4\ United States
Environmental Protection
tl *m Agency
EPA/600/R-19/039
July 2018
www.epa.gov/ord
ScenCompare
WMOST Climate Scenario Viewer and
Comparison Post Processor
(Version 1: July 31, 2018)
WMOST
&
<>V*
Tool -
Watershed Manage^e^1
Office of Research and Development
National Health and Environmental Effects Research Laboratory
-------
-------
JRk United States EPA/600/R-19/039
tnp\ Agencym0ntal Pr°teCti°n July 2018
www.epa.gov/ord
ScenCompare
WMOST Climate Scenario Viewer and
Comparison Post Processor
(Version 1: July 31, 2018)
Naomi Detenbeck
Atlantic Ecology Division
National Health and Environmental Effects Research Laboratory
Narragansett, Rhode Island 02882
Chris Weaver
Exposure Analysis and Risk Characterization Group
National Center for Environmental Assessment
Office of Research and Development
Research Triangle Park, NC 27709
National Health and Environmental Effects Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Atlantic Ecology Division
Narragansett, Rhode Island 02882
-------
Notice and Disclaimer
The views expressed in this ScenCompare User Guide are those of the authors and do not necessarily
reflect the views or policies of the U.S. Environmental Protection Agency. This document was subjected
to the Agency's ORD review and approved for publication as an EPA document. Mention of trade names
or commercial products does not constitute endorsement.
ScenCompare Instructions and User Guide, July 31st, 2018
-------
Table of Contents
Notice and Disclaimer ii
1 Workbook Organization 1
2 Controls 2
2.1 Load WMOST Scenario Data 2
2.2 Compare Decision Variables Across Scenarios 3
2.3 Compare Overall Costs across Scenarios 4
2.4 Compare Time Series Variables Across Scenarios 5
2.5 Compare Land Management Variables Across Scenarios 5
3 Example Application for Wading-Threemile Watershed 8
3.1 Getting Started 8
3.1.1 Run WMOST Scenarios 8
3.1.2 Load WMOST Data 10
3.2 Baseline and Climate Scenarios 12
3.2.1 Compare Model Input Data 12
3.2.2 Compare Cost and Decision Variables Across Scenarios 13
3.2.3 Compare Time Series Variables Across Scenarios 14
3.3 Land Use Optimization Scenarios 16
3.3.1 Compare Robustness of Land Management Variables Decisions Across Scenarios 16
3.3.2 Compare Land Management Variables Across Scenarios 18
References 19
ScenCompare Instructions and User Guide, July 31st, 2018
-------
-------
ScenCompare
WMOST Climate Scenario Viewer and
Comparison Post Processor
(Version 1; July 31, 2018)
Instructions
ScenCompare is a MS-Excel application designed to view and compare WMOST scenario results.
ScenCompare is compatible with MS-Excel (versions 2010, 2013 and 2016). The tool is specifically
intended to allow comparison of WMOST results for different climate scenarios, but ScenCompare more
generally allows comparison and evaluation of any sets of WMOST results to understand the effects of
varying climate, land use, and other model inputs on the set of management actions selected by
WMOST to meet the specified management goal at the lowest cost. For example, ScenCompare can
assist users interested in applying WMOST as part of Robust Decision Making (RDM) approaches for
identifying vulnerabilities, and managing goals and risks, in the face of uncertain future conditions.
Under RDM-type approaches, the outcome of a prescribed management strategy, such as a Watershed
Implementation Plan (WlP)to meet Total Maximum Daily Load (TMDL) requirements, can be tested
against multiple scenarios of future changes in temperature and rainfall and then analyzed within
ScenCompare to determine under what conditions the strategy might be expected to fail to meet
performance requirements. Further, ScenCompare provides WMOST users access to all outputs
generated by a WMOST run, thereby expanding on the set of standard outputs visible in the WMOST v3
interface.
These instructions focus on the process for loading WMOST results into ScenCompare, generating
summary tables comparing decision variables, and generating time series plots of user-selected
variables across scenarios. ScenCompare users should already be familiar with WMOST model outputs
being processed. Please refer to the WMOST documentation for details on variables and modeled
components included in WMOST output files and available for processing with ScenCompare.
1 Workbook Organization
ScenCompare is an Excel workbook that uses customized Visual Basic for Applications (VBA) code to
automate key tasks. The initial workbook includes:
• Introduction tab: Describes the purpose of ScenCompare and allows the user to navigate to
the Controls, Variable Definitions and Loaded Scenarios tabs (the user can also navigate
to the tabs by clicking on them at the bottom of the screen).
• Controls tab: Provides access to steps in compiling and analyzing WMOST data.
l
ScenCompare Instructions and User Guide, July 31st, 2018
-------
• Variable Definitions tab: Provides a description of the WMOST variables in the output data.
• Loaded Scenarios tab: Provides the inventory of output files imported by the user
• Model Results tab: Provides the WMOST model results output for each loaded file
• Model Input Data tab: Provides the model input data for each loaded file
Section 2 describes the steps involved in compiling and analyzing WMOST data. Note that the user can
also use all standard Excel functions and capabilities from within ScenCompare to customize graphics or
tables, perform calculations, etc.
2 Controls
2.1 Load WMOST Scenario Data
To load WMOST Scenario Input Data and Results into ScenCompare, click on the leftmost button on the
Controls tab labeled "Load WMOST Scenario Data." When prompted, select the model Specifications
and Results file. Note that the Specifications and Results file is a log file generated by WMOST during a
model run and saved to the same file folder that contains the WMOST model. The log file contains the
model inputs and results for the run.
ScenCompare adds the WMOST scenario outputs to the Model Results tab, the WMOST input data to
the Model Input Data tab, and the name of the file and other model details to the Loaded Scenarios tab.
Repeat these steps to load the data for other WMOST model runs as many times as needed to import
the desired data. Note that you may add results to ScenCompare at any time.
ScenCompare verifies that any data file loaded after the first data file has the same number of land uses,
land use sets, water users and time steps. Differences in these variables leads to mismatches in the
variable order and incorrect comparisons of data. If a difference is found, ScenCompare gives the option
to keep or discard the data that was just loaded. It is recommended that you discard the data.
ScenCompare automatically fills in the average annual precipitation and temperature, if you used the
hydrology module when developing the WMOST scenarios or used the Hydro-Climate Automation
module (HCAM). If you developed the data manually, you can enter precipitation and temperature
statistics on the Loaded Scenarios tab. ScenCompare uses this information for some of the graphs
(see Section 2.3).
2
ScenCompare Instructions and User Guide, July 31st, 2018
-------
Figure 2: Example of filter selecting variables that take different values among three scenarios. In this example, the
scenarios show differences in total operating costs, management approaches selected (e.g., stormwater BMPs and ASR) and
level of implementation, and associated costs.
Home Insert
ijji] From Access : jh
Page Layout
Tn
Formulas Data
| © Connections
Properties
WMOST ScenCompare_012517.xlsm - Microsoft Excel'
Review l&k """Nuance PDF
I w m [§|] Data Validation -
% From Web
From Other Existing Refresh
^ From Text sources " Connections All - "y ^it Links
Get External Data
Connections
pply Consolidate
Text to Remove
'anced columns Duplicates What-If Analysis -
I Data Tools
& Group *¦ *i If^D
^ Ungroup ~ ™| S
Hi Subtotal
Outline rs ;
H162
A
B
C
D
E F ^
1
2
35
Variable
objective
DALu33
Description [ ^
Objective cost
Land Area - 0.6" Infiltrat
Units |
$
ac
wASR [~
18186396.09
468.9155099
No Brockton [HwlBT
466450.5466 1635461.18^
0 0
Data Difference |»T
— —«f
J
36
DALU43
Land Area - 0.6" Infiltrat ac
188.44859
0 0 1
37
DALUS3
Land Area - 0.6" Infiltrat ac
56.29038155
0 52.40900839 1
43
DALU113
Land Area - 0.6" Infiltrat ac
2.463176949
0 0
66
DALU45
Land Area -1" Infiltratio ac
133.4065532
0 0
67
DALU55
Land Area -1" Infiltratio ac
4.51856203
0 0
73
DALU115
Land Area -1" Infiltratio ac
-3.26569E-16
0 0
111
DQAsrAddl
Additional ASR capacity MGD
186.1776694
0 0
125
CLuSet3
0.6" Infiltration trench -
$/yr
160377.4126
0 62698.78515
127
CLuSetS
1" Infiltration trench - Lc $/yr
70982.79257
0 0 1
135
CWtp
Total cost of potable wa
$/yr
466450.5466
466450.5466 283139.1895 1
146
CCAsr
Capital cost of aquifer s1 $/yr
16315025.63
0 0 1
147
CAsr
Total cost of aquifer stoi $/yr
17488585.34
0 0 1
150
ClbtW
Total cost of potable int $/yr
0
0 183311.3571 1
154
CWMake
Penalty for water deficit $/yr
0
0 1106311.856 1
156
TIP: You can use standard Excel filter tools (accessible through the Data menu) to hide/show a subset of
variables you are interested in, or to show only the variables that have Data Difference of 1.
2.3 Compare Overall Costs across Scenarios
Clicking on the "Make Climate Graphs" on the Table Comparison tab (see Section 2.2) creates a new tab
called ClimateGraph objective This tab contains three climate plots:
1) a scatterplot that charts the objective cost versus average annual precipitation,
2) a scatterplot that charts the objective cost versus average annual temperature, and
3) a bubble plot that charts the objective costs versus total annual precipitation and
average annual temperature.
The "objective" value is the total annualized cost of all watershed management actions taken to meet
the specified objective (i.e., meeting water demand subject to physical constraints and water quantity
and/or quality targets).
As discussed in Section 2.1, if you did not use the hydrology module in WMOST, themodel Specifications
and Results file does not automatically include statistics for average annual precipitation and average
annual temperature. This means that ScenCompare cannot automatically extract the information
needed to create the climate graphs. To rectify this issue, you can enter the precipitation and
temperature statistics for each scenario on the Loaded Scenarios tab before creating the tables and
ScenCompare will use those values to create the climate graphs.
4
ScenCompare instructions and User Guide, July 31s", 2018
-------
2.4 Compare Time Series Variables Across Scenarios
ScenCompare can also be used to compare time series variables across scenarios. These are variables
representing time-dependent flows or stocks in the modeled watershed components, listed on the
Controls tab.
To compare time series variables across scenarios, place the cursor in the first variable of interest in the
list under Step 3 on the Controls tab, and click on the button labeled "Create Tables and Graphs for
Selected Variables"2 to the right of the data columns. This action creates a new tab called Table nn
where nn is the name of the selected variable. This new tab provides summary statistics for the selected
variables for each scenario in ScenCompare (e.g., minimum, maximum, average, and number of
observations greater than 0 (as the default threshold)) along with values for each time step.
This action also creates three plots on the time series variable tab:
1) a time series plot of the variable over the time period,
2) a histogram of the count of time steps (e.g., number of days if WMOST was run using a
daily time step) for which the variable takes a value greater than the "Count Threshold"
for each scenario, and
3) a box-and-whisker plot that shows the minimum, 1st quartile, median, 3rd quartile, and
maximum for each scenario.
You can use the button on the right, "Make Climate Graphs", to create a new tab called
ClimateGraph nn that contains three climate graphs that compare the average value over the entire
time series across the scenarios. Refer to Section 2.3 for more details on the climate graphs and how
they are created.
TIP: You can change the "Count Threshold" from its default value of 0 to any number, and the tab
adjusts the count statistics and histogram to reflect the new threshold.
TIP: You can use standard Excel tools to add or modify the formatting of the basic plot generated
automatically by ScenCompare.
2.5 Compare Land Management Variables Across Scenarios
ScenCompare can also be used to compare land management decisions - such as decisions to conserve
land as undeveloped or to implement stormwater BMPs - across scenarios. The comparison is done
using the Table Comparison tab.
As further described in the WMOST documentation, WMOST represents land use using hydrologic
response units (HRUs) and land use management options using a series of HRU sets. Land use
management option variables have units in acres.
2 Note that you may select multiple variables to be processed at the same time by holding the control key and
clicking on all desired variable names before clicking on the button.
ScenCompare Instructions and User Guide, July 31st, 2018
5
-------
The names of land use management variables all begin with "DALu" and are followed by two numeric
identifiers. The first identifier is the HRU number and the second identifier is the HRU set number. For
example, DALul2 represents the land use allocation of HRU 1 in HRU Set 2. The convention is the same
for double digit HRU or HRU set numbers. For example, DALulOlO represents the land use allocation of
HRU 10 in HRU set 10. The variables are described in the second column on the Table Comparison tab
and in the Variable Definitions tab based on the management option and the HRU. For example, Land
Area - 0.6" Infiltration trench, "Medium to low density residential, Sand and Gravel" contains the acres
of medium to low density residential land, on sand and gravel (the HRU represented by the combination
of land use and soil type) on which WMOST decided to implement a 0.6" infiltration trench.
Each of the land use management options (HRU sets) has an associated cost variable that contains the
cost associated with the decision. The variables are named CLuSet#, where # is the number of the HRU
Set.
To compare land use allocation variables, you must look at each land use management set individually.
By convention, the first land use management set represents land conservation decisions. All other sets
are related to stormwater management decisions.
• Land Conservation: One of the management options available in WMOST is the decision to
conserve undeveloped land. The decision essentially reallocates baseline land use to
undeveloped land uses, keeping the total land area the same. The final land area allocation is
reported through the first set of DALu variables (all DALu variables for HRU set 1). To determine
whether land area was conserved, you can first look at the CLUSetl variable to see if it is greater
than $0 for any of the scenarios, which would indicate that WMOST incurred costs to conserve
land.3 To determine how much land area was conserved, you should then look at the DALu
variables and compare values to the baseline acres you had specified in your WMOST run.4 The
resulting difference is the change due to land conservation. A positive difference means more
land was conserved, and a negative difference means the land was converted to undeveloped
areas or conserved.5
• Stormwater Management: WMOST may also implement stormwater BMPs on developed HRU
areas. The areas managed using stormwater BMPs are reported in the remaining DALu variable
sets. The values represent the number of acres receiving the type of stormwater BMP defined
by the HRU set, e.g., acres of medium to low density residential on sand and gravel managed
using a 0.6" Infiltration trench. The number of acres will be a portion of the HRU area reported
in the first management set described above. WMOST may select multiple stormwater BMP
types for any given scenario but the total acres managed across the HRU sets cannot exceed the
total area, i.e., stormwater BMPs are mutually exclusive and WMOST applies only one type of
BMP to any given parcel of land.
3 This presumes that you specified non-zero costs to acquire land for conservation in your WMOST inputs.
4 Note that the baseline HRU acres are reported in the Model Input Data tab, using the variable ALuBase and HRU
number.
5 Note that decisions to conserve land will not be flagged as a change in the Data Difference column in
Table_Comparison tab unless the allocations differed across scenarios.
6
ScenCompare Instructions and User Guide, July 31st, 2018
-------
TIP: To only view results that relate to land use management, you can use MS-Excel filter tool to select
the relevant variables: DALu## and CLuSet#.
Figure 3: Example of selected variables related to land use management decisions. The screen shows differences in the
number of acres managed using 0.6" infiltration trenches for several HRUs.
|y IT ScenCompare_012517.xlsrTr" Microsoft Excel
ll Home P Insert Page Layout Formulas Data Review View Nuance PDF
:
m *
Calibri ~ 11 "A" a'
— = = 3?-*" 3^ Wrap Text
General
m m
b
B™
a*
m
i4£ -
Paste
^ j
B / U -~ gj t ^ -
m w m i
- =F HI Merge & Center -
$
_ 0/ , *.0 .00
" fo y .oo >.o
Conditional Format as
Formatting - Table T
Cell
Styles -
Clipboard ft
Font ft
Alignment
ft
Number ft |
Styles
H3 - &
S>
B
C
D
E
F
G
Variable-^ Description
*• Unil 7 wASR
U
No Brockton w wlBT
w | Data Difference
*Tand Area - Land Area with Conservation, "Forest, Sand and Gravel" ac
833.6582144
833.6582144
833.6582144
°L
4
DALu21
Land Area - Land Area with Conservation, "Open nonresidential, Sar ac
74.36874457
74.36874457
74.36874457
0
5
DALu31
Land Area - Land Area with Conservation, "Medium to low density re: ac
597.9906384
597.9906384
597.9906384
0
6
DALu41
Land Area - Land Area with Conservation, "High-density residential, ac
321.8551432
321.8551432
321.8551432
0
7
DALu51
Land Area - Land Area with Conservation, "Commercial-industrial-trc ac
60.80894358
60.80894358
60.80894358
0
8
DALuGI
Land Area - Land Area with Conservation, "Agriculture, Sand and Gr ac
10.34096625
10.34096625
10.34096625
0
9
DALu71
Land Area - Land Area with Conservation, "Forest, Till & fine-grainec ac
161.0035535
161.0035535
161.0035535
0
10
DALu81
Land Area - Land Area with Conservation, "Open nonresidential, Till ac
1.739357974
1.739357974
1.739357974
0
11
DALu91
Land Area - Land Area with Conservation, "Medium to low density re: ac
28.63970334
28.63970334
28.63970334
0
12
DALu101
Land Area - Land Area with Conservation, "High-density residential, ac
61.81480419
61.81480419
61.81480419
0
13
DALu111
Land Area - Land Area with Conservation, "Commercial-industrial-trc ac
2.463176949
2.463176949
2.463176949
0
14
DALu121
Land Area - Land Area with Conservation, "Agriculture, Till & fine-gr ac
12.6231485
12.6231485
12.6231485
0
15
DALu131
Land Area - Land Area with Conservation, "Cranberry bogs. Combine ac
340
340
340
0
16
DALu141
Land Area - Land Area with Conservation, "Forested wetland, Combit ac
64.2
64.2
64.2
0
17
DALu151
Land Area - Land Area with Conservation, "Nonforested wetlands, Cc ac
47.3
47.3
47.3
0
18
DALu12
Land Area - 0.6" Bioretention with UD, "Forest, Sand and Gravel"
ac
0
0
0
0
19
DALu22
Land Area - 0.6" Bioretention with UD, "Open nonresidential, Sand ai ac
0
0
0
0
20
DALu32
Land Area - 0.6" Bioretention with UD, "Medium to low density reside ac
0
0
0
0
21
DALu42
Land Area - 0.6" Bioretention with UD, "High-density residential, San ac
0
0
0
0
22
DALu52
Land Area - 0.6" Bioretention with UD, "Commercial-industrial-transp ac
0
0
0
0
23
DALu62
Land Area - 0.6" Bioretention with UD, "Agriculture, Sand and Gravel ac
0
0
0
0
24
DALu72
Land Area - 0.6" Bioretention with UD, "Forest, Till & fine-grained de| ac
0
0
0
0
25
DALu82
Land Area - 0.6" Bioretention with UD, "Open nonresidential. Till &
fi ac
0
0
0
0
26
DALu92
Land Area - 0.6" Bioretention with UD, "Medium to low density reside ac
0
0
0
0
27
DALu102
Land Area - 0.6" Bioretention with UD, "High-density residential, Till ac
0
0
0
0
28
DALu112
Land Area - 0.6" Bioretention with UD, "Commercial-industrial-transp: ac
0
0
0
0
29
DALu122
Land Area - 0.6" Bioretention with UD, "Agriculture, Till & fine-graine ac
0
0
0
0
30
DALu132
Land Area - 0.6" Bioretention with UD, "Cranberry bogs. Combined" ac
0
0
0
0
31
DALu142
Land Area - 0.6" Bioretention with UD, "Forested wetland. Combined' ac
0
0
0
0
32
DALu152
Land Area - 0.6" Bioretention with UD, "Nonforested wetlands, Combi ac
0
0
0
0
33
DALu13
Land Area - 0.6" Infiltration trench, "Forest, Sand and Gravel"
ac
0
0
0
0
34
DALu23
Land Area - 0.6" Infiltration trench, "Open nonresidential. Sand and I ac
n
35
DALu33
Land Area - 0.6" Infiltration trench, "Medium to low density residents ac
468.9155099
0
0
1
36
DALu43
Land Area - 0.6" Infiltration trench, "High-density residential. Sand a ac
188.44859
0
0
1
37
DALu53
Land Area - 0.6" Infiltration trench, "Commercial-industrial-transports ac
56.29038155
0
52.40900839
1
38
DALu63
Land Area - 0.6" Infiltration trench, "Agriculture, Sand and Gravel"
ac
0
0
0
0
39
DALu73
Land Area - 0.6" Infiltration trench, "Forest, Till & fine-grained depos ac
0
0
0
0
40
DALu83
Land Area - 0.6" Infiltration trench, "Open nonresidential. Till & fine- ac
0
0
0
0
41
DALu93
Land Area - 0.6" Infiltration trench, "Medium to low density residents ac
0
0
0
0
42
DALu103
Land Area - 0.6" Infiltration trench. "Hiah-densitu residential. Till & fi ac
0
0
0
0
H
< ~ ~!
Controls Variable Definitions
Loaded Scenarios Model Results Model Input Data Table
_Comparison ClimateGra
di|<
Ready Filter Mode | ||Bil
ScenCompare Instructions and User Guide, July 31st, 2018
7
-------
3 Example Application for Wading-Threemile Watershed
Below is an example of an application of ScenCompare using a WMOST case study on the Wading-
Threemile River Watershed in the Taunton Basin in Massachusetts. The example provides a guide for
setting up and loading scenario runs in ScenCompare (Section 3.1), using the tool's functions and
evaluating the scenario data (Section 3.2), and analyzing the various land use management decisions in
WMOST (Section 3.3).
3.1 Getting Started
The specific purpose of ScenCompare is to provide users with an interface and tool for comparing
WMOST results for different future climate scenarios. Therefore, this example details the differences in
WMOST management decisions between the baseline and future climate scenarios.
However, in general, the functions of this tool can be used to evaluate any set of WMOST results and
help you to understand the effects of model inputs on the management actions selected by the WMOST
optimization model, which meets specified management goals at the lowest cost. The following section
discusses the development of the WMOST scenarios for the Wading-Threemile River Watershed, and
how those scenarios are loaded into ScenCompare to prepare the tool for analysis.
3.1.1 Run WMOST Scenarios
Data for this example come from the Wading-Threemile subwatershed in the upper Taunton River basin
in Massachusetts (https://www.mass.gov/service-details/taunton-river-watershed). The Taunton River
watershed is the second largest watershed in Massachusetts and the largest freshwater contributor to
Narragansett Bay. The Taunton River is the longest undammed tidal river in New England, supporting
the largest herring run in the state. In 2009 it was designated as a Partnership Wild and Scenic River by
the National Park Service. In these Partnership Wild and Scenic Rivers communities protect their own
outstanding rivers and river-related resources through a collaborative approach. Challenges faced by
communities in the Taunton include protection of outstanding natural resource areas, flooding, sea level
rise and storm surges, water body impairments related to eutrophication, water supply constraints, and
the need to protect the downstream Hope Bay (RTI 2014). This case study was developed in
cooperation with a consortium of regional development agencies (Southeast Regional Planning and
Development District, SERPDD and the Metropolitan Area Planning Council, MAPC) and
nongovernmental organizations (Manomet, the Nature Conservancy, and Mass Audubon) which had
received funding from EPA Region 1 from the Healthy Communities Grant Program to assess the
benefits of green infrastructure within the watershed and to educate the public about those benefits.
The EPA ORD team has applied WMOST v3 to the two subwatersheds within the upper Taunton. A two-
stage objective was established: first, to minimize costs (capitol plus operations and maintenance) for
near term planning, and second, to minimize future costs under projected growth and climate scenarios.
Goals and constraints considered in Stage 1 included ecoregional targets for total phosphorus in lakes
and flowing waters, a reduction in total nitrogen loads to the Mt Hope Bay, and maintenance of
minimum low flows for a stable water supply and to support fish populations. Management options
8
ScenCompare Instructions and User Guide, July 31st, 2018
-------
under consideration include land conservation, stormwater best management practices (BMPs,
including green infrastructure), forested riparian buffer restoration, repair of water infrastructure leaks,
upgrades in wastewater treatment, water conservation, and aquifer storage and recharge. A
comparison of different traditional ("gray") and nature-based ("green) stormwater BMPs showed that
infiltration basins were the most cost-effect option to meet water quality goals. Initial results were
shared with the Resilient Taunton Watershed Network (RTOWN).
WMOST v3 is now being applied to future growth and climate scenarios to identify the most cost-
effective management actions. Future projections of mean annual temperature and mean annual
precipitation were obtained from the general circulation models (GCMs) included in the 5th Coupled
Model Intercomparison Project (CMIP5) for two of the representative concentration pathways (rcp-4.5 and
rcp-8.5) adopted by the Intergovernmental Panel on Climate Change (IPCC) in its 5th Assessment Report
(IPCC 2014). The pathways in this report correspond to changes in radiative forcing relative to pre-
industrial values (i.e. +2.6, +4.5, +6.0, +8.5 W/m2) that are possible in year 2100 based on projections of
greenhouse gas emissions (IPCC 2014). These data were corrected for bias and statistically downscaled
to a regional scale (Brekke et al. 2013). Four combinations of changes in temperature and precipitation
(AT, AP) were selected for this study to roughly bound the extremes of AT and AP reflected by the
collection of GCMs, thereby representing a range of possible future climate scenarios, and an average of
these scenarios was also calculated. A new set of input hourly temperature and precipitation data was
generated for each scenario by uniformly adjusting the baseline temperature and precipitation records by
the corresponding AT (absolute) and AP (percentage) values, respectively. Using the adjusted
temperature and precipitation data, hourly runoff rates were generated for each scenario using SWMM.
Similarly, the temperature, precipitation, and runoff data were used in SWMM to generate four new sets
of hourly nitrogen and phosphorus loading rates. In the following, climate change scenarios are labeled
as General Circulation Model (GCM) AT (°F)/AP (%).
One of the objectives of the WMOST Wading-Threemile case study is to analyze the robustness of
WMOST management decisions over these future climate scenarios. To do this, we created a series of
WMOST runs based on a historical dry year (2002) and five climate scenarios (one median projection
and four bounding scenarios). Bounding scenarios were identified using the US EPA LASSO tool
(Morefield 2016), focusing on 21 of the models that had been shown to perform well for New England in
hindcasting exercises (Sheffield et al. 2015). The average projection was AT = +4.4°F/AP = +10.1%. The
bounding scenarios were based on the FGOALS-s2, realization 3 (GCM AT = +6.2°F/AP = +19.4%), IPSL-
CM5A_LR, realization 1 (GCM AT = +5.0°F/AP = -1.7%), MPI-ESM-LR realization 2 (GCM AT = 3.7°F/ AP =
-2.2%), and CSIRO-Mk3-6-0 realization 1 (GCM AT = +3.3°F/AP = +19.5%) model runs6.
6 CSIRO-Mk3-6-0 is from the Commonwealth Scientific and Industrial Research Organisation in collaboration with
the Queensland Climate Change Centre of Excellence; FGOALS-s2 is from the LASG, Institute of Atmospheric
Physics, Chinese Academy of Sciences; IPSL-CM5A-LR is from Institut Pierre-Simon Laplace; MPI-ESM-LR is from
Max Planck Institute for Meteorology (MPI-M).
ScenCompare Instructions and User Guide, July 31st, 2018
9
-------
We ran the historical climate arid five future climate scenarios (four bounding and one median) for three
types of comparisons:
1. Baseline (baseline land use with no land management options);
2. Optimal stormwater BMP implementation for 2002 (fixed set of optimal stormwater land use
BMPs for climate scenario runs); and
3. Optimal riparian zone implementation for 2002 (selection of 10 potential riparian buffer land
conversions from developed land to forest).
The TN loading target (1,156 lbs N) was turned off for the baseline run (for the historical scenarios and
the climate scenarios), as well as in the climate scenarios for the stormwater BMP and riparian zone
runs to see if the future scenarios would meet the target given optimal BMPs selected for 2002. This set
of runs was designed to test how robust the original solution was.
In addition to the three comparisons listed above, we also compared the differences in decisions
between the historical baseline and two climate scenarios (the median and extreme) with the TN
loading target turned on and a stormwater BMP set available (Section 3.3.2). This set of comparisons
evaluated whether optimal management practices would change given climate change scenarios.
3.1.2 Load WMOST Data
Once you have completed your scenario runs and prepared the Scenario Log Files in
WMOST (see WMOST V3 User Guide7 for more details), open the ScenCompare
application. From the introduction tab, navigate to the Controls tab. You can also use
the Introduction tab to navigate to the Variable Definitions tab and the Loaded
Scenarios tab.
Go to Controls
Go to Variable
Definitions
Go to Loaded
Scenarios
Step 1: Load data from selected WMOST log files
Load WMOST Scenario
Data
select the Scenario Log File you
created for your baseline run and
click "Open".
After opening your file, another
dialog box will pop up, prompting
you to name this scenario. Enter a
Scenario Name
Name for this scenario (optional):
OK
Cancel j
|Baseline2002j
On the Controls tab, click the "Load WMOST Scenario Data"
button to open a file selection dialog box. In the dialog box,
01 File Open
Ms-r
Wading_Threemile ~ Baseline
Search Baseline
Organize ~ New folder
a
OneDrive
^ Libraries
H Documents
ill My Docurr
Public Doc
^ Music
B Pictures
B Videos
Computer
£, OSDisk (CO
if^jj Wad in g3 M i I e_b asel i n e3.3.9.5_SpecsResu Its. csv
E3 i Wading3Mile_baseline3.7m2.2_SpecsResults.C5v
Wading3Mile_baseline4.410.l_SpecsResults.csv
l§J| Wading3Mile_baseline5.0ml.7_SpecsResults.csv
Wading3Mile_baseline6.2L9.4_SpecsResults.csv
Type
Microsoft Excel Comma
Microsoft Excel Comma
Microsoft Excel Comma
Microsoft Excel Comma
Microsoft Excel Comma
0 j Wading3mile_baselineTN2002_SpecsResults.c5v
Microsoft Excel Comma
File name: Wading3mile_baselineTN2002_Sp« ~ | All Files (*¦*)
7 Available from https://www.epa.gov/ceam/wmost-30-download-page
10
ScenCompare Instructions and User Guide, July 31s", 2018
-------
3.2.3 Compare Time Series Variables Across Scenarios
Finally, we look at the comparisons available for the results time series variables. Under Step 3, on the
Controls tab, you can select one or more variables you would like to compare on their own tab. Select
the variables of interest in the "Series Variables" column, and click the "Create Tables and Graphs from
Selected Variables" button to generate a new tab for each variable In this case, we selected DQSwExt,
the flow time series of surface water flowing outside the watershed. The flow regime from surface
water to the external watershed is an indicator of watershed health because it represents the volume of
water available to the stream and downstream watersheds after the water is used for human demand.
Step 3. Compare time series across scenarios.
Series Variables (select below)
CateEorv
Description
I
COmAsr
O&M Costs
Aquifer storage and recovery (ASR)
f Create Tables and
COmESep
O&M Costs
Enhanced septic treatment
1 Graphs from Selected 1
COmGwPump
O&M Costs
Groundwater pumping
Variables M,
COmlbtW
O&M Costs
Interbasin transfer (IBT) potable water
*
COmlbtWw
O&M Costs
IBT wastewater
COmNpdist
O&M Costs
Nonpotable distribution system
COmOS
O&M Costs
Operation cost of offline storage use
Clear Tables and
COm Res
O&M Costs
Reservoir management
COmSwPump
O&M Costs
Surface water pumping
COmWrf
O&M Costs
Water reuse facility (WRF)
COmWtp
O&M Costs
Water treatment
COmWwtp
O&M Costs
Wastewater treatment
DQCSOS
Flow
Combined sewer to offline storage
DQGwExt
Flow
Groundwater to external
DQGwMake
Flow
Groundwater deficits
DQGwWtp
Flow
Groundwater to water treatment plant
DQIbtWUseNp
Flow
IBT potable water to nonpotable water use
DQIbtWUseP
Flow
IBT potable water to potable water use
DQOSWwtp
Flow
Offline storage to wastewater treatment plant
DQResAsr
Flow
Reservoir to ASR
DQResWtp
Flow
Reservoir to water treatment plant
DQSwAsr
Surface water to ASR
DQSwExt
Surface water to external
DQSwWtp
Surface water to water treatment plant
After selecting the button, a new tab is created entitled Table_DQSwExt. On the left side of this tab,
there is a table of the time series for all scenarios, as well as the minimum, average, and maximum
statistics for the time series, and a time step count threshold. The count threshold defaults to zero, but
it can be edited to calculate the count of time steps above a certain value. In the example below, we can
see that the flow out of the watershed exceeds zero for all days in the time period.
jSurface water to exter
Baseline2002 Baseline GCM-3.3 Baseline GCM-3.7 Baseline GCM-4.4 Baseline GCM-5.0 Baseline GCM-6.2
MIN
4.91604281
4.923838995
4.923838995
4.852392972
4.629905929
4.890133854
AVERAGE
113.7807581
140.0302234
140.0302234
124.4070025
105.762474
136.9017693
MAX
394.1073415
486.2379146
486.2379146
439.6757833
382.8236989
483.7298046
COUNT>0
365
365
365
365
365
365
Count Threshold (edit):
o|
1/1/2002
131.4222063
131.1075806
131.1075806
131.3151699
131.2154625
131.1097389
1/2/2002
98.65000245
98.94297298
98.94297298
98.70411448
98.82435458
98.80215908
1./3/2002
105.5583947
105.3437909
105.3437909
105.1340121
105.0347991
105.0600119
1/4/2002
99.60279215
99.29465405
99.29465405
98.97426677
98.91342454
98.89097536
1/5/2002
93.50104626
93.18199199
93.18199199
92.67792243
92.67217684
92.67959862
1/6/2002
92.91998335
93.10734564
93.10734564
92.24913241
92.28638843
92.52314579
1/7/2002
129.9370666
141.3680892
141.3680892
134.5667786
126.534404
140.5475422
As an example, we changed the count threshold to 114 MGD (the average flow of the baseline scenario)
and found that the majority of the scenarios had more time steps with flow out of the watershed
exceeding the baseline average compared to the baseline. However, the GCM 5.0 to -1.7 scenario did
not follow this trend, with fewer time steps exceeding the baseline average compared to the baseline
scenario.
14
ScenCompare instructions and User Guide, July 31s", 2018
-------
Variable *
Description
Unit ~
OptRipj *
OptRipc '
OptRipc ~
OptRip; " OptRipc '
OptRipc '
Data Difference
objective
Objective cost
S/yr
574.0
5533.1
5533.1
5533.1
5533.1
5533.1
1
CRipSetlll
LU Conv: From HRU2 To HRU1 - Cost of riparian buffer land use convers
on set for loads group 1
S/yr
3.6
3.6
3.6
3.6
3.6
3.6
0
CRipSetll2
LU Conv: From HRU2 To HRU1 - Cost of riparian buffer land use convers
on set for loads group 2
S/yr
62.0
62.0
62.0
62.0
62.0
62.0
0
CRipSetllB
LU Conv: From HRU2To HRU1 - Cost of riparian buffer land use convers
on set for loads group 3
S/yr
0.0
0.0
0.0
0.0
0.0
0.0
0
CRipSetl21
LU Conv: From HRU3 To HRU1 - Cost of riparian buffer land use convers
on set for loads group 1
S/yr
3.6
3.6
3.6
3.6
3.6
3.6
0
CRipSetl22
LU Conv: From HRU3 To HRU1 - Cost of riparian buffer land use convers
on set for loads group 2
S/yr
210.0
210.0
210.0
210.0
210.0
210.0
0
CRipSetl23
LU Conv: From HRU3 To HRU1 - Cost of riparian buffer land use convers
on set for loads group 3
S/yr
D.D
0.0
c.c
0.0
0.0
0.0
0
fRinSptH?
CRipSetl33
LU Conv: From HRU4To HRU1 - Cost of riparian buffer land use convers
on set for loads group 3
S/yr
0.0
48.0
48.0
48.0
48.0
48.0
1
LKipbetl4I
lu conv: from HKUi lo hkui - tost ot riparian butter land use convers
on set tor loads group 1
w
CRipSetl42
LU Conv: From HRU5 To HRU1 - Cost of riparian buffer land use convers
on set for loads group 2
S/yr
184.0
184.0
184.0
184.0
184.0
184.0
0
CRipSetl43
LU Conv. From HRU5 To HRU1 - Cost of riparian buffer land use convers
on set for loads group 3
S/yr
0.0
0.0
0.0
0.0
0.0
0.0
0
CRipSetl52
LU Conv From HRU6 To HRU1 - Cost of riparian buffer land use convers
on set for loads group 2
S/yr
12.5
12.5
12.5
12.5
12.5
12.5
0
CRipSetl53
LU Conv. From HRU6T0 HRU1 - Cost of riparian buffer land use convers
on set for loads group 3
S/vr
44.7
44.7
44.7
44.7
44.7
44.7
0
CRipSetl61
LU Conv From HRU8 To HRU7 - Cost of riparian buffer land use convers
on set for loads group 1
S/yr
4.1
4.1
4.1
4J.
4.1
4.1
0
CRipSetl62
LU Conv: From HRU8 To HRU7 - Cost of riparian buffer land use convers
on set for loads group 2
S/yr
62.0
62.0
62.0
62.0
62.0
62.0
0
CRipSetl63
LU Conv: From HRU8 To HRU7 - Cost of riparian buffer land use convers
on set for loads group 3
S/yr
0.0
0.0
0.0
0.0
0.0
0.0
0
CRipSetl71
LU Conv: From HRU9 To HRU7 - Cost of riparian buffer land use convers
on set for loads group 1
S/yr
12.9
12.9
12.9
12.9
12.9
12.9
0
CRipSetl72
LU Conv From HRU9To HRU7-Cost of riparian buffer land use convers
on set for loads group 2
S/yr
122.3
122.3
122.3
122.3
122.3
122.3
0
CRipSetl73
LU Conv: From HRU9 To HRU7 - Cost of riparian buffer land use convers
on set for loads group 3
S/yr
0.0
0.0
0.0
0.0
0.0
0.0
0
CRipSetlSl
LU Conv From HRUlOTo HRU7 - Cost of riparian buffer land use conver
on set for loads group 1
S/yr
1.9
1.9
13
1.9
1.9
1.9
0
CRipSetl82
LU Conv: From HRUlOTo HRU7 - Cost of riparian buffer land use conver
on set for loads group 2
S/yr
45.8
45.8
45.8
45.8
45.8
45.8
0
CRipSetl83
LU Conv: From HRUlOTo HRU7 - Cost of riparian buffer land use conver
on set for loads group 3
S/yr
0.0
0.0
0.0
0.0
0.0
0.0
0
CRipSetlSl
LU Conv: From HRU11 To HRU7 - Cost of riparian buffer land use conver
on set for loads group 1
S/yr
1.6
16
1.6
1.6
1.6
1.6
0
CRipSetl92
LU Conv From HRU11 To HRU7 - Cost of riparian buffer land use conver
on set for loads group 2
S/yr
103.4
103.4
103.4
103.4
103.4
103.4
0
CRip5etl93
LU Conv: From HRU11 To HRU7 - Cost of riparian buffer land use conver
on set for loads group 3
S/yr
0.0
0.0
0.0
0.0
0.0
0.0
0
CRipSetll02
LU Conv: From HRU12 To HRU7 - Cost of riparian buffer land use conver
on set for loads group 2
S/yr
5.5
55
5.5
5.5
5.5
5.5
0
fRraSetlKK
CRipTotal
Total cost of applying riparian buffer land management sets
S/yr
960.6
1008.5
1(308.5
1008.5
1008.5
1008.5
1
3.3.2 Compare Land Management Variables Across Scenarios
In this section, we provide an example of how to compare stormwater BMP decisions using the
historical baseline and two future climate scenarios (GCM 4.4/10.1 [the median scenario] and GCM
6.2/19.4 [an extreme bounding scenario]) when the optimization model is allowed to decide how much
land area to allocate to a stormwater BMP. In this run, the stream loadings target for TN was applied for
all scenarios.
We found the different stormwater BMP decisions made by using the "Compare Scenario Decisions"
button to create the Table Comparison tab. After filtering the "Data Difference" column for values of
"1", we found that, for all scenario runs, the model selected an Infiltration Basin with a design depth of
0.6" to be implemented on the Commercial/lndustrial/Transportation land use on the sand-and-gravel
soil type and till-and-fine-grained deposits soil type (DALu52 and DALull2, respectively).
Variable
Description [j Units
BMPs_Baseline200; T
BMPs_GCM-4.4 h BMPs_
GCM-6.2 \z_
Data Difference j^l
objective
Objective cost $/yr
6427.4676
6543.0085
6649.0394
1
DALu52
Land Area - 0.6" Infill ac
O.OODQ174
53.16350656
114.7482124
1
DALull2
Land Area - 0.6" Infill ac
1090.555717
1104.442104
1104.442104
1
CLuSet2
0.6" Infiltration Basil S/yr
1894.536918
2010.916898
2117.786745
1
CGwPump
Total cost of groundv $/yr
479.1857416
483.2004344
489.1591315
1
CSwPump
Total cost of surface $/yr
120.8216609
121.1416244
121.4615879
1
ClbtW
Total cost of 1BT pots $/yr
362.3282006
357.1545276
350.0369083
1
When analyzing the decision variables more closely, we see that the land use allocation for this BMP
changed between all three climate scenarios. The baseline scenario selected the BMP only on the till and
fine-grained deposits soil type (the value for DALu52 is negligible), and the future climate scenarios
selected the BMP for both the sand-and-gravel and till-and-fine-grained deposits soil types, with a larger
allocation of overall land area for the BMP in the extreme climate scenario (GCM 6.2/19.4).
We also see that the objective cost increased from the baseline scenario to the future climate scenarios,
with the highest objective cost occurring in the extreme climate scenario. The climate comparison
graphs show an extremely close linear relationship between objective cost and increasing precipitation
and temperature with r2 values of 0.9999 and 0.9554, respectively.
18
ScenCompare instructions and User Guide, July 31s", 2018
-------
s>EPA
United States
Environmental Protection
Agency
Office of Research and Development
National Health and Environmental
Effects Research Laboratory
Atlantic Ecology Division
Narragansett, Rl 02882
Official Business
Penalty for Private Use
$300
EPA/600/R-19/039 | July 2018
Recycled/Recyclable
Printed with vegetable-based ink on paper that
contains a minimum of 50% post-consumer fiber
and is processed chlorine free
------- |