided by the ship's differential GPS and, in some
cases, by a relative positioning system, the Track Point II, that gives the relative position
of the sampler with respect to the ship.
The measured physical properties were: water content, grain size distribution,
density, magnetic susceptibility, and P-wave velocity. The water content is given as the
ratio between the wet and dry weights of sediment samples and the grain size distribution
was obtained on bulk sediments using a laser sediment analyzer. The density, magnetic
susceptibility, and P-wave velocity were determined using a GEOTEK TM multi-sensor
track instrument (MST). A Computerized Axial Tomography Scanner (CAT-scan) was
also used to evaluate the density (Holler and KOgler 1990, Boespflug et al. 1995, Cr6mer
et al. 2002).
For the purpose of this study, the following geotechnical properties were measured:
intact and remolded undrained shear strength (C, and Cur) using the Swedish fall cone,
and liquidity index as a function of both the water content and the Cur. The liquidity
index is a measure of the sediment consistency. Details of the measurement protocols are
given by Maurice et al. (2000).
A study of the distribution of the organisms within the flood layer was conducted to
characterize its biological signature. As the abundance and diversity of benthic
organisms were modified by the flood event, the trace of their activity (i.e. bioturbation)
in the surface and sub-surface sediment layer should be visible in the sediment column.
De Montety et al. (2000) evaluated the quantity of biogenic structures, i.e. burrows and
tunnels in sub-surface sediments using CAT-scan images.
The geochemical and isotopic parameters used to identify the catastrophic layer of
1996 were the organic and inorganic carbon content, and the 2~~ and 137Cs activity.
The analytical methods to measure the carbon and nitrogen contents are described by StOnge et al. (1999). Data reported herein are expressed in percent of the total dry weight
of sediments. The 21~ and 137Cs activity measurements are described by Savard (2000).
A major problem occurring in the present study, especially for temporal evolution
of certain properties is the variability of the 1996 flood layer thickness. To evaluate it, a
spatial variability study was conducted on 29 box cores taken on two square areas of 90
m x 90 m each, which were close to each other. The sites selected for this study are
located at station VI and V2 (Fig. 1). They have been selected because at these locations,
the flood layer is easily recognizable by visual observation due to a well-defined sandy
layer at its base.
TREMBLAY ET AL. ON CATASTROPHIC FLOOD SEDIMENT LAYER
91
Signature of the 1996 Layer
The mean thickness of the flood layer in the investigated area varies from 10 to 60
cm (Maurice et al. 2000). In the first year following the flood, the 1996 flood layer was
easily differentiated from pre-flood deposits by its light color, its low consistency, and
the absence of bioturbation.
Physical and Geological Signature
The 1996 flood layer is characterized by a normally graded bed with a sandy layer
at its base. The thickness of the sandy layer varies from zero to a few centimeters. In the
1996 sandy layer, 75% of particles are distributed between 36 and 120 p.m with a mean
size of 58 p.m. In the layer deposited by the flood, the mean size of particles is 21 p.m
and 75% of the particles have a diameter >16 p.m indicating a broad spectrum of
particles.
Some physical measurements on a core taken in 1997 close to station 9 are
presented in Figure 2. The lower contact between the underlying sediments and the 1996
layer is characterized by discontinuities in density, magnetic susceptibility and P-wave
velocity. The high values correspond to the sandy layer found at 20 to 27 cm depth.
Figure 2 - Profiles of some physical

properties of a sample taken near station
9 in 1997 (from Savard 2000).
Figure 3 - Tomographic intensity profile

of a core taken near station 9 in 1997
(from CrOmer et al. 2002).
92
Figure 3 shows a profile of a sample taken near station 9 in 1997 (Cr6mer et al.
2002). The tomographic intensity (TI) profile (left) reveals a discontinuity of the
sedimentation regime marked by a higher TI throughout the flood layer, and a
progressive decrease towards the surface. The discontinuity is also visible on the gray
scale image (right). On this image, we note that the flood layer was deposited in several
episodes marked by a sandy layer at the base that contains laminations typical of a
deposition by traction (Cr6mer et al. 2002) and a sharp basal contact that denote the
presence of erosion at the top of pre-flood sediments. Each episode is associated with
high TI in the sandy layer, followed by lower TI values. According to Cr6mer et al.
(2002), the high TI values in the sandy layer are related to a high carbonate content. This
inorganic carbon originates from the exposed marine clays eroded during the flood. Its
concentration in the flood layer is up to 0.6% compared to the underlying sediments
where it is about 0.08% (Mucci et al. 2000a, St-Onge et al. 1999). As carbonate is nearly
absent in the finer pre-flood sediments, their TI is lower.
Geotechnical Signature
From a geotechnical point of view, the 1996 layer is differentiated by its higher
water content, its higher liquidity index that corresponds to a very low consistency, and
by its lower undrained shear strength both in intact (Cu) and remolded state (Cur)
(Maurice et al. 2000) compared to the underlying pre-flood deposits. In general, as
shown on a typical profile of a sample taken at the station 9 (Fig. 4), the water content
decreases with depth, as does the liquidity index, and Cu and Cur, all due to self-weight
consolidation process. The interface between the flood and the pre-flood sediments is
marked by a lower water content at the interface, and/or an important increase of water
content just below the interface. The lowest water content values are associated with the
sandy layer that contains less water than finer sediments. The interface is also
identifiable by an increase of Cu and Cur below the sandy layer. Note that it is impossible
to measure this parameter in the sandy layer because the method of measurement, the
Swedish fall cone, is adapted for sediment finer than sand. This is why there is no value
between 26 and 33 cm in Figure 4. Below the interface, the Cu and Cur increase mainly
because of bioturbation. Maurice et al. (2000) noted that bioturbation influenced the
consolidation behavior of sediments. In fact, the burrowing organisms drain and mix the
sediments contributing to an increase of the shear strength by reducing the water content
as well as liquidity index. Then, where burrowing organisms are active, mainly in the
first 5 cm (Fig. 5), we might expect an increase of the water content with depth as the
organisms activity decreases (not shown). Maurice et al. (2000) also inferred that if the
thickness of the layer was greater than the limit depth of activity for the organisms, which
is 40 cm for the Saguenay Fjord, a high water content and a weak resistance would
persist below this limit.
water
content
(%)
100
liquidity
Index
0 1 2 3
200
0 ~-r-q.~ I
10
[,
C u (kPa)
e~
13 30
bioturbation
structures (%)
0 20 40 60 80
1 2 3
I'i'l'
,Ii;u
20
93
10:
7
2
E" 2O
.s::
Q..
so
40
~o
,I,
,
1
Cur (kPe)
Interface between the 1996 layer
and the pre-flood sediments
Figure 4 - Profile of geotechnical
properties for station 9 in 1998.
-,i,l,l,
60
and the background sediments
Figure 5 - Profiles of percentage of

bioturbation structures for station 9 taken
in 1999 (from de Montety et al. 2000).
Biological Signature
The large amount of sediment transported and deposited in the upstream portion of
the fjord represented a short-term disaster for the benthic communities. In September
1996, two months after the flood, Pelletier et al. (1999) reported that in the Baie des Ha!
Ha!, the meiofauna had been reduced in population by more than 97%. One of the
dominant pre-flood species (nauplii) was absent in all the stations. Two families of
meiofauna (kinorhynches and ostracodes) were found in only one station located
downstream of the area of influence. The same authors also registered a dramatic
reduction of the macrofauna. In two of the stations monitored two months after the flood
and located about 6 km west of station 9, the macrobenthic organisms were totally absent.
Even if the populations were abundant in most of the investigated stations, they were less
abundant than those reported by Boss6 et al. (1996) before the flood.
Figure 5 shows the relation between the percentage of biogenic structures and
depth for a core taken in 1999 at station 9. These were determined with a CAT-scan.
This tool makes it possible to quantify the occupation rate of organisms through sediment
profiles, and then to identify the old sediment surface that is marked with an increase of
occupation rate compared to the base of the 1996 sediment layer. The counting of trace
fossils of organisms was done on every layer of 1 to 3 cm. As shown in Fig. 5, in the first
5 cm, the biogenic structures comprise a mean of 60%. The majority of benthos found
94
between 0 and 1 mm are nematodes. Polychetes are more abundant in subsurface layers.
On this figure, it can also be observed that the biogenic structures rapidly decrease with
depth. As a result, between 8 and 30 cm, bioturbation is almost absent, but a peak
appears at 30 cm, which peak corresponds to ancient biogenic structures present at the
surface of the pre-flood sediments. It can be noted here an evidence of erosion during the
flood. In fact, the top of the pre-flood sediments contains a lower level of benthic
activities than the top of the flood deposit.
Geochemical and Radioisotopic Signatures

The pre-flood sediments of the Saguenay Fjord generally consist of gray-black
anoxic sandy muds with organic carbon concentrations (Corg) ranging from 0.5 to 3% (by
weight) although more than 5% can be found in the sandy muds of the Bras Nord (Smith
and Walton 1980). The sedimentary organic matter is dominantly of terrestrial origin and
much of the recent accumulations are from upstream pulp-and-paper and sawmill outfalls
(Pocklington and Leonard 1979, Louchouarn et al. 1997).
Dissolved oxygen is depleted within the first five millimeters below the sedimentwater interface (Mucci and Edenborn 1992, Deflandre et al. 2002). Sulfate reducing
conditions are encountered within the first two centimeters, as evidenced by porewater
sulfate depletion and/or the appearance of acid volatile sulfides (AVS) in the solid
sediments (Mucci et al. 2000b). The shallow oxygen penetration depths and the rapid
onset of sulfate reduction refect the high oxidant demand of these sediments or, in other
words, the amount and accumulation rates of degradable organic matter. Under steadystate conditions, reactive (i.e., extractable in IN HCI) Fe and Mn accumulate in the thin
layer of oxic sediments. The concentration of authigenic Fe and Mn oxides near the
sediment-water interface results from the reductive dissolution of authigenic and detrital
oxides buried below the oxygen penetration depth and the diffusion of dissolved Fe(II)
and Mn(II) to the oxic zone where they are oxidized and precipitated as oxides (Mucci et
al. 2000a).
From a geochemical point of view, the 1996 flood layer can be most easily differentiated
from the indigenous sediment by its inorganic carbon content. The flood material
contains up to 0.6% Ci,org whereas it is nearly undetectable in the normal Fjord sediments
(-0.08%) (Fig. 6). The inorganic carbon content of the pre-flood sediments is low
because any calcareous material that settles at the sediment-water interface is readily
dissolved during oxic diagenesis (Mucci et al. 2000a). In contrast, the Cinorgdelivered
with the flood material will be preserved under the anoxic conditions that rapidly
developed throughout most of the deposit (Mucci et al. 2000a). Thus, the Ci,org is likely
to represent a useful test for the flood layer identification in the future.
The Corg content of the flood deposit is about one third to half of the pre-flood
sediment and its Corg/N molar ratio (-18) is similar to that of the indigenous sediment and
also attests to its predominantly terrestrial origin. In some cases, a peak of Corg is
observed at and just below the base of the flood deposit and may be related to the rapid
burial of organisms trapped at the original sediment-water interface (Deflandre et al.
2002).
Ctot (%)
0 1 234
10
~" 20
e"
Corg (%)
0 1 234
0.00.20.40.6
2O
o
,,...
-It,,
0246810
10
!>
5o--,lll
21~ activity 1370s activity

(dpm/g)
(dpm/g)
Cinorg (%)
tD
o 30
40
95
I, I i
,I,151a

Figure 6 - Geochemical profiles of cores

taken at station 9 in 1999 in 1997 (from
Savard 2000).
Q.
O
t:)
30
40
60 7 lll,l,ll
Figure 7 - Radioisotope profiles of cores

taken near station 9 (from Savard 2000).
The vertical distribution of Z~~ and the 137Cs w a s measured in a core at station 9.
The results reveal that the activity of these radioisotopes is a good indicator of the
position of the flood layer interface in the Bale des Ha! Ha! (Fig. 7), where the lower
contact is sharp. In contrast, this contact is difficult to identify in cores recovered in the
Bras Nord (not shown), because of a mixing of the layer due to bioturbation (Savard
2000). In the flood layer, the activity of these radioisotopes is low compared to the preflood sediments. The main reason why the 137Cs activity is low in the flood layer is
because the major source of 137Cs is the atmospheric thermonuclear tests started in 1952
(Eisenbud 1963). As the sediments eroded and transported down to the Saguenay Fjord
following the accidental event are older than these tests (deposited about 10 000 years
ago), they are almost exempt of that contamination compared to the pre-flood sediments
that are contaminated by atmospheric deposition of 137Cs. The 210Pb, which has a natural
origin, is also lower in the flood deposits because in this layer 2t~ is more decomposed
than in the recent pre-flood sediments.
Local Spatial Variability

The variability of properties can result from different aspects inducing errors on the
parameters we want to compare. The origin of variability is mainly natural and technical.
The natural variability comes from the settling of sediments from the flood. The
96
lateral variability of the thickness of the 1996 layer is very important (Maurice et al.
2000). This is primarily due to the presence of sedimentary bed forms such as channels,
dunes, mounts, etc. on the seafloor. Hill et al. (unpublished) observed in seismic records
that the major variations in sedimentation of the flood sediments are caused by of mesoscale bottom roughness in the order of meters that existed before the flood, and that were
also created during the flood. The flood layer was deposited by gravity involving sliding
and by turbidity current flow (CrEmer et al. 2002). These sedimentation processes cause
divergent-fill style of sediment accumulation where sedimentation occurs preferentially
in topographic lows. According to Hill et al. (unpublished), the meso-scale roughness on
the seafloor of the Baie des Ha! Ha! may have an impact on the sedimentological and
geotechnical properties of the bed (because of the potentially large variability in
depositional conditions and grain size), erodability (because of the hydrodynamic effects
of the meso-scale topography on the bottom boundary layer), biological recolonization
(because of the inherent heterogeneity of the new seafloor and the hydrodynamic effect
on the boundary layer) and consequently on bioturbation, bio-irrigation and geochemical
processes. To evaluate the natural variability of the flood layer thickness, a spatial
variability study was conducted on 29 box cores at stations VI and V2. The results
indicate that the thickness of the 1996 layer varies from 20 to 44 cm with a mean value of
28 cm and a standard deviation of 5.1 cm.
The technical variability concerns essentially the accuracy of the positioning system
that increases the variability of data. The positioning of the ship was provided by a
DGPS, which has an accuracy of 5 m. In addition, another variation is introduced by the
drifting of the sampler that is non negligible with water depth of about 100 to 200 m, as
found in the Saguenay Fjord. This drifting has been evaluated by comparing the position
of the ship and the position given by an acoustic positioning system, the Trackpoint II,
that gives the underwater position of a target, here the sampler, relative to reference point
on the ship. The accuracy of this system is 0.5% of slant range, which correspond to
about 1 m for a water depth of 200 m. The difference in position given by the ship and
the Trackpoint II varies from 1 to 75 m, with a mean value of 33 m. Thus, the precision
of the position of a sampling station is evaluated to be 33 m for the drifting of the sample
plus the 5 m accuracy on the position of the ship, then 38 m. This means that, even if the
position given by the ship is always the same for a given station, the sample can be off
the desired station of an average of 38 m. This is large enough to introduce variability in
the measured characteristics, as discussed previously. When using the Trackpoint II to
position the sampler underwater relatively to the ship, the error on position of the sampler
is reduced to about 6 m. In addition, the real position of all sampling stations, which is
supposed to always be the same for a given station are, in reality, as distant as 880 m for
Station 9. Therefore, it is clear that the sediment characteristics and the flood layer
thickness vary drastically inside that distance range.
Comparing Criteria for Defining the 1996 Layer

From all data acquired in this project, it seems possible for each discipline to use
one or more properties to define the 1996 flood layer. Due to the presence of the sandy
TREMBLAY ET AL. ON CATASTROPHICFLOOD SEDIMENT LAYER
97
layer and its light color compared to the pre-flood sediments, it is usually possible to
identify the flood layer by visual observation alone, especially in the Baie des Ha! Ha!
From a physical point of view, density appears to be the most useful criteria to
evaluate the position of the interface between the indigenous sediments and the flood
sediments. The densities, measured with MST or CAT-scan, reflect the water content of
sediment, its grain size distribution, degree of compaction and its mineralogy.
Furthermore, a density contrast exists even when there is no sandy layer at the base of the
flood layer, or when the sediments are bioturbated. The best geotechnical parameters are
the undrained shear strength and the liquidity index, which are both related to the density
state of sediments. Once again, these indicators of the interface between the old and the
new sediments are more evident with the presence of a sandy layer and when there is no
or little activity from benthic organisms. From a biological point of view, the
quantification of the biogenic structures using the CAT-scan technology is a valuable tool
to identify the interface between the background and the flood sediments. However, this
signature can also be erased by the bioturbation below the interface. Finally, the best
geochemical parameter to define the flood layer is indisputably the inorganic carbon
content. As mentioned earlier, in a regular sedimentation regime, the inorganic carbon is
dissolved during oxic diagenesis, but is preserved in a rapid sedimentation event as
during a flood.
From a physical and geotechnical point of view, the presence of the sandy layer
creates a sharp variation on profiles caused by the different properties of sandy sediments
compared to clayey sediments. Because the sandy layer is almost absent in the Bras
Nord, these properties are not always reliable in that sector of the fjord. In addition,
bioturbation can erase physical, geotechnical and biological signatures with time as
generally observed in the Bras Nord (results not shown), but they will be preserved if the
thickness of the 1996 layer is greater than the depth limit of organism activities. Thus,
the Cinorg seems to be the most reliable tool to identify the 1996 flood layer in the
Saguenay Fjord environment because it does not seem to be affected by bioturbation and
is still present in the absence of a sandy layer. Since the flood layer is thinner in the Bras
Nord, it will be more difficult to identify it in the future. However, in most of the area
covered by the deposit in the Baie des Ha! Ha!, the signature of the flood layer should be
identifiable using the diagnostic criteria discussed above.
Influence of the Variability on the Results and on the Temporal Evolution of the
1996 Layer
One of the primary objectives of the Saguenay Project was to monitor the evolution
of the 1996 flood layer. To do so, samples were taken each year and sometimes twice a
year at the same station and their profiles were compared. To compare adequately the data
from year to year, it is imperative that samples share the same geographical position,
sedimentation regime, and flood layer thickness. Unfortunately, comparisons are not
possible due to the significant spatial variability of the bulk sediment characteristics
mentioned above.
A solution to this problem is to normalize the sediment depth profile to the
98
measured thickness of the flood layer. An example is shown in Figure 8a for the
concentration of porewater or dissolved iron at station 9. The interface between flood
and pre-flood deposits was determined from the vertical distribution of Ci,o~g, but only the
depth is shown on Figure 8a (horizontal lines). As we see, the thickness of the flood
layer varies considerably at this station, from 22 to 33 cm. Consequently, an
interpretation of the temporal evolution of a given sediment property, for example the
concentration of iron, is complicated by the variability in the thickness of the 1996 layer.
After the normalization, the thickness of the flood layer is always the same and
dimensionless (Fig. 8b). Following this transformation, it becomes easier to evaluate the
temporal evolution of sediment properties because the scale of comparison is the same.
Thus, on Figure 8b, we see that the normalized porewater iron peak increases and rises
over with time. This simple solution, however, does not allow for the fact that the
behavior of a property could differ according to the thickness of the deposit. For
example, a contaminant could migrate to the new sediment-water interface if the flood
layer is thin or, could be confined within it if it is thick enough. Even if the flood layer
thickness normalization does not solve all the variability problems, it greatly helps to
constrain the effects of its spatial variability on the analysis-interpretation of the profiles
of diverse properties measured through sediment cores.
Fe (ppm)
1996
0
80
JllIIII
1997
160
1998
80
0
II
1999
160
2000
80
2001
160
II
II
10
x: 20
a
30
40 7 1 1 1 1 1 1
IIIJlll
/lllJIII
I I J I I I I
(a)
Interface between the 1996 layer and the pre-flood sediments
0.0 _ l l l l l t /
o
III
80
I I
160
80
160
80
160
ill
"c 0.5
CI
1.0
1.5 - I I 1 [ 1 1 1
f~l]
III
I~l]lll
I I I [ I I I
(b)
Figure 8 - Time- series analyses of dissolved iron at station 9 between 1996 and 2001 as
a function of a) depth and b) normalized depth.
TREMBLAY ET AL. ON CATASTROPHICFLOOD SEDIMENT LAYER
99
Conclusions
Physical, geotechnical, biological, and geochemical data suggest a discontinuity
between the 1996 flood deposit and pre-flood deposit. Most important, is the fact that all
the investigators involved in this project reached similar conclusions with respect to the
thickness of the layer, irrespective of the parameter measured. As a first approximation,
visual observations provide very reliable estimates in most of the sectors covered by the
flood layer, particularly in the Baie des Ha! Ha!, where the layer is generally thicker than
in the Bras Nord and where it is marked by a sandy layer at its base. In the Bras Nord,
bioturbation through the thinner flood deposit has intermixed both the upper and lower
boundaries between the flood layer and the pre-flood sediments, which makes the
discontinuity more difficult to recognize. Nevertheless, some properties such as density,
undrained shear strength, degree of bioturbation, and inorganic carbon content are useful
to differentiate the discontinuity in most of the covered area. Among them, it seems that the
inorganic carbon content is the most useful parameter to distinguish flood deposit from preflood sediments in the Saguenay Fjord, and in contrast to the other aforementioned
parameters, it is unaffected by the presence of a sandy layer at the base of the deposit.
This paper confirms that the variability of the thickness of the flood layer is
important and must be considered in temporal studies. A normalization of the depth
relative to the 1996 layer thickness may be a way to cope with its variability.
Acknowledgment
This project was funded by a strategic grant from the Natural Sciences and
Engineering Research Council of Canada (NSERC) and through the sponsorship of Alcan
Inc. The authors wish to thank all the staff and students involved in this project for their
assistance in the field and in the laboratory as well as the captains and crew of the
research vessels used during this study. Special thanks go to G. St-Onge, J.-F. Crrmer, L.
de Montety, and J. Savard for their contribution to this paper.
References
Boespflug, X., Long, B.F. and Occhietti, S., 1995, "CAT-scan in Marine Stratigraphy: a
Quantitative Approach." Marine Geology, Vol. 122, pp. 281-301.
Bossr, L., Ste-Marie, B., and Foumier, J., 1996, "Inventaire et Biogrographie des
Invertrbrrs des Fonds Meubles du Fjord du Saguenay." Rapport Technique Can.
Sci. Halieutiques Aquat. 2132.
Crrmer, J.-F., Long, B., Desrosiers, G., de Montety, L., and Locat, J., 2002, "Application
de la Scanographie ~ l'l~tude de la Densit6 des Srdiments et ~ la Caractrrisation des
Structures Srdimentaires: Exemple des Srdiments Drposrs darts la Rivirre
100
Saguenay (Qu6bec, Canada) Apr~s la Crue de Juillet 1996." Canadian

Geotechnical Journal, Vol. 39, pp. 440-450.
Deflandre, B., Mucci, A., Gagn6, J.-P.,Guignard, C., and Sundby, B., 2002, "Early
Diagenesis Process in Coastal Marine Sediments Disturbed by a Catastrophic
Sedimentation Event." Geochimica et Cosmochimica Acta. in press.
de Montety, L., Long, B., Desrosiers, G., Cr6mer, J.-F., and Locat, J., 2000,
"Quantification des Structures Biog&nes en Fonction d'un Gradient de Perturbation
dans la Baie des Ha! Ha! ~ l'Aide de la Tomodensitom6trie Axiale." Proc. of the
53 ra Canadian Geotechnical Conference, Montreal, pp. 131-135.
Drake, D.E. ,1999, "Temporal and Spatial Variability of the Sediments Grain-Size Distribution
on the Eel Shelf: the Flood Layer of 1995." Marine Geology, Vol. 154, pp. 169-182.
Eisenbud, M., 1963, "Environmental Radioactivity." McGraw-Hill Book Co., New York, 430 p.
Hill, P., Cervantes, O., Beneton, G. and C6t6, P., Unpublished, "Bottom Roughness
Characteristics of the Baie des Ha! Ha!, Saguenay Fjord, Quebec, Canada: a
Constraint on Mapping Bed Thickness of a Natural Capping Layer."
Holler, P. and K/Sgler, F.C., 1990, "Computer Tomography: a Nondestructive, High
Resolution Technique for Investigation of Sedimentary Structures." Marine
Geology, Vol. 91, pp. 263-266.
Leithold, E.L. and Hope, R.S., 1999, "Deposition and Modification of a Flood Layer on
the Northern California Shelf: Lessons From and About the Fate of Terrestrial
Particulate Organic Carbon." Marine Geology, Vol. 154, pp. 183-195.
Louchouarn, P., Lucotte, M., Canuel, R., Gagn6, J.-P. and Richard, L.-F., 1997, "Sources and
Early Diagenesis of Lignin and Bulk Organic Matter in the Sediments of the Lower StLawrence Estuary and the Saguenay Fjord." Marine Chemistry, Vol. 58, pp. 3-26.
Maurice, F., Locat, J., and Leroueil, S., 2000, "Caract6ristiques G6otechniques et
l~volution de la Couche de S6diment D6pos6e Lors du D61uge de 1996 dans la Baie
des Ha! Ha! (Fjord du Saguenay, Qu6bec)." Proc. of the 53 ra Canadian
Geotechnical Conference, Montreal, pp. 123-130.
Mucci, A. and Edenborn, H.M., 1992, "Influence of an Organic-Poor Landslide Deposit
on the Early Diagenesis of Iron and Manganese in a Coastal Marine Sediment."
Geochimica et Cosmochimica Acta, Vol. 56, pp. 3909-3921.
Mucci, A., Guignard, C., and Olejczyk, P., 2000a, "Mobility of Metals and As in
Sediments Following a Large Scale Episodic Sedimentation Event." Proc. of the
53 ra Canadian Geotechnical Conference, Montreal, pp. 169-176.
101
Mucci, A., Richard, L.-F., and Lucotte, M., 2000b, "The Differential Geochemical
Behavior of Arsenic and Phosphorus in the Water Column and Sediments of the
Saguenay Fjord Estuary, Canada." Aquatic Geochemistry, Vol. 6, pp. 293-324.
Pelletier, E., Deflandre, B., Nozais, C., Tita, G., Desrosiers, G., GagnE, J.-P., and Mucci,
A., 1999, "Crue l~clair de Juillet 1996 dans la REgion du Saguenay (QuEbec). 2.
Impacts sur les SEdiments et la Biote de la Baie des Ha! Ha! et le Fjord du Saguenay."
Canadian Jounzal Fisheries and Aquatic Science, Vol. 56, pp. 2136-2147.
Perret, D., Locat, J., and Leroueil, S., 1995, "Strength Development With Burial in FineGrained Sediments From the Saguenay Fjord, Quebec." Canadian Geotechnical
Journal, Vol. 32, No. 2, pp. 247-262.
Pocklington R., and Leonard, J.D., 1979, "Terrigenous Organic Matter in Sediments of
the St-Lawrence Estuary and the Saguenay Fjord." J. Fisheries Research Board
Can., Vol. 36, pp. 1250-1255.
Savard, J., 2000, "I~volution Temporelle des Couches Accidentelles (Crues, SEismes,
Glissements de Terrain) REcentes du Fjord du Saguenay d'Apr~s des Profils de
Radio-Isotopes de Courtes PEriodes CZ8ThF32Th,21~ 137Cs)."M.Sc Thesis,
Universit6 du QuEbec h MontrEal, MontrEal, QuEbec.
Schafer, C.T., Smith, J.N., and Seibert,G., 1983. "Significanceof Natural and Anthropogenic
Sediment Inputs to the Saguenay Fjord, Quebec." Sedimentary Geology, Vol. 36, pp. 177-194.
Schafer, C.T., Smith, J.N., and CEtE, R., 1990, "The Saguenay Fjord: a Major Tributary
to the St. Lawrence Estuary. In Oceanography of Large-Scale Estuarine System: the
St-Lawrence." Vol. 39, ed. M.I.E1-Sabh and N. Sylverberg, pp. 378-420.
Smith, J.N. and Walton, A., 1980, "Sediment Accumulation Rates and Geochronologies
Measured in the Saguenay Fjord Using the Pb-210 Dating Method." Geochimica et
Cosmochimica Acta, Vol. 44, pp. 225-240.
St-Onge, G., Leduc, J., Bilodeau, G., de Vernal, A., Devilliers, R., Hillaire-Marcel, C.,
Loucheur, V., Marmen, S., Mucci, A., and Zang, D., 1999, "CaractEristiques des
SEdiments REcents du Fjord du Saguenay (QuEbec) ~ Partir de Traceurs Physiques,
GEochimiques, Isotopiques et MicropalEontologiques." G~ographie physique et
Quaternaire, Vol. 53, no. 3, pp. 339-350.
St-Onge, G. and Hillaire-Marcel, C., 2000, "Inventaire des Couches Accidentelles REcentes
dans le Fjord du Saguenay h Partir des PropriEtEs Physiques de Carottes de Forages."
Proc. of the 53rd Canadian Geotechnical Conference, Montreal, pp. 145-151.
Yu, W., Lin, C.A. and Benoit, R., 1997, "High Resolution Simulation of the Sever
Precipitation Event Over the Saguenay Region, Quebec, in July 1996." Geophysical
Research Letters, Vol. 24, pp. 1951-1954.
Alex,]. Zeman I and Timothy S. Patterson I
Characterization of Contaminated Sediments in Hamilton

Harbour, Lake Ontario
Reference: Zeman, A. J. and Patterson, T. S., "Characterization of Contaminated Sediments
in Hamilton Harbour, Lake Ontario" Contaminated Sediments: Characterization, Evaluation,
Mitigation/Restoration, and Management Strategy Performance, ASTM STP 1442, J. Locat, R.
Galvez-Cloutier, R. C. Chancy, and K. R. Demars, Eds., ASTM International, West
Conshohocken, PA, 2003.
Abstract: Sediments in many harbors and connecting channels o f the Great Lakes have been
negatively impacted by industrial, agricultural and municipal discharges and by atmospheric
deposition. Contaminated sediments may pose risks to both human health and to aquatic
environments. All 42 U.S. and Canadian Areas of Concern (AOCs), designated by the
International Joint Commission (IJC), have problems associated with the occurrence of
contaminated sediments. Remedial Action Plans, which have been prepared for these areas,
include sediment remediation plans. Appropriate and cost-effective sedimentological,
geotechnical and geochemical investigations are required to determine the degree and extent of
contamination. The methodology used in sediment characterization includes in-situ penetration
tests, core sampling, acoustic bottom-classification systems, side-scan surveys and in-situ
porewater samplers. Contaminated sediments are typically characterized
for chemical
concentrations of contaminants of interest in terms o f both areal extent and vertical distribution.
Site specific examples of sediment analyses and characterization for remediation projects are
discussed for three sites within Hamilton Harbour, Ontario, where very heterogeneous sediments
exist.
Keywords: Great Lakes, sediment, remediation, toxicity, capping, dredging
=Aquatic Ecosystem Management Research Branch, National Water Research Institute, 867 I_akeshore Road,
Burlington, Ontario L7R 4A6.
102
www.astm.org
ZEMAN AND PATTERSONON HAMILTON HARBOUR, LAKE ONTARIO 103
Introduction
Drafting an adequate cleanup strategy for the remediation of a harbor or river
containing contaminated sediments requires detailed information on the sediment
properties and the nature and extent of contamination. There are many technologies
available for characterizing the necessary geotechnical and geochemical information from
sediments, both in-situ and in the laboratory. There are also several options for the
remediation of contaminated sediments, each with its own available technologies that
may be implemented at any given site. The characterization techniques, remedial options
and equipment utilized for the remediation of sites containing contaminated sediments
will be briefly described. Within the Great Lakes, the International Joint Commission
(IJC) designated 42 areas in both Canada and the U.S. containing contaminated sediments
as Areas of Concern (AOCs). The IJC has identified remediation of contaminated
sediments as a priority for investigation. Three sites that have been under study within
Hamilton Harbour (a high-priority AOC) will be briefly described in this paper.
Remediation Technologies
Sediment remediation techniques can be generally subdivided into two broad classes:
a) non-removal technologies which do not rely on dredging, and b) removal technologies
in which sediment is dredged from a site and is either stored elsewhere without treatment
or treated and stored in a confined disposal facility (CDF). Removal technologies
typically use environmental dredging (as opposed to navigation and construction
dredging), which is focussed on removing sediment from targeted regions with great
accuracy while restricting the amount of resuspension of contaminated sediment.
Non-Removal Technologies
Natural and Enhanced Sediment Isolation - In many cases where contaminated
sediments in harbors or bays have been investigated, the committees or government
bodies involved in organizing potential clean up plans have elected to leave the problem
alone. Sometimes the very delays involved in cleanup strategies (lack of funding,
disagreements between organizations) continue for so long that in effect, the natural
attenuation approach is implemented, even if not by choice. The natural attenuation
approach follows the principle that natural sedimentation will gradually cover over and
isolate contaminated sediments from the water column and eventually from the benthic
invertebrates. Although bioturbation (feeding, burrowing, and locomotory activities of
sediment-dwelling benthos) produces vertical sediment mixing, the zone of bioturbation
(in the Great Lakes region typically less than 10 cm) will be eventually moved into clean
sediments. In addition to sedimentation as a function of natural remediation, toxic
organic compounds are left to natural biodegradation. The natural sedimentation process
may be accelerated with the addition of a thin layer of clean sand (about 15 cm thick)
which is known as enhanced natural recovery. These passive methods ofremediation are
only effective where bottom currents and wave vortices are weak enough that surficial
sediments are not disturbed, and where no other sources of sediment disturbance occur,
104
such as heavy shipping traffic. The continued success of natural and enhanced sediment
isolation is dependent on the prevention of further contaminated discharges into the
aquatic environment.
In-Situ S e d i m e n t T r e a t m e n t - Sediments in both marine and freshwater environments
have been treated in-situ as an economical means for remediation, where the natural
attenuation approach has been rejected. The type of in-situ remediation chosen is
dependent on the nature of contamination. PAHs and total petroleum hydrocarbons
(TPHs) have been treated in-situ by aeration and the addition of nutrients mixed with the
oxidant calcium nitrate (Murphy et al. 1995) which caused significant biodegradation of
PAHs. Where freshwater eutrophication is a concern, sediments may be treated with
chemicals such as aluminum sulphate (alum), lime, calcite and ferric chloride (Klapper
1991).
In-Situ C a p p i n g - In-situ capping is the covering over of contaminated sediments
with a clean layer (usually sand) thick enough to isolate the sediment from benthos and
other aquatic organisms (Palermo et al. 1998). This remedial alternative is suitable for
situations when the sediment is contaminated with a wide range of compounds, e.g. with
several persistent organic contaminants and toxic metals. The application of the capping
layer must be placed so as not to significantly disturb the surficial sediment (i.e. the
capping material must not be too coarse or too heavy so that it will not penetrate the very
soft sediment). This has been shown to be achievable using medium to coarse sand as the
capping material over very soft silty clay (Zeman and Patterson 1997). Once the capping
material has been placed, the cap may be further protected from water currents by an
additional overlying layer of gravel or boulders.
R e m o v a l Technologies
Improved and newly designed dredges have been created for environmental cleanup
projects with the goals of minimizing sediment resuspension, and providing for the more
accurate and complete removal of contaminated sediments (Environment Canada - OR
1998; Dredging Supply Company Inc. 2002). Several of these dredges are herein briefly
described.
Cable A r m E n v i r o n m e n t a l B u c k e t - This Canadian dredge is designed to remove
contaminated sediment on a flat plane, using a precision clamshell aggregate bucket, so
that a level, horizontal cut is completed as opposed to a curved cut, typical of other
dredges. Cables anchored at the four comers of the bucket ensure this type of cut.
M u d C a t - The Mud Cat is a U.S. made dredge designed to operate in shallow marine
environments. It features an effective sediment removal system consisting of a boommounted horizontal auger and a centrifugal slurry pump. There is also dual convergence
at the auger head. The dredge is designed to minimize the resuspension of the sediment
while maximizing the solids content of the dredged material. An hydraulic forward tilt
and manual transverse tilt of a truss boom allows for use on sloping underwater sediment.
ZEMAN AND PATTERSON ON HAMILTON HARBOUR, LAKE ONTARIO 105
P n e u m a P u m p - The Pneuma Pump was developed in Canada and consists of a barge

mounted dredge utilizing guide cables operated above the water surface. The dredge is
pulled by a winch system with a speed of about 4-5 m per minute. The Pneuma Pump
was successfully used in the cleanup of Collingwood Harbour, which became the first
AOC in North America to become delisted.
V i s o r G r a b - The Visor Grab was developed in the Netherlands, and utilizes a sealed
dredge bucket. It has a standard backhoe bucket with a controlled visor, operated by
hydraulic cylinders on each side of the bucket. The bucket can be used on standard
excavators.
A m p h i b e x - The Amphibex is a Canadian designed dredge using a barge equipped
with spuds, stabilizers and an excavator. An hydraulic arm and bucket with attached
pumps allows the Amphibex to hydraulically dredge sediment without requiring any
suction. Positive pressure is used instead, which increases the solids content and
minimizes excess water. The Amphibex has proved to be less expensive than
conventional dredges, meeting regulatory requirements.
V i s c o u s E x c a v a t o r - This U.S. designed dredge uses an enhanced rotary cutterhead
element that removes viscous sediment more efficiently than conventional dredges. The
hollow cutterhead uses circular rings that rotate and pass the dredged material through it.
Sediment Toxicity
For areas of Canadian jurisdiction in the Great Lakes, the Ontario Ministry of
Environment and Energy (MOEE) sets the provincial sediment quality guidelines
(PSQGs) for sediment toxicity. Determining what can be classified as toxic sediment in
legal terms is no simple task. Toxicity must be determined in relation to the effects
contaminants have on organisms for any given area. The obvious disadvantage to this is
the fact that different species of organisms react differently to contaminants. As a result,
the MOEE does not rely on single-species data to develop PSQGs. Another disadvantage
is the fact that some species develop a level of immunity to contaminants over time. The
tolerance of species so exposed will thus differ to that of unexposed communities of the
same species. The species of organism to be used for a standard, and the mortality rate
from exposure to any given chemical or element must thus be established with great care.
This includes the determination of a time of exposure and a percentage rate of mortality
after this elapsed time. Based on the long term effects that contaminants in sediment
have on sediment dwelling organisms, the MOEE further classifies sediment into three
categories (Persaud et al. 1994):
Fish and sediment dwelling organisms are not affected.
Majority of sediment dwelling organisms are not affected.
L e v e l - "Likely to affect the health of sediment dwelling organisms".
No Effect Level -
Lowest Effect Level Severe Effect
106
Sediment with this rating requires further testing to determine if it is acutely toxic.
Acute toxicity is generally measured as the percentage of an effluent in clean water that
causes a 50% lethality rate to test organisms over a 96-hour period (Ontario Ministry of
the Environment 1988).
An approach created by researchers at the National Water Research Institute (NWRI)
classifies sediment toxicity based on the effects observed on four benthic invertebrate
species ( Chironomus riparius, Hexagenia limbata, Hyalella azteca and Tubifex tubifex).
Software has been developed to classify sediment into three categories based on this
method, namely, non-toxic, potentially toxic and toxic (Reynoldson and Pascoe 2000).
For comparison, in nearly all of U.S. National Oceanic and Atmospheric
Administration's (NOAA's) sediment toxicity studies, three separate bioassays are used
to measure toxicity. The three bioassays ensure that a range of species and assessment
end points (e.g., mortality, reproductive impairment, physiological stress) would be tested
to provide a wide perspective on adverse biological effects. The NOAA's bioassays are
used to determine: (1) amphipod survival following a 10-day exposure to bulk (or) whole
sediment; (2) successful fertilization of sea urchin (or other invertebrates) following
exposure of gametes to pore water; and (3) decreased light production, indicative of
metabolic stress, by a luminiscent marine bacterium, Vibrofisheri, when exposed to an
organic extract of sediment (Hameedi et al. 1999).
Mapping of Contaminated Sediments
Acoustical Mapping
The texture of sediments is an important factor for mapping sediment beds as
contaminants are usually found in much higher concentrations in finer sediments. In
addition, core samples are more difficult to retrieve from sandy sediments. A map of
surficial sediment texture may thus not only be a good indication of where contaminated
sediments are more likely to be found, but may show areas of sediment where core
samples can be more easily retrieved. If areas containing sandy, surficial sediments are
known, the planning of sampling at these sites may be changed accordingly (e.g. the use
of vibracoring instead of gravity coring).
Over the years, various technologies have been used for mapping marine and
freshwater beds using acoustical equipment, primarily for navigational and fisheries
purposes. As these technologies expanded and improved, some were utilized for the
mapping of sediment texture and type. Some systems presently in use were developed
especially for this purpose. An example of this is the acoustic marine-bed classification
system known as RoxAnn TM,which has been successfully used within the Great Lakes on
various projects in determining surficial sediment type (Rukavina and Caddell 1997).
RoxAnn works on the principle that various textures of surficial sediment reflect acoustic
signals differently. The type of echo received by RoxAnn is interpolated and instantly
classified within the software system used in conjunction with the unit.
ZEMAN AND PATTERSON ON HAMILTONHARBOUR, LAKE ONTARIO 107
An acoustical coring analysis method was developed by Caulfield Engineering to

classify sediments for dredging operations (Tuchman et al. 1997). Acoustic Core TM is an
acoustic remote sensing tool which was jointly field tested with the U.S. Army Corps of
Engineers R&D Center and which is now commercially available. Both core samples and
acoustical measurements were made at several sites in Great Lakes harbors and rivers to
compare the accuracy of the acoustical profiles. The profiles of sediment distribution
were reported to have an accuracy of 95% in estimating density in marine sediment.
A multiple-frequency echo sounding system known as DSLP (Detection of Sediment
Layers and Properties) can provide high-resolution analysis of the stratification of
suspended sediments and individual sediment layers of unknown physical and
sedimentological properties (Eden et al. 1999, Kern and Westrich 1999). DSLP uses a
complex digital signal to determine sediment type. The DSLP method can be used to
determine required volumes of sediments to be dredged for sediment remediation
projects.
Chemical Mapping
Chemical mapping of contaminated sites is best done in three dimensions, which
requires the sampling of sediment below the sediment/water interface. This is done by
taking core samples, whereby a tube usually made of clear acrylic is either dropped or
pushed through the sediment. Core samples may then be cut and subsampled at any depth
measured from the top of the core. Where duplicate core samples of the upper sediment
from the same location are desired, a large box corer with a retractable bottom is often
used to scoop up almost 1 m 3 of sediment. Two or more short core tubes are then pushed
through the sediment by hand and pulled out.
The mapping of chemical compounds within contaminated sediments has proven to
be a challenge at many sites. Deposited compounds from spills or dumped material tend
to be very site specific and initially cover a small area, often spreading out into a thin
layer of only several millimetres thick. Over several years, natural sedimentation will
bury such a layer. Subsampling cores at intervals that are too far apart may result in
contaminated layers being missed, especially if the sediment color and texture are very
similar to the neighboring sediment within the core. Contaminated sediments also tend to
be very fine, and are thus easily disturbed by currents, propeller wash, and anchor
dragging from passing ships. In industrial areas, contaminant spills not only come from
various sources, but usually differ in chemical makeup, frequency of occurrences and
volume. As a result, the composition of sediments in industrial areas is often very
heterogeneous and, in extreme cases, can significantly vary among sediment samples
collected only several metres away. Accurate three-dimensional chemical mapping is
thus very difficult in cases like this, and the site sediment can only be properly mapped
with a high density of sampling locations. The constant disturbance, migration, and
settlement of sediments results in the rendering of chemical maps obsolete over several
years, depending on the rate of change.
108
Chemical mapping may be done for hundreds of different elements and chemical
compounds found within contaminated sediment. The sampling technologies used for
extracting and measuring these chemicals vary in cost and procedures. One or more sets
of chemicals may be identified as being the primary contaminants in industrial sediment
(e.g. PAHs), but the extent of contamination of other chemical compounds may be
unknown and even unsuspected. Since budget limitations usually prevent the analyses for
all known hazardous substances that may be found in sediment, priority contaminants
must be established based on aquatic ecosystem health and budgetary limitations. The
determination of such priority contaminants is usually based on previous biological
monitoring (e.g. deformity and mortality rates of fish and animal species, and
determinations of sediment toxicity from bioassays of benthic invertebrates). It is wise to
test a small percentage of samples for hazardous compounds that are not known to be in
high concentrations at the site, but are potentially present. If unexpected high
concentrations of additional compounds are found, further testing for these compounds
may be deemed prudent on other cores.
Samples taken from an initial set of sampling sites usually reveal areas of sediment
that require more detailed mapping, and thus a secondary set of samples. It is often the
case where one sample from a set may reveal unusually high concentrations for some
contaminants in comparison to other samples taken nearest to it. This may indicate a
small localized spill or dump site of contaminated sediment, which may be deemed
important enough to map in greater detail to aid in determining a remedial strategy.
In principle, the density of sampling stations should be as great as possible, but
usually the total number of cores and samples is determined by time requirements and
budgetary constraints. For smaller surveys, the position of sampling stations depends
primarily on the size and shape of the water body. Different sampling strategies and
sampling grid designs (e.g. square grids, nested grids, parallel line sampling grids with
equal and unequal spacing, ray sampling grids) can be chosen for more complex surveys
(Golterman et al. 1983).
Once areas of contaminated sediments are sufficiently mapped to reveal calculated
volumes and measured concentrations, remedial options may be considered. Volumes are
determined by setting a contaminant concentration limit, and measuring the vertical range
in each core where concentrations meet or exceed the limit, in addition to the horizontal
distances between the cores. Computer software programs may then be used to calculate
the volume and three dimensional profile of the contaminated sediment deposit(s).
Chemical mapping of contaminated sediments is closely tied to risk assessments of
contaminated sites, considering both ecological risks quantified on the basis of laboratory
toxicity experiments and human-health risks, where assessments result e.g. in fish
consumption guidelines or cancer risk levels (FOrstner 2002).
ZEMAN AND PATTERSON ON HAMILTON HARBOUR, LAKE ONTARIO
109
Geotechnieal Investigations
In-Situ Tests
Physical in-situ testing of the geotechnical properties of sediment is usually

conducted to determine the strength and consolidation of the sediment (Richards and
Zuidberg 1985). Sediment is composed of solid particles and pore water, often with the
presence of gas, and is always under smile pressure from the overlying water column.
Geotechnical tests look at one or more of these attributes and their relationships.
Cone penetration testing may be conducted on in-situ sediments to obtain a quick and
economical general understanding of shear strength tendencies of site sediments. A
device known as a free-fall penetrometer, called STINGTM (Jasco Research Ltd. 2001)
which is shaped like a long rod with a fin, has been used for determining approximate
shear strength values at sites where contaminated sediments exist. The penetrometer is
dropped from a boat and allowed to fall and sink into the sediment bed. An electronic
measuring device attached to the instrument measures the rate of reduced speed upon
impact of the tip with the sediment. This gives an indication of the shear strength of the
sediment, which can aid in the interpretation of sediment grain size. For example, if the
penetrometer penetrates through 50 cm of sediment with an abrupt stop at 50 cm, it may
be assumed that the upper 50 em sediment layer is comprised of soft, fine-grained
sediments underlain by a different type of sediment with much higher shear strength (e.g.,
an overconsolidated glaciolacustrine sediment). A general knowledge of the site
sediment stratification is necessary, however to avoid misinterpretation. Both sand and
hard clay layers will greatly slow down or stop the penetrometer.
Specialized cone penetrometers have been equipped with electrodes to measure
sediment temperature, electrical conductivity, resistivity and shear wave velocity. Other
penetrometers have been designed to sample pore water and to be used for outflow
hydraulic conductivity testing (Campanella et al. 1993). Dialysis samplers, known as
"peepers" (Hesslein 1976) have been deployed into sediment at study sites to extract pore
water through a porous membrane into chambers filled with distilled water (Zeman and
Patterson 2000). The peepers remain vertically oriented in the sediment, over a time
period of several weeks. The diffusive equilibrium that results within the peepers allows
for the capture of major ions and nutrients.
In-situ testing of sediment also includes the use of field vane tests, porewater
pressure tests and seepage measurements. Field-vane tests are used for determining the
undrained shear strength and sensitivity of soft cohesive sediments. A four-bladed vane
fastened to the bottom of a vertical rod, is gently pushed into the sediment and is then
rotated. The relationship between torque and angular rotation is recorded so that
undrained shear strength may be calculated. Porewater pressure may be determined by
the use of differential piezometers, which are used to measure the extent of
underconsolidation in soft fine-grained sediments. Seepage meters may be installed in
nearshore sediments to monitor the interaction between groundwater and open water
110
bodies. This allows for measurement of the horizontal and vertical sediment
permeability, and the quantity and direction of groundwater flow (Zeman and Patterson,
2002).
Laboratory Tests
Where in-situ testing is deemed to be too costly or impractical, laboratory testing on
samples is required to determine geotechnical parameters. One of the most basic
classification tests performed on sediments is grain size. Sediment samples are generally
classified by their grain size and plasticity according to the Unified Soil Classification
System (Terzaghi and Peck 1968, ASTM Standard D 2487, Standard Classification of
Soils for Engineering Purposes). Under this system, sediments are classified into three
major groups: coarse-grained, fine-grained, and highly organic (peaty). The grain size of
sediment samples factors into the results of many laboratory tests. Such tests include
moisture content, shear strength, plasticity, consolidation, viscosity and specific surface
area. The moisture content of sands is around 30%, whereas the moisture content of soft,
fine grained sediment may be as high as 400% or greater. Shear strength can range from
< 1 kPa for very soft sediments, to > 80 kPa for hard clays. The specific surface area for
clay minerals varies from 5-20 m2/g for kaolinite to 700-800 m2/g for smectite group
minerals with expandable internal surfaces (Mudroch et al. 1997).
Site-Specific Examples From Hamilton Iiarbour, Ontario

Overview of Hamilton Harbour
Hamilton Harbour is located on the western end of Lake Ontario and has an area of
approximately 31 square kilometres (Fig. 1). The harbor has been designated as one of
the 42 AOCs within the Great Lakes by the IJC. It contains contaminated sediment that
exceeds the MOEE sediment quality guidelines at the "severe effect" level for both
metals and organic compounds (Persuad et al. 1994). The southern shore of the harbor
borders mainly on industrial land. The eastern shore is comprised of a narrow sand bar
used primarily as a transportation corridor. The north shore has a mix of land use, being
mainly residential and recreational. PAHs and heavy metals are most concentrated near
the industrial shoreline.
ZEMANANDPATTERSONONHAMILTONHARBOUR,LAKEONTARIO 111
Ca00,ngS,te
Red Hill
Creek
Scale
(kin)
:~
:3
Windermere Arm / \ ~ ~ ~ ' ,

Study Area
Windermere Basin
Figure 1 - Study areas in Hamilton Harbour
Hamilton Harbour ln-Situ Capping Project

A one hectare pilot size capping project was completed near the north shore of the
harbor in 1995 (Zeman and Patterson 1997). The contaminants o f concern were trace
metals and PAHs that had migrated northward from the south shore industrial area. A
sand cap averaging about 34 cm in thickness was successfully placed over very soft
contaminated sediments. Because of the sand cap, post-capping cores had to be taken by
vibracoring. The cores showed a very sharp sediment/cap interface and a significant
difference in trace metal concentrations between the sediment and the sand cap (Fig. 2).
112
Metal Concentrations
Core 1-97
Depth (cm)
0 m
Sand Cap
Sand Cap
...................
20 I
40
22..F
8060~
Sediment
.
Sediment
100
1
0
~
2000
100 200 300 400 500 600
Concentration(ug/g)
i
Y "' ~ ~
4000
6000
[~
Figure 2 - Metal concentrations in core from capping site in Hamilton Harbour
20
50
Figure 3 - Bathymetric changes after capping
1 13
Multibeam echo-soundings were taken of the site before, during and after cap
placement (Fig. 3). This consisted of sounding intervals that were approximately 5 m
apart, together producing a grid pattern of depths. Changes in bathymetry were most
evident within the capping area.
The placement of the cap was clearly delineated by using the acoustic sounder
RoxAnn. As with the multibeam echo soundings, RoxAnn mapping was conducted
before, during and after capping.
Randle Reef Site

The Randle Reef "hot spot" is located near the south shore of Hamilton Harbour
(Fig. 1). It is considered to contain the most highly contaminated sediment for PAH
concentrations in the harbor. Due to the softness of the sediment, several sets of gravity
cores were easily collected. Subsamples for PAHs were taken at varying intervals from
the cores and mapped using a contouring software program (Fig. 4). Since the sediment
at the site was found to be extremely heterogeneous, chemical maps of the site showed
widely varying concentrations within a relatively small area. The maps significantly
changed in appearance depending on how the extent of contamination was expressed (e.g.
maximum PAH concentrations found vs. maximum depths of highly contaminated
sediment). The maximum concentration of PAHs found at the site was 17 000 ppm near
the middle of the hot spot area.
Easing
NorthAmericanDatum1983
Figure 4 - Maximum PAH concentrations regardless of depth (A), and maximum depth of
PAH concentrations > 700 ppm (B), Randle Reef Study Site
114
Discussions on potential remedial measures for the Randle Reef hot spot have been
ongoing for several years. Acoustical, penetration, and chemical mapping conducted at
the site was instrumental in determining the extent of the hot spot, both in area and
volume. Due to the area coverage and volume of contamination, the remedial process is
greatly limited by financial resources available. Various options of remediation have
been considered, including dredging. For this option, it was proposed to remove
approximately 20 000 m 3 of the most toxic sediments with sum PAH concentrations
greater than 700 pg/g PAH minus naphtalene. Several ex situ sediment treatment options
have been assessed for applicability (Environment Canada 1999).
A final remedial project for Randle Reef has yet to be undertaken, although
stakeholders (committee members) have tentatively agreed to a confinement and capping
plan. A structure is to be built at Randle Reef that will be large enough to confine the onsite sediment in addition to dredged sediment from other contaminated sites in the harbor.
The structure will likely be capped above the water level to form an island or peninsula
(McGuirmess 2002).
Windermere Arm Site
Windermere Arm is a narrow channel situated in the southeast portion of the harbor,
covering an area of approximately 80 hectares (Fig. 1). The western shoreline of the Arm
is industrial. The northern section of the Ann is a well used shipping corridor. Red Hill
Creek flows into a basin which empties into the Arm from the south. An old landfill site
and an operating sewage treatment plant are situated alongside the Red Hill Creek.
Although sediment contamination in the Arm is not as severe as that found in Randle
Reef, concentrations of PAHs and PCBs are well above the MOEEs "lowest effect level"
guidelines, whereas concentrations of at least seven trace metals exist above the "severe
effect" level. Digital photographs taken of opened cores (Fig. 5) do not always give a
clear indication of where the most highly contaminated sediment may be found. The
sediment in Windermere Arm is not as heterogeneous as sediment found at Randle Reef,
but chemical contamination is still extensive, and chemical compounds often reach their
peak values independent of each other at various depths in the sediment (Table 1).
Figure 5 - Sediment core from Windermere Arm
1 15
Full chemical characterization of the sediment at Windermere Arm has yet to be

finalized. It is expected that a management strategy will be completed in 2004.
Table 1 - PAH, PCB and lead comparisons between two cores from Windermere Arm.
Core
Depth, cm
PCBs, ppm
Lead, ppm
0
10
41
PAHs,
ppm
22.99
113.81
0.28
A6A
1.84
.....
Not Detected
91.38
2,302.00
<2.25
A8
0
30
50
3.93
0.25
0.07
0.34
159.00
Not Detected
<5.0
<5.0
<5.0
Conclusions
The cleanup of AOCs and other sites with contaminated sediments require adequate
mapping and characterization of contaminants within the sediment. This is to ensure that
the volume of contaminated sediments and its geotechnical and geochemical properties
are known. From this information, the appropriate remedial technologies may be selected
for any cleanup plan, subject to budget limitations. Authoritative sediment toxicity
guidelines are necessary to determine whether contaminant levels in sediments at any site
warrant a cleanup plan.
From the three sites examined in Hamilton Harbour, it is clear that acoustical
mapping along with geotechnical and geochemical testing have provided significant
information for site delineation and sediment composition. Accurate placement of the
sand cap for the 1995 capping project was confirmed through both acoustical mapping
and cores collected at the site. The tentative cleanup strategy for Randle Reef was based
largely on sediment data collected from the site. Ongoing sampling and testing of
Windermere Arm sediments is expected to identify areas within the Arm to be selected
for a likely future cleanup project.
116
R e f e r e n c e s
Campanella, R. G., Davies, M. P., and Boyd, T., 1993, "Characterizing Contaminated
Soil and Groundwater Systems with In-Situ Testing," Proc.1993 Joint ASCECSCE National Conference on Environmental Engineering, July 12-14,
Montreal, Qurbec.
Dredging Supply Company Inc., 2002,
URL: http://www.dscdredge.com/viscousexcavator.html, Reserve, Louisiana.
Eden, H., Mialler,V., and Vorrath, D., 1999, DSLP Method, a New Surveying
Technology. In: DVW-SchriftenreiheBand 37, '~3ew~ssermessungund
Hydrographische Informationssysteme",pp. 106-115, Konrad Wittwer Verlag,
Stuttgart.
Environment Canada, Environmental Protection Branch Ontario Region, Great Lakes
2000 Cleanup Fund, 1998, The Remediation Technologies Program Great Lakes
2000 Cleanup Fund, Technical Report, Downsview, Ontario.
-
Environment Canada, Environmental Protection Branch Ontario Region, Great Lakes

2000 Cleanup Fund, 1999, Randle Reef Sediment Remediation Project,
Environmental Screening Report, Draft Report, Downsview, Ontario.
-
Frrstner, U., 2002, "Sediment Sampling, Sample Preparation, Grain Size Corrections,
and Chemical Criteria", In: Environmental Monitoring Handbook, Burden, F.R.,
McKelvie, I., Frrstner, U. and Guenther A. (eds.), McGraw-Hill Companies, Inc.,
New York, pp. 14.1-14.35.
Golterman, H. L., Sly, P. G., and Thomas, R. L., 1983, "Study of the Relationship
Between Water Quality and Sediment Transport", Technical Papers in Hydrology
No. 26, UNESCO, Paris.
Hameedi, M. J., Long, E. R., and Harmon, M. R., 1999, "Sediment Toxicity", In:
NOAA's State of the Coast Report, National Oceanic and Atmospheric
Administration, Silver Spring, MD. URL: .http://state-ofcoast.noaa.gov/bulletins/html/sed_15/sed.html
Hesslein, R.H., 1976, "An In-Situ Sampler for Close Interval Pore Water Studies,"
Limnology and Oceanography, Vol. 21, pp. 912-914.
Jasco Research Ltd., 2001, STING Mk.II. Seabed Terminal Impact Newton Gradiometer.
Product Specifications, URL: http://www.jasco.com/downloads/sting.pdf,
Victoria, BC.
1 17
Kern, U., and Westrich, E., 1999, Methoden zurErkundung, Untersuchung und
Bewertung von Sedimentablagerungen und Schwebstoffen in Gewiissern (Methods
for Mapping, Investigation and Assessment of Sediment Deposits and Suspended
Sediment in Water Bodies), Schriften, Deutsche Vereinigung ftir Wasserwirtschaft
(DVWK), No. 128, Bonn. Germany.
Klapper, H., 1991, Control of Eutrophication in Inland Waters, Ellis Horwood, New
York.
McGuinness, E., 2002, "Harbour Stakeholders Agree to Cap Hot Spot," The Hamilton
Spectator, April 25, 2002, Hamilton, Ontario.
Mudroch, A., Azcue, J. M. and Mudroch P. (eds.), 1997, "'Manual of Physico-Chemical
Analysis of Aquatic Sediments ", CRC Press Boca Raton, FL.
Murphy, T., Moiler, A., Brouwer, H., 1995, "In-Situ Treatment of Hamilton Harbour
Sediment," Journal of Aquatic Ecosystem Health, Vol. 4, pp. 195-203.
Ontario Ministry of the Environment., 1988, "Toxicity of Pulp and Paper Effluent in
Ontario," Queen's Printer, Toronto, Ontario.
Palermo, M., Maynord, S., Miller, J., and Reible, D., 1998, "Guidancefor ln-Situ
Subaqueous Capping of Contaminated Sediments," EPA 905-B96-004, Great
Lakes National Program Office, Chicago, Illinois.
Persuad, D., Jaagaumagi, R., and Hayton, A., 1994, "Guidelinesfor the Protection and
Management of Aquatic Sediment Quality in Ontario," ISBN 0-7729-9248-7,
Ontario Ministry of the Environment, Water Resources Branch, Toronto, Ontario.
Reynoldson, T. R., and Pascoe, T., 2000, "Development and Application of the BEAST
Softwarefor Using Biological Sediment Guidelines Developed for the Great
Lakes. Abstracts," 35 th Central Canadian Symposium on Water Pollution
Research, Burlington, Ontario.
Richards, A. F., and Zuidberg, H. M., 1985, "In-Situ Determination of the Strength of
Marine Soils," Strength Testing of Marine Sediments: Laboratory and In-Situ
Measurements, ASTM STP 883, R.C. Chaney, K.R. Demars, Eds., Philadelphia,
PA, pp. 11-13.
Rukavina, N.A., and Caddel, S., 1997, "Applications of an Acoustic Sea-Bed
Classification System to Freshwater Environmental Research and Remediation in
Canada," Proceedings of the Fourth Conference on Remote Sensingfor Marine
and Coastal Environments, Orlando, Florida, pp. I-317-1-329.
1 18
Terzaghi, K., and Peck, R. B., 1968, Soil Mechanics in Engineering Practice, 2 nd Ed.,
First Corrected Printing, Wiley, New York.
Tuchman, M., Bolattino, C., and Miller, J., 1997, "Solving Great Lakes Contaminated
Sediment Problems," Proceedings of the National Conference on Management
and Treatment of Contaminated Sediments. Cincinnati, Ohio, May 13-14, 1997.
Zeman, A. J., and Patterson, T. S., 1997, "Results ofln-Situ Capping Demonstration
Project in Hamilton Harbour, Lake Ontario," International Symposium on
Engineering Geology and the Environment, Athens, Greece, June 23-27, 1997.
Zeman, A. J., and Patterson, T. S. 2000, "Results of Physical, Chemical and Biological
Monitoring at the In-Situ Capping Site in Hamilton Harbour," Contribution 00088, National Water Research Institute, Burlington, Ontario.
Zeman, A. J., and Patterson, T. S., 2002: "Sediment Physical Parameters and
Techniques", in: "Environmental Monitoring Handbook", Burden, F. R.,
McKelvie, I., F6rsmer, U., Guenther, A. (Editors), McGraw-Hill Companies, Inc.,
New York, pp. 15.1-15.20.
Ann-Laure Moreau, 1 Jacques Locat, 2 Philip Hill, 3 Bernard Long,4 and Yvon Ouellet 5
In situ Flume Measurements of Sediment Erodability in Saguenay Fjord (Quebec,

Canada)
Reference: Moreau, A-L., Locat, J., Hill, P., Long, B., and Ouellet, Y., " I n situ Flume
Measurements of Sediment Erodability in Saguenay Fjord (Quebec, Canada),"
Contaminated Sediments: Characterization, Evaluation, Mitigation/Restoration, and
Management Strategy Performance, ASTM STP 1442, J. Locat, R. Galvez-Cloutier, R. C.
Chaney, and K. R. Demars, Eds., ASTM International, West Conshohocken, PA, 2003.
Abstract: The erodability of surficial sediment in Saguenay Fjord, Quebec, was
investigated over a three-year period at fourteen sites in Bale des Ha! Ha! and the North
Arm using a benthic flume, Miniflume. Previous studies showed that Saguenay Fjord
sediment characteristics vary greatly on small and large scales. The critical shear stress
(re) reflect this variability, with re between 0.07 and 0.48 Pa. The measured critical shear
stress and erosion rates were consistent with those of other in situ studies in other areas.
The variability in the results is linked to the great spatial variability of benthic sediments
rather than to temporal variability. Some of the results were integrated in the Parchure
and Mehta (1985) erosion law since the upper portion of the active sediment layer is
typically a stratified bed. The identification of erosion parameters (critical shear stress,
erosion rate, parameters a and Efin erosion law) and their high variability is the first step
in defining the erosion susceptibility hazard in Saguenay Fjord.
Keywords: in situ measurements, erodability, erosion rate, critical shear stress, Saguenay
Fjord, contaminated sediments, annular flume.
1 Ph.D. student, Department of Geology and Geological Engineering, Laval University,

Quebec, Canada G1K 7P4.
2 Professor, Department of Geology and Geological Engineering, Laval University,
Quebec, Canada G1K 7P4.
3 Marine Sedimentologist, Natural Resources Canada, Geological Survey of Canada, P.O.
Box 6000, Sidney, B.C., V8L 4B2.
4professor, INRS-ETE, 880 Chemin Ste-Foy, C.P. 7500, Quebec, Canada G1V 4C7.
5 Professor, Civil Engineering Department, Laval University, Quebec, Canada G1K 7P4.
119
www.astm.org
120
Introduction
A flash flood that occurred in July 1996 in the Saguenay area (Quebec) (Figure 1),
deposited a bed of cohesive sediment, on averagel0-60 cm thick (but locally up to 5 m
thick in the proximal delta environment) over an area of 65 km 2 of the deep fjord basin of
the upper reaches of the Saguenay Fjord (Pelletier et al. 1999). This new layer, called the
1996 layer, is expected to act as a capping layer over underlying contaminated sediments.
Industrialization of the upper reaches of the Saguenay valley during the 19th and 20th
centuries resulted in significant pollution of the benthic environment in the fjord, with
organic matter, heavy metals and polyaromatic hydrocarbons (PAH) being the principal
pollutants (Schafer et al. 1990). This research is part of a multidisciplinary research
project aimed at determining the efficiency of the 1996 layer to act as a capping layer.
This paper reports on studies of the erodability of surficial sediment in Saguenay Fjord.
The underlying hypothesis is that contaminants that diffuse through the 1996 layer from
underlying sediments may be re-suspended through current-induced erosion of the
surficial layer.
Over most of the study area the 1996 layer consists of cohesive silty clay. There are
two possible approaches to the study of cohesive sediment erosion: laboratory methods
(Villaret 1987, Dennett 1995) and in situ methods (Amos et al. 1992, 2000, Amos and
Gibson 1994, Maa et al. 1993, 1998). Due to the complexity of the natural bed sediments
and dependence of the sediment properties on physical, chemical and biological
properties, it was decided to study erodability in situ so as to avoid the need to transfer
sediment from the field to the laboratory. To achieve our objectives a medium size
annular flume, called Miniflume, was used (Amos et al. 2000). Constant geometry,
existence of a completely-developed boundary layer above the sediments, no pumps to
break down the suspended aggregates (such as fecal pellets and flocculated sediment) are
the principal advantages of an annular flume (Widdows et al. 1998). Paradoxically, the
principal disadvantage results from the geometry since the curvature of annular flumes
brings about the occurrence of secondary radial flows (Maa et al. 1993, Ravens and
Gshwend 1999). However this disadvantage can be minimized by reducing the channel
width (Fukuda and Lick 1980). The medium size of the Miniflumeallows easy
deployment and provides measurements that take into account the most significant
physical and biological parameters affecting erodability (Widdows et al. 1998).
Erodability is often characterized by the erosion threshold (~'c) (Pa) and erosion rate
(E) (kg.m2.s -~) although there is not a uniform definition for the erosion threshold
(Young and Southard 1978, Amos et al. 1992, Maa et al. 1998, Ravens and Gschwend
1999). In the same way, erosion rate (E) has been described by several mathematical
expressions. Ariathurai and Arulanandan (1978) used linear expressions, Mehta et al.
(1982) and Parchure and Mehta (1985) proposed an exponential expression, while
Kusuda et al. (1984) used a power-law expression. All these expressions are based on
excess bed shear stress. There are also power law expressions based on applied bed shear
stress (Lavelle et al. 1984) that imply that a critical value of the bed shear stress does not
exist, and a generalized exponential expression based on bed shear strength (Lee et al.
1994, Dennett 1995). Some of these expressions are appropriate for a uniformly
deposited bed (linear expression) while others apply to stratified deposits (exponential
MOREAU ET AL. ON IN SITU FLUME MEASUREMENTS
121
expression based on excess bed shear stress). 1996 layer corresponds to several pulses of
sedimentation due to turbidity currents (Cr6mer et al. 2002). It is composed of sediments
(sable and silt) set up by traction overlain by sediments (silty clay) set up by settling
(Cr6mer et al. 2002), which correspond to a stratified deposit (Villaret 1987).
In this study, the results of three annual deployments are presented and discussed.
The spatial and temporal trends are examined for stations distributed over the study area
(Figure 1). Some of the results were compared to the Parchure and Mehta (1985) erosion
law since the upper portion of the active sediment layer is typically a stratified bed.
Figure 1 - Location of Saguenay Fjord and Miniflume Deployment Sites. (a) ST19
Position in 1999, 2000 and 2001 and (b) ST36 Position in 1999, 2000 and 2001.
Environmental Setting
The Saguenay Fjord is the largest fjord in eastern Canada. It joins the St. Lawrence
estuary at Tadoussac about 180 km northeast of Quebec City (Figure 1). The Fjord
occupies a drowned valley, 93 km long and 1-6 km wide, incised into the crystalline
rocks of the Canadian Shield. It has a typically U-shaped cross-section and has high rock
walls from 366 m above to 274 m below sea level. The Fjord may be divided into three
sections: the main Fjord, Bale des Ha! Ha! and the North Arm. The study area includes
the North Arm and Baie des Ha! Ha!. It is characterized by sandy, silty clay sediments
and water depths ranging from 40 to 180 m. The North Arm is the area where fresh
122
waters from the Saguenay River encounter the marine waters from the Fjord (Drainville
1967). Ninety percent of the freshwater runoff into the Fjord is derived from this river,
which is about 63 km long and has a mean monthly discharge rate of 1500 m3s ~. The
natural flow regime of the river has been regulated by a series of hydro-electric dams
built between 1923 and 1943 (Schafer et al. 1990).
Miniflume
Description and Data Obtained
The Miniflume is an annular benthic flume made of acrylic. It is 30 cm high, has a

29 cm outer diameter and a 20 cm inner diameter, resulting in a 4.5 cm wide channel with
a total bed area of 0.0346 m 2 and a maximum flow volume about 0.01 m 3 (Figure 2). The
flume is open at the base and totally enclosed at the top. Below the top lid, there are four
square equidistant paddles embedded in a second, rotating lid. The flume, a batteries pack
housing and an electronics housing are mounted on a tubular frame (Figure 2), which is
placed on the sediments. This frame provides stability for the flume, while four springs
maintain a downward pressure on the flume (10 kg force) once deployed, thus allowing
the flume to encapsulate the local bed properties as well as those of the near bed waters.
A camera (Figure 2) can be installed on the frame, but during this study it only
functioned in 1999 and only for some stations. A Seapoint| Optical Backscatter Sensor
(OBS) monitors turbidity at a height of 12 cm above the base. At the same height,
opposite the OBS, a sample port was installed for purposes of sensor calibration.
Figure 2 - Photo of Miniflume on Its Frame and Schematic Diagram Showing Miniflume
in Cross Section.
MOREAU El" AL. ON IN SITU FLUME MEASUREMENTS
123
The raw data recorded during an experiment include motor voltage (mV) and OBS
response (mV). The relationship between the motor voltage (M), lid rotation (R), and
azimuthal current speed 10 cm above the bed (ulo) is: R = 1.22 -4 M (m.s-I), and ulo =
0.54 R (m.s-l). Due to the non-linear response of the motor for a given gear ratio, the
minimum speed of the lid is 0.16 m.s l . Before and after each deployment, the lid rotation
was verified. Determination of friction velocity (u,) (m.s 1) and bed shear stress rb from
the Miniflume are described in Amos et al. 2000.
Experimental Procedure
An identical four-phase experimental protocol was used over the three years of
deployment. The first phase incorporates a start delay. The depth of the deployment sites
varied between 47 and 192 m, and even if the flume was paid out very slowly to
minimize seabed disturbance on landing, a sediment cloud would be created on impact.
In order to obtain data which correspond as closely as possible to sediment erodability in
undisturbed conditions, a one hour delay was maintained before recording the OBS data.
This delay was determined using video recording. They show that after one hour the
cloud had completely disappeared. The second phase was a period of recording during a
still-water episode in order to determine the initial concentration in the flume without
applied shear stress and thus the offset of the OBS (Figure 3). The duration of this period
was 5 - 30 minutes. In the third phase (erosion), flow was increased in a series of eight
steps. The first four steps were of equal duration with times between 5 and 30 minutes.
The final tour steps were twice as long. The final lid speed was selected based on the
expected erosion characteristics of each site. Since the minimum lid speed is
0.16 m.s l,the first steps of erosion phase can present the same value for shear stress
applied on the bed (Figure 3). The last phase was a settling period of five steps, four with
approximately 0.16 m.s 1 lid speed and one with a 0 m.s -~. All steps had a duration
between 5 and 30 minutes. Between each settling step, the lid was accelerated to the
maximum speed for 1 to 3 minutes in order to resuspend settled material.
Determination of Erosion Rate (E) and Critical Shear Stress ( ~.)
Erosion rate (E) (kg.m-2.s -l) was evaluated as the rate of change of suspended mass
(S) in the flume. In order to determine the mass concentration of sediment in suspension
(S), the OBS sensor was calibrated in a laboratory with in situ sediment from each
deployment site because many characteristics of the sediments, such as color (level of
blackness), mineralogy (variation of refraction index of particles), shape of the particles,
percentage of fines compared to sand, can influence OBS response (Sutherland et al.
2000, Hatcher et al. 2000). Suspended sediment concentration was increased gradually
and, for each increase, samples were collected from the sample port. Each sample was
filtered through pre-weighed, 25 ram, glass fibre GFC| filters, dried for 48 hrs at 40~
and re-weighed. Each calibration curve consists of 5 to 7 points and the correlation
124
coefficient (r 2) was between 0.97 and 1. Knowing the concentration of sediment in

suspension (S) the erosion rate (E) can be determined from:
E=Am_
At
S,+~ - S , x V
At
a
(1)
where m is the eroded dry mass, V is the flume volume and a is the flume bed area. Study
of trends in the erosion rate through time (Amos et al. 1997), allows determination of the
erosion type (I or II) as defined by Mehta and Partheniades (1982). Type I erosion shows
an asymptotic decrease in erosion rate with time whereas Type II erosion shows
continuous erosion rate with time.
At the beginning of the erosion process, the particle size of suspended sediment is
so small that they are not easily detectable and determination of the threshold of erosion
is not possible, even with a sophisticated apparatus. An alternative method is the use of a
mathematical extrapolation. One of the calculation methods consists of applying a stress
on the bed until erosion ceases (Amos et al. 1997). It is supposed that erosion continues
until the critical shear stress of the new exposed surface exceeds the stress induced by the
moving fluid. The speed of the liquid is then increased in a series of steps. A regression
of sediment concentration (OBS response) with bed shear stress is applied. Logarithmic
plots are used because they provide a better fit for the lower levels of erosion (Amos et al.
1997). The intercept of the straight line regression with the offset of the OBS during the
still-water episode, or the fit of the regression line for a series of steps gives the critical
shear stress (re) (Amos et al. 1994).
Site Deployment
Over three years, 50 in situ deployments were made at 14 different stations: 6
stations in the North Arm and 8 in Baie des Ha! Ha! (Figure 1). Nine deployments gave
unusable results due to a bad connection between the OBS, motor and computer or open
communication between water in the flume and the ambient water. Forty one
deployments gave usable results. These correspond to results, which allowed the
determination of an erosion threshold. Twenty six present a concentration plateau during
each erosion step. For 6 stations, ST19, ST34, ST21, ST41, ST47, ST44, we obtained
usable results for the three years. For several stations several deployments were carried
out during the same year. Two usable results were obtained at stations ST19, ST34, ST36
in 2000 and 2001, at stations ST21, ST40, ST37, HZ in 2001, at station ST41 in 1999 and
three at station ST47 in 2001 (Table 1). Typically, a station corresponds to a circular area
with a diameter between 45 and 416 m except for ST21. The deployment at ST21 in 2000
is about 1600 m from deployments in 1999 and 2001 due to a positioning error. For the
other stations, the difference between deployment locations is related to the difficulty in
deploying the Miniflume in deep water.
ST29
STI9
STI9
ST20
ST26
ST34
ST34
ST21
ST21
MFI
MFI
ST36
ST36
ST36
MF2
ST41
ST41
ST47
ST47
ST47
HZ
HZ
ST44
ST40
ST40
ST37
ST37
ST37
station
408.9-1820.7 1.44E-04
455.6-1820.7 1.61E-04
423.9-1637.3
296.7-1637.3
218.5-1637.3
387-2146.7
202.2-1802.2
180-2041.9
187.1-2066.6
328.1-2066.6
82.7-1709.9 2.65E-04
44.1-1709.9 8.57E-05
0.8
1.09
0.88
1.15
0.8
0.65
0.61
1.09
1.15
1.42
0.88
1.09
1.01
1.09
0.19 1.15
0.07
0.16
0.14
0.42
0.17
0.22
0.22
0.23
1.42
1.15
0.8
1.41
1.15
1.42
1.15
1.42
2.59
0.14 2.59
0.21 1.42
85.7-1894.5
1.09
0.8
0.65
0.61
0.48 1.42
0.26 0.8
0.20 1.15
0.27 0.8
2.59
0.19 2.05
1.42
0.30 1.42
0.25 1.15
2.18E-04
2.06E-04
1.23E-04
8.36E-05
1.54E-04
1.12E-04
1.61E-04
3.12E-04
32.1-1388.9 9.20E-05
1.49E-04
2.40E-04
7.28E-05
1.13E-04
1.72E-04
378.1-2220.5
193-1327.2
616-1538.4
334.1-1538.4
0.91
1.15
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
1.15
1.15
1.15
1.15
0.09 1.15
0.12 1.15
1.15
1.15
1.15
0.16 1.15
0.15 1.15
0.29 1.15
1.15
1.15
1.15
0.33 1.15
0.31 1.15
0.35 1.15
1.42
0.8
1.15
1.09
1.15
0.34 1.42
0.14 1.42
0.23 1.15
0.22 1.09
0.29 1.15
Em~,~
type "t~ Xn,,~ real Xm~
(kg.m2.s l)
(Pa) (Pa) (Pa)
191.6-2237.6 5.13E-05
2001
conc
(mg.I t)
2.59
1.15
0.1
"c,: "cm~ real Xm~

(Pa) (Pa) (Pa)
1.65E-05
157.2-907.4 2.77E-05
161-2066.6 9.90E-05
174.2-870.4 3.58E-05
102.2-1562.9 7.35E-05
40.5-1388.9 6.76E-05
157.4-245.8 2.42E-06
157.8-253.7 5.02E-06
102.3-394.3
3.4-1004.9 5.47E-05
20.2-2220.5 8.55E-05
194.7-1529.1 8.29E-05
B
A
A
A
B
B
B
A
A
0.15 2.22
1.42
0.8
0.8
1.8
0.38
0.38
0.13
0.31
1.42
0.38
0~87
1.09
0.8
0.8
0.7
1.42
154.3-2041.9
0.9-1388.9
10.6-279.4
333.9-1538.4
1.68E-04
1.78E-04
1.37E-05
1574.9-2220.5 1.15E-04
A
A
B
A/B
1999
Em~.~ type X~ Xm:~, real '~m,,,~
cone
Em~,
type
(kg.m".s ~)
(Pa) (Pa)
(Pa)
(mg.l")
(kg.m".s1)
0.42 0.8
0.8
354-639.1
4.88E-05 A
102.3-298.9 5.05E-06 B 0.35 2.22
1.7
99.2-2237.6 3.42E-04 A
102.8-135.5 6.17E-07 B
0.24 0.8
0.8
181-5-774.8 7.70E-05 A
2000
eonc
(mg.l")
T a b l e i - Critical Shear Stress ( ~.) (Pa), Maximum Bed Shear Stress ( ~,~) (Pa) Applied during each Miniflume Deployment.
Real g,~a. Correspond to the Last Bed Shear Stress using for Determined ( ~.), After Real r,,,u~xthere was Saturation (Pa). E,,~,,,
is the Mean Erosion Rate for all the Erosion Phases (kg.m~sl ), conc is the Minimum and Maximum Concentration Recorded
during Erosion Phases, Type is the SSC Response Type.
ol
...L
m
Z
t~
m
rrl
(2;
.rIS)
Z
O'J
C
m
--4
126
Results
Suspended Sediment Concentration Profiles

For each profile, the time dependent erosion response is characterized by a
maximum erosion rate immediately after application of a higher shear stress (Piedra
Cueva and Mory 2001). The OBS response, i.e. the suspended sediment concentration
(SSC), following an acceleration step can be divided into two groups: most of the profiles
(Type A) show a rapid increase of SSC followed by a concentration plateau or a slower
increase until the next step (Figure 3 and Table 1). Ten profiles (Type B) show an
increase at the beginning of the step of the SSC followed by a decrease while the rotation
speed was held constant (Figure 3 and Table 1). Nine of the ten Type B profiles were
obtained in 2000 (STI9, ST34, MF1, ST36, ST47, ST44, ST37) and one in 1999 (ST47).
1.2
0.6
.--.
0.0
100o
500
o
~ 2E-3
~." IE-3
0E+0
t~ -1E-3
0
50
100 150
time (min)
200
250 0
50
100 150
time (rain)
200
250
Figure 3 - Bed Shear Stress ('rb, Pa) and Suspended Sediment Concentration (SSC, mg.lt )
(Pa) Obtained using Laboratory OBS Calibration and Erosion Rate (kg.m2.s~) (Equation
1)for an Example of Type A SSC Profile (ST41, 2000) (left) and Type B SSC Profile
(ST44, 2000) (right).
The same station may show different SSC profiles, Type A or Type B, between
deployments in the same year (Table 1: ST34 in 2000; Figure 4a) or from one year to the
next (Table 1: ST 19, MF1, ST36, ST47, ST44; Figure 4d). For deployments at the same
station in the same year, with identical bed shear stress profiles, the suspended sediment
concentration response is quantitatively similar (Figure 4c) with the exception of
experiments at station ST34 in 2000 and at station ST41 in 1999. At station ST34 (2000),
the same bed shear stress profile was applied but the duration of steps was doubled. The
concentration obtained for the longer steps were significantly greater (Figure 4a). At
station ST41 (1999), a much higher concentration increase was observed during the first
step for one of the experiments, presumably due to the presence of a very loose layer at
the seabed (Figure 4b). From one year to the next, however, the suspended sediment
127
concentration at the same station may differ by an order of magnitude, as is the case with
ST19, M F I , ST36 (Figure 4d), ST47 and ST44 (Table 1). Concentrations obtained for
Type B response are always lower than the concentration for Type A (Table 1).
1
0.0
2000 -
[7- -~-~- -
_ i
if- - --
\
I '
0
].6
..
_ ~c)
'
100
i
'
/ - - -
200
~
08
0.0
'
300 0
50
100 150
I ,~,, I ' I '
~u/
l
'~
2000 0
'
100
200
time (rain)
250
200
250
'- '- ' ~ ~-~--~,-,

I
200
'
'
f-
'
300 o
50
100 150
time (min)
Figure 4 - Bed Shear Stress ( z'b, Pa) and Suspended Sediment Concentration (SSC, mg.l/ )
Obtained for Different Stations: (a) Station ST34 for Two Deployments in 2000 with
Duration of Steps Doubled Between the Two. Dashed Line Represent the Second
Deployment [ST34-00(2)]; (b) ST41 for Two Deployments in 1999, Dashed Line
Represent the Second Deployment [ST41-99(2)]; (c) Station ST34 for Two Deployments
in 2001;for the Same Bed Shear Stress, Concentrations are Similar; (d) ST36 in 2000
(Solid Line) and in 2001 (Dashed Line) (ST36-01(3)). Type B Shows Concentration Ten
Times Smaller than Type A.
Erosion Rate
For all the stations, the erosion rate shows an asymptotic decrease with time for each
step (Figure 3) corresponding to Type I erosion as defined by Mehta and Partheniades
(1982). The peak erosion rate (E,,u=) and mean erosion rate (Em) were evaluated for each
acceleration step. The peak erosion rate is the m a x i m u m value within any increment of
flow (Amos et al. 1992, 1996, Widdows et al. 1998, Houwing 1999). It varies between
1.02x 10-5 and 2.55x 10 "3 kg.m'2.s 1. In contrast to Amos et al. (1996), no significant
relationship was found between E,,~x and the applied bed shear stress. Em appeared to
show a relationship with the applied current speed (Uto) taking the exponential form:
128
E,, = 6 . 3 1 x i 0 -7. 10 (1147 U~o), r 2 = 0.63, n = 101
(2)
Considering the diversity of sites, the correlation is notable. Amos et al. (1998)judged
the correlation for a mudflat in the Humber estuary with a r2 = 0.55 and n = 100 to be
noteworthy. The mean erosion rate for the total erosion phase (Emean) varies between
3.42x10 -4 and 6.17x 107 kg.m-2.s t and is smaller for Type B response (Table 1). Stations
in the North Arm always show a Em~,, higher than 2x10 5 kg.m-~.s-I. Figure 5 shows a
plot of E,~,m and (r~.) as a function of along-axis distance, measured from La Baie city in
Baie des Ha! Ha! and from station ST37 in the North Arm. The graph shows no relation
between E , , ~ and distance. Similarly, there is no temporal trend in erosion rate (Table 1,
Figure 5).
0.6
[]
0
0.4
[]
t..,~
o
~
0.2
[]
IE-30 I
L3
--
[]
[]
q
[]
c,
r~ :_
[]
~b IE-5
F
,3
9
m"~ IE-6
'
IE-7
4000 8000 !2000

o
0
4000 8 0 0 0 12000
Distance from
Distance fi'om the upstream
La Baie (m)
station in North Arm (m)
Figure 5 - Erosion Threshold and Mean Erosion Rate Along Axis Distance are Measured
from La Baie city in Baie des Ha t Ha! and from the Upstream Station in the North Arm
(Station ST37-01(3)). Crosses Represent 1999 Data, Circles 2000 Data and Squares
2001 Data.
Critical Shear Stress
For each usable result, the critical shear stress (rc) (Pa) and bed shear stress of the
last step before saturation (real r,,~) were determined (Table 1). The real r,,~ varies
between 0.61 and 1.7 Pa and ~. between 0.07 and 0.48 Pa. The average value of rc is 0.24
Pa with a standard deviation of 0.10. The value of ~'cfor stations ST21, M F I , ST47 is
greater when they show a Type B response (Table 1) and is of the same order for Type A
MOREAU ET AL. ON IN SlTU FLUME MEASUREMENTS
129
and B responses for station ST34, ST36 and ST44. For the same station, ~. values in Baie
des Ha! Ha! are more variable than in the North Arm. As for erosion rate, there is no
temporal trend in critical shear stress (~), the spatial variability is important on small and
large scale (Figure 5).
Erosion law
The time-rate of increase of suspended sediment mass per unit bed area, m, is given
in the functional form:
dm
dt
= f ( r h - r ,, v,, v2,...v,)
(3)
where rb is the bed shear stress, ~ is the bed shear strength, (rb- r~) is the excess bed shear
stress, a n d , v~, v2..... v, are parameters which define erosion resistance (Mehta 1989).
The bed shear strength may be used interchangeably with the critical shear stress re. From
experiments on unconsolidated cohesive sediments in stratified deposits, Parchure and
Mehta (1985) determined an exponential expression based on excess shear stress:
ln-ff--E=ot('rb
E/
r ),/2
--
(4)
s -
where Ef is defined as the floc erosion rate (kg.m-2.sl), tr is an empirical coefficient

(m.N-V'), ~ or ~'ct:.)is the bed shear strength (Pa), and E is the erosion rate over ten minutes
for each acceleration step. The value of At used in (1) is always equal to 10 minutes in
order to homogenize the results (Hydraulics Research Station 1980, Amos et al. 1996). In
equation (7) r h - ~'~ is determined from concentration profiles that showed a plateau for
each erosion step. A concentration plateau indicates that there is no more erosion and that
the bed shear stress equals the critical shear stress. Twenty six deployments demonstrate
this behavior (Table 2).
The ocvalues from the twenty six tests vary between 3.92 and 23.92 m.N v2, and the
floe erosion rate between 3.74 .9 and 1.56 -5 kg.m2.s ~ (Figure 6). Floc erosion rate
and ocare connected by an exponential relation (5):
E / = 1.10 I0 -~ exp[--0.43 Ct], r 2 = 0.92, n = 26
(5)
130
Table 2 - Values of Erosion Threshold ~. (Pa), Floc Erosion Rate Ef (kg.me.s -t ) and
Empirical Coefficient tz (m.N" ,/2) for Stations
.
. Concentration Plateau for Erosion
havmg
Steps. -99 Results obtained in 1999, -00 in 2000 and-01 in 2001. (1) Results obtained
for the First Test, (2) the Second and (3) the Third at the "Same" Station.
Station
r,.
a
El
Station
rc
ot
Ef
ST19-99
STI9-1
ST20-99
ST34-99
ST34(1)-01
ST34(2)-01
ST21-00
ST21(2)-01
MF1-01
ST36(2)-01
ST36(3)-01
MF2-00
ST41(2)-99
0.35
0.1
0.24
0.38
0.48
0.26
0.23
0.27
0.19
0.30
0.25
0.16
0.13
3.92 8.77E-06
5.04 1.28E-05
12.72 5.41E-07
12.06 2.04E-07
7.42 7.81E-06
5.45 1.56E-05
6.19 6.56E-06
9.35 8.25E-06
12.08 4.36E-07
5.65 1.10E-05
8.98 4.57E-06
12.60 3.94E-07
10.83 4.24E-07
ST41-00 0.15
ST41-01 0.19
ST47(2)-01 0.16
ST47(3)-01 0.14
ST44-99 0.15
ST44-01 0.22
HZ(I)-01 0.42
HZ(2)-O1 0.17
ST40-00 0.12
ST40(1)-01 0.22
ST40(2)-01 0.23
ST37(2)-01 0.14
ST37(3)-01 0.21
11.99 4.67E-07
12.55 3.88E-07
16.85 2.68E-07
12.92 1.06E-06
11.97 5.88E-07
12.25 3.14E-07
7.97 4.35E-06
13.07 4.58E-07
23.62 4.45E-09
23.92 3.74E-09
13.82 3.64E-07
7.72 4.79E-06
14.78 1.31E-07
The erosion law determined by Mehta et Parchure 1985 for stratified sediment fits quite
well for each set of results, but a general law would have a determination coefficient (r2)
equal to 0.58:
E=4.98.10-6
r'=0.58,
n=101
(6)
Discussion
Type B Response
Similar Type B behavior has been observed in other in situ studies in lower
Chesapeake Bay and Baltimore Harbor (Maa et al. 1993, 1998). These investigators
suggested several possible explanations: (1) a leakage from the flume; (2) decrease of the
skin friction shear stress due to suspension of surface sediment; (3) greater annulus
acceleration at the beginning of a step resulting in a larger initial bed shear stress than can
be maintained during the rest of the step, resulting in redeposition of the larger sediment
particles. Maa et al. 1998 concluded that leakage was the predominant factor in their
experiments, caused by a dynamic pressure difference and imperfect sealing between the
rotating ring and the two side walls. In the present study, Type B profiles are always
associated with lower SSC (Figure 3, Figure 4d and Table 1) suggesting that a decrease of
the skin friction shear stress due to the suspension of surface sediment is not a significant
factor. Leakage from the Miniflume, while possible given the often irregular meso-scale
morphology of the Fjord bottom, is considered unlikely. The VIMS Sea Carousel used by
131
Maa et al. (1993, 1998) is much larger than Miniflume, and probably allows more
leakage. It is more likely that mechanical limitations of the motor, resulting in a larger
initial bed shear stress (hypothesis 3) explains the Type B occurrences. A small decrease
in concentration during the first step in Type A profiles, when annulus acceleration is
maximal, supports this interpretation.
Erosion Rate and Critical Shear Stress
Two different interpretations of erosion Types have traditionally been suggested.

Type I erosion (depth-limited) is manifested as a decrease in the erosion rate as a function
of depth while Type II erosion (unlimited) occurs as a continuous function of time.
Conceptually, Type I erosion has been viewed as a surface phenomenon caused by
entrainment of small aggregates and flocs (Amos et al. 1992, 1997). Type II erosion is
characterized by failure along thin planes of weakness within the sediments, probably
related to the sediment microfabric generated after deposition (Amos et al. 1992),
enlargement of surface irregularities, and subsequent undercutting and release of large
aggregates. Type II erosion is typically associated with erosion of deeper sediment.
However, Sanford and Maa (2001) suggest a formulation that relates the two types. If the
time scale of the shear stress step is long compared to the time scale for depletion of
erodible sediment, then the erosion will appear to be Type I (Parchure et Mehta 1985,
Maa et al. 1998, Piedra Cueva et Mory 2001). Otherwise, it will appear to be Type II
(Gust et Morris 1989, Tolhurst et al. 2000). The same sediment bed can exhibit Type I
erosion or Type II erosion, or some intermediate state (Amos et al. 1996) depending on
the nature of the forcing (Sanford and Maa 2001).
Since all the experiments in the present study show Type I erosion whatever bed
shear stress is applied, it is likely that the time scale for depletion of erodible sediment is
shorter than the time scale of the erosion step of the experiment and the bed shear stress
applied was only sufficient to cause shallow erosion.
E,,~,,~ and ~. are variable, both in time for single stations and between different
stations; there is no obvious trend. Variability is more noticeable in Baie des Ha! Ha! but
probably only because more measurements were obtained in this area (Figure 5). To the
authors' knowledge, this is the deepest water in which in situ experiments have been
made. The quality of the in situ flume measurements at these depths is more difficult to
establish than for shallower water. Nevertheless, for the experiments when the video
camera was recording, it was observed that the frame maintained a stable base for the
flume and that there was no communication between the flume and the ambient water.
While, in spite of all the precautions, some results were still unusable and may be
attributed to equipment problems, it is more likely that the natural heterogeneity of
surficial sediment in Saguenay Fjord is a significant factor in the variability. Locatet al.
(2000) defined several geomorphologic units in Saguenay Fjord, which contributes to
highly variable meso-scale topography (Hill et al. 1999). Heterogeneity in bioturbation of
the recently deposited bed (Maurice 2000) also strongly influences surficial roughness
resulting in local heterogeneity of the physical properties (water content, particle size,
reflectivity) of surficial sediment (Schmitt 2001).
132
Prediction of Suspended Sediment Concentration in Miniflume
Comparison with previous studies - The o~and Ey values determined for the different
sites in Saguenay Fjord have a large range (Table 2, Figure 6) but they have been
determined over a small range of square root of excess shear stress (0.14 to 0.67 Pa)
(Figure 7). In order to compare these results with previous investigations, published data
from both in situ and laboratory studies were used (Table 3).
Table 3 - Values of Floc Erosion Rate Ey (kg. m2. s-l) and Empirical Coefficient ot (m. Ar '/:)
for Previous Studies.

Investigators
Bed
Partheniades (1965)
bay mud
Lee (1979)
lake mud
- Thorn and Parsons (1977)
estuarial mud
Thorn and Parsons (1979)
estuarial mud
Dixit (1982)
kaolinite
27 Parchure et Mehta (1985)
kaolinite+ tap water
28 Parchure et Mehta (1985)
kaolinite+ salt water (35 ppt)
29 Parchure et Mehta (1985) lake mud + varying salinity water
33 Caywood (1999)
kaolinite
34 Caywood (1999)
Chalk bluff
37 Caywood (1999)
mix
38 Ca~,wood(1999)
sand mix
-
--
1E-3
'
'
'
-7- ~ T - -
"
~
1E-5 -
1E-7_
IE-9
;4-+.~-+~ o o
~ ,~
o
~ -~ + +
~
~
#
' I ' I ' I ' I ' I '/
0
5
10 15 20 25 30
O
9
ID
~
o
~
,,/
E!
8.3
8.3
8.3
4.2
25.6
18.4
17.2
13.6
2.48
2.07
2.01
2.34
6.67E-08
7.00E-07
7.00E-07
3.10E-06
1.00E-06
8.33E-07
2.33E-06
5.33E-06
2.36E-04
8.86E-05
4.58E-05
2.18E-05
Partheniades(1965)
]
/
Thorn et Parsons(1977) /
/
Thorn and Parsons(1979)]
Lee (1979)
/
Dixit(1982)
/
ParehureetMehta(1985) /
Amosetal.(1992)
Dennett (1995)
Caywood (1999)
+ Saguenay (this study)
Figure 6 - Floc Erosion Rate (Ef, kg.m2.s -1) Versus o~(m.N"/2)for This Study (Crosses)
and Previous Studies.

The ~ and Ef values from previous investigators correspond to sediment deposited in
shallower water. Parchure and Mehta 1985 explained the differences between a'values
for the different authors in terms of a n u m b e r of factors, including the bed preparation
procedure, the characteristics of the apparatus used, the influence of temperature, the
effect of relatively high sediment concentrations on bed shear stress, the influence of
sediment type and the errors in estimating the magnitude of bed shear stress. The range of
133
values obtained with Miniflume suggests that sediment characteristics (bioturbation, grain
size...) are most important. Since, during this in situ study the same apparatus was used,
there was no bed preparation, all stations presented the same temperature characteristics,
relatively high sediment concentrations had no major effect on bed shear stress. Indeed,
the algorithm of Amos et al. (1997) shows that drag reduction increases with both
increasing SSC and applied shear stress and predicts a reduction in the friction velocity
(u.) of 1.24 % related to the maximum SSC (2.24 g.l l ) and shear stress (1.7 Pa).
Therefore, during the experiments drag reduction would not be significant.
The erosion rate for surficial sediments in Saguenay Fjord varies with the square
root of excess shear stress within the same range as previous studies (Figure 7).
1.E+00
I.E-01
&=====_....,...
1.E-02
31
_..--_-__"--- 32
.
~. I.E-03
~ 1.E-04 29 ~
t~ I.E-05 28 ;:
27 "
6
30
38
1.E-06
1.E-07
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Figure 7 - Erosion Rate (E) versus Square Root of the Excess Stress Determined for
Twenty Six Deployments in Saguenay Fjord (Continuous Lines; Table 2) and for
Previous Studies (Dashed Lines; Table 3). * Data from Amos et al. 1996 (adapted from
Dennet 1995 and Caywood 1999).
Conclusions
The variability of the physical, chemical and biological properties of the area led to
prefer an in situ study rather than a laboratory study in spite of the great depth of the
Fjord. For the first time at such depths, a circular benthic flume, Miniflume, was used to
measure the erodability of cohesive sediments. Typically, a station corresponds to a
circular area within a diameter between 45 and 416 m.
Over three years, 41 deployments gave results which allowed the determination of
an erosion threshold (~). The values of rc ranged between 0.07 and 0.48 Pa.
Two types of SSC profiles were obtained. Type A, the most common profile, shows
a concentration plateau for each acceleration step. Type B shows an increase at the
beginning of the step followed by a decrease as the lid speed was held constant. Type B
profiles were interpreted to be related to variations in lid rotation during the erosion step.
134
Excess lid acceleration at the beginning of a step appears to suspend larger particles,
which redeposit during the latter period of the step.
For all the stations, the erosion rate shows an asymptotic decrease with time for
each step corresponding to Type I erosion. The time scale for depletion of erodible
sediment is shorter than the time scale of the erosion step and the bed shear stress applied
was only sufficient to cause shallow erosion.
Mean erosion rates were between 3.42 10.4 and 6.17x 10.7 kg.m-2.s 1 and were
smaller for Type B profiles.
The erosion law determined by Mehta et Parchure 1985 for stratified sediment fits
quite well for each Type A profile, but a general law would have a determination
coefficient (r2) equal to 0.58. Erosion rate versus square root of the excess stress for
surficial sediments in Saguenay Fjord varies in the same range as previous studies.
Variability of Erosion thresholds, parameters o~and Ef in erosion laws are partly
related to the experimental conditions but mostly to the diversity of benthic sediment in
Saguenay Fjord. No significant temporal variations were noticeable.
If surficial sediments contain contaminants diffused upwards from underlying
sediments, they might be re-mobilized if the bottom current speed were sufficient.
Erodability of surficial sediments and their associated pollutants in Saguenay Fjord is
therefore of concern. Determination of benthic hydrodynamic conditions will be the
second step in defining the erosion susceptibility hazard in Saguenay Fjord.
Acknowledgments
The authors thank the Geological Survey of Canada (Atlantic) for loan of the
Miniflume and C. L. Amos for taking the time to demonstrate its use on the R/V Alcide
C. Horth and during a cruise in the Woods Hole region. Financial support was provided
by the Natural Sciences and Engineering Research Council of Canada, and Alcan of
Canada Ltd. We greatly appreciate the assistance of the captain and crew of the R/V
Alcide C. Horth. Geological Survey of Canada Contribution No. 2002110.
References
Amos, C. L., Daborn, G. R., Christian, H. A., Atkinson, A., and Robertson, A., 1992, "In
situ erosion measurements on fine-grained sediments from the Bay of Fundy,"
Marine geology, Vol. 108, pp. 175-196.
Amos, C. L., and Gibson, J., 1994, "The stability of dredge material at Dumpsite B,
Miramichi Bay, New Brunswick, Canada," Geological Survey of Canada Open
File Report, N ~ 3020.
Amos, C. L., Sutherland, T. F., and Zevenhuizen, J., 1996, "The stability of sublittoral,
fine-grained sediments in a subarctic estuary," Sedimentology, Vol. 43, pp 1-19.
135
Amos, C. L., Feeney, T., Sutherland, T. F., and Luternauer, J. L., 1997, "The stability of
fine-grained sediments from the Fraser River Delta," Estuarine, Coastal and Shelf
Science, Vol.45, No.5, pp.507-524.
Amos, C. L., Brylinsky, M., Sutherland, T. F., O'Brien, D., Lee, S., and Cramp, A., 1998,
"The stability of a mudflat in the Humber estuary, South Yorkshire, UK,"
Sedimentary. Processes in the Intertidal zone. Geological Society, London, Special
Publications, Vol. 139, pp. 25-43.
Amos, C. L., Sutherland, T. F., Cloutier, D., and Patterson, S., 2000, "Corrasion of a
remoulded cohesive bed by saltating littorinid shells," Continental shelf Research,
Vol. 20, pp. 1291-1315.
Ariathurai, R., and Arulanandan, K., 1978, "Erosion rates of cohesive soils," Journal of
Hydraulics Division, ASCE, Vol. 104, No. HY2, pp. 279-283.
Caywood, C. L., 1999, "Determination of erosion rates of cohesive sediments in the
Carson and Truckee Rivers," Master of Science, University of Nevada, Reno, 135
pp.
CrEmer, J-F., Long, B., Desrosiers, G., de Montety, L., and Locat, J., 2002 "Application
de la scanographie b_l'6tude de la densit6 des s6diments et ~ la caractErisation des
structures s6dimentaires: exemple des s6diments d6pos6s dans la rivi~re
Saguenay, Qu6bec, Canada) apr~s la crue de juillet 1996," Journal Canadien de
G~otechnique, Vol. 39, pp. 440-450.
Dennet, K. E., 1995, "Flume studies of the erosion of cohesive sediments," Thesis,
Georgia Institute of Technology, 238 pp.
Dixit, J. G., 1982, "Resuspension potential of deposited kaolinite beds," Master of
Science, University of Florida, Gainesville.
Drainville, G., 1967, "l~tude oc6anographique et ichtyologique du fjord du Saguenay,"
M~moire de maftrise, UniversitCdu Qudbec h Montreal, 166 pp.
Fukuda, M. K., and Lick, W., 1980, "The entrainment of cohesive sediments in
freshwater," Journal of Geophysical Research, Vol. 85, No. C10, pp. 2813-2824.
Gust, G., and Morris, M. J., 1989, "Erosion thresholds and entrainment rates of
undisturbed in situ sediments," Journal of coastal Research, Vol. 5, special issue,
pp. 87-100.
Hatcher, A., Hill, P., Grant, J., and Macpherson, P., 2000, "Spectral optical backscatter of
sand in suspension: effects of particle size, composition and color," Marine
Geology, Vol. 168, pp. 115-128.
136
Hill, P. R., C6t6, P., Locat, J., and the students of the UQAR Mission Stage, 1999,
"Evidence for gravity-flow processes as the mechanism for basin-sedimentation in
Baie des Ha! Ha!, Saguenay fjord, from sidescan sonar data," Abstract, Canadian
Meteorological and Oceanographic Society, Annual Meeting, Montreal, May 31 June 1, 1999.
Houwing, E-J., 1999, "Determination of the critical erosion threshold of cohesive
sediments on intertidal mudflats along the Dutch Wadden sea coast," Estuarine,
Coastal and Shelf Science, Vol. 49, pp. 545-555.
Hydraulics Research Station, 1980, "River Scheldt surge barrier. Study of estuarine
sediments," Internal Report, No. EX928.
Kusuda, T., Umita, T., Koga, K., Futawatari. T., and Awaya, Y., 1984, "Erosional
process of cohesive sediments," Water Science and Technology, Vol. 17, pp. 891901.
Lavelle, J. W., Mofjeld, H. O., and Baker, E. T., 1984, "An in situ erosion rate for a finegrained marine sediment," Journal of Geophysical Research, Vol. 89, No. C4, pp.
6543-6552.
Lee, D. Y., 1979, "Resuspension and deposition of lake Erie sediments," Master of
Science, Case Western Reserve University, Cleveland, Ohio.

Lee, S. C., Mehta, A. J., and Parchure, T. M., 1994, "Cohesive sediment erosion: Part ITest devices and field instrument assemblies; Part II- Relationship between the
erosion rate constant and bed shear strength," UFL/COEL/MP-94/02, Coastal and
Oceanographic engineering, University of Florida.
Locat, J., Urgeles, R., Schmitt, T., Martin, F. and Hoareau L., 2000, "The morphological
signature of natural disasters in the upper Saguenay fjord area, Qu6bec, Canada,"
53rd Canadian Geotechnical Conference, Vol. 1, pp. 109-I 16.
Maa, J. P. Y., Wright, L D., Lee, C-H., and Shannon, T. W , 1993, "VIMS sea carousel :
a field instrument for studying sediment transport," Marine Geology, Vol. 115,
pp. 271-287.
Maa, J. P. Y., Sanford, L., and Halka, J. P., 1998, "Sediment resuspension characteristics
in Baltimore Harbour, Maryland," Marine Geology, Vol. 146, pp. 137-145.
Maurice, F., 2000, "Caract6ristiques gEotechniques et 6volution de la couche de
s6diments d6pos6e lors du d61uge de 1996 dans la Baie des Ha! Ha! (fjord du
Saguenay, Qu6bec), " M~moire de maftrise, d~partement de g~ologie et gdnie
gdologique, UniversitdLavaL
Mehta, A. J., 1989, "On estuarine cohesive sediment suspension behavior," Journal of
Geophysical Research, Vol. 94, pp. 14306-14314.
137
Mehta, A. J., and Partheniades, E., 1982, "Resuspension of deposited cohesive sediment
beds," Eighteenth Conference Coastal Engineering, pp. 1569-1588.
Mehta, A. J., Parchure, T. M., Dixit, J. G., and Ariathurai, R., 1982, "Resuspension
potential of flocculated fine sediments," Estuarine Comparisons, V.S. Kennedy,
Academic Press, New York.
Parchure, T. M., and Mehta, A. J., 1985, "Erosion of soft cohesive sediment deposits,"
Journal of Hydraulics Engineering, ASCE, Vol. 111, No 10, pp1308-1326.
Partheniades, E., 1965, "Erosion of cohesive sediments," Journal of Hydraulics division,
ASCE, Vol. 91, No. HY 1, pp. 105-139.
Pelletier, E., Deflandre, B., Nozais, C., Tita, G., Desrosiers, G., GagnE, J-P., et Mucci, A.,
1999, "Crue 6clair de juillet 1996 dans la region du Saguenay (QuEbec). 2.
Impacts sur les sediments et le biote de la baie des Ha! Ha! et du fjord du
Saguenay," Canadian Journal of Fisheries and Aquatic Sciences, Vol. 56, pp.
2136-2147.
Piedra Cueva, I., and Mory, M., 2001, "Erosion of a deposited layer of cohesive
sediments," MCAnally, W.H. Mehta, Coastal and estuarine fine transport
processes, Elsevier, Amsterdam, pp. 41-52.
Ravens, T. M., and Gschwend, P. M., 1999, "Flume measurements of sediment
erodibility in Boston harbour," Journal of Hydraulic Engineering, Vol. 125, Nu.
I0, pp. 998-1005.
Sanford, L. P., and Maa, J. P. Y., 2001, "A unified formulation for fine sediments,"
Marine Geology, Vol. 179, pp. 9-23.
Schafer, C. T., Smith, J. N., and C6tE, R., 1990, "The Saguenay Fjord : A major tributary
to the St. Lawrence Estuary," Coastal and Estuarine studies, Vol. 39, pp. 378419.
Schmitt, T., 2001, "CaractEristique acoustique et Evolution des sediments superficiels de
la couche de 1996 dans la Baie des Ha! Ha! et le Bras Nord, fjord du Saguenay,
QuEbec," M~moire de maitrise, dgpartement de g~ologie et gdnie gdologique,
Universitd Laval.
Sutherland, T. F., Lane, P. M., Amos, C. L., and Downing, J., 2000, "The calibration of
optical backscatter sensors for suspended sediment of varying darkness levels,"
Marine Geology, Vol. 162, pp. 587-597.
Thorn, M. F. C., and Parsons, J. G., 1977, "Properties of Grangemouth Mud," Report No.
EX 781, Hydraulics Research Station, Wallingford, England.
138
Thorn, M. F. C., and Parsons, J. G., 1979, "Properties of Belawan Mud," Report No. EX
880, Hydraulics Research Station, Wallingford, England.
Tolhurst, T. J., Black, K. S., Paterson, D. M., Mitchener, H. J., Termaat, G. R., and
Shayler, S. A., 2000, "A comparison and measurement standardization of four in
situ devices for determining the erosion shear stress of intertidal sediments,"
Continental Shelf Research, Vol. 20, No. 10-11, pp. 1397-14 ! 8.
Villaret, C., 1987, "l~tude exp6rimentale et num6rique des lois d'6rosion pour des
s6diments coh6sifs," th~se, Universitd scientifique, technologique et m3dicale de
Grenoble.
Widdows, J., Brinsley, M. D., Bowley, N., and Barrett, C., 1998, "A benthic annular
flume for in situ measurements of suspension feeding/biodeposition rates and
erosion potential of intertidal cohesive sediments," Estuarine, Coastal and Shelf
Science, Vol. 46, pp. 27-38.
Young, R. N., and Southard, J. B., 1978, "Erosion of fine-grained marine sediments : seafloor and laboratory experiments," Geological Society of American Bulletin, Vol.
89, pp. 663-672.
Section I1: Mitigation and Restoration

Methods
Thiam Soon Tan, i Godakuru P Karunaratne, i Victor Choa, z and Myint Win Bo 3
Reclamation Using Waste Sediment by Sand Capping

Technique
Reference: Tan, T. S., Karunaratne, G. P., Choa, V. and Bo M. W., "Reclamation
Using Waste Sediment by Sand Capping Technique," Contaminated Sediments:
Characterization, Evaluation, Mitigation/Restoration, and Management Strategy
Performance, ASTM STP 1442, J. Locat, R. Galvez-Cloutier, R.C. Chaney, and K. R.
Abstract: Disposal of unwanted seabed sediment arising from dredging activities is an

increasingly intractable problem. One solution is to use the sediment as fill material for
land reclamation, as a replacement for sand, the traditional fill. As the water content of
sediment is high, even a small fill will ensure rapid consolidation. The underlying
philosophy is to place a layer of sand on top of the slurry to accelerate consolidation.
This paper summarizes the research carried out in Singapore. A pilot test and a largescale project will he described and the results from monitoring discussed. A physical
framework will then be provided to explain the sand capping of the sediment, followed
by the introduction of a simplified model to connect the two key aspects of the mechanics
of sand capping, namely the accumulation of sand on the sediment and the way the slurry
gained shear strength to resist this load. This simple model was able to produce results
consistent with that observed in the field.
Keywords: slurry, sedimentation and consolidation, yield shear stress, sand capping
Introduction
Singapore was a very small country of area 580 square kilometers at the time of its
independence in 1965. To alleviate this highly restrictive condition, land reclamation
was a top priority national effort in the last 37 years, As indicated in Figure 1, which
shows the increase in land area since 1965, the city-state has grown to an estimated 680
square kilometers today, an increase of 17%, and still growing by the day. Ongoing and
already approved projects will add at least another 50 square kilometers in the next ten
l Associate Professor, Centre for Soft Ground Engineering, National University of

Singapore, Kent Ridge Crescent, Singapore 117576
2 Professor, Nanyang Technological University, Singapore
3 Executive Geotechnical Engineer, SPECS Consultant Pte, Singapore
141
www.astm.org
142
years while those being planned will add another 50 square kilometers in the following
decade. The fill volume, mainly sand, needed to support such activities is enormous.
750
"~.
700
6- 650
r
600
<
550
500
1950
1960
1970
1980
1990
2000
2010
2020
Year
Figure 1 - l n c r e a s e in Land Area of Singapore

On the other hand, there has been a significant increase in the supply of dredged
sediment and marine clay as a result of reclamation activities, dredging activities to
maintain shipping lanes and exploitation of subterranean space, the disposal of which is
becoming an increasingly intractable problem. In recognition of the fact that these two
problems provide an ideal situation for marriage, significant research activities have
focused on how to use such materials directly as alternate fills in reclamation works. As
reclamation volume is always very large, it is important that any technique developed
will require minimal re-handling.
Two approaches have been developed to make use of such soils, in accordance to the
type of sediment obtained. If the sediment is very soft, in most cases with water content
much higher than liquid limit, then the sediment in the form of slurry is used directly in
the reclamation and the important challenge is to find ways to apply a small load on the
slurry as soon as possible so as to accelerate the consolidation process. In this case,
research has focused on finding an effective way to place a thin layer of capping sand on
the clay slurry, a method referred to as sand capping (Townsend et al. 1989). If the
sediment is sufficiently stiff for it to be in the form of lumps, the second method will use
these lumps directly in the reclamation, and the challenge is to develop effective post
treatment after the formation of the reclaimed land to ensure large inter-lump voids are
effectively closed.
Both techniques have been tested in pilot projects and then subsequently used in large
reclamation projects. A 40 hectares pilot project was carried out from 1988 to 1990 to
evaluate both methods and also to develop specific techniques to execute such a
reclamation project (Karunaratne, et al. 1990 and 1991). This pilot project was carried
out after a smaller scale project was carried out with the aim to build a fill with alternate
clay and sand layers, often referred to as the sandwich method (Lee et al. 1987). At
around the same time, to form a sand key for the building of a containment bund for a
reclamation project in the northeastern part of Singapore, seabed sediment was dredged
and, mainly in the form of lumps, was disposed in a reclamation site. This site was
characterized very recently using a radio-isotope cone (Karthikeyan et al. 2001 and
TAN ET AL. ON SAND CAPPING TECHNIQUE
143
2002). Another project was the creation offshore o f a 180 hectares containment bund in
eastern Singapore in 1986 to contain the hydraulic dumping of fine material washing
from sand quarrying activities. This area was reclaimed in the mid-1990s as part of the
reclamation works for the extension of Changi International Airport (Bo et al. 1998).
More recently, since 1999, a new mega project has commenced work, whereby dredged
and excavated clay lumps are being used directly as a major source of fill material.
The use of marginal soil as reclamation fill represents one of the most effective
approaches towards sustainable development of infrastructure related to the use and
generation of earth. Seabed sediment, regarded as waste previously and has to be
disposed in increasingly scarce land or offshore disposal sites when generated, is now
effectively a material of value when used as fill. Though there are two different
approaches depending on whether the sediment is very soft or in reasonable stiff lumps,
this paper will only concentrate on the topic of using very soft sediment as a fill, and will
discuss the approach adopted in Singapore. The use of clay lumps, which is the subject
of intense research activities presently, will not be described. The paper will first
describe results from two actual projects where the sand capping method was applied.
Using the results from these projects as a basis, the mechanics of the treatment method
will then be discussed. A simple model will then be introduced to link together the four
key parameters in the operation, namely the amount of sand that is being spread each
time, its grain size distribution, depth of water it has to descend to reach the clay and the
shear strength of the slurry. These four parameters will, in turn, affect the two key
aspects of the mechanics of sand capping, namely the rate of accumulation of sand on the
sediment and the formation of a permanent sand layer or otherwise o f the sand
accumulating. These ideas will then be used to explain the field performance described.
Changi South Bay Project (CSB Project)

This pilot test was carried out from 1988 to 1990 and had been reported by
Karunaratne et al. (1991). The project site was located to the south of Changi
International Airport. Figure 2 shows a plan of this project. A central dyke was formed
to partition the whole area into two large compartments and two sub-dykes were also
built for the purpose of facilitating the sand spreading operation. A thin sand mat, 0.3 m
to 0.5 m thick, was first laid on top the seabed. Then clay slurry was discharged from a
single outlet in Segment C 1 for Pond C and one in Segment A2 for Pond A.
After adequate time was set aside for sedimentation, typically about one week, a sand
layer of thickness varying from 5 cm to 20 cm was spread onto the pumped sediment. To
control this spreading, a perforated pipe system was used. Sand with water was pumped
through the perforated pipe, and a winch system was then used to move this spreader on
the surface of the pond in a controlled fashion. The thickness of sand spread was
controlled by the speed the pipe system was towed and the rate of pumping of the sand.
The purpose was to form a thin sand cap on top of this slurry. To ensure that the sand
could be formed on top of the slurry, the sand was spread in thin layers and in many
passes to form the cap. The selection of the thickness of sand to spread and the number
of passes needed is a crucial design consideration and is a key focus of this paper. After
adequate time of consolidation by the sand cap formed, a second layer of clay slurry was
144
pumped into the pond and after that, another layer of sand was spread on top of it in a
similar fashion. To monitor the effectiveness of the spreading, the density profile was
measured after each spreading using a gamma-ray probe. As the fill was a slurry, this
probe could be lowered directly into the sediment using its own weight. The accuracy of
this particular density probe, which used a 8.3 mCi Cesium 137 source, was 10 kg/m ~
Orlgln;tl
line
Shore
100 m
mI
SQale:
SQale: 1100
i
I
Figure 2 - Layout of Pond in the Changi South Bay Project
~"
rj
._
=
2.0 ~
-- ~ - A3 (N500)
----
1.o. l ~ i
~ ,~=~A4
-
__i
(N600)
6~o) _ 4
~oo-
--~
-10.
~
i
-2.0
....
1.1
; ....
115
; ....
12
~ ....
125
~ ....
13
; ....
135
14
Density (Mg/m3)
Figure 3 - Density Profiles at Different Distancesfrom Pumping Outlet
One problem encountered, when the clay slurry was pumped out from a single outlet,
was the dispersion of the clay slurry with coarser particles settling near to the pumping
point and finer particles depositing further away. This is clearly reflected in the density
profiles at different distances away from the pumping, which is shown in Figure 3 for
points in Segment A. Note that at Segment A2, the density at the surface is about 1.17
145
Mg/m 3, that is, a water content of just over 300% but shows an obvious increase with
depth. In Segment A5, the furthest away, the density at the surface is also about 1.17
Mg/m 3, but the density shows only a slight increase with depth. Subsequently in this
project, instead of pumping from a single outlet, the clay was pumped using the same
perforated pipe system as the one used for spreading sand. In that case, a nearly uniform
distribution was obtained.
The next important aspect of the operation was the formation of the sand cap on top
of the clay slurry. As the density profile of the pumped slurry clay changed with
distance, as shown vividly in Figure 3, the sand spreading sequence had to be adjusted to
take this into account. As shown in Figure 4, in Segment C2, the density of the slurry
was about 1.17 Mg/m 3 (just over 300% water content) but showed a significant increase
with depth, just as for Segment A2 in Figure 3. Thus for this segment, the spreading was
carried out using first a 5 cm pass followed by a 10 cm pass, then a 15 cm pass and
finally a 20 cm pass.
As shown in Figure 4, before spreading, the density at the surface is about 1.17
Mg/m 3 just as in Figure 3. However, the density increases with depth. After the first
spreading of 5 cm in Segment C2, it is clear that sand was trapped above the elevation of
-0.5 mCD (meter above chart datum). After the second spreading of l0 cm, a new sand
layer with a little mixture with slurry was clearly built on top of the first layer and the
bulk of the sand was now trapped above 0.0 mCD. Note that the density profiles are not
stationary as in between one spreading and the next; the slurry would have started
consolidating and compressed due to the trapped sand, and thus the surface of the sand
trapped after the first spreading would have settled some amount. Unfortunately, no data
was collected for the third spreading of 15cm and Figure 4 only showed the result for the
fourth spreading of 20 cm. At this stage, it was clear a clean sand layer was formed
above the elevation of 0.5 mCD, as indicated by the very high density of around 1.8
Mg/m 3, achievable only with sand.
1.5
,~
,~ -o.5
-1
,
-1.5
1.2
"%
" ~ ~ - - ~ - 2 n d spread (lOcm)l[

~ , [ a 4thspread(2Ocm) l[
1.4
1.6
Density (Mg/m3)
1.8
Figure 4 - Density Profiles at C2 after Different Stages of Sand Spreading
146
4.0
30
qlt ~ z ~ . . ~ - A a , ~ 4 - ~ "
o
>
X . \..
~m,
/L ~ o.~,.~
.
J~,_,
- ,k
Afiertwo 5cmSpreads
)1~(~.. ~,
~
0.0
. ." . .Before spreading
'~,~ ~,
1.o
.~e~"
- -)K-- After six 5cm spreads
h<~ ,,x
-'-Z~-- At~er ten 5era spreads

- ~ - After twelve 5cm
~:
~fleea~twelve5cm and
two 10crn spreads
-1.0
1.2
1.4
1.6
1.8
2.2
Density (Mg/m 3)
Figure 5 - Density Profiles at C6 after Different Stages of Sand Spreading

For Segment C6, which was the furthest away from the pumping outlet, the density
near the surface was also close to 1.17 Mg/m s, but showed negligible increase with depth,
just as for Segment A5 in Figure 3. Clearly, the slurry was much weaker, and thus the
sand had to be spread in thin layers. As this was an actual project, 5 cm was considered
the thinnest layer practical to control. As such, at this location, the sand was spread in
passes of 5 em until a layer was built up. After that, the spreading thickness was
increased. The density results after selective passes were shown in Figure 5. After 2
passes of 5 era, the density profile showed that most of the sand penetrated into the clay
slurry. However, even though the sand had totally penetrated, it could also be observed
that there was a slight increase in density, indicating that some amount of the sand spread
was trapped at the upper portion of the slurry. This sand that was trapped at the upper
portion would play an important role in arresting the penetration of future passes. This
was now clearly seen in the results after an additional four passes, that is, after 6 passes
totaling 30 cm of sand, there was a much more significant built up of sand, especially
towards the top now. However, at this stage, there was still no visible sand layer being
formed. Subsequently, the behavior became much more similar to that observed in
Segment C2 and discussed above. After an additional 4 passes of 5 cm each, a sand layer
mixed with slurry was clearly formed at the surface with density of around 1.4 Mg/m 3, or
a water content of about 120%. After this, a sand layer was formed near the surface and
finally, a clean sand layer, with density of about 1.8 Mg/m 3 could be seen to form after 80
cm of sand had been spread in twelve passes of 5 cm and 2 passes of 10 era.
147
The Silt Pond Project (SP Project)

In 1986, a containment sand bund was constructed around a previously burrowed pit
in the sea to provide a disposal area for silt and clay washings from sand quarrying
activities. This containment is nearly rectangular in shape with length of about 2000 m
and width of 750 m at one end to 1050 m at the other end. The total area is about 180
hectares. Pumping activities stopped towards the end of the 1980s when such quarrying
activities were phased out, and the slurry was allowed to consolidate under its own selfweight. Around 1995, reclamation of this pond commenced. To a large extent, the
lessons learned from the previous pilot test were applied to this project, which could be
viewed as a full scale implementation of the pilot test.
..................L................................................................
.......
,
Loc"o.A ............
L -,- LocationC
-10
- - - -
-15
0.8
1.2
1.4
--- *1,6
1.8
Density (Mg/m3)
Figure 6 -lnitial Density Profiles at Different Locations in the Silt Pond Project
Sub-surface investigations were carried out first in 1992 and then in 1994. These
tests indicated that the density near the surface is around 1.2 Mg/m 3 or a water content of
270%, with an average value o f about 1.3-1.35 Mg/m 3 or about 150% to 170% as shown
in Figure 6, which gave typical density profiles at three different locations in the silt
pond. As can be seen, the clay slurry, after many years of self-weight sedimentation and
consolidation, had settled to an elevation o f - 4 to -5 mCD. As these were field
measurements, it was difficult to obtain very precise values. Nevertheless, it could be
seen that the surface water content was ranging from 1.17 Mg/m 3 to 1.2 Mg/m 3 (water
content of 300% to 270%). More importantly, the density could be seen to increase very
rapidly with depth from a value of about 1.2 Mg/m 3 to a value of 1.4 to 1.5 Mg/m 3 within
a depth of 5 m. The difference between this profile and that observed in the CSB project
shown in Figure 3 was principally due to the different stages of consolidation as a result
of differing length of time the slurry was left alone.
As the water content of the slurry is still considerably large, it was recognized that
conventional sand dumping or hydraulic pumping would not work as the large quantity of
sand would break easily through such soft clay. But this profile was ideally suited for the
placement of thin sand layers as shown in the results from the pilot test at the CSB
148
project. Thus, the method developed at the pilot test in CSB project was applied here,
namely spreading the sand in thin layer through a perforated pipe system that is being
moved on the surface of the pond. As this was an actual and large scale project, the
density distribution was determined only a few times, and a somewhat less detailed
record was available. The density distributions at three different stages after dumping
were shown in Figure 7. Though the exact detail of the spreading was not well
established, but the sand was spread in thickness of about 20 cm in each pass.
As shown in Figure 7, it could be observed that the bulk of the sand spread just before
the first measurement on 9 January 1995 had been trapped at the upper portion near the
surface, but at this first measurement, no clean sand layer had been formed yet.
However, by the second measurement date, a clean sand layer, indicated by density as
high as 1.9 Mg/m 3 could be clearly seen. Subsequent spreading increased the thickness
of the sand layer formed. Also, from the bottom of the sand layer formed, the amount of
compression occurring between 20 March and 9 April, a total of 20 days, could be clearly
observed to be about 0.5 m. It is also important to note that the formation of a nearly 2 m
thick sand layer had occurred without any breaking in of the sand layer.
..............
-10
. ~
0.9
1.2
I -k-Apt 10, 95
1.5
1.8
._A
2.1
Density (Mg/m3)
Figure 7 - Density Profiles at One Location after Different Stages of Sand Spreading
Basic Principles in the Mechanics of Sand-capping

As can be seen from the above description, the use of soft sediment as a fill to be
treated by providing a sand-cap has proven to be a technically and economically feasible
approach even for a very large reclamation. A number of different aspects of mechanics
are involved in the entire process and these will be discussed next.
To be hydraulically transported, the slurry clay typically has water content 3 to 6
times of the liquid limit, at which state, the slurry has virtually no shear strength. On its
own, this slurry will undergo combined self-weight sedimentation and consolidation.
The theory of sedimentation and consolidation under self-weight for one-dimensional
problem is well established (Tan et al. 1990 and Tan 1995) and will not be discussed
149
further. At the end of the sedimentation period, the self-weight consolidation process
would have just started. At the CSB project, the sedimentation phase was observed to
take about a week to 10 days. At the end of this stage, at the segment furthest away from
the pumping outlet, the slurry was nearly uniform with a water content of about 270% 300% (density from 1.17 Mg/m 3 to 1.2 Mg/m3), as was observed in Segment A5 in
Figure 3 and Segment C6 in Figure 4 for the CSB project. Such a profile is an indication
that sedimentation is over but consolidation has just begun and thus the density has not
adjusted to the self-weight gradient imposed by gravity. On the other hand, at A2 where
the particle sizes were larger, the density profile showed a more significant increase with
depth compared to that at A5 where the particle sizes were much finer. Clearly, the rate
of sedimentation and consolidation at these two locations were different.
On other hand, for the SP project, the slurry had been left to settle under its own selfweight for a number of years, and at the end of that period, the density profile is
increasing with depth. This is a good indication that sedimentation is over and to a large
degree, the soil has consolidated under its own self-weight (Tan et al. 1990a). The very
rapid increase with depth of the density in the slurry in the upper portion of the slurry
highlights another very important aspect of the consolidation behavior of such a soft clay.
That is, a very small load can induce a very large compression when the water content is
very high. In effect, the success of the sand capping method depends on this simple
observation. Nevertheless, the water content at and near the surface was still very high,
principally because the effective stress available there was nearly zero. A measurable
yield shear stress o f a few Pascals can now be achieved for this slurry. Incidentally, the
water content at the surface of the slurry is about 3 times liquid limit, and this ratio seems
to be close to a boundary between sedimentation and consolidation observed by Tan et at.
(1990a), where the yield shear strength of a slurry shows a sudden change around this
ratio (Tan et al. 1990b).
The time needed for the slurry to gain sufficient strength by self-weight is too long
for most practical purposes. Even if adequate time is set aside for self-weight
consolidation, the absence of effective stress at the surface in a self-weight situation, will
result in a very weak surface layer, such as that seen in the results from the SP project,
and a way has to be found to increase the shear strength, especially at the surface. If the
void ratio versus pore pressure relationship of a slurry is examined (Tan et al. 1988), then
it can be seen that only a small load is needed to quickly induce a large increase in the
density. In the approach adopted in Singapore, this has been successfully achieved by
providing a sand cap through careful spreading of thin passes of sand through a moving
perforated pipe system.
Inoue et al. (1991) had found that the controlling factors in the mechanics of sand
capping were the grain-size of the sand spread, the thickness of the sand spread, the
height of supematant water and the shear strength of the slurry. When a sand mass is
spread onto the seawater, initially the motion of the sand particles is highly turbulent, but
the sand particles will quickly reach their respective terminal velocities and settle, as
shown in Figure 8. Particulate segregation will occur as heavier particles are falling
faster than lighter ones. The front of the falling sand mass will then impinge on to the
surface of the clay slurry. As shown in Tan et at. (2002), as individual particles, most of
them will be trapped and a very thin sand layer is then formed on top of the clay surface.
Trailing sand particles coming down and impinging on the trapped front particles will
150
then impose additional load. Depending on the interaction between the strength of the
slurry, and the rate of loading imposed by the trailing sand, this thin layer of sand will
either hold with increasing thickness or "break" into the clay. In the latter case, the
penetrated sand particles will either be arrested at the upper portion of the slurry or sink
to the bottom of the slurry. As shown in the field projects, the trapped sand also played
an important role.
120
~.
"6"
.~100
9 3.2cm
---~.--6.4 cm
, 95. cm
80
"E
60
"6
.~ 40
o
~
>
20
0.0
0.5
1.0
1.5
Depth of front (m)
2.0
Figure 8 - Velocity of Front of Falling Sand Particles for Different Thickness of Sand
Spread
The precise mode the thinly formed sand layer would break was difficult to assess,
but in all the experiments conducted in the laboratory, the failure was observed to be not
a punching type such as in the case for a shallow foundation. Instead, at a certain stage, a
weak point eaves in and the sand particles then flow into this plume, suggesting that this
is an instability problem. Thus far, no stability analysis could be found to take into
account this complex interaction described above where the loading and resistance are
both changing with time, and the resistance itself is also a function of the loading.
The above description identifies two key aspects in the mechanics of sand capping,
namely the rate of loading of the sand layer and the rate the slurry clay will gain strength
at the surface over the initial yield shear strength as a result of the ongoing consolidation.
On the other hand, the rate the slurry clay gained strength over its initial value is
governed by the load imposed and time. The important point is whether the slurry can
have enough strength right from the start, and if not, whether it can gain strength fast
enough to be able to hold the increasing load imposed by the falling sand.
Thus a key parameter to establish is the rate of sand accumulating on top of the slurry
surface. Usually, when this method is used, the sand is spread in thin layer, and to a large
degree, particulate settling is occurring. Based on Inoue et al. (1991) observation that
particles reach terminal velocity very quickly, Tan et al. (2002) has developed a simple
model to explain this accumulation. The terminal velocity of a particle is determined
using an extended Stokesian model of a spherical particle settling in a fluid, but taking
into account the actual drag coefficient if the flow is not in low Reynolds Number regime
(Batchelor 1973). To take into account the fact that the sand particles are actually not
spherical, an approach recommended by Zimmels (1986) was adopted. To account for
the grain size distribution, the sand particles are divided into smaller range of particle
151
sizes and the average terminal velocity in each range is determined. Thereafter, the
dispersion of a column of sand can be calculated based on the terminal velocity
determined for each range and the amount of particles in that range based on the grain
size distribution curve. In Tan et al. (2002), this model was shown to be able to describe
the behavior of this sand accumulation in experiments, a typical set of comparison is
shown in Figure 9. The important point to note is that this simplified model is able to
take into account three of the four key parameters, namely grain size distribution and
thickness of the sand spread and the height of supematant water above the slurry.
120
100
.~ 60 -
a. 40
-- c---Analytical model
20
Measured using cell
Estimated using video
01
0
10
20
30
40
50
Time (see)
Figure 9 - Sand Accumulation Experiment." Height of Water = 1.5m, Thickness of Sand

Spread = 7.5cm
To explain the interaction between the two aspects of the mechanics, namely the rate
the sand is accumulating on top of the slurry surface and the rate the slurry is gaining
shear strength, a simple "bearing capacity type" of concept was introduced by Tan et al.
(2002) to explain the subsequent break-in of the sand under the right conditions. From
the results of sand accumulation tests by Tan et al. (2002) and the simplified model
introduced, and noticeable in Figure 9, in most cases the rate of accumulation of sand is
nearly linear until near the end of the accumulation period. If this linear relation is
assumed, the loading on the clay surface with time, crL(ta) is:
CrL(ta)= kl rL ta
(1)
where to is the elapsed time after the sand layer begins to accumulate on the clay surface
and kj is a constant to account for the interaction with other particles.
The shear strength of the slurry, especially at the surface, will also increase with the
overburden load applied as a result of consolidation. Thus the shear strength increase is
also a function of two factors; time and the rate of sand accumulation, rL. A faster rate of
sand accumulation would mean a larger surcharge is applied in a given time and, in turn
the surface, which will consolidate almost immediately, will be able to gain strength
152
faster. As a first order estimate, this resistance, or "bearing capacity" of the slurry is
assumed to be also linear and of the form:
(~R(ta) = k2 (17yq-- k3 rL ta)
(2)
where k2 is a "bearing capacity factor" and k3 is to account for the increase in shear
strength due to an increase in vertical load. "~yis the yield shear stress of the slurry
initially. For failure to occur, the rate of loading must be significantly faster than the rate
of increase in shear resistance so as to exceed the combined shear resistance provided by
the initial yield shear stress and the subsequent increase due to consolidation, and the
amount of sand spread must be enough for this to occur. These two conditions are
specified as follows:
(a)
~L(t~f)-> OR(tar)
(3a)
(b)
tot~,,
(3b)
where t.fis the time to failure and t~. is the time needed for the entire amount of sand
spread to accumulate at the surface at a constant rate ofrL. Figure 9 shows that the
residual amount of sand deposited after time tau is small. At the onset of failure, OL(t.f) =
oa(taf) and a simple manipulation will give:
tal = a r ~
rL
where
ct -
(4a)
k2
k I -k2k 3
(4b)
If rL is expressed in load/unit area/time, then c~ is dimensionless. The above

/
derivation showed that the ratio r'~/< holds the key to understanding sand capping. In
the conceptual model developed, the idea is very simple. If the sand is deposited
completely at the rate rE in a time that is less than tar, then failure cannot occur, else
failure would occur. This simple idea is illustrated graphically in Figure 10. The three
unknown parameters kt, ke and ks are combined into a single parameter, which facilitates
back-analysis. Though Eq. 4a is a very simple equation, it actually encompasses all the
key ideas discussed previously. It relates the two key aspects of the mechanics of sand
capping, namely the rate of sand accumulating at the surface and the rate the shear
strength is increasing. The rate of accumulation of sand has taken into account the grain
size distribution, the amount of the sand used and the height of water the sand has to fall
through. Thus, this simple model is able to bring all the different ideas together. Clearly,
the most important limitation in this derivation is the assumption of linearity for increase
in loading and shear strength. However, this does not distract from the key idea that the
instability can be describe by the interaction between the two key aspects of mechanics.
I. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
L o a d i n g from
Sand (No
Failure)
.:o
- - L o a d i n g from
Sand (Failure)
" "r
~" "" ~ , . ' ' "
153
" ~ ' ~ " ' ~ Resistance from slurry

i
,.d
:
U
!
taf
tau
Time
Figure 10 - Schematic Interpretation of the Interaction between Sand Loading and Gain
in Shear Resistance by Slurry

Though the line.ar assumptions are expected to be true for only very short duration,
nevertheless, it can be used to provide a rough guide to interpret the field results. For a
given spread per pass in S Pascal (product of thickness per pass and the effective unit
weight), tau is given by:
to. = (S)~rj
(5)
where rL is in Pa/s. If the ratio oft~4t,u is now calculated based on Eqs. 4a and 5, then,
this will yield
t.j
t~,
ry
-
(6)
In Eq. 6, the ratio has to be greater than 1 for failure not to occur. To do this, either Xy
has to be large or S has to be small. In the former, this means the slurry has to be
stronger (lower water content) and in the latter, it means that sand has to be spread in
thinner layer. In Tan et al. (2002), back analysis of results from 37 tests indicates a value
ofct of 260. A 5 cm pass, equivalent to about 400 Pa under submerged condition, will
require the slurry to have a yield shear stress larger than 1.5 Pa. On the other hand, a 20
cm pass, equivalent to 1600 Pa will need a yield shear stress higher than 6 Pa for failure
not to occur. The clay slurry at the CSB project has a liquid limit of about 80%, and was
shown in Tan et al. (1991) to have a yield shear stress of about 1.5 Pa at 300% whereas to
achieve 6 Pa, its water content has to reduce to about 240%. These two values are in
reasonable concurrence with the observed behaviors shown in Figures 4 and 5.
It is recognized that the present treatment is simplified. The rate of loading by the
sand spread is likely to be non-linear when the water depth is great. The gain in yield
154
shear stress as a result of the ongoing consolidation is also unlikely to be a linear function
of time for significant period of time. Nevertheless, the above derivation does provide a
coherent way to demonstrate the importance of the ratio of the initial yield shear stress of
the slurry over the weight/unit area of the sand spread in controlling sand capping.
Conclusions
The problem of treating a large volume of slurry clay is increasingly important due to
the scarcity of land in major urban centers. In Singapore, the solution adopted is to use
such dredged material as reclamation fill, and to treat this, a sand cap is built on top of the
slurry. In this paper, results from two actual projects were described. These two
projects, one for 40 hectares and the other for 180 hectares, clearly showed the feasibility
of this method.
The mechanics of sand capping is discussed. The controlling parameters in sand
capping are the thickness and grain size of sand spread, height of water and the initial
yield shear strength of the slurry (or water content of the slurry). The first three
parameters govern the rate the sand is accumulating on top of the slurry while the rate the
slurry will gain strength over its initial value is a function of this loading and time. If the
loading is fast enough, failure will occur with the sand layer caving into the slurry. If not,
then a clean sand layer will be formed as can be seen from the result from the two
projects reported. A simple conceptual model was presented to demonstrate this idea.
A simplified model is able to show that the key parameter is the ratio of the initial
yield shear stress over the expected loading. When this is applied to the field project, it
seems to provide a reasonable guide as to what to expect in the field. Conclusions cannot
be firmer than this at the present stage as the assumptions involved are highly simplified
and also the field conditions are not as simplified as that assumed in the derivation.
However, this does not distract from the fact that the above derivation and subsequent
results do clearly demonstrate what are essential in the mechanics of sand capping.
Acknowledgments
The assistance of Dr. R. G. Robinson and Mr. M. Kathikeyan in preparing the figures
is gratefully acknowledged. The results of the laboratory works reported here were
basically drawn from the works of two former students, Dr. T. Inoue and Dr. T. C. Goh,
of the senior author.
References
Bo, M. W., Arulrajah, A., Choa, V., and Na, Y. M., 1998, "Land reclamation on slurry
like foundation," Proceedings, International Symposium on Problematic Soils, ISTohoku, A.A. Balkema, pp. 763-766.
155
Inoue, T., Tan, T. S., and Lee, S. L., 1990, "An investigation of Shear Strength of Slurry
Clay," Soils and Foundations, Vol. 30, No. 4, pp. 1-10.
Inoue, T., Tan, T. S., and Lee, S. L., 1991, "Sand Penetration into Clay Slurry," GEOCOAST '91, Sept, 1991, Yokohama, pp. 317-322.
Karunaratne, G. P., Yong, K. Y., Tan, T. S., Tan, S. A., Liang, K. M., Lee, S. L., and
Vijiaratnam, A., 1990, "Layered clay-sand scheme reclamation at Changi South Bay,"
Proceedings of the l Oth Southeast Asian Geotechnical Conference, 16-20 April 1990,
Taipei, ROC, Vol. 1, pp. 71-76.
Kartmaratne, G. P., Yong, K. Y., Tan, T. S., Tan, S. A., Lee, S. L., and Vijiaratnam A.,
1991, "Land reclamation using layered clay-sand scheme," GEO-COAST '91, Sept,
1991, Yokohama, pp. 335-340.
Tan T. S., 1995, "From sedimentation to consolidation: A geotechnical perspective,"
Proceedings, Compression and Consolidation of Clayey Soils, May, 1995, Hiroshima,
Japan, Yoshikuni and Kusakabe (eds), Balkema, pp. 937-948.
Tan, T. S., Yong, K. Y., Leong, E. C., and Lee, S. L., 1990a, "Sedimentation of Clayey
Slurry," Journal of Geotechnical Engineering, ASCE, Vol. 116, No. 6, pp. 885-898.
Tan, T. S., Yong, K. Y., Leong, E. C., and Lee, S. L., 1990b, "Behavior of Clay Slurry,"
Soils and Foundations, Vol. 30, No. 4, pp. 105-118.
Tan, T. S., Goh, T. C., Karunaratne, G. P., Inoue, T., and Lee, S. L., 1991, "Yield Stress
Measurement by a Penetration Method," Canadian Geotechnical Journal, Vol. 28,
pp. 517-522.
Tan, T. S., Goh, T. C., Karunaratne, G. P., and Lee, S. L., 1994, "Shear Strength of very
soft clay sand mixes," Geotechnical Testing Journal, ASTM, 17, No. 1, pp. 27-34.
Tan, T. S., Goh, T. C., and Karunaratne, G. P., 2002, "A study of the mechanics of sand
spreading and penetration into a slurry," Submitted to the Canadian Geotechnical
Journal.
Townsend, F. C., McVay, M. C., Bloomquist, D. G., and McClimans, S. A., 1989, "Clay
waste reclamation by sand/clay mixture or capping," Journal of Geotechnical
Engineering, Vol. 115, No. 11, pp. 1647-1666.
Zimmels, Y., 1986, "Accelerated and steady particle flows in Newtonian Fluids,"
Encyclopedia of Fluid Mechanics, Gulf Publishing Company, Houston, Vol. 5: Slurry
Flow Technology, pp. 93-153.
Dave Ta-Teh Chang, I Chao-Ping Sung, 2 Bor-Ling Chen, 2 and Nuan-Hsuan Ho 3

Durability Study for Geotextile Tube Use in Talimu River Sediment Control
Reference: Chang, D. T.-T., Sung C.-P., Chen B.-L., and Ho N.-H., "Durability Study
for Geotextile Tube Use in TaUmu River Sediment Control," Contaminated
Sediments: Characterization, Evaluation, Mitigation~Restoration, and Management

Strategy Performance, ASTMSTP 1442, J. Locat, R. Galvez-Cloutier, R. C. Charley, and
K. R. Demars, Eds., ASTM International, West Conshohocken, PA, 2003.
Abstract: To rebuild dikes for Talimu River regulation project, insitu sediments is
considered to be contained by geotextile tube ( geotube ) for use as dike core. Stress
analysis and matrix shape analysis are required to determine the geotube size. Based
on experience and the local climatic conditions of Xing Jiang, China, durability of the
geotubes is considered the most critical factor for sediment control. The geotube
material used in this study is woven geotextile made in Taiwan - GTM-1 and GTM-2.
The tests were conducted in accordance with ASTM Test Method (D 4355) for
deterioration of geotextiles from exposure to ultraviolet light and water. After UV
exposure for 200 and 500 hours, the results showed that the residual tensile strengths of
the geotextiles are greater than 90 % and 70 % of the ultimate tensile strength,
respectively. More results showed that reduction in tensile strength is insignificant
even after UV exposure for over 600 hours. Another durability test is the geotextile's
resistance to freezing and freeze-thaw cycles. Initially the geotextiles were placed in a
temperature chamber and cured at -20 ~ -30 ~ -40 ~ for 6 days. The results
showed that the geotextiles still retained greater than 90% of its ultimate tensile strength.
For the freeze-thaw cycles (-40 QC--40 ~ -10 ~
~ results also indicate that
the selected geotextiles provide good freeze-thaw resistance.
Keywords: sediment control, geotextile tube (Geotube), UV resistance, freezing
resistance
IProfessor,2Research Assistant, Civil Engineering Department of Chung Yuan Christian

University, 22 Pu-Jen, Pu-Chung Li, Chung Li, Taiwan, ROC 320.
3Ph.D candidate, Department of industry engineering and management, Chunghua
University, and also Director General of Bureau of Taiwan High Speed Rail,
Ministry of Transportation and Communications, R.O.C.
156
Copyright9 2003 by ASTM International
www.astm.org
CHANG I=T AL. ON GEOTEXTILE TUBE USE
157
Introduction
The Talimu River at Xing Jiang flows over 1200 km, making it one of the longest
rivers in China. In the last 30 years, the Talimu River has suffered from reduced water
flow, channel sedimentation, and a damaged shipping channel resulting in overflow of
river water and severe loss of water. Thousands of hectares of agriculture development
are facing land degradation problems, with more than 130 000 hectares of agriculture
land having become desert. The situation is expected to deteriorate if the problem is
not addressed properly. A local government environmental committee was established
to study the rehabilitation of the Talimu River and the ecology system on both sides of
the river. So far, a number of hydraulic structures have been built. The rehabilitation
has been undertaken in two phases. An initial short-term investment of up to USD 40
millions is called for to urgently implement a regulation to manage the river water flow
so as to restore the total volume conveyed to 550 millions cubic meters per day.
Long-term investment is forecast in the USD 150 millions bracket in order to provide a
total volume o f 750 millions cubic meters conveyance and eventually to guarantee the
supply of water for both industrial and household use.
Channels used in this project are mostly of conventional design where an
embankment dike was built on both sides of the river. A conventional design of this
type has the drawback of forming a wide but shallow channel. The region sits at the
border of a desert and the major soil profile here is fine, loose sand on which it is
difficult for dikes of this type to remain stable. As such, liquefaction is commonly
encountered here even if the slope of the embanked dike is reduced from 1:2 to 1:3.5 ;
see Figure 1 (Casagrande I975). Through a gradual process, the embankment dike will
erode back towards the river, with sediment blocking the waterway and causing the river
to overflow, resulting in a significant drop in the conveyance capacity. The committee
decided after extensive discussions to using geotubes filled with sediments as the dike
core. This method has two advantages: more stable embankments are constructed, and
at the same time, the waterway is dredged because the sediments are pumped into the
geotube. The structures, once completed, are more resistant to erosion and are thus
easier to maintain. This study focused on the durability of geotube and its use as the
dike core.
Original Designed Channel Cross Section

Embankments along the Talimu River were built by hydraulic fill. Initially, the
slope designed was 1:2.0. After experiencing total collapse o f quite a few sections of the
embankment dike, the gradient of the slope was reduced to 1:3.5. There are three
different types of channel design cross-sections which contain a 6 - - 8 cm concrete plate,
a 0 . 4 - 0 . 5 mm thick geomembrane, a 3--20 cm layer of coarse sand/gravel layer, and a
geotextile filter.Typical sediments in this region is classified as ML or SM based on its
grain size and for its non-plasticity o f fines (see Figure 2). This type of soil has poor
resistance to erosion. At the same time, capillary action and freeze-thaw conditions
could adversely affect the slope, foundation and core stability. Using twenty percent of
the effective grain size (D10) as the estimate effective pore size, the height of capillary
rise can be obtained, which is approximately 16 m. There is a substantial variation in
temperature in this region. During winter, the temperature can typically drop to -20 ~
sometimes go as low as - 4 0 ~ with the wind chill; in the summer, the temperature can
be as high as 40 ~
An embankment core consisting o f geotubes can isolate or at least
moderate the freeze-thaw effect, reduce capillary action, enhance the restraining forces
o f soil and increase erosion resistance.
158
Figure
1-
Liquefaction in Loose Sand Adjacent to a Waterfront (Casagrande 19 75)
CHANG ET AL. ON GEOTEXTILE TUBE USE
159
100
90
80
~O
Dro : 0.105 mm
70
60
D3o : 0.072 mm
D~o : 0.008 mm
~" 50
Cr : 6.171
40
Ca : 13.125
30
Ik
2O
10
1
0.1
0.01
0.001
Grain Size(ram)
Figure 2 -- Grain Size Distribution Analysis for Typical Sediment Sample
Research Approaches
To ensure long-term performance of the geotube, it is essential that the geotube be
made of geotextile material of sufficient strength and durability. In this study,
geotextile materials developed and manufactured by Taiwanese manufacturers were
used to conduct the optimum design, aging test and durability test. It is anticipated that
results from the analysis will serve as guidelines for future design, manufacture and
application of geotubes.
Optimum Design for Tube Matrix Size

At the initial or intermediate stage of the installation, the geotube has a cylindrical
cross section. As can be seen in Figure 3(a), when the geotube is flat, the height of the
installed geotube (H) approaches zero (H = 0), the radius of both comers (rounded ends)
of the geotube approaches zero ( R ' = 0 ) and the larger R, representing a arc of an
infinitely large circle approaches infinity (R---~oo). The tensile modulus of the
geotextile under this condition equals zero. On the contrary, when the geotube
becomes cylindrical in shape, H/2 = R = R', the tensile modulus of the geotextile
theoretically approaches infinity. The tube height-to-width ratio is normally kept at
below 0.8, so that the designed tensile modulus of the geotextile is not too high. The
matrix shape of geotube also can be expressed in a simplified manner as in Figure 3(b).
The cross section shows that both ends are semi-circular in shape while the upper and
lower sides are fiat.
160
Tensile Strength of
Geotextile (kN/m)
Matrix Ratio
2 0 0 ~
Height
of
Geotube
~
.,
1 0 =
IIII
III
I-,.
I"
I I
I I
II
II
I -
Geotube Width after Filling (w)

Geotube Width before Filling (FBW)
J v tI
w'l
b,~
(a) Relationship of Stress Tensile Strength of Geotextile and Shape of Tube
Id
bl
(b) Simplified Tube Cross Section
Figure 3 --Size and Shape of Geotubefor Calculation (from Leshchinsky et a11996)
Aging and Durability Evaluation of Geotube Material

The geotube is made of a woven geotextile. In accordance with ASTM Test
Method for Tensile Properties of Geotextiles by the Strip Method (D 1682), one can
measure and observe the changes in tensile strength. Aging conditions were simulated
by exposure to ultraviolet light and the use of a temperature chamber (Schneider 1988,
Sprague and Goodrum 1987). The operating range of the temperature chamber used in
this study is between -40 ~ and 40 ~
The tests were conducted in accordance with
ASTM Test Method for Deterioration of Geotextiles from Exposure to Ultraviolet Light
and Water (D 4355). Duration of UV exposure can be as long as 1000 hours, after
which changes in tensile behavior were analyzed. Both wet and dry woven geotextiles
were subjected to freeze-thaw cycles to study the effects of temperature on the tensile
strength of the geotextiles.
Basic Properties of Woven Geotextiles

Materials used in this study are Taiwan products, GTM-1 and GTM-2 denoted as
Sample A and Sample B, respectively. Basic properties of the woven g~otextiles are
given in Table 1.
Table 1 --Basic Properties of Woven Geotextiles Used in the Laboratory Study
Testltems
TestMethod
Thickness (mm)
ASTM D 5199
1.095
1.128
Mass per unit area (g/m 2)
ASTM D 5261
345.2
361.1
Strip tensile strength and

elongation (kN/m; %)
GTM-1(sampleA) GTM-2 (sampleB)
ASTM
D1682
MD~
CMD 2
66 (19.7)
64 (18.6)
82 (23.6)
70 (15.9)
Trapezoid tearing strength ASTM

(kg)
D4533
MD1
CMD 2
118.6
104.7
133.7
124.2
ASTM
D4632
MDI
CMD 2
258.5
224.5
270.5
234.5
Wild strip tensile strength ASTM

and elongation (kN/m; %) D4595
MDJ
CMD 2
57 (23.1)
56 (22.8)
71 (26.8)
62 (17.7)
Grab breaking load (kg)
Apparent opening size

(Ixm; 095)
ASTM D 4751
229
248
Coefficient of permeability
(cm/sec)
ASTM D 4491
2.4x 10.2
2.7x 10-2
~MD = machines direction

2CMD = cross machine direction
161
162
Durability Behavior Studies for Geotextiles
UV Resistance
After exposing the woven geotextiles of Sample A and Sample B to UV light for
200 hours in accordance with ASTM D 4355, the residual strip tensile strengths of the
geotextiles were measured to be 9 1 % and 93 % respectively, o f the tensile strength
before UV exposure. After 500 hours of UV exposure, the residual strip tensile
strengths were 76 % and 78 % respectively. This complies with the requirements of
typical design guide (Holtz et al. 1998), wherein residual tensile strength after 500 hours
of UV exposure shall not be less that 70 % of the ultimate tensile strength. Figure 4
shows the residual tensile strengths under UV exposure. From the curves, it is found
that for UV exposure beyond 600 hours, the additional strength loss is insignificant.
This is probably because of the white powder deposits formed on the geotextile surface,
which blocked the effect of further UV exposure. It is recommended that the chemical
composition of the white power deposit be examined and to understand the mechanism
of weakening of the effect of UV light. According to Raumann's findings (1982)
untreated polypropylene is sensitive to UV light, a suitable stabilizer should be added
for improving UV resistance. In this study, carbon black is added to the raw material
of geotextile A and B. Selection of proper additives is another important topic in
manufacturing stable geotextile material.
o
1oo
"~
90
MD
.t
CMD
70
60
50
100
200
300
400
500
600
700
800
900
IOOO
UV Exposure Period (hr)
(a) Sample A
I00
9- . e - -
80
'~
70
MD
~ C M D
6o
5O
100
200
300
400
500
600
UV Exposure Period (hr)
700
800
900
(b) Sample B
Figure 4 - The Strength Residual under UV Exposure
1O120
163
Strength Stability Subject to Temperature Variation

Wet geotextile and air-dried geotextile were cured at - 2 0 ~ -30 ~ and -40 ~ for
2, 4 and 6 days. The cured specimens were returned to room temperature (21 l~
before the strip tensile test was carried out to identify the effects of freezing condition on
geotextile tensile strength. The annual temperature ranges in Xing Jiang Province
varies from -40 ~ to 40 ~
The temperature at Urumqi City ranges from -10 ~ in
winter to 20 ~ in summer. Wet and air-dried geotextiles were again subjected to
freeze-thaw cycles from -40 ~ to 40 ~ and from - 1 0 ~ to 20 ~
When the
temperature chamber attained the intended temperature, it was maintained for at least 4
hours before switching to another temperature. This study used a specially designed
temperature chamber that can achieve a temperature gradient of 1 ~
min.
Geotextiles were treated with 15, 30, 45 and 60 freeze-thaw cycles. The tensile
strength of the geotextiles was measured after these curing cycles.
Referring to the table at the right hand side in each of Figures 5 through 10,
definitions for all digits can be described below, the first digit in the test designator is
either "A" or "B'. "A" or "B" denotes the type of geotextile used ( refer to Tablel ).
The second digit, either "1" or "2" denotes the machine direction, either 1 = MD or 2 =
CMD ( cross-machine direction ) denotes machine direction (MD) and "2" denotes
cross-machine direction (CMD). The third digit is either "D" or "W", wherein "D"
denotes dry condition and "W" denotes wet conditions. The last two digit is either of
"20", "30" or "40"; in which "20" denotes a curing temperature o f - 2 0 ~
"30"
denotes a curing temperature o f - 3 0 ~ and "40" denotes a curing temperature o f - 4 0
~
Each plotted data is an average value obtained from these six specimens. A total
of 852 specimens were prepared for strip tensile tests.
As can be seen from Figures 5 through 8, the effect o f temperature and curing
period on tensile strength is insignificant. As the number o f curing days increases, the
tensile strength reduction becomes less dependent on the curing time. For dry and wet
conditions, (D versus W specimens) there was no significant difference noted in the
tensile strength reduction curves. For all of these tensile strength reductions, the
results indicated the residual strength still remains above 90% of the ultimate tensile
strength. From Figures 9 and 10, it can be seen that with an increasing number of
freeze-thaw cycles, the strength reduction per cycle tends to increase. The strength
reduction of wet geotextile is slightly higher than that of air-dried geotextile. This
perhaps occurs as the freeze-thaw cycle causes water to squeeze and stretch the pore size
of the geotextile and hence undermines the matrix structure of the geotextile. In all
cases, the tensile strength still remains above 80%. It must be noted that the adverse
effects of freeze-thaw cycles on tensile strength are greater than that of freezing only.
100
100
_ ~ _
* AI-D-20
N 96
.i . . .i. . . .i . . .i. . . .~. . . . . .
"~ 96
"~94
~ 94
~92
~ 92
Liiii
.
9O
9o
Curing Period, Days
Curing Period, Days
Figure 5 --Sample A with MD Tensile Strength under Varied Curing Temperature
164
100 i
\- - - \ " ~ ...... I ' " ' ~ ~ ~ ~
iiiii
oo
I--\'~---1
* ,~-w-3o/i
._~_! . . . . . . . . . ~
......
t,.,
O
*~90
I ,
2
4
6
Curing Period, Days
Curing Period, Days
Figure 6-- Sample A with CMD Tensile Strength under Varied Curing Temperature
100
I00
BI-D-20
~ . _ _
---\'K ...... I-'-Bl~3~
_!_a?......... ~
90
o~
iiiii
2
4
6
CuringPeriod,Days
" ~ Bl-W-20
. . . . [ - - ~ BI-W-30
9O
8
~ _ ~od Day6
Figure 7 -Sample B with MD Tensile Strength under Varied Curing Temperature

~100
+ B2-D-20
--- -m-B2-D-30
~100 ~
] -m-B2-W-20
~
- " - N ' ~ . . . . . 1 ,It' B2-W-30
--~-B2-D-4o,~
~9{}
ii~.w.4o
NgO
2
4
6
CuringPeriod,Days
'
'
2
4
6
CuringPeriod,Days
Figure 8 -- Sample B with CMD Tensile Strength under Varied Curing Temperature
CHANG ET AL. ON GEOTEXTILETUBE USE
165
100
.x
A10(40-40)
....
la
)-
A1-W(-40.-.40) . . . . . . . . .
80
0
15
30
45
Cycles
60
~ 80
75
" ,A2-W(,-40.-.40![. . . . i . . . .
0
15
30
45
Cycles
60
75
Figure 9 - S a m p l e A with MD and CMD Tensile Strength under Freeze-Thaw Cycles

100
100
19o
-r
D(10--,20)
-.m- B l-W(-10-20)
at BI-D(-40--40)
~
BI-W(-40---40)
B2-D(- 10--20) . . . . .
B2-W(-10.--20) ._
9 B2-D(-40--40)
---(b)- B2-W(-40--40)
i
-- {-a ) - .....
80
80
0
15
30
45
Cycles
60
75
15
30
45
Cycles
60
75
Figure 10--Sample B with MD and CMD Tensile Strength under Freeze-Thaw Cycles
Review of Requirement for Geotube Tensile Properties

The geotubes used in this study were unseamed. To cope with the machine width
(5.5 m), the perimeter o f the geotube is set at 5 m. Based on the simplified cross
section as shown in Figure 3(b), a preliminary dimension can be estimated. In the case
of a single geotube, if the perimeter is 5 m, the water pressure can be assumed to be 0.3
kg/cm 2, then the perimeter stress of the geotube is 1.8 MPa (tensile stress of geotextile
requires 17.5 kN/m), the limit stress is 7.16 MPa (maximum tensile stress requires 70
kN/m); the geotube maximum height is 1.4 m, while the H/W ratio is 0.8. With
grouped tubes, the interlocking mechanism between geotextile and soil shall be
considered, such as consolidated drain strength, shear strength, normal stress and
surcharge. The resulting tensile strength for geotextile much lower than single tube in
factored in. According to typical analysis results as given above, a suitable geotube
material, sample B, is therefore selected for ongoing project details.
166
Conclusions and Recommendations
1. Carbon-black treated geotextile manufactured in Taiwan after subjection to UV

exposure still possesses a tensile strength that fully complies with relevant design
manual and is considered applicable in the Xing Jiang area, which has very harsh
climatic conditions.
2. The effect of freezing temperatures on the tensile strength of geotextile A and B is
insignificant. For a curing period of up to 6 days, the tensile strength remains
greater than 90%.
3. The effect of freeze-thaw cycle on the tensile strength of geotextile is insignificant.
In all cases, the tensile strength remains greater than 90%, except for Sample A after
subjected to 60 cycles of the -40 ~
~ cycle. In this particular case, the
tensile strength drops to slightly lower than 90 %.
4. For single tube and mechanically calculated and group tubes, all stress levels meet
the properties of Sample A and Sample B.
5. It is suggested that Sample B be used in the project as it has a higher tensile strength.
Reference
Casagrande, A., 1975, "Liquefaction and Cyclic Deformation of Sands, a Critical

Review," Proceedings of the Fifth Panamerican Conference on Soil Mechanics and
Foundation Engineering, Buenos Aires; reprinted as Harvard Soil Mechanics Series,
No.88, pp. 27.
Holtz, R. D., Christopher, B. R., Berg, R. R., 1998, "Geosynthetic Design and
Construction Guidelines Participant Notebook," Publication No. FHWA HI-95-038.
Federal Highway Administration, U.S. Department of Transportation.
Leshchinsky, D., and Leshchinsky O., Ling H. I., and Gilbert P. A., 1996, "Geosynthetic
Tubes for Confining Pressurized Slurry: Some Design Aspects," Journal of
Geo-technical Engineering, ASCE, Vol. 122, No.8.
G., 1982, "Outdoor Exposure Tests on Geotextiles," Proc.Second
International Conference. Geotextiles, Las Vegas, NV, IFAI, pp.541-546.
Raumann,
Shneider H., 1988, "Long-Term Performance of Polypropylene Geosynthetics,"

Geosynthetic Research Institute, Drexel University, Philadelphia, P.A., pp. 95-109.
Sprague C.J and Goodrum R. A., 1994, "Selecting Standard Test Methods for
Experimental Evaluation of Geosynthetic Durability," Transportation Research
Record, No. 1439, TRB, National Research Council, Washington, D.C., pp. 32-40.
Sibylle Dueri~and Ren6 Therrien 2
Factors Controlling Contaminant Transport Through the Flood Sediments of the

Saguenay Fjord: Numerical Sensitivity Analysis
Reference: Dueri, S. and Therrien, R., "Factors Controlling Contaminant Transport

Through the Flood Sediments of the Saguenay Fjord: Numerical Sensitivity
Analysis, ') Contaminated Sediments: Characterization, Evaluation,
Abstract: In July 1996, two days of intense rainfall caused severe flooding in the
Saguenay region and led to the discharge by rivers of several million cubic meters of
clean sediments to the Fjord. This turbidite layer, composed of clean sediments,
represents a potential barrier for the migration of heavy metals and PAHs from the
underlying contaminated sediments towards the new sediment-water interface. A
numerical model has been developed to simulate the vertical migration and
remobilization of dissolved contaminants in the capping layer. The model includes the
main physical, chemical and biological factors affecting the contaminant flux in the cap.
Calibration of the model has been achieved through comparison with the data collected
during the field and laboratory characterization of the sediments of the Saguenay Fjord. A
detailed sensitivity analysis, based on factorial design, shows that the model parameters
associated with bio-irrigation have the greatest impact on the model output and by
extension on the effectiveness of a capping layer.
Keywords: Saguenay Fjord, sediment cap, contaminant migration, numerical model,

sensitivity analysis, factorial design
PhD Student, Drpartement de Grologie et g6nie grologique, Universit~ Laval, Qu6bec,

Canada G 1K 7P4.
- Professor, Drpartement de Geologle et grnie grologique, Universit6 Laval, Qurbec,
Canada G I K 7P4.
9
167
www.astm.org
168
Introduction
The capping of contaminated sediments with clean material is a promising

alternative for the isolation of contaminants from the water column and the prevention of
further pollution. This approach has the advantage of being applicable even if the
contamination extends over a large surface. A capping layer is designed to physically
isolate the contaminants from the benthic fauna, stabilize the sediment, and prevent or at
least reduce the contaminant flux towards the new water-sediment interface. Any attempt
to predict the long-term effectiveness of the cap requires an understanding of the
physical, chemical and biological factors affecting the migration of contaminants in the
sediments.
The TRANSCAP- 1D numerical model was developed to simulate the migration of
dissolved contaminants through a subaqueous capping layer. The model considers the
physical, chemical and biological processes affecting the fate of a dissolved contaminant
in a sedimentary environment and can be used to predict the effectiveness of a sediment
cap in isolating the contaminant from the water column. The transport equation includes
advection, diffusion, chemical reactions and the effect of bio-irrigating worms.
The model has been applied to the Saguenay Fjord, where a major flood led to the
deposition of a natural capping layer over contaminated sediments. Since the catastrophic
flood, which occurred in 1996, the geotechnical and geochemical characteristics of the
capping layer, as well as the recolonisation of the benthic fauna, have been extensively
studied in the context of a multidisciplinary research project. The objective of i.his
research project is to assess the effect of the new capping layer on the migration of
contaminants towards the new sediment-water interface. A large database was available
for the model calibration and the response of the numerical model was successfully
matched to measured concentrations of dissolved arsenic in the Saguenay Fjord. These
results will be presented in a separate paper that focuses on the calibration of the model.
Sensitivity analysis is commonly used in engineering to assess the influence of a
model input on the output. The simplest form of sensitivity analysis consists in increasing
or decreasing the value of one input parameter at a time by some fraction, or perturbation,
and then evaluate the corresponding variation in the model output. This procedure is
repeated for each parameter tested and the parameter causing the greatest modification in
the model output is deemed the most important. Although this approach is simple and
commonly used, there is no guideline for the selection of parameters to evaluate and for
the magnitude of the perturbation to impose. In addition, since parameters are evaluated
one by one, any significant combined effect of two or more parameters is never detected.
Kabala (2001) presents an overview of sensitivity analysis, with an application to
pumping tests in aquifers. The definition of sensitivity of model output to input
parameters given by Kabala (2001) is more precise than the general definition given
above, because it is expressed in a mathematical form as the differential in the model
output with respect to selected input parameters. By using this definition, sensitivities can
be calculated with more precise mathematical techniques, such as the direct method that
requires differentiating the governing equations, or the adjoint method that requires
solving the governing equation and its mathematical adjoint.
DUERI AND THERRIEN ON NUMERICALSENSITIVITYANALYSIS
169
Other approaches have been developed in chemical engineering to determine the

combined effect of several parameters on model output. An example is a two-level
factorial design, which is used to determine the influence of n parameters on some
observed value with a minimum number of experiments (Box et al. 1978). Two levels, or
values, are selected for each parameter and a total of 2 n experiments are sufficient to test
the impact of each parameter, as well as every possible combination of parameters, on the
experimental output. By limiting the number of trials and still considering all combined
effects, the method is well suited for a rapid identification of the main parameters for the
model, which could then be studied or characterized in more detail. The application of
this method to the sensitivity analysis of numerical models has been very limited,
especially in the environmental domain, but it is well suited for that purpose. A good
example of the application of the factorial design to simulation results is presented by
Elliott et al. (2000), where the method was used to determine the kinetic parameters with
a significant effect on an anaerobic treatment process.
A large variability and uncertainty associated with several physical, chemical and
biological parameters was observed for the Saguenay Fjord. This situation requires a
detailed approach in order to identify the crucial parameters affecting the fate of
contaminants in sediments. To investigate the sensitivities of the model parameters that
are either variable or uncertain, a factorial design approach is presented here for the
model simulating the migration of contaminants in sediments of the Saguenay Fjord. The
purpose of the sensitivity analysis presented here is to deepen our understanding of the
effect of variable input values on the model output and to identify the factors that have
the greatest influence on the contaminant migration towards the new sediment-water
interface.
Model Equations and Parameters
The TRANSCAP-ID model simulates advection, diffusion, chemical reactions and

the effect of the burrowing activity of the benthic fauna. The consolidation of the
sediment and associated fluid advection are not included. The sediment is represented as
a dual porosity medium composed of sediment-pores and bio-irrigated tubes. The
mathematical formulation consists of two partial differential equation.,, corresponding to
the contaminant transport in the two media. These equations are coupled via a non-local
exchange term that represents the mass transfer of contaminant between the sediment and
the tubes. The formulation is mathematically similar to the one describing solute
transport in a dual-porosity medium, in the context of groundwater flow (tbr example,
Zheng and Bennett 2002).
The equation describing solute transport in the sediment porewater is given by:
(1)
R
where Cs and CT are the solute concentration [ML -3] in the sediments and the tubes,
respectively, ns is the sediment porosity [-], v~ is the fluid velocity in the sediment [LT-t],
170
R is the solute retardation factor [-], and fl is a first-order mass transfer coefficient [Tl].
The dispersion coefficient of the solute in the sediments, Ds [LZTt], is given by:
D s = D d +0t, .v s
(2)
where Dd is the effective diffusion coefficient [L2Tn] and aL is the sediment dispersivity
[L]. Note that when the fluid velocity is small in the sediment, which is the case for lowpermeability material, the dispersion coefficient is approximately equal to the diffusion
coefficient.
The equation describing solute transport in the tubes is given by:
nT_~ = nTDmc]2CT
~
~(Cs_CT)+ S
az-' -n~vT a--z-+
(3)
where nr is the porosity of the tubes [-], D,, is the dispersion coefficient for the tubes
[L2T-~], given by an expression similar to equation (2), and vr is the fluid or irrigation
velocity in the tubes [LTt]. In equation (3), S is a general source or sink term [ML3T -l]
representing the release of contaminant from mineral dissolution associated to bioirrigation. This term is defined later in this section.
The mass transfer coefficient fl and the porosity of the tubes nr decrease
exponentially with depth, following the distribution of bio-irrigated tubes (Martin and
Banta 1992). The mass transfer coefficient fl can be split iq two coefficients fl/[T l] and f12
[Ll], representing the maximal value of the mass transfer coefficient at the surface of the
sediment and the coefficient of exponential decrease respectively:
13(z) = 13,. exp~(z-
Zm..)]
(4)
As illustrated by Boudreau (1997), the value of the non-local term fll [T -l] at the surface
of the sediment can be estimated using the parameters of Aller's cylindrical diffusion
model (Aller 1980):
2. D d .r I
131 = (r~ - rl~)(8 - rI )
(5)
where r/is the inner radius of the tubes/burrows [L], r2 is the half-distance between two
tubes/burrows [L] and 8is the distance [L] from the burrow axis to a point where the
concentration equals the horizontally integrated value.
The porosity of the tubes n~ at the surface is calculated using cylinders to
approximate their geometry:
2
n~ = m . r I .n
where m represents the quantity of tubes per unit a r e a [L-2]. The exponential decrease
with depth of the tube porosity can be represented with the same coefficient f12 [L-~]:
(6)
DUERI AND THERRIEN ON NUMERICAL SENSITIVITY ANALYSIS
nv(z ) = n.r ~ . exp[[32(z - Zm~) ]
171
(7)
The source term S of the equation for the tubes (3) represents the chemical reactions
responsible for the remobilization of contaminants. The representation of the chemistry is
very simplified and the model only considers the dissolution caused by the reoxidation of
reduced contaminants adsorbed or coprecipitated with iron-monosulfide (FeS). The
dissolution is a consequence of the exposition of the anoxic sediment, containing sulfides,
to oxygenated water advected in the tubes by bio-irrigation. As observed by Furukawa
(2001) the irrigated burrow walls are the interface between oxic-anoxic milieus and are
characterized by steep geochemical gradients and rapid chemical mass transfer. In the
model, the dissolution term is composed of a kinetic factor 7 [L-l] and a coefficient L
[ML-3], representing the maximal dissolved contaminant concentration for a given FeS
concentration in the sediment. The representation of the chemical reaction was
specifically adapted for the dissolution and the remobilization of arsenic (As), which was
extensively studied in the Saguenay Fjord. In the anoxic zone, As is adsorbed or
coprecipitated with FeS, but the oxidation of the Fe-sulfides leads to the remobilization of
this contaminant. The coefficient L is derived from the linear correlation between the
released As and initial solid FeS that was observed during resuspension tests of Saguenay
sediments (Saulnier and Mucci 2000). The source term can be expanded in the following
way:
S = 7(L - CT)
(8)
L = kICF,s +-I%
(9)
where k/and k2 are regression parameters that can be viewed as dissolution coefficients
and CFes [ML 3] is the concentration of FeS.
Initial and boundary conditions are required to solve both equations (1) and (3). To
impose the initial conditions one specifies the initial solute concentration in the two
domains, sediments and tubes. Boundary conditions are required at the top and bottom of
the domain and can be either a prescribed concentration or prescribed mass flux for both
domains.
The two governing equations (1) and (3) are discretized with the finite volume
method. The equations are coupled through the mass exchange term and a simultaneous
solution for the concentration in the sediment and tube is obtained for each finite volume.
The system of equations is thus solved in a fully-coupling fashion, avoiding the use of
iteration that is necessary when the equations are decoupled. The block tridiagonal matrix
resulting from the discretization and assembly of the terms is solved with the Thomas
algorithm adapted to block matrices. The numerical solution was tested by comparing to
the semi-analytical solution MPNE of Neville et al. (2000), which has been developed for
one-dimensional solute transport in a dual-porosity medium with multiple nonequilibrium processes. The numerical model reproduces almost perfectly the
concentrations computed with the analytical solution (results not presented).
172
Design of the Sensitivity Analysis

The first step of the two-level factorial design is to determine the number of factors,
or parameters, that have to be tested and define the two levels, which are the minimum
and maximum value for each input parameter. Each factor is assumed independent of the
other factors, but a factor can be composed of a set of dependent input values. For n
factors at two levels each, a total of 2" simulations are necessary for a full factorial
design.
The choice of the parameters to be tested is based on the available information
about the processes affecting the migration of contaminants in the sediments of the
Saguenay Fjord. At the study site the sediments show a low permeability but are
intensively bio-irrigated. Thus bio-irrigation, dissolution and retardation are supposed to
influence the migration of contaminant. Consequently, a factorial design with five factors
was selected: three factors related to bio-irrigation (number of tubes per square meter,
depth of the tubes and bio-irrigation velocity) one factor related to the reactions
(dissolution coefficients) and the retardation factor. The fluid velocity in the sediment v~
was not considered for the sensitivity analysis because in the sediments of the Saguenay
Fjord there is no evidence for such advection. This parameter was therefore set equal to
zero for all simulations.
The minimum and maximum values of each tested parameter were determined from
the available information, which includes data measured at the Saguenay Fjord and data
derived from the literature. The values of the parameters are discussed in the following
paragraphs and are presented in Table 1.
The depth of bio-irrigation is represented in the model by two parameters: the
maximum depth attained by the tubes, biox, and the corresponding coefficient fie, where
fie describes the exponential decrease of both the mass transfer between tubes and
sediment fl and the tube porosity nr. The two parameters, biox and ,82, are related since
the maximum depth corresponds to the location where the mass transfer,8 and the tube
porosity nr exponentially tend to zero. According to field observations and to measured
bioturbation intensities (De Montety et al. 2000), the minimal value of the bio-irrigation
depth biox was set to 0.06 m and the maximal value to 0.26 m. This corresponds to values
of,82 equal to 70 m ~ and 16 m ~, respectively.
Table 1 - Minimum and Maximum Values f o r the Tested Factors
Factor
Parameters
Name (Units) MinimumMaximum
Bio-IrrigationDepth Bi
Depth of Tubes
Mass TransferCoefficient
biox (m)
132(m l)
0.06
70
Number and Dimensionof

Tubes B_,
Porosity of Tubes at Surface

Mass Transfer Coefficient
nr0(-)
131(dayl)
0.00039 0.00471
0.0157 0.119
0.26
16
Retardation Factor R
R (-)
45
Bio-IrrigationVelocityV
VT(m/day)
0.2
DissolutionCoefficientsD
kl (g-'mol i L-i) 1
k_,(p.g Lt)
1
1.7
1.3
DUERI AND THERRIEN ON NUMERICALSENSITIVITYANALYSIS
173
The number and the dimension of the tubes in the upper sediment layer are an
indicator of the intensity of bio-irrigation in this zone. This factor affects two input values
of the model: the tube porosity at the surface n~ and the mass transfer coefficient flj. To
calculate these parameters, one must first set the minimum and maximum values of the
inner radius of the tubes r~, as well as the number of tubes per square meter, m. The value
of m also defines the half-distance between two tubes r2 according to:
r2
1
_,-2~/m
(10)
The value of rj results from observation on sediments sampled with a box-corer in the
Saguenay Fjord during summer 1999 and 2000 and varies between 0.0005 m and 0.001
m. The minimum and maximum value of the number of tubes m, visually determined
from photographs of the undisturbed bottom sediments taken during the summer 2001,
are equal to 500 and 1500 tubes per square meter, respectivelly. Thus, the mass transfer
coefficient r / m a y vary between 0.0157 day ~ and 0.119 d a y , whereas the tube porosity
n~ varies between 0.00039 and 0.00471.
The value for the retardation factor, R , has been estimated for model calibration on
two measured arsenic concentration profiles and taking into account the average of the
range of values presented by Fuller (1978) for arsenic. The minimum and maximum
values of R have thus been set to 5 and 45, respectively.
The values for the bio-irrigation velocity vr were derived from data measured
during laboratory tests carried out on polychaete families living in shallow water
(Riisgard 1989, Riisgard 1991). In the calibrated model, the irrigation velocity i~ equal to
1 m day ~. The minimum and maximum values for the factorial design were determined
by modifying this value by a factor of five. Thus, we obtain a minimum bio-irrigation
velocity of 0.2 m day -I and a maximum value of 5 m day ~.
The minimum and maximum values of the dissolution coefficients were derived
from the values of dissolved As and solid FeS measured at two sampling stations of the
Saguenay Fjord in 1998 (Saulnier and Mucci 2000). To define the levels of these
parameters, it has been assumed that the dissolution coefficients, determined by means of
the measured data, correspond to the minimum and maximum values of the coefficient.
Thus, we assume that kj varies between 1 and 1.7 and that k2 varies between 1 and 1.3.
Four series of simulations were carried out for the sensitivity analysis (Table 2).
The first two series of simulations considered five variable factors, including a variable
dissolution coefficient, and were thus composed of 25 = 32 simulations. The series 3 and
4 simulated the response of the model for a non-reactive contaminant and the dissolution
coefficient was therefore equal to zero. This means that there were only four factors to
test and the factorial design was composed of 2 4 = 16 simulations. To evaluate the
influence of the thickness of the capping layer, two of the series were performed with a
cap thickness of 20 cm (series ! and 3) and the other two with a cap thickness of 10 cm
(series 2 and 4). The initial concentration of contaminant was extrapolated from the
arsenic concentration measured in the sediments of the Saguenay Fjord in 1998.
Each series was composed of simulations representing every possible combination
of the minimal and maximal values of each factor (Table 3). The results of the simulation
of each series were sorted in a standard order and thereafter the Yates algorithm was
174
applied in order to calculate the effect o f each factor and o f the c o m b i n a t i o n o f factors
(Box et al. 1978). T h e s e effects represent the change in the response as w e m o v e from the
m a x i m u m value to the m i n i m u m value o f a factor.
Description of the Series of Simulations for the Sensitivity Analysis
Table 2 -
Table 3 -
Series of
Simulations
Thickness of
Cap
Reactions
Number of
Simulations
I
2
3
4
0.2 m
0.1 m
0.2 m
0.1 m
Yes
Yes
No
No
32
32
16
16
Simulation Protocol for the Factorial Design of Four Tested Factors: Variable
Input Values Usedfor a Set of 16 Simulations
Simulation
Factors
Depth of Bio-Irrigation Quantity and Dimensions Retardation Bio-lrrigation
Bj
of Tubes B2
Factor R
Velocity V
biox
1~2
nro
I~u
R
VT
1
2
3
4
5
6
7
8
9
10
I1
12
13
14
15
16
0.06
0.26
0.06
0.26
0.06
0.26
0.06
0.26
0.06
0.26
0.06
0.26
0.06
0.26
0.06
0.26
70
16
70
16
70
16
70
16
70
16
70
16
70
16
70
16
0.00039
0.00039
0.00471
0.00471
0.00039
0.00039
0.00471
0.00471
0.00039
0.00039
0.00471
0.00471
0.00039
0.00039
0.00471
0.00471
0.0157
0.0157
0.119
0.119
0.0157
0.0157
0.119
0.119
0.0157
0.0157
0.119
0.119
0.0157
0.0157
0.119
0.119
5
5
5
5
45
45
45
45
5
5
5
5
45
45
45
45
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
5
5
5
5
5
5
5
5
Results
The sensitivity analysis was p e r f o r m e d over a simulation time o f ten years after the
capping event. T w o types of model output were c o n s i d e r e d for the analysis: the total
mass o f contaminant that left the system from the upper boundary, and the concentration
o f contaminant in the upper sediment layer at the end o f the simulation. The results o f the
sensitivity analysis were represented in a normal probability plot. This type o f plot is
175
generally used to test the normality of a distribution, but in our case it allows a
straightforward identification of the factors with a significant influence on the output of
the model. Each point on the graph corresponds to the effect of one or a combination of
factors. The effects represent the average variation of the response for a changing factor
over all conditions of the other factors and does not have units (Box et al. 1978). The
points that fall on or near the straight line are normally distributed and thus these effects
cannot be distinguished from the normally distributed error. On the opposite, the effects
that deviate from the straight line and therefore from the normal error distribution point
out significant factors. The sum of squares indicates the deviation of the effect from the
normal distribution and thus this value can be used to quantify the importance of a factor
for the response.
Contaminant Mass Released from the Upper Boundar 3, of the Sediment After 10 Years
The result of the first set of simulations, representing the migration of a reactive
contaminant through a cap of 0.2 m, is displayed in Figure 1. The graph illustrates that
the factors with the greatest influence on the released mass are those associated to bioirrigation (irrigation velocity V, number and dimension of tubes B2 and depth of the tubes
BD. There is also a second group of factors affecting the output of the model. This group
is composed of the combination of bio-irrigation factors with the factor of dissolution D.
V
+
B,V
B~V 6 ~
B~B:V ' \ .
~D
+
+
Ba
\ B,
>
BEB2
0
R-effects
-It
-2
-200000
200000
Effects
--
---~--]
400000
600000
Figure 1 - Effect of Tested Factors on the Mass Release of a Reactive Contaminant. Cap
Thickness of 20 cm. Significant Effects Identified by the Variable Name.

As these effects show a smaller deviation from the normal distribution, their influence on
the result is less significant. The result of the second series of simulations, representing a
reactive contaminant migrating through a cap of 10 cm, shows a very similar trend to the
176
previous series. The similarity is best illustrated by comparing the sums of squares,
presented in Table 4.
Table 4 - Sums of Squares for a Reactive Contaminant Migrating Through a Cap of 20
cm (Series 1) and a Cap of 10 cm (Series 2).
Sum of Squares %
Series I
Series2
V
B2
BzV
Bi
BtV
BtB2
BIB2V
22.47
19.10
16.13
11.96
10.08
8.46
7.23
22.53
19.01
16.17
11.79
10.10
8.57
7.25
SOFS Total
95.43
95.41
Figure 2 shows the results of the third series of simulations, representing a nonreactive contaminant migrating through a sediment covered by a cap of 20 cm. In this
case the contaminant release is strongly affected by the bio-irrigation depth BI, followed
by the combination of bio-irrigation factors B1, B2 and V. The retardation factor R also
exhibits a significant deviation from the normally distributed values. Figure 3 illustrates
the results of series 4, representing the effects for a non-reactive contaminant migrating
through a cap of 10 cm. In this case, the depth of the tubes Bt is the only important factor
identified on the graph.
+B~
+
B~V
1
e.~
+++ VBIB_,
B~V
t~
>
-1
R
+
-2
-8000 -4000
0
4000
Effects
8000
' 7
12000
Figure 2 - Effect of Tested Factors on the Mass Release of a Non-Reactive Contaminant.

Cap Thickness of 20 cm. Significant Effects Identified by the Variable Name.
2/
177
131
4-
//
| --
.ID
>
;/
/+
0-
.E
0
Z
-1
/7~//
/
/+
-2 -
- 10000
! 0000
20000
Effects
Figure 3 - Effect o f Tested Factors on the Mass Release o f a Non-Reactive Contaminant.
Cap Thickness o f 10 cm. Significant Effect Identified by the Variable Name.
D
+
B~
+g
++
132
i-
>
Z
-2-3
-4000
i
0
t
4000
Effects
i
8000
i
12000
Figure 4 - Effect o f Tested Factors on the Surface Concentration o f a Reactive

Contaminant. Cap Thickness o f 20 cm. Significant Effects Are Listed.
Contaminant Concentration in the Upper Sediment Layer After 10 Years
In this section, the sensitivity analysis was used to determine the significant factors
affecting the contaminant concentration in the upper sediment layer, from the four series
of simulations described previously. Figure 4 shows the result of the first series of
178
simulations, considering a reactive contaminant and a cap thickness of 20 cm. The graph
shows the significant influence of the dissolution coefficients D, the bio-irrigation depth
B~, the number and dimensions of the tubes B2 as well as the retardation factor R.
3
D
B~B2
"9--
+
4-
>
0-
B2 +~+
-2
-3 I
-20000
- 10000
0
Effects
10000
20000
Figure 5 - Effect of Tested Factors on the Surface Concentration of a Reactive
Contaminant. Cap Thickness of I0 cm. Significant Effects Are Listed.
/++/
>,
0i
-I I
-2 J +/,
-6000
-4000 -2000
0
Effects
r2000
4000
Figure 6 - Effect of Tested Factors on the Surface Concentration of a Non-Reactive
Contaminant. Cap Thickness of 20 cm.

Shown in Figure 5 are the results of the factorial design for a reactive contaminant
migrating through a cap of 10 cm. In this set of simulations the most significant factor is
the bio-irrigation depth Bt, followed by the dissolution coefficients R. The combination
179
of the depth and the number and dimensions of the tubes BIB2 is also an important factor,
whereas the retardation factor R does not play a significant role.
The series representing a non-reactive contaminant are illustrated in Figures 6 and
7. The graphs do not display any outliers and none of the factors shows a significant
deviation from the normally distributed population, represented by the straight line. Thus,
they do not reveal any significant factor affecting the concentration in the surface layer.
2~
.,..
0
4-
-2 !
-12000 -8000
-4000
0
Effects
4000
8000
Figure 7 - Effi~cts of Tested Factors on the Surface Concentration of a Non-Reactive
Contaminant. Cap Thickness of 10 cm.

Table 5 summarizes the significant effects highlighted by the sensitivity analysis.
The first two columns define the variable characteristics of the physical system (cap
thickness) and of the contaminant (possibility of dissolution). The third and the fourth
columns review the significant effects determined by the factorial design, depending on
the considered output variable, i.e., the released mass or the surface concentration.
Table 5 - Review of the Factors Having a Significant Effect on the Ot,tput of the Model.
Parameters
Response
Cap Thickness Dissolution
Released Mass
SurfaceConcentration
0.2 m
0.I m
0.2 m
0.1 m
V, B2, BI
V, B2, Bi
B~, V, R
B~
D, B2, Bi, R
Bi, D, B2
Yes
Yes
No
No
The contaminant mass released over a period of ten years is mainly affected by bioirrigation factors (BI, B2, V). The dissolution factor D does not display a significant
effect on this type of output. The comparison between the series with a cap of 10 cm
versus a cap of 20 cm indicates that the results depend upon the reactivity of the
180
contaminant. For the ranges of cap thickness tested in this study and considering a
reactive contaminant, the thickness variation does not lead to relevant differences in the
results of the factorial design. Conversely, the effect of the cap thickness is evident for a
non-reactive contaminant. The results for the surface concentration of contaminant ten
years after capping show that in the case of a reactive contaminant, a variable cap
thickness slightly influences the results of the factorial design. For a cap thickness of 20
cm, the most important factor is the dissolution coefficient D, whereas if the cap is only
10 cm thick, the depth of bio-irrigation BI becomes the most important factor. In both
cases, the bio-irrigation velocity V does not affect the results. Comparing the results of
both output variables, one observes that for a reactive contaminant the bio-irrigation
velocity V dominates the mass release, whereas the surface concentration is regulated by
the dissolution coefficients D. The model is generally very sensitive to bio-irrigation
parameters, except for the surface concentration of a non-reactive contaminant.
Conclusions
The factorial design has been successfully used to determine the most significant
parameters for the transport of dissolved contaminant through a capping layer. Two types
of responses were considered for the sensitivity analysis: the total mass of contaminant
that left the system from the upper boundary in the first ten years after capping and the
concentration of contaminant in the upper sediment layer after ten years. Both responses
represer,t a potential harm for the aquatic fauna, since they indicate an exposure pathway
for the organisms living in the marine environment. The results of the study pointed out
that for the studied conditions the significant parameters are a function of the response
variable, the contaminant type (reactive or non-reactive) and to a lesser extent the cap
thickness. The sensitivity analysis showed that the bio-irrigation parameters are generally
very important for the prediction of the released contaminant mass. It also demonstrated
that the concentration of contaminant in the upper sediment layer depends on the bioirrigation parameters as well as the dissolution coefficient.
The minimum and maximum values of each tested parameter were derived from the
data collected at the Saguenay Fjord and from the available information reported in the
literature;. The present study identifies the input values with a significant effect on the
result, thus highlighting the parameters that should be studied carefully in order to reduce
the uncertainty and get a better approximation of the performance of the capping layer.
However the most accurate characterization will not be able to eliminate the variability of
the input values. Therefore, an uncertainty analysis will be performed to determine the
distribution of the response considering the variability of the input parameters.
181
References
Aller, R. C., 1980, "Quantifying Solute Distributions in the Biotubated Zone of Marine
Sediments by Defining an Average Microenvironment," Geochimica et
Cosmochimica Acta, Vol. 44, pp. 1955-1965.
Box, G. E. P., Hunter, W. G., Hunter, J. S., 1978, Statistics for Experimenters: an
Introduction to Design, Data Analysis, and Model Building. John Wiley, New
York.
Boudreau, B. P., 1997, Diagenetic Models and Their Implementation: Modelling
Transport and Reactions in Aquatic Sediments. Springer-Verlag, Berlin Heidelberg.
De Montety, L., Long, B., Desrosiers, G., Cr6mer, J. F. and Locat, J., 2000,
"Quantification des Structures Biogbnes en Fonction d'un Gradient de Perturbation
dans la Baie des Ha! Ha! ~t l'Aide de la Tomodensitom6trie Axiale," Proceedings of
the 53th Canadian Geotechnical Conference, Vol. I, pp. 169-175.
Elliott, M., Zheng, Y. and Bagley, D. M., 2000, "Determining Significant Anaerobic
Kinetic Parameters Using Simulation," Environmental Technology, Vol. 21, pp.
1181-1191.
Fuller, W. H., 1978, Investigation of Landfill Leachate Pollutant Attenuation by Soils, US
EPA, Municipal Environmental Research Laboratory, Cincinnati, OH.
Furukawa, Y., Bentley, S. J. and Lavoie, D. L., 2001, "Bioirrigation Modeling in
Experimental Benthic Mesocosms," Journal of Marine Research, Vol. 59, pp. 417452.
Kabala, Z. J., 2001, "Sensitivity Analysis of a Pumping Test on a Well with Wellbore
Storage and Skin," Advances in Water Resources, Vol. 24, pp. 483-504.
Martin, W. R. and Banta, G.T., 1992, "The Measurement of Sediment Irrigation Rates :A
Comparison of the BR- Tracer and 22ZRN/Z26RADisequilibrium Techniques,"
Journal of Marine Research, Vol. 50, pp. 125-154.
Neville C.J., Ibaraki, M. and Sudicky, E.A., 2000, "Solute Transport with Multiprocess
Nonequilibrium: a Semi-Analytical Solution Approach," Journal of Contaminant
tlydrology, Vol. 44, pp. 141-159.
Riisgard, H. U., 1989, "Properties and Energy Cost of the Muscular Piston Pump in the
Suspension Feeding Polychaete Chaetopterus Variopedatus," Marine Ecology
Progress Series, Vol. 56, pp. 157-168.
Riisgard, H. U., 1991, "Suspension Feeding in the Polychaete Nereis Diversicolor,"
Marine Ecology Progress Series, Vol. 70, pp. 29-37.
182
Saulnier, I. and Mucci, A., 2000, "Trace Metal Remobilization Following the
Resuspension of Estuarine Sediments : Saguenay Fjord, Canada," Applied
Geochemistry, Vol. 15, pp. 191-210.
Zheng, C. and Bennett, G.D., 2002, Applied Contaminant Transport Modeling. 2 nd
Edition, John Wiley, New York.
Osman A. E1-Rayis l
Steps for Restoration of a Polluted Egyptian Closed Lagoon on the Alexandria
Mediterranean Coast
Reference: E1-Rayis, O. A.," Steps for Restoration of a Polluted Egyptian Closed

Lagoon on the Alexandria Mediterranean Coast," Contaminated Sediments.

Performance, ASTM STP 1442, J. Locat, R. Galvez-Cloutier, R. C. Chaney, and K. R.
Abstract: Lagoon (or Lake) Maryout Main Basin is one of the lagoon's four sub-basins,
polluted with organic and metals contaminants from untreated sewage and industrial
effluents from Alexandria City. Since 1993, these effluents have become primary treated
before discharging into this basin through two point sources; one source is at a new site,
as a step in rehabilitating the lagoon. Surface sediment samples were collected 3 years
later, from the Main Basin as well as from the other three basins of the lagoon to
determine the level of nine priority metal pollutants as well as organic carbon (Org-C).
The data reveal that the sediments of the Lagoon Main Basin, particularly that in vicinity
of the new point source, became spoiled and remarkably enriched with all the studied
contaminants. In addition, some of the studied metals are at levels comparable to those of
the median effect-range of Long and Morgan recommended by the EPA. This means that
either diversion or advanced treatment is necessary to eliminate these effluents.
Keywords: restoration, polluted Lagoon Maryout, sediments, organic carbon and metals,
Alexandria, Egypt
Introduction
Lagoon Maryout is a closed lagoon without a natural connection to the
Mediterranean Sea. Most of the lagoon is situated south of Alexandria and is completely
separated from the sea by a high (to - 20 m), rather well lithified carbonate ridge of late
Pleistocene age (Ali and West 1983). The long, narrow extension of the lagoon southwest
of Alexandria occupies the lowest part of the depression between two carbonate ridges
(named Abu Sir and Gable Maryout ridges). This lagoon extension parallels the coastline.
Lagoon Maryout is presently a small remnant of its precursor and now is artificially
divided with dikes into four sub-basins named Main, Fisheries, Northwest, and Southwest
Basins (Figure 1). The respective areas of these basins are 6000, 1000, 3000 and 7500
acres. The average depth of the lagoon is about 1 m.
1Professor, Oceanography Department, Faculty of Science, University of Alexandria
Moharrem Bek 21511, Alexandria, Egypt.
183
www.astm.org
184
East Treatment Plant

ETP
AN
MEDITERRANE
SEA
ALEXANDI
City
Center
EIMex
Bay~
Kalaa u,s=n
}" ~"~"_.,~c ~
/'~[
o1~ 9
06
CULTIVATED
LAND
L = LAKE
MARYOUT
M = Main
I~
N= Northwest ~
15/
S = Southwest[ _~
F = Flshertes
.
,,-'
After EL-RAYI$ (Petaorlal Cqmm,]
~'
'/
,=_,w..,..ut..
since July 1993.
o. . . 2.
~km
Figure 1- Lagoon Maryout and Its Four Basins and Locations of the Sampling Stations.
The Main Basin, since 1993, receives flow from three main sources: (1) Kalaa
(agricultural) Drain that also carries the discharge from a treatment plant called East
Treatment Plant (ETP), (2) Directly from the West Treatment Plant (WTP) outlet, and (3)
Omoum (agricultural) Drain. The other three basins receive flows from a variety of canals
and drains but do not receive arty effluent from the ETP or WTP. It is worth mentioning
that the water level in the lagoon and in the Omoum Drain downstream portion must be
regulated and always kept at 3.7 m below the sea surface level to allow for the drained
waters mainly from the neighboring agricultural lands, of Alexandria and Bohaira
Provinces, to flow gravitationally into them. This is accomplished by the El Mex
Pumping Station (Figure 1), which pumps the water from the downstream portion of the
drain across the lithified ridge on to the sea at El Max Bay. Before 1993, the lagoon Main
Basin received most of the raw industrial and sewage effluents of Alexandria City,
without treatment, through four drains (Figure 1), named Industrial, Gheit EI-Enab and
Kabbary Drains at the northern side of the lagoon beside the Kalaa Drain.
The historical sediment quality data and metal concentrations in the sediments from
Lagoon Maryout, before 1993, have been reviewed by EI-Rayis and EI-Nady (1997). The
bottom of the lagoon in 1968 (El-Wakeel and Wahby 1970) was covered with a badly
sorted sand-silt-clay sediment. The lagoon sediments were grayish to blackish gray in
color, having smelled of hydrogen sulfide in regions covered with rooted water plants.
The average median grain size of sediments in the Main, Northwest, Southwest and
EL-RAYIS ON RESTORING EGYPTIANCLOSED LAGOON
185
Fisheries Basins are 58, 6, 28 and 26 microns, respectively. These sediments were derived
mainly from the suspended load of silt and clay carded into the lagoon through the drains,
enriched with calcareous shells and constituting a sand-silt clay admixture. The sand is
constituted of the calcareous shells and the shell fragments of lamellibranches, gastropods
and tubeworms. These cover large areas of the lagoon bottom. The carbonate (as calcium
carbonate) content of the lagoon sediments averaged 41%. The organic matter content in
1968 averaged 8%, reflecting the high productivity of the lagoon in that date.
Over time, especially after 1960 -after industrialization of Alexandria - it seems
likely that the nature of the bottom of at least the Main Basin was altered from its original
condition. This by the continuous supply of sewage and industrial wastes, carrying large
loads of suspended matter, from the above mentioned four main point sources (Figure 1).
The respective suspended matter loads from the Kalaa, Industrial, Gheit E1-Enab and
Kabbary Drains were 42500, 433100, 30100 and 110400 kg/d (Environmental Impact
Statement CDM, 1978). After settling, this suspended matter eventually formed the
sludge bed on the bottom of the Main Basin. In 1987, EI-Rayis and E1-Sabrouti (1997)
studied the organic matter percentage (see Figure 3k) of the sediments of the Main and
Northwest Basins and found respective averages 19.8 and 8.5%. Also, they found that the
sediments containing high organic matter content (>18%) are those at the eastern half of
the Lagoon Main Basin particularly those in front of the fore mentioned four point
sources. The concentration of metals (Cu, Zn, Fe and Mn) in the sediments collected from
the Lagoon Main Basin in 1979 was measured by E1-Rayis and Saad (1990). They showed
that the metal concentrations tend to be also higher at the far eastern side of the lagoon
mimicking the 1987 distribution of organic matter (ElrRayis and E1-Sabrouti 1997). This
deterioration of the lagoon conditions especially the evasion of the malodorous hydrogen
sulfide gas has impacted the fish diversity, quality and quantity. This has led to a real
social problem to the local fishermen besides the loss of a considerable source of animal
protein.
The present work is a study of the level and distribution of eleven elements including
Cu and Zn and Org-C in the sediments of the Main and the other basins of Lagoon
Maryout collected in 1996. This is as a part of a pilot study to provide an understanding
and evaluation of the existing conditions in the lagoon Main Basin after the erection of
the ETP and WTP. The water quality of the lagoon is studied and mentioned elsewhere
(E1-Rayis et al 1998), the distribution of some of the water quality parameters in the
lagoon waters are presented in Figure 2. The present results, however, beside those for the
water quality of the lagoon may be of value to the decision makers in choosing a solution
for next phase of restoration of this important lagoon.
Twenty-three locations were sampled on Lagoon Maryout. Fourteen of the stations

were distributed throughout the Main Basin, LM, and there were three stations in each of
the Fisheries, LF, Northwest, LN, and Southwest, LS, Basins. The locations of the
sampling stations are shown in (Figure 1). All sediment samples were collected with a 6"x
6" petite Ponar dredge that sample an area of 0.023 square meters and to a depth of about
20 cm. The geotechnical data collected from borings in the lagoon to gather information
on stratigraphy of its sediments indicated that up to 30 cm sludge is found on the bottom
186
of the Main Basin, accumulated along the last forty years, (Geotechnical Report 1996).
The solid sediment (macrophytes were discarded) was placed into an appropriately
labeled plastic container and stored at 4~ until ready for shipment to EAESTI's
Analytical Laboratory in Sparks, Maryland, USA. Analyses were performed using
Standard Operating Procedures Manual (SOPs) based on guidelines provided by the USEnvironmental Protection Agency (SW 846 Method) and APHA, 1995. Method detection
limit for the studied trace elements was < 0.9 ug/g dry weight sediment. Blanks were
carried through the entire sample preparation and analytical process. Also duplicate
samples were processed on a routine basis to determine precision. In addition spiked
samples and reference materials were employed to determine accuracy. Finally the
"
II
."
mg/,
9r 0 o ~
Figure 2 -Distribution of the Values of Some of the Water Quality Parameters in the
Surface Waters of Lagoon Maryout, 1996.
187
EL-RAYIS ON RESTORING EGYPTIAN CLOSED LAGOON
concentration o f all calibrated standards were verified against a quality control check
sample obtained from outside source. The reference materials 2704 Buffalo River
sediments and 2710 and 2711 Montana Soil o f high and normal levels respectively were
used.
Analytical Results (and Matrix Interrelation Coefficients) of the Studied

Metals (mg/kg Dry WeighO and Organic-C (OC %) in the Sediments of Lagoon
Maryout.
Table 1 -
Stat.No
Hg
Sb
Ag
Ni
Cr
Pb
Zn
Cu
Cd
OC
.......L.MO.! .............. !..!.0 ..............3...3..0............ ~4.Zo ............. 6.6.8 ........... !.05.0....o....:3.700...o....10..5.0 .......... 62.5.. ............ 2.6.0 ..............9:.6. .....
LM02
350
.......i2~SYTiii;i
10.7
.......................F....i.~.i~....T...i.~.6.-[.....5~.~......T....i..:)5i~....I......~.6.g...T..i.gF....T...i..~6..-1......i.~.~..i.
....
..... E~i;iT
.......................i .......................T.......................t ........................i ...................... T.......................i .....................T......................i .......................!i~i~
......f~ii~i
...............................................i .......................i .......................................................................iii~g
LM06
LM07
LM08
! 32
........... i~i ............... ;i~'''''"~'71
....
......
110 ~.i......................
190 i~ .......................................................................
355
432
672 i~ ..........................................................................................
240
139
94
10.6
~
! 129
77
41
8.6
140
220 i 519
545
847 i 213
129
69
13.4
.............................. ~...............................................
36
......LMO.9.. .............. .2.8 ...............2.20 ............. !..!.0 ..............2.!.7 ................3...1.9............. .42! ............. !.,82.. ............. 7.2.. ................ 6.8. ................9:o ........
LM 10
73
240
140
442
662
3200
986
395
226
22.4
......LM.!..! ...........................................................................................................................................................................................................................
LM12
130
370
18
35
27
13
16
4..9. ......
11.2
......L.M..!..3_ .............. .2.:! .................4.1. ............... _1.60 ............. 400 ..............451. ............. 2..8.0 ............ !..1.5............... 4.8. ................ 32.. ............. 3...7..:.8......
......LNO.! ........................................ 3.6 ............... !,5,.0 ............ 2.,0.6 ............. 2.2.8 ............. .!..!.4 .............. 8.6 ............... 2...!.........................................9....4......
LN02
17
52
160
163
178
101
74
23
14.7
.....L'gi'i)3...............3] ................-6].............25"0................7"i...............i(i7 ..............75"'T"'5 g................2'5.......................................f ' i 3
LF01
15
48
110
85
133
95
! 81
4.5
.......L F02 ............... 2+o .................69 ............. t So ............... 92 ................ 95 .............. 2o4 .............. s4 ................ V ........................................ 6:s ........
LF03
19
73
130
73
89
110
101
21
5.2
.......L SOJ ................ ts .................95 ............... 2o0 ............. 37 ..................66 ................ ~2 ............... 43 .................. 5 .......................................... 43 ........
LS02
18
100
31
21
42
21
3.0
.......L.S.O.3. ................ !.7 .................4..[ .............. !.40 ............. !..5.6 ...............!.9.8 ............. !.3.2 .............. .6.8................. !..S ........................................6.:2 ........
H~
Sb
Ag.
1.0
0.85
0.57
1.0 i 0.54
i 1.0
0.83
0.63
0.43
0.91
0.76
0.50
0.89
0.86
0.50
0.90
0.84
0.46
0.90
0.86
0.58
0.90
0.84
0.35
0.22
-0.04
0.05
......... N! .......................................................................................... l:O ..............oR8 ........... 0:79 .......... 0:78 .......... .o..82 .......... 0-!.! ........... 0:48 ......
.......
Cr
1.0 i 0.89
0.89
0.91
0.88
.......... ,.Z.n.............1.......................i.............................................. [ ................................................i .............................. t:..~176
Cu
...........
i
........... .......................
OC
i
.......................
i
.......................
0.41
...........o...:g ......
i 1.O i 0.96 . 0.23
i ...............................................
i ............................................
! .......................
......
i 1.0
188
The results of the lagoon sediment analyses for different elements are presented in
(Table 1) and as areal distributions shown in (Figure 3).
-2
x}O mg/kgDW
.....
-Z
xlO mg/kg DW
< 2O
-1
-1
xl0 mg/kg
xl0 m g / k 9
@
,,~0
<IO0
rng/kg
-2
le mg/k9
~.
Figure 3 a - f - Areal Distribution of Some Metals and Organic Carbon in the Surface
Sediments of Lagoon Maryout, 1996
189
/// / h
mg/kg
Figure 3 g-k - Areal Distribution of the Other Metals and Organic Carbon in 1996 (and
of the Organic Carbon Studied in 1986) in the Surface Sediments of Lagoon Maryout.
From the figure, one can notice that most of the highest metal concentrations are
t~ound in the sediments of the lagoon Main Basin compared with those in the sediments of
190
the other three basins. In addition, within the Main Basin these high values particularly
those of the metals are generally found in two quite distinct areas. One at its eastern side
in front of the outlet of Kalaa Drain and the other at the northwest comer, in vicinity of
the WTP discharge site. This distribution pattern also correlated with the water quality
parameters (Figure 2). Silver showed a slight deviation in the position of the polluted
area in front of the WTP as it become more westward in the sediments close to the course
of Nobaria Canal (Figure 3). The high values of the Org-C in the Lagoon Main Basin are
mainly for the sediments at the northwestern side off the WTP outlet. This is reflected on
the interrelation coefficients between the studied parameters in the sediments (Table 1).
Apart from these, and as mentioned, the other sediments of the other basins are less
contaminated with these studied elements.
Comparing the current figures of the Org-C and the metals Cu and Zn (Figures 3)
with their counterparts obtained before 1986 (EI-Rayis and Saad 1990 and E1-Rayis and
EI-Sabrouti 1997) one can easily note that: (1) The sediments contaminated with metals
after 1993, the erection time of the WTP and its discharge site, become not only restricted
to the eastern side of the Lagoon Main Basin but also to those in a new sediment area at
the northwestern side in front of the new point source. (2) The high concentrations of the
Org-C after 1993 become restricted more to the sediments offthe new source WTP outlet,
while the sediments at the eastern side of the Main Basin, particularly those off the old
source Gheit EI-Enab, besides those off the Kalaa Drain, show quite lower contents of
Org-C. This means that after 1993 the distribution pattern of the Org-C actually is not
mimicking that oftbe metals (Table 2), unlike before 1993. These results refer to the high
impact of the direct discharge from WTP, compared to the other discharge from the ETP
that is indirectly reaching the lagoon through Kalaa Drain, in the nearby sediments. In
addition, there is a degradation of considerable proportion of the organic matter from the
spoiled sediments in front of the three old point sources after about 3 years from blockage.
-It is worthy to assure that the ETP does not dispose its effluent directly into the lagoon
but on to Kalaa Drain. It discharges its effluent into the Kalaa agricultural Drain, at a site
distant about 2 km landwards from the drain outlet, i.e. it mixes with the agricultural
drainage water before reaching the lagoon-. Obviously, this process beside the dilution
effect it gives a better chance for occurrence of a partial degradation to the organic matter
load (its biodegradable part, BODs, on average represents about 40 % of the total, COD,
bio plus non-bio degradable, E1-Rayis et al 1998) with a possible precipitation of
considerable part of the suspended particulate organic matter to the bottom of the drain
before reaching the lagoon.
The noticeable good association of most of the studied metals (significance p > 0.5)
with each other rather than with organic matter confirms their immobilization and the
degradation of the organic matter in the sediments.
The levels of the nine studied metals here are compared with those previously
reported in literature for Lagoon Maryout. For example, Sb concentrations in the Main
Basin in the present work (1996) ranged from 0.6-6.1 mg/kg compared to WWCG (1994)
ranges of 20-71 mg/kg in 1993. The other basins of the lagoon in 1996 exhibited a
concentration range of 0.41-1.2 mg/kg compared to the values of 13 and 21 mg/kg
reported in WWCG (1994). Arsenic in sediments of Main Basin and other basins ranged
from 0.35-11.3 and 2-4.7 mg/kg compared to 0.06-0.1 reported by Mostafa (1994) and 4-
1 91
16 mg/kg by WWCG (1994). Cadmium in 1996 shows concentrations ranged from

0.42-2.6 mg/kg in Main Basin and from 0.38-0.53 mg/kg in the other basins. There
are no literature concentrations for this metal. Chromium concentrations in the Main
Basin and other basins in 1996 ranged from 3.5-105 mg/kg and 2.1-22.8 mg/kg
respectively. These concentrations are similar to those found earlier and reported in the
literature (for example WWCG 1994). Copper in the sediments of the Main Basin in 1996
ranged from 5.6-625 mg/kg. This concentration range is similar to literature data except
for that reported in WWCG (1994). Copper measured in other basin sediments in 1996
ranged from 3.4-31.4 mg/kg, similar to literature data. Lead concentrations in sediments
of Main Basin in 1996 ranged from 2.7-370 mg/kg and from 4.2-20.4 mg/kg in other
basins. Similar to the discussion related to literature reported Pb in sediments, the
concentrations reported by WWCG (1994) of 110-3200 mg/kg is considerably higher than
those found in 1996. Mercury concentration range in 1996 in the Main Basin was 0.28-2.1
mg/kg that are similar to 1991-1992 and 1993 data that show a concentration range of
0.18-1.77 mg/kg. In 1996, three samples in other basins exceeded the detection limit and
ranged from 0.37-0.60 mg/kg. Nickel in 1996 sediments of the Main Basin ranged from
1.8-66.8 mg/kg that is closest to the data reported by the United States laboratory results
in WWCG (1994) of 31-107. Nickel in other basins ranged from 4.6-20.6 mg/kg. Silver in
Main Basin in 1996 ranged from 2.1-8.9 mg/kg, slightly lower than Ag in 1993 that
ranged from 3.4-11.8 mg/kg. Relatively high concentrations of Ag found in the sediments
of the Southwest Basin (2-13.5 mg/kg). Finally, Zn in 1996 the main Basin shows
concentrations ranged from 12.6-1050 mg/kg while other basin concentrations ranged
from 13-116 mg/kg. Main Basin literature Zn concentrations were similar to those found
in 1996 (100-1000 mg/kg, EI-Rayis and Saad 1990; WWCG 1994).
The elevated concentrations for the metals Hg, Ni, Pb, Cd, Zn and Cu (Table 2) are
found to be at levels above those corresponding of the median effect-range of Long and
Morgan recommended by the EPA. This may have possible effect on bottom fauna and
other organisms (including edible fish) that live in this polluted basin. Subsequently this
situation makes the Main Basin as a possible source health-hazard for fish and humans
who consume them.
Table 2 - Levels of Priority Metal Pollutants in the Lagoon Maryout Main Basin
Sediments with Respect to Those of Median (ER-M) and Long (ER-L) Effect-Range
9f Long ant
~n Recommended the EPA.
Element No.of
No. of
ER-L
No.
ER-M
No.
Highest
results
detects
mg/kg
above
mg/kg
above
value
ER-L
ER-M
mg/kg
Sb
15
12
2.00
6
25.00
0
6.10
As
15
15
33.00
0
85.0
0
11.30
Cd
15
10
5.00
0
9.00
0
2.60
Cr
15
15
80.00
1
145.00
0
105.00
Cu
15
15
70.00
12
390.00
2
625.00
Pb
15
15
35.00
12
110.00
5
370.00
Hg
15
12
0.15
12
1.30
2
2.20
Ni
15
15
30.00
7
50.00
3
66.80
A8
15
9
1.00
9
2.20
8
8.90
Zn
15
15
120.00
12
270.00
7
1050.00-"
m ,
192
The previous and aforementioned two important findings seem to point to the
following conclusions: (1) Most of the organic wastes in the lagoon are biodegradable.(2)
Most of the metals in the sediment are immobilized and most probably non-bioavallable,
unlike organic matter. Therefore, dredging of the spoiled sediments highly contaminated
with the toxic metals is seriously recommended. (3) The waste effluent from the WTP
must be either further treated (at least secondary treated) prior to discharge into the
lagoon or diverted.
The suggested diversion of the sources in addition to dredging of the lagoon's
spoiled sediments is adopted by the government for the next step for rehabilitation of the
lagoon. Further, the diverted effluents of the two treatment plants will be subjected to a
secondary treatment process before reuse for inland (neighboring desert) irrigation.
Acknowledgment
This work is carried out for the Alexandria General Organization for Sanitary
Drainage (AGOSD) and the U.S Agency for International Development (USAID) as part
of Alexandria Wastewater Project - Phase II (USAID Contract No. 263-0100-C -005117-00). The grant offered by USAID for carrying out this research work is gratefully
acknowledged.
References
Ali, Y. A. and West, I. A., 1983, "Relationships of Modem Gypsum Nodules in Subkhas
of Loess to Compositions of Brines and Sediments in Northern Egypt", Journal of
Sediment Petrology, Vol. 53, pp. 1151-1168.
APHA (American Public Health Association), 1995, Standard Methods for the
Examination of Water and Wastewater, 19th ed.
EI-Rayis, O. A. and EI-Nady, F. E., 1997, Chapter 3 in: Technical Report of Chemical and
Biological Characterization of Lagoon Maryout, Alexandria Wastewater Project
Phase II. Prepared for Metcalf & Eddy International.
EI-Rayis, O. A., EI-Nady, F. and Hinckely, D., 1998, Existing Environmental Conditions
in Lagoon Maryout South of Alexandria, Egypt: 1- Water Quality. Proc. 8th Int.
Conf. "Environ. Prot. Is a Must ". Alexandria 5-7 May 1998, NIOF, Euro-Arab
Coop. Center (VEA), Int. Sci. Assoc. (ISA) & Social Fund for Development (SFD),
pp. 33-43.
EI-Rayis, O. A. and Saad, M., 1990, "Heavy Metal Pollution in Lagoon Maryout, on the
Southeastern Coast of the Eastern Mediterranean Sea", Journal of King Abdulaziz
Universityfor Marine Sciences, Jeddah, Vol. 1 pp. 17-26.
E1-Rayis, O. A. and EI-Sabrouti, M. A., 1997, "Lake Maryout: Pollution Problems and
Proposals for Restoration", Fresenius Environmental Bulletin, Vol.6, pp.598-604.
EL-RAYIS ON RESTORING EGYPTIANCLOSED LAGOON
193
E1-Wakeel, S. K. and Wahby, S. D., 1970, "Texture and Chemistry of Lake Mariut
Sediments", Archive .Hydrobiology, Vol. 67, pp.368-395.
Environmental Impact Statement (CDM), 1978, Environmental Impact Statement for the
Alexandria Wastewater Plan Study. Vol. I, Prepared by Camp Dresser and Mckee
(CDM) Inc., Environmental Planning Division with Marine Environmental Services
CHAS. T., Main International Included Arab Technology and Economic Consulting
Office.
Geotechnical Report, 1996, Technical Memorandum No. 3. Volume 2 "Lake Maryout
Geotechnical Data Report", Alexandria Wastewater Project - Phase II, USAID
Project No. 263-0100.
Mostafa, A. H., 1994, Environmental Studies on the Water and Sediments of Lake Mariut,
with Reference to Pollution Chemistry, Ph.D. Thesis Alexandria University, 396 p.
US-Environmental Protection Agency, 1986, Acid Digestion of Sediments, Sludge and
Soils, SW 846 Method, Third Edition, September 1986.
WWCG (Wastewater Consulting Group), 1994, Report on Preliminary Investigations of
Lake Maryout Water Quality: Past, Present and Future, 1994.
Moir D. Haug, I S. Lee Barbour, l and Chris A. Johns 2

Rehabilitation of Brine-Saturated Sediments
Reference: Haug, M. D., Barbour, S. U, and Johns, C. A., "Rehabilitation of BrineSaturated Sediments," Contaminated Sediments: Characterization, Evaluation,

Locat, R. Galvez-Cloutier, R. C. Chancy, and K. R. Demars, Eds., ASTM International,
Abstract: A large quantity of brine-saturated fine-grained sediments (insolubles) is
produced each year as a by-product of potash mining. These sediments have traditionally
been disposed of in clay-lined ponds to minimize their environmental impact. Many of
these ponds are now nearing capacity and there is a desire to find techniques to wash the
brine from these sediments so they can be taken outside the waste management areas and
used for other purposes. Laboratory studies have shown that brine can be removed from
these insolubles by flushing with fresh water; however, the feasibility of de-brining by
this method had not been established at field scale. As a result, a field study of three
prototype de-brining systems was conducted at the Potash Corporation of Saskatchewan
Inc. (PCS) Allan Potash mine. The three treatment systems investigated were: a gravity
drainage pond, a reverse-flushing pond with a sand and gravel filter, and a reverseflushing pond with a geotextile filter.
Keywords: potash, reclamation of insolubles, slimes, brine removal, gravity drainage
pond, reverse-flushing pond
IProfessor, Department of Civil Engineering, 57 Campus Drive, University of

Saskatchewan, Saskatoon, SK, S7N 5A9.
a Geo-Environmental Engineer, MDH Engineered Solutions, 111 Research Drive,
Saskatoon, SK, S7N 3R2.
194
www.astm.org
HAUG ET AL. ON BRINE-SATURATEDSEDIMENTS
195
Introduction
The disposal of waste products from potash mining is facing increased attention
from both industry and regulators. Laboratory studies have shown that brine can be
removed (flushed) from insolubles with freshwater permeation; however, the feasibility
of de-brining by this method has not been established at field scale. This paper describes
the results of a pilot-scale field study conducted to evaluate the effectiveness of different
pond drainage system designs for brine removal. The primary issues explored during this
study involved the determination of how efficiently such a pond drainage system would
perform in the field, and which type of drainage/pond design is best suited to maximize
performance.
The approach that was used for this research was to design, construct and
instrument three test cells at PCS Allan, Saskatchewan (Figure 1). One of the test cells
was designed as a gravity drainage pond in which the slimes were surcharged with fresh
water to flush the brine into an underdrain system. The other two test cells were designed
as reverse-flushing systems in which the underdrain system was pressurized with fresh
water to flush the slimes in an upward direction. These test cells were designed based on
seepage and contaminant transport modeling to evaluate field performance over a
relatively short time period. The layout of the treatment system is shown in Figure 2.
Figure 1 - Photograph o f a Test Cell at PCS Allan, Saskatchewan
196
Background
Traditionally, in the treatment and disposal of potash mine waste, the waterinsoluble fine-grained wastes (slimes) have been mixed together with brine and coarse
salts and pumped as a slurry to tailings piles for deposition. This practice was generally
considered to be the most economical method of waste disposal. Potash mining in the
province has grown however, and several mines now produce significant amounts of
slimes as a by-product of potash production. These slimes contain high concentrations of
sodium chloride (salt) brine as well as water insolubles. According to Hart (1988), the
insoluble portion of this waste can be up to 10% of the total waste mass. These water
insolubles come from "clay" seams mined adjacent to the ore and consist of sand, silt and
clay sized particles. It is now understood that the long-term disposal of the slimes at
surface facilities would be more easily accomplished if the salt concentrations were
reduced. Recent unpublished research conducted by PCS Inc. has shown that it may be
possible to reclaim the insolubles portion of the waste stream if the sodium chloride
saturated brine is separated from the insolubles.
Figure 2 - Slimes Washing Pilot Project Site Plan, Including Slimes Sampling Points
HAUG ET AL. ON BRINE-SATURATED SEDIMENTS
197
Characteristics of Slimes
The characteristics of potash insolubles have been the focus of several studies. A
study investigating the potential for using insolubles as a component in a tailings pile
cover material was carried out by Haug (1988). Insolubles from the PCS Allan mine
were tested to determine the composition of slimes. Table 1 summarizes the results of
this analysis, which found that the major mineral components of potash insolubles at PCS
Allan are dolomite, halite, and sylvinite.
Table 1 - XRD Analyses of PCS Allan lnsolubles (after Haug 1988)
Component/Mineral
Proportion (%)
Quartz
Sylvinite
Dolomite
Halite
Anhydrite
Kaolinite
Illite
10.8
19.9
27.4
21.3
12.5
6.2
1.9
Brine from the thickener discharge pipe at the PCS Allan mine was also analyzed
as part of the 1988 study. Table 2 shows the results of this analysis, which found that the
major mineral constituents of the brine are sylvite and halite.
Table 2 - Chemical Analysis of PCS Allan Process Brine (after Haug 1988)
Constituent
Percentage
Sylvite
Halite
Calcium
Magnesium
Sulphate
Water
10.8
18.7
0.17
0.13
0.17
70.03
The hydraulic conductivity (k) of these "washed" slimes was measured with a
constant head petmeameter and found to be in the range of 9 x 1 0 -6 cm]s at a density of
1070 kg/m 3 and a water content of 29%. This test was run on 270 mm diameter samples
under a gradient of 2.7 for approximately 2.5 hours. In contrast, the k of the natural
(containing salt) insolubles using fresh water was found to be 3 x 10 -5 cm/s on a similar
sample under a gradient of 4.2 for 8 hours. Theoretical calculations of k based on
consolidation testing of the washed insolubles suggested that the hydraulic conductivity
of the insolubles should be in the 3 x 1 0 .7 cm/s range. The slimes layer hydraulic
198
conductivity during treatment in the test ponds was expected to play a critical role in the
effectiveness of each system.
Drainage Systems Design and Operation
Three treatment systems were evaluated in this project: a gravity drainage pond, a
reverse-flushing pond constructed with an aggregate filter, and a reverse-flushing pond
constructed with a geotextile filter. The gravity drained pond provided the simplest
design and effort of operation. However, the reverse-flow systems had several attractive
features including the following:
It was expected that there would be reduced consolidation of the slimes during
treatment and that the hydraulic conductivity of the sediments would remain
higher than those in the gravity pond. This would be most advantageous for
long-term treatment of thick layers of slimes.
If additional untreated slimes layers were added to the upflow system, the clean
water would flow through the cleanest slimes first. This would result in fresh
untreated sediments being added without adverse contamination of the cleaned
sediments below.
This system would allow for staged filling-flushing operations. The hydraulic
head applied to the underdrain system would be the only limiting factor to the
depth of slimes that could be effectively flushed.
Upward flushing away from the drainage system would minimize clogging of
the drainage system with fine-grained wastes or precipitates.
All three ponds were constructed with base dimensions of three meters by five
meters, to a depth of approximately three meters (Figure 3). The pond side slopes were
cut to an angle of 2 horizontal: 1 vertical. The sidewalls and the bottom of the ponds were
lined with a 60 mil high density polyethylene (HDPE) geomembrane liner. Detailed
design of the aggregate filter drain and geotextile followed the published
recommendations provided by Terzaghi (1929), U.S. Army Corps of Engineers (1955,
1975), and the U.S. Bureau of Reclamation (1987).
2:1 SS
A[' 2:1SS
3
/ ~
2:1 SS
-1A'
~-5.00
~ ss2 m~
: t
0.30 m~ ~
0.45 m
Note: 5 % slope on bottom of Pond #l
- Not to Scale Figure 3 - Upward and Downward Flow Systems - Plan View
HAUG ET AL. ON BRINE-SATURATEDSEDIMENTS
199
The ponds were constructed and filled with slimes in 1998. The three ponds were
instrumented with 1" diameter piezometers with screened intervals at three different
depths within the slimes layer, as shown on Figure 2. Monitoring of the porewater
quality and hydraulic head in the system commenced in December 1998. Porewater
samples and head measurements were obtained approximately every two weeks until
May 2000.
Gravity Drainage System Pond (Pond 1)

The gravity drainage system was composed of a specially lined pond with an
underdrain bedded in an aggregate filter. The insolubles were placed over top of the
drainage system. After slimes placement, the excess brine was decanted and the
insolubles surcharged with a fresh water cover. The head of fresh water flushed the brine
within the insolubles downwards into the underdrain system. The underdrain system
conducted the effluent to a collection basin (Culvert 1 on Figure 2).
Figure 4 shows the construction details of the gravity drainage pond. This system
consisted of approximately 0.60 m of screened drainage rock (25.4 mm) covered with
0.15 m (6") filter sand. A 0.15 m (6") diameter slotted HDPE pipe was installed in the
gravel to conduct the effluent to the collection basin. Approximately 0.57 m of slimes
were added to Pond 1.
Figure 4 - Cross section of Pond I (Gravity Flow)
Reverse-Flow Ponds (Ponds 2 and 3)

The reverse-flushing treatment systems were also composed of specially lined
treatment ponds with slimes placed over top of underdrain structures; however, an
adjacent fresh water supply pond was connected to the underdrain system at a higher
hydraulic head than the treatment pond. This caused the fresh water to flow upwards
through the slimes layer, with the brine effluent collecting in the treatment pond above
the slimes layer. For effective treatment, the upward gradient across the slimes must be
sufficient to promote adequate flushing, but not so great as to cause piping and short
circuiting through the slimes layer.
Pond 2 is a reverse-flow system that was constructed with a geosynthetic filter
separating the slimes from the inflow piping. Figure 5 shows the design details of Pond
2. The fresh water inlet pipe was fed from a large supply pond adjacent to the treatment
pond. Fresh water flowed through the piping system underneath the slimes and was
flushed upwards through an approximately 0.79 m slimes layer. A 0.4-m-thick screened
200
drainage rock layer was placed around the slotted HDPE inlet pipe and separated from
the overlying slimes by a non-woven geotextile (Amoco 4553).
Figure 5 - Cross section of Pond 2 (Reverse-Flow with Geotextile Filter)

Pond 3 is a reverse-flow system that was constructed with an aggregate filter
separating the slimes from the inflow piping. Figure 6 shows the design details of Pond
3. The underdrain system was comprised of a 0.13-m-thick layer of filter sand placed
over 0.63 m of drainage rock that surrounded the inlet pipe. This pond contained
approximately 0.61 m of slimes.
Figure 6 - Cross section of Pond 3 (Reverse-Flow with Sand Filter)
Sampling and Analyses

Water samples were taken every two weeks from the sampling points shown on
Figure 2. Samples were analyzed for a variety of parameters including electrical
conductivity, density, total dissolved solids, and chlorides. The test period for slimes
porewater analysis commenced in December 1998 and was completed in June 2000. The
various methods of testing the slimes porewater in the laboratory are shown in Table 3.
201
Table 3 - Laboratory Methods for Slimes Porewater Analyses
Parameter
Method
Density
Chloride
TDS
Pipetting/Gravimetric
Titration with Silver Nitrate
Major Ions (Carb-bicarb-SO4-Na-Ca) analyzed
by ion chromatography or titrated with EDTA
EC meter
max = 200 000 uS/cm
Electrical
Conductivity
Porewater samples were tested by trained technicians at an analytical laboratory.

Selected samples were also submitted to a separate accredited analytical laboratory for
comparison. Samples of treated slimes were collected from each pond in June 2000 at
the end of the testing period. Subsamples of the slimes were tested for several parameters
including electrical conductivity, density, chlorides, pH, sodium, calcium, magnesium,
and potassium content.
Results of Testing
The water chemistry results indicated that all three systems removed brine from
the insolubles. The gravity drainage pond (Pond 1) was the most effective of the three
brine flushing systems, but this was the result of difficulties operating the reverse-flow
ponds during freezing conditions.
Pond 1: Gravity Drainage Pond

Table 4 contains a summary of the averaged data collected from the three
piezometers in the slimes layer of Pond 1. The gravity drainage system removed more
than 92% of the brine from the porewater in the slimes. A plot showing the change in
chloride concentration with time for the slimes porewater, the influent (Pond 1), and the
effluent (Culvert 1) is shown in Figure 7. The surface of the gravity flow pond was
frozen between January and March 1999, but the results indicate that significant brine
removal occurred in that time interval.
Table 4 - Summary of Porewater Chemistry (Pond 1)
Parameter
Initial Concentration
Final Concentration
Removal
EC (uS/cm)
Chloride (mg/L)
TDS (mg/L)
Density (mg/cm3)
200 000
170 000
320 000
1.20
15 000
3400
6600
1.01
92.5%
98.0%
97.9%
95.0%
202
200 000
* POND#1
9 CULVERT#1
9 AI ---x-. B1
X ClJ
Downward Flow I
150 000
Treatment Time
276 days
loo ooo
e.
50 000
Dec-98 Feb-99 Apr-99 Jun-99 Aug-99 Oct-99 Dec-99 Feb-00 Apt-00 Jun-00
Figure 7 - Pond 1: Chloride Versus Time (98% Reduction, Downward Flow)
In June 2000, the slimes layer was sampled and analyzed for several parameters.
Figure 2 shows the location of the slimes samples obtained from Pond 1. Table 5
contains the average chemical analysis results with depth for each of the subsamples
tested.
Table 5 - Summary of Slimes Chemical Analysis (Pond 1)
Sample
EC
(uS/cm)
Dry Density
(g/cm 3)
Chloride (mg/L)
Sum of Cations (Mg,

Na, K, Ca) (mg/L)
P
P
P
P
P
6 180
5 950
14 710
6 570
6 290
0.998
1.032
1.052
n/a
0.992
9 020
9 250
33 200
n/a
n/a
8 630
8 342
27 150
n/a
n/a
1-I
1-2
1-3
1-5
1-6
The latei'al consistency in the EC results indicate that brine was being removed
evenly across the pond, with the exception of the extreme edge of the pond at P 1-3. The
EC, TDS, and chloride results for the slimes samples are similar to the porewater
chemistry results. This indicates that the porewater chemistry results reflect bulk changes
in the slimes chemistry and not local effects near the piezometer tips.
203
The chloride concentration data presented on Figure 7 indicates that the system
reached an adequate level of treatment after approximately 276 days of operation. A
field evaluation of the system and laboratory k testing of the sampled slimes indicates
that the hydraulic conductivity of the slimes layer was approximately 2 x 10-6 cm/s in
Pond 1 at the end of the testing period. The volume of fresh water that flushed through
the system over the effective treatment time was estimated to be 91 m 3, or approximately
5 pore volumes.
Pond 2: Reverse-Flushing Pond with Geotextile Filter

Table 6 summarizes the porewater chemistry data collected from the piezometers
in Pond 2. There was an approximately 60% reduction in brine concentration in the
treated insolubles, most of which occurred at the end of the treatment period when ice
was off the pond surface (Figure 8).
Parameter
Final Concentration
Removal
EC (uS/cm)
Chloride (mg/L)
TDS (mg/L)
Density (mg/cm 3)
200 000
170 000
316 000
1.20
165 000
67 000
128 000
1.08
17.5%
60.6%
59.5%
60.0%
One of the reasons for the relatively poor results were the operational problems
encountered during cold weather operation. The surface of the effluent pond was frozen
during the winter months, which impeded the measurement of head levels in the pond.
This made it difficult to maintain the optimum gradient across the slimes layer. In one
instance, the critical gradient was exceeded and piping was observed in the slimes layer
next to one of the piezometers. The piping appeared to "self heal" when the hydraulic
gradient was reduced. The installation of an air bubbler system or a wind powered
turbine to keep ice off the water surface would significantly reduce these operational
difficulties.
Sample
EC
(uS/cm)
Dry Density
(g/cm 3)
Chloride (mg/L)
Sum of Cations (Mg,

Na, K, Ca) (mg/L)
P 2-2
46 400
1.045
87 800
80 275
Figure 2 shows the sample locations for Pond 2, while Table 7 shows the average
results of the slimes chemical analysis measured at the end of treatment.
204
200 000
~_* POND #2
CULVERT#2
A2""x'" B2 X C2 t
"--x..~...., "'x'"x"'x
"-.
150 000
.~ 100 000
.to
50 000
Dec-98 Feb-99 Apr-99 Jun-99 Aug-99 Oct-99 Dec-99 Feb-00 Apr-00 Jun-00
Figure 8 - Pond 2: Chloride Versus Time (60% Reduction, Upward Flow)
Pond 3: Reverse-Flushing Pond with Aggregate Filter

Table 8 summarizes the porewater analytical data collected from the piezometers
in Pond 3. The brine removal efficiency was near 75% (Figure 9), approximately 15%
lower than Pond 1. Pond 3 experienced the same operational difficulties as Pond 2;
however, the area of piping failure appeared to be somewhat larger. It is unclear whether
the more significant piping problem was due to the difference in filter construction or the
result of a locally thinner slimes layer over the drain. Again, a method of maintaining an
ice free area on the pond surface would have improved the pond operations and likely the
effectiveness of brine removal. Figure 2 shows the sampling locations in Pond 3 and
Table 9 shows the average results of the slimes chemical analysis. The measurements of
electrical conductivity, TDS and chloride were similar to Pond 2 results.
The presence of a denser fluid overlying a less dense fluid, such as occurred in
Pond 2 and 3, could also have lead to instability in the plug flow front resulting in
"fingering" across the fresh water/brine interface. This may have acted to enhance the
onset of piping as described above. This possibility was anticipated in the original
design; however, it was felt that the relatively low hydraulic conductivity of the slimes
would restrict the development of an unstable interface between the fresh water and the
brine.

Parameter
Final Concentration
Removal
EC (uS/cm)
Chloride (mg/L)
TDS (rag/L)
Density (mg/cm 3)
200 000
170 000
304 000
1.20
102 000
36 000
72 000
1.05
49.0%
78.8%
76.3%
75.0%
200 000
e POND#3
9 CULVERT#3
150 000
c3 r
'* A3 --'o-- B3
et0
,ooooo
t~
_o
'.l oO" 9
50 000
,o
'~1 --Dec-98 Feb-99 Apr-99 Jun-99 Aug-99 Oct-99 Dec-99 Feb-00 Apr-00 Jun-00
,
Figure 9 - Pond 3: Chloride Versus Time (75% Reduction, Upward Flow)

Sample
EC
(uS/cm)
Dry Density
(g/cm 3)
Chloride (mg/L)
Sum of Cations (Mg,

Na, K, Ca) (mg/L)
P 3-2
36 700
n/a
83 000
75 525
205
206
Conclusions
In this paper, the feasibility of the-brining by flushing with fresh water has been
established based on a field pilot-scale treatment system operated at the PCS Allan
Potash mine in Saskatchewan. Of the three treatment systems investigated, the gravity
drainage pond provided the simplest design and operation, and achieved greater than 95%
removal of chlorides and total dissolved solids. The brine reduction in the gravity pond
was achieved by flushing with approximately five pore volumes of fresh water over 276
days. Two reverse-flow systems were evaluated that have several attractive features;
however, these systems encountered operational difficulties that adversely affected the
treatment system's effectiveness. The surface of all the ponds froze during the winter,
but only the reverse-flow ponds became difficult to operate because accurate control of
the gradient across the thin slimes layer was not possible. The reverse-flow ponds
achieved 60 to 75% removal of chlorides and TDS, but because an undetermined quantity
of freshwater flushed through each reverse-flow pond, the results are not directly
comparable to the gravity flow pond results. Detailed evaluation of the reverse-flow
system would require the maintenance of an ice free zone in the ponds to permit more
accurate operational control.
The demonstrated ability to remove brine from potash insolubles by flushing with
fresh water presents a potential design alternative for the permanent disposal of potash
insolubles.
Acknowledgments
This research project was funded by the Saskatchewan Potash Producers
Association. Jeff Stoicescu, P.Eng., was instrumental in the successful treatment system
construction and the development of the field monitoring program. Field sampling and
monitoring were diligently carried out by April Slotsve and several assistants. This paper
was written by Brenda Bews, P.Eng., based on reports produced by MDH Engineered
Solutions.
References
Hart, R.T., 1988, "Preliminary Observations and Investigations into the Suitability of
Insols as a Capping Material for Potash Tailings Piles," Internal Report # ID3749K, Hart Environmental Consulting, Research at PCS Cory, SK for the
Saskatchewan Potash Producers Association (SPPA), 38 pp.
Haug, M.D., 1988, "Potential Cover Materials," Internal Report on the University of
Saskatchewan Research Program for SPPA, Saskatoon, SK.
Terzaghi, K., 1929, "Effect of Minor Geological Details on the Safety of Dams,"
American Institute of Mining and Metallurgical Engineers, Technical Publication
215, pp. 31-44.
207
U.S. Army Corps of Engineers, June 1955, "Drainage and Erosion Control-Subsurface
Drainage Facilities for Airfields," Part 13, Chapter 2, Engineering Manual,
Military Construction, Washington, D.C., 15 pp.
U.S. Army Corps of Engineers, October 1975, "Guide Specification for Plastic Filter
Cloth," CW 02215.
U.S. Bureau of Reclamation, 1987, "Design of Small Dams," United States Government
Printing Office, Denver, Colorado.
Catherine N. Mulligan =and Behnaz Dahr Azma =

Use of Selective Sequential Extraction for the Remediation of Contaminated
Sediments
Reference: Mulligan, C.N. and Dahr Azma, B."Use of Selective Sequential

Extraction for the Remediation of C o n t a m i n a t e d Sediments," Contaminated
Sediments: Characterization, Evaluation, Mitigation/Restoration, and Management

Strategy Performance, ASTMSTP 1442, J. Locat, R. Galvez-Cloutier, R. C. Chaney, and
Abstract: The interaction of contaminants with sediments is a very complex

phenomenon and means are required to understand this matter more fully. To determine
the speciation of metals in sediments (the distribution of elements among chemical
forms or species), specific extracts are used in a process called sequential selective
extraction. This method can be used to determine if heavy metals are removable by
remediation techniques or to predict removal efficiencies. Sequential extraction can be
employed for the evaluation of the most appropriate sediment remediation technology
and for monitoring remediation procedures. A biosurfactant, rharnnolipid, was used to
remove organic-bound copper and carbonate-bound zinc. Exchangeable, carbonate,
reducible oxide and organic fractions are amenable to washing techniques, and
residually bound contaminants are not economical or feasible to remove. This
information is important in designing the most appropriate conditions for sediment
washing.
Keywords: Sequential extraction, sediment washing, remediation, heavy metals,
biosurfactants
Introduction
Cadmium, copper, lead, mercury, nickel and zinc are considered the most
hazardous heavy metals and are included on the EPA's list of priority pollutants
(Cameron 1992). Recently, the EPA has announced that the decontamination of
sediments will receive the highest priority. Sources of metals include domestic and
industrial effluents, the atmosphere, runoff and the lithosphere. Once heavy metals are
IAssociate Professor and Ph.D. student, respectively, Department Building, Civil and
Environmental Engineering, Concordia University, 1455 de Maisonneuve Blvd W.,
Montreal, Quebec, Canada H3G 1M8.
2O8
www.astm.org
MULLIGAN AND AZMA ON SELECTIVESEQUENTIALEXTRACTION
209
allowed to pass through the municipal waste treatment facility, they return to the
environment where they are persistent, cannot be biodegraded and can thus follow a
number of different pathways. The metals can adsorb onto the soil, run off into rivers or
lakes or leach in the groundwater, an important source of drinking water. Exposure to the
heavy metals through ingestion or uptake of drinking water (particularly where water is
reused) and foods can lead to accumulation in animals, plants and humans.
Sediments dewatering is frequently necessary after dredging to remediate and treat
contaminants. Methods include draining of the water in lagoons with or without
coagulants and flocculants, or using presses or centrifuges. Treatment methods are
similar to those used for soil and include pretreatment, physical separation, thermal
processes, biological decontamination, stabilization/solidification and washing (Mulligan
et al. 2001 a). However, compared to soil treatment, few remediation techniques have
been commercially used for sediments. Solidification/stabilization techniques are
successful but significant monitoring is required since the solidification process can be
reversible. In addition, the presence of organics can reduce treatment efficiency.
Vitrification is applicable for sediments but expensive. Only ira useful glass product can
be sold will make this process be economically viable. Thermal processes are only
applicable for removal of volatile metals such as mercury and costs are high. Biological
processes are under development and have the potential to be low cost. Since few low
cost metal treatment processes for sediments are available, there exists significant
demand for further development~ Pretreatment may be one of the methods that can
reduce costs by reducing the volumes of sediments that need to be treated. Methods to
determine the requirements for remediation are necessary for optimal design.
Methods of Determining Speciation

Solid- State Methods
The term speciation is related to the distribution of an element among chemical
forms or species. Heavy metals can occur in several forms in water and soils. Interest has
increased in these techniques to relate the degree of mobility with risk assessment (i.e.,
the more mobile the metal is, the more risk associated with it (Bourg 1995)). Not only is
total metal concentration of interest, but it is now accepted that understanding the
environmental behavior by determining its speciation is of paramount importance.
Metal speciation of solid-state sediments may be obtained by spectroscopic
methods such as X-ray absorption fine structure in sediments (XAFS) (Helz et at. 1996).
Detection limits as low as ~tg/g can be achieved at ambient atmospheric temperatures for
wet sediments. Since synchrotron radiation is required for XAFS, only a few researchers
have access to this type of equipment. X-ray absorption near-edge structure (XANES) or
extended X-ray absorption fine structure (EXAFS) can be used to determine the valence
states of elements.
Peterson et al. (1997) were able to determine Cr(III) and Cr(VI) in contaminated
soils at levels of 20 to 50 ~tg/g by XANES. Szuiczewski et at. (1997) compared the
XANES analysis and extraction of different species of chromium They found that only
10% of the total Cr was in the form of Cr(VI) and that only small amounts of this form
could be extracted as exchangeable and with a phosphate buffer. Chemical methods
210
supported that only a small amount was extractable but were not able to detect pools of
Cr(VI) identified by XANES. Efforts are continuing to correlate chemical extractability
and solid-state characterization.
Selective Sequential Extraction

To determine the speciation of metals in soils, various methods are used. One
method is to use specific extractants. By sequentially extracting with solutions of
increasing strengths, a more precise evaluation of the different fractions can be obtained
(Tessier et al. 1982). A soil or sediment sample is shaken over time with a weak
extractant, centrifuged and the supernatant is removed by decantation. The pellet is
washed in water and the supernatant removed and combined with the previous
supernatant. A sequence of reagents is used following the same procedure until, finally,
mineral acid is used to extract the residual fraction. Heavy metal concentrations are then
determined in the various extracts by atomic absorption, inductively coupled plasma
(ICP) or other means. Numerous techniques and reagents have been developed and have
been applied to soils (Shuman 1985), sediments (Tessier et al. 1982), sludge-treated soils
(Petrozelli et al. 1983) and sludges.
These methods are not standardized and even the results can vary with the same
reagents, pH, temperature, extractant strength and solid to volume ofextractant ratio.
None of the extractions is completely specific, however the extractants are chosen in an
attempt to minimize solubilization of other fractions and provide a distribution of the
partitioning of the heavy metals. The chemical agents are used to destroy the bonds of the
heavy metals to the various sediment components. The extracting agents increase in
strength throughout the sequence.
To extract the exchangeable fraction, ammonium acetate, barium chloride or
magnesium chloride at pH 7.0 is generally used (Lake 1987). They cause the
displacement of the ions in the soil or sediment matrix bound by electrostatic attraction.
Pickering (1986) showed that magnesium chloride leached low quantities of other
sulfides, organic matter, aluminum and silicon. Other agents such as calcium chloride,
potassium nitrate and sodium nitrate can also be used (Yong 2001 ).
The reducible phase (iron and manganese oxides) is extracted with hydroxylamine
hydrochloride with acetic acid at pH 2.0. The hydroxylamine hydrochloride reduces the
ferrous and manganese hydroxides to soluble forms. Other components such as organic
matter and clay components are not solubilized to any great extent (Tessier et al. 1979).
The carbonate phase (calcite and dolomite) is extracted at pH. 5.0 with sodium acetate
acidified with acetic acid. This solubilizes the carbonates, releasing carbonate-entrapped
metals. Organic matter, oxides or clay components are not solubilized.
Hot hydrogen peroxide in nitric acid is used to oxidize the organic matter. The
oxidized organic matter then releases metals that are complexed, adsorbed and chelated.
These agents are used so that the silicates are not affected by this treatment (Yong 2001).
In the final step, the silicates and other materials are dissolved by strong acids at high
temperatures. This residual fraction is usually used to complete the mass balances for the
metals.
MULLIGAN AND AZMA ON SELECTIVESEQUENTIALEXTRACTION
211
Uses of Speciation
Monitoring of Soil Remediation
Kabata-Pendias (1992) demonstrated that the speciation of trace metals in natural
soils depends on the physical and chemical characteristics of the soil. Soil pH, redox,
organic, carbonate, clay and oxide contents all influence metal speciation and mobility.
Simple and complex cations are the most mobile, exchangeable cations in organic and
inorganic complexes are of medium mobility and, chelated cations are slightly mobile.
Metals in organic or mineral particles are only mobile after decomposition or weathering
and precipitated metals are mobile under dissolution conditions (eg., change in pH).
Kabata-Pendias (1992) also showed the specialion of trace metals such as zinc, copper,
cadmium and lead. Zinc and cadmium are mostly organically bound, exchangeable and
water soluble. Copper is mainly organically bound and exchangeable. Whereas, lead is
slightly mobile and bound to the residual fraction. Chlopecka (1993) showed, however,
that the cadmium and zinc speciation of the soils depended significantly on the
application of sewage sludge on the soil. The speciation of the metals changed from the
oxidizable to the acetic acid extractable once the sludge was added to the soil due to
microbial oxidation of the organic matter in the sludge (Hooda and Alloway 1994).
Fertilizer addition, water and air pollution can also affect speciation.
The effect of soil pH, constituents and heavy metal types were evaluated. In a study
by Ramos et at. (1994), sequential extraction techniques were used to evaluate the
mobility of cadmium, zinc, lead and copper in contaminated soil in a national park.
Cadmium was found to be the most mobile and would likely be the most bioavailable
A potential method to determine if the heavy metals can be removed by remediation
techniques or predict removal efficiencies is to determine speciation with selective
extractive techniques. It is believed that exchangeable, carbonate and reducible oxide
fraction may be amenable to soil washing techniques (Li et at. 1995). Removal of
organically and residually bound fractions may not be economical to recover or necessary
due to lack of bioavailability. Ravishankar et al. (1994) evaluated several sludges for AI,
Cu, Fe, Mn and Zn speciation to predict bioleaching processes. They concluded that more
stabilized sludges contained higher contents of organically bound metals and that sludges
vary considerably making bioleaching prediction difficult. Gombert et at. (1994) used
sequential extraction to determine if cesium, cobalt and chromium could be removed by
soil washing. Since less than 20% was extracted after dissolving 20% of the soil mass,
soil washing was abandoned as an option.
Barona and Romero (1996) used sequential extraction to show that 0.1 M EDTA
could extract lead from various fractions of a contaminated soil by multiple regression
analysis. Van Benschoten et at. (1997) also assessed lead removal during batch washing.
Electrokinetic remediation of copper was monitored by Ribiero and Mexia (1997). The
immobilization by zeolite, hydroxyapetite and an iron-oxide waste of cadmium and lead
in an arable soil at an industrial site has been evaluated (Chlopecka and Adriano 1997).
Recently, sequential extraction techniques have been studied as a tool in various
applications. Yong et al. (1993) examined sequential extraction to obtain a better
appreciation of the ability of clay soil barriers to contain contaminants in landfill barriers.
Dahlin et at. (2001) examined the speciation of heavy metals in soils to determine if they
212
were mobile, extractable and plant-available. Samples with lead, arsenic and chromium
were collected from two Superfund sites. They found that using Tessier's sequential
extraction procedure with data from optical and scanning-electron microscopy, X-ray
diffraction and chemical analyses was the best approach during early stages of site
analyses. Man-made materials and large particles can make the use of sequential
extraction unsuitable.
Monitoring of Sediments by Sequential Extraction

To develop remediation strategies, the ease of release of the contaminants must be
investigated. In sediments, speciation has been determined by selective sequential
extraction by Galvez-Cloutier (1994). The effect of pH on lead was examined. As the pH
increased to 4 from 1.3, precipitation of hydroxyl and carbonate species occurs.
Stabilization takes place at pH 4.5 since the speciation is stable. Most of the lead
remained in the residual fraction which represents the lattice of the silicate minerals.
Yong (1995) compared an illite soil with sediments. Sequential selective extraction
showed that lead can be released easily from the exchangeable ion fraction. This fraction
was negligible for the sediment sample at pH values greater than 4. The use of more
aggressive extractants in the selective extraction process for the removal of oxide and
carbonate fractions suggests that these conditions may not be normally found in the
natural environment of the sediments for release of lead. However, for remediation of
dredged sediments this information is highly useful. The fraction associated with the
residual fraction would require total destruction of the sediment and thus is not likely to
occur naturally and would not be practical for remediation. Further work by GalvezCloutier and Dube (1998) indicated that sequential extraction results of zinc and
cadmium, in particular, varied according to the pH and that particle size was not a major
consideration for sediments.
Characterization of sediments was performed by Savvides et al. (1995). Heavy
metal contamination included Cu, Cr, Ni, Pb, Zn and Fe. Using sequential extraction,
they found that although the significant amounts of the metals were found in the residual
fraction, acid extraction was able to remove the heavy metals from the organic and
carbonate.
Dredged sediments can be appfied to soils to enhance fertility. If these sediments
are contaminated, there is the possibility that metals may leach from the sediments once
they are removed from the reducing environment and placed in an oxidizing one.
Sequential extraction can be utilized to evaluate the likelihood of metal mobilization
before application and over time to identify changes in metal mobility (Singh et al. 1998).
Fitch and Burken (1998) studied metal removal by lab-scale wetlands. Influent
levels of lead and zinc were I00 ppb and 200 ppb, respectively and removal rates were
90% for lead and 70 to 80% for zinc. Sequential extraction of the sediments was
employed to determine the removal mechanism. It showed that coprecipitation with iron
oxides led to substantial removal of lead and zinc.
Recently, Chattier et al. (2001) showed that sequential extraction could be used to
determine which metal fractions were extracted from sediments before and after
biological treatment. They showed that between 90 and 100% ofPb, Zn, and Cd that was
biologically removed was from the fractions bound to carbonate and oxide fractions and
MULLIGAN AND AZMA ON SELECTIVE SEQUENTIAL EXTRACTION
213
70 to 80% of the Ni from the same fractions. Approximately, 63% of the copper, on the
other hand, removed the organic and sulfide fractions, which were the major copper
fractions, Almost none of the Cr was removed.
Selective Sequential Extraction for Remediation of Contaminated Sediments with

Biosurfactants
Clearly, more work is needed in the field of sequential extraction for the evaluation
of remediation technologies. The use of sequential extraction procedures to characterize
the sediments and the sequential extractions were performed after washing with
biosurfactants. These surfactants are less toxic and more biodegradable than most
synthetic ones available. They are commercially available and have been studied for
various environmental applications. The most notable is the application of rhamnolipids
for the release of three times as much oil as water alone from the beaches of the Alaska
after the Exxon Valdez tanker spill (Harvey et al. 1990). Due to their anionic and
hydrophilic/hydrophobic nature, biodegradable surfactants including rhamnolipids,
surfactin, and sophorolipids, by-products of bacteria or yeast, have been able to remove
metals from an oil-contaminated soil by disruption of the pollutant/soil bonds (Mulligan
et al. 1999). Sequential extraction has been used to develop strategies and monitor the
remediation of the soil through multiple washing steps (Mulligan et al. 200 lb). In this
paper, sequential extraction was used to monitor remediation of sediments by the
biosurfactant rharnnolipids. Different concentrations ofrhamnolipid with and without
sodium hydroxide were used to extract the heavy metals and sequential extraction was
used before and after surfactant washing to determine which metal fractions were
removed by the surfactants. Sodium hydroxide was added as it was previously shown to
enhance the biosurfactant washing step (Mulligan et al. 1999).
Experimental Methods
Characterization of the Sediment

The metal-contaminated sediment sample was obtained from a canal area which
was surrounded by metal and steel industries. The sample was air-dried. The grain size
distribution of the sediment indicated 10% sand, 70% silt and 20% clay. X-ray analysis
indicated the presence of quartz (30%), feldspar (36%), illite (2%), kaolinite (27%),
chlorite (3%) and carbonate (0.5%) as performed by Mulligan et al. (2001c). Total
organic matter was 20% (w/w). The sediments were digested by the method
recommended by Environment Canada (1990) and then analysed by a Perkin Elmer
Atomic Absorption Aanalyst 100 Spectrophotometer for heavy metal content. These
results are summarized in Table 1.
Sediment Washing Procedures

Batch washing experiments were used to evaluate the feasibility of using
biosurfactants for the removal of heavy metals from sediments. The rhamnolipid
biosurfactant (JBR215) was obtained as a 15% solution (w/v) from Jeneil Biosurfactant
214
Co. It is produced from sterilized and centrifuged fermentation broth. Two major types of
rhamnolipids, RLL (R1) and RRLL (R2), are present in the solution. The molecular
weight of RLL (C26H4809.rhamnolipid type I) is 504 and ofRRLL (C32H58013~
rhamnolipid type II) is 650 as determined by the supplier.
Table 1 - Characterization qf Sediment Sample for l otal Heavy Metal Content

Parameter
Concentration (mg/kg)
Chromium
Copper
Nickel
Lead
Zinc
145
140
76
572
4~854
The biosurfactant solution was diluted with distilled water or 1% NaOH as required.
A quantity of 1.5 g from the sediment of each particle size was placed in individual 50mL
vials. 15.0mL of surfactant solution was added to each vial. The samples were kept in the
temperature incubator at constant temperature (25~ for 1 day. Blank samples including
15.0mL distilled water and 1.5 g sediment were provided. Vials were placed on the side
in order to achieve the maximum contact surface between the sediment particles and the
biosurfactant. The pH of the rhanmolipid with NaOH was 10 and without NaOH was 6.0.
Blanks included the same additives as for the biosurfactant studies without the presence
of the biosurfactant. The series of washings was performed by washing the soil for 24 h,
and then removing the supernatants (3,000 x g, 30 min). All experiments were performed
in triplicate. Results were reproducible to + 10%. The solutions from each sample were
collected. The samples were digested and the concentrations of copper and other metals
in each sample were measured by the Atomic Absorption Spectrophotometer. In order to
release the heavy metals trapped in the biosurfactant micelles, the organic copper colloids
were oxidized by adding 30% H202 slowly until no reaction was observed. The method
chosen for biosurfactant digestion was 3030E, which is recommended by APHA (1995)
and approved by the EPA.
Sequemial Extraction Procedures

The procedure used for the sequential extraction was similar to that of Yong et at.
(1993) shown in Table 2. Sediment samples (1.5 g) were washed with the surfactant
solutions and controls and then subsequently dried prior to sequential extraction. Each of
the fractions was collected and the concentrations of heavy metals were determined by
atomic absorption spectrometry. The amounts of copper, lead, zinc, chromium and nickel
extracted from each of the extractants were then calculated.
215
Table 2 - SequemialExtraction Procedure (Yong et al. 1993)

Sequence
Chemical Reagents
Fraction
Extraction of metals by surfactants and controls

overnight with 15 mL of solution
Soluble
Extraction of metals with 8 mL of 1 M MgCI2 (pH

7) for 1 h
Exchangeable
Extraction of metals with 8 mL of 1 M NaOAc I

adjusted to pH 5 with acetic acid for 5 h
Carbonates
Extraction of metals with 20 mL of 0.04 M

NH2OH.HCI in 25% (v/v) acetic acid (pH 2.5) at
96~ for 6 h
Oxides and
hydroxides
Extraction with 3 mL of0.02M HNO3 and 5 ml of

30% H202 (pH 2) for 2 h at 85~ followed by 3
mL of 30% H202 (pH 2) at 85~ for 3 h and then 5
mL of 3.2 M NHaOAcI in 20% (v/v) HNO3
diluted to 20 mL at room temperature for 30 min.
Organic
matter
Digestion at 90~ with 25 mL of dilute aqua regia

(50 mL HCI, 200 mL HNO3 and 750 mL water)
for3 h
Residual
fraction
1 Ac - denotes acetate
Sltrfactant Washing
A series of washings was performed on the sediments using 0.5%, 1% and 2%
rhamnolipid solutions with and without 1% NaOH. The control was 1% NaOH. Copper,
zinc, lead, nickel and chromium were analyzed in the supernatants. All are plotted in
Figure 1 with the exception of chromium since no chromium removal was detected. This
was because chromium is usually found in the more mobile anionic form (CrO4~) and
therefore the anionic biosurfactant had no affinity for the chromium. Removal rates for
zinc and nickel increased as the rhanmolipid concentration increased in an almost linear
fashion without NaOH. However for lead and copper, highest removal rates occurred for
0.5% rhamnolipid with 1% NaOH. Increasing the biosurfactant concentration when
NaOH was added did not enhance removal rates. Maximal removals were 7.3% nickel,
4.9% zinc, 1.7% lead and 15.6% copper. However, since zinc was in the highest
concentration, this metal was removed the most. For example, for a 2% rhamnolipid
216
concentration, approximately 250 mg/kg of zinc were removed compared to 20 mg/kg of

copper (0.5% rhamnolipid with NaOH). If an average molecular weight of 575 is
assumed for the rhamnolipid, then it can be calculated that 10 moles of the biosurfactant
removed 1 mole of zinc. Significant amounts of zinc (129 mg/kg) were also removed
with 0.5% rhamnolipid and NaOH.
Figure 1 - Percentage of Copper, Nickel, Zinc, and Lead Removedfrom Sediments by

Various Washing Agents, 0.5%, 1% and 2% Rhamnolipid With and Without 1% NaOH.
1% NaOH Was Also Evaluated.
Sequential Extraction of the Sediment Before Washing

Sequential extraction procedure s were used on the sediment to determine the
speciation of the heavy metals before surfactant washing. The exchangeable fractions of
all metals were small. The carbonate and oxide fractions accounted for over 60*/o of the
zinc present in the sediments (Tables 3 and 4). The organic fraction constituted over 80%
of the copper, and 50% of the chromium and lead. A major proportion of nickel was
found in the residual fraction. In a study by Ramos et al. (t994), copper, lead, cadmium
and zinc were sequentially extracted from soils with organic contents of approximately
15%. Similar to this research, they found that copper was associated with the organic
matter or residual fractions, lead and zinc were primarily found in the oxide fraction.
Sequential Extraction of the Sediments after Biosurfactant Washing

Sequential extraction of the sediments was performed after washing with the
biosurfactant. The results of these experiments for copper, zinc and nickel are shown in
Figures 2, 3 and 4. Rhamnolipid could remove the organically-bound copper when NaOH
was added and the oxide fraction when no NaOH was used since the pH was 6 and more
appropriate for copper removal associated with the oxide fraction. For zinc, there was a
major decrease in the carbonate fraction when no NaOH was added with the
biosurfactant, since this is removed more easily than the oxide fraction. In the presence of
MULLIGAN AND AZMA ON SELECTIVE SEQUENTIALEXTRACTION
217
NaOH and the biosurfactant, there was removal from the organic fraction of zinc. For
nickel, the oxide and carbonate fractions decreased as the biosurfactant concentration
increased. There were also significant reductions in the organic fraction, upon addition of
NaOH with 0.5 and 1.0% surfactant. Chartier et at. (2001) also showed similar trends.
Table 3 - Sequential Extraction Characterization of Metai
Contaminants m Sediment Before Washing With the Biosurfactant
Fraction (% of total )
Metal
Exchangeable
Carbonate
Oxide
Organic
Residual
1
1
1
0
4
1
l
8
9
18
21
4
34
23
46
51
86
54
29
22
25
12
2
39
10
Chromium
Copper
Lead
Nickel
Zinc
Table 4 - Heavy Metal Affinity for Various Fractions m the Sediments
(pH of Each Matrix Was Not Adjusted)

Heavy metal
Chromium
Copper
Lead
Nickel
Zinc
Affinity for each fraction

Organic>residual> oxide> carbonate, exchangeable
Organic>residual> oxide > carbonate, exchangeable
Organic> oxide >carbonate> residual>exchangeable
Residual>organic> oxide>carbonate> exchangeable
Oxide> organic> carbonate>residual >exchangeable
Figure 2 - Sequential Extraction of Cu after Washing Sediments with Each Agent, 0.5%,
1% and 2% Rhamnolipid, With and Without 1% NaOH and 1% NaOH. The Fractions are
Indicated as Solubfor Soluble, Exch for Exchangeable, Carbfor Carbonate, Oxidfor
Oxide, Orgfor Organic and Res for Residual.
218
F i g u r e 3 - Sequential Extraction of Zinc After Washing Sediments with Each Agent,

0.5%, 1% and 2% Rhamnolipid (Sur39,Withand Without 1% NaOH and 1% NaOH. The
Fractions are Indicated as Solub for Soluble, Exchfor Exchangeable, Carbfor
Carbonate, OxidJbr Oxide, Orgfor Organic and Res for Residual.
4 - Sequential Extraction of Nickel after Washing ,Sediments With Each Agent,

Water, 0.5%, 1% and 2% Rhamnolipid (Surfl), With and Without 1% NaOH and 1%
NaOH. The Fractions are Indicated as Solubfor Soluble, Exch for Exchangeable, Carb
for Carbonate, Oxidfor Oxide, Orgfor Organic and Res for Residual.
Figure
219
Treatability Protocol
When it has been determined that a dredged sediment sample has been
contaminated with heavy metals, an assessment must be performed to determine how to
remove the metals, and how much will be removed. Therefore, a scheme of assessment
can be used as shown in Figure 5. A selective sequential extraction procedure can be
integrated into the scheme with the characterization of the sediments. The most efficient
pH values and types of reagents can be devised for extracting the heavy metals from the
various fractions based on the selective sequential extraction and bench scale tests can be
carried out. An overall remediation plan can then be designed followed by scale-up to
field tests.
Inorganic contaminants
Characterization of sediments
Nature and concentration
of contaminants
Sediment properties
SelecUve sequential
extraction tests
Column and bench tests
,,,,j
I Contaminated sediment 1
Assessment
Determination of feasibility
of and type of remediation
with respect to desired
final contaminant concentration
1
Figure 5 - Protocol for Desig?ring the Most Appropriate Reraediation Strategy for H eav.l"-
Metal Contaminated Sediraent

The purpose of the sequential extraction studies on sediments that have been
previously washed with the surfactants was to determine from what fraction these
surfactants are removing the metals. This information can then be used to determine if
sediment washing is useful and if surfactant use is potentially effective and what
surfactants might be useful and under what conditions. For example, if the metals are
retained mainly in the residual fraction, it may be very difficult, if not impossible to use
any of these surfactants to remove these metals.
In this research, only the residual fraction was untouched during soil washing
procedures. All fractions under different conditions could be decreased. For example,
under acidic conditions, the carbonate phase would be released. However, under basic
220
conditions (co-addition of the surfactants with NaOH) copper, zinc and nickel could be
released since they were found in the organic phase.
Conclusions
The results in this study show that the organic phase-associated metals can be
removed by rhamnolipid with sodium hydroxide. Conditions without NaOH addition
were effective for removing the zinc and nickel in the oxide and carbonate phases.
Residual fractions, the most difficult to remove, were not affected during the surfactant
washing studies. This information is important in designing the appropriate conditions for
soil washing. A combination of acidic and basic conditions~ therefore~ would increase the
amounts of metals removed in the case of carbonate, oxide and organic-associated metals.
Sequential extraction procedures can provide indications but not specific
information on chemical binding of metals. This information can then be used to
determine remediation procedures and to monitor the procedures during the treatment
phase. It was shown that copper could be removed from the organic-bound fractions and
zinc could be removed from the oxide-bound fractions during washing with rhamnolipid
biosurfactants. More exact information from surface analytical methods such as
electromicroscopy, X-ray absorption spectroscopy (XANES, EXAFS) and X-ray
diffraction can be obtained but these equipments are not widely available. The
combination of sequential extraction with analytical techniques such as chromatography
would lead to more selectivity and improvements in treatability testing and in the design
of remediation techniques.
References
American Public Health Association (APHA) American Water Work Association

(AWWA) Water Pollution Control Federation (WPCF), 1995, Standard Methods
for the Examination of Water and Wastewater, 17th Edition.
Barona, A. and Romero, F., 1996, "Fractionation of Lead in Soils and Its Influence on
the Extractive Cleaning with EDTA," Em,'ironmental Technology, Vol. 17, No. 1,
pp. 63-70.
Bourg, A.C.M., 1995, Speciation of Heavy Metals and Implications For Their Mobili(y.
HecayMetals, Springer, Berlin, pp. 19-32
Cameron, R.E, 1992, Guide to Site ~ld Soil Descriptionfor Hazardous Waste Site
Characterization. Volume 1 Metals. Environmental Protection Agency, 1992,
EPA
Chattier, M., Mercier, G., and Blais, J.F., 2001. "Partitioning of Trace Metals Before and
After Biological Removal of Metals from Sediments," WaterResearch, Vol. 35,
No. 6, pp. 1435-1444.
Chlopecka. A., 1993, "Forms of trace metals from inorganic sources in soils and amounts
found in spring barley," Water, Air and Soil Pollution, Vol. 60, No. 1-2, pp. 127134.
Chlopecka, A. and Adriano, D.C., 1997, "Influence of zeolite, apatite ad Fe-oxide on Cd
and Pb uptake by crops", Science of the Total Environment, Vol. 207, No. I-2,
pp. 195-206.
MULLIGAN AND AZMA ON SELECTIVE SEQUENTIALEXTRACTION
221
Dahlin, C.L., Williamson, C.A., Collins, W.K., and Dahlin, D.C., 2001, "Can Sequential
Extraction Determinations Effectively Define Heavy Metal Species in Superfund
Site Soils," Annual International Conference on Soils, Sediments cold Water,
University of Massachusetts.
Environnement Canada, 1990, Service de conservation et protection Region Quebec
Methode d'analyse des metaux darts sediments, ies boues et les sols.
Fitch, M.W., and Burken, J.G., 1998, Metals Removal by Constructed Wetlands: Final
Report to The Doe Run Company, The Doe Run Company, May 26, 1998, 12 pp.
Galvez-Cloutier, R., 1994, "A Study of Heavy Metals Accumulation Mechanism in
Lachine Canal sediments," Ph.D. thesis, Department of Civil Engineering, McGiU
University, Montreal, Canada.
Galvez-Cloutier, R., and Dube, J-S., 1998, "An Evaluation of Fresh Water Sediments
Contamination: The Lachine Canal sediments Case, Montreal, Canada. Part II:
Heavy Metal Particulate Speciation Study," Water, Air atut Soil Pollution, Vol.
102, No. 3, pp. 281-302.
Gombert II, D., 1994, "Treatability Testing to Evaluate What Can Work", Nuclear
Technology, Vol. 108, No. 1, pp. 90-99.
Harvey, S., Elashi, I., Valdes, J.J., Kamelu, D., and Chakbarty, A.M., 1990,
"Enhanced Removal of Exxon Valdez Spilled Oil From Alaskan Gravel by a
Microbial Suffactant." Bio..Technology, Vol. 8, No. 3, pp. 228 -232.
Helz, G.R., Miller, C.V., Charnock, J.M., Mosselmans, J.F.W., Pattrick, R.A.D., Garner,
C.D.,and Vaughn, D.J., 1996, "Mechanism of Molybdenum Removal from the
Sea and its Concentration in Black Shales: XAFS Evidence," Geochimica
Cosmochimica Acta, Vol. 60, pp. 3631- 3639.
Hooda, P.S., and AIloway, B.J., 1994, "Changes in operational fractions of trace metals
in two soils during two years of reaction time following sewage sludge
treatment," International dournal of Environmental AnalA,tical Chemistry, Vol.
57, pp. 289-311.
Kabata-Pendias, A., 1992, "Trace Metals in Soils in Poland - Occurrence and
Behaviour," Trace Substances #1 Environmental Health. Vol. 25, pp. 53-70.
Lake, D.L.. 1987, "Chemical Speciation &Heavy Metals in Sewage Sludge and Related
Matrices. Heavy Metals in Wastewater and Sludge Treatment Processes. Volume
I." Sources, Analysis and Legislation. (Lester, J.N., ed.), CRC Press Boca Raton,
pp. 125-153.
Li, W., Peters, R.W., Brewster, M.D., and Miller, G.A., 1995, "Sequential Extraction
Evaluation of Heavy-Metal Contaminated Soil: How Clean is Clean?,"
Proceedings of the Air and Waste Management Association, 88th Annual Meeting
and Exhibition, San Antonio, Texas, June 18-23.
Mulligan, C.N., Yong, R.N., and Gibbs, B.F., 1999, "On the Use of Biosurfactants
for the Removal of Heavy Metals from Oil-Contaminated Soil," Environmental
Progress, Voi. 18, No. 1, pp. 50-54.
Mulligan, C.N., Yong, R.N., and Gibbs, B.F., 2001a, "An Evaluation of Technologies
for the Heavy Metal Remediation of Dredged Sediments," Journal of Hazardous
Materials. Voi. 85, No. 1-2, pp. 145-163.
Mulligan, C.N., Yong, R.N., and Gibbs, B.F., 2001b, "The Use of Sequential Extraction
Procedures for Soil Remediation," Clay Science for Engineering (K.Adachi and
222
M. Fukue, Eds.,) Balkema, Rotterdam, pp. 377-384.

Mulligan, C.N., Yong, R.N., and Gibbs, B.F., 2001c, "Heavy Metal Removal from
Sediments by Biosurfactants," Journal of Hazardous Materials, Vol. 85, No. 12, pp. 111-125.
Pickering, W.F., 1986, "Metal Ion Speciation- Soils and Sediments (a Review)," Ore
Geology Review, Vol. 1, pp. 83-146.
Peterson, M.L., Brown, G.E., Jr., Parks, G.A., and Stein, C.L., 1997, "Differential Redox
and Sorption of Cr(III/VI) on Natural Silicate and Oxide Minerals: XAFS and
XANES Results," Geochimica Cosmochimica Acta, Vol. 61, pp. 3399-3413.
Petrozelli, G., Giudi, G., and Lubrano, L., 1983, Proc. Int. Conf. Heavy Metals in the
Environment, p. 475.
Ramos, L., Hernandez, L.M., and Gonzalez, M.J., 1994, "Sequential Eractionation of
Copper, Lead, Cadmium and Zinc in Soils From or Near Donana National Park,"
Journal of Environmental Quality, Vol. 23, No. 1, pp. 50-57.
Ravishankar, B.R., Auclair, J.-C., and Tyagi, R.D.,1994," Partitioning of Heavy Metals
in Some Quebec Municipal Sludges," Water Pollution Research Journal of
Canada, Vol. 29, No. 4, pp. 457-470.
Ribiero, A.B., and Mexia, J.T., 1997, "A Dynamic Model for the Electrokinetic Removal
of Copper for Polluted Soil," Journal of Hazardous Materials, Vol. 56, No. 3, pp.
257-271.
Sawides, C., Papadopoulos, A., Haralambous, K.J., and Loizidou, M., 1995, "Sea
Sediments: Contaminated with Heavy Metals: Metal Speciation and Removal,"
Water, Science and Technology, Vol. 32, No. 9-10, pp. 65-73.
Shuman, L.M., 1985, "Fractionation Method for Soil Microelements," Soil Science. Vol.
140, pp. 11-22.
Singh, S.P., Tack, F.M., and Verloo, M.G., 1998, "Heavy Metal Fractionation and
Extractability in Dredged Sediment Derived Soils," Water Air and Soil Pollution,
Vol. 102, No. 3, pp. 313-328.
Szulczewski, M.D., Helmke, P.A., and Bleam, W.F., 1997, "Comparison of XANES
Analyses and Extractions to Determine Chromium Speciation in Contaminated
Soils," Environmental Science and Technology, Vol. 31, No. 10, pp 2954-2959.
Tessier, A., Campbell, P. G. C., and Bisson, M., 1979, "Sequential Extraction Procedure
for the Speciation of Particulate Trace Metals," Analytical Chemistry, Vol. 51,
No.7, pp. 844-851.
Tessier, A., Campbell, P. G. C., and Bisson, M., 1982, "Particulate Trace Speciation in
Stream Sediments and Relationships With Grain Size: Implication for
Geochemical Exploration," Journal of Geochemical Exploration, Vol. 16, No 2,
pp. 77-104.
Van Benschoten, J.E., Matsumoto, M.R., and Young, W.H., 1997, "Evaluation and
Analysis of Soil Washing Seven Lead-contaminated Soils," Journal of
Environmental Engineering - ASCE, Vol. 123, No. 3, pp. 217-224.
Yong, R.N., Galvez-Cloutier, R., and Phadingchewit, Y., 1993, "Selective Extraction
Analysis of Heavy-Metal Retention in Soil," Canadian Geotechnical Journal,
Vol. 30, No. 5, pp. 834-847.
Yong, R.N., 1995, "The Fate of Toxic Pollutants in Contaminated Sediments, "Dredging,
223
Remediation attcl Containment of Uontaminated Sedimettts, ASIM S]P 1293,

K.R. Demars, G.N. Richardson, R.N. Yong, and R.C. Chancy, Eds. American
Society for Testing and Materials, Philadelphia, PA, pp. 13-39.
Yong, R.N., 2001, Geoenvironmental Engineering, Contaminated Soils, Pollutant Fate
and Mitigation, CRC Press, Boca Raton, FL.
Ansm Kabir ~and Miguel Picomell 2

Effects of Overburden Stresses on Soil Reclamation by Eleetroosmosis
Reference: Kabir, A. and Picornell, M., "Effects of Overburden Stresses on Soil

Reclamation by Electroosmosis," Contaminated Sediments: Characterization,
Evaluation, Mitigation/Restoration, and Management Strategy Performance, ASTM

STP 1442, J. Locat, R. Galvez-Cloutier, R. C. Chaney, and K. R. Demars, Eds., ASTM
Abstract: A fine-grained soil stock was cleaned of organic matter and soluble salts. Then
the slurry was repeatedly washed with a barium-chloride solution to ensure a known pore
solution chemistry. Test specimens were prepared by first consolidating the slurry in
triaxial cells under varying cell pressures, and secondly by trimming the specimens to be
placed into electroosmotic cells enclosed in rubber membranes with the same confining
pressure applied on the membrane as had been used in the consolidation phase. The
electroosmotic treatment was performed at constant voltage gradient and was continued
until the electrical conductivity of the effluent was similar to that ofdeionized water. The
"postmortem" investigations of the treated specimens show that the confining pressure
has a significant effect on the removal of barium by the electroosmotic treatment.
Keywords: electroosmosis, effects of overburden, soil reclamation, barium
Introduction
Soil reclamation by electroosmosis is a technique that has been used to remove heavy
metals from fine-grained soils. This process is an alternative to the conventional flushing
of soil, with significant economic and technical advantages for fine-grained soils.
Electroosmosis is the transport of water under the influence of an electric current
between a cathode and an anode. The method consists of applying an electric gradient
across a soil mass to generate water flow. The application of DC current through the
electrode results in the development of electrical, hydraulic and chemical gradients within
the soil mass. It is assumed that the concentration of ions is initially uniform throughout
the specimen and the application of electric potential creates a uniform electric gradient.
The process results in the movement of ions due to the electric field and the flow of water
towards the cathode well. During the process the pore fluid gets acidic (pH 1 to 4) at the
anode and basic at the cathode (pH 11 to 13.5).
The existing technical literature abounds in research performed with laboratory
specimens and bench-top experiments using low-charge soil media, such as sand and
slurries of kaolinite (Lorenz 1969; Yong and Mourato 1990; Acar et al. 1994; Haran et al.
i Research Fellow, LMI, 2000 Corporate Ridge, McLean, Virginia 22120.

2 Professor of Civil and Environmental Engineering, North Carolina A&T State
University, Greensboro, NC 27411.
224
www.astm.org
KABIR AND PICORNELLON ELECTROOSMOSIS
225
1995). The references published using low-charge materials are too numerous to be
referenced exhaustively. There is a consensus that although kaolinite is commonly used
in laboratory experiments, charged clays might retard ion transport (Chen et al. 1999).
The presence of electrically charged clay minerals, as indicated by the cation-exchangecapacity "CEC" and the clay mineral content, has been demonstrated in laboratory
experiments (Vane and Zang 1997; Puppala et al. 1997) and in field trials (Chen et al.
1990). The effect of the electric charge in the clay particles is to reduce the usefulness or
efficiency of the electroosmosis treatment to remove electrolytes from contaminated soils
(Puppala et al. 1997; Chen et al. 1999).
If the electric charge of the clay particles reduces the efficiency of electroosmosis, an
increase of the density of the soil would result in smaller pore sizes. Thus the cations
would have to move closer to the charged clay particles. Since the electrostatic attraction
is inversely proportional to the square of the distance between electric charges, it appears
reasonable to expect that at higher densities (for identical soil media) the efficiency of the
electroosmosis treatment would be even lower.
In this manner, the depth of burial of the contaminated soil (that is, the applied
confining pressure) should have an effect on the density of the soil stratum, and thus, on
the efficiency or feasibility of electroosmosis decontamination of the stratum. Kelsh and
Parsons (1997) have cataloged sites of the Department of Energy indicating burial depths
of the contaminated layer. In most cases, the contaminated stratum is confined to the
surficial soils to depths of 15 m (50 ft). In some isolated cases, the depth of burial of the
contaminated stratum exceeds 152 m (500 ft). The effect of this large range of burial
depths (and thus the associated confining pressures) can be expected to have a
considerable influence on the density of the soil media, and thus on the efficiency of the
electroosmotic treatment.
This paper describes the effects of confining pressure on the removal of Barium from a
clayey soil (with a significant CEC) by electroosmosis treatment in a laboratory setting
Barium chloride reagent grade and soil from the flood plain of the Rio Grande in E1
Paso, Texas were the two main materials used in the study. Barium chloride was used to
contaminate the fine grained soil due to following considerations: (1) Barium is a toxic
metal common in industrial and nuclear waste, (2) Compounds of barium are highly
toxic, and (3) The valence of barium is two, and thus the electrostatic interaction of the
ions with electrically negative clay particles would be more apparent.
The collected soil sample was first dried and pulverized. Then the sample was washed
through a number 200 sieve and the soluble components were removed by dissolution
with diluted hydrochloric acid solutions. Then hydrogen peroxide, 30% by volume, was
added to the soil suspension to oxidize the organic matter. These additions were repeated
at day intervals until the peroxide did not react with the soil suspension. At this point the
soil suspension was titrated to a pH 4 to complete the removal of carbonates present.
These steps were implemented to ascertain that no other cat ion species would be present
and thus interfere with the removal of barium by the electromigration treatment.
226
At this point, the soil suspension was repeatedly washed with solutions of barium
chloride 0.01 N. For this purpose, the soil suspension was thoroughly mixed with some
barium chloride solution and allowed to settle for periods ranging from 3 to 5 days. The
resulting clear supernatant was then decanted and fresh barium chloride solution was
added to replace the decanted supematant and the suspension mixed thoroughly again.
This process was repeated until the electrical conductivity of the clear supematant
decanted was nearly identical to the electrical conductivity of the fresh barium chloride
solution.
The soil stock so prepared was characterized using identification tests such as the
hydrometer analysis and the Atterberg Limits. The soil stock is made up of 78% silt size
particles and 22% of clay size particles. The clean soil suspension exhibited a liquid limit
of 48 and a plastic limit of 21. The contaminated soil stock exhibited a liquid limit of 54
and a plastic limit of 25. These results were the average of three determinations for each
type o f soil suspension.
For the preparation of specimens for consolidation, the stock slurry was concentrated
by centrifugation at 3500 rpm for 45 minutes. The clear supematant was discarded and
the soil cake was scooped out of the centrifuge bottle and was thoroughly mixed with a
spatula on a glass plate. This material was then placed in molds holding a rubber
membrane to form specimens of 71.1 mm (2.8 in.) in diameter and 152.4 mm (6 in.) in
length. The specimen enclosed in the membrane was placed in a triaxial cell, and a
predetermined confining pressure applied on the membrane, while the specimen was
allowed to undergo consolidation. The outflow of pore water was monitored and the
process was continued until the outflow had essentially stopped. At this point, the
specimen was removed from the triaxial cell and the specimen was trimmed down to 51
mm (2 in.) diameter and 127 mm (5 in.) in length. The trimmings were kept to analyze
for the initial concentration of barium, the initial soil pH, the initial water content of the
specimen, and the initial electrical conductivity of the soil.
The trimmed down specimens were placed in an electroosmotic cell. The specimens
were isolated with a rubber membrane with the only exception o f one specimen at zero
confining pressure that was compacted and tested inside an acrylic tube. A sketch of the
electroosmotic cell is shown in Figure 1. The electroosmotic cell was made of Plexiglas
to facilitate the visual observation of the specimen during treatment. The cell was divided
in three parts: the two end caps and the central cylindrical portion o f the cell wall. The
end caps had 50 ml capacity reservoirs associated with the electrode wells. At the cathode
well, the reservoir was connected to a tube with a valve with the purpose of allowing the
drainage o f the well to measure the volume of fluid collected and the barium removed on
a daily basis.
To avoid the trapping of gas bubbles forming at the electrodes by electrolysis of water,
the electrodes were separated from the specimen with spacers. The specimens were in
contact with filter paper and these were confined with porous stones. The spacers
separated the electrode from the porous stone, and coincided within the electrode wells.
The wells were equipped with access ports to monitor pH, electrical conductivity, and
temperature o f the fluid in the reservoir, and permitting the escape o f gas forming.
A constant hydraulic head was maintained in the anode well using a Mariotte bottle as
the source of de-ionized water to replenish the water moved by the electroosmotic flow
227
Confining Pressure
!
Anode
Well
,
.~
Cathode
Well
'
/r
-/Porous
,,Plexiglass Cell Wall
Stone
ii
~ Cathode
11
Anode
Specimen
[ 4
Spacer,'
U._]""Rubber Membrane
Power Supply
I'
Ammeter
I t!
!
Drain Valve
Figure 1 - Sketch of Electroosmotic Cell Used in this Study

towards the cathode. The water levels in the two wells were controlled to maintain a
nearly zero hydraulic gradient during the electroosmotic treatment.
The cell wall was connected to a pressure control board. The pressure inside the cell
was kept at the same pressure that had been used for the consolidation of each particular
specimen. In this manner, four different tests were performed: test 1 for the specimen
with no confining pressure applied; test 2 for a confining pressure of 172 kPa (25 psi);
test 3 for a confining pressure of 241 kPa (35 psi); and test 4 for a confining pressure of
482 kPa (70 psi.)
An electrical gradient of 39 volt/m (1 volt/in.) was imposed on all specimens using a
constant voltage power supply through graphite electrodes. The electrical gradient was
kept constant for all the tests performed. The pH, the electrical conductivity, the
temperature, the change in volume experienced by the specimen and the amount of
effluent reaching the cathode well were monitored on a daily basis. The specimens were
kept under electroosmotic treatment for periods of time ranging from one and a half
months up to three months.
Upon completion of the treatment, the specimens were sliced into ten discs 12 mm
(0.5 in.) thick. The slices were numbered from 1 to 10 starting at the anode side. The
concentration of barium, pH, electrical conductivity, and moisture content was
228
determined for each slice. The barium concentrations were determined on extracts from
the slices obtained by the digestion method 3050 of the USEPA. The concentrations in
the extracts were obtained analyzing the samples using a flame or a furnace Atomic
Absorption Spectrophotometer.
Four different tests were performed to investigate the potential effects that overburden
could have on the remediation of barium contaminated soils. The only difference between
the specimens was the confining pressure used to consolidate and confine the specimen
during electroosmotic treatment. The test 1 specimen was formed into an acrylic tube by
compacting the soil slurry; no confining pressure was applied to consolidate or confine
the specimen during treatment; this is referred as zero confining pressure. The initial
moisture content of the soil trimmings was measured at 29.9%. The specimen for test 2
was consolidated and confined during treatment under 172 kPa (25 psi) confining
pressure; before the initiation of the treatment phase, the soil had a moisture content of
24.4%. The specimen for test 3 was prepared and treated under 241 kPa (35 psi)
confining pressure; the moisture content of the specimen was 22.8% upon completion of
the consolidation phase. The specimen for test 4 was prepared and treated under a
confining pressure of 482 kPa (70 psi) and resulted in a soil with a moisture content of
19.5% after the consolidation phase.
The differences in moisture content at the beginning of the treatment period indicate
that the specimens had large differences in density, void ratio and pore sizes that could
only be attributed to the confining stresses acting on the specimens. The treatment phase
lasted 41 days for test 1 and the three remaining tests were all treated for 95 days. The
complete set of all the data recorded has been presented elsewhere (Kabir 1996) and the
more relevant findings are summarized in the remaining sections of this paper.
Acid~Basic Conditions in the Electrode Wells

Before the initiation of the treatment, the pH in both wells was nearly a pH of 7. As
the treatment progressed, the pH in the anode well progressively decreased to a pH of
1.93 for test 1 and to a pH of 2.0 for test 3 and test 4. The only exception was test 2 in
which the anode well stabilized at a pH of 3.This can probably be attributed to problems
encountered for the full insertion of the pH probe into the anode well during this test.
By way of contrast, the pH in the cathode well increased as the treatment progressed,
eventually stabilizing at a pH of 11.8 for test 1 and test 2, at pH of 12 for test 3, and at a
pH of 13 for test 4. These results show a consistent pattern of increasing pH as the
density of the soil specimen increased. This fact supports the contention that the higher
densities restricted the diffusion of the hydroxyl ions into the soil mass resulting in higher
concentration of hydroxyl ions in the cathode well.
Measurements of Electroosmotic Flow

The electroosmotic flow peaked early in the treatment phase and then decreased to
non-measurable values towards the end of the process. During the first five days of
229
treatment, the accumulated effluent collected in the cathode well decreased from 5.3 ml
for test 1, to 4.4 ml for test 2, to 3.8 ml for test 3, and to 3.7 for test 4. These results show
a consistent pattern of decreasing effluent for increasing density of the specimen with the
associated decrease in porosity. Thus the higher the confining pressure, the lower will be
the induced water flow by the same voltage gradient.
The lower electroosmotic flow rates translate to smaller coefficients of electroosmotic
permeability. The coefficients of electroosmotic permeability in centimeters square per
second and volt calculated from the above measurements decreased from 2.0* 10-6 for test
1, to 1.96"10 -6 for test 2, to 1.6"10 -6 for test 3, and to 1.4"10 -6 for test 4. These results are
in general agreement with earlier reported findings (Casagrande 1952.)
Energy Consumption During the Treatment Process

The accumulated consumption of energy (in kilowatts hour per cubic meter of soil
treated) at the end of the treatment process for each specimen did consistently increased
from 3.0 for test 1, to 8.5 for test 2, to 8.7 for test 3, and to 9.1 for test 4. A further
difference between the specimens was that the rate of energy consumption had leveled
off for test 1, while it was still increasing at nearly constant rates for the remaining three
tests.
These results by themselves indicate that the energy expenditures for cleaning the soil
should increase as the soil contamination occurs in deeper layers. Nevertheless, these
results have to be seen in light of the removal efficiency described later. These later
results indicate that the removal of barium is, not only more costly, but the removal of
barium is much less effective; in other words, the effect is more pronounced than could
be inferred from the energy expenditures listed above.
A comparison of the energy expenditures of the three tests is presented in Figure 2.
These results show that the energy expense per unit mass of barium removed increased
for increasing confining pressures; that is, for increasing depths of burial of the
contaminated soil. The energy consumption increased by 20% from test 2 to test 3 (for
confining pressure increasing from 172 kPa to 241 kPa) and by 60% from test 2 to test 4
(for confining pressure increasing from 172 kPa to 482 kPa.)
Specimen Volumetric Changes Induced by the Electroosmotic Treatment

The specimens were initially trimmed down to 51 mm (2 in.). Upon completion of the
treatment and extraction from the cell, the diameter of the specimen was measured at
intervals of 12 mm (0.5 in.) along the length of the specimen. These results showed that
the specimens consistently decreased in diameter near the ends of the specimen, and
increased diameter towards the center of the specimen. The combined effect was a
general decrease of the volume of the specimen on the order of 6 ml to 7 ml. Since the
specimen expanded upon removal of the confining pressure, the 6 to 7 ml volume
decrease is a lower bound of the actual volume changes experienced by the specimens.
The effluent volume collected for tests 2, 3, and 4 was on the order of 12 ml. This
effluent would be the result of fluid supplied by the anode well and fluid extracted from
the pores of the specimen. Thus, 12 ml is an upper bound of the volume lost by the
specimens during treatment. Both indicators show that the specimens lost some volume,
230
1.80E-02
"~
1.80E-02
=
.~
Test 4
1.40E-02
%,,,,
1.20E-02
1.~-o2
Test 3
~'~
8.~E-~
N
o
~
6.1~E-03
2.1~E-~
.1~E-0,3
0.00E+00
0
10
2o
30
40
50
60
7o
8o
90
100
Time of Treatment (days)

Figure 2 - Cumulative Energy Spent per Unit Mass of Barium Removed
from 6 ml to about 12 ml. This implies that the water removed at the cathode well was
not thoroughly replaced by the constant pressure water supply. Thus in real applications,
the soil will tend to shrink and to accommodate this shrinkage strain, the soil mass would
have a tendency to crack.
Moisture Content of the Soil after Treatment

Upon removal of the specimen from the electroosmotic cell, the specimen was cut into
ten slices and a portion of each slice was used to determine the water content of the soil
after treatment. The results of all of these determinations are summarized in Table I. This
table also shows the moisture contents determined on the trimming of each specimen,
before the treatment. The results shown in Table 1 are in conflict with the volumetric
changes discussed earlier. The moisture content of the specimens, appear to have
increased during treatment. This increase seems to be larger near the ends of the
specimen and smaller towards the center of the specimen.
Although there is no conclusive evidence for this explanation, it appears that the
differences in water content within the specimen are the result of the specimen
preparation procedure. In this manner, the consolidation phase was probably not
completed and gradients in water content existed within the specimen before the
electroosmotic treatment was started. A similar explanation could be advanced to
reconcile the moisture contents of the trimmings of the specimens before being subjected
to the treatment process and the moisture contents of the specimens after treatment.
231
Table 1 - Moisture Content (%) of the Soil Slices and Trimmings

Slice Number ~
1
2
3
4
5
6
7
8
9
10
Trimmings
Test 1
Test 2
Test 3
Test 4
43.2
42.3
34.0
33.9
37.2
36.4
35.6
34.8
50.0
32.8
32.6
28.2
27.9
27.1
28.9
28.9
31.1
23.6
30.3
32.8
21.5
20.4
20.8
18.5
20.8
22.7
20.3
20.5
24.7
26.6
19.2
19.6
19.7
20.7
21.2
20.0
19.9
19.8
19.6
15.4
29.9
24.4
22.8
19.5
~Slice 1 is at the anode well, and slice 10 is at the cathode well
Acid~Basic Conditions Along the Treated Specimens

A portion of each slice was mixed with de-ionized water in a 1:1 soil/water ratio by
weight, and the pH of the supernatant was recorded. A summary of these measurements
is presented in Table 2. The initial pH of the soil stock suspension was measured to be at
a pH of 6.5. The measurements shown in Table 2 indicate very minor changes of the pH
of the soil within the specimen. The trend appears to be that for the higher confining
pressures, the central portion of the specimen appears not to have been affected, and only
a limited zone around the electrode well appears to have experienced any significant pH
changes.
The effects at the cathode well appear to be even more limited and no significant
difference appears to have occurred beyond the first slice. These results are in stark
contrast to the findings reported in the literature (Acar et al. 1990) about the penetration
of the basic front being about one third of the specimen, and the penetration of the acid
front being the remaining two thirds of the specimen. Most of the results published in the
technical literature have been obtained using low charge clay, such as kaolinite, and thus
have neglected the effect of the clay charge on the electro migration process. The results
shown in Table 2, suggest a fairly strong pattern of increased isolation of the central part
of the specimen as the pores become smaller and the movement of charged cat ions
becomes more difficult.
Evaluation of Free Barium Removal by the Electroosmotic Treatment

As a result of the initial saturation of the soil stock with 0.1 N barium chloride
solutions, the exchange complex of the clay was saturated with barium and the pore fluid
consisted of a 0.1 N barium chloride solution. Thus the specimens contained free barium
232
Table 2 - p H Measurements 1of Spectmen

.
,s Shces
.
Slice Number 2
Test 1
Test 2
Test 3
Test 4
1
2
3
4
5
6
7
8
9
10
4.88
4.22
5.92
6.06
6.02
6.39
6.41
6.46
6.76
,7.49
4.45
5.85
6.34
6.82
6.83
6.84
6.87
6.94
6.93
6.94
4.46
5.49
5.88
6.42
6.58
6.64
6.71
6.72
6.80
7.30
3.54
4.23
6.09
6.68
6.66
6.69
6.74
6.77
6.88
9.31
1The pH measurements were performed on a I: 1 soil extract

2Slice 1 is at the anode well, and slice 10 is at the cathode well
in the pore water, and barium fixed on the exchange sites o f the clay particles. A portion
of each slice was used to determine the free barium in the pore water after the treatment.
For this purpose, the soil was dispersed into de-ionized water and the volume brought to
100 ml. The concentration of barium in the water in equilibrium with the soil was
considered as free barium. This amount o f barium relative to the initial free barium
(corresponding to the 0.1 N barium chloride solution) is an estimate o f the barium left
over by the electroosmotic treatment. The results obtained for all specimens are
summarized in Table 3.
These results show a clearly defined trend of increasing percentage o f free barium not
removed by the electroosmotic treatment as the confining pressure on the specimen
increases. The free barium left over in the specimen o f test 1 is only about 1% o f the
initial barium present in the pore fluid. As the confining pressure increases for the tests 2,
3 and 4, the left over barium increases to 2%, 4%, and 7%, respectively. Furthermore, the
free barium left near the cathode well increases significantly above the overall rates
described above.
The removal of free barium was expected to be easier than the removal of the barium
absorbed on the clay exchange sites. These results, show that most of the free barium is
removed for all specimens; although, the confining pressure has an effect on the time
frame of the cleanup operation and the energy expenditures necessary to accomplish the
removal of free barium. The more plausible explanation is that the electric charge of the
clay particles and the smaller pores existing in the specimen as the confining pressure
increases forces a closer interaction between the cat ions and the clay charges. This
appears to reduce the mobility of the barium ions and requires higher-energy
expenditures to force the flow of ions towards the cathode.
Efficiency of Electroosmotic Treatmentfor Total Barium Removal
A portion of each slice was digested following the sludge digestion procedure 3050 of
the USEPA, and the total concentration o f barium was measured using an atomic
KABIR AND PICORNELL ON ELECTROOSMOSIS
233
Table 3 - PercentageI of Free Barium Remaining After Treatment

Slice Number2
Test 1
Test 2
Test 3
Test 4
0.7
0.8
1.4
1.4
1.4
1.3
1.5
1.0
1.3
5.3
1.3
1.6
2.2
2.3
2.1
2.1
2.0
2.6
2.0
5.6
2.0
4.0
2.8
4.1
4.2
4.0
5.8
6.2
7.2
9.9
0.7
5.0
4.4
4.0
5.1
5.9
7.1
8.4
11.6
52.2
2
3
4
5
6
7
8
9
10
IThe percentage is referred to the concentration of barium in the 0.01 N solution

2Slice 1 is at the anode well, and slice 10 is at the cathode well
absorption spectrophotometer. The amount of barium measured under these conditions
includes the free barium in the pore water and the barium absorbed on the clay exchange
sites. The efficiency was calculated as the percentage of barium removed with the
treatment (this was calculated as the difference between initial barium minus the barium
determined on the digested slice) referred to the total initial barium present in the
specimen before the electroosmotic treatment. A summary of all the results for the four
tests is presented in Table 4. These results show clearly that the confining pressure
influences drastically the efficiency of barium removal. This implies, that under field
conditions, the cleanup of deeper soil layers will require longer treatment periods and
larger energy expenditures.
All the tests show a consistent trend of higher efficiency at the anode that then
decreases gradually towards the cathode. This is the result of the transport of barium
towards the cathode; thus, the first region to show signs of barium removal would always
be near the anode well. The results for test 4 show an efficiency of about 72% at the
anode, while the efficiency near the cathode is very nearly zero. The results for this test 4
indicate that the cleanup was in the beginning stages when the treatment process was
stopped after 95 days.
At this point, it is worthwhile to recall that the energy expenditures for the
electroosmotic treatment of test 1 were about one third of the energy expenditures in test
4. Furthermore, the clean up process in test 1 had lasted less than 10 days, while the
cleanup process for tests 2, 3 and 4 was continued for 95 days and remained noticeably
active for periods in excess of 50 to 60 days. Considering that the results shown in
Table 4 indicate removal efficiencies for test 1 in excess of four times those achieved in
tests 2, 3, and 4, the large effect of confining, or overburden, pressure is patently clear.
The implication of these findings is that the depth of overburden above the contaminated
soil will drastically affect the cleanup of clayey soils with non-negligible cat ion
exchange capacities. These soils are by far the largest majority of soils found in nature.
234
Table 4 - Efficiency of the Treatment as Percentage of Total Barium Removed

Slice Number ~
Test 1
1
2
3
4
5
6
7
8
9
10
96
93
93
90
90
91
91
91
91
89
Test 2
77
45
44
22
22
20
19
18
17
0
Test 3
Test 4
73
44
43
21
22
20
21
10
3
1
72
37
28
20
20
20
14
2
1
0
Slice 1 is at the anode well, and slice 10 is at the cathode well

Conclusions and Recommendations
The results of the presently described laboratory program show conclusively that the
overburden stresses can have a significant effect on the rate and the cost of the removal of
contaminating cat ions from naturally occurring soil deposits by the electro migration
imposed by an applied direct electric current. This is especially true for soil deposits with
non-negligible exchange capacities.
Although the set of results available at this time is limited and does not permit a
quantification of the effects in terms of cost increases and time delays that would be
incurred, it appears that these two factors can be significantly affected by the cat ion
exchange capacity of the soil in question, the ionic valence of the cat ion involved, and
the magnitude of the overburden stresses. An important implication is that the
extrapolation of bench trials in the laboratory cannot be used to predict the actual field
performance unless the effect of the confining stresses on the cleanup process is
adequately accounted for.
An additional concern associated with the extrapolation of results to other cases is
related to the different behavior of each ionic species with the differences in the pH
environment of the pore water. This is the result of the electrolysis of water, associated
with the treatment process, that causes changes in the pH and these in turn cause
differences in the solubility of the different ionic species.
The presently described study was limited since a single cat ion was used, for one soil
type, and a constant voltage gradient was applied in all tests. The results described seem
to indicate the need to extend the work to include more variables and permit clarification
of the roll of all possible interferences caused by the presence of overburden stresses on
the soil deposit undergoing restoration.
235
References
Acar, Y. B., Gale, R. J., Putnam, G. A., Hamed, J., and Wong, R. L., 1990,
"Electrochemical processing of soils: Theory of pH gradient developed by
diffusion, migration, and linear convection," Environmental Science and
Health, Vol. 25 (6), pp. 687-714.
Acar, Y. B., Hamed, J., Alshawabkeh, A. N., and Gale, R. J., 1994, "Removal of Cd (II)
from saturated kaolinite by application of electric current," Geotechnique, Vol. 44,
pp. 239-254.
Casagrande, L., 1952, "Electroosmotic stabilization of soils," Journal of the Boston
Society of Ovil Engineers, Vol. 39, pp 51-83.
Chen, J. L., AI-Abed, S. R., Bryndzia, L. T., and Murdoch, L., "Cation transport and
partitioning during a field test of electroosmosis," Water Resources Research, Vol.
35, No. 12, pp. 3841-3851.
Kabir, A., 1996, "Electroosmotic removal of barium from contaminated soils," Thesis
presented to the Graduate School of The University of Texas at E! Paso, in partial
fulfillment of the requirements for the degree of Master of Science.
Haran, B. S., Zheng, G., Popov, B. N., and White, R. E., 1995, "Electrochemical
decontamination of soils: Development of a new electrochemical method for
decontamination of hexavalent chromium from sand," Electrochemical Society
Proceedings, Vol. 95, No. 12, pp. 227-251.
Kelsh, D. J., and Parsons, M. W., 1997, "Department of Energy sites suitable for Electrokinetic remediation," Journal of Hazardous Materials, Vol. 55, pp. 109- I 16.
Lorenz, P. B., 1969, "Surface conductance and electrokinetic properties of kaolinite
beds," Clays and Clay Minerals, Vol. 17, pp. 223-231.
Puppala, S. K., Alshawabkeh, A. N., Acar, Y. B., Gale, R. J., and Bricka, M., 1997,
"Enhanced electrokinetic remediation of high sorption capacity soil," Journal of
Hazardous Materials, Vol. 55, pp. 203-220.
Vane, L. M., and Zhang, G. M., 1997, "Effect of aqueous phase properties on clay
particle zeta potential and electro-osmotic permeability: Implications for electrokinetic soil remediation processes," Journal of Hazardous Materials, Vol. 55,
pp. 1-22.
Yong, R. N., and Mourato, D. 1990, "Influence ofpolysaccharides on kaolinite structure
and properties in a kaolinite-water system," Canadian Geotechnical Journal,
Vol. 27, pp. 774-788.
Thomas C. Sheahan, I Akram Alshawabkeh, l Loretta A. Fernandez, 2 and Karen S. Henry3
A Reactive Geocomposite to Remediate Contaminated, Subaqueous Sediments
Reference: Sheahan, T. C., Alshawabkeh, A., Fernandez, L. A., and Henry, K. S., "A
Reactive Geocomposite to Remediate Contaminated, Subaqueous Sediments,"
Contaminated Sediments: Characterization, Evaluation, Mitigation/Restoration, and

Management Strategy Performance, ASTMSTP 1442, J. Locat, R. Galvez-Cloutier, R. C.
Chaney and K. R. Demars, Eds., ASTM International, West Conshohocken, PA, 2003.
Abstract: This paper describes a new geocomposite overlay, consisting of filter fabric
layers surrounding a chemically reactive core, that is being developed to contain and
remediate contaminated, subaqueous sediments. This geosynthetic-based reactive
material will provide the same isolation as that of traditional soil caps, but will have the
added advantages of: providing a remediation layer to adsorb and neutralize target
contaminants; block resuspension of fine sediment particles; and serve as a foundation
material during remediation. In addition, the geocomposite could be used in a tube
configuration to simultaneously store and remediate dredged sediments. The paper
describes the results of a two-phase, bench-top experimental program to assess the
feasibility of this idea and to help model its scalability for sediment remediation in the
field. This testing consisted of two parts: one-dimensional consolidation testing on
layered systems of very soft soils and geotextiles, both woven and non-woven; and
sorption isotherm testing of lead ions (pb+2), in solutions of various pH and salt
concentration, for both soil and zeolite. Results show the effects of the two geotextile
types on the soil-geotextile system during consolidation, including changes in
coefficients of consolidation and hydraulic conductivity. The adsorption tests show the
effectiveness of the zeolite in binding the lead ions, and the soil's potential for lead
adsorption under a variety o f p H and salt concentrations.
Keywords: contaminants, soils consolidation, adsorption, geosynthetics, sediment,
dredge spoil
~Associate Professors, Department of Civil and Environmental Engineering, Northeastern

University, Boston, MA 02115.
2Former undergraduate student, Department of Civil and Environmental Engineering,
Northeastern University, Boston, MA 02115.
3 Research Civil Engineer, U.S. Army Corps of Engineers, Cold Regions Research and
Engineering Laboratory, Hanover, NH 03755.
236
www.astm.org
SHEAHAN ET AL. ON SUBAQUEOUS SEDIMENTS
237
Introduction
Contaminated, subaqueous sediments present one of the most serious environmental

challenges on both the national and global levels. In the U.S. alone, the Environmental
Protection Agency (EPA) estimates that 1.2 billion cubic meters in the upper 50 mm of
subaqueous sediments is contaminated enough to pose health risks to the aquatic food
chain (U.S. EPA 1997). The existing state of practice to treat such sediments is either
subaqueous capping with some combination of natural soil and/or geosynthetic layers, or
removal of the sediment for ex situ treatment and/or storage. However, capping systems
are relatively thick layers of material prone to erosion and other deterioration, and these
layers isolate the sediment without actively treating it; sediment removal is expensive and
can result in significant contaminant resuspension in the overlying water during removal.
In this study, experimental work was performed to investigate a new approach to
remediate these sediments using a reactive geocomposite material. Such a material (U.S.
Patent Office 2002) will consist of one or more filtering layers surrounding a reactive
layer containing chemicals that can react with potential contaminants in the sediment.
Figure 1 shows two possible configurations for the reactive geocomposite: a horizontal
overlay in large sheets (Fig. la) and in smaller strips like wick drains (i.e., drainage strips
pressed vertically into the soil commonly used to accelerate consolidation) used in a
vertical configuration (Fig. 1b). For the horizontal overlay configuration (Fig. I a), large
sheets would be placed on the surface of a saturated or near-saturated sediment deposit
(either subaqueous or terrestrial). A layer of surcharge material (e.g., rip-rap, gravel or
sand) would be placed on top of it to consolidate the sediment, producing expulsion of
pore fluid that contains target contaminants. As the pore fluid flows through the
composite, the middle, reactive layer of the geocomposite will react with the
contaminants (e.g., adsorb them), resulting in isolating and concentrating the chemicals
within that layer. The middle layer may also retain fine particles on which chemicals are
adsorbed. For the wick drain configuration (Fig. lb), a tube of filter material would
surround the reactive or adsorbing layer in a sock-like configuration. An additional
configuration that could be used with the reactive geocomposite would be as a tube such
as that currently being used for a variety of waterway stabilization projects (e.g., Land
and Water 2001).
Pore fluid flow
/
Geosyntheticfilter
iii:i:i!i:i:i:i:i:i!iii!iii:ii!
Reactive or adsorbing .~...'.iiiiiiii~ii.i i i i i i i i i i i i i i i i i i i . i . i . ~

layer "~1.~.. ..'_ :'_ :'.':':'.'::':'.':':..:':':'[~.. .I"
Ge~
filter " ~ ~ i
iii i ! ! ii ii iii i iii i ! iii iii iii !
. ". ". ". ". ." ContaminatedSediment . . . . .
Figure 1a - Schematic of Layers in Reactive Geocomposite (Horizontal or Overlay
Configuration)
238
~ ~ ~, ~, ~ ~
Surchargelayer
Contaminated
sediment
Fluid flow
Fluid flow
Reactive
geocomposite
Figure lb - Reactive Composite in Vertical or Wick Drain Configuration
In the overlay configuration, the reactive geocomposite will provide the same
isolation as that of traditional soil caps. The additional functions of the overlay include
providing a remediation layer that will adsorb and effectively neutralize target
contaminants from the underlying sediment; blocking resuspension of fine sediment
particles that may be closely bound to contaminants; and serving as a stable, protective
foundation material during remediation.
This paper describes the benchtop laboratory testing that has been performed to date
to determine the feasibility of this technique and to help model its scalability for sediment
remediation in the field. This testing is considered essential prior to necessary pilot
testing. One-dimensional consolidation tests were conducted on layered systems of very
soft soils and geotextiles, both woven and non-woven. In addition, sorption isotherm
tests were performed on lead ions (Pb in solution for both soil and a proposed
adsorption agent (zeolite) in the reactive portion of the geocomposite. The adsorption
testing also included examining the effects ofpH and salt concentration on the sorption
isotherms.
Background
Use of Geosynthetics in Subaqueous Sediments
The U.S. EPA program on Assessment and Remediation of Contaminated Sediments

(ARCS; U.S. EPA 1996) identifies potential functions of geosynthetics in in situ cap
designs as 1) providing a bioturbation barrier, 2) stabilizing the cap, 3) reducing
contaminant flux, 4) preventing mixing of cap materials with underlying sediments, 5)
promoting uniform consolidation, and 6) reducing erosion of the capping materials.
ARCS guidelines include a geotextile for separation and stability purposes, i.e., to
prevent fines from migrating into the sand layer of the cap and presumably as a
bioturbation barrier (both separating functions), and to provide a more stable foundation
for the overlying soil layers (reinforcement).
SHEAHAN ET AL. ON SUBAQUEOUSSEDIMENTS
239
Geomembranes have been used to cap subaqueous sediments but the problem of gas
formation and water entrapment under the geomembrane has led to problems with its
bulging and/or lifting away from the sediment (e.g., Savage 1986 and Cooke et al. 1993).
Whenever there is flow of water through fine-grained sediments in series with a
geotextile, clogging of the geotextile should be considered. Filters in geotechnical
structures are required to retain soil particles while allowing water to pass so that
excessive water pressure does not build up in the retained (protected) soil. Granular or
geosynthetic filters are usually installed in structures where the flow is in one direction
only (e.g., pavement edge drains), and the design criteria are well established (Giroud
1996; Holtz et al. 1997).
Henry et al. (1995) described a geocomposite developed and tested for placement on
pond bottoms to prevent access of wildlife to contaminated pond sediment. The
contaminant was particles of white phosphorous, a neurotoxin that caused thousands of
waterfowl deaths annually in Eagle River Flats, Alaska. A laboratory study evaluated
whether white phosphorus particles become suspended in the overlying water because of
1) upward water flow through contaminated sediment capped with geotextile, or 2)
tapping of the top of the geotextile or sediment and then vigorously stirring the water to
simulate waterfowl feeding and swimming. The geotextile formed an excellent cap to the
underlying sediment and white phosphorus for both experimental conditions.
Henry and Stark (1997) evaluated the ability of geocomposite that had very large
pores interspersed in a geotextile that was overlain by a large-pored drainage net to retain
0.25- and 0.15-mm diameter particles in the presence of turbulent flow and impact
loading. The geocomposite was placed in a pond (< 0.5 m deep) and overlain with two
different soil covers and without a soil cover. The geosynthetics significantly reduced the
amount of fine soil suspended by the stirring and loading compared to the bare pond
bottom for all three cases, and they did not float to the surface as did the geotextiles in a
previous study.
Koemer and Ko (1982) studied flow through soil-geotextile systems to assess the
fabric's long-term flow performance. In flow tests on a variety of these layered systems,
they found that the flow rate tended to decline over time and stabilize at a lower flow rate
than initially measured. Aydilek et al. (2000) performed consolidation tests on high
water content wastewater sludge layered with a woven geotextile, and modeled the
results using large strain numerical analysis.
The use of geocomposites for various sediment treatments has developed to the
commercial stage. For example, Bishop et al. (1994) describe sand-filled geocomposites
for surcharging and stabilizing subaqueous sediments. Zeolite-impregnated geotextiles
have been developed for treating liquid and gaseous wastes (GSA Resources, Inc. 2001).
Design of Reactive Filtersfor Contaminant Remediation

The type of reactive geosynthetic filter will vary depending upon the type of
contaminant. The material introduced should provide the necessary reactive conditions
for either collection and/or breakdown of the target contaminants. For heavy metal
contaminated sediments, the best option is to extract and adsorb the contaminants in a
separate phase of the geotextile. Zeolite has a significant negative charge and ion
exchange capacity so that it is ideally suited for this reactive purpose.
240
Zeolites are hydrous aluminosilicates that exist in many natural formations, thus its
cost is not prohibitive. Studies have evaluated the potential use of zeolites for soil and
water treatment (e.g., Leppert 1990, Misaelides and Godelitsas, 1995, Shanableh and
Kharabsheh 1996, Reyes et al. 1997, Sakadevan and Bavor 1998). Others have attempted
to modify the sorption potential of zeolites for organic contaminants by treating the
zeolites with surfactants (Haggerty and Bowman 1994, Li 1998 and Li et al. 2000).
Jacobs and Forstner (1999) discussed the potential of using zeolite as a media for reactive
barriers. They evaluated lead sorption on four types of zeolite and demonstrated its
capability for immobilizing contaminants.
Sorption rate in any environment is dependent upon the changes in the concentration
of the ion in solution. In general, sorption rate is given by
& &&
. . . .
Ot & ~ t
(1)
where s is the adsorbed concentration (mg/kg), and c is the concentration in solution

(mg/l). The reversible term ds/dt is often used to describe the sorption rate. The
equilibrium partitioning between the adsorbed phase and the aqueous phase of the
chemical components are commonly measured under controlled temperature and applied
pressure, and the resulting correlation of s versus c is often described by several
equilibrium models (linear, Freundlich, and Langmuir models). Assuming instantaneous
equilibrium in sorption reactions and linear isotherms, the change in concentration of the
adsorbed phase of species is linearly related to the change in concentration of the aqueous
phase,
0s
~-7 = K,
(2)
where I~ is the distribution coefficient (L/kg). Other methods have also been used to
account for sorption because the previous method ignores the effect ofpH, ionic strength,
redox reactions, competitive adsorption, and the mechanism of adsorption. These
methods include isotherm equations, mass action models, and surface complexation
models proposed by Langmuir (1987), Kirkner and Reeves (1988), Yeh and Tripathi
(1989), Mangold and Tsang (1991), and Selim (1992), among others.
To develop a reactive geocomposite overlay for contaminated sediments, it is
necessary to assess the effect of salinity and pH on zeolite performance for collecting
lead. In this study the sorption isotherms are evaluated for lead on kaolinite and zeolite
using different solutions (deionized water, synthetic sea water and synthetic brackish
water) under various pH conditions (3, 5 and 7). The distribution coefficient is calculated
for each case to compare sorption characteristics under various conditions. The results
will help to understand the potential of lead collection by a zeolite reactive geosynthetic
barrier in seawater and brackish water conditions.
241
Experimental Studies
One-dimensional Consolidation Tests on Soil-Geotextile Systems
To characterize the consolidation behavior of a layered system consisting of a
geotextile and a very soft sediment (such as that expected near the surface of a
subaqueous sediment), a series of one-dimensional (l-D) consolidation experiments was
performed. The primary goal of this set of experiments is to determine how the
sediment's consolidation behavior will be affected by the addition of a geotextile filtering
layer, including the geotextile's clogging potential during the consolidation process.
A kaolinite powder, produced by U.S. Silica Co., was used as the baseline soil. The
powder has a mean particle size of 1 ~tm, with 95% finer than 10 lam and 65% finer than
2 lam. This powder was mixed at a water content, w = 50%, just above its liquid limit (LL
--- 41%), overlain by a piece of geosynthetic, was placed in an incremental loading (IL)
consolidation device. Testing was performed in accordance with ASTM Test Method for
One-Dimensional Consolidation Properties o f Soils (D2435). The consolidation test
specimens were allowed to drain only at the top surface, i.e., at the geosynthetic; the
bottom surface was kept impermeable to imitate the relatively low hydraulic conductivity
o f underlying sediments. Two types of geosynthetic, one woven and one non-woven,
were selected for use based on the results from the computer program Geofilter
developed by TC Mirafi, which takes as its primary input the soil type being used under
the filter. Table 1 gives the two geotextiles selected and their respective opening sizes
and flow characteristics. It is noted that while the apparent opening size is approximately
the same for both geotextile types, there is one order of magnitude difference in the three
flow characteristics given.
Table 1 -Properties of Geotextiles Selected for 1-D Consolidation Testing
Geotextile
type
Non-woven
Apparent
opening size,
mm
0.18
Permittivity,
sec I
1.2
Permeability,
cm/sec
0.2 l
Flow rate,
1/min/m z
3866
Woven
0.212
0.28
0.01
733
After mixing the soil at approximately 50% water content (which produced a thick
slurry consistency in the soil), it was placed in the consolidation cell confining ring that
was already mounted on the impermeable ring base. For the tests in which a geotextile
was used, the selected geotextile was cut into a circular piece with approximately the
same diameter as the inside diameter of the confining ring holding the soil specimen.
This piece was soaked in distilled water, excess moisture removed, and it was placed on
the soil surface, after which the top porous stone and the loading platen were placed on
the layered system. For tests in which only the kaolinite slurry was tested, a piece of
filter paper was moistened and placed directly on the soil, and the stone and platen then
placed on the top.
242
All specimens were incrementally loaded starting at a vertical effective stress, O'v =
2.45 kPa, up to cy'v= 98.1 kPa using a load increment ratio of 1.0. Each load was left on
for 24 hours before the next load increment was applied. For each load level, specimen
vertical deformation was measured using a displacement transducer and a computerbased data acquisition system. Measurements were taken at 5, 10, 20, 40, 80, 160 and
300 seconds (5 minutes) and then every 5 minutes until the end o f the 24-hour period.
Each increment was analyzed to determine the deformation at the end of primary
consolidation, as well as the coefficient of consolidation (c~) and the coefficient of
hydraulic conductivity (kv) in the vertical direction (logt method). Comparisons were
made between the consolidation behavior o f the kaolinite without any geotextile layered
on top (i.e., only filter paper between the soil and the top porous stone), and layered
systems consisting o f kaolinite with either woven or non-woven geotextile. The number
of tests for each specimen type are as follows: 8 tests on kaolinite alone (test designation,
K); 4 tests on kaolinite with non-woven geotextile (K-NW); and 3 tests on kaolinite with
woven geotextile (K-W).
Figure 2 shows the comparison ofcv among the three specimen types, K, K-NW, and
K-W, at different cr'v levels during the consolidation tests. The points plotted represent
average values for a particular specimen type at that ~'~ level. There are no data for Cr'v
levels less than 25 kPa because the logarithm of time curves did not exhibit a sufficient
"break" to apply the graphical construction technique for determining c~.
These results show that the K and K-W specimens exhibit virtually identical Cv
values at the 3 stress levels, and that the K-NW specimens show consistently higher Cv
values than the other two specimen types (typically a difference o f about 1 x 10
cm2/sec). It is important to note that the Cv values for all the tests increase monotonically
and linearly with respect to O'v, even at about 100 kPa.
The k~ values are computed using the equation
k~ = C~mvyw
(3)
where mv is the coefficient o f volume change, ASv/Acr'v, and Ywis the unit weight of
water. Figure 3 shows the comparison ofkv values (average value at each stress level for
a given specimen type, K, K-NW and K-W) for the three specimen types versus o'v. For
all specimen types, the trend is the same: while the Cv values increase across this stress
range (Fig. 2), the stiffening o f the soil during consolidation causes a decrease in my. The
result is an initial drop in kv, after which the values level off. As with the cv data, the kv
for the kaolinite with non-woven geotextile shows the highest hydraulic conductivity at
all stress levels.
The reason for the different cv and kv data between K-NW and K-W is explained by
the geotextile properties given in Table 1. The GeoFilter program is heavily biased in its
geotextile selection for a given application on the grain size distribution input. Thus,
referring to Table 1, while the two geotextiles selected have similar apparent opening
sizes (0.18 mm for the non-woven and 0.212 for woven), the flow characteristics
(permittivity, permeability and flow rate) are all different by about an order of magnitude
for the two geotextiles. Since the permeability o f the woven geotextile is closer to that o f
the kaolinite, it has less impact on the layered system's consolidation characteristics. On
the other hand, the non-woven represents a layer o f relatively high hydraulic
243
conductivity; it is also somewhat thicker than the woven material. This combination
produces a layer of much higher flow transmission that aids consolidation of the layered
system.
,.--.
12
0')
C'q
F: 10
o
,,-
6
7
0ffl
c
0
o
0
20
40
60
80
100
120
Vertical effective stress, dv (kPa)

Figure 2 - Comparison of ev Values During 1-D Consolidation Tests on Kaolinite
and Kaolinite-Geotextile Layered Systems
Lead Sorption Tests on Zeolite and Soil
The lead sorption testing was conducted using a Varian SpectrAA-220 Atomic
Absorption (AA) spectrometer, a Fisher Accumet pH meter, and a Fisher Accumet 30
conductivity meter. Materials used include kaolinite, zeolite, synthetic sea water (product
CRC007 from Aqua Solutions Inc.), 50 ml centrifuge tubes, lead nitrate, and nitric
acid.The kaolinite was the same as that used for the consolidation tests, and was selected
for its "average" properties in terms of its cation exchange capacity and ability to adsorb
ions. In other words, it lies between silt and montmorillonite in terms of its potential
adsorption behavior. Kaolinite has a specific surface area of 10 to 20 mZ/g, and a cation
exchange capacity of 3 to 15 meq/100 g (Mitchell 1993). The zeolite used (product
ZS500H from GSA Resources Inc.) had a specific surface area of 531 m2/g and a cation
exchange capacity of 250 meq/100g. Zeolite was selected as the adsorption medium due
to its high ion exchange and adsorption capacity, and for its ability to immobilize heavy
244
metal ions. Brackish water was prepared by mixing sea water and deionized water at 1: 1
ratio.
The sorption capacities were determined by evaluating the partitioning of lead
between lead nitrate solution and the soil. A stock of 0.001 M lead nitrate solution was
prepared by mixing 331.2 mg of lead nitrate powder with deionized water to make 1 liter
of solution. In addition to the 0.001 M stock solution, five diluted solutions (5X10"4M,
3x10aM, lxl04M, 5x10"SM, and 2.5x10 5 M) were prepared by mixing specific volume
of deionized water with 500 ml of the stock solution.
.-.
12
E 10
GO
o
v
x
>
o
~
L_
"o
20
40
60
80
100
120
Vertical effective stress, ~'v (kPa)

Figure 3 - Comparison of kv Values During 1-D Consolidation Tests on Kaolinite and
Kaolinite-Geotextile Layered Systems

The adsorption tests were conducted by mixing 3 grams of the material (kaolinite or
zeolite) with 30 ml of lead nitrate solution. A few drops of concentrated nitric acid
solution were added to adjust the pH to three specific values (3, 5 and 7). The mixed
samples in the centrifuge tubes were stored for at least 24 hours with frequent shaking by
hand (in accordance with EPA procedures). The solid and solution were then separated
by placing the specimens in a centrifuge. AA analysis was conducted on the solutions to
evaluate lead concentration using Varian SpectrAA-220 Atomic Absorption
spectrometer. The pH and electric conductivity of each specimen were measured.
Since lead nitrate is a toxic substance that can be absorbed through the skin, care was
taken for safe handling of the chemical and to minimize the potential for cross-
245
contamination. All glassware and plastic centrifuge tubes used for the preparation and
storage of standard solutions are soaked in an acid bath of 10% nitric acid for
approximately 24 hours. Furthermore, the glassware and tubes are rinsed in deionized
water three times and allowed to dry before use.
The results of sorption isotherms are displayed in Figure 4 for zeolite and kaolinite.
A summary of the distribution coefficient for these cases is also provided in Table 2.
Table 2 - Distribution Coefficients for Lead on Kaolinite and Zeolite (L/kg)
Solution
Material
Kaolinite
Zeolite
pH=7
pH=5
600
3333
pH=3
333
5333
Synthetic Brackish Water
Kaolinite
Zeolite
60
520
60
866
24
639
Synthetic Sea Water
Kaolinite
Zeolite
11
244
27
333
19
300
Deionized Water
The significance of zeolite sorption capacity is clear in the case of deionized water.
The solution concentration was less than 1 rag/1 even though the total adsorbed lead
exceeded 2000 mg/kg. The results showed that Ka=5333 L/kg at pH=3 and Ka=3333 L/kg
at pH=5. Such values are significantly high; however, it is noted that, at a pH of 3, lead
sorption was higher than that at a pH of 5. Although kaolinite sorption is less than zeolite,
it also showed significant sorption capacity. The distribution coefficient was 600 L/kg at
pH=5 and 333 L/kg at pH=3. As expected, the sorption potential was higher at a pH=5
than at a pH=3. The results show that lead sorption is one order of magnitude higher on
zeolite than kaolinite when deionized water is used. The distribution coefficient was
highest at pH=5 (higher than at pH=3 or 7).
The impact of brackish water was to reduce the sorption potential for both zeolite
and kaolinite. For kaolinite, the distribution coefficient decreased to less than 10% of that
in deionized water. This shows the effect of ionic content of the solution on limiting the
sorption of lead via ion exchange at the clay surface. Although reduced, lead sorption is
still high for the case of zeolite in brackish water. Lead concentration in the solution was
less than 2 mg/1, while the sorbed concentration was up to 800 mg/kg. Distribution
factors greater than 500 L/kg were measured, and the highest is 866 L/kg at pH=5.
Synthetic sea water provides extreme conditions that can restrict ion exchange and
lead sorption on zeolite and other ion exchange materials. This is obvious in the results
for kaolinite and zeolite displayed in Figure 4 and Table 2. Kaolinite sorption capacity
was limited and Ka decreased to 11 L/kg at pH=7. Zeolite capacity was reduced by one
order of magnitude compared to the deionized water case. The highest distribution
coefficient is 333 L/kg measured at pH=5.
A summary of the results clearly demonstrates the significance of using zeolite as a
sorption media to collect lead from contaminated sediments. Brackish and sea water will
contain significant concentration of ions that will compete with lead for sorption sites on
246
250O
2500
9 pH=3
OpH=5
~ 2000
E
@pH=3
OpH=5
Zeolite: DI Water
9
2000
Kaolinite: DI Water
1500
~1500
1000
500
@0
1000
O9
03
_.J
0
0
5OO
0.2
0.4
0.6
0.8
800
1000
Zeolite: BrackishWater
9 pH=3
0 pH=5
"
m 800
E
[]
O pH=7
600
o~ 600
o
Kaolinite:BrackishWater
[]
IpH=3
OpH=5
[] pH=7
400
"O
400
"a
200
t~
[]
0
0.0
'
0.5
1.0
1.5
2.0
9 pH=3
-~ 1200
1000
10
15
1400
Zeolite: Sea Winter
20
Kaolinite: Sea Water
@pH=3
[]
1200
0 pH=5
03
2.5
1400
[]
200
O3
- OpH=5
1000 - []pH=7
[] pH=7
03 800
800
-o
Q
600
600
400
03
"1o
m
_1
200
200
[]
oi
0
Lead Concentration (mg/L)
O@
[]
400
Q
5
10
Lead Concentration (mg/L)
Figure 4 - Lead Sorption Isotherms on Kaolinite and Zeolite Under Various Conditions
15
247
zeolite. However, the results showed zeolite sorption capacity is still significant (although
reduced) at such conditions. Acidification did not seem to negatively impact the lead
sorption on zeolite. The highest capacity was obtained at a pH=5.
Conclusions
The consolidation tests showed that the non-woven geotextile can be an effective
filter that has minimal clogging potential and promotes consolidation of the underlying
sediment layer. Further testing is needed to assess the long-term clogging potential of the
layered systems.
A summary of the adsorption test results clearly demonstrates the significance of
using zeolite as a sorption medium to collect lead from contaminated sediments. Brackish
water and sea water will contain significant concentrations of ions that will compete with
lead for sorption sites on zeolite. However, the results showed zeolite sorption capacity is
still significant (although reduced) under such conditions. While acidification did not
negatively impact the lead sorption on zeolite, the highest capacity was obtained at pH =
5.
The next steps in this research include additional benchtop testing using an illitic
clay, and a field test using a prototype reactive geocomposite.
Ac~ow&dgmen~
The research described in the paper was supported primarily by a grant from the
National Science Foundation, CMS-0086733, awarded under the New Technologies for
the Environment program. Direct and in-kind support was also provided by Ten Cate
Nicolon and its Construction Products Division, which produces Mirafi geosynthetics.
The authors appreciate the support provided. Consolidation testing was performed by
Mr. Ming Zhang, former graduate student at Northeastern University.
References
Aydilek, A. H., Edil, T. B., and Fox, P. J., 2000, "Consolidation Characteristics of
Wastewater Sludge," Geotechnics of High Water Content Materials, ASTM STP
1374, T. B. Edil and P. J. Fox, Eds., American Society for Testing and Materials,
West Conshohocken, PA, pp. 309-323.
Bishop, D., Johannssen, K., and Kohlhase, S., 1994, "Recent Applications of Modem
Geosynthetics for Coastal, Canal and River Works," Proceedings, 5th lntl. Conf.
on Geotextiles, Geomembranes and Related Products, Singapore, pp. 545-548.
Cooke, D., Welch, E. B., Peterson, S. A., and Newroth, P., 1993, Restoration and
Management of Lakes and Reservoirs, Lewis Publishers, Tokyo.
Giroud, J. P., 1996, "Granular Filters and Geotextile Filters," Proceedings, Geofilters
'96, J. Lafleur and A. L. Rollin, Eds., Bitech Publications, Richmond BC, Canada,
pp. 565-680.
248
GSA Resources, Inc. (2001). Personal Communication.

Haggerty, G. M. and Bowman, R. S., 1994, "Sorption of Inorganic Anions by Organozeolites," Environmental Science and Technology, Vol. 28, pp. 452-458.
Henry, K. S., Collins, C. M., and Racine, C. H., 1995, "Use of Geosynthetics to Prevent
White Phosphorus Poisoning of Waterfowl in Eagle River Flats, Alaska,"
Proceedings, Geosynthetics "95, Nashville, TN, Vol. 2, pp. 483-496.
Henry, K. S. and Stark, J. A., 1997, "Geosynthetic Barriers to Prevent Poisoning of
Waterfowl," Proceedings, 14th Intl. Conf. on Soil Mechanics and Foundation
Engineering, Hamburg, A.A. Balkema, pp.1819-1922.
Holtz, R. D, Christopher, B. R., and Berg, R. R., 1997, Geosynthetic Engineering, BiTech
Publishers, Ltd., Richmond, BC, Canada, 452 pp.
Jacobs, P. H. and Forstner, U., 1999, "Concept of Subaqueous capping of Contaminated
Sediments with Active Barrier Systems (ABS) Using Natural and Modified
Zeolites," Water Research, Vol. 33, No. 9, pp. 2083-2087.
Kirkner, D. J. and Reeves, M., 1988, "Multicomponent Mass Transport with
Homogeneous and Heterogeneous Chemical Reactions: Effect of the Chemistry
on the Choice of Numerical Algorithm, 1: Theory," Water Resources Research,
Vol. 24, No. 1, pp. 1719-1729.
Koerner, R. M. and Ko, F. K., 1982, "Laboratory Studies on Long Term Drainage
Capability of Geotextiles," Proceedings, Second International Conference on
Geotextiles, Las Vegas, Nevada, Vol. 1, pp. 91-96.
Land and Water, 2001, "The Use of Geotextile Tubes for Spoils Containment," Land
and Water, Sept./Oct., pp. 41-42.
Langrnuir, D., 1987, "Overview of Coupled Processes with Emphasis on Geochemistry,"
Coupled Processes Associated with Nuclear Waste Repositories, Tsang, C. F.,
Ed., Academic Press, San Diego, Calif.
Leppert, D., 1990, "Heavy Metal Sorption with Clinoptilolite Zeolite: Alternatives for
Treating Contaminated Soils and Water, Journal of Mining Engineering, Vol. 42,
pp. 604-609.
Li, Z., 1998, "Chromate Extraction from Surfactant-Modified Zeolite Surfaces," Journal
of Environmental Quality, Vol. 27, Nos. 1-2, pp. 240-242.
Li, Z., Burt, T., and Bowman, R. S., 2000, "Sorption of Ionizable Organic Solutes by
Surfactant-modified Zeolites," Environmental Science and Technology, Vol. 34,
pp. 3756-3760.
249
Mangold, D. C. and Tsang, C., 1991, "A Summary of Subsurface Hydrological and
Hydrochemical Models," Review of Geophysics, Vol. 29, Feb. 1, pp. 51-79.
Misaelides, P. and Godelitsas, A., 1995, "Removal of Heavy Metals from Aqueous
Solutions Using Pretreated Natural Zeolite Material: The Case of Mercury (II),"
Toxicol. Environmental Chemistry, Vol. 51, pp. 21-30.
Mitchell, J. K., 1993, Fundamentals of Soil Behavior, John Wiley and Sons, New York.
Reyes O., Sanchez, E., Pelion, A., Borja, R., Colmenarejo, M. F., Milan, Z., and Cruz,
M., 1997, "A Comparative Study of Sand and Natural Zeolite as Filtering Media
in Tertiary Treatment of Wastewater from Tourist Areas," Journal of
Environmental Science and Health A, Vol. 32, Vols. 9-10, pp. 2483-2496.
Savage, C., 1986, "Chicago Area CDF Synthetic Liner and Sand Blanket Experience," in
A Forum to Review Confined Disposal Facilities for Dredged Materials in the
Great Lakes, report of the Dredging Subcommittee to the Great Lakes Water
Quality Board, International Joint Commission, Windsor Ontario.
Sakadevan, K. and Bavor, H. J., 1998, "Phosphate Adsorption Characteristics of Soils,
Slag and Zeolite to be Used as Substrates in Constructed Wetland Systems,"
Water Research, Vol. 32, No. 2, pp. 393-399.
Selim, H. M., 1992, "Modeling the Transport and Retention of Inorganics in Soils,"
Advances in Agronomy, Vol. 47, pp. 331-384.
Shanableh, A. and Kharabsheh, A., 1996, "Stabilization ofCd, Ni and Pb in Soil using
Natural Zeolite," Journal of Hazardous Materials, Vol. 45, Nos. 2-3, pp. 207-218.
United States Environmental Protection Agency, 1997, "The Incidence and Severity of
Contamination in Surface Waters of the United States," Vol. 1, National Sediment
Quality Survey, Rpt. No. EPA 823-R-97-006, Office of Water,
URL: http://www.epa.gov/OST/cs/congress.html
United States Environmental Protection Agency, 1996, "Guidance for In-Situ
Subaqueous Capping of Contaminated Sediments," Rpt. No. EPA 905-B96-004,
Great Lakes National Program Office (GLNPO),
URL: http://www.epa.gov/glnpo/sediment/iscmain
United States Patent Office, 2002, "Reactive Geocomposite for Remediating
Contaminated Sediments," patent application submitted by U.S. Army Corps of
Engineers, pending.
250
Yeh, G. T. and Tripathi, V. S., 1991, "A Model for Simulating Transport of Reactive
Multispeeies Components: Model Development and Demonstration," Water
Resources Research, Vol. 27, No. 12, pp. 3075-3094.
Section II1: Monitoring and Performance
Homa J. Lee, l Marlene A. Noble, 2 and Jingping Xu 2
Sediment Transport and Deposition Processes Near Ocean O u t f a l l s

in Southern California
Reference: Lee, H. J., Noble, M. A., and Xu, J. P., "Sediment Transport and
Deposition Processes near Ocean Outfalls in Southern California," Contaminated
Sediments: Characterization, Evaluation, Mitigation/Restoration, and Management
Abstract: An urbanized coastal ocean that has complex topography and large-scale
atmospheric and oceanographic forcing can contain a variety of sediment and pollutant
distribution patterns. For example, the central southern California Bight has two large
embayments, Santa Monica and San Pedro Bays, that are connected by a short, very
narrow shelf off the Palos Verdes peninsula. The complex topography causes quite
different oceanographic and sediment distribution patterns in this fairly small region of
the coastal ocean. In addition, three sewage outfalls discharge material over the outer
shelf. A large suite of sedirnent cores was obtained and analyzed for contaminants,
physical properties, accumulation rates, and grain sizes. Arrays of instrumented
moorings that monitor currents, waves, water clarity, water density and collect
resuspended materials were deployed. The data and models developed for the Palos
Verdes margin suggest that a large reservoir of DDT and its byproducts exists in the
coastal ocean sediment and will continue to be exhumed and transported along the shelf
for a long time. On the Santa Monica shelf, very large internal waves, or bores, are
generated at the shelf break. The near-bottom currents associated with these waves sweep
sediment and the associated contaminants from the shelf onto the continental slope. On
the San Pedro margin an initial examination of recent data collected in the coastal ocean
does not suggest that bacterial contamination on local beaches is primarily caused by
transport of material from the adjacent ocean outfall.
Key Words: waste disposal, ocean, sewer outfalls, Southern California, environment,
physical oceanography, DDT
t Civil Engineer, U.S. Geological Survey, 345 Middlefield Road, Menlo Park, CA 94025.
2 Oceanographer, U.S. Geological Survey, 345 Middlefield Road, Menlo Park, CA
94025.
253
www.astm.org
254
Introduction
The central Southern California Bight, which is dominated by the Los Angeles
metropolis, is a heavily urbanized section of the coastal ocean. More than 12 million
people live along the coastal corridor between Point Dume at the northern end of Santa
Monica Bay and Newport Beach at the southern end of San Pedro Bay (Fig. 1). The short
and very narrow shelf off the Palos Verdes peninsula connects these two large bays.
Outfalls that presently discharge treated effluent and other materials into the coastal
ocean are located in the middle of Santa Monica Bay, at the eastern end of the Palos
Verdes shelf and at the southeastern end of Santa Pedro Bay. Near all three sites, there
are disused outfalls that historically discharged sewage and/or effluents with relatively
high levels of contamination. Storm-water runoff is discharged into the coastal ocean at
numerous sites along this margin. Historically, these several types of discharges have
been a source of chemical and biological contaminants into the coastal ocean system.
~ lma
Verd~s
,-..
:~. ,. i22 2'~':.~~.............
_e
,.~..... .
llun~n
........
,-,,,,,,,,;
Figure 1 - - General circulation patterns in the Southern California Bight (after Hickey
1992).
Once contaminants enter the coastal ocean, they are transported by a variety of
processes. Many chemical (e.g. DDT) and some biological contaminants sorb onto free
sediment and are initially deposited near the source. However, this free sediment and
associated contaminants is commonly resuspended by surface waves, internal waves, or
coastal currents and transported far from the source regions by currents in the mid to
lower portions of the water column. Other effluent material remains suspended in the
LEE ET AL. ON OCEAN OUTFALLS
255
water column for longer periods of time and is transported by currents near the surface or
in the middle of the water column.
Currents over the continental slope and on the eastern margins of the adjacent basins
in the bight tend to flow toward the northwest (Fig. l, Hickey 1992). These northwesterly
slope currents are fed by the southeasterly flowing California Current System and by the
Undercurrent that flows along the continental slope northwestward from Baja California.
Although the transport over the slope is primarily toward the northwest in the Bight, the
complex topography in the region is one factor that causes the current, sediment transport
and sediment distribution patterns over the adjacent shelf to vary widely. Strong currents
transport suspended material primarily to the northwest over the Palos Verdes peninsula.
On the Santa Monica Shelf, just 25 km to the northwest, there is no preferred direction
for transport in the winter months (Fig. 2). Hence, it is difficult to predict the ultimate
fate of contaminates that enter into coastal waters in even this relatively small geographic
area.
Currents over the Santa Monica Shelf

Mid-shelf current
20
cm/s _2001 ~,' I~.,L ~ J

20
|along] shore
' . "direction
off snore
direction
Shelf-break~rrent
L
,.
Dec
1998
Feb
1 999
Currents over the Palos Verdes peninsula

40
cm/s
Mid-shelfcurrent
| ",,,
-40-]
" t " " 1"0"

,
'
'
'
'
40 -1 Shelf-break current
cm/s
-40M~iy
1992
'
'
'
'
~ebrua~
1993
Figure 2 ~ Mid-water, subtidal currents at two sites in the Southern California Bight.
256
CONTAMINATED
SEDIMENTS
The U. S. Geological Survey (USGS), together with many allied agencies, began a
series of programs in the early 1990s to determine the dominant processes in the Southern
California Bight that transport sediment and suspended contaminants. The region of the
coast from Santa Monica to San Pedro Bay was selected for initial studies because both
the topography and some contaminant distribution patterns, particularly the pattern of
DDT distribution, suggested that the transport pathways in these regions were connected
(Fig. 3). Three primary studies were located near existing ocean outfalls (Fig. 4). Arrays
I
~
- ~
-.
.
~
I
30 kin
~l|
; . . .
,.~-.;.,
9 ..
"~ ....
'%
-i
.':..
:" ~
: S( 9
Figure 3 -- The distribution of DDT in the sediment of the southern California Bight
(modified from Schiff 1998). The major source of DDT in the bight is the ocean outfall
off the Palos Verdes Peninsula. This figure shows that sediment from this outfall is being
transported far into Santa Monica Bay where it interacts with sediment affected by the
City of Los Angeles outfall, which discharges near the center of the bay.
of instruments measured currents, waves, water clarity, water density, and collected
samples of material suspended in the water column. Bottom sediment samples were
collected at several hundred sites for physical properties, geochronology, grain size,
biological activity and contaminant level analyses. High-resolution seismic and
bathymetric surveys (Fig. 5) determined both the seabed morphology and Holocenr
sediment thickness along much of the margin. Numerical models allowed us to
extrapolate the understanding developed locally to larger temporal and spatial regions.
These many sources of information have been and are continuing to be jointly analyzed
257
so that a refined understanding o f the sediment and contaminant transport and distribution
patterns can be developed for this region.
Figure 4 - - Locations of major ocean outfalls and USGS moorings, including dates of
deployment, in the Southern California Bight. Contour interval for bathymetry is 100 m.
Transport Processes Over the Palos Verdes Margin
In addition to metals and a number o f other compounds, DDT and its derivative
products, primarily p,p'-DDE, are known contaminants in the central southern California
Bight (Fig. 3). Much o f this DDT entered the coastal waters over the Palos Verdes shelf
in effluent from the Whites Point outfall as a result o f industrial production at a nearby
facility. The contaminant was sorbed onto fine sediment and much of it was deposited
near the outfall on the Palos Verdes shelf. Over 9 million cubic meters o f effluentaffected sediment lie within a 40 square kilometer area near the outfall (Lee et al. 2002.
Fig. 6). This sediment contains over 100 metric tons ofp,p '-DDE (Lee et al., 2002). DDT
is very persistent and accumulates in the marine food chain; at the highest trophic levels
it inhibits reproduction in birds that prey on fish and is found at concentrations deemed
unsafe for human consumption in white croaker and other endemic fish. Because it sorbs
onto f'me sediment, normal coastal ocean processes distribute DDT/p,p'-DDE over a wide
258
Figure 5 - - Oblique view of multibeam bathymetry and a digital elevation model Of

Santa Monica Bay, the City of Los Angeles, and the Palos Verdes margin (courtesy of J.
V. Garner and P. Dartnell, U.S. Geological Survey). Distance across the base of the
image is about 40 km.
region. Because the production of DDT in the United States was halted in the early
1970s, sediment deposited since that time has lower levels of contamination than do
deeper deposits. Accordingly, the maximum concentrations of DDT/DDE on the Palos
Verdes shelf commonly occur 20-30 cm below the seafloor (Fig. 7). However, over many
square kilometers, DDT/p,p'-DDE is found in significant concentrations (5-10 ppm) near
the surface of the seafloor because biological and physical processes transport some of
the heavily-contaminated sediment upwards.
The USGS and allied agencies designed a multiyear program in 1992 to determine 1)
the amount and distribution of DDT/p,p'-DDE on the Palos Verdes margin, 2) the
transport processes that move the contaminants, and 3) the fate of the contaminants.
Because the highest concentration of DDT/p,p'-DDE is buried, sedimentological studies
focused on how animals that live in the sediment transport contaminants upwards.
Investigators measured both the physical and biodiffusive rates for contaminant transport.
Other investigations focused on how water column processes resuspend and transport
contaminants. These results show that contaminants are indeed being mixed to the
surface of the seafloor, primarily through biodiffusion. Once they reach the seabed,
surface-wave-induced currents resuspend them and subtidal currents transport them.
259
Figure 6 - - Thickness (in cm) of effluent-affected sediment on the Palos Verdes margin
determinedfrom density logs of core samples. An integration of these data within the
context of a GIS yields a total volume of 9 million cubic meters (Lee et al. 2002).
The presence of surface wave currents large enough to resuspend contaminated
sediment was not associated with a preferred direction for subtidal currents on the shelf.
This is because the currents over the Palos Verdes shelf are not dominated by storms or
other wind-driven processes (Noble et al. 2002). The subtidal mid-depth and near-bed
currents flow primarily toward the northwest (Fig. 2), so contaminants are preferentially
transported toward Santa Monica Bay. In the process, some contaminants move off the
shelf, toward the slope, along the northwestern portion of the peninsula. Both the regional
DDT distributions (Fig. 3) and the footprint of the contaminated sediment deposit (Fig. 6)
on the Palos Verdes margin are consistent with the inferred sediment-transport patterns.
Numerical models that use the sediment properties, the contaminant properties, the
sediment-diffusion coefficients, contaminant sorption/desorption coefficients, and
measured oceanographic conditions were developed in order to determine the fate of the
contaminated deposit over time (Sherwood et al. 2002, Wiberg et al. 2002). These models
suggest that shelf sediment is primarily transported toward the northwest. Because
transport rates of the f'me-grained effluent-affected sediment are greater than transport
rates of natural sediment found southeast of the outfalls, erosion is likely to occur on the
southeast, or upstream end, of the effluent deposit. Accordingly, highly contaminated
sediment is likely to be eroded in that region. Thus, both models and measurements
suggest that DDT/p,p'-DDE will continue to be exhumed and transported from the Palos
Verdes shelf. There is still a large reservoir ofDDT/p,p '-DDE on the shelf. This
260
reservoir likely will be a potential source of contaminants to the region for a long time,
since measurements over the last 18 years suggest that the total amount of DDT/p,p 'DDE on the shelf has not decreased substantially (Lee et al. 2002). Hence a quick, natural
recovery for the region is unlikely. Studies of remediation methods (e.g. capping) are
underway.
, e , , l ~ l , , , l ~ l , , ,
lO
-:
-'..
20
~" 30
40
60
. . . .
. . . .
50
. . . .
i .
100 150
200
i . . . i . . . i . . p l . ,
1.0 1.2 1.4 1.6 1.8
p,p'-DDE (ppm)
2.0
Bulk density (g/cc)
Figure 7 - - Profiles o f p,p "-DDE concentration and bulk density vs. depth measured on a
sediment core obtained at a 60-m-deep site on the Palos Verdes shelf. The profiles show
that the most heavily contaminated sediment is buried by several tens o f cm although
sediment closer to the surface is still contaminated. The density profile shows a low
density layer extending to about 45 cm. This layer represents effluent affected sediment
deposited afierh the construction o f the ocean outfall on the Palos Verdes margin.
Transport Processes Within Santa Monica Bay
The USGS and allied local agencies initiated a multi-year program in 1997 to
determine sediment distributions and transport processes in Santa Monica Bay. The
sediment-distribution patterns on the Santa Monica shelf are somewhat unusual because
sediment grain-size distributions along similar isobaths change markedly over relatively
short distances. Fine sediment in the northern part of Santa Monica bay is abundant in the
middle and outer shelf regions (40-70 m). The amount of fine sediment at these water
depths decreases toward the southeast. In particular, the widest portion of the shelf, which
is in the southern part of the bay, has relatively little fine-grained sediment in the 50-70 m
depth ranges.
261
Two sites were occupied in the winter of 1998-1999 at locations just to the northwest
of the Los Angeles city sewage outfall in the middle of Santa Monica Bay (Fig. 4). One
site was near the shelf break, in 70 m of water; the other site was 2-3 miles inshore of the
shelf break, in 35 m. The temporal structure of the subtidal currents in the middle and
outer shelf regions was typical of patterns found in other regions of the continental shelf.
The currents tended to flow parallel to the isobaths, with a slight tendency for flow
toward the southeast (Fig. 2). Hence, one would expect that resuspended material and the
associated contaminants would move along the shelf, parallel to the isobaths. However, at
sites near the shelf break, resuspended material in the lower portion of the water column
is transported offshore because strong currents and large internal depressions of the
isotherms are generated over the shelf break during the maximum and falling phases of
spring tides. Isotherm deflection amplitudes were 20-40 m in water depths of 70 m. Nearbottom current speeds reached 35-40 em/s, amplitudes more than sufficient to resuspend
and transport fine sediment. The near-bed current direction was exclusively offshore
during these events. Hence, any material resuspended and transported by these events
leaves the shelf and is deposited on the slope. Bottom stress from these strong shelf-break
internal waves was much larger than bottom stress generated by surface waves (Fig. 8).
On the mid shelf, bottom stress was caused mainly by surface waves, a situation that is
more typical of continental shelf processes (Fig. 8).
The sediment grain-size distribution supports the inference that large internal wave
events keep the region swept free of free material. Photographs from a camera towed
across the continental shelf in this area shows that t'me material is found over the upper
portions of the slope and over the middle region of the shelf. However, cobbles and
gravel are the dominant sediment type near the shelf break. The exclusive offshore
direction of near-bed transport from these internal oscillations suggests that any
contaminants deposited from the nearby outfall will be transported off the shelf and out
of the human-impacted coastal system.
Transport processes on the mid-shelf, in 35 m of water, are similar to those found on
the Palos Verdes peninsula. The large offshore pulses associated with internal waves are
absent. Surface waves primarily resuspend fme materials and any associated
contaminants. Presumably, these materials are advected along the shelf by the current
field.
Transport Processes Within San Pedro Bay

Coastal contamination issues are highly visible in the southern part of San Pedro
Bay. Huntington Beach, which is adjacent to the Orange County Sanitation District ocean
outfall, was posted or closed over most the summer of 1999 due to bacterial
contamination. Contamination has continued to be a problem in succeeding summers
although the source of the contamination was unknown. Several studies examined the
likely sources on land, such as broken sewer pipes or discharges from storm-water drains.
They also identified a possible new source associated with coastal marshes (Barnett 2000,
Grant and Sanders 2000). Another hypothesis is that the beach was contaminated because
effluent from the outfall was transported across the shelf by subsurface coastal-ocean
transport processes. The primary hypothesis was that strong internal tides could move
material from the shelf break to the inner shelf. Once in the nearshore region, the
262
contaminants would then be transported to the beach by cross-shore currents or by

entrainment within flows caused by anthropomorphic structures. Other hypotheses were
that effluent was transported toward the coast through an adjacent submarine canyon or
was mixed to the surface and transported by diurnal wind forcing.
3
shear stress 'i ' current shear stress
dyne/cm2 1.5 f wave shear stresst
35 m site 1'
t
7 0 m ;ite
current shear stress
--r
.......
shear stress
dyne/cm2 1.5
0
18
shear stress !
dyne/cm2 1.51
15
211
wave shear stress
12
70 m site
i'
t
1
9, F " A ';'~'," . '~;', ,t:,'~~',,.... , ,,~,,t~, ,\,.~,'., .....,/~,~ .~. ]

4
Dec
1998
12
Feb
1999
26
Mar
1999
Figure 8 - Bottom stresses generated by surface waves and currents at a mid-shelf(35

m) and a shelf-break (70 m) site. Note that the stresses generated by currents is
generally much stronger that stresses generated by surface waves at the shelf-break site.
In the summer of 2001, the USGS and allied agencies deployed arrays of instruments
at more than 15 sites on the 5-6 mile wide section of the continental shelf at the southern
end of San Pedro Bay. The primary purpose of the arrays was to determine whether there
are mid-water cross-shelf transport processes that move contaminants, possibly from the
outfall located near the shelf break, onshore to local beaches during stratified summer
conditions.
263
Initial results from this program suggest that although coastal ocean processes can
move water and suspended materials 5- 6 km across the shelf, there is no connection
between coastal ocean processes and bacterial contamination on the beaches. In
particular, internal tides can move water with the same temperature as the effluent plume
(12 ~) toward the inner shelf (Fig. 9). However, no beach pollution events were
associated with this process in the entire summer of 2001. Six hydrographic surveys that
mapped the plume during the times when the beach was contaminated also showed that
there was always a spatial gap between the plume and the contaminated beaches. The
plume was seen to enter the submarine canyon with low bacterial levels, but this was not
associated with beach contamination events.
0
0 hours
~
60 m
0km
0 L
7km
6.2 hours
60J
7km
0km
0m
60 m
Okm
12.4 hours
7km
Figure 9 - - Transport o f cooler water on and off the shelf due to the internal tides.
Water cooler than 12 ~ is the same temperature as the effluent plume.
Conclusions
Clearly, even in this relatively small region of the western U.S. continental margin
(alongshore extent is less than 100 kin), many processes in the coastal ocean move free
264
sediment, associated contaminants, and suspended contaminates through the local coastal
system. In regions with complex topography, the local processes may change markedly
over small spatial scales. Surface waves dominate resuspension events on the Palos
Verdes peninsula and in the middle of the Santa Monica shelf. Yet, internal waves are
more likely to resuspend material along the shelf break near the middle of Santa Monica
Bay. These processes transport material across and off, rather than along the shelf.
Preliminary evidence suggests that internal wave resuspension is dominant only over
certain regions of the shelf break in Santa Monica and San Pedro Bays. However, if those
regions are near a sewage outfall, the presumed direction of transport pathways may
change markedly. We would expect that the transport of contaminants and other material
that remain suspended in the water column will also be affected by local changes in
transport associated with rapid changes in the shelf topography. The large data set
collected recently in the southern part of San Pedro Bay does show that there are erossshelf transport processes that advect suspended material. However, in this region these
processes are seldom strong enough to bring contaminants to the local beaches.
Developing an understanding of the many coastal ocean processes that transport
suspended and resuspended material in the coastal ocean is essential before realistic
predictive models can be developed to help manage the coastal ocean ecosystem.
Acknowledgments
Many agencies and institutions participated in portions of the above studies. The
authors would like to thank the City and County of Los Angeles, the National Oceanic
and Atmospheric Administration, Orange County Sanitation District, the Southern
California Coastal Water Research Project, and individuals at the Naval Postgraduate
School, University of Delaware, University of Southern California, Skidaway Institute of
Oceanography, Scripps Institution of Oceanography and SAIC.
References
Barnett, A. M., 2000, "Huntington Beach Closure: Relationships between High Counts
of Bacteria on Huntington Beachs and Potential Sources." Report prepared for
the Orange County Sanitation District.
Grant, S. B., and Sanders, B. F., 2000, "Tidal Transport of Bacteria Between the Talbert
Watershed and the Ocean." Interim Report 1 for the UCI Coastal Runoff Study
(CRIS). Report prepared for the National Water Research Institute, County of
Orange, Cities of Huntington Beach, Fountain Valley, Coast Mesa, Santa Aria,
Newport Beach.
Hickey, B. M., 1992, "Circulation Over the Santa Monica-San Pedro Basin and Shelf,"
Prog. Oceanogr. Vol. 30, pp. 37-115.
Lee, H.J., Sherwood, C. R., Drake, D. E., Edwards, B. D., Wong, F., and Hamer, M.,
2002, "Spatial and Temporal Distribution of Contaminated, Effluent-Affected
265
Sediment on the Palos Verdes Margin, Southern Califomia," Continental Shelf
Research, Vol. 22, Nos. 6-7, pp. 859-880.

Noble, M. A., Ryan, H. F. and Wiberg,, P. L., 2002, "The Dynamics of Subtidal
Poleward Flows over a Narrow Continental Shelf," Palos Verdes CA. Continental
Shelf Research, Vol. 22, Nos. 6-7, pp. 923-944.
Schiff, K., 1998, "Sediment Chemistry on the Mainland Shelf of the Southem California
Bight." Southern California Coastal Water Research Project 1997-98 Annual
Report, pp. 76-88.
Sherwood, C. R., Drake, D. E., Wiberg, P. L., Wheatcrofl, R. A., 2002, "Prediction of
the Fate ofp,p '-DDE in Sediment on the Palos Verdes Shelf, California, USA,"
Continental Shelf Research, Vol. 22, Nos. 6-7, pp. 1025-1058.
Wiberg, P. L., Drake, D. E., Harris, C. K., and Noble, M. A., 2002, "Sediment
Transport on the Palos Verdes Shelf over Seasonal to Decadal Time Scales,"
Continental Shelf Research, Vol. 22, Nos. 6-7, pp. 923-944.
Patrick J. Fox |
Numerical Model for Contaminant Transport in Consolidating Sediments
Reference:
Fox, P. J., "Numerical Model for Contaminant Transport in
Consolidating Sediments," Contaminated Sediments: Characterization, Evaluation,

Abstract: A numerical model for coupled, one-dimensional, large strain consolidation
and contaminant transport, called CCTI, is presented. The large strain consolidation
algorithm includes all the capabilities of a previous code, CS2. The contaminant
transport algorithm includes advection, dispersion, and linear equilibrium sorption with a
spatially varying distribution coefficient. Advection-dispersion transport is consistent
with spatial and temporal variations of porosity and seepage velocity in the consolidating
layer. The key to the transport model is the definition of two Lagrangian columns of
elements that follow the motions of the fluid and solid phases separately. Verification
checks of CCT1 show excellent agreement with analytical solutions for classical transport
problems involving advection, diffusion, dispersion, and sorption in rigid porous media.
The performance of CCTI is illustrated for an example problem involving consolidationinduced transport of a dilute contaminant species in a soft sediment layer.
Keywords: consolidation, contaminant transport, large strain, advection, dispersion,
sorption, numerical model
Introduction
Classical solutions for contaminant transport through porous media (Ogata 1970,
Bear 1972, Freeze and Cherry 1979) are derived strictly for rigid media in which no
deformation occurs during the transport process. This assumption is fully acceptable for
many cases involving flow through natural groundwater systems and waste containment
facilities because applied loads are small to nonexistent or the media has low
compressibility. There are, however, cases in which significant volumetric strains may
occur during the transport process. Such cases include disposal of contaminated slurry
wastes, simultaneous consolidation and mass transport through clay liners, consolidation
of contaminated clay deposits, and the consolidation of contaminated sediments due to
~Associate Professor, Department of Civil and Environmental Engineering and Geodetic

Science, The Ohio State University, Columbus, Ohio, USA.
266
Copyright9 2003by ASTMInternational
www.astm.org
FOX ON CONSOLIDATING SEDIMENTS
267
continuing sedimentation or the application of a subaqueous soil capping layer. Although

two published works provide theoretical discussion and analyses of the coupled problem
(Potter et al. 1994, Smith 2000), a contaminant transport model has yet to be developed
for a porous medium that has general constitutive properties and undergoes significant
volumetric strain.
This paper presents a numerical model for coupled, one-dimensional, nonlinear,
large strain consolidation and contaminant transport called CCT1 (_C_onsolidation and
Contaminant Transport 1). CCTI is based on the piecewise-linear approach (Yong et al.
1983, Fox and Berles 1997, Fox 2000, Fox et al. 2002) in which all variables pertaining
to problem geometry, material properties, fluid flow, and effective stress are updated at
each time step with respect to a fixed coordinate system. CCTI was adapted from a
previous code, CS2 (Fox and Berles 1997), which simulates vertical consolidation of a
soil layer subjected to surcharge loading. The development of CCT1 is first presented,
followed by verification checks and an example problem illustrating the performance of
the model for contaminant transport within a consolidating sediment layer.
Model Description
Geometry
A saturated soil layer of initial height 11o contains an arbitrary distribution of a
single dilute contaminant species. The soil layer is treated as one-dimensional and is
represented using two Lagrangian columns of elements--one for the solid skeleton
(including voids) and one for the fluid moving through the skeleton. Solid and fluid
elements move independently except that they are coupled via values of solid element-tosolid element relative fluid velocity that govern the movement of each system. The
Lagrangian analysis framework follows the motion of both fluid and solid phases
throughout a simulation.
The initial geometry of the soil layer, prior to the application of a vertical stress
increment (time t < 0), is shown in Figure l(a). Vertical position coordinate z and
element coordinatej are defined as positive upward from a fixed datum at the bottom of
the layer. Initially, both fluid and solid phases are represented as a column of Rj
elements. E a c h f h element has unit cross sectional area, initial height Lo, a central node
located at initial elevation Zoj, and an initial void ratio eoj. Fluid in t h e f h element has
initial volume VroJ = Loeo. j/(1 + eo.j), initial dissolved contaminant concentration co.j
[mass contaminant/volume fluid], and initial dissolved contaminant mass CIo,j = CojVio,~.
The initial concentration of sorbed contaminant on the solids of each element is assumed
to be in equilibrium with the corresponding fluid concentration. Top and bottom
boundaries of the column can be specified as drained or impermeable and, if drained, are
assigned individual constant total head values, h, and hb, taken with respect to the datum.
Constant contaminant concentrations, c, and cb, are likewise assigned to the top and
bottom boundaries. The distribution of Co4 can be specified by the user or calculated by
CCTI (normally consolidated case only) using an iterative procedure such that it is
consistent with the initial overburden effective stress qo, the compressibility and self-
268
Figure
1 - Geometry for CCTI: (a) initial configuration (t < 0), and

(b) after application of vertical stress increment (t > 0).
FOX ON CONSOLIDATINGSEDIMENTS
269
weight of the soil, and seepage forces due to steady flow across the layer (if h, r h b).
A time-dependent vertical effective stress increment Aq' is applied to the column
beginning at time t = 0 and moves with the upper boundary thereafter. Excess pore
pressures created as a result of the stress increment cause fluid flow from the interior of
the soil mass to all drainage boundaries. Soil deformation occurs in response to the net
fluid outflow from each solid element. Contaminant transport occurs by advection and
dispersion in the fluid phase, and sorption onto moving solid elements. The configuration
of the system at some later time t is shown in Figure l(b). Fluid elements have become
disconnected from their original solid elements and move in response to the local seepage
velocity field. Fluid elements are numbered according to a new coordinate m defined as
positive upward from the datum. The height of the layer is H ' , the height of t h e f h solid
element is L'sj, and the height of the m th fluid element is L~,m. Solid elements and fluid
elements have node elevations z~,j and z'i,m, respectively, and are updated at each time
step. Fluid elements have concentration e'm and dissolved contaminant mass C~,,,. The
heights of fluid elements increase as the porosity of the associated solid elements
decreases. Fluid elements move through, and possibly out of, the consolidating layer in
response to boundary fluid outflow from solids compression and any external hydraulic
gradient across the layer (if h, r h b). Thus, unlike the number of solid elements which is
always constant at Rj, the number of fluid elements may decrease/increase during the
compression/swelling process. The top and bottom fluid elements at any given time are
m[ and re'b, respectively. The volume of solids within each solid element is invariant
throughout the consolidation process (Fox and Berles 1997). The volume of fluid within
each fluid element Vj,, is also invariant unless the element is adjacent to a drainage
boundary, in which case the element will gain or loose fluid in response to the flow across
the boundary. A fluid element is eliminated from the top or bottom of the fluid column
when its volume is fully depleted at an outflow drainage boundary. Likewise, a new fluid
element is created at an inflow drainage boundary when the existing boundary fluid
element is filled to capacity. The capacity of new fluid elements added at the bottom of
the fluid column is Vfo,l and the capacity of new fluid elements added at the top of the
fluid column is Vfo,Rj.
Constitutive Relationships
Constitutive relationships for the compressible layer, shown in Figure 2, are
specified using discrete data points. The compressibility curve (Figure 2a) is defined by
Rs > 2 pairs of corresponding vertical effective stress ~-' and void ratio ~. The hydraulic
conductivity relationship (Figure 2b) is defined by Rt > 2 pairs of corresponding void ratio
~' and vertical hydraulic conductivity k . A solid element becomes overconsolidated if
t,
the vertical effective stress decreases below the preconsolidation stress Crp,j
(i.e.,
maximum past vertical effective stress) for the element. In this case, unloading and
reloading follow an identical path defined by crp,j", ~*hecorresponding preconsolidation
270
o~
>
ejt
e~,j
I
i
I
I
<~jt
a't.
P,J
Vertical Effective Stress, ~'
(a)
~9
,-~
kj
t~
;>
ejt
Void Ratio, e
(b)
Figure 2 - Soil constitutive relationships: (a) compressibility, and
(b) hydraulic conductivity.

Ae
t
void ratio ep,j,
and a constant recompression index Cr - mAlogcr'
271
CCT1 uses the same
hydraulic conductivity relationship for normally consolidated and overconsolidated

conditions, which is consistent with the findings of A1-Tabbaa and Wood (1987) and
Nagaraj et al. (1994).
Aside from unload/reload effects, a one-to-one correspondence is assumed for each
constitutive relationship in Figure 2. Thus, CCT1 does not account for the effects of
strain rate, secondary compression, or aging on the compressibility or hydraulic
conductivity of the soil. The constitutive relationships can, however, take essentially any
desired form by choosing the appropriate number of data points.
Stress, Pore Pressure, Flow, and Settlement

CCT1 calculates stresses, pore pressure, fluid flow, and settlement in the same
fashion as does CS2. Fox and Berles (1997) presented detailed information on these
aspects and only a brief review is provided here. The vertical total stress at node j, crj, is
computed from the overburden stress on the layer and the self-weight of overlying
elements as,
t
crj'
qo + Aq' +(h t - H t )y,~
L'4ZJ
9
2
R/
' '
+ '~-~L,,,y,
i=j+l
j = 1, 2,
..., Rj
(1)
where y~ is the saturated unit weight of solid element j, e~ is the corresponding void
ratio, and y,,, is the unit weight of water (constant). The specific gravity of solids G, is
constant for the soil column and e'j is taken as constant within each solid element for any
given time increment.
Vertical effective stress cry' is computed from void ratio e~ by interpolating
l
between data points in Figure 2(a) if ejt <-em
or using the recompression curve if
ej'> ep,j'.
Vertical hydraulic conductivity k~' is computed from e jr by interpolating
between data points in Figure 2(b). The pore pressure u'j is the difference between total
and effective stresses and the relative discharge velocity (positive upward) v'r f , j between
solid element nodesj andj+l (Figure 3a) is,
t
v;r,j =-k~,j hj+, - h j
z;,j+, -
j = 1, 2 ..... Rj-1
(2)
z;.j
where the total head h~ and the equivalent series hydraulic conductivity k~.j are,
t
t = t
hj
Uj
-s.j + Yw
t
(3)
j = l , 2 ..... RTI
(4)
kj.lk j (Ls,j+I + L~,j )

k~,j- L' k'+f~ k'
s,j+l
j = I, 2, ..., Rj
~s,j'*j+l
~o
i
991
r-
--I~
r-
t-.
4-
r I
m
z
.-t
o)
m
0
0
0
z-q
.,,4
t~
273
Once the relative discharge velocities are known, a new height is computed for each solid
element from the net fluid outflow over time increment At,
t+~t
t
Ls, j = L s u - ( p ~ f , j - V t r f , j _ l ) A t
1,
2. . . . .
Rj
(5)
where At is calculated based on numerical stability and a c c u r a c y constraints (Fox and

Berles 1997). New void ratios, layer height, settlement, and average degree of
consolidation are, respectively,
ej
,+w
'+~'(1 + e~
Ls,j
(6)
j = 1, 2 . . . . . R j
Lo
RI
'
(7)
j=]
S ,
(8)
= H o - H,+~
SI+&t
(9)
s,
where St is the total settlement at the end of consolidation. Fox and Berles (1997) and
Fox (2000) show that numerical solutions obtained using the above method are
essentially indistinguishable from analytical and numerical solutions obtained using
material coordinates (Gibson et al. 1967, Lee and Sills 1979, Schiffman et al. 1984).
Contaminant
Transport
Mass transport of contaminant occurs by advection o f fluid elements, dispersion

between fluid elements, and sorption onto solid elements that are also moving in response
to the consolidation process. The Lagrangian framework of CCT1 automatically accounts
for advection transport in the fluid phase and sorption transport in the solid phase.
Dispersion transport is calculated between contiguous fluid elements with variable
porosity and seepage velocity in the associated solid elements taken into account. The
dispersion mass flux between fluid element nodes m and m+l is (Figure 3b),
t
Jim = - D '
Z~,m
c,.+, - c a
'
m -- m b', mb+l,
...,
m t' - 1
(I0)
Z;,m+l -- Z;,m
where the equivalent series longitudinal hydrodynamic dispersion coefficient D~,~ is,
O t
z~,,
.t
-- Zt
"f,m+l
f.m
N
m - -_r o b ,,
mb+l
t '
..-, mt
'-1
(11)
The value of N depends on the relative positions of fluid elements m and m+l and the
associated solid elements. Solid elements that contribute to the value of N satisfy one of
the following four possible cases,
Case 1:
Zs,' j +
L~o
L'
~ Zf,m+.
' I and z~,j
' - ~'
i-'" s_< z',,.
2
t
N=n,
Z f ,m+ 1 -- Z f , m
(12)
274
Vt
Vs
= ff,:-I +
Vrf,j -- Prf,j-I
it rt
I eff,./Ls,j
t
neff ,j
Z f ,m
Lts, j
Case 2: -~'s , j
-1-
N= Z
(13)
s,) +
t
< Z f', m + I a n d
2
,
z'
Ls'J
'
"': - 2 > zf.~
(14)
DE:`'
Vrf,j q- Vrf,j_ I
Vs ~
t
05)
2n ~g,y
+L;,j
Case 3: gs.jt
2
zt
, I and z.,t J +
<-~Zf,m+
.+
"':
L],j
'
t
t
Ls,j
_ f
2 >- zr ' and Zs': - 2 < f,m
- z:,~
N= Z .;::,,(D. + a~lv.I)
v,
Vrf,j
'
~ neff ,j
Vrf,j-I
'' "
neff ,j
(16)
rf,j -- Vrf,j-I
t
t
neff , j Z s , j
L~,j
Case 4: z~,j +
r
, - s
_>zy,m+~
and Z.,y
Zt
:,,.+, N= Z
vs
Zt
(17)
f,m--Z,,: +
'
'
,
and z.,j
' - L.,j
_< zy.m.~
- 9_> z:,~
2
2
4- Lts'j
z,j - T
(18)
n:yj(D'+ctL[V;[)
-- V r f , j - I
n'
eff,j
rf,j --
rf,j-I
+ ~ '
L'
Aneff,j
where nef a
(19)
zY"*z - zs'j +
s,j
is the effective porosity o f the f h solid element, s is a constant
(< 1), D* (= rDo) is the effective diffusion coefficient, r = apparent tortuosity factor, Do =
free-solution diffusion coefficient, aL = longitudinal dispersivity o f the medium, vs =
equivalent seepage velocity between nodes m and re+l, and the vertical bars signify
absolute value. Although ~, aL, and D* may possibly change with media porosity during
consolidation, these values are taken as constants in CCT1. The latter assumption is
supported by test data o f Shackelford and Daniel (1991) for compacted clays.
Mass Balance and Sorption-Desorption
t+~
The updated total contaminant mass in the m th fluid element C:.~
is computed as,
t+At
CI, m =C'f, m - ( J ' .
..[.
Jm_,)At
m = m b + l , mb+2 , ..., m~-I
(20)
275
where the time increment At is limited by numerical stability constraints for dispersion
(Zhu and Fox 2000). Corresponding mass balance expressions for boundary fluid
elements require an additional term to account for advective transport across the
boundary. CCTI accounts for sorption-desorption reactions by defining Rk subcells
within each solid element (Figure 4). The elevation of the node for subcell k within solid
elementj is z;c,lk . The mth fluid element exchanges contaminant mass with the solids of
any subcell j k if the node of the subcell lies within the boundaries of the fluid element
(i.e.. z'c'Jk
'
, ___E/.'~
and z',c,jk < zf,m
' +
> zLm
2
). The total mass of sorbed contaminant
on the solids associated with the m th fluid element is,

RI Rk
'
t ~t
Zt
j'~--t~--I' ~
,,:
m =
'
m b,
mb+l .....
m,'
(21)
t
l
where sjk
is the sorbed contaminant mass per mass of solid for s u b c e l l j k , Yd.~
is the dry
unit weight of solid element j, and g is the gravitational constant. The total contaminant
mass in the fluid and solid phases associated with the mth fluid element C~ ~' is,
t+At
t+At
C~a'=C~+.~ +Cf,,,
m = mb
t+At
. m. b . .+1,
.
t+~t
mt
(22)
CCT1 then calculates new fluid element concentrations assuming linear equilibrium
sorption between fluid and solid phases,
,+A, ~"
Cm
C'm+a'
RJ Rk K
vt+At
Lm
+~-,E
j = l k=l
m = m b,.A, , m b,+a, +l, ..., m,'+~'
t+At rt+At
d , j k Y d , j l~s,j
(23)
gR,
where Kd.# is the equilibrium distribution coefficient (sorbed concentration/aqueous

concentration) for subcelljk. CCTI treats K d as a spatially varying parameter to allow for
a non-uniform distribution of soil sorption characteristics, such as organic content. The
sorbed concentration on the subcells associated with the m th fluid element is then,
,+At= l~.d,jkC
. . . . m at
(24)
Sjk
m = m a,+A,, mt+~+
b
1, .... m t,+~,
Solid
Element j
k=R k
9
z
Sj.Rk Kd,j,Rk I
k+l
k
S,i,k+ 1
sjk
k-1
Sj,k. I Kd,i,k_l
k= 1
Sj, l
I
Kd,i,k+1
Ka,jk
Kd,~,l
?i
IL:j
t
sc.jk
Figure 4 - S o r p t i o n s u b c e l l s f o r s o l i d e l e m e n t j .
.!_
276
CCT1 assumes instantaneous linear reversible sorption for simplicity, although it

has been demonstrated for aquifers that, when a system is perturbed by a sudden increase
in advection rate, nonequilibrium sorption effects must be taken into account (Rabideau
and Miller 1994, Kong and Harmon 1996). Such effects may be important during the
early stages of consolidation when the advection rate may be high in comparison to the
rate of contaminant sorption/desorption in the system.
Model Performance
CCTI and CS2 produce identical results for large strain consolidation problems
without contaminant transport. The performance of CCTI was evaluated for several
classical transport problems without consolidation (Aq -- 0). Consider an initially
uncontaminated stiffsoil layer with H o = 6 m, Gs = 1, e = 0.8, ( = l, k = 2 x I0 -s m/s, and
drained top and bottom boundaries. At t = 0, the concentration of a dilute contaminant at
the top boundary is increased to ct and held constant thereafter. The concentration at the
bottom boundary is maintained at zero. Five simulations were performed with CCT1 (Rj
= 100, R, = 4) corresponding to the following cases: I) advection, 2) diffusion, 3)
advection + diffusion, 4) advection + dispersion, and 5) advection + dispersion +
sorption. The transport parameters for these simulations are provided in Table 1. The
retardation factor Rf (= l + 7dKd ] for the final simulation is 1.625. Each simulation
was
gne~ J
terminated at t = 6 x l08 s. Contaminant concentration ratio (c/ct) profiles are plotted vs.
elevation z along with corresponding analytical solutions for transport in a semi-infinite
domain (Domenico and Schwartz 1990) in Figure 5.
The Lagrangian formulation of CCTI reproduces exactly the simple case of plug
flow advection (Simulation 1). CCTI creates new fluid elements with c = ct as fluid
enters the layer at the top boundary. These elements move downward at the constant
seepage velocity and, in the absence of dispersion and sorption, maintain their initial
concentration. This produces a step concentration profile, from c = 0 to c = c, at the top
of fluid element m = Rj. It is worthwhile to note that an earlier version of CCT1 did not
utilize the Lagrangian formulation for fluid elements and calculated advection-dispersion
transport based on seepage velocities between contiguous solid elements (i.e., treating the
fluid in each solid element as well-mixed). This approach produced unacceptable
numerical dispersion and was unable to properly simulate plug flow transport.
The remaining plots in Figure 5 show that CCTI accurately calculates mass
transport for diffusion, dispersion, and sorption under steady fluid flow conditions.
Errors are small and tend to increase with depth for each simulation, which is likely due
to the inconsistent bottom boundary condition for the analytical and numerical solutions.
At c/ct = 0.5, the errors in concentration ratio for Simulations 2 through 5 are 0.016%,
0.16%, 0.039%, and 0.40%, respectively. If the simulations are repeated at a higher
numerical resolution o f Rj = 200, these errors decrease to 0.0040%, 0.040%, 0.0075%,
and 0.34%.
A final example illustrates the performance of CCTI for a hypothetical problem
involving simultaneous large strain consolidation and mass transport of a nonspecific
sorbing contaminant in a soft sediment deposit. The deposit has an initial thickness of 5
m and the contaminant has an initially uniform concentration of 100 ~g/L between eleva-
Table 1 Variable
h,, hb (m)
vs (m/s)
D* (m2/s)
Transport parameters for five simulations.
Simulation 1
Advection
Simulation 2
Diffusion
6.5, 6.0
3.75 x 10 "9 ,~
0
6.0, 6.0
aL (m)
Ka(mL/g)
0
0
R:
277
Simulation 3
Advection +
Diffusion
Simulation 4
Advection +
Dispersion
5 x 10-I~
6.5, 6.0
3.75 x 10 -9 ,~
5 x 10"~~
6.5, 6.0
3.75 x 10 "9 ,~
5 x 10-I~
0
0
1
0
0
1
0.2
0
1
'
'
Simulation 5
Advection +
Dispersion +
Sorption
6.5, 6.0
3.75 x l 0 "9 ~,
5 x 10-~~
0.2
0.5
1.625
~4~
3
N
~
//
,~ , g /
//v
1 ~
r
~
,. , [
Analytical_~
Solutions I
L
0
~
0
[]
- -
0.2
0.4
Diffusion
Advection + Diffusion
Advection + Dispersion
Ad_vec~ispersion
+ Sorption
CCTI
0.6
0.8
1.0
c/c t
Figure 5 -
Comparison of analytical and numerical solutions for contaminant transport in

a rigidporous medium under steadyfluidflow conditions.
278
tions 3.5 and 4.0 m. The effective diffusion coefficient D* is 8 x 10~~ m2/s and the
distribution coefficient Kd is 4 mL/g. The normally consolidated sediment has a~. = 0.5
m, Gs = 2.78, ( = 1, and the following compressibility and hydraulic conductivity
relationships (based on values reported by Fox (1996) for a high plasticity clay slurry),
e = 4.3 - 1.01ogcr'(kPa)
(25)
logk(m/s) = -11 + 0.77e
(26)
Pore pressures are initially hydrostatic and the bottom of the deposit is impermeable. The
sediment deposit is to be capped with a granular soil layer having a buoyant unit weight
o f 11 kN/m ~ Information is needed regarding the possible migration o f contaminants
into the cap as a result of the capping process.
Simulations were performed (Rj -- 50, Rk = 5, Rs = Rt = 31) for cap thicknesses of 1
and 2 m, corresponding to effective stress increments of 11 and 22 kPa (applied
instantaneously). The top boundary of the sediment is assumed to maintain zero
concentration. Figure 6(a) shows settlement of the sediment and contaminant mass
outflow per unit area at the top boundary vs. log time. Total settlements for the 11 and 22
kPa stress increments are 0.494 and 0.713 m, respectively, yielding average final vertical
strains of 10% and 14%. At U,,,g = 99.9% (t = 11,309 days), mass outflows of
contaminant are 2.78 x 10 -2 and 3.84 10-2 g/m 2 for dq = 11 and 22 kPa, respectively,
which amount to 18% and 24% of the initial total contaminant mass within the sludge.
Contaminant outflow is greater for the 22 kPa stress increment due to the larger seepage
velocities and total fluid outflow from the sediment. Figure 6(b) shows concentration
profiles in terms of relative elevation z/H at U,,vg = 30%, 60%, and 90% for both stress
increments. Consistent with Figure 6(a), contaminant concentrations in the sediment
decrease with increasing time and stress increment.
The location of maximum
concentration moves upward with time relative to the solid phase in response to advective
flow toward the top boundary. With respect to absolute elevation z, the maximum
concentration point moves downward with time due to the large strains occurring in the
layer.
Conclusions
CCT1 is a numerical model for analysis of coupled, one-dimensional, nonlinear,
large strain consolidation and contaminant transport. The consolidation algorithm of
CCT1 accounts for vertical strain, soil self-weight, general constitutive relations,
changing hydraulic conductivity and compressibility during consolidation, timedependent loading, unload/reload effects, and steady flow across the layer due to unequal
boundary head conditions. Constitutive relationships for the soil are defined using
discrete data points and can take nearly any desired form. The contaminant transport
algorithm of CCTI accounts for advection, dispersion, and linear equilibrium sorption
with the distribution coefficient as a spatially varying parameter. Advection-dispersion
transport is consistent with spatial and temporal variations of porosity and seepage
velocity in the consolidating layer.
Two Lagrangian columns of elements, one for the solid phase and the other for the
fluid phase, are defined in CCTI to properly simulate transport in a consolidating sedi
sediment layer. The columns move independently of one another except that they are
coupled via values of solid element-to-solid element relative fluid velocity that govern the
.........
.......
........
........
279
5.0
0.1
4.0
~~
0.2
"-" 0.3
r162
3.0
/i
0.4
9~
0.5
9
D -S, A q = 2 2 k P a
~,,'
2.0
G}
- - Q - - CMO, Aq = 11 kPa
0.6
--[]- cMo
[]
--22
1.0
0.7
(a)
~'o
. . . . . . . . , . . . . . . . . , . . . . ~ o" . . . . .
0
10
100
1000
10000
Time, t (d)
0.8
~.
~
1.0
0.8
0.6 ~
m
Initial
Uavg = 30%, Aq = 11 kPa
bq
0.4
[]
Uavg = 60%, Aq = 11 kPa
t,
Uavg = 90%, Aq = 11 kPa
- -|
0.2
- Uavg = 30%, Aq = 22 kPa
- -13- - U,vg = 60%, Aq = 22 kPa

(b)
,
0
0
- -'~'-- Uavg = 90%, Aq = 22 kPa

~
20
40
60
80
100
120
Concentration, c ( g g / L )
Large strain consolidation of a hypothetical contaminated sediment deposit:

(a) settlement and contaminant mass outflow, and (b) dissolved contaminant
concentration profiles.
Figure 6 -
280
movement of each system. The Lagrangian analysis framework follows the motion of
both fluid and solid phases throughout a simulation.
CCT1 provides the following quantities as a function of time: 1) settlement of the
consolidating layer, and 2) total mass outflow of contaminant at each boundary. CCTI
provides the following quantities as a function of time and position within the
consolidating layer: 1) void ratio, 2) pore pressure, 3) vertical effective stress and
preconsolidation stress, and 4) contaminant concentration, including dissolved and sorbed
fractions.
The performance of CCTI for transport in a uniform porous medium (i.e., no
consolidation) was evaluated for five simulations involving: 1) advection, 2) diffusion,
3) advection + diffusion, 4) advection + dispersion, and 5) advection + dispersion +
sorption. Contaminant concentration profiles for each simulation were in excellent
agreement with corresponding analytical solutions for transport in a semi-infinite domain.
The performance of CCT1 was then illustrated for an example problem involving
consolidation-induced transport of a sorbing contaminant in a soft sediment deposit.
Simulations indicate that contaminant mass outflow from the layer increases with
increasing applied stress increment.
Acknowledgments
Financial support for this investigation was provided in part by Grant No. CMS9622644 from the Geomechanical, Geotechnical, Geoenvironmental Systems Program of
the U.S. National Science Foundation and is gratefully acknowledged. The author thanks
Prof. Thomas C. Harmon of UCLA for several enlightening discussions on mass transport
in sorptive porous media.
References
A1-Tabbaa, A. and Wood, D. M., 1987, "Some Measurements of the Permeability of

Kaolin," Geotechnique, Vol. 37, No. 4, pp. 499-503.
Bear, J., 1972, Dynamics of Fluids in Porous Media, Elsevier, New York, 764 pp.
Domenico, P. A. and Schwartz, F. W., 1990, Physical and Chemical Hydrogeology, John
Wiley & Sons, New York, 824 pp.
Fox, P. J., 1996, "Analysis of Hydraulic Gradient Effects for Laboratory Hydraulic
Conductivity Testing," Geotechnical Testing Journal, Vol. 19, No. 2, pp. 181-190.
Fox, P. J. and Berles, J. D., 1997, "CS2: A Piecewise-Linear Model for Large Strain
Consolidation," International Journal for Numerical and Analytical Methods in
Geomechanics, Vol. 21, No. 7, pp. 453-475.
Fox, P. J., 2000, "CS4: A Large Strain Consolidation Model for Accreting Soil Layers,"
Geotechnics of High Water Content Materials, Special Technical Publication 1374,
T. B. Edil and P. J. Fox, eds., ASTM, pp. 29-47.
Fox, P. J., Di Nicola, M., and Quigley, D. W., 2003, "Piecewise-Linear Model for Large
Strain Radial Consolidation," Journal of Geotechnical and Geoenvironmental
Engineering, in press.
281
Freeze, R. A. and Cherry, J. A., 1979, Groundwater, Prentice-Hall, Englewood Cliffs,

New Jersey, 604 pp.
Gibson, R. E., England, G. L., and Hussey, M. J. L., 1967, "The Theory of OneDimensional Consolidation of Saturated Clays, I. Finite Non-Linear Consolidation
of Thin Homogeneous Layers," Geotechnique, Vol. 17, No. 3, pp. 261-273.
Kong, D. and Harmon, T. C., 1996, "Using the Mutiple-Cell Balance Method to Solve the
Problem of Two-Dimensional Groundwater Flow and Transport with
Nonequilibrium Sorption," Journal of Contaminant Hydrology, Vol. 23, pp. 285301.
Lee, K. and Sills, G. C., 1979, "A Moving Boundary Approach to Large Strain
Consolidation of a Thin Soil Layer," Proceedings, 3rd International Conference on
Numerical Methods in Geomechanics, Aachen, Balkema, Vol. 1, pp. 163-173.
Nagaraj, T. S., Pandian, N. S., and Narasimha Raju, P. S. R., 1994, "Stress-StatePermeability Relations for Overconsolidated Clays," Geotechnique, Vol. 44, No. 2,
pp. 349-352.
Ogata, A., 1970, "Theory of Dispersion in a Granular Medium," Professional Paper 411/, U.S. Geological Survey.
Potter, L. J., Savvidou, C., and Gibson, R. E., 1994, "Consolidation and Pollutant
Transport Associated With Slurried Waste Disposal," Proceedings, 1st International
Congress on Environmental Geotechnics, pp. 525-530.
Rabideau, A. J. and Miller, C. T., 1994, "2-Dimensional Modeling of Aquifer
Remediation Influenced by Sorption," Water Resources Research, Vol. 30, No. 5,
pp. 1457-1470.
Schiffman, R. L., Pane, V., and Gibson, R. E., 1984, "The theory of one-dimensional
consolidation of saturated clays. IV. An Overview of Nonlinear Finite Strain
Sedimentation and Consolidation," Proceedings, Symposium on Sedimentation/
Consolidation Models: Prediction and Validation, R. N. Yong and F. C.
Townsend, eds., San Francisco, ASCE, pp. 1-29.
Shackelford, C. D. and Daniel, D. E., 1991, "Diffusion in Saturated Soil. II: Results,"
Journal of Geotechnical Engineering, Vol. 117, No. 3, pp. 485-506.
Smith, D. W., 2000, "One-Dimensional Contaminant Transport Through a Deforming
Porous Medium: Theory and a Solution for a Quasi-Steady-State Problem,"
International Journal for Numerical and Analytical Methods in Geomechanics, Vol.
24, No. 8, pp. 693-722.
Yong, R. N., Siu, S. K. H., and Sheeran, D. E., 1983, "On the Stability and Settling of
Suspended Solids in Settling Ponds. Part I. Piece-Wise Linear Consolidation
Analysis of Sediment Layer," Canadian Geotechnical Journal, Vol. 20, No. 4, pp.
817-826.
Zhu, Y. and Fox, P. J., 2000, "Smoothed Particle Hydrodynamics Model for Diffusion
Through Porous Media," Transport in Porous Media, Vol. 43, No. 3, pp. 441-471.
Antonina Degtiareva, l Maria Elektorowicz, I and Taghi Ebadi I
Assessment o f the Lead Release from Cables B u r i e d in Sediments into

the W a t e r C o l u m n
Reference: Degtiareva, A., Elektorowicz, M., and Ebadi, T., "Assessment of the Lead
Release from Cables Buried in Sediments into the Water Column," Contaminated
Sediments: Characterization, Evaluation, Mitigation~Restoration, and Management

Abstract: The study investigates the change in the water column quality in a situation
where submarine communication cables cross the water bodies. In spite of precautions in
technological achievements it is possible that these cables can be subjected to potential
corrosion processes. This study was performed for a particular area of the St. Lawrence
River. The impact of cables buried in sediments on lead speciation in the water column
was shown under different conditions (anaerobic and aerobic, at different partial
pressures of CO2, presence of organic acids). Equilibrium in the water was calculated
taking into account gas, water and solid phases. Programs Equilibrium and EPH from
FACT were utilized in this study to calculate Eh-pH diagrams and the equilibrium in
water. The maan form of lead was found to be PbOH (3.34x10 M/kg), the concentration
of the free lead ion was two orders of magnitude lower (2.5x10 "It M/kg). The
calculations, in which solid phases where considered, demonstrated that the water was
oversaturated with respect to dolomite and iron(Ill) hydroxide. It was speculated that
under these conditions dolomite could form a protective layer around the cable, which
can delay the corrosion. Under anoxic conditions cerussite and lead sulfide precipitated.
When Pco2 increased, the concentration of the free ion of lead increased.
9
-9
Keywords: lead, cable, corrosion, speciation, water quality, quality of sediments, Eh-pH
diagrams
1PDF, Associate Professor, and PDF, respectively, Department of Building, Civil, and
Environmental Engineering, Concordia University, 1455 Blvd de Maisonneuve W,
Montreal, PQ, H3G 1M8, Canada
282
www.astm.org
DEGTIAREVA ET AL. ON LEAD RELEASE
283
Introduction
Communication cables often cross water bodies in industrial areas. They consist of
copper base, lead and iron protective layers and asphalt covering. These cables could
possibly be subjected to the corrosion process. The process of corrosion and the release
and fate of corrosion compounds in the water bodies are complex and affected by various
factors.
Most metals have a tendency to corrode due to a reaction with the surrounding media,
such as water and air. An important way of considering corrosion processes is the
thermodynamic one, which shows under which conditions a metal will or won't corrode.
If the corrosion in moist environments is considered, redox-potential-pH (Pourbaix)
diagrams are necessary to estimate distribution of corroded species in these particular
environmental conditions (Wranglen 1985).
This study was conducted to investigate the possibility of corrosion and the possible
water quality change due to corrosion of the cable containing lead in the St. Lawrence
River. However, this methodology can be applied to other cases. A specific section of the
St. Lawrence River (north-shore mix) that represents the area of concern for
communication industries was studied. The average values of the concentration of
dissolved species present in the river were taken from the State of the Environmental
Report on the St. Lawrence River (St. Lawrence Centre 1996a, 1996b).
The construction and use the Eh-pH diagrams help to estimate conditions of possible
corrosion. In the present study the program EPH from FACT (Pelton et al. 1998) was
used to construct Eh-pH diagrams. To construct a diagram that would closely represent
the existing conditions of the St. Lawrence River, the following ions were taken into
account: SO42, C I and Ca 2+. The activities of the SO42, C I and Ca 2+ ions introduced into
a program were considered equal to the concentrations of these ions in the St. Lawrence
River. Concentrations of these ions above average values for the St. Lawrence River
were also considered in some cases.
Effects of the different factors such as physico-chemical characteristics of water
affecting the speciation and release of corrosion compounds were studied. The
equilibrium in water of the St. Lawrence River was calculated using the program
Equilibrium from software FACT (Pelton et al. 1998).
Impact of Oxic, Anoxic and Acid-base Conditions on Lead Corrosion
Impact of C02 Partial Pressure, Presence of Sulfur, Chloride and Calcium

Corrosion of cables proceeds differently, depending on a number of factors among
them the depth of burial. Cables lying on the sediments surface where water can be in the
equilibrium with air are subjected to higher concentrations of dissolved oxygen than the
cables buried in sediments. If cables are buried in sediments under anaerobic conditions,
there is an insufficient amount of oxygen in pore water, values of redox potential are
below zero, and corrosion could proceed differently.
284
The following systems were considered and the Eh-pH diagrams were calculated using
the program EPH from FACT to simulate the potential conditions to which
communication cables can be subjected:
Case 1 - Water was considered to be in equilibrium with air, therefore a partial
pressure was assumed to be Pco2 = 0.0003 atm. First, the simple system Pb-C-H20 was
considered to better illustrate the influence of changing conditions and different
concentrations of ions in water. Figure 1 shows the areas of predominance of ions Pb 2
and PbOH and the thermodynamic ranges where corrosion occurs (pH varies from 2 to
8.75). In the predominant areas of solid precipitates, PbCO3, PbO2 and Pb, corrosion is
still possible, but precipitated solids may (under suitable conditions) form a partially or
fully protective coating that delays corrosion. On the bottom of water bodies, especially
in the presence of microorganisms, higher concentrations of COz (Pco2 = 0.01 atm) may
be present.
Increased concentrations of carbon dioxide decrease the possibility of corrosion,
which is shown in Figure 2, where an increase of carbon dioxide concentration was
considered (Pcoz = 0.01 atm). The results demonstrate that the predominant area of lead
carbonate shifts down to pH = 7.25.
Pb-C-H20, 298.15 K
log10(P(CO2)(G))= -3.52E+00
..-. 1
iO.5I f
PI~*
'
PLO) i.l+ ~
10
pH
Figure 1 - Eh-pH diagram (Pcoz =0.003 atm). The total dissolved Pb species: 10 "6 M.
Case 2 - In this case, an additional element (S) was added to calculations performed
by the same program. The activity of the SO42 ion was considered equal to the
concentration of this ion in the St. Lawrence River. The following system was
considered: Pb-C-S-H20. The area of the solid lead sulfide appears on the Eh-pH diagram
(Fig. 3). As mentioned previously, precipitation of solids may delay the corrosion
process. As sulfides are the least soluble species, it could be expected that a low content
of soluble lead in the water develops under reduced conditions. A small increase of the
sulfate ion concentration does not seem to change the areas of predominance on the
diagram, however, if its activity increases up to 0.1 M/kg (mole per kilogram of the water
column), solid lead sulfate becomes predominant under low levels of pH (from 2 to 7)
and can also form a protective layer as is shown in Figure 4.
285
Pb-C-H20, 298.15 K
Iogl0(P(CO2)(G)) = -2.00E+00
1.0
4.+ 0.5
O
>
1+o
J+
I r
I I
0.0
-0.52
6 pH
....
Pbco+,
8~
-10
Figure 2 -Eh-pH diagram (Pco2 = 0.01 atm). The total dissolved Pb species: 10-6 M.
Pb-C-S-H20, 298.15 K
Iogl0(P(CO2)(G)) = -3.52E+00, Iogl0(A(SO4[--])(AQ)) = -3.56E+00
(s)
pH
10
Figure 3 - Eh-pH diagram for Pb-C-S-H20. The total dissolved Pb species: 10.6 M.
286
Pb-C-S-H20, 298.15 K
Iogl0(P(CO2)(G)) = -3.52E+00, I o g l 0 ( A ( S O 4 [ - ] ) ( A Q ) ) = -1.00E+00
1.0
0.5
- ' ~
. . . .
0.0
-0.5
pH
Figure 4 - Eh-pH diagram for Pb-C-S-H20 (activity O.1 M/kg). The total dissolved Pb
species: 10.6 M.
Case 3 - In this case, CI was introduced into calculations. As it was done previously
with SO4 the activity of CI- was considered equal to its concentration in the St. Lawrence
River. The diagram for the following system: Pb-C-CI-H20 was calculated. Figure 5
demonstrates that chloride ion at low concentrations does not affect the corrosion;
although it can form complexes with lead and subsequently, increase the concentration of
the total soluble lead in water.
Pb-C-CI-H20, 298.15 K
log10(P(CO2)(G)) = -3.S2E+00, Iogl0(A(CI[-])(AQ)) = -3.25E+00
,.o
9"*
O
>~
0.5
u,I
-0.5
10
pH
Figure 5 - Eh-pH diagram for Pb-C-CI-H20. The total dissolved Pb species." 10.6 M.
287
Chloride ions in the absence of strong cations can also decrease pH that would
accelerate the corrosion. This fact can be easy illustrated by the following calculations,
which were performed using the program Equilibrium from software FACT (Pelton et al.
1998):
Case 3a - The following system was considered: Pb+ 1000gH20+CO2+O2. The
following assumption has been made: Pco2 is 0.003 atm (ten times higher than Pco2 in
water in equilibrium with CO2 in air) and 02 is in equilibrium with air (Po2=0.21 atm). In
these conditions, pH of this system would be 6.25 and concentrations of predominant
lead ions would be as follows:
pb2+= 2.33x10 5 M/kg,
PbOH + = 2.80x 10-5 M/kg.
Case 3b - Upon addition of chloride to the system 3a (total dissolved chloride was 20
mg/L as in the St. Lawrence River) the value of pH decreased to 5.74 and changed
subsequently the lead speciation:
Pb z+ = 2.4x 10 "4 M/kg,
PbOH+ = 9.1x10 5 M/kg,
PbCI+= 5.5 x107 M/kg.
The calculations showed that the concentration of the total dissolved lead increased
by an order of magnitude; the concentration of the lead ion increased by one order of
magnitude and this ion found to be a predominant species.
When the activity of CI" increases up to 0.1 M/kg, the complex PbCI + becomes the
predominant one at low pH. This fact can be illustrated by diagram in Figure 6. To
calculate this diagram the activity of C1- introduced to calculations was assumed to be
equal to 0.1M/kg. The diagram shows that the corrosion occurs in these conditions.
Pb-C-CI-H20, 298.15 K
Iogl0(P(CO2)(G)) = -3.52E+00, Iogl0(A(CI[-])(AQ)) = -1.00E+O0
0.5
0.0
-0,5
10
pH
Figure 6 - Eh-pH diagram f o r Pb-C-CI-H20 system (activity O. 1 M/kg). The total

dissolved Pb species." 10 -6 M.
Case 4 - In this case, the system including Pb, C, C1, S, H20 at Pco2 = 0.003 atm was
considered. To calculate a diagram that represents conditions similar to those in the St.
Lawrence River, activities of SO4 and C I were assumed to be 0.000275 M/kg and
288
0.0005686 M/kg respectively. The applied conditions were close to those observed on the
bottom of the St. Lawrence River. In these conditions, PbO2 precipitating at high Eh,
PbCO3 at high pH and PbS at low Eh are the predominant lead forms (Fig. 7). They could
form a protective layer and delay the corrosion. If this layer is damaged (for example, due
to coarse sediments movement), the corrosion will occur. An increase of concentrations
of CO2 around the cable, which is possible due to the presence of microorganisms, may
decrease the corrosion.
Pb-C-CI-S-H20, 298.15 K
Iogl0(P(CO2)(G)) -- -2.52E+00, Iogl0(A(CII-I)(AQ)) = -3.25E+00,
Iog10(A(SO4[-])(AQ)) = -3.56E+00
0, i
~"
o.~ : ~ - ' ~
pH
PbC'O~s~
10
Figure 7 - Eh-pH diagram for Pb-CI-S-H20. The total dissolved Pb species." 10 -6 M.

Case 5 - The following system, which included calcium was considered: Pb-C-CaH20. The results of calculations (Fig. 8) demonstrate that the presence of calcium does
not seem to affect the corrosion at activities close to concentrations of Ca found in the St.
Lawrence River (up to 0.1 M/kg).
At the same time, the presence of calcium and calcium carbonates in water may raise
pH; therefore, decreasing the content of lead in the solution, which can be demonstrated
in the following calculations:
Case 5a - Upon addition of calcium (the total concentration of dissolved Ca = 25
mg/L as in the St. Lawrence River) to the system 3a, pH of the resulting solution
becomes 7.56 and the concentrations of predominant lead ions are:
Pb2+= 5.57x10 "8 M/kg,
PbOH = 1.37x 10-6M/kg.
The results of calculations show that the total concentration of soluble lead is
approximately 10 times lower than in the case 3a. The concentration of free lead ion
became 3 orders of magnitude lower while lead hydroxide ion becomes highly
predominant in the system.
Although Eh-pH diagrams are a valuable tool (Stumm and Morgan 1996) to predict
conditions of the possible corrosion, Nordstrom and Munoz (1986) pointed out that they
can not be used for quantitative chemical modeling because of the following reasons: a)
constant activities of ions are used to construct diagrams; b) noideality corrections are not
289
made; c) the redox potential of a natural water is not a clearly defined. Therefore, in order
to investigate the fate of corrosion products and the possibility of lead release into the
water column, the water-mineral equilibrium computations were applied in this study.
Pb-C-Ca-H20, 298.15 K
Iog10(P(CO2XG))= -3.52E+00, Iogl0(A(Ca[++I){AQ)) = -1.00E+00
1.0
fi
0.5
,,.,,
I11
0.0
-0.5
pH
10
Figure 8 - Eh-pH diagram Pb-C-Ca-I-I.,O (activity 0.1 M/kg). The total dissolved Pb
species: 10-6 g.
Fate of Corrosion Products Under Oxic and Anoxie Conditions and Their Influence
on Water Quality
Fate of Inorganic Corrosion Products

The equilibrium in the water of the St. Lawrence River has been calculated using the
program Equilibrium from FACT (Faculty for Analysis of Chemical Thermodynamics).
Investigations conducted to study corrosion products showed that in the aquatic
environment the main products of corrosion were lead carbonates (cerussite and
hydrocerussite) and lead oxides (Nordstrom and Munoz 1986, Pintado and Montero
1992, A1-Khafari and Badawy 1998). Therefore, in the present study the phase that
controls the solubility of lead in the water column near the cable was assumed to be
cerussite. With increase of Pco2 near the bottom of the river and in sediments (close to
cables), cerussite apparently is more likely to be formed than oxides. At the same time,
being in a more soluble phase, it may control the solubility of lead in the water column
near the cable.
In order to investigate the speciation of lead in the water column, the following
elements were taken into account: Ca, Mg, Na, K, S, C1, Pb, Fe and H20, CO2, 02.
Subsequently, the following parameters were considered:
290
3.38xl0 9 Pb + 55.5H20 + CO2 + 02 + 5.686x 104 C1 + 7.75x10 4 Ca + 2.92x10 4 Mg +

4.65x10 4 Na + 2.75x10 4 S + 1.23x10 5 Fe
The concentrations of all species are given in moles per kilogram of water (M/kg).
It was assumed that CO2 and 02 were in equilibrium with gases in air, where partial
pressures Pco2 = 0.0003 atm and Poz = 0.21 atm. No solids were allowed to precipitate.
The calculations showed:
- The ionic strength was 0.0039;
- The calculated values of Eh and pH were: Eh = 713 mV, pH = 8.5
The speciation of lead was calculated and the chemical forms of lead present in water
were listed in Table 1.
Table 1 - Lead speciation in the St. Lawrence River, M/kg
PbOH +
Pb 2+
PbCI +
HPbOz-
PbClz ~
Pb3(OH)42
3.37x10 -9
1.43x10 l l
3.29x10 -~3
5.09x10 -14
2.97xi0 -16
2.04x10 -22
Table 1 shows that the main form of lead was lead monoxide dipositive ion (PbOH+).
The concentration of the free lead ion was two orders of magnitude lower, while
complexes with chloride ion were also present.
The calculations, in which the solid phases were allowed to precipitate, showed that
the water was oversaturated with respect to dolomite and iron(l]/) hydroxide. Dolomite
can form a protective layer around cables, which would delay the process of corrosion.
Equilibrium with the above mentioned solid phases were calculated. These calculations
showed:
The ionic strength of the solution was 0.0030.
- The calculated values of Eh and pH were: Eh = 728 mV, pH = 8.3.
The lead speciation is shown in Table 2:
-
Table 2 - Lead speciation in the St. Lawrence River, M/kg, where solid phases were
allowed to precipitate.
PbOH
Pb 2
PbCI
HPbO2
PbCI2 ~
Pb3(OH)42
3.34x10 -9
2.5x10 l l
5.81x10 13
1.6x10 -14
5.25x10 q6
1.14x10 2z
The pH level was lower, than in the previous case (without solids) and this situation
influenced the lead speciation. The concentration of the free lead ion was a little higher,
and there were the higher concentrations of the lead chloride ion[+] and lead chloride
(aq).
To describe the situation where the cable is placed in the water column (on the
291
surface of the sediments), the following parameters were considered:

3.38x 109 Pb + 55.5H20 + CO2 + Oz + 5.686x10 -4 C1 + 7.75x10 4 Ca + 2.92x10 4 Mg +
4.65x10 4 Na + 2.75x10 4 S + 1.23x10 5 Fe + Pb,
The calculations showed:
The ionic strength of the solution became 0.0032.
- The calculated values of Eh and pH were: Eh = 729 mV, pH = 8.3.
The lead speciation changed as shown in Table 3.
-
Table 3 - Lead speciation in the water column (St. Lawrence River), M/kg, in the
presence of cables
i
PbOH
Pb 2+
PbCI +
HPbO2
PbC120
Pb3(OH)42
2.60x10 6
2.01x10 8
4.6x10 q~
1.19x10 t l
4.16x10 -13
5.20x10 -H
The calculations showed that cerussite (PbCO3) precipitated from the solution. Since
cerussite is formed around the cable from dissolution of lead, it may control the solubility
of lead in the water column. The total concentration of soluble lead in the water column
(0.54 rag/l) became 3 orders of magnitude higher, than that in the St. Lawrence River
(0.0007 mg/l). The total concentration of soluble lead in the water column became
equivalent to the 90 days LCs0 = 0.55 mg/L for fish (El-Egamy 1994).
The calculations showed that the relative amount of lead tetrahydroxide dipositive ion
also increased.
To find the speciation of lead under anaerobic conditions (cables buried in
sediments), the equilibrium was calculated for the same system without oxygen. Due to
lack of the data for pore water of the sediments, the concentrations of the major elements
were considered the same as in the St. Lawrence River. Subsequently, calculations were
performed using the following parameters:
3.38x10 9 Pb + 55.5H20 + CO2 + 5.686x10 -4 C1 + 7.75x10 4 Ca + 2.92x10 4 Mg +
4.65x10 4 Na + 2.75x10 4 S + 1.23x10 5 Fe + Pb
Two cases were considered: 1) Pco2 = 0.0003 atm, 2) Pco2 = 0.003 atm (ten times
higher as an increase of Pco2 in pore water and around the cable can be expected; one of
the reasons could be an activity of biota).
The calculations showed that in both cases cerussite and lead sulfide precipitated
from the solution.
Calculations performed for the first case (lower Pco2) showed:
- The ionic strength: 0.0023.
- The calculated values of Eh and pH such as: Eh = - 343 mV, pH = 8.4.
292
Lead speciation for the above-mentioned conditions is shown in Table 4. Since lead
carbonate was still formed in these conditions, it might control the solubility of the lead.
The total concentration of lead in the solution was 1.98x 10-6 M/kg (less than in aerobic
conditions). The distribution of chemical forms of lead was almost the same.
Table 4 - Lead speciation in the water column (St. Lawrence River), M/kg, in the
presence of cables, in anaerobic conditions at different values of Pco2, atm
Pc02
PbOH
pb2+
PbC1+
HpbOz"
PbCIz~
Pb3(OH)42"
1.90xl0 6
0.0003
0.003
1.05x10 6
1.15x10 8
2.64x10 -L~ 1.58xi0 -t3
2.39x10 13
2.98x10 14
3.29x10 8
7.55x10 l ~
6.82x10 t3
8.51x10 -j6
2.96x10 13
In the second case, where higher Pco2 was considered, the following values were
obtained:
- The ionic strength was 0.0036.
- The values of Eh and pH were: Eh = - 285 mV, pH = 7.69, both values were lower
than in the first case.
The lead speciation is shown in Table 4. Because of pH increase (by 8%) comparing
to the first case, the concentration of free lead was approximately 300% higher.
Concentrations of the complexes with chloride ion were also higher. The total
concentration of lead in the solution (1.08x 10-6 M/kg) was lower than in the first case.
The increase of Pco2 around the cable may decrease the total concentration of the soluble
lead in the water column, but in the same time increase the concentration of free ions.
Fate of Organo-metallic Corrosion Products

Pintado and Montero (1992) studied degradation of sheath covers of communication
cables and demonstrated that the asphalt cover can release some carboxylic acids. In
addition, it is known that organic acids are usually present in the bottom sediments of
shallow waters. Due to these facts, it was worthwhile to evaluate the impact of organic
acids on the fate of corrosion products.
To investigate the influence of organic acids on lead speciation three organic acids:
acetic, formic and oxalic (containing two atoms of carbon) were introduced to the
mentioned above system in the presence of cables. Various concentrations of acids were
considered in aerobic and anaerobic conditions.
The calculations showed that in aerobic conditions when the partial pressures of CO2
around the cable were assumed to be 0.0003 and 0.003 atm, acetic, formic and oxalic
acids did not influence very much the speciation of lead, redox potential and pH in the
water column at the concentrations up to 10-I M/kg ( 3 . 3 x i 0 2 each).
0.003
0.003
-321
-321
-330
Eh
7.7
7.7
8.4
pH
Pb 2+
PbCI +
1 . 0 5 x 1 0 -6 1 . 0 5 x 1 0 6 7 . 5 5 x l 0 -l~
1 . 0 5 x 1 0 -~' 3 . 2 9 x 1 0 8 7 . 5 5 x 1 0 t ~
1 . 9 0 x 1 0 -6 1 . 1 5 x 1 0 8 2 . 6 4 x 1 0 -I~
PbOH +
8.1x10 lz
6 . 7 x 1 0 -12
1 . 2 x 1 0 -j'-
Pb(C2H302) +
6 . 8 2 x l 0 -13
6 . 8 2 x 1 0 13
2 . 3 9 x 1 0 -13
PbCI2 ~
2 . 9 x 1 0 13
2.9x1043
1 . 6 x 1 0 -jl
HPbOz
Speciation of lead in the presence of organic acids in anaerobic conditions, M/kg
1,2 - c o n c e n t r a t i o n s o f o r g a n i c a c i d s are 3 . 3 x 1 0 7 M / k g e a c h ( 1 0 .6 M / k g is] s u m ) .

3 - c o n c e n t r a t i o n s o f o r g a n i c a c i d s are 3 . 3 x 10 .2 M / k g e a c h ( 1 0 -I M / k g ill s u m ) .
0.0003
atln
Pco2,
Table 5 -
1.9x10 14
1.7x10 14
5 . 8 x i 0 -15
Pb(CHO2)
r,o
r
r
oo
m
m
f-.
wI
I""
.r"
O
z
m
<
>
m
),
m
63
.--I
294
Repeated calculations for anaerobic conditions showed that in the presence of the
above-mentioned acids in the water column, the values of Eh decreased (Table 5). Even
at acid concentrations of 3.3 x l 0 7 M/kg, lead formed complexes with acetic and formic
acids. The concentration of complexes with oxalic acid was below 10-20M/kg. These
values were comparable with the concentrations of other complexes. It was concluded
that presence of organic acids could increase the total concentration of soluble lead in
water column.
At the same time, according to Poubaix diagram for lead (Fig. 1), in the absence of
sulfur in the system at low values of Eh, the complex with acetic acid becomes
predominant. Subsequently, in these conditions, acetic acid will increase the possibility of
corrosion and a protective layer consisting of solid precipitates might not be formed.
Another study considered volatile compounds present in the environment that can
represent sediment in shallow waters. In this case, the calculations showed that part of the
organic acids was transformed into methane. One part of methane was dissolved in the
water column and the other was in the gaseous phase. It was concluded that in shallow
waters under anaerobic conditions, after some time, a part of the organic acids could be
transformed into methane, which eventually will lead to decrease in the total soluble lead
concentration.
Conclusion
The corrosion of the cable is a very complex process from the thermodynamic point
of view and is influenced by many factors. These factors can affect corrosion directly, for
example, increasing the concentration of soluble lead by:
9 Forming soluble complexes; as it was shown above, chloride and organic acids
can form soluble complexes with lead and increase the concentration of soluble
lead;
9 Forming solids precipitates; carbonates form solid precipitates such as lead
carbonates and can decrease the lead concentration in solution.
Other factors can affect corrosion indirectly, for example, by:
9 Changing pH of the solution; the addition of Ca (and other strong cations)
increases pH.
9 In hard waters with high concentrations of calcium and magnesium, their
carbonates can form a protective layer around cables.
The above-mentioned factors must be considered together since their impacts are
interdependent. To study corrosion of submarine cables buried in sediments, the
composition of pore water of these sediments has to be taken into account. In addition,
information about the solid phase, which controls the solubility of lead have to be
supplied.
The chemical corrosion of the cable containing lead can possibly increase in the total
concentration of soluble lead in the water column up to 3 orders of magnitude compared
to that existing in the St. Lawrence River (0.54 rag/L). This value is equivalent to 90 days
LCso for non-salmonid fish (0.55 mg/L) (Firouseh 1999). Free lead ion Pb 2+ is
predominant below pH = 6, so any change in pH towards acidity will lead to increase in
the concentration of free lead ion, which is more toxic to biota. Organic acids such as
acetic and formic can increase the total concentration of soluble lead in a water column.
295
The methodology described in this paper can be successfully applied to the

assessment of the fate of other metals released from sediments in to the water column.
References
AI-Kharafi, F.M., Badawy, W.A., 1998, Effect of oxygen and oxidizers on the corrosion
and passivation of lead in chloride-containing acidic solutions. Corrosion prevention and
control, vol. 43(3), pp. 75-85.
EI-Egamy, S.S., 1994, Corrosion and passivation behavior of lead in aqueous carbonate
media. Bulletin of Electrochemistry, vol. 11(9), pp. 418-423.
Firouseh, A.S., 1999, "An investigation about the effects of cable corrosion on water
creatures", Internal Report, Concordia University, Montreal, Canada, pp. 21.
Nordstrom, D.K, Munoz, J.L., 1986, Geochemical thermodynamics. Blackwell Scientific
publications, pp. 477.
Pelton, A.D., Bale, C.W., Tompson, W.T., The FACT system, l~cole Polytechnique,
Montreal, Canada, http://www.crct.polymtl.ca, March 1999.
Pintado, J.L., Montero, F., 1992, "Underground biodeterioration of medium tension
electric cables". J. International Biodeterioration and Biodegradation, vol.29, pp. 357365.
St. Lawrence Centre, 1996a, "State of the Environment", Report on the St. Lawrence
River. Volume 1: The St. Lawrence Ecosystem. Environment Canada-Quebec Region,
Environmental Conservation, and Editions MultiMondes. " St. Lawrence UPDATE"
series.
St. Lawrence Centre, 1996b, "State of the Environment", Report on the St. Lawrence
River. Volume 2: The State of the St. Lawrence. Environment Canada-Quebec Region,
Environmental Conservation, and Editions MultiMondes. "St. Lawrence UPDATE"
series.
Stumm, W., Morgan, J.J., 1996. Aquatic chemistry. Chemical equilibria and rates in
natural waters, Third ed., Willey-Interscience publications, New-York, pp. 1022.
Wranglen, G., 1985, An introduction to corrosion and protection of metals. Chapman and
Hall, New York.
Sebnem E19i,1 and Paul A.Work2
Numerical Modeling of Hydrodynamic Circulation and Cohesive Sediment

Transport in HartweU Lake, South Carolina/Georgia, USA
Reference: El~i, $. and Work, P. A., "Numerical Modeling of Hydrodynamic

Circulation and Cohesive Sediment Transport in Hartwell Lake, South
Carolina/Georgia, USA," Contaminated Sediments: Characterization, Evaluation,
Abstract: This paper describes numerical modeling of hydrodynamic circulation and
cohesive sediment transport in Hartwell Lake, South Carolina/Georgia, a U.S. Army
Corps of Engineers (USACE) hydropower and flood control reservoir. A U.S.
Environmental Protection Agency (EPA) "Superfund" site is located on a tributary to the
lake because of high concentrations of polychlorinated biphenyls (PCBs) in the lake
sediments. The primary objective of this study involves prediction of depositional zones
for sediments transported within the reservoir.
The EFDC (Environmental Fluid Dynamics Code) model developed by Hamrick
(1996) is used to describe lake hydrodynamics and sediment fate. Historical records of
wind and flow were used to determine frequencies of occurrence and representative
conditions for prediction of long-term deposition zones for sediment transported by the
flow.
Sensitivity of hydrodynamic processes to model parameters were investigated and
wind was found to be the major force driving the circulation. Model results for cases with
realistic long-term forcing indicated likely zones of sediment deposition, useful for
mitigation of pollution problems as well as predictions of reservoir lifetime and
development of maintenance schemes.
Keywords: Hartwell Lake, EFDC, deposition, sedimentation
1Graduate Research Assistant, School of Civil and Environmental Engineering, Georgia

Tech Regional Engineering Program, Savannah, GA 31405.
2 Associate Professor, School of Civil and Environmental Engineering, Georgia Tech
Regional Engineering Program, Savannah, GA 31405.
296
Copyright9 2003by ASTMInternational
www.astm.org
EL(~I AND WORK ON CIRCULATION AND SEDIMENT TRANSPORT
297
Introduction
HartweU Lake, a U.S. Army Corps of Engineers reservoir, is located on the Savannah
River, between Anderson, South Carolina, and Hartwell, Georgia, USA (Figure 1). The
reservoir was built between 1955 and 1963, with joint goals of flood control, power
production, water supply, and recreation. It covers 227 km 2, has a shoreline length of
1548 km with an average depth of 20 meters. The terrain consists primarily of gentle
rolling hills and pine forest near the southern terminus of the Appalachian mountain
chain. Sediments contain high fractions of silt and clay.
High concentrations of polychlorinated biphenyls (PCBs) are found in the lake and in
Twelve-Mile Creek, a tributary, resulting from the operation of a capacitor manufacturing
facility in the Twelve Mile Creek Watershed between 1955-1976. PCBs are known to
cause cancer in animals and can cause non-cancer health problems, such as reduced
ability to fight infections, and low birth weights in humans (Clearwater 1997). PCBs are
hydrophobic, and typically bond to sediments. They also exhibit an affinity for finegrained sediments, such as silts and clays (EPA 1991). The U.S. EPA performed and
sponsored research on PCBs in Hartwell Lake (Elzerman et al. 1994), and EPA's selected
remedy was to rely on burial by natural sedimentation processes. Over time, incoming
"clean" sediments should bury the contaminated sediments, providing a clean sediment
cap on top and gradually reducing the health risks. Numerical modeling of sediment
transport in Twelve-Mile Creek has also been performed to support this conclusion,
focusing on non-cohesive sediments (EPA 1991). However, the degree of contamination
of the main pool of the lake, due to transport of (primarily) cohesive sediments from
Twelve-Mile Creek, has not been studied previously and is addressed in this paper.
Twelve
, oma n
'r
GEORGIA
L ~ . )
Savannah
River
Figure 1 - Map of Study Site. Model Domain Shows the Main Pool of the Lake Modeled
to Describe the Circulation and Sediment Deposition Patterns.
298
The primary objectives of this study are thus to describe hydrodynamic circulation and
sedimentation in the main pool of Hartwell Lake. The cohesive sediments are of primary
interest because of their potential to travel greater distances and to transport PCBs.
The hydrodynamic model of Hartwell Lake is developed using the Environmental
Fluid Dynamics code (EFDC), developed by Hamrick (1996). EFDC was developed to
simulate flows and transport processes in surface water systems, including rivers, lakes,
estuaries, wetlands and coastal areas. The structure of the EFDC model includes four
major modules: (1) a hydrodynamic model, (2) a water quality model, (3) a sediment
transport model, and (4) a toxics model. EFDC is capable of simulating both cohesive
and noncohesive sediment transport, near-field and far-field discharge dilution from
multiple sources, eutrophication processes, and the transport and fate of toxic
contaminants in the water and sediment phases. For a detailed description of EFDC, the
reader is referred to Hamrick and Wu (1997).
The model has been applied to several water systems. Applications of the model
include a study of tidal intrusion and its impact on larval dispersion in the James River
estuary (Shen et al. 1999a), modeling an estuarine front and its associated eddy (Shen and
Kuo 1999b), investigation of bottom shear stress in estuaries (Kuo et al. 1996),
hydrodynamic and sediment transport modeling in the middle Atlantic Bight (Kim et al.
1997). Simulation of Lake Okeechobee hydrodynamic and thermal processes (Jin et al.
2000, 2001, 2002), hydrodynamic and water quality modeling in Peconic Bay (Tetra
Tech 1999), and simulation of wetting and drying processes of hydrodynamic and
sediment transport in Morro Bay (Ji et al. 2000, 2001a) are among other applications.
Hydrodynamic Model Setup

Hydrodynamic circulation patterns in the main pool of Hartwell Lake are mainly
controlled by wind, inflows and outflows. Daily inflow and outflow data for Hartwell
Lake were obtained from the U.S. Army Corps of Engineers. USACE derives inflow
values from volume balance of the reservoir and calculates outflows by converting the
power measured at the power plant into discharge. Thus the published inflow values
include the changes in the reservoir including rainfall, evaporation, and infiltration. Wind
data were obtained from the National Oceanic and Atmospheric Administration for
Greenville-Spartanburg Airport, SC, located 65 kilometers northeast of the lake.
Available data include average and peak hourly wind speed and corresponding direction,
and cover the period 1962 to the present.
EFDC has been applied to Hartwell Lake to simulate lake response to wind forcing
and inflows/outflows and to simulate long-term sediment deposition patterns. For this
purpose a careful examination of available data is necessary since it is not practical to
simulate and obtain results for every single wind forcing and flow condition. To
determine which forcing combinations should be simulated to represent long-term
deposition patterns, a statistical analysis of daily mean values of inflow, outflow and
wind data for ten years (1990-1999) was performed. Figure 2 is a histogram based on 10
years of flow data, and shows that 20% of the time no water was released at the dam, and
about 40% of the time outflows were in the range of 75-150 m3/s. Whereas about 30% of
ELCI AND WORK ON CIRCULATIONAND SEDIMENTTRANSPORT
299
the time inflows in the range of 25-75 m3/s were observed. Mean outflow and inflow
values were the same: 120 m3/s.
Histogram of daily mean wind speeds is shown in Figure 3, indicating that 46% of the
time observed wind speeds were in the range of 2-3.5 m/s. Two dominant wind directions
were observed in the study area, northeast and southwest. 36% of the time winds were
blowing to northeast with a mean speed of 1.7 m/s and 35% of the time winds were
blowing to southwest with a mean speed o f 2.6 m/s.
The computational mesh was constructed using Digital Elevation Model data and is
shown in Figure 4. The computational grid had 51 x94 horizontal cells and five vertically
stretched layers, each with a fractional thickness of 20% of the water depth. Horizontal
discretization was 150 m on a side. Bathymetry of the lake is shown in Figure 5,
22
20
18
*~ 16
I.
~
12
10
'i
4~
,M
9'
0
flow (m3/s)
Figure 2 - Histogram of Daily Mean Flows for Ten Years (1990-1999).

20
18
16
14
lo
8
b
wind speed (m/s)
Figure 3 - Histogram of Daily Mean Wind Speed Measured at 10 m Height for
Greenville-Spartanburg Airport, SC, for Ten Years (1990-1999).
300
The river inflows and controlled outflow at the dam were simulated as sources and
sinks in the model system. At upstream boundaries, the river inflow rates were uniformly
distributed according to the depth of each computational cell in horizontal and according
to thickness of vertical layers in vertical. The controlled outflow at the dam was specified
at one cell at the downstream boundary.
Sensitivity Analysis
Several variables including lake water level, grid size, number and thickness of
vertical layers, wind direction and inflow and outflow magnitudes were systematically
altered as part of a sensitivity analysis. The analysis of daily water level data from 19642001 (Figure 6) showed that minimum and maximum water levels in Hartwell Lake have
varied by up to six meters. When the same analysis was conducted for 1990-1999,
minimum and maximum water levels have varied by three meters. In order to investigate
the sensitivity of the model results to seasonal changes in lake level, the model was run
using two water levels where depth was changed by five meters. During the simulation,
zero initial velocity was considered in a closed basin.
Figure 4 - Finite Difference Computational Mesh for Hartwell Lake. Computational
Cells are 150 m x 150 m. A, B, and C Represent Arbitrarily Selected Locations for
Comparison of Model Results.
301
2000
4000
6000
x (m)
Figure 5 - Bathymetry of Hartwell Lake Used in Simulations. Depth Ranges between 0 to

50 Meters. Contour Interval is 10 Meters.
In shallow parts of the lake, significant differences in surface layer velocities were
observed because of drying of cells that were wet in the initial run. The root-mean-square
(RMS) value was used to evaluate the differences in simulation results and calculated by
equation (1).
(1)
where
HI, He = the value of same variable obtained from first and second simulations
N
= the number of different data locations being compared
302
203
202
201
200
.~ 199
198
'~ 197
196
195
..
.~176
. . . . . . . . . . . . .
. ~
.~
~176176
~176
.~
~
. . . . . . . . . . . . . .
10
11
12
month
l
maximum
"
minimum . . . . . . averagej
Figure 6 - Variation in Maximum, Minimum and Average Monthly Water Levels

throughout the Year on Hartwell Lake for the Years 1964-2001.
The RMS of change in flow speed magnitude was 0.02 cm/s (about 0.1% of maximum
velocity observed) when the cells initially shallower than 5 m were excluded (N = 2405).
Whereas the RMS change increased to 2.5 cm/s when all cells were included in the
calculation (N = 2468). In the analyses velocity components were compared separately.
Based on the results above it has been concluded that for deeper parts of the lake (>5
meters), which in fact are of interest for purposes of this study, the simulation results are
not sensitive to seasonal changes in water levels.
Sensitivity of model results to grid cell size was investigated using two different grid
configurations as given in Table 1. In the main pool calculated velocities were almost the
same with an RMS change of 0.01 crn/s, whereas in the shallower parts of the lake (<5
meters) the results changed significantly with a RMS change o f 1.52 crn/s. It is concluded
that Hartwell Lake can be simulated using 51 x 94 grid configuration.
Table 1 - Computational Grids Used in Calculations
Cases
Grid
configuration
Cell width
(m)
No. of
vertical cells
51 94
103 188
150
75
Selection of vertical layers was found to be important for Hartwell Lake. When three
vertical layers were simulated and velocities were compared to simulation results with ten
vertical layers, RMS change of velocities increased to 6 era/s, whereas RMS change
obtained from comparison of five layer and ten layer simulations was 0.8 cm/s. Therefore
in all runs five or more vertical layers were selected. Another purpose of the sensitivity
analyses was to evaluate the sensitivity o f bottom layer (10% o f depth in each case)
303
velocities, which is important for sediment transport studies, to the total number of
vertical layers. The results showed that when the same relative thickness was selected for
the bottom layer, bottom layer velocities were not sensitive to the number of layers.
Lakewide circulation patterns are very sensitive to wind direction and magnitude,
inflow and outflow magnitudes. Wind is typically the major force driving the circulation.
Surface currents followed the dominant wind direction in each case. When wind speed
was reduced from 6.8 to 2.8 m/s and the flows were kept constant, we observed an
average difference of 3.5 cm/s in surface velocities. RMS change was calculated as 4
cm/s.
Sensitivity of model results to inflows was investigated by comparing three different
inflow magnitudes. In all simulations wind was kept at zero. The second and the third
tests of sensitivity to inflow included two times and five times the initial inflow in
magnitude. Increase of inflow by two times resulted in an increase of two times, while an
increase of inflow by five times resulted in an increase of six times the average of surface
layer velocity in the model domain.
Long-term deposition patterns for the sediments were investigated for representative
wind and flow conditions. Joint histograms of the inflow, outflow and wind speed data
were investigated. The most frequently observed conditions were determined and used
for numerical simulations of hydrodynamic processes in Hartwell Lake. The selected
cases discussed in this paper are listed in Table 2. The values in case 1 are the most
frequently observed values when individual histograms are considered, values in cases 2,
3 and 4 are selected based on results of joint histograms.
Table 2 - Values of Parameters Used in Model Simulations.
Cases
1) Average inflow, no outflow, dominant
northeast wind with average speed
2) High inflow, high outflow, dominant
northeast wind with high speed
3) High inflow, high outflow, dominant
southwest wind with high speed
4) Moderate inflow, moderate outflow,
dominant northeast wind with moderate speed
Inflow
(m3/s)
Outflow
(m3/s)
Wind
Speed
Wind
Direction
68
2.8
55
465
447
6.8
55
465
447
6.8
235
96
30
3.6
55
(m/s)
(o)
Discussion of Results
The top and the bottom layer velocity distributions within Hartwell Lake after five
days of simulation with forcing conditions specified as case 3 in Table 2 are given in
Figures 7 (a) and 7 (b). At least five days of simulation was required because it takes four
days for sediments to settle 50 meters with the specified settling velocity. A 90 day
simulation with measured time series of input was conducted and it was observed that
304
CONTAMINATED
SEDIMENTS
velocity distribution and sedimentation patterns were similar to the five day simulation
with average values of wind and flows as input data. Figure 7(a) shows that surface flow
velocities closely match the wind direction except where the vortices were present.
Maximum surface velocities (approximately 5 crn/s) were about 1% of wind speed. The
bottom layer velocities shown at Figure 7(b) are much lower in magnitude than the
surface velocities and formed a return flow.
Sediment deposition within Hartwell Lake with the conditions corresponding to case 3
in Table 2 is shown in Figure 8, indicating that sediments have mostly deposited in the
deeper parts of the lake closer to the dam. When extrapolated, 3.6 m of deposition in 40
years (0.09 m per year) is estimated. Two bathymetric surveys of the lake were conducted
by USACE; one in 1959 before completion of the reservoir, and one in 1973. Although
the 1959 survey includes several cross sections within the main pool, surveys from both
years were mostly available for the upstream region of the model domain. Results of
these surveys suggested deposition rates of 0.29 m per year in shallow regions, and 0.05
m per year in deeper regions of both the Seneca and the Tugaloo Rivers. A surveyed
1Annn
' "
120001-/
(m/s) 7 ; ~-~i/~j, '

"~"/~/,.r
~ --._~ .
0 ~
"~"~-S_-'-
-~\\ ,,,,
_ 1 "'
'<'<'<'<'r
~ ~-
Win~F ~-.~ ,~.:_'2,

_ (mls) _,__.~\. :~-~.--.
5 crn]s
.,;
,,t~.
I'
-6-3
/
-.
-~
. " ,' , ' - ' . . - / "
4000 -
"
9, ' , < "
~./J,,~,
--
../~
.7.. ~ - ~
A - ~-=a/i
,'-~,
ZUUU
:..~_~..~.~'
,"5
~ ~ ..~ -~ ~
~--.-,
.,
..
;,l
~l
//~_
. , _: .., / /
~'~=_
,~'~
~ 9
l/
. ~-..,
,
,
00-
2000
l/
4000
x (m)
(a)
-"f-
, I
6000
8000
2000
I- " -I"-",'l
4000
8000
8000
x (m)
(b)
Figure 7 - Simulated (a) Top Layer and (b) Bottom Layer Cell Velocity D&tribution for
Case 3 in Table 2 (Inflow = 465 m3/s, Outflow = 447 m3/s, Wind Speed = 6.8 m/s).
305
section within the model domain boundaries shown by the dotted line in Figure 8
indicated a deposition rate of 0.1 m per year that agrees well with the estimated
deposition rate of 0.09 m per year. An ongoing field study conducted by the authors will
reveal the deposition rates after 40 years of operation of reservoir.
When results for cases with lower wind speeds were generated, it was observed that
sediments followed the thalweg and were deposited in the deeper parts of the lake
regardless of the magnitudes of inflows and outflows. Local effects of inflows and
outflows were observed. When the inflow magnitude was increased for instance, higher
sediment deposition was observed in the uppermost part of lake, the throat, although the
deposition patterns remained the same elsewhere, indicating that higher inflow affected
the bottom layer velocity distribution, thus the deposition pattern there.
Simulated velocity profiles at 3 different locations represented by points A, B, and C
in Figure 4 are shown in Figures 9 and l0 for east-west and north-south components
respectively. Corresponding water depths at these locations are 34 m, 31 m, and 24 m. In
all cases the velocity profiles showed variability in directions at surface and bottom
layers indicating reverse flow. This was expected, since, the shear force developed by the
wind at the surface would be balanced by the hydrostatic force of reverse flow. These
results agree well with bottom layer velocity vector distributions.
Summary and Conclusions

In this study, application of a numerical model for simulation of hydrodynamic
circulation and sediment transport in a lake with respect to wind forcing and
inflows/outflows is described. The motivation is related to a PCB contamination problem
in the upper reaches of the lake. Remediation of the pollution problem requires
knowledge of the level of sedimentation in the lake, its rate of change, and the fate of
contaminated sediments.
This study also describes a sensitivity analysis conducted to assess the response of the
lake hydrodynamics to different variables. The model assessment was quantified in terms
of RMS values. Several variables, including lake levels, grid size, number and thickness
of vertical layers, wind direction and inflow/outflow magnitudes, were investigated
through a sensitivity analysis. The results of the sensitivity analysis indicated that the
calculated velocities in the main pool were not sensitive to lake levels and grid size.
Selection of the number and thickness of the vertical layers, on the other hand, were
found to be critical, and selection of at least 5 vertical layers was suggested. Although not
investigated in this paper, thermal stratification can be used as a reference for selection of
vertical layer boundaries. Lakewide circulation patterns were very sensitive to wind
direction and magnitude, inflow and outflow magnitudes. Wind was found to be the
major force driving the circulation.
The zones where contaminated sediments are likely to deposit and the deposition rates
were presented for one set of forcing conditions. We observed that sediments have mostly
deposited in the deeper parts of the lake closer to the dam. It can be stated that, wind
direction and magnitude primarily control the bottom layer velocity distribution, which
306
220
210
a)
190
~180
~7o
"~160
150
140
200
400
[
600
station (m)
800
1000
1200
1959 . . . . . 19731
b)
2000
4000
6000
x (m)
Figure 8 - Sediment Deposition in Millimeters within Hartwell Lake after 5 Days of

Simulation for Case 3 in Table 2 (Inflow = 465 m3/s, Outflow = 447 m3/s, Wind Speed =
6.8 m/s). Dashed Line in Part b) Shows the Cross-Section Surveyed by USACE in 1959
and 1973.
west
1.0
1.0
0.8
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0.0
0
1 2
east
I
0.8
0.6
0.4
0.0
-2 -1
307
0.2
0.0 4 - - - -
-4
u (cm/s)
-2
-2
u (era/s)
u (cm/s)
Figure 9 - Velocity Profiles for East- West Component of Velocity Computed by the
Model at Three Locations (.4, B, C Shown in Figure 4)for Case 3 in Table 2 (Inflow =
465 m3/s, Outflow = 447 m3/s, Wind Speed = 6.8 m/s).
B
south
C
north
1.0
1.0
1.0
0.8
0.8
0.8
0.6
0.4
------
0.6
0.6
0.4
0.4
0.2
0.2
0.2
0.0
0.0
0.0
-4
-2
v (cnVs)
-4
-2
v (cm/s)
-4
-2
v (crrds)
Figure 10 - Velocity Profiles for North-South Component of Velocity Computed by the

Model at Three Locations (.4, B, C Shown in Figure 4)for Case 3 in Table 2 (Inflow =
465 m3/s, Outflow = 447 m3/s, Wind Speed = 6.8 m/s).
308
directly controls the depositional patterns. For high wind speeds, deposition zones were
shifted away from the thalweg according to dominant wind direction. For low wind
speeds, sediments simply followed the thalweg and were deposited in deeper parts of the
lake. Inflows and outflows had local effects at bottom velocity vectors. Increased inflow
rate coming from the tributaries introduced vortices in the upstream regions of the model
domain, resulting in higher deposition rates.
Maximum deposition rate was estimated as 0.09 m per year for a selected forcing
combination representing commonly observed winds, inflows and outflows. This
deposition rate was compared to available surveys conducted in 1959 and 1973 by
USACE. Two surveys indicated a maximum deposition rate of 0.1 m per year that agreed
well with the estimated deposition rate of 0.09 m per year by the numerical model. An
ongoing field study conducted by the authors will reveal the deposition rates after 40
years of operation of the reservoir.
Acknowledgements
The work summarized in the present paper was supported by the South Carolina
Water Resources Center (SCWRC), the Georgia Water Research Institute (GWRI), and
the U.S. Geological Survey (USGS).
References
Clearwater, 1997, PCB Contamination Of The Hudson: addressed at

http://www.clearwater.org/news/hazard.html.
Elzerman, A. W., Farley, K. J., Dunnivant, F. M., and Cooper, C., 1994, "Predicting the
Future Fate of PCBs in Lake Hartwell," Technical Completion Report, G1588-03,
submitted to U.S. Department of the Interior, U.S. Geological Survey Reston, VA.
EPA, 1991, Sangamo- Weston Inc.~Twelve-Mile Creek~Lake Hartwell PCB
Contamination: addressed at
http://www.epa.gov/oerrpage/superfund/sites/npl/nar486.htm
Hamrick, J. M., 1996, "User's Manual for the Environmental Fluid Dynamics Computer
Code," Special Report No. 331, The College of William and Mary, Gloucester
Point, VA.
Hamrick, J. M., and Wu, T. S., 1997, "Computational Design and Optimization of the
EFDC/HEM3D Surface Water Hydrodynamic and Eutrophication Models," Next
Generation Environmental Model and Computational Methods, G. Delich and M.
F. Wheeler, eds., Society of Industrial and Applied Mathematics, Philadelphia,
pp. 143-161.
309
Ji, Z. G., Morton, M. R., and Hamrick, J. M., 2000, "Modeling Hydrodynamic and
Sediment Processes in Morro Bay," Estuarine and Coastal Modeling:
Proceedings, 6th International Conference, M. L. Spaulding and H. L. Butler,
eds., ASCE, New York, pp. 1035-1054.
Ji, Z. G., Morton, M. R., and Hamrick, J. M., 2001a, "Wetting and Drying Simulation of
Estuarine Processes," Estuarine, Coastal Shelf Science, 53, pp. 683-700.
Jin, K. R., Hamrick, J. H., and Tisdale, T., 2000, "Application of a Three-Dimensional
Hydrodynamic Model for Lake Okeechobee," Journal of Hydraulic Engineering,
126 (10), pp. 758-771.
Jin, K. R., and Ji, Z. G., 2001, "Calibration and Verification of a Spectral Wind-Wave
Model for Lake Okeechobee," Journal of Ocean Engineering, 28, pp. 571-584.
Jin, K. R., Ji, Z. G., and Hamrick, J. M., 2002, "Modeling Winter Circulation in Lake
Okeechobee, Florida," Journal of Waterway, Port, Coastal, and Ocean
Engineering, ASCE, 128 (3), pp. 114-125.
Kim, S., Wright, D. L., Maa, J., and Shen, J., 1997, "Morphodynamic Responses to
Extratropical Meteorological Forcing on the Inner Shelf of the Middle Atlantic
Bight: Wind Wave, Currents and Suspended Sediment Transport," Estuarine and
Coastal Modeling: Proceedings, 5th International Conference, M. L. Spaulding
and H. L. Butler, eds., ASCE, New York, pp. 456--466.
Kuo, A.Y., Shen, J., and Hamrick, J. M., 1996, "The Effect of Acceleration on Bottom
Shear Stress in Tidal Estuaries," Journal of Waterway, Port, Coastal, and Ocean
Engineering, ASCE, 122 (2) pp. 75-83.
Shen, J., Boon, J., and Kuo, A. Y., 1999a, "A Numerical Study of a Tidal Intrusion Front
and Its Impact on Larval Dispersion in the James River Estuary, Virginia,"
Estuaries, 22 (3), pp. 681-692.
Shen, J., and Kuo, A. Y., 1999b, "Numerical Investigation of an Estuarine Front and Its
Associated Eddy," Journal of Waterway, Port, Coastal, and Ocean Engineering,
ASCE, 125 (3), pp. 127-135.
Tetra Tech, 1999, "Theoretical and Computational Aspects of Sediment Transport in the
EFDC Model," Technical Report Preparedfor U.S. Environmental Protection
Agency, Tetra Tech, Inc., Fairfax, Va.
Rosa Galvez-Cloutier, l Myriam Muris, 2 Jacques Locat, 3 and Cyril Bourg 2
Retention o f H e a v y M e t a l s in t h e P o s t ' 9 6 Flood Sediment L a y e r

Deposited in the Saguenay River, Quebec, Canada
Reference: Galvez-Cloutier, R., Muris, M., Locat, J., and Bourg, C. "Retention of
Heavy Metals in the Post '96 Flood Sediment Layer Deposited in the Saguenay
River, Quebec, Canada," Contaminated Sediments: Characterization, Evaluation,
West Conshohocken, PA 2003.
Abstract: In July 1996 during a 50 year event flood, a new layer of sediments composed
of debris, gravel and fine materials was transported and deposited over an ancient, one
meter contaminated layer of sediments in the Saguenay River Fjord and Ha! Ha! Bay
floors. During the Industrial Revolution years, various metallurgic, plastic, aluminum and
pulp and paper production industries discharged their wastes on these waters, resulting in
the high contamination of both water and sediments. This contamination limited the
exploitation of fish and seafood. Given that the new layer is composed essentially of
cleaner material, the zone is presenting important changes in the direction of a healthier
environment. The Canadian Government and some of the surrounding industries aim to
assess the new potential of the zone and its environmental safety. At present, it
represents over a million-dollar study.
This part of the Saguenay project aimed to recognize and evaluate the capacity of
the new layer to contain and retain the contaminants left at the bottom layer. Particular
interest is given to mercury and to heavy metals such as Pb, Zn, Cd, and Ni and to their
geochemical distribution among natural adsorbing materials such as clays, oxides,
carbonates and organic matter.
The paper presents the recognition and sampling mission on the Alcide Hortb Ship,
the contamination profiles given in two dimensions (length and depth), the geochemical
distribution of heavy metals on the contaminated layer, transition layer and new layer as
well as the evolution of their retention and transfer. Discussion and relations with
common sediment characteristics such as grain size, cation exchange capacity and surface
area are also given. Sequential selective extraction has been used jointly with scanning
electron microscopy (SEM) to study heavy metal species.
Keywords: heavy metals, sediments, speciation, retention
1 Associate Professor, Department of Civil Engineering, Laval University, Quebec,
Canada, G I K 7P4.
2 Graduate students, Ecole Nationale de Travaux Public de l'l~tat, Lyon, France.
3 Professor, Department of Geology, Laval University, Quebec, Canada, G 1K 7P4.
310
www.astm.org
GALVEZ-CLOUTIER ET AL. ON RETENTION OF HEAVY METALS
311
Introduction
On July 1996, the Saguenay region north of Quebec City in Canada was affected by a
50-year diluvial rain that caused the flood of the Saguenay River. During 3 days, 200 mm
of precipitation caused the erosion and transport of six million tonnes of sediments
composed of debris, gravel and fine materials that were deposited over an ancient, about
one meter contaminated layer of bottom sediments in the Sagnenay River Fjord and the
Ha! Ha! Bay. During the Industrial Revolution years, various metallurgic, plastic,
aluminium and pulp and paper production industries discharged their wastes in to these
waters, resulting in the high contamination of water and sediments. Mercury and heavy
metals such as Pb, Zn, Cd, Ni, Hg and PAH (Poly Aromatic Hydrocarbons) and MAH
(Mono Aromatic Hydrocarbons) have been detected in the Saguenay zone during earlier
studies before the flood (Gagnon et al., 1993; Fortin and Pelletier 1995). This
contamination limited the exploitation of fish and seafood. Given that the new layer is
composed essentially of cleaner material, the zone is presenting important changes in the
direction of a healthier environment. The Canadian government and some of the
surrounding industries are aiming to assess the new potential of the zone and its
environmental safety. At present, it represents over a million-dollar study.
Sediments are an important and integral component of aquatic ecosystems
because they provide a substrate and habitat for a wide variety of organisms as well as
species that are important in the food chain (Galvez-Cloutier et al., 1995). Sediments are
also recognized contaminant reservoirs and constitute, under certain conditions, a source
for contaminant release. Upon release into the interstitial water, they may diffuse to the
overlying water, affecting benthic and higher organisms. The aim of this project was 1) to
diagnose the quality of the bottom and upper layer sediments, and 2) to assess the
potential for retention or transfer of contaminants into the interstitial water.
Site Characteristics and Sampling Strategy

The Saguenay Fjord is located 200 km north east of Quebec (see Figure 1) and it
expands linearly from Chicoutimi to Tadoussac. The Saguenay River takes its sources at
the Saint-Jean Lake and flows at 1000 m3/s. The Saguenay fjord is a narrow corridor, 1 to
3 km large, perpendicular to the Saint-Lawrence River Estuary. It is approximately, 70
km long and 280 m deep. At the East end of the fjord, the river splits in two portions
forming the Ha! Ha! Bay and the "northern branch". The sedimentation rate varies from
0.1 cm/year in the bay to 7 cm/year in the northern branch (Smith and Walton, 1980).
For this study, 3 sampling locations were chosen: the SAG5 in the northern branch, the
B04 in the Ha! Ha! Bay and the B08 where the bay and river melt its waters.
Respectively, these locations were representative of high erosion-high advection, low
erosion-low currents and average conditions.
During a recognition and sampling mission on the Alcide Horth Ship during the
summer of 2000, core and box sediment samples were taken at the three locations
described above. Core samples were extracted using a Lehigh core equipment that
allowed the extraction of about 2 m long, 10 crn diameter cores.
312
I
'I
70050'
'-48~162
-"_
" ~
\ /
7.,
ea~ de.s
H,, H,! . /t ~ Y "
k
RN. oes
V'dle/de Ha! Ha!
La Bale
70050 ,
l,
70~
i $ km
I
..... I
70~
. . . . . . .
~alnte-Hose--OU-NOrO
\-- J ~ " ~ ~
"
QUI~BEC . ~ .
~ ]~g""./'~,....-"~Atlantiqu,
'--'~.r-- ~./~.~_~
_" .~
e,
i~ternit~'~'~ ' -~'~
i
70o40 ,
I
70o30 '
I
vers/e~wJveSa~.L~--~A
I
Figure 1 n Sampling Sites at the Saguenay Fjord

Physico-chemical, mineralogical and contaminant characterization were run on core
samples. Boxes of 50 cm x 50 cmx 60 cm were used to obtain smaller core samples that
were placed on Plexiglas columns to run the diffusion tests.
General Geochemical Characterization of Sediments
General characteristics included the measurement of water content according to

ASTM standard D-2216, alkaline metals and mineralogy by X-ray diffraction according
to procedures described in ASA, (1986), carbonate content by acid titration (AFN, 1982),
organic matter by combustion (BNQ 1993), buffeting capacity by acid/base titration
(Yong, 1990), Cation and Anion Exchange Capacities (CEC & AEC) according to
Duquette and Hendershot (1993) and Specific Surface Area (SSA) according to the blue
methyl method (BNQ, 1993).
Based on the characteristics measured in the core samples, the Saguenay and Ha! Ha!
Bay floor may be schematized as presented in Figure 2. Depending on the location (Bay,
Northern branch or Melting sector) the thickness profile of the cores varied. The interface
layer was located at 60 em deep on the northern branch but only 21 cm deep on the Bay
and 25 cm on the Melting sector. A typical profile indicating these limits can be viewed
from the water content profile shown in Figure 3.
For samples taken at depths before the flood, titration results showed that all locations
presented important buffering capacities. Higher values were found in upper layers that
were richer in carbonates. CEC determinations yield 15 to 20 meq/100g for all locations
showing higher values to locations with higher pH. SSA values fall between 12 and
30m2/g that are characteristics of soils with poor clay materials. In this case, kaolinite and
illite were found at trace levels.
All metals (alkaline, heavy metal and transition metals) were measured by AAS
following the procedure described by CSL (1993). As ear) be seen in Figure 4, for
elements such as Mg, Na, Fe, Mn and K, the trace of the flood resulted in a decrease of
313
concentration. Opposite to this, Ca presented a peak that corresponded to the '96 layer
that is associated with higher carbonate concentrations (i.e. for the Ha! Ha! Bay: 0.70 in
the ancient layer and 3.5 in the '96 layer) (Ize, 2000) and to the higher buffeting capacity
discussed previously.
Figure 2 --Typical Sediment Profiles after the '96 Flood

Evaluation of Retention
Various studies (Yong and Phadungchewit 1993; Galvez-Cloutier et al. 1995) have
shown the importance of metal partition and distribution in the determination of metal
availability for transfer and transport. Changes in the physico-chemical conditions of the
immediate surrounding environment have to be considered in the evaluation of the final
fate of contaminants. As shown in Figure 5, the mechanisms of retention of cations are
adsorption at the surface of clay minerals, oxides and hydroxides, organic matter, the
(co)precipitation and with secondary minerals, complexation and coagulation with
organic matter, and the penetration of metals in the crystalline structure of primary
minerals (Galvez-Cloutier, 1995). Heavy metal retention by the solid components of the
sediments results from the heterogeneous equilibrium set among water, sediments
components and contaminants. Reactions such as oxidation-reduction, precipitation and
changes in pH or Eh may modify the geochemical form and bioavailability of heavy
metals.
314
Water Content % m/m

0
20
40
60
80
100
120
0
20
---.~_~--
40
60
.:--.
)-
- ~..:~ ~ - = - -
. :_-f;-~-;-~':
............
Interface BHH
Interface Confl.
Interface BN
80
100
120
140
~ t-f-'.'x
.}::5
-
" " Bras No
160
180
Baie de..
-~.-.,,
- Conflue=
200
Figure 3 - - Water Content Profiles

During this study, heavy metal retention was evaluated using a sequential selective
extraction (SSE) method. SSE relies on the concept that by using appropriate chemical
reagents the different metal fractions can be released once the binding phase is destroyed.
The extracting reagents in most of the cases are taken from routine individual soil analysis.
According to the method used in Yong et al. (1993) a maximum of five metal fractions can
be separated by sequential extractions associated to: (1) exchangeable, (2) to carbonates, (3)
to metal oxides, (4) to organic matter and (5) to residual mineral crystals.
A metal mass balance was obtained by comparing total metal concentrations to SSE
results. Total concentrations were obtained by acid digestion using nitric acid, hydrogen
peroxide and Aqua-Regia reagent (Centre Saint Laurent, 1993) and were dosed using a
3110 Model Perkin-Elmer AAS instrument. Mercury was transformed to hydride before
its detection by cold vapour in AAS.
GALVEZ-CLOUTIER ET AL. ON RETENTIONOF HEAVY METALS
315
Figure 4 - - Typical Metal Profiles, Case of the Northern Branch
As shown in Figure 6, the 3 sites showed high concentrations of mercury exceeding

the SEM (Minimal Effect Level) value for all samples within the first 60 cm with a peak
at the interface layer. It also presented higher values than those reported by Barbeau et al.
1981. Concentrations decreased with depth except for a deeper peak that was found about
1.35 cm depth below the melting zone that exceeded the SET (Toxic) level. This may
have corresponded to a more ancient deluge most probably that one at St-Vianney
(Quebec) that had similar effects to the '96 flood.
Figure 7 shows the concentration profiles for Cr, Cu, Pb and Zn for the 3 sampling
sites (Figs. 7a, 7b and 7c respectively). As it can be seen, concentrations were higher
within the interface layer, which is mostly a mixture of the most superficial portion of the
ancient layer.
316
Figure 5 - - Heavy Metal Partition. Source: Galvez-Cloutier (1995)

Indeed, when the new layer was transported it caused an abrasive front that scavenged
the ancient layer forming an "interface layer" that presented various depths ranging from
almost zero to 50 cm.
Lower concentrations were found in the Northern branch followed by the Bay and
finally the melting zone. In general, the order of decreasing concentrations was
Zn>Pb>Cr>Cu for the 3 sites. All metals concentrations exceeded the SEM values of the
MEO (1990) and MEC (1992) guidelines at the interface layer and decreased with depth.
The Figure 8 presents the distribution ofCu, Zn and Pb found at the interface zone for
the sites at the Ha! Ha[ Bay and at the Northern branch (which were the most
contaminated). Each sample was analyzed for ESS in triplicate following the procedure
described in Yong et al, 1993. From Figure 8 it can be seen that only 4 of the six phases
analyzed were present in significant manner. Also, it can be seen that the distribution
was similar for both sites.
At the northern branch Cu and Zn presented, in decreasing order of importance,
associations to: organic matter > residual > oxides and hydroxides > carbonates >
exchangeable. For Pb the order was: oxides and hydroxides > residual > organic matter >
carbonates > exchangeable. At the Bay, the pattern was similar for Pb and Cu but
slightly different for Zn in which the fraction associated to OM was lower.
G A L V E Z - C L O U T I E R ET AL. O N R E T E N T I O N O F H E A V Y M E T A L S
317
Hg (m g / k g )
0
20
-|
40
_'1
60
-*
8(1
100
ql!
120
'i
4
140
160
~I
I. BHH
L CN
-
- A --
"~
I. BN
I
,- ~
f
-,
Baie des Ha!Ha!

~Confluence
.L _ B r a s N o r d
interface
BHH
minterface
Confl.
- -interface
BN
........
SSE
: 0,1mg/kg
....
SEM
:0,2 mg/kg
- - S E T
:2 mglkg
180
2O0
Figure 6 - - Mercury Profiles at the 3 Sites

Since Cu and Zn are essential elements to living organisms, their affinity to organic
matter is confirmed. Pb can form easily Pb carbonate or oxide and Zn can form easily
ZnO and thus, these associations were confirmed by the EES results. Despite the
abundance o f carbonate, Pb or even Zn did not associate largely with carbonates. This is
explained by the fact that sediments were slightly anoxic and that a relatively large
portion of these elements was associated to the residual fraction. Indeed, from Figure 6 it
can be seen that there is an important background concentration (visible at about lm
depth and deeper) for Zn > Pb > Cr but almost absent for Cu.
318
Figure 7a -7c-- Heavy Metal Contamination Profiles at the 3 Sites
GALVEZ-CLOUTIER ET AL. ON RETENTIONOF HEAVY METALS
319
Figure 8 ---Heavy Metal Mass Distribution Patterns for the Northern Branch and
Ha.t Ha! Bay
As presented earlier in Figure 5 and in Yong (1995), these fractions do not present
equal potentials for release. Exchangeable, the most easily releasable metals, due to the
fact that alkaline ions (Na, K, Ca, Mg) in enough concentrations can exchange with
heavy metals, are not present in great amounts. Carbonates, the second more easily
extractable fraction is important for Pb but in a small proportion compared to other
retention mechanisms. The strongest association is with the residual fraction, which for
Zn and Cu counts for almost 50% of the total concentration in the Bay and 35% for the
northern branch. Given the poor oxic (over 50 m depth) and saline conditions of the
Saguenay fjord the forms with higher potential for release would be Cu and Zn. However
these fractions do not represent more than 30 to 40% of the total metal concentration,
which will always fall under the SSE (no effect) level. Included in the residual phase,
sulphur minerals such as pyrite can be included. If conditions change to more oxygenated
320
status, AVS (acid volatile sulphides) and their associated metals can be released, this may
increase metal bioavailability.
Conclusion
Before the '96 flood, heavy metals were present in various concentrations higher than
the background value but lower than the minimal effect level. On the contrary, Hg has
exceeded the toxic level in each interface layer. This is more critical for the bay, where
the interface layer is close to the surface (20 em deep). The Northern Branch has its
interface located at 60 can deep.
This study allowed the comparison between the ancient layer (old state of
contamination) and the new layer plus interface (new state of contamination) after 4 years
of the flood. According to our results and analysis, heavy metals are trapped within the
ancient and interface layer and are not in direct contact with the recent layer or upper
water column. If the physical conditions do not change, the adsorbing materials present in
the sediments should retain heavy metals. It would seem that its retaining capacity has
not been exceeded. Carbonates and oxides are present in abundance and offer
precipitation and complexation sites for further retention of heavy metals.
This study did not evaluate other forms of transport mechanisms such as
bioturbation, which gave new health to the zone and should develop greatly in the near
future. Other parts of the Saguenay project focused on this issue.
References
American Society of Agronomy (ASA), 1986, "Methods of soil analysis", C.A.

Black Editor.
ASTM International, 1998, Special Procedure for Testing Soil and Rock for
Engineering Purposes, West Conshohocken, PA, pp. 101-103.
Association Franqaise de Normalisation (AFN), 1982, Evaluation des carbonates.
norme exp6rimentale X31-105.
Barbeau, A., Bougie, R., and C6t6, J., 1981, "Temporal and Spatial Variations of
Mercury, Lead, Zinc and Copper in Sediments of the Saguenay Fjord" Canadian Journal
of Earth Sciences, 18, pp. 1065-1074.
Bureau de Normalisation du Qu6bec, 1993, Granulats-essal au bleu de m6thyl6ne.
BNQ 2560-255.
Centre Saint Laurent, 1993, "Methods Manual for Sediment Characterization" Plan
d'action Saint- Laurent.
Duquette, M. and Hendershott,W., 1993, Soil Surface Evaluation by Back titration.
Soil Science Society of America Journal, 57, pp. 1222-1228.
Fortin G.R and M. Pelletier, 1995, "Synth6ses des connalssances sur les aspects
physiques et chimiques de l'eau et des s6diments du Saguenay" Rapport technique,
Zones d'intervention prioritaires 22 et 23, Centre Salnt-Laurent, Environnement Canada
Gagnon C., Pelletier, E. and Maheux S., 1993, "Distribution of Trace Metals and
Some Major Constituents in Sediments of the Saguenay Fjord, Canada" Marine Pollution
Bulletin, Vol. 26, N~ pp. 107-110.
Galvez-Cloutier, R., Yong, R., Chan, J., and Bahout, E., 1995, "Critical analysis of
Sediment Quality Criteria and Hazard/Risk Assessment Tools." ASTM STP1293. Proc.
321
of the 1995 Int. Symposium on Dredging, Remediation and Containment of

Contaminated Sediments, pp. 306-316.
Galvez-Cloutier, R., 1995, "Study of Heavy Metal Accumulation Mechanisms in the
Lachine Canal Sediments". PhD Thesis, Department of Civil Eng. McGill University,
Montreal. 195 pages.
Ize, S., 2000, Rapport de stage : R6sultats sur la caract6risation physico-chimique des
s6diments du Saguenay, ENTPE, France.
Ministry of Environment Canada (MEC), 1992, "Review and Recommendations for
Canadian Interim Environmental Quality Criteria for Contaminated Sites" Inland Water
Directorate, Water Quality Branch, Scientific Series No 197.
Ministry of Environment Ontario (MEO), 1990, "Provincial Sediment Quality
Guidelines," p. 21.
Smith, J. and Walton, A., 1980, "Sedimentation Accumulation Rates and
Geochronologies Measured in the Saguenay Fjord using the Pb-210 dating method"
Geochimica et Cosmochimica Acta, 44, pp. 225-240.
Yong, R., Galvez-Cloutier, R., and Phadungchewit, Y., 1993, Selective Sequential
Extraction Analysis of Heavy Metal Retention in Soil. Canadian Geotechnical Journal,
30, pp. 834-847.
Yong, R. and Phadungchewit, Y., 1993, pH Influence on Selectivity and Retention of
Heavy Metals in Some Clay Soils. Canadian Geotechnical Journal, 30, pp. 821-833.
Yong, R.,1990, Buffer Capacity and Lead Retention in Some Clay Materials. Water,
Air and Soil Pollution J., 53, pp. 53-67.
Yong, R., 1995, The fate of Toxic Pollutants in Contaminated Sediments. ASTM STP
1293 Proceedings of the 1995 Int. Symposium on Dredging, Remediation and
Containment of Contaminated Sediments, pp. 13-19.
STP1442-EB/Jan. 2003
Author Index
A
Johns, Chris A., 194
Alshawabkeh, Akram, 236

Alves, Maria F. C.. 72
Arsenault, Sylvain, 35
Azma, Behnaz Dahr, 208
Kabir, Ansm, 224

Karunaratne, Godakuru P., 141
Kesavadas, V., 53
Balachandran, K. K., 53
Barbosa, Jane R., 72
Barbour, S. Lee, 194
Bo, Myint Win, 141
Bourg, Cyril, 310
Brin, Marie-Eve, 35
L
Lee, Homa J., 253
Leroueil, Serge, 35
Locat, Jacques, 3, 87, 119, 310
Long, Bernard, 119
M
Moreau, Ann-Laure, 119

Morgenstern, Peter, 19
Mucci, Alfonso, 3, 87
Mulligan, Catherine N., 208
Muffs, Myriam, 3 l0
Chang, Dave Ta-Teh, 156

Chen, Bor-Ling, 156
Choa, Victor, 141
D
Degtiareva, Antonina, 282

Desrosiers, Gaston, 3, 87
Dominguez, Gerardo, 35
Dueri, Sibylle, 167
Nair, K. K. C., 53
Nair, Maheswari, 53
Noble, Marlene A., 253
Ebadi, Taghi, 282

EI-Rayis, Osman A., 183
Elqi, ,~ebnem, 296
Elektorowicz, Maria, 282
Ouellet, Yvon, 119

P
Paimpillil, Joseph Sebastian, 53

Paschke, Albrecht, 19
Patterson, Timothy S., 102
Pelletier, I~.milien,3, 87
Picornell, Miguel, 224
Pillar, Valeffo D., 62
F
Fernandez, Loretta A., 236
Fox, Patrick J., 266
Freyer, Klaus, 19
G
Raya-Rodriguez, Maria Teresa, 62

Ricking, Mathias, 28
Rodrigues, Mafia Lucia K., 62, 72
Galvez-Cloutier, Rosa, 35, 310

H
Haug, Moir D., 194

Henry, Karen S., 236
Hill, Philip, 119
Ho, Nuan-Hsuan, 156
S
Schubert, Michael, 19
Schulze, Tobias, 28
323
Copyright* 2003 by ASTM lntcrnational
www.astm.org
324
Sheahan, Thomas C., 236

Sung, Chao-Ping, 156
T
Tan, Thiam Soon, 141
Teixeira, Elba C., 72
Therrien, Rend, 167
Thresiamma, Joseph, 53
Tremblay, Hd~ne, 3, 87
W
Weiss, Holger, 19
Wennrich, Rainer, 19
Work, Paul A., 296
X
Xu, Jingping, 253

Z
Zeman, Alex J., 102
STP1442-EB/Jan. 2003
Subject Index
A
Electroosmosis, 224
Environmental chemistry, 28
Environmental Fluid Dynamics Code,
296
Erodability. 119
Eutrophic sediments, 35
Adsorption, 236
Advection, 266
Alexandria, Egypt, 183
Annular flume, 119
Anthropogenic influence, 53
ASTM D 4355, 156
Factorial design, 167

Flood layer, 3, 87, 167, 310
Freeze-core techniques, 28
Freezing resistance, 156
Barium, 224
Biogeochemical barrier, 3
Biological signature, 87
Biosuffactants, 208
Bioturbation, 3
Brine removal, 194
G
Geochemical distribution, 72, 310
Geochemical signature, 87
Geocomposite, 236
Geological signature, 87
Geosynthetics, 236
Geotechnical signature, 87
Geotextiles, 156, 194, 236
Geotube, 156
Gravity drainage, 194
Great Lakes, 102
C
Cable corrosion, 282
Cal River basin, 62
Candiota Region, 72
Capping, 102
CCTI, 266
Coal, 72
CO2-LANCE, 28
Communication cables, 282
Consolidation, 14 l, 266
Contaminant migration, 167
Contaminant transport, 266
Corrosion, 282
Critical shear stress, 119
Cuddalore, 53
Hamilton Harbour, 102

Hartwell Lake, 296
Heavy metals, 19, 35, 62, 167, 208, 310
geochemical distribution, 72
High-resolution analysis, 28
Hydrodynamic circulation, 296
DDT, 253
Deposition, 296
Dike core, 156
Discriminatory power, 53
Dispersion, 266
Dredge spoil, 236
Dredging, 102
Durability, 156
In situ measurements, 119

Insolubles, 194
L
Lagoon Maryout, 183
Land reclamation, 141
Large strain consolidation. 266
Layer instability, 3
Lead. 282
LINDE-LANCE device, 28
EFDC model, 296

Eh-pH diagrams, 282
325
326
M
Matrix shape analysis, 156

Mercury, 310
Metal remobilization, 3
Metals, 183
Meteorological event, extreme, 3
Mobilization, 19
Monsoon effect, 53
Multivariate statistics, 62
N
Numerical model, 167, 266, 296
Nutrients, 35
O
Ocean, coastal, 253
Organic carbon, 183
Overburden stress, 224
Oxidation, 19
Sediment cap, 167

Sediment control, 156
Sediment distribution, 253
Sediment quality, 282
Sediment transport, 253, 296
Sediment washing, 208
Sensitivity analysis, 167
Sequential extraction, 72, 208
Sewer outfalls, 253
Slimes, 194
Slurry, 141,224
Soft sediments, 28
Soil consolidation, 236
Soil reclamation, 224
Sorption, 266
Southern California, 253
Spatial variability, 87
Speciation. 282, 310
Stress analysis, 156
Sulphates, 19
Palos Verdes, 253

Physical oceanography, 253
Polychlorinated biphenyis, 296
Polycyclic aromatic hydrocarbons, 3, 167
Potash, 194
T
Talimu River, 156
Tensile strength, 156
Toxicity, 102
Trace metals, 3, 53
Turbidite layer, 167
Randomization tests, 62
Reclamation of insolubles, 194
Remediation, 102, 208, 236
Restoration, 183
Retention, 310
Reverse-flushing pond, 194
Rhamnolipid, 208
S
Saguenay Fjord, 3, 87, 119, 167, 310
St. Augustin Lake, 35
St. Lawrence River, 282
Sand capping, 141
Santa Monica, 253
Sedimentation, 14 I, 296
U
UV resistance, 156
W
Waste disposal, 253
Waste sediment, 141
Water column, 282
Water quality, 282
Weathering, 19
Y
Yield shear stress, 141
Z
Zeolite, 236

Contaminated Sediments PDF

Uploaded by

Copyright:

Available Formats

Contaminated Sediments PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Contaminated Sediments PDF

Uploaded by

Copyright:

Available Formats

STP 1442

Jacques Locat, Rosa Galvez Cloutier, Ronald Chaney,

ASTM Stock Number: STPI442

Library of Congress Cataloging-in-Publication Data

Peer Review Policy

Printed in Ann Arbor, MI

SECTIONI: SEDIMENT CHARACTERIZATION

The Weathering Behavior of Contaminated Industrial Sediments after Their Exposure

Deep-Freeze Sampling Methods for Soft Sediments--MATHIASRICKING

Trace Metal Levels in Nearshore Sediments Close to Industrial Discharges off

Study of the Geochemical Distribution of Heavy Metals in Sediments in Areas

Characterization of a Catastrophic Flood Sediment Layer: Geological, Geotechnical,

In Situ Flume Measurements of Sediment Erodability in Saguenay Fjord

SECTIONII: MITIGATION AND RESTORATION METHODS

Factors Controlling Contaminant Transport Through the Flood Sediments of the

AND RENI~ THERRIEN

Steps for Restoration of a Polluted Egyptian Closed Lagoon on the Alexandria

Rehabilitation of Brine-Saturated Sediments--MOiR D. HAUG,S. LEEBARBOUR,

AND CHRIS A. JOHNS

Use of Selective Sequential Extraction for the Remediation of Contaminated

Effects of Overburden Stresses on Soil Reclamation by ElectroosmosiS--ANSM KABIR

A Reactive Geocomposite to Remediate Contaminated, Subaqueous

SECTIONIll: MONITORING AND PERFORMANCE

Numerical Model for Contaminant Transport in Consolidating Sediments-PATRICKJ. FOX

Numerical Modeling of Hydrodynamic Circulation and Cohesive Sediment Transport

Section I: Sediment Characterization

Characterization, Evaluation, Mitigation~Restoration, and Management Strategy

the BenthicArea Directly Affected by the 1996 Flood.

PELLETIER ET AL. ON FLOOD CAPPING LAYER

Sediment Deposition and Capping Process

Figure 2 - Temporal Evolution of the Backscatter Data Over Multibeam Sonar

Surveys (modified from Urgeles et al. 2002)

PELLETIER ET AL. ON FLOOD CAPPING LAYER

Total PAIls (mg.kg "I)

Changes in Hg, Cd and PAHs concentrations in a sediment core collected in the

The Performance of the Capping Layer

PELLETIER ET AL. ON FLOOD CAPPING LAYER

0.0 2.5 5.0

Figure 4 - Geotechnical Profiles o f the 1996 Turbidite

-Ill-- Solid Fc(HCI) (wt%)

Solid As(tot) (ppm)

Solid Hg(tot) (ppb)

- - AVS gmol g ' )

Figure 5 - Steady State Diagenetic Conditions in the Fjord as Represented by the

Manganese remobilization occurred almost unimpeded following the emplacement

PELLETIER ET AL. ON FLOOD CAPPING LAYER

-a-- Solid Fe(HCI) (wl%)

-4-- Solid As(to0 (ppm)

---a- Solid Mn(HCI)(ppm)

1000 2000 3000

Solid Hg(to0 (ppb)

100 200 300 400

Solid As(tot) (pOre)

Solid Hg(tot) (0pb)

Figure 6 - Evolution of Sediment and Porewater Geochemistry at Station 9 in the Bale

PELLETIER ET AL. ON FLOOD CAPPING LAYER

Figure 7- Distribution of Macrobenthic Organisms Found in Sediments of Five Stations

of the Baie des Ha!Ha!from 1996 to 2001.

Occurrence of PAHs in the New Layer