Quantification Environmental Impacts or Urban Green Roofs

Quantification of the Environmental Impacts
of Urban Green Roofs
A thesis submitted to The University of Manchester for the degree of Doctor

of Philosophy in the Faculty of Humanities
2013
Andrew Francis Speak
School of Environment and Development

CONTENTS
CONTENTS ii
LIST OF FIGURES viii
LIST OF TABLES xii
ABSTRACT xiii
DECLARATION xiv
COPYRIGHT STATEMENT xv
ACKNOWLEDGEMENTS xvi
GLOSSARY OF ABBREVIATIONS xvii
CHAPTER 1 INTRODUCTION 1
1.1 Research rationale 1
1.2 Involvement of Manchester City Council 3
1.3 Aims and Objectives 4
1.4 Methodological Framework 6
1.5 Study Site – Manchester 7
1.6 Significance of Research 9
1.7 Structure of Thesis 10
CHAPTER 2 CURRENT KNOWLEDGE ON CLIMATE 14

CHANGE AND URBAN CLIMATES
2.1 Climate Change 14
2.1.1 Definition and broad-scale concepts 14
2.1.2 Climate change in the UK 17
2.2 Urban Climates 20
2.2.1 General overview 20
2.2.2 Urban Heat Islands 21
2.2.3 Precipitation 24
2.2.4 Air Pollution 25
2.2.5 Water Pollution 27
2.3 Adaptation 28
2.3.1 Introduction to adaptation themes 28
ii
2.3.2 Adaptation with greenspace 32
2.4 Conclusion 36
CHAPTER 3 CURRENT KNOWLEDGE ON GREEN ROOFS 37

3.1 Green roofs – an introduction 37
3.2 Green roofs as adaptation strategies for high temperatures 41
3.3 Green roofs as SUDS 49
3.4 Green roofs as remediators of air pollution 52
3.5 Green roofs and runoff water quality 54
3.6 Other green roof benefits 55
3.6.1 Biodiversity 55
3.6.2 Carbon sequestration 56
3.6.3 Noise 59
3.6.4 Electromagnetic Radiation 57
3.6.5 Aesthetics 57
3.6.6 Engineering benefits 57
3.7 Conclusions 57
CHAPTER 4 METHODOLOGY 59
4.1 Overview 59
4.2 Study site 61
4.3 Air pollution remediation 66
4.3.1 Experiment overview 66
4.3.2 Site selection 67
4.3.3 Magnetic techniques 69
4.3.4 Leaf area estimation 70
4.3.5 Scanning Electron Microscopy (SEM) 72
4.3.6 Upscaling 72
4.4 Urban Heat Island impacts 73
4.4.2 Monitoring stands 74
iii
4.5 Runoff quantity 76
4.5.2 Monitoring 76
4.5.3 Soil analysis 80
4.6 Runoff quality 81
4.6.2 Water analysis 81
4.6.3 Substrate analysis 82
4.7 Analysis 83
CHAPTER 5 URBAN PARTICULATE POLLUTION 85
REDUCTION BY FOUR SPECIES OF GREEN

ROOF VEGETATION IN A UK CITY
5.1 Abstract 86
5.2 Introduction 87
5.3 Methodology 89
5.3.1 Site Description 89
5.3.2 Species Selection and Sampling 92
5.3.3 Magnetic Measurements 94
5.3.4 Elemental Analysis 94
5.3.5 Scanning Electron Microscopy (SEM) 95
5.3.6 Statistical analyses 95
5.3.7 PM10 removal quantification 96
5.4 Results 97
5.4.1 Site comparisons 97
5.4.2 Species comparisons 101
5.5 Discussion 109
5.5.1 Spatial differences in PM10 capture 109
5.5.2 Species differences in PM10 capture 111
5.5.3 Implications of wider green roof implementation 112
5.6 Conclusion 114
iv
CHAPTER 6 REDUCTION OF THE URBAN COOLING
EFFECTS OF AN INTENSIVE GREEN ROOF 115
DUE TO VEGETATION DAMAGE
6.1 Abstract 116
6.2 Introduction 117
6.3 Methodology 120
6.3.2 Monitoring 123
6.3.3 Data analysis 124
6.4 Results 125
6.4.1 Vegetation dynamics 125
6.4.2 Healthy green roof 125
6.4.3 Damaged green roof 135
6.5 Discussion 136
6.5.1 Localised cooling effect 136
6.5.2 Effects of green roof damage 138
6.5.3 Local climate modification 139
6.6 Conclusion 142
CHAPTER 7 RAINWATER RUNOFF RETENTION ON AN 143
AGED INTENSIVE GREEN ROOF

7.1 Abstract 144
7.3 Methodology 148
7.3.2 Monitoring 152
7.3.3 Data Analysis 155
7.4 Results 158
7.4.1 Climate and Weather 158
7.4.2 Soil properties 159
7.4.3 Runoff initiation 160
v
7.4.4 Inter-roof comparison 163
7.4.5 Storm events 167
7.5 Discussion 169
7.5.1 Retention efficiency 169
7.5.2 Implications for Manchester city centre 172
7.6 Conclusion 173
CHAPTER 8 METAL AND NUTRIENT DYNAMICS ON AN 175

8.1 Abstract 176
8.3 Methodology 182
8.3.2 Sample collection and analysis 184
8.3.3 Data analysis 190
8.4 Results 191
8.4.1 Substrate properties 195
8.4.2 Water chemistry 196
8.5 Discussion 201
8.5.1 General runoff characteristics 201
8.5.2 Elevated metal concentrations 202
8.5.3 Implication of findings 207
8.6 Conclusion 208
CHAPTER 9 CONCLUSION 209

9.2 Summary of findings 209
9.3 Significance of findings for green roof research 211
9.3.1 Significance for climate change adaptation 218
9.3.2 Recommendations for Manchester City Council 219
9.4 Strengths and limitations 220
vi
9.5 Research recommendations 222
REFERENCES 225
APPENDIX I 249
APPENDIX II 250
APPENDIX III 253
APPENDIX IV 257
WORD COUNT: 67,448
vii
LIST OF FIGURES
Fig. 1.1 – Location of Greater Manchester (Author created) 7

Fig. 1.2 – Location of The Corridor in relation to Manchester city centre 8
(source: Corridor Manchester 2012)
Fig. 1.3 – Outline of the thesis structure (Author created) 12
Fig. 2.1 – Change in global surface temperatures relative to 1951 – 1980 14
climate average (Source: NASA 2011)
Fig. 2.2 – European summer temperatures for 1500 – 2010. The top panel 15
shows the statistical frequency distribution of summer land-temperature
anomalies (relative to the 1970 – 1999 period) for the 1500 – 2010
period in Europe. The five coldest and warmest summers are
highlighted. The lower panel shows a running decadal frequency of
extreme summers, defined as those with a temperature exceeding the
95th percentile of the 1500 – 2010 distribution. (Source: Barriopedro et
al. 2011)
Fig. 2.3 – 10, 50 and 90% probability levels of changes to the mean daily 18
minimum temperature in winter (top) and summer (bottom), by the
2080s, under the Medium emissions scenario. (Source: Murphy et al.
2009)
Fig 2.4 – 10, 50 and 90% probability levels of changes to annual (top), winter 19
(middle) and summer (bottom) mean precipitation, by the 2080s, under
the Medium emissions scenario. (Source: Murphy et al. 2009)
Fig. 2.5 – A generalised Urban Heat Island. (Source: Erell et al. 2011) 21
Fig. 2.6 – Regional Climate Model climate change over the UK (2050s 22
minus 1980s, SRES A1b) due to (left) greenhouse gases, (middle)
greenhouse gases and increased energy use in urban areas, and
(right) increased energy use only. (Source: McCarthy et al. 2010)
Fig. 2.7 - Maximum surface temperature for the 98th percentile summer day 34
in town centres, with current form and when 10% green cover is
added or removed. Dashed line shows the temperature for the
1961-1990 current form case (source: Gill 2007
Fig 3.1 – Intensive green roof on the Precinct bridge in Manchester, UK 38
(Author image)
Fig. 3.2 – Extensive green roof in Manchester, UK (Author image) 39
Fig. 3.3 – Schematic diagram of the layered structure of a typical green roof 40
system Source: (American Wick Drain Corp, 2009)
Fig. 3.4 – Schematic diagram of the different processes occurring during 42
the day (top) and at night (bottom) which characterise the thermal
performance of a conventional roof (left) and a green roof (right).
(Author created)
Fig 3.5 – Comparison of the energetic exchanges of a dry or wet green roof 43
with a traditional roof, starting with 100 incident solar radiation units
– summer session (Source: Lazzarin et al. 2005)
viii
Fig. 3.6 – Cumulative runoff from a green roof and non-green roof in 50
Belgium during the 24 hour period of a 14.6 mm rain shower
(Source: Mentens et al. 2006)
Fig. 3.7 – Annual runoff as a percentage of total annual rainfall for intensive 51
green roofs (‘int’, n=11), extensive green roofs (‘ext’, n=121), gravel-
covered roofs (‘gravel’, n=8) and non-greened roofs (‘trad’, n=5).
(Source: Mentens et al. 2006
Fig. 4.1 – The locations of the study roofs (Author created) 62
Fig. 4.2 – 3-dimensional view of the Oxford Road Corridor with roof 63
locations (Source: Landmap)
Fig. 4.3 – Scale drawing of the Precinct Building roof (UoM 2007 - larger 64
version in Appendix 1)
Fig. 4.4 – Roof plan of the precinct (Author created) 65
Fig. 4.5 – Trays of S. album (left), A. stolonifera (right) and F. rubra (far 67
right, front) ready to be transplanted on to the study roofs (Author
image)
Fig. 4.6 – Panoramic view to the south west of Roof 1 (Author image) 67
Fig. 4.7 – Aerial photograph of Roof 1 (outlined) with directional wind rose 68
for the period 21/06/2011 – 04/08/2011 (Source: Googlemaps and
author created)
Fig. 4.8 – Firs Botanical Gardens (outlined) within the local setting (Source: 69
Googlemaps)
Fig. 4.9 – Simplification of the sedum leaves for 3-dimensional surface area 71
calculation (Author created)
Fig. 4.10 – A selected area of Manchester city centre and the Corridor for 73
analysis of flat roof area using ArcMap software (Source: Landmap)
Fig. 4.11 – Tinytag temperature sensor to scale (Source: Tinytag 2013) 75
Fig. 4.12 – Radiation shield and stand (Author image) 76
Fig 4.13 – Schematic diagram for the measuring of runoff (Author created) 77
Fig. 4.14 – V-notch weir pot and pressure transducer (Author image) 78
Fig. 4.15 – Regression plot of rainfall and runoff for the glass atrium roof, 79
with 1:1 line included
Fig. 5.1 – Map showing locations of the study sites with Firs control site to 91
the south of the city centre and the locations of Roofs 1 and 2 next
to Oxford Road. The shaded overlay, highlighting the city centre
and major roads, indicates the current extent of the Manchester
AQMA (Source: Defra, 2011)
Fig. 5.2 – Average ambient PM10 levels between 11am and 1pm at each of 97
the three sites on selected days throughout the study and daily
average at the Piccadilly monitoring station
Fig. 5.3 – Filter SIRM from 1 hour pumped air samples taken on 27 July 98
2011, showing the lower levels of magnetisable particulate matter at
the Firs control site.
ix
Fig. 5.4 – Leaf area-normalised SIRM values of the study green roof 99
species; (a) A. stolonifera, (b) F. rubra, (c) P. lanceolata, and (d) S.
album at the three study sites (note the differing y-axis scales).
Upward trends over the study period are noticeable, with Roof 1
capturing more PM10 than Roof 2
Fig. 5.5 – SIRM enrichment ratio of the two roadside sites to the less 100
polluted control site showing the higher enrichment on Roof 1 than
Roof 2
Fig. 5.6 – Leaf area-normalised SIRM values of the study species at each 101
of the three sites; (a) Control, (b) Roof 1, and (c) Roof 2 (note the
differing y-axis scales). Upward trends over the study period are
noticeable with A. stolonifera displaying the highest particle capture
Fig. 5.7 – SEM images of leaf surfaces of the four study species (X800 – 103
1600 mag.) with higher magnification images of selected spherules
(X12800 – 25600 mag.)
Fig. 5.8 – EDX for P. lanceolata abaxial leaf surface indicating distribution 104
of iron and silica within a particle cluster
Fig. 5.9 – Daily rainfall totals and leaf area-normalised SIRM values of the 108
‘permanent’ green roof species and the roof-height tree canopy
which show a clear increase in the drier second half of the study
period
Fig. 6.1 – Map to show the location of Precinct study site and the monitor 121
locations on the roof
Fig. 6.2 – Boxplots of the monthly air temperatures at 300 mm above the 127
surface recorded on the three study roofs. G1 is the undamaged
green roof, G2 the damaged green roof and BARE is the adjacent
conventional roof surface.
Fig. 6.3 – Temperature difference (δT) of the monthly medians (G1 = Bare 128
roof-G1, G2 = Bare roof-G2)
Fig. 6.4 – Average hourly temperatures for the period 28/03/2011 to 129
31/08/2012
Fig. 6.5 – Diurnal trends of the temperature difference (δT) between the 131
green roof areas and the conventional roof area for each season
Fig. 6.6 – Temperature profile for the three roof areas during one of the 3- 133
day 95th percentile exceedance periods (20 – 22 April 2011, with a
day either side)
Fig. 6.7 - Temperature profile for the three roof areas during one of the 3- 133
day 95th percentile exceedance periods (28 – 30 September 2011,
with a day either side)
Fig. 6.8 – Thermal image of the Bare roof, with green roof area G1 behind, 134
taken at 14:00 on 26/07/2011. Image is inclusive of any reflected
radiation incurred by non-perpendicular angle of photograph
Fig. 6.9 – Thermal image of G2 and concrete slab path taken at 14:00 on 134
26/07/2011
Fig. 6.10 – Thermal image of G2 and concrete slab path taken at 14:00 on 134
14/08/2012
Fig. 7.1 – Location of the study roof 149
x
Fig. 7.2 – Cross section of the green roof (left) and bare roof (right) showing 150
layer depths
Fig. 7.3 – Plan of the Precinct roof showing drainage catchment areas for 153
the bare and green roofs
Fig. 7.4 – Rainfall characteristics for the 69 analysed rain events with return 156
period estimates of 1 to 5 year events for Manchester from the
Flood Estimation Handbook cd-rom
Fig. 7.5 – Monthly rainfall expressed as a percentage of the climate 157
average for Manchester (1981 – 2010)
Fig. 7.6 – Rainfall and runoff for two events (a) Dec 2011 (b) Sept 2012 161
Fig. 7.7 – Regression plot of rainfall and runoff depth for (a) green roof and 162
(b) bare roof, with 1:1 lines included
Fig. 7.8 – Boxplot to show mean average runoff retention of all rainfall 163
events for both roofs
Fig. 7.9 – Mean runoff retention comparisons between the green roof and 164
bare roof for (a) season, (b) rainfall depth, (c) rainfall duration, (d)
peak rainfall intensity, (e) ADWP and (f) precipitation anomaly
Fig. 7.10 – Regression plots of runoff retention against (a) ADWP and (b) 165
rainfall duration, for the green roof
Fig. 7.11 – Regression tree for runoff retention percentage on the green 166
roof. ‘Yes’ and ‘no’ splits are assigned to left and right respectively
Fig. 8.1 – Location of the study roof 182
Fig. 8.2 – Cross section of the green roof (left) and bare roof (right) showing 183
layer depths
Fig. 8.3 – Plan of the Precinct roof showing drainage catchment areas for 187
the bare and green roofs
Fig. 8.4 – Monthly rainfall expressed as a percentage of the climate 191
average for Manchester (1981 – 2010)
Fig. 8.5 – Vertical soil profiles for Cu, Pb and Cr in the intensive green roof 193
Fig. 8.6 – Spatial distribution of Pb, Cd, Cu and Zn using ordinary kriging 196
with 55 data points (Source: BGS G-BASE project)
Fig. 8.7 – Boxplots to show water quality parameters for Bare roof runoff (B) 198
Fig. 8.8 – Boxplot of Pb concentrations in the runoff from the atrium (A) 204
Fig. 8.9 – A model of the atmospheric inputs, stores and outputs of Pb for 205
(a) the entire life of the green roof and (b) the annual flux based on
present-day data
Fig. 9.1 – The multiple benefits afforded by the three main compartments of 212
a green roof (Author created)
Fig. 9.2 – 36th highest daily value of PM10 in 2010 (Adapted from European 213
Environment Agency (EEA 2013))
xi
LIST OF TABLES
Table 3.1 – Summary of findings from studies on the cooling effect of green 42
roofs worldwide. (Köppen classification from Kottek et al. 2006)
Table 4.1 – Regression coefficients for the linear relationship between dry 71
weight and leaf area for the study species
Table 5.1 – Percentage increase in SIRM values and iron concentration 105
over the duration of the study, showing clear increases in SIRM but
variable patterns for iron concentration
Table 5.2 – Correlation matrix for the SIRM values of the study species (A 105
= A. stolonifera, F = F. rubra, P = P. lanceolata, S = S. album, Pl =
P. acerifolia, Sy = S. novae-angliae, H = H. inodorum) at each site
(C = Control, 1 = Roof 1, 2 = Roof 2) showing wide variability. Bold
values indicate significance at the 0.05 level.
Table 5.3 – Correlations of area-normalised leaf SIRM values with rolling 107
average ambient PM10 concentration (A=A. stolonifera, F=F. rubra,
P= P. lanceolata, S=S. album, Pl=P. acerifolia, Sy=S. novae-
angliae, H=H. inodorum)
Table 5.4. Estimates of quantities of PM10 captured by the study species 113
with ± 5% sensitivity to winter reduced rate. Higher removal rates
were achieved by the grasses A. Stolonifera and F. rubra
Table 6.1 – Kruskal-Wallis χ2 values (* = significant at p<0.05, 126

** = significant at p<0.01)
Table 7.1 – Characteristics of the dominant green roof vegetation 152

Table 7.2 – Selected soil characteristics and comparison to the FLL 159
guideline values for an intensive green roof
Table 7.3 – selected characteristics of the 6 storm events greater than 1 168
year return period
Table 8.1 – Characteristics of the rainfall-runoff events sampled 186

Table 8.2 - Selected soil characteristics and comparison to the FLL 192
guideline values for an intensive green roof
Table 8.3 – Elemental composition of the green roof substrate and bare 194
roof dusts for selected elements, including the ambient background
levels for England (Barraclough, 2007) and CLEA Soil Guideline
Values (SGVs), where available, for comparison
Table 8.4 – Output flux expressed as a percentage of input flux for the 200
three rainfall events where flux calculation was possible. Bold
results indicate >100 %
xii
ABSTRACT
Urban populations worldwide are expanding rapidly and consequently a large

number of people are becoming exposed to hazards inherent in cites.
Phenomena such as the urban heat island can exacerbate the effects of
heatwaves, and land surface sealing can lead to flash flooding. Cities are also
the sites of enhanced air and water pollution from non-point sources such as
concentrated motor vehicle use. Climate change predictions for the UK include
increased winter precipitation and an increase in frequency of summer
heatwaves. This will put further pressure on urban residents and infrastructure.
Roof greening can be used within climate change adaptation schemes because
green roofs have a range of environmental benefits which can help urban
infrastructure become more sustainable.
This thesis empirically quantifies several of these benefits, and the processes
influencing them, by monitoring real green roofs in Manchester. A number of
novel discoveries were made. Green roofs act as passive filters of airborne
particulate matter. 0.21 tonnes of PM10 (2.3% of the inputs) could be removed
from Manchester city centre in a maximum extensive green roof scenario.
Species and site differences in particle capture were exhibited and related to
morphology and proximity to sources respectively. An intensive green roof was
able to lower the monthly median overlying air temperature at 300 mm by up to
1.06 oC. A combination of drought and mismanagement caused damage to the
vegetation on one of the green roofs, with a subsequent reduction in the cooling
effect. Daytime air temperatures were higher than over an adjacent bare roof for
a larger proportion of the day than over the undamaged roof, and lower cooling
was observed at night.
A site-specific methodology was devised to monitor the rainwater runoff from an

intensive green roof and an adjacent bare roof. Average runoff retention of 65.7%
was observed on the green roof, compared to 33.6% on the bare roof. Season
and rainfall amount had significant impacts on retention, however, many other
explanatory variables such as Antecedent Dry Weather Period (ADWP) and
peak rainfall intensity had no demonstrable, significant impact. Intensive roof
construction on 10% of the rooftops in Manchester city centre would increase
annual rainfall retention by 2.3%. The runoff was characterised with regards to
heavy metals and nutrients. Nutrient levels were found to be not a significant
problem for water quality, however, Environmental Quality Standards (EQS)
values for protection of freshwater were exceeded for concentrations of Cu, Pb
and Zn. High metal concentrations within the sediments may be acting as
sources of pollution, particularly in the case of Pb. The age of the green roof
means that past atmospheric deposition of Pb could be contributing to the runoff
quality.
The multi-benefit aspect of green roofs is discussed in the light of the results of
this thesis and recommendations made for policy makers and the green roof
construction industry.
xiii
DECLARATION
No portion of the work referred to in this thesis has been submitted in support
of an application for another degree or qualification of this or any other
university or institute of learning.
Andrew Speak
xiv
COPYRIGHT STATEMENT
(i) The author of this thesis (including any appendices) owns certain
copyright or related rights in it (the ‘’Copyright’’) and he has given The
University of Manchester certain rights to use such Copyright, including for
administrative purposes. Copies (by any process) either in full, or of extracts,
may be made only in accordance with instructions given by the author and
lodged in the John Rylands University Library of Manchester. Details may be
obtained from the Librarian. This page must form part of any such copies
made. Further copies (by any process) of copies made in accordance with
such instructions may not be made without the permission (in writing) of the
author.
(ii) The ownership of any intellectual property rights which may be described
in this thesis is vested in The University of Manchester, subject to any prior
arrangement to the contrary, and in contractual agreement with the CASE
partner, Manchester City Council, and may not be made available for use by
third parties without the written permission of the University, which will
prescribe the terms and conditions of any such agreement.
(iii) Further information on the conditions under which disclosures and

exploitation may take place is available from the Head of School.
xv
ACKNOWLEDGEMENTS
Thank you to my supervisors, Dr Sarah Lindley, Dr Claire Smith and Dr

James Rothwell for providing support, knowledge, encouragement, and
advice throughout the entire PhD process.
This work was funded by NERC with CASE contributions from Manchester
City Council. Thanks must go to the various people who provided equipment,
without whom the study could not have gone ahead. The use of the SEM
and ICP-MS was provided, with assistance from Richard Cutting and Paul
Lythgoe, by the School of Earth, Atmospheric and Environmental Sciences at
the University of Manchester. Thanks to Simon Hutchinson and the
University of Salford School of Environment and Life Sciences for use of their
Molspin equipment. The use of the IR camera was provided by Andrew
Pinkerton, from the School of Mechanical, Aerospace and Civil Engineering
at the University of Manchester. Advice and assistance was gratefully
received from John Moore, Gregory Lane-Serff, Clive Agnew and Gordon
McFiggans at the University of Manchester, Virginia Stovin at the University
of Sheffield, and Barbara Maher at Lancaster University. Thanks also to The
UK Sedum Society for advice and cuttings of Sedum album. The botanical
survey of the roof was carried out by Cameron Crook.
Thank you to the music I streamed while writing the various parts of the
thesis for making the writing process more bearable.
Thank you to all my colleagues in Geography and the rest of the School of
Environment and Development for inspiring me, listening to me, advising me,
and taking me to the pub.
Thank you to Fergus for his love and support and being there at the start,
and to Liam for his love and support and being there at the end.
xvi
GLOSSARY OF ABBREVIATIONS
ADWP Antecedent Dry Weather Period

AQMA Air Quality Management Area
ARCC Adaption and Resilience in a Changing Climate
ASE Average Standard Error
AURN Automatic Urban and Rural Network
BGS British Geological Survey
BOD Biochemical Oxygen Demand
BREEAM Building Research Establishment Environmental Assessment Method
BVOC Biogenic Volatile Organic Compound
CART Classification and Regression Tree
CASE Collaborative Awards in Science and Engineering
CLEA Contaminated Land Exposure Assessment
DOC Dissolved Organic Carbon
EDX Energy Dispersive X-ray
EQS Environmental Quality Standard
FLL Forschungsgesellschaft Landschaftsentwicklung Landschaftsbau e
GI Green Infrastructure
GIS Geographical Information Systems
GMCCS Greater Manchester Climate Change Strategy
HEI Higher Education Institute
ICP-MS Inductively Coupled Plasma – Mass Spectrometry
ICP-OES Inductively Coupled Plasma – Optical Emission Spectrometry
IPCC Intergovernmental Panel on Climate Change
LAI Leaf Area Index
LMM Linear Mixed Model
LWA Light Weight Aggregate
MACF Manchester: A Certain Future
MCC Manchester City Council
MTC Manchester Technology Centre
NERC Natural Environment Research Council
PAH Polyaromatic Hydrocarbon
PM Particulate Matter
RDS Road Deposited Sediment
SEM Scanning Electron Microscopy
SGV Soil Guideline Value
xvii
SIRM Saturated Isothermal Remanent Magnetisation
SUDS Sustainable Urban Drainage Systems
SVF Sky View Factor
TSS Total Suspended Sediment
UCL Urban Canopy Layer
UDP Urban Diffuse Pollution
UFORE Urban Forest Effects
UHI Urban Heat Island
UKCIP United Kingdom Climate Impacts Programme
UMT Urban Morphology Type
VOC Volatile Organic Compound
WFD Water Framework Directive
WHO World Health Organisation
XRF X-ray Fluorescence
xviii
CHAPTER 1 INTRODUCTION
1.1 Research rationale
Urban populations worldwide are steadily rising. Currently 50% of the world’s
population is classed as urban and this is predicted to become 70% by 2050
(UN 2009). There are a number of pressures exerted on the urban system
that can act independently or in conjunction with each other to produce
impacts with potentially serious consequences for urban residents (Gill et al.
2004). For instance, urbanisation increases the amount of surface coverage
of materials with a high thermal capacity such as concrete and asphalt (Taha
2004), which is one of the contributing factors to the Urban Heat Island (UHI)
phenomenon (Oke 1982). It is widely recognised that built areas can be
warmer than surrounding rural areas, up to 5 - 6°C have been reported (Oke
1982) and when combined with high summer temperatures, human
physiological stress can occur. An estimated 70,000 heat-related fatalities
have been linked to the 2003 European heatwaves (Robine et al. 2008), of
which over 2,000 were in England and Wales (Larsen 2006).
A further pressure on the urban system arises from the impervious nature of
the built environment which can lead to pluvial (surface water) flooding during
heavy rainfall events. The hydrochemical quality of this runoff can also
become degraded by highly persistent and toxic heavy metals (Pizzol et al.
2011), or nutrients which can lead to eutrophication in receiving water bodies
that are already nutrient-rich (Ellis and Mitchell 2006). The processes behind
the inputs of these substances to the urban environment are themselves a
direct result of urbanisation, such as airborne dusts from high volumes of
road traffic (Robertson et al. 2003), and wet and dry deposited air pollution
from anthropogenic emissions (Fowler et al. 2007). There is an associated
deterioration of air quality in cities from these emissions. Particulate Matter
smaller than 10 microns (PM10), ozone (O3) and oxides of nitrogen (NOx) are
elevated in cities due to high vehicular emissions. While there are some
signs of abatement due to UK vehicle usage levelling off in the past three
years (DfT 2013), concentrations can still exceed air quality standards and do
not always seem to reduce in line with emissions estimates (NAEI 2012).
1
Some of the predicted effects of climate change will be felt most acutely by
the inhabitants of cities (Wilby 2007). It is estimated that there will be an
increase in the frequency of summer heatwaves (Murphy et al. 2009), which
when combined with the UHI effect, means that cities will become quite
uncomfortable places to live for some parts of the year. Changes to
precipitation are also expected, with increases of up to 33% by the 2080s
under a medium emissions scenario, with western UK being hardest hit
(Murphy et al. 2009).
These pressures and their predicted increase in frequency /severity from

both climate change and increasing urban populations means there is an
urgent need for adaptation strategies. One adaptation technique that seeks
to reduce the extent and magnitude of the UHI is to increase the amount of
urban vegetative cover (Hardin and Jensen 2007). This has numerous
additional potential benefits such as the passive filtration effect of the
vegetation and subsequent improvements in air quality (Powe and Willis
2004). A problem with extensive greening of cities is that there is competition
for expensive building space and the fundamental layout of cities is unlikely
to change (Wong et al. 2003). Green roofs overcome this barrier by utilising
the space afforded by rooftops which can account for 35% of the land cover
(MacMillan 2004).
Green roofs broadly consist of a drainage layer underneath a substrate layer

which supports vegetation. They offer a number of benefits such as localised
cooling, stormwater retention and air pollution reduction, which will be
detailed in Chapter 3. As green roofs are a relatively new technology there is
a need, particularly in the UK, to quantify and evaluate the benefits they
afford. This will realistically allow their usefulness within climate change
adaptation schemes to be demonstrated. There is a need to investigate how
well green roofs perform in a range of climates and urban morphologies. The
impacts of climate change vary widely in type and severity depending upon
location so green roofs can be tailored for a particular location in its current
and projected states. A large amount of green roof research has been
focussed on Germany (Köhler 2005; Mentens et al. 2006) so it is necessary
to understand how green roofs perform elsewhere in the world, for instance
2
in the tropics or in cooler maritime climates which are predicted to get
warmer in a future climate. There is also a need to bridge the gap between
academia and practitioners so that green roof policies can be developed and
so that the green roof industry can respond to scientific research in a way
that allows the best technology to be developed.
1.2 Involvement of Manchester City Council
This work was funded through a Natural Environment Research Council

(NERC) Co-operative Awards in Science and Engineering (CASE)
studentship (Grant reference: NE/G012393/1). The CASE scheme aims to
promote partnerships between Higher Education Institutes (HEIs) and other
bodies such as private industry or the public sector. Manchester City Council
(MCC), specifically the Green City team within the Environmental Services
Department, are the CASE partner for this thesis. The Green City team
oversee the delivery of Manchester’s climate change action plan –
Manchester: A Certain Future (MACF) – which is part of the Greater
Manchester Climate Change Strategy (GMCCS). The strategy sets out plans
to reduce carbon emissions by 48% by 2020 and provides an integrated
approach to tackling climate change (MCC 2011). One of the four headline
visions of GMCCS is to be prepared for and actively adapt to a rapidly
changing climate, and one of the themes of MACF is adaptation using green
and blue infrastructure.
The results of the thesis will feed into the evidence base for the Green
Infrastructure (GI) Strategy that MCC is developing. The GI strategy
provides a framework that brings together disparate GI strategies and
research (such as the Biodiversity Action Plan and Tree Audit data) into a
single place. The strategy establishes the economic, social and
environmental business case for investing in new, or improving the quality of
existing, GI across the city, highlights the multifunctional benefits that GI can
offer and includes examples of how GI can be used to support the specific
aims in each of the City's Strategic Regeneration Framework Areas
(Personal Communication from Neil Jones, Environmental Strategy Officer,
MCC, 16 May 2013).
3
In conjunction with the business case, a 15 year action plan and a
Geographical Information System (GIS) browser has been developed which
will identify future GI projects and research opportunities, as well as map the
ongoing evolution of the City's GI. The research will feed into these
workstreams and can be used to identify further areas of research in the
future. The initial research objectives of the CASE partnership were very
extensive and broad (see Appendix I) and the thesis necessarily evolved by
determining which aspects allowed the biggest scope for novelty and were
also feasible given budget and time constraints.
A green roof feasibility study was carried out by MCC which proposed a work
programme for Greater Manchester which, upon implementation, would
result in the increased installation of green roofs across the region whilst
providing a guidance document for the installations (MCC 2009).
Stakeholders for the study included the University of Manchester,
Manchester Metropolitan University, a commercial property company named
Bruntwood and Red Rose Forest, a community forest partnership initiative.
Five buildings were chosen to act as green roof demonstrator projects,
however, due to a number of setbacks which included buildings failing to
meet structural standards for being able to support a green roof, and the
global financial crisis, only one of the exemplar roofs was constructed. This
was installed on the Whitworth Art Gallery in 2010 and consisted of an
extensive-type roof planted with sedum. A short video of the construction
can be viewed at http://vimeo.com/14006655.
The Green City team are at the heart of a campaign to help Manchester
become Britain’s greenest city in terms of air quality, biodiversity, climate
change, energy and the built environment. Green roofs can play an
important role in each of these areas so this thesis will provide valuable
information for this campaign, whilst furthering knowledge within urban
environmental science under the NERC umbrella.
4
1.3 Aims and Objectives
This thesis seeks to quantify the environmental impacts of green roofs using
a multivariate, empirical monitoring approach. The thesis is directed by four
main aims which are related to the four environmental impacts of interest.
The first aim is to assess what quantities of air pollution, specifically PM10,
can be removed from the urban atmosphere by green roof vegetation. The
second aim is to see how green roofs can impact the local thermal
microclimate relative to a conventional roof surface. The third aim is to
quantify the rainwater retention efficiency of an intensive green roof. The
fourth aim is to assess the impacts on quality of rainwater runoff exerted by
an intensive green roof.
When assessing the multi-beneficial nature of green roofs, it is not possible

to sufficiently encompass all the benefits within one thesis, so this thesis will
focus on the benefits which previous work has indicated require more
research.
Specific objectives are:
 To quantify the amounts of PM10 that can be captured by green roof

vegetation
 To investigate differences in PM10 capture arising from roof location

and green roof vegetation composition
 To estimate the potential impact of green roofs on PM10 pollution

under theoretical future construction scenarios using upscaling
methods
 To measure the cooling impact of an intensive green roof on the

overlying air relative to a conventional roof surface
 To measure and compare the rainwater runoff retention efficiency of

an aged intensive green roof with that of a conventional roof surface
5
 To measure and compare the rainwater runoff quality of an aged
intensive green roof with that of a conventional roof surface and with
water quality guidelines
 To draw together results from experimental studies in order to assess

the environmental impacts of green roofs
During the data collection period, circumstances allowed for a further

objective:
 To investigate how the cooling effect of an intensive green roof can be

impaired by damage to the vegetation layer
1.4 Methodological Framework
Site selection was informed largely by issues of access to the roofs.

Rooftops can be highly unsafe zones when safety features such as railings or
raised edges are absent, therefore suitability for inclusion in the study
depended on these features. Potential green roofs and bare roofs, owned by
the stakeholders, were identified and access granted.
The four main chapters, wherein lie the results and discussion of the thesis,
each required their own monitoring methodologies. For Chapter 7, a novel
approach to monitoring roof runoff was called for, due to the problems
associated with adopting an experimental approach on a pre-existing roof
with a fixed structure that could not be altered. The monitoring periods for
each of the four chapters overlapped temporally. An empirical monitoring
approach was chosen, as opposed to a modelling approach. It can be
argued (see Chapter 4 – Methodology) that there is a need for more
empirical work within urban climatology. A large proportion of urban
climatology studies tend to rely on modelling approaches. Models are useful
tools for representing and investigating the dynamics of the various fluxes
within cities such as energy or pollutants, but there is a need for monitoring
work to obtain real-world data from different unique urban setups which can
then feed into and validate the modelling work.
6
1.5 Study Site – Manchester
The city of Manchester, located in North West England (see Fig. 1.1), is used
as a case study for this thesis. The Greater Manchester urban area is the
third largest in the UK, with a population of 2.7 million (UK Census 2011).
Manchester is representative of a large conurbation both in the UK and in
Northern Europe. It is of a sufficient size to manifest characteristics of
interest to urban climatologists, such as an Urban Heat Island (Smith et al.
2011), and significant transport-related air pollution (HFAS 2011). The
results of the thesis will therefore be of particular interest to not just
academics, but practitioners such as town planners, architects and local
authorities in Northern Europe.
Fig. 1.1 – Location of Greater Manchester. (Author created)
Sites were chosen within ‘The Corridor’ – a spatially-defined area of

economic development centred on the Oxford Road corridor of Manchester –
7
see Fig. 1.2. The Corridor is a major transport link stretching from the city
centre to south Manchester and is a strategically important economic
development area. A Corridor Partnership has been set up, which includes a
number of the stakeholders of the aforementioned green roof feasibility study
who are located around Oxford Road. As part of the partnership, a £1 billion
investment programme aims to maximise the opportunities arising from future
developments in the area, whilst supporting growth through improvements in
infrastructure. The environment is a key part of these infrastructure
improvements and The Corridor Partnership is keen to reduce pollution,
emissions and noise through incorporating sustainable urban planning, for
instance improved public transport and tree planting (MCSP 2009).
Fig. 1.2 – Location of The Corridor in relation to Manchester city centre. (Source:
Corridor Manchester 2012)
8
The Corridor is also of interest for urban climate studies as it has featured in
previous studies, including a characterisation of the variability of metal
contamination in Road-Deposited Sediment (RDS) (Robertson and Taylor
2007), and an investigation into the capability of The Corridor to adapt to
future climate warming by increasing the amount of green land cover in a
model simulation (Cavan and Kazmierczak 2011). However, an extensive
investigation of green roofs in the area has not been attempted.
Manchester was one of the world’s first industrial cities, and has a long
history of the climatic effects of urbanism, such as smog (MacKillop 2012). It
is also the site of one of the first observations on the uniqueness of urban
climates. The physicist, Sir Arthur Schuster, set up a weather observatory at
the University of Manchester which allowed comparisons with another station
located two miles north, and he wrote in 1893:
“The remarkable differences which appear in the temperature records

in different parts of the city furnish an additional proof, if proof were
wanted, that observations taken in or near a large town cannot be
taken to represent correctly the meteorological character of the
surrounding districts, but it by no means follows that these
observations are of no value. On the contrary, they may lead to some
important conclusions on what may be called town weather as
distinguished from country weather.” (Schuster, 1893 168)
It is apt, therefore, that Manchester continues to be a well researched city

within urban climatology.
1.6 Significance of Research
This study, by its empirical multivariate approach, has allowed a number of

novel green roof discoveries. It is the first study:
 to use magnetic techniques to quantify particle capture by green roof

vegetation
9
 to quantify, using particle counting, how much urban particulate
pollution can be expected to be removed from an urban area, annually,
in response to green roof installation (see Chapter 5)
 to reveal how the location, and plant species composition, of green

roofs can influence the capture of urban particulate pollution (see
Chapter 5)
 to assess the extent to which damage to the vegetation of an intensive

green roof allowed the impairment of the relative cooling properties to
be quantified, where previous studies have only postulated this effect
(Gill 2007) (see Chapter 6)
 to quantify runoff retention capacity of an intensive, aged, in-situ green

roof during a period of above average rainfall (see Chapter 7)
 to assess the effects of age on the runoff quality of an intensive green

roof (see Chapter 8)
 to undertake a chemical and physical analysis of green roof soil which

allowed a simple budget for Pb to be calculated. This identified a
possible link between historic atmospheric deposition and present day
runoff contamination (see Chapter 8)
 to assess the multi-benefit aspect of green roofs, within Manchester, in

terms of the services they provide as green infrastructure in cities, for
instance with respect to climate change adaptation (see Chapter 9)
10
1.7 Structure of Thesis
Fig. 1.3 shows a thesis outline diagram. This thesis is presented in an

‘alternative format’. This means that the findings of the thesis are presented
as articles suitable for publication in peer-reviewed journals. Three of the
articles have already been published, and the remaining article is under
review. A brief summary of the articles published is as follows:
Chapter 5 Published in ‘Atmospheric Environment’.

Speak, A. F., Rothwell, J. J., Lindley, S. J. & Smith, C. L. (2012).
Urban particulate pollution reduction by four species of green
roof vegetation in a UK city. Atmospheric Environment 61(0):
283-293.
Chapter 6 Published in ‘Urban Climate’.

Reduction of the cooling properties of an intensive green roof
due to vegetation damage. Urban Climate 3: 40 - 55
Chapter 7 Published in ‘Science of the Total Environment’

Rainwater runoff retention on an aged intensive green roof.
Science of the Total Environment 461-462: 28 - 38
Chapter 8 Submitted to ‘Environmental Pollution’

Metal and nutrient dynamics on an aged intensive green roof.
Submitted to Environmental Pollution 14 May 2013
11
Fig. 1.3 – Outline of the thesis structure. (Author created)
12
Chapter 1 has detailed the rationale, aims and significance of this research
and outlined the involvement of MCC as a CASE partner. Chapters 2 and 3
will provide a review of the state of current knowledge of relevance to the
thesis, and explain some of the different types and properties of green roofs.
Chapter 2 will discuss current climate change predictions for the UK and
introduce concepts from urban microclimatology and climate change
adaptation with green infrastructure. This will allow the existing knowledge
base on green roofs to be critically analysed in Chapter 3 and identify gaps in
the knowledge. Chapter 4 will outline the methodological approach used,
along with details of the empirical setups used in each of the studies.
Chapter 5 investigates the ability of green roofs to remediate PM10 pollution

by using longitudinal leaf sampling techniques. Chapter 6 uses air
temperature data collected over a green roof and an adjacent bare roof to
assess the relative cooling properties of the green roof. Chapters 7 and 8
deal with the green roof hydrology. In Chapter 7, rainwater runoff is
quantified to see how green roofs increase the rainfall runoff retention over
that of a bare roof, and in Chapter 8, a chemical analysis of the runoff gives
insights into how the quality of the runoff is affected by the presence of the
green roof. The findings of this research are then summarised in Chapter 9,
along with a discussion on the multi-benefit nature of green roofs and
suggestions of future research possibilities.
A number of typographic changes were made to Chapters 5 to 8 in order to

change the format from that of a published or submitted journal article to one
which matches the formatting of the thesis. The section numbering was
changed to coincide with chapter numbers and the table and figure numbers
were changed to continue the numbering of the previous chapters, otherwise
papers are as published/submitted.
13
CHAPTER 2 CURRENT KNOWLEDGE ON CLIMATE
CHANGE AND URBAN CLIMATES
2.1 Climate Change
2.1.1 Definition and broad-scale concepts
The fourth assessment report from the Intergovernmental Panel on Climate

Change (IPCC) highlighted that warming of the Earth’s climate system is
unequivocal and is evidenced from increasing global average air and ocean
temperatures and widespread melting of snow and ice (IPCC 2007a). Global
warming is the term given to the increase in the average temperature of the
oceans and near-surface air since the start of the industrial era (around 1750)
due to anthropogenic emissions of greenhouses gases (IPCC 2007b).
Carbon dioxide (CO2) is the most important greenhouse gas, and records
from the Scripps Institution of Oceanography show that global average CO 2
concentration has increased from 315 ppm in 1958 to 395 ppm in 2013
(Keeling et al. 2013). Currently around 50 Gigatonnes of greenhouse gasses
(CO2 equivalent) are emitted into the atmosphere each year (Rogelj et al.
2013), mostly from the combustion of fossil fuels for energy, and indirectly
from changes in land use. The elevated levels of greenhouse gasses have
led to an increase in forcing from solar radation in the lower atmosphere
resulting in a global average temperature rise (Fig. 2.1) of about 0.6°C over
the past 100 years (Hulme 2002)
Fig. 2.1 – Change in global surface temperatures relative to 1951 – 1980 climate
average temperatures. (Source: NASA 2011)
14
Climate models such as those produced by the Hadley Centre are able to
simulate the recent temperature changes when they incorporate both natural
and anthropogenic factors. Natural external forcing factors include changes
in the energy output of the sun and volcanic eruptions. These, when
considered alone and excluding increasing greenhouse gas concentrations,
are unable to account for the warming of the past 50 years (Stott et al. 2006).
This analysis, amongst others, led the IPCC to state that most of the warming
is due to humans, with a 95% confidence level (IPCC 2007b).
The Copenhagen Accord of 2009 stresses the International Community’s

commitment to ‘hold the increase in global climate change below 2 degrees
Celsius’ (Copenhagen Accord 2009). It has been criticised for not including
any binding emissions targets, and focussing on the ‘peaking of global and
national emissions as soon as possible’ when the focus should be more on
accounting for cumulative emissions which are much more important due to
the long residence time of CO2 (Anderson and Bows 2011). Revised
estimates, which incorporate the continued growth in emissions by both
Annex 1 and non-Annex 1 nations indicate a world where a 4oC global
average rise is likely (Anderson and Bows, 2008).
The current and projected impacts of this warming are numerous. Increasing
global temperature leads to melting of the ice caps, glacial retreat and
subsequent rise in sea levels (Hock et al. 2009). An increased intensity and
frequency of extreme weather events has been observed and related to
increasing sea-surface temperatures (Webster et al. 2005). Warming leads
to more evaporation and thus surface drying, so an increased intensity and
duration of droughts is predicted (Coumou and Rahmstorf 2012). Warmer air,
by being able to hold more moisture, may enhance precipitation extremes,
and provide more latent energy to drive storms (Coumou and Rahmstorf
2012). Additionally, changes in rainfall distribution are expected, with water
scarcity in some areas and floods in others (Boyd and Ibarraran 2009).
The IPCC projections for Europe indicate the largest warming is likely to be in
northern Europe in the winter and in the Mediterranean area in the summer
(IPCC 2007b). Fig. 2.2 shows that the five warmest European summers of
15
the past 500 years have occurred in the past decade. Upper extremes of
daily precipitation are very likely to increase in northern Europe and the risk
of summer drought is likely to increase in central and southern Europe
(Christensen et al. 2007).
Fig. 2.2 – European summer temperatures for 1500 – 2010. The top panel shows the
statistical frequency distribution of summer land-temperature anomalies (relative to the
1970 – 1999 period) for the 1500 – 2010 period in Europe. The five coldest and
warmest summers are highlighted. The lower panel shows a running decadal frequency
of extreme summers, defined as those with a temperature exceeding the 95th percentile
of the 1500 – 2010 distribution. (Source: Barriopedro et al. 2011)
What is clear is that an increase in global average temperature will have

profound and varied effects on a local scale, impacting areas with dense
human populations such as cities. Warming over land areas will be greater
than the global mean due to less evaporative cooling and a smaller thermal
16
inertia than the oceans (Christensen et al. 2007). Cities in Europe will have
to find ways to adapt to significant temperature increases. For example, the
maximum summer temperatures in Paris could rise by as much as 5°C by the
end of the century (Lemonsu et al. 2013).
2.1.2 Climate change in the UK
In the United Kingdom, the UK Climate Impacts Programme (UKCIP) is

responsible for developing a comprehensive climate model for predicting
changes to the climate of the British Isles and the production of reports to
help policy makers, local authorities and businesses plan and adapt, the
latest of which is UKCP09 (Murphy et al. 2009). UKCP09 gives probabilistic
projections for a number of atmospheric variables with different temporal and
spatial averaging, by several future time periods, under three future
emissions scenarios (Jenkins et al. 2009). It is an improvement on previous
model projections in a number of ways such as enhanced spatial resolution,
inclusion of carbon cycle feedback data and modelling uncertainty
exploration (Jenkins et al. 2009). Fig. 2.3 shows one of the main climate
changes predicted for the 2080s under the medium emissions scenario.
Under this scenario, at 50% probability level, there would be an increase in
summer mean temperatures (greatest in parts of southern England – up to
4.2°C) and an increase in mean daily minimum temperature in winter, by on
average 2.1°C to 3.5°C depending on location.
17
Fig. 2.3 – 10, 50 and 90% probability levels of changes to the mean daily minimum
temperature in winter (top) and summer (bottom), by the 2080s, under the Medium
emissions scenario. (Source: Murphy et al. 2009)
The predicted patterns for precipitation are shown in Fig. 2.4. An increase,
by up to 33%, of winter precipitation, is expected in western parts of the UK
and mean summer cloud amount decreases by up to -18% in parts of
southern UK (Murphy et al. 2009). The UKCP09 projections describe a
future with a higher frequency of extreme events - droughts and heatwaves in
the summer, and floods in the winter. In summer, a possibility is that rainfall
could become concentrated into intense downpours, with an increase in the
18
Fig. 2.4 – 10, 50 and 90% probability levels of changes to annual (top), winter (middle)
and summer (bottom) mean precipitation, by the 2080s, under the Medium emissions
scenario. (Source: Murphy et al. 2009)
frequency of 20, 30, 50 and 100 year return period events for the north west
UK (Sanderson 2010).
Recent instances of weather extremes exemplify the damage to

infrastructure and threat to human life that is possible. The 2003 European
19
heatwave caused an estimated 70,000 excess deaths (Robine et al. 2008), of
which over 2,000 were in England and Wales (Larsen 2006), and it was the
hottest summer for at least 500 years (Luterbacher et al. 2004). The wettest
autumn on record in England and Wales occurred in 2000 causing £1.3
billion in damages to property (Pall et al. 2011), and the period from May to
July in 2007 was also a record-breaking wet period with 406 mm rain (WMO
2009). These synoptic weather events can have greater impacts locally,
depending on the nature of the receiving environment. The next section
explores how urban areas can manifest unique climates which make them
potentially more vulnerable to the increased frequency of extreme events
predicted by climate change scientists.
2.2 Urban Climates
2.2.1 General overview
Urban climate phenomenon will now be discussed in detail, and related to the
study site of Manchester. Urban areas can generate their own local climate
conditions due to the nature of the physical aspects of the built-up
environment in comparison to more ‘natural’ rural areas, and due to the
human activities that occur within cities (Oke 1981; Wilby 2007). The
process of urbanisation replaces vegetated surfaces with impervious built
surfaces, and the result is a modification of the local climate – both within the
built-up area, and in the atmosphere above the city and beyond its
boundaries (Erell et al. 2011). Localised climates arise from the different
surface elements of urban areas forming a patchwork of paved areas and
some green space with contrasting radiative, thermal, aerodynamic and
moisture properties. These possess diverse energy budgets leading to
mutual interactions by radiative exchange and small-scale advection (Arnfield
2003). As well as being associated with urban-rural differences, aggregation
of these fundamental morphological units can create distinct climatic zones
within cities (Arnfield 2003). For instance, building walls and the streets
between them define the urban canyon which can affect local temperature
differences and wind and turbulence patterns at street level (Souch and
Grimmond 2006). These distinct zones can even influence the composition
20
of plant species living in that zone, based on the species’ temperature
preferences, to an extent which allows floristic mapping of phenomena such
as the Urban Heat Island (UHI) (Bechtel and Schmidt 2011).
2.2.2 Urban Heat Islands
Urban areas are subject to relatively high temperatures compared to their

rural hinterlands and it is widely recognised that built up areas can develop
UHIs which can be up to 5-6°C warmer than surrounding rural areas (Oke
1982).
Fig. 2.5 – A generalised Urban Heat Island. (Source: Erell et al. 2011)
A UHI arises when there is a difference in temperature between a certain

urban location and a given reference point in a nonurban location (Taha
2004). Fig. 2.5 shows how this temperature difference evolves as one
moves from outside a city into the urban core. They can appear due to a
combination of several factors:
• Urban areas are abundant in materials with a high thermal capacity

such as concrete and asphalt (Taha 2004).
• Urban geometry results in a higher total surface area able to reflect,

absorb and re-release solar radiation. The release of this extra heat
contributes directly to warming the air (Taha 2004).
21
• Urban areas have a smaller Sky View Factor (SVF, portion of visible
sky) which helps create night time UHIs as radiative cooling of urban
surfaces to the cooler sky is hindered by obstructions such as walls.
Radiated heat is subjected to multiple reflections within ‘urban canyons’
(streets cutting through dense, high buildings) which warms the surfaces and
air (Oke 1981;Holmer et al. 2007).
• The albedo of urban surfaces can be lower than that of the vegetated
surfaces abundant in rural areas (Taha 1997).
• The ‘sealed surfaces’ of urban buildings encourages rainfall runoff,

and this combined with a lower vegetative cover means less heat loss from
the system as evapotranspiration (Whitford et al. 2001).
• Urban areas have greater inputs of heat as a result of the high density
of energy use in cities such as heating of commercial buildings and a
concentration of vehicular transport (Davies et al. 2008; Hamilton et al. 2009;
Smith et al. 2009).
Fig. 2.6 – Regional Climate Model climate change over the UK (2050s minus 1980s,
SRES A1b) due to (left) greenhouse gases, (middle) greenhouse gases and
increased energy use in urban areas, and (right) increased energy use only.
(Source: McCarthy et al. 2010)
22
All these factors contribute to an altered surface energy balance and
consequent elevation of sensible heat in cities via the absorption of radiated
energy by the surrounding air. This localised effect of land-use change is
highly significant and has led to criticism of the climate change community’s
major focus on mitigation of greenhouse gas emissions and globally
averaged temperature increases rather than tackling the much greater
temperature increases occurring in cites (Stone 2012). Much more tangible
climate impacts will occur in urban areas, affecting a majority of the human
population. Cities are no longer considered by climatologists as sources of
anomalous data distortions within synoptic grids, but are being incorporated
into global climate policy (Hebbert and Jankovic 2013).
Fig. 2.6 shows the results of representing the urban land surface in the UK
Met Office Hadley Centre Regional Climate Model and demonstrates the
impact that urban centres such as London and Manchester can have on local
temperature changes. An increase in computer processing power is allowing
climate models to operate at increasingly finer resolution which allows cities,
and their unique climates, to be represented within general circulation models
(Hebbert and Jankovic 2013).
The effects of UHIs are varied but the main one of concern to city dwellers is
the increased ambient temperature which is typically of the magnitude of 2°C
but has been recorded to be as high as 10°C (Livada 2002). A study in
Turkey found the UHI of Istanbul was even responsible for significant
warming of urban groundwater (Yalcin and Yetemen 2009). Manchester’s
UHI has been relatively well researched and has been found to be 3°C in the
day and 5°C at night under ideal UHI conditions in summer (Smith et al.
2011). A study using fixed point monitoring stations found urban-rural
differences of up to 8°C in summer and 10°C in winter (Cheung 2011). The
UHI intensity had a negative linear relationship with distance from city centre,
evapotranspirative fraction, wind speed and cloud cover (Cheung 2011).
UHIs are generally more of a problem in summer months, when peak
temperatures become prevalent. In winter, a UHI can serve to reduce
heating costs (Akbari and Konopacki 2005).
23
A 10°C heat island was also observed in a study which utilised a large
number of simultaneous, but perhaps less reliable, measurements in March,
and found high temperatures correlating well with low vegetation cover
(Knight et al. 2010). A model to investigate the anthropogenic heat fluxes,
which would contribute to this UHI effect in Manchester, quantified mean heat
emissions of 23 W/m2 in city centre areas, with buildings accounting for 60%
of the total emissions (Smith et al. 2009). Manchester therefore represents a
city with a demonstrable UHI linked to the altered land use.
2.2.3 Precipitation
Increased cloudiness and rainfall can occur as a result of enhanced

convective activity within heat plumes generated by UHIs (Taha 2004).
Additionally, higher atmospheric particulate matter or aerosol concentrations
can provide condensation nuclei for rain droplets (Bonan 2002). In one of the
first studies on this phenomenon, the areas of increased precipitation ranged
from directly over the urban centre to 80 km downwind (Huff and Changnon
1973).
Of particular interest, however, is what happens to the precipitation, once it

falls, within the built environment. Hydrological processes in cities are
altered with respect to those within rural areas. Cities have large areas of
impervious surfaces, so that there is an increase in the rate and volume of
surface rainwater runoff (Whitford et al. 2001). Cities have even been
likened to a karst topography, with sewers performing the same function of
limestone caves leading to rapid through-flow of rain runoff and leaving a
parched physical environment in city centres (Bunge 1973). A high spatial
diversity in urban soil moisture characteristics arises with intensely built-up
areas being virtually devoid of replenishment by infiltration. Subsequently,
parks and gardens, by providing infiltration routes for surface runoff, may
then have more infiltration than adjacent rural areas (Douglas 1991). The
artificially created (sewers, gutters, storm drains) or human-modified
(culverted streams) drainage channels that exist in cities carry greater
volumes of runoff than in rural areas, and the water movement is
concentrated into bigger storm peak discharges (Douglas 1991).
24
A general lack of landscape features, such as greenspace and water bodies,
which can intercept, store, and infiltrate rainwater is one of the main causes
of flooding in towns and cities (Defra 2008). Flooding resulting from rainfall
can be caused by different patterns of rainfall, such as small but frequent rain
showers steadily moving a groundwater system to full capacity, or one heavy,
prolonged rain shower causing a flash flood (Dale 2005). Surface water
(pluvial) flooding can be more important than fluvial flooding in cities (Defra
2008) and the priority given to it in recent flood management legislation has
been recently improved due to the fact that 2 million people in UK urban
areas are at risk of a 1 in 200 year event (Houston et al. 2011).
In Manchester, flooding has the potential to cause huge amounts of damage

to buildings and critical infrastructure (Kazmierczak 2011), and surface runoff
from urban areas can also impact negatively on local receiving waters by
increasing bank erosion of streams (Hilten et al. 2008) and raising the
temperature of aquatic systems (Frazer 2005). Flooding has been suggested
to be the dominant climate impact in the Manchester region, over
temperature extremes, high winds, fog and drought, in a study which used
archival research and statistical analyses of past climate data (Smith and
Lawson 2012). There is also some evidence that flood events in the region
have been becoming more frequent over recent decades and have shifted
from riverine to pluvial in nature (Smith and Lawson 2012).
2.2.4 Air Pollution
Manchester, due to its industrial past, has had a long history of urban air
pollution. The Manchester and Salford Police Act of 1792 emphasised a
need for industrial chimneys to reduce the smoke from coal combustion
(Brimblecombe and Maynard 2001). The Clean Air Act of 1956 regulated
domestic and industrial sources of air pollution and brought about a
significant decrease in levels, mostly due to a switch from coal combustion to
natural gas in domestic heating. Since then, there have been a succession
of Clean Air Acts and EC Directives to combat air pollution by setting
emissions limits, for example (NEGTAP 2001). This substantial political
action to reduce the effects of the major pollutants responsible for
25
acidification, smog and eutrophication has resulted in a significant decline of
emissions of sulphur dioxide (SO2) and NOx (NEGTAP 2001). The UK has
extensive monitoring networks in place to observe changes in concentration,
deposition and effects on soils and freshwater, to ensure the targets are met
(Fowler et al. 2006).
While regional and transboundary air pollution levels are falling, high
concentrations of air pollutants can exist at a more local scale in urban areas
(Brimblecombe and Maynard 2001). This is due to a concentration of sources
of pollution in cities. Road traffic, for example, has increased five-fold since
the Clean Air Act of 1956 and is only very recently showing signs of levelling
off, in terms of vehicle miles, possibly as a consequence of the financial crisis
(DfT 2013). This has an associated decrease in air quality due to an
increase in emissions of NOx, benzene, O3 and particulate matter (PM) from
fuel combustion and vehicle component wear (AQEG 2005; Petroff et al.
2008; Bukowiecki et al. 2010).
Urban areas often manifest specific spatial and temporal distributions of air
pollution. For instance, in Manchester, identifiable peaks in ultra-fine PM
concentration occur with greater magnitude in the city centre and coincide
with the twice-daily rush hour vehicle increases (Harris et al. 2009). Strong
urban-rural gradients in air pollutant concentrations are well-documented
(Lovett et al. 2000; Harner et al. 2004) and concentrations are often positively
correlated with building density (Weng and Yang 2006). Road network
density is also important, with vehicle-sourced pollution being very high at the
kerbside (Birmili 2006).
Atmospheric inversion events can lead to the development of persistent

pollution episodes by trapping pollutants within the urban boundary layer and
reducing the speed of beneficial winds which normally ventilate cities and
disperse the pollution. Climate change could become responsible for an
increase in the frequency of stagnant air masses in summer (AQEG 2007).
Also, higher summer temperatures lead to a higher frequency of
photochemical smog episodes and there could also be an increase in the flux
of O3 from the stratosphere to the troposphere (AQEG 2007).
26
Mitigation measures for urban air pollution are currently a high priority
because levels in Europe, particularly for PM, remain high (Pascal et al.
2013). A recent study of 25 European cities highlights the public health
burden of PM and O3 in Europe, with significant health and monetary gains
from reducing PM concentrations to the World Health Organisation (WHO)
guideline values (Pascal et al. 2013). Newly emerging evidence suggests
possible effects on premature births, lung-function development in children
on top of established short and long-term effects on cardiovascular and
respiratory diseases (The Lancet 2013). The effects may exist at low levels
of air pollution, with no safe threshold level, but a linear concentration-
response relationship (The Lancet 2013). The majority of air pollution control
measures are rightly focussed on reducing emissions, but techniques that aid
in the removal of existing pollutants from the atmosphere, such as tree-
planting (Morani et al. 2011), are also highly desirable.
2.2.5 Water Pollution
A further problem caused by air pollution is related to the fate of the

pollutants within cities. Air pollutants, such as NOx and PM containing heavy
metals such as Cd, Cu and Zn from vehicular sources, accumulate on
impervious surfaces during the processes of wet and dry deposition and are
subsequently taken up in rainfall runoff and enter surface waters (Forster
1999). A first flush phenomenon often manifests, whereby the concentration
of pollutants is substantially higher during the initial stages of runoff initiation
than in later stages (Lee et al. 2002). Once in water bodies, pollutants cause
problems for aquatic life as they bioaccumulate and exert toxic effects
(Tipping et al. 2008; Pizzol et al. 2011). The key contaminant parameters for
water pollution potential are Total Suspended Solids (TSS), Biochemical
Oxygen Demand (BOD), bacteria, heavy metals (Zn, Pb and Cd) and
polyaromatic hydrocarbons (PAH), with urban open areas, particularly roads,
being important sources (Ellis and Mitchell 2006).
Nutrients, such as atmospherically deposited NO, NO2 and NH3, can also be
a problem when urban surface waters discharge to water bodies that are
already nutrient-rich, resulting in eutrophication effects (Ellis and Mitchell
27
2006). Characteristic problems associated with eutrophic systems include
the development of algal blooms which can often be toxic (Heisler et al.
2008). The algal blooms limit light to benthic aquatic vegetation and can
produce anoxic conditions when it decomposes which has adverse effects on
macrofauna populations e.g. fish (Heisler et al. 2008; Carey et al. 2013). The
damage costs of freshwater eutrophication in England and Wales have been
estimated to be up to £114 million/yr due to factors such as reduced value of
waterfront dwellings, drinking water treatment costs, reduced amenity value
of water bodies and negative ecological effects on biota (Pretty et al. 2002).
Urban areas can also demonstrate legacy pollution, whereby pollution exists
from past activities (Albanese and Cicchella 2012). Examples included
highly localised Cr (VI) contamination of soil on land once occupied by
tanneries (Stepniewska and Bucior 2001), and Pb contamination from a more
diffuse source of past use of lead additives in petrol (Laidlaw et al. 2012).
Leaching of these contaminants by infiltration of rainwater then causes
problems for nearby aquatic systems. These zones of contamination often
occur in brownfield sites, which are abandoned, vacant or derelict areas
within cities where the costs of dealing with the contamination have
prohibited development (Iverson et al. 2007).
The European Community passed its Water Framework Directive (WFD,

European Community 2000) in 2000 which requires management of waters
on a whole river catchment basis. One requirement of the WFD is the
identification and quantification of diffuse sources of urban water pollution in
terms of the land-use activities that generate it and the associated discharge
loadings (Ellis and Mitchell 2006). The sources of this Urban Diffuse
Pollution (UDP) can then be controlled in order to preserve water resources
from the deleterious effects of urbanisation (Defra 2011).
2.3 Adaptation
2.3.1 Introduction to adaptation themes
This section will explore some of the options open to city planners for
adapting cities to climate change and thereby reversing some of the
28
deleterious effects of urbanisation discussed in the previous section.
According to the IPCC (2007a), responses to climate change can be divided
into:
• Adaption – Initiatives and measures to reduce the vulnerability of

natural and human systems against actual or expected climate change
effects.
• Mitigation – with respect to climate change, mitigation means

implementing policies to reduce greenhouse gas emissions and enhance
sinks.
The focus for this thesis is adaptation, specifically urban adaptation.
The unique climates that manifest in cities have the potential to combine with
climate change projections to make cities hazardous places to live.
Additionally, half the world’s population currently live in cities, and this is
projected to increase to 70% by 2050 (UN 2009). Most European countries
have well over half their populations living in cities, and the UK urban
population is currently over 60% of the total population (Pateman 2011).
Therefore, simply from the higher population densities it is reasonable to
expect that climate change impacts will be felt most severely in cities (Wilby
2007), but the nature of the built environment enhances the risks.
For example, the large proportion of impervious surfaces in cities increases

the risk of surface water flooding. An increased volume and flow-rate of
rainwater runoff can put pressure on heavily modified and channelised urban
watercourses which compromises their ability to deal with excess water
(Whitford et al. 2001) with resultant flood damage to infrastructure. Rising
temperatures associated with climate change can combine with the high
temperatures attributed to the UHI effect, especially in the summer months
when temperature extremes are common. When the temperature is close to
human body temperature of 37°C, physiological stress can occur. Negative
health effects from climate change also arise from communicable disease,
lack of food and water security, stress and population migration (Lundgren et
al. 2013). Impacts on cities can also have different scales of effects
29
depending on the relative vulnerabilities of ecosystems, infrastructure and
people such as the elderly or infirm (Lindley et al. 2004; Hunt and Watkiss
2011).
The potential of predicted climate changes to cause an increase in

biophysical hazards in cities has prompted research into strategies to
respond. The planning response is the focus of the Building Knowledge for a
Changing Climate programme, which urges urban designers, architects and
planners to incorporate ‘climate-proofing’ in new developments and
refurbishments (Shaw et al. 2007). Recently in the UK, the ARCC
(Adaptation and Resilience in a Changing Climate) research network
contains multiple projects with the aim to develop tools for analysis of
adaptation options in urban areas, with particular focus on making
infrastructure, the built environment and transport systems resilient to
environmental change (ARCC 2013).
Passive, soft engineering, adaptation strategies that seek to reduce the

extent and magnitude of the UHI, and adapt to future climate change in urban
areas include increasing urban vegetative cover (Hardin and Jensen 2007),
which will be discussed in section 2.3.2, and using higher-albedo surface
materials (Pomerantz et al. 1999; Akbari and Matthews 2012). The latter
works by reflecting a higher proportion of incoming solar radiation, thus
absorbing, and later re-radiating, less energy to the surrounding air. Albedo
can be increased by using lighter coloured materials for construction, painting
existing surfaces with light paints or using new cool coloured materials that
are highly reflective in the near-infrared band (Synnefa et al. 2007). There is
a great potential for reducing the net absorption of solar radiation and
modelling studies have shown that significant urban temperature reductions
and consequent energy savings can be achieved by albedo modification
(Solecki et al. 2005; Synnefa et al. 2008). A problem with the use of highly
reflective materials in cities is that the reflected light can cause discomforting
glare (Synnefa et al. 2007) and the reflected radiation is simply diverted
elsewhere, enhancing the radiative forcing on nearby buildings. Also, in high
latitude cities, increasing the albedo of roof surfaces can reduce solar heating
30
in winter and thus offset energy savings from a decrease in air conditioning
with an increase in space heating (Oleson et al. 2010).
Passive adaptation strategies are preferable to active strategies (such as

mechanical air conditioning in homes and work places) because of the
reduced energy demand and thus reduced greenhouse gas and air pollutant
levels from the burning of fossil fuels by electricity-producing facilities to meet
this demand (Papadopoulos et al. 2003; Gupta and Gregg 2012). The use of
air conditioning has increased rapidly in the developed world with 87% of US
housing units having central or room air conditioning (Sivak 2009). Buying
and using air conditioning is prohibitively expensive for many urban poor
(Sivak 2009) and the increased summer electricity demand is significant.
Peak urban electric demand has been found to rise by 2% – 4% for each 1oC
rise in daily maximum temperature above a threshold of 15oC to 20oC (Akbari
and Taha 1992).
The main focus of climate change adaptation literature is the effect that rising
temperatures will have on energy use and human wellbeing, however,
adaptations that seek to also reduce the impacts of flooding and air pollution
are also receiving a lot of attention and are the subject of much research.
Adaptations that can meet multiple targets are much more beneficial than
those, such as high-albedo roofs, which only adapt for one climate impact.
Urban greenspace can often meet these multiple targets, whilst providing
additional benefits not directly related to climate change adaptation.
Adaptation appears to be a low priority issue for city planners and governors
in Europe at present (Carter 2011), however progress is being made in
overcoming barriers within policy, governance frameworks and uncertainty in
climate science. An example of good practice which utilises urban
greenspace comes from Frieburg in Germany. The city’s Land Use Plan
2020 prioritises landscape protection over building development, and
encourages the protection of cool air corridors into the city (Carter 2011).
31
2.3.2 Adaptation with greenspace
Vegetation in urban areas can have significant impacts on microclimate

(Wilmers 1988; Dimoudi and Nikolopoulou 2003; Gómez et al. 2004; Gomez
et al. 2008). Urban vegetation can take many forms - from large, landscaped
inner city parks (James et al. 2009) to roadside trees, residential gardens,
green facades (Köhler 2008) and green roofs (Grant 2006). These patches
of green exist at different spatial scales and at differing levels of
interconnectedness (Heidt and Neef 2008) but can impact on the urban
environment by affecting thermal dynamics (Takebayashi and Moriyama
2007), wind speed (Taha et al. 1991), air quality (Taha 1996) humidity (Liu et
al. 2008) and noise levels (Van Renterghem and Botteldooren 2008) on a
localised scale (Huang et al. 1987; Vargo et al. 2013) and at a city-wide scale
(Oke 1989). These benefits are often discussed in terms of ‘ecosystem
services’ which are benefits to mankind provided by nature (Robinson and
Lundholm 2012), such as the ones at the heart of this thesis.
With regard to lowering the high air temperatures synonymous with the UHI
effect, urban vegetation achieves its effect by:
• Reduction of solar heat gains in urban surfaces through leaf shading

(Hardin and Jensen 2007)
• Increase of solar reflectance due to the higher albedo of leaf surfaces

compared to most building materials (Grant et al. 2003)
• Increase of latent cooling by adding moisture to the air through

evaporation from soils and leaf surfaces and transpiration from plants
(Solecki et al. 2005)
Alexandri and Jones (2008) found that covering the entire building envelope
with vegetation can reduce daytime average temperatures in the urban
canyons of hot and arid Riyadh by 9.1oC. Courtyards in southern Israel
treated with shade trees and grass yielded a daytime air temperature
reduction of 2.5oC relative to a non-vegetated, unshaded courtyard
(Shashua-Bar et al. 2009). The local cooling effects of green areas can be
correlated with human comfort indices (Gómez et al. 2004).
32
Surface and air (globe) summer temperatures were measured at plots
consisting of a grass or concrete surface and with or without tree-shade, in
Manchester (Armson et al. 2012). Surface temperatures were greatly
affected by surface type and shade, with grass reducing maximum surface
temperatures by 24oC and tree shade reducing them by 19oC. Grass had
little effect on air temperatures, however, while tree shade reduced them by
up to 5oC - 7oC. Thus, grass and trees can have significant effects on
Manchester’s UHI by reducing surface temperatures, however, only the tree
shade can provide human thermal comfort benefits. Human thermal comfort
is strongly related to ambient air temperature and humidity, but also to solar
radiation, infrared radiation and windspeeds (Herrington and Vittum 1977).
Tree planting in Manchester can be used to climate-proof high density

residential areas (Hall et al. 2012) with reductions of maximum building
surface temperature of between 0.5oC and 2.3oC possible. However,
additional planting could only increase tree cover by between 2.8% and 5.3%
due to limitations on available space (Hall et al. 2012).
33
Gill (2007) characterised the urban environment of Manchester into distinct
Urban Morphology Types (UMTs) and used this to approximate the amount of
evapotranspiring surfaces. This information was used in an energy
exchange model which was run for the baseline 1961-1990 climate and low
and high emissions scenarios for the 2020s, 2050s and 2080s of the
UKCIP02 climate prediction conditions. Adding or removing 10% green
cover had a dramatic effect on surface temperature (Fig. 2.7), with addition
resulting in surface temperatures being kept at or below the 1961-1990
baseline temperatures up to, but not including, the 2080s high emissions
scenario (Gill 2007). 10% removal of green cover from the town centre UMT
resulted in a surface temperature increase of 8.2oC by the 2080s high
emissions scenario compared to a rise of only 4.4oC if no change was made
to surface cover. An assumption made in this study, however, was that there
is no drought stress on the vegetation. This would reduce the
evapotranspiration cooling effect due to a lack of water in the substrate and
plant tissues.
Fig. 2.7 - Maximum surface temperature for the 98th percentile summer day in town
centres, with current form and when 10% green cover is added or removed.
Dashed line shows the temperature for the 1961-1990 current form case. (Source:
Gill 2007)
34
When investigated as an adaptation strategy for UHIs it must be noted that
any strategy that simultaneously reduces urban temperature and greenhouse
gases via the sequestration of carbon in the increased plant biomass is
preferred (Solecki et al. 2005). Strategies of this nature, such as urban
afforestation programmes, are termed ‘adaptive mitigation’ and are highly
valuable for nations committed to reaching climate goals as set out by the
Kyoto protocol, among others (Stone 2012).
Urban greenspace can also help reduce the impact of climate change
predictions for increased winter precipitation. Sustainable Urban Drainage
Systems (SUDS) is the name given to a range of low impact development
techniques, such as permeable pavings and rain gardens, that aim to
manage surface water drainage and prevent flooding (CIRIA 2007). Natural
systems are replicated as much as possible so that surface water runoff can
be collected, stored and even cleaned before it is released slowly back into
the environment i.e. into groundwater and watercourses. Parks and other
urban GI can act as stores or access routes to the groundwater (Frazer
2005), however different soil types have different qualities, with clay soils
being more impervious to heavy downpours. Urban vegetation increases
evapotranspiration, thus allowing significant amounts of water to go back to
the atmosphere.
SUDS components work in a number of ways: they enable infiltration into the
ground (porous pavings), convey flow into a watercourse (swales), provide
on-site storage which can be temporary (rain gardens) and attenuate the flow
of water (green roofs) (CIRIA 2007). Source control is one of the more
important principles of SUDS design because it relieves pressure on the
more terminal storage techniques, such as ponds, whose capacities can be
exceeded in heavy rainfall events. Source control techniques include green
roofs and rainwater harvesting (CIRIA 2007).
Strategic planting schemes can also help to reduce air pollution. Vegetation,
with its high leaf surface area, can intercept gaseous pollutants (SO 2, NOx,
O3) and PM, thus enhancing deposition rates (Fowler et al. 1989). The
pollution can be retained on the leaf surface until it is washed off by rains or
35
removed from the system via deciduous leaf-fall. This has prompted a
number of quantification studies. Parks in Shanghai could contribute to 9.1%,
5.3% and 2.6% removal of PM, SO2 and NO2 respectively (Yin et al. 2011).
Species differences were displayed in a study by Beckett et al. (2000) which
found particle trapping efficiencies were higher for conifer species,
presumably due to the finer, more complex foliage structure.
Urban green areas can have multiple benefits in addition to climate

modification, pollution reduction and stormwater buffering such as the strong
relation between green space and perceived general health of city residents
(Maas et al. 2006), generation of positive social cohesion (Fuller and Gaston
2009) and native biodiversity protection (Millard 2008). These are added
incentives to adopt urban greening as an adaptation method for a changing
climate which should be included in urban planning approaches utilising
strategies such as plant species evaluations of suitability for the urban
environment (Monterusso et al. 2005). Green space is more than just a
luxury and aesthetic component of cities and should be considered an
integral part of urban planning (James et al. 2009).
2.4 Conclusion
This chapter has outlined some of the projected impacts of climate change
on cities and their inhabitants, with a particular emphasis on those relevant to
a large city in Northern Europe such as Manchester. Cities are where the
impacts of climate change will be most severely felt due to the high
population density and the nature of the built environment. High
temperatures may combine with the urban heat island effect to create a risk
of thermal discomfort and heat stress mortality. The impervious nature of
cities means that surface water flooding can have adverse effects on
infrastructure. High vehicle use in cities means air pollution threats are
present. Deposition of these pollutants onto the urban surface results in a
deterioration of the water bodies receiving runoff from these surfaces.
Increasing the amount of greenspace within cities can address these
problems, often simultaneously, and there is a growing interest in using green
infrastructure to climate-proof the world’s cities against future hazards.
36
CHAPTER 3 CURRENT KNOWLEDGE ON GREEN ROOFS
3.1 Green roofs – an introduction
One problem with extensive greening of urban centres is the competition for
expensive building space and the fact that the fundamental layout of cities is
unlikely to change for some years (Wong et al. 2003). It has already been
noted in section 2.3.2 how tree planting schemes aimed at climate-proofing
Manchester’s building stock are limited by the availability of ground space,
which is often restricted to intra-pavement planting (Hall et al. 2012). A
solution is to utilise the available space afforded by rooftops. Estimates of
urban rooftop coverage vary greatly but can be as high as 35% of the land
cover (MacMillan 2004) and 40 – 50% of the impervious area (Dunnett and
Kingsbury 2004).
The use of green roofs has a long tradition worldwide (Köhler et al. 2002)
with turf roofs developed in Viking times still being used on pitched roofs in
Scandinavia (Dunnett and Kingsbury 2004) and roof gardens have been
identified in the ruins of Roman Herculaneum, buried in the eruption of Mount
Vesuvius in AD 79 (Grant et al. 2003). The contemporary green roof
movement can be said to have begun in the early 20th century in Germany,
where extensive green roofs were developed for use on residential buildings
as a form of fire protection (Köhler et al. 2002), however it is possible that
these roofs were not initially planned to be greened because the initial fire
protection strategy was to just use tarboards and then a layer of a sand and
gravel, which consequently became colonised by spontaneous vegetation
(Werthmann 2007).
37
Fig. 3.1 – Intensive green roof on the Precinct bridge in Manchester, UK (Author
image)
Green roofs are broadly classified as intensive or extensive. Intensive green

roofs (Fig. 3.1), often called roof gardens, generally have a substrate depth
greater than 150-200mm (Oberndorfer et al. 2007) and are characterised by
an ability to support a wide variety of plants but with the need for high levels
of maintenance and possibly artificial irrigation. Extensive green roofs (Fig.
3.2) have a thinner substrate, require less maintenance and are
characterised by ground coverage plants of low height such as grasses and
sedum - a succulent genus chosen for its tolerance to drought and its shallow
substrate adaptability (Monterusso et al. 2005; VanWoert et al. 2005).
38
Fig. 3.2 – Extensive green roof in Manchester, UK (Author image)
Brown roofs are a less expensive type of extensive roof consisting of a gravel
substrate which is allowed to be colonised by wind and bird distributed plants.
39
Fig. 3.3 – Schematic diagram of the layered structure of a typical green roof system
(Source: American Wick Drain Corp 2009)
The species composition of a green roof is determined by a number of

factors such as design intent, aesthetic appeal, environmental conditions,
media depth and composition, roof accessibility and maintenance
requirements (Getter and Rowe 2008). Succulent plants such as sedum are
preferred for extensive green roofs due to their low maintenance
requirements and ability to withstand the often harsh conditions on a rooftop,
however, there is a lot of research underway into the suitability of many other
plant taxa with a preference for native taxa due to their adaptation to the
existing climate (Monterusso et al. 2005; Nagase and Tashiro 2012; Davies
et al. 2012). Green roof species compositions are also dynamic in nature,
and succession has been demonstrated on both extensive sedum roofs
(Rowe et al. 2012) and on intensive roofs planted with grasses and herbs,
often due to wind and bird-carried invasive species (Dunnett et al. 2008a).
40
Construction methods vary but a general outline can be seen in Fig. 3.3.
Initial construction costs can be high and a level of structural assessment is
needed when retrofitting green roofs, however the numerous benefits
afforded by them can outweigh the initial cost when a Life Cycle Analysis
(LCA) is performed (Saiz et al. 2006; Kosareo and Ries 2007). Installation
costs range from £60 m-2 for extensive roofs to £140 m-2 for intensive (Green
Roof Centre 2013). There may also be costs for structural assessment work
on retrofitted roofs (~ £1,500), and maintenance costs, which can be
substantial for intensive roofs. Energy use reduction is the critical factor for
environmental impact and long-term cost saving in LCAs (Kosareo and Ries
2007). Green roof technology and development are strongest in the
German-speaking countries of Europe, however, even there, interest in green
roofs as a serious architectural component only really took hold in the 1980s,
spurred on by the large social force that is the rise of environmentalism
(Werthmann 2007). Architects worldwide are increasingly aware for the need
for environmentally responsible design and green roofs are a useful resource
for meeting environmental standards (Stang and Hawthorne 2005). They
can also help buildings gain higher stature in sustainable building rating
systems (Kula 2005) such as the Building Research Establishment
Environmental Assessment Method (BREEAM) system used worldwide,
mostly due to their ability to reduce summer air conditioning energy use
(Sfakianaki et al. 2009).
3.2 Green roofs as adaptation strategies for high temperatures
The cooling properties of green roofs are generally less studied than those of
parks and trees (Bowler et al. 2010), however they have been researched
using both empirical (Köhler et al. 2003a; Spolek 2008; Teemusk and Mander
2009) and modelling studies (Barrio 1998; Niachou et al. 2001; Alexandri and
Jones 2008) to assess their role as adaptation strategies for high urban
temperatures. Climate modification mechanisms are similar to those of
urban vegetation discussed previously; raising the albedo, shading of
building materials from solar radiation and evapotranspirative cooling. Fig.
3.4 is a schematic diagram of these processes. On a conventional roof (left),
shortwave solar radiation is absorbed by the building material and re-radiated
41
as longwave radiation at night, contributing to the night-time UHI. Little water
is evaporated on the conventional roof, however the water content of green
roofs (right) allows for greater evapotranspirative cooling (blue arrow) which
lowers the sensible heat flux by partitioning more of the inputs into latent heat.
Plant leaf albedo can result in greater reflectance of the shortwave solar
inputs than on a conventional roof.
Fig. 3.4 – Schematic diagram of the different processes occurring during the day
(top) and at night (bottom) which characterise the thermal performance of a
conventional roof (left) and a green roof (right). (Author created)
42
Fig. 3.5 – Comparison of the energetic exchanges of a dry or wet green roof with a
traditional roof, starting with 100 incident solar radiation units – summer session
(Source: Lazzarin et al. 2005)
The results of a modelling and experimental study are summarised in Fig. 3.5
(Lazzarin et al. 2005). Solar absorption is simply the absorption, and
utilisation for photosynthesis, of incoming radiation by the vegetation (Posada
et al. 2012), and adduction is a term the author uses to sum convection and
radiation heat fluxes. The relative contribution of the main processes –
reflection, absorption and evapotranspiration – to the reduction in outside and
inside (building) heat flux can be clearly seen with, for example, outside
adduction for wet and dry roofs being 15% and 28% (respectively) of that on
a traditional roof. The soil heat flux is slightly complicated because higher
soil moisture levels in the wet green roof increase heat flux, but simultaneous
higher evapotranspiration rates are possible which lowers the heat flux. A
well-watered, intensive green roof could have considerable
evapotranspirative cooling during the day, as the rate of evapotranspiration
can be 70% to 80% of the evaporation from a free-water surface, such as a
lake (Erell et al. 2011). Extensive green roofs, however, would exhibit less
43
evapotranspirative cooling due to the thinner substrates being more water-
limited, but also because the plants used in these types of roofs tend to be
succulents which use Crassulacean Acid Metabolism (CAM) as part of their
photosynthesis. This means the plant stomata only open for gas exchange
at night, to reduce plant water loss, therefore closed stomata during the day
reduce evapotranspiration rates (Black and Osmond 2003).
A total of 22 papers investigating the cooling properties of green roofs are

presented and summarised in Table 2.1. Significant reductions were seen in
roof surface, ambient air and indoor temperatures when green roofs were
compared to standard roofs. Surface temperature reductions of up to 45 oC
(Liu and Baskaran 2003) were reported, with under-platform air temperature
reductions of 18oC (Simmons et al. 2008) and air temperature at 10cm
reductions of 0.7oC possible (Jim 2012) Most studies were conducted in
temperate zones with just one study in the northern latitudes of the Baltic Sea
area (Teemusk and Mander 2009) and a few studies in the tropics (Wong et
al. 2003; Wong et al. 2007; Jim and Peng 2012; Liu et al. 2012).
44
Country Location Climate Method Roof type Main finding Reference
(Köppen classification)
Canada Ottawa Snow. Empirical, roof Extensive - grass Green roofs lower the roof Liu and
membrane temperature by up to Baskaran 2003
Fully humid, warm summer o
45 C and also modify the
temperature fluctuations
experienced
Halifax Snow, warm summer Empirical, roof Extensive – grass Roof surface temperature reduced Macivor and
o
continental and wildflower by 3.44 C on average lundholm 2011
China Taiwan Humid, subtropical Empirical, roof Extensive – Maximum surface temperature Liu et al 2012
simulation various difference between bare roof and
o
under vegetation of 17.1 C
Estonia Tartu Snow. Empirical, roof Extensive – sedum Summer temperature fluctuations Teemusk and
decreased by green roofs. Mander 2009
Fully humid, warm summer
Germany Neubrandenberg Warm temperate. Fully Empirical, roof Extensive – sedum Surface temperature differences of Köhler et al.
o
humid, warm summer 6 C between green and gravel 2003a
roofs
Greece Athens Warm temperate, hot, dry Empirical and Extensive – Energy savings from reduction in Santamouris et
summer modelling, roof wildflower cooling load al. 2007
45
Empirical and Extensive – mixed Energy savings from reduction in Sfakianaki et
modelling, roof cooling load al. 2009
Empirical, roof Extensive Lower indoor temperature of Niachou et al.

buildings with green roofs 2001
Hong Hong Kong Humid, subtropical Empirical, roof Extensive Surface temperature reduction of Jim & Peng
o
Kong 5.2 C, and air temperature at 10 2012
o
cm reduced by 0.7 C
India Yamuna Nagar Warm temperate. Dry winter, Empirical and Extensive – herbs Indoor air temperature reduced by Kumar and
o
hot summer modelling, roof 5.1 C when green roof compared Kaushik 2005
to bare roof
Italy Trieste Warm temperate, hot, dry Empirical, roof Extensive Maximum temperatures under the Nardini et al.
summer simulation green roof modules were 15 – 2012
o
20 C lower than on the bare roof.
Japan Osaka Warm temperate, fully humid, Empirical, roof Extensive Increased humidity and decreased Harazono et
hot summer simulation temperature on green roof al. 1990
Empirical, roof Extensive – grass Surface temperature of roof Onmura et al.

o
material lowered by 30 C on green 2001
roof
Tokyo Warm temperate, fully humid, Empirical and Extensive – turf Inside air temperatures lower for Takakura et al.
hot summer modelling, roof green roof than bare roof 2000
simulation
46
Kobe Warm temperate, fully humid, Empirical, roof Extensive – turf Sensible heat flux on green roof Takebayashi
hot summer simulation lowered by evapotranspiration and Moriyama
2007
Singapore Singapore Equatorial, fully humid Empirical, roof Extensive – Ambient temperature differences of Wong et al.
o
various 6 C between before and after 2007
installation
Empirical, roof Intensive – various Ambient air temperatures reduced Wong et al.
o
by 4.2 C by presence of plants 2003
UK Manchester Warm temperate. Fully Modelling Unknown Predicted surface temperature Gill 2007
o
humid, warm summer decreases of up to 7.6 C (2080s
high emissions scenario, town
centres)
USA New York Warm temperate. Fully Empirical, roof Extensive – sedum Temperature fluctuations Gaffin et al.
humid, hot summer experienced by the roofing 2009
membrane reduced by green roofs
Austin Warm temperate. Fully Empirical, roof Extensive – native Inside air temperature reduced by Simmons et al.
o
humid, hot summer simulation perennials 18 C under green roof 2008
Portland Warm temperate. Dry, warm Empirical, roof Extensive – Ecoroof - heatflux reduced by 72% Spolek 2008
summer various in summer
o
Michigan Snow, warm summer Empirical, roof Extensive - sedum 5 C surface temperature difference Getter et al
continental to gravel roof 2011
Table 3.1 – Summary of findings from studies on the cooling effect of green roofs worldwide. (Köppen classification from Kottek et al. 2006)
47
Only one study looked at an intensive green roof (Wong et al. 2003). This
reflects the dominance of extensive roofs due to lower installation and
maintenance costs, however more research is needed on the cooling
properties of intensive roofs to demonstrate the expected higher cooling rates
resulting from greater evapotranspiration. The specific performance
expectations of intensive roofs means further research on them is required.
Six of the studies used constructed simulations of green roofs (Harazono et
al. 1990; Takakura et al. 2000; Takebayashi and Moriyama 2007; Simmons et
al. 2008; Liu et al. 2012; Nardini et al. 2012), which may not adequately
represent the processes of a real green roof or anticipate modification of
cooling effects as a result of real world conditions such as a combination of
poor maintenance and a long, dry weather period.
The Gill (2007) paper which focuses on the climate of Manchester found a
dramatic effect of roof greening on surface temperatures, preventing
temperatures exceeding the 1961-1990 baseline for all time periods and
future emissions scenarios in the model. There is no mention, however, of
the details of the green roof parameters used in the model such as whether
extensive or intensive roof designs were used, whether artificial irrigation was
included for summer months, which green roof taxa were considered or the
spatial increase of green roof cover over present day levels. The work of
Cavan and Kazmierczak (2011), which used an energy exchange model and
a ‘deep green’ scenario consisting of significant increases in greenspace in
the Oxford Road Corridor, found rising surface temperatures resulting from
climate change were significantly ameliorated. Interviews with Corridor
stakeholders suggested there is a good chance for implementation of some
elements of the ‘deep green’ scenario.
The cooling effect of green roofs discussed above reduces heat transfer
through the roof, thus reducing the heating and cooling energy consumption
in a building (Martens et al. 2008), i.e. they provide an insulation function.
The indoor air temperature reductions seen in table 3.1 can provide direct
energy savings, as every decrease in internal temperature of 0.5 oC may
reduce electricity use for air-conditioning by up to 8% (Dunnett and Kingsbury
2004). Because buildings represent 65% of total urban electricity
48
consumption, significant energy savings could be achieved from the wide-
scale greening of roofs (Kula 2005). Urban electricity use in summer could
also be reduced by the effect that large scale roof greening may have on the
UHI, with a 25% reduction in electricity use suggested by Akbari and
Konopacki (Akbari and Konopacki 2005) from the indirect effects on the UHI
of cool roof installation and increasing the amount of urban vegetation. This
highlights the influence green roofs may have on the local microclimate, as
opposed to reduced heat flux into individual buildings, and this area is largely
under-studied.
3.3 Green roofs as SUDS
Roof surfaces form a large proportion of the urban impervious area

(VanWoert et al. 2005) and vegetated roofs can help reduce the increased
rainwater runoff that this causes by acting as a buffer for heavy rainfall,
storing it in the soil substrate and returning moisture to the atmosphere via
evapotranspiration. A large amount of research has been undertaken in
Germany with the first mention of the water-retention capacity of green roofs
appearing in 1985 (Ernst and Weigerding 1985). The substrate layer confers
the bulk of the storage capacity and 25 litres of water can be retained per m 2
of a 60 mm thick extensive green roof (Scholz-Barth 2001). The storm water
runoff from green roofs is not just reduced in volume but the runoff event is
extended in duration, well beyond the duration of the rainfall itself (VanWoert
et al., 2005) which can help lesson the damaging effects of flash floods in
built up areas. This is clearly seen in Fig. 3.6, with significant delay to the
onset of runoff and significant reduction in overall runoff compared to the
traditional roof. A delay in initiation of runoff of around four hours can be
seen in Fig 3.6. Runoff peak flows compared to bare roof peaks have also
been found to be delayed up to 10 min for 57% of results (Berndtsson 2010).
Outflow delay, calculated as difference between hyetograph and hydrograph
centroids, ranged from 50 to 148 min for events producing an outflow peak
greater than 1 l/s (Fioretti et al. 2010).
49
Fig. 3.6 – Cumulative runoff from a green roof and non-green roof in Belgium during
the 24 hour period of a 14.6 mm rain shower. (Source: Mentens et al. 2006)
Fig 3.7 shows the runoff retention capabilities of four different roof surfaces
as determined in a meta-analysis by Mentens et al. (2006). Note the
dominance of studies on extensive over intensive green roofs as evidenced
by sample sizes. The green roofs clearly allow less of the annual rainfall to
become runoff, with the intensive roofs outperforming extensive ones. The
meta-analysis of data from 18 German publications allowed the derivation of
empirical models which found the annual rainfall-runoff relationship is
strongly determined by the depth of the substrate (Mentens et al. 2006)
which has been confirmed in other studies (Getter and Rowe 2006; Rowe et
al. 2006). Further studies on intensive roofs, in different climates, could
increase the evidence base for the increased retention afforded by deeper
substrates.
50
Fig. 3.7 – Annual runoff as a percentage of total annual rainfall for intensive green
roofs (‘int’, n=11), extensive green roofs (‘ext’, n=121), gravel-covered roofs (‘gravel’,
n=8) and non-greened roofs (‘trad’, n=5). (Source: Mentens et al. 2006)
A number of factors influence the runoff quantity such as slope of roof (Getter
et al., 2007), and green roof plant composition (Dunnett et al. 2008b) due to
differing evapotranspiration rates (Farzaneh 2005). The Dunnett et al (2008b)
study was an attempt to investigate how vegetation type can influence green
roof function, however it had to rely on roof simulations, due to the lack of
species diversity on existing green roofs and the need for control.
Runoff retention is also influenced by the size of the rain event with greater
retention for small rainfalls (Hilten et al. 2008; Simmons et al. 2008; Stovin et
al. 2012) and time of year as evapotranspiration is higher in summer leading
51
to shorter ‘recharge’ periods of the retention capacity (Mentens et al. 2006;
Stovin et al. 2012). Rainwater retention benefits are expected to be of major
importance in the tropics where simultaneous high temperatures and high
rainfall are common (Simmons et al., 2008).
While the retention capabilities of individual roofs are impressive, their

performance at the urban catchment scale becomes more modest, due to
limitations of the number of buildings that can support a green roof and
therefore the proportion of surrounding impervious surfaces dominates the
catchment hydrology. Mentens et al. (2006) found that runoff at the city scale
could be reduced by 2.7% if 10% of the roofs of Brussels were greened with
an extensive green roof (100 mm substrate depth). Therefore green roofs
are increasingly considered as SUDS for storm water management alongside
other technologies such as porous pavements and ponding basins (Frazer
2005) which aim to lessen the impact of urban development on receiving
water bodies (Carter and Jackson 2007).
3.4 Green roofs as remediators of air pollution
Green roofs have the potential to be an effective sink for air pollutants by
providing a large surface area for wet and dry deposition (Hill 1971).
Subsequent pH-dependent sequestration of pollutants such as heavy metals
within the soil layer (Kandpal et al. 2005) or sequestration within plant tissues
as a form of phytoremediation (Vangronsveld et al. 2009) can remove these
pollutants from the urban system. The vegetation canopy-atmosphere
interface is reflected in a parameter called the Leaf Area Index (LAI) which
incorporates the varied leaf morphologies, sizes, orientations etc. (Breda
2003). The vegetation interacts with the turbulent transfer of the pollutants
within the air mass (Lovett 1994). Particles and aerosols are then captured
by the vegetation surfaces through the processes of Brownian diffusion,
interception, impaction and sedimentation under gravity (Petroff et al. 2008).
Once adhered to leaf surfaces, pollutants can then be removed by
resuspension to the atmosphere, washed off in rains, or enter the substrate
layer when a leaf senesces (Nowak et al. 2006). Gaseous air pollutants are
primarily taken into the plant via leaf stomata, where they diffuse into
52
intercellular spaces and are absorbed by water films to form acids or can
react with inner-leaf surfaces (Smith 1990). The number of factors that can
potentially regulate pollutant removal is enormous, ranging in scale from
meteorology and aerosol chemistry through to leaf micromorphology and
canopy structure (Lovett 1994). For this reason the factors that control
deposition are conceptualised as resistances that reflect transport rates
through vegetation canopies (Lovett 1994). Intensive green roofs, with their
larger vegetation canopies, would be expected to have higher resistances
and higher pollution capture rates.
Previous studies on air pollution removal by green roofs are scarce and rely
heavily on modelling, with the parameters for pollutant deposition velocities
estimated from other studies on grass and ivy which were themselves not
determined in a green roof setting (Pugh et al. 2012), or estimating
deposition velocity as the inverse of plant canopy aerodynamic resistance
(Yang et al. 2008).
These modelling studies have nonetheless indicated promising results.

Yang et al. (2008) found that 1675 kg of air pollutants, such as NO2, SO2 and
PM10 , was removed by 19.8 ha of green roofs in one year with O3 accounting
for 52% of the total. A study in Toronto found that 58 metric tons of air
pollutants could be removed if all the roofs in the city were converted to
green roofs, however intensive green roofs have a higher impact than
extensive green roofs, with shrubs being almost as effective as trees at
pollutant removal within the Urban Forest Effects (UFORE) model (Currie
and Bass 2008). Pugh et al. (2012) included green roofs, as well as green
walls, in a model which found street-level concentrations of NO2 and PM10
could be reduced by 40% and 60% respectively. As previously mentioned,
the processes involved in pollution capture and removal from the urban
atmosphere are complex so there is a need to validate these modelling
studies with empirical data from real green roofs. For instance LAI is often
reduced to a single component in models. In reality this masks the interplay
of many processes such as species differences in leaf morphology, leaf
senescence and growth throughout the year, which can only be effectively
captured in empirical studies.
53
3.5 Green roofs and runoff water quality
Green roofs are effective at capturing air pollutants, however if there is

incomplete sequestration of the pollutants within the sediment layer then
there is the possibility of merely trading air pollution for water pollution as the
pollutants enter the rainwater runoff (Rowe 2011). Additionally, green roofs
often require artificial fertilisation, which can impact negatively on the nutrient
content of the runoff (Monterusso et al. 2004; Emilsson et al. 2007).
Subsequently, studies on the quality of the runoff from green roofs have
produced mixed results.
A number of studies have found green roofs can improve the chemical quality
of roof runoff by absorbing the pollutants of wet and dry deposition and by
uptake of nutrients such as nitrogen (Berndtsson et al. 2006). Some soil
minerals can encourage the sequestration of metals such as Cu and Pb via a
number of mechanisms; from weak electrostatic attraction to stronger
covalent surface complexes (Sposito 1989). Soil characteristics that
influence sequestration include pH, cation exchange capacity, grain size and
organic matter, clay and CaCO3 contents (Hooda and Alloway 1998). With
respect to green roofs these characteristics are dependent on the age of the
green roof, substrate composition and level of fertilisation. Substrate depth,
plant composition and local pollution sources are also influential on the
sequestration ability of green roofs (Berndtsson et al. 2006; Teemusk and
Mander 2007). Köhler et al. (2003b) found 94.7% of Pb and 87.6% of Cd
inputs were being retained by green roofs with a recycled soil substrate
consisting of a loamy sand with rock fragments.
Conversely, green roofs have been demonstrated to be sources of heavy

metals (Berndtsson et al. 2006; Mendez et al. 2011). This can occur when
sites for sequestration within the substrate are limited or the material used to
construct the substrate layer is itself acting as a source of pollutants. A
recent study found commercial soil mixes were responsible for elevated
levels of Fe, Al and Cu in newly constructed roofs (Vijayaraghavan et al.
2012).
54
High total N and total P can be found in the runoff from green roofs with a
high organic content in the substrate (Moran et al. 2003) due to the high
solubility of their compounds and the dependence on the vegetation to
rapidly uptake these nutrients after application (Aitkenhead-Peterson et al.
2011). Controlled-release fertilisers are often recommended to overcome
this problem (Berndtsson et al. 2006). Dissolved Organic Carbon (DOC) can
also reach significant levels of up to 42 mg/l from leaching of solubilised
humic acids and the activity of soil bacteria (Berndtsson et al. 2009;
Aitkenhead-Peterson et al. 2011; Mendez et al. 2011).
Green roofs therefore should not necessarily be seen as a tool for improving
the water quality of runoff by retaining pollutants of wet and dry deposition.
Construction elements of green roofs and the use of fertilisers should be
considered and tested thoroughly (Berndtsson et al. 2009) to ensure
secondary treatment methods are not necessary to remediate poor runoff
water quality. However it must be noted that overall runoff quantity
reductions can serve to lower pollutant fluxes compared to a conventional
roof.
3.6 Other green roof benefits
3.6.1 Biodiversity
Most extensive green roofs are inaccessible to the public and therefore
provide an undisturbed habitat for microorganisms, insects and birds (Getter
and Rowe 2006). This has led to green roofs being considered as sites for
urban reconciliation ecology whereby the anthropogenic environment may be
modified to encourage biodiversity preservation without compromising
societal utilisation (Francis and Lorimer 2011). In a study of 17 green roofs in
Basel, Switzerland, 78 spider and 254 beetle species were identified.
Eighteen percent of the spider species were listed as endangered or rare
(Brenneisen 2003). Coffman and Davis (2005) found 29 insect and seven
spider species on a large green roof in Michigan, USA, within two years of
construction. Invertebrate populations provide food for birds and bats (Grant
et al. 2003; Baumann 2006) and in the UK there is growing interest in the
potential of green roofs to provide replacement habitat for the rare black
55
redstart, Phoenicurus ochruros, which has a Species Action Plan under the
UK Biodiversity Action Plan (Frith 2000).
In a study of collembolan populations on green roofs in Hanover, a highly

dynamic diversity was observed, similar to that found in extreme soil
environments (Schrader and Boning 2006). It can certainly be a tough
environment on a roof with severe droughts, extreme temperatures, little
shade and high wind speeds, and certain plant species are better adapted to
these conditions (Oberndorfer et al. 2007). This limits the choice of plant and
animal species that can thrive on a green roof and hence limits the
ecosystems that develop. Earthworms struggle to survive, for instance,
because they cannot migrate to cooler depths in hot summer periods
(Brenneisen 2006). Green roofs can promote urban biodiversity to some
extent but are by no means a replacement for nature as a functioning habitat
(Schrader and Boning 2006).
3.6.2 Carbon sequestration
Getter et al. (2009) found extensive green roof systems in the US could
sequester 375g C/m2 in above and below ground biomass and substrate
organic matter over two years. This depended on net primary production
exceeding decomposition, a process which is yet to be quantified for green
roofs, but extrapolation of the results to a scenario where all the commercial
and industrial roofs (ca. 7000 ha) in the Detroit metropolitan area are
greened showed 55,252 t of carbon could be sequestered (Getter et al. 2009).
This represents a small but significant potential for sequestering carbon in
urban environments and could potentially include green roofs under the
classification of climate change mitigation as well as adaptation.
3.6.3 Noise
It has been estimated that up to 20% of the population of the European Union
(80 million people) are subjected to noise levels which health experts
consider to be above acceptable limits (European Commision 1996). Green
roofs can play a part in mitigating high noise levels by absorbing sound
waves due to the nature of the substrate and vegetation (Getter and Rowe
56
2006). Studies are scarce, however a reduction of 10 dB is theoretically
attainable for an extensive green roof relative to an acoustically rigid roof
(Van Renterghem and Botteldooren 2008) and a 10 cm deep green roof on
Frankfurt airport, Germany, reduced noise levels by 5 dB (Dunnett and
Kingsbury 2004).
3.6.4 Electromagnetic Radiation
Electromagnetic fields associated with power supplies and microwave and

radio transmissions are considered to have negative health effects
(Mateescu et al. 2008) and as vegetation can absorb radiation, green roofs
can be considered as a strategy to reduce human exposure (Grant et al.
2003).
3.6.5 Aesthetics
Where access to a green roof is possible, usually in the case of intensive

green roofs, the leisure and amenity aspect of roof greening can be
considered a major benefit. Viewing green plants and nature has the positive
health effects on humans of reduced blood pressure and muscular tension
and increased positive feelings (Getter and Rowe 2006) which has been
found to aid recovery from surgery for example (Ulrich 1984).
3.6.6 Engineering benefits
By protecting the roof membrane from large diurnal temperature fluctuations

of up to 70oC, as experienced in summer in New York City (Gaffin et al. 2009),
green roofs can prolong the lifetime of these expensive roof membranes.
Many of the materials used in green roof construction such as aggregate and
compost for the substrate layer and plastics for the drainage mats can come
from recycled materials, thus avoiding charges incurred for disposing of
these materials in landfill (Grant et al. 2003).
3.7 Conclusions
Green roofs have important benefits for urban areas with relation to lowering
the high temperatures associated with the UHI, acting as a buffer for heavy
rain events and thus lowering the frequency of pluvial flooding and
57
reducing/sequestering air pollution. As a form of climate change adaptation
their benefits cannot be overlooked by city planners and this thesis aims to
quantify these impacts so that an assessment of the multi-beneficial nature of
green roofs can be made. A synopsis of the results will provide evidence of
green roof benefits for people who work in policy and practice.
58
CHAPTER 4 METHODOLOGY
This chapter will provide an overview of the thesis methodology and details of
the generic experimental methodologies used in the studies. To avoid
repetition, the complete methodologies for each experiment are only included
in the relevant chapter (Chapters 5 to 8). An experiment overview is
provided at the start of each section to briefly introduce the reader to the
main purpose of each study.
4.1 Overview
An empirical approach was used for this thesis and consisted of the following:
 Data collection via installation of monitoring equipment on green and

conventional roofs to record a suite of information such as air
temperature, runoff volume and atmospheric PM10 concentration.
 Data collection via spot monitoring and analysis of green roof soil
properties in-situ.
 Collection of samples of green roof soil and runoff
 Laboratory analysis of samples for a wide range of characteristics,
such as pH and heavy metal concentrations.
 Secondary data were collated from various external sources such as
air quality from a Defra monitoring station and aerial photography from
Landmap
The data generated by these techniques were then cleaned and statistically
analysed to provide quantitative assessments of some of the environmental
impacts of green roofs.
It was deemed suitable to use an empirical approach for this subject because
green roofs represent an alteration of traditional urban surfaces, which
produce consequent alterations to the local microclimate. These alterations,
such as changing local air temperatures and reducing levels of air pollution,
are large enough, and occur on a sufficient time-scale, to be measured with
equipment available currently. The locus for this study is the urban-canopy
layer (UCL), which extends from ground level to the height of buildings, at the
59
point where it blends into the roughness sub-layer of the urban atmosphere
(Erell et al. 2011). The scale of the inherent heterogeneity of the UCL,
arising from the proportions of building-density, urban vegetation etc.,
determines the micro-scale level of climate studies in this zone. This study is
operating at a scale which avoids the heterogeneity and complexity that
arises in urban climatological studies at larger or aggregate scales (Arnfield
2003).
Since Howard (1818) established himself as the pioneer of urban climate

studies with his comparisons of urban and rural temperature, leading to a
description of the magnitude of the London UHI, field measurements have
played a large part in urban climatology. The advancement of technology
over the years has allowed parameters such as wind velocity, turbulence,
evapotranspiration, albedo and air pollution levels to be quantified, thus
allowing the unique urban climate to be studied. Field measurements can be
costly, time consuming and influenced by a large number of factors (Mirzaei
and Haghighat 2010), but their importance for quantifying urban phenomena,
ground-truthing remote sensing work and validating modelling work is
unquestionable.
An empirical methodology allows for a number of research questions around

the impacts of green roofs to be addressed, thus enabling the main thrust of
this thesis to be a multi-impact investigation of green roofs. Measuring
parameters of interest, along with as many suitable explanatory parameters
as possible from the urban environment, forms the basis of this investigation.
All four of the traditional earth systems – lithosphere, atmosphere,
hydrosphere and biosphere (Montello and Sutton 2006) - are sampled within
an urban context in this thesis. A full empirical description of a system can
allow predictions to be made and potentially provide explanations for
phenomena observed.
Furthermore, by investigating green roofs in situ increases the external

validity of the thesis. Empirical measurements made under the exact natural
conditions of the green roof setting allows the verisimilitude, or ‘ecological
validity’, of the study to be enhanced by allowing phenomena to express
60
themselves within the totality of the context in which they exist (Montello and
Sutton 2006). This allows generalisations to green roofs in similar climates
and urban settings to be made, therefore increasing the relevance and
usefulness of the thesis.
Green roofs represent a dynamic system. In order to capture this, a

developmental design which uses repeated measures through time in a static
location is considered optimum. Additionally using a between-case research
design for some of the green roof impacts allows comparisons with the
microclimate impacts of traditional roof surfaces to be made. The
conventional roof provides an urban baseline over which green roofs can
produce their measurable impacts.
Finally, this thesis does not employ the modelling approach common to urban
climatology studies in order to avoid the abstraction and simplification of the
real world inherent in modelling (Huggett 1993). Mathematical models have
practical limitations and uncertainties resulting from the subdivision of spatial
domains and the choice of mathematical methods for solving system
equations for example (Huggett 1993).
4.2 Study site
The location for all the experimental investigations of this thesis is the city of
Manchester. As mentioned in Chapter 1, Manchester is representative of a
large conurbation in the UK and in Northern Europe, which has been shown
to exhibit an UHI and significant transport-related air pollution. Roofs were
chosen within the Oxford Road Corridor because it provides an inner-city
focus for the study but also because The Corridor Partnership is keen to
invest in green infrastructure improvements in the area (Chapter 1) so the
results will be of local interest.
61
Fig. 4.1 – The locations of the study roofs. (Author created)
The locations of the roofs, and of the Firs greenhouse which was used in one
of the studies (Chapter 5), are shown in Fig. 4.1. Roof 1 is a 3 storey office
block with a conventional, bitumen type, roof membrane. Its location next to
a busy inner-city motorway made it a suitable choice for the study in Chapter
5. Roof 2 is known as the Precinct, because it consists of an indoor
shopping precinct which spans Oxford Road within a bridge. The 2 storey
roof is ideal for comparative studies (Chapters 6 to 8) because it consists of
an intensive green roof adjacent to a large paved roof surface. Both sections
receive the same solar radiation, winds, thermal inputs from the building
below, and neither roof lies in the rain shadow of taller buildings. The SVF
for each section was found to have a value of one using a Canon 350D
digital camera equipped with a Canon wide angle fisheye lens. This confirms
62
the open roof top location free from the physical influence of surrounding
buildings.
Fig. 4.2 shows a 3-dimensional impression of the Corridor with the locations
of roofs 1 and 2, produced using LiDAR data (Landmap) within the ArcScene
software (version 9). The roofs can be seen to be approximately similar in
height to surrounding buildings so no major effects from localised altered
wind characteristics are expected
Fig. 4.2 – 3-dimensional view of the Oxford Road Corridor with roof locations. Author
created)
Scale diagrams of the precinct roof layout were provided by the Estates
Department of the University of Manchester (Fig. 4.3). A larger version is
available in Appendix I, along with a roof plan for Roof 1. The diagrams were
used to calculate the area of the roof sections which are serviced by drains
investigated in Chapter 7. The simplified roof plan in Fig. 4.4 shows the
drainage catchments. Based on the dates on the roof plan in Appendix I and
personal communication from The Estates Department, the year of roof
construction is estimated at 1970, making the roof 43 years old at the time of
writing.
63
The green roof on the Precinct is classed as an intensive roof because it has
a soil layer 170 mm deep which supports a wide variety of flora and fauna. A
full botanical survey was carried out and the results can be found in Appendix
II. The vegetation is dominated by grasses (Agrostis stolonifera and Festuca
rubra) and herbaceous species (Ranunculus sp., Plantago lanceolata,
Senecio jacobaea, Anthoxantum odoratum and Rumex crispus amongst
others) with large individuals of Buddleja davidii and Rubus fruticosus.
Fig. 4.3 – Scale drawing of the Precinct Building roof (UoM 2007 - larger version in
Appendix I)
64
Fig. 4.4 – Roof plan of the precinct. (Author created)
65
4.3 Air pollution remediation
4.3.1 Experiment overview
To quantify the particulate pollution capture by green roof vegetation, trays of

four plant species (Fig. 4.5) were grown from seed or washed cuttings in a
magnetically clean greenhouse at Firs botanical gardens (Fig. 4.1). The
species chosen were Agrostis stolonifera, Festuca rubra, Plantago lanceolata
and Sedum album plus three in-situ species from both study roofs. The four
main species were chosen because they represent some of the vegetation
types that may be found on real green roofs. Sedum, especially, was chosen
due to it being the main plant used in the extensive green roof industry, and
therefore its ability to capture air pollution will be of great interest to green
roof professionals. Once established, the trays were placed on one of two
roofs in Manchester city centre, or left in the greenhouse as a control.
Samples were taken periodically over five weeks and analysed using
magnetic analysis techniques to quantify the particles captured on the leaf
surfaces. Scanning Electron Microscopy (SEM) of leaf surfaces enabled a
qualitative assessment of the micro-morphology to be made. Elemental
analysis of leaf material and tray soil, at the end of the study, was carried out
to provide information on the presence of heavy metals in these
compartments.
66
Fig. 4.5 – Trays of S. album (left), A. stolonifera (right) and F. rubra (far right, front)
ready to be transplanted on to the study roofs. (Author image)
4.3.2 Site selection
Roofs 1 and 2 were chosen for this study. The two roofs were chosen
because their locations allow the effects of proximity to vehicular sources of
pollution to be investigated.
Fig. 4.6 – Panoramic view to the south west of Roof 1. (Author image)
Roof 1 is in close proximity, and similar height, to an inner city elevated

motorway located to the south west (Fig. 4.6). A Kestrel 4500 weather meter
67
was placed on Roof 1 throughout the study to record the wind speeds and
directions. South-westerly winds dominate, as can be seen in Fig. 4.7.
These winds would be expected to carry vehicle-derived pollutants from the
motorway to Roof 1. Roof 2 is not located next to such a major source, thus
represents urban background pollution levels.
Fig. 4.7 – Aerial photograph of Roof 1 (outlined) with directional wind rose for the
period 21/06/2011 – 04/08/2011. (Source: Googlemaps and Author created)
The location of the control greenhouse, at Firs Botanical Gardens, was

considered suitable because of its peri-urban location, 200 m from a major
road. The land-use around the greenhouse is mostly residential, sports fields
and minor access roads with low traffic levels (see Fig. 4.8).
68
Fig. 4.8 – Firs Botanical Gardens (outlined) within the local setting. (Source:
Googlemaps)
4.3.3 Magnetic techniques
To investigate the quantities of PM10 captured by the leaves of the study

plants, magnetic analytical methods were chosen because a strong
correlation has been demonstrated between leaf Saturated Isothermal
Remanent Magnetisation (SIRM) and the presence of anthropogenic
ferromagnetic particles from vehicular combustion and industry (Matzka and
Maher 1999; Maher et al. 2008; Mitchell and Maher 2009). This allows area-
normalised leaf SIRM to be used as a proxy for ambient PM10 levels. The
technique is fast, sensitive and cheap and has been used to produce
unprecedentedly high spatial density pollution level data (Matzka and Maher
1999; Muxworthy et al. 2003; Szonyi et al. 2008; Kardel et al. 2012) and can
also be used for source attribution (Hansard et al. 2011). The method
involves imparting SIRM on the samples in a high magnetic field of 1 Tesla in
a Molspin high-field pulse magnetizer (Walden 1999). The SIRM imparted on
the samples is then measured with a Molspin magnetometer. Three readings
were obtained for each sample, and an average taken. After every fifth
69
sample the magnetometer was re-calibrated with a magnetic standard of 779
mA/m.
The duration of plant exposure is 5 weeks. This duration is similar to the

lengths of previous studies (Maher et al. 2008; Mitchell and Maher 2009) and
allows for sufficient time for particle deposition processes to occur, however a
longer exposure time would allow for equilibrium processes to be more fully
resolved. The exposure time was limited primarily by the amount of samples
obtainable from the trays of vegetation. A summer exposure time was
chosen because of the higher probability of warm, dry weather with
associated increases in air pollution levels. Other pollutants such as NOx
and O3 were not considered due to financial constraints on air and leaf
monitoring techniques for gaseous pollutants.
4.3.4 Leaf area estimation
In order to minimise handling of the leaf samples, a regression equation

relating dry leaf weight to leaf area for the seven plant species was obtained
from 10 samples of each species before the study began. Leaf area was
estimated by placing the leaf samples flat under glass on millimetre graph
paper. A photograph was taken from directly above with a high resolution
digital camera and the area calculated from the images via pixel counting
using Adobe Photoshop (CS version), similar to the method used by Patil and
Bodhe (2011). True leaf area is calculated by using the number of pixels in a
1 mm2 area of the graph paper in each image as a calibration. The samples
were dried for two days in an oven at 40oC. S. album was dried for three
days due to it being a succulent and thus highly efficient at retaining its water
content. S. album has cylindrical shaped leaves so a correction factor was
derived to convert surface area as viewed from above to the actual surface
area of the tapered cylinder based on a simplification of the leaf structure as
a cylinder with hemispherical ends (Fig. 4.9). Length and width were
measured for 50 leaves and used to derive the correction factor (multiply by
3.16) which converts the flat surface area (as estimated with the pixel-
counting method) to a more accurate three dimensional surface area.
70
Fig. 4.9 – Simplification of the sedum leaves for 3-dimensional surface area calculation
(Author created)
The R2 coefficients in table 4.1 suggest the linear regression equations were
acceptable for obtaining leaf area from the samples. The lower values for F.
rubra and S. album reflect difficulties of estimating the leaf area for these
species as a consequence of leaf morphology. The area-estimation method
for S. album described above may not account for natural variability in leaf
shape. The leaves of F. rubra were very thin and difficult to efficiently flatten
under the glass.
Species Regression R2
Agrostis stolonifera 0.96
Festuca rubra 0.75
Plantago lanceolata 0.95
Sedum album 0.81
Platanus acerifola 0.89
Symphyotrichum novae-angliae 0.99
Hypericum inodorum 0.99
Table 4.1 – Regression coefficients for the linear relationship between dry weight and
leaf area for the study species
71
4.3.5 Scanning Electron Microscopy (SEM)
At the end of the study, single leaves of the four main study species were
collected from Roof 2, which represents urban background levels of PM10.
The samples were immediately taken to the laboratory for analysis to avoid
desiccation and subsequent alteration of leaf morphology (Stabentheiner et
al. 2010). The samples were mounted on carbon stubs and sprayed with a
~20 nm thickness gold layer in preparation for Environmental SEM (FEI
XL30). This method is suitable for preserving the natural morphology of cells
and tissues in fresh samples (Pathan et al. 2010). Energy Dispersive X-ray
(EDX) analysis of a cluster of particles was undertaken to provide information
on the spatial distribution of elements within the particles.
Around 25 images were created from randomly selected areas of the leaf
surface of each species (upper and lower leaf surfaces both included) at X
3000 magnification level. These images were used to count the particles per
unit area of leaf surface, which provided a different estimate of particle
capture rates to the SIRM method. The number of particles per unit leaf area
accumulated over the time period that the plants were on the roof, assuming
no particles were present before roof placement and minimal particles
washed off, gives a capture rate. The leaf area per unit roof area is
estimated from the above-ground biomass in the trays by weighing the plant
material from a given area of tray and using the weight/area relationships
already determined. The particle capture rate and leaf area per unit area of
roof can then be used to estimate total particle capture by the total green roof
area in an upscaling scenario.
4.3.6 Upscaling
An area of Manchester, which comprises the city centre and the Oxford Road
Corridor, was selected for an investigation into the potential effects of
greening all the flat roofs in this area on the local air quality. A visual
analysis was carried out using ArcMap (version 9), which consisted of
drawing polygons to correspond with flat roof area in a geo-referenced aerial
photograph (Landmap). Fig. 4.10 shows the distribution of flat roofs in the
selected area. The area of these flat roofs was used to calculate the area of
72
potential roof greening which equalled approximately 50 hectares (15.3% of
the selected area). The basis for selection was simply if the roof was flat and
no consideration given to the ability of the roof to support the weight of a
green roof.
Fig. 4.10 – A selected area of Manchester city centre and the Corridor for analysis of
flat roof area using ArcMap software. (Source: Landmap)
4.4 Urban Heat Island impacts
Air temperature was measured every 15 minutes at 300 mm over an

intensive green roof and over an adjacent bare roof to quantify differences in
the cooling properties of the different surfaces. The temperatures were
73
monitored for a total of 17 months. Spot monitoring was also undertaken
using a thermal imager to quantify differences in surface temperature.
4.4.2 Monitoring stands
Monitoring stands were constructed to house Tinytag temperature sensors

(Fig. 4.11) inside a radiation shield based on the traditional Stevenson shield
(Fig. 4.12). The Tinytag sensors (2PLUS TGP-4500) were selected because
they have an acceptable level of accuracy (±0.4oC) while being affordable
and small. The design of the shield is similar to that in the low-cost approach
employed by Cheung et al. (Cheung et al. 2010) which was found to greatly
reduce any influence of solar radiation, and performed well in comparison to
a standard Stevenson shield. The stands were constructed from 2” X 2”
wood and painted white to minimise convective heat inputs into the radiation
shield. The sensors hang from a hook inside the shield so that they are 300
mm above the roof surface. This height was chosen as it has been used in a
number of previous studies on urban air temperature (Wong et al. 2003;
Gaffin et al. 2009) and is considered suitable for measuring the near-surface
air temperature without being too close so as to be dominated by convective
near-surface heat flux or too far to be dominated by mesoscale urban air
temperatures. The stands were placed centrally within their respective roof
area to avoid edge effects and to minimise advection influences from
adjacent surfaces.
74
Fig. 4.11 – Tinytag temperature sensor shown to scale. (Source: Tinytag 2013)
At the start of the study all the Tinytag sensors were checked against each
other for measuring consistency. All the batteries were replaced and the
sensors were kept in the same outdoor location for one week. The average
standard deviation was 0.14oC which is well within the accuracy level. A
Rotronics Hygroclip, which has a higher accuracy of 0.2oC, was placed in the
same environment for a comparison. The correlation R between the data
from the Rotronics and from the Tinytags was 0.98 on average and the
values differed by just 0.01oC – 0.4oC with an average difference of 0.1oC.
This was considered an acceptable level of intra, and inter monitor, variation.
Humidity data from the Tinytags were prone to measuring error, frequently
reporting 0% and 100% humidity, so data were not used.
A recording frequency of 15 minutes is suitable because it generates

sufficient data to capture the diurnal air temperature cycles. Higher recording
frequencies can produce an excessive amount of data for analysis which can
outweigh the benefits of the higher resolution provided. The monitoring
period of 17 months allowed for some seasons to be monitored more than
once thus enabling a wider variety of weather types to be encompassed
within the study.
75
Fig. 4.12 – Radiation shield and stand. (Author image)
4.5 Runoff quantity
Rainfall runoff was quantified over a 13 month period on the intensive green
roof and adjacent bare roof on the precinct roof. Using rainfall input data,
estimates of the retention efficiency of the green roof can be compared to
those for the bare roof. A large part of the hydrological properties of a green
roof are linked to the substrate layer, therefore physical and chemical soil
characteristics were also determined.
4.5.2 Monitoring
Due to the fixed nature of the roof, it was necessary to devise a method for
monitoring runoff by adapting to the existing roof structure. This is because it
was not possible to collect runoff from the bottom of the drain pipes or to
76
divert the flow into a measuring set-up. Therefore a novel approach was
called for. Two general approaches have been previously used to monitor
green roof runoff flow (Taylor Associates 2006):
 Calibrated primary flow devices (tipping buckets, weir boxes, flumes,

orifice restrictor devices)
 In-line flow meters
With calibrated weir boxes, the water level is monitored using a pressure
transducer. This method was used in the present study, however a
modification of the v-notch weir system was developed which allowed the
weir to sit within a roof drain in the form of a modified pot. Fig. 4.13 shows a
schematic diagram for the v-notch pot and Fig. 4.14 is a photograph of the
apparatus. Two pots were constructed – one for each of the study roofs.
Fig 4.13 – Schematic diagram for the measuring of runoff. (Author created)
77
Fig. 4.14 – V-notch weir pot and pressure transducer. (Author image)
Calibration of the pots was carried out in a laboratory by recording the

transducer pressure readings for a range of known water flows. The
calibration equations are second-order due to the widening of the v-notch at
greater heights of water above the datum. The R2 coefficients for the
calibration equations are good, with R2 = 0.91 for the green roof pot and R2 =
0.80 for the bare roof pot. See Appendix III for the regression plots. To
obtain the flow rate for each rain event, the pressure reading at H0, the height
of water in the pot just before flow starts, is adjusted to equal the pressure
reading at H0 for the calibration experiments, and then the calibration
equations are applied to convert the pressure reading into a flow rate.
Integration of the area under the runoff hydrograph gives the total volume of
runoff for that rain event. Discharge from the pots on the roofs was free.
Both roof drainage areas include sections of a central glass atrium on the
roof (Fig. 4.4). This means that the green roof drainage area is not entirely
green and there is a significant input (20% of the area) from a non-green
78
surface. In order to estimate the amount of runoff from this sloping glass roof
that is contributing to runoff on the green roof, rainfall-runoff relationships
were determined. This was undertaken for a range of different size rain
events. The runoff was measured using a tipping rain bucket. For each rain
event in the study the atrium runoff was subtracted from the overall runoff
measured in the drain to gain the runoff attributable to the green component
of the roof. Fig. 4.15 shows the rainfall-runoff relationship for the atrium and
the high R2 coefficient.
Fig. 4.15 – Regression plot of rainfall and runoff for the glass atrium roof, with 1:1
line included
79
4.5.3 Soil analysis
The following methods were used to characterise the green roof substrate
layer:
 Soil texture – Organic material was destroyed by hydrogen peroxide

and the remaining mineral soil dispersed by shaking with sodium
hexametaphosphate. The soil was analysed by sieving in a nest of
sieves (Rowell 1994). Silt and clay fractions smaller than 63 µm were
further categorised in a Malvern Instruments particle size analyser
(Hydro 2000 G).
 Field capacity – Water-saturated soil samples were allowed to drain
under suction in a Buchner funnel for two hours and then the
gravimetric water content was determined by weighing pre and post
oven drying at 105 oC (Veihmeyer and Hendrickson 1949; Rowell 1994)
 Dry bulk density – Small cores of sediment were taken with a metal
cylinder of known volume. Bulk density is the mass of oven-dry soil
divided by sample volume (Rowell 1994)
 Porosity – Particle density was first found by measuring the volume of
a known mass of particles by heating a sample in water to remove any
air and then weighing a known volume of this suspension. Porosity is
then calculated by:
porosity = 1 – (bulk density / particle density) (Rowell 1994)
 Organic matter content – oven-dry soil was heated to 500oC and the
mass lost corresponds to the organic matter that has been burnt off
(plus a negligible amount of water trapped within crystalline matrices)
(Dean 1974; Rowell 1994)
 Saturated hydraulic conductivity – A Guelph permeameter measures
in-situ hydraulic conductivity. The apparatus uses a constant-head
principle to measure the steady-state influx rate of water into a
saturated zone adjacent to a small borehole (Salverda and Dane
1993). Due to the spatial variation in hydraulic conductivity that can
80
exist, ten boreholes were augured for measurement and an average
taken
4.6 Runoff quality
Water samples were collected from the runoff from an aged intensive green
roof and an adjacent bare roof for a number of rain events. The quality of the
runoff was analysed in a laboratory for a number of parameters. Soil
samples were taken from the green roof substrate and samples of dust were
taken from under paving slabs on the bare roof. This was to investigate the
potential of the sediments to act as sources of contaminants in roof runoff via
element leaching. The samples were analysed in triplicate using X-Ray
Flourescence (XRF) techniques for metal concentrations.
4.6.2 Water analysis
All water samples were collected in new sample bottles. Bulk rain samples
were collected via a large funnel with a rim diameter of 30 cm, which
channelled rain into a 1 litre sample bottle that had been acid washed (25%
acetic acid) and rinsed in deionised water. Samples were filtered using 0.45
µm syringe filters (VWR cellulose acetate) to remove the fractions classed as
solid or particulate and leave the dissolved fraction (Hall et al. 1996). All
samples were stored in a refrigerator in the dark ready for water analysis
which in all cases was carried out as soon as possible and, after allowing
samples to reach room temperature, consisted of the following:
 pH – A Hanna (HI 9124) pH meter was first calibrated with buffer

solutions of pH 4, pH 7 and pH 10. The electrode was cleaned with
distilled water between measurements.
 Conductivity – A Hanna (HI 9033) conductivity meter was first
calibrated in a solution with conductivity of 84 µS/cm to prepare the
meter for samples with a relatively low expected conductivity. The
electrode was cleaned with distilled water between measurements.
81
 Anions – Anions were measured using ion chromatography in
accordance with US EPA method 300 (Metrohm 882 Compact IC plus)
and the analytes consisted of Cl, NO3, PO4 and SO4.
 Colour – Visible light absorbance over the wavelength range 190 –
700 nm was recorded using a spectrophotometer (Hach Lange
DR5000). Absorbance at 400 nm can be used as a proxy for
Dissolved Organic Carbon (DOC) in accordance with other work
(Worrall et al. 2002; Wallage et al. 2006). The ratio of the absorbance
at 465 nm and 665 nm , also known as the E4/E6 ratio, can give an
indication of the nature of the DOC, with ratios of 8 – 17 indicating
fulvic acids and 2 – 5 characteristic of humic material (Thurman 1985;
Wallage et al. 2006).
 Metals - Subsamples of 20ml were acidified, after filtration, with high
purity nitric acid, to stabilise the samples ready for analysis for metal
ions using Inductively Coupled Plasma Mass Spectrometry (ICP-MS)
(Agilent 7500cx). This method has suitably low detection limits for
concentrations of metals in µg/l, as are expected from the runoff.
4.6.3 Substrate analysis
Twelve soil samples, a representative number for the size of the roof, were
taken from the surface of the green roof, with the sampling locations
maximally spaced over the green roof. These samples were analysed in a
laboratory for the following:
 pH and conductivity – Soil samples were shaken vigorously in

deionised water for 15 minutes and then pH and conductivity
measured as in 4.6.2
 Anions – Soil samples are freeze dried and then shaken thoroughly in
deionised water for an hour. After filtering through 540 grade
hardened ashless filter paper and then through 0.45 µm syringe filters,
the solutions are analysed for water-soluble Cl, NO3, NO2, PO4 and
SO4 using the method in section 3.5.2.
82
Green roof soil samples and dust samples from the bare roof were dried at
40oC and sieved to 125 µm to remove stones/twigs/leaves and homogenised
in a ball grinder (Fritsch Spartan pulverisette 0) and then analysed for metals
using XRF in accordance with US EPA method 6200. Each sample was
analysed three times with a Niton XL3t XRF analyser. Three 40 mm
diameter cores of the full substrate depth were taken from central areas of
the green roof sections (See Fig. 4.4 for locations). These were split into ten
equal fractions, prepared as described above, and analysed separately using
XRF to provide a depth profile of soil elements to see if elemental
composition changes with depth.
Energy dispersive isotope-source XRF is a rapid, non-destructive total

element analysis method which is sufficiently accurate to be used in
environmental research (Boyle 2000). The speed of the technique allows a
high number of samples to be measured repeatedly which overcomes
problems with heterogeneity, common in sediments and soils. Ex situ
detection limits with field portable XRF range from 10 to 100 mg/kg (Raab et
al. 2005).
4.7 Analysis
Most monitoring equipment generated data in a .csv (comma separated

values) format which was then exported into an Excel spreadsheet. Some
equipment, such as the Molspin magnetometer, required manual entry of the
results into a spreadsheet. Data were then checked for errors and
anomalous results using graphical methods. Data cleaning was carried out,
mainly with the large temperature and runoff volume time series, to get the
data into a form more conducive for analysis. Formula calculations, such as
normalisation of the SIRM results by leaf area, were carried out in Excel.
Data preparation was also carried out in Excel. For example, runoff data
were collected every minute and needed to be averaged into 10 minute
brackets to coincide with the rainfall data which was recorded every 10
minutes by the Whitworth weather station.
Data were exported to statistical software packages for normality checking

and statistical analysis. Data were generally not normal, therefore non-
83
parametric methods were called for. SPSS (version 16) was used initially,
however the bulk of the analysis was carried out using R (version 2.13.2).
Graphs were produced using Excel or R.
84
CHAPTER 5 URBAN PARTICULATE POLLUTION
REDUCTION BY FOUR SPECIES OF GREEN
ROOF VEGETATION IN A UK CITY
Authors:
Speak, A.F.1, Rothwell, J.J.1, Lindley, S.J.1, Smith, C.L.2
Contribution of main author = 90%. Co-authors aided via supervisory

guidance and proof-reading / minor editing of drafts
Affiliations:
1
Geography, School of Environment and Development, The University of
Manchester, Arthur Lewis Building, Oxford Road, Manchester, M13 9PL, UK
2
College of Science and Engineering, University of Leicester, Bennett
Building, University Road, Leicester, LE1 7RH, UK
Atmospheric Environment 61(0): 283 – 293

http://dx.doi.org/10.1016/j.atmosenv.2012.07.043
85
5.1 Abstract
Urban particulate pollution in the UK remains at levels which have the

potential to cause negative impacts on human health. There is a need,
therefore, for mitigation strategies within cities, especially with regards to
vehicular sources. The use of vegetation as a passive filter of urban air has
been previously investigated, however green roof vegetation has not been
specifically considered. The present study aims to quantify the effectiveness
of four green roof species – creeping bentgrass (Agrostis stolonifera), red
fescue (Festuca rubra), ribwort plantain (Plantago lanceolata) and sedum
(Sedum album) – at capturing particulate matter smaller than 10µm (PM10).
Plants were grown in a location away from major road sources of PM10 and
transplanted onto two roofs in Manchester city centre. One roof is adjacent
to a major traffic source and one roof is characterised more by urban
background inputs. Significant differences in metal containing PM10 capture
were found between sites and between species. Site differences were
explained by proximity to major sources. Species differences arise from
differences in macro and micro morphology of the above surface biomass.
The study finds that the grasses, A. stolonifera and F. rubra, are more
effective than P. lanceolata and S. album at PM10 capture. Quantification of
the annual PM10 removal potential was calculated under a maximum sedum
green roof installation scenario for an area of the city centre, which totals 325
ha. Remediation of 2.3% (±0.1%) of 9.18 tonnes PM10 inputs for this area
could be achieved under this scenario.
86
5.2 Introduction
The adverse effects on health due to atmospheric particulate pollution have

been the subject of a growing number of studies in recent years (Daniels et al.
2004; Tchepel and Dias 2011). Urban residents are at particular risk from
certain anthropogenic sources such as traffic, and a rapidly increasing urban
population worldwide gives these studies additional importance because
higher numbers of people will be exposed to urban particulate pollution.
Particles smaller than 10 µm in diameter (PM10) can penetrate deep into the
pulmonary passages where any transition metals present, such as iron and
copper, can release free radicals in lung fluid and cause cellular inflammation
(Birmili 2006).
Modern anthropogenic sources in urban areas include vehicular traffic fuel

combustion, particularly diesel, and vehicular component wear (Petroff et al.
2008; AQEG 2005; Bukowiecki et al. 2010). The environmental burden from
component wear such as brakes, and resuspended road dust particles, is
high because the particles are more likely to contain metals, polyaromatic
hydrocarbons (PAHs) and sulphides which are toxic or carcinogenic. UK
mitigation techniques have led to a decline in industrially sourced emissions
but vehicle usage is continually increasing (DfT 2009), thus offsetting the
impact of vehicular emission controls (AQEG 2005). Air Quality Management
Areas (AQMA) are implemented in areas where air quality objectives are not
likely to be met, wherein mitigation procedures are required (AQEG 2005).
Persistent pollution episodes in cities are more likely to develop in winter in

the Northern hemisphere because of a rise in space heating and the
frequency of atmospheric inversions (Birmili 2006). Climate change could
also cause an increase of episodes in summer, as one of the predicted
weather changes is an increase in the likelihood of stationary air masses
(AQEG 2007).
Cessation of vehicular sources is impractical so cities in developed countries

will have an ongoing need to mitigate PM10. Transitional nations, with
increasing traffic amounts, will need to plan for the future health of their
citizens. Using plants as passive filters is being seen as a practical
87
remediation method (Beckett et al. 2000; Freer-Smith et al. 2005; Litschke
and Kuttler 2008; Peachey et al. 2009). Trees and woodland have been
found to be significant sinks for gaseous, aerosol, particulate and rain-borne
pollutants (Fowler et al. 1989). Four processes are responsible for
deposition onto the large surface area provided by leaves – sedimentation
under gravity, diffusion and turbulent transfer giving rise to impaction and
interception (Petroff et al. 2008). The large leaf area and turbulent air
movement caused by their structure makes trees particularly effective for
particle removal (Fowler et al. 1989). However, space in cities is at a
premium and there are limited opportunities to implement urban greening
programmes with trees. Greening roofs is a viable and attractive alternative
solution as roofs can form up to 35% of the urban land area (MacMillan 2004).
While a large amount of studies have focussed on trees, this study will look
at the potential contribution of green roofs for PM10 mitigation which is a
preferable strategy to the exclusion of vehicles from cities (Litschke and
Kuttler 2008).
There is already some evidence for the potential of green roof vegetation for
air pollution removal. Yang et al. (2008) found that 1675 kg of air pollutants,
such as NO2, SO2 and PM10 , were removed by 19.8 ha of green roofs in one
year, with PM10 accounting for 14% of the total. This was a modelling study
which used deposition velocities for shrubs determined in a non-green roof
environment. A study in Toronto found that 58 metric tonnes of air pollutants
could be removed if all the roofs in the city were converted to green roofs,
with intensive green roofs having a higher impact than extensive green roofs
(Currie and Bass 2008). Intensive roofs have a deeper soil substrate than
extensive roofs, which allows for a larger above-ground biomass and a wider
variety of plants to be grown. While these studies offer promising results,
they are based on modelling alone and, to the authors’ knowledge, no
empirical investigation of green roof removal of air pollution has been
published.
Mitchell and Maher (2009) recently confirmed the magnetic characteristics of

tree leaves to be a suitable proxy for ambient air pollution. This method can
be used to increase spatial and temporal pollution monitoring resolution and
88
has been used a number of times in magnetic biomonitoring studies involving
trees (Matzka and Maher 1999; Maher et al. 2008; Hansard et al. 2011). To
date the technique has not been applied to quantify PM10 removal by green
roof vegetation.
The objectives of the study are twofold. Firstly to investigate spatial

differences in particulate load to two roofs that differ in their proximity to
pollution sources and secondly to elucidate any species differences in
particulate capture rates between several common types of green roof
vegetation used in the UK. With the choice of study sites, it is intended to
ascertain whether differences in particle capture exist when a significant
traffic source is located in close proximity to a roof, elevating local
concentrations, compared to a roof subject solely to urban background
concentration levels. The results will further be used to quantify the potential
removal of metal-containing PM10 by green roofs in an area of the city centre
of Manchester. Green roofs are increasingly being recognised as
investments that can help address many of the challenges facing urban
residents and, in recognition of this, Manchester City Council is developing a
green roof policy guidance document (MCC 2009). In addition, this work will
support a burgeoning literature on the numerous benefits of green roofs such
as reduction of Urban Heat Islands (Spolek 2008) and their role as
Sustainable Urban Drainage Systems (SUDS) (Scholz-Barth 2001).
5.3 Methodology
5.3.1 Site Description
Manchester is a large city situated in north-west England. The Manchester

city district, which includes the centre of the Greater Manchester conurbation,
has a population of 498 000 (MCC 2010). It has seen a gradual improvement
in its air quality since its industrial past; however PM10 levels remain high
near roadsides and in urban centres (HFAS, 2011). Indeed, the whole of
Manchester city centre and the Oxford Road corridor (Fig. 5.1) are within an
AQMA (Defra 2011a). Oxford Road is a key transport route within an area of
high economic activity and has an annual average weekday traffic flow of
11,529 motor vehicles (GMTU 2010). In 2007, the Manchester district
89
produced 257 tonnes of PM10, with roads being the major source (HFAS
2011).
Two roofs situated on the Oxford Road corridor were chosen as locations for
this study. Roof 1, Manchester Technology Centre (MTC), is a 3 storey office
block with a conventional bare roof, situated adjacent north to the point
where a busy inner city motorway, the Mancunian Way, crosses Oxford Road
(Fig. 5.1). Traffic flows on the motorway reach over 60,000 vehicles a day
(GMTU 2010). Tyre and brake wear emissions account for ~23% of total
road transport emissions and, along with resuspended road dust, appear to
be more important for heavy duty vehicles than cars (AQEG 2005). Urban
canyon effects can strongly influence urban wind patterns and the Oxford
Road corridor might channel these winds northwards. The location of Roof 1,
combined with prevailing south-westerley winds, means this site is expected
to receive a high load of road derived pollution and thus represents a roof
with a strong local source.
90
Fig. 5.1 – Map showing locations of the study sites with Firs control site to the south
of the city centre and the locations of Roofs 1 and 2 next to Oxford Road. The
shaded overlay, highlighting the city centre and major roads, indicates the current
extent of the Manchester AQMA (Defra, 2011)
91
Roof 2, the precinct bridge, crosses Oxford Road and consists of an
extensive paved area over the road. This 2 storey roof was chosen for its
proximity to the busy Oxford Road and represents a roof with an urban
background pollution source. A control site at the Firs Botanical Gardens is
situated 3.5 km south of the roof monitoring sites (see Fig. 5.1). This peri-
urban site is in a location 200 metres away from major road-sourced PM10 in
an area of high tree-planting. This control site is used to make comparisons
between the other sites by utilising enrichment ratios.
5.3.2 Species Selection and Sampling
Four perennial species were chosen for this study – S. album, F. rubra, A.
stolonifera and P. lanceolata. The Sedum species is a common choice of
plant for commercial extensive green roofs. F. rubra and A. stolonifera are
both common British grasses, which can grow on turf roofs, and P. lanceolata
is a common invasive species on green roofs (Dunnett et al. 2008a). Trays
of the four species were grown from seed, or washed cuttings (S. album), in
a ‘magnetically clean’ greenhouse at the control site. John Innes number 3
compost was used. Trays of mature plants were then placed on the two
study roofs on 03/07/2011. Additional trays were left in the greenhouse at the
control site. Trays containing just compost were also placed on Roof 2 and
in the control greenhouse to investigate soil metal concentration changes
when plants are absent. In dry periods, plant trays were watered with pre-
collected rainwater direct to the soil layer to avoid artificial rain effects of
washing particles from the leaf surfaces.
Leaf samples were collected from mature plants. The sampling frequency
was three times a week and all samples were collected in the mornings.
Sampling began on 5/07/11 and the final samples were collected on 03/08/11.
Enough leaves were collected to provide material for two samples per
species, with each sample sufficient to fill a 10 ml sample pot. Very young
leaves were not collected to ensure exposure times covered the entire study
period. The samples were transported to the laboratory in plastic bags where
they were dried in ovens for two days at 40°C (three days for S album),
weighed, and stored in plastic film for subsequent magnetic measurement.
92
Samples were also taken from vegetation growing on or near the roofs, to
investigate particulate capture by plants with prolonged exposure at the site
locations. In the case of Roof 1, this was from the roof-height tree canopy of
Platanus acerifolia and, for Roof 2, samples were taken from established
plants of Symphyotrichum novae-angliae and the evergreen Hypericum
inodorum, growing in an adjacent green roof.
Surface areas of leaves, for area-normalising of results, were obtained by

pixel counting images of scanned leaves and calculating regression
equations from the relationship between area and dry weight. Subsequent
surface areas could be calculated from sample weight. Estimates of above
surface biomass in the trays, useful for up-scaling pollutant removal
efficiencies, were made by visual calculation of number of samples per tray.
Ambient PM10 concentrations were recorded for half an hour at most

sampling events at all three locations using a TSI Sidepak AM510 personal
aerosol monitor. To determine the relative magnetic PM10 levels of the three
locations, pumped air samples of 100 l were collected from all three locations
at a rate of 1.4 l/min using the monitor in conjunction with an IOM sampler
head fitted with glass filters (1µm pore size Whatmann). The filters were
folded and placed in plastic pots ready for magnetic measurements (see
section 5.3.3). Ambient PM10 concentration data were also acquired from
Defra’s Automatic Urban and Rural Network (AURN) located in Piccadilly
Gardens in Manchester City Centre (Fig. 5.1). These data can be compared
with ambient PM10 levels recorded by the personal monitor to investigate any
temporal trends in PM10 occurring throughout the study and to assess site
differences. Rolling averages were calculated from the inter-sampling
periods prior to each sampling event to see if the atmospheric PM 10 levels
preceding sample collection influence leaf levels.
A Kestrel device was situated on Roof 1. This records wind speed and
direction, useful for determining prevailing wind patterns and hence
identifying upwind sources. Meteorological data were also available from the
Whitworth Observatory (UoM 2012), located adjacent to Roof 2. Rainfall was
93
recorded because rainfall can influence trends in leaf pollutant load as
particulate matter can be partially washed off.
5.3.3 Magnetic Measurements
Dried leaf samples were placed in 10 ml cylindrical styrene sample holders

and immobilized with packing film. Saturated Isothermal Remanent
Magnetisation (SIRM) was imparted on each sample in an applied field of 1
Tesla, using a Molspin Instruments’ high-field ‘Pulse Magnetizer’ (Walden
1999). SIRMs were then measured with a Molspin Instruments’ ‘Minispin’
fluxgate magnetometer. SIRM provides a measurement of the total
concentration of magnetic particles within a sample (Muxworthy et al. 2003),
once remanence values are normalised for leaf surface area (Matzka and
Maher 1999). Recent works have shown that leaf SIRM can be used as a
proxy for ambient PM10 levels (Mitchell and Maher 2009; Kardel et al. 2011).
5.3.4 Elemental Analysis
Following the magnetic measurements, subsets of the leaf samples from the
start and end of the study were prepared for elemental analysis via ICP-MS
(Agilent 7500cx). Leaves were digested in high purity nitric acid (15.6M) in
closed vessels using a microwave apparatus (MARS Xpress, CEM)
according to US EPA method 3051A. Increases in metal concentrations over
the study period may signify capture of metal-containing particles on the leaf
surfaces, with Fe being of particular interest due to the ferrimagnetic
properties of Fe containing minerals and hence influence on leaf SIRM.
Vehicle derived PM10 contains high levels of Fe due to conversion of Fe
impurities in fossil fuels on combustion (Muxworthy et al. 2003).
Soil samples were also prepared for elemental analysis by ICP-OES to

investigate potential metal concentration changes as particulates are washed
into the substrate by rains. Samples were taken at the end of the study from
one tray of each species at all three locations, and from trays of bare soil,
with samples taken from the surface soil layer and deeper in the trays. Soil
samples were also taken at the start of the study before the seeds were sown.
The soil was dried at 40°C and sieved to 63 µm to remove
94
stones/twigs/leaves and homogenised in a ball grinder (Fritsch Spartan
pulverisette 0) prior to acid digestion (as described above).
5.3.5 Scanning Electron Microscopy (SEM)
Single leaves of the four main species in the study were collected at the end
of the study from Roof 2. These were transferred in plastic sampling bags to
prevent contamination, and subsequently mounted on carbon stubs for
analysis via SEM (FEI XL30 Environmental Scanning Electron Microscope).
This was carried out on the same day to prevent desiccation and subsequent
alteration of leaf surface micro morphology (Stabentheiner et al. 2010).
Element mapping was undertaken via energy dispersive X-ray analysis (EDX)
on clusters of particle grains to investigate any significant elemental
distributions.
5.3.6 Statistical analyses
Correlations were carried out between average ambient PM10 concentrations

at the sites and with average PM10 recorded at the same times from the
AURN station in Piccadilly gardens. Rolling averages of ambient PM10 over
the pre-sampling periods at the AURN station were calculated and correlated
against leaf SIRM values. Data were grouped by species or site, which
allows for inter-site and inter-species comparisons to be made respectively.
Repeated Measures ANOVA (RMANOVA), with the Greenhouse-Geisser
correction, was used to test significance with site or species as a between-
subject factor. Linear Mixed Models (LMM), with site and species as random
effects, were fitted to any trends seen from the start of the study to the end.
Statistical analysis was carried out using SPSS software (SPSS 16).
95
5.3.7 PM10 removal quantification
Particle numbers were counted in SEM images of the same leaf samples
described in section 2.5, sampled at random, and averaged per unit leaf area.
The counts (N) were used to calculate particle volume by:
where D is average particle diameter and particles are assumed to be

spherical
Particle mass is then given by:
where ρ is the particle density, assumed to be 1.3 g/cm3 (Held et al. 2006).
The mass per unit leaf area can then be used to quantify PM10 removal by
estimating average leaf area per unit area of roof based on measurements of
above surface biomass of the experimental trays. To gain an annual removal
figure, it was assumed that while the study species are evergreen, growth is
compromised in the winter months, so a rate of 50% of the summer rate was
chosen. Sensitivity to this rate was assessed by calculating removal at ±5%.
Roof 2 was considered suitable for the quantification work as it represents an
‘average’ roof, subject to urban background PM10 levels.
96
5.4 Results
5.4.1 Site comparisons
Average ambient PM10 concentration at different days of sampling were

significantly correlated between the sites (control with Roof 1, r = 0.87;
control with Roof 2, r = 0.59; Roof 1 with Roof 2, r = 0.85) indicating that
while the magnitude of PM10 differs between the sites, the overall trends are
broadly similar (Fig. 5.2). Fig. 5.1 shows their locations, with the maximum
distance being between the control site and Roof 1 (3.5 km).
45
Control
40
Roof 1
35 Roof 2
Ambient PM10 (µg m-3)
30 Piccadilly
25
20
15
10
Date
Fig. 5.2 – Average ambient PM10 levels between 11am and 1pm at each of the three
sites on selected days throughout the study and daily average at the Piccadilly
monitoring station
97
Correlations between average ambient PM10 with average PM10 from the
AURN station (control, r = 0.59; Roof 1, r = 0.71; Roof 2, r = 0.64) showed
good agreement, with the lower value for the control site due to the increased
distance from the city centre. The filter SIRM values (Fig. 5.3) were lower
for air sampled at the control site than on the study roofs.
1.8
1.6
1.4
Filter SIRM (X10-6 A)
1.2
1
0.8
0.6
0.4
0.2
0
Control Roof 1 Roof 2
Site
Fig. 5.3 – Filter SIRM from 1 hour pumped air samples taken on 27 July 2011,
showing the lower levels of magnetisable particulate matter at the Firs control site.
98
Grouping the data by species (Fig. 5.4) shows SIRM was influenced by time,
with an increase seen from start of study to end. This effect of time was
significant. (RMANOVA, F=12.914, p = 0.01). There was a general trend
noticeable whereby the highest SIRM values are recorded on Roof 1,
followed by Roof 2. The lowest SIRM values were seen at the relatively
unpolluted control site. This influence of site location on SIRM was also
significant as evidenced by between-subjects effects of site (F=103.27,
p=0.01).
Fig. 5.4 – Leaf area-normalised SIRM values of the study green roof species; (a) A.
stolonifera, (b) F. rubra, (c) P. lanceolata, and (d) S. album at the three study sites
(note the differing y-axis scales). Upward trends over the study period are
noticeable, with Roof 1 capturing more PM10 than Roof 2
99
The increases in SIRM over the study period were found to be significantly
different between sites (LMM, F=55.76, p=0.01), with Roof 1 producing the
strongest effect on the SIRM increase. Fig. 5.5 shows the SIRM enrichment
ratios, the ratio of each roof over the control site, at each sampling event.
The plants on Roof 1 were clearly capturing more PM10 than those on Roof 2.
The two roofs displayed a similar trend for the latter half of the study.
4
Roof 1
3.5
Roof 2
3
2.5
Enrichment Ratio
1.5
0.5
-0.5
-1 Date
Fig. 5.5 – SIRM enrichment ratio of the two roadside sites to the less polluted control
site showing the higher enrichment on Roof 1 than Roof 2
Wind rose analysis of data from Roof 1 showed that the prevailing wind
direction throughout the study was predominantly from the south west,
especially in the drier final two weeks. These winds would potentially be
carrying large amounts of road derived pollution from the proximal, southerly
located Mancunian Way, which would explain the higher SIRM values at this
site.
100
5.4.2 Species comparisons
Grouping the data by site (Fig. 5.6) shows a trend in the ability of the different
species to capture magnetic PM10 could be seen to follow the order A.
stolonifera > F. rubra > P. lanceolata > S. album. Between-subjects effects
showed significant effects of species on the leaf SIRM levels (RMANOVA,
F=65.93, p=0.01), so some species were clearly collecting particles faster
than others. Significant effects of species (LMM, F=34.99, p=0.01) were
noted, with A stolonifera producing the best estimates, as well as higher
SIRM values.
Fig. 5.6 – Leaf area-normalised SIRM values of the study species at each of the
three sites; (a) Control, (b) Roof 1, and (c) Roof 2 (note the differing y-axis scales).
Upward trends over the study period are noticeable with A. stolonifera displaying the
highest particle capture
101
The SEM images in Fig. 5.7 show the different leaf surface
micromorphologies of the four green roof species and the predisposition of
the particles to accumulate within parallel grooves, especially for the grasses,
A stolonifera and F. rubra. P. lanceolata and S. album also have grooved
surfaces but the grooves are randomly distributed at the boundaries of
tessellating plate-like structures. The grooves are more pronounced with P.
lanceolata than S. album, and the latter also has a waxy surface, which may
explain the higher particle capture efficiency of P. lanceolata. Small barbs on
the surface of A. stolonifera leaves could also enhance particle capture.
102
Fig. 5.7 – SEM images of leaf surfaces of the four study species (X800 – 1600 mag.)
with higher magnification images of selected spherules (X12800 – 25600 mag.)
103
The grains present on the leaves were of varying shapes and sizes.
However, a number of grains were found to be the spherule shape
characteristic of vehicle exhaust produced PM10, formed from cooled droplets
(Maher et al. 2008). Spot EDX analysis of these spheroidal particles yielded
high Fe concentrations. Elemental mapping of a cluster of particles also
showed a high occurrence of Fe – rich particles in the 2 – 5 µm range with
larger particles being rich in silica (Fig. 5.8), indicating a geogenic source for
these larger particles (De Berardis et al. 2007).
Fig. 5.8 – EDX for P. lanceolata abaxial leaf surface indicating distribution of iron
and silica within a particle cluster
The results of the ICP-OES analysis of the upper and lower soil horizons
from the plant trays showed no significant spatial or temporal differences in
levels of Fe, Al, Ni, Pb or Zn.
Leaf Fe concentrations did not increase over the duration of the study despite
a fairly high correlation between leaf SIRM values and Fe concentrations (r =
0.67, p=0.01). Table 5.1 shows this relationship, with clear SIRM increases
for all species and sites, but a rather erratic pattern for leaf Fe concentrations.
The extent of these increases from start of study to finish manifests with the
grasses showing higher average SIRM increases than P. lanceolata and S.
album. Fe concentration increases were mostly apparent on the roofs, with
the control site showing decreases or a minimal increase in the case of F.
rubra.
104
% increase SIRM % increase iron concentration
Control Roof 1 Roof 2 Average Control Roof 1 Roof 2 Average
A. 278.76 114.36 133.90 175.67 -24.79 35.04 -11.81 -0.52

stolonifera
F. rubra 28.02 295.69 243.27 188.99 0.76 139.66 90.93 77.12
P. 93.84 74.78 180.65 116.42 -60.88 86.28 124.36 49.92

lanceolata
S. album 70.62 14.79 4.82 30.08 -41.20 -32.80 7.41 -22.2
Table 5.1 – Percentage increase in SIRM values and iron concentration over the
duration of the study, showing clear increases in SIRM but variable patterns for iron
concentration
CA 1A 2A CF 1F 2F CP 1P 2P CS 1S 2S 1Pl 2Sy
1A .833
2A .462 .166
CF .478 .253 .200
1F .468 .426 .569 -.092
2F .774 .701 .740 .398 .579
CP .622 .594 .539 .200 .682 .823
1P .307 .401 .378 -.081 .747 .611 .804
2P .462 .400 .713 .414 .478 .825 .815 .627
CS .388 .192 .637 .384 .569 .521 .626 .404 .719
1S .749 .741 .156 .193 .484 .576 .538 .440 .185 .114
2S .085 -.204 .571 -.116 .341 .143 -.067 .185 .126 .099 -.077
1Pl .303 .471 .045 -.194 .554 .382 .801 .741 .441 .327 .462 -.310
2Sy .680 .653 .614 .284 .799 .874 .918 .815 .794 .606 .546 .101 .648
2H .609 .711 .490 .222 .712 .808 .816 .787 .773 .448 .458 .110 .623 .935
Table 5.2 – Correlation matrix for the SIRM values of the study species
(A = A. stolonifera, F = F. rubra, P = P. lanceolata, S = S. album, Pl = P. acerifolia,
Sy = S. novae-angliae, H = H. inodorum) at each site (C = Control, 1 = Roof 1, 2 =
Roof 2) showing wide variability. Bold values indicate significance at the 0.05 level.
105
The correlations in Table 5.2 show varying inter-species relationships, with S.
album generally having the lowest correlation with other species (average r =
0.31), and the two permanently located species on Roof 2, S. novae-angliae
and H. inodorum, having the highest correlations. Their correlation with each
other was very high (r = 0.94). Their permanence at the site could mean that
they are well-equilibrated with atmospheric PM10.
The best correlations between rolling averages of ambient PM10 against leaf
SIRM values (Table 5.3) were seen with the Roof 1 samples. This is possibly
due to their location closer to Piccadilly gardens. Good correlations were
also seen with the permanent shrubs, S. novae-angliae and H. inodorum, on
Roof 2, again indicating a potential equilibration with atmospheric PM10.
106
Site
Control Roof 1 Roof 2
A F P S A F P S Pl A F P S Sy H
r 0.39 -0.16 0.55 0.07 0.65 0.69 0.62 0.45 0.65 0.02 0.42 0.24 -0.14 0.64 0.70
Table 5.3 – Correlations of area-normalised leaf SIRM values with rolling average ambient PM10 concentration (A=A. stolonifera, F=F. rubra, P=
P. lanceolata, S=S. album, Pl=P. acerifolia, Sy=S. novae-angliae, H=H. inodorum)
107
18 20
Rainfall (mm)
16 18
P. acerifolia
S. novae-angliae
16
14
H. inodorum
14
12
Leaf SIRM (x10-6 A)
12
Rainfall (mm)
10
10
8
8
6
6
4
4
2 2
0 0
Date
Fig. 5.9 – Daily rainfall totals and leaf area-normalised SIRM values of the
‘permanent’ green roof species and the roof-height tree canopy which show a clear
increase in the drier second half of the study period
The increase of the SIRM values of the permanently located green roof
species and roof height tree canopy in the drier second half of the study
period (Fig. 5.9) suggested increasing PM10 levels in dry conditions, as would
be expected. However, the SIRM decrease after rainfall events exceeding
2mm, noted in Mitchell et al. (2010), did not exhibit a particular pattern.
Although the results suggest a non uniform response to rainfall events by
species type, it can be seen that the larger rainfall events within the period 16
– 22/07/2011 produce a decrease in the SIRM values for most species at
most locations (Fig. 5.4). The reverse was also apparent, i.e. a lack of
rainfall allowed the SIRM values to increase, as seen in the drier period
108
towards the end of the study. This dry period also produced a rise in ambient
PM10 as seen in Fig. 5.2.
5.5 Discussion
5.5.1 Spatial differences in PM10 capture
The site specific trends were found to be significant with Roof 1 being the site
with the highest particle capture rates, due to its location downwind of a
major source. Roof 1 is slightly different from Roof 2 in that it is adjacent to a
raised busy inner city motorway which is only a few metres lower, and
prevailing winds cross this motorway before reaching the roof. This could
imply that Roof 1 represents a receiving environment for a local source of
PM10 whereas Roof 2 is more dominated by urban background PM10
deposition. The high Fe contents of the particles found by EDX on samples
from Roof 2 might indicate the vehicular source of the matter in this location,
which confirms the suggestion for a considerable input from transport related
PM10 at the sites. The agglomerated clumps of particles found were quite
common and are potentially due to magnetic interactions between particles
(Mitchell and Maher 2009).
Spatial patterns of PM10 exist in the data with the three urban centre
locations, Roof 1, Roof 2 and Piccadilly, displaying similar trends in ambient
PM10 and Roof 1 plants having the highest correlation with the rolling
average Piccadilly ambient PM10 data. These spatial patterns support
previous work that suggest the utility of biomagnetic monitoring for capturing
fine-scale spatial variations in PM10 concentrations (Mitchell and Maher 2009;
Hansard et al. 2011). However the green roof species in this study, while
useful due to their evergreen nature, would need further study into the time
taken to equilibrate, as the results suggest equilibrium is yet to be reached.
This delay to equilibration is evidenced by the moderate leaf SIRM

correlations between the species, and the fact that the highest correlation of r
=0.95 was seen between the two permanent species on Roof 2, indicating
that they have had enough time to equilibrate. Underlying species differences
in the time taken to equilibrate could also be influencing the correlations.
109
Kardel et al. (2011) also found a progressive increase in leaf SIRM of
deciduous trees over a whole growing season, even when relative change in
ambient PM10 was negative. Equilibration times as short as 6 days have
been found for roadside trees (Mitchell et al. 2010), however Lehndorff et al.
(2006), investigating needle SIRM for the evergreen Pinus nigra, observed
26 months taken for equilibrium to be reached, with an increase in SIRM for
the first 20 months. This has implications for any conclusions made on PM10
removal quantities based on the results of this study. Net particle to leaf
deposition will be expected to remain positive after the end of the study so
calculations of PM10 removal might be underestimates of the true capture
efficiency.
The lack of equilibration means the increase in SIRM values over the limited
study time period can be viewed, for Roofs 1 and 2, as species and site
specific temporal trends triggered by moving from a low pollution
environment to a high one. For the control site, incremental SIRM values are
due to the fact that while the site is relatively unpolluted and remote from
major road sources, there is still a peri-urban background PM10 signal.
Ambient PM10 levels inside the greenhouse at the control were very similar to
those at the study roofs (Fig. 5.2), however, a larger fraction of this will be
non-magnetic, biogenic PM10 such as pollen and condensed Biogenic
Volatile Organic Compounds (BVOCs) (Litschke and Kuttler 2008). The filter
SIRM values in Fig. 5.3 confirm this, with a lower magnetic signal identified
from air sampled at the control site.
Rooftops, both by their distance from, and obliqueness to, urban streets,
represent an environment slightly removed from the vehicular source of
urban PM10. Roadside PM10 concentrations have been found to be 100%
higher at 0.3 m than at 1.5 m height (Mitchell and Maher 2009) indicating a
decrease with height. This means that PM10 deposition to rooftop surfaces is
dependent on aerial transport mechanisms via winds within the Urban
Canopy Layer (UCL) and gravitational deposition of particles from within this
mixed layer. Litschke and Kuttler (2008) stated that to be an effective filter of
urban air masses vegetation must be located as near as possible to the
emission source.
110
5.5.2 Species differences in PM10 capture
The SEM images (Fig. 5.7) show differences in the micromorphology of the
leaf surfaces which may go some way in explaining the species differences in
PM10 capture. The parallel grooves on the leaves of the grasses can trap
particles and prevent their resuspension. Hairy Tilia leaves have been found
to be more efficient at particle capture than smooth leaved Tilia species
(Kardel et al. 2011). Mitchell et al. (2010) observed plant species with hairy
or ridged surfaces have a significantly higher leaf SIRM than those with a
waxy or smooth surface, such as P. lanceolata and S. album.
The macro morphological influences on particle capture must also be

considered, with the grasses having dense arrangements of long thin blade-
like leaves which may affect the near-surface air flow and enhance
deposition (Litschke and Kuttler 2008). P. lanceolata, with its radial
arrangements of lanceolate leaves, and S. album with its tendency to form
low-lying mats and cylindrical leaves, may increase the near-surface
roughness length to a lesser degree than the grasses.
The weak correlation between SIRM and Fe concentrations is potentially due

to the nature of the SIRM measurement technique which is an indication of
total concentration of magnetic particles (Muxworthy et al. 2003). Different
Fe-containing minerals such as haematite and magnetite, have differences of
mass specific SIRM at least an order of magnitude apart at the 1 – 10 µm
size range (Walden 1999). This unaccounted for variation has the potential
to weaken the relationship between SIRM and presence of Fe minerals. A
possible solution to this would be to use incremental ARM acquisitions to
characterise the proportions of Fe containing minerals present (Walden
1999). The decoupling between SIRM and Fe may be due to species
dependant effects from cycling of plant intracellular Fe, via soil uptake
(Marschner et al. 1986) or surface absorption through the cuticle (Peachey et
al. 2009). Further work could explore whether stronger correlations with
SIRM may be achieved with Fe concentrations in leaf surface deposits,
obtained through leaf-washing (Freer-Smith et al. 2005).
111
The soil metal concentrations did not change during the course of the study.
This is potentially due to the brevity of the study being unable to capture the
long-term nature of what is presumed to be a small flux of metal-containing
particulates to soil.
5.5.3 Implications of wider green roof implementation
The species differences in particulate capture have implications for any roof-
greening projects that have the specific aim of pollution remediation, with a
necessity for increasing the near-surface roughness apparent. Intensive
green roofs, which have a deeper substrate able to support the larger above-
surface biomass, would be preferable, however higher installation and
maintenance costs are associated with these (Clark et al. 2008). Green roof
impacts may be mostly operating via reduction of urban background levels
within the UCL, but certain locations adjacent to strong sources, as with Roof
1, indicate that strategic planting informed by location may be a key
consideration.
Quantifying particulate removal by green roof vegetation at the city scale can
give an indication of the potential benefits of green roof installation. The
removal rates achieved by grasses in Table 5.4 are higher than the 1.12
g/m2/year for short grass and 1.52 g/m2/year for tall herbaceous plants
reported by Yang et al. (2008). Figures for trees in US cities are in the range
0.4 – 11.2 g/m2/year with an average of 3.8 g/m2/year (Nowak et al. 2006). In
comparison to trees, green roof species do not perform as well, however the
grasses, especially F. rubra, come close.
112
Species g/m2/year T/year under max green
roof scenario
A. stolonifera 1.81 ± 0.06 0.9 ± 0.03
F. rubra 3.21 ± 0.1 1.61 ± 0.05
P. lanceolata 0.49 ± 0.02 0.25 ± 0.01
S. album 0.42 ± 0.01 0.21 ± 0.01
Table 5.4. Estimates of quantities of PM10 captured by the study species with ± 5%
sensitivity to winter reduced rate. Higher removal rates were achieved by the
grasses A. Stolonifera and F. rubra
The green roofing potential of Manchester city centre and the Oxford Road
corridor, 326 ha in area, was quantified using aerial photography in ArcGIS
software. Polygons were constructed for all potential green roofs, based
wholly on whether the roof was flat, and no consideration given to structural
considerations. The total area of potential green roof coverage was found to
be approximately 50 ha (15.3% of the selected area). A ‘maximum extensive
green roof scenario’ was postulated, where every possible flat roof in this
selected area of city centre Manchester has an extensive sedum roof.
Emissions of PM10 for this area are estimated to be 9.18 tonnes per year,
based on an emissions inventory (HFAS 2011), and the area is mostly
classed as an AQMA (Defra 2011a).
The potential removal of particulate matter under this scenario is 0.21 metric
tonnes a year (Table 5.4). This equates as 2.3% (±0.1%) of the PM10
emitted by the scenario area, which is a considerable removal amount.
Larger quantities could be removed with grass roofs evidenced by the 9.8 –
17.5% in the scenario, for A. stolonifera and F. rubra respectively.
Consequently, intensive greening can be beneficially employed in critical
locations adjacent to PM10 sources, and extensive green roofs installed in
background locations. Sequestration of the metal-containing particles within
113
the substrate layer of the green roof is expected (Berndtsson et al. 2009).
Further research is needed, however, into the long term fate of these
particles and whether green roof substrates would reach pollutant saturation.
Despite the distance from sources, the results presented here suggest that
green roof vegetation has a sizeable impact on particulate matter levels in
the mixed layer. The benefit of reduced particulate air pollution in cities is
further evidence for the multi-beneficial nature of green roofs.
5.6 Conclusion
 The grasses A. stolonifera and F. rubra were found to have higher

particle capture efficiency than the invasive weed P. lanceolata and
the commercial extensive green roof species S. album. This should
be considered if green roofs are being constructed with the aim of air
pollution reduction.
 Both macro and micro morphological reasons for differences in

capture efficiency were posited.
 Spatial differences in leaf magnetic SIRM of the green roof species

were observed with proximity to PM10 sources being a main factor.
 Green roofs act as passive filters of airbourne particulate matter.

While not as effective as street trees, due to lower surface roughness
lengths and increased distance from sources, they can be considered
for remediation of urban air pollution because their construction does
not require major upheaval of the urban built environment, as tree-
planting schemes often do.
 0.21 tonnes of PM10 a year were removed from Manchester city centre
in a scenario involving all flat roofs within a chosen area being
installed with an extensive green roof. This is the equivalent of 2.3%
of the PM10 inputs of this area. Larger quantities can be removed with
grass roofs.
114
CHAPTER 6 REDUCTION OF THE URBAN COOLING
EFFECTS OF AN INTENSIVE GREEN ROOF
DUE TO VEGETATION DAMAGE
Authors:

Affiliations:
1
2
Urban Climate 3: 40 - 55
http://dx.doi.org/10.1016/j.uclim.2013.01.001
115
6.1 Abstract
Green roofs are increasingly being utilised as urban climate adaptation

measures given good evidence of their passive cooling effects. Shading of
conventional building surfaces, higher albedo and latent cooling from
evapotranspiration mean that vegetated surfaces can be used to significantly
lower surface and air temperatures in cities. The extent to which cooling
properties may be affected by poor maintenance, green roof damage and the
general health of the vegetation is not as well known. This study
investigates the ability of an intensive green roof in Manchester, UK, to cool
the air above it relative to an adjacent conventional, concrete roof. The
monthly median air temperature was found to be 1.06ºC lower at 300 mm
over the green roof. Diurnal trends were apparent, with strongest cooling of
1.58ºC, occurring at night when Urban Heat Islands are most common. A
section of green roof was damaged by mismanagement in a drought period,
which allowed an investigation into the impairment to the cooling effect when
large areas of bare substrate are present. Daytime temperatures were
higher over this section of the roof than over an adjacent healthy green roof,
being a maximum 0.63ºC warmer than the bare roof on average, in the late
morning. Cooling still occurred at night, albeit with a lower maximum
average relative cooling of 0.78ºC. Implications of the spatial influence of
green roof cooling are discussed along with the need to undertake careful
maintenance and irrigation of the roofs for the benefits to be maintained.
116
6.2 Introduction
Green roofs have been widely researched to assess their role as adaptation
strategies for high urban temperatures using both empirical (Köhler et al.
2003a; Spolek 2008; Teemusk and Mander 2009), and modelling studies
(Barrio 1998; Niachou et al. 2001; Alexandri and Jones 2007). The need to
ameliorate the Urban Heat Island (UHI) effect is given added importance
when the effects of climate change are factored into high urban temperatures.
An increase in the frequency of summer heat wave events (Murphy et al.
2009) implies that UK urban areas could become thermally uncomfortable
places to live in the future. The European heat wave of 2003 claimed an
estimated 70 000 lives (Robine et al. 2008), of which over 2000 were in
England and Wales (Larsen 2006). A large number of the fatalities were
located in urban areas due to the high population densities there, with half
the world’s population living in cities (UN 2009), and also due to the nature of
the built environment.
Urban areas are abundant in materials with a high thermal capacity such as
concrete and asphalt (Taha 2004). This, along with anthropogenic thermal
inputs (Smith et al. 2009) and urban canyon effects (Landsberg 1981), is one
of the contributors towards the UHI phenomenon, whereby cites in the UK
can be up to 7°C warmer than adjacent rural areas (Wilby 2008).
Conventional urban surfaces are also distinctly lacking in evapotranspiring
surfaces leading to differences in the partitioning of short-wave radiation in
an energy budget (Oke 1982). A lack of vegetated surfaces means less
incoming solar energy is used for latent cooling. It can be argued that UHIs
are currently not as big a problem for UK cites as for other international cities
because they reduce winter heating costs (Landsberg 1981), but predicted
higher temperatures in a future climate may exacerbate summer warm
conditions. Additionally, regions presently with cooler average temperatures,
such as North West UK, have lower adaptation thresholds, in terms of high
temperature-mortality relationships (Armstrong et al. 2011).
High temperatures can occur in areas home to vulnerable populations

causing an increase in the risk of human health impacts in cities (Donaldson
117
et al. 2001; Lindley et al. 2004). When the temperature is close to human
body temperature of 37°C, physiological stress can occur (Larsen 2003). At
the extreme end of the UKCP09 probabilistic predictions there is a 10%
likelihood that the Manchester conurbation will experience temperatures of
20-22°C during warm summer nights (Cavan 2010). It is at night that the
effects will be most felt because city residents will be indoors with an
associated reduced ability to seek refuge from overheating. Quantifiable
climate risk indices were investigated for Manchester (Smith and Lawson
2012) with threshold exceedance of meteorological variables, such as days
with maximum daily temperature greater than or equal to 29.2°C, found to be
indicative of weather-related impacts. In a future UK climate, four more heat
wave events per year are predicted, along with a summer rainfall decrease of
17% (Murphy et al. 2009).
Adaptation to rising temperatures is becoming as important as mitigation of

greenhouse gas emissions, and includes techniques such as designing
buildings to minimise incident solar gain or increase ventilation (Gupta and
Gregg 2012). Passive cooling from increasing the amount of greenspace in
the built environment is increasingly being seen as a beneficial adaptation
technique, preferable to increasing reliance on mechanical air conditioning.
The latter increases urban electric demand in summer (Akbari et al. 2009)
and also contributes to increased greenhouse gas and air pollutant levels
from the burning of fossil fuels by electricity-producing facilities to meet this
demand (Solecki et al. 2005). Microclimate modification by vegetation is
achieved in three ways. Plants shade buildings, and there is a localised
increase of latent cooling via evapotranspiration (Solecki et al. 2005) and
additionally there may be an increase in solar reflectance due to the higher
albedo of leaf surfaces compared to building materials (especially in
temperate and cold climates) (Grant et al. 2003).
Green roofs are being utilised increasingly in climate change adaptation

strategies and as a means to reduce the costs associated with mechanical
air conditioning (Castleton et al. 2010). Large scale roof greening was
suggested to have the indirect benefit of a 25% reduction in urban electricity
use by reducing UHIs (Akbari and Konopacki 2005), on top of direct energy
118
savings for individual buildings (Sailor 2008). Green roofs have an additional
benefit of needing no major upheaval of the existing urban infrastructure in
order to accommodate the vegetation. A hindrance to extensive greening of
urban centres is the competition for expensive building space and the fact
that the fundamental layout of cities is unlikely to change for some years
(Wong et al. 2003). A study into the potential of trees to climate-proof
Manchester found that the limited space available for tree planting meant that
tree cover could only be increased by up to 5.3%, which was not sufficient to
keep surface temperatures at current levels even up the 2020s (Hall et al.
2012). A solution is to utilise the available space afforded by rooftops.
Estimates of urban rooftop coverage vary greatly but can be as high as 35%
of the land cover (MacMillan 2004), and 15% of the city centre area of
Manchester is composed of flat roofs that could potentially support a green
roof (Speak et al. 2012).
Urban greening schemes often require careful planning and installation,

incorporating design techniques to maximise irrigation for instance, and the
resulting green space can need certain maintenance procedures. In the case
of green roofs this entails regular weeding and irrigation in dry periods. The
cooling properties are dependent on the lush vegetation and any impairment
to this cover can reveal large areas of soil substrate which can alter the
thermal properties of the roof. Data from Wong et al. (2007) indicate surface
temperature increases of up to 10°C and overlying air temperature increases
of around 2°C after a new extensive green roof was constructed in Singapore,
and incomplete foliage cover was a major reason for this. Bare substrate
was also responsible for surface warming on a green roof during warmer
months at higher latitudes (Teemusk and Mander 2010). These studies have
brought to light this deleterious effect of the bare substrate and a need for
further research into this area is apparent. The shallow substrates of green
roofs mean they are prone to drying out, and climate projections for less
summer rainfall indicate that drought may become a threat to plant survival if
artificial irrigation is not in place. Modelling studies often assume no drought
stress on the green roof vegetation component (Gill 2007), so correct
119
parameterisation of the reduced cooling during drought periods will improve
green roof modelling capabilities.
The aim of this study is to investigate how the cooling effects of an intensive
green roof are affected by adverse conditions. The objectives are thus
twofold. Firstly, to quantify the cooling effects of the green roof when
compared to an adjacent bare roof. Secondly, to look at how this cooling
effect is affected when a combination of mismanagement and drought
severely damages the green roof vegetation. In situ monitoring of air
temperatures above an aged green roof, can reveal thermal characteristics
that will be useful when considering the impact of green infrastructure on
UHIs both now, and in the future when climate change projections manifest.
A recent meta-analysis by Bowler et al. (2010) revealed that the majority of
studies comparing air temperature measurements of green and non-green
urban areas were focussed on parks and trees, with green roofs being less
studied. Most green roof studies concentrate on surface temperature
reductions and subsequent lowered heat transfer into buildings and reduction
in cooling loads (Niachou et al. 2001; Sfakianaki et al. 2009). By
investigating the effect green roofs have on the local microclimate, this study
will provide novel data on how they can aid UHI reduction and improve
thermal comfort in cities. Additionally, investigating their performance when
damaged can help provide valuable information on how their usefulness
within urban greening schemes can be compromised, which is important
information for greenspace managers.
6.3 Methodology
Manchester is a large city in north-west England with a population of 498 000

(MCC 2010). As one of the world’s first industrial cities, Manchester has a
long history of the climatic effects of urbanism such as smog, and today aims
to be a ‘green’ and ‘sustainable’ city (MacKillop 2012). So it is apt that urban
climate research is ongoing in the city. Manchester’s UHI was mapped using
ground and air transects in summer and found to be ~3°C during the day and
up to 5°C at night (Knight et al. 2010; Smith et al. 2011). Analysis using
120
stationary monitors, found the UHI of Manchester to be as high as 8°C in
summer and 10°C in winter, with a higher probability of UHI occurrence in
summer (Cheung 2011).
A green roof within the University of Manchester campus, on the Precinct

building, was chosen for the study. The roof was chosen because it has a
conventional roof area (900 m2), consisting of concrete paving slabs,
adjacent to a large (408 m2) intensive green roof, which is 43 years old. The
area is classified as open midrise, characterised by a fairly open
arrangement of buildings of 3 – 9 storeys with some trees, typical of an inner
city university campus area (Stewart and Oke 2012). All areas of the roof
receive approximately the same amount of direct sunlight throughout the day
until late evening when the paved area becomes shaded by adjacent
buildings. The concrete paving slabs are not common components of
Manchester’s roof surfaces, which tend to be dominated by low albedo clay
tiles on older Victorian buildings, or lightweight bituminous layers on modern
flat-roofed building stock. Therefore, the temperature data gained from this
roof are to be used in a comparative study against the adjacent green roof,
and not to be used in generalisations about other low-albedo roofs.
The green roof has a soil layer 170 mm deep, which supports a wide variety
of flora, including grasses, shrubs, rushes and small trees. The roof has,
however, suffered from poor maintenance and some parts of the roof are
dominated by invasive species such as Rubus fruticosa, Buddleja davidii and
Plantago lanceolata. The green roof is L-shaped with two distinct sections as
can be seen in Fig. 6.1. One area, G1, is dominated by patches of Juncus
species and Aster novi-belgii. G2 is dominated by Plantago lanceolata,
Senecio jacobaea, and the grasses, Agrostis stolonifera and Festuca rubra,
and has a narrow path running through the middle made of the same paving
slabs as the bare roof.
121
Fig. 6.1 – Map to show the location of Precinct study site and the monitor locations
on the roof
The green roof is of particular interest due to its age and the fact that the roof
was not constructed specifically for the study. Green roof studies with an
experimental component are often carried out on specially constructed roofs
or prefabricated green roof panels, i.e. the green element was under the
122
control of the investigator (Harazono et al. 1990; Takebayashi and Moriyama
2007). It can be argued that studies of this type can over or underestimate
the benefits of actual green roofs. Investigating a well established, real roof
garden which is subject to local weather patterns, will reveal green roof
characteristics that can be sensibly applied to the real world. It must be
noted, however, that the study focuses on an intensive green roof, which
cannot be supported by a lot of existing buildings without a structural refit at
some cost to the building owner.
6.3.2 Monitoring
Temperature sensors (Tinytag 2PLUS TGP-4500) with an accuracy of ±0.4ºC

were housed 300 mm above the roof surface in Stephenson shields. Stands
were painted white to reduce convective heat inputs into the shields above.
Dry bulb temperature was recorded every 15 minutes synchronously
between all sensors. All the sensors were initially checked for accuracy
against each other by keeping in the same outdoor environment for a week
and comparing with a Rotronics Hygroclip hand held monitor, which has an
accuracy of ±0.2ºC and resolution of 0.004ºC. Monitoring commenced with
one stand on the bare roof and one on the green roof area G1 (Fig. 6.1) on
05/06/2010. Two additional stands were added on 28/03/2011, with one
placed on green roof area G2 and one placed as an additional monitor on the
bare roof. Bare roof air temperatures were subsequently calculated as an
average of the two monitors. The stands were placed centrally within the
roof areas to avoid edge effects and minimise advection influences from
adjacent surfaces. Advection effects are further minimised by the long-term
monitoring approach and the open nature of the site. Sky View Factor is 1
for each stand, reflecting the open rooftop location.
Conclusions about the relationship between the air temperature above the
roof and the sensible heat flux must be made carefully, due to the potential
differences between the roof types. These arise from differences in flow over
the roof surfaces and eddy interference within the green roof vegetation
canopy. For the purpose of this comparative study, differences between the
roofs are indicative of differences in the thermal impact of the roof surfaces
123
on the overlying air mass. Flux differences, measurable by eddy covariance
techniques, are assumed to be minimal, and air temperature is thus a
suitable descriptor of thermal differences between roofs as found in other
studies (Wong et al. 2003; Getter et al. 2011).
Spot monitoring of surface temperatures was undertaken in the second and

third summers of the study with a thermal imager (Fluke Ti25) on 27/07/2011
and 14/08/2012. Image analysis was carried out with Smartview software
(Version 3.2). Similar weather conditions were sought as much as possible
for these two comparison days, however, summer 2012 in the UK was
characterised by a lot of rain and below average solar radiation due to
persistent cloud, making the choice of day for spot monitoring difficult.
Nonetheless, the two days were very similar in amount of solar radiation
received and humidity. Soil moisture conditions were fairly similar, with the
2011 monitoring occurring 5 days after rainfall (3 mm) and the 2012 occurring
7 days after significant rainfall (4.3 mm) but with a small shower (1.2 mm) two
days previous. Peak daytime wind speeds in 2011 were 3.5 m/s compared
with 5 m/s in 2012 which contributed to maximum temperatures of 25ºC in
2011 compared with 23ºC in 2012.
A septum tensiometer (Skye Instruments) situated between G1 and G2

logged data from November 2010. This was converted to soil moisture
content using data from extensive soil characterisation experiments (carried
out and personally communicated by C. Agnew, 15/08/2012). Meteorological
data were available from the Whitworth Observatory (UoM 2012), situated
close to the Precinct roof (Fig. 6.1).
6.3.3 Data analysis
Data were found to be not normal using the Anderson-Darling test, thus non-
parametric data analysis was implemented. Monthly median temperatures
were compared between the three roof areas, as were daily temperature
cycles for the whole study period. Diurnal trends in the temperature
differences between the two green roofs and the bare roof were also
investigated. Data were organised by season for the latter analysis, defined
as winter (Dec, Jan, Feb), spring (Mar, Apr, May), summer (Jun, Jul, Aug)
124
and autumn (Sep, Oct, Nov). Spring 2011 consisted of fewer readings due to
the new monitors only being placed on G2 and the bare roof in late March.
Statistical analyses were carried out using R (version 2.13.2) and SPSS
(version 16 for Windows).
6.4 Results
6.4.1 Vegetation dynamics
A major characteristic of this study is that it was carried out on an existing,

and aged, green roof compared experimentally to a bare roof. As a result, a
number of elements related to the structure of the green roof were out of
experimental control. Those of relevance to the study are the vegetation
types and their dynamics. Near the start of the study, there was a prolonged
dry period, during which G2 was intensively mowed, without irrigation, by the
building owners. This severely damaged the green roof vegetation, which
was predominantly grass, and resulted in large areas of substrate becoming
uncovered. Over the next two growing seasons, this was gradually colonised
by numerous wind-blown plant species and mosses. By summer 2012 G2
resembled a meadow, with some small bare patches remaining. G1 was not
mowed because grass was not dominant here, however, over the three
growing seasons of the study, G1 experienced a gradual spread of A. novi-
belgii. This herbaceous perennial forms dense thickets of growth up to 120
cm tall. By summer 2012 the monitor on G1 was surrounded by this growth.
6.4.2 Healthy green roof
The Spearman Rank Order Correlation between recorded temperature

values from each season was calculated. The temperatures are highly
significantly correlated (p<0.01) with R2 values between 0.980 and 0.997.
The air temperature distributions by month are shown in Fig. 6.2. The
outliers represent the natural skewness of the monthly-averaged temperature
data, which tends to be positive in warmer months. The median temperature
is consistently lowest on G1, while the bare roof is most frequently the
highest. The difference between the two is 0.43ºC on average (data range of
differences of 4.2ºC cooler to 3.1ºC warmer, S.D. 0.7ºC), with a minimum
125
average monthly difference of 0.1 ºC in January 2012 and a maximum
average monthly difference of 1.06 ºC in August 2012. Differences are
statistically significant for 9 of the 17 months, using the Kruskal-Wallis test
(Table 6.1). Post-hoc tests using Mann-Whitney test with Bonferroni
correction showed significant differences exist primarily between G1 and
Bare and secondly between G1 and G2. The magnitude of this cooling effect
in comparison to the bare roof can be seen in Fig. 6.3. G1 is consistently
cooler than the bare roof in every month. G2 displays similar trends, with a
lessened cooling effect in winter months, and in April 2012 which was
unseasonably cold. The cooling effect on G1 increases to a maximum at the
end of the study.
2011 Apr May Jun Jul Aug Sep Oct Nov Dec
2.56 2.03 2.52 6.95* 9.1** 9.15** 6.4* 11.7** 7.68*
2012 Jan Feb Mar Apr May Jun Jul Aug
5.69 2.33 4.87 1.82 1.38 15.08** 19.71** 30.01**
Table 6.1 – Kruskal-Wallis χ2 values (* = significant at p<0.05, ** = significant at

p<0.01)
126
Fig. 6.2 – Boxplots of the monthly air temperatures at 300 mm above the surface recorded on the three study roofs. G1 is the undamaged green
roof, G2 the damaged green roof and BARE is the adjacent conventional roof surface.
127
Fig. 6.3 – Temperature difference (δT) of the monthly medians (G1 = Bare roof-G1,
G2 = Bare roof-G2)
128
Fig. 6.4 shows the daily cycle of the above-roof air temperatures averaged
hourly for the whole period that all four monitoring stands were in place. G1
displays the lowest temperatures throughout the 24 hour cycle, with a
maximum difference from the Bare roof between 21:00 and 22:00 of 1.18 ºC.
However, the Bare roof has slightly lower temperatures, differing only by
0.1ºC, from 11:00 to 12:00. G2 displays the highest temperatures during the
day, becoming cooler than the Bare roof in the late evening.
Fig. 6.4 – Average hourly temperatures for the period 28/03/2011 to 31/08/2012
The diurnal patterns of the temperature difference (δT), calculated as green

roof temperature minus bare roof temperature, can be seen in Fig. 6.5,
organised by season. A broadly similar trend in all seasons can be seen
throughout the study whereby the air above the green roofs is much cooler
than over the bare roof at night, and either less cool, or warmer, during the
day.
129
This cycle has a larger amplitude in warmer seasons with a maximum cooling
effect of -1.58 ºC seen on G1 at the 21:00 time period of summer 2012. The
timing of this maximum cooling effect in the late evening is also seen on G2
in the summers of 2011 and 2012 and spring 2012. δT values for G1 are
more likely to become positive in warmer seasons, however the timings of
the switch to positive values and when the peak positive value occurs vary.
The larger amplitudes in summer start with large daytime positive δT peaks
in Summer 2010 but this becomes dampened in subsequent years, with G1
being cooler than the Bare roof for a larger proportion of the time in summer
2012. The peak cooling effect of the green roofs tends to occur in the late
evening at around 22:00 hours, followed by a steady return of δT towards
zero until the rising sun starts the diurnal thermal budget cycle again.
130
Fig. 6.5 – Diurnal trends of the temperature difference (δT) between the green roof areas and the conventional roof area for each season
131
The peak in positive δT in the cycle occurs in late morning and tends to start
moving back towards negative values after midday, however there is no
consistent seasonal pattern to when this occurs. Temperatures above G1
are very similar to those above the Bare roof during winter 2010/11 (winter
2010 in Fig. 6.5) and that winter was characterised by being exceptionally
cold for the UK, with prolonged sub-zero temperatures in December 2010.
δT values in winter 2011/12 are quite erratic, and also do not stray far from
zero. Less solar forcing and persistent damp conditions in temperate winters
would explain the lack of temperature differences between the green and
bare roofs.
Spring 2011 was unusually warm and dry, with an absence of rain from
11/04/2011 until 05/05/2011. The tensiometer on the green roof registered
soil water content falling steadily from the usual 10% to 3.5% vol.
The upper percentiles of air temperature for each season were calculated to
identify the warmest days in each period. Of the days identified there were
two separate occasions with three consecutive warm days (over the 99 th
percentile) that could warrant particular attention. These were 20-22/04/2011
and 28-30/09/2011. Temperature profiles of these periods plus a day either
side can be seen in Figs 6.6 and 6.7. The general trend is upheld, with the
green roof areas being cooler at night and warmer during the day with the
exception of autumn 2011, where G1 also remains cooler during most of the
day, except for a brief high in the late mornings.
Fig. 6.8 shows the differences in surface temperature between the Bare roof
and G1 with a difference in the mean of the selected areas of 18.5ºC.
Surface temperatures were also lower on G2 than the path that runs through
it (Figs. 6.9 and 6.10).
132
Fig. 6.6 – Temperature profile for the three roof areas during one of the 3-day 95th
percentile exceedance periods (20 – 22 April 2011, with a day either side)
Fig. 6.7 - Temperature profile for the three roof areas during one of the 3-day 95th
percentile exceedance periods (28 – 30 September 2011, with a day either side)
133
Fig. 6.8 – Thermal image of the Bare roof, with green roof area G1 behind, taken at
14:00 on 26/07/2011. Image is inclusive of any reflected radiation incurred by non-
perpendicular angle of photograph
Fig. 6.9 – Thermal image of G2 and concrete slab path taken at 14:00 on 26/07/2011
Fig. 6.10 – Thermal image of G2 and concrete slab path taken at 14:00 on
14/08/2012
134
6.4.3 Damaged green roof
The monthly median temperatures above G2 were not significantly different

from those over the Bare roof, with the Kruskal-Wallis test (data range of
differences of 2.9ºC cooler to 8.3ºC warmer, S.D 0.56 ºC). Temperatures
were actually highest of all three roofs over this damaged section of roof from
07:00 to 15:00 in the average hourly temperature data (Fig. 6.4), reaching a
maximum average difference over the Bare roof of 0.63ºC at 11:00.
Temperature profiles from the 99th percentile events (Figs. 6.6 – 6.7) also
display this daytime warm air temperature maximum over G2. The
temperature becomes cooler than over the Bare roof by late evening,
however the cooling effect is not as pronounced as with G1. The maximum
difference in Fig. 6.4 is only 0.41ºC, between 21:00 and 23:00, compared to
1.18ºC for G1.
Spring 2011 was the first season that a monitor was deployed on G2 and it is
also when the roof was in its worst state of damage. δT values for this
season display a similar cycle to other seasons with a cooling effect at night
and warmer air temperatures during the day (Fig 6.4). For every season δT
values are consistently positive during the day reaching a peak before
midday of around 0.5ºC. In summer 2012, the peak reached 0.75ºC and δT
values were positive for approximately half of the 24 hour cycle. δT values
throughout the whole study are only negative for 61.8% of the time,
compared to 80% for G1. Peak night time cooling over G2 barely exceeded
half a degree, apart from an anomalous dip in the δT values in winter
2011/12 and in summer 2012 when an average peak cooling of -0.78ºC was
recorded at 22:00.
Surface temperatures on the damaged grass in Fig. 6.9 were on average

13.3 ºC (70%) lower than those on the paved path. The grass was 13.1ºC
(67%) cooler than the path in 2012 (Fig. 6.10), which was not as warm as the
path in the 2011 image. The selected area over the living individual of P.
lanceolata in Fig. 6.9 shows average temperatures 18.5 ºC (58%) cooler than
the paving, with a similar surface temperature of just over 25ºC to the grass
in 2012.
135
6.5 Discussion
6.5.1 Localised cooling effect
Cooler air was found over both green roof sections when compared to the
adjacent bare roof. This effect was more pronounced for G1 and the
difference in the monthly median values was often significant. This localised
cooling, of up to 1.06ºC difference in the monthly median, (and 4.2ºC
maximum of the range) is a notable influence of green roof vegetation on the
internal boundary-layer (IBL) that occurs when air passes from one type of
surface over another (Erell et al. 2011). This is comparable to the 0.94ºC
cooler temperatures found in parks in the meta-analysis by Bowler (2010),
although their figure is for daytime cooling.
The largest cooling effect tended to occur in the late evening, with maximum
negative average δT values between 21:00 and 22:00. This can be
explained as the time when long wave radiation of thermal energy
accumulated in the bare roof during the day (Meyn and Oke 2009) would be
at its greatest relative to the green roofs, rather than an increase in latent
cooling on the green roofs at this specific time. The long wave radiation
would cause a corresponding rise in the sensible heat of the overlying air
mass, especially in periods of stable atmospheric conditions. One way of
looking at the effects of urban vegetation is that it warms the environment
less, rather than cools more (Dimoudi and Nikolopoulou 2003). Wong et al.
(2003) also found the timing of significant cooling of the air at 300 mm above
a tropical, intensive green roof was from sunset to sunrise, with a maximum
difference of 4.2ºC at the start of this period.
As the study takes place on a rooftop there will be no canyon effects on

nocturnal cooling, thus the thermal properties of the surfaces become more
important for the differences seen between roofs (Erell et al. 2011).
Evaporative fluxes are weak at night so the green roofs must have cooled
soon after sunset in order to explain the difference between them and the
bare roof, which will be slowly releasing stored energy.
136
Conversely δT values were at a minimum during the daytime and often
moving into positive values, indicating warmer air over the green roofs. The
timing of this peak positive δT tended to be in the late morning to midday.
This could be due to differences in the thermal capacity of the surfaces, with
the bare roof taking longer to warm up than the relatively thin soil substrate
and plant biomass on the green roofs (Wong et al. 2007). Uncovered
substrate could also be adding to these daytime higher temperatures and this
will be discussed further in section 6.5.2.
The predicted pattern is for cooling in the daytime when plant stomata are
open and evapotranspiration rates are high, thus creating an increase in
latent heat fluxes and a consequent decrease in sensible heat. Previous
studies that focussed on surface temperatures found daytime cooling on
green roofs (Liu and Baskaran 2003; Türkeri et al. 2010) and the assumption
is that this will translate into cooler air temperatures. Surface temperatures,
on the warm sunny days chosen for spot monitoring, were lower in this study
yet air temperatures were not lower. A study on tropical green roofs found
both daytime and night time cooling of overlying air, with maximum day
temperatures 1.7ºC below the control and minimum night temperatures 0.9ºC
below the control for a grass roof at 200 mm above the surface (Jim, 2012).
A lack of predicted day time cooling on the green roofs can be hard to explain.
Vegetated surfaces can be quite complex in their thermal behaviour due to
lower wind speeds caused by friction from the plant biomass, more water
available for evapotranspiration, and plant canopy interactions with both
incoming solar and outgoing long wave radiation (Erell et al. 2011). Jim
(2012) found the 700 mm high canopy of an intensive shrub roof actually
trapped air, which acts as a heat sink giving heat accumulations at around
600 mm. Similar trapping of air could explain a diminished cooling effect on
G1 during the day, however this would only apply in summer and autumn
when the perennial A. novi-belgii is in full growth.
137
6.5.2 Effects of green roof damage
Median monthly air temperatures on G2 did not differ statistically from those
on the bare roof. This is highly likely due to the fact that the grass was
damaged by drought and mismanagement. A cooling effect was apparent in
the evenings, however, the cooling effect of G2 was lower than that of G1
while following roughly the same pattern (Fig. 6.3). The cooling effect
appeared to become stronger toward the end of the study when plant cover
was increasing. The high cooling effect seen in April 2011 is unusual,
however, as the roof was quite damaged at this point. That month was
characterised by a prolonged period of unusually warm weather, with UK
average temperatures being 3.7ºC over the 1971-2000 average (Met Office
2011). This may have greatly raised air temperatures above the bare roof,
masking the warming experienced on G2. Cooling by evaporation from the
soil is unlikely as the tensiometer data (not shown) indicate quite severe
water loss during this dry period.
Wong et al. (2007) found extensive green roof systems to be ineffective at

reducing ambient air temperature during the day. This was due to poor
vegetation coverage on the newly constructed roofs, which revealed large
areas of dark substrate that subsequently displayed surface temperatures 4 -
5ºC warmer than the roof surface before the roof garden was installed. They
concluded that the high surface temperatures of the substrate may have
been responsible for ambient air temperatures measured above vegetated
areas being higher than those measured above exposed surface. Higher
surface temperatures were not seen on G2 at the spot monitoring events.
Thus, while dark soil and its relatively low thermal capacity may be
responsible for some heating of the air, some other process may be
contributing to the higher temperatures seen. There is a concrete path that
runs through G2, which could be adding sensible heat to the local air mass,
however, it is only a narrow path taking up 9% of the area of G2, and can by
no means account for the air temperatures over G2 being similar to, if not
higher than, over the bare roof. The vegetation types that colonised G2 may
have different cooling abilities than the types on G1. A previous study found
some species displayed differences in their effect on the air temperature at
138
300 mm, however, differences were only noticeable in the afternoon of days
with high extreme temperature conditions (Blanusa et al. 2010). Structural
differences certainly exist, with A. novi-belgii forming dense thickets on G1
and S. jacobaea growing more sparsely on G2 with areas of moss and
exposed substrate in between. Some species are common to both roofs
however – such as the grasses A. stolonifera and F. rubra.
The daily cycles in Fig. 6.5 show that the maximum positive δT values for G2
were in the late morning. As with G1 this could be due to the bare roof taking
longer to heat up than the thin soil substrates. On G2 the process tends to
start sooner, reach higher maxima, and move back into negative values later
than with G1. The maximum value reached in the seasonal averages is
0.81ºC in spring 2011. As the vegetation grows back, and the substrate
becomes more shaded, one would expect this daytime warming to diminish
but a fairly high value of 0.75ºC was reached at the end of the study in
summer 2012. An increase in evapotranspirative cooling from the greater
plant biomass also apparently fails to result in the temperature profile of the
air above G2 become more similar to G1.
Night time cooling was not as strong for G2 as with G1. Maximum cooling
was reached in summer 2012 with -0.78ºC. Higher temperatures were
reached during the day in summer 2012 but cooler temperatures at night.
Average δT minimum of -0.81ºC was reached in winter 2011/12, however
this was more likely due to relative warming of the bare roof, which came
earlier than in the previous year, in this season. Vegetation could be
delaying daytime warming when the blanketing effect of snow and frost is not
present in the mild winter of 2011/12.
6.5.3 Local climate modification
The intensive green roofs in this study were able to reduce the local air
temperature when compared to an adjacent conventional roof. This may
indicate a reduced convective sensible heat flux from the rooftop
environment to the urban boundary layer which would translate into a
lowered UHI effect. The cooling effect was at a maximum in the night time
when the UHI effect is at its strongest so green roofs can help to improve the
139
thermal comfort of cities, especially at night. A question to be asked,
however, is how much spatial influence can individual green roofs have?
The effect of the size of vegetated areas sufficient to produce a measurable
effect is of great interest to planners (Erell et al. 2011). Green roofs exert a
very local effect on the climate and can be seen as separate islands. Roof
greening on a larger scale would potentially have a larger impact than
isolated separate roofs by providing an effect greater than the sum of its
parts. Research on grass plots and trees suggests that many small plots will
be more effective than fewer large ones (Armson et al. 2012). A 1ºC
lowering of air temperatures above the roof could magnify up to quite
significant UHI impacts. Cooling of 0.8ºC was proposed in a modelling study
by Dimoudi and Nikolopoulou (2003) from just a 10% increase in the ratio of
green to built urban surfaces.
The zone of influence of green roofs on ambient air temperature, and how
that can be scaled up to the city level in modelling studies, is an important
factor for future studies to determine. Vertical influences can be quite
pronounced, with the air at 100 cm over a gravel roof found to be warmer
than above a green roof (Getter et al. 2011). Cooling effects do become
less marked with height, however, and no significant influence of green roofs
was found at 160 cm above the surface (Jim 2012). Winds affect the
horizontal component of a green roof’s influence, with cooler air being found
at the leeward edge (Dimoudi and Nikolopoulou 2003). Green roofs in large
enough numbers could potentially improve the thermal comfort of pedestrians
at street level (Gomez et al. 2008) by cumulative cooling impacts on the
urban air mass.
An important point that this study has shown, is that urban green roofs, if they
are to have the desired effect, need to be carefully managed. The cooling
properties of the green roof can be quite severely impaired when the
vegetation is damaged and bare substrate dominates. Therefore, irrigation in
dry periods may be necessary and maintenance workers need to be made
aware of the correct time to cut back vegetation. The FLL green roof
guidelines recommend that contractual maintenance arrangements are made,
with an extensive list of tasks to be considered for intensive roofs, such as
140
watering, mulching and pruning 8–10 times a year (FLL 2008). Lawn roofs
need extra attention, including mowing, aerating and repeat seeding 2–12
times a year. If more than 5% of the plant coverage has failed then
replacements are needed (GRO 2011). Additionally, the intensive green
roofs in this study are not common due to the additional installation and
maintenance costs associated with them. Extensive roofs, which are much
more widespread, may have less of a cooling effect on the overlying air mass
due to lower evapotranspirative cooling from the thinner substrate and
reduced plant biomass.
Climate change predictions, such as summer rainfall decreases and more

frequent heatwaves, will put severe drought stress on urban vegetation. If
green infrastructure is to be used as climate change adaptation then it needs
to be coupled with sensible installation and maintenance plans, for instance
storage of rainwater in wetter months for irrigation in drier months.
141
6.6 Conclusion
 An intensive green roof, dominated by perennial shrubs, was able to

lower the monthly median overlying air temperature at 300 mm by up
to 1.06ºC when compared to an adjacent conventional paved roof.
 The cooling effect was strongest at night with an average difference of

1.58ºC.
 Some daytime warming was observed on the green roofs with a peak
difference relative to the bare roof in late mornings. Differences in the
thermal capacity of the roof materials and vertical mixing
characteristics could explain this.
 A combination of drought and mismanagement caused severe

damage to the vegetation on one of the green roofs. The cooling
effect was impaired on this roof with daytime air temperatures being
higher than over the bare roof for a larger proportion of the day than
over the other, undamaged, green roof. Cooling still occurred at night,
however, although with a lower maximum average relative cooling of
0.78ºC observed.
142
CHAPTER 7 RAINWATER RUNOFF RETENTION ON AN
Authors:

Affiliations:
1
2
Science of the Total Environment 461-462: 28 - 38
10.1016/j.scitotenv.2013.04.085
143
7.1 Abstract
Urban areas are characterised by large proportions of impervious surfaces

which increases rainwater runoff and the potential for surface water flooding.
Increased precipitation is predicted under current climate change projections,
which will put further pressure on urban populations and infrastructure. Roof
greening can be used within flood mitigation schemes to restore the urban
hydrological balance of cities. Intensive green roofs, with their deeper
substrates and higher plant biomass, are able to retain greater quantities of
runoff, and there is a need for more studies on this less common type of
green roof which also investigate the effect of factors such as age and
vegetation composition. Runoff quantities from an aged intensive green roof
in Manchester, UK, were analysed for 69 rainfall events, and compared to
those on an adjacent paved roof. Average retention was 65.7% on the green
roof and 33.6% on the bare roof. A comprehensive soil classification
revealed the substrate, a mineral soil, to be in good general condition and
also high in organic matter content which can increase the water holding
capacity of soils. Large variation in the retention data made the use of
predictive regression models unfeasible. This variation arose from complex
interactions between Antecedent Dry Weather Period (ADWP), season,
monthly weather trends, and rainfall duration, quantity and peak intensity.
However, significantly lower retention was seen for high rainfall events, and
in autumn, which had above average rainfall. The study period only covers
one unusually wet year, so a longer study may uncover relationships to
factors which can be applied to intensive roofs elsewhere. Annual rainfall
retention for Manchester city centre could be increased by 2.3% by a 10%
increase in intensive green roof construction. The results of this study will be
of particular interest to practitioners implementing greenspace adaptation in
temperate and cool maritime climates.
144
7.2 Introduction
In the UK, green roofs are increasingly being recognised for their role as
Sustainable Urban Drainage Systems (SUDS) (White and Alarcon 2009).
SUDS are defined as management practices designed to drain surface water
in a more sustainable way than conventional systems (CIRIA 2007). Green
roofs reduce the rate and volume of runoff, and are located close to the
source, thus helping to improve stormwater management. This, along with
other benefits such as air pollution reduction (Speak et al. 2012), local urban
cooling (Takebayashi and Moriyama 2007), and creation of habitats for
wildlife (Oberndorfer et al. 2007), means that green roofs are becoming a
more prominent factor in local government planning guidelines (MCC 2009).
Urban areas are characterised by large proportions of their surface area

being impervious to rainfall runoff, which can lead to pluvial (surface water)
flooding during heavy rainfall events. Recent flood management legislation
in the UK has improved the priority given to pluvial flooding, as 2 million
people in UK urban areas are at risk of a 1 in 200 year event (Houston et al.
2011). Climate change could lead to an increase in flooding events,
because winter precipitation increases of up to 33% are predicted for the UK
by the 2080s under a medium emissions scenario, especially in western
parts (Murphy et al. 2009). Summer precipitation is expected to decrease,
however, rainfall could become concentrated into intense downpours, with an
increase in the frequency of 20, 30, 50 and 100 year return period events for
the north west UK (Sanderson 2010).
Adaptation techniques that aim to promote infiltration and restore the urban
hydrological balance include the creation of green areas such as parks,
bunds and swales (CIRIA 2007). Due to development pressures the extent
to which new green spaces can be established in urban areas is generally
very limited. Green roofs have the benefit of not requiring upheaval of the
existing urban form, and rooftops can constitute up to 50% of the impervious
area in densely built-up urban centres (Dunnett and Kingsbury 2004).
Green roofs consist of three layers – vegetation, substrate and drainage –

and two main types exist, defined by the depth of the substrate layer, with
145
extensive roofs being generally less than 150mm and intensive roofs having
deeper soils. The deeper substrates on intensive roofs allow for a greater
diversity and biomass of plants in the vegetation layer. The overall runoff
reduction process consists of: (i) delaying the initiation of runoff; (ii) reducing
the quantity of runoff; and (iii) distributing the runoff over a longer time period
via slow release of excess substrate pore water (Mentens et al. 2006). The
amount of rainfall that is retained is of interest to urban hydrologists and flood
prediction managers.
Research on the hydrological properties of green roofs has revealed a range

of average rainwater retention efficiencies. For extensive green roofs these
are 45% (DeNardo et al. 2005; Mentens et al. 2006) and 60% (Moran et al.
2003), and cumulative annual retentions of 50% (Stovin et al. 2012) and 60%
(VanWoert et al. 2005). Carter and Rasmussen (2006) report downfall
dependent retentions of 90% for small storms and 50% for large storms.
Substrate depth has a major influence due to the direct relationship between
deeper substrates and higher retentions due to the enhanced storage
capacity, with retentions of 75% possible on intensive roofs (Mentens et al.
2006).
A number of factors can affect the retention efficiency. The season can have
a large effect, with lower rainfall totals and higher evapotranspiration rates in
warmer months and therefore shorter retention capacity recharge times
between rainfall events. A meta-analysis of seasonal runoff data showed
runoff was significantly higher during winter (80% of winter rainfall becoming
runoff and 53% in summer) (Mentens et al. 2006). Antecedent Dry Weather
Period (ADWP) is another factor related to the inter rain event recharge
potential, and Stovin et al. (2012) found low ADWPs often produce low
retention, however a high ADWP does not guarantee high retention due to
the finite retention capacity of the roof and the influence of weather
conditions during the ADWP. The intensity and duration of the rainfall is
important ,with small showers (<10mm) being fully absorbed in a study in
Texas (Simmons et al. 2008). Stovin et al (2012) found the mean retention
for 21 significant large storms to be 43%, however, the total depth retained
was only 29.3% of total rainfall due to lower retention in larger storms.
146
Slope of green roof (Getter et al. 2007), vegetation composition (Dunnett et al.
2008b) and roof position, vegetation coverage and local climate (Berndtsson
2010) have all been stated as having an influence on green roof hydrological
performance. VanWoert et al. (2005) claim the physical characteristics of the
substrate layer are more important than the vegetation, and studies should
attempt to investigate the properties of the substrate. Higher organic
contents in mineral soils confer higher infiltration rates and holding capacities
(Brady and Weil 2008). The age of the roof can therefore become an
important factor as substrate properties change over time due to build up of
organic material, and macropore creation by vegetation roots or tunnelling
invertebrates, which would increase retention capacity, and soil compaction
which would decrease it (Getter et al. 2007). There are few examples, to
date, of studies which link the soil characteristics of aged green roofs to
retention capacity. This is important, as it allows estimates to be made of the
future performance of green roofs as adaptation strategies for the increased
flood risks caused by climate change.
The majority of green roof hydrological studies are carried out on artificial
extensive green roof test beds (VanWoert et al. 2005; Dunnett et al. 2008b).
There is a need therefore for research to focus on real, intensive green roofs
to characterise the benefits afforded by investing in deeper substrates and to
see how those benefits are maintained in a real world situation, in different
seasons, and when subjected to extreme rainfall events. Similarly, as green
roofing is a relatively novel technology in the UK, there is a lack of literature
on how older green roofs perform, and, therefore, the implications of their use
as longer-term adaptation strategies in UK urban areas is not discussed.
This study aims to quantify the rainfall retention properties of an aged,
intensive green roof in Manchester city centre. A comprehensive monitoring
approach will be used to make a comparison between the green roof and an
adjacent conventional paved roof surface, which will allow any differences in
rainwater retention to be investigated for a number of different rain events.
147
7.3 Methodology
Manchester is a large city in north-west England with a population of 498 000

(MCC 2010). The main source of flood risk to the Greater Manchester sub-
region is from fluvial flooding (Kazmierczak 2011), however surface water
flooding can become more important in urban areas (Defra 2008). Flooding
was found to be the dominant climate impact in the region, over temperature
extremes, high winds, fog and drought, with some evidence that flood events
have been becoming more frequent over recent decades (Smith and Lawson
2012). The majority of pluvial flooding events occurred during the summer
months, and this may indicate the significance of short duration, heavy
rainfalls characteristic of this season, and/or an increase in the impervious
nature of the landscape leading to increased surface runoff (Smith and
Lawson 2012). Surface water flooding events have the potential to impact
seriously on Manchester’s critical and transport infrastructure (Kazmierczak
2011), which can result in loss of power to homes and services, and large
financial costs for cleanup and recovery.
148
Fig. 7.1 – Location of the study roof
A green roof within the University of Manchester campus, on the Precinct

building, was chosen for the study (Fig. 7.1). The area is classified as open
midrise, characterised by a fairly open arrangement of buildings of 3 – 9
storeys with some trees, typical of an inner city university campus area
(Stewart and Oke 2012). The roof was chosen because it has a conventional
149
roof area (900 m2), consisting of concrete paving slabs, adjacent to a large
(408 m2) intensive green roof, which is 43 years old, and has an average
depth of 170 mm. The roof is not within rain shadows of any adjacent taller
buildings. Fig. 7.2 shows cross sectional representations of the two study
roofs. The green roof is of fairly standard construction with the vegetation
and substrate layers divided from the ‘egg box’ design plastic drainage layer
by a fibrous membrane. The roof itself is protected by a tough geotextile
membrane. The bare roof is a conventional roof surface consisting of
concrete paving. The 60 x 60 cm paving slabs sit on top of an insulating
polystyrene cushion and a plastic foam membrane that are impermeable to
water.
Fig. 7.2 – Cross section of the green roof (left) and bare roof (right) showing layer
depths
150
The green roof is of particular interest due to its age, the fact that the roof
was not constructed specifically for the study, and due to it having a mineral
soil substrate rather than the more usual, prefabricated, light weight
aggregate (LWA) based substrate. Green roof studies with an experimental
component often use specially constructed extensive green roof test rigs. It
can be argued that studies of this type can over or underestimate the benefits
of actual green roofs. For example, Stovin et al. (2012) recognised that the
small size (3 m2) of their green roof setup could underestimate the lag and
attenuation of their runoff hydrograph. Artificial experimentation studies can
also overestimate the benefits because the test rigs are 100% green
coverage, whereas in reality green roofs often have quite high proportions of
conventional roof surface, at the periphery, or to provide access for
maintenance. Investigating a well established, real green roof which is
subject to local weather patterns, will reveal green roof characteristics that
can be sensibly applied to the real world.
Table 7.1 reveals the vegetation on the green roof is dominated by the
grasses Agrostis stolonifera and Festuca rubra, and the invasive weed
Plantago lanceolata, which are generally lower in height and with thinner
leaves than the less common species. While contributing to less above
ground biomass per unit area, they form a substantial lower canopy. The
less common species form a secondary canopy above the low lying primary
canopy and in summer the vegetation can reach considerable heights over 1
m. In winter the shrubs die back and the biomass is mostly dominated by the
base canopy plants and mosses. Average surface coverage in summer is 97%
and remains as high as 85% in winter due to the mosses. There are
individuals of larger plants – Buddleja davidii (2 m) and Rubus fruticosus (1 m)
in the southern part of the roof, and infrequent occurrences of up to 30
different forb species. Plant height has been shown to be significantly
positively correlated with rainfall retention (Dunnett et al. 2008b) and the
vegetation in the present study, with its high spatial coverage and double
canopy, would be expected to have a considerable influence on retention.
151
Species Surface Average Leaf form Duration
coverage height
summer (cm)
(%)
Agrostis stolonifera 30 15 Narrow Perennial

graminoid
Festuca rubra 20 20 Narrow Perennial

graminoid
Plantago lanceolata 20 10 Rosette Perennial

lanceolate
Senecio jacobaea 15 60 Pinnately lobed Biennial
Mixed: 15
Epilobium ciliatum 100 Lanceolate Perennial
Symphyotrichum novi- 100 Lanceolate Perennial
belgii 20 Tri-foliate Perennial
Trifolium pratense 30 Tri-foliate Perennial
Ranunculus sp. 40 Lanceolate Perennial
Rumex crispus 2 moss Perennial
Various moss species
Table 7.1 – Characteristics of the dominant green roof vegetation
7.3.2 Monitoring
Most green roof hydrological studies employ methods which divert runoff
from green roof sections and quantify it via tipping rain buckets (VanWoert et
al. 2005), collect it in graduated tanks (Nardini et al. 2012), or monitor the
depth of water in collection tanks using a pressure transducer (Stovin et al.
2012). The structure of the precinct roof drainage system was such that
collection of runoff was not practical, therefore a novel approach was
employed. Square, plastic pots were constructed to sit in the top of two
drainage downpipes that have a diameter of 150 mm – one draining the
green roof section and one draining the adjacent conventional roof section
(Fig. 7.3). Runoff entering the drains was first channelled into the pots using
specially constructed plastic collars fixed into the drain openings with
waterproof sealant, and water could then leave the pots via a 60° v-notch cut
152
into one side. Pressure transducers (Hobo water level logger U20-001-04)
were placed in the pots and calibration relationships determined in a
laboratory to relate pressure to runoff (R2 = 0.91 for the green roof pot, R2 =
0.8 for the bare roof). Pressure was logged every minute, starting
08/09/2011 and finishing 12/10/2012. Gaps exist in the data due to logger
failure from 24/09/2011 to 08/10/2011 on the bare roof and from 20/05/2012
to 06/06/2012 on both roofs.
Fig. 7.3 – Plan of the Precinct roof showing drainage catchment areas for the bare
and green roofs
Effective drainage areas for the two drains were determined from
consultation with the University campus estate manager and reference to the
original architectural roof survey plans. The area of the green roof catchment
was estimated to be 384 m2 and the area of the bare roof was estimated at
228 m2. Both roof drainage areas include sections of a central glass atrium
on the roof (Fig. 7.3). This means that the green roof drainage area is not
entirely green and there is a significant input (20% of the area) from a non-
153
green surface. Rainfall-runoff relationships for this sloping glass roof were
determined for a range of different size rain events using a tipping rain bucket
and subsequently used to estimate the amount of runoff on the green roof
that is attributable to the atrium for each rain event. This was then subtracted
from the overall runoff measured in the drain, on an event by event basis, to
gain the runoff attributable to the green component of the roof.
Rainfall data were obtained from the Whitworth Observatory (UoM 2012),
situated 150 m from the precinct roof (Fig. 7.1), which employs a Theis laser
distrometer, with an accuracy of >90% and resolution of 0.001 mm/hr. The
data were available as rainfall totals in 10 minute intervals, thus logger data
were averaged into the same 10 minute interval for analysis.
Twenty samples of the substrate layer were taken and analysed in a

laboratory to determine the following parameters: soil texture (particle size
analysis); field capacity (Water content after a water-saturated soil sample is
drained with suction in a Buchner funnel (Veihmeyer and Hendrickson 1949));
dry bulk density (cylinder sampling method); porosity (from bulk density and
particle density); and organic matter content (loss on ignition). All methods
used are standard soil analysis techniques (Rowell 1994). A Guelph
permeameter was employed on the green roof to determine the saturated
hydrologic conductivity, a measure of permeability. This method is
acceptable once the spatial variability of soil permeability is accounted for
with multiple measurements (Salverda and Dane 1993), thus ten
measurements were averaged.
154
7.3.3 Data Analysis
Rainfall events are defined as being separated by continuous dry periods of

at least six hours, in accordance with a number of previous researchers
(VanWoert et al. 2005; Getter et al. 2007; Stovin et al. 2012). Consequently
the study period contained 254 rainfall events, ranging from 0.01 mm (the
lower limit of rainfall detection) to 56.08 mm, of which 178 both produced
runoff and occurred outside the periods when the loggers malfunctioned.
The sensitivity of the pressure transducer to long term changes in the water
levels in the runoff capture pots, and gaps in the data, means that runoff
volumes were more sensibly calculated on an event by event basis, thus 69
(39%) of the events that produced runoff, and had data, were investigated.
The events were classified as small (< 2 mm), medium (2 – 10 mm) and
large (> 10 mm), using the criteria originally proposed by Getter et al (2007).
The events selected for analysis were composed of 25% small, 56% medium
and 19% large. This compares favourably to the distribution of events in the
whole study period, with 54% small, 30% medium and 17% large. The lower
proportion of small events in the analysed data subset arises because 54%
of the small rain events did not produce runoff. Events were sampled
approximately equally for each month of the study to ensure a good
representation of different seasons and weather conditions. Large rain
events had the potential to become significant in that they may exceed the 1
year return period rainfall for Manchester, obtained from the Flood Estimation
Handbook cd-rom (NERC 1999). In Fig. 7.4 it can be seen that six events fall
between the 1 and 2 year return period estimation curves, and these events
were looked at more closely to see how the green roof functions in extreme
events.
155
Fig. 7.4 – Rainfall characteristics for the 69 analysed rain events with return period estimates of 1 to 5 year events for Manchester from the
Flood Estimation Handbook cd-rom
156
Data were organised by season for analysis, defined as winter (Dec, Jan,
Feb), spring (Mar, Apr, May), summer (Jun, Jul, Aug) and autumn (Sep, Oct,
Nov). Individual months were also classified as wet, dry or average, based
on their relation to the 30 year climate (Fig. 7.5) with those over 140%
average rainfall being a wet month, less than 60% a dry month, and the rest,
average months. Antecedent Dry Weather Period (ADWP), rainfall duration,
rainfall depth and peak 10-minute storm intensity were recorded for each
rainfall event. These were used to divide the rainfall retention data for
analysis. For example, events were divided into those with a short ADWP of
6 – 20 hours and those over 20 hours in accordance with Stovin et al. (2012).
The division thresholds of 8mm/hr for peak intensity and 10 hrs for duration
were chosen by examining the spread of the data and choosing values which
sensibly split them.
Fig. 7.5 – Monthly rainfall expressed as a percentage of the climate average for
Manchester (1981 – 2010)
157
Data were checked for normality using the Anderson Darling test and green
roof runoff data were normal, however the bare roof data and explanatory
variables were not, hence non-parametric analyses were employed. All
statistical investigations were carried out using R (version 2.11.1).
7.4 Results
7.4.1 Climate and Weather
Greater Manchester’s climate is maritime and temperate. Climate data for

the period 1981 – 2010 show average annual rainfall to be 828.8 mm, and
these data are referred to henceforth as ‘normal rainfall’. October to
December is generally the wettest period (80.7 – 92.5 mm) and February to
May the driest (51.4 – 61.2 mm). Fig. 7.5 shows the monthly rain totals for
the study period, with November, March and May having half the normal
rainfall, but five of the months having approximately double. These wet
periods mostly occurred in summer 2012, which means the study period
contains an uncharacteristically wet summer. The total rainfall for the study
period was 1249.6 mm which is 126% of the 1981 – 2010 climate average
rainfall expected for the same 14 month period.
158
7.4.2 Soil properties
Permeability pH Field Bulk Particle Porosity Organic

(cm/sec) Capacity Density Density content
(g/cm3) (θ)
3
(max (g/cm ) (g/l)
water
capacity
for FLL)
Study 1.68 x 10-3 6.5 38% 1.03 2.47 0.58 202.54

roof
FLL 5 x 10-4 6– 45 – - - - < 90

guideline 8.5 65%
value
Table 7.2 – Selected soil characteristics and comparison to the FLL guideline values
for an intensive green roof
The soil on the intensive green roof is a sandy loam, according to the UK
particle size distribution classification system. Table 7.2 displays the
characteristics of the substrate layer, and, where possible, a comparison to
the Forschungsgesellschaft Landschaftsentwicklung Landschaftsbau (FLL)
guidelines for an intensive green roof (FLL 2008). The saturated hydraulic
conductivity, at 0.0017 cm/sec, is quite low when compared to values from
previous work, but still corresponds to a soil type with good drainage
(Terzaghi et al. 1996) and is higher than the FLL permeability guideline value
of 5 x 10-4 cm/sec. Previously published values include 0.03 cm/sec
(Bengtsson 2005), and 0.4 cm/sec (Nardini et al. 2012), however these are
for extensive green roofs which display much higher permeability due to
having substrates dominated by LWAs such as expanded slate and pumice.
The bulk density was fairly low at 1.03 g/cm3, with sandy loams being
normally around 1.44 g/cm3 (Palla et al., 2012). The low bulk density is
probably attributable to the high organic content, which is much higher than
the FLL guideline suggestions, and at 9% by weight, is quite high for a
mineral soil. This value lies in the middle of published values, where
159
available, which range from 3.5 % (Nardini et al. 2012) to 16% (Palla et al.
2012). It must be stressed that the extensive roof substrates in those studies
tend to be carefully designed, prefabricated mixes of LWA material with
sphagnum peat additions for the organic content, as opposed to mineral soils.
The soil pH of 6.5 is normal and porosity of 0.58 is fairly average.
The field capacity, at 38%, is moderate. This is a laboratory determined

value and in practice will not reach this. By using the same logic as
Bengtsson (2005), the difference between the field capacity and the wilting
point represents a volumetric storage capacity for the soil. Using a value of 9%
for the wilting point of a sandy loam (Saxton et al. 1986), this means 49 mm
storage is theoretically possible in the 170 mm deep substrate. Overall the
soil characteristics indicate a fairly normal sandy loam mineral soil without
much compaction and with high organic matter. Observations throughout the
year indicate a healthy ecosystem on the roof, with earthworms, ants,
butterflies, bees and birds being seen in large numbers.
7.4.3 Runoff initiation
No statistically significant difference was seen between the two roofs with
regards to commencement of runoff due to the influence of the sloped glass
atrium which forms a significant part of the drainage areas of both roofs, and
tends to generate runoff soon after rainfall starts. The lag between the
timings of peak rainfall and peak runoff was calculated for each event,
however the effect of the atrium is expected to impact upon this also. Due to
the discretisation of the rainfall data, the analysis is restricted to a 10 minute
resolution. Timings of runoff commencement ranged from 0 to 90 minutes on
both roofs with a mean of 17.6 minutes on the green roof and 15.2 on the
bare roof (medians of 10 minutes for both roofs).
Fig. 7.6 shows rainfall and runoff for the two roofs for two events – one in
December 2011 and the other in September 2012. The similar lags before
onset of runoff can be observed.
160
Fig. 7.6 – Rainfall and runoff for two events (a) Dec 2011 (b) Sept 2012
The smallest rain event to produce runoff was 0.26 mm on 04/08/2012. The
ADWP for this event was only 9.5 hours and while the previous rain event
was only 0.04 mm, there had been rain on six previous days, so the roofs
would have been close to field capacity. The largest rain event that did not
produce runoff was 1.81 mm on 26/09/2012. The ADWP for this event was
161
only 14 hours and the previous rain event was 1.18 mm, but this was at the
start of a very warm period, so the atrium is assumed to have been quite dry.
Mean ADWP for all analysed events was 30 hours.
Fig. 7.7 – Regression plot of rainfall and runoff depth for (a) green roof and (b) bare
roof, with 1:1 lines included
162
7.4.4 Inter-roof comparison
Rainfall depth versus runoff depth of the 69 analysed events for the two roofs
(Fig. 7.7) shows runoff being generally lower on the green roof than on the
bare roof. This is indicated by regression line slopes of 0.56 on the green
roof and 0.68 on the bare roof, and a visual comparison with the 1:1 line.
Correlations between rainfall and runoff were high, with r=0.85 for the green
roof and r=0.97 for the bare roof (Spearman’s rank correlation, p<0.01). The
box plot in Fig. 7.8 shows the retention, expressed as percentage of rainfall
input to the roof that is retained. Median retention is higher on the green roof
(65.7%) than on the bare roof (33.6%) and the difference is highly significant
(paired Wilcoxon test, V=2408, p<0.01). Average runoff retained on the
green roof was 4 mm, or 2.4% of the substrate depth. There is a lot of
variation within the retention data, with the green roof retaining between 22%
and 100% and the bare roof between 8% and 72%.
Fig. 7.8 – Boxplot to show mean average runoff retention of all rainfall events for
both roofs
163
Fig. 7.9 – Mean runoff retention comparisons between the green roof and bare roof
for (a) season, (b) rainfall depth, (c) rainfall duration, (d) peak rainfall intensity, (e)
ADWP and (f) precipitation anomaly
Fig. 7.9 shows the results of separating the data by the six factors, outlined in
the methodology, which may influence the retention. The only significant
retention influences arose on the green roof with respect to season and
rainfall quantity – with retention being significantly lower in autumn (Kruskal-
Wallis X2 = 8.876, p<0.05) and for large rainfall events (X2=7.633, p<0.05).
While not significant, it is interesting to note the similar pattern on both roofs
in Fig 7.9(f) of decreasing retention moving from wet months to dry months,
contrary to what the expected trend would be. ADWP and rainfall duration
164
also show the opposite trend to that expected, with longer durations
producing higher retention (fig. 7.9c) and longer ADWPs producing lower
retention (Fig. 7.9e). These differences are small, however, and within the
standard deviations.
Fig. 7.10 – Regression plots of runoff retention against (a) ADWP and (b) rainfall
duration, for the green roof
165
Regression analysis using these explanatory variables was employed to see
if it could be utilised as a robust tool for predicting the runoff response to any
given rainfall. The scatter plots for ADWP and rainfall duration, in Fig. 7.10,
show a large amount of scatter. Linear and non-linear (logarithmic, power
and quadratic) regressions were attempted using ADWP, rainfall depth,
duration and peak intensity as explanatory variables, however, they all
resulted in poor R2 values and no significance. The large variation at smaller
values of the explanatory variables masks any relationships that may exist.
Similarly multiple regression with ADWP, peak rainfall and duration was
found to be non significant (R2 = 0.04, F=1.014, p=0.39).
Fig. 7.11 – Regression tree for runoff retention percentage on the green roof. ‘Yes’
and ‘no’ splits are assigned to left and right respectively
166
The data were subjected instead to non-parametric CART (Classification and
Regression Tree) analysis. Models were fitted using binary recursive splitting
of the data set into increasingly homogeneous subsets and the resulting tree
can be seen in Fig. 7.11. Rainfall is seen to be the primary determinant of
retention with larger rainfalls producing low retention, and smaller rainfalls
leading to further branches. Peak intensity was excluded by the analysis.
ADWP and duration both produce conflicting branches which reflects the
complicated interactions within the data. For instance, to get to the highest
retention figure of 86.1% requires a rainfall below 8.6 mm, ADWP greater
than 12.8 hours, and a duration less than 8.8 hours, which make sense, but a
duration greater than 18 hours is also indicated.
7.4.5 Storm events
The six events with rainfall/duration characteristics that exceed those of 1

year return period events are summarised in Table 7.3. The largest retention
depth in the study was 22.07mm for the storm event starting 02/12/2011.
The bare roof also had quite high retention for this event, but the rainfall
duration was very long, at 105 hours, so the rainfall was very prolonged with
brief dry periods (< 6 hrs) within it during which the roof could dry out a little.
This maximum depth equates to 13% of the substrate depth, which is much
lower than the 25% achieved in Stovin et al (2012). Four of the events
occurred in the summer period, and represent convectional storms.
167
Retention
10 min % depth
peak % Retention Retention Retention
Total Rain rainfall Retention bare green as % of depth
Rain duration intensity ADWP green roof substrate bare roof
Event (mm) (hh:mm) (mm/hr) (hh:mm) roof roof (mm) depth (mm)
02/12/2011 56.08 104:30 18.66 9:30 39.36 33.85 22.07 12.98 18.98
06/07/2012 22.30 12:00 9.06 8:20 73.22 45.59 16.33 9.61 10.17
19/07/2012 30.35 30:40 22.5 18:20 57.15 41.12 17.34 10.20 12.48
24/08/2012 10.19 4:00 9.96 18:30 58.65 28.80 5.97 3.51 2.93
29/08/2012 19.31 9:50 68.22 26:00 42.33 11.81 8.17 4.81 2.28
23/09/2012 33.87 27:20 6.36 49:50 36.58 20.88 12.39 7.29 7.07
Table 7.3 – selected characteristics of the 6 storm events greater than 1 year return period
168
The mean retention of the storm events on the green roof is 51.2%, which is
lower than the mean of all the study events, however, retentions are quite
variable and one significant storm event (22.3 mm) displayed retention of
73.2%. Bare roof retentions, at 30% on average, are similar to those in the
rest of the study. Further evidence that ADWP is not a very good predictor of
retention capacity comes from the 23/09/2012 event, which had the longest
ADWP of the six storms at just over two days, but the lowest retention,
probably because the rainfall was the second largest of the study.
7.5 Discussion
7.5.1 Retention efficiency
The median retention of 65.7% achieved by the intensive green roof is lower
than the median of eleven intensive roof studies of 75% found in the meta-
analysis undertaken by Mentens et al (2006). This is potentially due to the
unusually wet summer in the study period, with above average rainfall from
June to September. Schroll et al. (2011) noted that cold, wet environments
can be challenging ones for green roof retention performance. The green
roof would not have had much chance to restore its retention potential
between rain events, despite expected higher evapotranspiration rates in
summer from the greater plant biomass and elevated air temperatures. None
the less, this figure is higher than retentions for extensive green roofs,
previously quoted. Therefore the age of the roof does not appear to have
deleteriously affected the retention ability of the green roof. This could be
attributed to the high organic matter content that will have accumulated over
43 years of plant growth. The maintenance regime on the green roof has
ensured minimum disturbance to the substrate layer. Getter et al (2007)
found organic matter doubled in a green roof over 5 years which increased
the porosity and the water capacity of the roof. Therefore green roofs might
be expected to improve some hydrological benefits with time, and indeed
many of the soil properties were indicative of a normal substrate (Table 7.2).
Anecdotal observations of a thriving green roof ecosystem provide evidence

that the green roof was in a healthy ecological state. Earthworms are vital for
the incorporation of fresh plant material into the soil structure but they are
169
rare on extensive roofs due to lack of protection from winter frosts in the thin
substrate. Additionally, deeper substrates are preferable because they
reduce drought stress on the vegetation and allow a broader selection of roof
plants, away from the traditional monocultures of sedum plants found on
extensive roofs. There is evidence that runoff is reduced more by taller
plants with denser root systems (Dunnett et al., 2008b) and grasses can be
more effective than sedum and forbs (Nagase and Dunnett 2012). The larger
biomass on intensive green roofs will also increase the other benefits such as
air pollution reduction (Speak et al. 2012).
The green roof retention is double the average retention of 33.6% on the
bare roof. The paved roof used in this study is not as common a roof surface
as other forms of conventional roof, such as gravel roofs with 24% retention
and traditional smooth roofs with 15% retention (Mentens et al., 2006). The
increased retention may arise from the roughness of the paving slab surfaces
but mostly from the increased surface area brought about by the cushions
and foam membrane underneath the paving slabs (Fig. 7.2). While this roof
is interesting so far as it provides information on a different form of roof
structure, it must be remembered that most typical urban roof surfaces have
less retention capacity than the bare roof reported here. Consequently, the
relative green roof retention demonstrated in this study would be greater
when compared to more common roof surfaces.
As this study was carried out on a real roof, certain peculiarities unique to this
roof had to be taken into account, especially the fact that the green roof drain
had a significant impact from the glass atrium roof. It is expected that the
minimum rain to produce runoff and maximum rain to not produce runoff
would have been quite different, without the atrium influence. It also had the
effect of removing the delay effect seen in most previous studies (VanWoert
et al. 2005; Berndtsson 2010; Fioretti et al. 2010). Urban drains often receive
runoff inputs from heterogeneous sources and it is interesting to note that just
20% of the catchment area being non-green was sufficient to bring the mean
peak delay on the green roof closer to that of the bare roof. However, it must
be noted that the paved roof in this study, with its apparent increased storage
capacity relative to other conventional roof surfaces, would likely delay runoff
170
initiation in comparison to these other surfaces. The evidence from this study
shows that green roof installations with the aim of delaying and attenuating
peak runoff must achieve in excess of 80% vegetation cover over the roof
catchment in order to have any effect. However, where this isn’t possible
there are still benefits to be gained from improved retention rates.
Bengtsson (2005) described runoff generation only beginning once field

capacity is reached in the substrate of an extensive roof, with storage being
equal to the field capacity minus the wilting point - when the moisture fraction
in the substrate is soil-bound. The theoretical holding capacity of around 49
mm of rain, predicted by this method, was not reached in practice, with
retentions of on average 4 mm found and a maximum of 12 mm. This could
be because the above average rainfall during a large part of the study
prevented the substrate layer from drying out to levels capable of retaining
such large inputs of rainfall. The intensive roof would be expected to retain
the majority of a given depth of rainfall, below 49 mm, but so many other
factors are involved that it has proved difficult to predict.
Significantly lower retention was seen on the green roof in autumn, when
there was above average rainfall. High rainfall events lowered the retention
capacity of the green roof as the substrate became close to field capacity.
Carter and Rasmussen (2006) also found large rain events produced lower
retention. Multiple regression with explanatory variables was carried out to
see if it could be a robust modelling tool for estimating the runoff response of
storm events on green roofs. Similar to Stovin et al. (2012), explanatory
variables did not have a clear influence on retention. ADWP should in theory
be a good predictor of retention capacity, as was found in a study on
extensive roofs displaying 82% average runoff reduction by event (Voyde et
al. 2010), but the interaction of the other factors, such as seasonal effects on
evapotranspiration, produces too much variation in the results to pick out
clear individual trends. This also manifested in the CART analysis, with
duration greater than 18 hours, which in theory should lessen retention,
included among the criteria which produce the largest retention. A
continuous simulation moisture flux model, proposed by Stovin et al (2012) is
preferable to regression based methods. This conceptual approach takes
171
into account rainfall inputs and moisture fluxes out of the substrate via runoff,
or evapotranspiration during ADWPs, to predict the available retention depth
in the substrate and therefore how it may be expected to respond to further
rain inputs.
Contrary to expectations, season was found to have not as strong an effect

on retention for the period of this study. Except for autumn, as discussed
above, the seasons displayed very similar retention averages. Schroll et al.
(2011) found a medium-only treatment did not significantly differ in retention
to an extensive green roof in winter but summer differences were significant
with 65% retention on the green roof. In summer the lush, tall vegetation on
intensive roofs should dramatically increase evapotranspiration rates,
however, in an anomalously wet summer this effect may be masked by the
above average rainfall or the plants may be adversely affected by
waterlogged conditions. The highest retention values of the events in the
present study were achieved in summer, with two instances of complete
retention on the green roof, and 72% retention on the bare roof. A median
retention of 70% in summer (Fig. 7.9a) is still an encouraging result, and
permits the consideration of green roofs as a measure to mitigate the impact
of heavy summer downpours.
The extreme events in the study enabled an assessment of an intensive

green roof’s ability to cope and the retention was considerable. The age of
the roof is not an issue affecting the effectiveness of green roofs as an
adaptation measure. The case may be different for extensive roofs, and
longitudinal studies could help to reveal any significant changes occurring in
LWA substrates.
7.5.2 Implications for Manchester city centre
Using the roof quantification work of Speak et al (2012) the green roof spatial
extent potential of Manchester city centre and the Oxford Road corridor, is
approximately 50 ha of the total 326 ha. A feasible green roof construction
scenario of 10% (14.9 ha) of Manchester’s roofs can be postulated. With
65.7% average rainfall retention for an intensive roof, and employing
retention estimates from Mentens et al. (2006) of 45% on extensive roofs and
172
15% on traditional roofs, the increased retention of green roofs over bare
roofs for an average year’s rainfall can be estimated. Intensive roofs retain
2.3% more of the rainfall falling on the selected inner city zone, and extensive
roofs 1.3%. This is comparable to the figure for an identical hypothetical
green roof construction scenario, in Brussels, which found runoff was
reduced by 2.7% using extensive roofs (Mentens, 2006).
Although these figures may at first seem low, it highlights that green roofs
alone cannot be relied upon to provide all the SUDS benefits in a city
because flat roofs are a limited resource within urban areas (Carter and
Jackson 2007). However, one must also remember that green roofs have
multiple benefits stretching beyond their hydrological function. Intensive
roofs, in particular, should be considered by city planners because they offer
a higher retention capacity compared to conventional roof covers, and the
potential for varied and lush vegetation.
7.6 Conclusion
 Hydrological monitoring of urban roof drainage was achieved using a

compact v-notch weir methodology.
 Average runoff retention of 65.7% was achieved on an intensive green

roof, compared to 33.6% on an adjacent paved roof.
 Organic matter content in the 43 year old roof substrate was relatively
high, which may increase the retention capacity of the roof.
 Season and rainfall amount had significant impacts on retention,

however, many other explanatory variables such as ADWP and peak
rainfall intensity had no demonstrable, significant impact.
 Intensive roof construction on 10% of the rooftops in Manchester city

centre would increase annual rainfall retention by 2.3%.
 Finite retention capacity of the substrate means that maximum

retention may be expected to decrease significantly for large storm
events. Actual retention will always be less than maximum and can
173
even be zero. Using green roofs as solutions to inner city flood risk
must therefore proceed with caution as attention needs to be paid to
the duration, intensity and rainfall depth of storm events in the context
of feasible retention provided by the substrates.
174
CHAPTER 8 METAL AND NUTRIENT DYNAMICS ON AN
Authors:

Affiliations:
1
2
Submitted to Environmental Pollution – May 2013
175
8.1 Abstract
The degradation of runoff water quality is a common problem in urban

environments. Potentially toxic pollutants entering runoff from the urban
watershed, such as heavy metals, and excess nutrients can have
deleterious impacts on receiving water bodies. Green roofs have been
suggested as a remediation method because wet and dry deposited
pollution might be expected to be retained by the vegetation and soil
substrate layers. There is a need to quantitatively validate this property of
green roofs. Rainfall and runoff quality was compared between an aged
intensive green roof, and an adjacent conventional roof surface, for pH,
conductivity, colour and concentrations of dissolved Cd, Cr, Cu, Fe, Mn,
Ni, Pb, Zn, Cl, NO3, PO4 and SO4. Metal concentrations were also
determined in the green roof substrate and in dusts from the conventional
roof. Nutrient levels in the runoff were mostly below Environmental
Quality Standard (EQS) values, however some PO4 data were found to
be higher. NO3 levels were low in the green roof runoff, which conforms
to the expectations of a green roof that has not been artificially fertilised.
Median concentrations of Cu, Pb and Zn were in excess of EQS values
for protection of surface water. Green roof runoff was significantly higher
in Fe and Pb (X2 = 76.1 and 31.2 respectively, p < 0.001) than on the
bare roof and in rainfall. Fluxes revealed the green roof to be a potential
source of Pb. High concentrations of Pb within the green roof soil and
bare roof dusts provide a likely source of Pb leaching. The origin of the
Pb is likely to be from historic atmospheric deposition from vehicular
combustion of leaded petrol, industrial sources, and vehicle component
wear. This implies that the green roof is acting as a store of legacy
pollution, which has implications for the use of green roofs as mitigation
tools for air pollution.
176
8.2 Introduction
One of the problems of growing urbanisation is the potential for high pollutant
loads in urban storm runoff (Zobrist et al. 2000; Rocher et al. 2004). Urban
runoff can have significant adverse ecological effects in receiving water
bodies. Heavy metals in particular have been found to be highly persistent
and toxic to aquatic flora and fauna at low environmental concentrations
(Pizzol et al. 2011). Nutrients can also be a problem when urban surface
waters discharge to water bodies that are already nutrient-rich, resulting in
eutrophication (Ellis and Mitchell 2006). The poor water quality of urban
runoff results from the accumulation of particulate matter and dissolution of
environmentally harmful substances as it is conveyed over the impervious
roofs and roads in the urban watershed (Lye 2009). On rooftops these
substances include: heavy metals leached from roof surface materials
(Rocher et al. 2004); dry and wet deposited air pollutants such as SO2 and
NOx (Fowler et al. 2007); airborne dusts from vehicle use, industry,
construction (Robertson et al. 2003; Gobel et al. 2007); and salts from road
de-icing (Lundmark and Olofsson 2007). The distribution and concentration
of pollutants in runoff is related to the nature of the surfaces (Mendez et al.
2011), as well as local patterns of wet and dry atmospheric deposition
(Forster 1999). This Urban Diffuse Pollution (UDP) can be difficult to identify,
measure, and control, but there is nonetheless a need to control UDP
sources in order to preserve water resources from the pressures of
urbanisation and climate change (Defra 2011b). One way to control UDP is
to employ catchment based measures such as Sustainable Urban Drainage
Systems (SUDS), which are preferable because they involve the multiple
agencies identified in the process of UDP generation (Defra 2012). They
also place a greater emphasis on improving the quality of the water resource
by filtering out pollutants where possible (CIRIA 2007). SUDS technologies
that improve the hydrological function of rooftops, such as green roofs, have
the potential to be a great benefit for alleviating the problem of UDP, because
roofs can account for 50% of the urban impervious area in the UK (Dunnett
and Kingsbury 2004).
177
The ability of green roofs to reduce the volume of stormwater runoff has been
frequently reported (Getter et al. 2007; Spolek 2008; Stovin et al. 2012) with
annual runoff retentions of between 50% to 100% (Rowe 2011). Green roofs
are also efficient at capturing air pollution such as NOx, SO2, O3(Currie and
Bass 2008), and PM10 (Speak et al. 2012) and it is generally expected that
the pollutants would be retained in the vegetated roof (Berndtsson et al.
2009). However, there is the possibility that air pollution that has been
captured by vegetation will eventually leach into the roof runoff, thus trading
air pollution for water pollution (Rowe 2011) unless runoff is treated.
Studies have revealed green roofs to be both sources of water contaminants,

as well as sinks. Berndtsson et al. (2006) found extensive sedum-moss roofs
to be a source of pollutants such as Cd, Cr, Cu, Fe, K, Mn, Pb, Zn, NO 3 and
PO4 when compared to a non-vegetated roof, however it was concluded that
the overall runoff quality can be considered good in relation to water quality
standards. Mendez et al. (2011) found elevated As and high Pb in the first
flush of an extensive green roof. The green roof substrate itself can be a
source of metals, as was found in a study with different commercial soil
assemblages based on either clay and peat mixtures or inorganic volcanic
material and compost. Significant quantities of Fe and Al were leached from
some substrate types and considerable quantities of Cu from another
(Vijayaraghavan et al. 2012). Most studies, however, find no substantial
release of heavy metals from green roofs. A comparison of a Swedish
extensive roof with an intensive roof in Japan found neither roof to be a great
source of metals (Berndtsson et al. 2009). One of the first studies to
investigate green roof runoff quality found percentage retention of
contaminant influx to be 95% for Pb and 88% for Cd (Köhler et al. 2002).
Metal concentrations found in runoff are generally similar to those in
precipitation, and consequently, when the reduced quantity of runoff by green
roofs is taken into consideration, the net effect is of a reduced metal flux
(Berndtsson 2010). Concentrations of heavy metals in runoff coming from
different urban surfaces, including tiles, concrete and copper roofs, revealed
green roofs to have the lowest amount of metals, except for Zn (Gobel et al.
178
2007). Green roofs also generally act to mitigate mild acid rain by raising the
pH from between 5 to 6 in rainfall to between 7 and 8 (Berndtsson 2010).
The most common impact on green roof runoff quality comes from N and P.
High nutrients have been frequently found in green roof runoff
(Vijayaraghavan et al. 2012; Gregoire and Clausen 2011; Teemusk and
Mander 2007; Monterusso et al. 2004) with the N and P amounts being
directly related to organic matter content (Moran et al. 2003). Additionally,
the use of artificial fertilisers on green roofs is a major source of nutrients.
Emilsson et al. (2007) demonstrated how conventional fertilisers cause high
nutrient concentrations in the runoff, and this was influenced by not only the
vegetation system type, but by the age of the vegetation mat. Old mats
reduced the risk of nutrient leaching, potentially due to temporary storage in
the substrate and enhanced uptake by the well-established vegetation.
Consequently, reduced application of fertilisers is often suggested by authors
to reduce the nutrient concentrations in green roof runoff (Emilsson et al.
2007; Teemusk and Mander 2011). Dissolved Organic Carbon (DOC) has
also been found to be high in green roof runoff due to the presence of
organic material (Mentens et al. 2006; Berndtsson et al. 2009). This can be
an issue because the discolouration can be problematic in situations where
runoff is collected for re-use (Berghage et al. 2007), and concentrations of
DOC over 8 mg/l can produce disinfection by-products which then require
post-disinfection treatment under US legislation (Mendez et al., 2011).
There are a number of factors which can influence the green roof runoff
quality such as the volume of rainfall, local pollution sources, plant selection,
and substrate composition (Rowe 2011). NO3 concentrations were found to
be higher in runoff from a sedum roof in comparison to herbaceous
perennials, and in runoff from shallower substrates (Monterusso et al. 2004).
The age of the green roof can have an effect on whether the roof behaves as
a source or a sink of contaminants. Kohler et al. (2002) found retention of
PO4 increased from 26% in the first year to 80% in the fourth year of
monitoring of an extensive green roof. Similarly Berndtsson et al. (2006)
stated that PO4 release was not a problem on mature roofs. The age of the
roof can affect the hydrological conductivity (Getter et al. 2007) and also the
179
contaminant retention due to uptake by the well established vegetation.
Conversely one might also expect saturation of contaminants, for instance
shallow soils can quickly become sites of significant N leaching as a result of
high throughfall inputs and limited retention capacity (Dise and Wright 1995).
Green roofs are a relatively new technology in the UK, thus studies on aged
green roofs are scarce. A benefit of such a study would be to reveal how
older green roofs influence runoff water quality and consequently to see if
both contemporary and historic loadings of metals in an urban environment
significantly impact on this quality. Although many green roofs are now being
established in relatively clean environments, this is not the case everywhere.
Installation of green roofs in polluted areas as part of an air quality
management strategy needs to be guided by an awareness of the issue of
legacy pollutants so that suitable recommendations can be issued. There is
evidence of legacy inputs impacting on water quality in other environments,
such as N input to forest ecosystems from chronic atmospheric N deposition
exceeding assimilation capacity and resulting in enhanced export of
dissolved inorganic N in runoff (Dise et al. 1998). Soils contaminated by
historical metal mining or past atmospheric metal deposition can also
influence surface water quality (Rothwell et al. 2008; Mayes et al. 2010).
Extensive roofs, by definition, have substrates less than 150 mm and there
are engineering limits to how deep the soil layer of intensive roofs can be due
to the load bearing capacity of buildings. These shallow substrates,
positioned within urban environments, have the potential to become
saturated with dry and wet deposited pollutants over time. Fine fractions of
Road Deposited Sediment (RDS) can also be carried on winds ultimately
settling on rooftops, and RDS can frequently be highly contaminated with
heavy metals such as Pb (Taylor and Robertson 2009). The main sources of
Pb in urban centres historically is from leaded petrol use in vehicles and
industries handling materials that bear Pb (Del Rio-Salas et al. 2012). While
the use of leaded petrol in the UK has been phased out since 1985, Pb can
persist in soils due to its long residence time (Tijhuis et al. 2002). Studies in
Manchester have found high Pb levels in RDS several years after the phase-
out of leaded petrol, especially in the finer fractions (Robertson and Taylor
180
2007) thus urban soils in Manchester, and possibly old inner city green roofs,
are likely to contain elevated Pb, as well as other metals associated with
urban anthropogenic inputs, such as Cu, Mn and Zn (Robertson et al. 2003).
There is an ongoing need for quantitative assessments of the impacts of

green roofs on water quality, especially if their potential as SUDS is to be
promoted. The objective of the current study is to compare the quality of
runoff from an aged intensive green roof with that from an adjacent bare roof.
The age of the roof will be an interesting factor to consider when investigating
the impacts on runoff quality, so that predictions for future performance of
green roofs can be better constrained.
181
8.3 Methodology
Fig. 8.1 – Location of the study roof
Manchester is a large city in north-west England with a population of 498,000

(MCC 2010). A green roof within the University of Manchester campus, on
the Precinct building, was chosen for the study (Fig. 8.1). The area is
classified as open midrise, characterised by a fairly open arrangement of
182
buildings of 3 – 9 storeys with some trees, typical of an inner city university
campus area (Stewart and Oke 2012). The roof was chosen because it has
a conventional roof area (900 m2), consisting of concrete paving slabs,
adjacent to a large (408 m2) intensive green roof, which is 43 years old
(dated to 1970 on the original blueprints), and has an average depth of 170
mm. The roof is not within rain shadows of any adjacent taller buildings. Fig.
8.2 shows cross sectional representations of the two study roofs. The green
roof is of fairly standard construction with the vegetation and substrate layers
divided from the ‘egg box’ design plastic drainage layer by a fibrous
membrane.
Fig. 8.2 – Cross section of the green roof (left) and bare roof (right) showing layer
depths
183
The roof itself is protected by a tough geotextile membrane. The bare roof is
a conventional roof surface consisting of concrete paving. The 60 x 60 cm
paving slabs sit on top of an insulating polystyrene cushion and a plastic
foam membrane that are impermeable to water.
The green roof is of particular interest due to its age, the fact that the roof
was not constructed specifically for the study, and due to it having a mineral
soil substrate rather than the more usual, prefabricated, light weight
aggregate (LWA) based substrate. Green roof studies with an experimental
component often use specially constructed extensive green roof test rigs.
While offering a certain degree of control over variables of interest, artificial
experimentation studies can overestimate the benefits. For instance, test
rigs are often 100% green coverage, whereas in reality green roofs often
have quite high proportions of conventional roof surface, at the periphery, or
to provide access for maintenance. Investigating a well established, real
green roof which is subject to local weather patterns, will reveal green roof
characteristics that can be sensibly applied to the real world.
Building managers were consulted and it was ascertained that no additions of

artificial fertilisers or fresh soil had been carried out since 1994, when it was
last landscaped. The vegetation on the green roof is dominated by the
grasses Agrostis stolonifera and Festuca rubra, and the invasive weed
Plantago lanceolata. In summer the vegetation can reach considerable
heights, especially the flowering shrub Senecio jacobaea which covers about
20% of the green roof and grows to a height of 60 cm. There are a few
individuals of larger plants – Buddleja davidii and Rubus fruticosus in the
southern part of the roof.
8.3.2 Sample collection and analysis
Ten rainfall events were sampled between May and October 2012, seven of
which were classed as full event sampling and three events (D1- D3) were
sampled as spot monitoring during a wet period in July 2012. See Table 8.1
for descriptions of the rain events. The number of rain events sampled is
similar to previous work (Berndtsson et al. 2006; Mendez et al. 2011). Rain
events are defined as being separated by a dry period of at least six hours in
184
accordance with previous green roof hydrological research (Stovin et al.
2012). Rainfall depth and duration were recorded, as well as the Antecedent
Dry Weather Period (ADWP). Square plastic pots were constructed to sit in
the top of two drainage downpipes that have a diameter of 150 mm – one
draining the green roof section and one draining the adjacent conventional
roof section (Fig. 8.3).
185
Event Date Rainfall Duration ADWP No. of Rain Volume runoff (l) Runoff retention (%)
(mm) (hh:mm) (hh:mm) samples collected
from Green roof Bare roof Green roof Bare roof
each roof
A 07/05/2012 0.32 01:30 68:10 5 1530 1665 54 15
B 07/06/2012 5.3 11:00 08:00 3 775 790 62 35
C 15/06/2012 12.76 10:20 06:10 13 Yes x 2 1875 1895 62 35
D1 06/07/2012 22.3 11:30 08:40 1
D2 10/07/2012 5.44 17:30 06:10 2 Yes
D3 11/07/2012 1.23 04:10 09:00 1
E 15/08/2012 13.31 05:20 64:50 14 Yes x 2 1635 2130 68 30
F 21/08/2012 1.34 03:40 30:00 4 150 210 70 30
G 10/09/2012 17.45 10:40 05:10 12 2665 2670 69 33
H 11/10/2012 21 12:50 15:10 7 Yes 3305 3300 74 31
Table 8.1 – Characteristics of the rainfall-runoff events sampled
186
Runoff entering the drains was channelled into the pots using specially
constructed plastic collars fixed into the drain openings with waterproof
sealant, and water could then leave the pots via a 60° v-notch cut into one
side. 100 ml of runoff were manually collected from the input flows to the
pots in plastic sample bottles. Samples were collected as near to the start of
runoff commencement as possible, and then at ten minute intervals, until the
runoff started to diminish once peak runoff flow was reached. Sampling
frequencies were then reduced to every 20 minutes or every hour, depending
on flow rate. This meant that some longer rain events generated a larger
number of samples than other events. Approximately 20% of the area of the
both roof catchments is taken up by a glass atrium roof (Fig. 8.3). For the
green roof, this means that a proportion of the runoff will be attributable to a
non-green surface.
Fig. 8.3 – Plan of the Precinct roof showing drainage catchment areas for the bare
and green roofs
Bulk rain samples were collected at the same time as four of the events
(Table 8.1) using a funnel of diameter 310 mm which had been thoroughly
rinsed with de-ionised water. Rainfall data were obtained from the Whitworth
Observatory (UoM 2012), situated 150 m from the precinct roof (Fig. 8.1),
187
which employs a Theis laser distrometer, with an accuracy of > 90% and
resolution of 0.001 mm/hr.
Water samples were promptly taken to the laboratory and filtered using 0.45
µm syringe filters (VWR cellulose acetate). Subsamples of 20ml were
acidified with high purity nitric acid, to stabilise the samples ready for analysis
for metal ions using ICP-MS (Agilent 7500cx). All samples were then stored
in a refrigerator in the dark ready for analysis which in all cases was carried
out as soon as possible. Benchtop meters were used to measure the pH
(Hanna HI9124) and conductivity (Hanna HI9033). Anions were measured
using ion chromatography in accordance with US EPA method 300 (Metrohm
882 Compact IC plus) and the analytes consisted of Cl, NO3, PO4 and SO4.
Visible light absorbance over the wavelength range 190 – 700 nm was
recorded using a spectrophotometer (Hach Lange DR5000). Absorbance at
400 nm can be used as a proxy for Dissolved Organic Carbon (DOC) in
accordance with other work (Worrall et al. 2002; Wallage et al. 2006). The
ratio of the absorbance at 465 nm and 665 nm , also known as the E 4/E6 ratio,
can give an indication of the nature of the DOC, with ratios of 8 – 17
indicating fulvic acids and 2 – 5 characteristic of humic material (Thurman
1985; Wallage et al. 2006).
A representative number of samples of the substrate layer were collected

and analysed in a laboratory to determine the following physical parameters:
soil texture (particle size analysis), field capacity (Buchner funnel method);
dry bulk density (cylinder sampling method); porosity (from bulk density and
particle density); and organic matter content (loss on ignition). A Guelph
permeameter was employed on the green roof to determine the saturated
hydrologic conductivity, a measure of permeability. Twelve extra soil
samples were taken from the surface of the green roof, with the sampling
locations maximally spaced over the green roof. These samples were
analysed in a laboratory to determine pH, conductivity, and water soluble Cl,
NO3, NO2, PO4 and SO4. On the bare roof, it was observed that substantial
accumulations of dust were located underneath the paving slabs. Four
samples of this dust were collected. The soil and dust samples were dried at
40°C and sieved to 125 µm to remove stones/twigs/leaves and homogenised
188
in a ball grinder (Fritsch Spartan pulverisette 0) and then analysed for metals
using X-ray fluorescence (XRF) in accordance with US EPA method 6200.
Each sample was analysed three times with a Niton XL3t XRF analyser.
Three 40 mm diameter cores of the full substrate depth were taken from
central areas of the green roof sections (See Fig. 8.3 for locations). These
were split into ten equal fractions, prepared as described above, and
analysed separately using XRF to provide a depth profile of soil elements.
Soil geochemistry data for Manchester city centre were made available by
the British Geological Survey (BGS) Geochemical Baseline Survey of the
Environment (G-BASE) project which aims to provide a national capability for
geochemical mapping (Johnson et al. 2005). Composites are made from five
surface samples (0.05 to 0.20 m) within a 20 m X 20 m square, and analysed
using XRF (Fordyce et al. 2005). 55 data points were used to produce
geochemical maps using ordinary kriging techniques in Arc-Map utilising a
box-cox transformation to normalise data. The Environmental Quality
Standards (EQS) for surface water quality under directive 2000/60/EC of the
Water Framework Directive were obtained from the UK Environment Agency
(EA 2011).
189
8.3.3 Data analysis
Runoff quantity and retention data were available for three of the fully
sampled events, C, E and H (Table 8.1), and this was used to estimate fluxes
of the analytes by calculating the product of the average event concentration
and the total runoff volume for both roofs. The runoff quantity data were also
used to investigate the presence of a first flush in the events by using
mass/volume relationships (Kaczala et al. 2011). Only event E was found to
have a significant first flush and these higher concentrations were excluded
when calculating averages and undertaking analysis. This event was the
only one to exhibit first flush, possibly for two reasons – event E had one of
the largest ADWP of all the events (Table 8.1) so there would have been a
relatively larger accumulation of dry deposited material on this roof, and the
nature of the manual sampling technique meant that the first few minutes of
runoff in other events may have been missed while preparing equipment at
the onset of rain.
Data were checked for normality using the Anderson Darling test and found
to be not normal. Non-parametric statistical techniques were thus employed.
All statistical investigations were carried out using R (version 2.11.1).
190
8.4 Results
The weather during the study period was unusually wet. 52% of the days
had rain and July 2012 had over double the average amount of rain,
calculated as the climate average for 1981 – 2010 (Fig. 8.4). The maximum
ADWP for the studied rainfall events was just under three days.
250
Monthly rainfall as percentage
200
of climate average (%)
150
100
May-12 Jun-12 Jul-12 Aug-12 Sep-12 Oct-12
50
0
Date
Fig. 8.4 – Monthly rainfall expressed as a percentage of the climate average for
Manchester (1981 – 2010)
191
Permeability pH Field Bulk Particle Porosity Organic Cl NO2 NO3 PO4 SO4
(cm/sec) Capacity Density Density content (mg/kg)
(g/cm3) (θ) (mg/kg) (mg/kg) (mg/kg) (mg/kg)
3
(max (g/cm ) (g/l)
water
capacity
for FLL)
Study 1.68 x 10-3 6.5 38% 1.03 2.47 0.58 202.54 77.9 2.49 50.84 3.35 8.51
roof
FLL 5 x 10-4 6– 45 – - - - < 90 - - - - -

guideline 8.5 65%
value
Table 8.2 - Selected soil characteristics and comparison to the FLL guideline values for an intensive green roof
192
Fig. 8.5 – Vertical soil profiles for Cu, Pb and Cr in the intensive green roof
193
Concentration (mg/kg)
As Cr Cu Fe Mn Pb Zn
Green roof Median 23.3 67.2 17 21,188 245 105 126

substrate
Mean 33.7 76.5 45.7 25,575 313 164 160
(n=36)
Range 13.4 – 90.2 49 - 118 0 - 195 11,070 – 53,077 164 - 638 59 - 463 66.5 - 402
Bare roof Median 46.2 142.9 171 38,058 838 705 1,942
dust
Mean 47.2 144 182 38,252 834 691 2,401
(n=12)
Range 43 – 53.4 126 - 163 124 - 262 34,928 – 41,961 719 - 942 567 - 789 1,347 –
4,375
UK ambient Mean 13.9 33.4 19.8 - 615 62 95

background
95 percentile 27.1 59.5 43.3 - 1,567 158 224
CLEA SGV Residential 32 200 - - - 450 -
Commercial 640 5000 - - - 750 -
Table 8.3 – Elemental composition of the green roof substrate and bare roof dusts for selected elements, including the ambient background
levels for England (Barraclough, 2007) and CLEA Soil Guideline Values (SGVs), where available, for comparison
194
8.4.1 Substrate properties
Table 8.2 shows the soil characteristics and, where possible, a comparison to
the Forschungsgesellschaft Landschaftsentwicklung Landschaftsbau (FLL)
guidelines for an intensive green roof (FLL 2008). For a full summary of the
physical soil properties see Speak et al. (2013). The soil is a sandy loam,
with quite high organic matter content. Soil chloride levels were comparable
to those from a low impact roadside plot (Goodrich et al. 2009). The water-
extractable soil nutrient levels are average (Sumner 2000). Fig. 8.5 shows
the depth profiles for three of the more important elements and no depth-
concentration relationship is apparent within the substrate layer.
The concentrations of selected elements are presented in Table 8.3, along

with the UK ambient background concentrations for comparison (Barraclough
2007). The mean and median elemental concentrations are consistently
higher in the bare roof dust than in the green roof substrate, however, the
range of values is wider on the green roof, for all elements with the exception
of Zn. The mean concentrations for both green roof soil and bare roof dusts
are higher than the mean background concentration, and also higher than the
95th percentile background value in the case of As, Cr, Cu and Pb. The only
exception is Mn in the green roof substrate which has a lower mean than the
background. UK Contaminated Land Exposure Assessment (CLEA) Soil
Guideline Values (SGVs) were not exceeded for Cr. Mean Pb values in the
bare roof dust exceeded the SGVs for residential use and some upper values
exceeded the SGVs for commercial use. Levels of As in the dust exceed
SGVs for residential soil use but fall below the commercial use guidelines.
The green roof substrate only exceeded the residential SGVs for Pb in the
upper range of results. The G-BASE data revealed a location of very high
metals within the University campus, close to the green roof. Spatial
distributions of Cd, Cu, Pb and Zn are shown in Fig. 8.6. Average Standard
Error (ASE) values for Cd, Cu, Pb and Zn were 0.6, 121.5, 246.5 and 290.5
mg/kg respectively. For all the metals there are elevated soil concentrations
to the south and east of the studied area, with a very obvious hot spot within
the University campus. The precinct roof is located very close to this area of
high metals. Exploratory data analysis was undertaken by removing the hot
195
spot data point with extreme values, and the trend of higher metals around
the University campus and to the South West of the city centre persisted.
Fig. 8.6 – Spatial distribution of Pb, Cd, Cu and Zn using ordinary kriging with 55
data points (source: BGS G-BASE project)
8.4.2 Water chemistry
Firstly pH and conductivity were significantly lower in the rainwater than in

either of the roof runoffs (pH Kruskal Wallis X2 = 13.4, p < 001, conductivity
X2 = 15.9, p < 0.001). Green roof runoff displayed higher Fe and Pb
concentrations than the bare roof runoff and rainwater, as can be seen in Fig.
8.7. The differences were highly significant for both Fe (X2 = 76.1, p < 0.001)
and Pb (X2 = 31.2, p < 0.001). Rainwater was significantly higher in Cu (X2 =
36.6, p < 0.001) and Zn (X2 = 71.4, p < 0.001) than in the runoff from either
roof, and also higher for Cd and Ni, however the differences for the latter two
were minimal. There were also no significant differences in the levels of Cr
and Mn between the roof runoffs and rainwater.
196
The fluxes of metals are presented in Table 8.4. A percentage flux, relative
to rainfall, of greater than 100% indicates the roof is acting as a source of the
contaminant. Both roofs therefore appear to act as sinks of the metals, with
the exception of Pb, which displayed very high enrichment factors in the
runoff of both roofs over the rainfall. Event E also showed the green roof to
be a potential source of Cr.
There were no significant differences in Cl, PO4 or SO4 between the roofs or
between the roofs and rainwater. NO3 levels were higher in the bare roof
runoff than in the green roof runoff and rainwater and this difference was
significant (X2 = 25.5, p < 0.001). A post hoc test (Mann-Whitney with
Bonferroni correction) shows the major difference exists between the two roof
runoffs, as opposed to rain and runoffs. The NO3 levels are below 50 mg/l,
which is the upper limit for surface freshwaters with regards to eutrophication
threat (EA 2012). Some of the higher recorded values for PO4 exceed the
environmental standards of waters classified as poor i.e. 0.5 mg/l (Defra
2010) but just over half of the data (54% for bare roof runoff and 53% for
green roof runoff), and the median, were below the detection limit. The
fluxes in Table 8.4 reveal the bare roof to be more frequently a source of the
anions in comparison to rainwater inputs, with SO4 being higher than in
rainfall for all three events. PO4 is always lower than the rainfall input on
both roofs and NO3 is lower on the green roof.
197
Fig. 8.7 – Boxplots to show water quality parameters for Bare roof runoff (B), Green
roof runoff (G) and rainfall (R). Red lines indicate the EQS values for protection of
surface water quality
198
The absorbance at 400 nm is significantly higher on the green roof runoff (X2
= 69.7, p < 0.001), being approximately double that from the bare roof runoff.
A post hoc test shows the major difference is between the two roof runoffs,
as opposed to between rain and runoffs. The green roof runoff was observed
to have a weak green/yellow colour, and the median value of 3 abs/m is
double the European Community standard of 1.5 abs/m acceptable for
drinking water (Mitchell 1990). The E4/E6 ratios were consistently low with an
average of 3.3 (SD = 0.89) and one high value of 18.5 during the first flush at
the start of event E.
199
Roof Event Cr Mn Fe Ni Cu Zn Cd Pb Cl NO3 PO4 SO4
Bare C 35.5 12.4 14.7 9.5 7.0 6.1 0.0 140.3 82.5 137.9 62.5 101.8
E 88.8 19.2 22.1 13.5 11.8 70.8 3.1 512.3 78.1 142.4 54.0 121.2
H 40.8 24.0 6.2 5.2 16.6 19.8 0.0 219.3 202.6 87.9 0.0 105.2
Green C 23.6 19.9 53.9 8.4 6.4 2.0 0.0 233.9 46.7 34.0 36.0 127.0
E 134.9 27.5 51.9 21.3 11.1 16.6 3.6 734.4 128.7 64.4 24.1 97.5
H 16.5 16.7 12.5 3.3 8.6 1.1 0.0 69.4 64.2 23.3 0.0 35.8
Table 8.4 – Output flux expressed as a percentage of input flux for the three rainfall events where flux calculation was possible. Bold results
indicate >100 %
200
8.5 Discussion
8.5.1 General runoff characteristics
The roof runoff data displayed comparable characteristics to other research

in many respects. The roofs had the effect of slightly raising the pH of the
rainwater, which is likely a result of the effect of dissolved solutes picked up
on the roof. This buffering effect on slightly acidic rain is common (Teemusk
and Mander 2011). Berghage et al. (2007) recommend artificial liming of
green roofs after 10 – 30 years because the pH buffering effect is predicted
to diminish, however, this was not apparent on the 43 year old roof in the
present study. The conductivity is raised by roughly 50 µS on each roof
relative to the rainfall conductivity as a result of the increase in dissolved ion
content (Fig. 8.7).
A source of PO4 and SO4 may be present on both roofs giving rise to the
higher concentrations found (Fig. 8.7) and the high fluxes of SO4 for some
events (Table 8.4). The SO4 values are much lower than in previous studies,
however (Teemusk and Mander 2007; Vijayaraghavan et al. 2012). PO4
values are also lower than some published values (Beck et al. 2011;
Vijayaraghavan et al. 2012) but may potentially contribute to eutrophication.
However, this threat may not be regarded as severe due to the fact that just
over half the data were below the detection limit. There are no sources from
artificial fertilisers on the green roof and older roofs tend to generate less PO 4
runoff consistent with other studies.
Lower NO3 in the green roof runoff compared to the bare roof could be due to
being scavenged by plants. One might expect N saturation on a limited
depth soil after 43 years of receiving atmospheric inputs (Dise and Wright
1995) but this does not appear to be the case even though annual N
deposition in Manchester in 2011 is moderate at 12 – 14 kg N/ha/yr (Defra
2013). Average NO3 levels in rainfall, at 1.9 mg/l, were higher than the UK
average of 0.35 mg/l for the period 1986 - 2000 (Hayman et al. 2001) and
higher than the 1.12 found in mid Wales (Neal et al. 2007). This could be an
indication of high local sources from fossil fuel combustion (Buss et al. 2005)
201
in vehicles in the city centre plus or enhanced deposition in the above
average rainfall for the study period.
The higher Cl concentrations in the roof runoffs than in rainfall, and the
indication of the roofs as sources from the flux data, could likely be coming
from airborne spread of road salting (Lundmark and Olofsson 2007) as
Manchester is not a coastal city so rainwater inputs will be small . The
median rainwater Cl concentration of 1.3 mg/l is at the lower end of the range
of values of 0.4 – 3000 mg/l for rainfall in England (White and Broadley 2001).
The green roof is also a further distance from Oxford Road, and the study
took place mostly over the summer when road salting was not occurring. Soil
water soluble Cl values for roadsides in Poland varied from 40 – 150 mg/kg
(Czerniawska-Kusza et al. 2004) so the present study values are comparable.
The absorbance data suggest there is higher DOC in the green roof runoff,
most likely from the soil organic matter. The E4/E6 ratios reveal this to be
mostly humic in nature (Wallage et al. 2006) with a brief pulse of younger
fulvic acids in the first flush of event E which could be coming from fresh leaf
litter on the green roof surface or collected in the drainage channel. The
water from this roof is not being re-used so colour can be considered to be
not an issue.
8.5.2 Elevated metal concentrations
It would appear that dissolved metal concentrations may be an issue for

green roof runoff quality for this aged green roof. Median Cu, Pb and Zn all
exceed the EQS values for protection of surface water (1, 7.2 and 8 µg/l
respectively) and thus may contribute to UDP. Fluxes, however indicate both
roofs to be sinks, with the major exception of Pb. This could be due to the
unusually large rainwater concentrations found in this study, especially for Cd,
Cu and Zn. Cu and Zn levels were comparable to those found in rainfall in a
large city in China (Zuo et al. 2012). This may be a result of the small
rainwater sample size or high local sources of metals in rain. Rainwater
concentrations at the Holme Moss station of the heavy metal monitoring
network (Fowler et al. 2006), which is closest to Manchester, recorded Cd, Cr,
Cu, Pb and Zn levels of 0.04, 0.23, 1.57, 2.27 and 12.2 µg/l respectively in
202
2005, which further highlights the extreme nature of the rainfall
concentrations recorded for Cd, Cu and Zn in this study. In comparison with
other studies, green roof runoff values for Cu, Cr, Fe and Zn are fairly similar
to those reported by Berndtsson (2006) on a one year old extensive green
roof, and the Zn values were lower than the 300 µg/l found by Mendez et al.
(2011).
The Pb values deserve further discussion, as the high values in the runoff of
both roofs and low values in the rainfall suggest a source of Pb on the roofs.
The highest concentrations were found on the green roof with a median value
of 15 µg/l, which is higher than the 10 µg/l (Berndtsson et al., 2006) and 4
µg/l (Mendez et al., 2011) found in previous studies. The indication of a
potential source of Pb in the green roof soil and bare roof dusts is certainly
corroborated by the high Pb values found in the solid phase within the
sediments/soils (Table 8.3). A mean of 164 mg/kg on the green roof and 691
mg/kg on the bare roof compare well to inner city RDS samples taken from
Manchester city centre, which varied seasonally but ranged from 71 – 660
mg/kg (Robertson and Taylor 2007). There is some lead flashing on the
atrium roof over which the rainfall can flow, and roof surfaces have been
sources of Pb in previous studies on conventional roof surfaces (Simmons et
al. 2001; Rocher et al. 2004). Flashing sometimes gives rise to elevated Pb
(Abbasi and Abbasi 2011). Rainwater that is acidic or contains high
concentrations of Cl can be plumbosolvent, but rainwater can also reduce
corrosion by forming a protective, insoluble coating of PbCO3 (Evans 1960).
Given such high Pb values in the sediments it is likely that the majority of the
Pb is being leached from the sediments. Additionally, the atrium contribution
to catchment runoff is only 20% on both roofs.
203
In order to separate the influence of the atrium runoff on Pb concentrations,
atrium runoff was diverted and sampled separately for a heavy rain event.
The results in Fig. 8.8 indicate that both the atrium and the green roof are
raising the Pb concentrations over that of rainfall. The median and range are
larger for the atrium, which would suggest the flashing may indeed be a
contributing factor, however the values for the whole study are higher for
green roof runoff.
Fig. 8.8 – Boxplot of Pb concentrations in the runoff from the atrium (A), the green
roof (G) and in rainfall (R) for one event
204
To further investigate the dynamics of Pb on the green roof, a simple
input/output model was proposed (Fig. 8.9). Inputs were calculated from
atmospheric deposition data (MacKenzie et al. 1998; Fowler et al. 2006; Cloy
et al. 2008) and extrapolations from trends in emissions data (MAPAC 1993;
NAEI 2013). Present day mean soil concentration and soil bulk density were
used to estimate the total amount of Pb stored within the green roof as 72 g.
Present day Pb output was calculated from the product of runoff volume and
present day mean Pb concentration, and total Pb output (Fig. 8.9a) is simply
total Pb deposited minus Pb stored in the soil. Assumptions made are that
the soil contained no Pb when the green roof was installed, rainfall retention
efficiency is 66% (Speak et al. 2013) and runoff concentration has remained
constant over the life of the green roof. The present day situation shows
there is more Pb coming out of the roof than is being deposited which implies
the green roof soil is acting as a source of Pb. An output of 38 g over 43
years suggests a rate of 0.88 g/yr which is not too dissimilar to the present
day empirically estimated rate of 1.34 g/yr, indicating the model has
applicability.
Fig. 8.9 – A model of the atmospheric inputs, stores and outputs of Pb for (a) the
entire life of the green roof and (b) the annual flux based on present-day data
205
Keeping present day atmospheric inputs and outputs steady means the
green roof would continue to be a source of Pb for 70 years. The model also
includes the contribution of the atrium, with its potential lead flashing source,
and it is clear that while concentrations are comparable to green roof runoff
concentrations, overall mass flux is much lower. The original source of the
Pb in the substrates is debatable but it is likely a remnant of Pb pollution from
vehicle use prior to the phasing out of leaded petrol in 1985, as contemporary
atmospheric deposition is minimal. This is evidenced by the low rainwater
concentrations and total Pb deposition in 2008 in Manchester was 10 – 15
g/ha/yr/ (Defra 2013) (cf. 50 – 75 g/ha/yr in 1995 – 1998 (Fowler et al. 2006)).
The bare roof dusts represent a primary form of the contaminated RDS,
carried to the rooftop by winds, whereas it is somewhat diluted within the
green roof soil. While there is a much smaller quantity of sediments under
the paving of the bare roof, the high Pb content is contributing significantly to
the runoff concentrations, with just under half the values exceeding the EQS
(Fig. 8.7). The green roof substrate is 43 years old and the original source of
the soil is also unknown so it is possible that Pb may have been high in the
soil when it was installed. Depth profiles for the metals were uniform. This
could be a result of high bioturbation on the roof or it could be the soils were
high in metals to start with. One other thing to consider is the hotspot of
contaminants located very close to the precinct roof (Fig. 8.6). Soil Pb levels
in this area exceed the SGVs for commercial use. Soils at this level of
contamination pose a risk to human health and may require treatment. There
is the possibility that some of the contamination on the precinct roof may
have arisen from aeolian dust transport of this nearby contaminated topsoil.
Resuspended soil has been previously linked to the long distance transport
of Pb (Young et al. 2002). Examination of past industries in the area has not
revealed any obvious source for this high contamination.
The soil and dust levels on the precinct roof were quite high for most of the
metals, exceeding the 95th percentile values of the UK ambient background
survey. The narrower range of contaminants in the bare roof dusts backs up
the notion that the green roof soil, with its wider ranges of concentrations, is a
diluted form of the RDS.
206
8.5.3 Implication of findings
The discovery of legacy metal contamination of urban soils is not a new

phenomenon, however this is possibly the first time it has been found in the
soils and runoff from a green roof. Source apportionment (Del Rio-Salas et
al. 2012) of the metals found within the green roof substrate and in the bare
roof dusts would enable more details about the processes behind the
contamination to be revealed. Pb isotopic analysis (Chiaradia and Cupelin
2000), for instance, could quantify the contribution of anthropogenic
processes to the Pb enrichment observed in the area and green roof
sediments. Nonetheless the 43 year old green roof is old enough to have
been exposed to direct atmospheric inputs as well as post leaded petrol ban
inputs from re-suspended RDS. The fact that a green roof can thus
potentially contain a store of Pb pollution must be taken into consideration
when constructing contemporary green roofs near to pollution sources.
Green roofs are rightly being heralded as remediators of air pollution but if
that pollution is accumulating to levels which impact the future runoff quality
then problems may ensue. The promotion of green roof runoff for re-use, for
instance as a source of potable water, especially needs to be carefully
considered.
This study investigated an old intensive green roof with a mineral soil of
unknown origin. Modern green roof technologies are currently based on
utilising various lightweight substrates, such as expanded clay, and
consideration is being given to the impact of these upon runoff
(Vijayaraghavan et al. 2012). Expanded shale has been found to be good for
retention of P, NH4-N and metals (Long et al. 2006). There are also
amendments to substrates, such as adding biochar, which can significantly
reduce nutrient leaching whilst simultaneously allowing the addition of
artificial fertilisers (Beck et al. 2011). This study has shown that the long
term performance as regards runoff quality, of these modern substrates,
should be at least estimated, or ideally longitudinal monitoring studies
commenced. The pollution reduction properties of green roofs in urban areas
can then be balanced with improvements in the runoff quality from the
substrates.
207
8.6 Conclusion
Runoff from a green roof and an adjacent conventional roof surface was
characterised with regards to heavy metals and nutrients, and linked to soil
chemical properties. The study roof is of interest because it is an aged
intensive green roof. The following findings were made:
 Nutrient levels in the runoff were largely not problematic for runoff
quality, with the exception of a number of PO4 data which could
contribute to eutrophication in receiving water bodies. NO3 levels
were low so the roof did not appear to be saturated with regards to
inorganic N, contrary to expectation.
 High concentrations, exceeding the EQS limits for protection of
freshwater, of Cu, Pb and Zn were apparent in the green roof runoff.
Overall fluxes for three rain events revealed Pb quantities to be higher
in runoff from both the green roof and the bare roof than in rainfall,
which suggests the existence of a Pb source within their respective
sediments. Lead flashing within the rainfall catchments of both roofs
may also contribute.
 Metal and metalloid concentrations, specifically As, Cr, Cu, Fe, Pb and
Zn, within these dusts exceeded the 95th percentile values of UK
ambient background, thus indicating contamination on both roofs.
 Possible sources of contamination include wind blown Road
Deposited Sediment (RDS) which has settled onto the roofs and
become diluted within the green roof substrate, and a local area of
highly polluted topsoil which may also contribute to wind blown
contaminated sediment load on the roofs.
 The age of the green roof, and the results of a budget analysis, mean
historic atmospheric deposition of Pb could be contributing to the Pb
signal seen, thus the green roof substrate may be acting as a store of
legacy pollution.
The study highlights the need for consideration of local atmospheric pollution
inputs when installing a green roof and the consequent deterioration of runoff
quality which may result as the roof ages.
208
CHAPTER 9 CONCLUSION
9.1 Introduction
The purpose of this final chapter is to summarise and bring together the key
findings of the thesis and discuss their significance within the fields of green
roof research and climate change adaptation. There then follows a brief
discussion of the strengths and limitations of the empirical work and
suggestions for future research.
9.2 Summary of findings
The research was divided into four parts, each related to a different green
roof benefit, and each with its own aim and objectives. The research can
thus be summarised by reflecting upon how the four aims were met:
 To assess what quantities of air pollution, specifically PM10, can be

removed from the urban atmosphere by green roof vegetation.
A review of the literature revealed a lack of published studies which

quantitatively investigated the pollution capture capabilities of green roofs.
Magnetic analysis techniques and counting of particles on SEM images were
employed, for the first time on green roofs, to quantify the PM10 capture
efficiency of four green roof plant species (Chapter 5). Species differences in
particle capture efficiency were apparent, with two grass species
outperforming S. album and P. lanceolata. Site differences were also
apparent, with proximity to vehicular source on Roof 1, next to the Mancunian
Way, being an important factor. Removal rates per unit area of green roof
were calculated and used to estimate the quantities of PM10 that could be
removed from Manchester city centre under a maximum sedum green roof
scenario. This amounted to 0.21 tonnes per year, which is 2.3% of the
annual PM10 inputs to the area.
209
 To see how green roofs can impact the local thermal microclimate
relative to a conventional roof surface.
Air temperature at 300 mm above the surface was recorded for a period of 17
months over an intensive green roof and an adjacent paved roof on the
Precinct roof (Chapter 6). The monthly median air temperature over the
green roof was up to 1.06oC cooler than the conventional roof. This cooling
effect was strongest at night with an average difference of 1.58oC and a
maximum difference of 4.2oC seen. The cooling effect was also observed in
periods of peak temperature. Conversely, a relative daytime warming was
observed over the green roof sections with the peak difference occurring in
the late morning. The vegetation on one section of the green roof was
damaged by a combination of drought and mismanagement. This resulted in
an impairment of the green roof cooling effect with a lower maximum average
relative cooling of 0.78 oC observed, and daytime temperatures were higher
than over the bare roof for a larger proportion of the day than over the
undamaged green roof section. The green roof in this study was observed to
have a complex diurnal pattern to its thermal microclimate impacts which
traditional theories (DiGiovanni et al. 2013), based on evapotranspirative
cooling, were not able to fully explain. Differences in the thermal capacity of
the roof materials and vertical mixing characteristics could be influencing the
patterns seen.
 To quantify the rainwater retention efficiency of an intensive green roof.
Rainwater retention efficiency was calculated from rainfall and runoff data for
69 rain events over a 14 month period (Chapter 7). Runoff was quantified for
an intensive green roof and an adjacent paved roof, on the Precinct roof, via
a calibrated pressure transducer positioned within the drains in modified
receptacles based on a v-notch weir methodology. Average runoff retention
of 65.7% on the green roof compares with 33.6% on the conventional roof.
High organic matter content in the green roof substrate may contribute to the
retention capacity. Explanatory variables such as peak rainfall intensity and
rainfall duration had no significant impacts on retention, however there was a
significant reduction in retention capacity for high rainfall events (>10 mm)
210
and in autumn, which had above average rainfall. Construction of intensive
green roofs on 10% of Manchester city centre’s rooftops would potentially
increase annual rainfall retention for the area by 2.3%, however flood risk
should be considered separately in terms of large storm events.
 To assess the impacts of an intensive green roof on the quality of

rainwater runoff.
Runoff from an aged intensive green roof and an adjacent conventional roof,
on the Precinct roof, was collected and analysed for heavy metals and
nutrients (Chapter 8) with mixed results for water quality impacts. Due to the
age of the roof, the substrate was expected to be saturated with respect to
NO3, however runoff concentrations were low. A number of PO4 data
indicated concentrations that may contribute to eutrophication issues in
receiving water bodies. Cu, Pb and Zn concentrations exceeded EQS values
for protection of freshwater. Fluxes of Pb were higher in the runoff of both
roof surfaces than in rainfall, suggesting a source of Pb within their
respective sediments. Lead flashing on a part of the catchments of both
roofs could also contribute. An analysis of the sediments revealed
contamination by As, Cr, Cu, Fe, Pb and Zn which may have originated in
wind blown road deposited sediments. Additionally, a local area of highly
polluted topsoil could also contribute to the pollutant load. A budget analysis
was carried out with respect to Pb using historic atmospheric deposition data
which indicated the green roof soil is currently acting as a source of Pb. The
age of the roof means Pb deposition from before the phasing out of leaded
petrol could be responsible for the elevated concentrations. Therefore, in the
case of this particular green roof, the impacts on runoff quality may not be
considered a benefit. This has implications for the use of green roofs as a
remediation technique for air pollution as pollutants sequestered in the
substrate may subsequently impact on runoff quality.
9.3 Significance of findings for green roof research
The current awareness of green roof benefits, and which component of the
green roof structure they are mostly associated with, is demonstrated in Fig.
9.1. This thesis has provided a number of novel discoveries within the
211
benefits listed to the left. In particular it has demonstrated particle capture
efficiency of green roof vegetation, net cooling effects on overlying air,
significant runoff retention achievable, and some runoff quality improvement
with respect to NO3. The benefits operate on different scales; from the local
building scale of the cooling effects and reduced noise for building
inhabitants and neighbourhood scale of reduced flood risk (excepting large
storm events when substrate is at field capacity), to the citywide impacts of
reduced air pollutant levels and global impacts of carbon mitigation.
Fig. 9.1 – The multiple benefits afforded by the three main compartments of a green
roof. (Author created)
The evidence for removal of significant quantities of urban airborne

particulate matter (Chapter 5) can be used in air quality models that aim to
simulate the effects of increasing the spatial coverage of green roofs. The
results will also be of interest to city planners interested in using green
infrastructure to reduce air pollution levels because they reveal the need to
think carefully about species selection and roof location in order to maximise
pollutant capture. The fact that sedum plants were outperformed by grass
species should be of great interest to the green roof industry. UK emissions
of particulate matter are projected to decline by 22% from 2005 to 2020,
212
however a slight rise is then expected to 2030 due to increases in emissions
from vehicle component wear and road abrasion (NAEI 2012). Green roofs
can be considered within remediation schemes, along with trees, for this
latter source of urban particulate matter. Air quality standards may then be
attained in the UK by targeting areas where particulate air pollution is highly
prevalent, such as in London. Fig. 9.2 shows that Manchester’s air quality
with respect to PM10 is relatively low, therefore green roofs may have a larger
impact and confer increased health benefits in places such as Poland and
Greece, where PM10 concentrations tend to be much larger.
Fig. 9.2 – 36th highest daily value of PM10 in 2010. (Adapted from European
Environment Agency (EEA 2013))
The results from the study on the thermal impacts of an intensive green roof
(Chapter 6) show how green roofs can have significant impacts on the
overlying air temperature and can thus potentially contribute to a reduction of
the urban heat island effect during peak temperature periods in summer.
This is a primary aim of climate change adaptation with greenspace. An
213
average localised cooling effect of 1.06oC on the overlying air mass, which
was observed to be as high as 4oC at times, suggests an impressive
contribution of green roofs to adaptation schemes. One would expect this
effect to be additive, so it is suggested that increasing the areal coverage of
green roofs in urban areas would potentially have a large impact. The
cooling of the air relative to a bare roof is indicative of the larger scale effects
expected from changing the thermal properties of a greater proportion of
urban surface. These effects manifest at the roof level, therefore will be
important for processes within the UCL and the roughness sublayer of urban
atmospheres. The study also quantified how drought damage and poor
maintenance can reduce the cooling effect, thus highlighting the importance
of irrigation during dry periods and correct maintenance practices, especially
with intensive roofs. Irrigation in dry periods to maintain the cooling effects
could be carried out with the aid of a cistern for collection and re-use of
excess rainwater (Hardin et al. 2012). Extensive roofs require less artificial
irrigation, if at all, but evapotranspirative cooling in the daytime is less due to
the prevalence of vegetation using CAM photosynthesis.
The green roof hydrology study (Chapter 7) provided further evidence for the
increased retention efficiency afforded by investing in the deeper substrates
of intensive green roofs. The study took place during an exceptionally wet
year and so comments can be made on how green roofs will cope with some
of the increased precipitation patterns expected under climate change
scenarios (Murphy et al. 2009; Sanderson 2010). The age of the roof did not
appear to have negative impacts on the roof retention which is useful for the
green roof industry as it will allow green roofs to come with some guarantee
that their benefits are not likely to lessen with age. However, the upscaling
work revealed a fairly large green roof installation project (10% increase in
intensive roof coverage) would reduce the annual retention for the city centre
by just 2.3%. This figure would be smaller if the scenario used extensive
green roofing, which is more likely due to the lower costs associated.
Therefore, green roofs should not be considered in isolation as storm
management techniques, but with other SUDS techniques such as porous
pavements, swales and rain gardens (see Section 2.3.2).
214
In Chapter 8, the aged green roof was observed to have low impacts on
water quality with regards to nutrients and was seen to increase the pH of
slightly acidic rain. Nitrate saturation was expected so this result is again
promising with regards to the age of the roof. The impact with respect to
heavy metals, in particular Pb, is interesting. Legacy pollution is common in
urban areas and for a green roof to act as a store of historic Pb deposition
shows how ubiquitous in all niches of the urban environment the problem can
be. The implications for the green roof industry are that green roofs may not
be suitable for remediation of air pollution that involves heavy metals with a
complicated geochemistry that results in them being retained and
subsequently released. This is an important result for countries with
significant atmospheric inputs of metals such as those that still use leaded
petrol, e.g. Afghanistan. This feature uncovers an aspect of green roofs that
needs to be included in design phases i.e. could the roof lead to a reduction
in runoff quality as a consequence of air pollutant accumulation? Runoff
quality is more often compromised in green roofs for instance when fertilisers
are used to enhance vegetation growth or novel lightweight substrates are
tested.
A commonly repeated notion within green roof research is that intensive

green roofs generally afford greater benefits than extensive ones. This is
because the deeper substrates not only retain more rainfall and sequester
more carbon etc. but they allow a greater biomass per unit area in the
vegetation layer, with associated increases in pollution capture and
evapotranspiration. Also, by allowing a greater diversity of plant species than
the xerophytic plants associated with extensive roofs, they offer more
possibilities for biodiversity and landscaping for aesthetic appeal. This thesis
provided a range of results for an intensive roof, thus increasing the evidence
base for their benefits such as cooling of the overlying air. It is hoped that a
larger evidence base for the enhanced benefits of intensive over extensive
roofs will encourage the industry to consider their implementation, despite the
attraction of extensive roofs being generally cheaper and benefitting from
technological advances in LWA substrates.
215
This thesis also demonstrated that the benefits are maintained even when
the roof has aged considerably. Age was only a problem for the intensive
roof in this study with respect to water pollution potentially caused by the roof
being constructed before the phasing out of leaded petrol, and leaded
materials used in the construction of the atrium section. Green roofs
constructed with modern roofing materials under a relatively clean
contemporary atmosphere may not display such an issue with runoff water
quality in the future.
This thesis has investigated a number of the impacts of green roofs and as
such it allows an evaluation to be made of the multi-benefit nature of green
roofs within a UK setting. One of the main selling points for green roofs is the
fact that they have a suite of benefits that raises them above simply being a
quirky, architectural feature destined to only be installed on showcase
buildings. They do in fact provide tangible, measurable services for urban
centres, which help to address some of the deleterious impacts of
urbanisation and should therefore be viewed as functional components of
vernacular architecture. The multi-benefit nature of green roofs favours their
installation over other building scale adaptations such as reflective glazing
and insulation for reducing thermal transmittance (Gupta and Gregg 2012).
For instance, cool roofs, which rely on high albedo roofing materials, have
proven benefits for reducing thermal loads in buildings with consequent
reductions in air conditioning demand (Akbari et al. 2009; Akbari and
Matthews 2012), but they do not offer a solution to storm-water management
issues, nor do they reduce air pollution or promote biodiversity. Other
benefits that are less tangible but no less valuable include increasing the
liveability of cities and adding aesthetic or monetary value to a building or
community space.
The multiple benefit aspect ought to be integral to the green roof construction
process rather than seeing them as an engineering tool to solve one problem.
The results of this thesis will therefore be of interest to the green roof
construction industry. Strong links already exist between the industry and
researchers, for example the Sheffield Green Roof Centre has the support of
ZinCo, a market leader in green roof construction (Green Roof Centre 2013).
216
Future advances in green roof technology and a higher level of sophistication
in green roof design can help to maximise these multiple benefits. For
example, installing deeper substrates will confer greater hydrological benefits
but introducing a varied microtopography to the substrate and including some
dead wood and logs on the surface will increase invertebrate biodiversity
(Grant 2006). Green roofs have a great potential as multifunctional urban
design that will hopefully be realised more in the UK in the future.
Finally, this thesis has demonstrated the potential to use green roofs as
‘ecosystem services’ providers. Ecosystem services are defined as services
provided by the natural environment that benefit people, such as food and
fuel provision and recreation but also climate regulation, flood protection and
air and water purification (Defra 2007). This thesis has provided firm
evidence that green roofs can contribute greatly to those last four services,
although perhaps with the exception of water purification – see below.
Ecosystem services contribute to economic welfare both directly, by the
generation of income, but also by the prevention of damages which inflict
costs on society. It is in this respect that green roofs can be promoted within
policies, or when demonstrating that investment in natural capital makes
economic sense, for example within the Millenium Ecosystem Assessment
which raised awareness of how much people rely on ecosystems for the
services they provide (Layke 2009). However, the use of green roofs as
ecosystem services will rely on their correct maintenance, as has been
demonstrated in this thesis. Benefits such as the cooling effect and air
pollution capture will be reduced when the vegetation layer is damaged.
Returning to the results of Chapter 8, it has been seen that green roofs
produce mixed, and potentially negative, results with regards to runoff water
quality. It is possible, therefore, that green roofs could be classed within
‘ecosystem disservices’, which are defined as ecosystem functions that are
perceived as negative for human well-being, such as vegetation damage to
physical structures, pollen allergies, and vectors for disease (Lyytimaki and
Sipila 2009). It is certainly unfortunate that this particular combination of
organic material within an urban catchment can have negative consequences
for water quality, however, the list of benefits of green roofs far outweighs this
217
potential disservice. If necessary, monitoring programmes can be instigated
to identify problematic green roofs and runoff treatment methods used.
9.3.1 Significance for climate change adaptation
A policy support for investing in GI comes from the Climate Change Act of
2008 which requires the Government to report on climate change adaptation
every five years and requires public bodies to undertake climate change risk
assessments and plan for addressing those risks (Natural England 2009).
Green roofs are often included in climate change adaptation strategies which
are based on increasing the proportion of greenspace within cites (Hunt and
Watkiss 2011). The primary climate change projections of concern for
Manchester are: increased temperatures and a higher frequency of heatwave
events; altered precipitation manifesting as wetter winters and a potential
increase in heavy summer downpours leading to increased flood risk. A
secondary effect may also be an increase in air pollution during periods of
static air masses. These have associated deleterious effects on human
health and amenity.
The results of this thesis, when upscaled to the city scale, indicate that green
roofs may prove useful as climate change adaptation techniques. While
upscaling with regards to the cooling effect was beyond the scope of this
thesis, a cumulative reduction of the UHI effect is expected from widescale
roof-greening. When combined with tree planting projects such as the
‘million trees’ project in New York City (Stone 2012) the thermal balance of
urban areas can be shifted away from one dominated by the re-radiation of
stored energy and a subsequent rise in sensible heat. Modelling work has
shown that when included with other green infrastructure they can keep
temperatures in Manchester the same as 1961 – 1990 baseline levels up to
the 2080s high emissions scenario (Gill 2007). The model used a very
simplified green roof component and the cooling effect may be increased
further if it was to use results from this thesis. As long as the green roofs
receive adequate irrigation during prolonged dry periods, the cooling effects
will be maintained.
218
While the cooling effects have been demonstrated in this study for a cool
temperate climate, green roofs may not be as useful for climate change
adaptation in arid, desert cites such as Phoenix, USA or Fes, Morocco.
Cities in these areas need different techniques that are not so reliant on
water resources, which are scarce in desert environments (Stone 2012). In
desert locations, high-density building plans, high albedo building materials
and cool roofs would be better suited. Therefore green roofs should not be
seen as a panacea for all environments but they can certainly aid in schemes
that consider other infrastructure adaptations in an appropriate climate. In
temperate climates, especially, they are able to assist urban planners reach
city targets for climate change adaptation.
9.3.2 Recommendations for Manchester City Council
In light of the findings of this thesis, there are a number of green roof
installation recommendations that can be made to the CASE partner, MCC:
 Intensive roofs can provide greater rewards in terms of benefits such

as pollution removal (Chapter 5), cooling (Chapter 6) and runoff
retention (Chapter 7).
 Green roofs intended to lower air pollution should be planted with
grasses, as they can capture more particulate pollution than traditional
sedum-planted roofs (Chapter 5).
 The location of the green roof can influence the quantities of
particulate pollution captured (Chapter 5).
 Intensive green roofs provide significant cooling of the overlying air
mass, but correct maintenance and irrigation in dry periods may be
required (Chapter 6).
 Care must be taken that air pollution does not get substituted for water
pollution (Chapter 8), and also that freshly installed substrates do not
significantly deteriorate runoff water quality.
219
9.4 Strengths and limitations
The primary strength of this study was the multiple impact aspect. By
demonstrating how green roofs can perform in Manchester with regards to
several functions, the study attained a considerable breadth. When this
breadth is combined with the local relevance, the findings of the study will
provide invaluable local evidence to the CASE partner, Manchester City
Council. The information will feed into the evidence base for the business
case of the GI Strategy that MCC is developing (Chapter 1), by highlighting
the multifunctional benefits of GI. Green roof installations in Manchester can
proceed by being better informed about the expected benefits but also with
some advice for installation such as intensive roofs better due to the
enhanced benefits they provide, and cautions such as the need for
appropriate maintenance and irrigation.
Another strength of the study was its use of empirical methods which
included a wide range of field and laboratory techniques to quantify the
benefits of green roofs. A wealth of data were generated from the
temperature sensors, weather station and the pressure transducers which
allowed the investigation of temporal trends on the intensive green roof. The
study benefitted from access to various laboratory equipment such as a
Molspin magnetometer, Scanning Electron Microscope, PM10 monitor,
thermal imaging camera, ICP-OES, ICP-MS, Guelph permeameter and
access to laboratory equipment which enabled a full soil analysis to be
carried out.
A further strength of the thesis is that the studies were able to be carried out
on a real, in-situ green roof, as opposed to artificial test rigs. This allowed
sensible generalisations to the real world situation to be made.
The main aim of the thesis was to look across several regulating functions of
green roofs during the same period. The scope of the study as a whole
therefore understandably limited the extent of detail possible in each of the
individual components. It is recognised that the investigation of the cooling
effects of green roofs would have benefitted from the monitoring of a number
of extra variables. The initial aim of the study was to monitor the effects of a
220
healthy green roof on air temperatures in comparison to a bare roof. The
experimental design was set up with this in mind but was then able to capture
the effects of green roof damage which occurred during the study period due
to circumstances outside of the researcher’s control. In retrospect, the
analysis of these results would have benefitted from a quantification of the
green roof damage via measurements of plant biomass per unit area or leaf
area index, at the start and at repeated intervals, to observe how vegetation
recovery affected the thermal properties.
Similarly, measurements of albedo, evapotranspiration and air humidity

would have provided more information on the underlying mechanisms of the
net cooling differences observed between the roof surfaces. This may then
uncover reasons for the unusual result of daytime relative warming over the
green roofs. These variables were considered for inclusion in the study but
there was a lack of access to necessary equipment and in the case of
humidity, data had to be ignored due to recording errors. Budget constraints
and equipment failure are common aspects of scientific research (Gomez
and Jones III 2010).
Another possible limitation is related to the length of the studies and the
heavy dependence of urban climate experiments on local weather patterns.
The runoff retention of a green roof is normally strongly related to the rainfall
duration, quantity and ADWP and the study was carried out in a year with an
unusually wet summer and autumn. This can certainly be viewed as a
benefit by enabling comments to be made on how green roofs can cope with
extreme events which may become more common in some climate change
predictions. However, a longer study may have been able to capture
weather patterns more representative of a normal year. Similarly, with the air
pollution reduction experiment, a study period of five weeks, while showing
significant increases in particulate capture over time and allowing significant
site and species differences to be seen, may not have been long enough to
demonstrate equilibrium with atmospheric PM10 concentrations and
fluctuations mentioned in previous work (Mitchell et al. 2010). This means
the work may have potentially underestimated the particle capture efficiency.
221
9.5 Research recommendations
This thesis has provided valuable data to increase the evidence base for the
impacts of green roofs. However there remains a need for more quantitative
studies on the benefits of green roofs in a number of different climates and
urban settings. With reference to the work carried out in the present thesis,
there exist a number of potential directions in which future studies could be
taken.
Chapter 5 filled a research gap with regards to quantifying the air pollution
reduction capabilities of green roofs, however there remains much work to be
done in this area. Source characterisation of the captured particulate
material via mineralogical or magnetic techniques (Zajzon et al. 2013) can
determine whether the particles originated from vehicular traffic, or smelting
and industries and domestic heating systems. Using Pb isotopes to
differentiate between geogenic and anthropogenic sources (Del Rio-Salas et
al. 2012) would be particularly informative on the Precinct roof used in this
thesis.
There is a need for cheap and reliable techniques to be devised to quantify

the amounts of other significant air pollutants such as NOx and O3 that can be
captured by green roof vegetation to see if differences exist for gaseous
pollutants. Work can then be carried out to assess the potential health
impacts of extensive roof greening schemes in cities affected by different
forms of air pollution.
Work could be done on the fate of the captured pollutants. One assumption
in this study is that captured particles are washed from the leaves by rain but
the ultimate fate of these particles is unknown. Tracer studies could be used
to elucidate the physical pathways that pollutants take through the green roof,
and their residence time within the substrate, for example. Related to this is
the effect of the pollutants on the green roof vegetation species. This has
implications for the health of green roofs which needs to be maintained in
order to sustain their benefits.
222
Green roofs are located away from the vehicular sources of air pollution in
cities, therefore it would be interesting to investigate the less common vertical
counterpart to green roofs which are called living walls or green facades.
Work has already shown how they can increase cooling in urban canyons
(Alexandri and Jones 2008), so there is a need to empirically investigate how
they can impact on urban air quality, beyond the modelling work (Pugh et al.
2012) that has been carried out so far.
This thesis was able to monitor the effects of green roof damage on the
cooling effects, however, future work could research this issue more directly
by controlling the damage amount and monitoring other variables such as
LAI, evapotranspiration and albedo, that can provide some explanation for
the effects that damage have. The spatial cooling influence around green
roofs also needs to be determined quantitatively so that comments can be
made on the cumulative cooling effects expected from large-scale green roof
projects. This would allow the relative contribution of green roofs to climate
change adaptation schemes to be assessed, so that their inclusion can be
better informed.
The results from Chapter 6 can be used within modelling studies. For
instance, the U-values for the Precinct building can be sought and used
within an energy balance such as Envimet, which may allow a fuller
discussion on the thermal processes behind the different impacts on
overlying air temperature.
The runoff retention efficiency of intensive and/or aged green roofs should be
tested in a range of climates and local weather patterns. Vegetation type and
biomass per unit area should always be recorded to assess the relative
influence of this green roof component on retention characteristics.
A feature of the study in Chapter 8 is that it identified more than one potential
source of the Pb in the runoff. Characterisation of the Pb in the substrates
would allow firmer conclusions that the Pb in the soil came from historic
atmospheric deposition, wind deposited RDS or locally contaminated soil.
This would then allow a more robust conclusion about the relative
contribution of the lead flashing to be made.
223
Green roof runoff quality studies in the future, linked with air pollution studies,
should investigate contemporary pollutants such as Volatile Organic
Compounds (VOCs) and PAHs for their sequestration potential,
biogeochemical characteristics, and their effects on green roof plants / biota.
_____________________________________________________________
This research has shown that the creative use of green roofs within GI
strategies for urban areas can reap multiple benefits. Reductions in
particulate air pollution, localised cooling effects and rainfall retention have
been demonstrated for the city of Manchester within this thesis. The
evidence presented strongly suggests that green roofs can be used to help
climate-proof cities and make them less susceptible to both the deleterious
effects of urbanisation and predicted climate changes. Advances in green
roof technology and innovative design, reductions in costs as the industry
grows, and bold moves by building owners and city planners to install more
intensive green roofs, could all combine to create a future where green roofs
are ubiquitous within UK cities.
224
REFERENCES
Abbasi, T. & Abbasi, S. A. (2011). Sources of Pollution in Rooftop Rainwater

Harvesting Systems and Their Control. Critical Reviews in
Environmental Science and Technology, 41(23), 2097-2167.
Aitkenhead-Peterson, J. A., Dvorak, B. D., Voider, A. & Stanley, N. C. (2011).
Chemistry of growth medium and leachate from green roof systems in
south-central Texas. Urban Ecosystems, 14(1), 17-33.
Akbari, H. & Konopacki, S. (2005). Calculating energy-saving potentials of
heat-island reduction strategies. Energy Policy, 33(6), 721-756.
Akbari, H. & Matthews, H. D. (2012). Global cooling updates: Reflective roofs
and pavements. Energy and Buildings, 55, 2-6.
Akbari, H., Menon, S. & Rosenfeld, A. (2009). Global cooling: increasing
world-wide urban albedos to offset CO2. Climatic Change, 94(3-4),
275-286.
Akbari, H. & Taha, H. (1992). The impact of trees and white surfaces on
residential heating and cooling energy use in 4 Canadian cities.
Energy, 17(2), 141-149.
Albanese, S. & Cicchella, D. (2012). Legacy Problems in Urban
Geochemistry. Elements, 8(6), 423-428.
Alexandri, E. & Jones, P. (2007). Developing a one-dimensional heat and
mass transfer algorithm for describing the effect of green roofs on the
built environment: Comparison with experimental results. Building and
Environment, 42(8), 2835-2849.
Alexandri, E. & Jones, P. (2008). Temperature decreases in an urban canyon
due to green walls and green roofs in diverse climates. Building and
Environment, 43(4), 480-493.
American Wick Drain Corp accessed online 01/05/2012
www.americanwick.com.
Anderson, K. & Bows, A. (2008). Reframing the climate change challenge in
light of post-2000 emission trends. Philosophical Transactions of the
Royal Society a-Mathematical Physical and Engineering Sciences,
366(1882), 3863-3882.
Anderson, K. & Bows, A. (2011). Beyond 'dangerous' climate change:
emission scenarios for a new world. Philosophical Transactions of the
Royal Society a-Mathematical Physical and Engineering Sciences,
369(1934), 20-44.
AQEG (2005). Particulate Matter in the UK. Defra. London.
AQEG (2007). Air Quality Expert Group. Air Quality and Climate Change: A
UK Perspective. London: Published for The Department for the
Environment, Food and Rural Affairs.
ARCC. (2013). Adaptation and Resilience to a Changing Climate - A
coordinating research network. [Online]. Available: http://ukcip-
arcc.org.uk/ [Accessed 09/03/2013].
Armson, D., Stringer, P. & Ennos, A. R. (2012). The effect of tree shade and
grass on surface and globe temperatures in an urban area. Urban
Forestry & Urban Greening, 11(3), 245-255.
Armstrong, B. G., Chalabi, Z., Fenn, B., Hajat, S., Kovats, S., Milojevic, A. &
Wilkinson, P. (2011). Association of mortality with high temperatures in
225
a temperate climate: England and Wales. Journal of Epidemiology and
Community Health, 65(4), 340-345.
Arnfield, A. J. (2003). Two decades of urban climate research: a review of
turbulence, exchanges of energy and water, and the urban heat island.
International Journal of Climatology, 23(1), 1-26.
Barraclough, D. (2007). UK soil and herbage pollutant suervey: UKSHS
report no. 1. Introduction and summary. Bristol: Environment Agency.
Barrio, E. P. D. (1998). Analysis of the green roofs cooling potential in
buildings. Energy and Buildings, 27(2), 179-193.
Barriopedro, D., Fischer, E. M., Luterbacher, J., Trigo, R. & Garcia-Herrera,
R. (2011). The Hot Summer of 2010: Redrawing the Temperature
Record Map of Europe. Science, 332(6026), 220-224.
Baumann, N. (2006). Ground-nesting birds on green roofs in Switzerland:
Preliminary observations. Urban Habitats, 4, 37-50.
Bechtel, B. & Schmidt, K. J. (2011). Floristic mapping data as a proxy for the
mean urban heat island. Climate Research, 49(1), 45-58.
Beck, D. A., Johnson, G. R. & Spolek, G. A. (2011). Amending greenroof soil
with biochar to affect runoff water quantity and quality. Environmental
Pollution, 159(8-9), 2111-2118.
Beckett, K. P., Freer-Smith, P. & Taylor, G. (2000). Effective tree species for
local air-quality management. Journal of Arboriculture, 26(1), 12-19.
Bengtsson, L. (2005). Peak flows from thin sedum-moss roof. Nordic
Hydrology, 36(3), 269 - 280.
Berghage, R., Beattie, D., Jarrett, A. & O'Connor, T. (Year). Greenroof
Runoff Water Quality. In: Greening rooftops for sustainable
communities, 2007 Minneapolis, April 29 - May 1.
Berndtsson, J. C. (2010). Green roof performance towards management of
runoff water quantity and quality: A review. Ecological Engineering,
36(4), 351-360.
Berndtsson, J. C., Bengtsson, L. & Jinno, K. (2009). Runoff water quality
from intensive and extensive vegetated roofs. Ecological Engineering,
35(3), 369-380.
Berndtsson, J. C., Emilsson, T. & Bengtsson, L. (2006). The influence of
extensive vegetated roofs on runoff water quality. Science of the Total
Environment, 355(1-3), 48-63.
Birmili, W., Hoffmann, T. (2006). Particulate and Dust Pollution, Inorganic
and Organic Compounds. Encyclopedia of Environmental Pollutants.
Elsevier Ltd.
Black, C. C. & Osmond, C. B. (2003). Crassulacean acid metabolism
photosynthesis: 'working the night shift'. Photosynthesis Research,
76(1-3), 329-341.
Blanusa, T., Vaz Monteiro, M. & Cameron, R. W. F. (2010). Alternatives to
Sedum on green roofs? Utilising plants’ cooling potential. Paper
presented at the World Green Roof Congress. London 2010.
Bonan, G. B. (2002). Ecological Climatology: Concepts and Applications.:
Cambridge University Press, Cambridge, UK.
Bowler, D. E., Buyung-Ali, L., Knight, T. M. & Pullin, A. S. (2010). Urban
greening to cool towns and cities: A systematic review of the empirical
evidence. Landscape and Urban Planning, 97(3), 147-155.
226
Boyd, R. & Ibarraran, M. E. (2009). Extreme climate events and adaptation:
an exploratory analysis of drought in Mexico. Environment and
Development Economics, 14, 371-395.
Boyle, J. F. (2000). Rapid elemental analysis of sediment samples by isotope
source XRF. Journal of Paleolimnology, 23(2), 213-221.
Brady, N. C. & Weil, R. R. (2008). The nature and properties of soils (14th ed.
ed.). Harlow: Prentice Hall.
Breda, N. J. J. (2003). Ground-based measurements of leaf area index: a
review of methods, instruments and current controversies. Journal of
Experimental Botany, 54(392), 2403-2417.
Brenneisen, S. (2003). The benefits of biodiversity from green roofs: Key
design consequences. In: In Proc. of 1st North American Green Roof
Conference: Greening rooftops for sustainable communities,, 2003
Chicago 29-30 May 2003. The Cardinal Group, Toronto., 323-329.
Brenneisen, S. (2006). Space for Urban Wildlife: Designing Green Roofs as
Habitats in Switzerland. Urban Habitats, 4(1), 27-37.
Brimblecombe, P. & Maynard, R. (2001). The Urban Atmosphere and its
Effects. London: Imperial College Press.
Bukowiecki, N., Lienemann, P., Hill, M., Furger, M., Richard, A., Amato, F.,
Prévôt, A. S. H., Baltensperger, U., Buchmann, B. & Gehrig, R. (2010).
PM10 emission factors for non-exhaust particles generated by road
traffic in an urban street canyon and along a freeway in Switzerland.
Atmospheric Environment, 44(19), 2330-2340.
Bunge, W. W. (1973). Geography. Professional Geographer, 25(4), 331-337.
Buss, S. R., Rivett, M. O., Morgan, P. & Bemment, C. D. (2005). Attenuation
of nitrate in the sub-surface environment. Science Report
SC030155/SR2 for the Environment Agency.
Carey, R. O., Hochmuth, G. J., Martinez, C. J., Boyer, T. H., Dukes, M. D.,
Toor, G. S. & Cisar, J. L. (2013). Evaluating nutrient impacts in urban
watersheds: Challenges and research opportunities. Environmental
Pollution, 173, 138-149.
Carter, J. G. (2011). Climate change adaptation in European cities. Current
Opinion in Environmental Sustainability, 3(3), 193-198.
Carter, T. & Jackson, C. R. (2007). Vegetated roofs for stormwater
management at multiple spatial scales. Landscape and Urban
Planning, 80(1-2), 84-94.
Carter, T. L. & Rasmussen, T. C. (2006). Hydrologic behavior of vegetated
roofs. Journal of the American Water Resources Association, 42(5),
1261-1274.
Castleton, H. F., Stovin, V., Beck, S. B. M. & Davison, J. B. (2010). Green
roofs; building energy savings and the potential for retrofit. Energy and
Buildings, 42(10), 1582-1591.
Cavan, G. (2010). Climate Change Projections for Greater Manchester. In:
EcoCities project, U. o. M. (ed.). Manchester, UK.
Cavan, G. & Kazmierczak, A. (2011). Urban greening to adapt urban areas to
climate change: Oxford Road Corridor case study. EcoCities project,
University of Manchester, Manchester, UK.
Cheung, H. K. W. (2011). An urban heat island study for building and urban
design. PhD Thesis, The University of Manchester.
227
Cheung, H. K. W., Levermore, G. J. & Watkins, R. (2010). A low cost, easily
fabricated radiation shield for temperature measurements to monitor
dry bulb air temperature in built up urban areas. Building Services
Engineering Research & Technology, 31(4), 371-380.
Chiaradia, M. & Cupelin, F. (2000). Behaviour of airborne lead and temporal
variations of its source efects in Geneva (Switzerland): Comparison of
anthropogenic versus natural processes. Atmospheric Environment,
34(6), 959 - 971.
Christensen, J. H., B. Hewitson, A. Busuioc, A. Chen, X. Gao, I. Held, R.
Jones, R.K. Kolli, W.-T. Kwon, R. Laprise, V. Magaña Rueda, L. &
Mearns, C. G. M., J. Räisänen, A. Rinke, A. Sarr and P. Whetton,
(2007). Regional Climate Projections. In: Climate Change 2007: The
Physical Science Basis. Contribution of Working Group I to the Fourth
Assessment Report of the Intergovernmental Panel on Climate
Change.
CIRIA (2007). C697, The SUDS Manual. Construction Industry Research
and Information Association. www.ciria.org.
Clark, C., Adriaens, P. & Talbot, F. B. (2008). Green roof valuation: A
probabilistic economic analysis of environmental benefits.
Environmental Science & Technology, 42(6), 2155-2161.
Cloy, J. M., Farmer, J. G., Graham, M. C., MacKenzie, A. B. & Cook, G. T.
(2008). Historical records of atmospheric Pb deposition in four Scottish
ombrotrophic peat bogs: An isotopic comparison with other records
from western Europe and Greenland. Global Biogeochemical Cycles,
22(2).
Coffman, R. R., and Davis, G. (2005). Insect and avian fauna presence on
the Ford assembly plant ecoroof. In: In Proc. of 3rd North American
Green Roof Conference: Greening rooftops for sustainable
communities,, 2005 Washington, DC. 4–6 May 2005.: The Cardinal
Group, Toronto., 457-468.
Copenhagen Accord (2009). FCCC/CP/2009/L.7. Copenhagen, Denmark:
United Nations Climate Change Conference.
Corridor Manchester (2012). Map of the Corridor. Available at
www.corridormanchester.com. Accessed 01/05/2013.
Coumou, D. & Rahmstorf, S. (2012). A decade of weather extremes. Nature
Climate Change, 2(7), 491-496.
Currie, B. A. & Bass, B. (2008). Estimates of air pollution mitigation with
green plants and green roofs using the UFORE model. Urban
Ecosystems, 11(4), 409-422.
Czerniawska-Kusza, I., Kusza, G. & Duzynski, M. (2004). Effect of deicing
salts on urban soils and health status of roadside trees in the opole
region. Environmental Toxicology, 19(4), 296-301.
Dale, M. (2005). Impact of climate change on UK flooding and future
predictions. Proceedings of the Institution of Civil Engineers-Water
Management, 158(4), 135-140.
Daniels, M. J., Dominici, F., Zeger, S. L. & Samet, J. M. (2004). The National
Morbidity, Mortality, and Air Pollution Study. Part III: PM10
concentration-response curves and thresholds for the 20 largest US
cities. Research report (Health Effects Institute), (94 Pt 3 ), 1-21;
discussion 23-30 Cambridge, MA:Health Effects Institute.
228
Davies, M., Steadman, P. & Oreszczyn, T. (2008). Strategies for the
modification of the urban climate and the consequent impact on
building energy use. Energy Policy, 36(12), 4548-4551.
Davies, R., Simcock, R. & Toft, R. (2012). Biodiversity opportunities for a NZ
indigenous living roof. World Green Roof Congress. Copenhagen, 18 -
21 September 2012.
De Berardis, B., Incocciati, E., Massera, S., Gargaro, G. & Paoletti, L. (2007).
Airborne silica levels in an urban area. Science of the Total
Environment, 382(2–3), 251-258.
Dean, W. E. (1974). Determination of carbonate and organic matter in
calcareous sediments and sedimentary rocks by loss on ignition -
comparison with other methods. Journal of Sedimentary Petrology,
44(1), 242-248.
Defra (2007). An introductory guide to valuing ecosystem services. Report
published by the Department for Environment, Food and Rural Affairs
Accessed online 17/05/2013
https://www.gov.uk/government/uploads/system/uploads/attachment_
data/file/181882/pb12852-eco-valuing.pdf.pdf.
Defra (2008). Making Space for Water. Urban flood risk & integrated
drainage (HA2). IUD pilot summary report. June 2008. Halcrow Group
Ltd & Defra, London.
Defra (2010). The River Basin Districts Typology, Standards and
Groundwater threshold values (Water Framework Directive) (England
and Wales) Directions 2010. Accessed online at
http://archive.Defra.gov.uk/environment/quality/water/wfd/documents/2
010directions.pdf 15/02/2013.
Defra (2011a). Local Air Quality Management - AQMA Maps.
Defra (2011b). Water for Life (White Paper). Defra, London. Accessed online
at http://www.official-documents.gov.uk/ 05/02/2013.
Defra (2012). Tackling water pollution from the urban environment.
Consultation on a strategy to address diffuse water pollution from the
built environment. Defra, London. Available online www.Defra.gov.uk
05/02/2013.
Defra. (2013). UK deposition data. Available at
http://pollutantdeposition.Defra.gov.uk/data accessed 25/02/2013.
Defra, London [Online]. [Accessed].
Del Rio-Salas, R., Ruiz, J., De la O-Villanueva, M., Valencia-Moreno, M.,
Moreno-Rodriguez, V., Gomez-Alvarez, A., Grijalva, T., Mendivil, H.,
Paz-Moreno, F. & Meza-Figueroa, D. (2012). Tracing geogenic and
anthropogenic sources in urban dusts: Insights from lead isotopes.
Atmospheric Environment, 60, 202-210.
DeNardo, J. C., Jarrett, A. R., Manbeck, H. B., Beattie, D. J. & Berghage, R.
D. (2005). Stormwater mitigation and surface temperature reduction
by green roofs. Trans. ASAE, 48 (4), 1491 - 1496.
DfT (2009). Department for Transport: Transport Trends 2009 Edition.
London: HMSO.
DfT (2013). Road traffic (vehicle miles) by vehicle type in Great Britain,
annual from 1949. Table TRA0101 Available at
www.gov.uk/government/organisations/department-for-
transport/series/road-traffic-statistics Accessed 16/03/2013.
229
DiGiovanni, K., Montalto, F., Gaffin, S. & Rosenzweig, C. (2013). Applicability
of Classical Predictive Equations for the Estimation of
Evapotranspiration from Urban Green Spaces: Green Roof Results.
Journal of Hydrologic Engineering, 18(1), 99-107.
Dimoudi, A. & Nikolopoulou, M. (2003). Vegetation in the urban environment:
microclimatic analysis and benefits. Energy and Buildings, 35(1), 69-
76.
Dise, N. B., Matzner, E. & Gundersen, P. (1998). Synthesis of nitrogen pools
and fluxes from European forest ecosystems. Water Air and Soil
Pollution, 105(1-2), 143-154.
Dise, N. B. & Wright, R. F. (1995). Nitrogen leaching from European forests
in relation to nitrogen deposition. Forest Ecology and Management,
71(1-2), 153-161.
Donaldson, G. C., Kovats, R.S., Keatinge, W.R., McMichael, R.J. (2001).
Heat- and cold-related mortality and morbidity and climate change. In:
In: Maynard, R. L. (ed.) Health effects of climate change in the UK. .
London: Department of Health.
Douglas, I. (1991). The urban environment - a sourcebook for the 1990s.
Applied Geography, 11(3), 245-246.
Dunnett, N. & Kingsbury, N. (2004). Planting Green Roofs and Living Walls.:
Timber Press, Portland.
Dunnett, N., Nagase, A. & Hallam, A. (2008a). The dynamics of planted and
colonising species on a green roof over six growing seasons 2001-
2006: influence of substrate depth. Urban Ecosystems, 11(4), 373-384.
Dunnett, N., Nagase, A., Booth, R. & Grime, P. (2008b). Influence of
vegetation composition on runoff in two simulated green roof
experiments. Urban Ecosystems, 11(4), 385-398.
EA. (2011). Environment Agency UK - Chemical Standards Database.
Available at http://evidence.environment-
agency.gov.uk/ChemicalStandards/Home.aspx. Accessed online
11/02/2013.
EA (2012). Method Statement for Nitrate Vulnerable Zone review - Eutrophic.
Environment Agency report to Defra and Welsh Government -
supporting paper for the implementation of the nitrates directive 2013 -
2016.
EEA (2013). PM10 36th highest daily value, 2010. Downloaded from
http://www.eea.europa.eu/data-and-maps/figures/pm10-36th-highest-
daily-value-2010. Accessed 03/05/2013.
Ellis, J. B. & Mitchell, G. (2006). Urban diffuse pollution: key data information
approaches for the Water Framework Directive. Water and
Environment Journal, 20(1), 19-26.
Emilsson, T., Berndtsson, J. C., Mattsson, J. E. & Rolf, K. (2007). Effect of
using conventional and controlled release fertiliser on nutrient runoff
from various vegetated roof systems. Ecological Engineering, 29(3),
260-271.
Erell, E., Pearlmutter, D. & Williamson, T. (2011). Urban Microclimate:
Designing the Spaces Between Buildings. London: Earthscan Ltd.
Ernst, W. & Weigerding, I. (1985). Oberflächenentwässerung,
Gewässerentlastung durch ökologische/ökonomische Planung.
Bundesblatt, 34(11), 722-732.
230
European Commision (1996). Future noise policy. European Commision
green paper. Brussels COM(96) 540 final.
European Community (2000). Directive 2000/60/EC of the European
Parliament and of the Council of 23 October 2000 establishing a
framework for Community action in the field of water policy. Official
Journal of the European Communities, L327, 1 - 72.
Evans, U. R. (1960). The corrosion and oxidation of metals: scientific
principles and practical applications: St. Martin's Press.
Farzaneh, R. (2005). Evapotranspiration rates from extensive green roof
plant species. Pennsylvania State University.
Fioretti, R., Palla, A., Lanza, L. G. & Principi, P. (2010). Green roof energy
and water related performance in the Mediterranean climate. Building
and Environment, 45(8), 1890-1904.
FLL (2008). Forschungsgesellschaft Landschaftsentwicklung Landschaftsbau
e. V. (FLL) Guidelines for the Planning, Construction and
Maintenance of Green Roofing. English language edition, available at
www.fll.de ed. FLL, Bonn.
Fordyce, F. M., Brown, S. E., Ander, E. L., Rawlins, B. G., O'Donnell, K. E.,
Lister, T. R., Breward, N. & Johnson, C. C. (2005). GSUE: urban
geochemical mapping in great Britain. Geochemistry-Exploration
Environment Analysis, 5, 325-336.
Forster, J. (1999). Variability of roof runoff quality. Water Science and
Technology, 39(5), 137-144.
Fowler, D., Cape, J. N. & Unsworth, M. H. (1989). Deposition of atmospheric
pollutants on forests. Philosophical Transactions of the Royal Society
of London Series B-Biological Sciences, 324(1223), 247-265.
Fowler, D., Cape, N., Smith, R., Nemitz, E., Sutton, M., Dore, T., Coyle, M.,
Crossley, A., Storeton-West, R., Muller, J., Phillips, G., Thomas, R.,
Vieno, M., Yang, S., Famulari, D., Twigg, M., Bealey, B., Benham, D.,
Hayman, G., Lawrence, H., Vincent, K., Fagerli, H. & Simpson, D.
(2007). Acid Deposition Processes: RMP 2258 Final report to Defra.
Centre for Ecology and Hydrology.
Fowler, D., McDonald, A. G., Crossley, A., Nemitz, E., Leaver, J. N., Cape, R.
I., Smith, D., Anderson, P., Rowland, P., Ainsworth, G., Lawlor, A. J.,
Guyatt, H. & Harmens, H. (2006). UK Heavy Metal Monitoring Network.
Project Number EPG 1/3/204 for Defra.
Francis, R. A. & Lorimer, J. (2011). Urban reconciliation ecology: The
potential of living roofs and walls. Journal of Environmental
Management, 92(6), 1429-1437.
Frazer, L. (2005). Paving Paradise. Environmental Health Perspectives, 113,
457-462.
Freer-Smith, P. H., Beckett, K. P. & Taylor, G. (2005). Deposition velocities to
Sorbus aria, Acer campestre, Populus deltoides X trichocarpa
'Beaupre', Pinus nigra and X Cupressocyparis leylandii for coarse, fine
and ultra-fine particles in the urban environment. Environmental
Pollution, 133(1), 157-167.
Frith, M. G., D. (2000). The black redstart in urban Britain: a conservation
conundrum? British Wildlife,, 8, 381-388.
Fuller, R. A. & Gaston, K. J. (2009). The scaling of green space coverage in
European cities. Biology Letters, 5(3), 352-355.
231
Gaffin, S. R., Khanbilvardi, R. & Rosenzweig, C. (2009). Development of a
Green Roof Environmental Monitoring and Meteorological Network in
New York City. Sensors, 9(4), 2647-2660.
Getter, K. L. & Rowe, D. B. (2006). The role of extensive green roofs in
sustainable development. Hortscience, 41(5), 1276-1285.
Getter, K. L. & Rowe, D. B. (2008). Selecting plants for extensive green roofs
in the United States. Extension Bulletin E-3047 Michigan State
University. Available at http://www.hrt.msu.edu/greenroof/ Accessed
05/05/2013.
Getter, K. L., Rowe, D. B. & Andresen, J. A. (2007). Quantifying the effect of
slope on extensive green roof stormwater retention. Ecological
Engineering, 31(4), 225-231.
Getter, K. L., Rowe, D. B., Andresen, J. A. & Wichman, I. S. (2011).
Seasonal heat flux properties of an extensive green roof in a
Midwestern US climate. Energy and Buildings, 43(12), 3548-3557.
Getter, K. L., Rowe, D. B., Robertson, G. P., Cregg, B. M. & Andresen, J. A.
(2009). Carbon Sequestration Potential of Extensive Green Roofs.
Environmental Science & Technology, 43(19), 7564-7570.
Gill, S., Pauleit, S., Ennos, A. R., Lindley, S. J., Handley, J. F., Gwilliam, J. &
Ueberjahn-Tritta, A. (2004). Literature review: Impacts of climate
change on urban environments.: Centre for Urban and Regional
Ecology, University of Manchester. Available at
www.sed.manchester.ac.uk/research/cure/downloads/asccue_litrevie
w.pdf. Accessed 27/02/2013.
Gill, S. E., Handley, J.F., Ennos, A.R. & Pauleit, S. (2007). Adapting Cities for
Climate Change: The Role of the Green Infrastructure. Built
Environment, 33(1), 115-133.
GMTU. (2010). Greater Manchester Transportation Unit Link Count - Oxford
Road, under Mancunian Way. Ref 1007027 [Online]. Available:
http://www.gmtu.gov.uk/surveys.html [Accessed December 2011].
Gobel, P., Dierkes, C. & Coldewey, W. C. (2007). Storm water runoff
concentration matrix for urban areas. Journal of Contaminant
Hydrology, 91(1-2), 26-42.
Gomez, B. & Jones III, J. P. (eds.) (2010). Research methods in geography:
John Wiley & Sons Ltd, Chichester.
Gómez, F., Gil, L. & Jabaloyes, J. (2004). Experimental investigation on the
thermal comfort in the city: relationship with the green areas,
interaction with the urban microclimate. Building and Environment,
39(9), 1077-1086.
Gomez, F., Montero, L., De Vicente, V., Sequi, A. & Castilla, N. (2008).
Vegetation influences on the human thermal comfort in outdoor
spaces: criteria for urban planning. In: Gospodini, A. & Brebbia, C. A.,
eds. 5th International Conference on Urban Regeneration and
Sustainability (The Sustainable City), Sep 24-26 2008 Skiathos,
GREECE. Wit Press/Computational Mechanics Publications, 151-163.
Goodrich, B. A., Koski, R. D. & Jacobi, W. R. (2009). Condition of Soils and
Vegetation Along Roads Treated with Magnesium Chloride for Dust
Suppression. Water Air and Soil Pollution, 198(1-4), 165-188.
Grant, G. (2006). Extensive green roofs in London. Urban Habitats, 4(1), 51 -
65.
232
Grant, G., Luke Engleback, and Barry Nicholson. (2003). Green roofs: their
existing status and potential for conserving biodiversity in urban areas.
In: Produced for English Nature by EcoSchemes, L. (ed.) English
Nature Research Report No. 498.
Green Roof Centre. (2013). Green Roof Centre Supporters [Online].
Available: http://www.thegreenroofcentre.co.uk/about_us/supporters
[Accessed 16 May 2013].
Gregoire, B. G. & Clausen, J. C. (2011). Effect of a modular extensive green
roof on stormwater runoff and water quality. Ecological Engineering,
37(6), 963-969.
GRO (2011). The GRO Green Roof Code - Green Roof Code of Best
Practice for the UK 2011. Groundwork Sheffield.
Gupta, R. & Gregg, M. (2012). Using UK climate change projections to adapt
existing English homes for a warming climate. Building and
Environment, 55(0), 20-42.
Hall, G. E. M., Bonham-Carter, G. F., Horowitz, A. J., Lum, K., Lemieux, C.,
Quemerais, B. & Garbarino, J. R. (1996). The effect of using different
0.45 μm filter membranes on ‘dissolved’ element concentrations in
natural waters. Applied Geochemistry, 11(1–2), 243-249.
Hall, J. M., Handley, J. F. & Ennos, A. R. (2012). The potential of tree
planting to climate-proof high density residential areas in Manchester,
UK. Landscape and Urban Planning, 104(3-4), 410-417.
Hamilton, I. G., Davies, M., Steadman, P., Stone, A., Ridley, I. & Evans, S.
(2009). The significance of the anthropogenic heat emissions of
London's buildings: A comparison against captured shortwave solar
radiation. Building and Environment, 44(4), 807-817.
Hansard, R., Maher, B. A. & Kinnersley, R. (2011). Biomagnetic monitoring of
industry-derived particulate pollution. Environmental Pollution, 159(6),
1673-1681.
Harazono, Y., Teraoka, S., Nakase, I. & Ikeda, H. (1990). Effects of rooftop
vegetation using artificial substrates on the urban climate and the
thermal load of buildings. Energy and Buildings, 15(3-4), 435-442.
Hardin, M., Wanielista, M. & Chopra, M. (2012). A Mass Balance Model for
Designing Green Roof Systems that Incorporate a Cistern for Re-Use.
Water, 4(4), 914-931.
Hardin, P. J. & Jensen, R. R. (2007). The effect of urban leaf area on
summertime urban surface kinetic temperatures: A Terre Haute case
study. Urban Forestry & Urban Greening, 6(2), 63-72.
Harner, T., Shoeib, M., Diamond, M., Stern, G. & Rosenberg, B. (2004).
Using Passive Air Samplers To Assess Urban−Rural Trends for
Persistent Organic Pollutants. 1. Polychlorinated Biphenyls and
Organochlorine Pesticides. Environmental Science & Technology,
38(17), 4474-4483.
Harris, P., Lindley, S., Gallagher, M. & Agius, R. (2009). Identification and
verification of ultrafine particle affinity zones in urban neighbourhoods:
sample design and data pre-processing. Environmental Health, 8.
233
Hayman, G., Hasler, S., Vincent, K., Baker, S., Donovan, B., Smith, M.,
Davies, M., Sutton, M., Tang, Y. S., Dragosits, U., Love, L., Fowler, D.,
Sansom, L. & Page, H. (2001). Operation and Management of the UK
Acid Deposition Monitoring Networks: Data Summary for 2000. AEA
Technology report for Defra reference AEAT/ENV/R/0740.
Hebbert, M. & Jankovic, V. (2013). Cities and Climate Change: The
Precedents and Why They Matter. Urban Studies, 50(7), 1332 - 1347.
Heidt, V. & Neef, M. (2008). Benefits of urban green space for improving
urban climate. Ecology, Planning, and Management of Urban Forests:
International Perspectives, 84-96.
Heisler, J., Glibert, P. M., Burkholder, J. M., Anderson, D. M., Cochlan, W.,
Dennison, W. C., Dortch, Q., Gobler, C. J., Heil, C. A., Humphries, E.,
Lewitus, A., Magnien, R., Marshall, H. G., Sellner, K., Stockwell, D. A.,
Stoecker, D. K. & Suddleson, M. (2008). Eutrophication and harmful
algal blooms: A scientific consensus. Harmful Algae, 8(1), 3-13.
Held, A., Nowak, A., Wiedensohler, A. & Klemm, O. (2006). Field
measurements and size-resolved model simulations of turbulent
particle transport to a forest canopy. Journal of Aerosol Science, 37(6),
786-798.
Herrington, L. P. & Vittum, J. S. (1977). Human thermal comfort in urban
outdoor spaces. U S Forest Service General Technical Report NE,
(25), 130-138.
HFAS (2011). Highways Forecasting and Analytical Services Report 1679 -
The Greater Manchester Emissions Inventory 2007 Update.
Hill, A. C. (1971). Vegetation - sink for atmospheric pollutants. Journal of the
Air Pollution Control Association, 21(6), 341-&.
Hilten, R. N., Lawrence, T. M. & Tollner, E. W. (2008). Modeling stormwater
runoff from green roofs with HYDRUS-1D. Journal of Hydrology,
358(3-4), 288-293.
Hock, R., de Woul, M., Radic, V. & Dyurgerov, M. (2009). Mountain glaciers
and ice caps around Antarctica make a large sea-level rise
contribution. Geophysical Research Letters, 36.
Holmer, B., Thorsson, S. & Eliasson, I. (2007). Cooling rates, sky view
factors and the development of intra-urban air temperature differences.
Geografiska Annaler Series a-Physical Geography, 89A(4), 237-248.
Hooda, P. S. & Alloway, B. J. (1998). Cadmium and lead sorption behaviour
of selected English and Indian soils. Geoderma, 84(1-3), 121-134.
Houston, D., Werritty, A., Bassett, D., Geddes, A., Hoolachan, A. & McMillan,
M. (2011). Pluvial (rain-related) flooding in urban areas: the invisible
hazard. Joseph Rowntree Foundation, York. Available at:
www.jrf.org.uk/publications.
Howard, L. (1818). The Climate of London, deduced from meteorological
observations, made at different places in the neighbourhood of the
metropolis - Volume 1. London: W. Phillips.
Huang, Y. J., Akbari, H., Taha, H. & Rosenfeld, A. H. (1987). The potential of
vegetation reducing summer cooling loads in residential buildings.
Journal of Climate and Applied Meteorology, 26(9), 1103-1116.
Huff, F. A. & Changnon, S. A. (1973). Precipitation modification by major
urban areas. Bulletin of the American Meteorological Society, 54(12),
1220-1232.
234
Huggett, R. J. (1993). Modelling the human impact on nature: Systems
analysis of environmental problems.: Oxford University Press, Oxford.
Hulme, M., Jenkins,G.J., Lu,X., Turnpenny,J.R., Mitchell,T.D., Jones,R.G.,
Lowe,J., Murphy,J.M., Hassell,D., Boorman,P., McDonald,R. and
Hill,S. (2002). Climate Change Scenarios for the United Kingdom: The
UKCIP02 Scientific Report,. Tyndall Centre for Climate Change
Research, School of Environmental Sciences, University of East A
nglia, Norwich, UK., 120pp.
Hunt, A. & Watkiss, P. (2011). Climate change impacts and adaptation in
cities: a review of the literature. Climatic Change, 104(1), 13-49.
IPCC (2007a). Climate Change 2007: Synthesis Report. Contribution of
Working Groups I, II and III to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change. In: [Core Writing Team,
P., R.K and Reisinger, A. (eds.)]. (ed.). IPCC, Geneva, Switzerland,.
IPCC (2007b). Climate Change 2007: The Physical Science Bases.
Contribution of Working Group I to the Fourth Assessment Report of
the Intergovernmental Panel on Climate Change.
Iverson, M. A., Holmes, E. P. & Bomke, A. A. (2012). Development and use
of rapid reconnaissance soil inventories for reclamation of urban
brownfields: A Vancouver, British Columbia, case study. Canadian
Journal of Soil Science, 92(1), 191-201.
James, P., Tzoulas, K., Adams, M. D., Barber, A., Box, J., Breuste, J.,
Elmqvist, T., Frith, M., Gordon, C., Greening, K. L., Handley, J.,
Haworth, S., Kazmierczak, A. E., Johnston, M., Korpela, K., Moretti,
M., Niemelä, J., Pauleit, S., Roe, M. H., Sadler, J. P. & Ward
Thompson, C. (2009). Towards an integrated understanding of green
space in the European built environment. Urban Forestry & Urban
Greening, 8(2), 65-75.
Jenkins, G. J., Murphy, J. M., Sexton, D. M. H., Lowe, J. A., Jones, P., Kilsby,
C. G. (2009). UK Climate Projections: Briefing report. Met Office
Hadley Centre, Exeter, UK.
Jim, C. (2012). Effect of vegetation biomass structure on thermal
performance of tropical green roof. Landscape and Ecological
Engineering, 8(2), 173-187.
Jim, C. Y. & Peng, L. L. H. (2012). Weather effect on thermal and energy
performance of an extensive tropical green roof. Urban Forestry &
Urban Greening, 11(1), 73-85.
Johnson, C. C., Breward, N., Ander, E. L. & Ault, L. (2005). G-BASE:
Baseline geochemical mapping of Great Britain and Northern Ireland.
Geochemistry: Exploration-Environment-Analysis, 5 (4), 347 - 357.
Kaczala, F., Marques, M., Vinrot, E. & Hogland, W. (2011). Stormwater run-
off from an industrial log yard: characterization, contaminant
correlation and first-flush phenomenon. Environmental Technology,
33(14), 1615-1628.
Kandpal, G., Srivastava, P. C. & Ram, B. (2005). Kinetics of desorption of
heavy metals from polluted soils: Influence of soil type and metal
source. Water Air and Soil Pollution, 161(1-4), 353-363.
Kardel, F., Wuyts, K., Maher, B. A., Hansard, R. & Samson, R. (2011). Leaf
saturation isothermal remanent magnetization (SIRM) as a proxy for
235
particulate matter monitoring: Inter-species differences and in-season
variation. Atmospheric Environment, 45(29), 5164-5171.
Kardel, F., Wuyts, K., Maher, B. A. & Samson, R. (2012). Intra-urban spatial
variation of magnetic particles: Monitoring via leaf saturation
isothermal remanent magnetisation (SIRM). Atmospheric Environment,
55(0), 111-120.
Kazmierczak, A. K., C. (2011). Risk of flooding to infrastructure in Greater
Manchester. In: EcoCities project, M., UK. (ed.).
Keeling, R. F., Piper, S. C., Bollenbacher, A. F. & Walker, S. J. (2013).
Atmospheric CO2 concentrations (ppm) derived from in situ air
measurements at Mauna Loa Observatory, Hawaii. . La Jolla,
California: Scripps Institution of Oceanography.
Knight, S., Smith, C. & Roberts, M. (2010). Mapping Manchester's urban heat
island. Weather, 65(7), 188-193.
Köhler, M. (2008). Green facades-a view back and some visions. Urban
Ecosystems, 11(4), 423-436.
Köhler, M., and M. Keeley (2005). Berlin: Green roof technology and
development,. In EarthPledge. Green roofs: Ecological design and
construction.: Schiffer Books, Atglen, Pa.
Köhler, M., Schmidt, M. and Laar, M. (2003a). Green roofs as a contribution
to reduce urban heat islands. RIO 3 - World Climate & Energy Event,
1-5 December 2003,. Rio de Janeiro, Brazil.
Köhler, M., Schmidt, M., Grimme, F.W., Laar, M., de Assunc-a˜o Paiva, V.L.,
Tavares, S. (2002). Green roofs in temperate climates and in the Hot-
Humid tropics- far beyond the aesthetics. Environmental Management
and Health, 13(4).
Köhler, M. a. S., M. (2003b). Study of extensive 'Green Roofs' in Berlin. Part
III - Retention of contaminants.
Kosareo, L. & Ries, R. (2007). Comparative environmental life cycle
assessment of green roofs. Building and Environment, 42(7), 2606-
2613.
Kottek, M., Grieser, J., Beck, C., Rudolf, B. & Rubel, F. (2006). World map of
the Koppen-Geiger climate classification updated. Meteorologische
Zeitschrift, 15(3), 259-263.
Kula, R. (2005). Green roofs and the LEED green building rating system.
Greening rooftops for sustainable communities conference
proceedings. Washington D.C. USA, May 4 - 6 2005.
Kumar, R. & Kaushik, S. C. (2005). Performance evaluation of green roof and
shading for thermal protection of buildings. Building and Environment,
40(11), 1505-1511.
Laidlaw, M. A. S., Zahran, S., Mielke, H. W., Taylor, M. P. & Filippelli, G. M.
(2012). Re-suspension of lead contaminated urban soil as a dominant
source of atmospheric lead in Birmingham, Chicago, Detroit and
Pittsburgh, USA. Atmospheric Environment, 49, 302-310.
Landsberg, H. (1981). The Urban Climate. Academic Press, NY.
Larsen, J. (2003). Record heatwave in Europe takes 35000 lives. Earth
Policy Institute.
Larsen, J. (2006). Setting the record straight: more than 52000 Europeans
died from heat in summer 2003. Earth Policy Institute, 2006 [Online].
236
Available: http://www.earth-policy.org/Updates/2006/Update56.htm
[Accessed].
Layke, C. (2009). Measuring Nature's Benefits: A Preliminary Roadmap for
Improving Ecosystem Service Indicators. WRI Working Paper. World
Resources Institute, Washington DC. Accessed online 13 May 2013
www.wri.org/project/ecosystem-service-indicators.
Lazzarin, R. A., Castellotti, F. & Busato, F. (2005). Experimental
measurements and numerical modelling of a green roof. Energy and
Buildings, 37(12), 1260-1267.
Lee, J. H., Bang, K. W., Ketchum, L. H., Choe, J. S. & Yu, M. J. (2002). First
flush analysis of urban storm runoff. Science of the Total Environment,
293(1-3), 163-175.
Lehndorff, E., Urbat, M. & Schwark, L. (2006). Accumulation histories of
magnetic particles on pine needles as function of air quality.
Atmospheric Environment, 40(36), 7082-7096.
Lemonsu, A., Kounkou-Arnaud, R., Desplat, J., Salagnac, J. L. & Masson, V.
(2013). Evolution of the Parisian urban climate under a global
changing climate. Climatic Change, 116(3-4), 679-692.
Lindley, S. J., Handley, J. F., Theuray, N., Peet, E. & McEvoy, D. (2004).
Adaptation strategies for climate change in the urban environment:
Assessing climate change related risk in UK urban areas. In: 13th
Annual SRA-Europe Conference, Nov 15-17 2004 Paris, FRANCE.
Routledge Journals, Taylor & Francis Ltd, 543-568.
Litschke, T. & Kuttler, W. (2008). On the reduction of urban particle
concentration by vegetation - a review. Meteorologische Zeitschrift,
17(3), 229-240.
Liu, J.-m., Li, S.-h. & Yang, Z.-f. (2008). Temperature and humidity effect of
urban green spaces in Beijing in summer. Shengtaixue Zazhi, 27(11),
1972-1978.
Liu, K. & Baskaran, B. (2003). Thermal performance of green roofs through
field evaluation,. In: In Proc. of 1st North American Green Roof
Conference: Greening rooftops for sustainable communities,, 2003
Chicago. 29–30 May 2003.: The Cardinal Group, Toronto., 273-282.
Liu, T. C., Shyu, G. S., Fang, W. T., Liu, S. Y. & Cheng, B. Y. (2012).
Drought tolerance and thermal effect measurements for plants suitable
for extensive green roof planting in humid subtropical climates. Energy
and Buildings, 47(0), 180-188.
Livada, I., M. Santamouris, K. Niachou, N. Papanikolaou, and G.
Mihalakakou (2002). The thermal island effect in the extended region
of Athens. Theor. Appl. Climatol., 71, 219-230.
Long, B., Clark, S. E., Baker, K. H. & Berghage, R. (2006). Green roof media
selection for the minimization of pollutant loading in roof run-off. In:
Proc. of the Water Environment Federation WEFTEC, 2006 Dallas, TX,
October 21 - 25.
Lovett, G. M. (1994). Atmospheric Deposition of Nutrients and Pollutants in
North America: An Ecological Perspective. Ecological Applications,
4(4), 630-650.
Lovett, G. M., Traynor, M. M., Pouyat, R. V., Carreiro, M. M., Zhu, W.-X. &
Baxter, J. W. (2000). Atmospheric Deposition to Oak Forests along an
237
Urban−Rural Gradient. Environmental Science & Technology, 34(20),
4294-4300.
Lundgren, K., Kuklane, K., Gao, C. S. & Holmer, I. (2013). Effects of Heat
Stress on Working Populations when Facing Climate Change.
Industrial Health, 51(1), 3-15.
Lundmark, A. & Olofsson, B. (2007). Chloride deposition and distribution in
soils along a deiced highway - Assessment using different methods of
measurement. Water Air and Soil Pollution, 182(1-4), 173-185.
Luterbacher, J., Dietrich, D., Xoplaki, E., Grosjean, M. & Wanner, H. (2004).
European seasonal and annual temperature variability, trends, and
extremes since 1500. Science, 303(5663), 1499-1503.
Lye, D. J. (2009). Rooftop runoff as a source of contamination: A review.
Science of the Total Environment, 407(21), 5429-5434.
Lyytimaki, J. & Sipila, M. (2009). Hopping on one leg - The challenge of
ecosystem disservices for urban green management. Urban Forestry
& Urban Greening, 8(4), 309-315.
Maas, J., Verheij, R. A., Groenewegen, P. P., de Vries, S. & Spreeuwenberg,
P. (2006). Green space, urbanity, and health: how strong is the
relation? Journal of Epidemiology and Community Health, 60(7), 587-
592.
MacIvor, J. S. & Lundholm, J. (2011). Performance evaluation of native
plants suited to extensive green roof conditions in a maritime climate.
Ecological Engineering, 37(3), 407-417.
MacKenzie, A. B., Logan, E. M., Cook, G. T. & Pulford, I. D. (1998).
Distributions, inventories and isotopic composition of lead in Pb-210-
dated peat cores from contrasting biogeochemical environments:
Implications for lead mobility. Science of the Total Environment,
223(1), 25-35.
MacKillop, F. (2012). Climatic city: Two centuries of urban planning and
climate science in Manchester (UK) and its region. Cities, 29(4), 244-
251.
MacMillan, G. (2004). York University Rooftop Garden Stormwater Quantity
and Quality Performance Monitoring Report. Greening Rooftops for
Sustainable Communities. Portland, June 2-4, 2004.
Maher, B. A., Moore, C. & Matzka, J. (2008). Spatial variation in vehicle-
derived metal pollution identified by magnetic and elemental analysis
of roadside tree leaves. Atmospheric Environment, 42(2), 364-373.
MAPAC ( 1993). A Breathing Space: Vehicle related air pollution in North
West England. Manchester Area Pollution Advisory Council.
Marschner, H., Romheld, V. & Kissel, M. (1986). Different Strategies In
Higher-Plants In Mobilization and Uptake of Iron. Journal of Plant
Nutrition, 9(3-7), 695-713.
Martens, R., Bass, B. & Alcazar, S. S. (2008). Roof-envelope ratio impact on
green roof energy performance. Urban Ecosystems, 11(4), 399-408.
Mateescu, C., Alecu, G. & Kappel, W. (2008). Electromagnetic field as
environmental factor affecting human health. Revue Roumaine Des
Sciences Techniques-Serie Electrotechnique Et Energetique, 53(2),
113-121.
238
Matzka, J. & Maher, B. A. (1999). Magnetic biomonitoring of roadside tree
leaves: identification of spatial and temporal variations in vehicle-
derived particulates. Atmospheric Environment, 33(28), 4565-4569.
Mayes, W. M., Potter, H. A. B. & Jarvis, A. P. (2010). Inventory of aquatic
contaminant flux arising from historical metal mining in England and
Wales. Science of the Total Environment, 408(17), 3576-3583.
MCC (2009). Greater Manchester Green Roof Programme: Feasibility Study.
Part 1 Report: Scoping & Work Programme Development.:
Manchester City Council.
http://neweconomymanchester.com/stories/840-other_publications.
MCC. (2010). Manchester's population historically [Online]. Available:
http://www.manchester.gov.uk/downloads/download/4220/corporate_r
esearch_and_intelligence_population_publications [Accessed January
2012].
MCC (2011). The Greater Manchester Climate Strategy 2011 - 2020.
Transformation, adaptation & competitive advantage. Accessed online
http://www.manchesterclimate.com/node/2305 on 25/02/2013.
McCarthy, M., Best, M. J. & Betts, A. B. (2010). Including cities in Met Office
Hadley Centre global and regional climate projections. In: Urban
Climate News. Quarterly Newsletter of the IAUC. Issue No. 36 June
2010. Available at www.urban-climate.org.
MCSP (2009). Strategic Development Framework. Manchester City South
Partnership (MCSP), Manchester.
Mendez, C. B., Klenzendorf, J. B., Afshar, B. R., Simmons, M. T., Barrett, M.
E., Kinney, K. A. & Kirisits, M. J. (2011). The effect of roofing material
on the quality of harvested rainwater. Water Research, 45(5), 2049-
2059.
Mentens, J., Raes, D. & Hermy, M. (2006). Green roofs as a tool for solving
the rainwater runoff problem in the urbanized 21st century?
Landscape and Urban Planning, 77(3), 217-226.
Met Office. (2011). http://www.metoffice.gov.uk/climate/uk/2011/april.html
[Online]. [Accessed].
Meyn, S. K. & Oke, T. R. (2009). Heat fluxes through roofs and their
relevance to estimates of urban heat storage. Energy and Buildings,
41(7), 745-752.
Millard, A. (2008). Semi-natural vegetation and its relationship to designated
urban green space at the landscape scale in Leeds, UK. Landscape
Ecology, 23(10), 1231-1241.
Mirzaei, P. A. & Haghighat, F. (2010). Approaches to study Urban Heat
Island - Abilities and limitations. Building and Environment, 45(10),
2192-2201.
Mitchell, G. N. (1990). Natural discoloration of freshwater: chemical and
environmental genesis. Progress in Physical Geography, 14, 317 -
334.
Mitchell, R. & Maher, B. A. (2009). Evaluation and application of biomagnetic
monitoring of traffic-derived particulate pollution. Atmospheric
Environment, 43(13), 2095-2103.
Mitchell, R., Maher, B. A. & Kinnersley, R. (2010). Rates of particulate
pollution deposition onto leaf surfaces: Temporal and inter-species
magnetic analyses. Environmental Pollution, 158(5), 1472-1478.
239
Montello, D. R. & Sutton, P. C. (2006). An introduction to scientific research
methods in geography: Sage publications, California.
Monterusso, M. A., Rowe, D. B. & Rugh, C. L. (2005). Establishment and
persistence of Sedum spp. and native taxa for green roof applications.
Hortscience, 40(2), 391-396.
Monterusso, M. A., Rowe, D. B., Rugh, C. L. & Russell, D. K. (2004). Runoff
Water Quantity and Quality from Green Roof Systems. Acta Hort., 639,
369 - 376.
Moran, A., Hunt, B. & Jennings, G. (2003). A North Carolina field study to
evaluate greenroof runoff quality, runoff quantity, and plant growth.
ASAE Paper 032303., Am. Soc. ofAgric. Eng., St. Joseph,MI.
Morani, A., Nowak, D. J., Hirabayashi, S. & Calfapietra, C. (2011). How to
select the best tree planting locations to enhance air pollution removal
in the MillionTreesNYC initiative. Environmental Pollution, 159(5),
1040-1047.
Murphy, J. M., Sexton, D.M.H., Jenkins, G.J., Boorman, P.M., Booth, B.B.B.,,
Brown, C. C., Clark, R.T., Collins, M., Harris, G.R., Kendon, E.J., Betts,
R.A.,, Brown, S. J., Howard, T. P., Humphrey, K. A., McCarthy, M. P.,
McDonald, R. E., & Stephens, A., Wallace, C., Warren, R., Wilby, R.,
Wood, R. A. (2009). UK Climate Projections Science Report: Climate
change projections. Met Office Hadley Centre, Exeter, UK. 192pp. .
Muxworthy, A. R., Matzka, J., Davila, A. F. & Petersen, N. (2003). Magnetic
signature of daily sampled urban atmospheric particles. Atmospheric
Environment, 37(29), 4163-4169.
NAEI (2012). UK Emissions Projections of Air Quality Pollutants to 2030.
National Atmospheric Emissions Inventory report number
AEA/ENV/R/3337.
NAEI. (2013). National Atmospheric Emissions Inventory (NAEI) UK
emissions data selector [Online]. Available:
http://naei.Defra.gov.uk/data/ [Accessed 28/03/2013].
Nagase, A. & Dunnett, N. (2012). Amount of water runoff from different
vegetation types on extensive green roofs: Effects of plant species,
diversity and plant structure. Landscape and Urban Planning, 104(3-4),
356-363.
Nagase, A. & Tashiro, Y. (2012). Investigation of native plants in Jogasaki
coast for semi-intensive green roofs in Japan. World Green Roof
Congress. Copenhagen, 19-20 September 2012.
Nardini, A., Andri, S. & Crasso, M. (2012). Influence of substrate depth and
vegetation type on temperature and water runoff mitigation by
extensive green roofs: shrubs versus herbaceous plants. Urban
Ecosystems, 15(3), 697-708.
NASA (2011). Global land surface temperature image. NASA Goddard
Institute for Space Studies. http://data.giss.nasa.gov/gistemp/
Accessed 05/05/2013.
Natural England (2009). Green Infrastructre Guidance. Report cat. no.
NE176. Accessed online 12 May 2013
www.naturalengland.org/publications.
240
Neal, M., Neal, C., Wickham, H. & Harman, S. (2007). Determination of
bromide, chloride, fluoride, nitrate and sulphate by ion
chromatography: comparisons of methodologies for rainfall, cloud
water and river waters at the Plynlimon catchments of mid-Wales.
Hydrol. Earth Syst. Sci, 11(1), 294 - 300.
NEGTAP (2001). Transboundary Air Pollution: Acidification, Eutrophication
and Ground-Level Ozone in the UK. Report prepared by the National
Expert Group on Transboundary Air Pollution (NEGTAP) on behalf of
Defra.
NERC (1999). (Natural Environment Research Council - Centre for Ecology
and Hydrology) The Flood Estimation Handbook (FEH) CD.
Niachou, A., Papakonstantinou, K., Santamouris, M., Tsangrassoulis, A. &
Mihalakakou, G. (2001). Analysis of the green roof thermal properties
and investigation of its energy performance. Energy and Buildings,
33(7), 719-729.
Nowak, D. J., Crane, D. E. & Stevens, J. C. (2006). Air pollution removal by
urban trees and shrubs in the United States. Urban Forestry & Urban
Greening, 4(3–4), 115-123.
Oberndorfer, E., Lundholm, J., Bass, B., Coffman, R. R., Doshi, H., Dunnett,
N., Gaffin, S., Kohler, M., Liu, K. K. Y. & Rowe, B. (2007). Green roofs
as urban ecosystems: Ecological structures, functions, and services.
Bioscience, 57(10), 823-833.
Oke, T. R. (1981). Canyon geometry and the nocturnal urban heat island -
comparison of scale model and field observations. Journal of
Climatology, 1(3), 237-&.
Oke, T. R. (1982). The energetic basis of the urban heat island. Quarterly
Journal of the Royal Meteorological Society, 108(455), 1-24.
Oke, T. R. (1989). The micrometeorology of the urban forest. Philosophical
Transactions of the Royal Society of London Series B-Biological
Sciences, 324(1223), 335-349.
Oleson, K. W., Bonan, G. B. & Feddema, J. (2010). Effects of white roofs on
urban temperature in a global climate model. Geophysical Research
Letters, 37.
Onmura, S., Matsumoto, M. & Hokoi, S. (2001). Study on evaporative cooling
effect of roof lawn gardens. Energy and Buildings, 33(7), 653-666.
Pall, P., Aina, T., Stone, D. A., Stott, P. A., Nozawa, T., Hilberts, A. G. J.,
Lohmann, D. & Allen, M. R. (2011). Anthropogenic greenhouse gas
contribution to flood risk in England and Wales in autumn 2000.
Nature, 470(7334), 382-385.
Palla, A., Gnecco, I. & Lanza, L. G. (2012). Compared performance of a
conceptual and a mechanistic hydrologic models of a green roof.
Hydrological Processes, 26(1), 73-84.
Papadopoulos, A. M., Oxizidis, S. & Kyriakis, N. (2003). Perspectives of solar
cooling in view of the developments in the air-conditioning sector.
Renewable and Sustainable Energy Reviews, 7(5), 419-438.
Pascal, M., Corso, M., Chanel, O., Declercq, C., Badaloni, C., Cesaroni, G.,
Henschel, S., Meister, K., Haluza, D., Martin-Olmedo, P., Medina, S. &
Aphekom, g. (2013). Assessing the public health impacts of urban air
pollution in 25 European cities: results of the Aphekom project. The
Science of the total environment, 449, 390-400.
241
Pateman, T. (2011). Rural and urban areas: comparing lives using
rural/urban classifications. Office for National Statistics, London.
Pathan, A. K., Bond, J. & Gaskin, R. E. (2010). Sample preparation for SEM
of plant surfaces. Materials Today, 12, 32-43.
Patil, S. B. & Bodhe, S. K. (2011). Betel leaf area measurement using image
processing. International journal on Computer Science & Engineering,
3(7), 2656 - 2660.
Peachey, C. J., Sinnett, D., Wilkinson, M., Morgan, G. W., Freer-Smith, P. H.
& Hutchings, T. R. (2009). Deposition and solubility of airborne metals
to four plant species grown at varying distances from two heavily
trafficked roads in London. Environmental Pollution, 157(8-9), 2291-
2299.
Petroff, A., Mailliat, A., Amielh, M. & Anselmet, F. (2008). Aerosol dry
deposition on vegetative canopies. Part I: Review of present
knowledge. Atmospheric Environment, 42(16), 3625-3653.
Pizzol, M., Christensen, P., Schmidt, J. & Thomsen, M. (2011). Eco-
toxicological impact of "metals" on the aquatic and terrestrial
ecosystem: A comparison between eight different methodologies for
Life Cycle Impact Assessment (LCIA). Journal of Cleaner Production,
19(6-7), 687-698.
Pomerantz, M., Akbari, H., Berdahl, P., Konopacki, S. J., Taha, H. &
Rosenfeld, A. H. (1999). Reflective surfaces for cooler buildings and
cities. Philosophical Magazine B-Physics of Condensed Matter
Statistical Mechanics Electronic Optical and Magnetic Properties,
79(9), 1457-1476.
Posada, J. M., Sievanen, R., Messier, C., Perttunen, J., Nikinmaa, E. &
Lechowicz, M. J. (2012). Contributions of leaf photosynthetic capacity,
leaf angle and self-shading to the maximization of net photosynthesis
in Acer saccharum: a modelling assessment. Annals of Botany, 110(3),
731-741.
Powe, N. A. & Willis, K. G. (2004). Mortality and morbidity benefits of air
pollution (SO2 and PM10) absorption attributable to woodland in
Britain. Journal of Environmental Management, 70(2), 119-128.
Pretty, J. N., Mason, C. F., Nedwell, D. B. & Hine, R. E. (2002). A preliminary
assessment of the environmental costs of the eutrophication of fresh
waters in England and Wales. Report commissioned byt The
Environment Agency.
Pugh, T. A. M., MacKenzie, A. R., Whyatt, J. D. & Hewitt, C. N. (2012).
Effectiveness of Green Infrastructure for Improvement of Air Quality in
Urban Street Canyons. Environmental Science & Technology, 46(14),
7692-7699.
Raab, T., Hürkamp, K. & Völkel, J. (2005). Detection and quantification of
heavy metal contamination in alluvial soils of historic mining areas by
field portable x-ray fluorescence (FPXRF) analysis. In: International
Conference on Problematic Soils, 2005 Eastern Mediterranean
University, Famagusta, N-Cyprus. 299 -306.
Robertson, D. J. & Taylor, K. G. (2007). Temporal variability of metal
contamination in urban road-deposited sediment in manchester, UK:
Implications for urban pollution monitoring. Water Air and Soil Pollution,
186(1-4), 209-220.
242
Robertson, D. J., Taylor, K. G. & Hoon, S. R. (2003). Geochemical and
mineral magnetic characterisation of urban sediment particulates,
Manchester, UK. Applied Geochemistry, 18(2), 269-282.
Robine, J.-M., Cheung, S. L. K., Le Roy, S., Van Oyen, H., Griffiths, C.,
Michel, J.-P. & Herrmann, F. R. (2008). Death toll exceeded 70,000 in
Europe during the summer of 2003. Comptes Rendus Biologies,
331(2), 171-178.
Robinson, S. L. & Lundholm, J. T. (2012). Ecosystem services provided by
urban spontaneous vegetation. Urban Ecosystems, 15(3), 545-557.
Rocher, V., Azimi, S., Gasperi, J., Beuvin, L., Muller, M., Moilleron, R. &
Chebbo, G. (2004). Hydrocarbons and metals in atmospheric
deposition and roof runoff in central Paris. Water Air and Soil Pollution,
159(1-4), 67-86.
Rogelj, J., McCollum, D. L., O'Neill, B. C. & Riahi, K. (2013). 2020 emissions
levels required to limit warming to below 2 degrees C. Nature Climate
Change, 3(4), 405-412.
Rothwell, J. J., Evans, M. G., Daniels, S. M. & Allott, T. E. H. (2008). Peat
soils as a source of lead contamination to upland fluvial systems.
Environmental Pollution, 153(3), 582-589.
Rowe, D. B. (2011). Green roofs as a means of pollution abatement.
Environmental Pollution, 159(8-9), 2100-2110.
Rowe, D. B., Getter, K. L. & Durhman, A. K. (2012). Effect of green roof
media depth on Crassulacean plant succession over seven years.
Landscape and Urban Planning, 104(3-4), 310-319.
Rowe, D. B., Monterusso, M. A. & Rugh, C. L. (2006). Assessment of heat-
expanded slate and fertility requirements in green roof substrates.
Horttechnology, 16(3), 471-477.
Rowell, D. L. (1994). Soil Science: Methods and Applications: Longman
Group UK Ltd., Harlow, Essex.
Sailor, D. J. (2008). A green roof model for building energy simulation
programs. Energy and Buildings, 40(8), 1466-1478.
Saiz, S., Kennedy, C., Bass, B. & Pressnail, K. (2006). Comparative life cycle
assessment of standard and green roofs. Environmental Science &
Technology, 40(13), 4312-4316.
Salverda, A. P. & Dane, J. H. (1993). An examination of the Guelph
Permeameter for measuring the soils hydraulic properties. Geoderma,
57(4), 405-421.
Sanderson, M. (2010). Changes in the frequency of extreme rainfall events
for selected towns and cities. For Ofwat, produced by the Met Office.
Santamouris, M., Pavlou, C., Doukas, P., Mihalakakou, G., Synnefa, A.,
Hatzibiros, A. & Patargias, P. (2007). Investigating and analysing the
energy and environmental performance of an experimental green roof
system installed in a nursery school building in Athens, Greece.
Energy, 32(9), 1781-1788.
Saxton, K. E., Rawls, W. J., Romberger, J. S. & Papendick, R. I. (1986).
Estimating Generalized Soil-Water Characteristics from Texture. Soil
Science Society of America Journal, 50(4), 1031-1036.
Scholz-Barth, K. (2001). Green roofs: stormwater management from the top
down. Environmental Design and Construction Magazine. San
Francisco, California: BNP Media.
243
Schrader, S. & Boning, M. (2006). Soil formation on green roofs and its
contribution to urban biodiversity with emphasis on Collembolans.
Pedobiologia, 50(4), 347-356.
Schroll, E., Lambrinos, J., Righetti, T. & Sandrock, D. (2011). The role of
vegetation in regulating stormwater runoff from green roofs in a winter
rainfall climate. Ecological Engineering, 37(4), 595-600.
Schuster, A. (1893). Comparisons of Observations of Temperature taken at
two stations in Manchester from December 4th, 1892, to January 28th,
1893 by J. Wilson. Communicated by Arthur Schuster, Ph.D., F.R.S.,
F.R.A.S. Memoirs and Proceedings of the Manchester Literary and
Philosophical Society, 4(7), 160 - 168.
Sfakianaki, A., Pagalou, E., Pavlou, K., Santamouris, M. & Assimakopoulos,
M. N. (2009). Theoretical and experimental analysis of the thermal
behaviour of a green roof system installed in two residential buildings
in Athens, Greece. International Journal of Energy Research, 33(12),
1059-1069.
Shashua-Bar, L., Pearlmutter, D. & Erell, E. (2009). The cooling efficiency of
urban landscape strategies in a hot dry climate. Landscape and Urban
Planning, 92(3-4), 179-186.
Shaw, R., Colley, M. & Connell, R. (2007). Climate change adaptation by
design: a guide for sustainable communities TCPA (Town and
Country Planning Association), London.
Simmons, G., Hope, V., Lewis, G., Whitmore, J. & Gao, W. (2001).
Contamination of potable roof-collected rainwater in Auckland, New
Zealand. Water Research, 35(6), 1518-1524.
Simmons, M. T., Gardiner, B., Windhager, S. & Tinsley, J. (2008). Green
roofs are not created equal: the hydrologic and thermal performance
of six different extensive green roofs and reflective and non-reflective
roofs in a sub-tropical climate. Urban Ecosystems, 11(4), 339-348.
Sivak, M. (2009). Potential energy demand for cooling in the 50 largest
metropolitan areas of the world: Implications for developing countries.
Energy Policy, 37(4), 1382-1384.
Smith, C. & Lawson, N. (2012). Identifying extreme event climate thresholds
for greater Manchester, UK: examining the past to prepare for the
future. Meteorological Applications, 19(1), 26-35.
Smith, C., Lindley, S. & Levermore, G. (2009). Estimating spatial and
temporal patterns of urban anthropogenic heat fluxes for UK cities: the
case of Manchester. Theoretical and Applied Climatology, 98(1-2), 19-
35.
Smith, C. L., Webb, A., Levermore, G. J., Lindley, S. J. & Beswick, K. (2011).
Fine-scale spatial temperature patterns across a UK conurbation.
Climatic Change, 109(3-4), 269-286.
Smith, W. H. (1990). Air Pollution and Forests: Springer, New York.
Solecki, W. D., Rosenzweig, C., Parshall, L., Pope, G., Clark, M., Cox, J. &
Wiencke, M. (2005). Mitigation of the heat island effect in urban New
Jersey. Global Environmental Change Part B: Environmental Hazards,
6(1), 39-49.
Souch, C. & Grimmond, S. (2006). Applied climatology: urban climate.
Progress in Physical Geography, 30(2), 270-279.
244
Speak, A. F., Rothwell, J. J., Lindley, S. J. & Smith, C. L. (2012). Urban
particulate pollution reduction by four species of green roof vegetation
in a UK city. Atmospheric Environment, 61(0), 283-293.
Speak, A. F., Rothwell, J. J., Lindley, S. J. & Smith, C. L. (2013). Rainwater
runoff retention on an aged intensive green roof. Science of the Total
Environment 461-462: 28 - 38.
Spolek, G. (2008). Performance monitoring of three ecoroofs in Portland,
Oregon. Urban Ecosystems, 11(4), 349-359.
Sposito, G. (1989). The Chemistry of Soils: Oxford University press, New
York, NY.
Stabentheiner, E., Zankel, A. & Polt, P. (2010). Environmental scanning
electron microscopy (ESEM)-a versatile tool in studying plants.
Protoplasma, 246(1-4), 89-99.
Stang, A., Hawthorne, C. (2005). The Green House: New directions in
sustainable architecture. New York: Princeton Architectural Press.
Stepniewska, Z. & Bucior, K. (2001). Chromium contamination of soils,
waters, and plants in the vicinity of a tannery waste lagoon.
Environmental geochemistry and health, 23(3), 241-245.
Stewart, I. D. & Oke, T. R. (2012). Local climate zones for urban temperature
studies. Bulletin of the American Meteorological Society, 93, 1879 -
1900.
Stone, B. (2012). The City and the Coming Climate. Cambridge: Cambridge
University Press.
Stott, P. A., Jones, G. S., Lowe, J. A., Thorne, P., Durman, C., Johns, T. C. &
Thelen, J. C. (2006). Transient climate simulations with the HadGEM1
climate model: Causes of past warming and future climate change.
Journal of Climate, 19(12), 2763-2782.
Stovin, V., Vesuviano, G. & Kasmin, H. (2012). The hydrological performance
of a green roof test bed under UK climatic conditions. Journal of
Hydrology, 414–415(0), 148-161.
Sumner, M. E. (2000). Handbook of Soil Science: CRC Press, Florida.
Synnefa, A., Dandou, A., Santamouris, M., Tombrou, M. & Soulakellis, N.
(2008). On the Use of Cool Materials as a Heat Island Mitigation
Strategy. Journal of Applied Meteorology and Climatology, 47(11),
2846-2856.
Synnefa, A., Santamouris, M. & Apostolakis, K. (2007). On the development,
optical properties and thermal performance of cool colored coatings
for the urban environment. Solar Energy, 81(4), 488-497.
Szonyi, M., Sagnotti, L. & Hirt, A. M. (2008). A refined biomonitoring study of
airborne particulate matter pollution in Rome, with magnetic
measurements on Quercus Ilex tree leaves. Geophysical Journal
International, 173(1), 127-141.
Taha, H. (1996). Modeling impacts of increased urban vegetation on ozone
air quality in the South Coast Air Basin. Atmospheric Environment,
30(20), 3423-3430.
Taha, H. (1997). Urban climates and heat islands: Albedo, evapotranspiration,
and anthropogenic heat. Energy and Buildings, 25(2), 99-103.
Taha, H. (2004). Heat Islands and Energy. In: Cleveland, C. J. (ed.)
Encyclopedia of Energy. New York: Elsevier Inc.
245
Taha, H., Akbari, H. & Rosenfeld, A. (1991). Heat-island and oasis effects of
vegetative canopies – micro-meteorological field-measurements.
Theoretical and Applied Climatology, 44(2), 123-138.
Takakura, T., Kitade, S. & Goto, E. (2000). Cooling effect of greenery cover
over a building. Energy and Buildings, 31(1), 1-6.
Takebayashi, H. & Moriyama, M. (2007). Surface heat budget on green roof
and high reflection roof for mitigation of urban heat island. Building
Taylor Associates (2006). Guidelines for monitoring the hydrologic and water
quality performance of green roofs in the Greater Seattle, Washington
region. Prepared for Seattle Office of Sustainability and Environment
by Taylor Associates Inc. .
Taylor, K. G. & Robertson, D. J. (2009). Electron microbeam analysis of
urban road-deposited sediment, Manchester, UK: Improved source
discrimination and metal speciation assessment. Applied
Geochemistry, 24(7), 1261-1269.
Tchepel, O. & Dias, D. (2011). Quantification of health benefits related with
reduction of atmospheric PM10 levels: implementation of population
mobility approach. International Journal of Environmental Health
Research, 21(3), 189-200.
Teemusk, A. & Mander, U. (2007). Rainwater runoff quantity and quality
performance from a greenroof: The effects of short-term events.
Ecological Engineering, 30(3), 271-277.
Teemusk, A. & Mander, U. (2010). Temperature regime of planted roofs
compared with conventional roofing systems. Ecological Engineering,
36(1), 91-95.
Teemusk, A. & Mander, U. (2011). The Influence of Green Roofs on Runoff
Water Quality: A Case Study from Estonia. Water Resources
Management, 25(14), 3699-3713.
Teemusk, A. & Mander, Ü. (2009). Greenroof potential to reduce temperature
fluctuations of a roof membrane: A case study from Estonia. Building
Terzaghi, K., Peck, R. B. & Mesri, G. (1996). Soil Mechanics in Engineering
Practice (3rd Edition ed.): John Wiley & Sons, New York.
The Lancet (2013). Air pollution: Europe's avoidable health risk. Lancet,
381(9870), 876-876.
Thurman, E. M. (1985). Organic Geochemistry of Natural Waters.
Nijhoff/Junk: Dordrecht.
Tijhuis, L., Brattli, B. & Sæther, O. M. (2002). A Geochemical Survey of
Topsoil in the City of Oslo, Norway. Environmental geochemistry and
health, 24(1), 67-94.
Tinytag (2013). Accessed online 05/005/2013 www.geminidataloggers.com
Tipping, E., Vincent, C. D., Lawlor, A. J. & Lofts, S. (2008). Metal
accumulation by stream bryophytes, related to chemical speciation.
Environmental Pollution, 156(3), 936-943.
Türkeri, N., Cem Altun, M. & Göçer, C. (2010). Field monitoring of thermal
performance of a retrofit green roof in Istanbul, Turkey. Paper
presented at the World Green Roof Congress. London 2010.
UK Census (2011) Population data. Office for National Statistics. England
and Wales
246
Ulrich, R. S. (1984). View through a window may influence recovery from
surgery. Science, 224(4647), 420-421.
UN. (2009). Urban and Rural areas wallchart [Online]. Available:
http://www.un.org/esa/population/publications/publications.htm
[Accessed].
UoM. (2012). University of Manchester - The Whitworth Meteorological
Observatory [Online]. Available:
http://www.cas.manchester.ac.uk/restools/whitworth/data/ [Accessed
March 2012].
Van Renterghem, T. & Botteldooren, D. (2008). Numerical evaluation of
sound propagating over green roofs. Journal of Sound and Vibration,
317(3-5), 781-799.
Vangronsveld, J., Herzig, R., Weyens, N., Boulet, J., Adriaensen, K., Ruttens,
A., Thewys, T., Vassilev, A., Meers, E., Nehnevajova, E., van der Lelie,
D. & Mench, M. (2009). Phytoremediation of contaminated soils and
groundwater: lessons from the field. Environmental Science and
Pollution Research, 16(7), 765-794.
VanWoert, N. D., Rowe, D. B., Andresen, J. A., Rugh, C. L. & Xiao, L. (2005).
Watering regime and green roof substrate design affect Sedum plant
growth. Hortscience, 40(3), 659-664.
Vargo, J., Habeeb, D. & Stone, B., Jr. (2013). The importance of land cover
change across urban-rural typologies for climate modeling. Journal of
Environmental Management, 114, 243-252.
Veihmeyer, F. J. & Hendrickson, A. H. (1949). Methods of measuring field
capacity and permanent wilting percentage of soils. Soil Science,
68(1), 75-94.
Vijayaraghavan, K., Joshi, U. M. & Balasubramanian, R. (2012). A field study
to evaluate runoff quality from green roofs. Water Research, 46(4),
1337-1345.
Voyde, E., Fassman, E. & Simcock, R. (2010). Hydrology of an extensive
living roof under sub-tropical climate conditions in Auckland, New
Zealand. Journal of Hydrology, 394(3-4), 384-395.
Walden, J., Oldfield, F., Smith, J. (Eds.). (1999). Environmental Magnetism A
Practical Guide. Quaternary Research Association Technical Guide
No. 6. London: Quaternary Research Association.
Wallage, Z. E., Holden, J. & McDonald, A. T. (2006). Drain blocking: An
effective treatment for reducing dissolved organic carbon loss and
water discolouration in a drained peatland. Science of the Total
Environment, 367(2-3), 811-821.
Webster, P. J., Holland, G. J., Curry, J. A. & Chang, H. R. (2005). Changes in
tropical cyclone number, duration, and intensity in a warming
environment. Science, 309(5742), 1844-1846.
Weng, Q. H. & Yang, S. H. (2006). Urban air pollution patterns, land use, and
thermal landscape: An examination of the linkage using GIS.
Environmental Monitoring and Assessment, 117(1-3), 463-489.
Werthmann, C. (2007). Green Roof - A Case Study. New York: Princeton
Architectural Press.
White, I. & Alarcon, A. (2009). Planning Policy, Sustainable Drainage and
Surface Water Management: A Case Study of Greater Manchester.
Built Environment, 35(4), 516-530.
247
White, P. J. & Broadley, M. R. (2001). Chloride in soils and its uptake and
movement within the plant: A review. Annals of Botany, 88(6), 967-988.
Whitford, V., Ennos, A. R. & Handley, J. F. (2001). "City form and natural
process"--indicators for the ecological performance of urban areas and
their application to Merseyside, UK. Landscape and Urban Planning,
57(2), 91-103.
Wilby, R. L. (2007). A Review of Climate Change Impacts on the Built
Environment. Built Environment, 33(1), 31-45.
Wilby, R. L. (2008). Constructing climate change scenarios of urban heat
island intensity and air quality. Environment and Planning B-Planning
& Design, 35(5), 902-919.
Wilmers, F. (1988). Green for melioration of urban climate. Energy and
Buildings, 11(1-3), 289-299.
WMO (2009). State of the Climate in 2007. World Meteorological
Organization. Accessed online
http://www.wmo.int/pages/publications/showcase/documents/1036_we
b.pdf 05/03/2013.
Wong, N. H., Chen, Y., Ong, C. L. & Sia, A. (2003). Investigation of thermal
benefits of rooftop garden in the tropical environment. Building and
Environment, 38(2), 261-270.
Wong, N. H., Puay Yok, T. & Yu, C. (2007). Study of thermal performance of
extensive rooftop greenery systems in the tropical climate. Building
Worrall, F., Burt, T. P., Jaeban, R. Y., Warburton, J. & Shedden, R. (2002).
Release of dissolved organic carbon from upland peat. Hydrological
Processes, 16(17), 3487-3504.
Yalcin, T. & Yetemen, O. (2009). Local warming of groundwaters caused by
the urban heat island effect in Istanbul, Turkey. Hydrogeology Journal,
17(5), 1247-1255.
Yang, J., Yu, Q. & Gong, P. (2008). Quantifying air pollution removal by
green roofs in Chicago. Atmospheric Environment, 42(31), 7266-7273.
Yin, S., Shen, Z. M., Zhou, P. S., Zou, X. D., Che, S. Q. & Wang, W. H.
(2011). Quantifying air pollution attenuation within urban parks: An
experimental approach in Shanghai, China. Environmental Pollution,
159(8-9), 2155-2163.
Young, T. M., Heeraman, D. A., Sirin, G. & Ashbaugh, L. L. (2002).
Resuspension of soil as a source of airborne lead near industrial
facilities and highways. Environmental Science & Technology, 36(11),
2484-2490.
Zajzon, N., Marton, E., Sipos, P., Kristally, F., Nemeth, T., Kis-Kovacs, V. &
Weiszburg, T. G. (2013). Integrated mineralogical and magnetic study
of magnetic airborne particles from potential pollution sources in
industrial-urban environement. Carpathian Journal of Earth and
Environmental Sciences, 8(1), 179 - 186.
Zobrist, J., Müller, S. R., Ammann, A., Bucheli, T. D., Mottier, V., Ochs, M.,
Schoenenberger, R., Eugster, J. & Boller, M. (2000). Quality of roof
runoff for groundwater infiltration. Water Research, 34(5), 1455-1462.
Zuo, X., Fu, D. & Li, H. (2012). Speciation distribution and mass balance of
copper and zinc in urban rain, sediments, and road runoff.
Environmental Science and Pollution Research, 19(9), 4042-4048.
248
APPENDIX I
Meeting held 21st October 2009 to discuss ideas for project development.
Attendees:
 Sarah Lindley - UoM

 Andrew Speak - UoM
 Corin Bell - MCC
 Neil Jones - MCC
Original objectives set for the work during the initial phase of the project by the
CASE partner Manchester City Council:
Investigate the green roof context of:
 Reduced surface temperatures

 Reduced urban heat islands
 Reduced building energy use (aircon)
 Rainwater runoff reduction
 Rainwater runoff quality impacts
 Fire risk
 Noise reduction
 Air pollution reduction
 Winter insulation
 Biodiversity
Using the five projected green roof installations and the following approaches:
 Collect baseline data before construction and monitor impacts after

construction
 Compare green roofs to adjacent bare roofs
 Use remote sensed data and GIS
249
APPENDIX II
Precinct roof scale diagram showing subsurface drainage catchments
250
Precinct roof scale diagram used for catchment area calculation
251
Manchester Technology Centre roof plan
252
APPENDIX III
University of Manchester Precinct Roof Botanical Survey
Cameron S Crook, July 2012
1. Vascular Plant Species Recorded on the Bare Roof Areas
*Relative
Taxon Abundance Date Spatial Reference
Galium aparine L. R 05.07.2012 53,46816N;2,23471W
Dryopteris filix-mas (L.) Schott R 05.07.2012 53,46834N;2,23483W
Salix cinerea L. R 05.07.2012 53,46808N;2,23489W
Salix caprea L. R 05.07.2012 53,46808N;2,23489W
Urtica dioica L. R 05.07.2012 53,46808N;2,23489W
Buddleja davidii Franch. R 05.07.2012 53,46808N;2,23489W
Conyza canadensis (L.) Cronquist R 05.07.2012 53,46808N;2,23489W
Geranium robertianum L. R 05.07.2012 53,46808N;2,23489W
Cirsium pratense (Huds.) Druce R 05.07.2012 53,46808N;2,23489W
Prunella vulgaris L. R 05.07.2012 53,46787N;2,23658W
Cardamine flexuosa With. R 05.07.2012 53,46796N;2,23570W
Juncus bufonius subsp. foliosus (Desf.)

Arcang. R 05.07.2012 53,46784N;2,23521W
Polygonum persicaria L. R 05.07.2012 53,46793N;2,23500W
Poa annua L. R 05.07.2012 53,46789N;2,23494W
Fraxinus excelsior L. R 05.07.2012 53,46789N;2,23494W
Polygonum arenarium Waldst. & Kit. R 05.07.2012 53,46789N;2,23494W
Cerastium fontanum Baumg. R 05.07.2012 53,46788N;2,23495W
Plantago major L. R 05.07.2012 53,46788N;2,23495W
Sedum acre L. R 05.07.2012 53,46788N;2,23495W
Plantago lanceolata L. R 05.07.2012 53,46788N;2,23495W
Taraxacum officinale (L.) Weber ex

F.H.Wigg. R 05.07.2012 53,46788N;2,23495W
Epilobium ciliatum Raf. R 05.07.2012 53,46788N;2,23495W
Anthoxanthum odoratum L. R 05.07.2012 53,46788N;2,23495W
253
Sonchus oleraceus (L.) L. R 05.07.2012 53,46788N;2,23495W
Galinsoga quadriradiata Cav. R 05.07.2012 53,46788N;2,23495W
2. Vascular Plant Species Recorded within Existing Plot (1) – G2
*Relative
Juncus effusus L. O 05.07.2012 53,46802N;2,23459W
Juncus conglomeratus L. O 05.07.2012 53,46802N;2,23459W
Acer platanoides L. R 05.07.2012 53,46807N;2,23443W
Arrhenatherum elatius (L.) P.Beauv. ex O(LF)

J.Presl & C.Presl. 05.07.2012 53,46807N;2,23443W
Hypericum androsaemum L. O(LF) 05.07.2012 53,46803N;2,23437W
Buddleja davidii Franch. O 05.07.2012 53,46733N;2,23304W
Sonchus oleraceus (L.) L. O 05.07.2012 53,46833N;2,23439W
Holcus mollis L. F(LD) 05.07.2012 53,46833N;2,23439W
Agrostis stolonifera L. F 05.07.2012 53,46837N;2,23429W
Rubus fruticosus G.N.Jones O(LF) 05.07.2012 53,46855N;2,23462W
Hypericum humifusum L. O(LF) 05.07.2012 53,46787N;2,23442W
Veronica serpyllifolia L. R 05.07.2012 53,46780N;2,23437W
Ranunculus acris L. F 05.07.2012 53,46780N;2,23437W
Hedera helix L. O(LF) 05.07.2012 53,46770N;2,23430W
Juncus bufonius subsp. foliosus (Desf.) O

Arcang. 05.07.2012 53,46753N;2,23415W
Sagina procumbens L. O(LF) 05.07.2012 53,46823N;2,23454W
Festuca rubra L. F(LA) 05.07.2012 53,46823N;2,23454W
Ranunculus repens L. F 05.07.2012 53,46823N;2,23454W
Epilobium ciliatum Raf. F(LA) 05.07.2012 53,46823N;2,23454W
Aster novi-belgii L. F(LD) 05.07.2012 53,46823N;2,23454W
Cirsium vulgare (Savi) Airy Shaw O 05.07.2012 53,46823N;2,23454W
Epilobium parviflorum Schreb. F 05.07.2012 53,46823N;2,23454W
Achillea millefolilium E.Mey. R 05.07.2012 53,46823N;2,23454W
254
Prunus avium (L.) L. O 05.07.2012 53,46830N;2,23454W
Anthoxanthum odoratum L. F(LD) 05.07.2012 53,46830N;2,23454W
Trifolium pratense L. F 05.07.2012 53,46830N;2,23454W
Cerastium fontanum Baumg. O 05.07.2012 53,46830N;2,23454W
Rumex crispus L. F(LA) 05.07.2012 53,46830N;2,23454W
Prunella vulgaris L. O 05.07.2012 53,46830N;2,23454W
Crepis capillaris (L.) Wallr. F 05.07.2012 53,46830N;2,23454W
Centaurea nigra L. O 05.07.2012 53,46833N;2,23454W
Trifolium repens L. F 05.07.2012 53,46834N;2,23454W
Plantago lanceolata L. F 05.07.2012 53,46834N;2,23454W
Senecio jacobaea L. F 05.07.2012 53,46834N;2,23454W
Trifolium campestre Schreb. F(LA) 05.07.2012 53,46834N;2,23454W
Hypochaeris radicata L. F 05.07.2012 53,46834N;2,23454W
Centaurium erythraea Rafn O 05.07.2012 53,46837N;2,23456W
Aethusa cyanpium L. R 05.07.2012 53,46780N;2,23437W
3. Vascular Plant Species Recorded within Existing Plot (2) – G1
*Relative
Juncus effusus L. O 05.07.2012 53,46802N;2,23459W
Juncus conglomeratus L. O 05.07.2012 53,46802N;2,23459W
Lotus corniculatus L. O 05.07.2012 53,46799N;2,23454W
Carex flacca Schreb. R 05.07.2012 53,46797N;2,23489W
Acer platanoides L. R 05.07.2012 53,46807N;2,23443W
Arrhenatherum elatius (L.) P.Beauv. ex O(LF)

J.Presl & C.Presl. 05.07.2012 53,46807N;2,23443W
Hypericum androsaemum L. A(LD) 05.07.2012 53,46803N;2,23437W
Buddleja davidii Franch. O (LF) 05.07.2012 53,46733N;2,23304W
Sonchus oleraceus (L.) L. O 05.07.2012 53,46833N;2,23439W
Holcus mollis L. F(LD) 05.07.2012 53,46833N;2,23439W
255
Agrostis stolonifera L. F 05.07.2012 53,46837N;2,23429W
Rubus fruticosus G.N.Jones O(LF) 05.07.2012 53,46855N;2,23462W
Ranunculus acris L. F 05.07.2012 53,46780N;2,23437W
Hedera helix L. O(LF) 05.07.2012 53,46770N;2,23430W
Juncus bufonius subsp. foliosus (Desf.) O

Arcang. 05.07.2012 53,46753N;2,23415W
Sagina procumbens L. R 05.07.2012 53,46823N;2,23454W
Festuca rubra L. F 05.07.2012 53,46823N;2,23454W
Ranunculus repens L. F 05.07.2012 53,46823N;2,23454W
Epilobium ciliatum Raf. F 05.07.2012 53,46823N;2,23454W
Aster novi-belgii L. F(LA) 05.07.2012 53,46823N;2,23454W
Cirsium vulgare (Savi) Airy Shaw O 05.07.2012 53,46823N;2,23454W
Epilobium parviflorum Schreb. O 05.07.2012 53,46823N;2,23454W
Achillea millefolilium E.Mey. R 05.07.2012 53,46823N;2,23454W
Prunus avium (L.) L. R 05.07.2012 53,46830N;2,23454W
Anthoxanthum odoratum L. F(LA) 05.07.2012 53,46830N;2,23454W
Trifolium pratense L. F 05.07.2012 53,46830N;2,23454W
Cerastium fontanum Baumg. O 05.07.2012 53,46830N;2,23454W
Rumex crispus L. F 05.07.2012 53,46830N;2,23454W
Prunella vulgaris L. O 05.07.2012 53,46830N;2,23454W
Crepis capillaris (L.) Wallr. O 05.07.2012 53,46830N;2,23454W
Centaurea nigra L. O 05.07.2012 53,46833N;2,23454W
Trifolium repens L. F 05.07.2012 53,46834N;2,23454W
Plantago lanceolata L. F 05.07.2012 53,46834N;2,23454W
Senecio jacobaea L. F 05.07.2012 53,46834N;2,23454W
Trifolium campestre Schreb. F 05.07.2012 53,46834N;2,23454W
Hypochaeris radicata L. O 05.07.2012 53,46834N;2,23454W
Centaurium erythraea Rafn R 05.07.2012 53,46837N;2,23456W
*Relative Abundance: D-Dominant; A-Abundant; F-Frequent; O-Occasional; R-Rare; L-Locally
256
APPENDIX IV
Regression plots of flow calibration for green roof (top) and bare roof (bottom)
257
Calibration was carried out in the laboratory. Pressure transducers were placed in
each of the two v-notch vessels and the vessels were filled with water up to the point
at which flow starts, h0. A showerhead was fitted to the water tap to minimise any
effects of turbulence on the pressure transducers. A third pressure transducer was
used to measure atmospheric pressure in the laboratory for subsequent adjustment
of readings using the dedicated function within the HOBO software. Different flow
rates were simulated and heights, indicated by increases in water depth over h0,
were obtained from the transducers.
When flow rate is plotted against pressure transducer height, a relationship between
the two can be deduced. A certain amount of scatter is noticeable in the plots,
which is not expected given the accuracy of the v-notch method. This was
potentially due to the small size of the vessels and turbulence arising from the influx
of water. The small size of the vessels means that the transducers were measuring
very small increases in water depth above the datum, h0. The high correlations are
both significant at p<0.01 thus despite the scatter, the quadratic regression
equations were taken as acceptable predictors of flow given transducer height
reading.
The error of the transducers was also considered to be acceptable. The

specifications for the HOBO U20-001-04 data loggers are:
Range: 0-4 m
Water level accuracy: ± 0.3 cm
Resolution: 0.14 cm
258

Quantification Environmental Impacts or Urban Green Roofs

Uploaded by

Copyright:

Available Formats

Quantification Environmental Impacts or Urban Green Roofs

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

Quantification Environmental Impacts or Urban Green Roofs

Uploaded by

Copyright:

Available Formats

Quantification of the Environmental Impacts

of Urban Green Roofs

A thesis submitted to The University of Manchester for the degree of Doctor

Andrew Francis Speak

School of Environment and Development

CHAPTER 2 CURRENT KNOWLEDGE ON CLIMATE 14

CHAPTER 3 CURRENT KNOWLEDGE ON GREEN ROOFS 37

CHAPTER 5 URBAN PARTICULATE POLLUTION 85

REDUCTION BY FOUR SPECIES OF GREEN

CHAPTER 7 RAINWATER RUNOFF RETENTION ON AN 143

AGED INTENSIVE GREEN ROOF

CHAPTER 8 METAL AND NUTRIENT DYNAMICS ON AN 175

CHAPTER 9 CONCLUSION 209

9.1 Introduction 209

WORD COUNT: 67,448

Fig. 1.1 – Location of Greater Manchester (Author created) 7

Table 6.1 – Kruskal-Wallis χ2 values (* = significant at p<0.05, 126

Table 7.1 – Characteristics of the dominant green roof vegetation 152

Table 8.1 – Characteristics of the rainfall-runoff events sampled 186

Urban populations worldwide are expanding rapidly and consequently a large

A site-specific methodology was devised to monitor the rainwater runoff from an

(iii) Further information on the conditions under which disclosures and

Thank you to my supervisors, Dr Sarah Lindley, Dr Claire Smith and Dr

ADWP Antecedent Dry Weather Period

1.1 Research rationale

These pressures and their predicted increase in frequency /severity from

Green roofs broadly consist of a drainage layer underneath a substrate layer

1.2 Involvement of Manchester City Council

This work was funded through a Natural Environment Research Council

When assessing the multi-beneficial nature of green roofs, it is not possible

Specific objectives are:

 To quantify the amounts of PM10 that can be captured by green roof

 To investigate differences in PM10 capture arising from roof location

 To estimate the potential impact of green roofs on PM10 pollution

 To measure the cooling impact of an intensive green roof on the

 To measure and compare the rainwater runoff retention efficiency of

 To draw together results from experimental studies in order to assess

During the data collection period, circumstances allowed for a further

 To investigate how the cooling effect of an intensive green roof can be

1.4 Methodological Framework

Site selection was informed largely by issues of access to the roofs.

Fig. 1.1 – Location of Greater Manchester. (Author created)

Sites were chosen within ‘The Corridor’ – a spatially-defined area of

“The remarkable differences which appear in the temperature records

It is apt, therefore, that Manchester continues to be a well researched city

1.6 Significance of Research

This study, by its empirical multivariate approach, has allowed a number of

 to use magnetic techniques to quantify particle capture by green roof

 to reveal how the location, and plant species composition, of green

 to assess the extent to which damage to the vegetation of an intensive

 to quantify runoff retention capacity of an intensive, aged, in-situ green

 to assess the effects of age on the runoff quality of an intensive green

 to undertake a chemical and physical analysis of green roof soil which

 to assess the multi-benefit aspect of green roofs, within Manchester, in

Fig. 1.3 shows a thesis outline diagram. This thesis is presented in an

Chapter 5 Published in ‘Atmospheric Environment’.

Chapter 6 Published in ‘Urban Climate’.

Chapter 7 Published in ‘Science of the Total Environment’

Chapter 8 Submitted to ‘Environmental Pollution’