EPA/600/R-ll/123a
November, 2011
BASINS and WEPP Climate Assessment Tools (CAT):
Case Study Guide to Potential Applications
NOTICE
THIS DOCUMENT IS A DRAFT. This document is distributed solely for the purpose ofpre-
dissemination peer review under applicable information quality guidelines. It has not been
formally disseminated by EPA. It does not represent and should not be construed to represent
any Agency determination or policy. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use
Global Change Research Program
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC 20460
Draft - Do not cite or quote
-------
DISCLAIMER
This document is distributed solely for the purpose of pre-dissemination peer review under applicable
information quality guidelines. It has not been formally disseminated by EPA. It does not represent and
should not be construed to represent any Agency determination or policy. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
Draft - Do not cite or quote
-------
FOREWORD
Draft - Do not cite or quote
-------
AUTHORS AND REVIEWERS
The National Center for Environmental Assessment (NCEA), Office of Research and Development, was
responsible for preparing this external review draft report. Preparation of an earlier draft report was
conducted by AQUA TERRA Consultants, under EPA Contract EP-C-06-029.
AUTHORS
AQUA TERRA Consultants, Decatur, GA
Paul Hummel
Paul Duda
Tong Zhai
Elizabeth Wolfram
U.S. Environmental Protection Agency, National Center for Environmental Assessment, Global Change
Research Program, Washington, DC
Thomas Johnson
Meredith Warren
REVIEWERS
We are very grateful for the many excellent comments and suggestions about improving this report
provided by EPA reviewers Steve Kramer, Philip Morefield, Julie Reichert, and Christopher Weaver.
Draft - Do not cite or quote
-------
CONTENTS
LIST OF TABLES 6
LIST OF FIGURES 8
LIST OF ABBREVIATIONS 10
EXECUTIVE SUMMARY 12
1. INTRODUCTION 14
2. BASINS AND WEPP CLIMATE ASSESSMENT TOOLS 17
3. CASE STUDIES 24
3.1. Introduction 24
3.2. Climate Change Impacts on Annual and Seasonal Nutrient Loading and Streamflow in the
Upper Mississippi River Basin, Raccoon River, IA 27
3.3. Managing Climate Change Impacts on Stormwater Runoff from an Urban Redevelopment
Site in Upper Roanoke River, VA 46
3.4. Climate change impacts on sediment erosion from fields under corn and soy production,
including BMP analysis, in Blue Earth County, MN 57
3.5. Precipitation Pattern Change Impacts on Water Quality in the Tualatin River, OR 67
3.6. Assessing the impact of increased drought in Sespe Creek, CA Watershed 74
3.7. Sensitivity of Stormwater to Changes in Precipitation Volume, Event Intensity, and
Impervious Cover in the Western Branch of the Patuxent River, MD 81
3.8. Model Limitations 89
4. CONCLUDING COMMENTS 91
REFERENCES 93
APPENDIX A 98
Draft - Do not cite or quote
-------
LIST OF TABLES
Table 3.1. Summary of case studies presented in this report 25
Table 3.2. Land use summary for Raccoon River watershed 29
Table 3.3. Raccoon River model calibration statistics for yearly results. NSE: Nash-Sutcliffe Efficiency
coefficient; PB: percent bias, RMSE: root mean square error 29
Table 3.4. Mean annual streamflow (cms) for all combinations of temperature and precipitation change.
Baseline condition shown in box 32
Table 3.5. Mean annual nitrogen load (kg/ha/yr) for all combinations of temperature and precipitation
change. Baseline condition shown in box 33
Table 3.6. Mean annual phosphorous load (kg/ha/yr) for all combinations of temperature and precipitation
change. Baseline condition shown in box 33
Table 3.7. Mean annual sediment load (tonnes/ha/yr) for all combinations of temperature and
precipitation change. Baseline condition shown in box 33
Table 3.8. Mean annual evapotranspiration (cm/yr) for all combinations of temperature and precipitation
change. Baseline condition shown in box 34
Table 3.9. NARCCAP climate models used to develop the case study scenarios 36
Table 3.10. Mean monthly streamflow (cms) for all scenarios 43
Table 3.11. Mean monthly nitrogen load (kg/ha) for all scenarios 43
Table 3.12. Mean monthly phosphorous load (kg/ha) for all scenarios 43
Table 3.13. Mean monthly sediment load (tonnes/ha) for all scenarios 43
Table 3.14. Total and aMean annual precipitation, temperature, streamflow, TN, TP, and TSS loads for all
scenarios 44
Table 3.15. Baseline land use summary for the commercial redevelopment site 47
Table 3.16. Event precipitation, mean total inflow, and concentrations of TP and TSS for the baseline and
climate change scenarios 50
Table 3.17. Event mean total inflow (cms) under different rainfall intensities and stormwater management
strategies. Flow values greater than the threshold are highlighted in orange 53
Table 3.18. Soy spring chisel plow land management specifications in WEPPCAT 59
Table 3.19. Corn spring chisel land management specifications in WEPPCAT 59
Table 3.20. Mean annual sediment yield (tonnes/ha/yr) for corn production under conditions of changing
climate. Scenarios named to reflect changes in precipitation volume and intensity; V =
volume, I = intensity, numerical value reflects percent change from baseline 60
Table 3.21. Mean annual sediment yields (tonnes/ha/yr) for soy production under conditions of changing
climate. Scenarios named to reflect changes in precipitation volume and intensity; V =
volume, I = intensity, numerical value reflects percent change from baseline 61
Table 3.22. Corn fall mulch till management characteristics in WEPPCAT 63
Table 3.23. Corn no-till management characteristics in WEPPCAT 63
Table 3.24. Sediment yield (tonnes/ha/yr) resulting from corn production under all climate change and
land use scenarios. NB=no buffer, GB3=3 meter grass (generic) buffer, GB=6 meter grass
(generic) buffer, GB9=9 meter grass (generic) buffer, FB3=3 meter forest buffer, FB6=6
meter forest buffer, FB9-9 meter forest buffer, V0=no change in volume, V20=20 percent
increase in volume, V20 +110=20 percent increase in volume and 10 percent increase in
rainfall intensity 64
Table 3.25. Sediment yield (tonnes/ha/yr) results from a 2°C increase and temperature and a 20 percent
increase in mean annual rainfall volume. Notes: NB=no buffer, GB3=3 meter grass (generic)
buffer, GB=6 meter grass (generic) buffer, GB9=9 meter grass (generic) buffer, FB3=3 meter
forest buffer, FB6=6 meter forest buffer, FB9-9 meter forest buffer. Grayed areas signify
tillage and filter strip combinations producing 6 tons/ha/year or less of sediment 65
Draft - Do not cite or quote
-------
Table 3.26. Tualatin River model streamflow calibration/validation results. Relative % Error is the
average of observed-simulated/observed comparisons. Median % Error is the median of
observed-simulated comparisons/average of observed values 69
Table 3.27. Tualatin River model water quality calibration/validation results 69
Table 3.28. Precipitation, streamflow and annual loadings of TN and TSS for all climate scenarios 71
Table 3.29. Land use summary for Sespe Creek watershed 75
Table 3.30. Streamflow volume (normalized by watershed area) calibration and validation results for
Sespe Creek model 76
Table 3.31. Annual mean flow and mean annual 7-day low flow during the period of drought 79
Table 3.32. Annual mean flow and mean annual 7-day low flow for drought duration and
duration/severity scenarios 79
Table 3.33. Land use summary for Western Branch of the Patuxent River watershed 82
Table 3.34. Selected Western Branch hydrology model calibration/validation statistics 83
Table 3.35. Land use summary by land use scenario, as portion of watershed in wercent 86
Table 3.36. Annual streamflow and TSS load characteristics for the Western Branch of the Patuxent
scenarios 86
Draft - Do not cite or quote
-------
LIST OF FIGURES
Figure 3.1. The Raccoon River watershed and major tributaries 28
Figure 3.2. BASINS CAT option window for making multiple changes within a user specified range. ... 31
Figure 3.3. BASINS CAT pivot table screen displaying results for mean annual streamflow at reach 2 of
the Raccoon River SWAT model 32
Figure 3.4. Contour plot showing percent change in mean annual streamflow for all combinations of
temperature and precipitation changes 35
Figure 3.5. BASINS CAT modify existing data form showing adjustment to January precipitation using
the Months/Years adjustment capability 37
Figure 3.6. Mean monthly precipitation, temperature, and flow rate for the NARCCAP and baseline
scenarios 39
Figure 3.7. Mean monthly TSS, TN, and TP loadings for the NARCCAP climate models and baseline
scenarios 40
Figure 3.8. Mean monthly precipitation, temperature, and flow rate for the CC_5, CC-6, and baseline
scenarios 41
Figure 3.9. Mean monthly TSS, TN, and TP loadings for the CC-5, CC-6 and baseline scenarios 42
Figure 3.10. The commercial redevelopment site in the Upper Roanoke River watershed 47
Figure 3.12. Rainfall versus total inflow (Q) dynamics at watershed outlet for all climate scenarios 50
Figure 3.13. BASINS CAT Endpoint definition dialog, where users can specify constituents, event and
monthly (inter-annual) statistics, and thresholds for color coding the results 52
Figure 3.14. Rainfall versus total inflow (Q) dynamics at the watershed outlet for all stormwater
management scenarios and climate scenarios 54
Figure 3.15. BASINS CAT result grid of the centralized management scenario, showing the color coded
endpoint (event mean total inflow, cfs) values that are above the specified maximum
threshold in Figure 3.13 55
Figure 3.16. Mean annual sediment yield for Lasa soil at 2 percent slope under corn production for all
climate scenarios 61
Figure 3.17. Mean annual sediment yield for all land use and three climate scenarios under corn
production 62
Figure 3.18. Mean annual sediment yield for Lasa soil under corn and soy production for three climate
scenarios 62
Figure 3.19. Sediment yield (tonnes/ha/yr) under corn fall mulch till with a 3, 6, and 9 meter grass buffer.
NB=no buffer, GB3=3 meter grass buffer, GB6=6 meter grass buffer, GB9=9 meter grass
buffer, V0=no adjustment to annual precipitation volume,V20=20 percent increase in annual
precipitation volume, V20+I10=20 percent increase in annual precipitation volume plus a 10
percent increase in the intensity of the largest 5 percent of events 65
Figure 3.20. Tualatin River Watershed Location 68
Figure 3.21. BASINS CAT form specifying increased precipitation in the top 30% of events 70
Figure 3.22. Time-series plot showing three alternative precipitation pattern change methods 71
Figure 3.23. Percent change in TSS loads for all climate scenarios. Percent change calculated relative to
baseline conditions (1980-2005) 72
Figure 3.24. Location of the Sespe Creek watershed 75
Figure 3.25. BASINS CAT form defining adjustments to precipitation 77
Figure 3.26. Mean monthly streamflow during drought period for all scenarios 80
Figure 3.27. The Western Branch of the Patuxent River watershed and its location with the Chesapeake
Bay watershed 82
Figure 3.28. Measured and simulated suspended sediment concentrations on the Western Branch of the
Patuxent River watershed in mg/1 84
Figure 3.29. Climate scenario specifications in BASINS CAT 85
Draft - Do not cite or quote
-------
Figure 3.30. Simulated sensitivity of stormwater runoff volume to changes in impervious cover,
precipitation volume, and precipitation intensity 87
Figure 3.31. Simulated sensitivity of sediment load to changes in impervious cover, precipitation volume,
and precipitation intensity 88
Draft - Do not cite or quote
-------
LIST OF ABBREVIATIONS
% percent
°C Celsius
°F Fahrenheit
7Q10 7-day average low streamflow with a 10-year return period
ARS Agricultural Research Service
BASINS Better Assessment Science Integrating point & Non-point Sources
BMP best management practice
bu bushel
CA California
CAT Climate Assessment Tool
CFM change factor methodology
cfs cubic feet per second
cms cubic meters per second
CTIC Conservation Tillage Information Center
EMC event mean concentration
GCM global climate model
GCRP Global Change Research Program
GIS geographic information system
ha hectare
HSPF Hydrologic Simulation Program-FORTRAN
HUC Hydrologic Unit Code
ICLEI ICLEI - Local Governments for Sustainability
IPCC Intergovernmental Panel on Climate Change
km kilometers
LULC land use land cover
MD Maryland
mg/1 milligram/liter
MGD million gallons per day
mm millimeter
MN Minnesota
NARCCAP North American Regional Climate Change Assessment Program
NCDC National Climatic Data Center
NLCD National Land Cover Data
NPS nonpoint source
NRCS Natural Resource Conservation Service
NSE Nash-Sutcliffe Model Efficiency Coefficient
PB Percent Bias
PET potential evapotranspiration
PLOAD Pollutant Loading Application
Draft - Do not cite or quote
10
-------
PRISM
RCM
RMSE
STATSGO
SWAT
SWMM
TN
TP
TSS
UMRB
USAGE
USDA
USEPA
USGS
VA
WEPP
yr
Parameter-elevation Regressions on Independent Slopes Model
regional climate model
Root Mean Square Error
State Soil Geographic Database
Soil Water Assessment Tool
Storm Water Management Model
Total Nitrogen
Total Phosphorus
Total Suspended Solids
Upper Mississippi River Basin
U.S. Army Corps of Engineers
US Department of Agriculture
US Environmental Protection Agency
U.S. Geologic Survey
Virginia
Water Erosion Prediction Project
year
Draft - Do not cite or quote
11
-------
EXECUTIVE SUMMARY
Climate is changing. During the last century, the global average temperature increased 1.4°F (IPCC,
2007). Changes in the form, amount, and intensity of precipitation have also been observed, although with
significant regional variability (IPCC, 2007; Groisman 2005). Climate modeling experiments suggest
these trends will likely continue or accelerate throughout the next century (IPCC, 2007; Karl et al., 2009).
There is increasing concern about the potential effects of climate change on water resources. Potential
effects of climate change include increased risk of flooding and drought, changes in the quality and
seasonal timing of runoff, loss of aquatic habitat, and ecosystem impairment (Bates et al., 2008; Karl et
al., 2009; U.S. EPA, 2008a).
Many communities, states, and the federal government are considering adaptation strategies for reducing
the risk of harmful impacts resulting from climate change. Challenges remain, however, concerning how
best to incorporate diverse, uncertain, and often conflicting information about future climate change into
decision making. Despite continuing advances in our understanding of climate science and modeling, we
currently have a limited ability to predict long-term (multi-decadal) future climate at the local and
regional scales needed by decision makers. It is therefore not possible to know with certainty the future
climatic conditions to which a particular region or water system will be exposed. In addition, most water
and watershed systems are already vulnerable to existing non-climatic stressors including land-use
change, point-source discharges, and habitat loss. The potential interaction of climate change with the
effects of other existing and future stressors are not well understood at the watershed scale.
Scenario analysis using computer simulation models is a useful and common approach for assessing the
vulnerability of water and watershed systems to a wide range of plausible but uncertain future conditions
and events, including the effectiveness of alternative management responses (Lempert, 2006, 2010;
Volkery, 2009). By exploring the implications of a wide range of uncertain but plausible future
conditions, we can identify how we are most vulnerable, and guide the development of management
strategies that are robust across a wide range of potential future conditions and events (Sarewitz et al.,
2000). To reduce the likelihood of future impacts, tools and information are needed for assessing the
potential implications of potential climate change, land-use change, and management responses in
specific watershed locations.
USEPA and partners recently developed two assessment tools, the BASINS Climate Assessment Tool
(CAT) and the Water Erosion Prediction Project Climate Assessment Tool (WEPPCAT). The tools are
each intended to facilitate application of existing simulation models for conducting scenario-based
assessments. Specifically, they provide flexible capabilities for creating and running climate change
scenarios to address a wide range of "what if questions about how weather and climate could affect
water and watershed systems. Combined with the existing capabilities of the BASINS and WEPP models,
the tools can be used to explore the combined effects of potential changes in climate and land use on a
range of streamflow and water quality endpoints, as well as the potential effectiveness of management
practices for reducing impacts.
This report presents a series of short, illustrative case studies using the BASINS and WEPP climate
assessment tools. Case studies are presented using BASINS CAT with the HSPF, SWAT, and SWMM
water models, and using WEPPCAT with the WEPP model. Each case study presents a real or plausible
issue in a specified location, and applies BASINS CAT or WEPPCAT to address or inform upon the
problem. Taken together, the six case studies illustrate the use of BASINS CAT and WEPPCAT to
address a range of practical, real-world questions of potential interest to water and watershed managers.
Draft - Do not cite or quote 12
-------
Case study simulations illustrate important differences in the sensitivity of streamflow and water quality
endpoints to changes in specific climate drivers. Generally, increased precipitation resulted in increased
streamflow and pollutant loads. The response to increased precipitation was found to be reduced or even
reversed, however, by increased evapotranspiration that resulted from increased annual temperatures.
Increased temperature combined with reduced precipitation consistently resulted in decreased streamflow.
An awareness of these subtleties in the response of different streamflow and water quality endpoints to
specific types of climate change highlights the need for improved understanding of system behavior, and
in turn, the difficulty in developing quantitative predictions of future change.
Responding to climate change will require that information about climate change be incorporated into
applicable facets of community and natural resource management and decision making. Considering
climate change in the context of a broad agenda allows communities to determine how climate change
risks rates against other activities and factors in the community and may also help to identify ways to
adapt for climate change using existing methods.
The scientific approach supported by these tools, i.e., scenario analysis, can be useful for understanding
system behavior, identifying vulnerabilities, and evaluating the effectiveness of management responses to
inform management decision making. The tools presented in this report, however, are just one step
forward in building our capacity for understanding and responding to climate change. Application of
hydrologic models in this way has limitations, many of which are not well understood (Ghosh 2010;
Ludwig, 2009; Najafi, 2011; Vaze et al., 2010). Further study is required to better assess, refine, and
develop our current modeling capabilities. Further study is also required to better address the challenge of
incorporating diverse, uncertain, and often conflicting information about future climate change into water
resources decision making.
Draft - Do not cite or quote 13
-------
1. INTRODUCTION
There is growing concern about the potential effects of climate change on water resources.
Climate has a direct influence on the occurrence, distribution, and management of water resources.
Changes in the amount, form, and intensity of precipitation, together with factors affecting evaporative
loss such as air temperature have a direct influence on the quantity, quality, and timing of available water
(Gleick and Adams, 2000). Water infrastructure is designed and operated to maintain safe and reliable
drinking water supplies, flood protection, wastewater treatment, and urban drainage under anticipated
climatic conditions. Climate change presents an increased risk of harmful impacts to these and other water
management goals.
It is now generally accepted that human activities including the combustion of fossil fuels and land-use
change have resulted, and will continue to result in, long-term climatic change (IPCC, 2007; Karl et al.,
2009). The 2007 Fourth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC)
states that "warming of the climate system is unequivocal, as is now evident from observations of
increases in global average air and ocean temperatures, widespread melting of snow and ice and rising
global average sea level" (IPCC, 2007). During the last century, the global average temperature has
increased 1.4°F (IPCC, 2007). Changes in precipitation patterns have also been observed, although with
significant regional variability (IPCC, 2007; Groisman 2005). Climate modeling experiments suggest
these trends will likely continue or accelerate throughout the next century (IPCC, 2007; Karl et al., 2009).
The effects of climate change will vary in different locations depending on the specific type of change
that occurs together with the attributes individual watersheds including physiographic setting, land-use,
and human use and management of water. Effects will vary in different regions of the nation but could
include increases or decreases in available water supply, changes in the seasonal timing of supply,
increased risk of flooding and drought, increased water temperature, changes in pollutant loading, loss of
aquatic habitat, and ecosystem impairment (Bates et al., 2008; Karl et al., 2009; U.S. EPA, 2008a). In
addition, in many areas water resources are stressed and vulnerable to existing, non-climatic stressors
including increasing demand, land-use change, point-source discharges, and habitat degradation. Climate
change will interact with these stressors in different settings in complex ways that are not well
understood. Where effects are large, current water management may not be adequate to cope with the
effects of climate change.
Responding to climate change is complicated by the scale, complexity, and inherent uncertainty of the
problem. Despite continuing advances in our understanding of climate science and modeling, current
climate models have a limited ability to predict long-term (multi-decadal) future climate at local and
regional scales (Sarewitz et al., 2000). It is therefore not possible to know with certainty the future
climatic conditions to which a particular location or water system will be exposed. This uncertainty
should not, however, be considered a barrier to taking action. Current global and regional climate models
(GCMs, RCMs) are excellent tools for understanding the complex interactions and feedbacks associated
with future emissions scenarios and identifying a set of plausible, internally consistent scenarios of future
climatic conditions. Historical observations and paleo records of climatic variability can also provide
useful information about the type and range of changes possible in different regions of the nation. By
exploring the implications of a wide range of uncertain but plausible future conditions, we can identify
how we are most vulnerable, and use this information to guide the development of robust strategies for
reducing risk (Sarewitz et al., 2000).
Vulnerability (to climate change) is defined by the IPCC (2007) as "the degree to which a system is
susceptible to, or unable to cope with, adverse effects of climate change, including climate variability and
extremes". Assessing the risks and impacts of climate change (vulnerability assessment) can take many
Draft - Do not cite or quote 14
-------
forms depending on the ecological or social resource of interest, decision context, and projected the
potential range of expected climate changes characteristics.
Scenario analysis using computer simulation models is a common approach (Lempert, 2006, 2010;
Volkery, 2009). A number of water simulation models are available that are capable of representing the
response of watershed systems to changes in climatic, land-use, and management drivers. Many of these
models, such as those currently available in EPA's BASINS modeling system (HSPF, SWAT, SWMM),
are well validated and already commonly used to support management decision making. Several excellent
references are available discussing the use of scenarios in assessing climate change impacts and
vulnerabilities to water (e.g., see IPCC TGICA, 2007; WUCA, 2010; Brekke et al, 2009; U.S. EPA,
201 la).
In an effort to support scenario-based assessments of climate change impacts on water, USEPA and
partners have developed two assessment tools, the BASINS Climate Assessment Tool (BASINS CAT)
and the Water Erosion Prediction Project Climate Assessment Tool (WEPPCAT). These tools facilitate
application of existing simulation models for conducting scenario-based assessments. Specifically, they
provide flexible capabilities for creating and running climate change scenarios to address a wide range of
"what if questions about how weather and climate could affect water and watershed systems (e.g., how
would increases in the intensity of rainfall events affect stormwater runoff, what type of climate change
would need to occur to increase stream water temperatures to a level harmful to fish?). Combined with the
existing capabilities of the BASINS and WEPP models, the tools can be used to explore the combined
effects of potential changes in climate and land use on a range of streamflow and water quality endpoints,
as well as the potential effectiveness of management practices for reducing impacts.
BASINS CAT was originally released in 2007 with EPA's BASINS modeling system version 4.0, and
was originally available only with the Hydrologic Simulation Program-FORTRAN (HSPF) watershed
model (Johnson and Kittle, 2006; Imhoff et al., 2007; U.S. EPA, 2009a). With the release of BASINS
CAT Version 2 in 2011, CAT capabilities will also be available with the Soil and Water Assessment Tool
(SWAT) and Stormwater Management Model (SWMM).
WEPPCAT was released in 2010 in partnership with the USDA Agricultural Research Service
(http://typhoon.tucson.ars.ag.gov/weppcat/index.php). WEPPCAT provides an online platform for
creating and running climate change scenarios to assess potential implications for soil erosion from
agricultural lands using the USDA ARS Water Erosion Prediction Pilot (WEPP) model.
About This Report
This report presents a series of short case studies using the BASINS and WEPP climate assessment tools.
The case studies are designed to address three general objectives. First, the case studies illustrate
conceptually how scenarios based on different types of climate, land use, and management information
can be used to address different questions about the potential implications of climate change on
watersheds. Climate change scenarios are created based on model projections as well as historical data
and past events. Land use change and management scenarios are also included to address questions
related to the relative effects of land use versus climate change, and the effectiveness of management
practices for reducing impacts. Second, the case studies illustrate selected capabilities of the tools when
used with different models.
Finally, while the primary intent of case studies is illustrative, the results are based on real simulations in
each study location. Results thus convey information about how watersheds in different parts of the nation
could respond to future changes in climate, land-use, and management practices. It should be noted that
due to the significant effort involved in developing new models, all simulations in this study used pre-
existing models. Additionally, while all models are calibrated, efforts to validate the models were limited
Draft - Do not cite or quote 15
-------
and models may not represent all local management and other factors in full detail. Results should thus be
considered qualitative and heuristic rather than absolute.
While the climate change scenarios evaluated in this case study were relatively simple, they provide a
screening-level understanding of stormwater runoff sensitivity to climate change, and the potential
effectiveness of stormwater management strategies for reducing climate change impacts. Evaluation of
more detailed climate change or management scenarios is also possible. The coupling of BASINS CAT
and hydrologic/hydraulic models, when calibrated, can facilitate development of rapid assessment
methods that provide timely and usable quantitative information. The flexible capabilities of BASINS
CAT for creating and running scenarios can aide and facilitate a wide range of analyses. These
capabilities can be an important addition to the tools used by stormwater managers to design, manage, and
maintain stormwater infrastructure.
The intended audience of this report is watershed or water utility managers, urban and regional planners,
agency officials, researchers, and other water professionals interested in conducting modeling studies of
the potential effects of climate change on water and watershed systems, including the coupled effects of
climate change, land-use change, and the effectiveness of management responses. The report may be of
particular interest to current users of BASINS or WEPP that want to extend the scope of their modeling to
include the potential effects of climate change. The intent of the information presented in this report is to
stimulate further creativity and exploration of the different ways scenario analysis can be used to support
management decision making.
Draft - Do not cite or quote 16
-------
2. BASINS AND WEPP CLIMATE ASSESSMENT TOOLS
This chapter provides a brief introduction to the BASINS and WEPP climate assessment tools. More
detailed documentation of BASINS CAT is available in the document "BASINS 4.0 Climate Assessment
Tool (CAT): Supporting Documentation and User's Manual" (USEPA, 2009a), and on the BASINS web
page (http://water.epa.gov/scitech/datait/models/basins/bsnsdocs.cfm). More detailed documentation of
WEPPCAT is available on the WEPPCAT web page
(http://typhoon.tucson.ars.ag.gov/weppcat/index.php).
BASINS CAT and WEPPCAT are each intended to facilitate application of existing simulation models
for conducting scenario-based assessments. The conceptual basis of these tools is simple; to provide
flexible capabilities for creating and running scenarios to address a wide range of what-if questions users
may have about the potential effects of climate change on water and watershed systems. It is important to
note that BASINS CAT and WEPPCAT do not provide climate change data for any particular region of
the United States. Rather, the tools simply provide a capability for users to create meteorological data
reflecting any type of change they wish to consider. In each case, climate change scenarios are created
using the change factor or delta-change approach, whereby historical meteorological data within a
selected baseline period (e.g., daily temperature, daily precipitation) are adjusted to create scenarios for
input to water models. These capabilities support a range of assessment goals, e.g., simple screening
analysis, systematic sensitivity analysis, or assessing more detailed scenarios based on climate model
projections.
Introduction to BASINS CAT
EPA's BASINS modeling system integrates environmental data, analytical tools, and watershed modeling
programs to support assessments of watershed land use change, pollutant discharges, and management
practices on water quality (U.S. EPA, 2001; U.S. EPA, 2007; http://www.epa.gov/waterscience/basins/).
BASINS consists of four components: (1) a comprehensive collection of national cartographic and
environmental databases, (2) environmental assessment tools and utilities (summarize results; establish
pollutant source/impact interrelationships; selectively retrieve data; import tool, download tool, grid
projector, post processor, and land use, soil classification and overlay tool); (3) automated watershed
characterization reports (for eight different data types); and (4) a suite of watershed models including
HSPF (Bicknell et al., 2005), SWAT (Neitsch et al., 2005), AQUATOX (Clough and Park, 2006),
SWMM (Rossman, 2010) and PLOAD (U.S. EPA, 2007). The main interface to BASINS is provided
through Map Window, anon-proprietary, open-source Geographic Information System (GIS). The GIS
provides a framework for linking BASINS modeling tools with environmental data (Figure 2.1).
Draft - Do not cite or quote 17
-------
BASINS
CIS
Web Data
Download
Tool
Pol I tiC.Hi
Boundaries
TIGER Line
and Census
Data
Monitoring
Digital
Elevation
Meteorological
Data (Weather
Stations)
BASINS 4.0 System Overview
Tools and Utilities
Watershed Report*
-
Watershed Delineation
(Automatic or Manual}
Parameter Estimation
Additional
User Supplied
Data
-
Watershed Modeling
HSPF/WlnHSPF
« *•» •-
Decision Making and
Analysis
Postprocessing
SenScn
-
Watershed Management
Sensitivity Analysts
Climate Analyst Tool
~:
Nutrient Management
Source Water Protection
TMDLs
UAAs
Figure 2.1. Overview of the EPA BASINS 4 Modeling System.
BASINS CAT is not a stand-alone modeling application. CAT is a BASINS plug-in available for use with
pre-existing, calibrated BASINS models. Development was intended to facilitate application of existing
BASINS models to assess the implications of climate variability and change. Given a pre-existing,
calibrated model within BASINS, BASINS CAT provides three capabilities (U.S. EPA, 2009a).
• a flexible scenario generation capability for creating meteorological time series using the change
factor approach reflecting any user determined change in temperature and precipitation for use as
input to the selected BASINS model (Table 2.1);
• managing the new climate data for input into BASINS models; and
• a post processing capability for calculating management targets (endpoints) useful to water and
watershed managers from model output (Table 2.1).
Table 2.1. Summary of BASINS CAT options for adjusting meteorological time series to create climate
change scenarios and assess endpoint values based on model simulation outputs.
Modifying
historical
precipitation
records
• Apply a multiplier to each value within selected months in a multi-year record
• Apply multiplier to each value within selected years in a multi-year record
• Represent storm intensification by applying multiplier to values (events) only
within a selected size class
• Represent changes in event frequency by adding or removing storm events to
observed historical precipitation time series
Draft - Do not cite or quote
18
-------
Modifying
historical air
temperatures
Add or subtract from each value within selected months in a multi-year record
Add or subtract from each value within selected years in a multi-year record
Creating
complex scenarios
• Combine multiple adjustments to precipitation and temperature time series to
create complex scenarios
• Create spatially variable climate change scenarios for multiple locations
• Create synthetic climate change scenarios within specified ranges
• Export BASINS CAT climate change scenarios as text (ASCII) files
Calculating
assessment
endpoints
• Calculate mean, max, min, sum and other summary values from model output
time series
• Calculate summary values for specified range of concern in model output time
series (e.g., selected months, years)
• Calculate duration-frequency events based on model output time series (e.g.,
100-year flood, 2-year flood)
Climate change scenarios (i.e., the adjusted meteorological time series created using the tool) are
contained within the same BASINS Watershed Data Management (WDM) file with the original, historical
meteorological records. In addition, BASINS CAT provides a view/export capability that can be used to
display the changes resulting from a specific adjustment or save the adjusted weather record as an ASCII
file.
Climate assessment capabilities are accessible for 3 BASINS models (BASINS CAT Version 2, released
September, 2011): the Hydrologic Simulation Program-FORTRAN (HSPF), the Soil and Water
Assessment Tool (SWAT), and Stormwater Management Model (SWMM). These 3 models provide
general capabilities for application to a wide range of issues in water management. Models differ in the
approaches used to represent key processes, input requirements, and endpoints simulated. It is the user's
choice to determine which model is most appropriate for a given assessment. The following is brief
summary of the 3 BASINS models accessible to BASINS CAT.
HSPF
HSPF is a process-based, basin-scale model that provides a comprehensive package for simulating
watershed hydrology and water quality for a wide range of conventional and toxic organic pollutants
(Shoemaker et al., (2005). The model simulates watershed hydrology, land and soil contaminant runoff,
sediment-chemical interactions, and in-stream fate and transport in one-dimensional stream channels. It
can be configured to represent all types of land uses, and offers the ability to include land use activities
and potential management controls. Since its inception in 1980, HSPF has been widely applied in the
planning, design, and operation of water resources systems, and is arguably one of the best verified
watershed models currently available. HSPF can be applied to most watersheds using available
meteorological, land use, hydrography, management, streamflow, and water quality data. The principal
model outputs include streamflow runoff and mass loads or concentrations of sediment, nutrients,
pesticides, and toxic chemicals at selected points within a watershed. The most recent release is HSPF
Version 12, which is distributed as part of the EPA BASINS (Better Assessment Science Integrating
Point and Nonpoint Sources) system.
HSPF represents a watershed as a group of various land uses all routed to a representative stream
segment. The modeling framework is defined in terms of subwatershed segments, one-dimensional stream
reach segments and well-mixed reservoirs/lakes. The spatial scale for simulation uses one-dimensional,
lumped parameters on a land-use or subwatershed basis. For overland flow, the model assumes one-
Draft - Do not cite or quote 19
-------
directional kinematic-wave flow. The receiving water bodies assume complete mixing along the width
and depth.
Processes simulations for pervious and impervious land areas include water budget, sediment generation
and transport, and the generation and transport of other water quality constituents. Hydrologic simulations
include consideration of interception, infiltration, evapotranspiration, interflow, groundwater loss, and
overland flow processes. Sediment production is based on detachment and/or scour from a soil matrix and
transport by overland flow in pervious areas, whereas solids buildup and wash-off is simulated for
impervious areas. HSPF also simulates the in-stream fate and transport of a wide variety of pollutants,
such as nutrients, sediments, tracers, dissolved oxygen/biochemical oxygen demand, temperature,
bacteria, and user-defined constituents.
Some key HSPF model strengths are as follows (Shoemaker et al., (2005)):
• HSPF can be set up as simple or complex, depending on application, requirements, and data
availability.
• HSPF is one the few watershed models capable of simulating land processes and receiving water
processes simultaneously.
• A variety of simulation time steps can be used, including sub-hourly to 1 minute, hourly or daily.
• The model includes capabilities to address a variable water table.
• The model enables user-defined model output options by defining the external targets block.
SWAT
SWAT is a widely applied, physically-based, watershed-scale model designed to assess long-term
changes to water quality and quantity as a result of resource management and land use changes (Neitsch
et al., 2005). SWAT uses a curve number approach for hydrologic simulation. It does not simulate event-
based changes. Utilizing weather and soils data, and information on vegetation, topography and land use,
SWAT can model the physical process associated with hydrology and sediment and nutrient transport
among other things. This enables the model to be used in large watersheds as well as small ungaged
streams. SWAT can also be modified for more specialized modeling.
SWAT is the result of more than 30 years of modeling investigations and research efforts conducted
primarily by the USDA Agricultural Research Service (ARS) and the Texas A & M University Blackland
Research and Extension Center (BREC) in College Station, Texas. SWAT is a public domain, basin-
scale, continuous simulation model that operates on a daily time step and is designed to predict the
nonpoint source loadings and resulting water quality impacts of water, sediment, and agricultural
chemicals (nutrients and pesticides) from a watershed. In addition, the model includes capabilities and
functionality to assess a wide variety of impacts of alternative management practices and land use
changes. The model is physically based, computationally efficient, and capable of continuous simulations
over long periods of time, ranging from days to decades. Major model components include weather,
hydrology, erosion/sedimentation, soil temperature, plant growth, nutrients, pesticides, bacteria,
agricultural management, stream routing and pond/reservoir routing (Gassman et al., 2007). The
simulation of these components is carried out within SWAT's basic building block, the Hydrologic
Response Unit (HRU). HRUs represent unique combinations of land use, soil characteristics, and
management within each sub-basin being modeled.
The SWAT model has comprehensive representation of all major watershed processes. It has a
particularly strong representation of agricultural land use. Hence, it is usually selected for assessing
nutrient loads from agricultural dominant watersheds. The model uses GIS technology, topography, soils,
precipitation, plant growth, and crop management information to form a complete deterministic
representation of the hydrology and water quality of a watershed.
Draft - Do not cite or quote 20
-------
SWAT has gained both national and international acceptance as an efficient and reliable watershed
modeling tool as demonstrated by hundreds of SWAT-related papers in the open technical literature,
presentations at international SWAT conferences, and its inclusion in EPA's BASINS modeling system.
Additional information regarding the development and use of SWAT can be found at:
http: //www .brc .tamus. edu/swat.
Some key SWAT model strengths are as follows:
• The model is physically based. Watersheds can be modeled to evaluate the relative impact of
changes in management practices, climate, and vegetation on water quality or other variables of
interest
• The model uses readily available inputs. The minimum data required to make a run are commonly
available from various government agencies.
• The model's ability to simulate crop and plant communities and provide crop yield and plant
biomass.
• The mathematical solutions within the model are computationally efficient. Simulation of very
large basins or a variety of management strategies can be performed expeditiously without
excessive investment of time or money.
SWMM
The EPA Storm Water Management Model (SWMM), first developed in 1971, is a rainfall-runoff
simulation model that can be used to simulate runoff quantity and quality from primarily urban areas on a
single event or long-term (continuous) basis (see http://www.epa.gov/ednnrmrl/models/swmm/). SWMM
is commonly used to inform decisions related to stormwater management, combined sewer overflows,
assessing nonpoint source pollution loads, and low impact development techniques. Typical SWMM
applications include the design and sizing of drainage system components for flood control, flood plain
mapping of natural channel systems, evaluating the effectiveness of BMPs for reducing wet weather
pollutant loadings, generating non-point source pollutant loadings for waste load allocation studies, and
designing control strategies for minimizing combined sewer overflows and sanitary sewer overflows
(Rossman, 2010).
SWMM operates on time steps ranging from seconds to years. SWMM accounts for spatial variability by
dividing a study area into a collection of smaller, homogeneous subcatchment areas, each containing its
own fraction of pervious and impervious sub-areas. Overland flow can be routed between sub-areas,
between subcatchments, or between entry points of a drainage system through a system of pipes,
channels, storage/treatment devices, pumps, and regulators. SWMM simulates the quantity and quality of
runoff generated within each subcatchment, and the flow rate, flow depth, and quality of water in each
pipe and channel during a simulation period comprised of multiple time steps.
Some key SWMM strengths are as follows:
• SWMM model accounts for all hydrologic processes that produce runoff from urban areas.
• Accounts for interruption in natural stream transport network such as nonlinear reservoir routing
of overland flow.
• SWMM contains a flexible set of hydraulic modeling capabilities dealing with industry standard
stormwater structures such as stormwater storage, divider, pumps, weirs and orifices etc.
• Can simulate different flow regimes such as such as backwater, surcharging, reverse flow, and
surface ponding.
• In addition to modeling runoff, can account for the production, transport, and treatment of
pollutant loads associated with runoff.
Draft - Do not cite or quote 21
-------
Introduction to WEPPCAT
The Water Erosion Prediction Project (WEPP) is a process-based soil erosion model developed in the
mid-1990s by the USDA Agricultural Research Service (ARS). It is currently one of the best known and
validated models available for simulating soil erosion from agricultural areas. WEPP can be used to
assess how erosion rates are impacted by precipitation events, soil type, vegetation type, topography and
number of common best management practices (BMPs) for reducing soil loss. Simulations can be run at
the hill slope or watershed scale (Flanagan and Nearing, 1995). Hill slope scale simulations are ideally
suited for assessing the effectiveness of BMPs in local settings such as a specific farm field. Watershed
scale applications consist of linking hill slopes via channels and impoundments (Flanagan and Nearing,
1995). Recent developments allow forested land cover, such as forested riparian buffers, to be represented
in WEPP. WEPP is available as a desktop model or through a web-based interface.
WEPPCAT is an online application of the WEPP model that provides flexible capabilities for creating
climate change scenarios to assess the potential effects of climate change on soil erosion using the WEPP
model. WEPPCAT was developed in partnership with the USDA ARS, and is available for use at
http://typhoon.tucson.ars.ag.gov/weppcat/index.php (Figure 2.2).
Water Erosion Prediction Project Climate Assessment Tool
WEPPCAT | Databases | Help j Tutorial | Acknowledgments 1 Scientific References | Link
-cAtr"
*yy4
^. Baseline Conditions
Assess Change
r
,-CI
name:
Modify Climate
rFie
Man
AlfE
| Viev
r F"
agement:
Ifa (w cuttings)
Manaaemem
^ .
Use Filter Strip
|-CO
Run Assess Change Sc
mpare Assess Change Scenaric
| Compare |
3
enario
Welcome to the WEPPCAT Simulation
Global warming is expected to lead to a more vigorous hydrological cycle, including more total
rainfall and more frequent high intensity rainfall events. Rainfall amounts and intensities increased
on average in die United States during the 20th century and: according to climate change
models, they are expected to continue to increase during the 21st century. These rainfall
changes, along with expected changes in temperature: solar radiation, and atmospheric CO2
concentrations, will have significant impacts on soil erosion rates. The processes involved in the
impact of climate change on soil erosion by water are complex, involving changes in rainfall
amounts and intensities; number of days of precipitation, ratio of rain to snow, plant biomass
production, plant residue decomposition rates, soil microbial activity, evapo-transpiration rates,
and shifts in land use necessary to accommodate a new climatic regime. WEPPCAT is a web-
based erosion simulation tool that allows for the assessment of changes in erosion rates as a
consequence of user-defined climate change scenarios. This tool is based on the USDA-ARS
Water Erosion Prediction Project (WEPP) erosion model. It has the capability of taking into
account all of the erosion-affecting processes listed above.
To work vdth WEPPCAT, run the Baseline Conditions first by selecting the inputs on the left
and pressing the "Run Baseline Conditions'1 button, \\~hen the simulation is complete this area
wiH contain a brief summary of the results- To view more detailed outputs from the model as
Figure 2.2. WEPPCAT opening screen
WEPPCAT simulations are limited to the hill slope scale only. The WEPPCAT online interface allows
users to input field characteristics including soil series, field size, slope steepness, slope shape, and land
management (e.g., alfalfa with cutting, bluegrass with grazing, etc.). Like the parent WEPP model, daily
meteorological data necessary to run WEPPCAT are generated using the stochastic weather generator
Cligen. WEPPCAT outputs include mean annual precipitation, runoff, soil loss and sediment yield. Users
can also generate spatial sediment loss data and a sediment particle size profile.
Baseline meteorological data for WEPPCAT simulations are generated using Cligen parameters based on
observed monthly average temperature and precipitation from NOAA National Climatic Data Center
(NCDC) weather stations. Climate change scenarios are created by adjusting Cligen parameters to reflect
potential changes of interest to users. Available adjustments include increases and decreases in mean
monthly temperature, precipitation volume, and the transition probabilities of a wet day following a dry
Draft - Do not cite or quote
22
-------
day, and a dry day following a dry day (i.e., number of wet days). These adjustments can be made either
uniformly among months of the year, or individual adjustments can be made to specific months of the
year. In addition to changing precipitation volume, Cligen parameters can also be adjusted to increase the
proportion of annual rainfall occurring in large magnitude events (i.e., to represent an increase in event
intensity independent of changes in total annual precipitation). WEPPCAT provides a capability to
increase the proportion of annual precipitation occurring in large magnitude events up to 25 percent1.
Adjustments in precipitation intensity are made by applying the user determined increase to the largest 5
percent of events, and simultaneously decreasing precipitation in the lower 95 percent events by the same
volume such that the adjustment results in negligible change in the volume of annual precipitation. This
adjustment can only be made to all events across the entire year. It is currently not possible to adjust the
intensity of events only in specific months of the year. Precipitation data can also be modified in
WEPPCAT based on elevation using the PRISM model climate database. Modifications are made by
selecting precipitation values or elevations for areas surrounding the selected weather station.
Adjustment of rainfall intensity is accomplished by altering the standard deviation of the distributions of daily precipitation
used by the climate generator. This approach results in a slight change in average annual rainfall even if changes to the overall
volume are not indicated in the model inputs.
Draft - Do not cite or quote 23
-------
3. CASE STUDIES
3.1. Introduction
This Chapter presents a series of six case studies that are designed to illustrate selected capabilities and
approaches for conducting scenario-based assessments using the BASINS CAT and WEPPCAT tools
(Table 3.1). Case studies are an effective way to demonstrate the utility of these tools to water managers
and others interested in assessing climate change impacts on their systems. The case studies in this report
encompass a range of spatial and temporal scales, climate and land-use change scenarios, and hydrologic
and water quality endpoints of concern. They were designed to include applications of either BASINS
CAT or WEPPCAT to assess the implications of future climate change in the context of changing land
use and management responses.
Case studies vary in the way different information about climate and land use change is used to develop
scenarios for exploring system sensitivity, vulnerability, and the effectiveness of management response.
Climate change scenarios can be developed based on any available information about climate change. The
IPCC Task Group on Data and Scenario Support for Impacts and Climate Analysis (TGICA) describes
three different types of climate scenarios: synthetic scenarios, analogue scenarios, and scenarios based on
outputs from climate models (see IPCC TGICA (2007) for a more complete discussion). Synthetic
scenarios are created by incrementally modifying climatic attributes within a predetermined, plausible
range of future change. For example, adjustments of historical temperatures by 1, 2, and 3°C and
historical precipitation by 5, 10, and 15 percent could be applied in various combinations to create 9
different climate change scenarios (IPCC TGICA, 2007). Analogue scenarios are constructed by
identifying a time or geographic location that has a climate similar to anticipated future conditions in the
location of interest. These records can be obtained either from the past (temporal analogues) or from a
different geographic location (spatial analogues). Model-based scenarios are developed using output from
GCM and RCM modeling experiments that simulate the response of the climate system to changes in
greenhouse gas emissions and other climate forcings. The case studies in this report illustrate applications
of either BASINS CAT or WEPPCAT using each type of scenario.
Many watersheds are currently stressed by a wide range of non-climatic factors including land-use
change, water withdrawals, and other factors. Water infrastructure and management also exerts a major
control on observed hydrologic and water quality conditions. Climate change will interact with existing
and future changes in non-climatic factors in complex ways in different locations. Understanding and
responding to climate change requires consideration of climate change in a holistic context.
A critical concern is the interaction of climate and land-use change on water. Land-use change can be
considered in a scenario analysis in much the same way as climate change. Land use scenarios can be
based on a range of context dependant information. Future land use and land cover conditions will be
influenced by population growth, land use regulations, and economic factors, among other things.
Understanding the potential effectiveness of alternative management strategies is likewise a critical
concern. In many areas existing infrastructure and management may be well capable of handling
anticipated future hydrologic change. In other areas, a greater risk may be present requiring some further
action. For example, assessing the impact of climate change on agriculture may require assessing various
types of cropping practices or inclusion of best management practices as scenarios. Stormwater runoff
assessments may require developing scenarios that depict community build-out conditions under current
zoning or projected population growth (EPA 2009b).
Ultimately, the scenarios used in an analysis should depend on the available information and, equally
important, the goals and requirements of a specific assessment activity. In each case, consideration of
multiple scenarios is desirable to capture the full range of underlying uncertainties associated with future
climate, land use, and management practices on water resources.
Draft - Do not cite or quote 24
-------
It should be noted that the primary intent of the case studies in this report is to illustrate the use of
BASINS CAT and WEPPCAT in a variety of applications. All case studies using HSPF, SWAT, or
SWMM were conducted using pre-existing, calibrated models. In certain cases minor modifications such
as performing additional calibration were made. In each case, however, models may not represent in full
detail all management practices and other factors influencing the hydrologic behavior of case study
watersheds. WEPP simulations did not require a pre-existing model, and were developed independently.
In addition, analysis and discussion of simulation results are brief and not comprehensive. Results should
therefore not be considered absolute.
Table 3.1. Summary of case studies presented in this report.
Section
Topic
Analysis Approach
3.2
Streamflow and water quality
sensitivity to climate change
in the Raccoon River, Iowa,
using BASINS CAT with
SWAT
PART A: Assess scenarios based on different
combinations of assumed temperature and precipitation
change within plausible ranges of future change; changes
uniform for each month of the year
PART B: Assess scenarios based on downscaled climate
model projections (NARCCAP) for temperature and
precipitation for mid-21st century; changes vary among
months of the year
3.3
Urban stormwater sensitivity
to rainfall change and
effectiveness of management
in the Upper Roanoke River,
VA, using BASINS CAT with
SWMM
PART A: Assess scenarios based on different assumed
changes in precipitation (single event) within a plausible
range of future change
PART B: Assess performance of 2 stormwater
management strategies under precipitation change
scenarios developed in PART A
3.4
Agricultural soil erosion
sensitivity to climate change
and management practices in
Blue Earth County, MN, using
WEPPCAT
PART A: Assess scenarios based on different
combinations of assumed changes in temperature,
precipitation volume, and precipitation event intensity;
changes uniform for each month of the year
PART B: Assess performance of land management
practices for reducing sediment loss from corn fields
under climate change scenarios developed in PART A.
3.5
Streamflow and water quality
sensitivity to changes in
precipitation amount,
frequency, and intensity in the
Tualatin River, OR, using
BASINS CAT with HSPF
Assess scenarios based on different combinations of
assumed increases in precipitation annual volume,
precipitation event intensity (proportion of annual total in
occurring in large magnitude events), and precipitation
event frequency (number of precipitation events per year)
Draft - Do not cite or quote
25
-------
3.6
Streamflow sensitivity to dry
weather events in Sespe
Creek, CA, using BASINS
CAT with HSPF
Assess scenarios based on targeted adjustments to a
historical period of dry weather; scenarios represent
increased severity of historical dry period, increased
duration of historical dry period, and increased severity
and duration of historical dry period
3.7
Streamflow and water quality
relative sensitivity to climate
change versus impervious
ground cover in the Western
Branch of the Patuxent River,
MD, using BASINS CAT with
HSPF
Assess scenarios based on different combinations of
assumed increases in precipitation annual volume,
precipitation event intensity (proportion of annual total in
occurring in large magnitude events), and impervious
ground cover
Draft - Do not cite or quote
26
-------
3.2. Streamflow and water quality sensitivity to climate change in the Raccoon River, Iowa,
using BASINS CAT with SWAT
Case Study Overview
This case study illustrates two different assessment approaches, a general sensitivity
analysis (PART A) using synthetic climate change scenarios and a more detailed scenario
analysis (PART B) using climate model projections to assess potential climate change
impacts on streamflow and total nitrogen (TN), total phosphorus (TP), and total suspended
solids (TSS) loads from agricultural lands using BASINS CAT with the SWAT watershed
model.
In PART A, the climate change scenarios were created by increasing historical mean
annual temperatures +0 to +5°C by increments of 1°C and adjusting mean annual
precipitation volume -10 to +20 percent by increments of five percent. Land use remained
constant and no management practices were assessed.
In PART B, climate change scenarios were created by applying adjustments to average
monthly temperature and precipitation volumes based on projections from four regionally
downscaled climate models. Two additional scenarios were created by synthetically
adjusting average monthly precipitation volumes for one of the climate model projections
to further explore the seasonal impacts on pollutant loading and streamflow. Land use
remained constant and no management practices were assessed.
Introduction
Nutrient pollution is an ongoing water quality issue in the Mississippi River basin, leading to such
problems as extensive algae growth and hypoxia in the Gulf of Mexico (Rabalais, 2001). The Upper
Mississippi River Basin (UMRB), a major agricultural region in the U.S., is a significant net exporter of
nutrients to the Mississippi River and Gulf of Mexico. Future climate change is projected to result in
warming temperatures and changes in precipitation regimes. Specific regional climate changes are
uncertain, but it is possible that changes in temperature and precipitation could influence pollutant loading
in the Mississippi River Basin (Rossi et al., 2009). Managers and decision makers interested in
quantifying future nutrient loads from the UMRB will likely need to consider the potential impacts of
climate change in addition to other factors that impact water quality (e.g. land use, public policy, pollution
abatement technology, etc.).
A watershed sensitivity study can help establish a general understanding of how climate changes may
interact with the landscape and alter hydrologic processes and water quality. In this case study, a SWAT
model of the Raccoon River in IA, a sub-basin within the greater UMRB, was used to simulate potential
watershed response to projected climate change. BASINS CAT was used to create an array of climate
change scenarios for model simulations and assess endpoints. The sensitivity of the Raccoon River was
explored in two different ways:
In PART A of this case study, adjustments to precipitation and temperature were applied uniformly to the
entire duration of the simulation using the BASINS CAT multiple changes within a user specified range
feature to assess potential changes to mean annual streamflow and pollutant loadings (TN, TP, and TSS).
Draft - Do not cite or quote 27
-------
However, changes to climate seldom follow a definitive and uniform pattern across the year, especially
for precipitation; therefore this case study was taken a step further in PART B.
PART B of this case explores how seasonal changes to precipitation volume and temperature can impact
monthly and mean annual streamflow and TN, TP, and TSS loads. Spatial variability was also accounted
for by applying distinct adjustments to temperature and precipitation data from the two meteorological
stations included in the model. The BASINS CAT months/years adjustment feature was used to develop
the climate change scenarios based on simulations for 4 RCMs, and the tool was used to assess sensitivity
of endpoints.
Location Description
The Raccoon River watershed encompasses an area of roughly 9,400 km2 in central Iowa (Figure 3.1). It
is comprised of two 8-digit Hydrologic Unit Codes (HUCs). The northern portion (HUC 07100006)
contains the North Raccoon branch and the main Raccoon River. The southern portion (HUC 07100007)
contains the Middle and South branches which flow into the main Raccoon River at the HUC outlet. The
Raccoon River watershed drains into the Des Moines River at the city of Des Moines, IA. Land use in the
watershed is predominantly agricultural, with minimal urban development and forests (Table 3.2).
7
N
A
| | Raccoon Rive
Figure 3.1. The Raccoon River watershed and major tributaries.
Draft - Do not cite or quote
28
-------
Table 3.2. Land use summary for Raccoon River watershed.
Land Use Portion of Watershed (%)
Corn 42
Soybeans 33
Other Ag 13
Urban/Developed 8
Forest 2
Wetland 2
Water Model Setup
A SWAT model of the Raccoon River was identified from a previous effort investigating the impacts of
ethanol corn production in the UMRB (USEPA, 2008a). A follow-on effort was performed to isolate the
Raccoon River basin and improve model calibration (USEPA, 2010). The Raccoon River SWAT model
uses the 2001 National Land Cover Data (NLCD) and 2004-2006 Cropland Data Layer for the land use
coverage and the USDA-NRCS STATSGO for soils data. Data from the Conservation Tillage
Information Center and the 1997 and 2002 USDA Census of Agriculture were used to identify the
cropping rotation and management practices for the agricultural land areas. Each sub-watershed was
assigned appropriate management and tillage practices. The model was set up to run using 1960-2001
weather data developed by Di Luzio (2008). These weather data, developed for modeling and
assessments, are gridded datasets of daily precipitation and temperature (maximum and minimum)
spatially interpolated using slope, elevation and aspect as spatial covariates. Grid cells are four km2 (two
km on each side) and cover the conterminous U.S.
Initial SWAT parameters for the Raccoon River model were acquired from a national database developed
as part of a previous UMRB SWAT model (USEAP, 2010). More detailed calibration of the Raccoon
River SWAT model was carried out using available streamflow, TN, TP and TSS data at the watershed
outlet at Van Meter, IA. The entire 42-year simulation duration, 1960-2001, was used to conduct the
calibration. Streamflow was reasonably well-calibrated while nutrient and sediment loadings showed
mixed calibration statistical values (Table 3.3). The model was limited by exclusion of point sources
which influences streamflow and pollutant loads. Quantitative results should therefore not be considered
absolute, but rather as indicative of the relative changes resulting from the scenarios considered in this
case study.
Table 3.3. Raccoon River model calibration statistics for annual results. NSE: Nash-Sutcliffe Efficiency
coefficient; PB: percent bias, RMSE: root mean square error.
Endpoint Streamflow TN TP TSS
R2
NSE
PB
RMSE
0.934
1
16.5
17
0.934
0.472
39.8
24657
0.903
0.485
37.4
2224
0.398
0.069
24.4
3880
Draft - Do not cite or quote 29
-------
PART A: Annual Sensitivity Analysis
Scenario Development: PART A
A total of 42 SWAT model simulations were completed. Climate change scenarios included one baseline
scenario and 41 climate change scenarios. No land use or management scenarios were included.
Climate Change Scenarios
The focus of PART A was to assess the sensitivity of the Raccoon River to changes in mean annual
precipitation and temperature. Information about potential future changes in temperature and precipitation
in the Raccoon River watershed was obtained from an ensemble of statistically downscaled climate
change data acquired from The Nature Conservancy's Climate Wizard web site (www.climatewizard.org).
Climate Wizard is a user-friendly portal for accessing and visualizing summary statistics for projected
future changes in temperature and precipitation at any location within the U.S. based on GCM projections
archived by the Program for Climate Model Diagnosis Coupled Model Intercomparison Project (CMIP3)
for 16 climate models and 3 greenhouse gas emissions scenarios (SRES A2, IB, and Bl). Summary
information is presented for 2 future periods, mid-century (2050s) and end- century (2080s).
In PART A of this case study, climate change scenarios were developed to fall within the ensemble range
of projected end- century (2080s) temperature and precipitation changes for the Raccoon River watershed.
Projected changes in mean annual temperature ranged from approximately 2 to 6.5°C, and projected
changes in annual precipitation volume from -22 to +30 percent. Spatial variability in projected changes
across the watershed was relatively small.
Climate change scenarios for the SWAT model were developed using BASINS CAT's ability to create
multiple changes within a user specified range (Figure 3.2). This feature automates the creation of
multiple climate adjustments for selected variables by specifying a range and step increment within the
range (e.g., to change temperature from 0 to 3°C by increments of 1°C). When 2 or more variables are
selected, this feature creates scenarios reflecting each possible combination of changes for selected
variables. The following adjustments were made to the Raccoon River temperature and precipitation
records for 1960-2001:
• Average daily temperatures increased by 0 to 5°C at increments of 1°C (i.e., 0, 1, 2, 3, 4, 5).
• Precipitation volume adjusted by -10 to +20 percent at increments of five percent (i.e., -10, -5, 0,
5, 10, 15,20).
A total of 42 climate change scenarios resulted from each unique combination of the six temperature and
seven precipitation adjustments. For simulation of each scenario, the SWAT model used the modified
temperature and precipitation inputs to internally re-compute potential evapotranspiration (PET) using the
Penman-Montieth algorithm. This differs from other BASINS CAT models (HSPF, SWMM) where PET
is re-computed external to the model by BASINS CAT, and then provided to the model as an input
variable in the same manner as temperature and precipitation.
Draft - Do not cite or quote 30
-------
Modify Existing Data
Precp
base 1 AREA (and 1 more)
Multiply Existing Values by a Number (eg Precipitation)
Modification Name:
Existing Data to Modify:
How to Modify:
Number to multiply existing data by
O Single Change '•*_) Multiple changes within specified range
Minimum |p.9
Maximum:
1.2
Increment: 0.05
multiplication factor
multiplication factor
Events
EH Vary values only in the following Events
Exceeding threshold
Allow gaps up to
Sum of values exceeding threshold
Total duration above
Months/Years
EH Vary only in selected
None
Ok
Cancel
Figure 3.2. BASINS CAT option window for making multiple changes within a user specified range.
Land Use Scenarios
Land use and land cover (LULC) data were held constant for all model runs. While it is unlikely LULC in
the Raccoon River will not change in the future, holding it constant allowed for the assessment of
potential impacts from climate change only.
Management Scenarios
Future management scenarios were not evaluated in this case study.
Endpoint Selection: PART A
The endpoints for this study consisted of mean annual streamflow and loadings of TN, TP, and TSS. A
cursory assessment of these constituents was considered appropriate given the goal of evaluating general
watershed sensitivity to climate change in a highly agricultural watershed.
Results: PART A
Model results for mean annual streamflow and loadings of TN, TP, and TSS are shown in Tables 3.4 to
3.7. The combination of zero percent change in precipitation and 0°C change in temperature represents
the baseline conditions of the model (historical climate). An effective method for analyzing results from a
series of climate scenarios created with BASINS CAT is the pivot table capability. The tool allows users
to interactively build pivot tables by specifying row, column, and cell variables (Figure 3.3). The pivot
tables are displayed in the BASINS CAT interface and may also be saved in a form readily available for
use by common spreadsheet tools.
Draft - Do not cite or quote
31
-------
Results show changes in streamflow were proportional to changes in precipitation, and inversely
proportional to changes in air temperature. Changes in pollutant loads were directly proportional to
changes in streamflow, increasing with increases in streamflow. For example, a five percent increase in
precipitation and a 0°C increase in temperature resulted in 56 cms in mean annual streamflow, while a
five percent increase in precipitation with a 5°C increase in temperature resulted in 33 cms in mean
annual streamflow.
1. Climate Assessment Tool 2.0 QPlfx]
File Edit Options
Model || Results Piv
Help
ot Table
Rows A
Columns P
Cells F
0.8
0 30.731
1 26.35
em Add Current Value
ecp Multiply Current Value
owOut Mean base REACH 2 FLOW_OUT
0.95 1 1.05 1.1 1.15
38.789 47.149 56.029 65.373 74.687
33.859 41.96 50.585 59.559 69.633
2 22.609 29.756 37.367 45.604 54.258 63.015
3 19.159 25.82 32.971 40.969 49.218 57.713
4 16.211
22.288 28.996 36.399 44.391 52.57
5 13.697 19.269 25.593 32.652 40.277 43.158
V
J
V
1.2
84.19
77.874
72.052
66.45
61.038
56.418
Figure 3.3. BASINS CAT pivot table window displaying results for mean annual streamflow at the outlet
of the Raccoon River SWAT model.
Table 3.4. Mean annual streamflow (cms) for all combinations of temperature and precipitation change.
Baseline condition is highlighted by the box in the first column.
Precipitation
Change, %
-10
-5
0
5
10
15
20
Temperature Change,
0 1
31
39
47
56
65
75
84
26
34
| 42
51
60
69
78
°C
2
23
30
37
46
54
63
72
3
19
26
33
41
49
58
66
4
16
22
29
36
44
53
61
5
14
19
26
33
40
48
56
Draft - Do not cite or quote
32
-------
Table 3.5. Mean annual nitrogen load (kg/ha/yr) for all combinations of temperature and precipitation
change. Baseline condition is highlighted by the box in the first column.
Precipitation
Change, %
-10
-5
0
5
10
15
20
Temperature Change
0 1
7.9
9.8
11.6
13.5
15.4
17.1
18.9
7.1
9.0
| 10.9
13.0
14.9
16.7
18.6
°C
2
6.3
8.1
10.0
12.1
14.1
15.9
17.8
3
5.6
7.4
9.2
11.3
13.5
15.5
17.5
4
5.0
6.8
8.6
10.8
13.1
15.2
17.3
5
4.4
6.1
7.9
9.9
12.2
14.4
16.6
Table 3.6. Mean annual phosphorous load (kg/ha/yr) for all combinations of temperature and
precipitation change. Baseline condition is highlighted by the box in the first column.
Precipitation
Change, %
-10
-5
0
5
10
15
20
Temperature
0 1
0.6
0.7
0.8
1.0
1.1
1.3
1.4
0.5
0.6
| 0.8
0.9
1.1
1.2
1.3
Change °C
2
0.5
0.6
0.7
0.9
1.0
1.1
1.3
3
0.4
0.6
0.7
0.8
1.0
1.1
1.3
4
0.4
0.5
0.7
0.8
1.0
1.1
1.3
5
0.4
0.5
0.6
0.8
0.9
1.1
1.3
Table 3.7. Mean annual TSS load (tonnes/ha/yr) for all combinations of temperature and precipitation
change. Baseline condition is highlighted by the box in the first column.
Precipitation
Change, %
-10
-5
0
5
10
15
20
Temperature Change, °C
012
0.6
0.8
1.0
1.2
1.5
1.8
2.1
0.5 0.4
0.7 0.6
| 0.8 0.7
1.1 1.0
1.3 1.2
1.6 1.4
1.9 1.7
3
0.3
0.5
0.6
0.8
1.1
1.3
1.6
4
0.3
0.4
0.6
0.8
1.0
1.2
1.5
5
0.2
0.4
0.5
0.7
0.9
1.1
1.4
Evapotranspiration for the climate scenarios was analyzed to determine its potential role on endpoint
values. BASINS CAT was used to generate apivot table of modeled evapotranspiration from the land
surface for each of the climate scenarios (Table 3.8). Evaporation from water surfaces was
inconsequential and not included. Evapotranspiration is much less sensitive to changes in precipitation
versus temperature, and is the likely cause of the decrease in annual streamflow and pollutant loads as
temperature increases from the baseline. As temperature increases, evapotranspiration increases, and the
amount of streamflow decreases. For example, as temperature increases from 0 to 5°C, holding the
precipitation increase constant at five percent, streamflow decreases from 56 cms to 33 cms, while at the
same time evapotranspiration increases from 60.5 cm/yr to 69.2 cm/yr.
Draft - Do not cite or quote
33
-------
Table 3.8. Mean annual evapotranspiration (cm/yr) for all combinations of temperature and precipitation
change. Baseline condition is highlighted by the box in the first column.
Precipitation
Change, %
-10
-5
0
5
10
15
20
Temperature Change, °C
0 1 2
58.7
59.4
60.0
60.5
61.0
61.4
61.8
60.4 61.9
61.3 62.9
] 61.9 63.7
62.5 64.4
63.1 65.0
63.6 65.6
64.0 66.1
3
63.3
64.4
65.3
66.1
66.9
67.5
68.1
4
64.6
65.9
66.9
67.8
68.7
69.4
70.0
5
65.8
67.1
68.2
69.2
70.2
71.0
71.7
In addition to pivot tables, contour plots can provide a visual display of results from the climate scenarios,
a presentation useful for climate vulnerability assessment and decision support. While BASINS CAT
does not directly generate contour plots, model output can be exported as text files for use with any
graphics and plotting software. Figure 3.4 is a contour plot of the simulated change in streamflow for each
combination of temperature and precipitation adjustments. Contours were generated by interpolation from
the original 42 scenario endpoints, indicated as dots on the plot, using DPlot software
(http://www.dplot.com). The impact of warming temperatures on mean annual streamflow can be seen by
moving vertically from the point labeled "Current Climate". Similarly, the impact of changes in
precipitation can be seen by moving horizontally in the plot.
Draft - Do not cite or quote
34
-------
Change in Mean Annual Flow, %
-50 5 10 15
Change in Mean Annual Precipitation. %
Figure 3.4. Contour plot showing percent change in mean annual streamflow for all combinations of
temperature and precipitation change.
PART B: Seasonal Sensitivity Analysis
Scenario Development: PART B
A total of 7 model simulations were completed. Scenarios included 1 baseline climate scenario and 6
climate change scenarios. No land use change or management scenarios were included.
Climate Change Scenarios
In PART A, climate change scenarios were created without consideration that climate change may vary
seasonally throughout the year. Adjustments were made uniformly to all temperature and precipitation
values within the historical baseline period. Although not well understood, it is possible that climate
change will vary seasonally throughout the year. For example, there may be increases in winter
precipitation with little change during the summer months. Similarly, greater warming may occur during
the winter months than summer. The effects of the seasonal timing of climate change on streamflow and
water quality could be great. BASINS CAT provides the capability to apply change factors to only
selected months of the year. This capability allows scenarios to be created representing changes that vary
on a seasonal basis.
Draft - Do not cite or quote
35
-------
For PART B, the climate scenarios were developed to explore how seasonal precipitation patterns can
impact mean monthly and annual endpoint values. As in PART A, climate scenarios were developed
using the change factor methodology (CFM) (Anandhi et al. 2011), often called the delta change method.
Scenarios were developed by adjusting the mean monthly values of temperature and precipitation for the
entire 1960 to 2001 simulation period. To create climate change scenarios that reflect plausible seasonal
variation of change, monthly deltas were developed using dynamically downscaled GCM model
projections developed by the North American Regional Climate Change Assessment Program
(NARCCAP) (http://www.narccap.ucar.edu). The NARCCAP projections are a series of high resolution
climate simulations developed by nesting RCMs within coarser resolution GCMs. The baseline
simulations cover 1971-2000 and the climate change simulations cover 2041-2070. The changes applied
to the Raccoon River SWAT model weather data represented the difference in monthly average values
between the baseline and the future simulations.
Temperature and precipitation data from four NARCCAP models were used to develop an initial set of
climate change scenarios (CC-1, CC-2, CC-3, CC-4) (Table 3.9). CC-3 was further modified to create two
additional climate change scenarios. CC-5 and CC-6 maintained the same mean annual rainfall and
temperature as CC-3, but monthly changes were altered to represent two different seasonal patterns of
changes that sum to the same net annual values.
Table 3.9. NARCCAP climate models used to develop the case study scenarios.
Climate
Scenario
CC-1
CC-2
CC-3
CC-4
CC-5
CC-6
NARCCAP
Climate Model
crcm cgcm3
rcm3 cgcm3
rcm3_gfdl
wrfg ccsm
rcm3_gfdl_(with synthetic monthly adjustments)
rcm3_gfdl (with synthetic monthly adjustments)
The BASINS CAT Months/Years adjustment feature was used to modify the monthly temperature and
precipitation climate data in the SWAT model for each of the simulations (Figure 3.5). Adjustments to
historical precipitation are made by specifying a multiplier that is applied to each record within a given
month in the precipitation time series. Temperature records are adjusted by specifying a constant degree
change within a given month in the temperature time series.
PART B of this case study also incorporates consideration of spatial variability of climate change within
the Raccoon River watershed. The Raccoon River SWAT model used in this case study receives
meteorological input from two locations, the northern and southern subwatersheds of the Raccoon River
basin (Di Luzio 2008). In PART A of this case study, identical change factors were applied to
temperature and precipitation data from each of these locations for each scenario considered. This
approach assumes the spatial variability of climate change within the study watershed is negligible.
Conversely, spatial variability in climate change can be represented in BASINS CAT by applying
different sets of change factors to meteorological data from stations in different watershed locations (e.g.,
in large or topographically complex watersheds). In PART B of this case study spatial variability was
represented by applying different monthly change factors to data from the northern and southern
subwatersheds of the Raccoon River basin. Using the BASINS CAT's Months/Years adjustment feature,
adjustments were first applied to temperature and precipitation data from one location, followed by
application of a different set of adjustments to data from the second location.
Draft - Do not cite or quote 36
-------
Modify Existing Data
i- n x
Modification Name:
Existing Data to Modify:
H ow to M odify: M ultiply E xisting Values by a N umber (eg Precipitation]
base PCP2 PR EC
Number to multiply existing data by
(*) Single Change O Multiple changes within specified range
Value
1.2102
multiplication factor
Events
Vary values only in the following Events
Exceeding threshold
Allow gaps up to
Sum of values exceeding threshold
Total duration above
Months/Vears
0 Vary only in selected
Month
Feb
Mar
Apr
May
Jun
ul
Aug
Sep
Uct
H ov
Dec
Ok
Cancel
Figure 3.5. BASINS CAT modify existing data window showing adjustment to January precipitation
using the Months/Years adjustment capability.
Land Use Scenarios
Same as PART A. See Section 3.2.1.
Management Scenarios
Same as PART A. See Section 3.2.1.
Endpoint Selection: PART B
The endpoints for this study consisted of annual streamflow and loadings of TN, TP, and TSS at monthly
and annual time steps.
Results: PART B
Model simulation results are shown in Figures 3.6 to 3.9 and Tables 3.10 to 3.13. The BASINS CAT
ability to specify monthly time subsets for endpoint analysis was used to extract the mean monthly values.
Table 3.14 presents the annual total for precipitation, TN, TP, and TSS and the mean annual streamflow
for all scenarios. In general, the contrast of the various monthly climate adjustments clearly demonstrate
Draft - Do not cite or quote
37
-------
that the distribution of rainfall and temperature changes within a year produced significantly different
outcomes in streamflow and pollutant loadings.
The simulation results as presented in Table 3.14 indicate that all climate scenarios have higher mean
annual temperature and annual total precipitation than the baseline. The mean monthly dynamics for
precipitation, streamflow and pollutant loadings deviate within Scenarios CC-1 to CC-4 from the baseline
(Figures 3.6 and 3.7 and Tables 3.10 to 3.13). In general, the climate scenarios have higher mean monthly
precipitation than the baseline in the spring and fall, but lower mean monthly precipitation in the summer.
Looking at the trends in detail, it is evident that monthly differences in precipitation in combination with
increased temperatures can have a significant impact on the endpoints. For example, the climate change
simulations indicate higher mean precipitation in the spring versus the baseline. However, the endpoint
values tend to be lower than the baseline, possibly the result of earlier plant growth and higher rates of
evapotranspiration caused by warmer spring temperatures.
The influence of monthly variation in precipitation and temperature is further demonstrated by the
comparisons of CC-3 and two synthetically adjusted versions, CC-5 and CC-6. All three scenarios have
the same mean annual rainfall and temperature, but the seasonal distribution of precipitation is modified
in CC-5 and CC-6 (Figure 3.8, top panel). The modified scenarios indicate additional potential impacts on
mean monthly and annual streamflow and pollutant loads. For example, in Scenario 6, more than
doubling the baseline precipitation in the winter and early spring resulted in a significant increase in all
endpoints during those same months (Figures 3.8 and 3.9) likely due to reduced evapotranspiration and
limited plant uptake. This scenario also had the highest mean annual streamflow and pollutant loadings,
indicating potential risk from higher precipitation volumes in the winter and spring (Table 3.14).
Draft - Do not cite or quote 38
-------
6 7
Month
10
11
12
Figure 3.6. Mean monthly precipitation, temperature, and streamflow rate for the NARCCAP and
baseline scenarios.
Draft - Do not cite or quote
39
-------
10 11
12
Month
Figure 3.7. Mean monthly TSS, TN, and TP loadings for the NARCCAP and baseline scenarios.
Draft - Do not cite or quote
40
-------
6 7
Month
10
11
12
Figure 3.8. Mean monthly precipitation, temperature, and streamflow rate for the CC-5, CC-6, and
baseline scenarios.
Draft - Do not cite or quote
41
-------
I 1 1 1
1 1 1 1 1 1 1
0.30 -i
10 11
12
Month
Figure 3.9. Mean monthly TSS, TN, and TP loadings for the CC-5, CC-6 and baseline scenarios.
Draft - Do not cite or quote
42
-------
Table 3.10. Mean monthly streamflow (cms) for all scenarios.
Climate
Scenario
baseline
CC-1
CC-2
CC-3
CC-4
CC-5
CC-6
Table 3.11.
Climate
Scenario
baseline
CC-1
CC-2
CC-3
CC-4
CC-5
CC-6
Table 3.12.
Climate
Scenario
baseline
CC-1
CC-2
CC-3
CC-4
CC-5
CC-6
Jan
22
32
37
25
32
31
68
Mean monthly
Jan
0.6
0.9
0.9
0.6
0.9
0.8
1.7
Mean monthly
Jan
0.04
0.1
0.1
0.06
0.06
0.07
0.18
Feb
34
26
28
21
33
30
101
nitrogen
Feb
0.7
0.6
0.6
0.5
0.8
0.6
2.1
Mar Apr
54 64
29 44
35 61
54 83
45 68
46 66
103 78
load (kg/ha)
Mar Apr
1 0.9
0.7 0.9
0.7 1.2
1.1 1.7
0.9 1.4
0.9 1.3
2.2 1.9
May
78
83
74
124
87
111
71
Jun Jul
91 63
113 51
94 48
94 38
70 43
141 65
48 26
Aug
31
17
21
15
20
24
8
Sep
37
19
44
30
63
27
26
Oct
38
21
34
41
60
31
40
Nov
31
23
27
20
53
34
46
Dec
22
20
22
17
37
30
57
Mean
47
40
44
47
51
53
56
for all scenarios.
May
1.6
2.3
2.1
3.5
2.5
3.1
2.3
phosphorous load (kg/ha) for all
Feb
0.08
0.07
0.07
0.05
0.08
0.07
0.28
Mar Apr
0.14 0.08
0.08 0.09
0.08 0.12
0.16 0.17
0.11 0.12
0.12 0.12
0.26 0.1
May
0.1
0.2
0.13
0.27
0.17
0.25
0.1
Jun Jul
2.3 1.8
2.9 1.8
2.6 1.6
3.1 1.8
2.3 1.5
3.9 2.3
1.8 1
scenarios.
Jun Jul
0.13 0.08
0.19 0.05
0.16 0.06
0.13 0.03
0.1 0.05
0.26 0.07
0.05 0.03
Aug
0.9
0.7
0.8
0.7
0.7
1
0.3
Aug
0.03
0.01
0.02
0.01
0.02
0.01
0.01
Sep
0.6
0.4
0.9
0.6
1.2
0.6
0.5
Sep
0.06
0.03
0.09
0.06
0.14
0.04
0.06
Oct
0.4
0.3
0.5
0.6
0.9
0.5
0.6
Oct
0.04
0.03
0.04
0.06
0.09
0.04
0.07
Nov
0.3
0.3
0.4
0.3
0.8
0.5
0.6
Nov
0.02
0.02
0.02
0.01
0.04
0.03
0.04
Dec
0.5
0.5
0.5
0.4
0.9
0.7
1.2
Dec
0.03
0.03
0.03
0.02
0.04
0.04
0.1
Mean
1
1
1.1
1.3
1.2
1.3
1.4
Mean
0.07
0.07
0.08
0.09
0.09
0.09
0.11
Table 3.13. Mean monthly TSS load (tonnes/ha) for all scenarios.
43
-------
Climate
Scenario
baseline
CC-1
CC-2
CC-3
CC-4
CC-5
CC-6
Jan Feb Mar
0.03 0.05 0.1
0.05 0.04 0.05
0.06 0.04 0.05
0.03 0.02 0.1
0.04 0.04 0.07
0.04 0.04 0.08
0.13 0.22 0.26
Apr May
0.11 0.14
0.07 0.18
0.11 0.14
0.17 0.29
0.12 0.18
0.12 0.25
0.14 0.13
Jun Jul
0.18 0.13
0.27 0.11
0.21 0.1
0.2 0.06
0.13 0.09
0.36 0.14
0.08 0.05
Aug
0.05
0.02
0.03
0.02
0.03
0.03
0.01
Sep
0.07
0.03
0.09
0.06
0.15
0.04
0.05
Oct
0.07
0.03
0.06
0.08
0.12
0.05
0.08
Table 3.14. Total and mean annual precipitation, temperature, streamflow, TN, TP, and TSS loads
Climate Scenario
baseline
CC-1
CC-2
CC-3
CC-4
CC-5
CC-6
Total Annual
Precipitation,
mm
813.0
835.9
853.5
857.9
884.0
884.2
884.3
Temp.
Annual Mean,
C
8.83
11.85
11.61
11.50
11.22
11.22
11.22
Mean Annual
Streamflow,
cms
47.1
39.8
43.7
46.9
51.0
52.9
55.7
Nov Dec
0.04 0.02
0.03 0.02
0.03 0.02
0.02 0.02
0.08 0.05
0.04 0.04
0.07 0.09
for all scenarios.
Mean
0.08
0.07
0.08
0.09
0.09
0.1
0.11
Total Annual Total Annual
TSS, tonnes/ha TN, kg/ha
0.98
0.88
0.94
1.07
1.11
1.23
1.30
11.6
12.4
12.9
15.1
14.7
16.2
16.2
Total Annual
TP, kg/ha
0
0
0
1
1
1
1
83
89
90
03
01
11
27
44
-------
Summary
This study assessed the sensitivity of a predominantly agricultural watershed, the Raccoon River, to
climate change. The primary BASINS CAT feature demonstrated in this study was the ability to automate
the adjustment of temperature and precipitation time series data for the SWAT watershed model on a
monthly, seasonal, and annual basis. This allowed an array of climate change scenarios to be
automatically generated and used as inputs for model simulations. The pivot table feature was used to
generate endpoint result tables for all temperature and precipitation change combinations considered. This
capability made it possible to quickly develop and evaluate watershed sensitivity to the climate change
scenarios. In PART B, the BASIN CAT ability to report mean monthly values for endpoints was used to
extract monthly endpoint outputs from the model simulations.
The climate scenarios applied in PART A and B were developed using synthetic adjustments and climate
model projections. While simple, they were effective in showing that the Raccoon River watershed is
indeed sensitive to changes in both precipitation and temperature. PART A presented a general watershed
sensitivity analysis whereby uniform adjustments to precipitation and temperature were applied across the
entire watershed and simulation time series. The results enabled the development of a contour plot, a
simple guide for assessing watershed response across a range of temperature and precipitation changes. In
the context of watershed management, a simple assessment such as the one presented may be adequate,
providing enough detail to inform the underlying watershed management goal. This type of modeling and
analysis approach can also be extended to assess the general sensitivity of land-use change or land
management practices.
PART B explored how seasonal shifts in climate change (mainly in terms of varied precipitation) can
affect mean monthly and annual endpoints results. While not well understood and somewhat uncertain,
seasonal shifts in precipitation due to climate change are likely. Further, a shift in the timing or
seasonality of climate patterns could lead to unexpected outcomes that are not visible when examining the
trends at an annual scale only. To illustrate the effect of seasonal variability, monthly adjustments to
precipitation and temperature in the form of delta changes were applied to the baseline Raccoon River
watershed SWAT model. The comparison of the climate scenarios to each other and the baseline
demonstrates the potential effects of seasonal shifts of climate on streamflow and pollutant loadings in the
Raccoon River watershed, especially precipitation as explored in Scenario CC-5 and CC-6. The
sensitivity of results to the seasonality of changes could have implications for management decision-
making. For example, if climate change in the UMRB results in higher precipitation volumes in the
winter months, higher runoff and pollutant loads are likely. Managers may have to look toward more
winter cover crops or modified fertilizer application protocols.
45
-------
3.3. Urban stormwater sensitivity to rainfall change and effectiveness of management in the
Upper Roanoke River, VA, using BASINS CAT with SWMM
Case Study Overview
Using BASINS CAT with SWMM, this case study demonstrates a general urban stormwater
sensitivity analysis (PART A), including precipitation change scenarios, to assess stormwater
volumes, nutrient and TSS concentrations from an urban redevelopment site. In PART B,
the ability of two management scenarios to meet stormwater goals is assessed under the same
precipitation change scenarios presented in PART A.
In PARTS A and B, a rainfall event was adjusted using synthetic adjustments across a range;
specifically event volume was increased by 10 to 30 percent by increments of 10 percent.
Management scenarios in PART B included distributed and centralized stormwater
management.
Introduction
Urbanized watersheds generally exhibit a higher sensitivity to rainfall and snowmelt events than pre-
development conditions. Impervious cover alters local hydrologic processes, creating amplified runoff
events and decreased baseflow in streams during periods of dry weather. Urban stormwater runoff carries
pollutants from roads, rooftops, and other impervious areas into stormwater systems and nearby streams.
Urban stormwater impacts on local aquatic ecosystems and public health have made it a high
management priority in many cities. Changes in precipitation regimes could significantly alter stormwater
runoff volumes and pollutant loading from urban environments. Water resource planners, engineers, and
others engaged in stormwater management should explore the interactions of climate change with the
urban environment in their planning and decision making.
PART A of this case study explores the sensitivity of stormwater runoff from a commercial
redevelopment site to precipitation change at the event scale. The Storm Water Management Model
(SWMM) was used to simulate the urbanized watershed's response, and BASINS CAT was used to
develop the precipitation change scenarios. The BASINS CAT capability to create multiple changes
within a user specified range was used to modify rainfall event volumes, creating an array of climate
scenarios for use in the model simulations. Endpoints analyzed included stormwater flow rate and event
mean concentrations (EMCs) of TP and TSS at the site outlet.
PART B assessed two additional SWMM models where stormwater BMPs were employed to explore the
benefits of alternative stormwater management scenarios under precipitation change. Such analysis can be
a cost effective way of evaluating climate change adaptation strategies. BASINS CAT was used to
develop the precipitation change scenarios and assess the results from the model simulations.
Location Description
The study site was a 0.2 km2 commercial redevelopment project located in the headwaters of the Upper
Roanoke River (HUC03010101) in southwest Virginia (Figure 3.10; Young et al. 2009). The site was
previously undeveloped except for a few small commercial buildings and a motel. Since 2008, it has been
undergoing two phases of redevelopment. Phase I involved the construction of a shopping mall, theater,
restaurants, and stormwater detention facility in the southern portion covering approximately 0.05 km2 of
the site. Phase II called for a big-box retail development in the northern portion covering approximately
0.1 km2 of the site. The entire site drains from the northwest to southeast corner. Baseline land use
46
-------
categories for the site are shown in Table 3.15. For this small urban watershed, there are three distinct
land use categories: green space, impervious surfaces, and rooftop. The impervious space represents any
paved surface such as roads or parking lot. Green space represents any naturally occurring or manmade
pervious land cover. Rooftop represents all rooftop surfaces, including both conventional and vegetated
rooftop.
Wesi Virginia
North Carolina
Reaches in 03010101
First and Main Site
HUC0301Q101
A/
D
0.125
I
0.25
0.375
I
Scale in Miles
Figure 3.10. The commercial redevelopment site in the Upper Roanoke River watershed.
Table 3.15. Baseline land use summary for the commercial redevelopment site.
Land Use Portion of Watershed, %
Green space
Impervious
Rooftop
43
41
16
Water Model Setup
Evaluation of a watershed for the purpose of storm water management is commonly completed on the
timescale of individual rainfall-runoff events. The SWMM model was developed for running event-based
simulations, unlike HSPF and SWAT which are typically run on a continuous, annual or longer time scale
(although SWMM can also be run on continuous time scales). This case study used a series of SWMM
models originally developed for evaluating an optimization tool designed to improve site development
and stormwater BMP selection in Virginia (Young et al., 2009). The model included subwatershed
delineation and hydrologic discretization of the SWMM hydraulic schematic. In the original evaluation,
47
-------
two alternative stormwater management scenarios were modeled in comparison to the baseline scenario,
pre-development conditions with no runoff control measures.
The SWMM model used in this case study was not calibrated; rather the baseline scenario was used as a
basis against which the two alternative stormwater management scenarios were compared. The model
uses a custom designed SCS Type II storm that has a 31.7 mm/hr rainfall intensity for a 1-hour duration.
The model simulation duration is two days with a 1-hour design event in the beginning hour of the
simulation. For this case study, temperature data were retrieved from a nearby National Climatic Data
Center (NCDC) weather station in Blacksburg, VA. Initial assessment of model simulations indicated that
temperature was not a significant factor given the short timescale of the event; therefore, changes to
temperature were not included in the simulations.
PART A: Runoff Sensitivity Analysis
Scenario Development: PART A
A total of 10 model simulations were completed. Scenarios included 1 baseline scenario and 9
precipitation change scenarios. No land use or management scenarios were included.
Precipitation Change Scenarios
Precipitation change scenarios were developed to fall within the ensemble range of projected end-of-
century (2080s) precipitation changes for this region based on statistically downscaled data from 16
CMIP3 climate models acquired from the Climate Wizard web site (see Section 3.2 for more information
on these models). The ensemble range of projected change was used to establish boundary conditions for
the design event adjustments. Projected changes in annual precipitation ranged from -17 to +27 percent.
The climate scenarios for input to SWMM were developed using the BASINS CAT capability to create
multiple changes within a user specified range. This feature automates the creation of multiple
adjustments for selected variables by specifying a range and step increment within the range (e.g., to
change temperature from 0 to 3°C by increments of 1°C). In this study, the design event rainfall intensity
was increased 10 to 30 percent by increments of 10 percent, which together with the baseline scenario
resulted in a total of four precipitation change scenarios.
The event rainfall intensity for the baseline and three precipitation change scenarios are shown in Figure
3.11. For each model run, BASINS CAT was used to generate a revised PET record based on the revised
temperature record using the Hamon method (Hamon, 1961). The revised PET record was then provided
as an input variable to the SWMM model in the same manner as the revised temperature and precipitation
records.
48
-------
200
150
=55 100
c
're 50
Baseline Rainfall Event (31.75 mm/hr)
10% Increase
20% Increase
30% Increase
10 20 30 40
Time (minute)
50
60
Figure 3.11. Rainfall event intensity for the baseline 1- hour rainfall event and three precipitation change
scenarios.
Land Use Scenarios
Part A of this case study does not include land use scenarios. LULC was held constant to assess the
impacts of precipitation change only.
Management Scenarios
Part A of this case study does not include management scenarios. Stormwater BMPs were not included to
assess the impacts of precipitation change only.
Endpoint Selection: PART A
The endpoints for this study included the event mean stormwater flow rate (cms) and the EMCs of TP and
TSS at the site outlet. The original study by Young et al. (2009) included TP and TSS, both reported as
key pollutants in the State of Virginia. Using BASINS CAT, the endpoint time series and desired
statistics at event or months/years timescales can be reported.
Results: PART A
Table 3.16, developed using the BASIN CAT pivot table capability, shows the resulting event values for
the baseline and three precipitation scenarios. Precipitation values are presented as sum totals for the
event, while flow, TSS, and TP are event means. The rainfall event dynamic and flow hydrograph for all
scenarios are shown in Figure 3.12. The increase in rainfall volume in the simulated design storm was
found to increase the flow rate and pollutant concentrations during the event. While the flow increases of
14, 28, and 38 percent followed a nearly linear response to the 10, 20, and 30 percent increase in rainfall
volume respectively, increases in pollutant concentration (4, 6.6, and 6.9 percent, respectively, for TSS;
11, 18, and 20 percent, respectively, for TP) diminished as precipitation volume increased. This response
is logical given that first flush of pollutants would be washed away at a much faster rate for larger events
and pollutant concentrations will become diluted as runoff volumes increase.
49
-------
Table 3.16. Event rainfall intensity, mean flow, and concentrations of TP and TSS for the baseline and
three precipitation change scenarios.
Scenarios
Baseline
+10%
+20%
+30%
Rainfall and En
Event rainfall
intensity
(mm/hr)
31.75
34.90
38.10
41.30
dpoints
Flow
(cms)
0.021
0.024
0.027
0.029
TSS
(mg/1)
0.981
1.020
1.046
1.049
TP
(mg/1)
0.423
0.471
0.501
0.509
20
40
80
100
120
140
160
180
Rain Intensity (mm/miimt
O i — * ' — L
Zi "wh CD
-------
Land Use Scenarios
Part B of this case study does not include land use scenarios. LULC was held constant to assess the
impacts of the precipitation change and management scenarios only.
Management Scenarios
A baseline SWMM model with no stormwater BMPs and two alternative SWMM models representing
two different stormwater management strategies were included in PART B:
• Baseline: Pre-redevelopment conditions.
• Centralized management: strategy that consists of installing a small number of conventional mass
storage/detention structures to collect storm runoff. These detention structures are designed to
capture runoff from large drainage areas within the entire watershed and release through control
structures such as weirs. Such a mass storage-delayed release approach serves both to reduce the
peak flow rate and to help reduce pollutant loading in runoff through filtration and gravitational
settling.
• Distributed management: strategy that consists of a large number of storage structures of low
capacity, combined with infiltration structures, such as pervious pavement and green rooftops,
throughout the headwater areas of the watershed. By spreading out multiple source-control BMPs
throughout the whole drainage basin, this approach can achieve in-situ runoff volume reduction
while minimizing pollutant movement off-site.
Endpoint Selection: PART B
PART B focused on stormwater flow from the redevelopment site outlet. An arbitrary management target
of maintaining stormwater runoff below 0.02 cms (0.7 cfs in Figure 3.13) was assumed to illustrate the
potential utility of this analysis for management decision making. BASINS CAT allows users to specify
such thresholds and color code endpoint values that exceed it (e.g., see Figure 3.13).
51
-------
E ndpoint N ame: T otal I nflow E vt
D ata set: | FandM_Centralized_M anual S D11 T otal I nflow
Attribute: Mean
Manage Attributes
Highlight Values
Default Color: White
M inimum Value: < none>
Color Lower Values: D eepS kyB lue
Maximum Value: 0.7
Color H igher Values: 0 rangeR
Events
Only include values in the following Events
Exceeding threshold
0
Allow gaps up to
Sum of values exceeding threshold 0
Total duration above
1
M onths/Years
O Only include values in selected
All
Ok
None
Cancel
Figure 3.13. BASINS CAT Endpoint window where users can specify endpoints, event and monthly
(inter-annual) statistics, and thresholds for color coding the results.
52
-------
Results: PART B
Changes in event rainfall intensity resulted in increased flows of stormwater across the baseline and two
management scenarios (Table 3.17). The simulation results indicated that the mean flow at the outlet is
the highest with the baseline, followed by centralized management, and distributed management. The
event rainfall dynamic and runoff hydrograph for the baseline and precipitation change scenarios are
shown in Figure 3.14. Increases in the design event rainfall intensity were shown to increase the flow rate
in an almost linear fashion for all three management scenarios. Both centralized and distributed
management helped reduce peak flow rate significantly compared to the baseline scenario with no
management practices in place. The centralized scheme also prolonged flow duration (at a very low flow
rate) to beyond 400 minutes after the onset of the design event (not shown in Figure) while flow
essentially ended after 180 minutes for the baseline and the distributed management scenarios. The
distributed management scenario utilized a series of source-control BMPs that have limited storage
capacity (in contrast to the large detention structures in the centralized management scenario), which can
be overcome in larger storms, leading to the delayed "second peak" in its hydrograph. Nevertheless, like
centralized management, it drastically reduced the flow rate under all precipitation change scenarios.
Figure 3.15 illustrates the option within BASINS CAT to display simulation results with endpoint values
color-coded based on a user-specified criterion. In this case, 0.02 cms was chosen (arbitrarily) as the
ceiling stormwater flow rate above which the cell containing the endpoint value is highlighted in orange.
Table 3.17. Event mean flow (cms) under all precipitation change and management scenarios. Flow
values greater than the threshold are highlighted in orange.
Scenario
Precipitation
31.7
Base
mm/hr)
34.9
+10%
38.1
+20%
41.3
+30%
Baseline
Centralized
Distributed
0.021
0.016
0.004
0.024
0.019
0.006
0.027
0.008
0.029
0.025
0.011
53
-------
20
40
60
80
100 120 140 160
180
Rain Intensity (mm/minut
0 i— ' <-*
=1 l-^i CD ^
=1 CD CD CD C
1 1
— _ _
r^L
-----
....
1 1
• +30% Rain - Flow
• +20% Rain - Flow
• +10% Rain- Flow
Baseline Rain - Flow
2.0
1.5
0.5
0.0
2.0
Baseline (no management)
£
u
0.5
0.0
2.0
1.5
1.0
0.5
0.0
0
Centralized management
Distributed management
20
40
60
120
140
160
ISO
80 100
Time (minute)
Figure 3.14. Rainfall versus flow dynamics at the site outlet for all precipitation change and management
scenarios.
54
-------
Climate Assessment Tool 2.0 - BASINSCAT SWMM
File Edit Options Help
Model Climate Data Assessment Endpoints Results
Pivot Table
Start
Refresh
Run Rain
Run the model
Refresh results from the last model run
TernpSep Total Inflow Evt
O Show Progress of Each Run
0 Clear Results on Start
Saved Results
J Multiply
Add
Mean
Current Value Current Value FandM Centralized Manual SD11 +
base
1
2
3
4
5
6
7
1.1
1.1
1.1
1.2
1.2
1.2
1.3
1.3
1.3
1.8
3.6
5.4
1.8
3.6
5.4
1.8
3.6
5.4
0.57263
0.67018
0.67018
0.67018
0.76938
0.76939
0.86622
G: \CAT\M ode!Centralized\M odif ied-001
G:\CAT \ModelCentralized\Modified-002
G:\CAT \ModelCentralized\Modified-003
G:\CAT\ModelCentralized\Modified-004
G:\CAT \ModelCentralized\Modified-005
G:\CAT\ModelCentralized\Modified-006
G:\CAT \ModelCentralized\Modified-007
G:\CAT \ModelCentralized\Modified-008
G:\CAT \ModelCentralized\Modified-009
Finished runs
Figure 3.15. BASINS CAT result grid window for the centralized management scenario, showing the
color coded endpoint (event mean flow, cfs) values that are above the specified maximum threshold
defined in Figure 3.13.
Summary
This study assessed the sensitivity of stormwater quantity and quality in a small urban watershed in
Virginia to precipitation change using SWMM. The assessment utilized BASINS CAT to process
adjustments to design event rainfall intensity. This allowed an array of precipitation scenarios to be
automatically generated and used as input for model simulations. The pivot table feature of BASINS CAT
was used to generate endpoint result tables for the precipitation changes considered. The endpoint
definition dialog allows the user to specify maximum and minimum threshold values by which the
resulting values can be color-coded accordingly in the result grid. This can be a very helpful feature in
quickly identifying the scenarios that exceed management targets when using BASINS CAT.
PART A was a general sensitivity analysis of the baseline site conditions that indicated increasing rainfall
event intensity increased both stormwater flow rate and pollutant EMCs. PART B assessed stormwater
sensitivity to climate change from a management perspective. The assessment included three models
representing the baseline with no stormwater BMPs, a centralized management, and a distributed
management approach. Increasing precipitation resulted in an almost linear increase in stormwater flow at
the outlet for all management scenarios. On an event basis, for any given rainfall intensity, the two
55
-------
alternative stormwater management approaches significantly lowered the peak runoff flow rate; the
centralized approach resulted in the longest duration of runoff flow.
While the climate change scenarios evaluated in this case study were relatively simple, they provide a
screening-level understanding of stormwater runoff sensitivity to climate change, and the potential
effectiveness of stormwater management strategies for reducing climate change impacts. Evaluation of
more detailed climate change or management scenarios is also possible. The coupling of BASINS CAT
and hydrologic/hydraulic models, when calibrated, can facilitate development of rapid assessment
methods that provide timely and usable quantitative information. The flexible capabilities of BASINS
CAT for creating and running scenarios can aide and facilitate a wide range of analyses. These
capabilities can be an important addition to the tools used by stormwater managers to design, manage, and
maintain stormwater infrastructure.
56
-------
3.4. Agricultural soil erosion sensitivity to climate change and management practices in Blue
Earth County, MN, using WEPPCAT
Case Study Overview
In PART A, this case study demonstrates, using WEPPCAT with WEPP, a general
sensitivity analysis of sediment yield from agricultural fields under different climate, land
use, and management scenarios. PART A climate scenarios included:
• Temperature increases of 0, 2 and 4°C
• Precipitation volume adjustments of-10, 0, +20 percent
• Precipitation intensity increases of 10 percent
Land use scenarios include a30mx30m field with either Lasa or Lerdal soil series with a
slope of 2 or 5 percent. Management scenarios include corn spring chisel plow and soy
spring chisel.
In PART B, a subset of the PART A climate and land use scenarios were paired with a series
of management scenarios including alternative tillage practices and sediment filter strips to
assess these practices in the context of climate change adaptation.
Introduction
Soil loss from agricultural fields can lead to decreased productivity and impacts on adjacent water bodies.
Soil erosion is affected by land use, soil characteristics, slope, and local climate. Field crop production
can be disruptive to soil structure, resulting in significant erosion and soil loss during runoff events. In
regions of the country where agriculture constitutes a significant percentage of land use, sediment erosion
can have a significant impact on water quality. Land managers and farmers often employ management
practices to reduce agricultural impacts on surface water, including cover crops, contour plowing, and
vegetated riparian filter strips that remove sediment from runoff.
Climate change, including changes to temperature and precipitation patterns, has the potential to increase
erosion and soil loss in agricultural areas. This case study explores the sensitivity of a field under corn
and soy production in Blue Earth County, MN and the potential climate adaptation benefits of common
erosion control practices including filter strips and alternative tilling methods.
In PART A, WEPPCAT was used to assess the general sensitivity of fields under conventional corn and
soy production in Blue Earth County, MN to potential changes in climate. Climate change scenarios
included adjustments to temperature, precipitation volume, and precipitation event intensity. PART B
expanded on the analysis in PART A to explore the potential effectiveness of selected management
practices for reducing climate change impacts on soil loss.
Location Description
Minnesota ranks among U.S. states as one of the top producers of corn and soy (USDA, 2010a). Blue
Earth County, MN is located in the south-central part of the state, the region producing the majority of
these crops for Minnesota (USDA, 2010a). The county is suitable for corn and soil production given the
generally flat topography, soil quality, and ample precipitation (USDA, 201 Ob). Local soils are generally
well drained and classified under hydrologic groups A and B. The county has a few poorly drained soils
categorized under hydrologic groups C and D that developed from glacial outwash in areas with little to
no slope (USDA, 201 Ob). Topographic slopes in the county range from 0 to over 15 percent.
57
-------
Model Setup
WEPPCAT was accessed online at http://typhoon.tucson.ars .ag. gov/weppcat/index.php. User-defined
model inputs include field characteristics (field length and width, shape, slope, soil type) field
management, riparian filter strip characteristics, weather station for baseline meteorological data, and
precipitation and temperature adjustments for creating climate change scenarios. The WEPP model does
not require calibration against observed data.
PART A: General Sensitivity Analysis
Scenario Development: PART A
A total of 109 model simulations were completed. Scenarios included 18 climate change scenarios, 4 land
use scenarios, and 2 management scenarios.
Climate Change Scenarios
Climate change scenarios were developed to fall within the ensemble range of projected end-of-century
(2080s) temperature and precipitation changes based on statistically downscaled data from 16 CMIP3
climate models acquired from the Climate Wizard web site (see Section 3.2 for more information on these
models). Projected changes in temperature ranged from approximately 2°C to 7°C, and projected changes
in precipitation volume ranged from approximately -23 percent to +33 percent.
Baseline meteorological data for WEPPCAT simulations are generated using Cligen. Weather parameters
are generated by Cligen based on observed monthly average temperature and precipitation information
from NOAA National Climatic Data Center (NCDC) weather stations. WEPPCAT creates climate change
scenarios by adjusting Cligen parameters to reflect potential changes of interest to users. Available
adjustments include increases and decreases in mean monthly temperature, precipitation volume, and the
transition probabilities of a wet day following a dry day, and a dry day following a dry day (i.e., number
of wet days). These adjustments can be made either uniformly among months of the year, or individual
adjustments can be made to specific months of the year. In addition to changing precipitation volume,
Cligen parameters can also be adjusted to increase the proportion of annual rainfall occurring in large
magnitude events (i.e., to represent an increase in event intensity independent of changes in total annual
precipitation). WEPPCAT provides a capability to increase the proportion of annual precipitation
occurring in large magnitude events up to 25 percent2. Adjustments in precipitation intensity are made by
applying the user determined increase to the largest 5 percent of events, and simultaneously decreasing
precipitation in the lower 95 percent of events by the same volume such that the adjustment results in
negligible change in the volume of annual precipitation.
WEPPCAT was used to modify precipitation volume, precipitation intensity, and temperature. Baseline
meteorological inputs for simulations were obtained from the NCDC Winnebago weather station given its
proximity to Blue Earth County and location to the soils series of interest. The climate scenarios consisted
of a series of synthetic adjustments to temperature and precipitation volume and intensity applied in
combination. The meteorological data were adjusted in the following manner:
• Temperature was increased annually by 0, 2 and 4°C.
• Precipitation volume was increased annually by -10, +0 and +20 percent. (Scenarios designated
as V-10, VO, and V-20, respectively.)
Adjustment of rainfall intensity is accomplished by altering the standard deviation of the distributions of daily precipitation
used by the climate generator. This approach results in a slight change in average annual rainfall even if changes to the overall
volume are not indicated in the model inputs.
58
-------
Precipitation was then adjusted to assess the effects of increased event intensity (i.e., as defined here,
increased intensity refers to an increased proportion of annual rainfall occurring in large magnitude
events). This was accomplished by first adjusting the annual precipitation volume by
-10, 0 and 20 percent, then increasing the intensity of the largest 5 percent of precipitation events by 10
percent. This intensity adjustment does not result in an increase in mean annual rainfall; rather it
redistributes the annual volume to generate larger storms in the upper 5th percentile3. A total of 18 climate
scenarios were included in this case study.
Land Use Scenarios
PART A included 4 land use scenarios for a 30m x 30m field. Characteristics included two Blue Earth
County soils types, Lasa and Lerdal at 2 and 5 percent uniform slopes. Lasa soil is a hydrologic group A
soil (high infiltration and water holding capacity) and Lerdal is a hydrologic group C soil (low infiltration
and water holding capacity).
Management Scenarios
The two management scenarios evaluated are corn spring chisel plow and soy spring chisel. Land
management options that can be represented in WEPPCAT simulations are predefined and fixed in terms
of tilling, planting and harvesting methods (Tables 3.18 and 3.19).
Table 3.18. Soy spring chisel plow land management specifications in WEPPCAT
Date
4/5
4/10
5/10
5/10
6/10
10/15
Table
Date
4/15
4/25
5/1
5/10
5/10
6/5
10/15
Operation Type
Tillage
Tillage
Tillage
Plant-Annual
Tillage
Harvest
3.19. Corn spring chisel
Operation Type
Tillage
Tillage
Tillage
Tillage
Plant-Annual
Tillage
Harvest
Operation Name
Chisel Plow
Field cultivator, secondary tillage, after duckfoot points
Planter, double disk openers
Soybeans - Medium Fertilization Level
Cultivator, row, multiple sweeps per row
Soybeans - Medium Fertilization Level
land management specifications in WEPPCAT
Operation Name
Chisel Plow
Field cultivator, secondary tillage, after duckfoot points
Tandem Disk
Planter, double disk openers
Corn, Jefferson IA, High production 125 bu/acre
Cultivator, row, multiple sweeps per row
Corn, Jefferson IA, High production 125 bu/acre
3 "Rainfall Intensification is accomplished here by altering the standard deviation of the distributions of daily precipitation used
by the climate generator." (WEPPCAT 2011). This approach results in a slight change in average annual rainfall even if changes
to the overall volume are not indicated in the model inputs. In this case study it resulted in a minor 1-2 % decrease in average
annual rainfall. For the example, in the V-lOscenario a 10 percent decrease in volume results in 26.8 inches of rain per year,
while a 10 percent decrease in rainfall plus a 10 percent increase in rainfall intensity in the largest 5 percent of storms results in
26.2 inches of average annual rainfall. This difference was deemed insignificant and actually resulted in more conservative TSS
loads due to a decrease in annual runoff versus the volume only adjustment in annual precipitation.
59
-------
Endpoint Selection: PART A
The endpoints simulated by WEPPCAT are the same as the original WEPP model: sediment loss and
sediment yield. Sediment loss is the total amount of soil displaced along the length of a field due to runoff
as measured at the bottom of the slope. Sediment yield is the amount of soil displaced (sediment loss) as
measured at the bottom the slope minus any retained by a filter strip (if applicable).
Results: PART A
Simulation results are shown in Tables 3.20 and 3.21 and Figures 3.16 to 3.18. Results suggest sensitivity
of sediment yield to increases in precipitation volume and intensity. The greatest change was observed for
the scenario V20 + 110, with simulated a sediment yield close to double the yield under the baseline
scenario. This illustrates the synergistic effect of increasing precipitation volume and intensity on
sediment yield. Results also suggest increases in volume have a greater impact on the overall increase in
sediment yield versus intensity alone. For example, under historic weather conditions (VO) the Lasa soil
at a 2 percent slope under corn production yielded 4.9 tons/ha/yr of sediment. Increasing the precipitation
volume 20 percent resulted in 7.4 tons/ha/yr sediment yield, a 51 percent increase, while the combined
effect of increasing the precipitation volume 20 percent and event intensity 10 percent resulted in 8.3
tons/ha/yr, a 69 percent increase.
Field slope, soil hydrologic group and crop type influenced sediment yield under all climate scenarios. As
expected, a 2 percent slope resulted in a lower sediment yield versus a 5 percent slope. Lasa soil also
resulted in a lower sediment yield versus Lerdal, likely due to the soil properties affecting infiltration and
water holding capacity. Finally, corn production resulted in a much lower sediment yield versus soy.
Table 3.20. Mean annual sediment yield (tonnes/ha/yr) for corn production under conditions of changing
climate4. Scenarios named to reflect changes in precipitation volume and intensity; V = volume, I =
intensity, numerical value reflects percent change from baseline.
Soil
Type&
Slope
Temp
Increase,
°C
Precipitation Scenarios
V-10 V-10+I10 VO VO+I10 V20
V20+I10
Rainfall, mm
656.6
641.9
725.2
712.95
869.75
852.6
Lasa 2%
Lasa 5%
Lerdal
2%
Lerdal
5%
0
2
4
0
2
4
0
2
4
0
2
4
3.8
3.8
4.0
7.6
8.5
9.4
5.6
5.6
6.1
8.7
10.1
11.9
4.3
4.3
4.5
8.5
9.4
10.5
6.3
6.5
6.9
10.1
11.7
13.9
4.9
4.9
5.2
9.9
10.8
12.1
7.2
7.4
8.1
11.2
13.0
15.5
5.4
5.6
5.8
10.8
12.1
13.5
8.1
8.3
9.0
12.6
14.6
17.5
7.4
7.6
7.8
14.6
16.1
18.2
10.8
10.8
11.7
16.4
18.6
22.4
8.3
8.3
8.7
15.9
17.7
19.7
11.9
11.9
12.8
18.2
20.8
25.1
' See Footnote 3 for explanation of discrepancy in annual rainfall values resulting from intensity adjustments.
60
-------
Table 3.21. Mean annual sediment yields (tonnes/ha/yr) for soy production under conditions of changing
climate5. Scenarios named to reflect changes in precipitation volume and intensity; V = volume, I =
intensity, numerical value reflects percent change from baseline.
Soil
Type&
Slope
Temp
Increase,
°C
Rainfall, mm
Lasa
2%
Lasa
5%
Lerdal
2%
Lerdal
5%
0
2
4
0
2
4
0
2
4
0
2
4
Precipitation
V-10
656.6
6.7
6.7
7.2
16.8
16.8
17.7
9.4
9.6
10.5
23.1
23.5
25.1
Scenarios
V-
10+110
641.9
7.4
7.4
7.8
18.4
18.4
19.3
10.5
10.8
11.9
26.0
26.7
28.5
VO
725.2
8.5
8.5
9.2
21.3
21.1
22.2
12.1
12.6
13.7
30.0
30.7
32.7
VO+I10
712.95
9.2
9.4
10.1
22.9
22.9
24.2
13.2
13.9
15.0
33.0
34.3
36.8
V20
869.75
12.6
12.8
13.7
30.7
30.9
32.5
17.9
18.6
20.4
45.3
46.6
50.2
V20+I10
852.6
13.7
13.9
14.8
33.2
33.2
35.0
19.5
20.4
22.2
49.3
51.1
54.9
in
* . 0 0 °
*rt
c
O r \
S ° j ^ A
1 J
0)
1 1
0 2 4
Change in Temperature °C
4V-10
V-lOtllO
A VO
VOH10
• V20
OV20H10
Figure 3.16. Mean annual sediment yield for Lasa soil at 2 percent slope under corn production for all
climate change scenarios.
' See Footnote 3 for explanation of discrepancy in annual rainfall values resulting from intensity adjustments.
61
-------
7Q Q
1 ft n
"^ 1 A n -
4/1 1 d n
c
S i 7 n
"^ 1 n n -
o>
>- & n
c
0> c n
^ -i n
«
-> n
On
o
Q Lasa V-10
Q A * Lasa VO
*Lasa V20
Lcrdal V-10
9 OLcrdal VO
OLcrdal V20
2 5
Fie Id Slope, %
Figure 3.17. Mean annual sediment yield for the Lasa and Lerdal soil under corn production for three
climate change scenarios.
3^ O
ro
;= 25.0 -
in
01
c
i 20.0
01
= 15.0
c
01
E 10.0
73
01
1/1 5.0
On -
A
A
A
A
A
8
2 5
Field Slope, %
A Corn V-10 (110
A Corn VOM10
ACornV20)llO
SoyV-10-HlO
ASoyVO+110
ACn\-\i~T\ i 11 n
Figure 3.18. Mean annual sediment yield for Lasa soil under corn and soy production for three climate
change scenarios.
PART B: Managing Soil Loss under Climate Change
Scenario Development: PART B
A total of 189 model simulations were completed. Scenarios included 1 baseline climate scenario, 9
climate change scenarios, 1 land use scenario, and 18 management scenarios.
62
-------
Climate Change Scenarios
PART B used a subset of the climate change scenarios evaluated in PART A. The climate scenarios
consisted of a series of synthetic temperature and precipitation adjustments. Mean monthly temperatures
were increased by 0, 2 and 4°C. Mean monthly precipitation volumes were increased by 0 and 20 percent.
Precipitation was further adjusted to assess the impact of increasing intensity. Similar to PART A, this
was accomplished by first increasing the precipitation volume 20 percent and then increasing the intensity
by 10 percent. A total of 9 climate scenarios were assessed.
Land Use Scenarios
One land use scenario was considered, a field with an area of 30 x 30 meters with Lerdal soil at a five
percent uniform slope. Lerdal is a hydrologic group C soil meaning it has both a low infiltration rate and
water holding capacity. This soil series and slope were selected since it represents a "worst case" scenario
for a field under crop production in Blue Earth County, MN.
Management Scenarios
PAPT B was designed to assess the potential benefit of alternative tilling practices and grass and forest
filter strips for climate change adaptation. The corn spring chisel plow land management scenario from
PART A was selected as a baseline scenario. Two additional land management options were included,
corn no-till and corn fall mulch till. The land management scenario characteristics in terms of tilling,
planting and harvesting were fixed and predetermined by the WEPPCAT model (Table 3.19, 3.23, and
3.24). WEPPCAT provides the option of including a filter strip to assess potential sediment yield
reductions. Grass and forest filter strips included in the model simulations were 3, 6, and 9 m wide x 30 m
long. A baseline scenario with no filter strip was also included for baseline comparisons under each
climate change scenario and tilling practice. A total of 18 management scenarios were assessed.
Table 3.22. Corn fall mulch till management characteristics in WEPPCAT.
Date
4/25
5/5
5/10
5/10
6/5
10/15
11/1
Table
Date
5/10
5/10
10/15
Operation Type
Tillage
Tillage
Tillage
Plant-Annual
Tillage
Harvest
Tillage
Operation Name
Field cultivator, secondary tillage, after duckfoot points
Tandem Disk
Planter, double disk openers
Corn, Jefferson IA, High production 125 bu/acre
Cultivator, row, multiple sweeps per row
Corn, Jefferson IA, High production 125 bu/acre
Chisel plow, straight with spike pts
3.23. Corn no-till management characteristics in WEPPCAT.
Operation Type
Tillage
Plant-Annual
Harvest
Operation Name
Planter, no-till with fluted coulter
Corn, Jefferson IA, High production 125 bu/acre
Corn, Jefferson IA, High production 125 bu/acre
Endpoint Selection: PART B
Same as PART A.
Results: PART B
The model simulations provide a broad picture of the potential sediment yield associated with varying
degrees of climate change and land management options (Table 3.25). Generally, increases in
precipitation volume and intensity and temperature resulted in increased sediment yields under all
63
-------
management scenarios. Sediment yield decreased as filter strip width increased; however, there were
diminishing returns with sediment reduction as the filter strip width increased from three to nine meters
(Figure 3.19).
Table 3.24. Sediment yield (tonnes/ha/yr) resulting from corn production under all climate change, land
use, and management scenarios. Buffers named to reflect cover; NB=no buffer, GB = grass buffer, and
FB = forest buffer, and numerical value signifies width. Precipitation scenarios named to reflect changes
in volume and intensity; V = volume, I = intensity, numerical value reflects percent change from baseline.
Temp
Increase
0°C
2°C
4°C
Buffer
NB
GB3
GB6
GB9
FB3
FB6
FB9
NB
GB3
GB6
GB9
FB3
FB6
FB9
NB
GB3
GB6
GB9
FB3
FB6
FB9
Corn Fall Mulch
VO
8.1
4.7
3.6
2.7
4.3
2.9
2.2
9.4
5.2
3.6
2.9
4.5
3.1
2.2
11.7
6.1
4.0
2.9
5.4
3.4
2.5
V20
11.7
7.2
5.4
4.3
6.3
4.5
3.6
13.5
7.8
5.6
4.3
6.9
4.7
3.6
17.0
9.2
6.3
4.7
8.1
5.4
3.8
V20+
110
13.0
8.1
6.1
4.9
7.2
4.9
4.0
15.2
9.0
6.3
4.9
7.8
5.4
4.0
18.6
10.3
7.2
5.2
9.2
5.8
4.3
Corn No Till
VO
1.3
1.3
1.3
1.1
1.3
1.3
1.1
1.3
1.3
1.3
1.1
1.3
1.3
1.3
1.6
1.6
1.3
1.3
1.6
1.3
1.3
V20
2.0
1.8
1.8
1.8
1.8
1.8
1.8
2.0
2.0
2.0
1.8
2.0
2.0
1.8
2.2
2.2
2.0
2.0
2.2
2.0
2.0
V20+
110
2.0
2.0
2.0
1.8
2.0
2.0
2.0
2.2
2.2
2.2
2.0
2.2
2.0
2.0
2.5
2.5
2.2
2.0
2.2
2.2
2.2
Corn Spring Chisel
VO
11.2
6.3
4.5
3.4
5.4
3.6
2.7
13.0
6.7
4.7
3.4
5.8
3.8
2.7
15.5
7.6
4.9
3.6
6.5
4.0
2.7
V20
16.4
9.6
6.9
5.4
8.5
5.6
4.3
18.6
10.3
7.2
5.4
9.0
5.8
4.3
22.4
11.9
7.8
5.6
9.0
6.3
4.5
V20+
110
18.4
10.8
7.8
6.1
10.3
6.5
4.9
21.1
11.9
8.3
6.1
11.2
6.7
4.9
25.3
13.5
9.0
6.5
13.0
7.2
4.9
64
-------
14.0
-vo
-V20
-V20+I10
0.0
GB3 GB6
Filter Strip Scenario
GB9
Figure 3.19. Sediment yield (tonnes/ha/yr) under corn fall mulch till with a 3, 6, and 9 meter grass buffer.
Buffers named to reflect cover; NB=no buffer, GB = grass buffer, and FB = forest buffer, and numerical
value signifies width. Precipitation scenarios named to reflect changes in volume and intensity; V =
volume, I = intensity, numerical value reflects percent change from baseline.
The simulation results can be used to characterize sensitivity to climate change (Table 3.25). For example,
the simulation for a field under corn spring chisel and current climate conditions without a buffer resulted
in a sediment yield of 11.2 tons/ha/yr (Table 3.25). If climate change resulted in a 20 percent increase in
annual rainfall (V20 scenario) and a 2°C increase in temperature, the sediment yield would be 18.6
tons/ha/yr, a 66 percent increase over current yields. A land owner could also use the model simulations
to determine potential options for not only maintaining current sediment yield, but also to identify ways to
reduce sediment yield under altered precipitation regimes. As indicated in Table 3.26, certain land
management practices and/or filter strips could meet both of these management goals. If the land owner
wanted to maintain a sediment yield of 6 tons/ha/yr or less under the V20 precipitation scenario, a number
of options may exist. No-till for corn production was by far the superior management practice for
reducing soil yield under the V20 scenario. A landowner could also maintain current tillage practices and
install a 6 to 9 m forest buffer or 9 m grass buffer and reduce sediment yields below 6 tons/ha/yr.
Table 3.25. Sediment yield (tonnes/ha/yr) results from a 2°C increase and temperature and a 20 percent
increase in mean annual rainfall volume. Buffers named to reflect cover; NB=no buffer, GB = grass
buffer, and FB = forest buffer, and numerical value signifies width. Grayed areas signify tillage and filter
strip combinations producing 6 tons/ha/year or less of sediment.
Buffer
NB
GB3
GB6
GB9
FB3
FB6
FB9
Corn Fall Mulch Corn No Till Corn Spring Chisel
V20 V20 V20
13.5 2.0 18.6
7.8 2.0 10.3
5.6 2.0 7.2
4.3 1.8 5.4
6.9 2.0 9.0
4.7 2.0 5.8
3.6 1.8 4.3
65
-------
Summary
In this case study, WEPPCAT was used to explore soil erosion sensitivity to climate change by efficiently
modifying weather data for input into the WEPP model. WEPPCAT enables users to run an almost
unlimited number of land management-climate change scenario combinations to explore potential
sediment yields. The ability to modify temperature and precipitation volume and intensity data, as well as
explore the benefits of filter strips and alternate land management options were all utilized in this case
study. While the model outputs are not actually predicting future conditions, they provide a useful
mechanism for comparing sediment yield across a range of potential futures in order to assess sensitivity
and vulnerability.
The results from PART A indicated a relatively high degree of sensitivity of agriculture land in Blue
Earth County, MN to climate change. The sediment yields from fields with both Lasa and Lerdal soil
were almost double compared to the baseline yields under the most extreme climate change scenarios
evaluated in this study. The finding also indicated that crop type and slope play a significant role in
determining sediment yield under all climate change scenarios.
In PART B, sediment management options were explored in the context of climate change adaptation.
Sediment yield from a corn field under alternative land management and climate change scenarios was
modeled. This type of information can help identify appropriate management practices for adapting
agricultural land to climate change. The findings indicated that sediment yields could potentially be
prevented or even reduced under the most extreme climate changes if management practices are
employed.
66
-------
3.5. Streamflow and water quality sensitivity to changes in precipitation amount, frequency,
and intensity in the Tualatin River, OR, using BASINS CAT with HSPF
Case Study Overview
This case study used BASINS CAT with HSPF to explore watershed sensitivity under
changing precipitation regimes with respect to flow, TN, and TSS. The BASINS CAT
capability to modify precipitation patterns was employed to increase precipitation by 10 and
20 percent using the following climate scenarios:
• constant percent increase applied to all precipitation events (constant increase).
• increase applied to the largest 30 percent of precipitation events (intensity increase).
• increase in total number of annual precipitation events (frequency increase).
Temperature was increased by 2°C for all scenarios. Potential evapotranspiration was
recalculated with BASINS CAT to reflect this change.
Introduction
It is projected that precipitation will be directly impacted by changes in atmospheric circulation and
increases in water vapor and evaporation associated with warmer temperatures due to climate change
(IPCC, 2007). While many regions are expected to see an overall increase in precipitation, there is
significant uncertainty with respect to changes in local and regional rainfall patterns leading to this
increase. Watershed responses to different precipitation patterns are also uncertain and will depend on
land use and water management among other drivers.
In this case study, the impacts of alternate precipitation patterns are assessed using an existing HSPF
model of the Tualatin River in Oregon. BASINS CAT was used to develop the climate change scenarios
and assess the water quality and quantity endpoints. This study used the BASINS CAT capabilities for
making adjustments to all precipitation events, precipitation events within a user specified size class, and
adding new events to the precipitation baseline time series.
Location Description
The Tualatin River (HUC 17090010) drains 712 square miles in northwest Oregon, and is a tributary of
the Willamette River (Figure 3.20). Land use and land cover ranges from the densely populated areas of
southwest Portland, Hillsboro, Tigard and Beaverton to agricultural areas near Scholls, Gaston, Banks,
Mountaindale and North Plains to the forests of Oregon's Coast Range, Tualatin Mountains and
Chehalem Mountains. Most of the fast-growing urban population, approximately 500,000 residents,
resides on 15 percent of the watershed's area. About 35 percent of the watershed is used for agriculture,
and about 50 percent of the watershed is forested.
67
-------
Legend
— Reaches
]J Tualatin Basin
Figure 3.20. Tualatin River Watershed Location.
Water Model Setup
A pre-existing HSPF model of the Tualatin River was acquired from a previous modeling study that
included the entire Willamette River (Johnson et al., 2011). Model segmentation of the watershed was
based on intersections of land use, hydrologic soil group, and available NCDC weather stations. Soils
data were taken from the STATSGO soil survey and LULC was based on the 2001 National Land Cover
Database (NLCD). Four point source inputs were included in the model using data from the U.S. EPA's
Permit Compliance System. Meteorological data (precipitation, air temperature, and potential
evapotranspiration) were drawn from the BASINS4 Meteorological Database (USEPA 2007), a
consistent, quality-assured set of nationwide data records disaggregated to the hourly time step typically
used by HSPF. Meteorological data from three National Climatic Data Center (NCDC) weather stations
(350595 - Beaverton, 351222 - Buxton, 352997 - Forest Grove) across the watershed were used as input
to the model.
The baseline model data was for 1980 through 2005. A hydrology calibration period of 10/01/1995 to
09/30/2005 and validation period of 10/01/1985 to 09/30/1995 were used for the Tualatin stream gage at
the basin outlet. The water quality calibration and validation periods were 10/01/1991 to 9/30/1995 and
10/01/1986 to 9/30/1990, respectively. A brief summary of calibration and validation results are provided
in Tables 3.26 and 3.27.
68
-------
Table 3.26. Tualatin River model daily streamflow calibration/validation results. NSE: Nash-Sutcliffe
Efficiency coefficient.
NSE
E'
R2
Calibration
('95 -'05)
0.799
0.731
0.726
Validation
('85 - '95)
0.811
0.702
0.769
Table 3.27. Tualatin River HSPF model monthly water quality calibration/validation results. Relative
Percent Error is the average of observed-simulated/observed comparisons. Median Percent Error is the
median of observed-simulated comparisons/average of observed values.
Endpoint Statistic
TSS Load Relative Percent Error
TSS Concentration Median Percent Error
Total N Load Relative Percent Error
Total N Concentration Median Percent Error
Calibration Validation
('95 - '05) ('85 - '95)
3
-7.8
2
-16.8
5
10
-6
-19.2
Scenario Development
A total of 7 model simulations were completed. Scenarios included 1 baseline climate scenario and 6
climate change scenarios. No land use or management scenarios were included.
Climate Change Scenarios
Climate change scenarios were developed to fall within the ensemble range of projected mid-century
(2050s) temperature and precipitation changes for this region based on statistically downscaled data from
16 CMIP3 climate models acquired from the Climate Wizard web site (see Section 3.2 for more
information on these models). Projected mid-century changes in temperature for this region ranged from
approximately 1°C to 2.5°C, and projected changes in precipitation ranged from approximately -10 to +18
percent.
Six climate change scenarios were created. Each scenario included an increase of 2°C applied to all daily
temperature values in the baseline record. Potential evapotranspiration (PET) records were revised using
the BASINS CAT Penman-Montieth option to account for temperature changes. Annual precipitation
volume was increased by 10 and 20 percent using three different capabilities available in BASINS CAT:
adjusting the magnitude of all events in the record, adjusting the magnitude of specific events, and
randomly adding/deleting events to change the number of precipitation events in the record (Figure
3.21). The 10 and 20 percent increases in precipitation volume were applied to the baseline precipitation
records in the following manner:
• constant percent increase applied to all precipitation events (constant increase).
• increase applied to the largest 30 percent of precipitation events (intensity increase).
• increase in total number of annual precipitation events (frequency increase).
69
-------
Modify Existing Data
Modification Name: Preciplntensity
E xisting D ata to M odify: CO M PU T E D 0 R 350595 PR E C (and 2 more)
How to Modify:
| Multiply large/small events by a number
Percent Change in Volume
O Single Change (*) Multiple changes within specified range
Minimum
Maximum:
10
20
Increment: 10
Events
0 Vary values only in the following Events
Exceeding threshold 0
Change 30 % of volume
Allow gaps up to
Sum of values exceeding threshold 0
Total duration above [o_
Months /Years
Vary only in selected
Ok
Cancel
Figure 3.21. BASINS CAT window specifying increased precipitation in the top 30 percent of events.
A time-series plot for a small portion of the modeled precipitation record for each climate change scenario
is shown in Figure 3.22 as an example of how a 20 percent increase in precipitation is achieved using
each of the three different precipitation adjustment capabilities. The plot contains curves for each pattern
change method: a dashed line for the constant increase, a dotted line for intensity increase, and a light
solid line for frequency increase. Each of the three precipitation events in Figure 3.22 demonstrates
different aspects of the three methods. The first event (hour 16 - 17) is a new event added by the increase
in frequency method. The second event (hour 0-2) shows the changes applied for a 2-hour event that is
in the top 30 percent of the original record. The lowest values (light solid line) represent the increased
frequency method, but this event is actually from the original record. The dashed line represents the
constant increase method and is thus 20 percent higher than the original storm volume. The dotted line
shows the intensity increase applied to this large event, one that falls within the largest 30 percent of
events. The third event (hour 7- 8) is a small event that is increased only by the constant change method.
The intensity modification does not apply since it does not fall within the largest 30 percent of events.
70
-------
Precipitation (inches)
ooooooo<
'-*• to Cl tt. 01 01 '•** C
Increase all events
ase large events
.__
-
i — i
14:00 16:00 18:00 20:00
1986 Oct28
22:00 0:00 2:00
Time
4:00 6:00 8:00
1986Oct29
10:00
Figure 3.22. Example of precipitation event distribution for the 3 climate change scenarios.
Land-Use Scenarios
Land use scenarios were not evaluated in this case study. While it is unlikely land use in the Tualatin
River watershed will not change in the future, land use was held constant in order to focus on potential
climate change impacts only.
Management Scenarios
Management scenarios were not specifically evaluated in this case study. BMPs or other management
practices may have been included in the original HSPF model, but no adjustments were made in order to
focus on potential climate change impacts.
Endpoint Selection
Two water quality constituents, TSS and TN, and one water quantity constituent, mean annual
streamflow, were selected as the analysis endpoints.
Results
Results for the selected endpoints from all model runs, including values for percent and absolute change
in annual precipitation, and the maximum precipitation event during the simulation are presented in Table
3.28. The two annual precipitation volume increases, 10 and 20 percent, are consistent for the three
adjustment methods, but the maximum precipitation event depths vary. Temperature adjustments are not
included in the table since a constant 2°C temperature increase was included in all model runs.
Table 3.28. Precipitation, streamflow, and loadings of TN and TSS for all climate scenarios.
Precipitation Annual Max Mean Annual Annual
Scenario volume precipitation, Precipitation Streamflow, load load
71
-------
Baseline
Frequency
Constant
Intensity
Frequency
Constant
Intensity
increase,% i
0
10
10
10
20
20
20
nm
,014
,115
,115
,115
,217
,217
,217
Event, mm
20.2
20.2
22.2
26.9
20.2
24.2
33.6
cms
34.5
38.4
39.3
38.6
44.2
45.3
43.9
TN,
kg/ha
17.7
19.3
19.5
18.8
21.5
21.6
20.2
TSS,
tonnes/
ha
0.50
0.57
0.66
0.78
0.67
0.86
1.19
The model results show a significant watershed response with all three endpoints increasing under all
climate change scenarios. The results also indicate that the different precipitation patterns have varying
degrees of impact on the endpoints. Streamflow shows the largest response to the constant increase,
followed by the frequency increase, and then the intensity increase. Similarly, TN is less impacted by the
intensity increase with more substantial, and very similar, responses from the constant increase and the
frequency increase.
TSS was found to be highly sensitive to the climate change scenarios, but responded differently than TN
and Streamflow (Figure 3.23). The frequency increases yielded a 14 and 35 percent increase in TSS for
the 10 and 20 percent scenarios, respectively. The responses to the constant increases are more substantial
(33 and 74 percent, respectively) and the increases in TSS in response to the intensity increases are nearly
double those of the constant increase (57 and 140 percent, respectively). These results indicate that all
precipitation scenarios increase TSS loads, however, increasing event intensity has the greatest potential
impact.
re
0
1/1
1/1
"re
3
c
c
OJ
OD
re
c
U
•M
C
01
11
D-
IfiO
140
i7n
i nn
80 -
60 -
40 -
?n
0 -
* Frequency
A • Constant
i A Intensity
+
1 '
10 20
Percent Increase in Average Annual Rainfall
Figure 3.23. Percent change in TSS loads relative to baseline for all climate change scenarios.
Summary
This case study demonstrated three different ways BASINS CAT can adjust precipitation records to
represent potential changes in climate. Mean annual volume increases of 10 and 20 percent were applied
to baseline precipitation records using three different methods: constant increase of all events, increase of
72
-------
event frequency, and increase in event intensity of the largest 30 percent of events. Temperature was also
adjusted to reflect projected changes.
This study illustrates the sensitivity of a watershed to changing precipitation patterns and highlights the
variation in response to these different adjustments. The results indicate that even if annual precipitation
volume remains constant, how and when it occurs makes a difference in watershed response. Of particular
note was the response of TSS loads to different levels of event intensification. Given the dramatic
response in TSS loading to precipitation intensity, additional scenarios could be run to further explore
these relationships including more detailed analysis of TSS response to changes in specific events, and the
effects of adding seasonal variability of changes (e.g., monthly)(see the Section 3.2). The analysis of
additional endpoints, either in the form of new constituents (e.g. total P) or hydrologic response (e.g. peak
flow value) may also provide further insights to watershed sensitivity to changing precipitation patterns.
73
-------
3.6. Streamflow sensitivity to dry weather events in Sespe Creek, CA, using BASINS CAT with
HSPF
Case Study Overview:
This case study used BASINS CAT with an HSPF watershed model of Sespe Creek, CA, to
examine the sensitivity of Streamflow to changes in magnitude and duration of dry weather
events (meteorological drought). A dry period in the historical observation record was
selected to create a series of climate scenarios. BASINS CAT was used to develop climate
scenarios that increased average annual temperatures by 2°C and altered precipitation in the
following way:
• Increased the magnitude of a historical dry period (1959-1961) by adjusting annual
precipitation volume by 0, -10, and -20 percent
• Extended the duration of the historical dry period to include the years 1959-1964 by
reducing precipitation in wet years that followed the 1959-1961 dry period
• Increased the magnitude and extend the duration of the dry period to include the
years 1959-1964
Introduction
Managing the impacts of drought on water supply is an important goal of watershed management.
Climate change in many parts of the nation could result in warmer, dryer conditions leading to increased
drought risk. Responding to this challenge will require an improved understanding of the implications of
climate change for drought, and the development of management strategies for reducing the impacts of
drought.
In this case study, BASINS CAT and an HSPF watershed model were used to assess the sensitivity of
water supply to increased severity of dry weather events (meteorological drought) in Sespe Creek, CA. In
the context of this case study, we defined drought severity to include the magnitude and duration of
precipitation deficit. The simulation endpoints evaluated include mean annual and low-flow Streamflow
statistics.
Location Description
The Sespe Creek watershed (Figure 3.24) covers an area of approximately 700 km2 in southwestern
California, with its headwaters near the Ventura-Santa Barbara county line. It flows east through
relatively pristine, mountainous and remote terrain of the Los Padres National Forest. Then it bends
south through a bedrock-confining gorge before widening out into a broad alluvial fan near the City of
Fillmore until its confluence with the Santa Clara River. Elevations range from more than 2,000 m in the
headwaters and upper reaches to about 120 m above mean sea level at the mouth. The typical hydrologic
pattern includes peak flows in late winter/early spring in response to winter rains and spring snowmelt
followed by very dry conditions in summer and fall. The watershed is primarily undeveloped, with
dominant land uses being forest and shrub land (Table 3.29).
74
-------
I Sespe Creek Watershed
3 6 12Mi!e
Figure 3.24. Location of the Sespe Creek watershed.
Table 3.29. Land use summary for Sespe Creek watershed.
Land Use
Portion of Watershed, %
Forest 14
Shrub 80
Open/Grassland 3
Agriculture 2
Developed 1
Model Setup
The Sespe Creek model was extracted from a larger HSPF model of the Santa Clara River (AQUA
TERRA Consultants, 2009). This model was developed as part of the Santa Clara River Watershed
Management effort, a joint effort by Ventura County Watershed Protection District, Los Angeles County
Department of Public Works and U.S. Army Corp of Engineers Los Angeles District. As part of this
effort, the Sespe Creek portion of the model was calibrated to historic streamflow data for the period of
10/1/1997 through 9/30/2005, and validated for 10/1/1987 through 9/30/1996. Statistical results of the
calibration/validation are shown in Table 3.30. For this case study, the model was run for the period 1952
- 2001 in order to include a wide range of hydrologic conditions and at least one significant period of low
flow.
75
-------
Table 3.30. Streamflow volume (normalized by watershed area) calibration and validation results for
Sespe Creek model.
Gage
Streamflow Simulated
(cms) Observed
Volume Error (%)
Daily R2
T)
Monthly 2
K
Daily Peak Difference (%)
Wheeler Springs
Calibration
(10/1/02 -
9/30/05)
24.9
22.6
9.5
0.95
0.91
0.98
0.97
4.7
Validation
(10/1/86 -
9/30/96)
18.3
18.5
-1
0.91
0.82
0.98
0.96
3.8
Fillmore
Calibration
(10/1/96 -
9/30/05)
26.2
27.7
-6.1
0.96
0.92
0.99
0.98
-5.5
Validation
(10/1/93 -
9/30/96)
26.7
24.9
7
0.92
0.84
0.97
0.94
9.6
The original Sespe Creek model made use of observed pan evaporation data as the potential
evapotranspiration (PET) input term required by HSPF. It was necessary to replace the observed pan data
with computed PET that could be regenerated by BASINS CAT for each future simulation in order to better
represent the impact of future temperature adjustments on PET. This was accomplished using the Penman-
Monteith option for estimating PET in BASINS CAT. The baseline model was run using the PET generated
by BASINS CAT in place of the observed pan evaporation data. Using BASINS analysis tools, results from
the modified model were compared to the original. Differences in total and mean Streamflow volumes were
less than 1%, differences in the lowest 10% of Streamflow was 3%, and the difference in the highest 1% of
Streamflow was 2%. The original model calibration was thus considered acceptable for use in the case study.
Scenario Development
A total of 6 model simulations were completed. Scenarios included 1 baseline climate scenario and 5
climate change scenarios. No land use or management scenarios were included.
Climate Change Scenarios
Climate change scenarios were developed to fall within the ensemble range of projected mid-century
(2050s) temperature and precipitation changes for this region based on statistically downscaled data from
16 CMIP3 climate models acquired from the Climate Wizard web site (see Section 3.2 for more
information on these models). Projected mid-century changes in temperature for this region ranged from
approximately 1°C to 3°C, and projected changes in precipitation ranged from approximately -40 to +35
percent.
A total of 5 climate change scenarios were created to represent increased drought severity including
increased magnitude and duration of precipitation deficit. Each climate change scenario included an
increase of 2°C applied to all daily temperature values in the baseline record from 1952-2001 to represent
projected changes in temperature for this region. Potential evapotranspiration (PET) records were revised
to account for temperature changes using the Penman-Montieth method. Analysis of annual observed
Streamflow in Sespe Creek from 1950 to present day showed that the period from 1959 - 1961
represented a prolonged period of low flow. The first 3 climate change scenarios represented changes in
the magnitude of dry weather events and were created as follows:
• Decreased precipitation by 20 percent during the observed low flow period 1959-1961, identified
hereafter as "Precip -20"
76
-------
• Decreased precipitation by 10 percent during the observed low flow period 1959-1961, identified
hereafter as "Precip -10"
• Maintained historic precipitation during the observed low flow period 1959-1961 (thus
representing the impact of the 2°C temperature change only), identified hereafter as "Precip 0"
Figure 3.25 shows the BASINS CAT window used to create the precipitation adjustments in the three
climate change scenarios. The three fields at the top define which input records will be adjusted and by
what method (i.e. "Multiply Existing Values ..."). The BASINS CAT capability for creating multiple
changes within a user specified range was made to produce the three adjustments. The bottom frame
defines the time of year (Months/Years) during which the adjustments were applied (e.g., 1959 - 1961).
BASINS CAT was then used to combine these three precipitation adjustments with the temperature and
PET adjustments described above to create three scenarios of varying drought severity.
1. Mo dif y Existi ng Data Q (Ml ®
Modification Name: HE
|
Existing Data to Modify: OBSERVED VC1 52 PR EC (and 4 more) [ View ]
H ow to M odify: M ultiply E xisting Values by a N umber (eg Precipitation)
Number to multiply existing c
O Single Change 0 Mu
Minimum 0.8
Maximum: 1
Increment: 0.1
Events
O Vary values only in the fo
Exceeding threshold
Allow gaps up to
Sunn of values exceedin
Total duration above
Months /Years
[**! Vary only in selected C
Itiple changes within specified range
multiplication factor
multiplication factor
llowing Events
0
0
3 threshold 0
0
Calendar Year v
1950 1955 B1965 1970 1975 1930 1385
1951 1956 |1366 1971 1976 1981 1986
1952 1957 1962 1967 1972 1977 1982 1987
1953 1958 1963 1968 1973 1978 1983 1988
1954 EE!s^Hl964 1969 1974 1979 1984 1989
< I"! I Li
All
None
Ok Cancel
Figure 3.25. BASINS CAT window defining adjustments to precipitation.
77
-------
The fourth scenario represented increased drought duration (hereafter referred to as "Duration"). The dry
period of 1959 - 1961 was followed by two years of more typical precipitation. This was followed by
another dry year in 1964. The original 3-year dry period was extended by adjusting the years 1962 - 1963
to follow the trend of the years directly preceding and following them. This was done by applying a
constant multiplier to all events in 1962 and 1963 such that their mean annual precipitation matched the
mean annual totals of 1961 and 1964. This adjustment, combined with the temperature and PET
adjustments described above, led to a simulated drought period of six years (1959-1964). The fifth and
final scenario represented both increased drought severity and increased duration (hereafter referred to as
"Duration/Severity"). This scenario was created using the same adjustments as in the fourth scenario to
increase duration, together with a 10% precipitation decrease applied to all six years of the extended
drought period to represent increased drought severity.
Land Use Scenarios
Land use scenarios were not evaluated in this case study. While it is unlikely that land use in the Sespe
Creek watershed will not change in the future, land use and land cover was held constant in order to focus
on potential climate change impacts.
Management Scenarios
Management scenarios were not specifically evaluated in this case study. BMPs or other management
practices may have been included in the original HSPF model, but no adjustments were made in order to
focus on potential climate change impacts.
Endpoint Selection
The simulation endpoints considered in this case study were mean annual streamflow and mean annual 7-
day low flow. Additionally, mean monthly streamflow values from each scenario were plotted for
comparison. For some analyses, endpoint values were computed only for the scenario's period of
intensified drought (1959 - 1961 or 1959 - 1964) to more clearly understand the scenario's impact.
Results
Endpoint values for all model simulations are presented in Tables 3.31 and 3.32. Table 3.31 shows results
for the drought severity scenarios, where temperature increase was applied to the entire run and
precipitation adjustments were applied only to the period 1959 - 1961. Table 3.32 shows results for
scenarios where the drought was intensified by increasing the duration (1959 - 1964) and then reducing
rainfall during this period by 10 percent. For both tables, endpoint results are reported only for their
respective drought periods as the mean annual endpoints were only minimally impacted over the length of
the entire simulation period (1952 - 2001).
Results from Table 3.31 show that changing only the temperature ("Precip 0" scenario) caused only a
slight decrease (roughly 5 percent) in the two endpoints. However, combining the temperature change
with decreases in precipitation during the dry period of 1959 - 1961 had a significant impact on the
endpoints. Decreasing precipitation by 10 percent ("Precip -10" scenario) led to decreases from the
baseline in mean annual streamflow and mean annual 7-day low flow of 38 and 30 percent, respectively.
For the 20 percent precipitation decrease ("Precip -20" scenario), decreases from the baseline were 62
percent for mean annual streamflow and 39 percent for mean annual 7-day low flow.
78
-------
Table 3.31. Simulation results for the Precip 0, Precip -10 and Precip -20 scenarios as applied to historic
period of low flow (1959-1961) .
Scenario
Baseline
Precip 0
Precip -10
Precip -20
Change in
Temp
°C
0
2
2
2
Change in
Precipitation
%
0
0
-10
-20
Mean Annual
Streamflow
(1959-1961)
cms
12.6
12.1
7.83
4.85
Mean Annual 7-Day
Low Flow (1959-
1961)
cms
0.612
0.584
0.427
0.374
Results in Table 3.32 also show dramatic change in endpoint values in response to the Duration and
Duration/Severity scenarios. Adjusting the precipitation record for the years 1962 - 1963 (normal rainfall)
to match the mean of years of 1961 and 1964 (low rainfall) and extending the dry period to 1959 to 1964
led to a significant decrease from baseline conditions for all simulated scenarios. Mean annual streamflow
in the Duration scenario decreased by 61 percent and mean annual 7-day low flow decreased by 43
percent for the extended drought period. Furthering the severity of the extended drought by applying a
0.9 precipitation multiplier (Duration/Severity scenario), led to an additional 13 percent decrease in mean
annual streamflow and an additional 12 percent decrease in mean annual 7-day low flow.
Table 3.32. Simulation results for the Duration and Duration/Severity scenarios as applied to the
extended period of low flow (1959-1964).
Scenario
Baseline
Duration
Duration and Severity
Change
in
Temp,
°C
0
2
2
Change
in Precipitation,
%
0
0
-10
Mean
Annual
Streamflow
(1959-1964) cms
48.6
19.0
12.8
Mean
Annual 7-Day
Low Flow
(1959-1964)
cms
1.41
0.80
0.63
Using the BASINS CAT option to save input and output files for all simulations, BASINS analysis tools
were used to generate a plot of mean monthly streamflow for each scenario (Figure 3.26). This plot
provides further insight into the results in Tables 3.31 and 3.32. While the Duration/Severity scenario
yielded the greatest decrease from baseline values, it is notable that endpoint values from this scenario
(final row of Table 3.32) were still higher than baseline values for the original drought period (first row of
Table 3.31). . The Duration and Duration/Severity scenarios were developed by decreasing all rainfall
events in 1962 - 1963 by the same amount, the early 1962 precipitation still remained substantial and led
to significantly increased streamflow (see 'Duration' and 'Duration/Severity' curves in plot). Thus, the
mean annual streamflow and mean annual stream flow and 7-day low flow values for these scenarios
remained at levels above the original baseline drought.
79
-------
100
E
o
Precip -10
Precip -20
Duration
Duration/Severity
0.01 -
1958 1959 1960 1961 1962 1963 1964
AVERAGE MONTHLY FLOW at RCH728
1965
1966
Figure 3.26. Mean monthly streamflow during drought period for all scenarios.
Summary
This case study illustrates the sensitivity of streamflow to potential increases in drought severity resulting
from climate change. A baseline low flow period was selected from the historic streamflow record. This
drought was then intensified in 3 different ways. First, the severity of the drought was increased by
increasing the mean annual temperature and decreasing annual precipitation. The drought duration was
then lengthened by reducing precipitation in two consecutive, relatively wet years occurring mid-way
within the selected period of drought. Finally, these two adjustments were combined by decreasing
precipitation during the entire, extended drought period. The endpoints of mean annual streamflow and 7-
day low flow responded as expected with significant decreases from baseline values for all simulations.
Several BASINS CAT features were used to create the intensified drought scenarios. First, the ability to
modify specific time span subsets allowed for precipitation changes to be applied only during drought
periods. Second, the capability for creating multiple changes within a user specified range was used to
decrease precipitation. Finally, the ability to combine adjustments was used to create the scenario where
both drought duration and severity were increased. BASINS CAT was used to assess changes in mean
annual streamflow and 7-day low flow in this study, but the tool has the ability to generate and report any
desired duration-frequency event (7Q10, 100-year flood) or any n-day high or low flow time series
percentile. BASINS CAT can also report such values for specific time periods.
80
-------
3.7. Streamflow and water quality relative sensitivity to climate change versus impervious
ground cover in the Western Branch of the Patuxent River, MD, using BASINS CAT with
HSPF
Cast Study Overview:
This study evaluated the role of impervious cover in managing stormwater under increased
precipitation volume and event intensity using BASINS CAT with an HSPF model of the
Western Branch of the Patuxent River, MD. The sensitivity of stormwater runoff generation
and pollutant loads, specifically TSS, was explored through a combinations of the following
land use and climate change scenarios::
• watershed impervious cover of 8.6, 15, and 25 percent
• increased precipitation volume of all events by 0, 10 and 20 percent
• increased event intensity of 70th percentile and greater events by 0, 10 ,and 20
percent
Introduction
Urban and suburban development of watersheds results in increased impervious cover in the form of
houses, parking lots, roads, and sidewalks. Impervious cover disrupts aspects of the pre-development
hydrologic conditions which affect the health and integrity of local waterways. Many of the effects are
interrelated and often difficult to quantify, but two of the most significant causes of the impairment of
urban streams are increased stormwater runoff and runoff pollution.
Changes in local and regional climate may increase the frequency and intensity of heavy storm events
leading to substantial increases in stormwater runoff, pollutant loads, and flooding. There is also the
potential for synergy among climate and land use changes, exacerbating the impact of urbanization,
leading to increases in the amount of runoff and water quality impacts on surface water. An investigation
of the relationships between climate change and urbanization can provide a simple, heuristic
understanding of how reductions in impervious cover could be used to compensate for increased
stormwater runoff associated with climate change (Pyke et al., 2011).
BASINS CAT was used with an HSPF model of the Western Branch of the Patuxent River, MD to assess
the relative sensitivity of stormwater to changes in precipitation volume, event intensity, and impervious
cover. The BASINS CAT capability for creating multiple changes within a user specified range was
used to create an array of climate scenarios for use as model inputs. The model simulations were used to
develop a simple heuristic model for exploring land use-based climate change adaptation options.
Location Description
The Western Branch of the Patuxent River (HUC 02060006) drains an area of 230 km2 east of
Washington D.C., and is a tributary of the Patuxent River and Chesapeake Bay (Figure 3.27). Land use is
mixed, consisting of agricultural, forest, barren, wetlands and urban land uses (Table 3.33).
81
-------
New York
Western Branch, Patuxent River
Reaches
Figure 3.27. The Western Branch of the Patuxent River watershed and its location with the Chesapeake
Bay watershed.
Table 3.33. Land use summary for Western Branch of the Patuxent River watershed.
Land Use Portion of Watershed, %
Forest 39
Urban/Developed 34
Agricultural 25
Wetland 2
Barren < 1
Model Setup
The Western Branch of the Patuxent HSPF model used in this case study is based upon a model of the
Patuxent River watershed developed during the early 1990s for the USGS and the state of Maryland
(AQUA TERRA Consultants, 1994) and subsequently modified for other projects. For this project the
NLCD 2001 land cover was used, replacing the GIRAS land use land cover data from BASINS used in
earlier versions of the model. The period from 10/1/1985 through 9/30/2005 was chosen for the
simulation based on available meteorological data.
82
-------
Model calibration and validation checks were performed since this case study required some
modifications to the original HSPF model. The calibration and validation efforts for this case study were
not as extensive as in the original full-modeling study given the scope of this project; they were
completed to check that the case study model would yield reasonable results. The calibration period was
kept the same as the earlier model, 10/1/1985 through 9/30/1988. The USGS gage on the West Branch
has a gap in observed data in the early 1990s so that period could not be used for calibration/validation
checks. Accordingly, available streamflow and meteorological datasets for the time period of 10/1/1995 -
9/30/2005 were used for model validation.
The calibration checks for streamflow show monthly R2 values of 0.74 for the calibration period and 0.81
for the validation period (Table 3.34). The hydrology simulation for the validation period appears better
than for the calibration period, which is likely a factor of the validation period being considerably longer.
The overall flow balances are very good, with errors in total volume less than 1 percent, and the storm
peaks are well simulated, with errors less than 6 percent for calibration and nearly 2 percent for
validation.
Table 3.34. Selected Western Branch hydrology model calibration/validation statistics. NSE: Nash-
Sutcliffe Efficiency coefficient
Daily R2
Daily NSE
Monthly R2
Monthly NSE
% Error in Total Volume
% Error in Storm Peaks
Calibration
0.50
0.47
0.74
0.73
-0.9
-5.8
Validation
0.56
0.52
0.81
0.81
0.9
2.1
Limited data was available for calibrating TSS on the Western Branch, but a time series plot of the
simulated and observed TSS concentrations shows that the simulation captures the overall range and
distribution of TSS concentrations at the sampling location (Figure 3.28). While these checks should not
be construed as a complete calibration, the model was deemed appropriate for representing the relative
change across various model input scenarios.
83
-------
3.000
1995
1996
Time
Figure 3.28. Measured and simulated TSS concentrations (mg/L) in the Western Branch of the Patuxent
River watershed.
Scenario Development
A total of 9 model simulations were completed. Scenarios included 6 climate change scenarios and 3 land
use scenarios. No management scenarios were included.
Climate Change Scenarios
Climate change scenarios for this case study represented changes to precipitation only. No temperature
adjustments were made. Climate change scenarios were developed using an ensemble of 16 CMIP3
statistically downscaled climate models presented in Climate Wizard (See Section 3.2 for more
information on these models). A plausible end-of-century range of climate change for this region was
determined to be 0 to 20 percent increase in mean annual precipitation. The scenarios were developed to
explore various changes in precipitation patterns. Three scenarios reflected changes in precipitation
volume:
• Increase event values by 0% for all years on record
• Increase event values by 10% for all years on record
• Increase event values by 20% for all years on record
Three scenarios reflected changes in event intensity, where the proportion of annual precipitation
occurring in events above the 70th percentile was increased while events below this threshold were
decreased to create no net change in annual volume:
• Increase volume of events above the 70th percentile event by 0%
84
-------
• Increase volume of events above the 70th percentile event by 10% and decrease volume of events
below threshold event to maintain mean annual precipitation
• Increase volume of events above the 70th percentile event by 20% and decrease volume of events
below threshold event to maintain mean annual precipitation
The creation of these scenarios was facilitated by the BASINS CAT capability for combining multiple
adjustments to meteorological time series to create complex scenarios. Figure 3.29 shows the selection of
the two adjustments used to develop the 20 percent increase in events above the 70th percentile while
maintaining the same total annual precipitation. A total of 6 climate scenarios were developed.
,__ Climate Assessment Tool 2.0 - BASINSCAT_HSPF_Patuxent
File Edit Options Help
Model
Climate Data
Assessment Endpoints Results Pivot Table
Base Model
New Model
G:\CAT\CS5 - PatuxentVWest Branch\CalibratedModel\WestBran.uci
Modified
Add
Remove
Edit
Copy
View
Prepared
PrecipMult Multiply from 1.1 to 1.2 step 0.1
Preciplnt Intensify 10
Preciplnt Multiply 0.909
Preciplnt Intensify 20
9 Preciplnt Multiply 0.833
Total iterations selected = 1 (0:00)
Figure 3.29. Climate change scenario specifications in BASINS CAT.
Land Use Scenarios
Three land use scenarios were developed to simulate current watershed impervious cover and two
potential impervious cover futures:
• Current percent impervious, 8.6 percent overall
• Increase the overall percent imperviousness to 15 percent
• Increase the overall percent imperviousness to 25 percent
The increases in impervious cover were obtained through a proportional decrease in each pervious land
use category. These percent increases in impervious land were obtained by shifting land use primarily
from forest and agriculture to urban, as well as an increase in the amount of urban land that is considered
impervious (as increased urban density). The choice of 15 and 25 percent imperviousness for the future
85
-------
scenarios is not based on local information, but falls within the range for moderate to highly developed
watersheds in the mid-Atlantic region (USEPA, 2009b). Table 3.35 shows the percentages of each land
use category for each of the three land use scenarios.
Table 3.35. Summary of W.B. Patuxent land use composition by category for each land use change
scenario.
Scenario Scenario Scenario
Land Use 8.6% Impervious 15% Impervious 25% Impervious
Forest 39 36 32
Urban/Developed 34 39 46
Agricultural 25 23 20
Wetland 222
Barren <1 <_1 <_1
Management Scenarios
Management scenarios were not evaluated in this case study.
Endpoint Selection
The endpoints for this study consisted of average annual streamflow and mean annual TSS loads. Given
the goal of assessing the relative sensitivity of stormwater to changes in precipitation volume, event
intensity, and site impervious cover, a focus on streamflow is appropriate along with TSS loads since both
are generally impacted as a result of these factors.
Results
Model simulation results for mean annual streamflow and TSS loads are shown in Table 3.36. Percent
changes are expressed relative to the baseline conditions: 0 percent increase in precipitation volume, 0
percent increase in precipitation intensity, and 8.6 percent impervious cover. Simulations indicated that
streamflow in the Western Branch watershed is the most sensitive to increases in precipitation volume. A
20 percent increase in overall precipitation volume leads to a 46.3 percent increase in mean annual
streamflow. Increases in impervious cover also resulted in significant changes to mean annual
streamflow. Increases in precipitation intensity did not have nearly as large an effect on mean annual
streamflow as did increases in precipitation volume and impervious cover.
These simulations indicated that TSS loads in the Western Branch watershed are the most sensitive to
increases in precipitation volume. A 20 percent increase in overall precipitation volume led to a 62
percent increase in annual TSS load. A major increase in TSS load was also shown resulting from
increases in precipitation intensity. Results also indicated that increases to impervious cover actually
decreased the annual TSS loading.
Table 3.36. Annual streamflow and TSS load characteristics for the Western Branch of the Patuxent
scenarios
Scenario
Precipitation
Volume
Increase,
%
0
10
20
Mean Annual
Streamflow,
cms
2.8
3.4
4.1
Change in Mean
Annual
Streamflow, %
22
46
Mean Annual
TSS Load,
tonnes/ha
0.06
0.07
0.09
Change in
Mean Annual
TSS Load (%)
31
62
86
-------
Precipitation
Intensity
Impervious
Cover
0
10
20
8.6
15
25
2.8
2.9
3.0
2.8
3.1
3.6
3
6
11
28
0.06
0.07
0.08
0.06
0.05
0.05
21
40
-4
-10
Figure 3.30 shows the changes in stormwater runoff associated with changes in impervious cover
(holding precipitation volume and intensity constant), precipitation volume (holding impervious cover
and precipitation intensity constant), and precipitation intensity (holding impervious cover and
precipitation volume constant). These results, though based on limited data, suggest that when expressed
on a constant percent basis, mean annual stormwater runoff is most sensitive to changes in precipitation
volume, followed by changes in impervious cover and precipitation intensity.
50
40 •
•Precipitation volume
•Precipitation intensity
Impervious cover
5 10 15 20
Change in Impervious Cover, Volume, and Intensity (%)
25
Figure 3.30. Simulated sensitivity of stormwater runoff volume to changes in impervious cover,
precipitation volume, and precipitation intensity.
Similarly, Figure 3.31 shows the changes in TSS load associated with changes in impervious cover,
precipitation volume, and precipitation intensity. The results suggested that when expressed on a constant
percent basis mean annual TSS loads are most sensitive to changes in precipitation volume followed by
changes in precipitation intensity. Mean annual TSS load is inversely related to changes in impervious
cover. It should be noted that TSS loading from developed land is complex, and the relationship seen here
is not universal. The observed inverse relationship in these simulations likely resulted from the
conversion of primarily agricultural land to urban/suburban land. The conversion of forested land to
urban/suburban is likely to increase TSS loads. Moreover, newly developed land with significant new
construction and exposed soil can result in equal or greater loading than agricultural land.
87
-------
Precipitation volume
Precipitation intensity
Impervious cover
Change in Impervious Cover, Volume, and Intensity (%)
Figure 3.31. Simulated sensitivity of TSS load to changes in impervious cover, precipitation volume, and
precipitation intensity.
Summary
This study assessed the relative sensitivity of streamflow and TSS pollution to changes in precipitation
volume, event intensity and site impervious cover for a mixed-use watershed in Maryland. The primary
BASINS CAT feature demonstrated in this study was the ability to synthetically adjust climate inputs for
watershed models. This allowed an array of climate change scenarios to be automatically generated and
used as input for model runs. This case study also demonstrates the approach of combining climate
change scenarios with land use change scenarios. While the climate change and land use scenarios applied
were relatively simple, they were effective in illustrating some basic points.
The results indicated that mean annual streamflow in the Western Branch watershed is more sensitive to
changes in precipitation volume and impervious cover than they are to event intensity. In addition, TSS
loads in this watershed are more sensitive to annual precipitation volume and event intensity versus
impervious cover. While the most extreme changes yielded dramatic changes in endpoint results, even
relatively minor changes had notable impacts on either mean annual streamflow or annual TSS load, or
both. It should also be noted that while increasing impervious cover leads to higher mean annual
streamflow, it may actually lead to decreased TSS loads, likely due to agricultural land being converted to
urban uses.
This case study illustrates the potential synergy of climate change and urbanization with respect to
stormwater runoff and water quality impacts. Management practices such as low-impact development
that reduce impervious cover could be used to compensate for increased stormwater runoff associated
with climate change. In short, this study illustrates an important concept for water planners: that improved
development strategies have the potential to reduce or offset the effects of climate change.
88
-------
3.8. Limitations of Case Study Simulations
The case studies in this report illustrate selected capabilities of the BASINS and WEPP climate
assessment tools for conducting scenario-based assessment of watershed response to changes in climate,
land-use, and management practices. The scientific approach supported by these tools, i.e., scenario
analysis, can be useful for understanding system behaviour, identifying vulnerabilities, and evaluating the
effectiveness of management responses to inform management decision making. The tools presented in
this report, however, are just one step forward in building our capacity for understanding and responding
to climate change. Application of hydrologic models in this way has limitations, many of which are not
well understood (Ghosh 2010; Ludwig, 2009; Najafi, 2011; Vaze et al., 2010). Further study is required to
better assess, refine, and develop our current modeling capabilities. The following discussion briefly
identifies several issues, current limitations, and future needs associated with the use of hydrologic
models for climate change impacts assessment.
Use of a calibrated hydrologic model to conduct analyses assumes that change scenarios do not alter
watershed behavior in a way that invalidates the model parameterization achieved through calibration.
This issue exists in any modeling analysis (Donigian 2002, 2003), but is of particular concern when
considering climate change scenarios (Vaze et al., 2010). In many cases, climate change scenarios will
fall outside the range of historical observations used to calibrate the model. It is reasonable to assume that
at some point, large changes imposed by scenarios will affect model calibration. It is difficult to know
where this point is, however, or what the implications are for results. BASINS CAT and WEPPCAT, for
the most part, impose no constraints on the type and magnitude of climate changes made. Users must
therefore be cautious to consider the validity of model simulations, particularly when assessing change
scenarios falling outside the range of observed climatic variability.
At the decadal scale, climate change may alter groundwater storage and recharge potentially impacting
streamflow and water quality. The HSPF and SWAT models contain simple representations of the
shallow groundwater system, including percolation recharge, storage, discharge to streams, losses to deep
aquifers, and loading of pollutants such as nitrate. While these representations are adequate for the
simulation of surface water hydrology in most watersheds, the models do not provide a complete
representation of groundwater pathways because exchanges with deeper aquifers are not explicitly
simulated. Where these exchanges with deeper aquifers are represented in the models they are typically
held constant and their sensitivity to climate change is not simulated. Thus, a complete picture of any
long-term trends attributable to future climate may not be fully represented by a given watershed model.
Future research is needed to belter understand the long-term impacts of climate change on groundwater,
and to better represent these effects in watershed models.
Evapotranspiration (ET) is a major component of the water budget that is directly sensitive to climate. ET
is also strongly influenced by land cover, which is in turn influenced by climate. Changes in ET have a
significant influence on the occurrence, distribution, and movement of water including soil moisture,
groundwater recharge, and streamflow. The method used to calculate ET, or more commonly the
reference potential evapotranspiration (PET), is thus a key process in simulating the watershed response
to climate change. The models in the case studies each have one or more options for representing PET.
Many watershed modeling efforts perform well with simplified approaches to estimating PET, such as the
Hamon method, which depends primarily on temperature. The robustness of watershed model calibrations
conducted with simplified PET is suspect under conditions of climate change, since a variety of other
factors that influence PET, such as wind speed, relative humidity, and cloud cover, are also likely to
change. It is advisable to use a full energy balance method for PET, such as Penman-Monteith PET
(Jensen et al. 1990), yet little is known about the proper specification of climate-altered input variables
such as wind and solar radiation. Further research is needed to improve these capabilities.
89
-------
Atmospheric concentration of carbon dioxide (CO2) is a direct driver of climate change. Atmospheric CO2
concentrations have increased steadily throughout the last century. The trajectory of future CO2
concentration will vary depending on human efforts to reduce emissions, but could plausibly exceed 500
ppm (per volume) by 2050 (compared to about 370 ppm per volume in 2000). Increases in atmospheric
CO2 concentrations effectively reduce the stomatal conductance of plant leaves, thus reducing water loss
through transpiration. Limited research has been conducted on the potential effects of increased
atmospheric CO2 on ecosystems, but the initial findings indicate that when CO2 levels increase, ET
decreases (Leakey et al. 2009). CO2 effects on plant growth could also influence nutrient uptake, litter
fall, and other processes affecting water quality. Therefore, incorporation of CO2 fertilization into a
model is a potentially significant factor affecting simulation of watershed response to climate change. The
SWAT watershed model includes a plant growth module that can account for the effects of increased
CO2, but the other models available with BASINS CAT and WEPPCAT do not. Further research is
needed to improve our understanding of how changes in atmospheric CO2 concentrations influence
vegetative processes and our capabilities in representing these processes in hydrologic modeling.
90
-------
4. CONCLUDING COMMENTS
Many communities, states, and the federal government are considering adaptation strategies for reducing
the risk of harmful impacts resulting from climate change6. Challenges remain, however, concerning how
best to incorporate diverse, uncertain, and often conflicting information about future climate change into
decision making. Tools and information are needed to help build capacity for understanding and
responding to climate change impacts at local scale.
This report is a guide to application of two modeling tools recently developed by U.S. EPA and partners,
BASINS CAT and WEPPCAT. BASINS CAT and WEPPCAT are not stand-alone models. Rather, these
tools facilitate application of existing water models (HSPF, SWAT, SWMM, and WEPP) to assess
questions about climate change impacts. More specifically, each tool provides flexible capabilities for
creating user-specified climate change scenarios to address a range of "what-if' questions about the
potential effects of climate change on water. This report presents six short case studies designed to
illustrate tool capabilities. Each case study presents a real or plausible issue in a specified location, and
applies BASINS CAT or WEPPCAT to address or inform upon the problem. Taken together, the six case
studies illustrate the use of BASINS CAT and WEPPCAT to address a range of practical, real-world
questions of potential interest to water and watershed managers.
Case study simulations illustrate important differences in the sensitivity of streamflow and water quality
endpoints to changes in specific climate drivers. Generally, increased precipitation resulted in increased
streamflow and pollutant loads. The response to increased precipitation was found to be reduced or even
reversed, however, by increased evapotranspiration that resulted from increased annual temperatures.
Increased temperature combined with reduced precipitation resulted in consistent decreases in
streamflow. An awareness of these subtleties in the response of different streamflow and water quality
endpoints to specific types of climate change highlights the need for improved understanding of system
behavior, and in turn, the difficulty in developing quantitative predictions of future change.
Some of the most difficult questions that water and watershed managers will likely face regarding climate
change involve the development of effective management responses for reducing climate risk. Case study
simulations, while intended only as illustrative, suggest the potential effectiveness of climate adaption
options involving practices that are readily available, easy to employ, and likely to provide benefits under
a range of current and future conditions. For example, agricultural management practices such as
alternative tillage practices, BMPs (filter strips) and planting different crops may be successful in
reducing the impacts of climate change on sediment yields from fields. A more distributed stormwater
management system that employs green infrastructure was found to be beneficial in reducing stormwater
impacts under conditions of increased precipitation at event and annual time scales. Adapting stormwater
management practices to climate change may be further enhanced by reducing watershed impervious
cover as explored in Section 3.7. Considering climate change in the context of broad community planning
will help determine how climate change risks rate against other priorities and can be incorporated into
existing decision making processes.
6 For examples see: Climate Change Adaptation Task Force
(http://www.whitehouse.gov/administration/eop/ceq/initiatives/adaptation) and Federal Agency Climate
Adaptation Planning Implementation Instructions
http://www.whitehouse.gov/sites/default/files/microsites/ceq/adaptation final implementing instructions 3 3.
pdf.
91
-------
The scientific approach supported by these tools, i.e., scenario analysis, can be useful for understanding
system behavior, identifying vulnerabilities, and evaluating the effectiveness of management responses to
inform management decision making. The tools presented in this report, however, are just one step
forward in building our capacity for understanding and responding to climate change. Application of
hydrologic models in this way has limitations, many of which are not well understood (Ghosh 2010;
Ludwig, 2009; Najafi, 2011; Vaze et al., 2010). Further study is required to better assess, refine, and
develop our current modeling capabilities. Further study is also required to belter address the challenge of
incorporating diverse, uncertain, and often conflicting information about future climate change into water
resources decision making.
92
-------
REFERENCES
Anandhi. A., A. Frei, D. Pierson, E. Schneiderman, M. Zion, D. Lounsbury, and A. Matonse. (2011)
Examination of change factor methodologies for climate change impact assessment. Water Resources
Research, 47, W03501, 10 pp.
AQUA TERRA Consultants. (1994) Patuxent River Basin Watershed Model. U.S.
Geological Survey, Towson, MD.
AQUA TERRA Consultants. (2009) Hydrologic Modeling of the Santa Clara River Watershed with the
U.S. EPA Hydrologic Simulation Program - FORTRAN (HSPF). July 2009. 149 pp.
Bates, B.C., Z.W. Kundzewicz, S. Wu and J.P. Palutikof, Eds. (2008) Climate Change and Water.
Technical Paper of the Intergovernmental Panel on Climate Change, IPCC Sesreetariat, Geneva, 210 pp.
Blanco, et al. (2009) Hot, congested, crowded and diverse: Emerging research agenda in planning.
Progress in Planning 72: 153-205.
Blanco, et al. 2009. Shaken, shrinking, hot, impoverished and informal: Emerging research agendas in
planning. Progress in Planning, 72:195-250.
Blue Earth County, MN. (2010) Blue Earth County Profile, http://www.co.blue-
earth.mn.us/about/profile.php. Accessed December.
Brekke, L.D., Kiang, J.E., Olsen, J.R, Pulwarty, R.S., Raff, D.A., Turnipseed, D.P., Webb, R.S., and
White, K.D. (2009) Climate change and water resources management-A federal perspective: U.S.
Geological Survey Circular 1331. Washington, DC. 65pp. (Also available online at
http://pubs.usgs.gov/circ/1331/.)
Brown, Casey. "Decision-scaling for Robust Planning and Policy under Climate Uncertainty." World
Resources Report, Washington DC. Available online at http://www.worldresourcesreport.org. Accessed
May.
CCSP. (2003) Strategic plan for the U.S. climate change science program: a report by the Climate Change
Science Program and the Subcommittee on Global Change Research, final report, July, 2003. Available
online at http://www.climatescience.gov/Library/stratplan2003/final/default.htm
Clough, JS; Park, RA. (2006) AQUATOX (Release 3) Modeling environmental fate and ecological
effects in aquatic ecosystems, addendum to Release 2 & 2.1 Technical Documentation. U.S.
Environmental Protection Agency, Washington, DC.
Di Luzio, M., G. L. Johnson, C. Daly, J. K. Eischeid, and J. G. Arnold. (2008) Constructing Retrospective
Gridded Daily Precipitation and Temperature Datasets for the Conterminous United States. Journal of
Applied Meteorology and Climatology, 47(2): 475-497.
Donigian A.S. (2002) Watershed Model Calibration and Validation: The HSPF Experience. Transactions
of the ASABE.
Donigian A.S. (2003) Sediment Calibration Procedures and Guidelines for Watershed Modeling.
Transactions of the ASABE.
93
-------
Flanagan B.C. and Nearing M.A. eds. (1995) USDA - Water Erosion Prediction Project Hillslope Profile
and Water shed Model Documentation. NSERL Report No. 10.
USDA-ARS National Soil Erosion Research Laboratory.
Ghosh, Subimal, Deepashree Raje and P. P. Mujumdar. (2010) Mahanadi streamflow: climate change
impact assessment and adaptive strategies. CURRENT SCIENCE, VOL. 98, NO. 8, 25 APRIL 2010
Gleick, P and Adams, DB. (2000) Water: the potential consequences of climate variability and change for
water resources of the United States. Report of the Water Sector Assessment Team of the National
Assessment of the Potential Consequences of Climate Variability and Change, Pacific Institute, Oakland
CA; ISlpgs.
Groisman, P; Knight, R; Easterling, D; et al. (2005) Trends in intense precipitation in the climate record.
J Climate 18:1326-1350
Hamon, W.R. (1961) Estimating potential evapotranspiration. Journal of the Hydraulics Division, ASCE
87(HY3): 107-120.
Imhoff, J.C.; Kittle, J.L., Jr.; Gray, M.R; Johnson, T.E. (2007) Using the climate assessment tool (CAT)
in U.S. EPA basins integrated modeling system to assess watershed vulnerability to climate change.
Water Sci. Technol. 56:49-56.
IPCC (Intergovernmental Panel on Climate Change). (2007) Climate change 2007: impacts, adaptation,
and vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change. Cambridge, United Kingdom: Cambridge University Press.
IPCC-TGICA. (2007) General Guidelines on the Use of Scenario Data for Climate Impact and
Adaptation Assessment. Version 2. Prepared by T.R. Carter on behalf of the Intergovernmental Panel on
Climate Change, Task Group on Data and Scenario Support for Impact and Climate Assessment, 66pp.
Jensen M.E., R.D. Burman, and R.G. Allen. (1990) Evapotransipiration and irrigation water requirements.
ASCE Manuals and Reports on Engineering Practice No. 70, Am. Soc. Civil Engr., New York, NY, 332
Johnson, T. and J. Kittle. (2006) Sensitivity Analysis as a Guide for Assessing and Managing the
Impacts of Climate Change on Water Resources. Water Resources IMPACT (8)5.
Johnson, T., and C. Weaver. (2009) A framework for assessing climate change impacts on water and
watershed systems, Environmental Management, 43:118-134
Johnson, T.E., J.B. Butcher, A. Parker, and C.P. Weaver, in press: Investigating the Sensitivity of U.S.
Stream Water Quality to Climate Change: The EPA Global Change Research Program's "20 Watersheds"
Project. Journal of Water Resources Planning and Management.
Karl, T.R., Melillo, J.M., and Peterson, T.C. (eds.). (2009) Global Climate Change Impacts in the United
States. Cambridge University Press.
Leakey, Andrew D. B., Elizabeth A. Ainsworth, Carl J. Bernacchi, Alistair Rogers, Stephen P. Long,
Donald R. Ort. (2009) Elevated CO2 effects on plant carbon, nitrogen, and water relations: six important
lessons from FACE. Journal of Experimental Botany, Vol. 60, No. 10, pp. 2859-2876.
94
-------
Lempert R.J., D.G. Groves, S.W. Popper, Steve C. Bankes. (2006) A general, analytic method for
generating robust strategies and narrative scenarios.
Ludwig, R., I. May, R. Turcotte, L. Vescovi, M. Braun, J.-F. Cyr, L.-G. Fortin, D. Chaumont, S. Biner, I.
Chartier, D. Caya, andW. Mauser. (2009) The role of hydrological model complexity and uncertainty in
climate change impact assessment. Adv. Geosci., 21, 63-71.
Milly, P.C.D., J. Betancourt, M. Falkenmark, R.M. Fiirsch , Z.W. Kundzewicz , D.P. Lettenmaier, and R.
J. Stouffer. (2008) Stationarity Is Dead: Whither Water Management? Science, Vol. 319, no. 5863 pp.
573-574. DOI: 10.1126/science.1151915
Najafi, M. R. , H. Moradkhani* and I. W. Jung. (2011) Assessing the uncertainties of hydrologic model
selection in climate change impact studies. HYDROLOGICAL PROCESSES
Neitsch, SL; Arnold, JG; Kiniry, JR; et al. (2005) Soil and water assessment tool: theoretical
documentation, version 2005. USDA ARS, Grassland, Soil and Water Research Laboratory, Temple, TX.
Available online at
http://www.brc.tamus.edu/swat/downloads/doc/swat2005/SWAT%202005%20theory%20final.pdf
Paul, M.J., J.L. Meyer. (2001) Streams in the urban landscape. Annu. Rev. Ecol. Syst. 32, 333-365.
Pyke, Christopher, Meredith Warren, Thomas Johnson, James LaGro Jr., Jeremy Scharfenberg, Philip
Groth, Randall Freed, William Schroeer, and Eric Main. (2011) Assessment of low impact development
for managing stormwater with changing precipitation due to climate change. Submitted to Landscape and
Urban Planning.
Quay, R. (2010) Anticipatory Governance, A Tool for Climate Change Adaptation. Journal of the
American Planning Association. 76(4):496-511.
Rabalais, Nancy N., Turner, R.Eugene, Wiseman, William J., (2001) Hypoxia in the Gulf of Mexico.
Journal of Environmental Quality, Vol. 30 No. 2, p 320-329.
Rossi, A., N. Massei, B. Laignel, D. Sebag, and Y. Copard, (2009) The response of the Mississippi River
to climate fluctuations and reservoir construction as indicated by wavelet analysis of streamflow and
suspended-sediment load, 1950-1975. Journal of Hydrology, v3 77(3-4): 23 7-244.
Rossman, L. (2010) Stormwater Management Model User's Manual Version 5.0. Water Supply and
Water Resources Division, National Risk Management Research Laboratory. U.S. Environmental
Protection Agency EPA/600/R-05/040. Revised July.
Sarewitz, D., Pielke Jr, R.A., Byerly Jr, R. (2000) Introduction: death, taxes, and environmental policy.
In: Sarewitz, D; Pielke, Jr, RA; Byerly, Jr. R; eds. Predictions: science, decision making, and the future of
nature. Washington, DC: Island Press; pp. 1-7.
Shoemaker, L., Dai, T., and Koenig, J. (2005). TMDL Model Evaluation and Research Needs.
EPA/600/R-05/149. USEPA National Risk Management Research Laboratory, Cincinnati, OH.
Neitch, S.L., J.G. Arnold, J.R. Kiniry, J.R. Williams. (2005) Soil and Water Assessment Tool Theoretical
Documentation Version 2005. http://swatmodel.tamu.edu/media/1290/swat2000theory.pdf
95
-------
U.S. EPA (Environmental Protection Agency) (2001) Better assessment science integrating point and
nonpoint sources: BASINS version 3 user's manual, EPA823H01001, June 2001. Available online at
http://www.epa.gov/waterscience/basins/bsnsdocs.html
U.S. EPA. (2007) Better Assessment Science Integrating point and Nonpoint Sources ~ BASINS Version
4.0. EPA-823-C-07-001. U.S. Environmental Protection Agency, Office of Water, Washington, DC.
Available at: http://www.epa.gov/waterscience/basins/.
U.S. EPA. (2008a) National water program strategy: response to climate change, U.S. Environmental
Protection Agency. EPA 800-R-08-001, September 2008. Available online at
http://www.epa.gov/water/climatechange
U.S. EPA. (2008b) Modeling Water Quality Impacts of Corn Production for Ethanol in the Upper
Mississippi River Basin. U.S. Environmental Protection Agency, Office of Water, Washington, DC.
U.S. EPA. (2009a) BASINS Climate Assessment Tool (CAT): Supporting documentation and users
manual,EPA/600/R-08/088, May, 2009.
U.S. EPA. (2009b) Land-Use Scenarios: National-Scale Housing-Density Scenarios Consistent with
Climate Change Storylines. Global Change Research Program, National Center for Environmental
Assessment, Washington, DC; EPA/600/R-08/076F. Available from: National Technical Information
Service, Springfield, VA, and online at http://www.epa.gov/ncea.
U.S. EPA. (2010) Model Calibration of Water Quality Impacts of Corn Production for Ethanol in the
Raccoon River Basin. U.S. Environmental Protection Agency, Office of Water, Washington, DC.
U.S. EPA. (201 la) Climate Change Vulnerability Assessments: Four Case Studies of Water Utility
Practices. EPA/600/R-10/077F.
U.S. EPA. (2011) Climate Change Vulnerability Assessments: A review of water utility practices. U.S.
EPA Office of Water Report.
Vaze, J., D.A. Post, F.H.S. Chiew, J.M. Perraud, N.R. Viney, and J. Teng, (2010) Climate non-
stationarity - Validity of calibrated rainfall-runoff models for use in climate change studies. J.
Hydrology, 3 94(3-4): 447-45 7
Volkery, A. and T. Ribeiro. (2009) Scenario planning in public policy: Understanding use, impacts and
the role of institutional context factors. Technological Forecasting and Social Change. 76(9): 1198-1207.
WUCA. (2010) Decision Support Planning Methods: Incorporating Climate Change Uncertainties into
Water Planning.
Young, K.D., T. Younos, R.L. Dymond, and D.F. Kibler. (2009) Virginia's Stormwater Impact
Evaluation: Developing an Optimization Tool for Improved Site Development, Selection and Placement
of Stormwater Runoff BMPs. VWRRC Special Report No. SR44-2009. Virginia Polytechnic Institute
and State University, Blacksburg, Virginia
USDA. (2010a) National Agricultural Statistics Service, . Accessed
December.
96
-------
USDA. (2010b) National Resource Conservation Service. Web Soil Survey. 2010.
http://websoilsurvey.nrcs.usda.gov/appAVebSoilSurvey.aspxAccessedDecember.
97
-------
APPENDIX A
Selected Sources of Climate Change Information
The case studies in this report utilized climate model projections and other information to develop the
climate scenarios. The information used in this report is just a small subset of the climate data currently
available. Selected additional sources of climate change information, data, and guidance concerning the
use of climate change data are listed below. Most sources provide climate change projections developed
from climate modeling experiments using GCMs or RCMs. This is not an exhaustive list of climate data
sources. Information and guidance about climate change in different parts of the country can be obtained
from additional sources including government agencies and universities. Overtime, additional
information about climate change will become available as climate models are improved, new modeling
experiments are conducted, new monitoring data becomes available, and research such as investigations
of paleo-climate better reveal historical patterns of climate variability and change.
Bias Corrected and Downscaled WCRP CMIP3 Climate and Hydrology Projections
http://gdo-dcp.ucllnl.org/downscaled_cmip3_projections/dcplnterface .html
ClimateWizard
http://www.climatewizard.org
Conservation International
http: //futureclimates. conservation. org/
Data Basin
http: //databasin .org/
Earth System Grid gateway
http://pcmdi3.llnl.gov/esgcet/home.htm
IPCC Data Distribution Centre (DDC)
http://www.ipcc-data.org/
Lawrence Livermore National Laboratory, Program for Climate Model Diagnosis and Intercomparison
http: //www-pcmdi. llnl .gov/ipcc/about_ipcc .php
Lawrence Livermore National Laboratory/Bureau of Reclamation/Santa Clara University
http://gdo-dcp.ucllnl.org/downscaled_cmip3_projections/dcplnterface .html
North American Regional Climate Change Assessment Program (NARCCAP)
http://www.narccap.ucar.edu/data/index.html
SERVIR - Regional Visualization and Monitoring System
http: //www. servir.net/
USDA Forest Service
http://www.fs.fed.us/rm/data_archive/dataaccess/contents_datatype.shtml
98
------- |