social vulnerability, green infrastructure, urbanization and climate
TRANSCRIPT
University of Massachusetts - AmherstScholarWorks@UMass Amherst
Dissertations Dissertations and Theses
9-1-2013
Social vulnerability, green infrastructure,urbanization and climate change-induced flooding:A risk assessment for the Charles River watershed,Massachusetts, USAChingwen ChengUniversity of Massachusetts - Amherst, [email protected]
Follow this and additional works at: http://scholarworks.umass.edu/open_access_dissertations
This Open Access Dissertation is brought to you for free and open access by the Dissertations and Theses at ScholarWorks@UMass Amherst. It hasbeen accepted for inclusion in Dissertations by an authorized administrator of ScholarWorks@UMass Amherst. For more information, please [email protected].
Recommended CitationCheng, Chingwen, "Social vulnerability, green infrastructure, urbanization and climate change-induced flooding: A risk assessment forthe Charles River watershed, Massachusetts, USA" (2013). Dissertations. Paper 781.
SOCIAL VULNERABILITY, GREEN INFRASTRUCTURE, URBANIZATION AND CLIMATE CHANGE‐
INDUCED FLOODING: A RISK ASSESSMENT FOR THE CHARLES RIVER WATERSHED,
MASSACHUSETTS, USA
A Dissertation Presented
by
CHINGWEN CHENG
Submitted to the Graduate School of the
University of Massachusetts Amherst in partial fulfillment
of the requirements for the degree of
DOCTOR OF PHILOSOPHY
September 2013
Regional Planning
© Copyright by Chingwen Cheng 2013
All Rights Reserved
SOCIAL VULNERABILITY, GREEN INFRASTRUCTURE, URBANIZATION AND CLIMATE CHANGE‐
INDUCED FLOODING: A RISK ASSESSMENT FOR THE CHARLES RIVER WATERSHED,
MASSACHUSETTS, USA
A Dissertation Presented
by
CHINGWEN CHENG
Approved as to style and content by:
__________________________________________
Elizabeth A. Brabec, Chair
__________________________________________
Robert L. Ryan, Member
__________________________________________
Qian Yu, Member
__________________________________________
Yi‐Chen Ethan Yang, Member
______________________________________
Elisabeth M. Hamin, Department Head
Department of Landscape Architecture and
Regional Planning
To
My Grandparents
&
My Parents
v
ACKNOWLEDGMENTS
This research is based upon work supported by the National Institute of Food and
Agriculture, the U.S. Department of Agriculture, the Massachusetts Agricultural Experiment
Station, the Department of Environmental Conservation under Project No. MAS009584 and
MAS00971, and the National Science Foundation under Grant No. BCS‐0948984. In addition,
this work is funded by the Water Resources Research Center (grant award year 2012‐2013) at
the University of Massachusetts Amherst. Finally, my PhD study has been generously supported
through several Teaching and Research Assistantships at the Department of Landscape
Architecture and Regional Planning in addition to a Study Abroad Scholarship from the Ministry
of Education in Taiwan.
The completion of my PhD study would not have been possible without the support of
my committee and many people. I would like to express my deepest gratitude to my advisor,
Professor Elizabeth Brabec. I thank her for giving me the opportunity to study and for leading
me into an academic career. Without her persistent encouragement and support, I would not
have started and finished my PhD journey smoothly. I would also like to express my very great
appreciation to Dr. Robert Ryan. I thank him for supporting me throughout my studies, advising
me in research design, providing feedback and comments throughout the process, and
mentoring me in my academic career. Advice given by Dr. Yi‐Chen Ethan Yang was a
tremendous help in building the model and analyzing the data for this research. I thank him for
giving his generous time, helpful insights and advice for raising the quality of the research. Last
but not least in my dissertation committee, I thank Dr. Qian Yu for her helpful guidance in
research methods and constructive advice whenever I needed it.
I would also like to express my deepest appreciation to Dr. Mark Hamin. I thank him for
being on my exam committee and providing generous and insightful feedback in my academic
vi
experiences. I am particularly grateful for the guidance and support given by Dr. Elisabeth
Hamin throughout the PhD program. I would also like to extend my greatest appreciation to all
the faculty and staff in LARP for all the assistance provided throughout my academic
experiences.
I would like to express my sincerest thanks to Dr. Paige Warren. I thank her for giving
me the opportunity and continuous support in the BMA ULTRA‐ex project. I also wish to
acknowledge the BMA ULTRA‐ex team who have provided regular feedback and help during the
development of the work―Dr. Robert Ryan, Dr. Craig Nicolson, Dr. Michael Strohbach, Dr. Jack
Ahern, Dr. Erick Strauss, Rachel Danford, Torrey Wolff and Dr. Andreia Quintas. I would also like
to express my gratitude to Dr. Henry Renski, Dr. Ken‐Hou Lin and Miao‐Shan Yen for their
helpful insights about my research design and methods as well as Dr. Susan Cutter for
generously providing her “SoVI recipe.”
Finally, I am tremendously grateful to my family and all my friends, for their company,
moral support and all the help they gave me when I needed it most. I would like to express my
gratitude to Frank Bryan for his help in meticulously proofreading every word in the dissertation.
I thank my dear family for their unconditional support throughout my years living abroad and
particularly their understanding and encouragement for finishing my PhD studies. I am
particularly grateful for my wonderful parents, Linche Cheng and Suli Tsai, for having set an
example of persistence and endurance in achieving their goals in addition to their passion and
great love of working for the community. I thank them for giving me the opportunity and their
encouragement to seize mine. Lastly, my very special and warm thanks to Róbert Gál for giving
me the strength I needed to have made this dissertation possible.
vii
ABSTRACT
SOCIAL VULNERABILITY, GREEN INFRASTRUCTURE, URBANIZATION AND CLIMATE CHANGE‐
INDUCED FLOODING: A RISK ASSESSMENT FOR THE CHARLES RIVER WATERSHED,
MASSACHUSETTS, USA
SEPTEMBER 2013
CHINGWEN CHENG
B.S., NATIONAL TAIWAN UNIVERSITY
M.L.A., UNIVERSITY OF MICHIGAN ANN ARBOR
PH.D., UNIVERSITY OF MASSACHUSETTS AMHERST
DIRECTED BY: ELIZABETH A. BRABEC
Climate change is projected to increase the intensity and frequency of storm events that
would increase flooding hazards. Urbanization associated with land use and land cover change
has altered hydrological cycles by increasing stormwater runoff, reducing baseflow and
increasing flooding hazards. Combined urbanization and climate change impacts on long‐term
riparian flooding during future growth are likely to affect more socially vulnerable populations.
Growth strategies and green infrastructure are critical planning interventions for minimizing
urbanization impacts and mitigating flooding hazards. Within the social‐ecological systems
planning framework, this empirical research evaluated the effects of planning interventions
(infill development and stormwater detention) through a risk assessment in three studies.
First, a climate sensitivity study using SWAT modeling was conducted for building a long‐
term flooding hazard index (HI) and determining climate change impact scenarios. A Social
Vulnerability Index (SoVI) was constructed using socio‐economic variables and statistical
viii
methods. Subsequently, the long‐term climate change‐induced flooding risk index (RI) was
formulated by multiplying HI and SoVI. Second, growth strategies in four future growth
scenarios developed through the BMA ULTRA‐ex project were evaluated through land use
change input in SWAT modeling and under climate change impact scenarios for the effects on
the risk indices. Third, detention under climate sensitivity study using SWAT modeling was
investigated in relation to long‐term flooding hazard indices.
The results illustrated that increasing temperature decreases HI while increasing
precipitation change and land use change would increase HI. In addition, there is a relationship
between climate change and growth scenarios which illustrates a potential threshold when the
impacts from land use and land cover change diminished under the High impact climate change
scenario. Moreover, spatial analysis revealed no correlation between HI and SoVI in their
current conditions. Nevertheless, the Current Trends scenario has planned to allocate more
people living in the long‐term climate change‐induced flooding risk hotspots. Finally, the results
of using 3% of the watershed area currently available for detention in the model revealed that a
projected range of 0 to 8% watershed area would be required to mitigate climate change‐
induced flooding hazards to the current climate conditions.
This research has demonstrated the value of using empirical study on a local scale in
order to understand the place‐based and watershed‐specific flooding risks under linked social‐
ecological dynamics. The outcomes of evaluating planning interventions are critical to inform
policy‐makers and practitioners for setting climate change parameters in seeking innovations in
planning policy and practices through a transdisciplinary participatory planning process.
Subsequently, communities are able to set priorities for allocating resources in order to enhance
people’s livelihoods and invest in green infrastructure for building communities toward
resilience and sustainability.
ix
TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS ...................................................................................................................... v
ABSTRACT ........................................................................................................................................ vii
LIST OF TABLES ............................................................................................................................... xiv
LIST OF FIGURES ............................................................................................................................. xvi
CHAPTER
1. INTRODUCTION ............................................................................................................................ 1
1.1 Planning in wicked social‐ecological systems .............................................................. 1
1.2 Social‐ecological systems research .............................................................................. 3
1.3 Research issues and background ................................................................................. 5
1.4 Research questions and hypotheses ............................................................................ 9
1.5 Social‐ecological systems planning framework and research ................................... 10
1.6 Organization of chapters ............................................................................................ 13
2. BACKGROUND ............................................................................................................................ 15
2.1 Planning in social‐ecological systems ........................................................................ 15
2.1.1 Resilience and climate change ................................................................... 15
2.1.2 Watershed as a social‐ecological planning unit ......................................... 18
2.2 Hazards, vulnerability, risk assessment ..................................................................... 19
2.2.1 Risks and disasters ..................................................................................... 19
2.2.2 Vulnerability and resilience ....................................................................... 20
2.2.3 Vulnerability assessment models .............................................................. 21
2.3 Urbanization impacts in social‐ecological systems .................................................... 24
2.3.1 Urban hydrology overview ......................................................................... 24
x
2.3.3 Urbanization impacts on streamflow and flooding ................................... 26
2.4 Climate change impacts on hydrology in New England ............................................. 30
2.5 Study context and study area .................................................................................... 30
3. LONG‐TERM CLIMATE CHANGE‐INDUCED FLOODING RISK ASSESSMENT ................................ 34
3.1 Introduction ............................................................................................................... 34
3.2 Background ................................................................................................................ 35
3.2.1 Flooding losses in the Boston Metropolitan Area ..................................... 35
3.2.2 Demographics and status of equity in the Boston Metropolitan
Area ......................................................................................................... 38
3.2.3 Social vulnerability and its measures ......................................................... 43
3.3 Methodology .............................................................................................................. 45
3.3.1 Risk assessment framework ....................................................................... 45
3.3.2 SWAT Modeling.......................................................................................... 47
Model description ................................................................................... 48
Model input and data source ................................................................. 49
Model calibration and validation ............................................................ 53
3.3.3 Climate change assessment methods ........................................................ 54
Climate sensitivity tests .......................................................................... 56
Climate change impact scenarios ........................................................... 61
3.3.4 Social Vulnerability Index (SoVI) ................................................................ 64
3.3.5 Long‐term climate change‐induced flooding risk index (RI) ..................... 68
3.4 Results ........................................................................................................................ 69
3.4.1 HI and SoVI spatial overlay ........................................................................ 69
3.4.2 Long‐term climate change‐induced flooding risk assessment ................... 70
xi
3.5 Discussion ................................................................................................................... 72
3.5.1 Climate change impacts on the HI ............................................................. 72
3.5.2 HI and SoVI HotSpots spatial mismatch ..................................................... 73
3.5.3 Growing risks in suburbs ............................................................................ 75
3.5.4 Limitation of this study .............................................................................. 76
3.6 Conclusion .................................................................................................................. 78
4. FUTURE LONG‐TERM CLIMATE CHANGE‐INDUCED FLOODING RISK ASSESSMENT
UNDER GROWTH SCENARIOS AND CLIMATE CHANGE SCENARIOS ................................. 80
4.1 Introduction ............................................................................................................... 80
4.2 Background ................................................................................................................ 82
4.2.1 MAPC community types and MetroFuture ................................................ 82
4.2.2 BMA ULTRA‐ex growth scenarios 2030 ..................................................... 84
4.2.3 Population and growth scenarios .............................................................. 85
4.3 Methodology .............................................................................................................. 88
4.3.1 Future long‐term climate change‐induced flooding risk
assessment framework ........................................................................... 88
4.3.2 Projected population to land use conversions .......................................... 89
Step 1: Identify developable lands ......................................................... 90
Step 2: Define housing density types ..................................................... 93
Step 3: Inner Core communities development assumptions ................. 94
Step 4: Non‐Core communities development assumptions ................... 95
Step 5: 2030 land use decision rules ...................................................... 97
4.3.3 SWAT 2030 land use .................................................................................. 99
4.4 Results ...................................................................................................................... 101
xii
4.4.1 Land use change ....................................................................................... 101
4.4.2 Future growth scenario impacts .............................................................. 103
4.5 Discussion ................................................................................................................. 109
4.5.1 Growth strategy and urban landscape change ........................................ 109
4.5.2 Land use and land cover change .............................................................. 110
4.5.3 Growing population in RI HotSpots ......................................................... 112
4.5.4 Climate change vs. urbanization impacts ................................................ 114
4.6 Conclusion ................................................................................................................ 115
5. EFFECTS OF DETENTION FOR MITIGATING CLIMATE CHANGE‐INDUCED LONG‐TERM
FLOODING HAZARDS ....................................................................................................... 117
5.1 Introduction ............................................................................................................. 117
5.2 Background .............................................................................................................. 118
5.2.1 Planning origins for green infrastructure in hazard mitigation ............... 118
5.2.2 Green infrastructure practices for flooding mitigation ........................... 120
5.3 Methods ................................................................................................................... 121
5.4 Results ...................................................................................................................... 124
5.5 Discussion ................................................................................................................. 128
5.5.1 Detention effects under climate change impacts .................................... 128
5.5.2 Detention effects on reducing long‐term flooding risks .......................... 128
5.5.3 Adaptive land use planning ...................................................................... 129
5.5.4 Resilient stormwater management design .............................................. 131
5.6 Conclusion ................................................................................................................ 133
6. CONCLUSION ............................................................................................................................ 134
6.1 Contribution to the field in planning ....................................................................... 134
xiii
6.1.1 On‐site detention innovations ................................................................. 136
6.1.2 Remove people from the risk equation ................................................... 137
6.1.3 Reduce social vulnerability in the risk equation ...................................... 137
6.1.4 Green equity ............................................................................................ 138
6.2 Future directions ...................................................................................................... 138
6.3 Final remarks ............................................................................................................ 139
BIBLIOGRAPHY ............................................................................................................................. 141
xiv
LIST OF TABLES
Table Page
3.1 Monetary loss in property damages (in 2011 dollars) in Massachusetts and the
Boston Metropolitan Area from 1953 to 2011…………………………………………………………… 36
3.2 Concepts and indicators with positive effects on social vulnerability and
corresponding Social Vulnarbility Index variables……………………………………………………… 44
3.3 SWAT 2005 land use categories………………………………………………………………………………… 52
3.4 Descriptive statistics of Social Vulnerability Index variables……………………………………… 65
3.5 Principal component analysis of Social Vulnerability Index ………………………………………. 67
3.6 Summary of descriptive statistics of long‐term climate change‐induced flooding risk
index (RI) among climate change impact scenarios…………………………………………………… 71
4.1 Summary of growth strategies applied in the growth scenarios for the Boston
Metropolitan Area through 2030………………………………………………………………………………. 85
4.2 Summary of projected population and housing units increase from 2000 to 2030
between Inner Core and non‐core communities among growth scenarios ………………. 87
4.3 MassGIS 2005 land use categorized for developable versus non‐developable land
uses in Inner Core and non‐core communities………………………………………………………….. 91
4.4 2030 land use categories and descriptions……………………………………………………………….. 98
4.5 2005 land use to 2030 land use conversion rules …………………………………………………….. 99
4.6 SWAT 2030 land use designation based on 2030 land use ……………………………………….. 100
4.7 Land use change areas from 2005 to 2030 in growth scenarios ……………………………….. 102
4.8 Summary of risk assessment indices descriptive statistics………………………………………… 104
4.9 Infill land change analyses among growth scenarios ………………………………………………… 110
xv
4.10 Mean and standard deviation of percent imperviousness in each SWAT 2005 land
use category for the Boston Metropolitan Area ……………………………………………………….. 111
4.11 Projected impervious increase from total land use change areas from 2005 to 2030
among growth scenarios ……………………………………………………………………………………....... 112
5.1 Pearson’s correlation coefficient between long‐term flooding hazard index (HI) and
the fraction of detention area modeled in sub‐watershed in climate change
conditions…………………………………………………………………………………………………………………. 126
5.2 Summary of regression coefficients and percent detention area required for
reaching hazard mitigation goals ……………………………………………………………………………… 127
6.1 Summary of research hypotheses and findings ..……………………………………………………... 135
xvi
LIST OF FIGURES
Figure Page
1.1 A conceptual framework of a social‐ecological system in urban ecology .....……………. 5
1.2 A social‐ecological systems planning framework ……………………………………………………... 12
1.3 Research study in the social‐ecological systems planning framework ……………………... 13
2.1 Water balance in an urbanized watershed ……………………………………………………………….. 25
2.2 Flood frequency curves in highly urbanized areas …………………………………………………….. 27
2.3 Number of overbank flooding events in a year in relation to urbanization ……………….. 28
2.4 Location of the Charles River Watershed study area ………………………………………………... 32
2.5 Current land cover types in percentages of the Charles River Watershed area ………… 33
3.1 Monetary losses compared between Boston Metropolitan Area (BMA) and Non‐
BMA in Massachusetts from 1953 to 2011 ……………………………………………………………….. 38
3.2 Difference in percentage of foreign‐born residents from 2000 to 2009 in BMA and
the Charles River Watershed ………………………………………………………………………………….... 39
3.3 Poverty rate from 2005 to 2009 in BMA and the Charles River Watershed .…………….. 40
3.4 The number of people met the Environmental Justice criteria in Massachusetts, the
Boston Metropolitan Area and the Charles River Watershed …………………………………… 41
3.5 Location of the Environmental Justice populations in the Boston Metropolitan Area
and the Charles River Watershed ……………………………………………………………………………… 42
3.6 Risk assessment framework for the long‐term climate change‐induced flooding
hazards and social vulnerability in the Charles River Watershed ………………………………. 46
3.7 Location and size of the subbasins compared to the census tract patterns in the
Charles River Watershed ………………………………………………………………………………………….. 51
3.8 SWAT calibration and validation hydrograph in the Charles River Watershed …………. 54
xvii
3.9 Histogram of the number of climate sensitivity tests in a range of long‐term flooding
hazard indices ……………………………………………………………………………………………………........ 57
3.10 Climate sensitivity tests results in long‐term flooding hazard index (HI) controlled by
mean temperature change ………………………………………………………………………………………. 58
3.11 Climate sensitivity tests results in long‐term flooding hazard index (HI) controlled by
precipitation variation change …………………………………………………………………………………. 59
3.12 Climate sensitivity tests results in long‐term flooding hazard index (HI) controlled by
mean precipitation change ………………………………………………………………………………………. 60
3.13 Projected annual mean temperature change under IPCC scenarios ………………………… 61
3.14 Annual average precipitation change from downscaled GCM models……………………… 62
3.15 Climate change impact scenarios …………………………………………………………………………….. 63
3.16 Spatial overlay between the long‐term flooding hazard index (HI) and the Social
Vulnerability Index (SoVI) …………………………………………………………………………………………. 70
3.17 Maps of the long‐term flooding risk index (RI) among climate change impact
scenarios ………………………………………………………………………………………………………............. 72
3.18 The Getis‐Ord Gi*Hot Spots spatial analyses revealed spatial mismatch between (a)
the long‐term flooding hazard index and (b) the Social Vulnerability Index ……………… 74
4.1 Charles River Watershed study area and the Metropolitan Area Planning Council
(MAPC) community types in the Boston Metropolitan Area …………………………….......... 83
4.2 Future long‐term climate change‐induced flooding risk assessment framework
integrating growth scenarios 2030 and climate change impact scenarios ………………… 89
4.3 Housing density decision tree for Inner Core communities ………………………………….…… 95
4.4 Housing density decision tree for non‐core communities ……………………………………….. 97
xviii
4.5 Percentage of watershed area change from lower density in 2005 land use to higher
density among growth scenarios ……………………………………………………………………………….103
4.6 The long‐term climate change‐induced flooding risk index (RI) mean and standard
deviation among climate change impact and growth scenarios ……………………………….. 105
4.7 Long‐term flooding hazard index (HI) density distributions among growth and
climate change impact scenarios ……………………………………………………………………………… 106
4.8 Maps of long‐term climate change‐induced flooding risk index (RI) in current climate
conditions among growth scenarios …………………………………………………………………………. 107
4.9 Maps of the long‐term climate change‐induced risk index (RI) in the Current Trends
growth scenario among climate change impact scenarios ……………………………………….. 108
4.10 Maps of allocation of projected increased population in growth scenarios overlaid
with current conditions of long‐term flooding risk index (RI) HotSpots ……………………. 114
5.1 Research framework for using green infrastructure for mitigating climate change
impacts on long‐term flooding hazards ..………………………………………………….……………... 122
5.2 Significant effects of detention in mitigating long‐term flooding hazards in ten
climate change conditions ..………………………………………………………………………………....... 127
5.3 Detention effects on long‐term climate change‐induced flooding risk index (RI) in
climate change impact scenarios compared to RI HotSpots in current climate
conditions …..…………………………………………………………………………………………………………… 129
1
CHAPTER 1
INTRODUCTION
“Man and nature are the twin agents of the perennial revolution which shapes
and reshapes the face of the earth and the character of man’s activities.”
(Gutkind 1952)
1.1 Planning in wicked social‐ecological systems
Planning is to re‐solve wicked problems inherent in society in which no optimum
solution can be found as a result of ill‐defined public policy issues (Rittel and Webber 1973),
regardless of whether the questions concern the location of a wind turbine, the adjustment of a
tax rate, or gun control. Dealing with planning issues in the coupled human and natural systems
adds another layer of complexity and uncertainty. Coping with the dynamics of change requires
integrated systems thinking and interdisciplinary conceptual frameworks to untangle wicked
planning issues such as sustainability and climate change. Resilience theory in the linked social‐
ecological systems has provided an alternative lens through which to view planning framework
for both problem‐setting and problem‐solving (Wilkinson 2012).
Human is part of the ecosystem. Like other animals, human shapes and is being shaped
by the physical environment; yet unlike other animals, human has the capacity to build beyond
an ecosystem’s capacity to sustain, which in turn is unable to sustain Homo sapiens. The
concept of the coupled human and natural systems (Liu et al. 2007) or the linked social‐
ecological systems (Folke 2006) indicates that ecology and society are intricately connected as
one system, much like a water molecule (H2O) which bonds hydrogen (H2) and oxygen (O2). Just
as the properties of water are entirely different from hydrogen or oxygen, the linked social‐
ecological systems are different from their component parts. We can study nature and society
2
as separate objects. Examples include a biologist studying the mechanism in a mantis shrimp’s
raptorial appendages that exhibit a powerful movement with a speed of 23 meters per second
under the water, a sociologist studying to what degree women receive less pay than men, and
economists arguing whether a threshold exists in national debts that indicate the needs for
austerity measures in a government. On the other hand, complex problems such as setting a
sustainable yield amount for wood harvesting ―an issue interlinked with the tree’s biological
growth in ecosystems, economic demand and supply curve, cultural values of environmental
ethics, social structures governing the management of land―that involve interactions between
people and ecosystems have revealed wicked planning issues in the intricately linked social‐
ecological systems. Sustainable development that planners struggle with for decades is not an
ecological problem alone, and not an economic problem alone, not a social problem alone; it is a
convergence of all three aspects integrated in the linked social‐ecological system (Holling 2001).
Climate change is another complex planning issue in the multi‐scale social‐ecological
system. Induced by anthropogenic greenhouse gases (IPCC 2007), climate change has not only
threatened habitats for wildlife such as polar bears but also livelihoods for Homo sapiens. The
increased sea levels would likely force people who live in coastal low‐lying areas and island
nations to migrate to higher areas or other countries or otherwise require necessary adaptation
measures that are likely unaffordable for the most impacted low‐income island nations such as
Indonesia, the Philippines, and other Pacific Island countries. The changing rainfall patterns and
early glacial and snow melt, in addition to sea water intrusion, would worsen water accessibility
and affect rain‐fed dependent agriculture such as the farms in northeastern Brazil. It would also
increase the burdens of women in households looking for alternative water sources in some
African countries. The increased range of uncertainty in weather‐related disasters would also
increase intensity and frequency of extreme weather such as heat waves, extreme cold‐days,
3
droughts, storms, hurricanes, tornados, and severe winter weather and cause major casualties
and property damage, particularly in already risk‐prone areas in low‐ and middle‐income
countries as well as in developed countries (Leary 2008). The scale and complexity in global
social‐ecological systems involving international politics are particularly wicked in order to cope
with anticipated environmental, social and economic impacts from climate change.
1.2 Social‐ecological systems research
The science of integrating ecology and sociology can be traced back to the 1920s at the
Chicago School of Sociology when Park and Burgess studied people’s interaction with urban
landscapes. Park introduced three ecological theories―competition, niche partitioning and
succession―to understand the drivers for spatial differentiation of people in cities (Humphries
2012). Since then, multiple disciplines have evolved to study the interactions between people
and ecosystems such as human ecology, urban ecology, ecological economics, landscape
architecture, urban design and regional planning. In order to have a comprehensive
understanding of linked systems, it is necessary for a synthesis and integration of several
different conceptual frames (Costanza et al. 1993).
A social‐ecological system can be illustrated as the dynamics between the biological and
human ecosystems integrating social, biological, physical, and built components interacting with
one another (Cadenasso and Pickett 2013)(Figure 1.1). Urban ecologists have studied social‐
ecological systems intensively with two approaches known as “ecology of cities” and “ecology in
cities” (Pickett et al. 2001). The “ecology in cities” provides lens in analyzing how human
settlement and activities affect the physical environment. The impact of urbanization on natural
ecosystem and environment can then be described and quantified. For example, impervious
4
surfaces derived from the built development have multiple impacts on the functions of
ecosystems in cities, including alteration of the natural water cycle with increased surface runoff
and reduced groundwater infiltration and recharge, as well as urban heat island effect that is
associated with alteration of micro climate (Pickett et al. 2001).
The “ecology of cities” provides a system‐oriented approach in examining cities as a
dynamic ecosystem. Classical theory of ecology considers that an ecosystem is a closed, self‐
contained and equilibrium system. Contemporary ecosystem theory, on the other hand,
proposes that an ecosystem is non‐equilibrium and a connected open system that can withstand
disturbances to the degree below the threshold in which the system would change to another
state (pick 2001, 2004, cade 2004, walker 2006). The contemporary school of thoughts opened
a window for a more comprehensive theory in connecting ecological, social, and physical
dimensions in urban ecology studies (Cadenasso and Pickett 2013).
In 1997 in the United States, the National Science Foundation (NSF)made the first
attempt to fund long‐term ecological research (LTER) projects in two cities—Phoenix and
Baltimore. Continuing the momentum, the Urban Long‐Term Research Areas program was
initiated with exploratory grants (ULTRA‐Ex) to start long‐term social‐ecological research in over
20 cities (Humphries 2012). The Baltimore Ecosystem Study, for example, aimed to serve
Baltimore metropolitan area, including Baltimore, Baltimore County and six other counties in
Maryland. The first phase of the project has focused on the Gwynns Falls Watershed, including
Baltimore and its region with a population of about 356,000 in 1990 (BES 2010). The goal of the
study was to understand the patch dynamics—patterns and changes in hierarchical spatial
heterogeneity and the effects of this heterogeneity on ecosystem function (Pickett and
Cadenasso 2006). The patch dynamics serve as a tool that integrates biophysical and social
5
components of the human ecosystem. In addition, the directional water flow, along with its
ability to transport materials and exchange energy, plays a key role in integrating multiple
functions of a watershed (Pickett and Cadenasso 2006). Therefore, a watershed is a logical
study unit for linked social‐ecological systems research.
Figure 1. 1 A conceptual framework of a social‐ecological system in urban ecology
(Cadenasso and Pickett 2013)
1.3 Research issues and background
“The time is ripe for some form of precautionary planning which considers
vulnerability of the population as the real cause of disaster – a vulnerability that
is induced by social‐economic conditions that can be modified by man, and is not
just an act of God. Precautionary planning must commence with the removal of
concepts of naturalness from natural disasters.” (O'Keefe, Westgate, and
Wisner 1976) p 567
Physical Biological
Built Social
The biological ecosystem concept
The human ecosystem concept
6
The tragedy of 2005 Hurricane Katrina was a human rather than a natural disaster.
Natural hazards have become intricately intertwined with technological and man‐made hazards
(Colten 2006; Walker and Burningham 2011) in the dynamically linked social‐ecological systems.
Flood is omnipresent in almost every city in the United States (White and Haas 1975; Platt
1999). Among the nation’s economic losses caused by natural disasters in the past fifty years,
more than half are due to flooding‐related damages—as much as 291 billion in 2009 dollars (Gall
et al. 2011). Under the impact of climate change, more frequent and intense floods as a result
of the increased intensity and duration of storm events are likely to affect the Northeast region
(IPCC 2007; Rock et al. 2001). The increased frequency of extreme storm events in recent
decades—Superstorm Sandy in 2012, Hurricane Irene in 2011 and serious floods in 2011, 2010,
and 2005—has coincided with climate change projections in the Northeast. The phenomenon
that the increased natural hazards that are likely to be aggravated by climate change is referred
to as “climate change‐induced” hazards in this study.
The Boston Metropolitan Area, consisting of 101 communities with a population of 3.16
million, is expected to grow about 10% by 2030 (MAPC 2009). Currently, the population is
aging, becoming more diverse in its younger cohort, increasing in inequality in socio‐economic
status, and increasing in needs for support for minority groups and immigrants. The current
demographics and socio‐economic structure exemplify with some of the key concepts of social
vulnerability. Socially vulnerable populations are place‐specific and include not only people who
are poor or have lower social status, but also women, children, the elderly, people who have
insufficient access to natural resources and capital investment, those who live under insecure
political regimes, and those who have at best a limited voice in the political realm (Bartlett 2008;
Douglas et al. 2008; Hardoy and Pandiella 2009; Maantay and Maroko 2009). Moreover, race
and immigration status, in addition to poverty, have been shown to be positively related to
7
social vulnerability in the United States (Colten 2006). Therefore, the root cause of social
vulnerability resides in social systems created by society (Beck 1992)and is inherent in the
process of urbanization.
As one of the NSF‐funded ULTRA‐ex projects, the BMA ULTRA‐ex project aims to
understand the socio‐economic (e.g., land use policy, population change, investment, social
capital) and bio‐physical (e.g., climate change) drivers that influence social‐ecological processes
(e.g., land use and land cover change, urban greening development and stewardship) that
interact within ecosystems and have impacts on social‐ecological outcomes (e.g., biodiversity,
water quality and quantity, public health) (BMA‐ULTRA).
Growth strategy, as a planning intervention, differentiates compact development and
redeveloping existing built areas (i.e., Infill development) from suburban sprawl that targets
clearing agriculture, forest, wetlands and other open space for new development (i.e., greenfill
development). To investigate land use and land cover change driven by population growth, the
BMA ULTRA‐ex team has developed four different growth scenarios that vary in growth strategy
in allocating projected population between the inner cities and suburbs.
Current Trends follows a suburban sprawl pattern. MetroFuture aligns with policies set
forth by the Metropolitan Planning Area Council (MAPC) for the region and focuses on Transit‐
Oriented Development and farmland preservation. Green Equity emphasizes allocating urban
greening (e.g., trees, green infrastructure) for underserved neighborhoods (e.g., low‐income,
minority) for environmental justice (see detailed findings on tree allocation in relation to
environmental justice in Boston (danford review). Finally, Compact Core investigates even
higher density and more infill development in the inner cities of the metropolitan area in order
to preserve open space in the suburban and rural areas. In addition, the BMA ULTRA‐ex growth
scenarios were developed through a transdisciplinary participatory planning process (Meppem
8
and Gill 1998) with stakeholders and multi‐institutions in a feedback loop of learning and
developing scenarios in deliberating multiple benefits toward a consensus of sustainable
development.
Green infrastructure is a planning intervention that is widely accepted in national
policies for sustainable development and stormwater management. In the report Towards A
Sustainable America: Advancing Prosperity, Opportunity, and a Healthy Environment for the 21st
Century written by the President’s Council on Sustainable Development (PCSD), green
infrastructure was identified as one of the five strategic areas of sustainable community
development (1999). In the same report, green infrastructure was defined as an interconnected
ecological network of natural and man‐made spaces that value different social, economic,
cultural, and ecological functions provided by ecosystems in order to guide sustainable
development (PCSD 1999). The benefits provided by green infrastructure are therefore
interlinked with social‐ecological systems. The United States Environmental Protection Agency
particularly emphasizes using green infrastructure for managing water quality and quantity (i.e.
stormwater management) (EPA 2013). As urbanization impacts on urban hydrology and climate
change‐induced flooding increasingly becomes an eminent threat to cities, the role of green
infrastructure in enhancing resilience in urban ecosystems and as a climate change adaptation
strategy in coping with effects from climate change impact is an emerging research area.
Furthermore, the way in which green infrastructure plays a role in serving the socially vulnerable
population to enhance social resilience is critical for contributing to the knowledge of the social
functions of green infrastructure performance as well as for the integration of ecological design
with community planning.
While global climate change poses impacts across temporal and spatial scales, for
regional planning on a scale such as the Boston Metropolitan Area or for local planning in the
9
city of Boston, there is a need for empirical study on a local scale using climate change
parameters for setting the planning framework. In addition, it is critical to understand the
interaction between society and ecology, the hydrological functions in urban systems, and the
livelihood of residents in relation to flooding hazards that can potentially be aggravated by
climate change. Finally, we need to evaluate the effects of planning interventions in the linked
social‐ecological systems in order to better plan for climate change towards resilience and
sustainability.
1.4 Research questions and hypotheses
It is with the understanding of the linked social‐ecological systems that this research is
intended to provide a resilience planning framework and empirical planning research for the
investigation of how planning interventions interact between people and ecosystems. The
empirical case study applied in this research aims to use flooding as a social‐ecological medium
to unpack the relationship between climate change‐induced flooding hazards and social
vulnerability in addition to the evaluation of planning interventions―growth strategy and green
infrastructure ―on the effects of reducing climate change‐induced flooding risks.
Charles River Watershed is one of the watersheds within the Boston Metropolitan Area
which has the most population and contains the most socially vulnerable groups. The
overarching research questions are: (1) To what degree does climate change become sensitive
to flooding hazards in the watershed? (2)What is the spatial relationship between flooding
hazards and social vulnerability under climate change scenarios? (3) In what way do growth
scenarios associated with land use and land cover change influence flooding hazards, in addition
to the impacts from climate change scenarios? (4) To what degree does green infrastructure
10
mitigate climate change‐induced flooding hazards? (5) What is the role of planning
interventions in social‐ecological resilience planning for climate change?
Key hypotheses in this research include:
Climate change as a result of increasing temperature and precipitation change
would increase flooding hazards
Populations with high social vulnerability are exposed to higher flooding hazards
induced by climate change
Suburban sprawl in the Current Trends scenario is associated with the most land use
and land cover change and increases flooding hazards
Suburban sprawl in the Current Trends scenario imposes more people in climate
change‐induced flooding risk areas
Detention has effects in mitigating climate change‐induced flooding hazards and can
serve as a climate change adaptation strategy
1.5 Social‐ecological systems planning framework and research
In coping with uncertainty and the dynamics of change in the linked social‐ecological
systems, a revolving learning‐by‐doing feedback loop for adaptive planning processes includes
identifying goals and objectives, plan formulation, plan implementation, and plan evaluation in
addition to plan monitoring (Kato and Ahern 2008). Transdisciplinary participation plays a
critical role in the sustainable development learning process (Meppem and Gill 1998). In
studying the dynamics of social‐ecological systems with an emphasis on the transdisciplinary
participatory approach, a social‐ecological systems planning framework in a feedback loop
process is proposed as following (Figure 1.2):
11
(1) The transdisciplinary participatory planning process involves interdisciplinary
researchers, local governance and planners, stakeholders, and the public through a
combination of various forms of participatory methods (e.g., workshops, charrettes,
surveys, interviews) that allow participants to achieve upper “ladder of citizen
participation” for engaging citizen power (Arnstein 1969).
(2) The transdisciplinary participatory planning process drives plan‐making
development to incorporate planning interventions and decide social‐ecological
drivers based on the goals and objectives in the planning agenda.
(3) Empirical research is conducted in the social‐ecological systems for the evaluation of
plans that are formulated through planning intervention and indicators identified in
the planning process.
(4) Document and monitor social‐ecological outcomes from the plan evaluation and
share findings and lessons learned with the transdisciplinary participants.
(5) Continue the transdisciplinary participatory planning process with the new insights
from social‐ecological outcomes and improve plans and/or modify social‐ecological
drivers as adaptive planning processes toward resilience and sustainability.
This research is performed under a larger research agenda set forth in the BMA ULTRA‐
ex project. Climate change and population change are the two key social‐ecological drivers
identified through the transdisciplinary participatory planning process. In addition, growth
strategies and green infrastructure are two key planning interventions applied for the
development of growth scenarios involving land use and land cover changes associated with
urbanization processes. Flooding is the social‐ecological medium being investigated in the
social‐ecological systems. Therefore, social vulnerability is introduced in order to understand
the socio‐economic characteristics created through the urbanization process and its interaction
12
with climate change‐induced flooding hazards. The long‐term climate change‐induced flooding
risk assessment―a product of multiplying the flooding hazard index (HI) and the Social
Vulnerability Index (SoVI)―is the social‐ecological outcome for the evaluation of planning
interventions. Finally, planning interventions serve as both climate change mitigation (e.g.,
reducing carbon emissions, minimizing impervious surfaces, reducing urban heat island effects)
and climate change adaptation strategies (e.g., enhancing resilience to climate change‐induced
flooding risks)(Figure 1.3).
Figure 1. 2 A social‐ecological systems planning framework
13
Figure 1. 3 Research study in the social‐ecological systems planning framework
1.6 Organization of chapters
This dissertation comprises three independent but inter‐linked empirical studies.
Chapter 1 and Chapter 2 provide the overarching overview of the research agenda and
background. Chapter 3, 4 and 5 illustrate the details of the three independent studies including
their own literature review, research methods, results, discussions and conclusions.
Chapter 1 introduces the background and purpose of the research as well as the
proposed research framework in the social‐ecological systems.
Chapter 2 provides literature reviews on five key research areas in addition to the
description of the study area of the Charles River Watershed: (1) an overview of the resilience
theory and its linkage to social‐ecological systems planning, (2) an overview of hazards,
14
vulnerability, and risk assessment focusing on the integration of using social vulnerability in the
place‐based risk assessment for flooding hazards, (3) an overview of urbanization impacts on
hydrology, focusing on the relationship between land cover change and its effects on
streamflow that is associated with baseflow and long‐term flooding, and(4) an overview of
climate change impacts on hydrology based on temperature and precipitation trends and
projections in the Northeast region.
Chapter 3 presents the long‐term climate changed‐induced flooding risk assessment in
current climate and land use conditions as the foundation and baseline for the following two
studies. This study defines the measurement of flooding hazards and social vulnerability with
focuses on the effects of climate change variables on long‐term flooding risks.
Chapter 4 presents the future long‐term climate changed‐induced flooding risk
assessment. Building on Chapter 3, growth strategies developed in the BMA ULTRAL‐ex growth
scenarios are evaluated. This chapter investigates the impacts from urbanization in addition to
climate change impact scenarios.
Chapter 5 investigates green infrastructure intervention based on the hydrological
model developed in Chapter 3 under current land use conditions with climate change sensitivity
studies for understanding the effects of detention in mitigating long‐term flooding hazards
under climate change variables.
Chapter 6 concludes with a summary of findings, insights and the limits of this research
as well as suggestions for further research.
15
CHAPTER 2
BACKGROUND
2.1 Planning in social‐ecological systems
2.1.1 Resilience and climate change
Embracing planning as a science in anticipating change in the intricately linked social‐
ecological systems, the concept of resilience has created a new frontier in planning theory.
Resilience theory, deeply rooted in ecology (Holling 1973), has provided system thinking in
dealing with complex issues such as climate change in the dynamic social‐ecological systems.
Resilience as a metaphor and as a frame of reference for studying shocks and stresses in human
systems has been applied to disaster studies, psychology, sociology, political science and
economics (Pendall, Foster, and Cowell 2010)as well as in urban and regional planning (Adger
2006; Wilkinson 2012).
Resilience is defined as a capacity or persistence in a system to absorb shocks and
disturbances that change the state of variables in the system and the ability to return to the
state of the same structure and functioning of the system (Holling 1973; Walker and Salt 2006).
There are three key concepts in resilience thinking: threshold, adaptive cycles, and anarchy. A
threshold refers to a tipping point when a system is unable to resist the change of the state of
variables which eventually causes regime shift. An adaptive cycle describes how a system
functions and reacts to disturbances in four phases: (1) the rapid growth phase (r phase) when
actors in a system eager to seize new opportunities and available resources seize them in the
early stages of development; (2) the conservation phase (k phase) when actors in a system move
away from a strict competition and establish a connection and steady growth under carrying
16
capacity; (3) the release phase (omega phase) occurs when a disturbance exceeds the resistance
and breaks the structure and functioning of a system; and (4) the reorganization phase (alpha
phase) when a system reorganizes and renews, and returns to the basic structure and
functioning of a system. The concept of anarchy emphasizes that a system is constantly
changing and includes multiple thresholds at multiple scales. When the cost of adaptability
exceeds the benefits of staying in the existing system or the system has lost a great deal of its
structure and functions to the point that restoring the system is not feasible or desirable, then
transformation of the system to another set of structures and functions would provide a greater
resilience in the future (Walker and Salt 2006).
Human is part of nature (Steiner 2002). The complex structure of ecological and social
systems is intertwined and indivisible. Nevertheless, the scale of anthropogenic change to the
environment has reached a level that threatens the survival of the human race. Climate change
is an example of an event caused by humans (Karl, Melillo, and Peterson 2009) when the
reciprocal changes between the ecological and social changes are so great that they pose a
threat to the survival of human beings as well as other living beings. As stated in the report, Our
Common Future, the goal of sustainable development is to sustain the abilities of future
generations to survive and thrive (Bruntland 1987). Climate change as a shock and disturbance
to sustainability in the dynamics of the inter‐linked social‐ecological systems can be perceived
from two perspectives. From an ecological point of view, how can the ecosystems in the built
environment be resilient to anthropogenic alteration? From a social perspective, how can
human beings be resilient and adaptive to drastic as well as persistent changing environments
that are shaped by bio‐physical characters and reshaped through anthropogenic influence such
as in the case of climate change?
17
Climate change is a complex phenomenon that possesses several characteristics. First,
climate change happens across a spatial scale. Its causes and effects cross neighborhoods,
municipalities, regions, and nations. It is a global issue as well as a regional and local issue.
Second, climate change happens across a temporal scale. Causes that occurred five decades ago
continue to have effects at this moment due to the lag time of accumulated effects of
greenhouse gases (Karl, Melillo, and Peterson 2009). Third, there is a great uncertainty in how
we understand and predict climate change. Since the Intergovernmental Panel on Climate
Change (IPCC) published the first climate change assessment report in 1990, the world has
gained more knowledge and momentum in facing this challenge. Nevertheless, work remains to
be done to trace to the past in order to understand climate change and project its causes and
effects into the future. Lastly, to our understanding, there is no single solution to resolve the
effects of climate change. This is due to the production of greenhouse gases, which are the
major causes of climate change, being intertwined with our social‐ecological system, which
ranges from energy, transportation, agriculture and manufacture to land use, urban
development patterns and buildings. It has links to every component of the system we live in.
Moreover, the choices we make today will affect climate change impacts in the future (Karl,
Melillo, and Peterson 2009).
Resilience theory is related to properties of climate change in several aspects: (1)
resilience theory deals with cross‐temporal and spatial scales where climate change occurs; (2)
resilience theory deals with complex social‐ecological systems that climate change has impacts
on; (3) resilience theory deals with uncertainty generated by the character of climate change.
The resilience framework should be robust rather than be searching for optimal control of the
social‐ecological systems and this framework is intended to inform policy and management
(Anderies, Walker, and Kinzig 2006).
18
2.1.2 Watershed as a social‐ecological planning unit
A watershed is a tangible and logical study unit for the understanding of interactions
between climate change, hydrology and urbanization in dynamic social‐ecological systems. A
watershed is a land area defined by geomorphologic boundaries within which hydrologic flow
interacts with human activities. A watershed can be defined as a bioregion, which is an area
with “rough boundaries determined by natural characteristics that are distinguishable from
other areas by particular attributes of flora, fauna, water, climate, soils, landforms, and by
human settlements and culture” (Steiner 2002)(p 106). In addition, a watershed provides a
natural management framework not only for water resources but also for the study of spatial
heterogeneity and human ecosystems (Pickett, Cadenasso, and Grove 2004). Therefore, it
instantly provides a tangible working unit for scientific study and a planning scale.
For example, one can apply landscape ecology principles and study landscape
structures, functions, and processes in a watershed. One can also monitor and assess the input
and output of hydrological cycles and nutrient flows and evaluate the environmental quality of
the watershed. In addition, watershed bioregions are larger scale ecosystems that are more
likely to provide buffers for external carrying capacity in order to overcome insufficiency and the
instability of smaller ecosystems when facing disturbances (Newman and Jennings 2008).
Moreover, human activities are interconnected with water flows. The water quality and the
amount of water usage from upstream cities have accumulated significant impacts on
downstream cities. Furthermore, water provides transportation accessibility and allows cities to
trade and exchange culture and social experiences. As a result, in considering the
environmental, economic, and social dimensions, watershed planning is particularly powerful as
a regional planning unit in addressing sustainability (Campbell 1996)(p 452).
19
2.2 Hazards, vulnerability, risk assessment
2.2.1 Risks and disasters
Disasters occur when human populations suffer from the consequences of natural or
man‐made hazards that interrupt ecological, social, cultural, and economic systems and paralyze
the functioning of a society materially (e.g., property damage, economic loss), physically (death
or injury) and mentally (distress or traumatized illness)(Oliver‐Smith 1986; O'Riordan 1986; Vale
and Campanella 2005). Disasters are time and space specific in particular social‐economic and
cultural context. The level of devastation as a result of disasters goes beyond death toll and the
dollar amount of damages or the losses of fixed structures and crops. It also accounts for the
social and political complexity involved in the disasters and post‐disaster recovery (Huq et al.
2007).
Human has invented the concept of risk and studied it in order to comprehend and cope
with uncertainties and potential threats in life, such as the ravages of disasters(Slovic 1999).
Risk for disasters involves hazard, exposure and vulnerability (Crichton 1999). Hazard and
exposure are correlated whereas vulnerability is intricately linked with society. For example,
floods would not have become hazards if people had not lived in flood‐prone areas and in turn
disasters would not have happened to people (Harding and Parker 1979). However, this seemly
over‐simplified argument fails to explain why people continue to live on floodplains and bear
the risks. Savini and Kammerer echoed White’s viewpoint that humans are eventually
responsible for disasters (White 1945) when they stated: “Man takes a calculated risk when he
builds his towns, cities, and other installations on the floodplain of a river. He, thus, is
responsible for flood damages, if not for the floods that cause them”(Savini and Kammerer
1961)(p 1591‐A). This “calculated risk” in each decision that is made during the process of
20
civilization is deeply rooted in particular cultural, social, political and economic contexts that are
place‐specific. Government takes risks to build dikes; investors take risks to develop floodplains;
individuals take risks to live on the flood‐prone properties; insurance agencies take risks to
compensate losses. This chain of links to risks has made our society more susceptible to
hazards. Consequently, when society does not have all the necessary means and resources to
prevent and mitigate disasters, the hardest hits are the ones with the least capacity to prepare,
respond and recover, forming the root of vulnerability (Oliver‐Smith 2003).
2.2.2 Vulnerability and resilience
Vulnerability is the key to understanding risks, managing disasters, and becoming
resilient. The concept of resilience describes a capacity of a system to absorb shocks from
disturbances and to return to the state of the same structure and functions (Holling 1973;
Walker and Salt 2006). In the risk management context for human society, resilience is further
defined as a “capacity to anticipate, prepare for, respond to, and recover from significant
threats with minimum damage to social well‐being, the economy, and the environment”(NRC
2010). To build upon resilience concept, Adger claimed that vulnerability is a sign for the
“erosion of the elements of social‐ecological resilience”(Adger 2006)(p 269). While one school
of thought argues vulnerability is an inverse of resilience (De Wrachien, Mambretti, and Sole
2008); the other suggests vulnerability includes three dimensions: (1) exposure to specific social
or environmental stresses, (2) sensitivity to those stresses, and (3) adaptive capacity to cope
with impacts from those stresses (Polsky, Neff, and Yarnal 2007; Turner et al. 2003). Moreover,
vulnerability to hazards is closely associated with social resources at multiple scales from an
individual to collective adaptive capacity (Birkmann 2006). The complex political and economic
systems during the urbanization processes often result in a “risk society” that creates socially
21
vulnerable groups who are more susceptible to various hazards (Blaikie 1994; Colten 2006).
When communities have insufficient coping capacity for the shocks and disturbances in the
coupled natural and human systems (Liu et al. 2007), they are likely to become more vulnerable
to the adverse effects of uncertainty and extreme variation, which climate change has promised.
Resilience theory is pertinent for studying the concept of vulnerability and managing
risks. Disasters happen when human society reaches the threshold and enters the release phase
and subsequent reorganization phase. Vulnerability is exposed during the transition between
the two phases and affects the adaptive capacity of resilience. Vulnerability links to social and
cultural contexts in political society and reflects individual as well collective adaptive capacities
in coping with uncertainty and calamity. Therefore, vulnerability reflects the dynamic phases in
resilience and applies to multiple levels of inter‐linked social‐ecological systems (Schneiderbauer
and Ehrlich 2006).
2.2.3 Vulnerability assessment models
Vulnerability assessment has progressed over the past twenty years for addressing
complex coupled human and natural systems in coping with disasters. Birkmann reviewed
several different frameworks and summarized six different schools of thought based on how
vulnerability is systematized into conceptual and analytical risk and resilience frameworks
(Birkmann 2006) (p 39):
The Double Structure of Vulnerability (Bohle 2001) framework considers “exposure” to
hazards as the “external” side of vulnerability, which is related to entitlement theory,
political economy approaches, and human ecology perspectives, whereas the capacity
of “coping” as the “internal” side of vulnerability, which is related to crisis and conflict
theory, action theory approaches, and models of access to assets.
22
The conceptual framework of disaster risk (Davidson 1997; Bollin et al. 2003) identifies
four components that define a disaster risk: (a) the probability and severity of hazards,
(b) the exposure of structures, population, and economy, (c) the physical, social,
economic and environmental vulnerability, and (d) the coping capacity and measures
from physical planning, social capacities and economic capacities.
The vulnerability assessment in the global environmental change community (Turner et
al. 2003) considers the drivers and consequences of vulnerability are cross spatial,
functional and temporal scales. In addition, the concept of vulnerability is composed of
three elements in the coupled human and environmental conditions: exposure,
sensitivity and resilience.
The Pressure and Release (PAR) model (Blaikie 1994) illustrates the dynamics in societies
that lead to formation of risks based on the equation: Risk=Hazard X Vulnerability.
Ideologies in political and economic systems that limit access to power, structures and
resources are the root causes that form dynamic pressures in society. Those pressures
appear in various forms (e.g., lack of appropriate skills, or local markets) and forces (e.g.,
press of freedom, rapid urbanization) and eventually generate unsafe conditions, such
as a fragile physical environment or local economy, vulnerable society and insufficient
public actions in coping with disasters.
The holistic approach to risk and vulnerability assessment (Cardona and Barbat 2000)
considers risk as a function of hazard and vulnerability, which result in potential social,
economic and environmental consequences. This model identifies three vulnerability
factors: (a) the exposure and physical susceptibility that is hazard dependent, (b) the
social and economic fragility that is non hazard dependent, and (c) a lack of resilience or
ability to cope and recovering.
23
The BBC conceptual framework (Cardona 1999; Bogardi and Birkmann 2004)links
vulnerability to sustainable development (i.e., environmental, social, economic) in a
feedback loop system. This model considers risks are the results of interaction between
hazards and vulnerability. The concept of vulnerability involves the exposed and
vulnerable elements and coping capacity. Additionally, vulnerability reduction is
integrated into the feedback loop that contributes to risk reduction.
One thread that ties the different schools of thought together is the recognition that risk
is correlated with the interaction between hazard and vulnerability. In addition, the coping
capacity in resilience thinking has become a critical measure of vulnerability. Finally,
vulnerability assessment encompasses environmental, social, and economic factors that
correspond to elements in sustainability.
Vulnerability that reflects the dynamic socio‐economic and cultural structure and fabrics
of a society also varies from place to place. In order to address dimensions of place‐based
context‐specific vulnerability based on Hazards‐of‐Place (HOP) model, Cutter (1996) developed a
Social Vulnerability Index (SoVI) (Cutter 1996; Cutter, Boruff, and Shirley 2003; Cutter and
Morath 2013). The HOP model integrates biophysical and social vulnerability at a specific
geographic location. The SoVI is a benchmark for measuring vulnerability in a systematic and
quantitative approach. It has been used successfully in several national and regional social
vulnerability studies and significant trends across temporal and spatial distributions were found
in the United States (Cutter, Boruff, and Shirley 2003; Borden et al. 2007; Cutter and Finch
2008). In recent years, the concept of resilience has been incorporated into vulnerability
assessments using this approach, such as the Disaster Resilience of Place (DROP) model (Cutter
et al. 2008), for the assessment of one of the most flood‐ and hurricane‐prone regions in the
southeast of the United States (Cutter, Burton, and Emrich 2010).
24
2.3 Urbanization impacts in social‐ecological systems
2.3.1 Urban hydrology overview
Urbanization, from a historical and world‐wide point of view, is seen as part of the
process of industrialization in which population in cities have been growing since the 1800s
(Davis 1955). In the context of this research, urbanization in the Boston Metropolitan Area is the
process of anticipating land development in order to accommodate population growth through
2030. The anthropogenic impacts on ecosystems through the urbanization process and its
associated use of resources have three phases: deforestation and clearing, construction and
post‐occupancy (Savini and Kammerer 1961). The hydrologic change is evident in the first phase
of the urbanization process related to land cover change (Lindh 1972) and human‐induced
disturbances such as dams, channelization and diversions of streams, and urban drainage
systems (NRCS 1998). There are many impacts from land development associated with land use
and land cover change on hydrology (Arnold and Gibbons 1996; Leopold 1968; Schueler 1994;
Shuster et al. 2005), including hydrologic functioning, water quality, and both in short‐term
flooding (i.e., flash floods) and long‐term flooding (i.e., overbank flooding).
In the natural water cycle, stream baseflow combined with subsurface and surface flow
is largely sustained by groundwater and stream outflow fluctuates with the input from
precipitation, evapotranspiration output and soil water availability (Thornthwaite and Mather
1955). Urbanization has altered natural water balances mainly through impervious and sewer
(piping) areas (Leopold 1968) in addition to inputs from piped water for irrigation and pipe
leakage (Bhaskar and Welty 2012; Tjallingii 2012). In addition, rainwater harvest (e.g., cisterns,
rain barrels) and detention and infiltration techniques used in stormwater best management
practices (e.g., greenroofs, bioswales) as well as natural interceptors through soil, surface water
25
bodies, and plants can affect water content in urban areas and consequently influence
streamflow (Tjallingii 2012)(Figure 2.1).
Natural water balance: Stream outflow = Baseflow (subsurface or groundwater flow) +
Surface flow + precipitation – evapotranspiration – soil water (1)
Figure 2. 1 Water balance in an urbanized watershed (modified from Tjallingii 2012;
Bhaskar and Welty 2012)
The change of land use and land cover from forests or wetlands to commercial or
industrial land use and housing developments with impervious surfaces such as roads, buildings,
roofs and parking lots significantly reduces the amount of infiltration and increases the surface
runoff. For example, compared to natural ground cover, which has 10% runoff and 50%
infiltration, urbanized areas that have surfaces over 75% impervious results in less than 15%
infiltration and contributes to more than 55% of runoff (NRCS 1998). In addition, impervious
surfaces have been highly associated with shorter lag times between the fall of precipitation and
runoff mass that instigates the increase of flashy variation of streamflow or a peak discharge in a
short period of time (Simmons and Reynolds 1982). In completely impervious basins, the lag
time can be as little as one eighth of that in the natural state (Anderson 1970), resulting in
Storage
Soil/Surface/Plants
Detention Harvest
Stream Inflow Stream Outflow
Precipitation Evapotranspiration
Infiltration Aquifer recharge
Seepage /Springs
Pumping Groundwater
Baseflow (subsurface or groundwater flow)
Surface flow Piped water
Stream Baseflow
Stormwater runoff
Piped water
Rainwater
26
significant increases in stormwater runoff and flash floods in urban areas (Sala 2003;
Hosseinzadeh 2005).
Furthermore, urbanization associated with the clearance of natural land cover which is
replaced with impervious surfaces has direct environmental impacts such as wildlife habitat loss,
degradation of ecosystem integrity and environmental quality (Allan 2004; Allan, Erickson, and
Fay 1997; Lammert and Allan 1999; Roth, Allan, and Erickson 1996) in addition to increased
stream temperature, concentrated pollutants from piping as well as non‐point source pollution
that impedes water quality in the watershed (Booth 1993; Brabec, Schulte, and Richards 2002;
Burton and Hook 1979; Ellis and HvitvedJacobsen 1996; Paul and Meyer 2001).
2.3.3 Urbanization impacts on streamflow and flooding
In Leopold’s (1968) seminal book: Hydrology for Urban Land Planning―A Guidebook on
the Hydrologic Effects of Urban Land Use, the importance of land use planning in minimizing
urbanization impacts on hydrologic functioning of the watershed as well as the critical linkage of
using hydrological parameters for understanding urbanization issues was demonstrated
(Leopold 1968). Flood frequency is one of the metrics that is widely used to quantify the effects
of urbanization on streamflow (NRC 2009). Flood frequency is defined as the probability of a
given flood event once in any given year. In Leopold’s (1968) influential publication, he
demonstrated significant correlations between increased annual flooding events and increased
urbanization, measured by percent of impervious and sewered areas based on 9‐year historical
data between 1942, 1944‐1951 in Brandywine Creek, Coatesville, Pennsylvania, with a mean
annual flood return period at 2.3 years (Figure 2.2). In this case, the same recurrence interval of
2.3 years in a 50‐50 urbanized state would result in a discharge rate at 200 cubic feet per second
(cfs) compared to 75 csf in a natural state. To understand urbanization impacts on long‐term
fl
be
an
fl
Fi
th
in
M
ar
be
C
ooding, Leop
etween the r
nd 67 csf as t
ooding event
igure 2.2
Since L
he correlation
ncreased mag
Moscrip and M
reas decrease
ecomes negli
harlotte, Nor
old (1968) tra
atio of overb
he natural ba
ts were nearly
Flood frequ
Leopold’s infl
n between inc
gnitude and fr
Montgomery 1
es upon the in
gible (e.g., a
rth Carolina (M
ansformed th
ank flooding
ankfull stage (
y four times m
uency curves
uential work
creased storm
requency of f
1997; Ng and
ncreases of fl
study showed
Martens 1968
27
he same data
and urbaniza
(Figure 2.3).
more than tho
in highly urb
in 1968, ther
mwater runof
flood events (
Marsalek 19
ood recurren
d a threshold
8). Therefore
into a diagra
ation effects u
In 50‐50 urba
ose in the nat
banized areas
re have been
ff associated w
(Allen et al. 19
989). Neverth
nce intervals (
of 50 years i
e, while the fr
m illustrating
using a 1.5 ye
anization ratio
tural state.
s (Leopold 19
numerous st
with impervio
979; Huang e
heless, the eff
(Hollis 1975) a
nterval in me
requency of s
g the relations
ar return per
os, overbank
968)
tudies suppor
ousness and t
et al. 2008;
fect of imperv
and eventual
etropolitan
mall and regu
ship
riod
rting
the
vious
ly
ular
fl
w
Fi
(L
th
na
th
In
ha
M
ar
ooding event
would cause se
igure 2.3
Leopold 1968
Moreo
he baseflow (
atural state o
he increased
n addition to i
ave inconsist
Moreover, the
reas could co
ts would incre
evere damag
Number of
8)
over, criticism
Hollis 1975; P
of the stream,
impervious su
imperviousne
ent impacts d
e leakage in th
ontribute to in
ease many tim
es are less sig
f overbank flo
m has risen in
Price 2011)—
, baseflow is
urfaces are as
ess associated
due to incons
he piping syst
ncreased base
28
mes more, the
gnificantly inf
ooding event
regards to th
the base of t
largely sustai
ssociated wit
d with urbani
sistency in eva
tems as well a
eflow. Furthe
e rare and ex
fluenced by u
ts in a year in
he understudi
he sustained
ned by groun
h decreased
zation, defor
apotranspirat
as input from
ermore, othe
xtreme floodin
rbanization (
relation to u
ied urbanizat
portion of st
ndwater; in an
baseflow (Sh
estation and
tion rate (Pric
m irrigation in
r human‐influ
ng events tha
Hollis 1975).
urbanization
ion impacts o
reamflow. In
n urbanized s
uster et al. 20
agricultural c
ce 2011).
the urbanize
uenced chang
at
on
n the
state,
005).
crops
d
ge in
29
the urbanization process such as rerouting streamflow, channelization, and climate change
impacts require further study due to their impact on baseflow.
From Price’s review, nevertheless, a clear correlation of decreased baseflow can be
found as a result of increased evapotranspiration and decreased infiltration. Long‐term flooding
is heavily influenced by a watershed’s ability to balance its water budget as a result of storm
events. Urbanization associated with increased imperviousness has impacts on both increased
stormwater runoff and decreased baseflow that could have either or both positive and negative
effects on long‐term flooding. While urbanization is often largely associated with impervious
surfaces, which remains a critical component for assessing hydrological functioning and stream
quality (Schueler, Fraley‐McNeal, and Cappiella 2009), it is not the sole indicator for long term
management (Rogers and DeFee Ii 2005).
In addition, it is critical to make the distinction between total impervious areas (TIA)
(i.e., all roads, roofs, building footprints that are impervious) and effective impervious areas
(EIA) (i.e., hydrologically connected impervious areas). Effective impervious areas are
hydrologically connected through curbs, gutters and pipes; by contrast, non‐effective
impervious areas drain to pervious areas, such as disconnected downspouts from roof areas to
adjacent lawns (Alley and Veenhuis 1983). Most studies have used total impervious area that
lump‐sums both effective and non‐effective impervious areas (Brabec, Schulte, and Richards
2002) and therefore overlooked the accuracy and effectiveness of using impervious area as an
indicator for stormwater management (Brabec 2009). Further studies are needed to
understand the “true” effects of urbanization―incorporating multiple indicators in addition to
effective impervious areas―on streamflow, particularly the baseflow, in order to develop
appropriate corresponding urban planning and policies for mitigating urbanization impacts on
hydrological functioning of the watershed.
30
2.4 Climate change impacts on hydrology in New England
Climate change is likely to increase the intensity of precipitation patterns in their
magnitude and duration and affects the global hydrologic cycle (Frederick and Major 1997; IPCC
2007). The effects of climate change intensify the intensity and frequency of storm events and
therefore magnify the urban hydrological impacts(Frederick and Major 1997; Wood,
Lettenmaier, and Palmer 1997). More frequent and intense storm events are likely to occur in
some areas such as the New England region (Rock et al. 2001). The consequences of irregular
and intensified flooding and drought events have significant impacts on populated urban regions
where current water infrastructure was designed based on past climate trends and conventional
knowledge (Ashley et al. 2005; Means, West, and Patrick 2005).
More frequent flooding and intensified storm events will cause more damage in
populated urban regions. Alternative structural stormwater design may be needed for
accommodating climate change effects (Semadeni‐Davies et al. 2008). Increased capital
investment in upgrading drainage infrastructure is therefore necessary under the current
infrastructure system (Arisz and Burrell 2006; Muller 2007). The under‐served population is
therefore more vulnerable to the impacts of climate change‐induced flooding (Bartlett 2008;
Douglas et al. 2008; Hardoy and Pandiella 2009; Maantay and Maroko 2009). Planning for
resilience in enhancing the capacity of absorbing urban flooding hazards is therefore a top
national priority in cities, particularly along the coast (Beatley 2009; Godschalk 2003).
2.5 Study context and study area
New England was the frontier for European settlements since the arrival of the
Mayflower in 1620 in the northeastern United States. The climate of generous precipitation, the
31
proximity to the ocean and the plentiful rivers on the one hand brought prosperity to dozens of
mill towns and coastal cities during the industrial era; on the other hand, they contributed to
several hurricanes and flooding disasters over centuries. Despite federal and state emergency
management agencies continuing their efforts in hazard mitigation and risk management, little
attention has been given to this region for disaster research. This is likely owed to its lower
vulnerability compared to other parts of the country (Borden et al. 2007; Cutter and Finch 2008)
and therefore lower priority for the nation. Until recently, increasing concerns in rising sea
levels as a result of climate change impacts have drawn some attention to vulnerability and
resilience research on coastal communities, particularly for large coastal cities like New York City
and Boston (Kirshen, Knee, and Ruth 2008). However, recent significant flooding events in 2010
and 2011 were mainly non‐coastal floods, suggesting further research on climate‐induced inland
flooding is needed. As Massachusetts’s Climate Change Adaptation Report stated, “the need to
perform risk and vulnerability assessments was widely recognized across all sectors” (Cash
2011)(p 3).
The Charles River Watershed comprises 778 km2 and intersects 35 municipalities,
including a large eastern portion of the Boston (Figure 2.4). The 182 km‐long main channel
flows eastward through 23 cities and towns and discharges into Boston Harbor. Within the 35
municipalities, all but two towns at the upstream edge of the watershed are not within the
Boston Metropolitan Area. Those two towns contain less than 1% of land within the watershed
boundary and are not included in the MAPC future growth projections; therefore, they are
omitted from this study.
32
Figure 2.4 Location of the Charles River Watershed study area
33
The watershed is 44% urbanized (i.e., commercial, industrial, residential, transportation,
institutional, junkyard, utilities) with little agriculture (3%) and recreational (3%) land uses.
Within the urbanized areas, impervious surfaces compose 21% of the watershed areas are
whereas other pervious areas (e.g., lawns, planting beds) compose 23%. The other half of the
watershed is dominated by natural areas, including forests (36%), wetlands (11%) and water
bodies (3%) (Figure 2.5).
The Charles River Watershed was chosen out of nine watersheds in the Boston
Metropolitan Area for several reasons. First, the entire watershed is predominantly within the
Boston Metropolitan Area with minimal coastlines since coastal flooding is not the focus of this
study. In addition, this watershed is the most densely populated and consists of the most
Environmental Justice populations that are related to social vulnerability, which is one of the key
elements in the long‐term climate change‐induced flooding risk assessment (see Chapter 3).
Figure 2.5 Current land cover types in percentages of the Charles River Watershed area
0
5
10
15
20
25
30
35
40
Forest Other perviousarea
Imperviousarea
Wetland Recreation Agriculture Water
% W
atershed Area
34
CHAPTER 3
LONG‐TERM CLIMATE CHANGE‐INDUCED FLOODING RISK ASSESSMENT
3.1 Introduction
Climate change has posed increased risks to environmental hazards. These will affect
more people who are already exposed to natural hazards (e.g., flooding, droughts, hurricanes,
severe storms), in addition to new challenges promised by climate change (e.g., early snow melt,
rising sea levels, heat waves). Previous climate change studies have suggested probable trends
of increasing temperatures and changing precipitation as well as increased intensity and
duration of storm events that are likely to result in more flooding in the Northeast region (IPCC
2007; Rock et al. 2001). Superstorm Sandy 2012 is another wake‐up call for many New
Englanders that extreme events have occurred more frequently and intensely in the past two
decades. Hurricane Irene in 2011 and serious floods in 2011, 2010, 2005 and 1996 correlated
with recent projections of climate change impacts. As a result, this study employs the term
“climate change‐induced” flooding hazards to reflect the assumption that a greater probability
of long‐term flooding hazards (in lieu of short‐term flash floods) will be part of climate change
impacts.
Natural hazards do not pose risks to society until human systems are vulnerable to
subsequent casualties and monetary losses. Compounded by population growth in most
metropolitan areas, more people and property are likely to be exposed to climate change‐
induced hazards. The ones hardest hit are often groups that are socially vulnerable as
exemplified in the tragedy of Hurricane Katrina in 2005 when historically disadvantaged
neighborhoods in New Orleans were the most affected (Colten 2006). The aftermath of
Hurricane Katrina was a tragic reminder that hazard study is an integration of both physical and
35
human variables in the coupled human and natural systems (White 1945; Hewitt and Burton
1971)and that risk of disaster is the interaction between hazards (i.e., floods, droughts, storms,
earthquakes) and vulnerability of society (i.e., demographics, social‐economic status, built
environment, infrastructure, economy)(Tate, Cutter, and Berry 2010; Schneiderbauer and
Ehrlich 2006; Cutter, Boruff, and Shirley 2003; Birkmann 2006). As a result, an understanding of
climate change‐induced risk assessment integrating hazards and social vulnerability studies is
therefore critical in planning for climate change.
This study aims to integrate climate change‐induced flooding hazards and social
vulnerability into flooding risk assessment using the Charles River Watershed within the
populous Boston Metropolitan Area as a case study. The research questions arise to what
degree does the Charles River Watershed become sensitive to climate change‐induced flooding
hazards? Are socially vulnerable people spatially correlated to the locations of the climate
change‐induced flooding hazards? The hypotheses were: (1) climate change derived from
increasing temperature and precipitation change will increase long‐term flooding hazards; (2)
populations with higher social vulnerability tend to live in places with higher flooding hazards.
The findings of this study contribute to place‐based understanding of climate change influence
on watershed‐specific hydrology in addition to implications of contributing to serving socially
vulnerable groups in the process of coping with climate change.
3.2 Background
3.2.1 Flooding losses in the Boston Metropolitan Area
Floods are present in almost every city in the United States (White and Haas 1975; Platt
1999). Among the nation’s economic losses caused by natural disasters in the past fifty years,
36
more than half were due to flooding related damages—as much as $291 billion in 2009 dollars
(Gall et al. 2011). Over the past 60 years in Massachusetts, severe storms and flooding related
hazards were among the top three causes of environmental disasters and were responsible for
$883 million (in 2011 dollars) and accounted for a third of total economic losses in property
damages (Table 3.1). 61% of flooding damages occurred within the five counties that contain
the Boston Metropolitan Area (BMA).
Table 3.1 Monetary loss in property damages (in 2011 dollars) in Massachusetts (MA)
and the Boston Metropolitan Area (BMA) from 1953 to 2011 (Data source: SHELDUS)
Hazards MA BMA BMA/MA
Wind 669,702,857 27% 249,635,129 37%
Severe Storms 460,106,827 18% 171,625,495 37%
Floods 392,809,831 16% 238,903,881 61%
Tornado 295,220,097 12% 6,208,713 2%
Hurricane 232,478,220 9% 77,517,773 33%
Winter Weather 219,585,698 9% 78,864,153 36%
Coastal Hazards 176,444,529 7% 84,036,222 48%
Lightning 56,117,791 2% 30,937,859 55%
Hail 20,430,050 1% 10,010,921 49%
TOTAL 2,522,895,900 100% 947,740,145 38%
Severe storms and flood‐related damages together accounted for 43% of total monetary
losses in the five counties—Essex, Middlesex, Suffolk, Norfolk and Plymouth—that contain the
BMA. Extreme storm events in particular contributed significantly to damages. For example,
severe flooding in 2010 alone accounted for 44% (171.5 million in 2011 dollars) of total flooding
damages from 1953 to 2011. Consequently, flooding has remained a major concern in BMA,
which accounted for 61% of total flooding losses in Massachusetts (Figure 3.1).
37
Figure 3.1 Monetary losses compared between Boston Metropolitan Area (BMA) and
Non‐BMA in Massachusetts from 1953 to 2011 (in 2011 dollars)(Data source: SHELDUS)
Aside from monetary losses, fatalities due to natural hazards accounted for eight
people, all caused by severe storms in Massachusetts. Only one occurred in the Boston
Metropolitan Area. Severe storms, along with flooding and hurricanes were the major causes of
injury. 70% of a total 106 recorded injuries in the past 60 years were located in the Boston
Metropolitan Area. One single severe storm in 1996, for example, injured 60 residents in
Plymouth County alone. In addition, 3 out of 5 people that were injured in the severe flooding
in 2010 were reported inside the Boston Metropolitan Area. The historical records of economic
damage, injury and fatality have provided an overview of the types, locations, patterns, and
economic impact indicators of natural hazards. The next section describes demographics and
social‐economic structures that are associated with increased social vulnerability to disasters.
$‐
$100
$200
$300
$400
$500
$600
$700
Millions
Non‐BMA
BMA
38
3.2.2 Demographics and status of equity in the Boston Metropolitan Area
The Boston Metropolitan Area (BMA) within the Metropolitan Area Planning Council
(MAPC) region consists of 101 municipalities and a population of 3.16 million with an estimated
11% increase through 2030 (MAPC 2009). Currently, its demographics show older population,
more diversity and foreign‐born people in younger age groups, as well as an increase in income
and education inequality (MAPC 2011). The portion of the population that is in age of 55 and
older has risen from 21% to 25% in the past decade and seniors over 65 are projected to be
more than 25% of the population in 37 communities in the BMA region. The aging population
primarily consists of non‐Hispanic white groups while younger (below 45 years old) groups are
dominated by non‐white groups (MAPC 2011)(p 11‐13). Furthermore, the population consisting
of foreign‐born residents has risen 3% from 2000 to 2009, coming mainly from Latin America
(37%), Asia (30%) and Europe (22%) (Figure 3.2). They are primarily distributed outside the city
of Boston in the metropolitan area where the Charles River Watershed is located. The highest
increases have been in the towns of Wellesley and Milford.
Moreover, there is an increasing income gap between the rich and poor. Between 1979
and 2006, the median income of the top 20% rich group has increased from 15 times to 20 times
more than the bottom 20% poor group (MAPC 2011)(p 26). The income disparity is also
reflected in racial groups—African Americans have between 52% and 57% of the income of
Asian and white groups while Latinos and other races have less than half of the income of the
wealthiest in the region.
Fi
(M
la
po
p
to
de
th
as
igure 3.2
MAPC 2011, p
Finally
abor force em
opulation con
rimarily in the
o the towns o
evelopment g
he Charles Riv
s in poverty r
Mi
Difference
p 16) and the
y, MAPC’s equ
mployment (M
nsists of abou
e city of Bost
of Wellesley a
growth (Figur
ver Watershe
ate.
W
ilford
in percentag
e Charles Rive
uity report sta
MAPC 2011) (p
ut 25% Latino
on and subur
and Milford th
re 3.3). It is w
ed experience
Wellesley
39
ge of foreign‐
er Watershed
ated that disc
p 77), resultin
os, 20% Africa
rban commun
hat are close t
worth noting t
ed the most in
Boston
born residen
d
crimination re
ng in people l
Americans, a
nities that are
to highways a
that the town
ncrease in for
ts from 2000
emains the m
iving below t
and 12% Asia
e densely dev
and are unde
ns of Wellesle
reign born po
Char
Char
0 to 2009 in B
major obstacle
he poverty. T
ns (p79) who
veloped in add
ergoing
ey and Milfor
opulation as w
rles River Wate
rles River
BMA
e for
This
o live
dition
d in
well
ershed
Fi
R
C
En
eq
th
as
M
EJ
C
M
igure 3.3
iver Watersh
In add
ommonwealt
nvironmenta
qual to or mo
han 65% of st
s English‐isola
Massachusetts
J criteria, and
harles River W
Massachusetts
Milf
Poverty rat
hed
ition to MAPC
th of Massach
l Justice (EJ) p
ore than 25%
tatewide aver
ated with a la
s has a total o
d the BMA com
Watershed is
s’s population
W
ford
te from 2005
C’s equity rep
husetts (Mass
populations b
of non‐white
rage ($40,673
ack of “very w
of 6.5 million
mprised 35%
the most pop
n) within the
BWellesley
40
to 2009 in B
port, the Offic
sGIS) has esta
based on thre
e minority gro
3 in 2010), an
well” English p
and about on
of Massachu
pulous waters
BMA and con
Boston
MA (MAPC 2
ce of Geograp
ablished a sta
ee major crite
oups, a media
d more than
proficiency. B
ne in three (3
usetts’s EJ pop
shed (about 9
ntains one thi
2011, p 78) an
phic Informat
atewide datab
eria at the uni
an household
25% of house
Based on 201
5%) residents
pulations (Fig
900,000 peop
ird of the BM
Charles
Charles
nd the Charle
tion for the
base for
t of block gro
income lowe
eholds identif
0 census data
s met one or
gure 3.4). The
ple, one sixth
A’s EJ
River Watersh
River
es
oup—
er
fied
a,
more
e
of
hed
41
populations, which are primarily located in the lower basin (Figure 3.5). EJ populations in
Massachusetts have increased from 29% in 2000 to 35% in 2010. It is worth noting that the
criteria used for identifying EJ populations were modified between 2000 and 2010. Based on
2000 census data, foreign‐born populations were included as an indicator in addition to the
remaining three criteria used in 2010.
Figure 3.4 The number of people (pop) met the Environmental Justice (EJ) criteria in
Massachusetts (MA), the Boston Metropolitan Area (BMA) and the Charles River Watershed
(CRW) (Data source: MassGIS and Census 2010)
‐
500
1,000
1,500
2,000
2,500
MA BMA CRW
EJ Population
Thousands
35% Total
MA pop
35% Total MA EJ pop
33% Total BMA EJ pop
42
Figure 3.5 Location of the Environmental Justice (EJ) populations in the Boston
Metropolitan Area (BMA) and the Charles River Watershed (CRW)(Data source: MassGIS EJ
Populations Census 2010)
In sum, the Boston Metropolitan Area is aging, becoming more diverse in its younger
population, increasing in socio‐economic inequality, and requiring increased resources for
minority groups and immigrants, which were part of the Environmental Justice population
identified by the state of Massachusetts. The current trends of the demographic and socio‐
economic structure correlate with some of the key indicators of Social Vulnerability described in
the next section.
43
3.2.3 Social vulnerability and its measures
As described in Chapter 2, vulnerability is a complex concept used in hazard studies in
association with the potential impacts on the livelihood of human society as a result of disasters,
and is often measured by casualties and monetary losses. Social Vulnerability provides a lens
through which to study the dynamic socio‐economic structure that affects the abilities of certain
populations to respond to, adapt to, mitigate, and recover from disasters and associated socio‐
economic impacts, particularly under climate change impacts (i.e., limited access to and higher
prices for food, water, energy, transportation)(Lynn, MacKendrick, and Donoghue 2011; IPCC
2012, 2007).
There is neither a single definition nor a single measure for social vulnerability
commonly used in the literature. Nevertheless, there is a consensus that race, gender,
education, immigration status, and socio‐economic factors have resulted in limited access to
information, financial resources and political power and are closely related to coping capacity to
disasters (Blaikie 1994; Colten 2006; Satterfield, Mertz, and Slovic 2004; Yohe and Tol 2002).
Specifically, environmental inequality is correlated with flooding hazards, which suggests that
even though higher socio‐economic status groups tend to be exposed to flooding hazards, it is
the lower socio‐economic status groups who are most at risk (Fielding and Burningham 2005).
Furthermore, other indicators such as single female households, large family structures,
insufficient medical and social infrastructures for the elderly, low‐income groups, people who
work in services and occupations which are dependent on natural resources extractions, and
both rural and urban populations tend to increase Social Vulnerability (Cutter, Boruff, and
Shirley 2003; Lynn, MacKendrick, and Donoghue 2011; Tate, Cutter, and Berry 2010)(Table 3.2).
44
Table 3.2 Concepts and indicators with positive effects on social vulnerability and
corresponding Social Vulnarbility Index (SoVI) variables (modified from Cutter et. al 2003,
Cutter and Morath 2013, Tate et al. 2010)
Concept Descriptions of indicators with positive effects on social vulnerability
SoVI Variables (see Table 3.3)
Age The elderly and children are more susceptible to harm and require special care in response to disasters
MEDAGE, QKIDS, QPOP65O
Gender Females have lower wage and less access to resources and decision power
QFEMALE, QFEMLBR
Race/Ethnicity/ Immigration Status
Non‐white, people with English and culture isolation and immigrants have less access to resources and decision power
QBLACK, QINDIAN, QASIAN, QSPANISH, MIGRA
Education People with less education have less access to resources and decision power
QED12LES
Family Structure Single female households and large families require more resources to cope with disasters
QFHH, PPUNIT
Social Dependence/ Infrastructure Provision
People with special needs or who rely on social benefits require more resources to cope with disasters, in addition to insufficient access to medical facilities and physicians
QSSBEN, NRRESPC, HOSPTPC
Socio‐Economic Status
Poor and transient residents have less access to resources and decision power
QRENTER, PERCAP, MHSEVAL, M_C_RENT, PHYSICN, QCVLUN, QRICH, QPOVTY, QMOHO
Occupation Laborer, services‐based, and resource‐extraction based occupations are more likely to be affected by natural disasters and subsequent economic recovery
QCVLBR, QAGRI, QTRAN, QSERV
Urban/Rural Increases in urban population and housing density complicates evacuation and disaster management resources; Rural population tends to have less capacity for emergency response and greater dependency on a single‐sector or local‐based resource extraction economy
HODENT, *QRFRM, *QURBAN
*QRFRM (percent rural farm population) and QURBAN (percent urban population) are omitted in this study due to unavailability in the 2010 census
The Social Vulnerability Index (SoVI) developed by Cutter and her colleagues (2003) has
been widely employed in risk assessment in the United States (e.g., hazard and risk assessment
in South Carolina (SCEMD 2008) and California (CalEMA, 2010), the Integrated Hazards
45
Assessment Tool (IHAT) used by the Carolina Emergency Management Division
(mapra.cas.sc.edu/ihat), NOAA’s Digital Coast (www.csc.noaa.gov/digitalcoast/tools/slrviewer)).
The SoVI was specifically designed for cross‐temporal and cross‐spatial scale comparison. After
years of sensitivity and validation testing in a range of studies on local, regional and national
scales in the United States (Borden et al. 2007; Cutter, Boruff, and Shirley 2003; Cutter and Finch
2008), 32 socio‐economic variables were exclusively identified for constructing the SoVI (Cutter
and Morath 2013). In order to increase generalizability and comparability to other places, SoVI
was employed for measuring social vulnerability in the Charles River Watershed.
3.3 Methodology
3.3.1 Risk assessment framework
Birkmann (2006) reviewed several different conceptual frameworks and summarized six
different schools of thoughts based on how vulnerability was systematized into conceptual and
analytical risk and resilience framework (Birkmann 2006) (see Chapter 2). A common thread
that ties the different schools of thoughts together is the recognition that the probability of risk
to disasters is a result of interaction between exposure to probable hazards and vulnerability
inherent in society and physical environment in response to hazards: Risk = (Hazard) X
(Vulnerability). Therefore, this formula provides the basic framework for the long‐term climate
change‐induced flooding risk assessment in this study.
The long‐term flooding hazard index (HI) was defined as a probability of daily stream
flow volume exceeding the bankfull discharge based on a 45‐year period. The bankfull discharge
volume—the water volume that represents the maximum holding capacity of the river
channel—was determined by the 2‐year return period (Reed, Johnson, and Sweeney 2002).
46
SWAT Modeling
OUTPUT
Long‐Term Flooding Hazard Index (HI)
INPUT
Long‐Term Climate Change‐Induced
Flooding Risk Index (RI)
X
Social Vulnerability Index (SoVI)
=
30 Variables Social
Vulnerability Indicators
SoVI
method
Climate Change Impact
Temperature mean
Precipitation mean
Precipitation variation
Daily stream flow volume was simulated through hydrological modeling using Soil and Water
Assessment Tool (SWAT). Climate change impacts from the mean temperature, mean
precipitation and precipitation variation inputs were incorporated in the SWAT modeling.
Days when Q > Qbankfull HI = P (Q > Qbankfull) = (1)
365 days a year X 45 years P: Probability
Q: Simulated Stream baseflow (mm)
Qbankfull: Stream bankfull (mm) under current climate condition
Based on the assessment of climate sensitivity tests, climate change impact scenarios
were subsequently selected for further assessment using the spatial correlation with SoVI.
Finally, HI and SoVI were converted into the same unit of analysis and were multiplied to
formulate the long‐term climate change‐induced flooding risk index (RI) (Figure 3.6).
Figure 3.6 Risk assessment framework for the long‐term climate change‐induced flooding
hazards and social vulnerability in the Charles River Watershed
47
3.3.2 SWAT Modeling
Soil and Water Assessment Tool (SWAT)(TAMU 2011) was selected for this study for
several reasons. First, SWAT is a physically‐based (i.e. based on readily observed and measured
information in lieu of conceptual models) and spatially distributed model designed to simulate
water balance and stream flows on the watershed scale using a continuous daily time step for
long‐term yield (Arnold and Allen 1996). Second, SWAT has been successfully used to study
climate sensitivity and impacts on hydrology. For example, Wu (2007) employed SWAT and
climate sensitivity studies to investigate climate change impacts on drought versus average
periods in the Great Lakes watershed (Wu and Johnston 2007). In addition, multiple General
Circulation Models (GCMs) and IPCC climate change scenarios have been incorporated in SWAT
for studying hydrologic cycles, stream flows and water availability (Bekele and Knapp 2010;
Takle, Jha, and Anderson 2005). Incorporating climate change data into stream flow impact
studies is critical for this study. Moreover, SWAT has the capacity to work with ArcGIS, which is
the primary platform for conducting spatial analysis with SoVI (ArcSWAT 2009.93.7b available at
swat.tamu.edu/software/arcswat/). Combined with GIS, SWAT can simulate the temporal and
spatial variability of the hydrological processes at the subbasin level defined in the model
(Santhi et al. 2008). Finally, SWAT was developed through the USDA Agricultural Research
Service, gained international recognition, and had applications for studying water resources
management issues worldwide. Even though SWAT was originally developed for rural
environments in order to predict the effects of agriculture best management practices on water
quality (Arnold and Allen 1996), there has been an increase in case studies using SWAT for urban
and suburban watersheds (Fohrer et al. 2000; Richards et al. 2008). The flexibility of SWAT
modeling is therefore a plus for this study.
48
Model description
SWAT has eight components―hydrology, weather, sedimentation, soil temperature,
crop growth, nutrients, pesticides, and agricultural management. For the purpose of this study,
only hydrology and weather were applied. Specifically, stream outflow at each subbasin and
climate variability in temperature and precipitation were the key variables. The hydrological
cycle is modeled based on the water balance equation:
(2)
SWt: final soil water content
SWo: initial soil water content
t: time (days)
Rday: amount of precipitation
Qsurf: amount of surface runoff
Ea: amount of evapotranspiration
Wseep: amount of percolation and bypass flow existing soil profile
Qgw: amount of return flow (i.e., base flow originating from groundwater)
Each component in the hydrologic process can be simulated through its respective water
balance equation and methodology for surface runoff, evapotranspiration, soil water (i.e.,
percolation, bypass flow, lateral flow), and ground water (i.e., shallow aquifer and deep aquifer
flow). Streamflow in a main channel is determined by four major sources within each subbasin:
surface runoff, lateral flow, base flow from shallow aquifers, and transmission losses through
channel routing. The channel outflow is spatially connected between subbasins: the outflow
from the upper stream subbasin feeds into the lower stream subbasin. Within each subbasin,
the water balance is simulated within Hydrologic Response Units (HRUs). Each stream channel
49
outflow volume is the sum of the water balance in all HRUs within the respective subbasin.
Therefore, HRUs are not spatially connected.
The primary method used to determine surface runoff and infiltration rate is a Soil
Conservation Service Curve Number (SCS CN). Soil properties also affect the amount of water
which can be held in the soil. For example, “hydric soil” type A has the highest infiltration rate
and the lowest runoff potential whereas “hydric soil” type D has the lowest infiltration rate and
the highest runoff potential. In addition, percolation rate is influenced by the soil temperature
and saturated conductivity in the soil layer, which also affects the lateral subsurface flow.
Furthermore, the groundwater flow contributing to stream flow is modeled by simulating
shallow aquifer storage which is recharged through percolation. Finally, transmission losses due
to abstractions reduce runoff. Channel losses are a function of channel width, length, and flow
duration. The detailed theory and methodology for each step of the simulation can be found in
SWAT Theoretical Documentation Version 2009 (Neitsch et al. 2009).
Model input and data source
The first step to set up a watershed simulation is to define subbasins. The grid‐based
Digital Elevation Model (DEM) data was used for initial delineation of watershed and subbasin
boundaries. DEM data was based on a 30‐m grid generated by the USGS National Elevation
Dataset (NED), which has the highest quality and is the prime source for elevation information in
the United States (ned.usgs.gov/Ned/). In order to better match the size of census tracts (the
final unit for constructing flooding Risk Index), additional outflow points were manually added
on to the main channel. The median size of census tracts in the watershed, 1 km2, was used as
the minimum size for additional subbasins. Eventually, a total of 54 subbasins were delineated,
ranging from 0.5 to 35 km2 with a mean of 14 km2, in order to align with the size pattern of
50
census tracts that appear to be larger in the suburban communities and smaller in Boston and
its surrounding communities (Figure 3.7).
The HRU was determined based on land uses and soils. SWAT has a built‐in database for
land use categories in urban areas and for various agricultural crops. Each SWAT land use has an
associated SCS CN varied by “hydric soil” type. Four major urban land use categories (UCOM,
URHD, URMD, URLD) and four non‐urban SWAT land use categories (AGRL, FRST, WETL, WATR))
adopted from the SWAT urban database were applied in this study. Land use data came from a
state‐wide dataset based on 0.5m resolution digital ortho imagery captured in April 2005. This
was the latest publicly accessible digital land use and land cover data available from MassGIS. A
total of 33 categories identified in the MassGIS Land Use 2005 dataset were then aggregated
based on the similarity of the development intensity and characteristics of urban and non‐urban
general land use (i.e. general agricultural and general forest mix) and land cover types (i.e.,
grassy, forest, wetlands and water) for SWAT land use categories. Since this watershed is highly
urbanized and plant and crop types were not critical to this study, especially considering that the
combined agricultural and recreational land areas comprised 6% of the watershed area,
agricultural and recreational land uses were combined and assumed to have similar grassy land
cover characteristics for the purpose of modeling the streamflow in this study (Table 3.3).
Fi
ce
igure 3.7
ensus tract p
Location an
atterns in the
nd size of the
e Charles Riv
51
e subbasins (l
er Watershed
abeled with
d
numbers) compared to th
he
52
Table 3.3 SWAT 2005 land use categories
SWAT Land Use
Descriptions MassGIS Land Use 2005 (LU05_DESC)
Total Area (acres/ha)
% Watershed Area
AGRL Agricultural; Non‐forest open space; Grassy land cover
• Agriculture – Orchard – Nursery – Cropland – Pasture
• Recreational – Participation Recreation – Spectator Recreation – Golf Course – Cemetery
11,943/ 4,833
6.3
FRST Forest • Forest – Forest – Bushland/Successional
68,271/ 27,628
35.8
WETL Wetlands • Wetlands and Water – Non‐Forested Wetland – Saltwater Wetland – Cranberry Bog – Forested Wetland
20,437/ 8,271
10.8
UCOM Commercial • Commercial – Commercial – Industrial – Transitional – Junkyard – Open Land
• Urban Public/Institutional • Transportation • Utilities
– Mining – Waste Disposal – Powerline/Utility
24,501/ 9,915
12.9
URHD Residential‐High Density
• Residential – Multi‐Family – High Density
22,816/ 9,233
12.0
URMD Residential‐Medium Density
• Residential – Medium Density
16,652/ 6,739
8.7
URLD Residential‐Low Density
• Residential – Low Density – Very Low Density
20,234/ 8,189
10.6
WATR Water • Water
5,634/ 2,280
3.0
53
Soil data was derived from the United States Department of Agriculture (USDA) Natural
Resources Conservation Service (NRCS). The SSURGO‐certified soil datasets have met all
standards and requirements approved by the NRCS and possess the most detailed information
developed by the National Cooperative Soil Survey. The hydric soil groups used for delineating
HRUs included hydric soil group A (35% of watershed area), group B (30% of water shed area),
group C (24% of water shed area), and group D (11% of water shed area). A total of 1470 HRUs
were then identified with a combination of SWAT land use type and hydric soil group properties.
Daily observed weather data between 1990 and 2011 was obtained from the National
Climatic Data Center (NCDC) in three stations— Walpole 2 (USC00198757), East Milton Blue Hill
Observatory (USW00014753), and Boston Logan International Airport (USW00014739). The
three variables used to calibrate the watershed were maximum temperature, minimum
temperature, and total precipitation. The completeness of daily records poses the greatest
constraint for gathering observed weather data. The weather stations were chosen because
the complete historical daily records in addition to their locations were immediately due south
and in close proximity to the upper, middle and lower basins respectively.
Model calibration and validation
Calibration is the process of adjusting model parameters to minimize the difference
between simulated output and observed values (NRC 2009). Validation is the process of using
part of the dataset as an input in the calibrated model in order to compare the calibrated model
results with the observed values. The observed daily streamflow data between 1990 and 2011
was obtained from the United States Geological Survey (USGS) database at stream gage number
01104500 located in Waltham, Massachusetts, at the lower basin along the Charles River main
stream. This study used a 2‐year warm‐up period (1990 to 1991), a 14‐year calibration period
(1
m
an
m
an
Fi
3.
fl
fr
a
Pa
Stream
flow
1992 to 2005)
model efficien
nd 0.93 for th
model in its re
nd 0.6 is cons
igure 3.8
.3.3 Climate
Two m
ooding hazar
rom General C
global scale,
anel on Clima
) and a 6‐yea
cy coefficient
he validation
semblance to
sidered a goo
SWAT calib
change asses
major approac
ds in river ba
Circulation M
which are illu
ate Change (I
r validation p
t (NSE) (Nash
period, which
o the basin pr
od model fit (M
bration and v
ssment meth
ches have bee
sins under cli
Models (GCMs
ustrated in th
PCC) (IPCC 20
54
period (2006 t
and Sutcliffe
h represented
roperties. A N
Moriasi et al.
validation hyd
hods
en employed
imate change
s) according to
he Special Rep
007; Prudhom
Yea
to 2011). The
e, 1970) were
d a high level
NSE value of 1
2007) (Figure
drograph in t
to study stre
e impacts. On
o the greenho
ports on Emis
mme et al. 201
ar
e results of th
0.81 for the
of confidenc
1 represents
e 3.8).
he Charles Ri
eam flows and
ne approach i
ouse gas emis
ssion by the In
10). The adva
he Nash‐Sutcl
calibration pe
ce in the simu
a perfect mat
iver Watersh
d associated
is scenario‐ba
ssion scenario
ntergovernme
antage of usin
iffe
eriod
ulated
tch
ed
ased
os on
ental
ng an
55
IPCC scenario‐based approach is that it has been widely accepted in science and policy realms as
a top‐down assumption for climate change. Therefore, using IPCC scenarios is considered a
defensible and politically accepted method for studying climate change impacts (Praskievicz and
Chang 2009). Nevertheless, spatial mismatch and uncertainty inherent in major GCMs are
known to scientists to exist (Carter et al. 1994; Gates 1985). In order to study local watersheds
as small as the Charles River Watershed (about 800 km2), recent efforts have been made to
understand the range of uncertainty among downscaling methods (e.g., Regional Climate
Models (RCMs), statistical downscaling (SD)) in order to reflect climate change on the local scale
(Chiew et al. 2003; Fowler, Kilsby, and Stunell 2007). For example, hydrological impacts on the
25,000 km2 Lule River Basin using both GCMs and RCMs were found similar patterns in the
increase of mean stream flow and earlier spring peak flow but also found variations in seasonal
dynamics (Graham et al. 2007). In addition, Chen et al. (2012) evaluated six different
downscaling methods on the hydrological impacts on a regional basin (24,600 km2) and found
that the results varied significantly depending on which downscaling method was chosen (Chen,
Brissette, and Leconte 2011). Therefore, the wide range of uncertainty among various GCMs
and within the different downscaling methods remains a drawback of this “top‐down” approach
(Brown et al. 2011; Praskievicz and Chang 2009).
The other approach is scenario‐neutral and uses synthetic weather generation data for
climate sensitivity tests (Prudhomme et al. 2010). This “bottom‐up” approach is advantageous
for a grounded understanding of climate variability impacts on the study subject such as
stormwater runoff and flooding hazards in a local basin (Nemec and Schaake 1982; Skiles and
Hanson 1994), which is suitable for a physical environment vulnerability assessment (Brown et
al. 2011). The disadvantage lies in a lack of reflection on the probable future global emission
scenarios and climate change projections (Praskievicz and Chang 2009).
56
This study adopted the merits of both approaches to first identify climate sensitivity and
trends in long‐term flooding hazards in the watershed, then designating climate change
scenarios based on IPCC scenarios and GCMs projections.
Climate sensitivity tests
A total of 150 climate sensitivity tests were generated from a combination of three
climate variables: mean temperature change (0, +1, +2, +3, + 4, + 5˚C), mean precipitation
change (0%, ±10%, ±20%) and variation in precipitation change (0%, ±10%, ±20%). The method
for generating the 150 weather data sets for a theoretical 50‐year timeframe is described in a
separate article (Steinschneider and Brown in review). The data set was applied to the
calibrated and validated SWAT model using a 45‐year daily time step in addition to a 4‐year
warm‐up period between 2050 and 2099. The outcomes of simulated stream outflow volumes
from 150 climate sensitivity tests were then used to construct the long‐term climate change‐
induced flooding hazard index (HI).
The results of HI illustrated that 104 out of 150 (69.3%) climate sensitivity tests had no
increase in probability of flooding hazards when compared to current climate conditions while
the remaining 46 climate sensitivity tests revealed positive effects, which excluded variables of
5˚C increase in mean temperature and ‐10% and ‐20% in mean precipitation change (Figure 3.9).
Increasing temperature had negative effects; when temperature increased to 5°C, HI became
lower than current climate conditions, regardless precipitation change. When mean
temperature was controlled at an incremental increase of 1°C, both increasing mean
precipitation and precipitation variation had positive effects (Figure 3.10). When precipitation
variation change was controlled, increasing mean temperature and precipitation had similar
patterns, which indicated slightly positive effects from precipitation variation (Figure 3.11). On
57
the other hand, mean precipitation illustrated significant effects when increasing mean
temperature and precipitation variation patterns altered drastically under the increased change
in the mean precipitation variable (Figure 3.12). It is worth noting that the effects of the
interaction between precipitation and temperature changes did not have consistent trends,
particularly when the mean temperature increased to 4°C and the precipitation variation
increased by 10%—HI values did not vary from a 0% increase in precipitation variation(Figure
3.10 and 3.11).
Figure 3.9 Histogram of the number of climate sensitivity tests in a range of long‐term
flooding hazard indices
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0
5
10
15
20
25
30
35
40
45
Number of Clim
ate Sensitivity Tests
Long‐Term Flooding Hazard Index (HI)
current climate conditions
58
Figure 3.10 Climate sensitivity tests results in long‐term flooding hazard index (HI)
controlled by mean temperature change
80% 90% 100% 110% 120%80%
90%
100%
110%
120%
80% 90% 100% 110% 120%80%
90%
100%
110%
120%
80% 90% 100% 110% 120%80%
90%
100%
110%
120%
80% 90% 100% 110% 120%80%
90%
100%
110%
120%
80% 90% 100% 110% 120%80%
90%
100%
110%
120%
80% 90% 100% 110% 120%80%
90%
100%
110%
120%
(a)
Long‐Term Flooding Hazard Index (HI)
(b)
(c) (f)
(e)
(d)
X axis: Precipitation variation changeY axis: Mean precipitation change
0.0%‐1.0% 1.0%‐2.0%2.0%‐3.0% 3.0%‐4.0%4.0%‐5.0%
mean temperature change at (a) +0°C (b) +1°C (c) +2°C (d) +3°C (e) +4°C (f) +5°C
59
Figure 3.11 Climate sensitivity tests results in long‐term flooding hazard index (HI)
controlled by precipitation variation change
0 1 2 3 4 580%
90%
100%
110%
120%
0 1 2 3 4 580%
90%
100%
110%
120%
0 1 2 3 4 580%
90%
100%
110%
120%
0 1 2 3 4 580%
90%
100%
110%
120%
0 1 2 3 4 580%
90%
100%
110%
120%
(a)
Long‐Term Flooding Hazard Index (HI)
(b)
(c)
(e)
(d)
X axis: Mean temperature change (°C)Y axis: Mean precipitation change
0.0%‐1.0% 1.0%‐2.0%2.0%‐3.0% 3.0%‐4.0%4.0%‐5.0%
precipitation variation change at
(a) ‐20%(b)‐10% (c) +0%
(d) +10%(e) +20%
60
.
Figure 3.12 Climate sensitivity tests results in long‐term flooding hazard index (HI)
controlled by mean precipitation change
0 1 2 3 4 580%
90%
100%
110%
120%
0 1 2 3 4 580%
90%
100%
110%
120%
0 1 2 3 4 580%
90%
100%
110%
120%
0 1 2 3 4 580%
90%
100%
110%
120%
0 1 2 3 4 580%
90%
100%
110%
120%
(a)
Long‐Term Flooding Hazard Index (HI)
(b)
(c)
(e)
(d)
X axis: Mean temperature change (°C)Y axis: Precipitation variation change
0.0%‐1.0% 1.0%‐2.0%2.0%‐3.0% 3.0%‐4.0%4.0%‐5.0%
mean precipitation change at
(a) ‐20%(b)‐10% (c) +0%
(d) +10%(e) +20%
C
sc
p
ha
sc
sc
2
3
Fi
20
U
th
si
es
limate chang
Since t
cenarios were
rojections in
ave an increa
cenario (IPCC
cenario (IPCC
100 from an a
.13).
igure 3.13
006, Figure 3
Accord
niversity of S
here exists a w
gnificant pos
stimated abo
ge impact sce
this study foc
e selected fro
the New Eng
ase in mean te
‐B1) and an in
‐A1fi) based o
average of se
Projected a
)
ding to the do
Southern Miss
wide range of
itive trend of
out an increas
narios
cused on flood
om climate va
land region.
emperature b
ncrease betw
on 1961‐1990
everal GCMs (
annual mean
ownscaled GC
sissippi, and t
f annual mea
f an increase
se of 10% bet
61
ding rather th
ariables with p
Based on IPC
between 1.3 a
ween 1.4 and
0 historical da
(i.e., GFDL, Ha
temperature
CMs prepared
the Nature Co
n precipitatio
of 0.125% a y
ween 2050 a
Yea
han drought h
positive effec
CC scenarios, t
and 2.8 ˚C (2.
5.3 ˚C (2.6 an
ata and data
adCM3 and P
e change und
d by the Univ
onservancy (c
on change fro
year from 195
nd 2100 (Figu
ar
hazards, clima
cts on HI in ad
the Northeas
.4 and 5˚F) in
nd 9.5˚F) in a
projection th
CM) (NECIA 2
der IPCC scen
ersity of Was
climatewizard
om 80% to 12
50 through 21
ure 3.14).
ate change im
ddition to GCM
st region is lik
a low emissio
high emission
rough the ye
2006) (Figure
arios (NECIA,
shington, the
dcustom.org),
0% with a
100. It is
mpact
Ms
kely to
on
n
ar
,
,
Fi
fr
m
fr
w
te
an
va
igure 3.14
rom climatew
In add
medium and h
rom this study
when the mea
emperature in
nd 3˚C variab
ariation in pre
Annual ave
wizardcustom
ition to curre
high impacts o
y and project
n temperatu
ncrease betw
bles were cho
ecipitation (+
erage precipit
m.org)
ent conditions
on HI based o
ted GCMs mo
re increased t
ween 1.3 and 5
sen for const
+10% and +20
62
tation change
s, three clima
on the integra
odels from the
to 4˚C in add
5.3˚C by the e
tructing scena
0%).
e from down
ate change sce
ation of empir
e literature. S
ition to GCMs
end of this ce
arios in addit
scaled GCM m
enarios were
rical climate s
Since the HI b
s’ projection
entury (NECIA
tion to increas
models (prod
identified at
sensitivity tes
became insen
of a probable
A 2006), only +
sed mean and
duced
low,
sts
nsitive
e
+1, 2
d
63
Climate change impact scenarios were then defined from the combinations of climate
variables as the following: current climate conditions with a combination of 0˚C in mean
temperature, 0% mean precipitation and 0% precipitation variation; Low Impact climate change
scenario with a combination of +3˚C, +10% mean precipitation and 0% precipitation variation;
Medium Impact climate change scenario with a combination of +2˚C, +10% mean precipitation
and +10% precipitation variation; High Impact climate change scenario with a combination of
+1˚C, +20% mean precipitation and +20% precipitation variation. Low Impact and Medium
Impact climate change scenarios are within the range of GCMs projections. High Impact climate
change scenario represents an extreme condition that will have significant effects on long‐term
flooding hazards (Figure 3.15).
Figure 3.15 Climate change impact scenarios
64
3.3.4 Social Vulnerability Index (SoVI)
Social Vulnerability Index (SoVI) was constructed using 30 variables from the latest 2010
census data and the 2010 American Community Survey (ACS) data at the census tract level.
Data for a total of 376 census tracts in the 35 municipalities that contain the Charles River
Watershed were collected. Percent urban and percent rural populations were two variables
omitted from the original list of 32 variables (Cutter and Morath 2013) due to unavailability
from the changes of survey methods in the 2010 census. Census tracts with missing data were
excluded from the calculation. As a result, a total of 352 valid census tracts were used for
constructing SoVI (Table 3.4).
Several steps were taken to construct SoVI using IBM SPSS v21.0.0. First, the variables
were normalized as either percentage, per capita or density function. Second, the variables
were standardized through z‐score with a mean of 0 and standard deviation of 1. Third,
principal component analysis was conducted using a varimax rotation and Kaiser criterion for
component selection. Then variables with highly correlated values—greater than 0.5 or less
than ‐0.5—were selected as influential variables and were listed within each component. Each
component illustrated a variance that helped to describe the concept of Social Vulnerability.
The first identified component was composed of the most variables and then the number of
variables decreased as more components were identified. The results revealed eight
components that explained 72% variance among the variables for describing Social Vulnerability
(Table 3.5). The final step was to sum all components based on their positive (add to) or
negative (subtract from) influence on Social Vulnerability from the identified concepts (Table
3.2). After constructing SoVI, 277 census tracts that intersected with the Charles River
watershed boundary were selected using the ESRI GIS “Clip” function. The SoVI scores ranged
65
between ‐30 (low Social Vulnerability) and 21 (high Social Vulnerability) with a mean of ‐0.4 and
a standard deviation of 6.2. It is worth noting that the SoVI scores are in relative values rather
than absolute. For instance, a census tract with a SoVI score of 10 has relatively higher—not ten
times higher—Social Vulnerability than a census tract with SoVI equals to 1.
Table 3.4 Descriptive statistics of Social Vulnerability Index (SoVI) variables (N: number
of valid census tracts; Sources: a) 2010 ACS‐5yr, b) Census 2010, c) Mass GIS 2007)
Variables Description N Min. Max. Mean SD Source
HODENT No. of housing units per sq. mile
369 .000 61247 7614 8980 A
HOSPTPC Per capita no. of community hospitals
369 .000 .0024 .0001 .0002 C
M_C_RENT Median monthly contract rent ($)
362 286 2000 1161 371 A
MEDAGE Median age 369 16 50 35 7 BMHVAL Median value of
owner occupied housing units
358 149600 1000000 491358 175570 A
MIGRA % foreign born citizens immigrating in 1990‐2000
369 .000 .405 .081 .069 a
NRRESPC Per capita residents in nursing homes
369 .000 .113 .005 .014 a
PERCAP Per capita income ($)
369 5572 124093 41042 20479 a
PHYSICN No. of persons per 100000 employed in healthcare & technical jobs
369 .000 .004 .001 .001 a
PPUNIT Ave. no. of people per household
363 1.300 4.110 2.392 .427 b
QAGRI % employment in farming, fishing, & forestry
369 .000 .032 .001 .003 a
QASIAN % Asian & Hawaiian Islanders
369 .000 .663 .088 .083 b
QBLACK % African American 369 .000 .942 .134 .213 bQCVLBR % population in the
labor force 369 14.600 95.700 69.139 10.559 a
QCVLUN % unemployed civilian
369 .000 23.400 4.900 3.666 a
66
QED12LES % population >= 25 years old with no high school diploma
363 .000 .604 .105 .108 a
QFEMALE % female population 369 .081 .735 .514 .064 bQFEMLBR % females in the
labor force 369 .081 .760 .521 .070 a
QFHH % single female headed households
368 .000 .513 .124 .108 a
QINDIAN % Native American 369 .000 .080 .002 .006 bQKIDS % population <5
years old 369 .000 .139 .053 .025 b
QMOHO % of mobile homes 368 .000 .031 .001 .003 aQPOP65O % population >= 65
years old 369 .000 .284 .114 .052 b
QPOVTY % population below poverty line
369 .000 .863 .145 .139 a
QRENTER % renter occupied housing units
368 .000 1.000 .501 .265 a
QRICH % households earning >=$100,000
369 .000 79.100 33.837 18.533 a
QSERV % employed in service industry
369 .000 1.000 .167 .118 a
QSPANISH % Hispanic 369 .006 1.197 .119 .139 bQSSBEN % population
collecting social security benefits
368 .000 .500 .211 .080 a
QTRAN % employed in transportation, communications, & other public utilities
369 .000 .119 .038 .023 a
67
Table 3.5 Principal component analysis of Social Vulnerability Index
Component Component Cardinality
Concepts (see Table 3.2)
Accumulative variance explained
Dominant variables
Correlation
1 (+) Socio‐Economic 27.07 PERCAP ‐0.87 Status QRICH ‐0.86 Race MHVAL ‐0.69 Education M_C_RENT ‐0.66 Occupation PHYSICN ‐0.63 QRENTER 0.60 QPOVTY 0.66 QBLACK 0.70 QCVLUN 0.70 QSPANISH 0.77 QFHH 0.79 QED12LES 0.85 QSERV 0.88
2 (+) Age/Socio‐ 39.95 HODENT ‐0.76 Economic QRENTER ‐0.68 Status MEDAGE 0.60 QKIDS 0.60
3 (+) Gender 47.35 QFEMALE 0.95 QFEMLBR 0.97
4 (+) Occupation 52.96 QCVLBR ‐0.88
5 (+) Age/Social 58.23 QPOP65O 0.52 Dependence NRRESPC 0.68
6 (+) Race/ 62.99 QMOHO ‐0.76 Immigration QASIAN 0.51 Status MIGRA 0.51
7 (+) Ethnicity/ 67.63 QINDIAN 0.55 Occupation QAGRI 0.64
8 (‐) Infrastructure Provision
71.94 HOSPTPC 0.61
68
3.3.5 Long‐term climate change‐induced flooding risk index (RI)
The long‐term climate change‐induced flooding risk index (RI) was calculated by
multiplying the Social Vulnerability Index (SoVI) with the long‐term flooding hazard index (HI).
Because SoVI was based on the unit of census tract and HI was derived from the SWAT modeling
at the unit of subbasin, several steps were taken to unify the unit of analysis at the census tract
level for constructing the RI. First, subbasin units were intersected with census tract units using
the ESRI GIS “Clip” function. Since census tract was based on population density varied between
1000 and 8000 people with an optimum of 4000 people (see glossary in US Census Bureau,
www.census.gov/dmd/www/glossary.html), the size of census tract varied widely with a range
from 0.04 and 46.54 km2 and a mean of 3.67 km2. Some census tracts are as large as one entire
community (e.g., Sherborn, Dover, Millis) in rural areas, while others are as small as a few blocks
in the cities such as Boston. An area‐weighted method was therefore applied in census tracts
that contained multiple subbasins for calculating respective HI values.
HICensus Tract Y = ∑xi [HISubbasin Xi*(Areasubbasin Xi in census tract/Areasubbasin Xi)] (3)
Census Tract Y = census tract that intersects with multiple subbasins
Subbsin Xi = Each subbasin intersected with Census Tract Y
Moreover, in order to have a positive score in hazard indices in order to reflect positive scores of
risk indices, SoVI scores were rescaled from 0 to 1 (mean=0.6, SD=0.1). Finally, the RI was
created by multiplying the weighted HI and the rescaled SoVI scores in each census tract.
Long‐Term Climate Change‐Induced Flooding Risk Index (RI) = HI * SoVI (4)
69
3.4 Results
3.4.1 HI and SoVI spatial overlay
The long‐term flooding hazard index (HI) at the outlet of each subbasin represented
accumulated streamflows from the entire drainage area that included hydrologically‐connected
upper‐stream subbasins. Under current climate conditions, communities at the upper and
lower ends of the stream such as Boston, Waltham, Lincoln and lower Brookline were exposed
to lower HI whereas higher HIs were clustered at the upper middle streams in Medway and
Millis as well as at the lower middle streams in Dover and most of Needham. Under climate
change impact scenarios, HI represented the same spatial patterns with increased intensity
across Low Impact, Medium Impact and High Impact climate change scenarios. High Impact
illustrated more than double the probabilities of long‐term flooding hazards than in other
climate change impact scenarios.
The Social Vulnerability Index (SoVI) in general was concentrated in urbanized areas. It
appeared to be lower in the upper Charles River Watershed in towns of Norfolk, Franklin, Millis,
Medway, Holliston and Medfield, and higher in the lower basin in the cities of Boston, Brookline,
Newton, Needham, Wellesley and Weston.
When overlaying SoVI with HI, areas with relatively higher SoVI at the lower basin
correlated with the highest HI along the main river channel. Under the High Impact climate
change scenario, the communities with relatively higher social vulnerability in the suburbs of
Boston would encounter the most long‐term flooding hazards induced by climate change (Figure
3.16).
70
Figure 3.16 Spatial overlay between the long‐term flooding hazard index (HI) and the
Social Vulnerability Index (SoVI)
3.4.2 Long‐term climate change‐induced flooding risk assessment
The long‐term climate change‐induced flooding risk index (RI) could be considered as a
SoVI‐weighed HI resulting from multiplying the rescaled SoVI (index values from 0 to 1) and the
HI two indices together. Therefore, census tracts with a SoVI score of zero represented the
lowest RI, implying that places with no social vulnerability were considered to be at minimal risk
71
because the concept of risk was only associated with the livelihood of human beings. On the
other hand, a high RI would reflect people who had a high Social Vulnerability and who were
also exposed to high long‐term flooding hazards. Since the HI score was based on probabilities,
a RI score could be interpreted as an absolute term. For example, RI increased by more than
double under the High Impact climate change scenario when compared to the current climate
conditions and Low Impact climate change scenarios (Table 3.6). High risk indices were
concentrated around suburban towns, including Weston, Wellesley and Needham at the middle
to lower basin under the current climate conditions and Low Impact climate change scenarios.
The upper basin had higher risks when greater impacts from climate change posed higher
flooding hazards under Medium and High Impact climate change scenarios (Figure 3.17).
Table 3.6 Summary of descriptive statistics of long‐term climate change‐induced
flooding risk index (RI) among climate change impact scenarios
Mean Median Standard Deviation
Minimum Maximum
current climate conditions 0.0049 0.0046 0.0021 0 0.0119
Low Impact 0.0055 0.0053 0.0023 0 0.0134
Medium Impact 0.0065 0.0062 0.0028 0 0.0165
High Impact 0.0112 0.0102 0.0058 0 0.0317
72
Figure 3.17 Maps of the long‐term flooding risk index (RI) among climate change impact
scenarios
3.5 Discussion
3.5.1 Climate change impacts on the HI
Results from the climate sensitivity tests have partially rejected the hypothesis that
increasing temperature and increasing precipitation would increase long‐term flooding hazards.
73
This sutdy revealed that the increasing mean temperature had negative effects and the
increasing mean precipitation had positive effects on steamflows. The findings were supported
by the previous studies that also used SWAT modeling and climate sensitivity tests (Bekele and
Knapp 2010; Fontaine et al. 2001). In addition, increasing temperature would increase
evapotranspiration rate and decrease water yield and baseflows (Praskievicz and Chang 2009),
regardless land cover types were forest‐dominated (Fontaine et al. 2001) or agriculture‐
dominated (Ficklin et al. 2009). Since the long‐term flooding hazard index in this study was
based on streamflow volume, the decreased streamflow would result in the reduced HI.
3.5.2 HI and SoVI HotSpots spatial mismatch
The long‐term flooding hazard index (HI) and the Social Vulnerability Index (SoVI) were
significantly clustered respectively. However, their “HotSpots” did not match spatially. Spatial
autocorrelation analysis was a function in the ESRI GIS which was used for testing significance in
the spatial patterns. The HI values between 54 sub‐watersheds under current climate
conditions (Moran’s I=0.34, p <0.001, z score=3.7 SD), Low Impact (Moran’s I=0.28, p<0.001, z
score=22.4 SD), Medium Impact (Moran’s I=0.30, p<0.001, z score=23.7 SD) and High Impact
(Moran’s I=0.34, p<0.001, z score=26.7 SD) climate change scenarios were all significantly
clustered. So did the significant spatial pattern in the SoVI (Moran’s I = 0.14, p <0.001, z score =
14.8 SD). The clustered HI pattern suggested higher long‐term flooding hazardous places tend
to locate in close proximity; the clustered SoVI pattern suggested that people with higher social
vulnerability tend to live near each other.
Getis‐Ord Gi*Hot Spots analysis, a spatial analysis function in the ESRI GIS using polygon
contiguity (First Order) and euclidean distance methodologies, was applied for testing spatial
relationship. The HotSpots represent high‐and‐high and low‐and‐low values that are clustered
sp
ch
B
cl
Fi
be
cl
re
18
po
su
W
ha
(a
patially. The
hannel in all c
oston’s Dorch
lustered in th
igure 3.18
etween (a) th
Statist
limate condit
evealed no sig
867.55). The
opulation wit
uggesting tha
Watershed cur
azardous are
a)
results illustr
climate chang
hester neighb
e town of No
The Getis‐
he long‐term
tically, there w
ions). In add
gnificant relat
refore, the sp
th high social
at the people
rrently do not
as.
rated that the
ge impact sce
borhoods out
orfolk in the u
Ord Gi*Hot S
flooding haz
was no strong
ition, spatial
tionship betw
patial mismat
vulnerability
who are cons
t live in close
(b)
74
e HI HotSpots
enarios. The h
tside the main
upper basin (F
Spots spatial
zard index an
g correlation
regression an
ween the two
tch between
y would live in
sidered the m
proximity wi
were located
high SoVI Hot
n channel and
Figure 3.18).
analyses reve
nd (b) the Soc
between the
nalysis using G
o indices (R‐sq
HI and SoVI r
n high long‐te
most socially v
ith the highes
d along the C
tSpots were c
d the low SoV
ealed spatial
cial Vulnerabi
HI and SoVI (
GeoDa (Anse
quared=0.02,
ejected the h
erm flooding h
vulnerable in
st long‐term f
harles River m
clustered in
VI HotSpots w
mismatch
ility Index
(r=‐.203 in cu
lin et al. 1996
Log likelihoo
hypothesis tha
hazard areas,
the Charles R
flooding
main
were
urrent
6)
d= ‐
at
River
75
It is worth emphasizing that this study is locally‐based and watershed‐specific
assessment. Boston is a historical harbor city and is not located along the main channel of the
Charles River. The densely developed cities such as Boston are susceptible to other types of
flooding such as flash floods, which are highly correlated with impervious surfaces in the
urbanized watersheds (Praskievicz and Chang 2009; Bodini and Cossu 2010); nevertheless, the
long‐term riparian flooding hazards occur most along the main stream. Therefore, even though
the city has the highest social vulnerability, this study suggests that the long‐term climate
change‐induced riparian flooding hazard is not the top concern for Boston in the Charles River
Watershed.
3.5.3 Growing risks in suburbs
Even though the HosSpots analyses revealed the mismatch between the HI and SoVI,
census tracts with the highest HI are among the ones with moderate to high social vulnerability
outside Boston and in the suburban communities showing high risk indices in the middle and
upper basin (Figure 3.17). Therefore, there is a greater concern for suburban and rural
communities about long‐term flooding hazards under climate change impacts. The Boston
Metropolitan Area has had an increase of foreign‐born populations and an increase of poverty
rate, which are highly correlated with SoVI, in the suburban and growing communities such as
the towns of Wellesley and Milford. If the trends continue, those areas could potentially
become high SoVI HotSpots.
By no means, it is critical to use the risk index as an integration of both hazards and
social vulnerability in risk assessment. For example, Wellesley is located in the high HI HotSpots
area with moderate SoVI, which then resulted in a high RI. The RI is then helpful to gauge the
climate change impacts on the long‐term flooding risks in the Charles River Watershed,
76
considering growing social vulnerability and growing hazards in the suburban and growing
communities. In addition to flooding, climate change related hazards such extreme heat have
been recorded as having increased impacts on the rural communities with higher social
vulnerability, which is correlated with higher percentage of aging population and lower access to
infrastructure and resources for risk management (Oven et al. 2012; Reid et al. 2009; Sheridan
and Dolney 2003). Furthermore, human disturbances (i.e. deforestation, grazing, urbanization)
resulting in land cover change contribute to alteration of hydrological cycle (Price 2011) that
eventually have impacts on flooding (Jones et al. 2012). Increasing urbanization in the suburban
and rural communities can therefore impose more long‐term climate change‐induced flooding
risks in those areas in the Charles River Watershed. The impact from urbanization is discussed in
Chapter 4.
3.5.4 Limitation of this study
Despite the findings have revealed effects of individual variable in temperature and
precipitation change on long‐term flooding hazards, the causality between the increased
temperature and the increased precipitation and the muticolinearity among climate variables
related to uncertainty in climate change were not included in this study. In addition, this study
has examined an average of a 45‐year period rather than seasonal effects on flooding hazards.
For example, research has shown that climate change has impacts on the increased flooding in
winter due to the early snow melt and precipitation falling as rain instead of snow, in addition to
the increased evapotranspiration and droughts in summer (Jones et al. 2012; Karl, Melillo, and
Peterson 2009). Furthermore, this study was limited to single hazard rather than multi‐hazards
risk assessment. Therefore, even though the most socially vulnerable people who live in the
southeast Boston neighborhoods are not most threatened by the long‐term flooding hazards,
77
those population who live mostly in low lying areas along the coasts are prone to other types of
climate change‐induced hazards such as rising sea levels and more frequent and intense
hurricanes and storm surges (Kirshen, Knee, and Ruth 2008).
Furthermore, the Social Vulnerability Index (SoVI) has its limitation of using publicly
available data that are in quantitative nature and lack of qualitative quality. Qualitative data is
essential in order to understand perceptions and behavior that are relevant to become
vulnerable to flooding and other types of climate change‐induced hazards. For example, people
tend to underestimate the probability of flooding hazards or having a lack of accurate
information on the risk due to false security by flooding control structures such as levees and
dams (Hung, Shaw, and Kobayashi 2007; Zhang 2010). In addition, the hazardous places such as
coastline and waterfront properties are often associated with natural beauty and desirable
places to live regardless known probability to disasters; as a result, the wealthy people who can
afford insurance tend to voluntarily live in hazard‐prone areas (Blaikie 1994; Platt 1999).
Therefore, future research for gathering qualitative data is critical to understand why people live
in hazard‐prone areas and how people take actions in coping with climate change.
In sum, further studies in people’s behavior and actions toward flooding mitigation and
climate change adaption strategies will play a critical role in the success of climate change
planning. Finally, integrating social vulnerability with local‐based and multiple climate change‐
induced hazards (e.g., flash floods, droughts, extreme heat and cold, tornados, hurricanes,
severe storms, sea level rises) is essential for the development of comprehensive climate change
adaptation strategies to address further needs for place‐based community planning.
78
3.6 Conclusion
Hydrological systems are complex and dynamic in the urbanized watersheds. The
multicolinear interactions between temperature, precipitation, land cover, soil, topography, and
channel characteristics make streamline planning strategies for climate change difficult
particularly at a local scale. It is critical to employ empirical studies to demystify climate change
impacts on local communities and using place‐based risk assessment in order to better
understand local‐based hazards and social vulnerability. Subsequently, planners and policy‐
makers can prioritize resources to where make most impacts for reducing climate change‐
induced risks. In addition, climate sensitivity tests used in this study have shown valuable
parameters for climate change planning decision‐making on a local basis. For example,
according to the GCMs projections, the mean temperature would increase to 3°C and mean
precipitation would increase to 10%, which fits the Low Impact climate change scenario defined
in this study. Based on the findings in this study, lower increase in long‐term flooding hazards
would likely occur under climate change impacts when land cover remains unchanged.
Therefore, planners can use the Low Impact climate change scenario as a framework for
scenario planning in developing strategies for long‐term flooding hazards mitigation. In
addition, planners can use the SoVI and the RI for identifying local communities who are at the
most risks and who are in high social vulnerability in order to prioritize resources and enhance
resilience in those communities. Planners can also use SoVI as a lens in integrating socio‐
economic and vulnerability data into the comprehensive planning process. As climate change
emerges as a top global as well as local agenda, its impact on local community needs to be
addressed in local climate change planning.
79
Finally, quantitative assessment is an initial step in serving as a factual basis for hazard
mitigation and climate change planning policies. In order to understand people’s attitudes and
their choices to live in hazard‐prone areas, as well as what actions people would take for climate
change adaptation, it is critical to integrate risk perceptions and local knowledge from engaging
the stakeholders and the public in the participatory planning processes. The next chapter will
investigate urbanization impacts on long‐term flooding hazards from stakeholder‐input growth
scenarios under climate change impact scenarios.
80
CHAPTER 4
FUTURE LONG‐TERM CLIMATE CHANGE‐INDUCED FLOODING RISK ASSESSMENT UNDER
GROWTH SCENARIOS AND CLIMATE CHANGE SCENARIOS
4.1 Introduction
Urban forms and development patterns affect biogeophysical characteristics and
functions of the ecosystem in cities while the land, water, air, flora and fauna, energy, and
nutrients in urban areas have been altered both intentionally and inadvertently (Pickett et al.
2001; Spirn 1984). They significantly correlate to environmental performance (Alberti 1999). In
addition, urbanization impacts associated with housing density in urban development patterns
have been linked to flooding damages (Brody et al. 2011) as well as to the urban heat island
effect in which the micro climate is altered and contributes to climate change (Pickett et al.
2001). Population growth is a driver for the urbanization process and subsequent land use and
land cover change that results in societal and environmental impacts. As population increases,
more housing and associated urban infrastructure such as water supply, stormwater
management, sewer systems, waste management, electricity, schools, fire stations, hospitals,
parks and recreation, etc. will be required for supporting the health, safety and welfare of the
newcomers. New development is often associated with land clearing, sealing, and compaction,
which results in an increase of impervious surfaces and alterations of the hydrological cycle in
the watershed (Booth and Jackson 1997; Brabec, Schulte, and Richards 2002; Lindh 1972).
The Boston Metropolitan Area (BMA) has a population of 3.6 million with an estimated
increase of 11 through 2030 (MAPC 2009). With the projected urban growth, the Metropolitan
Area Planning Council (MAPC) has outlined growth management strategies that aim to address
efficient transportation, natural resources and land conservation, improvement of health and
81
education, and an increase in equitable economic development opportunities in the proposed
MetroFuture Plan for 101 municipalities in the BMA (MAPC 2009). In terms of urban form, the
MetroFuture scenario concentrates population growth along transit systems in lieu of the
suburban sprawl in the Current Trends scenario. To further investigate urban form and
development patterns in relation to social equity and environmental performance, the Boston
Metro Area Urban Long‐Term Research Area‐Exploratory (BMA ULTRA‐ex) project team at the
University of Massachusetts Amherst has engaged with stakeholders in the BMA to develop two
additional future growth scenarios: the Green Equity scenario, which focuses on greening
investment such as trees and community gardens in low‐income neighborhoods, and the
Compact Core scenario, which emphasizes concentrated development in the Inner Core
communities identified by MAPC. Using the same number of projected population in 2030, each
scenario differs in its distribution of population growth between Inner Core and non‐core
communities.
Building upon the long‐term climate change‐induced flooding risk assessment in the
Charles River Watershed, which was developed in the previous chapter, this chapter aims to
answer the following questions: 1) In what way does the distribution of population growth
influence urban form and development pattern in terms of land use change? 2) To what degree
do the four future growth scenarios affect long‐term flooding hazards? 3) To what degree does
climate change have impacts on future long‐term flooding hazards in the future growth
scenarios? 4) What is the relationship between social vulnerability and long‐term climate
change‐induced flooding hazards in the future growth scenarios? Several hypotheses were: 1)
Suburban sprawl results in greater land disturbance in the watershed, including land cover
change from forest and agricultural lands to urban development as well as increased
development in low density areas. 2) Suburban sprawl results in an increase in flooding hazards.
82
3) Under higher climate change impact, suburban sprawl results in more long‐term flooding
hazards in more places. 4) Suburban sprawl in the Charles River Watershed results in more
socially vulnerable people under the risk to the long‐term climate change‐induced flooding
hazards.
4.2 Background
4.2.1 MAPC community types and MetroFuture
The Metropolitan Area Planning Council (MAPC) is one of the 13 Regional Planning
Agencies (RPAs) erected in the state of Massachusetts. The mission of MAPC is to “promote
smart growth and regional collaboration in Metropolitan Boston” under the direction of the
regional plan— MetroFuture: Making a Greater Boston Region (MAPC 2009). Based on land use
and housing patterns, recent growth trends and projected development patterns, MAPC
identified four types of communities and outlined planning strategies for each type of
community in the MetroFuture plan. The Inner Core communities are intensively built with less
than 1 % vacant developable lands; therefore, any new development is essentially through reuse
or redevelopment of existing sites. There are 16 cities and towns identified as Inner Core,
including the high density cities of Boston, Cambridge, Somerville, Chelsea, etc. The Regional
Urban Centers are outside the Inner Core communities and characterized by downtown urban
centers with multiple‐story and mixed‐use blocks surrounded by moderately dense residential
areas. These urban communities are experiencing less growth rates than suburbs. Maturing
Suburbs are moderately developed and experiencing moderate growth rates whereas
Developing Suburbs are less developed and experiencing the most growth rates with mostly
83
single family housing development. The MAPC planning area comprises 101 municipalities; 33
of them are intersected with the boundary of the Charles River Watershed (Figure 4.1).
Figure 4.1 Charles River Watershed study area and the Metropolitan Area Planning
Council (MAPC) community types in the Boston Metropolitan Area
BOSTON
IPSWICH
STOW
ACTON
SHARON
FRANKLIN
SUDBURY
BOLTON
CONCORD
HOPKINTON
DUXBURY
CANTON
NATICK
LYNN
DOVER
NORWELL
NEWTON
MILLIS
PEABODY
LINCOLN
BEVERLY
MILFORD
CARLISLE
HUDSON
MEDWAY
HINGHAM
MARSHFIELD
WALPOLE
QUINCY
ESSEX
PEMBROKE
WESTON
FRAMINGHAM
WRENTHAM
MILTON
SCITUATEHOLLISTON
WAYLAND
LITTLETON
NORFOLKHANOVER
WEYMOUTH
HAMILTON
BELLINGHAM
MEDFIELD
LEXINGTON
FOXBOROUGH
SHERBORN
WOBURN
DANVERS
BEDFORD
WALTHAM
MARLBOROUGH
ASHLAND
SAUGUS
WILMINGTON
STOUGHTON
NEEDHAM
BRAINTREE
MIDDLETON
DEDHAM
SALEM
TOPSFIELD
READING
GLOUCESTER
LYNNFIELD
NORWOODRANDOLPH
WESTWOOD
ROCKLAND
BURLINGTON
COHASSET
SOUTHBOROUGH
WENHAM
WELLESLEY
MEDFORD
NORTH READING
REVERE
BOXBOROUGHWAKEFIELD
HOLBROOK
ROCKPORT
MALDEN
CAMBRIDGE
STONEHAM
BROOKLINE
MANCHESTER
MAYNARD
HULL
BELMONT
MELROSEWINCHESTER
ARLINGTONEVERETT
SOMERVILLE
MARBLEHEAD
WATERTOWN
CHELSEA
SWAMPSCOTT
WINTHROP
NAHANT
Legend
Charles River
Charles River Watershed
Boston Metro Area
MAPC
Inner Core
Maturing Suburbs
Developing Suburbs
Regional Urban Centers
´ 0 5 10 15 202.5Miles
84
MetroFuture was initiated by MAPC through more than five years of participatory
planning processes with regional and local stakeholders as well as the public, in addition to
intensive data analyses and assumptions in projected growth for each MAPC community type in
the Boston Metro Area. MetroFuture plan has laid out 65 goals addressing sustainable growth
patterns, housing choices, community vitality, prosperity, transportation systems, and natural
resources management (i.e., energy, air, water, and wildlife) in the region. In terms of the
growth patterns that this study focuses on, MetroFuture adopted Smart Growth principles and
techniques such as focusing on growth along public transit lines and redeveloping built areas
where infrastructure systems are already in place. Where new development is needed,
clustered development patterns are strongly encouraged to preserve large open space and
working farmlands (MAPC 2009).
4.2.2 BMA ULTRA‐ex growth scenarios 2030
The BMA ULTRA‐ex team has developed four stakeholder‐input future growth scenarios
extended from MAPC’s plans for the Boston Metropolitan Area—Current Trends (CT),
MetroFuture (MF), Green Equity (GE) and Compact Core(CC) (Ryan et al. 2013). Each scenario
illustrates one possible future that reflects policy assumptions in growth management strategies
between Inner Core and non‐core communities.
As described by MAPC, the Current Trends serves as a “business as usual” scenario
compared to the alternative growth management strategies laid out in the MetroFuture.
Currently, the 16 Inner Core communities have been more than 99% built‐out and less than 1%
of undeveloped land has zoning that allows for housing development (based on currently
available 2005 land use data and 200X zoning information from MassGIS). Therefore, growth
strategies in the Inner Core communities were assumed to be strictly where lands have been
85
developed and infrastructure has already been put in place (i.e. infill development) whereas in
non‐core communities, growth strategies include both infill and greenfill. Greenfill is a growth
management strategy that allow for new development on agricultural and forest lands. In
Current Trends, greenfill was the least restrictive with a lack of concern on farmland protection
or open space conservation. In contrast, MetroFuture emphasized more infill along transit lines
and limited growth through greenfill, particularly for the purpose of preserving working
farmlands. Green Equity was designed to achieve a particular goal in raising environmental
justice in terms of distribution of urban open space and ecosystem services in relation to low‐
income neighborhoods (Danford et al. in review). The assumption was that Green Equity would
put less growth pressure on the Inner Core in order to allow more green space in urban
neighborhoods. On the other hand, Compact Core was assumed to have the most population
growth in the Inner Core using the most intensive efforts in infill development in order to relieve
development pressures on suburban communities that build into farmlands and forest lands
(Table 4.1).
Table 4.1 Summary of growth strategies applied in the growth scenarios for the Boston
Metropolitan Area through 2030
Growth Strategy Current Trends (CT)
Metro Future (MF)
Green Equity (GE)
Compact Core (CC)
Greenfill Much More Less More Much Less
Urban Densification (Infill) Much Less More Less Much More
Farmland Preservation Much Less Much More Less More
Environmental Equity Much Less More Much More Less
4.2.3 Population and growth scenarios
Population projection through the year 2030 in each Traffic Analysis Zone (TAZ) was
provided by MAPC for the Current Trends and the MetroFuture scenarios. A TAZ is defined in
86
relation to traffic‐related data (e.g., journey‐to‐work) and is therefore not consistent with
census tract boundary (see TAZ metadata in census.gov). Based on MAPC projections, Current
Trends has a 9.5% population increase from the 2000 census, which is slightly less than the
projected growth of 11% for the MetroFuture. In exploring more urban greening in the Inner
Core communities for environmental justice in lieu of population growth, a midpoint of the
population growth rate between Current Trends and MetroFuture was given for Green Equity
with a factor of 0.72 in MetroFuture’s projection (Danford et al. in review). For Compact Core
scenario, population is assumed to have even more growth than that of MetroFuture while the
focus is on redevelopment in the Inner Core communities.
According to the historical population data since 1900, Boston had a climax of 800,000
population in 1950 and since then has been experiencing a decline to less than 600,000 in 1960
and is now growing incrementally through the 2010 census when it reached 617,594.
Considering the past urban capacity in Boston, the project team assumed that a population
change of about 150,000 people—about 1.45 times MetroFuture’s projected population
change—was a reasonable growth for the Compact Core scenario. It is worth noting that the
population projection analyses were conducted in the early 2000s. As a result, the projected
change was based on the 2000 census instead of the latest 2010 census data.
CT 2030 pop = 2000 pop + projected pop change in CT 2030 (1)
MF 2030 pop = 2000 pop + projected pop change in MF 2030 (2)
GE 2030 pop = 2000 pop + (0.72 X projected pop change in MF 2030 in Inner Core)
+ (1.314 X projected pop change in MF 2030 in non‐Core) (3)
CC 2030 pop = 2000 pop + (1.45 X projected pop change in MF 2030 in Inner Core)
+ (0.496 X projected pop change in MF 2030 in non‐Core) (4)
pop: population
87
Based on the 2010 census data, the average household has 2.59 people. Therefore, the
projected housing unit change from 2000 to 2030 can be calculated by dividing the projected
increase population change by 2.59. The increased population and housing units were then
used for land use change and hydrological modeling in the Charles River Watershed for this
study (Table 4.2).
Projected increase of housing units = Projected increase pop change / 2.59 (5)
Table 4.2 Summary of projected population and housing units increase from 2000 to
2030 between Inner Core and non‐core communities among growth scenarios
Growth Scenarios Current Trends* Metro Future* Green Equity Compact Core
POP 2000 Total BMA 3,069,685 same same same
Inner Core 1,350,661
non‐core 1,719,024
POP 2030 Total BMA 3,360,752 3,413,000 3,413,000 3,413,000
Inner Core 1,426,845 1,711,760 1,660,975 1,793,380
non‐core 1,933,907 1,701,240 1,752,025 1,619,620
POP Increase Total BMA 291,067 343,315 343,315 343,315
Inner Core 78,003 181,376 130,591 262,996
non‐core 213,064 161,939 212,724 80,319
Housing Units Increase BMA 112,418 132,554 132,554 132,554
Inner Core 30,117 70,029 50,421 101,543
non‐core 82,264 62,525 82,133 31,011
POP: population; a: percent of total population; b: percent population increase from POP 2000 *Population projections for Current Trends and MetroFuture scenarios were provided by MAPC
88
4.3 Methodology
4.3.1 Future long‐term climate change‐induced flooding risk assessment framework
This chapter has built on Chapter 3 and integrated additional urbanization impacts in
the long‐term climate change‐induced flooding risk assessment using BMA ULTRA‐ex future
growth scenarios for the Charles River Watershed. The detailed methodology for SWAT
modeling, climate change assessment, climate change impact scenarios, long‐term flooding
hazard index (HI), Social Vulnerability Index (SoVI), and long‐term flooding risk index (RI) were
described in Chapter 3. This chapter focuses on methods for developing future growth
scenarios through 2030, applying future scenarios to SWAT modeling, and assumptions for
building future long‐term climate change‐induced flooding risk indices.
Among the 30 variables used in SoVI, the number of housing units per square mile
(HODENT) was the only variable that could possibly be updated based on projected population
and housing unit increase in growth scenarios. Nevertheless, housing density was only one of
the four variables that contributed to the explanation in one of the eight principal components
(a correlation of ‐.76 to component 2) (Table 3.5). Without other supporting projections such as
renter occupancy rate, employment rate, migration rate, income or demographic composition
change, SoVI was assumed to maintain the same pattern as the 2010 census data and was
therefore unchanged from Chapter 3 in this study.
Adopted from the long‐term climate change‐induced flooding hazard index (HI) defined
in Chapter 3, future HI in growth scenarios was calculated by comparing simulated future
streamflow to bankfull volume for current land use (MassGIS Land Use 2005) under current
climate conditions defined in Chapter 3. Climate change impact scenarios for Low Impact,
89
CT: Current Trends
MF: MetroFuture
GE: Green Equity
CC: Compact Core
Low Impact
Medium Impact
High Impact
Climate Change Impact
Scenarios
SWAT Modeling
Growth Scenarios 2030
OUTPUT
Long‐Term Flooding Hazard Index (HI)
INPUT
Long‐Term Climate Change‐Induced
Flooding Risk Index (RI)
X
Social Vulnerability Index (SoVI) =
SoVI Social
Vulnerability
SoVI
Method
Moderate Impact, and High Impact were also taken from Chapter 3 and used as weather input
for SWAT modeling for each future growth scenario of 2030 SWAT land use input (Figure 4.2).
Days when Q > Qbkful05 HI = P (Q > Qbkful05) = (6)
365 days a year * 45 years P: Probability
Q: Streamflow
Qbkful05: Based on 2005 land use current climate stream bankfull discharge
Figure 4.2 Future long‐term climate change‐induced flooding risk assessment framework
integrating growth scenarios 2030 and climate change impact scenarios
4.3.2 Projected population to land use conversions
Population growth is the driver in the urbanization process. Several steps were involved
in transforming MAPC’s population projection into SWAT land use for SWAT modeling.
90
Step 1: Identify developable lands
According to MAPC community types, Inner Core communities have limited vacant land
available and therefore the growth strategy focuses on infill vacant lots, abandoned buildings or
cleaning up unused industrial lands (i.e. brownfields) for new housing and offices in addition to
reusing and redeveloping existing buildings and sites for accommodating increased population
growth. Therefore, developable lands for Inner Core give priority to current residential,
commercial, and industrial lands in addition to vacant lands. Transportation and institutional
land uses in general are not used for additional market‐based housing units because the lands
are usually privately owned and in use as part of transit, educational or civic service systems.
However, MAPC has projected a large amount of population growth in several TAZs that
comprise primarily transportation or institutional land uses. Therefore, 15% of the existing
transportation or institutional lands were assumed to be re‐developable and part of the
secondary urban infill category if the primary infill lands do not suffice new housing demands.
All other land uses in Inner Core such as wetlands and waters, recreational, utilities, agricultural
and forests are non‐developable land uses that are not included in the calculation of land
changes in all scenarios.
In non‐core communities (i.e., Regional Centers, Mature Suburbs, Developing Suburbs)
new development was assumed to happen less in the already built lands as infill development
but to occur more on undeveloped lands as greenfill development. Therefore, in addition to
urban infill potential, agriculture and forest lands that allowed for residential and mixed‐use
development under current zoning regulations were included in the developable land
categories. On the other hand, transportation and institutional lands were assumed not to be
91
used for new housing development (Table 4.3). Several assumptions for selecting developable
lands in non‐core communities were based on zoning definitions provided by MassGIS:
1. CP (conservation/passive recreation): zoning areas were excluded from any land use for
development
2. PRIM_USE2: this category was used for zoning selection because it reflected more
current conditions of built lands
3. Agricultural lands: considering full potential for mixed use and including all commercial
(CB, GB, HB, LB), industrial (GI, IN, LI),mixed use (MU), residential (MH, ML, MM, R1, R2,
R3, R4, R5, RA) and not‐zoned (NZ) zoning
4. Forest lands: considering less pressures of mixed use on commercial and industrial
zoning areas and assuming primarily residential (MH, ML, MM, R1, R2, R3, R4, R5, RA),
mixed use (MU), and not‐zoned (NZ) areas would be sufficient and more likely to be
encouraged for new residential development
Table 4.3 MassGIS 2005 land use categorized for developable versus non‐developable
land uses in Inner Core and non‐core communities
Inner Core Communities Developable Non‐Developable
• Residential – Multi‐Family Residential – High Density Residential – Medium Density Residential – Low Density Residential – Very Low Density Residential
• Commercial – Commercial – Industrial – Transitional – Junkyard – Open Land
• Urban Public/Institutional (15%) • Transportation (15%)
• Wetlands and Water – Non‐Forested Wetland – Saltwater Wetland – Water – Cranberry Bog – Forested Wetland
• Recreational – Participation Recreation – Spectator Recreation – Water‐based recreation – Saltwater Sandy Beach – Golf Course – Marina – Cemetery
• Utilities
92
– Mining – Waste Disposal – Powerline/Utility
• Agriculture – Orchard – Nursery – Cropland – Pasture
• Forest – Forest – Bushland/Successional
non‐core Communities Developable Non‐Developable
• Residential – Multi‐Family Residential – High Density Residential – Medium Density Residential – Low Density Residential – Very Low Density Residential
• Commercial – Commercial – Industrial – Transitional – Junkyard – Open Land
• Agriculture (residential, commercial, industrial zoning, MU and NZ)
– Orchard – Nursery – Cropland – Pasture
• Forest (residential zoning, MU and NZ) – Forest – Bushland/Successional
• Urban Public/Institutional • Transportation • Wetlands and Water
– Non‐Forested Wetland – Saltwater Wetland – Water – Cranberry Bog – Forested Wetland
• Recreational – Participation Recreation – Spectator Recreation – Water‐based recreation – Saltwater Sandy Beach – Golf Course – Marina – Cemetery
• Utilities – Mining – Waste Disposal – Powerline/Utility
• Agriculture (CP) – Orchard – Nursery – Cropland – Pasture
• Forest (CP, Commercial and Industrial zoning)
– Forest – Bushland/Successional
93
Step 2: Define housing density types
Housing needs are directly connected to population growth (Table 4.2). After
identifying the developable land areas in each TAZ, the next step is to calculate the housing
density that is projected in each TAZ. Housing density is defined as the number of housing units
on developable land areas (i.e. gross housing density). A total of seven housing density types
were identified to reflect probable housing development in the Boston Metropolitan Area. In
addition, “negligible” and “unreasonable” categories are defined below:
a. Very Low Density (VLD): housing on more than one acre lot, such as large estates in
suburban and rural areas that aremost likely to be associated with agricultural and
forest greenfill development
b. Low Density (LD): housing density of between 1 to 4 units per acre, mostly single‐family
residential development
c. Low Medium Density (LMD): housing density of between 4 to 12 units per acre, such as
small lot single‐family or duplex houses
d. Medium Density (MD): housing density of between 12 to 32 units per acre, such as town
houses and multi‐family housing
e. Medium High Density (MHD): housing density of between 32 to 60 units per acre, low‐
rise building type with two to five floors for apartment or condominium development
f. High Density (HD): housing density of between 60 to 120 units per acre, mid‐ to high‐rise
building type with four to ten floors, often with mixed uses at lower floors
g. Very High Density (VHD): housing density of between 120 to 200 units per acre, high‐rise
building type with more than ten floors and with mixed uses at lower floors
94
h. Unreasonable: housing density of more than 200 units per acre in the Inner Core
communities or projected housing units exceeding probable housing density using the
assumptions in the non‐core communities (see steps 3 and 4)
i. Negligible: TAZs with projected population of less than 10 people, including negative
projections, were considered negligible considering possible errors from projections. In
addition, the assumption was made that less than four housing units in one TAZ could
be absorbed through existing housing infrastructure and therefore would not alter
urban landscapes.
Step 3: Inner Core communities development assumptions
Inner Core communities are expected to grow only through infill or redevelopment of
existing built areas. However, it is unrealistic to assume that every existing building would be
reused or torn down and rebuilt or that every inch of brownfield would be redeveloped. In
addition, in order to tell the social and environmental outcomes from each scenario apart,
different amounts of developable lands were assigned for infill development (Figure 4.3). The
Current Trends scenario has the least infill potential (25%) under the assumption that population
growth takes place primarily in the suburban area and through greenfill development. The
MetroFuture and Green Equity scenarios assume the same amount of land (35%) for infill
development. Compact Core has the most aggressive redevelopment goal, using 50%
developable land for infill development. After allocating infill development land availability,
housing density is calculated based on projected new housing units for each scenario. The
primary infill lands can be developed up to a maximum of 120 housing units per acre. After the
high density capacity, the secondary infill lands would be used for up to 200 housing units per
95
acre. If projected housing units exceed the capacity, then those TAZs were marked as having an
unreasonable projection.
Figure 4.3 Housing density decision tree for Inner Core communities
Step 4: Non‐Core communities development assumptions
Projected development patterns for non‐core communities occur in two forms: Infill and
Greenfill. Similar to the assumptions in the Inner Core, different percentages of developable
lands for urban infill were assigned for each scenario (Figure 4.4). Current Trends was assumed
to have the least infill opportunity and was given only 10% of redevelopment land potential.
The MetroFuture and Compact Core scenarios were considered to have the most aggressive
urban growth strategy through infill development and were therefore given the most potential
land (30%) among all the scenarios. Green Equity was assumed to have moderate infill
redevelopment goals with 20% of developable lands. Based on characteristics and development
ΔPOP ΔHU
Yes
Negligible
Urban (Primary) Infill
CC Scenario50% Land
@ 120 HU/AC
GE Scenario35% Land
@ 120 HU/AC
MF Scenario35% Land
@ 120 HU/AC
CT Scenario25% Land
@ 120 HU/AC
No
Zero Agricultural
Greenfill
Unreasonable
enough land?
up to 200
HU/AC
Yes No
Zero Forest
Greenfill
Urban (Secondary) Infill15% (Transportation + Institutional) Lands
Very High
Low Medium
MediumHigh
High Medium Low
120<HD<=200
4=<HD<=12
32<HD<=60
60<HD<=120
1=<HD <4
12<HD<=32
HD>200 or Infill land = 0 ac
ΔPOP <10
96
trends defined in the MAPC community types, various maximum allowable housing densities for
each community type were further delineated to reflect probable outlooks of growth patterns
by 2030. Regional Centers received a maximum of 32 units per acre, assuming the trends of
low‐rise mixed use buildings continue in downtown core areas as well as in existing residential
areas. Maturing Suburbs and Developing Suburbs have similar urban landscapes with a slightly
smaller maximum allowable housing density of 24 and 16 units per acre respectively.
After urban infill reaches the maximum allowable housing density, the remaining
housing units were then allowed to be allocated on the developable agriculture and forest lands
following different assumptions varied by community types and scenarios. Firstly in each TAZ,
agriculture lands greater than 40 acres were considered to have conservation values as working
farms and not allowed to be used for housing development. On the other hand, agricultural
lands less than 40 acres were to have full potential for greenfill development. Current Trends
was given the most agriculture greenfill opportunity at 80% whereas Compact Core was
assumed to be the most progressive in agriculture conservation and therefore only allowed 10%
of developable agriculture lands for new development. Based on one of the MetroFuture’s goals
in protecting farmlands, 20% of agriculture lands for greenfill development were assigned to it
whereas Green Equity received a moderate 40%. In addition to the difference among scenarios,
maximum allowable housing density differed in each community type. Regional Centers
received a maximum of 16 units per acre whereas Maturing Suburbs and Developing Suburbs
were given 12 and 8 units per acre respectively for agricultural greenfill development potentials.
The last piece of land to be developed was forested land. After filling urban infill and
agriculture greenfill, the remaining housing units were allowed up to a maximum housing
density in developable forest lands that varied in each community type based on assumptions of
the outlook in development patterns. Regional Centers received a maximum of 12 units per
97
acre whereas Maturing Suburbs and Developing Suburbs were assumed to have 8 and 6 units
per acre respectively as a maximum density allowance in forest greenfill development.
Figure 4.4 Housing density decision tree for non‐core communities
Step 5: 2030 land use decision rules
Land use change in 2030 was assumed to happen through densifying existing urban land
uses (infill) or in agriculture or forest lands (greenfill). For example, low density residential land
use in the MassGIS 2005 Land Use category could become moderate density residential land use
in 2030 after converting projected population change into housing units and housing density
type changes in each TAZ. In order to determine under what circumstances one land use would
change from one category in 2005 to another category in 2030, the threshold housing density in
each land use category was designated through the following procedures.
ΔPOP ΔHU
Yes
Negligible
No
Urban Infill Only
Agricultural Greenfill (>40ac; if<=40ac then 100%)
CC Scenario10% Land @ 16 du/ac-RC@ 12 du/ac-MS@ 8 du/ac-DS
GE Scenario40% Land @ 16 du/ac-RC@ 12 du/ac-MS@ 8 du/ac-DS
MF Scenario20% Land @ 16 du/ac-RC@ 12 du/ac-MS@ 8 du/ac-DS
CT Scenario80% Land @ 16 du/ac-RC@ 12 du/ac-MS@ 8 du/ac-DS
enough land?
Urban + Ag
Urban + Ag + Forest
Unreasonable
enough land?
Yes No
enough land?
Yes No
Forest Greenfill
CC Scenario@ 12 du/ac-RC@ 8 du/ac-MS@ 6 du/ac-DS
GE Scenario@ 12 du/ac-RC@ 8 du/ac-MS@ 6 du/ac-DS
MF Scenario@ 12 du/ac-RC@ 8 du/ac-MS@ 6 du/ac-DS
CT Scenario@ 12 du/ac-RC@ 8 du/ac-MS@ 6 du/ac-DS
Urban InfillCC Scenario30% Land @ 32 du/ac-RC@ 24 du/ac-MS@ 16 du/ac-DS
GE Scenario20% Land @ 32 du/ac-RC@ 24 du/ac-MS@ 16 du/ac-DS
MF Scenario30% Land @ 32 du/ac-RC@ 24 du/ac-MS@ 16 du/ac-DS
CT Scenario10% Land @ 32 du/ac-RC@ 24 du/ac-MS@ 16 du/ac-DSΔPOP
<10
HD type
HD type
HD type
98
First, current housing density was estimated based on the descriptions of MassGIS 2005
Land Use residential categories. High density and multifamily residential land uses were
described as housing on lots of less than ¼ acres, which translated into a housing density greater
than 4 dwelling units per acre. Medium density residential land use was defined as housing on
¼ to ½ acre lots, which resembled housing density between 2 and 4 dwelling units per acre. Low
density residential land use was defined as housing on ½ to 1 acre lots, which translated into a
housing density of 1 to 2 dwelling units per acre. Very low density residential land use was
defined as housing on lots that are larger than 1 acre, which resembled a housing density of less
than 1 dwelling unit per acre.
Next, infill versus greenfill development was differentiated in 2030 land use in order to
reflect differences in growth strategies. Residential land use types for 2030 were simplified into
four levels of density: Mixed Use (MU, >60 du/ac), High Density (HD, 33‐60 du/ac), Medium
Density (MD, 9‐32 du/ac) and Low Density (LD, <=8 du/ac). These categories were used instead
of seven housing density types in step 2 for a better match with 2005 land use categories (Table
4.4). Finally, 2030 land use was determined based on the current housing density defined in
2005 land use plus projected housing density allocated in each TAZ (Table 4.5).
Table 4.4 2030 land use categories and descriptions
2030 Land Use Symbol Growth Strategy Descriptions
UHMU Infill Urban Mixed Use URHD Infill Urban Residential High Density URMD Infill Urban Residential Medium Density URLD Infill Urban Residential Low Density RRHD Greenfill (after Infill) Rural Residential High Density RRMD Greenfill (after Infill) Rural Residential Medium Density RRLD Greenfill (after Infill) Rural Residential Low Density Unreasonable Unreasonable Unreasonable projected growth Unchanged Negligible;
Undevelopable No land use change from 2005
99
Table 4.5 2005 land use to 2030 land use conversion rules
2005 Land Use Projected Density Type in TAZ (see steps 2, 3, 4)
2030 Land Use
Commercial Any density UHMU
High Density/Multifamily Residential MD, MHD, HD, VHD UHMU
(>4du/ac) VLD, LD, LMD URHD
Ag, Fo RRHD
Medium Density Residential HD, VHD UHMU
(2‐4du/ac) MD, MHD URHD
Ag, Fo RRHD
VLD, LD, LMD URMD
Low/Very Low Density Residential HD, VHD UHMU
(<2du/ac) MD, MHD URHD
Ag, Fo RRHD
LMD URMD
VLD, LD URLD
Agriculture Ag‐MD RRHD
Ag‐LMD RRMD
Ag‐LDD/VLD RRLD
Regional Centers Fo RRHD
Maturing Suburbs Fo RRMD
Developing Suburbs Fo RRLD
Forest Fo‐LMD RRMD
Fo‐LD/VLD RRLD
Transportation/Institutional Inner Core VHD UHMU
Ag: any density type and agricultural lands; Ag‐Density Type: density type defined in step 2 Fo: any density type on forest lands; Fo‐Density Type: density type defined in step 2
4.3.3 SWAT 2030 land use
SWAT land use categories were derived from 2005 land use. Since each streamflow
output at each subbasin is a lump sum outcome from all Hydrologic Response Units (HRUs) (see
section 3.3.2), 2030 land use in each TAZ was converted to SWAT land use and then calculated
as a future land use input at each HRU for each growth scenario. Several steps were involved in
this conversion. First, GIS analysis was employed to intersect the boundaries of TAZs with
subbasins . Each subbasin retained information about the amount of land area that experienced
100
change from land use in 2005 to projected land use in 2030. Subsequently, 2030 land use
categories were recategorized into SWAT 2030 land use for each growth scenario (Table 4.6).
Then the new ratio of SWAT 2030 land use in each subbasin became the new HRU fraction.
Finally, the new HRU fraction of SWAT 2030 land use was equally distributed if different
hydric soil types were present. For example, subbasin 16 consisted of five SWAT 2005 land use
categories: AGRL, FRST, WATR, UCOM, and URHD. While AGRL, FRST, and WATR were projected
to be unchanged, UCOM was projected to have a 0.5% increase derived from URHD in the
Current Trends scenario. In addition, UCOM contained two HRUs with hydric soil types A and B
respectively and three HRUs in URHD with hydric soil types A, B, and C. Subsequently, the 0.5%
increase in UCOM was divided by two so that each HRU received an increase of 0.25%. The
decrease of URHD was divided by three so that each HRU received a decrease of 0.17%. Neither
total land gain or loss nor new SWAT land use was created in each subbasin. The only change
was the fraction of each HRU that reflected each SWAT land use change from 2005 to 2030 in
each future growth scenario.
Table 4.6 SWAT 2030 land use designation based on 2030 land use
SWAT 2030 Land Use
2030 Land Use Descriptions
UCOM UHMU Extensive developed areas primarily through mixed use and urban infill on commercial lands. Housing density is greater than 60 units per acre.
URHD URHD/RRHD High density residential areas. Housing density is between 33 and 60 units per acre.
URMD URMD/RRMD Medium density residential areas. Housing density is between 9 and 32 units per acre.
URLD URLD/RRLD Low density residential areas. Housing density is less than or equal to 8 units per acre.
101
4.4 Results
4.4.1 Land use change
SWAT land use change in growth scenarios compared to 2005 land use resulted in the
Current Trends scenario having the most land area change with the largest increase in higher
density residential and use the largest amount of agriculture and forest lands. MetroFuture and
Compact Core had similar infill development patterns that essentially densified existing land use
types, but not to the extent of changing land use type (Table 4.7).
Overall, Current Trends experienced the most land changes: 7.6% of watershed area,
compared to MetroFuture at 2.1%, Green Equity at 3.5%, and Compact Core at 1.3%. Current
Trends primarily gained high density and mixed use residential lands from converting lower
density residential lands in addition to greenfill development on agriculture and forest lands.
Green Equity appeared to have similar pattern as Current Trends, but on a much smaller scale, as
it converted existing low and medium density to higher density residential lands. MetroFuture
and Compact Core had similar infill patterns yet MetroFuture had five times more greenfill than
in Compact Core (Figure 4.5).
102
Table 4.7 Land use change areas from 2005 to 2030 in growth scenarios
LU 2005 LU 2030 SWAT LU 2030
Growth Strategy
Current Trends
Metro Future
Green Equity
Compact Core
Land Use 2005 change through Greenfill (acre)
Agriculture RRHD URHD Greenfill 0 6 6 9
Agriculture RRLD URLD Greenfill 35 2 36 0
Agriculture RRMD URMD Greenfill 61 17 16 0
Forest RRLD URLD Greenfill 1,334 1,387 1,583 260
Forest RRMD URMD Greenfill 484 0 0 0
Subtotal 1,914 1,412 1,641 269
2005 Land Use change through Infill (acre)
HD Residential UHMU UCOM Infill 1049 703 1045 813
MD Residential UHMU UCOM Infill 0 0 0 0
MD Residential URHD URHD Infill 4,979 1,544 2,212 1,081
LD Residential UHMU URMD Infill 0 1 0 1
LD Residential URHD URHD Infill 510 77 137 52
LD Residential URMD URMD Infill 5,557 0 1,381 66
Subtotal 12,095 2,325 4,775 2,012
No land use change (acre)
HD Residential URHD URHD Infill 10,587 12,267 11,604 12,405
MD Residential URMD URMD Infill 8,673 8,819 7,786 8,162
LD Residential URLD URLD Infill 13,150 16,224 14,785 15,675
Commercial UHMU UCOM Infill 6,269 7,827 7,626 7,868
Transportation UHMU UCOM Infill 67 0 17 86
Institutional UHMU UCOM Infill 197 366 338 401
Subtotal 38,942 45,502 42,155 44,597
103
Figure 4.5 Percentage of watershed area change from lower density in 2005 land use to
higher density among growth scenarios
4.4.2 Future growth scenario impacts
The definition of long‐term flooding hazard index (HI) in future growth scenarios was
the probability of the event exceeding current conditions. Therefore, when comparing across
scenarios, an increase or decrease in HI value reflects a particular scenario that could be worse
or better than another one. The long‐term climate change‐induced flooding risk index (RI)
reflected the variation of HI among the growth scenarios. Overall, the results reveal a greater
variance in RI through climate change scenarios than among growth scenarios (Table 4.8).
Current Trends had the highest RI among growth scenarios.
When compared across climate change scenarios, all growth scenarios underwent a
similar pattern of RI increasing from current climate conditions to Low, Medium, and High
Impact climate change scenarios (Figure 4.6). Nevertheless, under the High Impact climate
0%
1%
2%
3%
4%
5%
6%
7%
8%
Current Trends MetroFuture Green Equity Compact Core
% watershed area
05HD
05MD
05LD
Greenfill
104
Table 4.8 Summary of risk assessment indices descriptive statistics
Index Growth Scenario
Climate Scenario
Mean Median Standard Deviation
Minimum Maximum
SoVI 0.5847 0.5863 0.121 0 1
HI Current Trends
Current 0.0077 0.0080 0.0035 0.0027 0.016
Low 0.0088 0.0088 0.0036 0.0028 0.018
Medium 0.0103 0.0102 0.0046 0.0037 0.022
High 0.0175 0.0180 0.0090 0.0057 0.040
HI Metro Future
Current 0.0076 0.0076 0.0033 0.0026 0.016
Low 0.0086 0.0082 0.0034 0.0027 0.018
Medium 0.0100 0.0097 0.0044 0.0036 0.021
High 0.0175 0.0180 0.0090 0.0057 0.040
HI Green Equity
Current 0.0076 0.0078 0.0034 0.0027 0.016
Low 0.0088 0.0084 0.0035 0.0029 0.018
Medium 0.0100 0.0097 0.0044 0.0036 0.021
High 0.0175 0.0180 0.0090 0.0057 0.040
HI Compact Core
Current 0.0076 0.0077 0.0033 0.0027 0.016
Low 0.0087 0.0083 0.0033 0.0029 0.017
Medium 0.0101 0.0100 0.0043 0.0038 0.021
High 0.0175 0.0180 0.0090 0.0057 0.040
RI Current Trends
Current 0.0044 0.0042 0.0020 0 0.011
Low 0.0050 0.0048 0.0022 0 0.013
Medium 0.0059 0.0056 0.0027 0 0.015
High 0.0100 0.0093 0.0052 0 0.029
RI Metro Future
Current 0.0043 0.0041 0.0019 0 0.011
Low 0.0049 0.0047 0.0021 0 0.012
Medium 0.0057 0.0055 0.0026 0 0.015
High 0.0100 0.0093 0.0052 0 0.029
RI Green Equity
Current 0.0044 0.0042 0.0019 0 0.011
Low 0.0051 0.0048 0.0021 0 0.012
Medium 0.0057 0.0055 0.0026 0 0.015
High 0.0100 0.0093 0.0052 0 0.029
RI Compact Core
Current 0.0044 0.0041 0.0019 0 0.011
Low 0.0050 0.0047 0.0020 0 0.012
Medium 0.0058 0.0055 0.0026 0 0.015
High 0.0100 0.0093 0.0052 0 0.029
105
Figure 4.6 The long‐term climate change‐induced flooding risk index (RI) mean and
standard deviation among climate change impact and growth scenarios
change scenario, there were no differences in mean HI between growth scenarios, implying the
presence of a threshold effect when the influence of climate change exceeded land use change
impacts.
The Kernel density distribution for RI (n=277) illustrated the overall HI distribution
pattern among growth and climate change scenarios (Figure 4.7). Comparing the growth
scenarios when climate was controlled at current climate conditions and Low Impact, Current
Trends had the widest range of HI values with more units with higher HI values when the Green
Equity came in second, MetroFuture came in third, and Compact Core had the narrowest range
with more units in the low hazard indices. Under the Moderate Impact climate change scenario,
MetroFuture and Green Equity had almost identical pattern. Under the High Impact climate
change scenario, all growth scenarios had the same patterns with nearly double the highest HI
values in some units than the values in the Medium Impact climate change scenario. The
0.0%
0.2%
0.4%
0.6%
0.8%
1.0%
CT MF GE CC CT MF GE CC CT MF GE CC CT MF GE CC
Current Climate Low Impact Medium Impact High Impact
RI m
ean
value
Growth Scenarios CT: Current Trends MF: MetroFuture GE: Green Equity CC: Compact Core
106
impacts from the climate change scenarios were clearly revealed in the distribution of HI when
the patterns were similarly across growth scenarios.
Current Climate HI Low Impact Climate Scenario HI
Medium Impact Climate Scenario HI High Impact Climate Scenario HI
Growth Scenarios CT: Current Trends, MF: Metro Future, GE: Green Equity, CC: Compact Core;
Climate Change Impact Scenarios L: Low Imapct, M: Medium Impact, H: High Impact
Figure 4.7 Long‐term flooding hazard index (HI) density distributions among growth and
climate change impact scenarios
0.00 0.01 0.02 0.03 0.04 0.05
02
04
06
08
01
00
12
0
de
nsi
ty
HICTHIMFHIGEHICC
0.00 0.01 0.02 0.03 0.04 0.05
02
04
06
08
01
00
12
0
de
nsi
ty
HICTLHIMFLHIGELHICCL
0.00 0.01 0.02 0.03 0.04 0.05
02
04
06
08
01
00
12
0
de
nsi
ty
HICTMHIMFMHIGEMHICCM
0.00 0.01 0.02 0.03 0.04 0.05
02
04
06
08
01
00
12
0
de
nsi
ty
HICTHHIMFHHIGEHHICCH
107
Figure 4.8 Maps of long‐term climate change‐induced flooding risk index (RI) in current
climate conditions among growth scenarios
´ 0 2 4 6 81Miles
Current Trends MetroFuture
Green Equity Compact Core
Legend
Charles River
Tributary
RI0-0.2%
0.2-0.4%
0.4-0.6%
0.6-0.8%
0.8-1%
1-1.2%
108
Figure 4.9 Maps of the long‐term climate change‐induced risk index (RI) in the Current
Trends growth scenario among climate change impact scenarios
Spatial distribution illustrated similar patterns among growth scenarios with slight
differences in a few census tracts (Figure 4.8). Overall, the highest RI was located in the middle
to lower Charles River while the lowest RI was located among the census tracts in the tributary
subbasins, particularly in the lower basin in Boston and Lincoln as well as in northern Waltham
and Weston. The same pattern occurred in each growth scenario. The spatial distribution
pattern remained the same in each growth scenario across all the climate change scenarios
(Figure 4.9). When the growth scenario was controlled, the impacts of climate change were
´ 0 2 4 6 81Miles
Current Climate Low Impact
Medium Impact High Impact
Legend
Charles River
Tributary
RI0-0.2%
0.2-0.4%
0.4-0.6%
0.6-0.8%
0.8-1%
1-1.2%
1.2-1.4%
1.4-1.6%
1.6-1.8%
1.8-2%
2-2.2%
2.2-2.4%
2.4-2.6%
2.6-2.8%
2.8-3%
109
more severe at the locations that already had a higher RI in the middle to lower Charles River.
Under the Low Impact climate change scenario, the watershed was under a 1% long‐term
flooding risk while the RI increased to 1.6% in the Medium Impact climate change scenario and
3% in the High Impact climate change scenario.
4.5 Discussion
4.5.1 Growth strategy and urban landscape change
This study supports the first hypothesis that suburban sprawl leads to greater land
disturbance from both greenfill development and a greater spread of infill in low density areas.
MAPC has projected 52,153 less people in the Current Trends scenario than the MetroFuture
scenario, which resulted in 20,136 fewer housing units based on this study. Nevertheless, the
Current Trends scenario consumed the most agriculture and forest lands while the Compact Core
scenario used 99% of the already developed areas with the least overall land disturbance in the
watershed. Since greenfill only occupied 1% of the overall watershed area, infill became the
determinant growth strategy for changing landscapes.
In addition, the growth decision rules for developable allowance in infill areas created
different dynamics in urban landscapes. The Current Trends scenario assumed a smaller focus
on infill with the least amount of allowable lands at the site development level. Consequently,
more land or an otherwise higher density of housing development would be needed to
accommodate the increased population. However, based on current development patterns in
MAPC community types, the Current Trends scenario was assumed to receive the least number
of maximum allowable housing densities in the decision rules. As a result, the Current Trends
had the most TAZs that experienced land change as well as the most unreasonable projections
110
than in other scenarios. The MetroFuture and Compact Core scenarios received three times
more percentage of infill lands that are allowed for redevelopment and resulted in 26% more
infill land changes than in the Current Trends scenario (Table 4.9).
Imagining such impacts on urban landscapes, the Current Trends scenario would have
more high‐rise tower buildings sporadically in downtown or commercial strips in addition to
scattered low‐ or mid‐rise multi‐family housing on currently single‐family lots. In contrast, the
MetroFuture and Compact Core scenarios would have more low‐rise and mid‐rise
redevelopment throughout current low or medium density residential areas with more high‐rise
mixed use buildings in downtown areas and along transit lines. Therefore, urban form and its
landscapes were heavily influenced by growth strategy and associated land use policies.
Table 4.9 Infill land change analyses among growth scenarios
Current Trends Metro Future Green Equity Compact Core
Total infill land area (ac) 7,119 15,774 11,102 14,252
% Infill in total land change 72% 91% 84% 98%
4.5.2 Land use and land cover change
Despite stream channel characteristics, urbanization associated with impervious
surfaces, which are significantly correlated with land use type, is the primary factor influencing
streamflow (Leopold 1968). Impervious surface has been commonly accepted as an
environmental indicator for land use planning (Arnold and Gibbons 1996). It is widely
recognized that 10% effective impervious area (EIA) results in significant degradation of aquatic
ecosystem quality (Booth and Jackson 1997; Schueler 1994; Schueler, Fraley‐McNeal, and
Cappiella 2009). Furthermore, urbanization particularly contributes to flash floods as a result of
excessive and intense stormwater runoff (Savini and Kammerer 1961) and the increase in
111
frequency of overbank flows (i.e., long‐term flooding). For example, the rate of flooding
frequency would double at about 20% urbanization (20% impervious surface and 20% sewered
area) and quadruple at 50% urbanization in the studied watershed (Leopold 1968). Schueler et
al. (1994, 2009) developed an Impervious Cover Model (ICM) that defined four types of urban
stream conditions based on the thresholds of watershed impervious cover: sensitive (<10%),
impacted (10%‐ 25%), non‐supporting stream (25%‐60%), and urban drainage (>60%) (Schueler
1994; Schueler, Fraley‐McNeal, and Cappiella 2009).
Currently, the Charles River watershed has 21% total impervious surfaces (see Chapter
2), suggesting that the watershed has reached the threshold of an impacted watershed.
However, the specific impact of imperviousness in relation to streamflow and long‐term riparian
floods needs further investigation. In order to understand land use change in relation to land
cover change, analyses were performed on imperviousness change. Based on the average
impervious surfaces in SWAT 2005 land use types, the percentage of imperviousness between
the highest and lowest density developed areas could vary as much as 40% on average in the
entire Boston Metropolitan Area (Table 4.10).
Table 4.10 Mean and standard deviation of percent imperviousness in each SWAT 2005
land use category for the Boston Metropolitan Area (Data source: MassGIS)
SWAT 2005 Land Use Mean SD
UCOM 65% 17%
URHD 43% 11%
URMD 30% 7%
URLD 26% 8%
By extracting the change of development density between 2005 and 2030 in each
growth scenario and using the average imperviousness for each SWAT land use type, there was
as much as 1% impervious surface increase in Current Trends, 0.6% for Green Equity, 0.4% for
112
MetroFuture, and 0.2% for Compact Core (Table 4.11). Unfortunately, the variation among
scenarios in terms of land cover change was not significant enough to see distinguishable
differences in long‐term flooding hazards, which helps to explain the resulting little variation
among the growth scenarios.
Table 4.11 Projected impervious increase from total land use change areas from 2005 to
2030 among growth scenarios
2005 Land Use SWAT LU 2030
Impervious Increase
Current Trends
Metro Future
Green Equity
Compact Core
Low Density Residential
UCOM 39% 2.2 1.6 1.4 1.6
URHD 17% 88.0 13.3 23.6 8.9
URMD 5% 253.6 0.0 63.0 3.0
Medium Density Residential
UCOM 35% 36.1 35.9 49.9 35.9
URHD 13% 631.1 195.7 280.4 137.0
High Density Residential
UCOM 22% 303.4 163.7 254.3 188.1
Agriculture/ Forest
UCOM 65% 11.8 10.1 17.4 12.5
URHD 43% 0.0 2.6 2.6 3.9
URMD 30% 164.2 5.2 5.0 0.0
URLD 26% 349.9 355.0 413.8 66.5
Total imperviousness area increase (ac) 1,840 783 1,111 457
Imperviousness increase (% watershed) 1.0% 0.4% 0.6% 0.2%
4.5.3 Growing population in RI HotSpots
Distribution of population in future growth has impacts on future flooding risks, not only
through land use change that would have impacts on flooding hazards but also through social
vulnerability. Allocating population to where are already in high risk areas would only aggravate
existing social vulnerability if no plans have been made to upgrade infrastructures and expand
human resources for meeting existing and future demands.
113
In order to understand to what degree each growth scenario distributes future
population in high risk areas, a GIS spatial statistics tool, Getis‐Ord Gi* HotSpot analysis was
used. HotSpot analysis examined the significance in the spatial distribution based on a Z score
of each feature under normal distribution. Z scores of ±1.65 represent a 90% confidence level,
±1.96 represent a 95%, ±2.58 represent a 99%, and between +1.65 and ‐1.65 represents no
significance. The positive sign represents a tendency of higher values to cluster spatially.
Therefore, census tracts with RI Z scores greater than +1.65 are considered long‐term climate
change‐induced flooding risk HotSpots. In other words, the risk HotSpots illustrated that high
long‐term climate change‐induced flooding risk areas, both higher long‐term flooding hazards
and higher social vulnerability, are distributed heavily in the middle to lower basin.
Through GIS spatial mapping with projected population among growth scenarios and
baseline (2005 land use and current climate conditions) RI HotSpots, the Current Trends scenario
distributed the most population growth in high risks areas whereas the Compact Core scenario
distributed the least. Keeping in mind that the Current Trends scenario had projected a slightly
smaller population increase in the overall metropolitan area, a greater percentage of growth
(4.5%) was allocated in the suburbs where higher long‐term flooding risks currently occur and
would be aggravated by climate change impacts. On the other hand, the Compact Core scenario
focused population growth in the Inner Core communities (e.g., Boston, Cambridge), which
currently have lower long‐term flooding hazards. In addition, it allocated the least growth (3%)
in RI HotSpots in the Charles River Watershed (Figure 4.10).
114
Figure 4.10 Maps of allocation of projected increased population in growth scenarios
overlaid with current conditions of long‐term flooding risk index (RI) HotSpots
4.5.4 Climate change vs. urbanization impacts
This study revealed that the Charles River Watershed responded to the long‐term
flooding hazards more sensitively under climate change impacts than to the land use and land
cover changes associated with urbanization impacts. The paired t‐test in between climate
change impact scenarios showed significant difference between HI in all pairs of scenarios
(p<0.0001). From climate sensitivity study results in Chapter 3, the increasing mean
´ 0 2 4 6 81Miles
Current Trends MetroFuture
Green Equity Compact Core
Legend
RI HotSpot
GiZScore
< -2.58 Std. Dev.
-2.58 - -1.96 Std. Dev.
-1.96 - -1.65 Std. Dev.
-1.65 - 1.65 Std. Dev.
1.65 - 1.96 Std. Dev.
1.96 - 2.58 Std. Dev.
> 2.58 Std. Dev.
1 Dot = 10 people
Population Increase
Total population increase on RI HotSpot (>1.65 Std. Dev. In all red units) Current Trends: 13,188 MetroFuture: 11,060 Green Equity: 11,527 Compact Core: 10,311
115
temperature to 1˚C when precipitation was controlled could result in a range of 0.02% to 0.67%
decreases in HI. In addition, under the High Impact climate change scenario, the difference in
land use and land cover changes among growth scenarios has diminished. This phenomenon
implied a “threshold” effect in the interaction between climate and urbanization change.
Further studies are needed to gauge the threshold in order to provide planning parameters in
land use and land cover change under climate change impacts.
From this study, the Charles River Watershed appears to be more susceptible to climate
change impacts when the increase in population through 2030 is primarily through infill
development. As a result, planning strategies for future growth should aim for Smart Growth
and Low‐Impact Development strategies in addition to infill development in order to minimize
land use and land cover change in the watershed. Finally, the findings also imply that climate
change adaptation should be the top priorities for local communities in order to cope with
projected climate change impacts.
4.6 Conclusion
This study has revealed several particular interests in urbanization impacts from
projected population growth. Firstly, growth strategy and land use policies play an important
role in changing urban form and altering urban landscapes. In the series of growth assumptions,
we have carefully catered to each MAPC community type in order to illustrate a “ground truth”
analysis on what could actually be happening in 2030. Secondly, the projected land use change
among growth scenarios in our study was not large enough to have significant impacts on long‐
term flooding risks. This may have been due to the fact that growth strategies for all scenarios
focused heavily on infill redevelopment, even in the Current Trends scenario. Thirdly, the
allocation of population growth was critical in planning for reducing climate change‐induced
116
environmental risks. Among the growth scenarios, the Current Trends scenario had not only the
most urbanization impacts on long‐term flooding hazards but also the potential to increase
social vulnerability by placing more people in areas currently exposed to higher long‐term
flooding hazards. The MetroFuture, Green Equity, and Compact Core scenarios had similar
growth strategies that allocate more people in clustered forms and less growth through
suburban greenfill development, where higher flooding risks occurred along the main Charles
River.
In addition to urbanization impacts, this study also demonstrated the importance of
using empirical study for the understanding of climate change impacts on long‐term flooding
hazards and the spatial relationships with social vulnerability in an urbanized watershed.
Besides the influences from different growth management strategies on land use and land cover
change, it is critical to plan for climate change mitigation and adaptation with people in mind
and prioritize resources in reducing social vulnerability as well as enhancing resilience in the
linked social‐ecological systems.
117
CHAPTER 5
EFFECTS OF DETENTION FOR MITIGATING CLIMATE CHANGE‐INDUCED LONG‐TERM FLOODING
HAZARDS
5.1 Introduction
Climate change has posed increased risks to environmental hazards (e.g., flooding,
droughts, hurricanes) in addition to new challenges under climate change impacts (e.g., early
snow melt, rising sea levels, heat waves). Floods are omnipresent in almost every city in the
United States and account for more economic losses than any other single geophysical hazard
(White and Haas 1975). Previous climate change studies have suggested promising trends of
increasing temperature and changing precipitation patterns as well as increased intensity and
duration of storm events that are likely to result in more flooding events in the Northeast region
(IPCC 2007; Rock et al. 2001). In the United States, flooding mitigation strategies have focused
on structured engineering solutions such as dams and dikes along streams and rivers since the
late 1910s (Godschalk 1999). In recent decades, scholars have called for “soft” and non‐
structural strategies such as green infrastructure (Thomas and Littlewood 2010) and land use
planning (Burby 1998; Godschalk 2004) to be integrated into planning and design interventions
for comprehensive hazard mitigation and stormwater management.
Detention is among the most prevalent stormwater management practices for flood
mitigation; however, the perceived benefits could be overestimated without empirical study
(Beecham et al. 2005). If growth development trends continue to increase impervious surfaces
and consume more floodplains, wetlands, forest and agricultural lands, fewer open lands would
be available for natural on‐site detention (i.e., depressional land areas designated for temporary
surface water runoff or flood storage) in order to mitigate flooding, particularly when the
118
impacts from climate change are at a higher level. For local planners and stakeholders to make
adaptive land use decisions about climate change, this paper aims to answers two key
questions: (1) To what degree and in what way does climate change have an impact on long‐
term flooding hazards? (2) How much detention area in the watershed would be needed for
mitigating riparian flooding hazards induced by climate change? The hypothesis was that
detention would reduce long‐term climate change‐induced flooding hazards. Therefore, green
infrastructure that includes detention functions can serve as a part of the climate change
adaptation strategy.
5.2 Background
5.2.1 Planning origins for green infrastructure in hazard mitigation
For the first 140 years, the United States had no plans for dealing with catastrophic
disasters. Several seminal disasters that occurred on grand scales cast shadows on the country:
the Chicago Fire (1871), the Boston Fire (1872), the Dam Break in Johnstown, Pennsylvania,
(1889), the Galveston Hurricane (1900), the Baltimore Fire (1904), the San Francisco Earthquake
(1906), the Miami Hurricane (1926), the Lower Mississippi Flood (1927), and the New England
Hurricane (1938). Policy responses after those disasters were sporadic and risk management
efforts in hazard mitigation were fragmented. Furthermore, disaster response and recovery
were primarily the responsibility of the local governments, social networks and influential
individuals (Platt 1999). Local governments subsequently established codes and regulations to
ensure the safety of public works, such as the Fire Limits Ordinance of 1872 in Chicago (Rosen
1986) and more stringent building and fire codes in New York City (Platt 1999). The first federal
disaster policy passed by the U.S. Congress was the 1917 Flood Control Act (Godschalk 1999).
119
Ever since, the U.S. Army Corps of Engineers undertook most flooding mitigation efforts using
structural solutions, such as levees and flood walls along major rivers (Burby 1998). By the late
1960s, flood control structures dominated riverine flood management (Godschalk 1999).
Since 1950, the federal government has progressively taken the lead on risk
management and shifted its mitigation focus from structural to non‐structural strategies such as
implementation of insurance programs, integration with land use planning, and supporting
building and construction codes. The environmental movement spurred by Rachel Carson’s
book, Silent Spring, in 1962, was a successful seminal social movement responding to the
environmental impacts caused by urbanization and economic development. The movement was
spurred on the awareness of “Land Ethics” which states that ecosystems have their intrinsic
values (Leopold 1949), and recognizes that it is critical to protect the environment which
provides basic quality of life, such as clean air, water, and soil, in addition to the benefits of
natural disaster mitigation.
The seminal Disaster Relief Act of 1950 marked the first general federal disaster
management policy (Godschalk 1999; Platt 1999). This policy opened the door for federal
disaster assistance to be accessible to local government for alleviating fiscal burdens on the
communities. After Hurricane Betsy in 1965, Congress requested that the Department of
Housing and Urban Development (HUD) to revisit a flood insurance program drafted in 1956.
Policy reports suggested diverting part of the federal disaster relief assistance to property
owners who would pay flooding insurance. Subsequently, Congress passed the National Flood
Insurance Act of 1968 to mandate that homeowners who live on floodplains purchase flooding
insurance. It also incorporated non‐structural mitigation by identifying and mapping of 100‐year
floodplains, encouraging new or substantially improved structures to meet a minimum federal
120
standard in order to be charged actuarial rates for flood insurance coverage, and providing
funding for the purchase of chronically floodprone properties (Platt 1999).
Together with the “Design with Nature” concept termed by Ian McHarg (1969),
environmental planning became an essential part of landscape and urban planning practices. It
focused on incorporating natural and social hazard assessment in the protection of natural
resources and open space (McHarg 1969). McHarg’s innovative and inclusive approach, which
used spatial analysis mapping, established the fundamental method in environmental planning.
The parks and open space reform led by Frederick Law Olmsted Sr. was a pioneering work of
green infrastructure that employs ecological services in multiple functions such as cleansing air
pollution, providing recreation, and mitigating floods in the cities (Fabos 1995). Central Park in
New York City (1857) and Boston’s Emerald Necklace park system with the “Muddy River”
proposed in 1880 are famous projects resulting from Olmsted’s visions for “bringing nature to
the city” (Fabos 1995). In the late 20th century, non‐structural mitigation was fortified through
the protection and restoration of ecological services and natural functions. Programs such as
the 1994 Unified National Program for Floodplain Management, the wetland mitigation
program, protection of coastal dunes and beaches, and riverine restoration (Godschalk 1999)
have played an important role in non‐structural natural hazard mitigation. The concept of
preserving natural areas and using ecological design for protecting, enhancing, and restoring
ecosystem services in order to improve environmental quality and provide hazard mitigation is
fundamental in the concept of green infrastructure.
5.2.2 Green infrastructure practices for flooding mitigation
Green infrastructure, in lieu of grey infrastructure, is a system that “uses natural
systems—or engineered systems that mimic natural processes—to enhance overall
121
environmental quality and provide utility services” defined by the U.S. Environmental Protection
Agency (EPA 2013). In acknowledging what ecosystem services can provide to achieve the same
functions that are designed to serve urban utility usage and to better improve environmental
quality, green infrastructure has been promoted nationwide as part of the alternative
stormwater best management practices (BMPs). The enhanced ecological functions
consequently help to increase the resilience of ecosystems in absorbing impacts from flooding
and climate change (Ahern 2011; Folke 2006). Therefore, green infrastructure is as a key
concept in integrating stormwater and risk management as a climate change adaptation
strategy in spatial planning (Gill et al. 2007).
Stormwater BMPs are tools to mitigate stormwater runoff impacts not only on water
quantity (i.e., volume and rate) but also water quality (EPA 2013). Stormwater BMPs embrace
both structural and non‐structural stormwater management strategies (Urbonas and Doerfer
2005). Structural BMPs emphasize ecological engineering design including infiltration trenches
and wells, vegetated swales and rain gardens, vegetated buffer strips, porous pavements,
infiltration basins, vegetated waterways, inlet straps and oil grit separators, detention basins
(wet and dry), constructed wetlands, in‐stream aeration, street sweeping, and greenroofs .
Non‐structural BMPs emphasize policy and regulations that help to alleviate the root of the
problem—urbanization—and engage the public. These include, but not limited to, land use
planning, natural resources management, streams and wetlands restoration, management of
household chemicals, on‐site programs of runoff management (Ellis and Marsalek 1996).
5.3 Methods
The goal of this research is to examine the effects of using green infrastructure as a
long‐term flooding hazard mitigation strategy. This study was built upon the SWAT model that
122
was developed previously for the Charles River Watershed and applied the same climate change
sensitivity test methodology for constructing the long‐term flooding hazard indices (see Chapter
3 for details). Based on the GCMs projections with increased temperature and precipitation
change illustrated in Chapter 3, this study employed 36 climate change conditions as climate
inputs in SWAT modeling. The climate sensitivity tests therefore included combinations of mean
temperature (0, +1, +2, +3˚C), mean precipitation (0, +10, +20%) and precipitation variation (0,
+10, +20%). In addition, the fraction of the land area used for detention in the sub‐watershed
area was the primary green infrastructure strategy tested in SWAT, serving as the independent
variable. The dependent variable was the long‐term flooding hazard index (HI) derived from the
output of SWAT modeling. Finally, land use was a control variable based on currently available
GIS data (MassGIS 2005 Land Use, see Chapter 3 for data descriptions) (Figure 5.1).
Figure 5.1 Research framework for using green infrastructure for mitigating climate
change impacts on long‐term flooding hazards
SWAT has an impoundment water routing function for modeling water that is
temporarily stored for the water supply or for flooding control and mitigation. Besides
reservoirs, wetlands, and ponds, which were controlled by land use in this study, the function of
depressions/potholes (referred to as “detention” hereafter) was employed to simulate the
function of stormwater management practices using detention areas.
Current condition (2005 Land Use)
36 Climate Change combinations
Climate Change Impact
SWAT Modeling
Urbanization Impact
OUTPUT
Long‐Term Flooding Hazard Index (HI)
INPUT
Fraction of detention
area in sub‐watershed
Green Infrastructure
123
Detentions are closed depression areas in the watershed which function as temporary
water storage areas. Surfacewater and precipitation are the main source of the inflow and
when storage exceeds the maximum volume assigned for each detention, the excessive volume
then becomes surfacewater and contributes to stream baseflow. In addition to water overflow,
detentions loose water through evaporation and seepage during the day. In SWAT, only one
detention in each subbasin was created through assigning one Hydrologic Response Unit (HRU,
see Chapter 3). To optimize water storage function in the model, HRUs with the largest AGRL
SWAT land use category were selected as detentions. In addition, 100% of the selected HRU
area was assigned as the drainage area for its respective detention (POT_FR=1). Furthermore,
for the consistency of the long‐term flooding hazard defined in this study, the maximum storage
for each detention was the volume of the bankfull discharge volume in its respective subbasin.
Steamflow involves a spatial directional relationship between upstream and
downstream subbasins. In order to account for a spatially connected concept, the term “sub‐
watershed” is introduced in this study. Not all subbasins have hydrological connections with
one another. For upstream subbasins with first‐order streams, the output of the water budget
is the result of the sum of each HRU in its own subbasin. For downstream subbasins with
second‐order streams, the streamflow output is accumulative from upstream subbasins. For
example, the outflow at subbasin No. 18 is the accumulation of the output from subbasin No. 18
itself plus upstream subbasins No. 17 and No. 20. Therefore, the effect of detention is the result
in its respective sub‐watershed. 54 subbasins delineated in SWAT modeling from Chapter 3
technically resulted in 54 sub‐watersheds, including sub‐watershed No. 6 that is actually the
entire Charles River Watershed. As a result, the total area of 3.2% (25 km2) of the Charles River
Watershed area was modeled as detention areas in this study. On average, each sub‐watershed
124
consisted of 2.9% of land for detention, with a range of 0.9% to 8.7% and a standard deviation of
1.5%.
Finally, Pearson’s correlation and linear regression were employed for analyzing the
relationship between the fraction of detention areas in sub‐watershed areas and the long‐term
flooding hazard index (HI) for each climate change combination.
Y = aX + b (2)
Y: HI of each sub‐watershed
X: Fraction of detention (pothole) area in sub‐watershed area
a: X variable coefficient
b: Intercept constant
5.4 Results
The Pearson’s correlation (Pr) analysis of the long‐term flooding hazard indices (HI) for
climate sensitivity tests and the presence of detention areas indicated that only ten climate
combinations have a significantly moderate to weak relationship (Pr between ‐.272 to ‐.390, p
value <0.05) (Table 5.1). Correlation coefficient values between ‐.2 and ‐.3 are considered weak
relationships; correlation coefficient values between ‐.3 and ‐.5 are considered moderate
relationships. The other 26 climate change combinations have no significant relationships with
detention (p>0.05). In general, climate change variables correlated with detention when
precipitation variation was controlled at a zero percent increase and when temperature
increased less than 3˚C. All 36 combinations are correlated negatively except for two
combinations, where the mean temperature increased 3˚C in combination with only the
precipitation variation change.
125
The top ten significantly correlated data sets were examined further with regression
analysis (Table 5.2). The results illustrated that detention area had a weak (R square ranges
from 0.07 to 0.15) yet significant (p<0.001) effect in reducing long‐term flooding hazard indices
from 0.06% to 0.28% for every 1% increase of detention area in the sub‐watershed. In addition,
increasing the mean precipitation resulted in a trend with a steep slope, while increasing the
mean temperature resulted in a trend with a gentle slope.
Considering the parameters of climate change and flooding hazards for planning, two
hazard mitigation policy goals were examined: (1) reduce long‐term flooding hazard indices (HI)
to zero, and (2) mitigate flooding impacts to the level in current climate conditions (HI=0.013,
which is the intercept of the regression equation) (Table 5.2). The results illustrated that an
average of 14% with a range of 13 to 18% of detention area in a sub‐watershed would be
needed for reaching the first goal of zero hazards, while an average of 5% and up to 8% of
detention area would be needed for reaching the second goal in mitigating climate change‐
induced hazards to the current climate conditions.
A steeper slope represents a greater effect from applying detention in respective
climate change combinations (Figure 5.2). For example, when the temperature increases 1°C in
combination with a mean precipitation increase of 20%, every 1% increase in detention area
could decrease HI between 0.25% to 0.28%. However, no detention needed for mitigating
climate change‐induced flooding hazards to current climate conditions when only temperature
increases with no precipitation change.
126
Table 5.1 Pearson’s correlation coefficient between long‐term flooding hazard index (HI)
and the fraction of detention area modeled in sub‐watershed in climate change conditions
Climate Change Variables HI Statistics Correlation Tmp mean
(+˚C) Pcp mean
(+%) Pcp var (+%)
Mean Std. Dev.
Pearson’s r (Pr)
Relationship to detention
0 0 0 1.0% 0.4% ‐.390** Significant
1 0 0 0.9% 0.4% ‐.366** Moderate
0 10 0 1.6% 0.7% ‐.359** to Weak
1 10 0 1.4% 0.6% ‐.358**
1 20 0 2.6% 1.2% ‐.346*
1 20 20 2.4% 1.1% ‐.344*
2 10 10 1.3% 0.5% ‐.341*
2 10 0 1.2% 0.5% ‐.329*
2 0 0 0.7% 0.3% ‐.304*
3 10 0 1.1% 0.4% ‐.272*
0 20 0 2.9% 1.4% ‐.263 Non‐
0 20 10 3.1% 1.4% ‐.255 significant
3 0 0 0.7% 0.2% ‐.255 Weak
0 20 20 3.2% 1.4% ‐.255
1 20 10 2.7% 1.2% ‐.242
0 10 10 2.0% 0.8% ‐.234
0 10 20 2.1% 0.8% ‐.233
1 10 10 1.6% 0.7% ‐.233
1 10 20 1.9% 0.8% ‐.233
2 20 0 2.3% 1.1% ‐.228
2 20 20 2.6% 1.1% ‐.215
2 20 10 2.4% 1.1% ‐.213
0 0 10 1.3% 0.4% ‐.212
0 0 20 1.4% 0.5% ‐.204
2 10 20 1.8% 0.6% ‐.188 Non‐
1 0 20 1.3% 0.4% ‐.180 significant
1 0 10 1.1% 0.4% ‐.174 Very Weak
3 20 0 2.2% 0.9% ‐.140
3 20 10 2.3% 0.9% ‐.129
3 20 20 2.5% 1.0% ‐.126
3 10 20 1.7% 0.6% ‐.098 Negligible
3 10 10 1.5% 0.5% ‐.085
2 0 20 1.2% 0.4% ‐.073
2 0 10 1.0% 0.3% ‐.061
3 0 10 1.1% 0.3% .035
3 0 20 1.0% 0.3% .036
T
re
r
*p
Fi
cl
Table 5.2
eaching haza
Clim
Pr rank
Tmpmean(+˚C
1 0 2 1 3 0 4 1 5 1 6 1 7 2 8 2 9 2 10 3
p<0.05 **p<
igure 5.2
limate chang
Summary o
rd mitigation
mate Change
p n )
Pcp mean (+%)
0 0 10 10 20 20 10 10 0 10
0.001 tmp: t
Significant
e conditions
C
of regression
n goals
Variables
Pcp var (+%)
0 0 0 0 0 20 10 0 0 0
emperature;
effects of de
Current Climat
127
coefficients
Regres
X variablea
‐0.101**‐0.084**‐0.162**‐0.146**‐0.281*‐0.247*‐0.119*‐0.109*‐0.057*‐0.076*
pcp: precipit
etention in m
e Conditions
HI = 1.3%
and percent
ssion Coeffici
Intercept b
* 0.013*** 0.011*** 0.021*** 0.018** 0.034** 0.031** 0.017** 0.015** 0.009** 0.013**
tation; var: va
itigating long
Clim+tm
tmpcppva
detention ar
ients %re
R Square
H
0.152 0.134 0.129 0.128 0.120 0.119 0.116 0.108 0.092 0.074
ariation
g‐term floodi
mate Change mp(˚C)+pcp(%
p: temperatp: precipitatar: precipitat
ea required f
% Detention aequired to reHI=0 HI=0.
13 0
14 0
13 5
13 4
13 8
13 8
15 4
14 2
16 0
18 0
ng hazards in
Conditions )+pvar(%)
ure mean tion mean tion variation
for
area ach 013
n ten
128
5.5 Discussion
5.5.1 Detention effects under climate change impacts
Climate change impacts on hydrology are complex and varied from watershed to
watershed (Praskievicz and Chang 2009). From the previous results in the climate sensitivity
assessment of the long‐term flooding hazard index (HI) under the current land use for the
Charles River Watershed, increasing temperature in general would result in lower HI due to
higher evapotranspiration, while increasing both mean and variation in precipitation would
result in a higher HI (Chapter 3). Higher temperatures resulting in a lower HI explain the
insignificant effects of applying long‐term flooding mitigation strategies. For example, detention
coefficients became positive and detention requirements became negative when the mean
temperature increased by 3˚C.
5.5.2 Detention effects on reducing long‐term flooding risks
To understand the effects of mitigation strategies on reducing the flooding risks that
take social vulnerability into consideration, effects of detention on long‐term climate change‐
induced flooding risk index (RI) are shown as the difference in RI derived from the RI value after
detention treatment the minus RI value before being processed by respective climate change
impact scenarios (derived from Chapter 3) and using current RI HotSpots (derived from Chapter
4) as a reference for areas currently with high RI. The maps in Figure 5.3 illustrate that in
general, detention is effective in reducing RI throughout the basin except in Lincoln and
Brookline (in red and orange). Comparatively, detention becomes more effective (a greater
difference after treatment) during higher levels of climate change impact scenarios.
129
Figure 5.3 Detention effects on long‐term climate change‐induced flooding risk index (RI)
in climate change impact scenarios compared to RI HotSpots in current climate conditions
5.5.3 Adaptive land use planning
This study focused on using detention as one of the stormwater Best Management
Practices for flooding hazards mitigation. Detention area requires depressional land areas that
can be inundated with water for a period of time. Applying this concept to landscape planning
and design, those detention areas could possibly be applied on public recreational land use
areas such as athletic fields and parks. Based on 2005 land use data, the Charles River
Watershed has 3.6% of its land for recreational uses, including cemeteries, golf courses, passive
130
and active recreation, marinas and beaches. Excluding privately owned golf courses and
cemeteries, only 1.7% of land areas are probable for using as detention areas, which is 2% to 7%
short of reaching the current climate conditions level of HI under selected climate change
conditions. Based on land use and land cover analysis, 44% of the Charles River Watershed is
urbanized (e.g., commercial, residential, utilities) (see Chapter 2). Most of the impervious areas
are streets, building footprints, parking lots, and staging areas that are mainly under the
categories of urban land use and comprise 21% of the total watershed area.
With limited natural open space and recreational land use areas that could possibly
allow for detention areas in an urbanized watershed, achieving policy goals for reducing long‐
term flooding hazards will require more innovative and aggressive land use planning and design
in both impervious and other pervious areas. For example, detention techniques could be
implemented on residential lots for site scale detention. In addition, a recent project in Chicago
has successfully implemented detention techniques underneath impervious road surfaces.
Therefore, retrofiting BMPs to provide holistic green infrastructure networks through urban
systems (Ellis, Arcy, and Chatfield 2002) plays a critical role in mitigating climate change‐induced
flooding hazards.
Finally, an integration of many structural and non‐structural stormwater management
strategies at multiple scales is essential in both climate change mitigation and adaptation.
Multi‐objective watershed planning (Yeh and Labadie 1997)and multi‐functional design (Miguez,
Mascarenhas, and Magalhaes 2005) with adaptive planning strategies are the keys for creating
more resilient water management systems for climate change planning.
131
5.5.4 Resilient stormwater management design
This experimental and opportunistic approach in dealing with uncertainties in planning
(Abbott 2005) is known as “learning as we plan, plan as we learn” (Michael 1973), “planning as
learning” (De Geus 1988) or “learning by doing” (Kato and Ahern 2008), serving as a
fundamental strategy in adaptive planning (Folke et al. 2005; Lynch 2008)for a more robust and
resilient stormwater management and green infrastructure system (Ahern 2011).
Resilience thinking suggests that a resilient world would value diversity, ecological
variability, modularity, acknowledging slow variables, tight feedbacks, social capital, innovation,
overlap in governance, and ecosystem services (Walker and Salt 2006). In addition to adaptive
learning, Ahern (2011) proposed resilient stormwater management BMPs to (1) achieve multi‐
functionality with economic efficiency and an inherent constituency of social or political
support; (2) include redundancy and modularization with risk‐spreading and back‐up functional
meta‐systems that are decentralized and adaptable and can “contain” disturbance with
characteristics of flexibility, adaptability, and spatial segregation; (3) embrace diversity with
differential responsed to disturbance, stress and opportunity, serving as a bio‐library of memory
and knowledge, and having complementarities of resource requirements; (4) build networks and
connectivity in meta‐systems with circuitry and redundancy, risk‐spreading, and design for
functions and flows; (5) enhance adaptive capacity, treating actions as opportunities for
experimentation and innovation, such as performing “learn‐by‐doing” and “safe‐to‐fail” design
experiments (Ahern 2011).
Stormwater management has evolved to look beyond imperviousness effects and
structural BMPs in order to examine their balance with pervious surfaces and non‐structural
BMPs (Brabec 2009). Urbanization associated with unsustainable development patterns and
132
living styles is the root cause of climate change and urban flooding. There is no single solution
to resolve the effects of climate change. This is mainly because the major causes of climate
change is the production of greenhouse gases, which are intertwined with our social‐ecological
system, ranging from energy, transportation, agriculture and manufacture, to land use, urban
development patterns and buildings (Blanco, Alberti, Forsyth, et al. 2009; Blanco, Alberti,
Olshansky, et al. 2009). As climate change an integral factor of anthropogenic influence in the
hydrologic cycles in this changing world, this is a critical time for rethinking stormwater
management.
From a structural stormwater management point of view, more trial‐and‐error and
adaptive learning process are likely to occur when dealing with uncertainties in hydrologic
change as a result of climate change effects (Griffiths, Pearson, and McKerchar 2009; He, Valeo,
and Bouchart 2006). It is therefore critical to allow such experiments to happen in communities.
The success of the Street Edge Alternatives (SEA Street) program in Seattle, Washington, is an
encouraging example of adaptive strategy for safe‐to‐fail. The key to its success lies not only in
the innovative design but also in its supportive leadership and effective public participation and
outreach. This experiment of integrating urban design and stormwater BMPs eventually led to
changes in zoning and stormwater ordinances in order to accommodate alternative street edge
treatment in Seattle. This practice has spread out to the rest of the country and is being
integrated in part into local stormwater management manuals and street design, such as the
Green Alley and Sustainable Streets initiatives in Chicago, Illinois.
133
5.6 Conclusion
This study has demonstrated a range of potential climate change impacts on long‐term
flooding hazards and the effectiveness of using detention areas for mitigating long‐term flooding
hazards. Since climate change has implications on long‐term environmental hazards associated
with water resources and management, the findings are particularly timely for landscape and
urban planning for climate change. Two hazard mitigation policy goals were examined for
achieving (1) zero hazards and (2) mitigating climate change‐induced hazards for reaching the
long‐term flooding hazard index (HI) equal to the level under current climate conditions.
Even though the zero percent chance of long‐term flooding hazards is an extreme policy
goal, it provides an upper boundary for developing policy frameworks with feasible intermediate
goals. In addition, it is worth noting that detention area alone has limited potential for flooding
mitigation and is no substitution for integrated land use and watershed management strategies
such as open space and floodplain protection, wetlands restoration, and clustered and compact
growth strategies (Brody et al. 2011). Moreover, engaging the stakeholders and the public to
integrated water management and land use policies that can “Make room for River,” such as the
examples from the Dutch experience (Woltjer 2007; Wolsink 2006), plays a critical role for the
success in both long‐term and short term flooding hazard mitigation.
Finally, innovation in planning and design in order to provide multiple uses in recreation
and public lands as well as detention and infiltration under impervious surfaces in urbanized
areas play a critical role in integrating stormwater management into green infrastructure system
networks for climate change adaptation.
134
CHAPTER 6
CONCLUSION
6.1 Contribution to the field in planning
This research has attempted to contribute to the field of planning in the areas of
theoretical framing of research issues, empirical research case studies, and knowledge gained
for building innovations in planning practices.
Chapter 3, 4, and 5 answered the first four research questions posed in Chapter 1
through findings that supported or rejected hypotheses (Table 6.1). Not all hypotheses were
supported. First, increasing temperature leads to reduced long‐term flooding hazard index (HI)
in this study. Even though some literature supported this finding by suggesting that increasing
temperature resulting in higher evapotranspiration and lower baseflow, the fact is that
temperature change alone cannot explain all the variations of climate change in terms of
precipitation and seasonal change. Therefore, there remains a range of uncertainty in climate
change in accounting for the interaction between temperature and precipitation change and its
combined impacts on flooding hazards.
Second, there is no significant spatial correlation between the long‐term flooding
hazard index (HI) and the social vulnerability index (SoVI). This is due to the location of the most
vulnerable populations not being on the main channel of the river. It reveals that risk
assessment is place‐specific and hazard specific. No transferability or generalizability can be
applied from this particular finding.
The last inconsistency was a surprising finding that a “threshold” effect exists that was
not fully anticipated. Results in Chapter 4 revealed an interaction between land change and
climate change impacts on long‐term flooding hazards. In the High Impact climate change
135
Table 6.1 Summary of research hypotheses and findings
Hypotheses (Chapter 1) Findings
Climate change as a result of
increasing temperature and
precipitation change would
increase flooding hazards
Ø Hypothesis partially rejected
Increasing temperature reduces HI (Chapter 3)
Increasing precipitation change increases HI (Chapter 3)
Populations with high social
vulnerability are exposed to
higher flooding hazards as well
as flooding hazards induced by
climate change
× Hypothesis rejected
No significant spatial correlation between HI and SoVI
(Chapter 3)
The growth strategy of suburban
sprawl in the Current Trends
scenario is associated with the
most land use and land cover
change and increases flooding
hazards
o Hypothesis supported
The Current Trends scenario experiences an increase of
more than 5 times in land use change than in the
Compact Core scenario (Chapter 4)
Current Trends is associated with the highest HI among
scenarios when climate change impacts are low
(Chapter 4)
In the High Impact climate change scenario, the mean
HI becomes no difference among scenarios (Chapter 4)
The growth strategy of suburban
sprawl in the Current Trends
scenario imposes more people in
the climate change‐induced
flooding risk areas
o Hypothesis supported
The Current Trends scenario placed about 3,000 more
people in risk HotSpot areas than in the Compact Core
scenario (Chapter 4)
Green infrastructure has effects
in mitigating climate change‐
induced flooding hazards
o Hypothesis supported
Detention area increased in 1% of drainage area, HI
decreased between 0.06% and 0.28% under climate
change conditions (Chapter 5)
136
scenario, there was no difference in mean HI among growth scenarios. These inconsistent
findings critically reinforce the importance of conducting empirical studies for understanding the
interaction between human and natural systems in the interlinked social‐ecological systems.
The last research question posed in Chapter 1 was “what is the role of planning
interventions in social‐ecological resilience planning for climate change?” The findings from the
evaluation of growth strategies (i.e., infill vs. greenfill) in Chapter 4 and green infrastructure (i.e.,
detention) in Chapter 5 revealed several implications in planning.
6.1.1 On‐site detention innovations
As infill development becomes the primary growth strategy in conserving farmlands and
open space, urban areas will become denser with limited land area for the necessary detention
for mitigating climate change impacts. Urban drainage systems designed to manage stormwater
runoff have been the most critical component of flood prevention. Rainfall is the source of
runoff. As Tjallingii put it,“storage is the key issue” (Tjallingii 2012)(p 95). There exist positive
effects of detention in mitigating flooding hazards as shown in Chapter 5. Therefore, on‐site
detention in many forms (Beecham et al. 2005) can serve as a critical role in coping with climate
change‐induced flooding hazards. How to harvest rainwater in the first place before it becomes
runoff is the key to the puzzle. As a result, innovations in green infrastructure urban design such
as greenroofs, cisterns, rainbarrels, and underground storage underneath pavement or buildings
can function as detention in mitigating climate change‐induced flooding hazards, which in turn
plays a role in adapting to climate change. While cities are projected to undergo renovation and
rebuilding for growing populations, the opportunity for redeveloping cities to be more resilient
and sustainable through innovations in planning and design must be seized.
137
6.1.2 Remove people from the risk equation
First and foremost, making a proactive planning strategy to reduce the number of
people exposed to climate change‐induced hazards should be the top priority in planning. In
order to do this, a comprehensive place‐based multi‐climate change induced‐hazard risk
assessment is critical for spatial planning and growth management for current and future
populations. In planning for future growth, housing (re)development in hazardous areas should
be discouraged through zoning regulation and policy disincentives such as charging higher
development fees for climate change adaptation measures or higher flood insurance rates. In
planning for current residence, a transdisciplinary participatory planning process should be
required for every community to share information about climate change–induced hazards and
probable climate change adaptation strategies including community retreat and relocation.
6.1.3 Reduce social vulnerability in the risk equation
When people cannot be removed from the risk equation (i.e., relocation is not desired
nor feasible option), the next critical step is to reduce social vulnerability. The Social
Vulnerability Index (SoVI) has provided a tool in gauging potential population distributed
spatially in relation to hazards as well as the potential needs for risk management. Planners can
then incorporate the spatial information of social vulnerability into planning practices. In
addition, in the comprehensive planning process, planners can share information of climate
change impacts and potential hazardous areas with the stakeholders as well as the communities
for identifying appropriate strategies for climate change adaptation. For example, a series of
workshops were conducted in East Boston and found that the target population do not have
climate change adaptation perspectives nor do they have adequate resources for hazard
138
mitigation and prevention as well as resources for responding to and recovering from disasters
(Douglas et al. 2012). Therefore, planners play an important role in ensuring equitable planning
processes as well as equitably distributed access to resources (Krumholz and Forester 1990;
Fainstein 2000) in order to reduce social vulnerability in the risk equation.
6.1.4 Green equity
While investing in green infrastructure has become an emerging planning innovation as
described earlier, the interplay between greening and reducing social vulnerability can be seen
as an environmental justice issue (Danford et al. in review). Allocating investment of green
infrastructure in which it can benefit the socially vulnerable populations the most should be
included in the planning agenda for resilient and sustainable development. For example, a
detention area for mitigating flooding should be prioritized to be in the location not only where
the most people would be removed from exposure to hazards but also where the exposure to
the socially vulnerable groups would be reduced the most. In other words, green infrastructure
investments in climate change adaptation strategies should be prioritized not simply by reducing
hazards but by reducing the exposure of the socially vulnerable populations to hazards in the
risk equation.
6.2 Future directions
This research has attempted to investigate the dynamic relationship between human
and natural systems in the social‐ecological systems study with several parameters that could be
further developed for future research. First, the structure of the Social Vulnerability Index (SoVI)
was limited to quantitative data and methods. It is lacking in intangible vulnerability measures
139
that require qualitative input. For example, the question of why people live on floodplains
requires an understanding of people’s perceptions and wide understanding of flooding risks and
factors that influence people’s behavior.
In addition, it is critical to recognize the concept that places with high hazard risks do
not equal places with low resilience. There is a growing body of research in building an index for
measuring resilience, which accounts for a community’s ability and capacity for risk
management and adaptation to climate change. Moreover, further investigation is needed for
understanding the dynamics in temporal and spatial relationships in the social‐ecological
processes. For example, in what way and to what extent does land cover change in upstream
communities influence downstream flooding hazards?
Finally, this research is limited to single‐hazard risk assessment and single‐treatment for
assessing multiple green infrastructure applications. Further research frameworks and modeling
methods for multiple climate change‐induced risk assessments and evaluations of
comprehensive planning intervention strategies will provide useful knowledge and
advancement of local planning practices.
6.3 Final remarks
Thinking globally, planning locally. While climate change has posed challenges in
planning on global, national, regional and local scales, it is at the local scale that planners can
see who are most impacted and in what circumstance. Empirical study at the local scale
provides social‐ecological outcomes that can inform policy‐makers and practitioners for setting
climate change parameters for seeking innovations in planning policy and practices. This
research demonstrated one example in understanding a range of various levels of impacts from
climate change on how much land might be needed to mitigate such impacts. Through a
140
transdisciplinary participatory planning process, communities are able to communicate and
have a mutual understanding and consensus in what level of risk is acceptable. Subsequently,
communities are able to set priorities for allocating resources to enhance people’s livelihoods
and invest in green infrastructure for building resilient and sustainable communities.
141
BIBLIOGRAPHY
Abbott, J. 2005. "Understanding and managing the unknown: The nature of uncertainty in
planning." SAGE Urban Studies Abstracts no. 33 (3).
Adger, W. Neil. 2006. "Vulnerability." Global Environmental Change‐Human and Policy
Dimensions no. 16 (3):268‐281.
Ahern, Jack. 2011. "From fail‐safe to safe‐to‐fail: Sustainability and resilience in the new urban
world." Landscape and Urban Planning no. 100 (4):341‐343.
Alberti, M. 1999. "Modeling the urban ecosystem: A conceptual framework." Environment &
Planning B: Planning & Design no. 26 (4):605.
Allan, J. D. 2004. "Landscapes and riverscapes: The influence of land use on stream ecosystems."
Annual Review of Ecology Evolution and Systematics no. 35:257‐284.
Allan, J. D., D. L. Erickson, and J. Fay. 1997. "The influence of catchment land use on stream
integrity across multiple spatial scales." Freshwater Biology no. 37 (1):149‐161.
Alley, W. M., and J. E. Veenhuis. 1983. "Effective impervious area in urban runoff modeling "
Journal of Hydraulic Engineering‐Asce no. 109 (2):313‐319.
Anderies, John M., Brian H. Walker, and Ann P. Kinzig. 2006. "Fifteen weddings and a funeral:
Case studies and resilience‐based management." Ecology & Society no. 11 (1):386‐397.
Anderson, Daniel G. 1970. "Effects of urban development on floods in northern Virginia." U. S.
Geological Survey Water‐Supply Paper:C1‐c22.
Anselin, Luc, Anil K. Bera, Commerce University of Illinois at Urbana‐Champaign. College of, and
Research Business Administration. Office of. 1996. Spatial dependence in linear
regression models with an introduction to spatial econometrics. Champaign: University
of Illinois at Urbana‐Champaign.
142
Arisz, H., and B. C. Burrell. 2006. Urban drainage infrastructure planning and design considering
climate change. Paper read at 2006 IEEE EIC Climate Change Conference, 2006, at Place
of Publication: Piscataway, NJ, USA; Ottawa, ON, Canada. Country of Publication: USA.
Arnold, C. L., and C. J. Gibbons. 1996. "Impervious surface coverage: The emergence of a key
environmental indicator." Journal of the American Planning Association no. 62 (2):243‐
258.
Arnold, J. G., and P. M. Allen. 1996. "Estimating hydrologic budgets for three Illinois
watersheds." Journal of hydrology. no. 176 (1‐4):57.
Arnstein, Sherry R. 1969. "A ladder of citizen participation." In The City Reader, edited by Richard
T. LeGates and Frederic Stout, 233‐244. London; New York: Routledge.
Ashley, R. M., D. J. Balmforth, A. J. Saul, and J. D. Blanskby. 2005. "Flooding in the future:
Predicting climate change, risks and responses in urban areas." Water Science and
Technology no. 52 (5):265‐273.
Bartlett, Sheridan. 2008. "Climate change and urban children: impacts and implications for
adaptation in low‐ and middle‐income countries." Environment and Urbanization no. 20
(2):501‐519.
Beatley, Timothy. 2009. Planning for coastal resilience: Best practices for calamitous times.
Washington DC: Island Press.
Beck, Ulrich. 1992. Risk society: Towards a new modernity. London; Newbury Park, Calif.: Sage
Publications.
Beecham, S., J. Kandasamy, M. Khiadani, and D. Trinh. 2005. "Modelling on‐site detention on a
catchment‐wide basis." Urban Water Journal no. 2 (1):23‐32.
143
Bekele, Elias G., and H. Vernon Knapp. 2010. "Watershed modeling to assessing impacts of
potential climate change on water supply availability." Water Resources Management
no. 24 (13):3299‐3320.
BES. 2010. Baltimore Ecosystem Study. Cary Institute of Ecosystem Studies 2010 [cited May 4
2010]. Available from http://www.beslter.org.
Bhaskar, A. S., and C. Welty. 2012. "Water balances along an urban‐to‐rural gradient of
metropolitan baltimore, 2001‐2009." Environ. Eng. Geosci. Environmental and
Engineering Geoscience no. 18 (1):37‐50.
Birkmann, Jörn. 2006. "Measuring vulnerability to promote disaster‐resilient societies:
Conceptual frameworks and definitions." In Measuring vulnerability to natural hazards
towards disaster resilient societies, edited by Jörn Birkmann, 9‐54. New York: United
Nations University.
Blaikie, Piers M. 1994. At risk: Natural hazards, people's vulnerability, and disasters. London;
New York: Routledge.
Blanco, H., M. Alberti, R. Olshansky, S. Chang, S. M. Wheeler, J. Randolph, J. B. London, J. B.
Hollander, K. M. Pallagst, T. Schwarz, F. J. Popper, S. Parnell, E. Pieterse, and V. Watson.
2009. "Shaken, shrinking, hot, impoverished and informal: Emerging research agendas in
planning." Progress in Planning no. 72:195‐250.
Blanco, Hilda, Marina Alberti, Ann Forsyth, Kevin J. Krizek, Daniel A. Rodríguez, Emily Talen, and
Cliff Ellis. 2009. "Hot, congested, crowded and diverse: Emerging research agendas in
planning." Progress in Planning no. 71 (4):153‐205.
BMA‐ULTRA. 2012. Boston metro ecological research. University of Massachusetts Amherst
January 15, 2013 [cited March 24 2012]. Available from www.umass.edu/urbaneco/.
144
Bogardi, J. , and J. Birkmann. 2004. "Vulnerability assessment: The first step towards sustainable
risk reduction." In Disaster and Society: From Hazard Assessment to Risk Reduction,
edited by D. Malzahn and T. Plapp. Berlin: Logos Verlag Berlin.
Bohle, Hans‐Georg. 2001. "Vulnerability and criticality: perspectives from social geography."
IHDP update no. 2 (01):3‐5.
Bollin, C. , C. Cardenas, H. Hahn, and K.S. Vatsa. 2003. Natural disaster network: Disaster risk
management by communities and local governments. Washington, D. C.: Inter‐American
Development Bank.
Booth, D. B., and C. R. Jackson. 1997. "Urbanization of aquatic systems: Degradation thresholds,
stormwater detection, and the limits of mitigation." Journal of the American Water
Resources Association no. 33 (5):1077‐1090.
Booth, Derek B. 1993. Consequences of urbanization on aquatic systems: Measured effects,
degradation thresholds, and corrective strategies. Paper read at Watershed '93, a
national conference on Watershed management, at United States.
Borden, Kevin A., Mathew C. Schmidtlein, Christopher T. Emrich, Walter W. Piegorsch, and
Susan L. Cutter. 2007. "Vulnerability of U.S. cities to environmental hazards." Journal of
Homeland Security and Emergency Management no. 4 (2).
Brabec, Elizabeth A. 2009. "Imperviousness and land‐use policy: Toward an effective approach
to watershed planning." Journal of Hydrologic Engineering no. 14 (4):425‐433.
Brabec, Elizabeth, Stacey Schulte, and Paul L. Richards. 2002. "Impervious surfaces and water
quality: A review of current literature and its implications for watershed planning."
Journal of Planning Literature no. 16 (4):499.
145
Brody, Samuel D., Joshua Gunn, Walter Peacock, and Wesley E. Highfield. 2011. "Examining the
influence of development patterns on flood damages along the Gulf of Mexico." Journal
of Planning Education and Research no. 31 (4):438‐448.
Brown, C., W. Werick, W. Leger, and D. Fay. 2011. "A decision‐analytic approach to managing
climate risks: Application to the upper great lakes." J. Am. Water Resour. Assoc. Journal
of the American Water Resources Association no. 47 (3):524‐534.
Bruntland, G. H. . 1987. Our common future: The world commission on environment and
development. Oxford.
Burby, Raymond J. 1998. Cooperating with nature: Confronting natural hazards with land use
planning for sustainable communities, Natural hazards and disasters. Washington, D.C.:
Joseph Henry Press.
Burton, Thomas M., and James E. Hook. 1979. "Non‐point source pollution from abandoned
agricultural land in the Great Lakes basin." Journal of Great Lakes Research no. 5 (2):99‐
104.
Cadenasso, Mary L., and Steward T. Pickett. 2013. "Three tides: The department and state of the
art of urban ecological science " In Resilience in ecology and urban design linking theory
and practice for sustainable cities, 29‐46. Dordrecht; New York: Springer.
Campbell, Scott. 1996. "Green cities, growing cities, just cities? Urban planning and the
contradictions of sustainable development." In Readings in Planning Theory, edited by
Scott Campbell and Susan S. Fainstein, 435‐458. Malden, MA, USA: Blackwell Publishing.
Cardona, O.D. 1999. "Environmental management and disaster prevention: Two related topics:
A holistic risk assessment and management approach." In Natural Disaster
Management, edited by J. Ingleton. Lundon: Rudor Rose.
146
Cardona, O.D. , and A.H. Barbat. 2000. El riesgo sısmico y su prevencion. Madrid: Calidad
Siderurgica.
Carter, T. R., M. L. Parry, H. Harasawa, and S. Nishioka. 1994. IPCC technical guidelines for
assessing climate change impacts and adaptations. In Special Report to Working Group
II. London: Department of Geography.
Cash, David. 2011. Massachusetts climate change adaptation report. edited by Executive Office
of Energy and Environmental Affairs and the Adaptation Advisory Committee.
Chen, J., F. P. Brissette, and R. Leconte. 2011. "Uncertainty of downscaling method in
quantifying the impact of climate change on hydrology." Journal of Hydrology no. 401
(3‐4):190‐202.
Chiew, F.H.S. , T.I. Harrold, L. Siriwardena, R.N. Jones, and R. Skrikanthan. 2003. Simulation of
climate change impact on runoff using rainfall scenarios that consider daily patterns of
change from GCMs. In Proceedings of the International Congress on Modelling and
Simulation. Townsville: MODSIM.
Colten, C. E. 2006. "Vulnerability and place: Flat land and uneven risk in New Orleans." American
Anthropologist no. 108 (4):731‐734.
Costanza, R., L. Wainger, C. Folke, and K. G. Maler. 1993. "Modeling complex ecological
economic systems: Toward an evolutionary, dynamic understanding of people and
nature." Bioscience no. 43 (8):545‐555.
Crichton, David. 1999. "The risk triangle." In Natural Disaster Management, edited by Jon
Ingleton, 102‐103. London: Tudor Rose.
147
Cutter, S. L., L. Barnes, M. Berry, C. Burton, E. Evans, E. Tate, J. Webb, vulnerabilities Local
evidence on, and change adaptations to global environmental. 2008. "A place‐based
model for understanding community resilience to natural disasters." Global
Environmental Change no. 18 (4):598‐606.
Cutter, Susan L. 1996. "Vulnerability to environmental hazards." Progress in Human Geography
no. 20 (4):529‐539.
Cutter, Susan L., Bryan J. Boruff, and W. Shirley. 2003. "Social vulnerability to environmental
hazards." Social Science Quarterly (Blackwell Publishing Limited) no. 84 (2):242‐261.
Cutter, Susan L., Christopher G. Burton, and Christopher T. Emrich. 2010. "Disaster resilience
indicators for benchmarking baseline conditions." Journal of Homeland Security and
Emergency Management no. 7 (1).
Cutter, Susan L., and Christina Finch. 2008. "Temporal and spatial changes in social vulnerability
to natural hazards." Proceedings of the National Academy of Sciences of the United
States of America no. 105 (7):2301‐2306.
Cutter, Susan L., and Daniel P. Morath. 2013. "The evolution of the Social Vulnerability Index." In
Measuring vulnerability to natural hazards (Second Edition), edited by Jörn Birkmann.
New York: United Nations University.
Danford, Rachel S., Chingwen Cheng, Michael W Strohbach, Robert L. Ryan, Craig Nicolson, and
Paige S. Warren. in review. "What does it take to achieve equitable urban tree canopy
distribution? A Boston case study." Cities and the Environment.
Davidson, R. . 1997. An urban earthquake disaster risk index. Standford: The John A. Blume
Earthquake Engineering Center, Standford University.
Davis, Kingsley. 1955. "The origin and growth of urbanization in the world." American Journal of
Sociology no. 60 (5):429‐437.
148
De Geus, Arie P. 1988. "Planning as learning." Harvard Business Review no. 66 (2).
De Wrachien, D., S. Mambretti, and A. Sole. 2008. "Risk analysis and vulnerability assessment in
flood protection and river basin management." In Flood Recovery, Innovation and
Response, edited by D. Proverbs, C. A. Brebbia and E. PenningRowsell, 3‐15.
Douglas, Ellen M., Paul H. Kirshen, Michael Paolisso, Chris Watson, Jack Wiggin, Ashley Enrici,
and Matthias Ruth. 2012. "Coastal flooding, climate change and environmental justice:
Identifying obstacles and incentives for adaptation in two metropolitan Boston
Massachusetts communities." Mitigation and Adaptation Strategies for Global Change
no. 17 (5):537‐562.
Douglas, Ian, Kurshid Alam, Maryanne Maghenda, Yasmin McDonnell, Louise McLean, and Jack
Campbell. 2008. "Unjust waters: climate change, flooding and the urban poor in Africa."
Environment and Urbanization no. 20 (1):187‐205.
Ellis, J. B., B. J. D. Arcy, and P. R. Chatfield. 2002. "Sustainable urban‐drainage systems and
catchment planning." Water and Environmental Management Journal no. 16 (4):286‐
291.
Ellis, J. B., and T. HvitvedJacobsen. 1996. "Urban drainage impacts on receiving waters." Journal
of Hydraulic Research no. 34 (6):771‐783.
Ellis, J. B., and J. Marsalek. 1996. "Overview of urban drainage: Environmental impacts and
concerns, means of mitigation and implementation policies." Journal of Hydraulic
Research no. 34 (6):723‐731.
EPA. 2012. Green Infrastructure. EPA, April 23, 2013 2013 [cited March 28 2012]. Available from
http://water.epa.gov/infrastructure/greeninfrastructure/.
Fabos, J. G. 1995. "Introduction and overview: The greenway movement, uses and potentials of
greenways." Landscape and Urban Planning no. 33 (1‐3):1‐13.
149
Fainstein, Susan S. 2000. "New directions in planning theory " In Readings in Planning Theory,
edited by Scott Campbell and Susan S. Fainstein, 173‐195. Malden, MA, USA: Blackwell
Publishing.
Ficklin, D. L., Y. Luo, E. Luedeling, and M. Zhang. 2009. "Climate change sensitivity assessment of
a highly agricultural watershed using SWAT." Journal of Hydrology no. 374 (1‐2):16‐29.
Fielding, J., and K. Burningham. 2005. "Environmental inequality and flood hazard." Local
Environment no. 10 (4):379‐395.
Fohrer, N., S. Haverkamp, K. Eckhardt, and H. Frede. 2000. Policy induced land use changes and
their effects on land degradation: a modeling approach. Paper read at Congress of ESSC,
at Valencia.
Folke, C. 2006. "Resilience: The emergence of a perspective for social‐ecological systems
analyses." Global Environmental Change‐Human and Policy Dimensions no. 16 (3):253‐
267.
Folke, Carl, Thomas Hahn, Per Olsson, and Jon Norberg. 2005. "Adaptive governance of social‐
ecological systems." Annual Review of Environment & Resources no. 30 (1):441‐473.
Fontaine, T. A., J. F. Klassen, T. S. Cruickshank, and R. H. Hotchkiss. 2001. "Hydrological response
to climate change in the Black Hills of South Dakota, USA." Hydrological Sciences Journal
no. 46 (1):27‐40.
Fowler, H.J., C.G. Kilsby, and J. Stunell. 2007. "Modelling the impacts of potential future climate
change on water resources in north‐west England." Hydrology and Earth System
Sciences (11):1115‐26.
Frederick, K. D., and D. C. Major. 1997. "Climate change and water resources." Climatic Change
no. 37 (1):7‐23.
150
Gall, Melanie, Kevin A. Borden, Christopher T. Emrich, and Susan L. Cutter. 2011. "The
unsustainable trend of natural hazard losses in the United States." Sustainability no. 3
(11):2157‐2181.
Gates, W. Lawrence. 1985. "The use of general circulation models in the analysis of the
ecosystem impacts of climatic change." Climatic Change no. 7 (3):267‐284.
Gill, S. E., J. F. Handley, A. R. Ennos, and S. Pauleit. 2007. "Adapting cities for climate change: The
role of the green infrastructure." Built environment. no. 33 (1):115‐132.
Godschalk, D. R. 2004. "Land use planning challenges: Coping with conflicts in visions of
sustainable development and livable communities." Journal of the American Planning
Association no. 70 (1):5‐13.
Godschalk, David R. 1999. Natural hazard mitigation: Recasting disaster policy and planning.
Washington, D.C.: Island Press.
———. 2003. "Urban hazard mitigation: Creating resilient cities." Natural Hazards Review no. 4
(3):136.
Graham, L.P. , S. Hagemann, S. Jaun, and M. Beniston. 2007. "On interpreting hydrological
change from regional climate models." Climatic Change (81):97‐122.
Griffiths, G. A., C. P. Pearson, and A. I. McKerchar. 2009. "Climate variability and the design flood
problem." Journal of Hydrology, New Zealand no. 48 (1):29‐38.
Gutkind, Erwin Anton. 1952. Our world from the air: an international survey of man and his
environment. Garden City, New York: Doubleday.
Harding, D. M., and Dennis J. Parker. 1979. "Natural hazard evaluation, perception and
adjustment." Geography.
Hardoy, Jorgelina, and Gustavo Pandiella. 2009. "Urban poverty and vulnerability to climate
change in Latin America." Environment and Urbanization no. 21 (1):203‐224.
151
He, J., C. Valeo, and F. J. C. Bouchart. 2006. "Enhancing urban infrastructure investment planning
practices for a changing climate." Water Science and Technology no. 53 (10):13‐20.
Hewitt, Kenneth, and Ian Burton. 1971. The hazardousness of a place: a regional ecology of
damaging events. Toronto: University of Toronto Press.
Holling, C. S. 1973. "Resilience and stability of ecological systems." Annual Review of Ecology &
Systematics no. 4:1‐23.
———. 2001. "Understanding the complexity of economic, ecological, and social systems."
Ecosystems no. 4 (5):390‐405.
Hollis, G. E. 1975. "Effects of urbanization on floods of different recurrence interval " Water
Resources Research no. 11 (3):431‐435.
Hosseinzadeh, S. R. 2005. "The effects of urbanization on the natural drainage patterns and the
increase of urban floods: Case study Metropolis of Mashhad‐Iran." Sustainable
Development and Planning II, Vols 1 and 2 no. 84:423‐432.
Humphries, C. 2012. "The science of cities: Life in the concrete jungle." Nature no. 491
(7425):514‐5.
Hung, Hoang Vinh, Rajib Shaw, and Masami Kobayashi. 2007. "Flood risk management for the
RUA of Hanoi: Importance of community perception of catastrophic flood risk in disaster
risk planning." Disaster Prevention & Management no. 16 (2):245‐258.
Huq, Saleemul, Sari Kovats, Hannah Reid, and David Satterthwaite. 2007. "Editorial: Reducing
risks to cities from disasters and climate change." Environment and Urbanization no. 19
(1):3‐15.
IPCC. 2007. "Mitigation of climate change." In Climate Change 2007. Cambridge; New York:
Cambridge University Press.
152
———. 2012. Managing the risks of extreme events and disasters to advance climate change
adaptation. Edited by V. Barros C.B. Field, T.F. Stocker, D. Qin, D.J. Dokken, K.L. Ebi, M.D.
Mastrandrea, K.J. Mach, G.‐K. Plattner, S.K. Allen, M. Tignor, P.M. Midgley. Cambridge,
England: Cambridge University Press.
Jones, Julia A., Irena F. Creed, Kendra L. Hatcher, Robert J. Warren, Mary Beth Adams, Melinda
H. Benson, Emery Boose, Warren A. Brown, John L. Campbell, Alan Covich, David W.
Clow, Clifford N. Dahm, Kelly Elder, Chelcy R. Ford, Nancy B. Grimm, Donald L. Henshaw,
Kelli L. Larson, Evan S. Miles, Kathleen M. Miles, Stephen D. Sebestyen, Adam T. Spargo,
Asa B. Stone, James M. Vose, and Mark W. Williams. 2012. "Ecosystem Processes and
Human Influences Regulate Streamflow Response to Climate Change at Long‐Term
Ecological Research Sites." bisi BioScience no. 62 (4):390‐404.
Karl, Thomas R., Jerry M. Melillo, and Thomas C. Peterson. 2009. Global climate change impacts
in the United States, A State of Knowledge Report from the U.S. Global Change Research
Program. New York, NY: Cambridge University Press.
Kato, Sadahisa, and Jack Ahern. 2008. "'Learning by doing': adaptive planning as a strategy to
address uncertainty in planning." Journal of Environmental Planning & Management no.
51 (4).
Kirshen, Paul, Kelly Knee, and Matthias Ruth. 2008. "Climate change and coastal flooding in
Metro Boston: impacts and adaptation strategies." Climatic Change no. 90 (4):453‐473.
Krumholz, Norman, and John Forester. 1990. Making equity planning work : leadership in the
public sector, Conflicts in urban and regional development. Philadelphia: Temple
University Press.
153
Lammert, M., and J. D. Allan. 1999. "Assessing biotic integrity of streams: Effects of scale in
measuring the influence of land use/cover and habitat structure on fish and
macroinvertebrates." Environmental management no. 23 (2):257‐270.
Leary, Neil. 2008. Climate change and vulnerability. London; Sterling, VA: Earthscan.
Leopold, Aldo. 1949. "The land ethic." In Sustainable Urban Development Reader, edited by
Stephen M. Wheeler and Timothy Beatley, 23‐32. New York: Routledge.
Leopold, Luna B. 1968. Hydrology for urban planning: A guidebook on the hydrologic effects of
urban land use. USGS Circular.
Lindh, Gunnar. 1972. "Urbanization: A hydrological headache." AMBIO: A Journal of the Human
Environment no. 1 (6):185‐201.
Liu, J. G., T. Dietz, S. R. Carpenter, M. Alberti, C. Folke, E. Moran, A. N. Pell, P. Deadman, T. Kratz,
J. Lubchenco, E. Ostrom, Z. Ouyang, W. Provencher, C. L. Redman, S. H. Schneider, and
W. W. Taylor. 2007. "Complexity of coupled human and natural systems." Science no.
317 (5844):1513‐1516.
Lynch, Amanda H. 2008. "Adaptive governance: how and why does government policy change?"
Ecos (146):31‐31.
Lynn, Kathy, Katharine MacKendrick, and Ellen M. Donoghue. 2011. Social vulnerability and
climate change: synthesis of literature. USDA Forest Service.
Maantay, Juliana, and Andrew Maroko. 2009. "Mapping urban risk: flood hazards, race, and
environmental justice in New York." Applied Geography no. 29 (1):111‐124.
MAPC. 2009. From plan to action: a MetroFuture summary edited by Jessie Grogan:
Metropolitan Area Planning Council.
———. 2011. The State of Equity in Metro Boston. edited by Mariana Arcaya and Jessie Grogan:
Metropolitan Area Planning Council.
154
McHarg, Ian L. 1969. Design with nature. Garden City, N.Y.: Published for the American Museum
of Natural History [by] the Natural History Press.
Means, E. G., N. West, and R. Patrick. 2005. "Population growth and climate change will pose
tough challenges for water utilities." Journal American Water Works Association no. 97
(8).
Meppem, Tony, and Roderic Gill. 1998. "Planning for sustainability as a learning concept."
Ecological Economics no. 26 (2):121‐137.
Michael, D. N. 1973. On learning to plan– and planning to learn. San Francisco, CA: Jossey‐Bass.
Miguez, M. G., F. C. B. Mascarenhas, and L. P. C. Magalhaes. 2005. "Multifunctional landscapes
for urban flood control in developing countries." In Sustainable Development and
Planning II, Vols 1 and 2, edited by A. Kungolos, C. A. Brebbia and E. Beriatos, 1579‐
1588.
Muller, Mike. 2007. "Adapting to climate change: water management for urban resilience."
Environment and Urbanization no. 19 (1):99‐113.
NECIA. 2006. Climate change in the U.S. Northeast. Northeast Climate Impacts Assessment.
Neitsch, S. L., J. G. Arnold, J. R. Kiniry, J. R. Williams, and K. W. King. 2009. Soil and Water
Assessment Tool theoretical documentation. Texas A&M University.
Nemec, J., and J. Schaake. 1982. "Sensitivity of water resources systems to climate variation."
Hydrological Sciences (27):327‐343.
Newman, Peter, and Isabella Jennings. 2008. "Modeling cities as ecosystems." In Cities as
sustainable ecosystems: Principles and practices, 92‐142. Washington D. C.: Island Press.
NRC. 2009. Urban stormwater management in the United States. Edited by Pollution National
Research Council . Committee on Reducing Stormwater Discharge Contributions to
Water. Washington, D.C.: National Academies Press.
155
———. 2010. Adapting to the impacts of climate change. Edited by National Research Council.
Washington DC: The National Academies Press.
NRCS. 1998. "Disturbance affecting stream corridors." In Federal Stream Corridor Restoration
Handbook. USDA Natural Resources Conservation Service.
O'Keefe, Phil, Ken Westgate, and Ben Wisner. 1976. "Taking the naturalness out of natural
disasters." Nature Nature no. 260 (5552):566‐567.
O'Riordan, Timothy. 1986. "Coping with environmental hazards." In Themes from the work of
Gilbert F. White, edited by Gilbert F. White, Robert W. Kates and Ian Burton, 272‐309.
Chicago: University of Chicago Press.
Oliver‐Smith, Anthony. 1986. The martyred city: death and rebirth in the Andes. Albuquerque:
University of New Mexico Press.
Oliver‐Smith, Anthony 2003. "Theorizing vulnerability in a globalized world: a political ecological
perspective." In Mapping vulnerability : disasters, development, and people, edited by
Greg Bankoff, Georg Frerks and Thea Hilhorst, 10‐24. London; Sterling, VA: Earthscan
Publications.
Oven, K. J., S. E. Curtis, S. Reaney, M. Riva, M. G. Stewart, R. Ohlemuller, C. E. Dunn, S. Nodwell,
L. Dominelli, R. Holden, and A. Geographic Perspective The Health Impacts of Global
Climate Change. 2012. "Climate change and health and social care: Defining future
hazard, vulnerability and risk for infrastructure systems supporting older people's health
care in England." Applied Geography no. 33:16‐24.
Paul, Michael J., and Judy L. Meyer. 2001. "Streams in the urban landscape." Annual Review of
Ecology and Systematics (32):333‐365.
156
PCSD. 1999. Towards a sustainable America: Advancing prosperity, opportunity, and a healthy
environment for the 21st century. Washington D.C.: The President's Council on
Sustainable Development.
Pendall, Rolf, Kathryn A. Foster, and Margaret Cowell. 2010. "Resilience and regions: Building
understanding of the metaphor." Cambridge Journal of Regions, Economy and Society
no. 3 (1):71‐84.
Pickett, S. T. A., and M. L. Cadenasso. 2006. "Advancing urban ecological studies: Frameworks,
concepts, and results from the Baltimore Ecosystem Study." Austral Ecology no. 31
(2):114‐125.
Pickett, S. T. A., M. L. Cadenasso, and J. M. Grove. 2004. "Resilient cities: Meaning, models, and
metaphor for integrating the ecological, socio‐economic, and planning realms."
Landscape and Urban Planning no. 69 (4):369‐384.
Pickett, S. T. A., M. L. Cadenasso, J. M. Grove, C. H. Nilon, R. V. Pouyat, W. C. Zipperer, and R.
Costanza. 2001. "Urban ecological systems: Linking terrestrial ecological, physical, and
socioeconomic components of metropolitan areas." Annual Review of Ecology and
Systematics no. 32:127‐157.
Platt, Rutherford H. 1999. Disasters and democracy: The politics of extreme natural events.
Washington, D.C.: Island Press.
Polsky, Colin, Rob Neff, and Brent Yarnal. 2007. "Building comparable global change vulnerability
assessments: The vulnerability scoping diagram." Global environmental change : human
and policy dimensions. no. 17 (3):472.
Praskievicz, Sarah, and Heejun Chang. 2009. "A review of hydrological modelling of basin‐scale
climate change and urban development impacts." Progress in Physical Geography no. 33
(5):650‐671.
157
Price, Katie. 2011. "Effects of watershed topography, soils, land use, and climate on baseflow
hydrology in humid regions: A review." Progress in Physical Geography no. 35 (4):465‐
492.
Prudhomme, C., R. L. Wilby, S. Crooks, A. L. Kay, and N. S. Reynard. 2010. "Scenario‐neutral
approach to climate change impact studies: Application to flood risk." Journal of
Hydrology no. 390 (3–4):198‐209.
Reed, Seann, Dennis Johnson, and Timothy Sweeney. 2002. "Application and national
geographic information system database to support two‐year flood and threshold runoff
estimates." Journal of Hydrologic Engineering (May/June):209.
Reid, Colleen, Marie O'Neill, Carina Gronlund, Shannon Brines, Dan Brown, Ana Diez‐Roux, and
Joel Schwartz. 2009. "Mapping Community Determinants of Heat Vulnerability."
Environ. Health Perspect. Environmental Health Perspectives (4).
Richards, C. E., C. L. Munster, D. M. Vietor, J. G. Arnold, and R. White. 2008. "Assessment of a
turfgrass sod best management practice on water quality in a suburban watershed."
Journal of Environmental Management Journal of Environmental Management no. 86
(1):229‐245.
Rittel, Horst W. J., and Melvin M. Webber. 1973. "Dilemmas in a general theory of planning "
Policy Sciences no. 4:155‐169.
Rock, Barrett N., Lynne Carter, Henry Walker, James Bradbury, S. Lawrence Dingman, and C.
Anthony Federer. 2001. Climate impacts on regional water. In The New England
Regional Assessment: University of New Hampshire.
Rogers, George Oliver, and Buren B. DeFee Ii. 2005. "Long‐term impact of development on a
watershed: early indicators of future problems." LAND Landscape and Urban Planning
no. 73 (2):215‐233.
158
Rosen, Christine Meisner. 1986. The limits of power: Great fires and the process of city growth in
America. Cambridge; New York: Cambridge University Press.
Roth, N. E., J. D. Allan, and D. L. Erickson. 1996. "Landscape influences on stream biotic integrity
assessed at multiple spatial scales." Landscape Ecology no. 11 (3):141‐156.
Ryan, Robert L., Paige S. Warren, Craig Nicolson, Chingwen Cheng, Rachel Danford, and Michael
Strohbach. 2013. Scenario planning for the Boston metropolitan region: Exploring
environmental and social implications of alternative futures. Paper read at Fábos
Conference on Landscape and Greenway Planning 2013: Pathways to Sustainability, at
Amherst, Massachusetts.
Sala, Maria. 2003. "Floods triggered by natural conditions and by human activities in a
mediterranean coastal environment." Geografiska Annaler. Series A, Physical Geography
no. 85 (3/4):301‐312.
Santhi, C., N. Kannan, J. G. Arnold, and M. Di Luzio. 2008. "Spatial calibration and temporal
validation of flow for regional scale hydrologic modeling." Journal of the American
Water Resources Association no. 44 (4):829‐846.
Satterfield, T. A., C. K. Mertz, and P. Slovic. 2004. "Discrimination, vulnerability, and justice in the
face of risk." Risk Analysis no. 24 (1):115‐129.
Savini, John, and J. C. Kammerer. 1961. "A review, classification, and preliminary evaluation of
the significance of the effects of urbanization on the hydrologic regimen." U. S.
Geological Survey Water‐Supply Paper:A1‐A42.
Schneiderbauer, Stefac, and Daniele Ehrlich. 2006. "Social levels and hazard (in)dependence in
determining vulnerability." In Measuring vulnerability to natural hazards towards
disaster resilient societies, edited by Jörn Birkmann, 78‐102. New York: United Nations
University.
159
Schueler, Thomas R. 1994. "The Importance of Imperviousness." Watershed Protection
Techniques no. 1 (3):100‐111.
Schueler, Thomas R., Lisa Fraley‐McNeal, and Karen Cappiella. 2009. "Is impervious cover still
important? Review of recent research." Journal of Hydrologic Engineering no. 14
(4):309‐315.
Semadeni‐Davies, A., C. Hernebring, G. Svensson, and L. G. Gustafsson. 2008. "The impacts of
climate change and urbanisation on drainage in Helsingborg, Sweden: Suburban
stormwater." Journal of Hydrology no. 350 (1‐2):114‐125.
Sheridan, S. C., and T. J. Dolney. 2003. "Heat, mortality, and level of urbanization: measuring
vulnerability across Ohio, USA." Clim. Res. Climate Research no. 24:255‐265.
Shuster, W. D., J. Bonta, H. Thurston, E. Warnemuende, and D. R. Smith. 2005. "Impacts of
impervious surface on watershed hydrology: A review." Urban Water Journal no. 2
(4):263‐275.
Simmons, D. L., and R. J. Reynolds. 1982. "Effects of urbanization on base‐flow of selected
South‐shore streams, Long Island, New York." Water Resources Bulletin no. 18 (5):797‐
805.
Skiles, J. W. , and J. D. Hanson. 1994. "Responses of arid and semiarid watersheds to increasing
carbon dioxide and climate change as shown by simulation studies." Climatic Change
(26):377‐397.
Slovic, P. 1999. "Trust, emotion, sex, politics, and science: Surveying the risk‐assessment
battlefield (Reprinted from Environment, ethics, and behavior, pg 277‐313, 1997)." Risk
Analysis no. 19 (4):689‐701.
Spirn, Anne Whiston. 1984. The granite garden: Urban nature and human design. New York:
Basic Books.
160
Steiner, Frederick R. 2002. Human ecology: Following nature's lead. Washington D.C.: Island
Press.
Steinschneider, S., and C. Brown. in review. "A semiparametric multivariate, multi‐site weather
generator with low‐frequency variability for use in climate risk assessments." Water
Resources Research.
Takle, Eugene S., Manoj Jha, and Christopher J. Anderson. 2005. "Hydrological cycle in the upper
Mississippi River basin: 20th century simulations by multiple GCMs." Geophysical
research letters. no. 32 (18):n.p.
TAMU. 2011. SWAT: Soil & Water Assessment Tool 2011 [cited December 11 2011]. Available
from http://swatmodel.tamu.edu/.
Tate, E., S. L. Cutter, and M. Berry. 2010. "Integrated multihazard mapping." Environ. Plann. B
Plann. Des. Environment and Planning B: Planning and Design no. 37 (4):646‐663.
Thomas, Kevin, and Steve Littlewood. 2010. "From green belts to green infrastructure? The
evolution of a new concept in the emerging soft governance of spatial strategies."
Planning Practice & Research no. 25 (2):203‐222.
Thornthwaite, C. W., and John Russell Mather. 1955. The water balance. Centerton, N.J.
Tjallingii, Sybrand. 2012. "Water flows and urban planning." In Sustainable urban environments :
an ecosystem approach, edited by Ellen van Bueren, Laure Itard, Hein van Bohemen and
Henk Visscher. Dordrecht [Netherlands]; New York: Springer.
Turner, B. L., R.E. Kasperson, P.A. Matson, J.J. McCarthy, R.W. Corell, L. Christensen, N. Eckley,
J.X. Kasperson, A. Luers, M.L. Martello, C. Polsky, A. Pulsipher, and A. Schiller. 2003. "A
framework for vulnerability analysis in sustainability science." Proceedings of the
National Academy of Sciences (14):8074‐8079.
161
Urbonas, B. R., and J. T. Doerfer. 2005. "Master planning for stream protection in urban
watersheds." Water Science and Technology no. 51 (2):239‐247.
Vale, Lawrence J., and Thomas J. Campanella. 2005. The resilient city: How modern cities recover
from disaster. New York: Oxford University Press.
Walker, Brian H., and David Salt. 2006. Resilience thinking: Sustaining ecosystems and people in
a changing world Washington, D.C.: Island Press.
Walker, G., and K. Burningham. 2011. "Flood risk, vulnerability and environmental justice:
evidence and evaluation of inequality in a UK context." Critical Social Policy no. 31
(2):216‐240.
White, Gilbert F. 1945. Human adjustment to floods: A geographical approach to the flood
problem in the United States, University of Chicago, Chicago, Illinois.
White, Gilbert F., and J. Eugene Haas. 1975. Assessment of research on natural hazards.
Cambridge, Mass.: MIT Press.
Wilkinson, Cathy. 2012. "Social‐ecological resilience: Insights and issues for planning theory."
Planning Theory no. 11 (2):148‐169.
Wolsink, Maarten. 2006. "River basin approach and integrated water management: Governance
pitfalls for the Dutch Space‐Water‐Adjustment Management Principle." Geoforum no.
37 (4):473‐487.
Woltjer, Johan Al Niels. 2007. "Integrating water management and spatial planning." Journal of
the American Planning Association no. 73 (2):211‐222.
Wood, Andrew W., Dennis P. Lettenmaier, and Richard N. Palmer. 1997. "Assessing Climate
Change Implications for Water Resources Planning." Climatic Change no. 37 (1):203‐228.
162
Wu, Kangsheng, and Carol A. Johnston. 2007. "Hydrologic response to climatic variability in a
Great Lakes watershed: A case study with the SWAT model." Journal of Hydrology no.
337 (1/2).
Yeh, C. H., and J. W. Labadie. 1997. "Multiobjective watershed‐level, planning of storm water
detention systems." Journal of Water Resources Planning and Management‐Asce no.
123 (6):336‐343.
Yohe, G., and R. S. J. Tol. 2002. "Indicators for social and economic coping capacity: Moving
toward a working definition of adaptive capacity." Global Environmental Change:
Human and Policy Dimensions no. 12 (1):25‐40.
Zhang, Yang. 2010. "Residential housing choice in a multihazard environment: Implications for
natural hazards mitigation and community environmental justice." Journal of Planning
Education and Research no. 30 (2):117‐131.