TECHNICAL
DOCUMENTATION FOR
THE TEMPERATURE
BINNING FRAMEWORK
Expert and Public Review
Draft
1
April 12, 2021 I EPA 430-R-21-004
2
3 THIS INFORMATION IS DISTRIBUTED SOLELY FOR THE PURPOSE OF
4 PREDISSEMI NATION PEER REVIEW UNDER APPLICABLE INFORMATION
5 QUALITY GUIDELINES. IT HAS NOT BEEN FORMALLY DISSEMINATED BY
6 THE ENVIRONMENTAL PROTECTION AGENCY IT DOES NOT REPRESENT
7 AND SHOULD NOT BE CONSTRUED TO REPRESENT ANY AGENCY
8 DETERMINATION OR POLICY
-------
9 FRONT MATTER
10 Acknowledgements
11 This technical documentation was developed by the U.S. Environmental Protection Agency's (EPA) Office of
12 Atmospheric Programs. As described herein, components of this Temperature Binning Framework are derived
13 from sectoral impact modeling studies produced by many externa I academic experts, consultants, and Federal
14 agencies, including the Department of Energy (DOE) and the National Ocean and Atmospheric Administration
15 (NOAA). Support for the technical documentation's production was provided by Industrial Economics, Inc. EPA
16 gratefully acknowledges these contributions.
17 • Peer Review and Public Review (placeholder)
18 • Recommended Citation (placeholder)
19
20 Data Availability
21 R-code and input/output data for theTemperature Binning Framework and Tool are publicly available at the
22 following site:
23 https://www.epa.gov/cira/temperature-binning-framework
24
-------
2s Technical Documentation For The
26 Temperature Binning Framework
27 Expert and Public Review Draft
28 EPA 4 3 0 - R - 2 1 -004
29 CONTENTS
30 ONE I INTRODUCTION 1
31 1.1 Objective ofthe Framework 1
32 1.2 Intended Use 2
33 1.3 Comparison to Existing Methods 3
34 TWO | TEMPERATURE BINNING METHODOLOGY 6
35 2.1 Methods Overview 6
36 2.2 Available Sectoral Impacts 12
37 2.3 Developing Impact Function Parameters 16
38 2.4 Economic Impacts CaIculation 22
39 2.5 Temperature Binning R Package 25
40 2.6 Interpretation and Key Limitations/Uncertainties of Results 25
41 THREE | CLIMATE IMPACT ANALYSIS USING TEMPERATURE BINNING 30
42 3.1 CONUS Economic Impacts of Climate Change: Results by Degree 30
43 3.2 Adjusting Economic I mpacts for Socioeconomic Conditions 32
44 3.3 Regional Economic Impacts of Climate Change: Results by Degree 33
45 3.4 Physical Impacts of Climate Change: Results by Degree 34
46 3.5 Risk Reduction through Ada ptation: Results by Degree 35
47 3.6 Economic Impacts for a Custom Temperature Trajectory 38
48 3.7 Economic Benefits of Emission Reduction 41
49 REFERENCES 46
50 APPENDIX A | DETAILS OF SECTORAL IMPACT STUDIES A-l
51 A. 1 Sector Data Overview A-l
52 A.2 Health Sectors Data Processing A-3
53 A.3 Infrastructure Sectors Data Processing A-21
-------
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
54 A.4 Water Resources Sectors Data Processing A-38
55 A.5 Electricity Sectors Data Processing A-43
56 APPENDIX B | R TOOL DOCUMENTATION B-l
57 B.l'ciraTempBin'Overview B-l
58 B.2 'ciraTempBin' Function Details B-3
59 APPENDIX C | CLIMATE IMPACT ANALYSIS: ADDITIONAL INFORMATION C-l
60 C.1 Climate Models a nd Scenarios C-l
61 C.2 Socioeconomic Scenariosand Input Data C-3
62 APPENDIX D | INFORMATION QUALITY D-l
63 Ensuring Information Quality D-l
64
-------
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
ONE | INTRODUCTION
The Temperature Binning Framework provides a method of utilizing existing climate change sectoral impact
models to create estimates of the physical and economic impacts of climate change by degree of warming.
These relationships between temperature and impacts in the United States (U.S.) can then be applied to
custom scenarios to efficiently estimate impacts and damages under different emission or policy pathways.
This technical document outlines the underlying theory, design, and structure of the Temperature Binning
Framework. The Framework is implemented by application of open-source code referred to as the
Temperature Binning Tool.
1.1 Objective of the Framework
The Temperature Binning Framework documented in this report builds on approaches demonstrated in
numerous previously published studies to produce physical and economic estimates of climate change
impacts in the contiguous United States (CONUS), for a broad range of the most economically important
impact sectors. The Framework is based on a recently published conceptual paper and demonstration of
the method (Sarofim etal., 2021a), and builds on previous analyses which have established strong
relationships between the effects of warming in CONUS and monetized damages (U.S. EPA 2017; Hsiang et
al., 2017; Martinich and Crimmins 2019; and Neumann etal., 2020).
The main objective of the Framework, and the Temperature Binning Tool which implements the approach,
is to provide estimates of the physical and economic impacts in the U.S. from 21st century trajectories of
temperature and sea level rise. The tool is parameterized using a set of underlying published literature
which relates climate change projections to:
1. Related environmental stressors (e.g., extreme temperatures, precipitation, floods, air quality) to
assess exposure to vulnerable individuals and physical assets;
2. Physical impacts of climate-driven environmental stressors, such as property damage, health
effects, or damaged infrastructure; and
3. Economic processesthat are important to understand the relationship between physical impacts
and economic outcomes, such as reduced economic welfare.
The term "temperature binning" refers to a concept of synthesizing results of climate model-specific
sectoral impact results by temperature change (sometimes using integer degree bins), described more fully
in Sarofim etal. (2021a). The basic concept is to identify the arrival years of a given quantity of warming
from a common baseline period (e.g., 1986 to 2005) for a particular climate model used in a sectoral impact
study and extract associated impact estimates using a broader period (e.g., 11-year bin) centered around
the arrival year. Impacts can then be compared across climate models, or general circulation models
(GCM)s, by quantity of warming. Temperature binning aids comparability of independent analyses by
producing an estimate of the physical or economic impact for a given amount of warming, without
consideration of when that warming occurred or which scenario or climate model was used to develop the
-------
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
estimate. For sectors where impacts are primarily driven by changes in sea level, a similar "binning"
approach is followed, however in these cases bins are defined by set thresholds of sea level rise (e.g., 25
centimeters globally) rather than temperature increments. References to "Temperature Binning" intend to
include both temperature and sea level rise binned results.
The Temperature Binning Framework has the flexibility for expansion of sectoral coverage, as new data
from additional studies become available and meet Framework requirements. For example, the Framework
was developed and originally implemented using nine sectors1 (Sarofim et al., 2021a) from the second
modeling phase of the U.S. EPA's Climate change Impacts and Risk Analysis (CIRA) project and its associated
technical report (EPA, 2017), however, the current version of the Temperature Binning Tool (vl.0)
incorporates the results of sectoral impact studies completed after the 2017 CIRA results, as well as studies
from other research groups. The Framework's flexibility will enable incorporation of additional sectoral
results over time.
1.2 Intended Use
The EPA developed the Temperature Binning Framework and Tool to provide a quantitative storyline of
physical and economic impacts of climate change in the U.S., by degree of warming or custom temperature
trajectory, region, and sector. These applications are intended to support analysis coordinated by EPA;
however, the Tool and its underlying damage functions may be of use toothers working in the field.
Defining the relationship between different levels of warming and the associated impacts is also of interest
to audiences outside the modeling community, including decisionmakers, planners, and the public.
Outputs of the Framework and Tool can readily synthesize the results of a broad range of peer-reviewed
climate change impacts projections and to support analysis of other climate change and socioeconomic
scenarios not directly assessed in the supporting literature. This information is intended to supplement and
complement more aggregate economic impact estimates derived from integrated assessment models, such
as the Social Cost of Carbon (SCC).
For certain sectors, the Temperature Binning Framework can also analyze the potential for adaptation to
reduce the physical and economic impacts of climate change. For sectors with available information, the
potential implications of no adaptation, reactive adaptation, and proactive adaptation response scenarios
can be evaluated.
The use of temperature bins to analyze and present results, rather than results limited to set time periods
and emission scenarios, enhances communication to non-expert audiences without compromising the
value of capturing any potential non-linear relationships between temperature and damages. Estimates of
impacts by degree are also adjusted to account for socioeconomic conditions in the Framework. This
capability allows for exploration of the importance of arrival times for certain thresholds (e.g. how would
impacts associated with 2-degrees of warming differ if 2-degrees of warming occurs in 2050 versus 2090?).
1 The nine sectors in Sarofim et al. (2021a) are Labor, Roads, Extreme Temperature Mortality, Electricity Demand and Supply, Rail, Coastal
Properties, Electricity Transmission and Distribution, Southwest Dust, and Winter Recreation.
Page 2
-------
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
Taking this feature a step further, the Framework is also able to analyze impacts of custom temperature
trajectories, which can be thought of as a series of temperature change and socioeconomic conditions (e.g.
1-degree in 2050, 2-degrees in 2070, 3-degrees in 2090). More details on the method used are provided in
Section 2, and example outputs are provided in Section 3.
In addition, although most of the economic impact literature on which the approach is based was
developed using a consistent setofGCMs, climate scenarios, and socioeconomic inputs, the approach, as
demonstrated in this documentation, is well-suited to incorporate results from other studies. This is
important as the Framework currently only includes a subset of the potential impacts of climate change in
the U.S. The Framework's flexibility to incorporate results from external studies drives a long-term
objective to populate the Tool with impact estimates and functions from the broader climate literature.
This will ensure that the Tool is informed by the best available data and methods, which can then be
revisited and updated over time as scientific and economic capabilities continue to advance.
Finally, this Framework is designed to quantify the sectoral impacts of climate change in the U.S., which
provides insight on how different levels of greenhouse gas(GHG) mitigation can reduce future risks. As
such, this Framework does not address the costs of reducing emissions, which have been well-examined
elsewhere in the literature (e.g., Energy Modeling Forum, 2021). Similarly, the health benefits associated
with reductions in other co-emitted air pollutants are beyond the scope of this Framework. The Framework
also does not capture interactions between sectors (such asthe land-energy-water nexus), including the
potential for compounding or cascading effects across sectors.
1.3 Comparison to Existing Methods
The modeling of climate change impacts typically begins with running a set of emissions or concentration
scenarios (IPCC 2000, Meinshausen 2011, Taylor et al., 2012, IPCC 2013, Hayhoe et al., 2017, Riahi et al.,
2017) through complex earth system models, followed by using the temperature and precipitation outputs
of those climate models as inputs to sectoral impacts models. Scenario-based analysis has been the default
approach to projecting future climate impacts for several decades, and has successfully served asthe
backbone of international and federal climate assessments and special reports (e.g. IPCC 2018, USGCRP
2018), modeling intercomparison efforts (e.g. Knutti and Sedlacek, 2013; Warszawski etal., 2013; Eyring et
al., 2016), and individual modeling studies. The Representative Concentration Pathways (RCP) (Moss etal.,
2010) and the Shared Socioeconomic Pathways (SSP) (Riahi etal., 2017) provide projections over the 21st
century of possible future climates ranging from low to high greenhouse gas concentrations and radiative
forcing, allowing for economic modeling to proceed concurrently with, rather than sequential to, physical
scientific modeling (van Vuuren etal., 2014). However, there are some important limitations and challenges
to relying primarily on the traditional scenario-based approach for driving climate impacts analysis.
One difficulty is that it is challenging for there to be a comprehensive set of scenarios that explore all
potential futures. Greenhouse gas emissions or atmospheric concentrations from these scenarios are used
as inputs to climate models with the goal of producing comparable results. However, when using climate
model output to drive impacts analyses, the differences in climate sensitivity between different models can
Expert and Public Review Draft-Do Not Cite
Page 3
-------
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
have a dominant effect, obscuring the role of other structural differences between the models (e.g.,
different responses of precipitation, cloudiness, stagnation events, or other climatic outcomes)
(Schleussner etal., 2016). An additional challenge is one of communication: scenario names can be
enigmatic for the public, whether it is "A1B" from the SRES scenarios, "RCP8.5" of the RCP scenarios, or
"SSP4-6.0" from the SSP/RCP based scenarios. Characterizing changes in impacts that track with
temperature rather than complex scenarios is more intuitive for non-technical audiences, and more easily
associated with the global temperature targets discussed in international negotiations (IPCC, 2018) or
reported in media stories (World Bank 2013; Plumer and Popovich 2018). Moreover, different research
groups and individual assessments highlight different scenarios that may not be directly comparable across
assessments, whereas temperature changes area stable metric.
To address these challenges, the most common alternative methodology is to discuss climatic impacts by
degree rather than by scenario. The National Research Council (NRC) "Climate Stabilization Targets"
assessment (NRC, 2011) presented most of its finding by degree, noting that "using warming as the frame
of reference provides a picture of impacts and their associated uncertainties in a warming world -
uncertainties that are distinct from the uncertainties in the relationship of C02-equivalent concentrations to
warming." The Intergovernmental Panel on Climate Change (IPCC) 1.5 degrees assessment presented a
comparison of impacts at 2 degrees and 1.5 degrees in order to inform global temperature targets (IPCC,
2018). The IPCC and its contributors also have a long history of presenting risks by degree in the "burning
embers" or "reasons for concern" diagram (Smith etal., 2009; Yohe 2010; O'Neill et al.,2017; IPCC, 2019).
Patterns of climate change are often presented normalized by temperature, asthose patterns are robust
when considering the magnitude of change or the scenario (Tebaldi et al.,2020). Wobus etal. (2018) and
Sanderson et al. (2019) presented future risks in the U.S. by degree of warming for the impacts of extreme
temperatures and extreme precipitation events respectively.
As demonstrated in Sarofim etal. (2021a), designing analyses with relational temperature-impact functions
for a given sector can improve comparability between analyses, yield results in a framework that is more
intuitive for communications purposes, and be used to inform simple computational models that can
rapidly and flexibly estimate impacts by sector for any desired scenario. Compared to other damage
function techniques that may rely on an opportunistic ensemble of climate scenarios and 21st century eras
(e.g., Neumann et al., 2020), the Temperature Binning Framework provides a more reliable fit for small
temperature changes from baseline, i.e., between 0 and 1 degrees C. Thus temperature binning is better
suited to estimating impacts from small changes in temperature from a baseline era (e.g., 1986-2005),
which is important for estimating the impacts of near-term climate change. In addition, the temperature
binning approach provides a capability to examine alternative socioeconomic impact scenarios, with
nuanced non-linear or combinatorial treatment of the effect of socioeconomic drivers on specific sectors,
which is not possible using some econometric techniques.
Work by researchers affiliated with the Climate Impact Lab is also focused on estimating economic impacts
of climate change for the U.S. (e.g., Hsiang et al., 2017; Houser et al., 2015). Similar to some econometric
sectoral analyses included in the Temperature Binning Tool, the Climate Impact Lab sectoral analyses rely
Page 4
-------
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
on interpretation of historical data to identify relationships between climate metrics or events and the
economic impacts that result, which are then applied to project economic impacts for future climate and
event forecasts. Other work in the Temperature Binning Tool relies on process-based simulation models
constructed to reflect physical and economic responses to climate stressors. A key advantage of the
Framework is that it can readily accommodate both types of studies, which provides an opportunity for
significant expansion of sectoral coverage beyond those in the CIRA project and specifically made ready for
incorporation in the Tool. The Climate Impacts Lab work includes some sectors (e.g., violent and property
crime) that should be amenable to inclusion in the Temperature Binning Tool at a future date. Flexibility to
accommodate different types of study methodologies also enables comparison to structural uncertainties
across impacts models estimating impacts for the same sector.
Another class of economic impact estimation tools that include components that are in some ways similar
to the Temperature Binning Framework are integrated assessment models (lAMs - e.g., PAGE, RICE and
DICE, FUND, IMAGE). Some lAMs are used to identify an economically optimal GHG mitigation pathway
which balances marginal costs of GHG abatement with marginal costs of GHG damage. To do so, marginal
abatement cost functions (and GHG offset pools and their costs) are needed, and a means for translating
GHG emissions into temperature pathways. lAMs are generally global in scope, although some estimate
impacts at finer spatial levels. The Temperature Binning Framework, by contrast, addresses only the
impacts associated with a defined temperature and socioeconomic pathway, and, in this application, only
for CON US. Overall, the Temperature Binning Framework provides an efficient and transparent damage
estimation approach that operates independently of lAMsand adds the flexibility to use other means of
determining temperature trajectories. The Framework and Tool also rely on a relatively rich, recent, and
peer-reviewed set of economic damage functions for a large number of U.S. sectors. For that reason, the
Tool can help in responding to relevant policy questions by estimating the effects of an incremental policy
to reduce GHGs, and thereby complement the types of analysis and outputs provided by lAMs.
Expert and Public Review Draft-Do Not Cite
Page 5
-------
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
TWO | TEMPERATURE BINNING METHODOLOGY
This section provides an overview ofthe methodology underpinning the Temperature Binning Framework,
including a discussion of the sectors currently processed for inclusion in the Framework, an overview how
sectoral impact model outputs are processed for the Framework, an outline of how economic impacts are
calculated, an introduction to the Temperature Binning R Tool, and finally, a discussion of key limitations
and uncertainties ofthe method.
2.1 Methods Overview
The Temperature Binning Framework produces economic impacts by degree of warming, which can be
useful for communicating the risks of climate change.2 In addition, by controlling for time dependencies
when creatingthe impact by degree estimates, the temperature binned impacts can be mapped to any
temperature pathway and socioeconomic scenario to create custom impact trajectories.
The first step in the Framework is to develop temperature binned impacts from the underlying sectoral
impact literature. The time dimension ofthe impacts is transformed by either pulling out known
socioeconomic scalars, such asGDP per capita or population, or developing atimeseries of multipliers to
capture complex interactions between non-climate time trends and impacts. Additional details are
provided later in this section. The binned estimates can then be used in a variety of ways to communicate
impacts by degree of warming or impacts associated with any trajectory of temperature change.
Figure 1 shows an example of the application of the Framework. The core concept of the Framework is to
sort economic impact information from the peer reviewed literature in integer degree bins (Box A), so that
it can then be deployed to any temperature pathway (Box B) to provide information on the economic
impacts of that temperature pathway (Box C). The results from a sectoral impact model are processed using
the Temperature Binning Framework to estimate impacts by degree and scaled to produce the data shown
in Box A ofthe graphic. This step, as an endpoint in itself, serves as an important output ofthe tool for
communication of impacts by degree. The temperature pathway being evaluated is defined in Box B
(degrees per year) and by mapping impacts by degree to degrees by year, the Framework produces impacts
by year (Box C).
2 The term 'impacts by degree' should be interpreted to include'impacts by sea level rise increment'for the select sectors where impacts
are driven by sea level rise (i.e., Coastal Property and High Tide Flooding).
-------
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
FIGURE 1. EXAMPLE APPLICATION OF THE TEMPERATURE BINNING FRAMEWORK
Box A shows the impacts per capita by degree of warming for an illustrative example sector and Box B shows an example
temperature trajectory. Box Cpresents the impact trajectory associated with the example temperature trajectory by
mapping impacts by degree to arrival times in the example scenario. The tables to the bottom left provide the example input
data shown in Boxes A andB.
2000 2020 2040 2060 2080 2100
Year
In addition, a custom socioeconomic scenario can then be applied to each of the resulting sector-specific
impacts (which are scaled, per unit - in this example, per capita) in what is called the "re-scaling" step to
produce atimeseries of total impacts. In the simplest case, this would involve multiplying impacts per
capita by population in each year. A more complicated case might also involve the application of additional
growth factors that scale non-linearly with population (e.g., electricity demand growth or road traffic). The
sections that follow provide more detail about the approach and the additional options that are available.
Scope of Temperature Binning Methodology
The Temperature Binning Framework provides estimates for seven regions of the contiguous U.S. at annual
timesteps across the 21st century (2010-2090).3 The underlying climate and impacts data are typically
sourced for years 2006 to 2100 for sectors influenced by temperature and precipitation stressors and 2000
to 2100 for sectors vulnerable to sea level rise. The 2006 start year is the earliest year included in a one-
degree temperature bin for the six core GCMs (i.e., the GCMs used by CIRA sectors) and the SLR sector
models run from the base year 2000. The underlying impact data in the tool covers a range of warming
from zero to six degrees, however higher degrees of warming can be extrapolated, with an understanding
3 The current Tool produces results through 2090 dueto the definition of era runs used to define early and late century estimates. Future
versions could extend to 2100.
Expert and Public Review Draft-Do Not Cite
Page 7
-------
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
of the added uncertainty. The Framework is not designed for estimating effects of cooling, or negative
changes in temperature, relative tothe baseline period, although it does not require temperatures to
monotonically increase over the analysis period.4
Currently, the Tool includes 16 sectors, however the method is designed to be flexible in accommodating
additional sector studies as they become available. Sectoral coverage of the Framework is described further
in Section 2.2.
Defining Binning Windows
The first step in the temperature binning method is to process the underlying sectoral impact model
results. Temperature binned damages are most often calculated from a time series of impacts with a known
associated time series of temperature changes, often defined by a particular GCM and forcing scenario
(e.g., RCP). A smoothed temperature pathway is first developed from the known temperature pathway
using 11-year averages over the period of analysis compared tothe baseline climate era (1986-2005). From
the smoothed pathway, 11-year windows are identified around the first arrival year of each integer one to
six degrees above baseline, and impacts are averaged within each window to represent the corresponding
integer degree of warming. Figure 2 shows the temperature binning windows for six GCMs used in sectoral
models from theCIRA project currently processed for the Framework. 5 Due to computational constraints,
the only scenario used is RCP8.5. RCP8.5 is a pathway with relatively high greenhouse gas concentrations,
leading to substantial warming by 2100. RCP8.5 was chosen to assess a wide range of future temperatures,
and the selection of a higher emissions scenario ensures that this temperature binning approach evaluates
the broadest range of sectoral impacts at higher levels of warming (e.g., 4 or 5 degrees C) in addition to
smaller levels (i.e., an RCP with considerably lower forcing may not reach higher degree bins, therefore
leading to data gaps on the sectoral impact response to higher levels of warming). It is important to note
that the selection of RCP8.5 does not imply a judgment regarding the likelihood of that scenario. Recent
research, such as Christensen et al. (2018) suggests that even in the absence of any global climate policy,
RCP8.5 has a higher forcing than the most likely future concentration pathway.
Sector impact models driven by other GCMs and/or emission scenarios function in the same way: 11-year
windows are defined for each integer degree based on the CONUS temperature trajectory defined by the
climate model employed, compared tothe 1986-2005 baseline era or a custom baseline used in the
relevant work.
4 The current Tool limits user temperature impactstoOto 10 degrees, providing a suggested limit to extrapolation on the high end. GMSL
inputs are limited to 0 to 150cms.
5 The six GCMs included in the CIRA framework were selected to provide a representative range of likely climatefutures - see EPA 2017
for details.
Page 8
-------
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
FIGURE 2. TEMPERATURE BINNING WINDOWS FOR SIXGCMS
This graphicshows the 11-year windows assignedto each integerCONUS temperature change by CIRA GCMforRCP8.5.
Arrival years, ortheyearat which the 11-year moving average reaches the given integer, are listed in each bin.
2010 2020 2030 2040 2050 2060 2070 2080 2090 2100
CanESM2
2011
2033
2048
2062
2076
2091
CCSM4
2011
2037
2059
2077
2091
GISS-E2-R
2026
2052
2082
HadGEM2-ES
2013
2029
2044 2055
2064
2077
MIROC5
2017
2033
2050
2067
2081
GFDL-CM3
2013
2032
2049
2061 2071
2087
Degrees of Warming
¦1 ml 3 4 ¦5 B6
Results are averaged for each degree of warming across all available climate models. Note that some GCMs
do not reach six degrees of warming by the end of the century. Impacts associated with higher degrees of
warming are therefore defined only by those GCMs that reach those levels of warming. For example, as
shown in Figure 2, impacts at six degrees are only available for three of the six CIRA GCMs (CanESM2,
HadGEM2-ES, and GFDL-CM3).6
Indexing impacts to CONUS degrees of warming streamlines the required climate data to run the
Framework compared to detailed impact models that might require more spatially or temporally refined
climate inputs. In doing so, however, the Framework assigns any variation in temporal or spatial
manifestations of climate linked to integer degrees of CONUS warming based on the underlying climate
scenarios used to produce the binned results. For example, Table 1 shows degrees of warming by National
Climate Assessment (NCA) region averaged over the six CIRA framework GCMs (RCP8.5) by integer of
CONUS warming. The bins are defined by average annual temperatures and an infinite combination of daily
or even hourly temperatures can reach the same average annual temperature and the Temperature
Binning Framework will not precisely capture that nuance. For example, a GCM not included in the
calibration of the Temperature Binning Framework may have a different distribution of extreme high and
low temperature days than any of the GCMs that were considered, which could have implications for the
resulting extreme temperature mortality. That is not to say extreme temperatures are not represented in
the Framework; they are present as defined by the underlying climate models. Table 1 also provides the
global mean temperature change from the 1986-2005 baseline, for comparison (see Section 2.4 for more
details on this conversion). Although in this application the Framework utilizes CONUS temperatures, some
audiences may be more accustomed to global temperature changes. Using CONUS temperatures allows for
6 The lack of GCM coverage at higher temperatures may in some cases present inconsistencies in the impacts by degree approach.
Changes in in daily or seasonal temperature, precipitation, and other climatic factors implicitly incorporated in the underlying sectoral
models where it is potentially important (e.g., Southwest Dust), but these patterns of climate hazards may be distinct to individual GCMs.
As a result, temperature bins that are based on different groups of GCMs are likely to display some differences when non-temperature
stressors are influential, such as for sectors that are driven by extreme events. Further research could be needed to assess the potential
importance of thisfactor, but it is also clear that the potential bias is likelyto be smaller for more moderate warming scenarios, where
more GCMs are available. More detail on this point can be found in Sarofim et al. (2021a).
Expert and Public Review Draft-Do Not Cite
Page 9
-------
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
a closer match between the climate variable and impacts but simplified conversion factors can be used to
translate between CONUS and global temperature changes for the purpose of communication.
TABLE 1. AVERAGE REGIONAL TEMPERATURES BY DEGREE OF WARMING
Temperature change by National Climate Assessment region and integer degrees of national (CONUS) warming (Celsius)
from 1986-2005 average baseline, sixGCM average, with corresponding global temperature change.
CON US AT
1
2
3
4
5
6
Midwest
1.367
2.344
3.430
4.466
5.611
6.626
Northeast
1.238
2.252
3.418
4.531
5.490
6.817
Northern Plains
1.081
2.079
3.089
4.192
5.407
6.331
Northwest
0.949
1.822
2.650
3.824
4.658
5.840
Southeast
1.051
1.919
2.900
3.770
4.627
5.548
Southern Plains
1.105
2.172
3.107
3.946
4.941
5.615
Southwest
0.905
1.991
2.872
3.842
4.680
5.738
Global AT
0.424
1.189
1.954
2.718
3.483
4.248
Note: Global temperatures increases from a pre-industrial baseline are 0.454 degrees C higher than the 1986-2005 baseline values presented above.
Precipitation patterns, and therefore precipitation driven impacts, are also represented by degrees of
CONUS temperature change. For precipitation-driven impact sectors, this can result in larger variations
between GCM-specific impacts by degree compared to temperature-driven sectors. Figure 3 shows the
percent change in precipitation compared to baseline for the six CIRA GCMs at two degrees of warming.
The suggested method in this Framework is to calculate impacts using several GCMs and use the average
for interpretation. An alternative method could be to rely upon results from a subset of GCMs that are
known to have similar climate patterns to the scenario of interest. For example, if there is interest in the
implications of a relatively wet future, an analysis using the CMIP5 CanESM2 GCM, the wettest of the CIRA
ensemble, could provide insights.
FIGURE 3. PERCENTAGE DIFFERENCE IN PRECIPITATION FOR THE 2-DEGREE TEMPERATURE BIN
CanESM2
HadGEM2-ES
CCSM4
MIROC5
GISS-E2-R
GFDL-CM3
30
20
10
0
-10
-20
-30
For sectors vulnerable to sea level rise, binning by degree of warming presents challenges for precisely
capturing the links between climate stressors and economic impacts. Degrees of warming are correlated
with sea level rise but non-linearities and time dependencies in the relationship make tying sea level rise
Page 10
-------
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
driven impacts to temperatures a suboptimal option. Instead, these sectors are binned based on 25cm
increments of SLR. The underlying CIRA sea level rise sector studies (Coastal Properties and High Tide
Flooding) estimate economic impacts for six global mean sea level (GMSL) scenarios based on Sweet et al.
2017, ranging from 30cm to 250cm of GMSL rise by the end of the century. Figure 4 shows the arrival times
for 25cm increments of GMSL rise, 25cm to 150cm by end of century. Arrival times for GMSL rise vary
dramatically across scenarios, particularly when comparing the 30cm or 50cm scenarios tothe 100cm to
250cm scenarios. Unlike the temperature models, which represent the same emission scenario with
different characterizations, the GMSL scenarios intend to represent different outcomes. Also note that
while temperature bins are based on CONUS temperature changes, sea level rise is based on global arrival
times. Measures of sea level rise specifically for the U.S. are not commonly used in the literature; therefore,
the Framework uses GMSL. As with the temperature bin indexing, regional and local sea levels are mapped
to GMSL based on the underlying Sweet et al. (2017) models, which include effects such as land uplift or
subsidence, oceanographic effects, and responses of the geoid and the lithosphere to shrinking land ice.
When custom sea level rise scenarios are used in the Framework, the relationship between GMSL and
regional sea levels, and ultimately regional impacts, are mapped implicitly based on the underlying
models.7
FIGURE 4. SEA LEVEL RISE BINNING WINDOWS FOR SIX SCENARIOS
This graphic shows the 11-year windows assigned to each 25cm increment for results from each GMSLscenario (Sweet et
al., 2017). Values calculated using a year2000baseline. Arrival years, or the year at which the 11-year moving average
reaches the given integer, are listed in each bin.
2020 2030 2040 2050 2060 2070 2080 2090 2100
30 cm
2080
50 cm
2052
100 cm
2040
2065
2083
150 cm
2036
2054
2068
2080
| 2092
200 cm
2033
2048
2060 2070
2079
2085
250 cm
2031
2045 2055
2064
2079
Increments of GMSL Rise
25 ¦ 50 l 75 ¦ 100 ¦ 125 ¦ 150
Treatment of Time Dependencies
One of the key characteristics of the Temperature Binning Framework is the ability to analyze changes in
temperature at different points throughout the century and produce impact estimates that scale with
custom population and GDP forecasts. Previous methods have allowed for scaling by presenting impacts as
proportional to GDP (see Hsiang etal., 2017 for example). This method, however, does not account for non-
linearities in the relationship between GDP, population, and impacts (e.g., the value of a statistical life,
which is valued using a non-linear elasticity of GDP per capita) and it does not capture how variations in
population demographics (particularly geographic distribution and age) affect impact estimates.
7 Analyses conducted to support Neumann et al. (2020), Yohe et al. (2020), and Lorieet al. (2020) showed that economic impact results
for the Coastal Property sector were consistent for like increments of SLR across SLR trajectories within about 10 percent tolerance, if
socioeconomic trends are controlled (socioeconomics drives a function for real property value appreciation in the National Coastal
Property Model).
Expertand Public Review Draft-Do Not Cite
Page 11
-------
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
The Framework improves on the traditional scalar approach by explicitly accounting for two components of
time dependencies that can broadly be thought of in terms of quantity and composition, where quantity is
the traditional scalar (e.g., damages per capita or as a percent of GDP) and composition refers tothe
changes invulnerability or exposure within a given population. For example, at a given temperature health,
labor, and recreation impacts in 2010 will differ from those in 2090 based on both the total population and
the demographic composition of the population. These differences are accounted for through a series of
multipliers, some of which link tothe custom GDP and population scenarios. The details of how
socioeconomic drivers are addressed in the underlying studies (Section 2.2) and in the Temperature Binning
Framework (Section 2.3) are described in more detail later in this section.
Future analyses can estimate impacts for a scenario defined by a trajectory of climate change and a
trajectory of "quantity" measures (i.e., GDP and regional population), and mapping the climate change
trajectory tothe timeseries of sector-specific multipliers defined within the Framework. In this way, the
model is flexible to any number of future socioeconomic scenarios but requires minimal data inputs as it
relies on previously determined compositions. See further discussion in Section 2.3.
2.2 Available Sectoral Impacts
The Temperature Binning Tool is a secondary data synthesis application that relies on existing primary
research quantifying sectoral impacts and is designed to accommodate a variety of impact estimates,
including those run with unique climate trajectories, socioeconomic assumptions, and temporal scopes.
Many of the sectoral studies currently processed for this Framework are part of the CI RA framework, and
therefore rely on a consistent set of climate models and socioeconomic scenarios (see Martinich and
Crimmins, 2019; EPA 2017; Neumann et al., 2020; and Sarofim et al., 2021a for more details on sector
studies and the CIRA framework). However, other studies with different climate or socioeconomic
projections can also be integrated into the Framework if the necessary information is available. Necessary
data and specifications include that the underlying study provides regional impacts by degree of warming
(or cm of SLR) that can be scaled for socioeconomic changes and adjusted for other time dependencies
unique to the sectoral impact function. Although ideally the introduced sectors meet all of these
qualifications, there may be instances where methods are adapted to allow for the inclusion of certain
studies and their results. For example, if a study only provides national estimates, impacts could be
distributed tothe regions based on population or another relevant proxy. See Section 2.3 for more
discussion of necessary information.8
The Temperature Binning Framework currently includes 16 sectoral impacts, many with multiple adaptation
scenarios and sub-impacts, as seen in Table 2. This list will continue to evolve as new sector studies are
8 EPA is currently working with study authors to add three additional research studies to the sectoral scope of the Tool: 1) Two new
sectors (violent and property crime; agriculture) and three sectors that overlap with estimates already in the Tool (labor, extreme
temperature mortality, and coastal property) from Hsiang et al. 2017; 2) Coastal wind damage from changes in tropical storm activity,
derived from Dinan (2017); and 3) Inland riverine flooding from an in-process update to Wobus et al. (2019). This expansion in sectoral
scope remains a high priority option for inclusion in a future revision to the Tool.
Page 12
-------
413 published and processed for temperature binning. See Appendix A for more details on the sectors currently
414 processed for the Framework, including citations for the underlying studies.
415 TABLE 2. SUMMARY OF SECTORAL IMPACTS IN THE TEMPERATURE BINNING FRAMEWORK
Sector
Impact Types3
Key Socioeconomic
Driver
Adaptation Scenarios
Air Quality
• Ozone Mortality
• PM2.5 Mortality
• Population
• GDP/capita (VSL)
• 2011 Air Pollutant Emissions Levelb
• 2040 Air Pollutant Emissions Levelb
Labor
• Lost Wages
• Population (high-risk
workers)
• GDP/capita (wages)
• No Adaptation
Roads
All Roads
• Road repair, user cost (vehicle damage), and
delay costs
• Population (traffic)
• No Adaptation
• Reactive Adaptation
• Proactive Adaptation
Asphalt Road
Maintenance*
• Asphalt road surface repairs, temperature
stress only
• None
• No Adaptation
Extreme Temperature
• Heat-related mortality (VSL)
• Cold-related mortality (VSL)
• Age-stratified city
population
• GDP/capita (VSL)
• No Adaptation
• Adaptation
Electricity Demand and
Supply
• Infrastructure expansion costs
• Electricity demand
forecast
• No Adaptation
Rail
• Repair (including equipment and labor) and
delay costs
• Population (passenger
traffic)
• GDP (freight traffic)
• No Adaptation
• Reactive Adaptation
• Proactive Adaptation
Coastal Properties
• Costs related to armoring, elevation,
nourishment and abandonment (including
storm surge impacts)
• GDP/capita (property
values)
• No Adaptation
• Reactive Adaptation
• Proactive Adaptation
High Tide Flooding
• Traffic delays, road elevation
• Population (traffic)
• No Adaptation
• Reasonably Anticipated Adaptation
• Direct Adaptation
Electricity Transmission
and Distribution
Infrastructure
• Repair or replacement of transmission and
distribution lines, poles/towers, and
transformers
• Electricity demand
forecast
• No Adaptation
• Reactive Adaptation
• Proactive Adaptation
Southwest Dust
• All Mortality
• All Respiratory Morbidity
• All Cardiovascular Morbidity
• Asthma ER
• Acute Myocardial Infarction Morbidity
• Age-stratified
population
• GDP/capita (VSL)
• No Adaptation
Urban Drainage
• Upgrading urban storm water infrastructure
• None
• Proactive Adaptation
Valley Fever
• Morbidity - Hospitalization Costs
• Morbidity - Lost Productivity
• Mortality
• Population
• GDP/capita (VSL)
• No Adaptation
Water Quality
• Lost recreational value
• Population
• No Adaptation
Wildfires
• Morbidity from air quality (hospitalization
costs and lost productivity)
• Mortality from air quality
• Response Costs
• Population
• GDP/capita (VSL)
• No Adaptation
Winter Recreation
• Snowmobiling revenues
• Alpine Skiing revenues
• Cross Country Skiing revenues
• Population (potential
recreators)
• Adaptation (defined by
snowmakingfor alpine skiing)
*Non-CIRA study
Blue rows are SLR sectors
Notes:
a. Impacts types refer to the sub-impacts processed for the Framework and available as outputs in the Tool.
b. The two emissions levels in the underlying air quality study are not strictly adaptation scenarios however they are entered into the
Framework using the same structure.
Expertand Public Review Draft-Do Not Cite
Page 13
-------
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
Adaptation Scenarios
The Framework accounts for adaptation by defining the adaptation assumptions for each sectoral impact
and allowing for multiple adaptation scenarios per sector, as available from the underlying sectoral study.
The last column in Table 2 identifies the available adaptation scenarios for each sector currently in the
Framework. The available adaptation options generally fall in three categories:
• No adaptation. The no adaptation scenario represents a "business as usual" scenario. For
econometrically based sectors (e.g., Labor), adaptation is included to the extent that adaptation is
currently occurring. For infrastructure sectors (i.e., Rail, Roads, Electricity Transmission and
Distribution Infrastructure, Coastal Properties, and High Tide Flooding), a no adaptation approach to
infrastructure management does not incorporate climate change risks into the maintenance and
repair decision-making process beyond baseline expectations and practice.
• Adaptation. The adaptation scenario explicitly accounts for some behavioral change in response to
changing climate. The infrastructure sectors include two adaptation scenarios, following Melvin et
al. (2017):
o Reactive adaptation, where decision makers respond to climate change impacts by
repairing damaged infrastructure, but do not take actions to prevent or mitigate future
climate change impacts (a variant on this scenario is the "Reasonably anticipated
adaptation" option for the High-Tide Flooding and Traffic sector, which is defined similarly
to the Reactive scenario); and
o Proactive adaptation, where decision makers take adaptive action with the goal of
preventing infrastructure repair costs associated with future climate change impacts. This
Proactive Adaptation scenario assumes well-timed infrastructure investments, which may
be overly optimistic given that such investments have oftentimes been delayed and
underfunded in the past, and because decisionmakers and the public are typically not fully
aware of potential climate risks (these barriers to realizing full deployment of cost-effective
adaptation are described in Chambwera et al.,2015).
The Framework estimates results for all available adaptation scenarios which allows for evaluation of
impacts associated with a temperature trajectory under a variety of adaptation assumptions. The general
adaptation scenarios considered in the Framework and Tool will not capture the complex issues that drive
adaptation decision-making at regional and local scales. As such, the adaptation scenarios and estimates
should not be construed as recommending any specific policy or adaptive action.
Climate Scenarios in Underlying Models
The CIRA sectors in the Temperature Binning Tool are parameterized based on a set of results from
underlying sectoral models that use one RCP that spans the largest range of future temperature projections
Page 14
-------
450
451
452
453
454
455
456
457
458
459
460
461
462
463
for the 21st century U.S.9 RCPs are identified by their approximate total radiative forcing (not emissions) in
the year 2100, relative to the year 1750. RCPs developed for the IPCC's Fifth Assessment Report released in
2014 include 2.6 W/m2 (RCP2.6), 4.5 W/m2 (RCP4.5), 6.0 W/m2 (RCP6.0), and 8.5 W/m2 (RCP8.5). The
baseline climatic data within the Temperature Binning Tool was created using RCP8.5 to ensure the
broadest possible range of application to both low and high temperature bins. RCP8.5 is a pathway with
relatively high greenhouse gas concentrations, leading to substantial warming by 2100. Note that RCP8.5
does not represent any particular national or global policy, and is used in the Tool because it covers a wide
range of warming levels (low to high). Results for RCP4.5 would likely be comparable, once binned into
comparable integer temperature bins, but RCP8.5 results are employed.10 Although RCP8.5 is preferred for
scenario-based result inputs to the Framework, potential new sectoral studies that are run at different RCPs
or other scenarios are not excluded from the Framework.
The CIRA sectors rely on six GCMs from the fifth phase of CMIP (CMIP5) shown in Table 3: CCSM4, GFDL-
CM3, GISS-E2-R, HadGEM2-ES, MIR05, and CanESM2.n
TABLE 3. GCMS USED BY CIRA SECTORS IN THE TEMPERATURE BINNING TOOL
Center (Modeling Group) Model Acronym References
Canadian Centre for Climate Modeling and Analysis
CanESM2
Von Salzen et al.,
2013
National Center for Atmospheric Research
CCSM4
Gent et al., 2011
Neale et al., 2013
NASA Goddard Institute for SpaceStudies
GISS-E2-R12
Schmidt et al., 2006
Met Office Had ley Centre
HadGEM2-ES
Collins etal., 2011
Davies et al., 2005
Atmosphere and Ocean Research Institute, National Institute for
Environmental Studies, and Japan Agency for Marine-Earth
Science and Technology
MIROC5
Watanabe et al., 2010
Geophysical Fluid Dynamics Laboratory
GFDL-CM3
Donner etal., 2011
9SeetheThird National Climate Assessment (2014)and Climate Impacts Group (2013) for useful descriptions of howthe RCPs compare
to other common scenarios. References: Walsh, J., D. Wuebbles, K. Hay hoe, J. Kossinet al., 2014: Ch. 2: Our Changing Climate. Climate
Change Impacts in the United States: The Third National Climate Assessment, J. M. Melillo, Terese (T.C.) Richmond, and G. W. Yohe, Eds.,
U.S. Global Change Research Program, 19-67. doi:10.7930/J0KW5CXT; Climate Impacts Group, 2013. Making sense ofthe new climate
change scenarios. University of Washington, available at: http://cses.washington.edu/db/pdf/snoveretalsok2013sec3.pdf.
10 In general, studies have found that the sensitivity of impacts for a given temperature level to the specific scenario is low compared to
other sources of uncertainty.
11 Sectors developed for EPA 2017 used only five GCMs (they did not includeGFCL-CM3). Several sectoral models (e.g. Water Quality,
Urban Drainage, etc.) were not updated since 2017 and therefore do not include results for GFDL. These sectors were generally ones with
smaller overall economic impacts.
12 Some of the GCMs in the CMIP5 archive were run multiple times to develop individual initializations for each climate model. In general,
the LOCA dataset provides projections usingthe first initialization of each GCM. However, for the GISS-E2-R model, the LOCA dataset
provided data for RCP4.5 using run #r6ilpl and run #r2ilpl for RCP8.5. The main reasoning for this difference is that the GCM
initializations (raw data from CMIP5) did not provideall ofthe climatedata necessary for doingthe LOCA constructed analog and bias
correction technique. Whilethe usage of different initializations for the GISS-ER-R model could introduce inconsistency, the statistical
differences across runs of the same GCM are dramatically lower than across models, and those differences are further dampened by the
LOCA bias correction. To evaluate the potential that these alternative initializations could introduce inconsistencies, an analysis was
completed com paring the raw #r6ilpl runs for both RCPs and the raw #r2ilpl runs for both RCPs. The results of this comparative
analysis confirmed that the differences are minimal and that it is reasonable to use the LOCA projections.
Expertand Public Review Draft-Do Not Cite
Page 15
-------
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
2.3 Developing Impact Function Parameters
Impact function parameters are sector-specific functions that define impacts by degree, which can then be
applied to any temperature and socioeconomic trajectories. Parameters must be 1) regional (NCA region),
2) scaled by sector-specific, tailored socioeconomic scalars (to allow for custom scenario inputs), 3)
adjusted for other time-dependent factors, where applicable, and 4) available by degree of warming.
The objective of this pre-processing step is to define regional impacts that can be scaled (e.g., impacts per
capita, impacts per road mile), are tied to degrees of warming (or cm of SLR), and can be adjusted for
additional time-dependent aspects of the impact function (e.g., demographic shifts and energy demand
shifts).
Regional Impacts
The Temperature Binning Tool is run at a subnational scale. Results are currently processed and presented
at the regional levels used in the NCA, of which there are seven across the CONUS (see Figure 5). The
underlying sectoral models may estimate impacts at a finer spatial scale (which are then aggregated up to
regions) or, less commonly, over larger regions that are disaggregated. The NCA regions are aggregations of
states, therefore most impacts estimated by administrative boundaries (e.g., county, state, zip code) sum
cleanly to regions. Physical boundaries, such as Hydrological Unit Codes (HUCs)—common in water
resource models, can also be attributed to regions using spatial weighting to account for areas that span
regions. It is not necessary for a sector study to include all regions to work in the Temperature Binning Tool.
Southwest Dust and Winter Recreation, for example, are two studies that are limited to specific regions of
the CONUS. National impacts are calculated by summing over the seven regions.
FIGURE 5. NATIONAL CLIMATE ASSESSMENT (NCA) REGIONS
To scale results, population is input at the regional level, and GDP at the national level. Additional scalars,
such as road or rail miles or property values, are also input at the regional level (see Section 2.4).
Northeast
Southeast
MiOwest
Northern ptams
Southern Plans
Southwes?
Northwest
Page 16
-------
488 The Temperature Binning Tool uses NCA regions for consistency, however there is no methodological
489 reason why another spatial scale could not be used, for example counties. Sectoral impact studies that only
490 produce national estimates can also be used in the Framework, either to produce national results or with
491 impacts allocated across regions using a proxy scalar such as population.
492 Socioeconomic Scalars
493 Total impacts per sector for a given period are a function of climate and socioeconomic drivers. For
494 example, a degree of warming in 2090 could be expected to result in larger health impacts than a degree of
495 warming in 2010 due to projected population growth. One objective of the Temperature Binning
496 Framework is to estimate impacts for custom combinations of climate and socioeconomic projections that
497 capture this difference. The Framework accomplishes this by separating climate effects from other
498 socioeconomic trends during pre-processing in a manner that allows the two to be reassembled in custom
499 scenarios during impacts calculation.
500 There are two ways the Temperature Binning Framework "reassembles" temperature change pathways and
501 socioeconomic trends in custom combinations:
502 1. Mapping impacts directly to regional population and/or GDP per capita, both of which projections
503 can be customized in the Framework; and
504 2. Scaling impacts by a set of time-dependent multipliers that may be influenced by population and
505 GDP, but not in a manner that can be separated outside of the underlying model.
506 The Framework is designed toaccept any defined regional population and national GDP projection over the
507 period of analysis (2010-2090). Table 4 identifies the current sectors for which impacts scale with the
508 defined population and/or GDP projections. Population scalars are applied linearly asa simple multiplier
509 and are most common for health impacts. Scalars representing GDP per capita can be simple multipliers
510 (e.g., wage rates for Labor and Valley Fever sectors, where impacts are multiplied by the ratio of the future
511 year GDP per capita to 2010 GDP per capita) or non-linear elasticities (e.g., VSL for Air Quality, Extreme
512 Temperature, Southwest Dust, Wildfire, and Valley Fever sectors).
513
Expertand Public Review Draft-Do Not Cite
Page 17
-------
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
TABLE 4. SECTORAL IMPACTS LINKED TO CUSTOM SOCIOECONOMIC SCENARIOS
Sector
Link with Regional Population Input
Linkwith GDP per Capita Input
Air Quality
X
X
Extreme Temperature
Xa
X
Laborb
X
Southwest Dust
xc
xe
Water Quality
X
Wildfired
X
xe
Winter Recreation
X
Valley Fever
X
xf
Notes:
a. Scaled to city populations.
b. The underlying laborstudy finds that the numberof high riskworkers is projected to remainconstantthroughout
the century; therefore, labor impacts do not scale with population.
c. Scaled to Arizona, Colorado, New Mexico, and Utah populations.
d. Wildfire mortality and morbidity impacts. Wildfire response costs do not scale with population or GDP per capita.
e. Mortality impacts scale with GDP percapita; morbidity impacts do not.
f. Mortality impacts and lost productivityscale with GDP per capita; morbidity impacts do not.
In some underlying impact studies, population, GDP and other socioeconomic trends are important drivers
of the results, but they cannot be cleanly extracted due to complexities in the underlying studies. For these
sectors, instead of linking tothe custom scenarios provided, impacts are adjusted using a time series of
scalars defined empirically in the underlying studies, shown in Table 5. For example, in the Coastal Property
sector, property values are projected to change over time, and therefore an efficient adaptation option late
in the century may not be efficient early in the century when property values are different. At the same
time, threats early in the century trigger adaptation actions, and therefore the property is no longer
vulnerable later in the century, which could cause damages to decrease over time. The Roads sector
provides another example. Under no adaptation, increases in population lead to increased road traffic
which, in combination with freeze/thaw patterns, drive road surface degradation. These complex factors
are captured in a series of annual scalars per region. Sectors that are population based (such asthe health
impact sectors) can also require time-dependent scalars. For these sectors, the underlying study calculates
impacts for a finer resolution of population than regional totals, and therefore the scalars are used to adjust
risk over time based on socioeconomic trends in the underlying models. For example, the Extreme
Temperature and Southwest Dust studies have age-stratified impact functions, and Winter Recreation
impacts vary by state. The time-dependent scalars provide a way of adjusting for these detailed
demographic shifts over time while still only requiring regional population totals as inputs. The final group
of time-dependent scalars can bethought of as demand growth factors. Impacts for these sectors are tied
to population and GDP growth, but in a spatially resolved and non-linear manner. Electricity demand and
rail and road traffic rise with increasing population but are also influenced by more detailed spatial
distribution of population and demand elasticity, which are part of the underlying process-based models.
Page 18
-------
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
TABLE 5. SECTORAL IMPACTS AND TIME-DEPENDENT SCALARS
Sector
Time-dependent Scalars
Scalar Construction
Electricity Demand and Supply
Electricity demand and supply growth factor
2010 vs. trajectory3
Electricity Transmission and Distribution
Infrastructure
Electricity demand growth factor
Rail
Rail traffic growth factor
Roads
Road traffic growth factor
Coastal Properties
Propertyvaluesand adaptation decisionmaking
2010 vs. 2090
scenario
interpolation15
High Tide Flooding
Road traffic and adaptation decision making
Extreme Temperature
Demographic composition
Southwest Dust
Demographic composition
Winter Recreation
Demographic composition
Notes:
a. Annual series of impacts with socioeconomic change are compared to a constant 2010 socioeconomic scenario
run.
b. Impacts are estimated using constant 2010socioeconomic conditionsand 2090 socioeconomic conditions,
then a ratio is taken between the two and interpolatedfor the interveningyears.
There are multiple methods for constructing time-dependent scalars from the underlying sectoral study
results. The first four sectors listed in Table 5 rely on multipliers constructed as the ratio of future annual
impact projections (i.e., changing climate and changing socioeconomics) versus impacts with a constant
2010 socioeconomic scenario (i.e., changing climate and constant socioeconomics). By comparing the two
runs in each year, the pre-processing can identify the contribution of socioeconomic drivers tothe overall
impact estimate.13 This type of information is most often provided for processed-based sectoral modeling,
where socioeconomic growth can be switched on and off. The last five sectors in Table 5 use time-
dependent scalars based on two runs with constant socioeconomic scenarios, defined by conditions in 2010
and 2090. The 2090 scalar is then calculated as the ratio of estimated impacts using 2090 population versus
2010 population. Scalars for years between 2010 and 2090 are interpolated between the two end points.
This option is less data intensive but does not provide the same level of detail asthe trajectory-based
scalars.14
Some sectors may incorporate both types of scalars (i.e., direct links to population and GDP, and time-
dependent multipliers). Extreme Temperature, Winter Recreation, and Southwest Dust sectors all utilize
both scalars. Other health sectors (including Wildfire, Valley Fever, Air Quality, and Labor) only include the
direct inputs link. This distinction is based on the underlying model structure. Other sectors, such as Urban
Drainage, and the response (suppression) cost portion of Wildfire impacts, do not use any temporal or
socioeconomic scalars.
13 Note that the Tool calculates trajectory-based scalars for every five years (not annually), but the method and Tool would support
annual scalars as well.
14 A possible extension could be to add more intermediate runs, such as 2050 scenario run to add detail to the interpolated scalars.
Expertand Public Review Draft-Do Not Cite
Page 19
-------
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
Economic Valuation Measures
The underlying sectoral models define economic impacts using a variety of valuation measures suited tothe
sector and underlying methods. For some sectors and sub-impacts, valuation represents direct costs, e.g.,
the medical cost to treat an illness, or the expense to repair a road or other physical structure damaged by
a climatic hazard. In other cases where no market transactions take place, such as when an individual dies
prematurely from a climatic hazard or when water quality is impaired, the economic valuation involves the
use of welfare economic techniques. These methodologies are often used to estimate what individuals
would be willing to pay to avoid the risk of an undesirable outcome. The VSL is one such measure used to
value mortality outcomes in many of the health sectors. Table 6 presents the valuation measures used for
each of the sectors and impacts currently in the Framework. The table also indicates which sectoral impact
models directly provide economic impacts, and for which sectors valuation occurs as a multiplier on a
binned physical or unitless impact. For example, VSL is applied to a modeled risk of premature mortality,
while many of the process-based sectors (e.g., Roads, Rail, and Coastal Property) directly estimate
economic impacts, which are binned during pre-processing. Multipliers are preferred, where possible, as
they provide the option to produce physical impacts (e.g., number of deaths avoided), which in turn
provide an alternative method for communicating climate impacts.
TABLE 6. ECONOMIC VALUATION MEASURES BY SECTORAL IMPACT
Sector
Impact
Valuation Measure
Valuation Application
Air Quality
Ozone mortality
VSL
Multiplier on premature
mortality
PM2.5 mortality
VSL
Coastal Properties
Coastal property damage
Propertydamage/adaptation
costs
Direct
Electricity Demand and
Supply
Change in power sectorcosts from
reference scenario
Capital,
operations/maintenance, and
fuel costs
Direct
Extreme Temperature
Extreme cold mortality
VSL
Multiplier on premature
mortality
Extreme heat mortality
VSL
Electricity Transmission
and Distribution
Infrastructure
Stress to transmission and
distribution infrastructure
Repair and replacement costs
Direct
High Tide Flooding
Traffic delays and adaptation costs
due to high tide flooding
Delay costs
Direct
Labor
Lost wages for high-risk occupations
Wages: annual, high risk
workers
Multiplier on hours lost
Rail
Rail impacts, risk of track buckling
Repair and delay costs
Direct
Roads
All Roads
Damage to paved and unpaved road
surfaces
Repair and delay cost
Direct
Asphalt Roads
Maintenance
Road impacts
Repair costs
Direct
Southwest Dust
Hospitalization (acute myocardial
infarction)
Hospitalization costs:
cardiovascular
Multiplier on incidences
Hospitalization (cardiovascular)
Hospitalization costs:
cardiovascular
All mortality
VSL
Page 20
-------
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
Sector
Impact
Valuation Measure
Valuation Application
Hospitalization (respiratory)
Hospitalization costs:
respiratory
Asthma ED visits
Hospitalization Costs: Asthma
Urban Drainage
Proactive costs of improving urban
drainage infrastructure
Repair costs
Direct
Water Quality
Water quality impacts
Lost welfare
Direct
Wildfire
Morbidity
Hospitalization costs
Direct
Mortality
VSL
Multiplier on premature
mortality
Response or suppression costs
Wildfire responsecosts
Multiplier on acres
burned
Winter Recreation
Lost ticket sales from alpine skiing
Lost ticket revenues
Direct
Lost ticket sales from cross-country
skiing
Lost ticket revenues
Lostticketsalesfrom snowmobiling
Lost ticket revenues
Valley Fever
Mortality
VSL
Multiplier on incidences
Morbidity
Cost of illness: Valley Fever
Lost wages
Wages: daily, all workers
Impacts by Degree
The Framework relies on sectoral impact results processed, or "binned", to degrees of warming. After
adjusting for socioeconomic and other time-dependent trends, impacts are mapped to degrees of warming.
Section 2.1 describes the binning process, whereby binning windows for each integer degree of warming
zero to six degrees are defined across atimeseries of impacts, specific totheGCM(s) used in the underlying
sectoral impact model. This process is used to estimate regional impacts by degree when sectoral impact
results are available annually.15
Not all sectoral impact studies produce annual results, either due to computational constraints or the
structure of the underlying model. For example, Urban Drainage and Water Quality, two sectors part of the
CIRA project that were not specifically simulated using the temperature binning arrival times, produce
results only ata set number of eras. Similarly, asphalt roads, a non-CIRA sector, also provide era-level
results. The Framework is flexible to these inputs provided the underlying climate projections are well-
documented and available. For these sectors, bins are defined by first constructing a time series of impacts
using the era-impact pairings, with an added pair for zero damages for the baseline period (1986-2005).
Years within known pairings are linearly interpolated and end of century results are extrapolated linearly
based on the latest two available pairings. Binning windows are defined for the synthetic time series of
impacts using the underlying climate data. This process adds uncertainty through imposing linear
interpolations between known points, and the level of uncertainty is higher when fewer eras of results are
15 The bins shown in Section 2.1 are specific to the six GCMs used in the CIRA framework, downscaled and bias corrected for the LOCA
dataset. When using non-CIRA sectors in the Framework, bins are defined following the same process, which requires access to the
climate data used in the underlying impact analysis. Note that new bins based on integer degree arrival times should be defined for all
outside climate models, even those using the same GCMs, unless they rely on the same LOCA downscaling and bias correction methods.
Expertand Public Review Draft-Do Not Cite
Page 21
-------
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
available (for example, Water Quality impacts rely on 2050 and 2090-era results only, while projections for
Urban Drainage impacts are available for 2030, 2050, 2070, and 2090 eras). Building a synthetic time series
potentially overstates confidence in the shape of the time series, but it allows for the inclusion of a wider
set of potential impact studies, particularly those developed outside of the CI RA framework.
A final consideration in defining impacts by degree is the assignment of baseline periods. The majority of
CIRA sectors use the default climate baseline (1986-2005), but outside studies and select CIRA sectors
define future climate change against different baseline periods. Where possible (i.e., where consistent
baseline data is available), the baseline is shifted to match the Framework default. This is not possible in all
cases, and in those instances, binning windows are developed based on the available baseline.
2.4 Economic Impacts Calculation
The Temperature Binning Framework can be used to estimate climate impacts in several ways, including
impacts by degree, impacts for a specified scenario, and the difference in impacts for two emission
scenarios. This section describes the process for estimating impacts for one or more climate scenarios,
summarized in Figure 6. The Framework begins with one or more emissions trajectories to be evaluated.
Emissions trajectories are translated to global mean temperatures and sea level rise trajectories, using a
reduced complexity climate model, such as Hector or FalRto calculate global mean temperature (Hartin et
al., 2015; Smith et al., 2018) and simple sea level rise models to calculate global mean sea level (e.g., BRICK
from Wong et al., 2017; or the semi-empirical model from Kopp et al., 2016). The Temperature Binning
Framework accepts global mean temperature and translates them to CONUS temperatures using a reduced
form function.16 Using the processed results data from the underlying sectoral studies and defined
socioeconomic scenarios, the Framework calculates regional damages per time step. Results are then
aggregated to the national scale, and when two or more scenarios are analyzed, physical and economic
impact projections under a mitigation scenario are compared to estimated impacts under a reference case.
The results can also be used as inputs to other post-processing analyses, such as economy wide models.
16 CONUS Temperature Change = 1.30764 x Global Temperature Change + 0.34057, based on an analysis of global and CONUS
temperatures from the six GCMs used in the CIRA project.
Page 22
-------
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
FIGURE 6. TEMPERATURE BINNING FRAMEWORK SUMMARY
The Temperature Binning Framework starts with an emission trajectory, translated to CONUS temperature changes and
global sea level rise—the primary inputs to the Temperature Binning Tool. Custom populations and GDP inputs also feed into
the tool. The inputs define analysis scenarios which are mapped to the binned sectoral modeling outputs to estimate
regional impacts. These results can be used in a number of post-processing analyses including comparisons of impacts
across policy scenarios or as inputs to economy wide models.
Policy/Scenario
to be Analyzed
Baseline Emissions
Trajectory
Uncertainty in
effectiveness
Baseline
uncertainty J ^
Input Data
Legend
Sources of
uncertainty
f Tool used in
each step J
Input Processing Steps
Temp. BinningTool Inputs
Temp. BinningTool Steps
Sectoral Modeling
Steps
Post-Processing Steps
~ Optional Step
Emissions (A)
GCMs/Climate
sensitivities
>1 r Climate ^ r SLR Scenarios/non- ^ i r Climate
Model (e.g., emissions-based I Model (e.g.,
J Hector) J ^ factors J BRICK) J
Global
Temperature (A)
K
Reduced
Form/GC/W
relationships
CONUS
Temperature (A)
[ *5Er ] [ )
Global SLR (A)
Climate variables at
~ sector model
( GCMs ) spatial scale
f Structural ^1 f Sector
^ Uncertainty J v ^ Models J
Custom Population
Regional Impacts
( Sector Models )
and GDP Inputs
(A)
(
N/A j TB T°o1
]
Difference in Damages
Sectoral Impact
(Baseline - Policy)
Aggregations
Impacts at sector
model spatial scale
Economy-wide modeling
or other post-processing
Defining Climate Scenarios
The Temperature Binning Framework aims to provide reliable climate impact estimates with limited input
requirements to support rapid assessment. To that end, the Framework is flexible in terms of the necessary
climate inputs. Impacts in the Framework are keyed to CONUS temperature change and global sea level
rise, however the minimum required input is global mean temperature change, which can then be
translated to the necessary climate variables within the Framework. The Framework runs on an annual
scale; however, it can work with any timestep of input data by interpolating between known points.
• Temperature Inputs: CONUS or Global temperature change, relative to a 1986-2005 baseline.
Temperature-driven sectors are indexed to CONUS degrees of warming, relative to the 1986-2005
baseline. An annual timeseries of temperatures is preferred, although interpolation (and
extrapolation) can be used to fill in a timeseries from a minimum of two points. CONUS degrees of
warming are used in the Framework because, relative to global temperatures, they provide a closer
Expert and Public Review Draft-Do Not Cite
Page 23
-------
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
link to the local climate stressors influencing the underlying models (Sarofim etal., 2021a). For
some climate models and other sources of temperature trajectories, CONUS degrees of warming
might not be a readily available, and instead the climate scenarios are defined by global
temperature change. The Framework includes a translation function to convert global changes in
temperature (from the 1986-2005 baseline) to CONUS changes in temperature, based on an
econometric relationship derived from the LOCA dataset.1718
• Sea Level Rise Inputs: Global mean sea level, relative to a 2000 baseline or no custom input. Sea
level-driven damages are indexed to global mean sea levels, relative to a 2000 baseline. Although
considered a separate input from the temperature pathway, the sea level rise inputs should be
consistent with the temperature pathway to maintain consistency across all sectoral results. In
some cases, the same models used to develop temperature trajectories might also produce sea
level rise pathways. In other cases, sea level rise pathway could be developed in a separate model
from the same emissions trajectory used to develop the temperature trajectory. Finally, if the input
climate scenario does not include a defined sea level pathway, the Framework includes a translation
function, based on data from Sweet et al. (2017), to estimate global mean sea level from global
temperatures.19
The Framework provides flexibility by allowing for the use of any number of climate models to produce the
temperature and sea level rise inputs. Reduced complexity climate models (Nicholls et al.,2020; Sarofim et
al., 2021b) work well in this setting as they can emulate some of the aggregate response characteristics of
GCMs within seconds, allowing for exploration into a range of scenarios, uncertainties, and small
perturbations to the climate system. Reduced complexity climate models are defined by a series of
parameters that can be optimized to emulate more complex GCMs, retaining the computationally efficiency
and ease of use while replicating the global mean outputs of these models.
An example (used in the case studies presented in Sections 3.6 and 3.7) uses Hector, a reduced-form global
climate carbon-cycle model, to develop temperature inputs from a custom emission scenarios.20
Defining Socioeconomic Trajectories
The Framework allows custom regional population and national GDP inputs, which influence impact
projections through the socioeconomic scalars described in Section 2.3. In the absence of custom scenarios,
the Framework applies default population and GDP projections that are consistent with theCIRA project's
scenarios (see EPA 2017 for more details), and therefore align with the scenarios used in many of the
17 U.S. Bureau of Reclamation, Climate Analytics Group, ClimateCentral, Lawrence Livermore National Laboratory, Santa Clara University,
Scripps Institution of Oceanography, U.S. Army Corps of Engineers, and U.S. Geological Survey, 2016: Downscaled CMIP3 and CMIP5
Climate Projections: Release of Downscaled CMIP5 Climate Projections, Comparison with Preceding Information, and Summary of User
Needs. Available onlineat http://gdo-dcp.ucllnl.org/downscaled_cmip_projections/techmemo/downscaled_climate.pdf. Data available at
http://gdo-dcp.ucllnl.org/downscaled_cmip_projections/.
18 Global to CONUS mean temperature change estimated as CONUS Temp =1.30764*Global Temp+0.34057.
19 GMSL = 15.908 x Global Temperature Change0-9716, where global temperature change is relative to 1986-2005 baselineand GMSL is
relative to a 2000 baseline. The function used in the temperature input stage totranslate global temperatures to CONUS temperatures is
inverted to produce global temperature from CONUS inputs when necessary.
20 For more information on Hector, see:
Page 24
-------
665 underlying sectoral impact studies. The default population scenario is based on the national-level UN
666 Median Population projection (United Nations 2015), disaggregated to the county-level using EPA's ICLUSv2
667 model (Bierwagen etal., 2010; EPA 2017b) and reaggregated to NCA regions for this analysis. GDP
668 projection is defined by the EPPA, version 6 model (Chen et al., 2015), using the aforementioned UN
669 Median population projection for the U.S. (United Nations 2015) and the 2016 Annual Energy Outlook
670 reference case (USEIA 2016) for the U.S. through 2040.
671 Defining Output Sets
672 The Framework calculates impacts across multiple dimensions: year, region, sector, sub-impact, and
673 adaptation scenario. Results can be aggregated across these dimensions to meet the needs of analysis, with
674 the exception of the adaptation scenarios, which represent different states of the world and should not be
675 summed.
676 The results can feed into a number of post-processing analyses, including comparisons across emission
677 policies or climate sensitivities, or fed into economy-wide models.
678 2.5 Temperature Binning R Package
679 The Temperature Binning Framework is implemented through the use of a process tool developed in R, a
680 popular free software environment for statistical computing and graphics. The Rtool consists of an R
681 Package available for download and installation at https://www.epa.gov/cira/temperature-binning-
682 framework.The Rtool allows users to import custom temperature, sea level rise, national GDP, and regional
683 population scenarios into R from Excel and CSV files, and to use these scenarios to project annual average
684 damages throughout the 21st century due to climate change for any and all sectors available in the tool.21
685 More specifically, the main output of the Rtool is a dataset of average annual economic damage estimates
686 at single year intervals from 2010 through 2100 for each sector, adaptation, impact type, model (GCM or
687 SLR scenario), and region.22 The Rtool also provides options for aggregation of outputs (i.e., summing all
688 impact types for each sector), calculating discounted damages (annual and cumulative), plotting damages
689 over time, and saving output tables.
690 Additional information in the R Package is provided in Appendix B.
691 2.6 Interpretation and Key Limitations/Uncertainties of Results
692 The Temperature Binning Framework provides a method of utilizing existing climate change sectoral impact
693 studies to create time independent estimates of the physical and economic impacts by degree of warming.
694 EPA designed the Framework and Tool to readily synthesize the results of a broad range of peer-reviewed
695 climate change impacts projections, and to support analysis of other climate change and socioeconomic
696 scenarios not directly assessed in the supporting literature. Projected physical and economic impacts from
697 the Tool are intended to provide insights about the potential magnitude of climate change impacts in the
21 The R tool, by default, calculates projected damages for all sectors in the tool. Alternatively, users have the option to select a specific
set of sectors for which to calculate damages.
22 The main output includes information about the underlying input scenario, for user reference.
Expertand Public Review Draft-Do Not Cite
Page 25
-------
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
U.S. However, none of the estimates should be interpreted as definitive predictions of future impacts and
damages. Instead, the intention is to produce estimates of future effects using the best available data and
methods, which can then be revisited and updated overtime as science and modeling capabilities continue
to advance.
The results provided by the Temperature Binning Framework and Tool should be used and interpreted with
consideration of the following limitations and uncertainties. These include:
• Coverage of Sectors and Impacts: The Temperature Binning Tool incorporates a subset of all known
climate change impacts, chosen based on current understanding, available data and methods, and
demonstrated magnitudes of economic effect. EPA (2017) further identifies additional sectors and
impacts not addressed in the broader CIRA project, including cross-sectoral impacts, and
incomplete coverage of effects within sectors-those are also omitted here. Examples of key
missing sectors include the impacts of climate change on agriculture, migration, and political
instability. Sectors that have already been modeled and incorporated into the Tool can be improved
to capture more of the physical and/or economic effects, such as by expanding the population
coverage and characterization of adaptation for extreme temperature-related mortality. Using
more than one sectoral model to estimate impacts for a given sector would also lead to increased
understanding of the results (and increased confidence, if the models are in agreement). Further,
the sectoral studies largely omit potentially important indirect effects (e.g., how does road and
electricity distribution infrastructure failure affect health and welfare, particularly during extreme
events?), and the potential for cascading failures. As a result, the scope of estimates included in this
tool very likely underestimates impacts that could be reasonably expected under future climate
scenarios.
• Uncertainty in Warming Arrival Time: As described in Section 2 of this technical documentation,
damage functions have typically been estimated using a single or limited number of emissions
scenarios, and a limited number of climate models. However, there may be differences in a 2-
degree scenario depending on how and when that level of warming is reached (Sarofim et al.,
2021a). Aspects of this question have been addressed by several researchers (Tebaldi and Knutti
2018, Ruane et al., 2018, Baker et al., 2018, Tebaldi et al., 2020): generally, these studies find that
the sensitivity of impacts for a given temperature level to the specific scenario is low compared to
other sources of uncertainty, but that there are important sensitivities in the C02 concentration,
aerosol concentration, and interannual variability across scenarios. One physical difference that can
arise when a temperature threshold is reached later in time is that the land-ocean differential
would generally be expected to be smaller as a scenario approaches stabilization: this potential
issue is partially addressed by using national rather than global temperatures for the binning. In
general, while use of global temperatures improves the ability to associate results with the
temperature targets discussed in climate policy, the use of national temperatures reduces scatter,
improves fit, and allows better emulation of GCMs that might not have been used to generate the
sector-specific damage functions. Note that there are some sectors where in theory an impact
Page 26
-------
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
would be better associated with global temperatures than national temperatures, where the
impacts are a function of large scale weather pattern or ocean circulation changes.
• Path Dependency- Sectors where the impacts are a function of cumulative exposure can be more
challenging to represent in a temperature binning context. For example, sea level rise is a function
of the integration of heat absorption by the ocean and melting of land ice, and so is a more complex
function of temperature over time, compared to health impacts from heat stress that occur in
direct response to local ambient weather. There are approaches to addressing some of these
difficulties: for example, financial smoothing is applied in the Framework for one-time adaptation
costs or threshold damages to avoid discontinuities in the relationship between temperature and
damages. Pre-loaded in theTool is an approach that estimates a relationship between centimeters
of sea level rise and damages, and then applies a separate function to relate time and temperature
to calculate sea level rise. For these reasons, the use of an external, custom SLR scenario which can
be derived from tools such as BRICK is encouraged, or methods that acknowledge the inherent
physical time dependency of SLR on a temperature path, rather than contemporaneous
temperature. The approach used in the Framework is appropriate for reduced complexity modeling
and benefits analysis but might not be as amenable for general communication purposes.
• Rate of Change and Direct Effects of GHGs - This approach does not capture impacts that are a
function of rate of change, rather than absolute change (though there is a paucity of studies on that
topic in general). Nor does it capture impacts that are a direct function of greenhouse gas
concentrations, such as ocean acidification, C02 fertilization, or ozone resulting from methane
oxidation in the atmosphere. Impacts that are sensitive tonon-GHG factors, such as aerosol
emissions or land-use changes, will also be challenging to emulate. Inter-sectoral interactions (such
as the land-water-energy nexus) and cascading risks would also be difficult to capture in this
framework. Some of these challenges are surmountable - for example, Schleussner etal. (2016)
shows temperature slices for coral reefs under assumptions of coral adaptation for both 2050 and
2100 in order to account for the ability of coral to adapt to slower rates of change, and O'Neal et al.
(2017) created reasons for concern figures for rate-of-change and C02 concentration as a
complement to the temperature-based reasons for concern - but require more complexity in
approach.
• Cross-Sectoral Impacts Modeling: With some exceptions, the sectoral impact models that were
simulated to develop functions used in the Temperature Binning Tool were run independently of
each other. Some sectors, however, could reasonably interact with each other. These intersectoral
effects are not reflected in the Tool.
• GHG Emissions and Climate Scenarios: While emissions and climate scenarios are inputs to the
Temperature Binning Tool, uncertainties in these components of climate impact studies should be
acknowledged as contributing to uncertainty in the outputs of this Tool. Further, only six GCMs are
used in most of the underlying sectoral impact modeling results that feed into the Tool. For those
sectors where there is little variation in impacts resulting from the different GCM, such as Winter
Recreation, there can be reasonable confidence when extrapolating to other, untested GCMs. For
Expertand Public Review Draft-Do Not Cite
Page 27
-------
775
776
111
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
other sectors with more GCM-to-GCM variability, such as for climate impacts on the Rail sector,
confidence in such extrapolation will be lower. More work understanding the causes of that
variability, such as whether it is related to GCM-specific changes in precipitation or temperature
changes in specific regions, could enable more sophisticated extrapolations.
• Variability in Societal Characteristics: The results in the Tool do not separately report impacts for
socially vulnerable populations, nor analyze how individual behavior affects vulnerability to climate.
Results from the Tool are aggregated across demographic groups.
• Feedbacks: The socioeconomic scenarios that drive the modeling analyses do not incorporate
potential feedbacks from climate change impacts to the socioeconomic system (e.g., changes in
albedo from land use change or increased GHG emissions resulting from vegetative changes) nor
from sectoral damages to the economy (e.g., significant expenditures on protective adaptation
measures, such as seawalls, would likely reduce available financial capital to the economy for other
productive uses).
• Geographic Coverage: The primary geographic focus of this Tool is the contiguous U.S., excluding
Hawai'i, Alaska, and the U.S. territories. This omission is particularly important given the unique
climate change vulnerabilities of these high-latitude and/or island locales. In addition, some sectoral
analyses assess impacts in a limited set of major U.S. cities (e.g., Extreme Temperature Mortality),
and incorporation of additional locales would gain a more comprehensive understanding of likely
impacts.
• Climate Drivers -The Temperature Binning Tool relies on estimation of impacts based on annual
temperature indexing. While changes in daily or seasonal temperature, precipitation, and other
climatic factors are used to drive the underlying sectoral models where it is relevant (e.g.,
Southwest Dust), these stressors other than annual temperature changes are only implicitly
included within the temperature bins developed from each of the six GCMs considered. More detail
on this point can be found in Sarofim et al. (2021a). Additionally, because not all GCMs reach six
degrees by 2100, average impacts at higher temperatures are driven by a subset of GCMs that may
not reflect average climate driver characteristics. This could lead to non-linearities at higher
temperatures that are driven by the mix of available climate models rather than non-linearities in
impact response.
• Adaptation - Depending on the sector, the Temperature Binning Tool includes impact estimates
that employ a variety of assumptions regarding adaptive responses to climate impacts. For some
sectors, the tool includes estimates that incorporate adaptation, in which they reflect the current
understanding of the climate risk mitigating effects of adaptation in the literature. Much of the
current literature reflects impact estimates developed for limited or no adaptation conditions. This
is in part because the historical experience of climatic conditions such as those expected to be
experienced in the future is limited, so mechanisms of adaptation maybe poorly understood. Asa
result, reliably quantified estimates of the impact of adaptation are not currently available for all
sectors addressed in this tool. In addition, in many sectors adaptive action to date has been
Page 28
-------
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
surprisingly slow, even where literature suggeststhat the economic benefits of taking action to
mitigate climatic risks exceed the costs-for example, in response to coastal risks of accelerated
storm surge and sea level rise (Lorie et al., 2020). For some sectors, including many of the
infrastructure sectors and the heat stress mortality sector, the tool provides the user an option to
assess impacts under alternative human response scenarios, including no adaptation, reactive
adaptation (to repair damage but without forward planning to avoid future damage), and proactive
adaptation (including action and investment in risk mitigation based on some level of foresight of
future conditions). For several sectors where the Tool does not provide options toassessthe effects
of alternative adaptation assumptions, such as labor or winter recreation, adaptation is partially
represented in the underlying results used to create the damage functions. For example, the
econometric methodology used in the labor analysis would capture any extreme temperature
adaptations employed by outdoor industries in the base period. Also, the winter recreation analysis
included the use of artificial snow creation/blowing.
The general adaptation scenarios considered in the analyses of the Framework will not capture the
complex issues that drive adaptation decision-making at regional and local scales. As such, the
adaptation scenarios and estimates presented in all sections of this report should not be construed
as recommending any specific policy or adaptive action.
Expertand Public Review Draft-Do Not Cite
Page 29
-------
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
THREE | CLIMATE IMPACT ANALYSIS USING TEMPERATURE BINNING
This section demonstrates the capabilities of the Framework to evaluate climate impacts for the 16 sectors
currently processed in the Tool. Specifically this section provides examples of the ability of the Framework
to estimate sectoral impacts by degree (for CONUS and by region, as economic and physical impacts, and by
adaptation assumption), adjust impacts for socioeconomic conditions, and finally, using theToolto
estimate the potential benefits of emissions reductions using two custom temperature trajectories. These
results are for illustration purposes only, do not reflect analysis of any particular policy or action, and
should be interpreted with a consideration of the uncertainties described in Section 2.6. See Appendix C for
more information on input scenarios used in this section.
3.1 CONUS Economic Impacts of Climate Change: Results by Degree
As discussed in Section 1, presenting impacts by degree of warming provides intuitive anchors for non-
technical audiences and supports comparison across modeling efforts. Estimating impacts by degree is also
the first step in developing impact trajectories for a custom temperature scenario (i.e., Box A in Figure 1).
Figure 7 shows CONUS-level annual impacts by degree for each of the 16 processed sectors for each of the
GCMs used in the underlying study. Results reflect the "primary" adaptation scenario for sectors with
multiple adaptation options available, which are chosen to best represent a continued "business as usual"
adaptation response (see Section 3.5 for more details). The figure shows the Framework's ability to capture
non-linearities in the relationship between temperature and impacts. While some sectors have consistently
increasing impacts as temperatures increase, others taper off or accelerate at higher temperatures,
particularly at 6-degrees. For most sectors there is a strong consistency across GCMs (see Table 7 for more
examples of the average and range of impacts across GCMs). Results across the GCMs generally have larger
differences (as a percent of mean and in absolute terms) at higher degrees of warming. By producing both
average impacts and GCM-specific results, the Framework allows for analysis of some of the uncertainties
listed in Section 2.6, particularly around arrival times for degrees of warming and GHG emissions and
climate scenarios.
Variation in results across GCMs is highest in sectors where impacts are driven by a climate stressor
correlated with, but not directly linked to, mean temperature. Examples include sectors vulnerable to
extreme temperatures (e.g., Extreme Temperature; Rail, which is sensitive to frequency of daily max
temperature above a threshold), and those vulnerable to precipitation (e.g., Air Quality, which is sensitive
to the frequency of days with rain, which affects particulate matter formation; Roads, where impacts are
driven by extreme precipitation and freeze-thaw cycles; Urban Drainage, which is driven by extreme
precipitation events; and Valley fever, which is sensitive to combinations of monthly temperature and
precipitation that lead to aridity). Inthese cases, GCM-specific projections of temperature and precipitation
can lead to differentiated results. For example, the GCM CanESM2 projects much wetter conditions than
other models in Western U.S. at higher levels of warming, leading to a reduction in aridity and ultimately a
lower projected Valley Fever impact than other GCMs. The sea level rise variability for Coastal Properties is
related to difficulties disentangling path dependencies in this process-based sector, as noted in Section 2.
-------
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
FIGURE 7. NATIONAL ECONOMIC IMPACTS BY DEGREE ($MILLIONS)
Impacts by CONUS degree of warming (Celsius) relative to the 1986-2005 baseline, under2020socioeconomic conditions, in
millions of $2015. Results for High Tide Flooding, Extreme Temperature, Roads, Rail, Coastal Properties, and Electricity
Transmission and Distribution Infrastructure reflect th e primary adaptation scenarios (see Section 3.5). Each series
represents the underlying GCM or sea level rise scenario (Coastal Properties and High Tide Flooding). Sectors are ordered by
their average 5-degree impacts; note that the y-axis scalar varies by row.
LO
O
C\l
in
S_
o
Q
CO
c
o
_co
o
CO
Q.
E
200,000
100,000
Impacts by Degrees and GCMs
Air Quality High Tide Flooding Extreme Temp
Roads
30
00£oastal Property Electricity Inf. Southwest Dust Urban Drainage
Valley Fever Winter Recreation Water Quality Asphalt Roads
N 'V 'b
-------
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
Section 2.4, and adjusting for average warming up tothe common baseline period (1986-2005) -0.45
degrees Celsius. 1.5 and 2.0 degrees of global warming relative to pre-industrial are equivalent to 2.9 and
3.5 degrees ofCONUS warming from the 1986-2005 baseline, respectively.
FIGURE 8. PROJECTED NATIONAL IMPACTS FOR GLOBAL TEMPERATURE CHANGES RELATIVE TO PRE-
INDUSTRIAL ERA
Impacts by sector for 1.5 and 2.0 degrees Celsius of warming globally relativeto a pre-industrial baseline(2.9 and3.5
degrees CO NUS relative to 1986-2005) under2090socioeconomic conditions. Dots represent the average estimate across
GCMs (or sea level rise scenarios) and bars represent the range of GCM-specific (sea level rise scenario-specific) results.
$20
$16
$12
$8
$4
-$o
f i
////////^
* #
/
\ef
r
3.2 Adjusting Economic Impacts for Socioeconomic Conditions
A key feature of the Framework is the ability to analyze impacts for any degree of warming that account for
changing socioeconomic conditions, for example, both total population and the demographic composition
of population at the time of evaluation. Table 7 shows annual economic impacts at 2-and 3-degrees of
CONUS warming under 2020 and 2090 socioeconomic conditions. In all sectors that include socioeconomic
adjustments (Asphalt Roads and Urban Drainage impacts are not influenced by socioeconomic conditions),
impacts are greater in 2090 than 2020. For health sectors where impacts are primarily driven by mortality,
a function of population and GDP per capita, estimated impacts approximately double over the century at
the same degree of warming.
Page 32
-------
900 TABLE 7. PROJECTED NATIONAL ECONOMIC IMPACTS AT 2- AND 3-DEGREES OF WARMING IN 2020 AND
901 2090 ($BILLIONS)
902 Impacts at 2-and 3-degrees of CONUS warming (Celsius) relative to the 1986-2005 baseline, under2020 and2090
903 socioeconomic conditions, in billionsof$2015. Low and high GCM projected values shown below the average estimate. Note
904 that impacts for Asphalt Roads and Urban Drainage a re not adjusted for any time dependencies.
Sector
2-Degrees
3-Degrees
2020
2090
2020
2090
Air Quality
$20.6
$15.6 to $25. 7
$42.3
$35.1 to $49.5
$25.2
$19.3 to $31.1
$52.9
$44.8 to $61.0
Asphalt Roads
$1.2
$1.1 to $1.3
$1.6
$1.4 to $1.7
Coastal Properties
$3.8
$2.8 to $4.8
$5.5
$3.5 to $7.4
$5.5
$4.4 to $7.8
$7.9
$4.8 to $12.0
Electricity Demand and Supply
$5.2
$3.2 to $7.4
$7.2
$4.5 to $10.4
$9.6
$6.3 to $12.3
$13.3
$8.7 to $17.0
Electricity Transmission and
Distribution
$5.7
$4.4 to $6.9
$8.8
$6.8 to $10.3
$7.4
$6.8 to $8.0
$11.4
$10.6 to $12.5
Extreme Temperature
$13.0
$7.2 to $15.7
$25.7
$14.7 to $31.8
$26.1
$14.3 to $34.0
$52.3
$28.8 to $68.8
High Tide Flooding and Traffic
$17.5
$11.7 to $29.2
$32.0
$21.3 to $53.9
$36.8
$31.3 to $48.8
$67.0
$57.5 to $88.0
Labor
$8.9
$6.4 to $11.0
$23.1
$16.5 to $28.6
$14.3
$10. 7 to $19.1
$37.2
$27. 7 to $49.6
Rail
$3.6
$1.8 to $5.5
$11.8
$6.0 to $18.2
$6.1
$2.1 to $16.1
$20.3
$7.0 to $53.2
Roads
$9.9
$6.8 to $15.5
$10.8
$7.7 to $17.1
$17.4
$12.1 to $29.4
$19.0
$13.1 to $31.8
Southwest Dust
$2.3
$1.4 to $3.0
$5.3
$3.2 to $6.9
$3.5
$2.6 to $4.4
$7.9
$6.1 to $10.0
Urban Drainage
$4.1
$3.2 to $5.8
$4.1
$2. 7 to $5.6
Valley Fever
$1.9
$1.8 to $2.0
$4.3
$4.0 to $4.4
$2.8
$2.5 to $3.1
$6.1
$5.5 to $6.9
Water Quality
$1.3
$1.1 to $1.5
$1.7
$1.4 to $2.0
$2.0
$1.7 to $2.5
$2.6
$2.2 to $3.2
Wildfire
$8.3
$6.2 to $11.0
$16.3
$12.1 to $21.9
$11.2
$8.3 to $14.0
$22.3
$16.3 to $28.1
Winter Recreation
$1.4
$1.1 to $1. 7
$1.7
$1.4 to $2.1
$2.0
$1.7 to $2.5
$2.4
$2.0 to $3.0
905 3.3 Regional Economic Impacts of Climate Change: Results by Degree
906 The Framework produces impact projections at the regional level which can help inform potential
907 adaptation planning and communicate risks. Figure 9 presents examples of impacts by degree for the
908 sectors with the highest impacts at 2-degrees of warming in each region. Air Quality, the sector with the
909 largest estimated national damages at 2-degrees, is the largest sector regionally for the Northeast,
910 Southwest, and Southeast. However, when looking within and across regions, the projected magnitude of
911 the largest sectors varies significantly: Air Quality impacts in the Southeast reach over $150 billion per year
912 by end of century, while the largest sector in the Northwest (Wildfire) reaches just over $6 billion annually.
Expertand Public Review Draft-Do Not Cite
Page 33
-------
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
913 The GCM averages in Figure 9 also highlight the ability of the Framework to capture non-linearities in the
914 relationship between temperature and economic impacts.
915 FIGURE 9. LARGEST PROJECTED REGIONAL ECONOMIC IMPACTS BY DEGREE ($BILLIONS)
916 This figure shows impacts by degree ofCONUS warming in Celsius relative to the 1986-2005 baseline (in billions of $2015 for
917 the largest economic impact sector in each region. Results representthe average across GCMs and a re shown for2020 and
918 2090 socioeconomic conditions. Note that the scales of the y-axes vary by panel.
g^g —•—2020 -*-2090
920 3.4 Physical Impacts of Climate Change: Results by Degree
921 The Framework also produces physical impact measures for sectors where economic impacts are estimated
922 through multipliers on physical outcomes (see the last column of Table 6 in Section 2.3). Physical impact
923 measures provide another method of communicating climate impacts: for example, premature mortality
924 can be an easier concept for some audiences to grasp compared tothe VSL. As with economic impacts,
925 physical impacts are also adjusted for socioeconomic conditions (primarily through population and
926 demographic composition). Table 8 shows the available physical impacts at each degree of warming under
927 2090 socioeconomic conditions. These impacts scale linearly with the analogous economic impacts either
928 through VSL (premature mortality), wildfire suppression costs (acres burned), or weather exposed (high-
929 risk) industry wages (work hours lost).
930
Page 34
-------
931 TABLE 8. PROJECTED NATIONAL ANNUAL PHYSICAL IMPACTS BY DEGREE: 2090 SOCIOECONOMIC
932 CONDITIONS
933 Available physical impacts in the Framework include premature mortality, acres burned, and work hours lost. Impacts
934 assume 2090 socioeconomic conditions. Impacts presented by CONUS degree change (Celsius)from thel 986-2005 baseline.
Physical
Degree Change (CONUS in Celsius)
Value
Sector
1
2
3
4
5
6
Total
2,150
4,542
5,962
9,295
12,923
23,143
Air Quality
Ozone
506
1,102
1,489
2,205
3,020
3,590
PM2.5
1,644
3,440
4,473
7,089
9,903
19,553
Premature
Mortality
Extreme
Temperature
Total
633
1,688
3,432
5,305
7,336
10,852
Cold-related
-32
-49
-57
-64
-70
-71
Heat-related
666
1,736
3,489
5,368
7,406
10,923
Southwest Dust
169
169
348
519
622
850
Valley Fever
19
134
276
395
459
531
Wildfire
69
481
1,007
1,386
1,650
1,769
Acres Burned
Wildfire
575,549
2,302,194
2,666,904
3,030,677
3,388,595
3,613,215
Work Hours
Lost (thous.)
Labor
146,480
306,183
493,191
700,401
943,401
1,284,056
Note: Values presented are direct outputs from the Tool. Results do not reflect an implied precision in the estimates or a
determination of significantfigures.
935 3.5 Risk Reduction through Adaptation: Results by Degree
936 As noted in Section 2.2, the Framework incorporates a capacity to generate and report analytically
937 consistent results by degree for multiple adaptation scenarios, to the extent adaptation scenarios were
938 analyzed and reported in the underlying literature. In general, adaptation options are available atthree
939 levels: No Adaptation (sometimes better characterized as historical levels of adaptation, depending on the
940 sector); Reactive Adaptation, where adaptative action is taken but without advance planning or foresight;
941 and Proactive Adaptation, where all cost-effective adaptations, including those involving planning and
942 foresight about future climate conditions, are undertaken. The general adaptation scenarios considered in
943 the analyses of this report will not capture the complex issues that drive adaptation decision-making at
944 regional and local scales. As such, the adaptation scenarios and estimates presented in all sections of this
945 report should not be construed as recommending any specific policy or adaptive action.
946 There are six sectors currently processed for the Framework where an adaptation option is operable, most
947 of these are infrastructure sectors: Coastal Properties, Electricity Transmission and Distribution
948 Infrastructure, Extreme Temperature, High Tide Flooding and Traffic, Rail, and Roads. The capacity to
949 consider adaptation scenarios enables analysis of the value of adaptation in reducing future climate
950 damages, reflecting the impact of different assumptions about how effectively society might adapt as
951 climate changes manifest. Similar to results summarized in Section 3.1, adaptation scenario results can be
952 generated by degree and GCM, for custom climate inputs, and for custom socioeconomic scenario inputs.
953 Illustrative results for the six sectors with adaptation options are presented in Table 9 below, and in bar
954 chart form in Figure 10. Both exhibits use 2020 socioeconomic scenario inputs, and present average results
Expertand Public Review Draft-Do Not Cite
Page 35
-------
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
across GCMs, but Table 9 provides results for six CONUS integer degree bins, while Figure 10 focuses on the
2-degree bin results. Shaded rows in Table 9 indicate the "primary" adaptation response assumption as
identified in the underlying literature. In the infrastructure sectors, a "No Adaptation" assumption is
generally considered to reflect little or no implementation of potentially cost-effective options to minimize
damage, so while it is informative, it may not be considered the most likely response in the long-term. On
the other end of the spectrum of adaptation response, a "Proactive" assumption requires collective
planning, upfront expense for future benefit (therefore requiring financing), and sometimes requiring
perfect foresight. Therefore, this scenario may not be considered the most likely response. For the Extreme
Temperature sector, the adaptation option is characterized as an illustrative sensitivity analysis, assuming
that all of the 49 largest U.S. cities are assumed to have the mortality incidence function of one of the
hottest and best adapted U.S. cities (Dallas, TX) - but without consideration of the likely costs incurred to
achieve lower susceptibility, such as increased deployment of air conditioning, or other physiological or
technical barriers to achieving the high level of adaptation capacity observed in Dallas. Asa result, the
Adaptation scenario for Extreme Temperature is not considered to be the primary result, or most likely
response, for all cities.
Results in Table 9 follow expected patterns of damage magnitude. Estimates are higher for higher degrees
of warming, and lower as adaptation effort increases. One exception is seen in the result for Reactive and
Proactive Adaptation in the Coastal Properties sector, for the 1-and 2-degree bins, where Proactive
Adaptation scenario results are slightly largerthan Reactive Adaptation scenario results. Inthis sector,
reactive adaptation is limited to structure elevation, which is a very cost-effective method for mitigating
storm surge risk, but which does not address permanent inundation of properties from gradual sea level
rise. Proactive Adaptation, however, includes the option to armor shorelines with seawalls - protecting
properties from both storm surge and permanent inundation, but at higher cost. In the low temperature
bins, the underlying model chooses armoring in the Proactive Scenario, but with low levels of SLR, the full
expected benefits are not realized unless higher temperatures (and sea levels) are realized. For 3-degree
and higher bins, however, the results revert tothe expected pattern, and Proactive Adaptation results
represent the lowest estimated damages. Overall, the results presented in Table 9 and Figure 10 support
the conclusion that adaptation assumptions are influential to damage results.
Page 36
-------
984 TABLE 9. PROJECTED ANNUAL IMPACTS BY ADAPTATION SCENARIO
985 This table presents annual impacts by sector and adaptation scenarios for integer degree changes in CONUS temperature (1
986 to 6 degrees Celsius) from the 1986-2005 baseline, under2020 socioeconomic conditions. Impacts are presented in billions
987 of $2015.
CONUS Degree Change in Celsius
Sector
Adaptation Scenario
1
2
3
4
5
6
Coastal
Properties
No Adaptation
$2.9
$7.0
$12.3
$18.4
$29.4
$41.3
Reactive Adaptation
$1.5
$3.8
$5.5
$7.0
$11.6
$18.7
Proactive Adaptation
$1.7
$4.1
$5.5
$6.3
$7.0
$7.6
Electricity
No Adaptation
$4.1
$6.0
$8.2
$10.4
$12.4
$14.6
Transmission
Reactive Adaptation
$3.9
$5.7
$7.4
$9.0
$9.0
$10.2
and Distribution
Proactive Adaptation
$2.9
$3.6
$4.1
$5.1
$5.5
$6.6
Extreme
No Adaptation
$4.8
$13.0
$26.1
$40.2
$55.5
$82.0
Temperature
Adaptation
$0.5
$2.2
$6.4
$12.5
$21.2
$37.3
High Tide
No Adaptation
$49.5
$121.4
$265.7
$462.2
$661.5
$808.8
Flooding and
Reasonably Anticipated Adaptation
$7.1
$17.5
$36.8
$67.7
$98.0
$122.2
Traffic
Direct Adaptation
$2.3
$5.7
$7.8
$6.9
$7.4
$6.9
No Adaptation
$1.7
$3.6
$6.8
$10.5
$21.0
$38.4
Rail
Reactive Adaptation
$1.9
$3.6
$6.1
$8.8
$16.8
$30.8
Proactive Adaptation
$0.1
$0.2
$0.5
$1.0
$1.3
$2.1
No Adaptation
$13.9
$64.8
$138.4
$244.4
$336.4
$421.1
Roads
Reactive Adaptation
$4.9
$9.9
$17.4
$29.0
$32.5
$47.9
Proactive Adaptation
$5.3
$7.4
$5.7
$6.4
$4.7
$4.9
Note: Shaded rowsare "primary" results, or best representative of a continued "businessas usual"
adaptation response.
988
Expertand Public Review Draft-Do Not Cite
Page 37
-------
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
FIGURE 10. PROJECTED ANNUAL IMPACTS BY ADAPTATION SCENARIO AS A PERCENT OF NO ADAPTATION
IMPACTS, 2-DEGREES IN 2020
Forsectors wherethe underlying sectoral study simulates multiple adaptation scenarios, theplotsin this figure present
impacts under each scenario as a percent of no adaptation impacts (e.g. where no adaptation equals 100 percent). Labels
show total impacts in billions of $2015, and forthe adaptation scenarios, labels show percent decrease in impacts relative to
no adaptation. Impacts are estimated for a 2-degree Celsius temperature change (CONUS) relative to the 1986-2005
baseline and 2020socioeconomic conditions.
Coastal Property
No Adaptation Reactive Proactive
Adaptation Adaptation
Electricity Trans, and Distribution
High Tide Flooding
100%
75%
50%
25%
No Adaptation Reactive Proactive
Adaptation Adaptation
1 $121.4
$17.5
(-86%)
$5.7
¦ ¦ I
No Adaptation Reasonably
Direct
Anticipated
Adaptation
Adaptation
Rail
Roads
Extreme Temperature
100%
75%
50%
25%
No Adaptation Reactive Proactive
Adaptation Adaptation
1 $64.8
$9.9
$7.4
(-85%)
(-89%)
No Adaptation Reactive
Proactive
Adaptation
Adaptation
100%
75%
50%
$13.0
$2.2
(-83%)
No Adaptation
Adaptation
3.6 Economic Impacts for a Custom Temperature Trajectory
The Framework can use the defined impacts by degree with socioeconomic adjustments to define an
impact trajectory (i.e., impacts associated with a set of combinations of socioeconomic conditions, defined
by year, and degrees of warming). Using the example process outlined in Figure 1 in Section 2.1, this step
involves defining a temperature trajectory (Box B) and calculating results by mapping impacts by degree of
warming to the defined trajectory (Box C).
The Tool accepts custom temperature trajectories informed by emissions scenarios from any climate
model. Reduced complexity climate models, such as Hector and FaIR (Hartin et al., 2015; Smith et al., 2018)
are designed to calculate global temperature trajectories based on a custom emissions scenario.
Specifically, the emissions are converted to concentrations where needed to calculate a change in global
radiative forcing from which a change in global mean temperature is calculated. Global mean temperature
changes are converted to CONUS temperature changes within the Temperature Binning Tool.
Page 38
-------
1009 Figure 11 shows an example temperature trajectory used to demonstrate this capability, defined as the
1010 GCAMv5.3 reference scenario (Calvin et al.,2019; JGCRI 2020), with five different climate sensitivities (ECS)
1011 calculated within Hector. The bold line represents the central scenario (ECS 3 - reflecting global warming of
1012 3°Cfora doubling of atmospheric C02 concentrations), and the dotted lines show temperature pathways
1013 for the same scenario under alternative climate sensitivities (ECS 1.5, ECS 2, ECS 4.5, and ECS 6). The
1014 efficiency of the Framework allows for evaluation of multiple temperature pathways, which supports
1015 uncertainty analysis, for example, across climate sensitivities.
1016 FIGURE 11. EXAMPLE TEMPERATURE PATHWAYS FOR IMPACT EVALUATION
1017 CONUS degrees of warming relative to a 1986-2005 baseline for the GCAMv5.3 reference scenario, across five climate
1018 sensitivities (ECS). ECS 3 is the central case and ECS 1.5 to ECS 4.5 represents the likely range in IPCC (2013).
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
Running each of the five scenarios through the Framework results in the impacts trajectories shown in
Figure 12. The impacts shown are annual impacts, summed across sectors and regions, that reflect the
combination of temperature and socioeconomic condition trajectories. Sea level rise results are estimated
using the temperature to sea level rise conversion function described in Section 2.4. The highlighted zone
represents the likely range of impacts based on the climate sensitivity results. The projections represent
impacts across 15 modeled sectors (excluding Asphalt Roads which is an alternative method to the Roads
sector). While the shapes of the impact curves appear similar to the shape of the temperature trajectories,
they are much steeper. For the central case (ECS 3), temperatures range from 0.7 degrees to 3.5 degrees of
warming through 2090 (a range of a factor of five), while impacts range from $33 billion to $454 billion (a
Expertand Public Review Draft-Do Not Cite
Page 39
-------
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
1029 factor of over 13) reflecting non-linearities of impacts in response to temperatures and changing
1030 socioeconomic conditions.23
1031 FIGURE 12. PROJECTED NATIONAL ANNUAL ECONOMIC IMPACTS ASSOCIATED WITH A CUSTOM
1032 SCENARIO ($B ,-LIONS)
1033 CONUS total annual impacts across 15 modeled sectors (excluding Asphalt Roads) for the GCAMv5.3 reference case with five
1034 climate sensitivities (ECS). ECS 3 is the central case and ECS 1.5 to ECS 4.5 represents the likely range (shown as the shaded
1035 orange region). Impacts presented in billions of $2015.
$1,000
$900
$800
IT $700
O
CO
52 $600
i/i
u
OJ
CL
E $500
CO
3
"Z.
S $400
lo
ZJ
c
< $300
$200
$100
$0
1036
1037 The Tool calculates impacts by year, region, sector, sub-impact, and adaptation scenario, allowing for
1038 analysis of custom scenarios across any of these dimensions. For example, Figure 13 shows projected
1039 annual impacts of the GCAMv5.3 reference scenario by sector.24 Under this illustrative scenario, Air Quality
1040 is the sector with the largest impacts, accounting for 25 percent of total modeled impacts in 2090. Health-
1041 related impacts of climate change on Air Quality experiences a 2.5-fold increase between 2050 and 2090
1042 while High Tide Flooding and Traffic, the second largest modeled sector, increases 4-fold over the same
1043 period. The largest seven sectors make up approximately 86 percent of total annual impacts in 2090,
23 Note also in Figure 11, in 2090 ECS 6 reaches temperatures above five degrees and in Figure 12 for thesame climate sensitivity and
time period there is a jump in projected impacts. This may reflect the influence of GCMs dropping out at the 5-degree mark (of the CIRA
GCMs, only CanESM2, HadGEM2-ES and GFDL-CM3 reach 6-degrees bytheend of the century).
24 Figure 13 and all presented results aggregated across sectors use the primary scenario, as identified inTable 9, for sectors with multiple
adaptation scenario runs. Asphalt Roads Maintenance impacts are excluded from aggregated results due to overlap with the Roads
sector.
Page 40
-------
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
though it is important to note that the modeled sectors only represent a portion of all impacts from climate
change.
FIGURE 13. CONUS ANNUAL ECONOMIC IMPACTS ASSOCIATED WITH A CUSTOM SCENARIO BY SECTOR
($BILLIONS)
CONUS annual impacts by modeled sectors (excluding Asphalt Roads) fortheGCAMv5.3 reference case for the central
climate sensitivity (ECS 3). Impacts presented in billions of $2015.
$500
$450
$400
$350
CO
c
O
= $300
CO
V)
£ $250
Q.
£
$200
c
c.
<
$150
$100
$50
$0
Others
Air Quality
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060 2065 2070 2075 2080 2085 2090
3.7 Economic Benefits of Emission Reduction
The Tool can be run for multiple temperature trajectories defined by multiple climate sensitivities to
explore uncertainty (demonstrated in Section 3.6) or emissions scenarios to estimate how impacts may
change in response toGHG emissions reductions. This section demonstrates the latter capability,
comparing the GCAMv5.3 reference scenario presented in the previous section to an illustrative scenario
defined by a 90 percent reduction in global C02 emissions (relative to the reference scenario) by 2100.
Relative to 2020 C02 emissions, emissions increase by 70 percent in the reference scenario and decrease by
69 percent in the hypothetical emissions reduction scenario by 2090. Figure 14 shows C02 emissions
pathways for each scenario.
Expertand Public Review Draft-Do Not Cite
Page41
-------
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
1060 FIGURE 14. GLOBAL ANNUAL C02 EMISSIONS UNDER THE ILLUSTRATIVE REFERENCE AND EMISSIONS
1061 REDUCTION SCENARIOS
1062 Global fossil fuel and industrial CO2 emissions under the GCAMv5.3 reference and 90 percent emissions reduction scenarios,
1063 in petragrams of carbon peryear(PgC/Yr). Emissions reductions begin in 2026 and reach a 90 percent red uction relative to
1064 the reference scenario in 2100.
orouDCTirMLncorH,^tr^oro^DcnrNjLncoTH'^trv-omti50^rMLncorH'd"r^o
rHrHTHTHrNi(N(NfOrnm,^t^^TtLnLnLnv£>^vx>r^.r*-r^.p-.oooocc
-------
1071 FIGURE 15. TEMPERATURE PATHWAYS FOR AN ILLUSTRATIVE EMISSIONS REDUCTION SCENARIO
1072 RELATIVE TO A REFERENCE
1073 Left, CONUS degrees of warming relative to a 1986-2005 baseline for the GCAMv5.3 90percent CO 2 emissions reduction
1074 scenario, across five climate sensitivities (ECS). Right, difference in CONUS degrees of warming between the GCMv5.3
1075 reference scenario (see Figure 11) and the 90percent emissions reduction scenario. ECS 3 is the central case and ECS 1.5 to
1076 ECS 4.5 represents the likely range (IPCC 2013).
S 25
I
1 2
GCAMv5.3 90% Emissions Reduction Scenario
o 0*
Temperature Difference (Reference -90% Reduction)
o
2010 2015 2020 2025 2030 2035 2040 204S 2050 2055 2060 2065 2070 2075 2080 2085 2090
0 — ¦' -
2010 2015 2020 2025 2030 7035 2040 2045 7050 2055 7060 2065 7070 7075 7080 7085 7090
— fC-S 1.5 — . FfS 2 ECS,3 FCS 4.5 GCSjjB - LCS.1.5 — • 1X5.2 USJ ¦ — LCSJV!> ECS_6
1078 Figure 16 shows the resulting difference in projected economic impact between the two illustrative
1079 emissions scenarios.25 The temperature pathways begin to diverge in 2030 and the projected difference in
1080 annual impacts reaches over $10 billion by 2050 (ECS 3). The net benefits of the hypothetical GHG
1081 reduction scenario increase over the century as the temperature pathways diverge. Also important is that
1082 the economic benefits per degree reduction from the reference scenario also increase over the century as
1083 impacts are non-linear, with more pronounced acceleration of benefits per degree at higher temperatures.
25 Note that differences in annual impacts are calculated as the difference in projected impacts, not the impacts associated with the
difference in temperature, the latter of which would not account for non-linearities in the sectoral impact functions.
Expert and Public Review Draft-Do Not Cite
Page 43
-------
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
1084 FIGURE 16. PROJECTED NATIONAL ANNUAL ECONOMIC EFFECTS OF EMISSIONS REDUCTIONS ($BILLIONS)
1085 Reduction in CONUS total annual impacts across 15 modeled sectors (excluding Asphalt Roads) for the hypothetical
1086 GCAMv5.3 90percent CO2 emissions reduction scenario relative to the reference case with five climate sensitivities (ECS).
1087 ECS 3 is the central case and ECS 1.5 to ECS 4.5 represents the likely range (shown as the shaded orange region). Impacts
1088 presented in billions of $2015.
$450
$400
$350 ^—
U1
1089
1090 Figure 17 shows the projected effects of emissions mitigation under the hypothetical 90% reduction
1091 scenario by sector. The High Tide Flooding and Traffic sector is projected to experience the largest benefits
1092 (i.e., avoided damages), with benefits to the Air Quality sector showing the second largest effects. A
1093 comparison of the distribution of benefits by sector to the distribution of impacts under the reference
1094 scenario in Figure 13 shows that High Tide Flooding and Air Quality swap places as the first and second
1095 largest sectors. This is a result of the difference in sensitivity of impacts for these two sectors for the
1096 specific changes in temperature explored in this comparison.
Page 44
-------
1097 FIGURE 17. PROJECTED NATIONAL ANNUAL ECONOMIC BENEFITS OF HYPOTHETICAL EMISSIONS
1098 REDUCTIONS BY SECTOR ($BILLIONS)
1099 Reduction in CO NUS annual impacts by modeled sectors (excluding Asphalt Roads) for the illustrative GCAMv5.3 90percent
1100 CO 2 emissions reduction scenario relative to the reference case for the central climate sensitivity (ECS 3). Impacts presented
1101 in billions of $2015.
1102
1103
1104
Expertand Public Review Draft-Do Not Cite
Page 45
-------
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
REFERENCES
Baker, H.S., Millar, R.J., Karoly, D.J.,et al. 2018. Higher C02 concentrations increase extreme event risk in a
1.5 °C world. Nature Climate Change, 8:604-608. h
Calvin, K., Patel, P., Clarke, L.,etal. 2019. GCAM v5.1: representing the linkages between energy, water,
land, climate, and economic systems, Geosci. Model Dev., 12:677-698.
https://doi.org/10.5194/gmd-12-677-2019.
Christensen P., Gillingham K., Nordhaus, W. 2018. Uncertainty in forecasts of long-run economic
growth. Proceedings of the National Academy of Sciences, USA. 115(21):5409-5414. doi:10.1073/pnas.
1713628115.
Chambwera, M., Heal, G., Dubeux, C., et al. 2014. Economics of Adaptation. Chapter 17 in the Working
Group II contribution to the IPCC Fifth Assessment Report Climate Change 2014: Impacts, Adaptation,
and Vulnerability.
Collins, W.J., Bellouin, N., Doutriaux-Boucher, M.,etal. 2011. Development and evaluation of an Earth
system model-HadGEM2. Geoscience Model Development, 4:1051-1075.
Davies, T., Cullen, M.J.P., Malcolm, A.J., et al. 2005. A new dynamical core for the Met Office's global and
regional modelling of the atmosphere. Quarterly Journal of the Royal Meteorological Society, 131:1759-
1782.
Dinan, T., 2017. Projected increases in hurricane damage in the United States: the role of climate change
and coastal development. Ecological Economics 138, 186-198.
https://doi.Org/10.1016/j.ecolecon.2017.03.03.
Donner, L.J.,Wyman, B.L.,Hemler, R.S.,etal. 2011. The dynamical core, physical parameterizations, and
basic simulation characteristics of the atmospheric component AM3 of the GFDL global coupled model
CM3. Journal of Climate, 24(13):3484-3519. https://doi.Org/10.1175/2011JCLI3955.l
Energy Modeling Forum. 2021. Stanford University, Stanford, California.
EPA. 2017. Multi-Model Framework for Quantitative Sectoral Impacts Analysis: ATechnical Report for the
Fourth National Climate Assessment. U.S. Environmental Protection Agency, EPA 430-R-17-001.
EPA. 2017b. Updates to the Demographic and Spatial Allocation Models to Produce Integrated Climate and
Land Use Scenarios (Iclus) (Version 2). U.S. Environmental Protection Agency, Washington, DC.
EPA/600/R-16/366F. Available online at:
htt ps://cfpub. epa.gov/ncea/icl us/recordisplay. cfm?deid=322479
Page 46
-------
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
Eyring, V., Bony, S., Meehl, G.A., etal. 2016. Overview of the Coupled Model Intercomparison Project Phase
6 (CMIP6) experimental design and organization. Geoscientific Model Development, 9(5):1937-1958.
Gent, P.R., Danabasoglu, G., Donner, L.J., etal. 2011. The community climate system model version 4.
Journal of Climate, 24:4973-4991.
Hartin, C.A., Patel, P., Schwarber, A.,et al. 2015. A simple object-oriented and open-source model for
scientific and policy analyses of the global climate system-Hector vl. 0. Geoscientific Model
Development, 8(4):939-955.
Hayhoe, K., Edmonds, J., Kopp, R.E.,etal. 2017. Climate models, scenarios, and projections. In: Climate
Science Special Report: Fourth National Climate Assessment, Volume I [Wuebbles, D.J., D.W. Fahey, K.A.
Hibbard, D.J. Dokken, B.C. Stewart, and T.K. Maycock (eds.)]- U.S. Global Change Research Program,
Washington, DC. USA. 133-160. doi:10.7930/J0WH2N54.
Hsiang, S., Kopp, R.,Jina, A., et al. 2017. Estimating economic damage from climate change in the United
States. Science, 356(6345): 1362-1369.
IPCC. 2000. Special Report on Emissions Scenarios. A Special Report of Working Group III of the
Intergovernmental Panel on Climate Change. Cambridge University Press. 570 pp.
IPCC. 2013. Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth
Assessment Report of the Intergovernmental Panel on Climate Change. 1535 pp., Cambridge
University Press, Cambridge, UK and New York, NY. doi:10.1017/CB09781107415324
IPCC. 2018. Summary for Policymakers. In: Global Warming of 1.5°C. An IPCC Special Report on the
impacts of global warming of 1.5°C above pre-industrial levels and related global greenhouse gas
emission pathways, in the context of strengthening the global response to the threat of climate
change, sustainable development, and efforts to eradicate poverty [Masson-Delmotte, V., P. Zhai, H.-
O. Portner, D. Roberts, J. Skea, P.R. Shukla, A. Pirani, W. Moufouma-Okia, C. Pean, R. Pidcock, S.
Connors, J.B.R. Matthews, Y. Chen, X. Zhou, M.I.Gomis, E. Lonnoy, T. Maycock, M. Tignor, and T.
Waterfield (eds.)]. World Meteorological Organization, Geneva, Switzerland. 32 pp.
IPCC. 2019. Summary for Policymakers. In: Climate Change and Land: an IPCC special report on climate
change, desertification, land degradation, sustainable land management, food security, and greenhouse
gas fluxes in terrestrial ecosystems [P.R. Shukla, J. Skea, E. Calvo Buendia, V. Masson-Delmotte, H.- O.
Portner, D. C. Roberts, P. Zhai, R. Slade, S. Connors, R.van Diemen, M. Ferrat, E. Haughey, S. Luz, S.
Neogi, M. Pathak, J. Petzold, J. Portugal Pereira, P. Vyas, E. Huntley, K. Kissick, M. Belkacemi, J. Malley,
(eds.)].
Joint Global Change Research Institute (JGCRI). 2020. Global Change Analysis Model (GCAM). Github
Repository. Available at:
Expertand Public Review Draft-Do Not Cite
Page 47
-------
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
Knutti, R.,and Sedlacek, J. 2013. Robustness and uncertainties in the newCMIP5 climate model
projections. Nature climate change, 3(4):369-373.
Kopp, R.E., Kemp, A.C., Bittermann, K., etal. 2016. Proceedings of the National Academy of Sciences.
113(11): E1434-E1441.
Lorie, M., Neumann, J.E., Sarofim, M.C.,etal. 2020. Modeling coastal flood risk and adaptation response
under future climate conditions. Climate Risk Management 29.
https://doi.Org/10.1016/i.crm.2020.100233
Meinshausen, M., Raper, S.C.B., Wigley, T.M.,etal. 2011. Emulating coupled atmosphere-ocean and carbon
cycle models with a simpler model, MAGICC6—Part 1: Model description and calibration. Atmospheric
Chemistry and Physics, 11:1417-1456.
Melvin, A., Larsen, P., Boehlert, B.,etal. 2017. PNAS, 114:122-131.
Moss, R.H., Edmonds, J.A., Hibbard, K.A., et al. 2010. The next generation of scenarios for climate change
research and assessment. Nature, 463(7282):747-756.
Neumann, J.E., Willwerth, J., Martinich, J., etal. 2020. Climate damage functions for estimating the
economic impacts of climate change in the United States. Review of Environmental Economics and
Policy, 14(1): 25-43.
National Oceanographic and Atmospheric Administration. 2017. Global and regional sea level rise scenarios
for the United States. NOAA Center for Operational Oceanographic Products and Services, Technical
Report. NOS CO-OPS 083.
Neale, R.B., Richter, J., Park, S., et al. 2013. The mean climate of the community Atmosphere Model
(CAM4) in forced SST and fully coupled experiments. Journal of Climate, 26:5150-5168.
Nicholls, Z.R., Meinshausen, M., Lewis, J., etal. 2020. Reduced Complexity Model Intercomparison Project
Phase 1: introduction and evaluation of global-mean temperature response. Geoscientific Model
Development, 13(11):5175-5190.
National Research Council (NRC). 2011. Climate Stabilization Targets: Emissions, Concentrations, and
Impacts over Decades to Millennia. Washington, DC. The National Academies Press.
https://doi.org/10.17226/12877.
O'Neill, B.C., Oppenheimer, M., Warren, R., et al. 2017. IPCC reasons for concern regarding climate change
risks. Nature Climate Change, 7(l):28-37.
Paltsev, S., Monier, E., Scott, J., et al. 2013. Integrated economic and climate projections for impact
assessment. Climatic Change, doi:10.1007/sl0584-013-0892-3.
Page 48
-------
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
Plumer, B.,and Nadja, P. 2018. "Why Haifa Degree of Global Warming Is a Big Deal" New York Times,
October 7, 2018. Accessed April 20, 2020 from:
https://www. nytimes.com/interactive/2018/10/07/climate/ipcc-re port-half-degree.html
Riahi, K., Van Vuuren, D.P., Kriegler, E. 2017. The shared socioeconomic pathways and their energy, land
use, and greenhouse gasemissions implications: an overview. Global Environmental Change, 42:153-
168.
Ruane, A.C., Phillips, M.M., Rosenzweig, C. 2018. Climate shifts within major agricultural seasons for +1.5
and +2.0 °C worlds: HAPPI projections and AgMIP modeling scenarios. Agric. For. Meteorol. 259:329-
344. doi:10.1016/j.agrformet.2018.05.013.
Sanderson, et al. 2019. Informing Future Risks of Record-Level Rainfall in the United States, Geophyscial
Research Letters, 46(7):3963-3972.
Sarofim, M.C., Martinich, J., Neumann, J.E., etal. 2021a. A temperature binning approach for multi-sector
climate impact analysis. Climatic Change, 165(1):1-18.
Sarofim, M.C., Smith, J.B., Juliana, A.S.,etal. 2021b. Improving reduced complexity model assessment and
usability. Nature Climate Change, 11( 1): 1-3.
Schmidt, G.A., Ruedy, R., Hansen, J.E., etal. 2006. Present-day atmospheric simulations using GISS ModelE:
Comparison to in situ, satellite, and reanalysis data .Journal of Climate, 19:153-192.
Smith, J.B.,Schneider, S.H., Oppenheimer, M., etal. 2009. "Assessing dangerous climate change through an
update of the Intergovernmental Panel on Climate Change (IPCC) "reasons for concern"". Proceedings
of the National Academy of Sciences . 106(11):4133-4137
Smith, C.J., Forster, P.M.,Allen, M., etal. 2018. FAIR vl. 3: a simple emissions-based impulse response and
carbon cycle model. Geoscientific Model Development, ll(6):2273-2297.
Schleussner, C.F., Lissner, T.K., Fischer, E.M., etal. 2016. Differential climate impacts for policy-relevant
limits to global warming: the case of 1.5 C and 2 C. Earth system dynamics, 7(2):327-351.
Taylor, K.E., Stouffer, R.J., Meehl, G.A., et al. 2012. An overview of CMIP5 and the experiment
design. Bulletin of the American meteorological Society, 33(4):485-498.
Tebaldi, C., and Knutti, R. 2018. Evaluating the accuracy of climate change pattern emulation for low
warming targets .Environmental Research Letters, 13(5):55006.
Tebaldi, etal. 2020. Climate model projections from the Scenario Model Intercomparison Project
(ScenarioMIP) of CMIP6. Earth Syst. Dynam. Discuss., , in review.
United Nations. 2015. World Population Prospects: The 2015 Revision. United Nations, Department of
Economic and Social Affairs, Population Division.
Expertand Public Review Draft-Do Not Cite
Page 49
-------
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft- Do Not Cite
U.S. Census Bureau. 2017. Population Estimates Program. Available online at:
https://www.census.gov/programs-surveys/popest.html
U.S. Energy Information Administration. 2016. Annual Energy Outlook.
USGCRP. 2018. Impacts, Risks, and Adaptation in the United States: Fourth National Climate Assessment,
Volume II [Reidmiller, D.R.,C.W. Avery, D.R. Easterling, K.E. Kunkel, K.L.M. Lewis, T.K. Maycock, and B.C.
Stewart (eds.)]. U.S. Global Change Research Program, Washington, DC, USA, 1515 pp.
doi:10.7930/NCA4.2018.
von Salzen, K., Scinocca, J.F., McFarlane, N.A.,et al. 2013. The Canadian fourth generation atmospheric
global climate model (CanAM4). Parti: representation of physical processes. Atmosphere-Ocean, 51,
104-125.
Van Vuuren, D.P., Kriegler, E., O'Neill, B.C.,et al. 2014. A new scenario framework for climate change
research: scenario matrix architecture. Climatic Change, 122(3):373-386.
Warszawski, K., et al. 2013. The inter-sectoral impact model intercomparison project (ISI-MIP): project
framework. Proc. Natl Acad. Sci. USA 111:3228-3232.
Watanabe, M., Suzuki, T., O'ishi, R., etal. 2010. Improved climate simulation by MIROC5: mean states,
variability, and climate sensitivity. Journal of Climate, 23:6312-6335.
Wobus, C., Zheng, P., Stein, J., Lay, C., Mahoney, H., Lorie, M., et al. 2019. Projecting changes in expected
annual damages from riverine flooding in the United States. Earth's Future, 7, 516-527.
https://doi.org/10.1029/2018EF001119.
Wobus, et al. 2018. Reframing Future Risks of Extreme Heat in the United States, Earth's Future,
6(9):1323-1335.
Wong, T.E., Bakker, A.M., Ruckert, K., etal. 2017. BRICK vO. 2, a simple, accessible, and transparent model
framework for climate and regional sea-level projections. Geoscientific Model Development, 10(7):2741-
2760.
World Bank. 2013. Turn Down the Heat: Climate Extremes, Regional Impacts, and the Case for Resilience. A
report for the World Bank by the Potsdam Institute for Climate Impact Research and Climate Analytics.
Washington, DC. World Bank.
Yohe, G., Willwerth, J., Neumann, J.E., et al. 2020. What the Future Might Hold: Distributions of Regional
Sectoral Damages for the United States — Estimates and Maps in an Exhibition. Climate Change
Economics, (2020) 2040002.
Yohe, G. 2010. "Reasons for concern" (about climate change) in the United States, A letter. Climatic
Change, 99:295-302. DOI 10.1007/sl0584-009-9797-6.
Page 50
-------
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
APPENDIX A | DETAILS OF SECTORAL IMPACT STUDIES
Appendix A Table of Contents
A.l Sector Data Overview A-l
A.2 Health Sectors Data Processing A-3
Air Quality A-3
Extreme Temperature A-6
Labor A-9
Southwest Dust A-ll
Valley Fever A-14
Wildfire A-17
A.3 Infrastructure Sectors Data Processing A-21
Coastal Properties A-21
High Tide Flooding and Traffic A-24
Rail A-26
Roads A-29
Asphalt Roads A-33
Urban Dra inage A-35
A.4 Water Resources Sectors Data Processing A-38
Water Quality A-38
Winter Recreation A-40
A.5 Electricity Sectors Data Processing A-43
Electricity Demand and Supply A-43
Electricity Transmission a nd Distribution Infrastructure A-45
A.l Sector Data Overview
This appendix provides additional detail on the sectoral studies currently processed for the Framework and
outlines the processing required to prepare the sectoral study results for inclusion in the Tool. This
appendix will be updated over time as additional sectoral studies and their functions are incorporated into
the Temperature Binning Framework. The sectors are presented in four groups: Health Sectors,
Infrastructure Sectors, Water Resources Sectors, and Electricity Sectors. Sectors within each group often
share data processing methods. Table A-l lists the 16 sectors by the four groups and summarizes the
regional coverage of the sectoral impacts as well as identifies the GCMs used in the sectoral impact models.
Table 2 (impact types, socioeconomic drivers, adaptation scenarios), Table 4 (links to population and GDP
inputs), Table 5 (time dependent scalars), and Table 6 (valuation measures) in the main text also provide
summarized information about the 16 sectors.
-------
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
Application of the Framework is not limited tothe sectors currently processed for the Tool. New sectors
that meet the requirements outlined in Section 2 can be added to the tool following the process
documented in this report. EPA is currently working with study authors to add three additional research
studies to the sectoral scope of the Tool: 1) Two new sectors (violent and property crime; agriculture) and
three sectors that overlap with estimates already in the Tool (labor, extreme temperature mortality, and
coastal property) from Hsiang et al. (2017); 2) Coastal wind damage from changes in tropical storm activity,
derived from Dinan (2017); and 3) Inland riverine flooding from an in-process update to Wobus etal.
(2019). This expansion in sectoral scope remains a high priority option for inclusion in a future revision to
the Tool.
TABLE A-l. REGIONAL COVERAGE AND GCMS USED BY SECTOR
Regional Coverage
Air Quality
Extreme Temperature
_c
Labor
(D
X
Southwest Dust
Valley Fever
Wildfire
Asphalt Roads
(D
i_
Coastal Properties
(J
High Tide Flooding and Traffic
to
TO
Rail
4—
_c
Roads
Urban Drainage
Water
Water Quality
Resources
Winter Recreation
Electricity
Electricity Demand and Supply
Electricity Transmission and Distribution
GCMs Used
-------
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
A.2 Health Sectors Data Processing
Air Quality
This sectoral study estimates mortality
risk associated with changing air
quality; specifically, ozone and fine
particulate matter (PM2.s)
concentrations.
UNDERLYING DATA SOURCES AND LITERATURE
Fann, N. L., Nolte, C. G., Sarofirn, M. C., Martinich, J., & Nassikas,
N.J. (2021). Associations between simulated future changes in
climate, air quality,and human health. JAMA Network Open, 4(1).
DohlO.lOOl/jamanetworkopen.2020.32064
This analysis uses air quality surfaces (i.e., concentrations in response to changes in meteorology and
emissions) and concentration-response functions employed by Fann etal. (2021) to quantify PM2.5-and
ozone-attributable premature mortality. Mortality is monetized using the value of statistical life (VSL). Two
simulated air pollutant emissions inventories are considered as adaptation scenarios: a 2011 dataset that
estimates unrestricted pollution burden from all sources as of that year, and a 2040 dataset that accounts
for the implementation of a suite of regulatory policies on stationary and mobile emissions sources.
Summaries of impacts by temperature bin degree in 2010 and 2090, the endpoints of socioeconomic
modeling, and emissions inventory are included in Figure A-l below.
FIGURE A-l. AIR QUALITY IMPACTS BY TEMPERATURE BIN DEGREE
A. 2010 SOCIOECONOMICS
Air Quality; 2011 Emissions; Ozone
140,000
120,000
10
§ 100,000
is 80,000
0
•5 60,000
tfi
J 40,000
1
20,000
0
2 3 4
Degrees of Warming
—* CCSM4
m
CanESM2
—•— GFDL-CM3
~
GISS-E2-R
—*— HadGEM2-ES
*
MIROC5
Air Quality; 2011 Emissions; PM2.5
2 3 4
Degrees of Warming
—"—CCSM4
~ Car>ESM2
GFDL-CM3
—~— GISS-E2-R
MiRr>r«;
* naUvCiVlt'CD
80,000
Air Quality; 2040 Emissions; Ozone
12 3 4
Degrees of Warming
—*—CCSM4
—#—
CanESM2
¦ GFDL-CM3
—~—
GISS-E2-R
» HadGEM2-ES
—A—
MIROC5
Air Quality; 2040 Emissions; PM2.5
2 3 4
Degrees of Warming
" CCSM4
GFDL-CM3
—~— GI5S-E2-R
® nauuCfvli'Cj
¦ - IVllr\Uuj
Expert and Public Review Draft-Do Not Cite
Page A-3
-------
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
B. 2090 SOCIOECONOMICS
25,000
20,000
15,000
10,000
5,000
0
-5,000
Air Quality; 2040 Emissions; Ozone
-10,000' '
0 12 3 456
Degrees of Warming
60,000
50,000
40.000
30,000
20.000
10.000
1 2 3 4 5 6
Degrees of Warming
Air Quality; 2011 Emissions; Ozone
—*—CCSM4
—•— CanESM2
•—GFDL-CM3
~ GISS-E2-R
> HadGEM2-ES
—"— CCSM4 —~— CanESM2
GFDL-CM3 —+— GISS-E2-R
—*—HadGEM2-E5 —A— MIROC5
Degrees of Warming Degrees of Warming
—"—CCSM4
GFDL-CM3
—*— HadGEM2-ES
Can£SM2
GISS-E2-R
MIROC5
—"—CCSM4
¦ GFDL-CM3
—*— HadGEM2-ES
Can£SM2
GISS-E2-R
MIROC5
Processing steps
Processing steps are illustrated in Figure A-2. Data inputs from Fann etal, (2021) are compiled using U.S.
EPA's Benefits Mapping and Analysis Program - Community Edition (BenMAP-CE) to generate results atthe
regional level. The original air quality data was provided by study authors with the dimension era—GCM—
36-km grid cell—pollutant (ozone/PM2.5)—emissions inventory (2011/2040). Data is available for four eras
(2030, 2050, 2075, 2095) and two climate models (CCSM4 and GFDL-CM3). Concentration-response
functions employed by Fann et al. (2021) are based on risk model information for those age 30-99 for PM2.5
and those age 0-99 for ozone. Within BenMAP-CE, impacts are aggregated to the regional level for each
era, GCM, pollutant, and emissions inventory scenario. The original results provided already account for
baseline incidence; therefore, no additional processing is needed to isolate climate impacts.
The third step divides total regional impacts by dynamic Integrated Climate and Land Use Scenarios, v2
(ICLUSv2) regional population to acquire per capita mortality estimates for each era. The calculations utilize
total regional population anticipating alternative population inputs would be unlikely to contain age-
stratified population projections. Next, era costs are assigned to the central year of the era (i.e., 2030, 2050,
2075, and 2095), and costs per era are transformed to annual costs by interpolating linearly between era
Page A-4
-------
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
impacts. Finally, yearly impacts for each pollutant impact type (ozone/PIV^s) are averaged across the GCM-
specific eleven-year windows around the first arrival times of integer degrees of warming relative to the
baseline.
FIGURE A-2. AIR QUALITY PROCESSING FRAMEWORK
C
to
CO
0J
U
O
Economic Damage: Multiply by region populations and VSL
(year/GCM/region/{O3/PM}/{2011 emissions/2040 emissions})
O
O
Final mortality estimates are produced by applying the per capita mortality rates to the input population
scenario and GDP input-adjusted VSLs. VSL is adjusted for changes in GDP per capita using an income
elasticity function26:
Limitations and Assumptions
• PM2.5-attributable premature mortality is quantified for those age 30 and older, and this analysis
assumes the impacts for those under 30 to be zero. Doing so underestimates the risk of premature
mortality experienced by those under 30. Additionally, doing so assumes that age demographics
remain proportional over the century.
• This analysis does not quantify morbidity effects associated with changes in PM2.5 and ozone, which
are likely to increase as temperature increases. Changes in air quality can provoke hospital
admissions for respiratory diseases and worsen other conditions.
26 This is a generic elasticity function that can be used in a time-series fashion, as used here, or for cross-sectional benefits transfers, as in
the example in Masterman and Viscusi (2018), "The Income Elasticity of Global Values of a Statistical Life: Stated Preference Evidence",
Journal of Benefit-Cost Analysis, 9(3):407-434.
Expert and Public Review Draft-Do Not Cite PageA-5
VSLt — VSL2 010 x (
2010
-------
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
For further discussion of the limitations and assumptions in the underlying sectoral modeling
approach, see Fann et al. (2021).
UNDERLYING DATA SOURCES AND LITERATURE
Mills, D., Schwartz, J., Lee, M,,Sarofim, M.,Jones, R., Lawson, M.,
Duckworth, M., & Deck, L (2014). Climate Change Impacts on
Extreme Temperature Mortality in Select Metropolitan Areas in
the United States. Climatic Change, 131,83-95.
Doi:10.1007/sl0584-014-115 4-8
Extreme Temperature
This sector addresses the impact of
extreme temperature on premature
mortality in 49 major U.S. cities.
Economic damages are based on extreme
heat and cold mortality rates, monetized by applying the VSL. The VSL trajectory through the simulation
period changes as a function of per capita income.
The underlying epidemiologic model includes runs with and without adaptation scenarios. The original
estimates are provided for 49 cities. Figure A-3 provides a summary of the results for heat and cold related
mortality, for both adaptation scenarios and six GCMs at the endpoints of the socioeconomic scenarios;
2010 and 2090.
FIGURE A-3. EXTREME TEMPERATURE IMPACTS BY TEMPERATURE BIN DEGREE
A. 2010 SOCIOEONOMICS
Extreme Temperature; No Adaptation; Hot
2 3 4
Degrees of Warming
—*—CCSM4
• CanESM2
—•— GFDL-CM3
~ GISS-E2-R
—*—HadGEM2-ES
A MIROC5
Extreme Temperature; Adaptation; Hot
2 3 4
Degrees of Warming
—h CCSM4
—•— CanESM2
—GFDL-CM3
~ GISS-E2-R
—*— HadGEM2-ES
—A— MIROC5
Extreme Temperature; No Adaptation; Cold
Extreme Temperature; Adaptation; Cold
2 3 4
Degrees of Warming
-CCSM4 —•—CanESM2
GFDL-CM3 —~— GISS-E2-R
- HadGEM2-ES A MIROC5
2 3 4
Degrees of Warming
-CCSM4 ~ CanESM2
GFDL-CM3 ~ GISS-E2-R
- HadGEM2-ES * MIROC5
Page A-6
-------
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
B. 2090 SOCIOECONOMICS
Extreme Temperature; No Adaptation; Hot
200,000 5
2 3 4
Degrees of Warming
—h—CCSM4
—•— Car>ESM2
GFDL-CM3
~ GI5S-E2-R
—*—HadGEM2-ES
A MIROC5
Extreme Temperature; Adaptation; Hot
2 3 4
Degrees of Warming
—*—CCSM4
—•—Car>ESM2
GFDL-CM3
—+— GISS-E2-R
~ HadGEM2-ES
—A— MIROC5
Extreme Temperature; No Adaptation; Cold
Extreme Temperature; Adaptation; Cold
2 3 4
Degrees of Warming
—*—CCSM4
~ CanESM2
GFDL-CM3
—~— GISS-E2-R
—*-= HadGEM2-ES
A MIROC5
io
-200- V
o
c?
o
-400
"c
O
-600
c
-800
Q
i
-1.000
-1^00
2 3 4
Degrees of Warming
—**—CCSM4
—•— CanESM2
GFDL-CM3
—~— GISS-E2-R
—HadGEM2-ES
A MIROC5
Processing steps
Processing steps are shown in Figure A-4. The original data was provided by the study authors with the
dimension degree - GCM - city - damage type (heat/cold mortality) - base population (2010/2090). The
first processing step was to sum the city damages to regional damages. Next, the incremental impacts of
climate change are isolated by subtracting theO-degree bin mortality results from each warming bin.
The model was run under two constant population assumptions: 2010 and 2090 estimates from ICLUSv2
(EPA 2017; Bierwagen et al., 2010). In addition to having different total populations, these two scenarios
vary in the distribution of population across modeled cities. In the third processing step, regional mortality
in each scenario is divided by total population in each scenario to obtain a mortality per capita estimate for
each population base. Both estimates, as well as an interpolation between the two, are available for impact
estimations.
Expert and Public Review Draft-Do Not Cite
Page A-7
-------
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
FIGURE A-4. EXTREME TEMPERATURE DATA PROCESSING FRAMEWORK
oo
c
CO
CO
CD
u
O
r
o
o
Economic Damage: Multiply by city populations and VSL
(year/GCM/region/{hot, cold}/{2010pop/2090pop})
Final economic damage estimates are produced by applying the per capita mortality rates tothe input
population scenario and GDP input-adjusted VSLs. Regional population inputs are translated to city
populations using factors derived from the ICLUSv2 population scenarios in 2010, 2050, and 2090, and
interpolated for years in between. VSL scales relative to changes in GDP per capita to general impact
estimates.
Limitations and Assumptions
• National per capita averages are based on the total population of modeled cities with heat and cold
impacts. There are certain cities in the Southeast (Atlanta, Broward-Ft. Lauderdale, Miami,
Orlando), Southern Plains (Austin, Dallas), and Southwest (Albuquerque, Los Angeles, Phoenix, San
Diego) regions that are modeled for adaptation to heat but are not modeled for adaptation to
extreme cold. It is assumed that these cities have minimal extreme cold damages, and therefore
their populations are included in the denominator as part of the total population over which cold
damages are averaged.
• This analysis only covers considers health impacts to individuals living in 49 cities within the
contiguous U.S., and therefore omits a large majority of the population vulnerable to extreme
temperatures.
• Cities that only experienced extreme cold in the historic period, notably those in the Northwest
region, do not show an increase in extreme-temperature related mortality in this analysis. This
result is an artifact of the methodology, which relies on observed temperature thresholds based on
a historic period. With increased temperatures, it is likely that many of these Northwestern cities
could experience heat-related mortality as well, which might be reflected if a different impact
estimation methodology had been applied.
• For further discussion of the limitations and assumptions in the underlying sectoral model, please
see Mills et al. (2014) and EPA (2017).
Page A-8
-------
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
UNDERLYING DATA SOURCES AND LITERATURE
Neidell, M., Graff-Zivin, J., Sheahan, M., Willwerth, J., Fant, C.,
Sarofim, M,, & Martinich, J. (In review). Temperature arid work:
Time allocated to work under varyingclimate and labor market
conditions.
Labor
The labor sector addresses economic
damages of changes in labor hours as a result
of climate change. The analysis estimates
changes in labor allocation, with both positive
and negative responses of changes in hours
worked in weather-exposed industries (e.g., agriculture, construction, manufacturing). The study finds the
relationship between temperature and hours worked is not significant during recession periods, and
therefore projected losses are adjusted to account for the probability of recession. Damages are based on a
physical measure of average hours worked by workers in high risk industries, which is monetized in Neidell
et al. (in review) by average wages across at-risk industries27. A summary of impacts by temperature bin
degree at 2010 and 2090, the endpoints of socioeconomic modeling, is included in Figure A-5.
FIGURE A-5. LABOR IMPACTS BY TEMPERATURE BIN DEGREE
A. 2010 SOCIOECONOM CIS
2 3 4
Degrees of Warming
—«— CCSM4
—•— GFDL-CM3
—*— HadGEM2-ES
CanESM2
GISS-E2-R
MIROC5
B. 2090 SOCIOECONOMICS
Degrees of Warming
—*—CCSM4
—•—CanESM2
¦ GFDL-CM3
—+— GISS-E2-R
~ HadGEM2-ES
—A— MIROC5
27 Hourly wages are based on average wages across at-risk industries: agriculture, forestry, fishing, hunting,
mining, construction, and manufacturing.
Expert and Public Review Draft-Do Not Cite
Page A-9
-------
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
Processing steps
Processing steps are shown in Figure A-6. Forgone wages due to climate-induced temperature increase are
calculated by combining an average hourly wage rate, high-risk worker population, and per high-risk worker
hours lost. The original data includes forgone hours for each year, GCM, and region combination. In the first
step, state-level data is summed to the NCA region level The second step is bypassed; these results already
account for baseline hours lost, so no additional processing is needed to isolate climate impacts. The third
step divides these estimates by high-risk worker population to acquire per high-risk worker estimates. The
population of high-risk workers varies by region but is assumed to remain constant over the course of the
century. Per high-risk worker hours lost are stored as the underlying impact, and final economic damage
estimates are produced by multiplying these rates by a regional population and an average wage.
FIGURE A-6. LABOR DATA PROCESSING FRAMEWORK
Input Data: Hours lost, states summed to regions
(year/GCM/region)
Isolate Climate Impact: Baseline already removed in Hours Lost input
Calculate per capita impacts: Hours lost divided by high risk worker population
(year/GCM/region)
Bin Results by Temperature: Average impacts over temperature bin eras by GCM
(degree/GCM/region)
Economic Damage: Multiply by wage rates and working populations
(year/GCM/region)
o
o
High Risk
Worker
L.
Population
Limitations and Assumptions
• High risk worker population is assumed to remain constant over the course of the century.
• This analysis does not evaluate the potential for new adaptations (behavioral or technological) by
workers or employers to mitigate the effects of extreme temperatures on labor allocation. Adaptations
present in the baseline period upon which the econometric analysis is based are assumed to be part of
the modeled response to future temperature changes, however, new adaptation behaviors or
technology are not evaluated.
• For further discussion of the limitations and assumptions in the underlying sectoral model see Neidell
et al. (in review).
Page A-10
-------
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
Southwest Dust
This sectoral study estimates health
burden and the economic value of that
burden resulting from changes in fine
and coarse airborne dust exposure due
to climate change in the Southwest.
Damages are based on the change in incidence of a range of morbidity and mortality outcomes, which are
monetized using direct hospitalization costs, indirect loss of income from hospitalization, costs of
emergency department visits, and (for premature mortality) the VSL.
Estimates of health costs are available by impact type: Emergency Department visits due to Asthma,
Cardiovascular, Respiratory, Mortality, and Acute Myocardial infarction. Asummary of impacts by
temperature bin degree in 2010 and 2090, the endpoints of socioeconomic modeling, is included in Figure
A-7.
UNDERLYING DATA SOURCES AND LITERATURE
Achakulwisut, P., Anenberg, S.C., Neumann, J. E., Penn,S. L, Weiss, N.,
Crimmins, A., Fann, N., Martinich, J., Roman, H.A., & Mickley, L. J,
(2019). Effectsof increasing aridity on ambientdustand publichealth in
the U.S. southwest under climate change. GeoHealth, 3(5), 127-144.
Doi:10.1029/2 019GH000187
FIGURE A-7. SOUTHWEST DUST IMPACTS BY TEMPERATURE BIN DEGREE
A. 2010 SOCIOECONOMICS
Southwest Dust; Acute Myocardial Infarction
Southwest Dust; All Cardiovascular
Degrees of Warming
1 2 3 4 5 6
Degrees of Warming
—*—CCSM4
• CanESM2
—¦— GFDL-CM3
~ GISS-E2-R
—»—HadGEM2-E$
—A— MIROCS
—"—CCSM4
—•—GFDL-CM3
—*— HadGEM2-ES
CanESM2
GISS-E2-R
MIROCS
Degrees of Warming
Southwest Dust; All Respiratory
—*—CCSM4
—•—GFDL-CM3
—*— HadGEM2-ES
CanESM2
GIS5-E2-R
MIROCS
—m—CCSM4
• CanESM2
GFDL-CM3
—GISS-E2-R
—«— HadGEM2-ES
—A— MIROCS
Expert and Public Review Draft-Do Not Cite
Page A-ll
-------
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
1481
1482
1483
Southwest Dust; Asthma ER
2 3 4
Degrees of Warming
—*—CCSM4
• CanESM2
—gfdl-cm3
~ GISS-E2.R
—«— HadGEM2-ES
A MIROC5
B. 2090 SOCIOECONOMICS
Southwest Dust; Acute Myocardial Infarction
2 3 4
Degrees of Warming
—*—CCSM4
—CanESM2
•- GFDL-CW3
~ GISS-E2-R
—*— HadGEM2-ES
A MIROC5
20,000
Southwest Dust; All Mortality
Degrees of Warming
-CCSM4 —*—CanESM2
GFDL-CM3 —~— GISS-E2-R
- HadGEM2-ES —A— MIROC5
Southwest Dust; All Cardiovascular
2 3 4
Degrees of Warming
—»«—CCSM4
~ CanESM2
—GFDL-CM3
~ GISS-E2-R
—4— HadGEM2-ES
A MIROC5
Southwest Dust; All Respiratory
2 3 4
Degrees of Warming
GFDL-CM3
—~— GIS5-E2-R
A MIROC5
1484
Page A-12
-------
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
Degrees of Warming
—*—CCSM4
• CanESM2
—•—GFDL-CM3
—~— GIS5-E2-R
—*—HadGEM2-ES
A MIROC5
Processing steps
Processing steps are illustrated in Figure A-8. The health burden and costs incurred asa result of increased
dust exposure due to climate change are calculated by combining average cost-of-illness and mortality
costs, population, and average number of cases.
The original results already account for baseline incidence; therefore no additional processing is needed to
isolate climate impacts. Original results are presented as increases in health impacts from baseline levels
for affected populations — for example, cardiovascular disease impacts are considered only for people over
65. To calculate costs across all considered impacts, damages per capita are calculated for the total
population by impact type; implicitly, damages for age groups not modeled are assumed to be zero. Per
capita estimates are calculated using total regional populations in 2010 and 2090. These damages are
produced across integer degree, health impact, and GCM.
Both estimates (2010/2090), as well as an interpolation between the two, are available for impact
estimation. Per capita estimates of health impacts are multiplied by regional population and average
medical cost. Medical costs are variable across health impacts. For all mortality, the VSL is utilized, which
scales relative to changes in GDP per capita. Final damages are calculated based off the temperature
trajectory assigned in the tool.
Expert and Public Review Draft-Do Not Cite
Page A-13
-------
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
FIGURE A-8. SOUTHWEST DATA PROCESSING FRAMEWORK
GDP
i
Medical
Costs
(constant)
Input Data: Percentage increase in incidence of health impacts
(year/GCM/impact)
Isolate Climate Impact: Baseline implicitly removed in Health Impacts input
Bin Results by Temperature: Average impacts over temperature bin eras by GCM
(degree/GCM/impact)
Scalable Tool Input: Multiply by impact-specific population, and divide by total
population
(degree/GCM/impact/{2010pop,2090pop})
Economic Damage: Multiply by southwest population and weighted average of
valuations by impact type
(year/GCM/{2010pop,2090pop})
O
O
Limitations and Assumptions
• While dust exposures are known to be large in the southwestern U.S., this analysis does not
consider health effects from coarse and fine dust in other regions of the U.S.
• This sector relies on population for a section of the Southwest region (Arizona, Colorado, New
Mexico, Utah) to calculate damages across impact types. The scaling of damages by this population
allows for custom inputs of socio-economic estimates but may introduce error if the age
demographics of the population are not roughly constant over the simulation period.
• For further discussion of the limitations and assumptions in the underlying sectoral model, see
Achakulwiset etal. (2019).
Valley Fever
This sectoral study estimates the health
burden and economic value associated with
climate change-related Valley fever
incidence. Valley fever is a prevalent disease
in the hot and dry Southwest region of the
U.S. but is expected to expand in geographic
scope with warming. Therefore, this analysis quantifies Valley fever impacts across the contiguous U.S.,
with most of the burden focused in the Southwest.
UNDERLYING DATA SOURCES AND LITERATURE
Gorris, M. E., Neumann, J. E., Kinney, P. L, Sheahan, M.,&
Sarofim, M. C. (2020). EconomicValuation of Coccidioidomycosis
(Valley Fever) Projections in the United States in Response to
Climate Change. Weather, Climate, and Society, 13(1), 107-123.
Doi:10.1175/WCAS-D-20-0036.1
Page A-14
-------
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
Impacts are based on the change in number of Valley fever cases and the probability of a range of
morbidity outcomes, which are monetized using direct hospitalization costs, costs of emergency
department visits, costs of physician visits, and indirect cost of lost productivity from hospitalization.
Mortality is valued using the VSL. Summary of impacts by temperature bin degree in 2010 and 2090, the
endpoints of socioeconomic modeling, are included in Figure A-9.
FIGURE A-9. VALLEY FEVER IMPACTS BY TEMPERATURE BIN DEGREE
A. 2010 SOCIOECONOMICS
Degrees of Warming Degrees of Warming
—"—CCSM4
—•— GFDI-CM3
—«— HadGEM2-ES
CanESM2
GISS-E2-R
MIROC5
—*—CCSM4
GFDL-CM3
—*—HadGEM2-ES
CanESM2
GISS-E2-R
MIROC5
Degrees of Warming
—*~—CCSM4
—¦— GFDL-CM3
—*— HadGEM2-ES
CanESM2
GISS-E2-R
MIROC5
B. 2090 SOCIOECONOMICS
Valley Fever; Lost Wages
2 3 4
Degrees of Warming
Valley Fever; Morbidity
2 3 4
Degrees of Warming
—*—CCSM4
—•— CanESM2
GFDL-CM3
—GIS5-E2-R
—*— HadGEM2-ES
—A— MIROC5
—*—CCSM4
~ CanESM2
GFDL-CM3
—GIS5-E2-R
* HadGEM2-ES
—A— MIROC5
Expert and Public Review Draft-Do Not Cite
Page A-15
-------
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
Degrees of Warming
—"—CCSM4
• CanESM2
GFDL-CM3
~ GIS5-E2-R
—*—HadGEMZ-ES
—A— MIROC5
Processing steps
Processing steps are illustrated in Figure A-10. Projected Valley fever incidence at the county-level was
provided by study authors for 10-year eras centered on 2030, 2050, 2070, and 2090, for six GCMs. In step 2,
baseline incidence by county is subtracted from projected incidence for Southwest counties that met an
endemicity threshold for Valley fever in the baseline period (112 Southwest counties out of 216). The
modeled CIRA baseline based on LOCA weather data is used in place of the Precipitation-Elevation
Regressions on Independent Slopes Model (PRISM) baseline from the underlying study. The PRISM baseline
provides total regional incidence and does not line up temporally with the modeled CIRA baseline. The CIRA
baseline for the period of 1995 (1986-2005) provides incidence by county and allows for comparison of
impacts by degree across sectors. A baseline of zero is assumed for all other counties with projected
incidence. This includes counties in the Southwest that did not meet the endemicity threshold in the
baseline period and counties outside of the Southwest region which, similarly, did not meet the endemicity
threshold in the baseline period. The resulting county-level incidence identifies cases of Valley fever
attributable to climate change.
Next, county-level impacts are summed tothe regional level, resulting in a total count of Valley fever cases
per region. In step 3, total impacts are divided by dynamic ICLUSv2 regional population to calculate a case
per capita value for each era,GCM, and region. Finally, an annual time series is constructed by linearly
interpolating between era values, and yearly impacts are temperature binned by GCM-specific eleven-year
windows.
A range of morbidity and mortality outcomes of varying severity associated with Valley fever cases are
valued to calculate a weighted average cost based on likelihood of outcome. Based on prior literature,
morbidity outcomes are expected to occur in 96 percent of Valley fever cases. Valued morbidity impacts
include direct hospitalization, emergency room visit with discharge, emergency room visit with
hospitalization, and physician visit. Lost productivity costs associated hospitalizations are monetized using
likelihood of outcome and wage rate. Finally, mortality is expected to occur in 4 percent of Valley fever
cases and is valued using VSL. To generate impact estimates, cases per capita are multiplied by regional
population and a weighted average of costs that accounts for likelihood and value of each outcome.
Page A-16
-------
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
FIGURE A-10. VALLEY FEVER PROCESSING FRAMEWORK
Input Data: Total number of Valley fever cases by county
(era/GCM/county)
Isolate Climate Impact: Subtract baseline incidence from projected incidence. Assume
baseline of 0 cases for counties that did not meet endemicity threshold during
baseline period
Calculate per capita impacts: Cases by county summed to regions, total cases divided
by total regional population
(era/GCM/region)
Bin Results by Temperature: Interpolate between eras to annual time series, average
impacts over temperature bin eras by GCM
(degree/GCM/region)
Economic Damage: Multiply by regional population and weighted average of
valuations by impact type
(year/GCM/region)
Medical
Costs
(constant)
o
o
Limitations and Assumptions
• This analysis assumes a baseline of zero cases for counties that did not meet the endemicity
threshold in the Southwest region during the baseline period as well as all counties with projected
Valley fever cases outside of the Southwest region.
• For further discussion of the limitations and assumptions in the underlying sectoral model, see
Gorris et al. (2020).
Wildfire
This sectoral study estimates health
impacts from wildfire emissions and
response costs from wildfire suppression.
Neumann etal. (2021) models changes in
wildfire activity for the western region of
CONUS. As such, response costs are
limited to this area, but this study models health impacts of the particulate matter from western wildfires
across the CONUS (as these emissions typically travel eastward across the continent.
Health impacts are based on the change in incidence of a range of morbidity and mortality outcomes, which
are monetized using direct hospitalization costs, costs of emergency department visits, lost productivity,
and (for mortality) the VSL. Response costs are estimated based on average wildfire response costs per acre
UNDERLYING DATA SOURCES AND LITERATURE
Neumann,J. E., Amend, M., Anenberg,S., Kinney, P. L, Sarofim, M.,
Martinich,J., Lukens, J., Xu, J., & Roman, H. (2021). Estimating
PM2.5-related premature mortality and morbidityassociated with
futurewildfireemissions in thewestern US. Environmental Research
Letters, 16(3). Doi:10.1088/1748-9326/abe82b
Expert and Public Review Draft-Do Not Cite
Page A-17
-------
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
burned, by NCA region. Summaries of impacts by temperature bin degree in 2010 and 2090, the endpoints
of socioeconomic modeling, are included in Figure A-ll below.
FIGURE
A.
A-ll. WILDFIRE IMPACTS BY TEMPERATURE BIN DEGREE
2010 SOCIOECONOMICS
Wildfire; Morbidity
2 3 4
Degrees of Warming
—*—CCSM4
—•— CanESM2
GFDL-CM3
—~— GISS-E2-R
—*— HadGEM2-ES
—A— MIROC5
Wildfire; Mortality
2 3 4
Degrees of Warming
—h—CCSM4
—•— CanESM2
m - GFDL-CM3
—~— GISS-E2-R
—•—HadGEM2-ES
—A— MIROC5
B. 2090 SOCIOECONOMICS
Wildfire; Morbidity
2 3 4
Degrees of Warming
—*—CCSM4
• CanESM2
—•—GFDL-CM3
—~— GIS5-E2-R
=*— HadGEM2-ES
A MIROC5
Wildfire; Mortality
2 3 4
Degrees of Warming
—*—CCSM4
—*— CanESM2
GFDL-CM3
—~— GIS5-E2-R
—«— HadGEM2-ES
A MIROC5
Processing steps
Processing steps are illustrated in Figure A-12. Data for each impact type (mortality, morbidity, and
response costs) are each processed separately, and summed to estimate total cost across impact types.
Regional mortality incidence attributable to climate change-related changes in PM2.5 concentrations
resulting from wildfires is provided by study authors for two 10-year eras centered on 2050 and 2090 and
five climate models. This analysis considers mortality estimated using a concentration-response function
based on risk model information specific to those age 30 and older. A "no wildfires" mortality baseline that
isolates the impact of projected climate (using the Localized Constructed Analogs, or LOCA data) on
wildfires is subtracted from projected incidence to isolate the health burden associated with climate-
induced changes in wildfire activity. The climate change-related mortality incidence is divided by dynamic
Page A-18
-------
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
ICLUSv2 regional population for each era to calculate mortality per capita for each era/GCM/region
scenario. Finally, an annual time series is constructed by linearly interpolating between era values, and
yearly impacts are temperature binned by GCM-specific eleven-year windows.
Regional morbidity incidence and valuation was provided by study authors for the same 10-year eras
centered on 2050 and 2090 and five climate models. This analysis sums valuation across a set of health
endpoints to determine one value associated with all morbidity impacts, representing cost of illness and
lost productivity for each era/GCM/region scenario.28 Baseline valuation is subtracted from 2050 and 2090
projected valuation to isolate the impact of climate change on wildfire-related morbidity. Morbidity
valuation is then divided by population, interpolated, and temperature binned as described above for
mortality.
Acres burned with baseline implicitly accounted for was available from authors with the dimension year—
GCM—region, excluding the Midwest, Northeast, and Southeast regions. Acres burned per region is
averaged across the GCM-specific eleven-year windows around the first arrival times of integer degrees of
warming relative to the baseline. Regional response costs per acre are inputted into the tool as a scalar
multiplied by acres burned to calculate total impacts. Response costs per acre remain constant across the
century.
Mortality incidence per capita is multiplied by total regional population and VSL to calculate total mortality
valuation. Morbidity valuation per capita is multiplied by total regional population to calculate total
morbidity valuation. Acres burned is multiplied by response costs per acre to calculate total response costs.
The total cost for each impact type is summed to calculate total regional impacts.
28 Full list of health endpoints includes: acute bronchitis, nonfatal acute myocardial infarction, asthma exacerbation (cough, wheeze,
shortness of breath), asthma emergency room visits, cardiovascular hospital admissions, asthma hospital admissions, chronic lung
disease hospital admissions (less asthma), respiratory hospital admissions, lower respiratory symptoms, upper respiratory symptoms,
work loss days, and minor restricted activity days.
Expert and Public Review Draft-Do Not Cite PageA-19
-------
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
FIGURE A-12. WILDFIRE PROCESSING FRAMEWORK
Input Data: Mortality incidence, morbidity valuation, and acres burned by region
(year/GCM/region/impact)
Isolate Climate Impact: Subtract baseline estimates from impact-specific projected
estimates (mortality, morbidity). Baseline implicitly removed in acres burned input
(year/GCM/region/impact)
Calculate per capita impacts: Mortality incidence and morbidity valuation divided by
population
(year/GCM/region/impact)
Bin Results by Temperature: Average impacts over temperature bin eras by GCM
(degree/GCM/region/impact)
Economic Damage: Multiply by regional population and weighted average of
valuations by impact type
(year/GCM/region)
O
O
Response
costs per acre
(Constant)
Limitations arid Assumptions
• Mortality incidence is quantified for those age 30 and older, and this analysis assumes the impacts
for those under 30 to be zero. Doing so underestimates the risk of premature mortality experienced
by those under 30. Additionally, doing so assumes that age demographics remain proportional over
the century.
• Similarly, the morbidity health endpoints included in this analysis are associated with various age
distributions. Total valuation is divided by total regional population, assuming that the health
burden outside of included age ranges is zero.
• For further discussion of the limitations and assumptions in the underlying sectoral model see
Neumann etal. (2021).
Page A-20
-------
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
A.3 Infrastructure Sectors Data Processing
UNDERLYING DATA SOURCES AND LITERATURE
Neumann,J. E., Chinowsky, P., Helman,J., Black, M., Fant,C., Strzepek,
K., & Martinich, J. (In review). Climate effects on US infrastructure:
the economics of adaptation for rail, roads, and coastal development.
Lorie, M., Neumann,J. E., Sarofim, M.C., Jones, R., Horton, R. M.,
Kopp, R. E., Fant, C.,Wobus,C., Martinich,J.,O'Grady, M., Gentile, L
E. (2020). Modeling coastal flood risk and adaptation response under
futureclimate conditions. Climate Risk Management, 29.
Doi:10.1016/j.crm.2020.100233
Coastal Properties
This sector study estimates future
property value damages as a result of
sea level rise combined with storm
surge attributed to climate change.
Damages are estimated for all real
properties (land and structure) in all
coastal counties that contain land with
a hydraulic connection tothe ocean
and containing property that is within
20 m elevation above sea level for the year 2000. Property values for potentially vulnerable structures and
land are "market adjusted" assessed values, from 2006, adjusted to 2017 value using data from Zillow.
These values are acquired for counties along the CONUS coast - see Lorie et al. (2020) for details.
The underlying damage simulation model includes cost estimates for no adaptation and two adaptation
scenarios (reactive and proactive), as defined in the underlying study. Under the no adaptation scenario,
properties are abandoned once inundated. Reactive adaptation loosely reflects structural adaptation
options that can be adopted without collective action (e.g., elevation of structures and land near
structures), while proactive adaptation includes consideration of options that likely require collective action
(such as beach nourishment and construction of seawalls). The model conducts a series of benefit-cost
calculations at the level of a 150m x 150m grid cell to assess where and when adaptation could be cost-
effective in mitigating property damage due to sea level rise and storm surge. Summaries of impacts by
integer degrees of warming in 2010 and 2090, the endpoints of socioeconomic modeling, and by
adaptation scenario is included in Figure A-13.
FIGURE A-13. COASTAL PROPERTIES IMPACTS BY INTEGER DEGREE OF WARMING
A. 2010 SOCIOECONOMICS
Coastal Properties; No Adaptation
Coastal Properties; Proactive Adaptation
2 3 4
Degrees of Warming
m 100 cm ~ 150 cm
—¦— 200 cm —~— 250 cm
—•—30cm —A— 50 cm
2 3 4
Degrees of Warming
—**— 100 cm —•— 150cm
—•— 200 cm # 250 cm
—*—30 cm —A—50 cm
Expert and Public Review Draft-Do Not Cite
Page A-21
-------
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
Degrees of Warming
—«— tOO cm
• 150 cm
—¦—200 cm
—#— 250 cm
B. 2090 SOCIOECONOMICS
—h—100 cm
—•— 150 cm
—•—200 cm
—~—250 cm
—•—30 cm
—A—50 cm
Processing steps
Processing steps are shown in Figure A-14. In step one, property damages are defined for each sea level
rise scenario, year, region, adaptation scenario, and for two static socioeconomic scenarios; 2010 and 2090.
Residential and commercial properties, as well as energy infrastructure, are considered tocalculate
potential damages in the model. Yearly damage estimates are then averaged for each global mean sea level
(GMSL) increment. For each SLR scenario and GMSL bin, damages are projected using a trajectory of yearly
sea level rise between 2000 and 2100. The binning approach is consistent with other sectors, except bins
are defined by thresholds of sea level rise rather than increments of temperature. Sea level rise bins are
defined as the ll~year window centered around the first arrival years of 25cm increments of GMSL from
Page A-22
-------
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
the baseline. For this sector, the baseline is anchored at 2000, as the National Coastal Property Model
(NCPM) starts with zero damages in this year.
These binned damage estimates are relied upon for impact estimate calculations. Because of the decision-
tree structure of the NCPM, population and GDPcannot be disentangled as drivers of impacts. A linear
interpolation between impact estimates associated with runs of the NCPM where socioeconomics are held
constant in 2010 and 2090 is used to model changes in socioeconomic drivers for this sector.
FIGURE A-14. COASTAL PROPERTIES DATA PROCESSING FRAMEWORK
Limitations arid Assumptions
• Damages are limited to land and structures within the study domain (i.e., flooding impacts to
structures inland of 20m elevation are not quantified), and exclude the value of public
infrastructure, which was not considered in the underlying sectoral study.
• Adaptation response decisions in the coastal zone are not typically made with strict cost-benefit
decision rules, particularly at the local level. Other factors may include local zoning bylaws, future
land use plans, the presence of development-supporting infrastructure, or proximity to sites with
high cultural value. However, the analytical framework of this coastal property model provides a
simple, benefit-cost decision framework that can be consistently applied for regional and national-
scale analysis.
• For further discussion of the limitations and assumptions in the underlying sectoral model see
Neumann et al. (in review), Lorie et al. (2020), and USEPA (2017).
Expert and Public Review Draft-Do Not Cite
Page A-23
-------
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
High Tide Flooding and Traffic
This sector study estimates the cost of delays to
passenger and freight traffic on coastal roads that
experience flooding due to combinations of high
tides and sea level rise, and costs of adaptation in
the form of infrastructure improvements.
Delay damages are in terms of passenger and freight vehicle-hours. These are monetized based on the
value of travel time savings (VTTS) for passenger traffic, and the National Cooperative Highway Research
Program's (NCHRP) inputs for cost of delay for freight traffic. Infrastructure improvements include building
sea walls or elevating the elevation ofthe roadway surface. Infrastructure improvement costs include
estimates of material, labor, and construction delays.
This sector considers three adaptation scenarios: no adaptation, reasonably anticipated adaptation, and
direct adaptation. These adaptation scenarios differ from scenarios modeled for other infrastructure
sectors. The no adaptation, reactive adaptation, and proactive adaptation scenarios of other infrastructure
sectors are based on infrastructure development for an unchanging, current, or future climate in a given
model time step. For this sector, the no adaptation scenario estimates costs of delays associated with
flooding of roadways with the assumption that drivers do not re-route and instead wait until the roadway is
clear to travel. The reasonably anticipated adaptation scenario assumes drivers re-route to avoid flooded
roadways, with only slight delay due to increased travel time. This scenario also includes ancillary
protection; in cases where flooded are roadways are near properties that would be protected by sea walls
or beach nourishment, this scenario assumes those roadways would also be protected and thus no longer
flood. In the direct adaptation scenario, where delay costs are high enough, roadways are either protected
from flooding through the construction of a sea wall or elevation ofthe road profile. Figure A-15 provides a
summary of results by integer degree of warming and adaptation scenario in 2010 and 2090, the endpoints
of socioeconomic modeling, and representing the static runs for each SLR scenario.
UNDERLYING DATA SOURCES AND LITERATURE
Fant, C., Jacobs, J. M., Chinowsky, P.,Sweet, W., Weiss,
N., Martinich,J. & Neumann, J. E. (In review). Mere
nuisance or growing threat? The physical and economic
impact of high tide flooding on US road networks.
FIGURE A-15. HIGH TIDE FLOODING AND TRAFFIC IMPACTS BY INTEGER DEGREE OF WARMING
A. 2010 SOCIOECONOMICS
High Tide Flooding and Traffic; Direct Adaptation
High Tide Flooding and Traffic; No Adaptation
<*_ 10.000
23 4
Degrees of Warming
rtuu.a;::'
700,000
=> 600,000
500.000
400,(WO
y, 300,000
= 200,000
5
100.000
2 3 4
Degrees of Warming
—*—100 cm
200 cm
—•—30 cm
150 cm
250 cm
50 cm
—*—100 cm
—•—200 cm
—~—250 cm
Page A-24
-------
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
Degrees of Warming
w 100 cm
• 200 cm
—•—30 cm
150cm
250 cm
50 cm
B. 2090 SOCIOECONOMICS
High Tide Flooding and Traffic; Direct Adaptation
2 3 4
Degrees of Warming
High Tide Flooding and Traffic; No Adaptation
1 2 3 4 5
Degrees of Warming
" 100 cm
200 cm
—•—30 cm
150 cm
250 cm
50 cm
—h—100 cm —•—160 cm
b 200 cm ~ 250 cm
—•— 30 cm —A— 50 cm
Tide Flooding and Traffic; Reasonably Anticipated Adaptation
Degrees of Warming
—"—100 cm
¦ 200 cm
—»—30 cm
150 cm
250 cm
50 cm
Processing steps
Processing steps are seen in Figure A-16. In step one, damages at the county level are aggregated to the
regional level. These damages are available for all SLR scenario, year, and adaptation scenario combinations
for both the 2010 and 2090 static socioeconomic runs.
Similar to the Coastal Properties sector, this sector "zeroes out" in 2000, and thus has no baseline for which
to adjust. Damages associated with climate change from the baseline period are then binned by 25cm
increments of GMSL value.
Expert and Public Review Draft-Do Not Cite
Page A-25
-------
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
FIGURE A-16. HIGH TIDE FLOODING AND TRAFFIC DATA PROCESSING FRAMEWORK
Limitations and Assumptions
• The underlying sectoral analysis is limited to road segments within the flood extent for the current
minor flood level. This extent is expected to migrate further inland as sea levels rise. This analysis
also omits consideration of impacts to underground roads.
• Flooding as a result of rainfall or riverine flooding is not modeled and may exacerbate flood events
or durations in the coastal zone if they occur simultaneously.
• Many direct adaptation options (e.g., hydrologic infrastructure) are not considered.
• For further discussion of the limitations and assumptions in the underlying sectoral model see Fant
et a I. (In review).
Rail
This analysis estimates repair, equipment,
and delay costs to rail infrastructure due to
rail track buckling or the risk of buckling
associated with elevated temperatures.
Damages are based on costs of repair,
including equipment and labor, and delay
costs. These costs are then scaled using
total track miles in each region of CONUS.
The analysis is completed for each of three adaptation scenarios: no adaptation, proactive adaptation, and
reactive adaptation. The no adaptation scenario incorporates no speed restrictions but results in a higher
UNDERLYING DATA SOURCES AND LITERATURE
Neumann, J. E., Chinowsky, P., Helman, J., Black, M., Fant, C.,
Strzepek, K., & Martinich,J. (In review). Climate effects on US
infrastructure:theeconomicsof adaptation for rail, roads,and
coastal development.
Chinowsky, P., Helman, J., Gulati, S., Neumann,J., & Martinich,J.
(2019). Impacts of climate change on operation of the US rail
network. Transport Policy, 75,183-191.
Doi:10.1016/j.tranpol.2017.05.007
Page A-26
-------
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
risk of track buckling associated with continued use of trains during high temperature events. Track
buckling events require repair that create delays. The reactive scenario considers reduced train speeds at
higher temperatures to reduce likelihood of track buckling. The proactive scenario includes installation of
temperature sensors to monitor probabilities of track buckling and modify train speeds as necessary (and
therefore prevent delays associated with their unexpected need for repair). A summary of results by
temperature bin degree and adaptation scenario in 2010 and 2090, the endpoints of socioeconomic
modeling, is included in Figure A-17.
FIGURE A-17. RAIL IMPACTS BY TEMPERATURE BIN DEGREE
A. 2010 SOCIOECONOMICS
JA
O
60.000
50.000
0
HadGEM2-ES —A— MIROC5
B. 2090 SOCIOECONOMICS
250,000
£ 200.000
8
ClO
1 150.000
Z
a
"5 100.000
CO
c
g
2 50.000
10.000
V- 8.000
o
(VI
(A
| 6.000
"5
0
"5 4.000
c
Q
1 2.000
Rail; Proactive Adaptation
2 3 4 5 6
Degrees of Warming
1 2 3 4 5 6
Degrees of Warming
Rail; No Adaptation
—**—CCSM4
—•— CanESM2
GFDL-CM3
~ GISS-E2-R
—•—HadGEM2-ES
—A— MIROC5
" CCSM4
—«— GFDL-CM3
—•— HadGEM2-ES
CanESht2
GISS-E2-R
MIROC5
Expert and Public Review Draft-Do Not Cite
Page A-27
-------
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
200,000
Rail; Reactive Adaptation
g 100,000
"5
o 150,000
= 50,000
2
2
! 3
Degrees of Warming
4
5
6
—h— CCSM4 —•— CanESM2
¦ GFDL-CM3 —~— GISS-E2-R
—•— HadGEM2-ES —A— MIROC5
Processing steps
Processing steps are shown in Figure A-18. In step one, the impact model for this sector is run at the
Climatic Research Unit (CRU) grid cell level, and damages are aggregated to the NCA region level. This
impact model assumes thatthe spatial extent and distribution of rail infrastructure remains constant across
the 21st century. In step two, baseline costs associated with delays and costs under a control climate
scenario are subtracted for all adaptation scenarios. Thus, damages presented are due to climate change.
Annual damages are then temperature binned. Total damages are then divided by total miles of rail within
a region to produce damages per mile.
Damages per mile are scaled with the number of rail miles in a region as well as a socioeconomic growth
scalar, with 2010 as the base year. The model assumes rail traffic increases linearly with GDP growth for
freight traffic, and linearly with population growth for passenger traffic. Freight traffic represents 96
percent of rail traffic, and passenger traffic the remaining 4 percent. This analysis assumes that rail traffic
can be modified with custom GDP and/or population projections.
Page A-28
-------
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
FIGURE A-18. RAIL DATA PROCESSING FRAMEWORK
Limitations and Assumptions
• The model assumes the number of rail miles is fixed and does not grow over time, though rail traffic
over the existing rail network grows with a weighted average of population growth (for the
passenger rail component) and economic growth (for the much largerfreight rail component).
• Equipment, labor, and repair supply costs are assumed to remain constant.
• For further discussion of the limitations and assumptions in the underlying sectoral model see
Neumann et al. (In review), Chinowksy et al. (2017), and EPA (2017).
Roads
This sector study estimates the cost of roa(
repair, user costs (vehicle damage), and
road delays due to changes in road surface
quality as a result of climate change
(specifically changes in temperature,
precipitation, and flooding).
Damages are based on the cost of repairs
and delays associated with either
deteriorated road surfaces or road
shutdowns to complete repairs. The per
mile impacts are then multiplied by total
regional road miles, and adjusted to reflect the likelihood of delay mitigation as proxied by an index of road
density in each >2 degree by >2 degree grid cell, to produce a total damage estimate in a region.
UNDERLYING DATA SOURCES AND LITERATURE
Neumann, J. E., Chinowsky, P., Helman, J., Black, M., Fant, C.,
Strzepek, K., & Martinich, J. (In review). Climate effects on US
infrastructure:the economics of adaptation for rail, roads, and
coastal development.
Neumann, J. E., Price, J., Chinowsky, P., Wright, L, Ludwig, L,
Streeter, R., Jones, R., Smith, J. B., Perkins, W., Jantarasami, L, &
Martinich, J. (2014). Climate change risks to US infrastructure:
impacts on roads, bridges, coastal development, and urban
drainage. Climatic Change, 131,97-109. Doi:10.1007/sl0584-013-
1037-4
Expert and Public Review Draft-Do Not Cite
Page A-29
-------
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
Similar to the rail and coastal properties studies, the analysis models three adaptation scenarios: no
adaptation, proactive adaptation, and reactive adaptation. In the no adaptation scenario, repairs to roads
are limited to historic repair budgets; damages in this scenario are based on the cost of repairs to road
surfaces, damage to vehicles associated with incompletely maintained roads, and delays associated with
repairs to road surfaces or speed limitations attributed to poorly maintained roads.29 Under the reactive
adaptation scenario, repair budgets are increased to repair all damages in a given year to re-establish the
pre-repair level of service. In the proactive scenario, roads are pre-emptively strengthened to prevent
damage with consideration of future climate changes in the design and materials used for repair. Under the
reactive and proactive adaptation scenarios, damages are based on the cost of repairs to road surfaces and
the delays associated with repairs or speed limitations due to poorly maintained roads. The model
considers three types of environmental stressors: temperature, precipitation, and flooding. Damages differ
by road surface; road surfaces are either unpaved, paved, or gravel. This impact model runs atthe quarter-
degree grid cell level, and each grid cell is assigned adaptation-scenario specific budget for repairs. Figure
A-19 provides a summary of results by temperature bin degree and adaptation scenario in 2010 and 2090
(the endpoints of socioeconomic modeling). Note that the proactive adaptation results generally reflect a
much lower damage estimate overall than no adaptation or reactive costs, but that in some scenarios the
timing of those costs may be accelerated (and actually be triggered by relatively modest levels of warming)
because of optimization of the capital cost of resilience investments and the high payoff to these
investments in terms of avoiding future repairs and delays.
FIGURE A-19. ROADS IMPACTS BY TEMPERATURE BIN DEGREE
A. 2010 SOCIOECONOMICS
Roads; No Adaptation
2 3 4
Degrees of Warming
»¦ CCSM4
• CanESM2
—•— GFDL-CM3
~ GISS-E2-R
—•—HadGEM2-ES
—A— MIROC5
Roads; Proactive Adaptation
/ /
2 3 4
Degrees of Warming
—**— CCSM4
—•— CanESM2
—•— GFDL-CM3
—~— GISS-E2-R
» HadGEM2-ES
—A— MIROC5
29 The budget constraint in the no adaptation scenario can be thought of as a resilience threshold. For small amounts of warming, roads
and their maintenance systems are adequate to meet increased stress. Once that resilience threshold is exceeded, costs increase quickly
as road damage occurs.
Page A-30
-------
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
Degrees of Warming
—«—CCSM4
—•— Can£SM2
—GFDL-CM3
~ G1SS-E2-R
—*— HadGEM2-ES
—A—MIROCS
B. 2090 SOCIOECONOMICS
Degrees of Warming
—H— CCSM4 —•— CanESM2
—•— GFDL-CM3 > GISS-E2-R
« HadGEM2-ES + MIROCS
Processing steps
Processing steps are seen in Figure A-20. In step one, damages at the quarter-degree grid cell level are
aggregated to the regional level. These damages are available for all GCM, year, and adaptation scenario
combinations.
Baseline damages are then subtracted from projected damages for each GCM to arrive at damages
associated with climate change from the baseline period for each adaptation scenario. Damages associated
with climate change from the baseline period are then temperature binned — temperature binned
damages are average annual damages from eleven-year windows around first arrival times of integer
degrees of warming from the baseline period for each GCM. Total damages are then divided by total miles
of road within a region to produce damages in terms of dollars per mile.
Expert and Public Review Draft-Do Not Cite
Page A-31
-------
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
Damages per mile are scaled using total regional road miles. To account for additional repair due to
increased traffic on damaged roads, a population-dependent scalar is applied to damage trajectories. This
scalar is based on the percent increase in damages across the century when the underlying model is run
with population growth compared to a run with static population. Impact estimates are calculated using
temperature binned damages multiplied by road miles and adjusted based on changes in road traffic. Note
that because repair under the proactive scenario strengthens road surfaces pre-emptively, before damage
occurs and with a planned road closure, delay times under the proactive adaptation scenario are
approximately half the projected delays for no adaptation and reactive adaptation- see Neumann et al. (In
review) for details.
FIGURE A-20. ROADS DATA PROCESSING FRAMEWORK
Input Data: Road damages, quarter-degree grids summed to regions
(GCM/year/region/{no adaptation, reactive adaptation, proactive adaptation})
Isolate Climate Impact: Subtract estimates from adaptation scenario-specific baselines
(GCM/year/region/{no adaptation, reactive adaptation, proactive adaptation})
Road Miles
(constant)
Bin Results by Temperature: Average damages over temperature bin eras by GCM
(degree/GCM/region/{no adaptation, reactive adaptation, proactive adaptation})
Scalable Tool Input: Divide by road miles
(degree/GCM/region/{no adaptation, reactive adaptation, proactive adaptation})
Economic Damage: Multiply by road miles and road traffic growth
(year/GCM/region/{no adaptation, reactive adaptation, proactive adaptation})
o
o
Limitations and Assumptions
• The model assumes a fixed capital and maintenance expense budget, which is usually exhausted at
some point under the no-adaptation scenario. This time dependency of the no adaptation scenario
is difficult to eliminate in the data processing steps, which could bias the estimate up or down,
depending on the speed of warming relative tothe underlying scenarios. This bias is expected to be
relatively small and the use of GCM average results minimizes this potential bias.
• Damages to vehicles associated with incompletely maintained roads are modeled only in the no
adaptation scenario; the model assumes roads are completed repaired and thus vehicles receive no
damage under the reactive and proactive adaptation scenarios.
• For further discussion of the limitations and assumptions in the underlying sectoral model see
Neumann etal. (In review), Neumann etal. (2014), and EPA (2017).
Page A-32
-------
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
Asphalt Roads
This sector study estimates the cost of
asphalt road maintenance associated with
climate change. Unlike the CIRA roads
sector, this sector does not model any
adaptation scenarios.
Future impacts are quantified by comparing
historical asphalt grades (values associating pavement temperature and performance) and those associated
with future climate projections. This analysis includes four roadway types: interstates, national routes, state
routes, and local roads. Impacts are based on the cost of maintaining the standard practice of material
selection for asphalt road maintenance ratherthan employing proactive pavement adaptation. Costs per
lane mile are multiplied by total regional asphalt lane miles to produce a total damage estimate in a region.
A summary of results by temperature bin is included in Figure A-21 below. Note that asphalt lane miles are
constant throughout the century, therefore only one set of impacts is shown in the figure.
FIGURE A-21. ASPHALT ROADS IMPACTS BY TEMPERATURE BIN DEGREE
Degrees of Warming
—*—CCSM4
• CanESM2
—•—GFDL-CM3
~ GIS5-E2-R
—*—HadGEM2-ES
A MIROC5
Climate Data Processing
This study was not part of the CIRA project and relies on different climate data that needed to be processed
for this sector to be included in the Framework. Underwood etal. (2017) select 19 climate models from
CMIP5, three of which (CanESM2, CCSM4, and MIROC5) overlap with the CIRA suite of GCMs, from the
archives of the Climate Analytics Group. Although this study used the same GCMs, the bias correction and
downscaling processes used by Climate Analytics Group differed from those used in the LOCA climate
dataset; therefore, new temperature bins are defined for the relevant new climate scenarios. RCP8.5
results are used for consistency with CIRA sectors. Maximum and minimum daily temperature data for
these three GCMs were processed in the 30-year periods employed by the study to determine future
annual temperatures associated with the era-level GCM-specific asphalt road damage estimates available
from the study. The CONUS baseline mean temperature was calculated from downscaled LOCA data that
matches the CIRA baseline period of 1986-2005, rather than the 1966-1995 U.S. Historical Climatology
UNDERLYING DATA SOURCES AND LITERATURE
Underwood, B.S., Guido,Z., Gudipudi, P.,& Feinberg, Y. (2017).
Increased costs to US pavement infrastructurefrom future
temperature rise. Nature Climate Change, 7,704-707.
Doi:10.1038/nclimate3390
Expert and Public Review Draft-Do Not Cite
Page A-33
-------
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
Network (USHCN) baseline data employed by Underwood et al. (2017). The USHCN baseline data is
available by station location and aggregatingto CONUS may introduce error relative tothe underlying
methodology. The LOCA data serves as the best available proxy for USHCN station data while also allowing
for comparison by degree across sectors, which use the same baseline. The LOCA baseline mean
temperature was subtracted from yearly projected temperature to identify GCM-specific integer degree
arrival years that were used for temperature binning.
Processing steps
Processing steps are seen in Figure A-22. In step one, total undiscounted costs and total lane miles by
weather station are aggregated to the regional level. These impacts are available for all GCMs and regions
for three eras: 2010 (2010-2039), 2040 (2040-2069), and 2070 (2070-2099), as well as a baseline era, which
are assigned to 1995 (1986-2005). Total costs, which refer tothe sum of impacts over each 30-year era, are
divided by 30 to reflect an annual cost.
Baseline impacts are then subtracted from projected impacts for each GCM to arrive at maintenance costs
associated with climate change for each era. Total costs attributable to climate change are divided by total
lane miles per region, resulting in a cost per lane mile for each era and GCM scenario. Next, era costs are
assigned tothe central year of the era (i.e., 2025, 2055, and 2085), and costs per era are transformed to
annual costs by interpolating linearly between era impacts. Finally, costs per lane mile are binned in GCM-
specific eleven-year windows around the first arrival times of integer degrees of warming relative to the
baseline. Costs per lane mile are scaled using total regional lane miles. Final impact estimates are calculated
based on regional lane miles and temperature-binned damages per road mile.
FIGURE A-22. ASPHALT ROADS DATA PROCESSING FRAMEWORK
do
c
cu
o
O
CU
Input Data: Total costs and total lane miles, weatherstations summed to regions
(era/GCM/region)
Isolate Climate Impact: Subtract baseline period costs from projected era cost
estimates
(era/GCM/region)
Calculate cost per lane mile: Costs divided by asphalt lane miles
(era/GCM/region)
Bin Results by Temperature: Interpolate between eras to annual time series, average
impacts over temperature bin eras by GCM
(degree/GCM/region)
Economic Damage: Multiply by lane miles
(year/GCM/region)
o
o
Page A-34
-------
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
Limitations and Assumptions
• The underlying study includes a suite of 19 climate models, three of which are part of the CIRA suite
ofGCMs (CanESM2, CCSM4, and MIROC5). These three models reach warmer temperatures more
quickly than the average across all 19 models in Underwood et al. (2017), and thus result in a higher
average estimate of damages compared to the results presented in the paper. However, compared
to the full suite of 38 CMIP5 GCMs, the three models are relatively close to the median temperature
change values in 2090.
• The model references, but does not quantify, impacts of a proactive adaptation scenario. Therefore,
uncertainty exists in how the modeled maintenance costs may be reduced as a result of adaptive
actions or technologies.
• For further discussion of the limitations and assumptions in the underlying sectoral model see
Underwood etal. (2017).
Urban Drainage
This sector study estimates the costs of
proactive adaptation for urban drainage
systems in 100 major coastal and non-
coastal cities of the contiguous U.S. to
meet future demands of increased runoff
associated with more intense rainfall
under climate change.
Adaptive actions focus on the use of best management practices to limit the quantity of runoff entering
stormwater systems and maintain current level of service (i.e., proactive adaptation to avoid damages),
instead of expanding formal drainage networks of basins and conveyance systems. These best management
practices generally include temporary storage above or below ground (e.g., bioswales, retention ponds), or
infiltration (e.g., permeable pavement), and are based on EPA guidelines and construction cost estimates
(see Price et al., 2014 for additional details).
Specifically, the analysis uses a reduced-form approach for projecting changes in flood depth and the
associated costs of flood prevention under future climate scenarios, based an approach derived from EPA's
Storm Water Management Model (SWMM). The approach assumes that the systems are able to manage
runoff associated with historical climate conditions and estimates the costs of implementing the adaptation
measures necessary to manage increased runoff due to climate change. Impacts are estimated in units of
average adaptation costs per square mile for a total of 100 cities across the contiguous U.S. for three
categories of 24-hour storm events (those with precipitation intensities occurring every 10, 25, and 50
years—metrics commonly used in infrastructure planning) and four future eras periods: 2030 (2020-2039),
2050(2040-2059), 2070(2060-2079), and 2090(2080-2099). A summary of results by temperature bin
degree is included in Figure A-23 below. Note that urban drainage impacts by degree are constant
throughout the century, therefore only one set of impacts is shown in the figure.
UNDERLYING DATA SOURCES AND LITERATURE
Price, J., Wright, L, Fant, C., &Strzepek, K. (2014). Calibrated
Methodology for Assessing Climate Change Adaptation Costs for
Urban Drainage Systems. Urban Water Jou rnal, 13 (4), 331-344.
Doi:10.1080/1573062X.2014.991740
Expert and Public Review Draft-Do Not Cite
Page A-35
-------
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
FIGURE A-23. URBAN DRAINAGE IMPACTS BY TEMPERATURE BIN DEGREE
Degrees of Warming
—*—CCSM4
~ CanESM2
GFDL-CM3
—~— GIS5-E2-R
—+—HadGEM2-ES
—A— MIROC5
Processing steps
Processing steps are seen in Figure A-24. First, the adaptation costs per square mile (weighted by area) for
the 50-year storm for each GCM, city, scenario, and era combination, are aggregated to the regions used in
NCA4 (and the Temperature Binning Tool).30 Unlike most other underlying studies, the Urban Drainage
study does not produce an annual time series of results, due in part to the impact of extreme events which
are not well-characterized at an annual scale. The binning process requires an annual time series from
which the 11-year windows of damages corresponding to each integer degree arrival by GCM can be pulled;
therefore, linear interpolation is used to create an annual time series of values for each GCM, scenario, and
region combination for the period 1995-2099, using the known damage values at each of the four eras.
Values are extrapolated for 2090-2099 using the linear trend observed between the 2070 and 2090 eras,
and values for years prior to 2030 are estimated by using 1995 as a baseline year; i.e., impacts are assumed
to be zero in 1995 and results are interpolated linearly between 1995 and 2030. Finally, impacts are binned
by integer degrees of warming for each GCM, scenario, and region combination. No physical or economic
scaling is required since analytic results are scaled to the NCA region level in Step 2.
30 For example, for a region with 2 cities, each with an area of 100 square miles, each city's area is divided by the sum of the areas,
resulting in a proportion value of 0.5 for each city. This proportion value is then multiplied by each calculation of per-square-mile
adaptation costs (calculated by storm, scenario, and year) to produce a weighted average adaptation cost per square mile.
Page A-36
-------
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
FIGURE A-24. URBAN DRAINAGE DATA PROCESSING FRAMEWORK
Limitations and Assumptions
• The underlying analysis assumes that the systems are able to manage runoff associated with
historical climate conditions and estimates the costs of implementing the adaptation measures
necessary to manage increased runoff due to climate change.
• Inclusion of all U.S. cities with stormwater conveyance systems would provide a more
comprehensive characterization of future impacts. Therefore the current estimates included for
this sector represent underestimates of potential damages.
• For further discussion of the limitations and assumptions in the underlying sectoral model see
Neumann et al. (2014), Price et al. (2014), and EPA (2017).
Expert and Public Review Draft-Do Not Cite
Page A-37
-------
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
A.4 Water Resources Sectors Data Processing
Water Quality
This analysis estimates damages in terms
of the change in willingness to pay to avoid
changes in water quality. This analysis
estimates climate change effects on water
quality at the eight-digit HUC scale of the
contiguous U.S. using the Hydrologic and
Water Quality System (HAWQS)
biophysical model. Note that the damages
estimated for this sector only cover the
change in value of recreation opportunities
and do not include the value of health
effects or other amenities associated with
clean water.
UNDERLYING DATA SOURCES AND LITERATURE
Fant, C., Srinivasan, R., Boehlert, B., Rennels, L, Chapra,S. C.,
Strzepek, K. M., Corona, J., Allen, A., & Martinich, J. (2017). Climate
change impacts on US water quality using two models: HAWQS and
US Basins. Water, 9(2), 118. Doi:10.3390/w9020118
Boehlert, B., Strzepek, K. M., Chapra,S. C., Fant, C., Gebretsadik, Y.,
Lickley, M., Swanson, R., McCluskey, A., Neumann,.!.,& Martinich,J.
(2015). Climate change impacts and greenhouse gas mitigation
effects on US water quality. Journal of Advances in ModelingEarth
Systems, 7,1326-1338. Doi:10.1002/2014MS000400
Yen, H., Daggupati, P., White, M. J., Srinivasan, R., Gossel, A., Wells,
D., & Arnold, J. G. (2016). Application of large-scale, multi-resolution
watershed modelingframeworkusingthehydrologicand water
quality system (HAWQS). Water, 8(4), 164. Doi:10.3390/w8040164
HAWQS advances the functionality of the widely used and accepted Soil and Water Assessment Tool
(SWAT), providing a platform for water quality modeling, primarily by minimizing the necessary initialization
time. Originally developed by the U.S. Department of Agriculture (USDA), SWAT has been the core
simulation tool for numerous U.S. national and international assessments of soil and water resources. The
use of HAWQS over SWAT improves the ease of application to national scale analyses while still simulating a
large array of watershed processes for a defined period of record.
The HAWQS model follows a broad modeling sequence: (1) the landscape phase, where the primary
processes are climate, soil water balance, nutrient and sediment transport and fate, land cover, plant
growth, farm management, and (2) the main channel phase, where the main processes are river routing,
and sediment and nutrient transport through the rivers and reservoirs.
The HAWQS model projects changes in water quality parameters and simulated changes in river flow for
five climate models under RCP8.5 and RCP4.5. These projections include future municipal wastewater
treatment plant loadings (point source) scaled to account for population growth. Changes in overall water
quality are estimated using changes in a Climate-oriented Water Quality Index (CWQI), a metric that
combines multiple pollutant and water quality measures. Four water quality parameters (water
temperature, dissolved oxygen, total nitrogen, and total phosphorus) are aggregated from the eight-digit
HUC level to the Level-Ill Ecoregions, weighted by area.31 Finally, a relationship between changes in the
CWQI and changes in the willingness to pay for improving water quality is used to estimate the economic
implications of projected water quality changes. For more information on the approach and results for the
31 Designed to serve as a spatial framework for environmental resource management, ecoregions denote areas within which ecosystems
(and the type, quality, and quantity of environmental resources) are generally similar. Ecoregions were originally created to support the
development of regional biological criteria and water quality standards, and to set management goals for nonpoint source pollution.
Page A-38
-------
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
water quality sector, please refer to Fant et al. (2017), Boehlert et al, (2015), and Yen et al. (2016).
Specifically, impacts are estimated as per capita change in the willingness to pay to improve water quality
for two future eras: 2050 (2040-2059) and 2090 (2080-2099). A summary of results by temperature bin
degree in 2010 and 2090 (the endpoints of socioeconomic modeling) is included in Figure A-25 below.
FIGURE A-25. WATER QUALITY IMPACTS BY TEMPERATURE BIN DEGREE
A. 2010 SOCIOECONOMICS
Degrees of Warming
—*—CCSf.M
—•— CanESM2
GFDL-CM3
~ G1SS-E2-R
—•— HadGEM2-E$
—A— MIROC5
B. 2090 SOCIOECONOMICS
Degrees of Warming
* CCSM4 • CanESM2~
¦ GfDL-CM3 • GISS-E2-R
—•— HadGEM2-ES —A— MIFtOCS
Processing steps
Processing steps are seen in Figure A-26. The per capita willingness to pay for each EPA Level 3 Ecoregion,
GCM, and era combinations are first aggregated to the state-level and then tothe regions used in NCA4
(and the Temperature Binning Tool). Like the Urban Drainage sector described above, the Water Quality
study also does not produce an annual time series of results. Therefore, an annual time series of damages
needs to be constructed for each GCM and region combination based on available data (i.e., for the 2050
and 2090 eras). Linear interpolation is used tocreate an annual time series of values for each GCM and
region combination for the period 1995-2099. Values are extrapolated for 2090-2099 using the linear trend
observed between 2050 and 2090, and values for years prior to 2050 are estimated by using 1995 as a
baseline year; i.e., impacts were assumed to be zero in 1995 and results are interpolated linearly between
1995 and 2050. Finally, impacts are binned by integer degrees of warming for each GCM and region
combination. Impact estimates are calculated by applying regional population as a physical scalar.
Expert and Public Review Draft-Do Not Cite
Page A-39
-------
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
FIGURE A-26. WATER QUALITY DATA PROCESSING FRAMEWORK
Limitations and Assumptions
• Decreases in water quality due to climate change will likely have adverse effects on human health
and the environment that are not represented in the results of this section. For example, climate
change impacts to water quality may affect ecological dynamics of freshwater systems, with
cascading effects on ecosystem services and recreational opportunities.
• This analysis only considers four water quality parameters, and omits other constituents, such as
sediment and heavy metals, that may be affected by changes in the climate system.
• The methods underlying the analysis do not consider the effects of climate change-induced extreme
events on water quality, such as increased siltation and runoff following wildfire events.
• The analysis considers only a subset of all use/non-use values linked to water quality changes,
therefore the damages reported here are likely underestimates of future impacts.
• By creating an annual time series for the period 1995 to 2100 based on values from 2050 and 2090
only, the Temperature Binning processing does not capture any non-linearities in the relationship
between damages and temperature, particularly in the early years of the century.
• For further discussion of the limitations and assumptions in the underlying sectoral model, see Fant
et al. (2017) and Boehlert et a I. (2015).
Winter Recreation
This sector estimates lost revenue due to climate
change to suppliers of three types of winter
recreation occurring at 247 sites across the U.S.:
alpine skiing, Nordic skiing, and snowmobiling.
Damages are based on the number of visits to
winter recreational sites, entrance fees, and state-
level average ticket prices. The model was run using both 2010 and 2090 ICLUSv2 population. A summary of
results by temperature bin degree in 2010 and 2090 (the endpoints of socioeconomic modeling) is included
in Figure A-27.
UNDERLYING DATA SOURCES AND LITERATURE
Wobus, C., Small, E. E., Hosterrrian, H., Mills, D., Stein, J.,
Rissing, M,, Jones, R., Duckworth, M., Hall, R., Kolian, M.,
Creason,J.,& Martinich,J. (2017). Projected climate change
impacts on skiing and snowmobiling: A case study of the
United States. Global Environmental Change, 45,1-14.
Page A-40
-------
2059
2060
2061
2062
2063
2064
FIGURE A-27. WINTER RECREATION IMPACTS BY TEMPERATURE BIN DEGREE
A. 2010 SOCIOECONOMICS
Winter Recreation; Alpine Skiing
Winter Recreation; Cross-Country Skiing
2 3 4
Degrees of Warming
2 3 4
Degrees of Warming
- CCSM4 —•— CanESM2
- GFDL-CM3 —~— GISS-E2-R
- H*tGEM2-ES • -A- - MIROCS
—«—CCSM4
• Can£SM2
=«— GFDL-CM3
» GISS-E2-R
—*—HadGEM2-ES
A MIROCS
Winter Recreation; Snowmobiling
2 3 4
Degrees of Warming
—h CCSM4
• Can£SM2
—•— GFDL-CM3
» G1SS-E2-R
—•— HadGEM2-ES
A. MIROCS
B. 2090 SOCIOECONOMICS
Winter Recreation; Alpine Skiing
Winter Recreation; Cross-Country Skiing
2 3 4
Degrees of Warming
2 3 4
Degrees of Warming
" CCSM4
—•— Can£SM2
• GFDL-CM3
—~— GISS-E2-R
—•—HacK3EM2-ES
A MIROCS
—h—CCSM4
• CanESM2
¦ GFDL-CM3
~ G1SS-E2-R
—~—HadGEM2-ES
A MIROCS
Winter Recreation; Snowmobiling
2 3 4 5
Degrees of Warming
2065
—"— CCSM4
—•— CanESM2
—•— GFDL-CM3
—~— GISS-E2-R
• Ht>3GEM2-ES
—A— MIROC5
Expert and Public Review Draft-Do Not Cite
Page A-41
-------
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
Processing steps
Processing steps are shown in Figure A-28. In step one, lost ticket sales are estimated across recreational
activities (alpine skiing, Nordic skiing, and snowmobiling), GCMs, degrees, and regions, for both a 2010 and
2090 population.
In the second step, both the 2010 and 2090 regional estimates produced in step one are divided by regional
population. Regional population estimates are based on ICLUSv2 population data. This produces a per
capita estimate of forgone ticket sales for alpine skiing, Nordic skiing, and snowmobiling. The per capita
cost estimates are multiplied against regional population to produce total lost revenue estimates.
FIGURE A-28. WINTER RECREATION DATA PROCESSING FRAMEWORK
Input Data: Forgone Ticket Sales: alpine skiing, nordic skiing, and snowmobiling
(degree/GCM/region/{2010pop,2090pop})
Forgone
Revenue per
Capita
(constant)
Isolate Climate Impact: Baseline already removed in Forgone Revenue input
Scalable Tool Input: Divide by Regional Population estimates
(degree/GCM/region/{2010pop, 2090pop})
Economic Damage: Multiply by lost ski revenue and regional populations
(year/GCM/region/{2010pop,2090pop})
O
o
Limitations and Assumptions
• The scope of winter recreation loss for the tool is derived only from analysis of the alpine skiing,
Nordic skiing, and snowmobile sub-sectors of the industry. Potential losses to other winter
recreation activities (e.g., tubing) are not quantified in this study.
• Potentially compensating adaptations from the lost opportunity to engage in winter recreation (for
example, with other forms of outdoor recreation, or with indoor recreation) are not considered.
• For further discussion of the limitations and assumptions in the underlying sectoral model, see
Wobus et al. (2017b) and EPA (2017).
Page A-42
-------
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
A.5 Electricity Sectors Data Processing
Electricity Demand and Supply
This sector estimates increases in system
costs to the power sector. These system
costs include capital, fuel, variable
operation and maintenance (O&M), and
fixed O&M costs.
Increased costs are based on projected
changes in demand for and supply of
electricity across generation types. Effects
on energy demand reflect the net impact of increased demand for residential, commercial, and industrial
space cooling during summer/warmer months, and decreased demand for space heating during
winter/cooler months. Effects on supply reflect the decreased production capacity of thermal power plants,
and transmission capacity of the transmission system, associated with higher temperatures.32 The complex
interplay of supply and demand, coupled with forecast changes in fuel and energy production technology
availability and prices, are modeled using the Global Change Assessment Model (GCAM-USA), a detailed
service-based building energy model with a 50-state domain.
Costs are provided for a static baseline run, in which climate is held as constant to the CIRA baseline while
socioeconomic variables are dynamic, and a projection run in which both climate and socioeconomic
variables are changing. Estimates of costs with- and without-climate change are provided in five-year
intervals. A summary of results by temperature binning degree in 2010 and 2090 (the endpoints of
socioeconomic modeling) is provided in Figure A-29 below.
FIGURE A-29. ELECTRICITY DEMAND AND SUPPLY IMPACTS BY TEMPERATURE BIN DEGREE
A. 2010 SOCIOECONOMICS
Electricity Demand and Supply
30,000
to 25,000
o
« 20.000
W
is
° 15,000
"5
£ 10,000
Q
S 5,000
0 1 2 3 4 5 6
Degrees of Warming
—*—CCSM4
—*— CanESM2
GFDL-CM3
—~— GIS5-E2-R
—*—HadGEM2-ES
A MIROC5
32 Note that the transmission system effects in this sector are separate from those modeled in the Electricity Transmission and
Distribution Infrastructure sector.
UNDERLYING DATA SOURCES AND LITERATURE
McFarland,J.,Zhou, Y., Clarke, L, Sullivan, P., Colman, J., Jaglom,
W.S., Colley, M., Patel, P., Eom,J., Kim, S. H., Kyle, G. P., Schultz,
P., Venkatesh, B., Haydel, J., Mack, C., & Creason, J. (2015).
Impacts of rising air temperatures and emissions mitigation on
electricity demand and supply in the United States: a multi-model
comparison. Climatic Change, 131,111-125. Doi:10.1007/sl0584-
015-1380-8
Expert and Public Review Draft-Do Not Cite
Page A-43
-------
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
B. 2090 SOCIOECONOMICS
Degrees of Warming
—*—CCSM4
—•— CanESM2
• GFDL-CM3
~ GISS-E2-R
—*—HadGEM2-ES
—A— MIROC5
Processing steps
Processing steps for this sector are shown in Figure A-30. System costs for the power sector are provided
for each GCM and a climate reference scenario for each region in 5-year intervals. Annual costs are
interpolated between the 5-year interval data and costs are binned by degrees of warming for each GCM
and region. To remove time dependencies, the final temperature binned estimates arethe percentage
change in costs from the reference scenario for each temperature bin, GCM, and region.
For a given input temperature trajectory, the percentage changes in system cost are multiplied by the
reference scenario costs to produce total cost estimates across the century. That is, damages in a given
year are dependent on warming, which maps a percentage change in costs from the reference scenario
based on temperature binned damages, and the baseline system costs in the reference scenario for that
year.
FIGURE A-30. ELECTRICITY DEMAND AND SUPPLY DATA PROCESSING FRAMEWORK
Page A-44
-------
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
Limitations and Assumptions
• Projected changes in heating degree days (HDD) and cooling degree days (CDD) are based on a
temperature set-point of 65°F, a common convention that may lead to a conservative energy
demand estimate.
• The temporal aggregation of the underlying electricity supply model is too coarse to assess the
impact of extreme temperature events that occur on only the very hottest days of the year. As a
result, the underlying study focuses on a single aspect of climate change: average ambient air
temperature, and therefore omits effects of extreme temperature effects on peak demands and the
loads required to meet those changes. Effects from future changes in the frequency and magnitude
of extreme temperatures may stress electric power systems, and these economic risks are not
captured in this study.
• For further discussion of the limitations and assumptions in the underlying sectoral model, see
McFarland et al. (2015).
Electricity Transmission and Distribution Infrastructure
This analysis estimates damages to the
electric transmission and distribution
infrastructure due to climate change. This
multi-dimensional analysis considers a
wide range of climate stressors, including
extreme temperature, extreme rain,
lightning, vegetation growth, wildfire
activity, and coastal flooding. Impact receptors include transmission and distribution lines, poles/towers,
and transformers.
Monetized damages for this sector are the costs of repair or replacement of damaged infrastructure. The
impact model for this sector was run under two infrastructure system scenarios: one with expansion of
infrastructure associated with demand growth, and one with static infrastructure. Increases in demand
growth may be due to population growth, or increased demand due climatic change — in particular,
warmer temperatures increase usage of air-conditioning. The model identifies changes in performance and
longevity of physical infrastructure, such as power poles and transformers, and quantifies these impacts in
economic terms. While certain climate stressors do cause power outages which have associated direct and
indirect economic costs, these damages are not included in damage estimates.
Like other infrastructure sectors, the analysis is based on three adaptation scenarios. These include
proactive adaptation, reactive adaptation, and no adaptation. Repair costs are also allocated based on the
activity being performed. These activities include transmission line capacity, wildfire repair, tree trimming,
substation sea-level rise, substation storm surge, wood pole decay, transmission transformer lifespan, and
distribution transformer lifespan. Figure A-31 below provides a summary of the results by temperature
binning degree and adaptation scenario in 2010 and 2090 (the endpoints of socioeconomic modeling).
UNDERLYING DATA SOURCES AND LITERATURE
Fant,C., Boehlert, B., Strzepek, K., Larsen, P., White, A., Gulati, S., Li,
Y., & Martinich, J. (2020). Climate change impacts and costs to U.S.
electricity transmission and distribution infrastructure. Energy, 195.
Doi:10.1016/j.energy.202 0.116899
Expert and Public Review Draft-Do Not Cite
Page A-45
-------
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
2163
2164
2165
2166
2167
2168
2169
FIGURE A-31. ELECTRICITY TRANSMISSION AND DISTRIBUTION INFRASTRUCTURE IMPACTS BY
TEMPERATURE BIN DEGREE
A. 2010 SOCIOECONOMICS
Electricity Transmission and Distribution; No Adaptation
Electricity Transmission and Distribution; Proactive Adaptation
2 3 4
Degrees of Warming
2 3 4
Degrees of Warming
—«—CCSM4
• CanESM2
—•— GFDL-CM3
~ GISS-E2-R
—•— HadGEM2-£S
—A— MIRCC5
—«—CCSM4
—•— CanESM2
—•— GFDL-CM3
» GISS-E2-R
—•—HadGEM2-ES
—A— MIROC5
Electricity Transmission and Distribution; Reactive Adaptation
2 3
Degrees of Warming
-CCSM4
GFDL-CM3 -
-HacG£M2-ES
-CanESM2
- GISS-E2-R
- MIROC5
B. 2090 SOCIOECONOMICS
Electricity Transmission and Distribution; No Adaptation
25.000
Electricity Transmission and Distribution; Proactive Adaptation
2 3
Degrees of Warming
2170
—"—CCSM4
—•— CanESV12
—GFDL-CM3
—~—GISS-E2-R
—HadGEM2-ES
—A— MIROC5
Page A-46
-------
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
Processing steps
Processing steps are seen in Figure A-32. The underlying impact model produces damage estimates for
each infrastructure type, GCM, and adaptation scenario. There are nine infrastructure types, seven of which
grow with electricity demand. Therefore, damage estimates for these seven infrastructure types are
influenced by both population and climate. To isolate damages associated with warming, damages
associated with static demand are scaled by growth of demand attributable to warming. Demand
attributable to warming is calculated based on the percentage increase in demand across the century for
each GCM from a baseline demand across the century with a constant climate, as seen in Figure A-33.
After damages associated with climate driven infrastructure growth are calculated, results are aggregated
for each GCM, infrastructure type, degree, and region combination. These costs are calculated for each
adaptation scenario. Costs are then aggregated from individual infrastructure types tothe sector total.
A population driven demand scalar is implemented to account for increases in demand for the grid as
population grows. Thus, final damage estimates include expansion of electric grid infrastructure associated
with a warming climate and with population growth. Note that because these damage estimates rely on an
empirical relationship between damages with and without infrastructure growth in the underlying impact
model, these damage estimates cannot be adjusted for custom input population trajectories.
FIGURE A-32. ELECTRICITY TRANSMISSION AND DISTRIBUTION INFRASTRUCTURE DATA PROCESSING
FRAMEWORK
Input Data: Infrastructure Repair and Replacement, states summed to regions
(GCM/year/region/{no adaptation, reactive adaptation, proactive adaptation})
CUD
c
o
O
Isolate Climate Impact: Scale damages without infrastructure growth to percent
change in damages with infrastructure growth from no-climate run to GCM runs
(GCM/year/region/{no adaptation, reactive adaptation, proactive adaptation})
Bin Results by Temperature: Average damages over temperature bin eras by GCM
(degree/GCM/region/{no adaptation, reactive adaptation, proactive adaptation})
Scalable Tool Input: Damages already scaled in Temperature Binned results
(degree/GCM/region/{no adaptation, reactive adaptation})
Economic Damage: Multiply by demand growth
(year/GCM/region/{no adaptation, reactive adaptation, proactive adaptation})
o
o
Expert and Public Review Draft-Do Not Cite
Page A-47
-------
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
FIGURE A-33. ELECTRICITY TRANSMISSION AND DISTRIBUTION INFRASTRUCTURE DEMAND GROWTH
SCALAR PROCESSING
Demand Growth Scalars
Limitations and Assumptions
• The model assumes that grid demand is controlled by population change and climatic factors; grid
demand is assumed to not be influenced by economic growth. Future changes in the design and
structure of electric grids are not considered in this study.
• Two of the nine infrastructure types considered in this study do not scale with changes in
population.
• For further discussion of the limitations and assumptions in the underlying sectoral model, see Fant
eta I. (2020)
Page A-48
-------
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
APPENDIX B | R TOOL DOCUMENTATION
The R Tool is an R package called 'ciraTempBiri, and consists of one main temperature binning function,
several pre- and post- processing functions, and additional helper functions. The package is currently available
for download through GitHub, and may eventually be submitted for disseminationthrough the Comprehensive
R Archive Network(CRAN).33The'ciraTempBiri package depends on several additional, widely used R
packages - 'dplyrVtidyr', and ' ggplot2\ Package installation requires 'devtools\ All package
dependencies are available through CRAN.
The Tool implements the Frameworkdescribed in this report. The first section in this appendix presents an
overview of the main elements of the 'ciraTempBiri package. The second section provides detailed
descriptions of each of the functions in the package. For more details on the underlying methodology, refer to
the main documentation report.
B.l ^ciraTempBin' Overview
The RTool includes one main function for temperature-andSLR-binning, and several additional pre-and post-
processing functions. The primary functions and their main arguments are highlighted in the table below. The
function dependencies are shown in the following figure.
Function Name
Function Type
Inputs
tempBin
Main temperature-and SLR-
binning function
Custom scenarios for temperature, global mean sea
level rise (GMSL), population, and gross domestic
product (GDP)
import inputs
Function for importing custom
scenarios from CSV files
Paths to CSV files containing custom scenarios
get plots
Create plots of temperature-and
SLR-binned impacts (heatmaps
and impacts over time)
Outputs of temperature-and SLR-binning
Function dependencies are shown in the Figure B-l.
33 For more information on CRAN, visit https://cran.r-project.org/.
-------
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
2221 FIGURE B-l FUNCTION DEPENDENCIES IN THE CIRATEMPBIN' R PACKAGE
Helper Functions Primary Functions
2222
2223 Main temperature- and SLR-binning function
2224 • "tempBinO* - Project annual average climate change impacts throughout the 21st century for
2225 available sectors.
2226 Pre-processing functions
2227 Primary functions:
2228 • " import_inputs ()' - Import custom scenarios for climate and socioeconomics (temperature and
2229 global mean sea level rise (GMSL), population, and GDP) from user-specified file names.
2230 Helperfunctions:
2231 • " convertTemps ()' - Convert contiguous U.S. (CONUS) temperaturestoglobal temperatures orvice
2232 versa.
2233 • "temps2slr ()' - Convert global temperature change in degrees Celsiustoglobal mean sea level rise
2234 (GMSL) in centimeters.
2235 Post-processing functions
2236 Primary functions:
2237 • "get_plots ()' - Create and save plots for summarizedtemperature binning outputs
Page B-2
-------
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
Helper functions:
• "aggregate_impacts ()" - Summarize and aggregate impacts from temperature binning. Calculate
national totals, average across models, sum impact types, and interpolate between impact year
estimates.
Included datasets (default outputs)
The RTool contains a dataset with default results "defaultResults", which contains annual impacts
produced by "tempBin ()" for the default scenarios (i.e. default temperature, GDP and regional population
trajectories), andean be loaded into the R workspace ('load (defaultRe suits) )".
B.2 ^cira TempBin' Function Details
B.2.1 Primary Functions
B.2.1.1tempBin
Project annual average climate change impacts throughout the 21st century for available sectors.
Description
This function allows users to project annual average climate change impacts throughout the 21st
century (2010-2090) for available sectors. Users may specify an optional list of custom scenarios using
"inputsList". The output of "tempBin ()" is an Rdata frame object containing annual average
impacts, by year, for each sector, adaptation, impact type, model (GCM or SLR scenario), and region.
Usage
tempBin(inputsList=NULL, sectorsList=NULL, aggLevels=NULL, pv=TRUE,
baseYear=2010, rate=0.03, silent=TRUE)
Arguments
inputsList
A list of named elements (names ("inputsList) =
c("templnput", "slrlnput", "gdplnput",
"poplnput")"), each containing dataframes ofcustom temperature,
global mean sea level rise (GMSL), gross domestic product (GDP), and/or
population scenarios, respectively, overthe period 2010 to 2090. For
more information, see "import inputs ()". Values for each scenario
type must be within reasonable ranges. For more information, seethe
details, below, for "tempBin ()" and documentation for
"import inputs ()".
sectorsLi st
A character vector indicating a selection of sectors for which to calculate
results. If "NULL", all sectors are included.
aggLevels
Levels of aggregation at whichto summarize data: one or more of
"c ( "national", "modelAverage", "impactYear",
"impactType", "all")". Defaultstoall levels (i.e.,
"aggLevels="all""). Uses the same aggregation levels as
"aggregate impacts ()").
pv
A "TRUE/FALSE" value indicating whetherto calculate present values
for the annual impacts. Defaults to "pv=TRUE". Present values (i.e.,
discounted impacts) are calculated as "discounted impacts=
Expert and Public Review Draft-Do Not Cite
Page C-3
-------
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
annual impacts/(1+rate)A(year-baseYear)\ Setan
annual discounting rateand a baseyearusing 'baseYear' and 'rate',
respectively.
baseYear
Base year used for calculating present values of annual impacts (i.e.,
discounting). Defaults to 'baseYear=2010'.
rate
Annual discount rate used in calculating present values (i.e., discounting)
annual impacts. Defaultsto'rate=0. 03' (i.e., 3% per year).
silent
A 'TRUE/FALSE' value indicating the level of messaging desired by the
user(default='FALSE').
Details
This function allows users to project annual average climate change impacts throughout the 21st
century for available sectors. 'tempBin ()' isthe main function in the R Tool. The Tool implements the
methods of the Temperature Binning Framework, as described in this report.
Users can specify an optional list of custom scenarios with'inputsList' and specify a selection of
sectors with 'sectorsList'. 'tempBin ()' uses default scenarios for temperature, population, and
GDP when no inputs are specified (i.e.,'inputsList' is'NULL') or for empty elements of the inputs
list. If the user does not specify an input scenario for GMSL(i.e., 'inputsList=list (slrlnput=
NULL)' , 'tempBin ()' first convertsthe CONUStemperature scenarioto global temperatures and
then convertsthe global temperaturestoa global mean sea level rise (GMSL) height in centimeters. For
more information on the conversion of CONUStemperaturestoglobal temperatures, see
'convertTemps () '. For more information on the conversion of global temperatures toGMSL, see
'temps2slr()'.
Values for input scenarios must be within reasonable ranges. Temperatures must be in degrees Celsius
and values must be greaterthanor equal to zero and less than or equal to 10 degreesof warming.
Values for GMSL must be in centimeters(cm) and values must be greaterthanor equal to zeroand less
than or equal to 150 cm. Population and GDP values must be greaterthanor equal to zero. Ifa user
inputs a custom scenario with values outside the allowable ranges, 'tempBin ()' will not run the
scenarios and will instead stop and return an error message.
If'inputsList=NULL', the temperature-and SLR-binning function 'tempBin ()' uses defaults for
all scenarios. Otherwise, 'tempBin ()' looks for a list object passedtothe argument 'inputsList'.
Within that list,'tempBin ()' looks for list elements 'templnput', 'slrlnput', 'gdplnput', and
'poplnput' containing dataframes with custom scenarios for temperature, GMSL, GDP, and regional
population, respectively. 'tempBin ()' will default backtothe default scenarios for any list elements
that are'NULL' or missing. In other words, running'tempBin (inputsList=li st () )' returns the
same outputs as running 'tempBin () \ For help importing custom scenarios from CSV files, refer to
the pre-processing function 'import_inputs () \
'tempBin ()' linearly interpolates between values for all (both custom and default) to the annual level.
Temperatures are interpolated using 1995 as the baseline year (i.e., the central year of the 1986-2005
baseline). GMSL is interpolated using 2000 as the baseline year. In other words, the temperature (in
degrees Celsius) is set to zero for the year 1995 and GMSL is set to zero for the year 2000. The
interpolated temperature and GMSL scenarios are bound together into a joined scenario of "drivers",
Page B-4
-------
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
with temperature and GMSL values combined into a column called 'driverValue', and additional
columns for year, the driver unit ("degrees Celsius" and "cm" for temperature and GMSL, respectively),
and the associated modeltype ("GCM" or "SLR").
The population scenario is interpolated at the annual level for each region. National totals for
population are calculated from the interpolated values. The interpolated population scenario is joined
with values for the national total population and the annually interpolated GDP scenario to create a
socioeconomic scenario. GDP per capita is calculated by dividing GDP by national population.
Each sector is associated with a model type ("GCM" or "SLR"). With different binned and scaled impacts
for each adaptation, impact type, impact yearestimate, region, model type, and model. 'tempBin ()'
initializes a results dataframeforall such combinations by joining the socioeconomic scenariowith the
annual driver values for the associated modeltype. The annual scaled impacts for each combination are
calculated by passing the associated driver value to associated scaled impact function and added to the
results dataframe.34
Each sector, adaptation, impact type, and impact year combination included inthe model isalso
associated with a specified physical scalar, a physical adjustment, a damage adjustment, an economic
scalar, and an economic multiplier. 'tempBin ()' joins the scalardata for each such combination
(updated to reflect user inputs) with the results dataframe.35 Annual impacts a re calculated by
multiplying the scaled impacts by the physical scalar(adjusted by the physical adjustmentand damage
adjustment) and economic scalars (adjusted by the economic multiplier). Physical impacts are calculated
by multiplying the scaled impacts by a physical scalar(adjusted by the physical adjustment and damage
adjustment) and are shown for selected impact types.
'tempBin ()' aggregates or summarizes resultsto levels of aggregation specified by the user (passed
to 'aggLevels') usingthe post-processing helperfunction 'aggregate_impacts ()' (see
'aggregate_impacts ()'). Users can specify a single aggregation level or multiple aggregation levels
by passing a single characterstringorcharactervectorto 'aggLevels'. Options for aggregation
include calculating national totals faggLevels="national"'), averaging across modeltypes and
models ('aggLevels="modelAverage"'), summing over all impact types
faggLevels="impactType "'), and interpolate between impact year estimates
faggLevels="impactYear "'). Users can specify all aggregation levels at once by specifying
'aggLevels="all"' (default) or no aggregation levels faggLevels="none").
For each of the'aggLevels', 'tempBin ()' performs the following summarization (using
'aggregate_impacts ()'):
• '"national"': Annual values are summed across all regions present in the data. I.e., data is
grouped by columns'"sector", "adaptation", "model_type", "model",
"year"' (and '"impactType"' and/or'"impactYear"' if they are present in the data)
and summed across regions. Years which have missing column data for all regions return as 'NA'
34 The R Tool includes a list of approximation functions for each sector, adaptation, impact type, impact year estimate, region, model
type, and model combination. These functions were created from thetemperature- and SLR-binned data described elsewhere in this
report.
35 't empBin () * includes data on physical scalars, physical adjustments, damage adjustments, economic scalars, and economic
multipliers. The scalar data is updated to reflect user inputs. More specifically "tempBin () * updates the physical scalar data for
regional population and updates the economic multiplierto reflect user inputs for population and/or GDP.
Expert and Public Review Draft-Do Not Cite
Page C-5
-------
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
(missing). The rows of the datafra me of national values (with column 'region="National
Total"') arethen bound to the regional values.
• '"modelAverage"': Annual results are averaged across all models present in the data, i.e.,
data is grouped by columns '"sector", "adaptation", "model_type",
"region", "year"'(and'"impactType"' and/or'"impactYear"' if they are present
in the data) and averaged across models. Averages exclude missing values. Years which have
missing column data for all models return as 'NA' (missing). The rows of the datafra me of model
averages (with column 'model="Average"') arethen bound tothe rows with individual
model results.
• '"impactType"': Annual results are summed across all impact types by sector present in the
data. I.e., data isgrouped by columns '"sector", "adaptation", "model_type",
"model", "region", "year"' (and '"impactYear"' if present in the data) and
summed across impact types. Mutates column 'impactType="all"' for all values. Years
which have missing column data for all impact types return as 'NA' (missing). If results are
aggregated across impact types, information about physical impacts (dataframe columns
'"physicalmeasure"' and'"physical_impacts"') are dropped.
• '"impactYear"': Annual results for sectors with only one impactyear estimate (i.e.,
'impactYear=="N/A"') a re separated from those with multiple impact yearestimates. For
sectors with multiple impact years (i.e. 2010 and 2090 socioeconomic runs), annual results are
interpolated between impact yearestimatesforapplicable sectors i.e., data isgrouped by
columns'"sector", "adaptation", "model type", "model", "region",
"year"' (and '"impactType"' if present in the data) and interpolated across years withthe
2010 run assignedtoyear2010 and the 2090 run assigned to year 2090. The interpolated values
are bound back to the results for sectors with a single impact year estimate, and column
'impactYear' setto'impactYear="interpolate"' for all values.
Users can choose to calculate present values of annual impacts (i.e., discounted impacts), by setting
'pv=TRUE'. Discounted impacts are calculated using a baseyearand annual discount rateas
'discounted impacts=annual impacts/(1+rate ) A (year-base Year)'. Set baseyear
and annual discount rate using'baseYear' (defaultsto 'baseYear=2010') and'rate' (defaults to
3% i.e., 'rate=0. 03'), respectively.
Outputs
The output of 'tempBin ()' is an Rdata frame object containing annual average impacts, by year
(2010-2090), for each sector, adaptation, model (GCM or SLR scenario), and region.
Examples
### View example scenario names
scenariosPath %>% list.files
### Temperature Scenario File Name
tempInputFile <- scenariosPath %>% paste("temp_scenariol.csv", sep="/")
### SLR Scenario File Name
slrlnputFile <- scenariosPath %>% paste ("slr_20Ocm. csv" , sep="/")
### Population Scenario File Name
popInputFile <- scenariosPath %>% paste("pop_iclusV5.csv", sep="/")
### Import inputs
example_inputsList <- import_inputs(
tempfile = tempInputFile,
slrfile = slrlnputFile,
Page B-6
-------
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
popfile = popInputFile
)
### Run custom temperature scenario and output impacts without aggregation
and with present values (default base year and discount rate)
df_tempExOut <- tempBin(inputsList= tempBin_inputs, aggLevels="none",
pv=TRUE, s ilent=TRUE) ~
B.2.1.2 importjnputs
Import custom scenarios for temperature- and SLR- binning from user-specified file names.
Description
This function enables usersto import data on custom scenarios for use with temperature binning. Users
specify path names to CSVfiles containing temperature, global mean sea level rise (GMSL), gross
domestic product (GDP), and population scenarios. 'import_inputs () 'reads in and format any
specified files as data frames and returns a list of dataframes for imported scenarios.
Usage
import inputs (tempfile=NULL, slrfile=NULL, popfile=NULL, gdpfile=NULL,
popform="spread", temptype="conus")
Arguments
tempf ile
A characterstring indicatingthe location of a CSVfile containing a
customtemperature scenario (first column contains years in the interval
2010 to 2090; second column contains temperatures, in degrees Celsius,
above the baseline).
slrfile
A characterstring indicatingthe location of a CSVfile containing a
custom sea level rise scenario (first column contains years in the interval
2010 to 2090; second column contains values for global mean sea level
rise (GMSL), in centimeters, abovethe baseline scenario).
popfile
A characterstring indicatingthe location of a CSVfile containing a
custom population scenario for NCA regions. The first column contains
years in the interval 2010 to 2090. The number of additional columns,
column names, and column contents of depend on the population
format set by 'popf orm'. For more details, see 'popfile \
gdpfile
A characterstring indicatingthe location of a CSVfile containing a
custom scenario for gross domestic product (GDP) (first column contains
years in the interval 2010 to 2090; second column contains values for
GDP, in total 2015$).
temptype
A characterstring indicating whet her the temperature values in the
temperature input file (specified by'tempf ile' represent global
temperature change ('temptype= "global"') or temperature
change for the contiguous U.S. ftemptype="conus"') in degrees
Celsius. Bydefault, the model assumestemperaturesareCONUS
temperatures (i.e., 'default="conus"').
popform
A characterstring indicating whetherthe populations in the population
input file specified by 'popfile' are spread across multiple columns
(i.e.,'popf orm="spread"') or are combined in a single column (i.e.,
'popf orm="gather "'). For both formats ('popf orm="spread"'
Expert and Public Review Draft-Do Not Cite
Page C-7
-------
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
or'popform="gather"'), the first column contains values for the
associated value year. If'popform="spread"' (default), the second
through eighth columns of' popfile ' must contain population values
for each NCA region, with the associated NCA region asthe column
name. If'popform="gather"', the second column must contain NCA
region names and the third column must contain values for the
associated region population.
Details
This function enables usersto import data on custom scenarios for use with temperature binning. Users
specify path names to CSVfiles containing temperature, global mean sea level rise (GMSL), population,
and gross domestic product (GDP) scenarios ftempfile', 'slrfile', 'gdpfile', and'popfile',
respectively). 'import_inputs ()' reads in and format any specified files as data frames and returns
a list of dataframes for imported scenarios. Users can specify whether the temperature input is for the
contiguous U.S. (CONUS) or global using 'temptype' and specify the format of the population scenario
using 'popform'. Users can specify whetherthe temperature input is for the contiguous U.S. (CONUS)
or global using 'temptype' and specify the format of the population scenario using'popform'.
Values for input scenarios must be within reasonable ranges. Temperatures must be in degrees Celsius
and values must be greaterthanor equal to zero and less than or equal to 10 degreesof warming.
Values for GMSL must be in centimeters(cm) and values must be greaterthanor equal to zeroand less
than or equal to 150 cm. Population and GDP values must be greaterthanor equal to zero. Ifa user
inputs a custom scenario with values outside the allowable ranges, 'import_inputs ()' will not
import that scenario and will instead stop and return an error message.
'import_inputs ()' drops missing values. The temperature-binning function 'tempBinO' linearly
interpolates missing values between available data points. For more information, see 'tempBin () '.
Ifthe temperature type is specified asglobal ftemptype="global"'), 'import_inputs ()'
converts input global temperatures in degrees Celsius fromthe 1986-2005 baseline ftemp_global',
below) to CONUStemperatures in degrees Celsius from the same baseline ftemp_conus', below). For
more information, see 'convertTemps () '.
Ifthe population input is spread across multiple columns (i.e., 'popform="spread"'), columns must
be named according to the NCA regions. Ifthe population input is in the gathered format, the region
value must be in the second column. The NCA region names for population inputs must be in the
following charactervector: 'c ( "Midwest", "Northeast", "Northern. Plains",
"Northwest", "Southeast", "Southern. Plains", "Southwest")'. All regions must
be present in the population input file.
'import_inputs ()' outputs a list of dataframesthat can be passedtothe main temperature-and
SLR-binning function 'tempBin ()' usingthe 'inputList' argument. For example, specify
'tempBin (inputsList=x)' to generate impacts for a custom scenario'x' (where 'x' isa list of
dataframessuchasthat output from'import_inputs ()') . For more information and examples, see
'tempBin () \ All inputs to 'import_inputs ()' are optional. Ifthe userdoes not specify a file path
for 'tempf ile\ 'slrfile', 'gdpfile', or'popfile' (or if there is an error reading in inputs from
those file paths), 'import_inputs ()' outputs a list with a 'NULL' value for the associated list
Page B-8
-------
2451 element. When the resulting list is passedasan argumenttothe temperature-and SLR-binning,
2452 'tempBin ()' defaults backtothe default scenarios for any list elements that are 'NULL' or missing. In
2453 other words, running 'tempBin (inputsList=list () )' returns the same outputs as running
2454 'tempBin ()'.
2455 Outputs
2456 'import_inputs()' returnsa listof named elements containing dataframes with custom scenarios for
2457 temperature, GMSL, GDP, and regional population, respectively:
tempinput
Dataframe containing a custom temperature scenario imported from the
CSV file specified by 'tempfile', with missing values removed,
'templnput' has twocolumns with names 'c ( "year",
"temp C")' containing the year and CONUStemperatures in degrees
Celsius, respectively.
slrlnput
Dataframe containing a custom GMSL scenario imported fromthe CSV
file specified by 'slrfile', with missing values removed, 'slrlnput'
hastwocolumns with names'c ("year", "sir cm")' containing
the yearand global mean sea level rise (GMSL) in centimeters,
respectively.
gdplnput
Dataframe containing a custom GDP scenario imported from the CSV file
specified by'gdpfile', with missing values removed.'gdplnput' has
twocolumns with names'c ("year", "gdp usd")' containing the
year and the U.S. national GDP in 2015$, respectively.
poplnput
Dataframe containing a custom temperature scenario imported from the
CSV file specified by 'popfile', with missing values removed,
'poplnput' has and three columns with names'c ("year",
"region", "reg pop")' containing the year, the NCAregion
name, and the NCA region population, respectively.
2458
2459 Examples
2460 ### Path to example scenarios
2461 scenariosPath <- system.file(package="ciraTempBin") %>%
2462 paste("extdata/scenariossep="/")
2463 ### View example scenario names
2464 scenariosPath %>% list.files
2465 ### Temperature Scenario File Name
2466 tempInputFile <- scenariosPath %>% paste("temp_scenariol.csv", sep="/")
2467 ### SLR Scenario File Name
2468 slrlnputFile <- scenariosPath %>% paste("slr_20Ocm.csv", sep="/")
2469 ### Population Scenario File Name
2470 popInputFile <- scenariosPath %>% paste("pop_iclusV5.csv", sep="/")
2471 ### Import inputs
2472 example_inputsList <- import_inputs(
2473 tempfile = tempInputFile,
2474 slrfile = slrlnputFile,
2475 popfile = popInputFile
2476 )
2477
2478 Asnapshot of the dataframes comprisingthe output list from the example above
2479 fexample_inputsList') is shown below:
2480
Expert and Public Review Draft-Do Not Cite
Page C-9
-------
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
example_inputsList$tempInput %>% head
##
year
temp_C
##
1
2010
0.7706176
##
2
2015
0.8879051
##
3
2020
1.0331185
##
4
2025
1.1999864
##
5
2030
1.3806500
##
6
2035
1.5674594
example_inputsList$slrInput %>% head
##
year slr_
_cm
##
1
1970
0
##
2
1980
2
##
3
1990
5
##
4
2000
8
##
5
2010
13
##
6
2020
19
example_inputsList$popIriput %>% head
##
year
region
reg_pop
##
1
2010
Northern.Plains
4956777
##
2
2020
Northern.Plains
5348266
##
3
2030
Northern.Plains
5879482
##
4
2040
Northern.Plains
6458399
##
5
2050
Northern.Plains
7090174
##
6
2060
Northern.Plains
7860014
B.2.1.4get_plots
Create and save plots for summarized temperature binning outputs.
Description
This function creates plots for the summarizedtemperature binning outputs. Results must be summed
across impact types. ~get_plots ()' returns a list with heatmapsfor model types present in the data
(GCMsand SLR scenarios) and annual results for all sectors and adaptations.
Usage
get plots(data, column="annual impacts", undiscounted=TRUE,
directory=NULL, save=FALSE, dpi=150, plotTypes="all")
Arguments
data
Dataframe of summarized outputs produced by 'tempBin () \ Do not
changecolumn namesof the'tempBin ()' output before running
'get plots ()\
column
A characterstring indicatingthe name of the column in the data for
which to create plots.
Page B-10
-------
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
undis counted
A 'TRUE/FALSE' value indicating whetherthe values in column
represent undiscounted values or discounted values (i.e., present
values). Defaults to 'undiscounted=TRUE'.
directory
A characterstring indicatingthe location of a directory in which tosave
the report objects. No default (i.e., 'directory=NULL').
save
A 'TRUE/FALSE' value indicating whetherto save results. Ifa directory
value is supplied (i.e.,'! is . null (directory)'), defaults to
'save=TRUE'. Otherwise, default is 'save=FALSE'.
dpi
Image resolution for saved PNG images. Users can control output
resolution, by supplying a numeric value up to 300 to 'dpi' (defaults to
'dpi=150').
plotTypes
Characterstring or character vector indicating which types of plots to
produce. Options are'c ( "heatmaps", "annual", "all")'. Set
'plotTypes="all"' (default) to produce both types of plots.
Details
This function processesthe results from temperature binning afterthe results have been summarized
for impact year estimates and impact types (e.g., as the summarized results from'tempBin ()' or
'aggregate_impacts ()').
By default,'get_plots ()' plots results from the'"annual_impacts"' column. Alternatively,
users can specify a column name in the data with 'column' (defaults to fcolumn=NULL').
The argument 'undiscounted' is used by 'get_plots ()' for plot labels and in file and directory
names for saving results.
'get_plots ()' produces separate heatmapsfortemperature-binned sectors (binned by GCM
models) and SLR-binned sectors (binned by end-of-century SLR scenarios). Colors in the heatmapare
determined from the underlying data. 'get_plots ()' also plots the average value and range of
results as a time series for each sector-adaptation-region combination, (average = mean, range =
minimum and maximum, acrosstheGCMs or SLR scenarios associated withthe sector).
If 'save=TRUE' and the usersupplies a path to a directory (i.e.,'! is . null (directory)'),
'get_plots ()' will try to save results to the specified directory. Separate directories are created
within the specified directory for heatmapsand annual results.
Users can setthe resolution of saved PNG images (in dots per inch, or "dpi"), by supplying a numeric
value up to 300 to 'dpi' (defaults to 'dpi=150').
Users can specify which plot types to produce by setting'plotTypes'. Set 'plotTypes="all"'
(default) to produce both heat maps and annual results) or specify a single type
fplotTypes="heatmaps"' and 'plotTypes="annual"', respectively).
Outp uts
List of heatmaps with list elements for all unique values for model types
present in data (i.e., "gem" for results calculated with temperature as
the driver value and "sir" for results calculated with GMSL as the driver
value.
Expert and Public Review Draft-Do Not Cite
PageC-11
-------
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
annual
List of lists of annual results. List of annual plots contains a list of sectors.
Each sector contains a list of adaptations for that sector. Each sector-
adaptation combination contains a list of nested regional plots.
Examples
### Create temperature binning scenario
df_tempExOut <- tempBin(aggLevels="none", pv=TRUE, silent=TRUE)
### Aggregate temperature binning summary across multiple columns
agg_tempExOut <- df_tempExOut %>%
aggregate_impacts(columns=c("annual_impacts", "discounted_impacts"))
### Create list of plots for aggregated results
agg_plotList <- agg_tempExOut %>% get_plots()
B.2.2 Helper Functions
.2.2.5 con vertTemps
Convert contiguous U.S. (CONUS) temperaturesto global temperatures or vice versa.
Usage
convertTemps(temps, from=c ("conus", "global'
temps
A numeric vector of CONUS or global temperatures in degrees Celsius.
from=c("conus",
"global")
A characterstring (one of 'c ("conus", "global")'), indicating
whether users a re converting from CON US to global temperatures
f f rom="conus"') or from global to CONUS('from="global"').
D es cri p ti on/ Detai Is
This pre-processing helper function converts a list of warming temperatures in degrees Celsius
('temps') from global to CONUS ff rom="global"') or vice versa ff rom="global"'). The
equations for converting between CONUS and global temperatures and back again are described
elsewhere in this report.
Outp uts
Outputs a numeric vector of temperatures in degrees Celsius.
Examples
### Create a numeric vector of CONUS integer temperatures from
### 1 to 7 degrees of warming (degrees Celsius)
conusTemps <- seq(l:7)
### Convert from CONUS temperatures to global temperatures
globalTemps <- conusTemps %>% convertTemps(from="conus")
B.2.2.6temps2slr
Convert global temperature change to global mean sea level rise (GMSL) in centimeters
Arguments
temps
A numeric vector of global temperatures in degrees Celsius.
Page ES-12
-------
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
D es cri p ti on/ Detai Is
This pre-processing helper function converts a vector of global temperatures to global mean sea level
rise (GMSL). The equation to convert global temperatures in degrees Celsius to global mean sea level
rise in centimeters is described elsewhere in this report.
Outp uts
Outputs a numeric vector of GMSL in centimeters.
Examples
temps2sir (1:7)
B.2.2.7 aggregatejmpacts
Summarize and aggregate impacts from temperature binning (calculate national totals, average across models,
sum impact types, and interpolate between impact estimate years).
Usage
aggregate impacts(data, columns= c("annual impacts"),
aggLevels=c("all"))
Arguments
data
Datafra me of results from temperature binning
columns
Charactervector of columns for which to aggregate results (defaultsto
columns=c("annual impacts"))
aggLevels
Levels of aggregation at whichto summarize data: one or more of
'c("national", "modelAverage", "impactYear",
"impactType", "all", "none")'. Defaultstoall levels (i.e.,
'aggLevels="all"').
D es cri p ti on/ Detai Is
This post-processing helper function aggregates and summarizes temperature binning results to levels of
aggregation specified by the user(passedto 'aggLevels'). Users can specifya single aggregation
level or multiple aggregation levels by passinga singlecharacterstringorcharactervectorto
'aggLevels'. Options for aggregation include calculating national totals faggLevels=
"national"'), averaging across model types and models faggLevels="model Aver age"'), summing
over all impact types faggLevels=" impactType"'), and interpolate between impact yearestimates
faggLevels=" impact Year"'). Users can specifya II aggregation levels at once by specifying
'aggLevel s="all"' (default) or no aggregation levels ('aggLevel s="none"l
Before aggregatingimpactsfornationaltotalsand/or model averages,'aggregate_impacts ()' will
drop any pre-summarized results (i.e., values for which'region="National Total"' and/or for
which 'model="average"', respectively) that are already present in the data 'and then reaggregate
at those levels.
For each of the'aggLevels', 'aggregate_impacts ()' performs the following summarization:
• '"national"': Annual values are summed across all regions present in the data. I.e., data is
grouped by columns'"sector", "adaptation", "model_type", "model",
"year"' (and '"impactType"' and/or'"impactYear"' if they are present in the data)
and summed across regions. Years which have missing column data for all regions return as 'NA'
(missing). The rows of the datafra me of national values (with column 'region="National
Total"' arethen bound tothe regional values.
Expert and Public Review Draft-Do Not Cite
PageC-13
-------
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
• '"modelAverage"': Annual results are averaged across all models present in the data. I.e.,
data is grouped by columns '"sector", "adaptation", "model_type",
"region", "year"'(and'"impactType"' and/or'"impactYear"' if they are present
in the data) and averaged across models. Averages exclude missing values. Years which have
missing column data for all models return as 'NA' (missing). The rows of the dataframeof model
averages (with column ,model="Average"' arethen bound tothe rows with individual
model results.
• '"impactType"': Annual results are summed across all impact types by sector present in the
data. I.e., data isgrouped by columns '"sector", "adaptation", "model_type",
"model", "region", "year"' (and '"impactYear"' if present in the data) and
summed across impact types. Mutates column 'impactType="all"' for all values. Years
which have missing column data for all impact types return as 'NA' (missing). If results are
aggregated across impact types, information about physical impacts (dataframe columns
'"physicalmeasure"' and'"physical_impacts"') are dropped.
• '"impactYear"': Annual results for sectors with only one impact year estimate (i.e.,
'impactYear=="N/A"') a re separated from those with multiple impact yearestimates. For
sectors with multiple impact years (i.e. 2010 and 2090 socioeconomic runs), annual results are
interpolated between impact yearestimatesforapplicable sectors i.e., data isgrouped by
columns'"sector", "adaptation", "model type", "model", "region",
"year"' (and '"impactType"' if present in the data) and interpolated across years withthe
2010 run assigned to year 2010 and the 2090 run assigned to year 2090. The interpolated values
are bound backtothe results for sectors witha single impact yearestimate, and column
'impactYear' setto'impactYear="interpolate"' for all values.
Note that 'aggregate_impacts ()' drops columns not used in grouping or aggregation.
Outp uts
Outputs a dataframeof results from temperature binning, summarized at the specified aggregation
levels.
Examples
### Create temperature binning scenario
df_tempExOut <- tempBin(aggLevels="none", pv=TRUE, silent=TRUE)
### Aggregate temperature binning summary across multiple columns
agg_tempExOut <- df_tempExOut %>%
aggregate_impacts(columns=c("annual_impacts", "discounted_impacts"))
Page B-14
-------
2638 APPENDIX C I CLIMATE IMPACT ANALYSIS: ADDITIONAL INFORMATION
2639 This appendix includes additional information on the data and underlying models used in the climate
2640 impact analyses presented in Section 3. Section C.1 provides more information on the climate scenarios
2641 and Section C.2 describes the socioeconomic assumptions.
2642 C.l Climate Models and Scenarios
2643 Description of Hector
2644 Hector is an open-source, object-oriented, reduced-form global carbon-cycle climate model (Dorheim et al.,
2645 2020; Hartin etal., 2015; Vega-Westhoff etal.,2019). Hector, like other reduced form climate models,
2646 calculates concentrations of greenhouse gases from a given emissions pathway while modeling the carbon
2647 cycle and other gas cycles. Global emissions of greenhouse gases (C02, CH4, N20, halocarbons) and aerosols
2648 (BC, OC, S02) are passed to Hector. Emissions are converted to concentrations where necessary, and are
2649 used to calculate radiative forcing and then a global mean temperature change along with other Earth
2650 system variables (Hartin et al., 2015).
2651 Hector has a three-part carbon cycle: atmosphere, land, and ocean. The atmosphere is treated as a single
2652 well-mixed box, where a change in atmospheric carbon is a function of anthropogenic fossil fuel and
2653 industrial emissions, land-use change emissions, and carbon fluxes between the atmosphere and ocean and
2654 the atmosphere and land. In Hector's default terrestrial carbon cycle, vegetation, detritus, and soil are
2655 linked to one another and to the atmosphere by first-order differential equations. Net primary production
2656 is a function of atmospheric C02 and temperature. Carbon flows from vegetation to the detritus and then
2657 down to soil, where some fraction is lost due to heterotrophic respiration.
2658 The surface ocean carbon flux is dependent upon the solubility of C02 within high and low latitude surface
2659 boxes which are calculated from an inorganic chemistry submodule (Hartin etal., 2016). Hector
2660 calculates pC02, pH, and carbonate saturations in the surface boxes; once carbon enters the surface boxes,
2661 it is circulated through the intermediate and deep ocean layers via water mass advection and exchanges,
2662 simulating a simple thermohaline circulation.
2663 Radiative forcing is calculated from each individual atmospheric constituent; C02, halocarbons, non-
2664 methane volatile organic carbons (NMVOCs), black carbon, organic carbon, sulfate aerosols, CH4, and N20,
2665 and forcing from tropospheric ozone and stratospheric water vapor. C02, CH4, N20, and halocarbons are
2666 converted to concentrations, while NMVOC, and aerosols are left as emissions (Hartin etal., 2015).
2667 Global atmospheric temperature is a function of a user-specified climate feedback parameter, which
2668 represents the equilibrium climate sensitivity for a doubling of C02 concentrations, total radiative forcing,
2669 and oceanic heat flux. Atmosphere-ocean heat exchange in Hector consists of a one-dimensional diffusive
2670 heat and energy-balance model, DOECLIM (Kriegler 2005; Vega-Westhoff et al., 2019). The code and
2671 detailed documentation can be found at
-------
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
Description of GCAM
The Global Change Analysis Model v5.3 (GCAM) is an open source model that represents the linkages
between energy, water, land, climate and economic systems (Calvin et a I., 2019). GCAM is a market
equilibrium model, global in scale and subdivided into 32 geopolitical energy and economic regions and 283
agriculture and land-use regions. GCAM is calibrated to a historical base year of 2010 and projects key
variables forward in timethrough 2100. GCAM is a partial equilibrium model, representing the supply,
demand and price for a variety of goods and services in the energy, agriculture and water sectors. A more
complete documentation of GCAM is available at
Prices of energy, agriculture, and forest products are adjusted until supply and demand are in equilibrium.
Asa dynamic-recursive model, decision-makers base decisions on present prices assuming they will remain
constant at those levels indefinitely as investment choices are made. GCAM computes emissions of 16
gases and short-lived species (C02, CH4, N20, F-gases, S02, BC, OC, NOx, CO, NMVOCs) from a variety of
human activities. These emissions are passed to a simple climate model, Hector, to calculate global mean
temperature among other climate variables.
Description of climate scenarios used in this case study
This technical documentation provides a case study to illustrate the function and capabilities of the
Temperature Binning Tool. The scenarios used in this case study are illustrative in nature and do not
represent any actual policies or programs at domestic or international levels.
The GCAM v5.3 reference used in this case study represents a scenario with no additional future climate
policies. When the emissions are run through Hector with an equilibrium climate sensitivity of 3°Cthe end
of century radiative forcing is 4.66 Wnr2 and global mean temperature is 2.69 °C relative to a 1986-2005
baseline. Population data are from the Shared Socioeconomic Pathway 2 (SSP2), "Middle-of-the-Road"
scenario, as developed by KC and Lutz (2017). Initial year GDP, labor productivity growth rates, and labor
force participation rates are derived to match external GDP data and forecasts from three sources: (1) the
U.S. Department of Agriculture (2015) for 1990, 2005, and 2010; (2) the International Monetary
Fund (2014) for 2011 to 2020; and (3) Delink et al. (2017), using SSP2, for 2021 through 2100 (Calvin et al.,
2019). Comparing the GCAM reference to EIA projections, GCAM fossil fuel and industrial C02 emissions are
up to 2 GtC higher in 2025.36
For the emission reduction scenario used in this case study, C02 emission reductions were linearly
interpolated between 2025 and 2100 based on the GCAM reference 2100 C02 emissions and run through
Hector v2.5. Reductions begin in the next time step after 2025 (i.e., 2026). Only fossil fuel and industry
emissions are reduced, not land use change emissions or other GHGs. The scenarios represent a reduction
36 World carbon dioxideemissions by region:
https://www. eia.gov/out I ooks/aeo/data/browser/? src=-fl #/?id=10-l E02019& region=0-
0&cases=Reference& st art=2010&end=2050&f=A& I inechart=~Reference-d080819.26-10-
IE02019&map=&ctype=linechart&sid=Reference-d080819.26-10-1 E02019&sourcekey=0
Page C-2
-------
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
in emissions by 90 percent in 2100. For all scenarios, equilibrium climate sensitivity was varied [1.5, 2.0, 3.0,
4.5, 6.0], with an ECS value of 3.0 C as the default Hector parameter.
FIGURE C-l. GCAM REFERENCE
Global mean radiative forcing (W m-2) and global mean temperature (°C) relative to a baseline of1986-2005for the GCAM
v5.3 reference scenario run with 6 different equilibrium climate sensitivities.
parameter_value
ECS_1.5
ECS_2
ECS_3
ECS_4.5
ECS_6
Global Mean Temperature
-2.5-
1900
1950
2000
2050
2100 1900
year
1950
2000
2050 2100
Radiative Forcing
Sea level rise components
Alternatives exist to the approach currently built into the Temperature Binning Tool, with two options
described here. Each option would build upon the core function within the Temperature Binning
Framework. Both methods are semi-empirical models of sea level change based on statistical relationships
between global sea level and global mean temperature or radiative forcing and provide an alternative to
process-based models to estimate future sea-level rise. The first is a semi-empirical method from Kopp et
al. 2016 that estimates global sea level change over the last 3000 years using a global database of regional
sea-level reconstructions. The semi-empirical model can be combined with temperature projections from
any scenarios to project future global sea level changes.
The second model is BRICK (Building blocks for Relevant Ice and Climate Knowledge), a modular semi-
empirical modeling framework where changes in global radiative forcing drive changes in global mean sea
level (Wong etal., 2017). BRICK simulates contributions to global sea level rise from the Greenland and
Antarctic ice sheets, thermal expansion, and glaciers and ice caps.
C.2 Socioeconomic Scenarios and Input Data
Expert and Public Review Draft-Do Not Cite
Page C-3
-------
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
2724 The example analyses in Section 3 use the default socioeconomic inputs stored in the Tool. Table C-l
2725 presents the Median Variant Projection of the United Nation's (UN) 2015 World Population Prospects
2726 dataset used to project future U.S. population changes between 2015-2100. This scenario is a mid-range
2727 population projection and allows for the reasonable incorporation of future population growth. Note that
2728 prior to 2014 historical data was acquired using U.S. Census data. Because the UN Median Population
2729 projection is only available at the national level, disaggregated population projections were produced at the
2730 county-level using EPA's ICLUSv2 model. The spatial pattern of population change in ICLUSv2 is dependent
2731 upon underlying assumptions regarding fertility, migration rate, and international immigration. These
2732 assumptions were parameterized using the storyline of SSP2, which suggests medium levels of fertility,
2733 mortality, and international immigration.
2734 TABLE C-l. NATIONAL AND REGIONAL POPULATION PROJECTIONS
2735 Default national (CONUS) and regional populations in the Tool, from ICLUSv2 andtheUN2015 World Population Prospects
2736 (Median Variant Projection).
Year
Popu
ation
CONUS
Midwest
Northeast
Northern
Plains
Northwest
Southeast
Southern
Plains
Southwest
2010
306,675,006
60,760,771
64,443,443
4,866,153
12,123,196
76,532,629
31,750,030
56,198,784
2015
319,615,728
62,531,700
66,262,356
5,143,618
12,500,441
79,584,504
33,914,631
59,678,477
2020
331,308,682
64,097,103
67,903,699
5,394,605
12,816,323
82,228,514
35,921,126
62,947,312
2025
342,770,319
65,655,846
69,515,267
5,640,953
13,124,230
84,797,999
37,898,560
66,137,464
2030
353,379,113
66,987,745
70,975,809
5,863,188
13,399,008
87,135,477
39,793,579
69,224,308
2035
362,816,645
67,985,079
72,264,293
6,050,737
13,622,298
89,177,634
41,541,842
72,174,762
2040
371,260,069
68,741,241
73,411,923
6,214,809
13,805,065
90,996,054
43,133,591
74,957,385
2045
378,916,029
69,343,598
74,455,788
6,361,550
13,960,224
92,638,619
44,583,572
77,572,679
2050
386,256,917
69,918,188
75,470,921
6,502,332
14,107,875
94,235,017
45,957,244
80,065,339
2055
393,468,477
70,505,868
76,489,049
6,641,433
14,254,356
95,838,507
47,283,831
82,455,432
2060
400,797,674
71,141,395
77,560,889
6,782,559
14,408,702
97,500,055
48,594,359
84,809,714
2065
408,139,446
71,800,214
78,677,245
6,922,707
14,564,856
99,183,481
49,871,975
87,118,968
2070
415,211,012
72,429,615
79,781,783
7,056,765
14,711,575
100,825,370
51,081,555
89,324,349
2075
421,784,560
72,988,264
80,822,953
7,180,349
14,842,861
102,363,714
52,200,684
91,385,736
2080
427,739,912
73,455,251
81,750,462
7,292,103
14,955,388
103,773,739
53,230,863
93,282,106
2085
433,153,810
73,846,353
82,580,397
7,394,173
15,053,044
105,068,620
54,183,348
95,027,873
2090
438,195,481
74,601,088
83,501,161
7,482,239
15,210,659
106,289,840
54,868,179
96,242,314
2737
2738 Using the UN Median population projection for the U.S. (United Nations 2015), the CIRA project team
2739 generated a projection of GDP growth with the EPPA, version 6 model (Chen et al., 2015), which is the
2740 default GDPscenario, presented in Table C-2. The team projected GDP by harmonizing EPPA-6 with the
2741 2016 Annual Energy Outlook reference case (USEIA 2016) for the U.S. through 2040. For periods after 2040,
2742 EPPA-6 and its baseline assumptions for other world regions and time periods is used. The projection
2743 reflects historical growth rates until about 2015, and projections thereafter. The annual rate of GDP growth
Page C-4
-------
2744 is just above 2 percent from 2010 to 2015 and begins the projection period at an annual rate of about 2.5
2745 percent. Over the remainder of the simulation, the rate of GDP growth steadily falls from about 2.5 percent
2746 in 2015 to 1.30 percent in 2099. Note, the impacts of climate change on economic activity (e.g., losses to
2747 labor supply or increased capital expenditures for adaptation) are not accounted for in this projection. As
2748 such, the usage of this scenario in impacts analysis may be overestimating GDPgrowth when considering
2749 the multi-sector damages. Additionally, the use of a single national-scale economic growth projection that
2750 omits region-specific socioeconomic changes may lead to inaccuracies on the regional level when national
2751 GDP is used as an indicator of regional economic activity.
2752 TABLE C-2. U.S. GDP PROJECTION
2753 Default national GDP ($2015) developedfortheCIRA project, based on a harmonization ofEPPA, version 6 model and the
2754 2016 Annual Energy Outlook reference case.
Year
GDP ($2015)
2010
$16,693,671,681,957
2015
$18,459,285,695,743
2020
$20,956,428,357,133
2025
$23,502,589,873,850
2030
$26,191,153,348,361
2035
$28,914,759,604,411
2040
$32,112,407,174,002
2045
$35,194,683,105,093
2050
$38,477,618,934,387
2055
$41,964,197,003,199
2060
$45,654,537,276,604
2065
$49,547,386,879,164
2070
$53,640,268,231,428
2075
$57,928,656,611,945
2080
$62,405,475,815,739
2085
$67,052,775,580,200
2090
$71,877,254,264,111
2755
Expert and Public Review Draft-Do Not Cite
Page C-5
-------
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
References
Calvin, K., Patel, P., Clarke, L., Asrar, G., Bond-Lamberty, B.,Cui, R.Y., Di Vittorio, A., Dorheim, K., Edmonds,
J., Hartin, C., Hejazi, M., Horowitz, R., Iyer, G., Kyle, P., Kim, S., Link, R., McJeon, H., Smith, S.J., Snyder,
A., Waldhoff, S., and Wise, M. 2019. "GCAM v5.1: Representing the Linkages between Energy, Water,
Land, Climate, and Economic Systems." Geoscientific Model Development 12(2):677-98. doi:
10.5194/gmd-12-677-2019.
Chen, Y.-H. H.,S. Paltsev, J. Reilly, J. Morris, and M. Babiker. (2015). The MIT EPPA6 Model: Economic
Growth, Energy Use, and Food Consumption. MIT Joint Program on the Science and Policy of Global
Change, Report 278, Cambridge, MA. Available online at
http://globalchange.mit.edu/re search/publications/2892
Dorheim, Kalyn, Robert Link, Corinne Hartin, Ben Kravitz, and Abigail Snyder. 2020. "Calibrating Simple
Climate Models to Individual Earth System Models: Lessons Learned From Calibrating Hector." Earth
and Space Science 7(ll):e2019EA000980. doi: https://doi.org/10.1029/2019EA000980.
EPA. 2017. Updates tothe Demographic and Spatial Allocation Models to Produce Integrated Climate and
Land Use Scenarios (Iclus) (Version 2). U.S. Environmental Protection Agency, Washington, DC.
EPA/600/R-16/366F. Available online at:
htt ps://cfpub. epa.gov/ncea/icl us/recordisplay. cfm?deid=322479
Hartin, C. A., P. Patel, A. Schwarber, R. P. Link, and B. P. Bond-Lamberty. 2015. "A Simple Object-Oriented
and Open-Source Model for Scientific and Policy Analyses of the Global Climate System - Hector vl.0."
Geoscientific Model Development 8(4):939-55. doi: https://doi.org/10.5194/gmd-8-939-2015.
Hartin, Corinne A., Benjamin Bond-Lamberty, Pralit Patel, and Anupriya Mundra. 2016. "Ocean Acidification
over the next Three Centuries Using a Simple Global Climate Carbon-Cycle Model: Projections and
Sensitivities." Biogeosciences 13(15):4329-42. doi: 10.5194/bg-13-4329-2016.
Kopp, Robert E., Andrew C. Kemp, Klaus Bittermann, Benjamin P. Horton, Jeffrey P. Donnelly, W. Roland
Gehrels, Carling C. Hay, Jerry X. Mitrovica, Eric D. Morrow, and Stefan Rahmstorf. 2016. "Temperature-
Driven Global Sea-Level Variability in the Common Era." Proceedings of the National Academy of
Sciences 113(11):E1434-41. doi: 10.1073/pnas. 1517056113.
Kriegler, Elmar. 2005. "Imprecise Probability Analysis for Integrated Assessment of Climate Change." 258.
United Nations. 2015. World Population Prospects: The 2015 Revision. United Nations, Department of
Economic and Social Affairs, Population Division.
U.S. Energy Information Administration. 2016. Annual Energy Outlook.
Page C-6
-------
2787 Vega-Westhoff, Ben, Ryan L. Sriver, Corinne A. Hartin, Tony E. Wong, and Klaus Keller. 2019. "Impacts of
2788 Observational Constraints Related to Sea Level on Estimates of Climate Sensitivity." Earth's Future
2789 7(6):677—90. doi: https://doi.org/10.1029/2018EF001082.
2790 Wong, Tony E., Alexander M. R. Bakker, Kelsey Ruckert, Patrick Applegate, Aimee B. A. Slangen, and Klaus
2791 Keller. 2017. "BRICK vO.2, a Simple, Accessible, and Transparent Model Framework for Climate and
2792 Regional Sea-Level Projections." Geoscientific Model Development 10(7):2741-60. doi: 10.5194/gmd-
2793 10-2741-2017.
2794
Expert and Public Review Draft-Do Not Cite
Page C-7
-------
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
APPENDIX D | INFORMATION QUALITY
Ensuring Information Quality
The technical documentation, Framework, Tool, and underlying analyses were conducted in accordance
with EPA's Guidelines for Ensuring and Maximizing the Quality, Objectivity, Utility, and Integrity of
Information Disseminated by the Environmental Protection Agency, which follows Office of Management
and Budget (OMB) guidelines and implements the Information Quality Act (IQA) (Section 515 of Public Law
106-554). The following section in this Appendix describes the independent peer review that was
performed on the technical documentation materials.
In accordance with OMB definitions, EPA defines the basic standard of information "quality" by the
attributes objectivity, integrity, utility, and transparency. For products meeting a higher standard of quality,
like this product, the Agency requires an appropriate level of transparency regarding data and methods in
order to facilitate the reproducibility of information by qualified third parties. The EPA uses various
established Agency processes (e.g., the Quality System, peer review requirements and processes) to ensure
the appropriate level of objectivity, utility, integrity, and transparency for its products is based on the
intended use of the information and the resources available.
Objectivity focuses on whether the disseminated information is being presented in an accurate, clear,
complete, and unbiased manner, and as a matter of substance, is accurate, reliable, and unbiased. The
technical documentation meets the standard for objectivity, due to activities described in the following:
a) The information disseminated was determined to be complete, accurate, and reliable based on
internal quality control measures adopted by the expert modeling teams. This included quality
checks throughout the chain of analytic steps, including developing and processing climate
projections, calibrating and validating the sectoral impact models, and checking data to ensure that
no errors occurred in the process to compile and summarize results.
b) The information disseminated was determined to be clear, complete, and unbiased based on
multiple rounds of internal review. Consistent with guidelines described in EPA's Peer Review
Handbook, the underlying sectoral modeling methodologies were peer-reviewed through scientific
journal publication processes. In addition, the Temperature Binning Framework was subject to an
external journal publication process. Citations for these publications can be found throughout the
main sector sections of the technical documentation and its appendices.
The technical documentation in full is being subject to a public comment period to ensure that the
information summarized by EPA was technically supported, competently performed, properly
documented, consistent with established quality criteria, and communicated clearly. This public
review period is also intended to provide feedback and comments on the utility of the Framework.
-------
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
Technical Documentation ForTheTemperature Binning Framework
Expert and Public Review Draft
The technical documentation in full is being subject to an independent, external peer review to
ensure that the information summarized by EPA was technically supported, competently
performed, properly documented, consistent with established quality criteria, and communicated
clearly.
Integrity refers to security of information, such as the protection of information from unauthorized access
or revision, to ensure that the information is not compromised through corruption or falsification. The
technical documentation, Framework, Tool, and underlying analyses meet the standard for integrity due to
the strategic steps taken to ensure that the data and information remained secure. These steps included
the use of password protected data storage repositories, password protected data transfer technology, and
multiple layers of data validation checks to ensure that the integrity was not compromised.
Utility is the usefulness of the information to the intended users. The technical documentation,
Framework, Tool, and underlying analyses meet the standard for utility because the information
disseminated provides insights (technical methods for quantifying physical and economic impacts)
regarding the potential magnitude of the impacts of climate change. Understanding the risks posed by
climate change can inform broader assessment reports and policy decisions designed to address these risks.
Transparency ensures access to and description of (1) the source of the data, (2) the various assumptions
employed, (3) the analytic methods applied, and (4) the statistical procedures used. The report and its
underlying analyses meet the standard for transparency for the following reasons:
a) The underlying datasets, sectoral impact models, and the methods supporting the Temperature
Binning Framework have been published with open access in the peer-reviewed scientific literature,
and are cited throughout the report. These papers, along with their online supplementary
materials, provide detailed information on the sources of data used, assumptions employed, the
analytic and statistical methods applied, and important limitations regarding the approaches and/or
how the results should be interpreted.
b) Appendix A for this Technical Documentation provides details on how results and output from each
sectoral impact model (or impacts study) are formatted and adapted for usage in the Framework
and Tool. This Appendix contains descriptions of the methodologies used in estimating impacts,
assumptions used, and citations to the underlying literature where the reader can go for more
information.
c) The technical documentation in full is being subject to a public comment period to ensure the
interested stakeholders had an opportunity to review and provide input on the methods of the
Framework and Tool.
d) All input and output data associated with the illustrative analyses of this Technical Documentation
have been posted on the following website. See I;
framework
e) The R package for the Temperature Binning Tool has been posted on the following website. See
https://www.epa.gov/cira/temperature-binning-framework
Page D-2
------- |