A STREAM CHANNEL STABILITY ASSESSMENT METHODOLOGY

BY

David L. Rosgen, P.H.
Wildland Hydrology
Pagosa Springs. CO 81147

ABSTRACT: Various definitions of stream channel stability are presented including "the natural stable channel",
the graded river, dynamic equilibrium, and regime channels, and a quantitative assessment methodology is presented
that distinguishes between stability states. The assessment procedure involves a stream channel stability prediction
and validation methodology on a hierarchical framework. The stream channel stability method develops field-
measured variables to assess: 1) Stream state or channel condition variables, 2) Vertical stability
(degradation/aggradation), 3) Lateral stability, 4) Channel patterns, 5) Stream profile and bed features, 6) Channel
dimension factor, 7) Channel scour/deposition (with competence calculations of field verified critical dimensionless
shear stress and change in bed and bar material size distribution), 8) Stability ratings (modified Pfankuch method)
adjusted by stream type, 9) Dimensionless ratio sediment rating curves by stream type and stability ratings, and 10)
Selection of position in stream type evolutionary scenario as quantified by morphological variables by stream type
to determine state and potential of stream reach.

The stability assessment is conducted on reference reach (stable) reaches and a departure analysis is performed when
compared to an unstable reach of the same stream type. The assessment procedure utilizes various hierarchical
levels for prediction and subsequent validation. Changes in the variables controlling river channel form, primarily
streamflow, sediment regime, riparian vegetation, and direct physical modifications can cause stream channel
instability. Separating the difference between anthropogenic versus geologic processes in channel adjustment is a
key to prevention/mitigation/restoration of disturbed systems

The adverse consequence of stream channel instability (dis-equilibrium) is associated with increased sediment
supply, land productivity change, land loss, fish habitat deterioration, changes in both short and long-term channel
evolution and loss of physical and biological function.

INTRODUCTION

Definitions: Within the scientific community, the terms "channel stability", "equilibrium", quasi-equilibrium and
"regime channels" evoke a deluge of various interpretations. Imagine the quantitative inconsistency of the field
observer in trying to implement a stream channel stability assessment procedure with which there is not common
agreement on what is meant by the term? Thus, it is not uncommon for journey-level professionals working with
rivers to disagree on a consistent working definition of what constitutes a stable river, even though they often use the
term "channel stability". A review of the literature provides insight into previous interpretation of terms, that all
appear to be synonymous, or at the least, have a common thread of similarity. Davis (1902), defined a "graded "
stream as the condition of "balance between erosion and deposition attained by mature rivers". Mackin (1948), as
reported by Leopold et al (1964), defined a graded stream as "one in which, over a period of years, slope is
delicately adjusted to provide, with available discharge and with prevailing channel characteristics, just the velocity
required for the transport of the load supplied from the drainage basin. The graded stream is a system in equilibrium;
its diagnostic characteristic is that any change in any of the controlling factors will cause a displacement of the
equilibrium in a direction that will tend to absorb the effect of the change." The controlling factors described by
Leopold et al (1964) were width, depth, velocity, slope discharge, size of sediment, concentration of sediment and
roughness of the channel. If any one of these variables were changed it sets up a series of concurrent adjustments of
the other variables to seek a new equilibrium. The central tendency of rivers to seek a probable state was described
by Leopold (1994). Strahler (1957) and Hack (1960), used the term "dynamic equilibrium" referring to an open
system in a steady state in which there is a continuous inflow of materials, the form or character of the system
remains unchanged. Equations showing river variables as a function of discharge were derived by Leopold and
Maddock, (1953), and by Langbein, (1963). These hydraulic geometry relations described adjustable characteristics
of open channel systems in terms of independent and dependent variables in quasi-equilibrium (not aggrading nor
degrading). Streams described to be "in regime" are synonymous with "stable channels" and equations describing
three dimensional geometry of stable, mobile gravel-bed rivers were presented by Hey and Thorne (1986).
Additional equations and discussion on stable river morphology were presented in Hey (1997). Regime channels, as
discussed by Hey (1997) allows for some erosion and deposition but no net change in dimension, pattern and profile
for a period of years. The following definition of stream channel stability was presented by Rosgen (1996): "is the
ability of a stream, over time, in the present climate, to transport the sediment and flows produced by its watershed
in such a manner that the stream maintains its dimension, pattern and profile without either aggrading nor
degrading". Processes of stream channel scour and or deposition have to occur in a natural stable channel, but over
time, if this leads to degradation or aggradation, respectively, then the stream would not be stable. This definition
summarizes many of the key points previously presented in the literature. This definition is predictable and
verifiable, and as such, was used in the development of the stream channel stability assessment methodology.

PRINCIPLES

River instability needs to be evaluated on spatial and temporal scales. It is also critical to recognize natural geologic
erosion and transport mechanics versus anthropogenic influences. Following major floods, due to requirements to
provide flood damage restoration plans, the author studied alluvial gravel-bed streams on slopes less than 0.02
where the pre and post-flood morphological variables were similar. Other reaches, however, that were in poorer
stability condition prior to the flood, received major damage by the same flow. The stable rivers became reference
reaches where data were collected on dimension, pattern, profile and channel materials. The1984 Lawn Lake flood
in Colorado inundated Fall River, a C4 stream type (for stream type descriptions see Rosgen, 1994, 1996) that was
in a stable meandering pattern. The extensive sediment load and corresponding "flood of record" did not create
instability. The stream maintained its dimension, pattern and profile and did not aggrade nor degrade.
Accumulations of sand occurred in the channel and within a few years the sand was routed through without the net
effect of aggradation. This stream is but one of many examples where the author has field evidence where post-
flood instability did not occur, even though these streams had potentially erodible material in their bed and banks.
Reference reaches such as this become a blueprint of the variables associated with stable natural channels. Field and
photographic evidence of channel change over time is an excellent reference procedure. Selection of the reference
reach involves collection of such evidence. Descriptions and applications of the reference reach methodology are
described in Rosgen (1998).

Stream channels that have been improperly
managed and have poor riparian vegetation are
subjected to accelerated streambank erosion and
corresponding channel adjustments leading to
instability. An example of instability that
occurred due to willow removal on a C4 stream
type on the Weminuche River in Southwestern
Colorado is dramatic not only for the magnitude
of change, but the consequence of change, as
well. The details of the combined effects of
willow spraying on stream channel instability and
changes in dimension, pattern and profile for this
reach are summarized in Rosgen (1996). The
consequence of a wide range of stream channel
instability can be described and quantified
through an evolution of stream types (Figure 1).
The evolution sequence that ensued on the
Weminuche River due to channel adjustment
following disturbance, created a change in
morphological stream types that is associated
with sequence category #3 in Figure 1. The
conversion changed the pre-disturbance C4
stream type (gravel-bed, meandering channel
with a floodplain), to D4 (braided), to G4
(incised gully due to avulsion), to an F4
(entrenched, meandering channel) and was
widening to eventually re-establish a C4 stream
Figure 1. Various Stream Type Evolution Scenarios
1. E C Gc F C E
2. C D C
3. C D Gc F C
4. C G F Bc
5. E Gc F C E
6. B G Fb B
7. Eb G B
8. C G F D
type, but at a lower elevation. Every tributary was rejuvenated due to the change in local base level, which created a
tremendous increase in sediment supply and transport and caused the water table to drop in the meadow, decreasing
productivity. Thus, the consequence of spraying the willows and induced stream channel instability was associated
with major loss of: land, vegetation productivity, fish habitat, visual values and loss of ability to handle future
floods. Increased sedimentation, both on-site and downstream occurred. Many other evolutionary scenarios
induced by channel instability and associated channel adjustment can occur. The author has observed at least eight
separate evolutionary scenarios as shown in Figure 1. One challenge in stability assessment is to determine the
evolutionary state and sequence of the stream. The cause of the instability is as important to understand as well as
the consequence.

OBJECTIVES

To prevent and or to correct stream channel instability, it is imperative to understand the mechanisms causing the
shift in morphological variables and stability indices. The diversity of opinion has made it difficult to conduct
consistent quantitative river stability assessments. It is not uncommon to have five individuals all "trained in these
matters", simultaneously standing on the same bank of a river, having five divergent and conflicting opinions.
Unless there are documented measurements, coupled with consistent, quantitative indices of stability, these
subjective opinions will persist. Understanding of these complex processes can only come with a program of
detailed measurements so that observations, stability indices and field assessment techniques can become effective.
To meet this objective, the author set up a river inventory hierarchy (Rosgen, 1996), (Figure 2). This would allow
an assessment at various levels
appropriate to the level of inquiry. All
4 levels are used initially, until
quantitative relations are established
with the prediction methods. Initial
stratification is accomplished at both
levels I and II. This is not done to
determine stability, but to stratify the
reach by valley and stream type.
Reference reach data is also obtained
from adjacent stable reaches of the same
valley and stream type. Reference
reaches do not have to be pristine or
relic sites, but meet the criteria of a
stable river. Prediction of stability is
made at level III, the "state" or
condition level. Level IV is the
validation inventory that requires the
greatest level of measurement detail
over a longer time period. For example,
one may estimate vertical stability or
bank erosion rate at level III, however
permanent cross-sections are re-
measured following runoff to verify bed
elevation shifts, and erosion pins/toe
pins are established at level IV to verify
the actual erosion that occurred. This
design allows prediction model
validation at level III, thus, the
prediction model can be extrapolated
without the need to always accomplish
level IV. Since these assessments
involve large areas and many miles of
river, this approach was designed to
provide a prediction methodology with
some credible validation.
DEPOSITIONAL
MATERIALS
FLUVIAL
PROCESS
STRUCTURAL
CONTROLS,
LITHOLOGY
CLIMATIC
INFLUENCE
BROAD
LIFE ZONES
BASIN RELIEF - LANDFORM S - VALLEY M ORPHOLOGY
W ATERSHED
CHANNEL SLOPE
Valley Slope / Sinuosity
CHANNEL SHAPE
Narrow- Deep
Wide - Shallow
GEOMORPHIC
CHARACTERIZATION
LEVEL I
StreamTypes
"A" through "G"
ENTRENCHMENT RATIO
WIDTH/DEPTH RATIO
SINUOSITY
RIPARIANVEGETATION
DEPOSITIONPATTERN
DEBRISOCCURRENCE
MEANDERPATTERN
Channel STABILITYRating
... SEDIMENTSUPPLY
... BEDSTABILITY
... W/DRATIO"STATE"
CHANNEL SLOPE
CHANNEL MATERIALS
STREAM "STATE" or CONDITION
LEVEL III
BankEROSIONPotential
StreamSIZE/ ORDER
FLOW REGIME
AlteredChannel "STATE"
... DIMENSIONS
... PATTERNS
... SLOPE
... MATERIALS
SEDIMENT
MEASUREMENTS:
Bedload Sediment
Suspended Sediment
STREAMFLOW
MEASUREMENTS:
Hydraulics
Resistance
Hydrographs
REACH SPECIFIC STUDIES
LEVEL IV
VALIDATION LEVEL
STABILITY:
Aggradtion/ Degradation
SEDIMENT:
Change in Storage &
Mtrs. Size Distribution
Bank Erosion Rates
Imbeddedness/ Distribution
Time Trends - Stability
CHANNEL PATTERNS
SingleThread
MultipleThread
Anastomosed
Sinuosity
Meander WidthRatio
DRAINAGE NETW ORK
MORPHOLOGICAL DESCRIPTION
LEVEL II
StreamTypes
A1 - A6 .......G1 - G6
Figure 2. The Hierarchical River Inventory stratifying stability
prediction and detailed river measurements by valley types and stream
types for various levels of inquiry


METHODOLOGY

This section of methodology is meant to be a sequence of suggested steps for the field practitioner to use in reaching
final conclusions and making recommendations for management and/or restoration. The stream channel stability
assessment methodology is broken into the following ten major categories. Based on field inspection and
measurements the categories of assessment are applied to the reference reach, as well as for impacted reaches. This
provides a consistent comparative analysis of departure and assists in selecting evolutionary shifts in stream type
and associated dimensionless sediment rating curves. A general summary of stability ratings and interpretations are
included at the end of these categories of assessment.

1) Stream Channel Condition or "State" Categories: Determine condition categories from field inspection and
measurement of stream channel condition characteristics. Specific categories are evaluated and documented based
on the criteria for each variable. Detailed descriptions and examples for each category are presented in Chapter 6
(Rosgen, 1996) which will help completing these assessments. The seven categories and associated variables
evaluated are: a) Riparian vegetation, (composition, density, and potential, climax riparian communities); b)
Sediment deposition patterns (8 patterns); c) Debris occurrence (includes large woody debris); d) Meander patterns
(8 patterns); e) Stream size/Stream order; f) Flow regime (perennial, ephemeral, intermittent, subterranean,
snowmelt, stormflow, rain-on-snow, spring-fed, glacial-fed, tidal, diversions, and reservoir regulated, and; g)
Altered states due to direct disturbance (dimension, pattern, profile and materials such as, channelization,
straightening, levees, concrete, rip-rap, etc.). These seven major condition states provide insight into specific
characteristics of the reference reach, as well as the stream type being assessed.

2) Vertical Stability/Degradation/Aggradation: From field measurements of bank height and entrenchment
ratios and documented observations of excessive erosion and/or deposition, determine vertical stability of the stream
reach. The degree of incision involves a measurement of bank height ratio (Table 1). It is measured as the ratio of
the lowest bank height of the cross-section divided
by maximum bankfull depth. For example a stream
could be incising and not yet abandoned its
floodplain or flood-prone areas. Bank height ratios
of 1.2 and 1.3 are characterized by both
streambanks eroding as the bank height is often
below the rooting depth of the riparian vegetation.
To determine if the stream has incised to the extent
that the stream has abandoned its floodplain is
determined by the entrenchment ratio, which indicates vertically containment. The entrenchment ratio is calculated
by first determining the elevation of the flood-prone area as measured at twice the maximum bankfull depth. The
floodprone area width at this elevation is then divided by the bankfull width. If the Entrenchment ratio is less than
1.4 (+or- 0.2), the stream is entrenched (Rosgen, 1994,1996). Additional indicators of incision/degradation are:
both left and right stream banks actively eroding, depositional features are being scoured, decrease in width/depth
ratio corresponding with increase in bank height ratio, and mobilization of largest size D-100 of bed material (see
category 8). The aggradation category is determined from a summary of the depositional patterns, coarse deposition
on floodplains and very high to extreme width/depth ratios. Longitudinal profiles of the reach showing elevations of
the bed, water surface, bankfull and lowest bank height indicate if the incision is advancing downstream or if a head-
cut is advancing from the downstream direction. Profiles and cross-sections should be permanently monumented
and read annually to verify the prediction of vertical stability.

3) Lateral Stability: Determine the degree of lateral containment (confinement) and potential lateral accretion.
The categories used for lateral stability are: a) Meander width ratio (degree of confinement) and b) Streambank
Erosion Hazard Index (BEHI) and Near-Bank Stress (NBS) (see Rosgen, 1996 and 2001, In Press). Meander width
ratio is the meander belt width (lateral containment of the channel within its valley) divided by bankfull channel
width. Values of meander width ratio by stream type are shown in Rosgen (1996, p.4 -9). Some streams can be
confined, but not entrenched. This provides insight into channel adjustment processes by stream type and degree of
confinement. Annual, lateral streambank erosion rates are multiplied times the bank height and stream length of
specific BEHI and NBS ratings along the reach. These values are converted to tons/year in order to apportion
Table 1. Conversion of bank height ratio (degree of
incision) to adjective ratings of stability
Stability Rating Bank Height Ratio
Stable (low risk of degradation) 1.0 - 1.05
Moderately unstable 1.06 - 1.3
Unstable (high risk of degradation) 1.3 - 1.5
Highly unstable >1.5
sediment supply sources. Many miles of stream can be evaluated using this method of prediction.. Level IV data
involves installing toe pins and cross-sections to accurately measure streambank erosion rates/lateral accretion. This
helps validate the model or revise estimates and better reflect actual rates. The validation work can also be used for
effectiveness monitoring prior to and following restoration and/or streambank stabilization.

4) Channel Pattern: Measure meander width ratios (meander length/bankfull width), ratio of radius of
curvature/bankfull width, sinuosity, meander width ratio (belt width/bankfull width), arc length and arc angle.
Convert all values to dimensionless ratios for comparative purposes. Additional assistance can be provided in
assessing channel pattern categories as shown in Chapter 6, Rosgen, (1996). Changes in pattern are compared using
dimensionless ratios when the reference reach data for the same valley and stream type may be of a different size.
Channel adjustment due to instability can often be interpreted from these variables such as accelerated down-valley
meander migration and excessive near-bank stress due to ratios of radius of curvature/width less than 2.0. Level IV
data utilizes aerial photo time/trends and cross-sections showing down-valley meander migration.

5) River Profile and Bed Features: A longitudinal profile is measured to determine changes in river slope
compared to valley slope which is very sensitive to sediment transport, competence and the balance of energy. Pool
to pool spacing, ratios of maximum depth of pools/mean depth of channel, and maximum depth of riffles/mean
bankfull depth are also obtained from longitudinal profile data. When pools start to fill (decrease in max
depth/mean depth ratio), and the stream is widening with a corresponding decrease in sinuosity and increase in
slope, the stream is becoming unstable. The reference condition for the same stream type will have dimensionless
ratios that are used for comparison of the magnitude of departure. Spacing of step/pools in steeper stream types are
inversely proportional to slope and directly proportional to width, and as such, are shown as a ratio of bankfull width
by slope categories. The total removal of large woody debris often increases the step/pool spacing and as a result
the excess energy increases the potential for channel degradation. Level IV validation of prediction estimates are
accomplished by installing permanent longitudinal profiles with bench marks tied into permanent cross-sections or
stationing pins. Measurements taken on a thalweg survey provides data on maximum bankfull depths, the various
bed features, including riffles and pools, and documents any change in slope. Elevation measurements of the bed,
water surface, bankfull, and low bank height also identifies changes in degree of incision along the profile as
presented above in assessment item 2). Data summaries including dimensionless ratios for bed features and river
profile can be recorded and analyzed in the "Reference Reach Field Book", (Silvey and Rosgen, 1998.

6) Channel Dimension Relations: Determine changes in the bankfull width and mean bankfull depth (width/depth
ratio). This ratio indicates departure from the reference reach and is very sensitive and diagnostic of instability.
Increases in width/depth ratio are often associated with accelerated streambank erosion, excessive sediment
deposition, streamflow changes, channel widening due to evolutionary shifts from one stream type to another (i.e.,
G4 to F4 to C4), and direct alteration of channel shape from channelization, etc. The degree of width/depth ratio
increases are shown as a departure from the reference condition of the stable stream type to establish stability
ratings (Table 2). A decrease in width/depth ratio departure analysis will have a proportionate reduction in
width/depth ratio values. This reduction from the reference
condition is only applied when the bank height ratio is greater
than 1.0. For example a "moderately unstable" rating for a
stream channel with a bank height ratio greater than 1.0 would
have a width/depth ratio decrease of 0.8 to 0.6 This is
associated with a width/depth ratio that is decreasing as the
stream is incising (i.e.,C4 stream type conversion to a type G4
). The corresponding reduced width/depth ratio creates excess
shear stress in an incising stream type, which is adjusting
toward an unstable condition. The level IV analysis
establishes permanent, monumented cross-sections to
determine the rate and extent of change in both the width/depth and bank height ratios.

7) Stream Channel Scour/Deposition Potential (Sediment Competence): Compute critical dimensionless shear
stress to determine the size of sediment particle that can be moved. Relations modified from Andrews (1984) and
Andrews and Nankervis (1995) are used for this computation. Subsequent calculations using a Shields relation
compares the existing slope and depth of a stream to be able to transport the largest size made available annually
(during bankfull stage) to the channel. The procedure involves sampling the bed material on the riffle to obtain d
50
,
Table 2. Conversion of width/depth ratios to
adjective ratings of stability from reference
conditions
Stability Rating Ratio of W/D Increase
Very stable 1.0
Stable 1.0 - 1.2
Moderately unstable 1.21 - 1.4
Unstable >1.4
excavate a core sample of bar material (located on the lower 1/3 of meander on the point bar midway between the
thalweg and the bankfull stage). The bar sample is used to obtain d
s50
of the relation as a surrogate of the sub-
pavement size distribution. Locations of this specific depositional feature and subsequent are shown in Chapter 7,
(Rosgen ,1996). The bar also provides an interpretation of the size distribution of bedload at the bankfull stage and
the largest size on the bar is used to obtain data representing the largest size of sediment frequently made available
to the channel. The following calculations are used to make the competence prediction:

ci
=.0834 (d
50
/ds
50
)
-.872

Where:
ci
=critical dimensionless shear stress
d
50
=median diameter of pavement or bed material on riffle
ds
50
=median diameter of bar sample (sub-pavement)

The following equation is used to predict the depth and slope to move the largest size of sediment made available to
the channel on a frequent basis:
ci =
__dS
,
(
s
) (Di)
transformed to: d
=
(
ci
) (
s
) (D
i
)
S
Where:
s
=submerged specific weight of sediment
D
i
=Largest diameter of particle on bar (use mm if depth is in meters)
d =mean bankfull depth of the channel
S =water surface slope at the bankfull stage
If the combination of depth and/or slope does not move the largest size, then potential aggradation or excessive
deposition and corresponding high width/depth ratio is anticipated. If the depth and or slope exceeds that required to
move the largest size, then potential degradation, or excess scour leading to incision has potential for instability.
This procedure is verified by three methods at level IV: 1) Measured bedload size distribution of bedload at the
bankfull stage, and corresponding slope and bankfull depth measurements at the bankfull stage, 2) Monumented
cross-sections and vertical scour chains are installed before and after runoff. The scour chains give the depth of
scour and subsequent change of particle size over chain. The largest particle over the scour chain exhibits the
largest size of particle moved for the corresponding shear stress of the flows responsible. The cross-section shows
net change of bed elevation, and specific changes over the scour chain, and 3) Annual replicate core samples at the
same location on the bar shows sizes moved for a back-calculated shear stress as well as shifts in size distribution of
bedload at the bankfull stage. It is the coarse fragment that determines channel morphology of gravel-bed streams
(Leopold, 1992), thus it is important to be able to move the largest sediment clasts frequently made available to the
river at the bankfull stage. These data can also be compared to measured bedload size distribution with the USGS,
Helley-Smith bedload sampler. This field method has been tested on many rivers by the author with excellent
success when compared to both scour chain and measured bedload data.

8) Stream Channel Stability rating (modified Pfankuch procedure): Determine channel stability ratings to
predict potential state from the stable reference reach of the same stream or potential evolutionary type. The stability
rating procedure evaluates the upper and lower banks and streambed for evidence of excessive erosion/deposition.
The procedure has been used for 25 years by the USDA Forest Service and other Federal Agencies, (Pfankuch,
1975). The system evaluates mass wasting potential adjacent to the channel, detachability of bank and bed
materials, channel capacity and evidence of excessive erosion and/or deposition. The larger the number, the greater
the risk for instability. The risk rating of the classification was later converted to ratings by stream type. This
modification was made to reduce the likelihood of applying the same numerical rating of "good" to C4 versus B4
stream types. Naturally, C4 stream types by their meandering nature, flatter slopes and point bars will obtain higher
channel stability numbers than the steeper B4 stream types, even though both streams are very stable. In contrast,
the channel stability ratings for a very stable B4, will be much lower than the C4 stream type, when both are stable.
To remedy this dilemma, relations were developed to place numerical categories in adjective ratings by stream type
(Table 3). For example a rating of "Good" for a B4 has a range of 40-64, whereas, the "good" rating for C4 stream
types is 60-95. Applications for stability and sediment supply have been related to measured sediment rating
curves. For example, the higher the stability rating number, the higher the intercept and steeper the slope of the
suspended sediment rating curve as shown in Figure 3 for Redwood Creek, California (Leven, 1977). A similar
analysis was performed on measured stream data in North Carolina (Coweeta Experimental Forest), Northern
California, Idaho, Montana and Colorado (Rosgen, 1980). This is used in conjunction with dimensionless ratio
sediment rating curves in the next category of assessment. Level IV verification involves the combination of
measured sediment rating curves, cross-sections, longitudinal profiles and channel material size distributions. This
level of assessment compares predicted to observed values of sediment and stability.



Table 3. Conversion of Stability Rating to Reach Condition by Stream Type
Stream Type A1 A2 A3 A4 A5 A6 B1 B2 B3 B4 B5 B6
Good (Stable) 38-43 38-43 54-90 60-95 60-95 50-80 38-45 38-45 40-60 40-64 48-68 40-60
Fair (Mod.
Unstable) 44-47 44-47 91-129 96-132 96-142 81-110 46-58 46-58 61-78 65-84 69-88 61-78
Poor (Unstable) 48+ 48+ 130+ 133+ 143+ 111+ 59+ 59+ 79+ 85+ 89+ 79+
Stream Type C1 C2 C3 C4 C5 C6 D3 D4 D5 D6
Good (Stable) 38-50 38-50 60-85 70-90 70-90 60-85 85-107 85-107 85-107 67-98
Fair (Mod.
Unstable) 51-61 51-61 86-105 91-110 91-110 86-105 108-132 108-132 108-132 99-125
Poor (Unstable) 62+ 62+ 106+ 111+ 111+ 106+ 133+ 133+ 133+ 126+
Stream Type DA3 DA4 DA5 DA6 E3 E4 E5 E6
Good (Stable) 40-63 40-63 40-63 40-63 40-63 50-75 50-75 40-63
Fair (Mod.
Unstable) 64-86 64-86 64-86 64-86 64-86 76-96 76-96 64-86
Poor (Unstable) 87+ 87+ 87+ 87+ 87+ 97+ 97+ 87+
Stream Type F1 F2 F3 F4 F5 F6 G1 G2 G3 G4 G5 G6
Good (Stable) 60-85 60-85 85-110 85-110 90-115 80-95 40-60 40-60 85-107 85-107 90-112 85-107
Fair (Mod.
Unstable) 86-105 86-105 111-125 111-125 116-130 96-110 61-78 61-78 108-120 108-120 113-125 108-120
Poor (Unstable) 106+ 106+ 126+ 126+ 131+ 111+ 79+ 79+ 121+ 121+ 126+ 121+

9) Dimensionless Ratio Sediment Rating Curves: Instability and the corresponding increase in sediment supply is
often reflected in measured sediment rating curves. The source of this increase in sediment supply often is
associated with channel adjustment, including degradation and lateral accretion (bank erosion). The variation in
sediment rating curves is shown in Figure 4, reflecting differences in sediment supply for various Colorado streams
(Williams and Rosgen, 1989, and Rosgen, 1996). Additional sediment rating curves by channel stability ratings
indicating changes in stability and associated sediment supply are shown in Rosgen, (1980). On the Hatchie River
in West Tennessee, Simon (1989) summarized the effects of channelization and corresponding stream stability
change comparing evolution stages of channels to measured upward shifts in the slope of the measured suspended
sediment rating curves. The sediment yields, from the Hatchie River, a stable, meandering, low width/depth channel
with a well developed floodplain (E6 stream type) was 62.9 tons/km
2
(163 tons/year/mi
2
) The South Fork Forked
Deer River which, following channelization, became incised (F6 stream types) with resultant sediment yields of
961.4 tons/km
2
(2,490 tons/yr/mi
2
). Simon (1989) was showing these changes in sediment yield associated with
channel instability and adjustments using the channel evolution model (Shumm, et al 1984 and Simon and Hupp,
1986). The channel evolution model and stages of adjustment are related to quantitative morphological values
corresponding to stream types (Rosgen, 1999). Both of these approaches are compatible at describing the
consequence of channel adjustment. As the E6 stream type incises and changes to a G6 and eventually F6, the
channel goes through an evolutionary adjustment of instability and associated stream type change. The evolutionary
sequence of the Hatchie and South Fork Forked Deer Rivers matches scenario #5 (Figure 1). When instability due
to change in energy, sediment supply or direct disturbance occurs, the severity is such that stream types can change.
This change reflects increases in sediment supply due to channel adjustment (streambed and streambank erosion)
and eventual increased sediment yield as shown by Simon (1989). This does not infer, however, that all F stream
types are unstable, such as the Colorado River in the Grand Canyon. The best assessment approach is a
combination of stability analysis with stream morphology necessary to establish potential departure from the
reference condition.

Reference reach sites representing stable stream reaches are used to establish dimensionless sediment rating curves
by stream type and stream stability and as such, can be used to ascertain departure (Troendle, et al, In Press).
Significant departure from the reference dimensionless sediment rating curves when comparing good and fair with
poor stability ratings. Stream types that become unstable to the extent that they change morphological type are
generally associated with poor stability and an increase in sediment supply. Sediment delivery ratios provide a
means of extrapolation of sediment rating curves for rivers of different geology, size, stability, and associated
morphology. These ratios are developed by taking the bedload and suspended sediment values and dividing them by
the same units of sediment values at the bankfull discharge. Their corresponding discharges are also divided by the
bankfull discharge to establish dimensionless ratio sediment rating curves. To convert these curves to actual
numbers following extrapolation, sediment and discharge measurements at bankfull are obtained at the most detailed
level of river stability assessment, then multiplied by the dimensionless ratios established for that stream type and
stream condition. Sediment and flow data should also be collected at a lower flow to insure the slope and intercept
of the dimensionless ratio sediment rating curve matches observed values. Confidence bands above the reference
reach for the same, but stable stream type using the dimensionless ratios give a preliminary range of departure. In
other words, the natural variability in sediment supply as shown in the sediment rating curves for the reference reach
is documented to avoid the unwise tendency of trying to establish sediment TMDL's as a fixed value for a given
stream. Also, natural geologic sediment rates can be established as reflected in the various stream types such as the
A3a+(steep, debris torrent channel incised in heterogeneous, unconsolidated landslide debris and or glacial till).
These stream types have periodic and catastrophic, naturally high to extreme erosion rates due to their unlimited
sediment supply and high energy. Sediment yields from these systems cannot and should not be altered, as the
entire fluvial system has adjusted over time to accommodate such sediment loads. Efforts to restore "stability" in
these channel types are fighting natural processes and face a high risk of failure.

10) Stream Type Evolutionary Scenarios: Determine the current state and evolutionary sequences as shown in
Figure 1. The use of this relation requires the field observer to select not only the stream type, but the location in a
particular sequence of evolution. This not only provides a current state evaluation, but provides an interpretation of
the physical potential of this reach. A stability assessment can assist those doing restoration design. Often, unstable
channels are "patched in place"unfortunately it is often the "wrong place"or perhaps, the wrong stream type.
Another use of this specific assessment protocol is to be able to identify the potential stable stream type as opposed
to the currently existing stream type. Restoration can speed up the adjustment or recovery period by obtaining the
morphological data used from the reference reach of the appropriate stream type to match the stable form. Another
application of stream type evolution is to specify a potential dimensionless sediment rating curve that would apply
associated with the stability of a particular morphological evolutionary state as depicted in figures 5 and 6.

Summary of condition assessment and stability ratings: The summary of the ten major stability rating categories
and condition variables of the level III prediction analysis are shown in Table 4. By completing each of the above
assessments, the field practitioner can see the pattern of channel change and note a change in one variable is
Figure 3. Suspended sediment rating curves by
channel stability ratings (from Leven, 1977 in:
Rosgen, 1980)
Figure 4. Bedload rating curves stratified by stream type
(Rosgen, 1996, data from Williams and Rosgen, 1989)
accompanied by changes in several others. The interpretation of the stability categories allows the observer to
conclude as to the overall stability and potential state of sediment supply.


Table 4. Summary of stability condition categories for the Level III inventory
Stream Name ____________________________ Observers _________________________________________
Location _________________________________ Stream Type ____________ Date ______________________
Riparian Vegetation, comp/density ____________ Flow regime _______________________________________
Stream size, Stream order ___________________ Depositional pattern _________________________________
Meander pattern ___________________________ Debris/channel blockages _____________________________
Channel stability rating (Pfankuch) _____________ Describe altered channel state _________________________
Stability category by stream type________ _________________________________________________
Sediment supply (check appropriate category) Dimension/shape:
Extreme __________________________ Width _____________________________________
Very high _________________________ Depth _____________________________________
High _____________________________ Width/depth ratio ____________________________
Moderate __________________________ Patterns (*show as function of W
bkf
)
Low ______________________________ Meander length* ____________________________
Streambed (vertical) stability Radius of curve * ____________________________
Bank Height ratio ___________________ Belt width* _________________________________
Aggrading _________________________ Sinuosity ___________________________________
Degrading _________________________ Arc angle __________________________________
Stable _____________________________ Arc length* _________________________________
Profile:
Width/depth ratio/condition: Water surface slope __________________________
Excellent (stable) ___________________ Valley slope_________________________________
Good _____________________________ Bed features: (Type and/or ratio max. depth/bankfull depth)
Fair ______________________________ Riffle ______________________________________
Poor ______________________________ Pool _______________________________________
Streambank erosion hazard: Step/pool (p/p spacing) ________________________
Bank erodibility: Near-bank stress: Convergence/divergence ______________________
Extreme ________ Extreme________ Riffle/pool spacing * _________________________
High ___________ High __________ Dunes/antidune/smooth bed ____________________
Moderate _______ Moderate ______ Describe channel evolution scenario:
Low ___________ Low __________ Evolution type number ________________________
Very Low _______ Very Low ______ Existing state (type) __________________________
Annual streambank erosion rate ____________ Potential state (type) __________________________
Length of banks studied _______________ Competence calculation:
Tons/year __________________________ Critical dimensionless shear stress _______________
Curve used _________________________ Largest particle on bar ________________________
Bankfull depth (existing) ______________________
Dimensionless Sediment rating curve: Bankfull depth required _______________________
Normal ____________________________ Slope (existing) _____________________________
Above normal _______________________ Slope required ______________________________





CONCLUSION

Although this stability prediction method may seem onerous, it has been applied in watershed management and for
geomorphological assessments for many years. The author has trained hundreds of individuals in this procedure that
have collected both level III and IV data to help improve and validate the prediction relations. An additional
application of this approach has been used for restoration proposals, where an understanding of the cause,
consequence and correction of the problem involves an inventory that isolates the processes associated with stream
channel instability. The recent requirement to establish TMDL's for clean sediment involves an understanding of
natural rates, natural variability and documenting departure conditions leading to adverse consequence of instability
and corresponding disproportionate sediment yields. If we understand the various processes of change, prevention
through good management and application of mitigation measures can be appropriately applied to the problem.
Continued field measurements are the key to improving upon these procedures and add to the collective
understanding of these complex and valuable river systems.

"A consistent chronicle of field observations and collected data is essential to the practice of hydrology. As with
opportunities, good data are available only once.". Luna B. Leopold.


Acknowledgements: The author wishes to thank Dr. Luna B. Leopold, Dr. Richard Hey, and Dr. Charles Troendle
for their review and to Hilton Lee Silvey and J im Nankervis for graphics and statistical assistance, and to J osh Kurz
for assistance with the field forms, review and computer support.


REFERENCES

Andrews, E.D., 1984. Bed-material Entrainment and Hydraulic Geometry of Gravel-Bed Rivers in
Colorado. Geol. Soc. of Am. Bull., 95, 371-378.
Andrews, E.D. and Nankervis, J ames M. 1995. Effective Discharge and the Design of Channel
Maintenance Flows for Gravel-Bed Rivers. 151-164, Natural and Anthropogenic Influences in
Fluvial Geomorphology, Geophysical Monograph 89, AGU, Wash. D.C.
Davis, W.M., 1902. Base Level, Grade, and Peneplain: Geophysical Essays, XVIII, pp. 381-412, Ginn,
Boston.
Hack, J .T. 1960. Interpretations of Erosional Topography in Humid Temperate Regions. Am. J our. Sci.
Vol.258A, 80-97.
Hey, R. D.,and Thorne, C.R. 1986. Stable Channels with Mobile Gravel Beds, J ournal of Hydraulic
Engineering, 112, 671-689.
Hey, R.D. 1997. Stable River Morphology. Applied Fluvial Geomorphology for River Engineering and
Management. Ed.,.C.T. Thorne, R.D. Hey and M.D. Newson. P223-236, J ohn Wiley and Sons.
Langbein, W. B. 1963. A Theory for River Channel Adjustment. Soc. Amer,. Civil Engineers Trans.
Leven,Richard, 1977. Suspended sediment rating curves from USGS and USDA Forest Service-Redwood
Creek. Six Rivers N. F., Arcata, Calif. In: EPA, 1980, An Approach to Water Resources
Evaluation of Non-point Silvicultural Sources.p, VI.40, EPA -600/8-80-012, Athens, Ga.
Leopold, L.B., amd Maddock T.J r., 1953. The Hydraulic Geometry of Stream Channels and some
physiographic implications. U.S. Geol. Survey professional. Paper 252.
Leopold, L.B., Wolman. R.G., and J .G. Miller, 1964. Fluvial Processes in Geomorphology. W.H.
Freeman, San Francisco.
Leopold, L.B. 1994. A View of The River, Harvard University Press, Cambridge, Mass. 292 pp.
Leopold, L.B.. 1992. Sediment Size the Determines Channel Morphology, 297-307, In: Dynamics of
Gravel-Bed Rivers. Ed. P.Billi, R.D. Hey, C.R. Thorne & P. Tacconi, J ohn Wiley and Sons Ltd.
Mackin, J .H. 1948. Concept of the Graded River. Geol. Soc. Am. Bull., Vol.59, p. 463-512.
Pfankuch, Dale J . 1975. Stream Reach Inventory and Channel Stability Evaluation. USDA Forest Service,
R1-75-002. Govt. Printing Office, #696-260/200, Wash. D.C. 26 pp.
Rosgen, David L. 1980. Total Potential Sediment, In: An Approach to Water Resources
Evaluation of Non-point Silvicultural Sources. Chapter 6, EPA -600/8-80-012, Athens, Ga.
Rosgen, David L. 1994. A Classification of Natural Rivers, Catena, Vol 22, 169-199, Elsevier Science,
B.C. Amsterdam.
Rosgen, David L. 1996. Applied River Morphology. Wildland Hydrology Books, Pagosa Springs,
Colorado,
Rosgen, David L. 1998. The Reference Reach-A Blueprint for Natural Channel Design. In: Proceedings
of Amer. Soc. Civil Engineers, Restoration of Wetlands and Rivers, Denver, Colorado.
Rosgen, David L. 1999. Development of a River Stability Index for Clean Sediment TMDL's. In:
Proceedings of Wildland Hydrology, Ed. D.S. Olsen and J .P. Potyondy, AWRA, Bozeman,
Montana, p. 25-36.
Rosgen, David L. 2001. In Press. A Practical Method of Computing Streambank Erosion Rate, 7
th
Federal Inter-
agency Sediment Conference, March 24-29, Reno, Nevada.
Silvey, H.Lee, and Rosgen, D.L., 1998. The Reference Reach Field Book. Wildland Hydrology Books, Pagosa
Springs, Colo. 81147, 210 pp.
Simon Andrew, and Hupp, C.R., 1986. Channel Evolution in Modified Tennessee Streams. In: 4
th
Federal
Inter-Agency Sedimentation Conference 2:71-82.
Simon, Andrew, 1989. The Discharge of Sediment in Channelized Alluvial Streams. Water Resources
Bull., AWRA vol. 25,No. 6, 1177-1188.
Strahler, A.N. 1957. Quantitative Analysis of Watershed Geomorphology. Am. Geophys. Union Trans.
V. 38, pp 913-920.
Troendle, C.A., Rosgen, D.L., Ryan, S., Porth L., and Nankervis,J . 2001. In Press. 7
th
Federal Inter-Agency
Sediment conference, March 24-29, Reno, Nev.
Williams, G.P. and Rosgen, D.L. 1989. Measured Total Sediment Loads (Suspended and Bedloads) for 93
United States Streams. US Geological Survey Open File Report 89-67, Denver, Colorado.128 pp.