Executive Summary Improving the Assessment and Valuation of Climate Change Impacts for Policy and Regulatory Analysis Modeling Climate Change Impacts and Associated Economic Damages and Research on Climate Change Impacts and Associated Economic Damages June 2011 Workshop Report

Improving the Assessment and Valuation o
Climate Change Impacts for Policy
and Regulatory Analysis j

Sponsored by

r-nA United Slates
Environmontai
^ W L—l Protection Agency

^0^ 0 *. DEPARTMENT Or

mENERGY





Executive Summary:

Improving the Assessment and Valuation of
Climate Change Impacts for Policy and
Regulatory Analysis

Modeling Climate Change Impacts and Associated Economic Damages

and

Research on Climate Change Impacts and Associated Economic Damages

June 2011

Workshop Sponsored by:
U.S. Environmental Protection Agency
U.S. Department of Energy

Workshop Report Prepared by:
ICF International

-------
I. Introduction

In 2009 and early 2010, the U.S. Environmental Protection Agency (EPA) and U.S. Department of Energy
(DOE) joined other U.S. government agencies in conducting an analysis of the social cost of carbon (SCC).
The interagency working group used the DICE, FUND, and PAGE integrated assessment models (1AM) to
estimate a range of values for the SCC from 2010 to 2050 for use in U.S. government regulatory impact
analyses. The U.S. government analysis concluded in February 2010 and the estimated SCC values were
first used in March 2010 in the analysis of DOE's Energy Conservation Standard for Small Electric Motors.
In preparation for future revisions to the U.S. government SCC analysis, EPA and DOE seek to improve
the understanding of the natural scientific and economic impacts of climate change. This enhanced
understanding is also intended to inform ongoing work of the U.S. government to improve regulatory
assessment and policy analysis related to climate change.

To further these objectives, the EPA National Center for Environmental Economics and Climate Change
Division and the DOE Office of Climate Change Policy and Technology sponsored a pair of invitational
workshops on November 18-19, 2010 and January 27-28, 2011. The November workshop focused on
conceptual and methodological issues related to modeling and valuing climate change impacts. It also
addressed the implications of these estimates for policy analysis. The January workshop reviewed recent
research on physical impacts and associated economic damages for nine impact categories (e.g., human
health, agriculture, sea level), with a particular focus on knowledge that might be used to improve lAMs.

This workshop summary was prepared by ICF International on behalf of EPA and DOE. It does not
represent the official position or views of the U.S. government or its agencies, including EPA and DOE,
nor has it been reviewed by the workshop speakers and other participants. The potential improvements
and key findings outlined below represent the perspectives of one or more participants, as expressed at
the workshops and summarized by the planning committee. However, these summaries do not
necessarily represent consensus views, since none was sought at these workshops. This Executive
Summary is organized into six sections: Physical Impacts Assessment; Valuation of Damages;
Representing Impacts and Damages in Models; Communication of Estimates; Research and
Collaboration; and Specific Impacts Sectors.

II. Physical Impacts Assessment

Participants made comments and suggestions related to impacts assessment, including the following:

• More fully incorporate uncertainty. Natural and social scientists should attempt to more fully
characterize the uncertainty in impacts assessments, including parametric, stochastic, and
structural uncertainty at all stages in the modeling process. Many of the current 1AM inputs and
parameters represent too narrow a range of possibilities. Complex and non-linear processes at
the high ends of the impacts probability distribution (i.e., "fat tails") should be better
characterized.

-------
• Consider both top-down and bottom-up approaches. Estimates from both top-down and
bottom-up approaches can help to estimate and bound the range of climate change impacts.
For bottom-up approaches, the appropriate scale and detail may be different for each sector.

• Incorporate threshold effects of physical and biological impacts. Mechanistic and process
models relying on basic principles (e.g., conservations of energy, plant biophysiology, ocean
biogeochemistry) should be used, when possible, to extrapolate responses to new conditions,
since statistical methods may not capture non-linear threshold effects of unprecedented levels
of change. When climate change impacts are expected to be within or close to the range of past
variations, statistical models are appropriate.

• Capture climate variables beyond global mean temperature. A better characterization of
multiple climate variables (e.g., precipitation, storms, seasonal and diurnal temperature
variations, rate of temperature change) and threshold effects on a geographically disaggregated
scale could improve model calibration and the accuracy of local damage projections.

• Focus research efforts on sectors that could have the largest influence on overall damage
estimates. This will include research on impact categories that could comprise a large share of
total damages but where relatively little information has been collected to date. Researchers
should not simply focus on issues that are easiest to approach. Research priorities should be
guided by the combination of potential consequences and uncertainty, not one or the other
alone.

• Increase focus on high-impact events, multi-century impacts. Existing studies tend to examine
the means of the impacts probability distribution, neglecting the low-probability, high impact
tails of the distribution, which can have a significant influence on 1AM results. Impact studies
should address this gap, recognize the potential for unexpected and unpredictable events, and
attempt to model very long-term impacts (e.g., beyond 2100), despite great associated
uncertainty. To do this, modelers should develop more complete multi-century projections for
socio-economic and climate inputs including estimates of socio-economic uncertainty.

• Rigorously test, compare, and evaluate impact models. Model intercomparison projects have
helped to improve physical climate models and could be used to improve impact models.

III. Valuation of Damages

Comments and suggestions related to damage valuation included the following:

• Consider alternate functional forms for damage functions. Representation of damages could be
improved by: evaluating the additive or multiplicative nature of impacts; better incorporating
discontinuities; better capturing natural capital and its interactions with physical and social
capital; and generally considering a broader set of functional forms. Alternate forms are

-------
particularly important given the challenges in extrapolating damage functions calibrated at 2-
3°C warming to considerably higher global mean temperature increases.

• Clearly incorporate human behavioral responses. Adaptation and technological development
should be more fully incorporated in estimates of climate change impacts, and the underlying
assumptions associated with those factors should be clearly articulated.

• Consider different ways of equity weighting when conducting social welfare analysis of
climate policies. Several workshop participants suggested considering different ways of
incorporating equity weights into the SCC or lAMs more generally. For example, most lAMs use
a utility function with a single parameter that controls preferences regarding intra-generational
equity, inter-generational equity, and risk aversion. Future research should explore alternative
functional forms that allow these effects to be disentangled.

• Fully account for non-market impacts and non-use values. This includes improving estimates of
impacts currently included in some models (e.g., health impacts) and incorporating impacts
currently missing from most models (e.g., ocean acidification, loss of cultural heritage). Revealed
and stated preference estimates and benefit-transfer methods should be improved and
estimated jointly to mitigate problems with each.

• Consider "outer measures" of climate damages. Developing a model for a highly simplified but
inclusive "outer" measure of climate change damages may help provide an upper bound on SCC
estimates. Current bottom-up models are "inner" measures that attempt to capture and sum
the individual components of climate damages. Since it is challenging to capture all of the
components and interactions between them, these models will tend toward underestimation.

IV. Representing Impacts and Damages in Models

Throughout both workshops, but especially during the first, participants made suggestions related to
integrating impacts and damages in models. These comments included the following:

• Improve both aggregated and disaggregated models while utilizing the strengths of each.

There are important roles for models across the spectrum of aggregation, as more or less
aggregation may be appropriate for different applications. Model type and analysis time scale
should be matched to analytical objective. Since aggregation can contribute to a bias in impact
estimates, some models should be less aggregated spatially, temporally, and sectorally to more
realistically represent impact mechanisms. Since disaggregated models can incorporate more
realistic impact mechanisms and use empirical data to estimate model parameters, they can be
used to calibrate components of more comprehensive aggregated models.

• Incorporate more sectors. lAMs should include a broader range of sectors. For example, no
lAMs currently represent ocean acidification.

-------
• Incorporate interactions between sectors. Interactions between sectors (and among climate
and non-climate stressors) may be synergistic or antagonistic, additive, multiplicative, or
subtractive, making cumulative impacts larger or smaller than the sum of the individual impacts.
Double-counting should be avoided.

• Use consistent scenarios. Consistent socio-economic and climate scenarios should be used in
impact and damage assessment to facilitate inter-comparison, integration, and combination of
estimates.

• Increase model flexibility to facilitate improvements. lAMs should be (re)designed to facilitate
updates to models or model components as new research develops. A more flexible or modular
structure would allow components to be individually updated or replaced.

• Conduct new empirical studies and better incorporate existing research. lAMs need new
primary impacts research from which to draw. Research needs include empirical studies on:
physical impacts, monetization of damages, decision making under uncertainty, adaptation-
related technological change, adaptive capacity, tipping points, and impacts beyond 2100. lAMs
could also be improved by drawing more on the existing body of research.

V. Identify metrics for model validation. Metrics and methods of
validation are needed to assess models and model results.
Communication of Estimates

Participants, particularly at the first workshop, made comments and suggestions related to the

communication of impacts and damages estimates. These comments included the following:

• Increase transparency. lAMs should be made more accessible and transparent, including their
key assumptions, structural equations, parameter values, and underlying empirical studies.

• Fully and clearly communicate uncertainty. Communication should help decision makers and
the public fully and clearly understand uncertainty and its implications. The full range of model
outputs should be communicated and used, rather than focusing on one central value from a set
of model runs.

• Consider other metrics. Multiple criteria, in addition to the SCC and cost-benefit analysis, should
be used for climate-related regulatory analysis, including additional cost-effectiveness
measures.

VI. Research and Collaboration

Comments and suggestions related to research and collaboration included the following:

-------
• Increase collaboration and communication between natural scientists, economists, and
modelers. Collaboration and communication should be increased between all parties involved in
impacts assessment, damages valuation, and integrated assessment modeling. Impacts
assessment and valuation efforts should be coordinated with existing efforts such as the
National Climate Assessment and international impacts and valuation efforts. 1AM data sources,
damage functions, and outputs should be reviewed by relevant members of the Impacts,
Adaptation, and Vulnerability (IAV) and economic valuation communities to ensure that lAMs
reflect the current state of the primary literature for each of the impact categories.

• Increase capacity to address challenges. Additional funding and staff are needed to help
address existing impacts and damages assessment challenges.

VII. Specific Impacts Sectors

The second workshop focused on the current state of research in nine impact categories. This section

highlights key research findings and recommendations for future research for each of the categories.

Storms and Other Extreme Weather Events

• Fewer tropical storms are expected in the future, but average wind speeds and precipitation
totals are expected to increase. The intensity of the strongest storms is expected to increase.

• Estimates in the literature for increases in cyclone property damages due to climate change
range from 0.002 to 0.006% of global GDP. Increases in property damages from all extreme
events (including cyclones) due to climate change under an A1B scenario, according to one
study, range from $47-$100 billion (2008 dollars) per year, or 0.008-0.018% of GDP, by 2100.

• Fatalities may increase or decrease due to climate change impacts on extreme events, as deaths
from tropical cyclones may decrease more than deaths from other extreme events (e.g., heat
waves) increase. Tropical cyclones are expected to continue to be the dominant cause of
extreme event-related damages.

Water Resources

• Water demand, supply, and management should be modeled on a river basin scale to effectively
estimate climate change impacts.

• National estimates from the literature of climate change damages to water resources range
from $12-$60 billion (2009 dollars) per year for the United States according to analyses in a
range of studies.

• Coupling approaches that model changes using regional hydrologic models and those using
regional economic models could help bridge some gaps in water resources damage estimation.

-------
Human Health

• The majority of climate change health effects result from diarrhea, malnutrition, and malaria.
The World Health Organization estimates that the costs to treat climate change-related cases of
diarrhea, malnutrition, and malaria in 2030 would be $4 to $13 billion under a scenario in which
C02 is stabilized at 750ppm by 2210. The study predicts a 3%, 10%, and 5% increase in cases of
diarrhea, malnutrition, and malaria, respectively.

• Health impact valuation depends largely on mortality valuation, particularly in developing
countries and particularly among children. Adjusting the value of a statistical life for income is
critical for accurate valuation.

Agriculture

• Estimates in the literature project the global range of yield changes in the 2050s to be
approximately -30 to +20% under a 2.3°C mean global temperature increase (relative to 1961-
1990).

• Average global effects of climate change on agriculture are expected to be positive in the short
term and negative in the long term. The location of the inflection points is unknown.

o C02 fertilization from increasing C02 concentrations will benefit some plants (C3 plants)
more than others (C4 plants). Elevated C02 concentrations especially benefit weeds.

• Agriculture contributes only 2-3% of U.S. GDP, but the highly inelastic nature of agricultural
demand means that even a small reduction in agricultural production from climate change could
result in large price changes and large welfare losses.

• Adaptation and technological change can help to mitigate the impacts of climate change on
agriculture. A key challenge will be producing heat and drought tolerant plants with high yields.

Sea Level Rise

• Climate-induced sea level rise will be compounded by both natural and human-induced
subsidence in many densely-populated coastal areas.

• Emissions abatement may stabilize the rate and ultimate total amount (in 100s of years) of sea
level rise, but not reduce the current significant commitment to sea level rise.

• The valuation of sea level rise damages depends heavily on wetland values and adaptation.

Marine Ecosystems and Resources

• Increasing atmospheric C02 concentrations cause ocean C02 concentrations to increase,
decreasing ocean pH, and decreasing saturation states for calcite and aragonite, which are used
by marine animals to produce calcareous parts (e.g., shells).

• Damages from decreased mollusk harvest revenues due to a 0.1-0.2 ocean pH decrease are
estimated at $1.7 to $10 billion in net present (2007) value losses through 2060. Under the A1FI

-------
scenario pH decreases of 0.1 and 0.2 are expected by approximately 2040 and 2060,
respectively.

• Assessments using bio-climate envelopes, minimum realistic models, and ecosystem and food
web models would be beneficial to estimate marine impacts.

• A wide variety of studies to estimate damages is needed, using both revealed and stated
preferences, to estimate total economic value of marine ecosystems and resources. Analyzing
the results available from multiple existing studies could be used in a benefit transfer study to
estimate economic value by transferring available information into the appropriate context.

Terrestrial Ecosystems and Forestry

• Three major types of terrestrial ecosystem impacts are expected: changes in vegetation
distribution and dynamics, wildfire dynamics, and species extinction risks. For example,
predicted global vertebrate extinctions due to land use and climate change range from over 30%
to nearly 60% for >2 degree warming.

• Understanding changes in pest outbreaks, interior wetlands, and snow pack are important gaps.

• Natural scientists and economists need to work together to identify biophysical impacts
assessment endpoints best suited for use in revealed and stated preference valuation studies.

Energy Production and Consumption

• Energy impacts may be beneficial for small to modest climate change, due primarily to
decreases in heating requirements for buildings, but are expected to be dominated by negative
impacts in the long-run and at higher levels of temperature change.

• More data and research are needed to evaluate the effects from wildfire and sea level rise on
power sector infrastructure, and temperature impacts on electricity production, transmission,
and distribution.

Socio-economic and Geopolitical Impacts

• Climate change-induced natural disasters, migration caused by sea level rise and other climate
factors, and increasing resource scarcity may promote conflict; however, the policy debate
regarding socio-economic and geopolitical impacts from climate change is well ahead of its
academic foundation, and sometimes even contrary to the best evidence.

-------
Improving the Assessment and Valuation
Climate Change Impacts for Policy
and Regulatory Analysis

Sponsored by

vvEPA

• U» OCPARTMENt OF

ENERGY

United Slates
Environmental
Protoctxjn Agency

January 27-28, 2011

Capital Hilton, Washington,

Workshop Report:

Improving the Assessment and Valuation of
Climate Change Impacts for Policy and
Regulatory Analysis - Part 2

Research on Climate Change Impacts and Associated Economic Damages

March 2011

Workshop Sponsored by:

U.S. Environmental Protection Agency
U.S. Department of Energy

Workshop Report Prepared by:
ICF International

-------
Table of Contents

Table of Contents	1

I.	Introduction	4

Context	4

Workshop Format	5

II.	Findings and Potential Improvements Identified by Workshop Participants	5

Cross-Cutting Comments	6

Comments related to impact assessment and valuation within sectors	6

Comments related to combining impact assessment and valuation from different sectors	8

Comments related to research and collaboration priorities	9

Sector-Specific Comments	10

Storms and Other Extreme Weather Events	10

Water Resources	11

Human Health	11

Agriculture	11

Sea Level Rise	12

Marine Ecosystems and Resources	12

Terrestrial Ecosystems and Forestry	12

Energy Production and Consumption	13

Socio-economic and Geopolitical Impacts	13

III.	Chronological Presentation of Workshop Proceedings	13

Workshop Introduction	13

Welcome	13

Opening Remarks	14

Questions	15

1

-------
Storms and Other Extreme Weather Events	16

Impact of Climate Change on Storms and Other Extreme Weather Events	16

Global Damages from Storms and Other Extreme Weather Events	17

Discussion: Storms and Other Extreme Weather Events	18

Water Resources	18

Hydrological/Water Resource Impacts of Climate Change	19

Estimating the Economic Impact of Changes in Water Availability	19

Discussion: Water Resources	20

Human Health	21

Climate-Associated Changes in Health Conditions/Diseases and Air Pollution	21

Estimating the Economic Value of Health Impacts of Climate Change	22

Discussion: Human Health	23

Agriculture	24

Biophysical Responses of Agro-ecosystems to Climate Change	24

Estimating the Economic Impact of Climate Change in the Agricultural Sector	25

Discussion: Agriculture	26

Sea Level Rise	27

Sea Level Impacts of Climate Change	27

Estimating the Economic Impact of Sea Level Rise	29

Discussion: Sea Level Rise	29

Marine Ecosystems and Resources	30

Modeling Climate and Ocean Acidification Impacts on Ocean Biogeochemistry	31

Modeling Climate and Acidification Impacts on Fisheries and Aquaculture	32

Economic Impact of Climate Change and Ocean Acidification on Fisheries	33

Nonmarket Valuation of Climate and Acidification Impacts on Marine Resources	34

2

-------
Discussion: Marine Ecosystems and Resources	35

Terrestrial Ecosystems and Forestry	36

Biological Responses of Terrestrial Ecosystems to Climate Change	36

Estimating the Economic Impact of Climate Change on Forestry	38

Valuing Climate-associated Changes in Terrestrial Ecosystems and Ecosystem Services	39

Discussion: Terrestrial Ecosystems and Forestry	40

Energy Production and Consumption	41

U.S. Energy Production and Consumption Impacts of Climate Change	41

Impacts of Climate Change on Global Energy Production and Consumption	42

Discussion: Energy Production and Consumption	43

Socio-economic and Geopolitical Impacts	44

Regional Conflict and Climate Change	44

Migration Impacts of Climate Change	45

Discussion: Socio-economic and Geopolitical Impacts	46

Panel Discussion: Incorporating Research on Climate Change Impacts into Integrated Assessment
Modeling	47

David Anthoff, University of California, Berkeley	47

Tony Janetos, Joint Global Change Research Institute, Pacific Northwest National Laboratory	48

Robert Mendelsohn, Yale University	49

Cynthia Rosenzweig, National Aeronautics and Space Administration	49

Gary Yohe, Wesleyan University	50

Panel Discussion	51

Closing Remarks	51

Summary Comments by U.S. Department of Energy	51

Summary Comments by U.S. Environmental Protection Agency	52

3

-------
I. Introduction

This report summarizes the January 27-28, 2011 workshop, Research on Climate Change Impacts and
Associated Economic Damages, sponsored by the U.S. Environmental Protection Agency (EPA) and U.S.
Department of Energy (DOE). This was the second in a series of two workshops, titled Improving the
Assessment and Valuation of Climate Change Impacts for Policy and Regulatory Analysis.

This report is organized as follows:

• The first section provides an introduction to the report and the workshop, including context and
workshop format.

• The second section provides a summary of the findings and potential improvements to climate
change impacts and damages assessment, as identified by workshop participants. This section
summarizes and categorizes the wide variety of cross-cutting and sector-specific
recommendations highlighted by individual participants over the course of the two-day
workshop.

• The third section provides a chronological presentation of the workshop proceedings, including
a summary of each presentation and discussion section.

• The appendix to the report provides the final workshop agenda, charge questions, participant
list, and extended abstracts of most speaker presentations.

This report serves as the EPA and DOE planning committee's summary of the workshop. It has not
received official endorsement from the workshop speakers and other participants.

Context

In 2009 and early 2010, EPA and DOE participated in the interagency working group on the social cost of
carbon (SCC). The interagency group used the DICE, FUND, and PAGE integrated assessment models
(1AM) to estimate a range of values for the SCC from 2010 to 2050 for use in U.S. government regulatory
impact analyses (RIA). The SCC working group reported their findings in February 2010 and the
estimated SCC values were first used in the analysis of DOE's Energy Conservation Standard for Small
Electric Motors.1 In preparation for future iterations of this process, EPA and DOE seek to improve the
natural science and economic understanding of the potential impacts of climate change on human well-
being. This enhanced understanding is also intended to inform ongoing work of the U.S. government to
improve regulatory assessment and policy analysis related to climate change.

To help inform this work, EPA's National Center for Environmental Economics and Climate Change
Division and DOE's Office of Climate Change Policy and Technology sponsored a pair of invitational
workshops in late 2010 and early 2011. The first workshop took place on November 18-19, 2010 and
focused on conceptual and methodological issues related to modeling and valuing climate change

1 See http://go.usa.gov/3fH.

-------
impacts. It also addressed the implications of these estimates for policy analysis. The second workshop,
which is the focus of this report, took place on January 27-28, 2011, and reviewed recent research that
examines the physical impacts and associated economic damages for a variety of impact categories (e.g.,
human health, agriculture, sea level rise), with a particular focus on knowledge that might be used to
improve lAMs.

Workshop Format

The workshop took place over two days, January 27-28, 2011, at the Capital Hilton Hotel in Washington,
DC. The workshop was attended by approximately 100 individuals, including representatives from
several U.S. federal government agencies, non-governmental organizations, academia, and the private
sector. A full list of workshop participants is available in the Appendix.

The workshop began with opening remarks about recent progress in estimating climate change impacts
and valuing climate damages. The vast majority of the workshop consisted of plenary sessions covering
research on the following specific impact categories:

• Storms and Other Extreme Weather Events

• Water Resources

• Human Health

• Agriculture

• Sea Level Rise

• Marine Ecosystems and Resources

• Terrestrial Ecosystems and Forestry

• Energy Production and Consumption

• Socio-economic and Geopolitical Impacts

Most sessions included presentations by at least one natural scientist and at least one economist,
followed by an open discussion with the audience. The workshop concluded with a panel discussion
about incorporating the research on climate change impacts into integrated assessment modeling and
brief summary comments by representatives of EPA and DOE.

II. Findings and Potential Improvements Identified by Workshop
Participants

Over the course of the two-day workshop, a number of research findings and suggestions for improving
the assessment of climate change impacts and damages were identified by the workshop participants.
This section aims to summarize and categorize those suggestions.

The participants' suggestions related to impacts and damages assessment both generally and within
specific impact categories. As such, the section is organized into two broad categories of comments:
cross-cutting comments and sector-specific comments.

-------
The potential improvements and key findings outlined below represent the perspectives of one or more
participants but, importantly, do not represent a consensus since none was sought at this workshop.

Cross-Cutting Comments

Throughout the workshop, several participants made comments or suggestions related to impact and
damage assessment that apply to more than one impact category. This section is organized into three
types of comments:

• Comments related to impact assessment and valuation within sectors

• Comments related to combining impact assessment and valuation from different sectors

• Comments related to research and collaboration priorities

Comments related to impact assessment and valuation within sectors

Throughout the course of the workshop, the participants' comments and suggestions related to impacts
assessment and valuation within sectors spanned a wide range of topics, including the following:

• Clearly incorporate the human response. Throughout the two-day workshop, numerous
participants emphasized the importance of clearly incorporating the human response to climate
change when estimating climate change impacts. The human response includes both adaptation
and technology development. Many participants emphasized that modeling needs to explicitly
account for the human response to climate change impacts, which will greatly affect the
magnitude of damages. Participants further emphasized that it is important to clearly articulate
what is assumed and included regarding the human response.

• Build on knowledge of non-climate change impacts and responses. During the conference,
several participants emphasized that climate change impacts (e.g., storms, health impacts) and
the corresponding management responses are not without analog. Participants noted that, in
many cases, climate change will simply enhance or reduce impacts of other anthropogenic
stressors, rather than introduce entirely new stimuli. Participants encouraged the assessment
and valuation communities to build on knowledge of existing impacts when modeling future
climate change impacts and responses.

• Use appropriate concepts to measure welfare impacts. Numerous participants emphasized that
GDP is not a measure of welfare, and recommended that alternate measures be used. For
example, one participant suggested that consumer surplus is a better measure of welfare.
Another participant suggested that direct costs of damage provide a reasonable estimate for
global welfare losses.

• Use appropriate scale and level of detail. Several participants discussed the importance of scale
and level of detail in impacts assessment. Participants suggested that the scale and level of
detail required to model each sector is somewhat unique.

o Consider both top-down and bottom-up approaches. Participants noted that both top-
down and bottom-up approaches should be considered to estimate climate change

-------
impacts. One participant noted that estimates from the two approaches should
converge, confirming the accuracy of the estimates.

• Better estimate non-linear impacts and damages. Workshop participants identified several
potential improvements related to the assessment of non-linear impacts and damages, including
threshold effects, non-linear scaling, and unprecedented changes.

o Incorporate threshold effects of physical and biological impacts. Numerous

participants highlighted the thresholds that characterize physical and biological impacts.
Participants emphasized the importance of accounting for thresholds, which pervade
almost every sector, to robustly estimate impacts. One participant noted that aggregate
representations of the natural science can lead to neglect of important threshold
effects.

o Recognize non-linear character of damages. Workshop participants emphasized that
climate change damages do not necessarily scale linearly with socio-economic variables
(e.g., population, income). Several participants highlighted examples of climate damages
that relate inversely or non-proportionally with socio-economic variables such as
population and income (e.g., sea level rise or health damages). Participants emphasized
that it is not appropriate to automatically assume linear scaling of damages.

o Use mechanistic approaches. Several participants recommended the use of mechanistic
and process models since statistical extrapolation may not capture non-linear effects of
unprecedented levels of change. Participants suggested that impacts modelers rely on
basic principles (e.g., plant biophysiology, ocean chemistry) to predict the responses to
new climate conditions. However, in cases where climate change impacts are expected
to be within or close to the range of past variations, statistical models are appropriate.
Participants noted that statistical modeling is more appropriate in some sectors (e.g.,
agriculture, wildfire impacts) than others.

• Increase focus on extreme climate events. Several workshop participants highlighted the need
for impacts and valuation assessment to examine the impacts and damages from extreme
climate events for incorporation in lAMs and the SCC. Several participants noted that impact
studies tend to examine temperature changes of 2.5 to 3 degrees Celsius, neglecting to evaluate
the impacts at higher temperature changes. Furthermore, impacts assessment has tended to
focus on the means of the probability distribution of impacts, neglecting to evaluate the low-
probability, high impact tails of the distribution. Participants emphasized the importance of
these high-impact events (e.g., extreme temperature increase, sea level rise) on 1AM results.

• Account for very long-term (beyond 2100) impacts. Several participants noted the need for
impacts and valuation assessment to examine the damages from very long-term (e.g., beyond
2100) impacts. Participants emphasized the importance of these long-term impacts (e.g., ice
sheet melting, possible deep ocean anoxia) for 1AM results. Participants noted that impact and
damage assessments tend to focus on more near-term impacts. They explained that it is critical

-------
to model long-term impacts, despite the great uncertainty associated with doing so.
Characterization of that uncertainty is important.

• Distinguish between committed and projected impacts. Several workshop participants
highlighted a need to distinguish between committed and projected impacts in models.
Participants emphasized that models need to correctly account for this notion of irreversibility in
terms of both positive and negative impacts and damages. They explained that due to the
inertia of natural systems, mitigation will not lead to an immediate cessation of some impacts.
For example, mitigation can stabilize the rate of sea level rise but cannot prevent the realization
of significant, already committed sea level rise.

• Fully account for non-market and non-use values. Several participants emphasized the
importance of fully incorporating non-market and non-use values to build robust estimates of
damages. Several participants recommended that revealed and stated preference estimates and
benefit-transfer methods be improved to value these impacts. One participant suggested that
problems with both revealed preference and stated preference methods can be mitigated by
joint estimation of revealed preference and stated preference data. Another participant
recommended collaboration between natural scientists and economists so that impacts
assessment endpoints correspond to the items assessed in valuation exercises.

• More fully incorporate uncertainty. Throughout the conference, several participants
recommended that scientists attempt to characterize the great uncertainty in impacts
assessments. lAMs can be used to evaluate the impact of some aspects of sector-specific
uncertainty on the SCC.

• Rigorously evaluate impact models. Several workshop participants emphasized the importance
of rigorously testing, comparing, and evaluating impact models. Several participants noted the
important role of Model Intercomparison Projects (MIPs) in improving physical climate models,
and suggested that these types of approaches be utilized for a wider range of models. One
example of an ongoing impact MIP is the Agricultural Model Intercomparison and Improvement
Project (AGMIP).

• Recognize the potential for unexpected and unpredictable events. Several participants
highlighted the fact that there will likely be unexpected impacts from climate change. For
example, one participant described an unexpected outbreak of a shellfish-caused
gastrointestinal disease in Alaska due to increases in water temperatures above a previously
unknown threshold. Another participant noted that new agricultural pests with unknown
characteristics will likely develop as a result of climate change.

Comments related to combining impact assessment and valuation from different sectors
Participants also suggested potential improvements in modeling interactions between different natural
and economic systems. These suggestions include the following:

-------
• Incorporate interactions between sectors. Throughout both days of the workshop, numerous
participants highlighted the importance of including interactions between sectors when
estimating impacts. Participants highlighted several examples of interacting sectors, including
sea level rise and extreme storms, heat-related health effects and space cooling demand, and
water resources and energy demand, among others. Participants noted that interactions and
feedbacks may be synergistic or antagonistic, additive, multiplicative, or subtractive. As a result,
the cumulative impacts may be larger or smaller than the sum of the individual impacts.
Participants further noted the importance of avoiding double-counting when integrating the
climate change impacts across sectors.

• Keep climate change in context. Over the course of the workshop, several participants
emphasized that climate change is only one of many other global, anthropogenic changes
impacting the sectors discussed. Participants highlighted the need to incorporate interactions
with non-climate stressors and to account for the risk of climate change in relation to other
global changes. For example, one participant presented the impacts of water allocation policy
on water resources; another participant highlighted needs (e.g., education) other than climate-
related health issues competing for investment in developing countries; and, a third participant
highlighted coastal management decisions that affect the impacts from relative sea level rise.

Comments related to research and collaboration priorities

Finally, participants made suggestions related to framing a research agenda. These suggestions include
the following:

• Focus research efforts on impacts with greatest magnitude. Several workshop participants
recommended that future research efforts focus on the impacts and damages with potentially
the greatest magnitude. One participant noted that it is more important to estimate a high cost
damage with some quantified but sizable measure of error, than a lower cost damage with high
precision. This recommendation applies both within and among sectors. For example, one
participant suggested that it is more important to improve estimates of mortality impacts than
to improve estimates of morbidity impacts since the monetized value of mortality impacts tends
to overwhelm the monetized value of morbidity impacts. Another participant suggested that
lAMs can be used to evaluate the relative magnitude of impacts in different sectors. Research
can then be targeted on improving estimates of those impacts with the greatest effect on the
SCC.

• Foster interaction between natural scientists, economists, and modelers. Throughout the
workshop, participants highlighted the importance of increasing collaboration between natural
scientists, economists, modelers, and all those involved in impacts assessment, damages
valuation, and integrated assessment modeling.

o Encourage trans-disciplinary review of 1AM damage functions. Multiple participants
recommended that 1AM data sources, damage functions, and outputs be reviewed by
relevant members of the Impacts, Adaptation, and Vulnerability (IAV) and economic

-------
valuation communities. In this way, the communities could ensure that lAMs reflect the
current state of the primary literature for each of the impact categories

o Link into existing efforts. One participant recommended that impacts assessment and
valuation efforts be coordinated among existing efforts such as the National Climate
Assessment, the United Nations Environment Programme (UNEP) Programme of
Research on Climate Change Vulnerability, Impacts and Adaptation (PRO-VIA), and other
international efforts to improve knowledge on impacts and valuation.

• Use consistent scenarios. Workshop participants suggested that consistent climate scenarios
should be used in impact and damage assessment. This would facilitate intercomparison of
impacts and damages estimates, and would also aid in the integration and combination of these
estimates into lAMs.

• Increase capacity to address challenges. Numerous participants highlighted a need for
additional funding and staff to help address existing impacts and damages assessment
challenges. A couple of participants highlighted that 1AM development is severely understaffed
and underfunded, particularly as compared to general circulation model (GCM) development.
One participant highlighted discrepancies in funding between different impact sectors, noting
the relative lack of funding for climate change health impacts.

Sector-Specific Comments

The vast majority of the workshop proceedings focused on the current state of research in nine impact
categories. This section aims to highlight the key research findings and recommendations for future
research for each of the nine impact categories, as identified by workshop participants. The summaries
below are based directly on the workshop presentations, and are not intended to be comprehensive.

Storms and Other Extreme Weather Events

• Fewer tropical storms are expected in the future, but the average wind speeds and precipitation
totals of the storms that do occur are expected to increase. The intensity of the strongest storms
is expected to increase.

• Estimates in the literature for increases in cyclone property damages due to climate change
range from 0.002 to 0.006% of Gross World Product (GWP). Increases in property damages from
all extreme events (including cyclones) due to climate change, according to one study, range
from $47-$102.5 billion (2008 dollars) per year, or 0.008-0.018% of GWP, by 2100.

-------
Water Resources

• Water demand, supply, and management must be modeled on a river basin scale in order to
effectively estimate climate change impacts.

• National estimates from the literature of climate change damages to water resources range
from $11.5-$60 billion (2009 dollars) per year for the United States.

• Coupling approaches that model changes using regional hydrologic models and approaches that
model changes using regional economic models could help to bridge some gaps in water
resources damage estimation.

Human Health

• The most significant health effects from climate change result from diarrhea, malnutrition, and
malaria.

• The World Health Organization (WHO) estimates that the costs to treat climate change-related
cases of diarrhea, malnutrition, and malaria in 2030 would be $4 to $13 billion. This would be a
3% increase in diarrhea cases, a 10% increase in malnutrition cases, and a 5% increase in malaria
cases.

• Health impact valuation depends largely on mortality valuation, particularly in developing
countries, and particularly among children. Appropriately adjusting the value of a statistical life
for income is critical for accurate valuation.

• Health impacts in developing countries must be considered in the context of other stressors.

Agriculture

• Estimates in the literature project the global range of yield changes in the 2050s to be
approximately -30 to +20%, compared with 1990, under an A2 SRES emission scenario, which
corresponds to a 2.3°C mean global temperature increase from base temperatures (1961-1990).

• Global effects of climate change on agriculture are expected to be positive on average in the
short term and negative in the long term. The location of the inflection points, with respect to
climate change severity over time, where impacts change from positive to negative are
unknown.

o C02 fertilization from increasing carbon dioxide concentrations will benefit some plants
(C3 plants) more than others (C4 plants). Elevated C02 concentrations especially benefit
weeds, which are particularly good at taking advantage of high C02 concentrations.

• Agriculture's contribution to only 2-3% of U.S. GDP is due, in part, to the paradox of value and
price, where rare, nonessential goods cost more than essential goods. However, the highly
inelastic nature of agricultural demand means that even a small reduction in agricultural
production from climate change could result in large price changes and large welfare losses.

-------
• Adaptation and technology change can help to mitigate the impacts of climate change on
agriculture. A key challenge will be producing heat and drought tolerant plants with high yields.

Sea Level Rise

• Climate-induced sea level rise will be compounded by subsidence in many densely-populated
coastal areas.

• The greenhouse gases emitted so far have committed the planet to a significant amount of sea
level rise that will be fully realized over coming centuries. Emissions abatement may stabilize the
rate of sea level rise, but not reduce the globe's current commitment to sea level rise.

• The valuation of sea level rise damages depends heavily on wetland values and on adaptation.

Marine Ecosystems and Resources

• Damages from decreased mollusk harvest revenues due to a 0.1-0.2 ocean pH decrease are
estimated at $1.7 to $10 billion in net present (2007) value losses through 2060.

• Assessments using the following three approaches would be beneficial to estimate marine
impacts: a bio-climate envelope, which evaluates species tolerance to changing environmental
variables, to provide key first pass estimate; minimum realistic models, which represent only the
most important components of a specific system, on high value fisheries; and ecosystem and
food web models to assess component interactions.

Terrestrial Ecosystems and Forestry

• Three major types of terrestrial ecosystem impacts are expected: changes in vegetation
distribution and dynamics, wildfire dynamics, and species extinction risks. Predicted global
extinctions range from relatively low levels up to 60% of species.

• Understanding changes in pest outbreaks, interior wetlands, and snow pack are important gaps
in impacts assessment.

• Natural scientists and economists need to work together to identify biophysical impacts
assessment endpoints that correspond to the items assessed in valuation exercises.

-------
Energy Production and Consumption

• Energy impacts may be beneficial for small to modest climate change, but are expected to be
dominated by negative impacts in the long-run.

• In the U.S. and across the group of industrialized countries, energy use and expenditures for
space conditioning is expected to decrease due to near-term warming, since decreases in energy
demand for heating are likely to be greater than increases in energy demand for cooling. Net
demand changes may be quite different from this conclusion over the long-term and for other
regions.

• More data and research are needed to evaluate wildfire and sea level rise impacts on power
sector infrastructure, and temperature impacts on electricity production, transmission, and
distribution.

Socio-economic and Geopolitical Impacts

• Climate change-induced natural disasters, migration caused by sea level rise and other climate
factors, and increasing resource scarcity may promote conflict.

• Estimates for the number of future environmental refugees range from 50 million by 2010 to 1
billion by 2050. These estimates include numerous environmental causes for displacement,
including climate change.

• The policy debate regarding socio-economic and geopolitical impacts from climate change is
running well ahead of its academic foundation, and sometimes even contrary to the best
evidence.

III. Chronological Presentation of Workshop Proceedings

This section presents the proceedings of the workshop in chronological order, including the following
components: workshop introduction; session presentations and discussion sessions; panel discussion;
and closing remarks. The following summary represents statements from one or more workshop
participants but, importantly, represents neither the views of EPA or DOE, nor a consensus, since none
was sought at this workshop.

Workshop Introduction

The workshop introduction included a welcome from Dr. Elizabeth Kopits of EPA, followed by opening
remarks from Dr. Michael Oppenheimer of Princeton University and Dr. William Cline of the Peterson
Institute for International Economics, as well as questions about the opening remarks.

Welcome

The workshop commenced with a welcome by Dr. Elizabeth Kopits. She noted that this workshop was
the second of two EPA- and DOE-sponsored workshops in preparation for future efforts to estimate the
social cost of carbon. She mentioned the previous meeting focused on integrated assessment models,
the 2009-2010 SCC process, and broad conceptual issues associated with integrated assessment. She

-------
explained that the second workshop would focus on a more detailed review of the quantitative research
on climate change impacts and associated damages, which underpins integrated assessment models.
She noted that in some sectors, the science and economics may have evolved enough to indicate ways
in which 1AM damage functions can be improved; while in other sectors, it may only be possible to
enumerate research gaps and priorities. She highlighted the desire to facilitate increased dialogue and
coordination between natural scientists and economists, noting that each session of the workshop
would pair a natural scientist with an economist and would include interdisciplinary dialogue.

Opening Remarks

Progress in estimating climate change impacts

Following Dr. Kopits' introduction, Dr. Michael Oppenheimer of Princeton University presented opening
remarks on past and potential future progress in estimating climate change impacts. He noted that the
systematic assessment and valuation of potential climate change impacts and damages dates back to
the 1970s. He explained that there have been recent advances in process-based and statistical modeling
of physical exposure and impacts. These advances include improvements in general circulation model
resolution and downscaling; statistical modeling of agriculture, migration, and conflict responses; and
deployment of GIS data.

In contrast, there has been slow and limited progress in accounting for adaptation capacity and human
responses. Dr. Oppenheimer noted that current approaches to incorporate adaptation responses are
obscure, that there is little understanding of the gap between adaptation capacity and implementation,
and that indirect effects of human responses (e.g., of migration) have not been assessed. He emphasized
the importance of incorporating the human response in climate change impacts modeling.

Dr. Oppenheimer then presented five emerging areas in estimating climate change impacts. First, he
presented the indirect and remote consequences of human responses, such as human population
migration or shifts in human activities (e.g., agriculture) that affect resources and populations at a
distance. Second, he presented the interacting effects of adaptation and mitigation, highlighting the
biodiversity, food price, and political ramifications of biofuel development and the political
reverberations of geo-engineering. Third, he presented the complexity of interacting systems and
stressors. For example, he noted how upstream water diversion can cause deltaic subsidence which
interacts with sea level rise. Fourth, he presented complexities associated with climate extremes and
disasters, such as the local specificity of exposure and vulnerability, and the learning that may occur as
rare events become more frequent. Fifth, he presented the dynamic nature of vulnerability, which
evolves with development and changes as learning competes with mal-adaptive and risk-shifting
behavior. He noted the complexities associated with diverse potential development pathways and the
existing gap between top-down and bottom-up estimates of impacts.

Progress in valuing climate damages

Next, Dr. William Cline of the Peterson Institute for International Economics presented opening remarks
on past and future progress in valuing climate damages. He discussed the different historical approaches
to estimating the SCC and emphasized the importance of catastrophic risk valuation and discount rate

-------
selection. Dr. Cline emphasized the role of the pure rate of time preference in Ramsey discounting,
noting the use of a zero pure rate of time preference in several estimates. Dr. Cline also emphasized the
effect of uncertainty, noting that insuring against catastrophe would warrant aggressive action even
without a zero pure rate of time preference.

Regarding catastrophic risks, Dr. Cline noted that some have been considered in damage valuation,
including collapse of the ocean conveyor belt, collapse of the West Antarctic Ice Sheet, and a runaway
greenhouse effect from methane release. Dr. Cline then presented the possibility of hydrogen sulfide
release from deep ocean anoxia due to anaerobic bacteria buildup caused by sea level rise and
shutdown of the ocean conveyor belt. He explained that the resulting elevated hydrogen sulfide levels
could be toxic to plants and animals and could cause mass extinction. He noted that this risk should not
be ignored, despite the fact that it would likely not occur for 2000 years or more. In order to account for
very long term but catastrophic damages, Dr. Cline argued that contingent valuation might be necessary,
as the use of even the lowest interagency discount rate produces numbers too low to justify action to
prevent a risk so far in the future. He suggested that super-contingent valuation may be helpful to value
such catastrophic risks. He noted that the contingent valuation implied by the Copenhagen pledge to
reduce emissions from developing countries far exceeds current 1AM estimates for the social cost of
carbon.

Regarding the discount rate, Dr. Cline noted that a descriptive approach using Treasury Inflation
Protected Securities yields a discount rate lower than what was used by the interagency process. Dr.
Cline then noted the importance of the elasticity of marginal utility when using a prescriptive approach,
suggesting that a value of 1.5 is more appropriate than a value of 1 as proposed by Stern or a value of 2
as proposed by Weitzman.

Dr. Cline noted that, if targets are set, the SCC is defined as the marginal abatement cost along the least
cost pathway to achieving the targets. Finally, Dr. Cline suggested that the interagency group should
consider an insurance approach to determining the SCC.

Questions

In response to a question from the audience, Dr. Oppenheimer noted his interest in statistical
approaches. He emphasized that estimates from top-down approaches and bottom-up approaches
should converge, thereby giving one indication of their reliability.

During the question and answer session, one participant suggested that the time scale of long-term C02
effects is tens of thousands years, two orders of magnitude larger than what is typically considered. Dr.
Cline noted that he may have contributed to the use of a scale of hundreds of years, but that typical
discount rates imply that what happens after 80 years is of little consequence. A second participant
agreed with Dr. Cline in preferring a precautionary or insurance approach over a cost-benefit approach.
The participant noted that as the marginal benefit of the abatement curve approaches vertical, the
precautionary benefits of abatement increases dramatically.

A third participant suggested that adaptation will occur on a global scale and asked how adaptation
could be incorporated into the SCC and U.S. policies. In response, Dr. Oppenheimer first noted the many

-------
levels of human response, from individual to international. He then suggested that many studies do not
consider the human response at all. Finally, he noted that a comprehensive solution to incorporate
adaptation may not yet be available, but that it is important to strive to do better and to incorporate the
human response to impacts at whatever level is possible.

Storms and Other Extreme Weather Events

Following the workshop introduction, there were nine sessions, each covering a specific impact
category. The first session covered the impacts and damages from storms and other extreme weather
events. The session was moderated by Dr. Alex Marten of the U.S. Environmental Protection Agency and
included presentations by Dr. Tom Knutson, National Oceanic and Atmospheric Administration (NOAA);
and Dr. Robert Mendelsohn, Yale University.

Impact of Climate Change on Storms and Other Extreme Weather Events

Dr. Tom Knutson, of NOAA's Geophysical Fluid Dynamics Laboratory (GFDL), presented the effects of
climate change on tropical cyclones. Most of his presentation focused on the findings from a World
Meteorological Organization (WMO) study. The study sought to determine two things: whether the
human impact on hurricanes is detectable and what future climate change implies for hurricane activity.

Regarding the detection and attribution of climate change in tropical cyclone activity, the study
concluded that it remains uncertain whether past changes in tropical cyclone activity exceed natural
variability levels. For example, while the frequency of past tropical cyclone activity is correlated with
increases in sea surface temperature, the trends disappear when storm counts are adjusted to account
for improved storm observation data.

To evaluate the future implications of climate change for tropical cyclone activity, the study considered
the changes associated with the Intergovernmental Panel on Climate Change (IPCC) Special Report on
Emissions Scenarios (SRES)2 A1B scenario. To predict future activity, the study used high resolution
atmospheric models, regional dynamical downscaling models, and statistical/dynamical techniques that
are able to reproduce historic interannual variability. The WMO expert team made five general
conclusions regarding the effect of future climate change on tropical storm activity.

First, the models almost unanimously predict that there will likely be fewer tropical storms globally, with
predictions ranging from no change in frequency to a decrease of 34%. There is greater uncertainty
regarding the frequency of storms in individual basins (e.g., the Atlantic), though storms in the southern
hemisphere decrease in frequency fairly consistently in the different models.

Second, the experts predict a likely increase in average hurricane wind speeds (intensity) globally, with
increases ranging from two to 11 percent. Dr. Knutson noted that this statistic does not necessarily
apply to individual basins.

Third, the experts predict that there is a greater than 50% chance that the frequency of very intense
hurricanes will increase by a substantial fraction in some basins. Fourth, climate change will likely result

2 http://www.ipcc.ch/ipccreports/sres/emission/index.htm

-------
in higher rainfall rates in hurricanes. The models predict a roughly 20% increase within 100 km of a
storm. Fifth, sea level rise is expected to exacerbate storm surge impacts, even if storms themselves do
not change.

Dr. Knutson summarized these conclusions by explaining that the experts predict fewer tropical storms
overall, but an expanded range of storm intensity. The combination of these two changes will result in
an increased number of intense storms. Throughout his presentation, Dr. Knutson emphasized the wide
range of estimates and disagreement among different models for various tropical cyclone metrics.

Dr. Knutson finished his presentation by noting several factors that might exacerbate the study's
projections. These factors include changes in the vertical profile of warming, the El Nino / La Nina
pattern of warming, wind shear, and ocean heat transport.

Global Damages from Storms and Other Extreme Weather Events

Dr. Robert Mendelsohn of Yale University expanded on Dr. Knutson's presentation by discussing the
effects of climate change on all extreme events, and by delving into the damages associated with these
effects. Dr. Mendelsohn examined cold events, drought, floods, hail, heat waves, tornadoes,
thunderstorms, tropical cyclones, and extratropical severe storms. His presentation aimed to describe
how climate change affects future extreme events, to reflect any underlying changes in future
vulnerability, to estimate damage functions for each type of extreme event, and to describe damages
from future extreme events caused by climate change.

First, Dr. Mendelsohn considered the change in damages and deaths due to changes in future income
and population, without considering climate change effects. He showed that damages increase greatly
due to these socio-economic changes, noting that damages from heat waves increase most dramatically
and damages from tropical cyclones are the greatest in magnitude. Next, he showed that deaths do not
follow the same pattern, with deaths associated with some extreme events (e.g., heat waves) predicted
to increase, while others (e.g., those associated with floods, tropical cyclones) predicted to decrease. He
noted that deaths increase with increased population, but decrease with increased income.

Dr. Mendelsohn next summarized estimates of changes in cyclone damages due to climate change. He
noted that increases in damages have been estimated at 0.002%-0.006% of Global World Product. He
noted that the tropical cyclone generator, which models cyclone development, estimates different
results in different basins, but shows a consistent increase in tropical cyclone power in the Atlantic and
Northwest Pacific basins.

Dr. Mendelsohn then regressed damages and deaths on different variables to develop damage
functions. He noted that, using U.S. data, damages increase with both income and population density,
but less than proportionally. Using international data, however damages increase less than
proportionally with income, but decrease as population density increases. Dr. Mendelsohn noted that
the impacts of climate change due to cyclones will be greatest in North America and Asia. He further
noted that the impacts from other extreme events are likely to be less significant by an order of
magnitude than those from tropical cyclones.

-------
Dr. Mendelsohn explained several limitations associated with this estimation process, including:
uncertainty of non-hurricane impacts; the coarseness (e.g., country-scale) of some of the key data; the
need for better data about damages, extreme events, and the relationship between them; the paucity of
information on ecosystem impacts; and the lack of explicit incorporation of adaptation in impact
models.

Dr. Mendelsohn concluded with a brief summary of results. He noted that predicted climate change
impacts from all extreme events (including tropical cyclones) range from $47 to $100 billion per year by
2100. This is equivalent to 0.008-0.018% of GWP by 2100. Finally, he noted that climate change has a
mixed effect on fatalities because tropical cyclone deaths may fall more than other deaths increase.

Discussion: Storms and Other Extreme Weather Events

During the question and answer session, a couple of participants asked about intersectoral impacts,
particularly the interaction between storms and sea level rise. Dr. Mendelsohn noted that the effects of
sea level rise and storms are at least additive, but may be interactive and more than additive.

A couple of participants questioned the result that the additive effects of climate change on tropical
storms will result in a decrease in deaths. Dr. Mendelsohn explained that this result is due to the
predicted decrease in tropical cyclone intensity in the Indian Ocean due to climate change, where
approximately 90% of current tropical cyclone deaths occur. He noted that deaths are predicted to
increase in every part of the world except for Southern Asia, but that the magnitude of the increases will
be small. He also clarified that his work did not monetize deaths.

Several participants asked about the estimation of deaths from heat waves. Dr. Mendelsohn explained
that with heat waves, it is the variance rather than the average temperature that matters, since humans
are able to adapt to higher average temperatures. He noted that his work does not include ecosystem
effects, such as mammalian die-off due to heat waves, nor does it include empirical work on human
labor changes due to temperature increases. One participant challenged the assertion that only variance
matters, noting that humans are not able to live at very high temperatures. Dr. Mendelsohn clarified
that only variance matters within the context of extreme events.

One participant asked whether a valuation of impacts on agriculture and famine was incorporated. Dr.
Mendelsohn noted that agriculture effects are incorporated but famine effects are not. Another
participant asked whether the predicted relative contribution to damages from different extreme events
will change significantly as more research is conducted. Dr. Mendelsohn noted that damages from
floods and perhaps heat waves may prove to be more important than currently predicted. However, he
explained that it was unlikely that they would increase by an order of magnitude, thus concluding that
tropical cyclones will continue to have the greatest impact.

Water Resources

The second session covered the impacts and damages to water resources. The session was moderated
by Dr. Robert Kopp of DOE and included presentations by Dr. Ken Strzepek, University of Colorado at
Boulder and Massachusetts Institute of Technology; and Dr. Brian Hurd, New Mexico State University.

-------
Hydrological/Water Resource Impacts of Climate Change

Dr. Kenneth Strzepek of the University of Colorado at Boulder and the Joint Program on the Science and
Policy of Global Change at the Massachusetts Institute of Technology commenced the second impact-
specific discussion by discussing how impacts to water resources can be incorporated into lAMs. Dr.
Strzepek explained that water resources encompass a broad spectrum of issues and appear in 18 of the
30 chapters of the IPCC Working Group 2 report.

He explained that, globally, the largest water use is agriculture, while, in the United States, the largest
use is thermal cooling for energy production. Dr. Strzepek emphasized the importance of using the
appropriate spatial and temporal scale when modeling impacts to water resources. He noted that
disaggregation is necessary in order to properly model water supply and demand. Dr. Strzepek explained
that a regional or national scale, as often used in lAMs, is far too course to properly model water
resources and crops, since the local location of water is critically important and aggregation averages
water excesses and water shortages. Using several examples, he illustrated how aggregation can
misrepresent the supply and demand of water resources. Instead, Dr. Strzepek suggested that a river
basin scale may be more appropriate.

Dr. Strzepek further suggested that water management systems must be modeled at the basin-level to
appropriately describe impacts and, especially, adaptation. He noted that water management systems
can adapt to changing water supply by increasing water storage capacity to level supply. He noted that
modeling water management is crucial since cross-sectoral impacts are greater than sectoral impacts,
since different uses are forced to compete for available water. Dr. Strzepek also explained that the
metrics used to model water resources are important. For example, different metrics of drought
produce very different results.

Next, Dr. Strzepek explained that flooding and storms are very important when considering climate
change impacts to water resources. He noted that a lot of work has been conducted on the effects of
drought, but that flooding is particularly harmful as it can cause serious damage to capital investments,
including infrastructure such as roadway bridges.

Finally, Dr. Strzepek emphasized that threats to water resources must be considered in the context of
other global changes. He illustrated that municipal and industrial water demand and environmental
policy threaten agricultural water supply more than climate change. He underscored that the effects of
climate change must be considered in relation to other uncertainties and stressors. He also mentioned
that addressing uncertainty is important.

In response to a question, Dr. Strzepek explained that water management and technological
improvements can ameliorate some impacts (e.g., by leveling supply using dams), however a lack of
water due to climate change could still have serious implications.

Estimating the Economic Impact of Changes in Water Availability

Following Dr. Strzepek's presentation, Dr. Brian Hurd of New Mexico State University presented
estimates of the economic impacts of climate-related changes in water availability. Dr. Hurd again
emphasized the complexity of water and water systems, noting the spatial and temporal variability as

-------
well as the variability in uses, infrastructure, and vulnerability. He noted that statistical modeling is
nearly impossible and that process models work better for estimating the magnitude of changes outside
of historical experience. He highlighted the importance of behavioral aspects such as adaptation and
optimization.

Dr. Hurd then presented national-level estimates from the literature of annual economic impacts of
climate change on water resources. Estimates in the literature range from $7 billion to $60 billion (2009
dollars), under varying assumptions. These estimates arise both from studies that use a hydro-economic
model to aggregate benefits and costs and from studies that use regional economic models to estimate
impacts on jobs, income, and GDP. He noted that despite very different methodologies, the estimates
are consistent in order of magnitude and share of GDP. He noted the counter-intuitive result from some
studies that GDP may increase where impacts are greatest, since disasters can increase the number of
jobs in certain sectors and locations.

Dr. Hurd concluded with a list of knowledge gaps that limit current estimates and provide areas for
future improvement. These gaps include: understanding changes in extreme events; the role of water
rights, and federal and state regulation; administrative constraints in adaptation; projections of market
prices and trade flows of agricultural and other water-intensive products; measuring, monitoring, and
modeling groundwater; water security and food security issues; and assessing and measuring economic
outcomes of water quality. He suggested that coupling approaches that model changes using regional
hydrologic models (hydro-economic) and approaches that model changes using regional economic
models (dynamic system simulation) approaches could help to bridge some gaps.

Discussion: Water Resources

During the question and answer session, one participant questioned the negative impacts for New York
State presented by Dr. Hurd, given her understanding that climate projections indicate that water will
increase in New York. Dr. Hurd clarified that the study in question only presented impacts from
projected drought scenarios without looking at projected increases in water availability. Dr. Strzepek
further noted that models indicate an increase in winter precipitation for New York. Without reservoir
capacity, New York would be unable to harness the additional water.

A second participant highlighted an example of non-climate global changes interacting with climate
change impacts. He noted that there is a movement by EPA to make closed-cycle cooling the standard
for thermoelectric power, which would eliminate the 48% of U.S. water currently used for power plants.
A third participant questioned whether water withdrawals or water consumption is more important
from a modeling and policy perspective. He noted that cooling towers reduce withdrawals but increase
consumption. Dr. Strzepek explained that this is a very complicated issue that is affected by temperature
and runoff in addition to consumption. He noted that within the United States the issue is most relevant
in the western United States.

A fourth participant asked for global estimates of damages to tie this sector into SCC estimates. Dr. Hurd
noted that he had focused the research for his presentation on U.S. impacts. Dr. Strzepek cited a series

-------
of studies by the World Bank that estimate adaptation costs for water resources at $80-100 billion per
year for developing countries. He noted that additional global work has been funded and is underway.

A fifth participant asked about the importance of variability with regard to water resources. Dr. Strzepek
emphasized that variability, as well as seasonality, is crucial. However, he noted that the GCMs have
trouble capturing this variability. Dr. Hurd further noted that the literature is not well developed on the
real impact and economic estimates of those changes. Another participant asked about the attitude of
water managers towards variability. Dr. Strzepek explained that water managers tend not to worry
about variability because engineers overbuild infrastructure to withstand 100-year events. Since some
water managers believe that the uncertainty about what constitutes the current 100-year event dwarfs
the potential effects of climate change, they also believe that if systems are prepared for current
variability, they are also prepared for climate change.

Human Health

The third session covered the impacts and damages to human health. The session was moderated by Dr.
Charles Griffiths of EPA and included presentations by Dr. Kristie Ebi, Carnegie Institution for Science;
and Dr. Maureen Cropper, Resources for the Future and University of Maryland at College Park.

Climate-Associated Changes in Health Conditions/Diseases and Air Pollution

Dr. Kristie Ebi of the Carnegie Institution for Science introduced the third impact category with her
presentation on climate-associated changes in health outcomes. Dr. Ebi presented several different
health conditions that will be affected by climate change, highlighting malnutrition, diarrheal disease,
and malaria. Since climate change is never the direct cause of death, she noted that deaths due to
climate change have to be modeled.

Dr. Ebi began by presenting an overview of the direction and magnitude of different climate change
health impacts, as presented by the IPCC Fourth Assessment Report (AR4). The only net positive impact
predicted by the report is a reduction in cold-related deaths. Meanwhile, the report predicts net
negative impacts from increases in malaria; malnutrition; deaths, disease, and injuries from extreme
weather events; cardio-respiratory disease from changing air quality; infectious disease; and diarrheal
disease.

Dr. Ebi noted that the current impacts of these diseases are enormous. For example, under-nutrition
results in 35% of child deaths, 11% of the total global burden of disease, and 21% of disability-adjusted
life-years (DALYs) for children younger than 5 years. When all the effects of malnutrition are considered
(including loss of cognitive function, poor school performance, and loss of future earning potential), the
total estimated costs of environmental risk factors could be as high as 8-9% of a typical developing
country's GDP in South Asia or Sub-Saharan Africa. Dr. Ebi noted that the scale of current impacts means
that even small increases in the impacts will have significant effects. For example, even small increases
in temperature could result in enormous increases in the number of mosquito vectors and thus in the
prevalence of malaria.

Dr. Ebi then discussed the major regional differences in impacts and their concomitant equity
implications. While diarrheal disease has impacts across the globe, the major burden exists in India and

-------
Africa. The burden of malaria lies almost exclusively in Africa. These distributional differences extend to
the country level, with different areas within a country experiencing very different rates of disease.

Dr. Ebi noted that current models do not account for complexities and interactions between different
impacts. For example, poor nutritional status promotes infectious disease and vice versa. She also
mentioned that higher temperatures can lead to increased ozone formation, which affects anyone with
compromised lung function or cardiovascular disease.

Dr. Ebi next presented estimates of the costs to treat climate change-related illness from the United
Nations Framework Convention on Climate Change (UNFCCC). The UNFCCC estimates that the costs to
treat climate change-related cases of diarrhea, malnutrition, and malaria in 2030 would be $3,992 to
$12,603 million. This accounts for a 3% increase in diarrhea cases, a 10% increase in malnutrition cases,
and a 5% increase in malaria cases. Dr. Ebi then noted that there will likely be unexpected climate
change health impacts. As an example, she described the unexpected outbreak of a shellfish-caused
gastrointestinal disease in Alaska, which was caused by increases in water temperatures above a
previously unknown threshold.

Dr. Ebi finished her presentation with an extensive list of research needs emphasizing the significant
need for work to understand exposure-response, to model impacts, to take into account other drivers,
and to understand adaptation. In response to a question, Dr. Ebi noted the need for research on
threshold effects. She noted that the health sector has been severely underfunded with regards to other
climate change effects, particularly in the United States.

Estimating the Economic Value of Health Impacts of Climate Change

Dr. Maureen Cropper of the University of Maryland and Resources for the Future built on Dr. Ebi's
presentation with a presentation on estimating the economic value of the health impacts of climate
change. Dr. Cropper noted that economists should value damages after adaptation, plus the costs of
adaptation. However, she focused her presentation on valuing the health impacts themselves.

Dr. Cropper noted that health impact valuation depends largely on mortality valuation, particularly in
developing countries, and particularly among children. She emphasized that, by far, the largest damages
are due to mortality. She underscored that it is better to estimate a crucial issue with some measure of
error (e.g., mortality), than a less important issue with high precision (e.g., morbidity).

Dr. Cropper next discussed two different approaches to valuing increased risks of mortality and
morbidity. First, the human capital-cost of illness (COI) approach values increases in mortality risk using
the present discounted value of forgone earnings, and values an injury by the associated medical costs
and lost productivity. Second, the value of a statistical life (VSL)-willingness to pay (WTP) approach
values increases in mortality risk using what people will pay for small reductions in risk of death, and
values an injury by adding the willingness to pay to avoid pain and discomfort to the COI value. Dr.
Cropper noted that the VSL can be estimated using revealed preference studies (e.g., based on
compensating wage differentials, purchase of safety equipment) or stated preference studies.

-------
Dr. Cropper noted that there have been dozens of VSL studies in high-income, and even middle-income
countries, but that there has only been one study in a low-income country (Bangladesh). She explained
that the VSL can be transferred from one country to another using the income ratio between the two
countries and the appropriate income elasticity. Dr. Cropper noted that the income elasticity is usually
assumed to be one, however, many other factors that affect the VSL differ between countries, including
risk preferences, life expectancy, and consumption. Dr. Cropper showed evidence from the literature
that the income elasticity should be greater than one and should increase as income falls. Based on a
couple of studies, she suggested that an income elasticity of 1.5 is more appropriate.

Dr. Cropper then discussed the difficulties associated with estimating the VSL for children. She explained
that an accepted method is the use of parents' willingness to pay to reduce risks to their children. In
high income countries, this method suggested that the VSL for a child is approximately twice that of an
adult. However, parents' WTP may be different in countries where one out of five children dies before
age five. She suggested that, in the interim, the same VSL be used for adults and children until a
sensitivity analysis can be conducted.

Finally, Dr. Cropper discussed valuing morbidity. She noted that most estimates capture the value of lost
productivity and the cost of medical treatment but that most estimates neglect the value of discomfort,
inconvenience, and pain. She again noted that morbidity damages are significantly smaller than
mortality damages, once more suggesting that it is most important to refine mortality estimates. She
finished by noting that there may be other relevant health impacts, such as the macroeconomic impacts
of malaria or the impacts of malnutrition on human capital formation, which could each affect economic
growth.

Discussion: Human Health

During the discussion session, a couple of participants challenged the use of WTP as the most
appropriate way to frame and value climate impacts. Instead, the participants suggested that willingness
to accept compensation be used. This suggestion is based on the fact that the people that cause climate
change (e.g., developed countries, older generations) tend not to be the people that are impacted by
climate change (e.g., developing countries, younger generations). Dr. Cropper defended the use of WTP,
emphasizing that WTP reflects market preferences.

During the discussion, Dr. Ebi and Dr. Cropper emphasized that climate change is just one of many
stressors on developing countries. Dr. Cropper suggested that it is important to keep climate change in
context and acknowledge that developing countries may prefer to invest in, e.g., education rather than
climate change mitigation or adaptation. She suggested that viewing climate change impacts
independently would result in over-allocation of resources to climate change issues. One participant
asked whether countries should invest in mitigation or in current health impacts. Dr. Ebi suggested that
the money for the two different issues does not tend to come from the same sources. In response to
another question, she emphasized that the smallest portfolio of funding is directed towards climate
change health issues, since it is difficult to attribute health impacts to climate change.

-------
One participant asked how demographic transitions affect health impacts of climate change. Dr. Ebi
explained that the issue is complex. She noted that many countries are already undergoing a
demographic transition. However, she further noted that there are limits on how much wealth can
address health impacts. As an example, Dr. Ebi explained that malaria and dengue control is extremely
difficult and requires discovering the right balance of components and maintaining efforts.

A participant asked about impacts on labor productivity. Another participant explained that he used a
biophysical model of the human body to estimate how much labor a person can produce at different
temperatures and humidity levels. He explained that normal non-air conditioned labor is not possible
above certain thresholds. He concluded that without adaptation, labor productivity could fall by 30
percent and GDP could fall by 10 to 15 percent. A final participant asked what reference case should be
used when evaluating the SCC and asked if and how current adaptation efforts will affect the reference
case. Dr. Ebi agreed that current adaption would change the reference case.

Agriculture

The fourth session covered the impacts and damages to agriculture. The session was moderated by Dr.
Charles Griffiths of EPA and included presentations by Dr. Cynthia Rosenzweig, National Aeronautics and
Space Administration (NASA) Goddard Institute of Space Studies; and Dr. Wolfram Schlenker, Columbia
University.

Biophysical Responses of Agro-ecosystems to Climate Change

Dr. Cynthia Rosenzweig of the NASA/Goddard Institute for Space Studies introduced the fourth impact
category by presenting the biophysical climate change effects on agro-ecosystems. Dr. Rosenzweig
began by presenting observed climate change impacts on agriculture, which include high temperature
effects on rice yield, earlier planting of spring crops, increased forest fires, change in pests, and declines
in livestock productivity.

Dr. Rosenzweig explained that studies show insects are emerging earlier, including agriculturally
beneficial insects such as bees, as well as pests such as the potato beetle. She noted that climate change
may cause new pests to emerge, one of the potential climate change surprises. Next, Dr. Rosenzweig
showed that increasing carbon dioxide concentrations benefit C3 plants (e.g., wheat, rice, soybean,
barley) more than C4 plants (e.g., corn, sorghum, sugarcane), as C4 plants are already able to concentrate
C02. She emphasized that increasing C02 concentrations benefit weeds in addition to crops, noting that
weeds are favored as they are particularly good at taking advantage of high C02 concentrations.

Next, Dr. Rosenzweig explained that increasing temperatures can speed up growth cycles. This
acceleration negatively impacts yields, as crops have less time to accumulate carbohydrates. She further
noted that high temperature stress during critical growth periods (e.g., pollination) could have
detrimental effects. Dr. Rosenzweig then described the effects of changes in precipitation. She noted
that both drought stress and excess water can be damaging to yields.

Dr. Rosenzweig presented temperature maps that show warming is expected to be greatest over land
and at the highest northern latitudes. Similarly, she showed maps that indicate increases in precipitation
are very likely in the high latitudes, while decreases are likely in most subtropical land regions. She

-------
showed that the most negative yield effects are expected in the lower latitudes, where developing
countries are, while the less negative or more positive effects will be in the higher latitudes.

Dr. Rosenzweig noted that globally, the literature consistently estimates the range of yield changes to be
approximately -30 to +20 percent. The estimates of the most negative effects range from -32 to -35
percent, while the estimates of the most positive effects range from 19 to 25 percent. Dr. Rosenzweig
then showed that global effects of climate change are expected to be positive in the short term and
negative in the long term. She noted that the location of the inflection points where impacts change
from positive to negative are unknown.

Dr. Rosenzweig then presented the three main approaches to model agricultural impacts, along with
advantages and disadvantages, data requirements, spatial resolution, and level of uncertainty for each
approach. First, statistical approaches use historical data to estimate statistical relationships between
crop and climate variables. These relationships are then used to project climate impacts on yield.

Second, expert system approaches use statistical relationships between observed crop yields and
observed climate variables to estimate production potential. Third, dynamic process crop models use
data and modeled relationships to explicitly simulate the various processes affected by climate. She
noted that the graphs that show increasing yield responses to low levels of warming were assembled
using largely incomparable data points from very different models and studies, using different
coefficients.

Following her presentation of modeling approaches, Dr. Rosenzweig discussed the ability of adaptation
and technology to modulate the biophysical impacts. She presented three levels of adaptation, each
with increasing benefit, as well as increasing complexity, cost, and risk. The first level includes adjusting
varieties, planting times, and spacing. The second level includes actions such as diversification and risk
management. The third level includes transformation from land-use or distribution change. She
demonstrated that adaptation is not always possible or complete.

Finally, Dr. Rosenzweig finished with a list of gaps and uncertainties related to the biophysical climate
change impacts on agriculture. She emphasized the importance of precipitation impacts, which are
critical but relatively unknown. Other gaps and uncertainties include: simulating extreme weather
events; interactions between warmer temperatures, C02, and ozone; interactions between
evapotranspiration, soil moisture, crop yield, and water availability; pests; scale effects; yield gaps and
plateaus; and multi-model comparisons and assessments. She emphasized the importance of rigorously
testing and comparing models, and noted AGMIP, the Agricultural Model Intercomparison and
Improvement Project, which is a relatively new effort to assess and ultimately improve agricultural
models.

Estimating the Economic Impact of Climate Change in the Agricultural Sector

Next, Dr. Wolfram Schlenker of Columbia University and the National Bureau of Economic Research gave
his presentation on estimating the economic impact of climate change in the agricultural sector. First,
Dr. Schlenker presented the fact that U.S. agriculture only accounts for two to three percent of U.S.
GDP, which might be interpreted to mean that agricultural impacts are negligible. He explained that the

-------
low contribution to GDP results from the paradox of value and price, where rare, nonessential goods
cost more than essential goods. Through a series of graphs he showed that GDP is not a welfare
measure and suggested that consumer surplus is a better option. He showed that because agricultural
demand is highly inelastic, a small reduction in agricultural production (e.g., from climate change) results
in large price changes and could lead to large welfare losses (i.e., reductions in consumer surplus).

Then, Dr. Schlenker discussed the global importance of U.S. agriculture. He explained that corn, rice,
soybeans, and wheat contribute 75 percent of the calories consumed by humans worldwide. World
caloric production has been trending upward, resulting in falling real prices over the 20th century. Dr.
Schlenker explained that the U.S. share of caloric production has been roughly constant at around 23
percent for the last 50 years. He noted that this share is larger than Saudi Arabia's share of oil
production, which means that impacts on U.S. yields have the potential to influence world markets.

Next, Dr. Schlenker presented a statistical analysis examining the link between temperature and yields.
He explained the highly non-linear relationship between yields and the number of exposures to
particularly cold or warm days (above 84-86°F). He noted that the negative slope of impacts at high
temperatures is ten times greater than the slope at low temperatures, which implies large yield declines
if maximum temperatures increase significantly. He concluded that the driving force behind climate
change impacts on agriculture is extreme heat, with impacts depending on both the baseline
temperature and the predicted increase.

Dr. Schlenker then explored the ability of technological progress to mitigate climate impacts. Through a
series of graphs, Dr. Schlenker presented the historic evolution of heat tolerance using data from
Indiana. He showed that while corn yields have increased continuously in the second half of the 19th
century by a total factor of three, the evolution of heat sensitivity is highly nonlinear, growing with the
adoption of double-cross hybrids in the 1940's, peaking around 1960, and then declining sharply as
single-cross hybrids were adopted. However, Dr. Schlenker questioned whether future innovation could
increase both yield and heat tolerance. He suggested that genetically modified crops may have the most
potential.

Dr. Schlenker then discussed the role of agriculture and land use change in contributing to or mitigating
climate change. He noted that land use change is responsible for approximately 20% of C02 emissions.
Dr. Schlenker specifically discussed ethanol, which converts agricultural land from food production to
energy production in an effort to mitigate climate change. He explained that the estimated food supply
elasticity is roughly twice as large as the demand elasticity. As a result, one third of the caloric input
diverted to biofuel production would be compensated with a reduction in food consumption while two
thirds would be compensated with increases in food production. He noted that the U.S. ethanol
mandate is predicted to lead to a decrease in food consumption of 1%, an increase in commodity prices
of 20%, and a possible expansion of agricultural areas.

Discussion: Agriculture

During the discussion session, a couple of participants asked questions about C02 fertilization effects.
One participant asked how C02fertilization should be incorporated into reduced-form models, such as

-------
those used to develop the SCC. Dr. Rosenzweig explained that AGWIP would hopefully be able to isolate
the C02 effects for incorporation. She expressed her belief that an average of current estimates is
correct. She believes that the high- and low- (zero) ends of current estimates are both incorrect. Dr.
Schlenker noted that there is a wide range of estimates found in the literature. Dr. Rosenzweig
suggested that a risk management approach is most appropriate, where ranges and uncertainties are
estimated and used, instead of a single number. Another participant asked whether C02 fertilization
effects are non-linear and characterized by plateaus. Dr. Rosenzweig noted that there are bursts and
ebbs in some processes but that effects continue up to concentrations of 700, and possibly even
800ppm.

One participant asked about the biophysical basis for climate change effects and whether biophysical
barriers might pose a limit to adaptation efforts. Dr. Rosenzweig first outlined the biophysical effects of
climate change, including temperature-caused speed up of the lifecycle, damage at critical growth
periods, and water stress. She reiterated that the easiest adaptation efforts include management
actions such as planting earlier. She noted that crop breeders are optimistic about the potential of
genetic improvements, though there is not a lot of plasticity in the genes controlling for certain growth
stages. Dr. Rosenzweig emphasized the challenge of pairing heat tolerance with high yields.

Another participant asked whether the speakers thought current estimates are optimistic (meaning
incomes will likely be worse than predicted) or pessimistic, particularly considering the existence of
known and unknown unknowns. Dr. Schlenker acknowledged that the unknowns pose a difficult
question. Dr. Rosenzweig suggested current estimates may be overly optimistic. She emphasized the
need for collaboration between climate scientists, agronomists, and economists. A different participant
suggested the need for greater interaction between economic models and crop models. Dr. Rosenzweig
agreed, noting that AGMIP facilitates a trans-disciplinary interaction and dialogue.

A final participant asked to what extent Dr. Schlenker's current statistical results could be used to
improve lAMs. He noted the need to separate the effects of temperature, C02, and precipitation in
lAMs. Dr. Schlenker emphasized the uncertainty in modeling precipitation, particularly extremes. He
concluded that using currently available estimates of extreme precipitation would not necessarily
improve agricultural impact estimates.

Sea Level Rise

The fifth session concluded Day 1 of the workshop and covered the impacts and damages from sea level
rise. The session was moderated by Dr. Robert Kopp of DOE and included presentations by Dr. Robert
Nicholls, University of Southampton; and Dr. Robert Tol, Economic and Social Research Institute.

Sea Level Impacts of Climate Change

Dr. Robert Nicholls of the School of Civil Engineering and the Environment and the Tyndall Centre for
Climate Change Research at the University of Southampton introduced the last impact category of Day
One, the sea-level impacts of climate change. Dr. Nicholls began by emphasizing the importance of sea
level rise, despite the coasts being a small proportion of the earth's surface. He showed that population
and economic density in the coastal zone is significantly greater than other areas of the earth's surface.

-------
Next, Dr. Nicholls explained that climate-induced sea level rise is caused by the thermal expansion of
seawater, as well as the melting of land-based ice (e.g., small glaciers in the Rockies or Alaska, the
Greenland ice sheet, the West Antarctic ice sheet). He showed that sea level was fairly stable in the 19th
century and that the rate of sea level rise has accelerated recently. He noted the great uncertainty
regarding future projections of sea level rise.

Dr. Nicholls emphasized the importance of keeping climate change in context. While climate change is
contributing to sea level rise, the coast is also experiencing other changes that contribute to changing
sea level (e.g., coastal management, water extraction). Dr. Nicholls emphasized that relative sea level,
which is determined by both sea level rise and subsidence, is what matters.

Dr. Nicholls then presented the impacts of sea level rise, which include: inundation, flood, and storm
damage; wetland loss and change; erosion; saltwater intrusion; and higher water tables and impeded
drainage. He noted that all five impacts are affected by interacting climate and non-climate factors. Dr.
Nicholls next showed the links between sea level rise impacts and socio-economic sectors, noting the
high number of strong links and lone potential benefit. Dr. Nicholls presented a series of images showing
observed impacts from sea level rise (including its interaction with storms) and maps identifying the
areas, cities, and assets exposed to future sea level rise.

Next, Dr. Nicholls presented a graph showing the limits of mitigation actions to control sea level rise. He
emphasized the globe's current commitment to sea level rise, noting that mitigation efforts are only
able to stabilize the rate of sea level rise. He emphasized that mitigation is still beneficial, while limited.
He noted that the globe's commitment to sea level rise indicates a need for adaptation action.

Dr. Nicholls explained that adaptation can include (planned) retreat from the coasts, accommodation of
assets (e.g., raising houses on stilts), and coastal protection using hard or soft barriers. Each impact is
associated with multiple possible adaptation responses. He noted that, generally, the relative cost of
adaptation is extremely low when compared to the coasts' value. Dr. Nicholls presented the optimistic
and pessimistic views of potential impacts from and adaptation to sea level rise. He noted that both
views are supported by reasonable arguments.

Dr. Nicholls finished with a series of concluding remarks, including the following. Climate-induced sea-
level rise is inevitable; the major uncertainty is its magnitude. Climate-induced SLR will be compounded
by subsidence in many densely-populated coastal areas. Risks are already increasing, and this will
continue. The worst-case (do nothing) impacts are dramatic. There are widely differing views concerning
the success or failure of adaptation. Mitigation of climate change and subsidence is needed to make the
problem more manageable. To adapt to dynamic coastal risks, proactive assessment is required.

Following Dr. Nicholls' presentation, one participant suggested that sea level rise studies be combined
with studies on storm length and intensity, citing the importance of winter storms in the Netherlands.
Dr. Nicholls agreed and suggested that specific drivers and key issues need to be evaluated for each
place on the earth's coast.

-------
Estimating the Economic Impact of Sea Level Rise

Dr. Richard S. J. Tol of the Economic and Social Research Institute in Dublin, Trinity College in Dublin, and
Vrije Universiteit in Amsterdam, continued the discussion of sea level rise by presenting the economic
impact of sea level rise. Dr. Tol presented the economic implications of sea level rise, focusing on direct
costs, adaptation, and general equilibrium effects.

Dr. Tol explained that, to estimate direct costs, economists typically estimate a unit cost and multiply
the unit cost by the impact estimates provided by natural scientists such as Dr. Nicholls. For example, to
estimate the costs of inundation, an economist would multiply the number of acres submerged by the
average acre value. Dr. Tol emphasized that average acre values should be used as opposed to beach
front values since property markets will adjust to coastal realignment. He noted the difficulty in
estimating average acre cost, citing a study that used nonmarket valuation to identify wetland values.

Next, Dr. Tol emphasized the importance of incorporating adaptation into estimates of climate damages.
He showed that populations with higher income generally suffer less and are less vulnerable to floods.
However, he noted that even fairly sophisticated models are only able to explain 60 to 70 percent of
vulnerability, due to a large amount of variation that is not understood. He noted that while optimal
adaptation models can be built, historically, adaptation has never been optimally implemented. He
showed numerous examples of suboptimal adaptation implementation, where adaptation efforts
indicate an under- or over-valuation of damages, as estimated under the IPCC Special Report on
Emissions Scenarios (SRES) A1B scenario. For example, the Dutch currently pay approximately 0.2% of
GDP for coastal protection while damages are estimated to be less than 0.1% of GDP. One participant
questioned the assumptions and results presented, questioning why Holland would be affected by the
impacts from the SRES A1B scenario.

Finally, Dr. Tol presented general equilibrium effects of sea level rise. He explained that land loss would
affect agriculture, and hence all other markets, and that coastal protection would affect construction
and capital. He presented the results from a static computable general equilibrium (CGE) model, first
with no protection and then with full protection. He noted that impacts only amount to fractions of a
percent and that losing capital is more important than losing land. Dr. Tol emphasized that increases in
GDP modeled under full protection are misleading since GDP is a measure of economic activity, not
welfare. Instead, Dr. Tol suggested that, globally, direct costs are a reasonable measure of welfare costs.

Dr. Tol finished with a series of conclusions. He noted that sea level rise is one of the better understood
impacts even though estimates contain significant uncertainty. He noted that the extent of saltwater
intrusion, future storm characteristics, wetland value, and adaptation are some of the largest sources of
uncertainties.

Discussion: Sea Level Rise

During the discussion session, several questions touched on the issue of timescales. One participant
noted the contradictory conclusions about impacts from warming that have been generated in different
studies. For example, politicians have identified 2°C of warming as problematic, FUND has identified net
benefits up to 3°C of warming, and other studies indicate that the Greenland ice sheet would collapse

-------
with 3°C of warming. He noted that sea level rise is one of the biggest impacts on a long timescale. Dr.
Tol clarified that SLR is not a large component of marginal impacts. He explained that the damages from
SLR due to melting of the Greenland ice sheet would depend on the timescale of the melting. If
complete melting occurred over two to three centuries, with sea level rising at three meters per century
or more, it would be very difficult to adapt, causing significant damages. However, if the rate of sea level
rise was two meters per century, it would be possible to raise dikes and adapt. He noted that rapid
collapse of the West Antarctic Ice Sheet would cause many coastal cities to be largely flooded.

In response to another question, Dr. Nicholls emphasized the importance of evaluating the distribution
of impacts over time, rather than focusing on expected annual damages. In particular, he noted the
importance of events such as storms. He noted that while climate change may increase storms, storms
drive coastal action today and the issues will be fundamentally the same in the future.

Another participant again raised the issue of interacting impacts, expressing his frustration that
workshop discussions have focused on storms without sea level rise and sea level rise without storms.
He emphasized the non-linear and interactive nature of climate change impacts. Dr. Tol noted that there
are other interacting impacts as well, including changes in wind and sedimentation patterns.

Several questions addressed the impacts in the Netherlands. In response to one question, Dr. Tol
explained that the Dutch are overprotecting against some predicted impacts. In fact, the speakers noted
that the Dutch conducted an economic analysis intending to justify their work, but got results that
indicated the work was not justified. In response to another question, Dr. Tol clarified that the Dutch are
spending about twice as much as they would have in the absence of predicted sea level rise, to prepare
defenses for 60-80 cm of SLR in the SRES A1B scenario. Dr. Nicholls noted that the SRES scenarios are
optimistic.

The last group of questions concerned extreme storms. One participant asked the speakers to confirm
that the Netherlands were building coastal infrastructure to withstand a 1 in 10,000 year event while
New Orleans is building for a 1 in 100 year event. Dr. Nicholls explained that the new defenses in New
Orleans are built for a 1 in 100 year event, but would probably withstand a 1 in 500 year event without
breach. Another participant asked the speakers to confirm that SLR was not substantively included in
rebuilding efforts after Hurricane Katrina. Dr. Nicholls confirmed that New Orleans may have done a
little to include SLR, but for the most part, SLR was not included.

Marine Ecosystems and Resources

After brief Day 2 opening comments from Elizabeth Kopits, the sixth session commenced, covering the
impacts and damages to marine ecosystems and resources. The session was moderated by Dr. Chris
Moore of EPA and included presentations by Dr. Sarah Cooley, Woods Hole Oceanographic Institute; Dr.
Paul McElhany, National Oceanic and Atmospheric Administration (NOAA); Dr. David Finnoff, University
of Wyoming; and Dr. John Whitehead, Appalachian State University.

-------
Modeling Climate and Ocean Acidification Impacts on Ocean Biogeochemistry

Dr. Sarah Cooley of the Woods Hole Oceanographic Institute initiated the discussion on marine
ecosystems and resources by presenting an overview of modeling changes in ocean biogeochemistry
due to ocean acidification and climate change. She organized her discussion into four sections.

First, Dr. Cooley presented an overview of the chemistry and observed impacts of ocean acidification.
She explained that a quarter of the anthropogenic C02 burden dissolves in the ocean, combining with
water to produce carbonic acid. She noted that the rate of present change in ocean acidification is too
fast to be compensated by rock weathering and other mechanisms. Dr. Cooley presented a series of
graphs that show increasing atmospheric C02 concentrations are associated with increasing ocean C02
concentrations, decreasing ocean pH, and decreasing saturation states for calcite and aragonite, which
are used by marine animals to produce hard parts (e.g., shells). She also showed that anthropogenic C02
has penetrated to ocean depths of thousands of meters.

Dr. Cooley noted that ocean acidification is likely to cause other changes in ocean biogeochemistry. For
example, nitrogen-fixing organisms such as phytoplankton thrive in the higher concentration of C02,
likely causing a shift in the nitrogen pool towards ammonia. Additionally, changing pH and/or C02
concentration will likely change metal ion speciation, increasing both copper (which is toxic) and iron
(which is a fertilizer). Dr. Cooley emphasized that ocean acidification is occurring along with numerous
other anthropogenic stressors, which could be antagonistic or synergistic to acidification-induced
change.

Second, Dr. Cooley discussed Earth system model simulations and their ability to predict future
conditions. She noted the use of data-model comparisons to evaluate model skill. She explained that it is
crucial to correctly model ocean physics and that biogeochemical parameterizations are under
continuous improvement. Dr. Cooley explained the use of model intercomparisons to create and
evaluate forecasts, including the Ocean Carbon-Cycle Model Intercomparison Project (OCMIP). She
explained that the most significant uncertainty in modeling ocean acidification is identifying future
atmospheric C02 concentrations.

Third, Dr. Cooley discussed biological responses to ocean acidification. She identified numerous
biological groups that will be directly or indirectly affected by ocean acidification, including corals,
mollusks, plankton, reef communities, and marine predators. She demonstrated that calcification
responses vary significantly among different organisms. She emphasized that individual, population, and
ecological implications, including follow-on food web effects, are not yet understood. She presented
evidence that calcifiers tend to vacate areas when conditions do not suit them.

Next, Dr. Cooley discussed the valuation of ecosystem services. She noted that most studies focus on
market values, but that non-market values, indirect use values, and non-use values must also be
incorporated in an informed analysis. She presented an estimate of damages assuming that decreases in
pH result in lower mollusk harvests. She estimated annual losses of $75 to $187 million in ex-vessel
revenues from a 0.1 to 0.2 pH decrease, amounting to $1.7 to $10 billion in net present value losses
through 2060. Dr. Cooley noted that valuation of impacts on coral reefs is driven by tourism effects.

-------
Fourth and last, Dr. Cooley discussed knowledge gaps and needs. She noted the need to properly link
three main models: physical, biological, and human/economic. She identified numerous relationships
within these models that are not well understood. Finally, Dr. Cooley presented the increasing level of
uncertainty associated with the progressing stages of ocean acidification impacts (e.g., changes in ocean
pH are more certain than effects on marine organisms, which are more certain than changes in
ecosystem services).

In response to a question, Dr. Cooley explained that there are a large number of studies examining the
observed impacts of historic ocean acidification. However, she explained that there are no good
baselines to ascertain when "normal" conditions are exceeded. She noted that numerous time series
stations are currently examining this question, which is high on the international research agenda.

Modeling Climate and Acidification Impacts on Fisheries and Aquaculture

Dr. Paul McElhany of the NOAA Northwest Fisheries Science Center expanded on Dr. Cooley's
presentation with a discussion of modeling climate change and acidification impacts on fisheries,
aquaculture, and other marine resources.

Dr. McElhany started by enumerating the impacts and impacted resources associated with climate
change and ocean acidification. He noted that impacts on capture fisheries will be complicated, while
aquaculture has some ability to adapt using relocation, control, and species switching. He further noted
that while the direct C02 effects on growth and survival are relatively well understood, the effects on
stratification and circulation are not known.

Next, Dr. McElhany described nearly a dozen different model types that are used to model impacts on
marine resources, including: fishery stock assessments, population viability analyses, food
web/ecosystem models, NPZ (nutrients, phytoplankton, zooplankton) models, minimum realistic
models, maximum unrealistic models, modeled range maps, individually-based models, life-cycle
models, bioenergetics, and expert systems. Dr. McElhany noted that IPCC-class Earth system models
must be downscaled to match the near-shore, small-scale processes at the biological scale. He noted
that the IPCC avoided modeling coastal ocean impacts due to their complexity. However, he emphasized
that biological action is concentrated in coastal regions and that these gaps must be addressed.

Dr. McElhany then presented several examples of marine resources modeling. His examples spanned a
wide range of scale and scope. Some examples only examined a single variable or a single species, while
other examples examined all climate change impacts or entire ecosystems. Dr. McElhany noted the vast
complexities associated with the life cycle of a single species and the greater complexities associated
with ecosystems. He emphasized the importance of modeling interactions between species. He
highlighted one study's results that indicate a general decline in fisheries, especially with all climate
change effects, and that range shifts will be the biggest impact. He noted the ambitious nature of the
Atlantis model which attempts to couple oceanography, ecology, and fisheries submodels. He also noted
a fairly comprehensive evaluation of fisheries using a bioclimate envelope.

Dr. McElhany provided a "reality check" that identified numerous big questions remaining regarding
marine resources. He identified significant unknowns including changes in the Gulf Stream, stratification,

-------
upwelling, and decadal oscillations, among others. He noted the possibility for positive changes, such as
improved fishing in some areas. Dr. McElhany further noted that details are critical in modeling marine
resources. For example, species interactions, phenology, synergistic effects, short-term variability, and
local circulation are all critical factors. He noted that lab studies do not necessarily scale to ecosystems.

Finally, Dr. McElhany suggested that coarse-scale impact assessment would be beneficial in the future.
He suggested the use of back-of-the-envelope estimates and assessment using three approaches: a
bioclimate envelope to provide key first pass estimates, minimum realistic models, which model only the
most important components of a specific system, on high value fisheries, and ecosystem and food webs
to look for interactions. He emphasized the importance of resolving the big climate questions.

In response to a question, Dr. McElhany noted that he was not aware of any studies examining the
effects of changing aragonite saturation states on fish. He noted one observational study that indicates
major oyster reproductive failures in the past several years, which are correlated to pH changes. A
different participant suggested that the reference case for marine resources needs to be carefully
considered and that acidification impacts need to be considered in the context of a variety of other
environmental stressors that affect the baseline.

Economic Impact of Climate Change and Ocean Acidification on Fisheries

Following Dr. McElhany's presentation, Dr. David Finnoff of the University of Wyoming commenced the
economic portion of the marine resources discussion, with his presentation on the economic impact of
climate change and ocean acidification on fisheries. Dr. Finnoff began by describing the potential
significance of ocean acidification, citing historic mass extinction events linked to ocean surface pH,
challenges for calcifying organisms, and Dr. Cooley's work that calculated net present value losses from
decreased mollusk harvests of $1.7 to $10 billion through 2060. He noted that Dr. Cooley's work, while
providing a useful initial estimate, is based on lost revenue rather than more appropriate measures of
welfare such as consumer surplus.

Next, Dr. Finnoff discussed the economic consequences of ocean acidification, noting that disruptions in
ecosystem services are material damages that imply welfare changes. He highlighted the reciprocity of
the relationship where ocean acidification is caused by human activity and, in turn, affects human
activity. Dr. Finnoff explained that assessment of material damage requires characterizing the changes in
production and consumption, determining the responses of prices, and identifying adaptation options.
He noted that changes in ocean acidification do have the potential to affect production possibilities, as
well as direct and indirect costs.

Dr. Finnoff explained that both reduced-form/partial equilibrium and structural/general equilibrium
representations have pros and cons. He emphasized the importance of identifying the appropriate
balance and utilizing both approaches. He explained that non-convexities and species interaction require
more detailed and comprehensive models. Through a series of simplified graphs, he demonstrated that
with problems characterized by non-convexity, it is necessary to understand the entire surface of
possibilities to be able to locate the global optimum.

-------
Using an illustrative example of the Bering Sea Food Web, he discussed a simplified model that might be
used in an 1AM. He demonstrated the non-linear, non-systematic results from shocks, and identified
non-convexities, non-monotonic changes, and problems with reduced-form aggregation. He concluded
that bio-economic harvests offish and crab are likely affected to varying degrees and magnitudes
depending on their location in the food web; non-harvested stocks may or may not have cascading
effects depending on their location in the food web; and to assess tradeoffs, it is necessary to assess
changes in flows and stocks simultaneously.

Dr. Finnoff concluded that welfare measurement of materials damages has some well-known
characteristics, but that for ocean acidification, a lot of issues remain unresolved. He suggested that a
clear understanding is needed of how ocean acidification affects production and consumption
possibilities in a consistent setting. He noted that using dose-response relationships of environmental
change from the natural sciences is crucial, but that it is not yet resolved how much detail is necessary
for a good understanding. Finally, Dr. Finnoff concluded that if problems are convex or well-behaved,
aggregate representations of the natural science may be sufficient for good economic assessments.
However, if problems have pervasive non-convexities, he noted that policy makers must expand the
scope of their analysis for good economic assessments. It may be necessary for the assessor to know the
entire possibilities surface.

Nonmarket Valuation of Climate and Acidification Impacts on Marine Resources

Dr. John Whitehead of Appalachian State University delivered the final presentation in the marine
ecosystems and resources impact category. He described nonmarket valuation of climate change and
ocean acidification impacts to marine resources.

Using the example of coral reefs, Dr. Whitehead described the different methods available to estimate
nonmarket values. He explained that use values may be estimated by the willingness to avoid climate
change due to use of affected resources. Direct uses of coral reefs include diving, snorkeling, and
viewing; indirect uses include fishing. He then explained that non-use, or passive use, values may be
estimated by the willingness to avoid climate change without the intent to use the affected resources.
Willingness to pay for nonuse values can be motivated by altruism, ecological ethic, or bequests.

Dr. Whitehead explained that use values can be estimated using revealed preference or stated
preference valuation methods, while non-use values can only be estimated using stated preference
methods. Dr. Whitehead further explained that revealed preference methods include hedonic price,
averting behavior, and travel cost methods, as well as producer surplus values. He noted that the travel
cost method is most appropriate when considering marine resources. Dr. Whitehead then described
stated preference methods, which include contingent valuation, choice experiments, and contingent
behavior. He explained that there are problems with both revealed preference and stated preference
methods, which can be mitigated by joint estimation of revealed preference and stated preference data.

Dr. Whitehead cited several examples in the climate change literature of revealed preference and stated
preference studies. He explained that no study to date explicitly addresses nonmarket valuation of
climate change and marine resources. Instead, Dr. Whitehead discussed a very simple nonmarket

-------
valuation based on data from national recreation surveys, where he regressed saltwater fishing
participation and fishing days on temperature and precipitation. He suggested a more complex
estimation would be possible using the recreational fisheries demand study.

Dr. Whitehead concluded that there is very little existing research with which to develop the SCC for
marine resources. He suggested that meta-analyses could be used in a benefit transfer study, using
values for coral reef recreation, outdoor recreation, and recreational catch. However, he noted that the
behavioral response to climate change is missing. Dr. Whitehead suggested that a wide variety of studies
is needed, using both revealed and stated preferences, to estimate total economic value, use value, and
non-use value. He suggested the most promising avenue is using existing revealed preference data. New
studies using stated preference data could differentiate between marine and other values and estimate
the behavioral response to climate change. Revealed preference and stated preference joint estimation
could differentiate between use and non-use value.

Discussion: Marine Ecosystems and Resources

During the question and answer session, one participant asked how changes in keystone species can be
incorporated in food web models. Dr. McElhany suggested that if food web models are built properly,
keystone species should be included. He noted that model results become more tenuous as conditions
change further away from the case in which the model was parameterized.

A second participant questioned the incorporation of thresholds and discontinuities into economic
models. He noted that economic models indicate small marginal changes, but that natural scientists
tend not to consider marginal changes, as they are more concerned with thresholds. Dr. Finnoff agreed
about the importance of thresholds and discontinuities, emphasizing the aspects of his presentation that
dealt with non-linearities. Dr. Finnoff explained that the economics literature knows how to handle
thresholds, in principle. He suggested the need for an approach to evaluate the proximity of thresholds.
He suggested using a recursive view and developing a model that can handle changes in states. Dr.
Whitehead added that there is a need for non-use values in a world very different from today. He
suggested the possibility that entire classes of opportunities could disappear. He noted the need for
modeling to address individuals' recreational choices. Dr. McElhany cited large scale ecological changes
in the North Pacific as a historical example of state changes that resulted in big community changes.

A third participant asked if the rate of ocean carbon uptake is constant or changing. Dr. Cooley noted
recent efforts to evaluate the ocean's ability to take up C02 in the long run. She cited evidence that
ocean uptake is slowing and will continue to slow due to chemical reasons and changes in ocean
circulation. She noted that the slow-down will not reverse or even significantly alleviate ocean
acidification.

Several participants asked about the interactions between different stressors. One participant asked
about climate change impacts other than ocean acidification, such as loss of phytoplankton biomass. Dr.
McElhany explained that the results he presented were based on a model generation previous to newer
data on changes in primary productivity. He emphasized there is ongoing and continued learning, as well
as remaining unknowns including changes in primary productivity, in ocean circulation, in temperature

-------
regimes, in stratification, and in availability of nutrients. He noted that each unknown could have
significant effects.

Another participant asked if productive areas of the ocean would be squeezed as warming-induced
range shifts move commercially valuable species pole-ward and ocean acidification pushes some species
toward the equator since ocean acidification happens more rapidly in colder water. Dr. McElhany agreed
that might happen, noting a lack of study on the interacting trends. Dr. Cooley agreed with the
participant's summary, noting the need to do lab experiments to better understand the interacting
effects.

A different participant asked whether coral reefs would be able to adapt to sea level rise by growing
towards the sun and whether ocean acidification would affect their ability to adapt. Dr. Cooley noted
that coral reefs can grow annually by millimeters or centimeters. However, she noted several interacting
factors that might impede the ability of corals to adapt, including the change in deep ocean chemical
conditions and the vertical and latitudinal shrinking of optimal waters. She explained that these
interactions are not well understood. She further noted that coral growth rates do not necessarily
correlate with vertical growth, due to the somewhat horizontal structure of corals.

A final participant asked about incorporating coral bleaching into lAMs. He noted that coral bleaching is
tied to warming and is an example of a non-linear, non-marginal impact. Dr. Cooley agreed with the
need to incorporate coral bleaching, disease, and destruction. She suggested that research on ocean
acidification is a necessary first step, since it is necessary to understand acidification before it is possible
to understand synergistic interactions. She further noted that ecosystem-scale studies are time- and
manpower-intensive, and expensive, resulting in a small number of existing studies. Dr. Whitehead
added that revealed preference studies would not address coral bleaching well, but that stated
preference studies could. Dr. Finnoff noted economic studies on previous large scale disasters might be
informative to this issue.

Terrestrial Ecosystems and Forestry

The seventh session covered the impacts and damages to terrestrial ecosystems and forestry. The
session was moderated by Dr. Steve Newbold of EPA and included presentations by Dr. Karen Carney,
Stratus Consulting; Dr. Brent Sohngen, Ohio State University; and Dr. Alan Krupnick, Resources for the
Future.

Biological Responses of Terrestrial Ecosystems to Climate Change

Dr. Karen Carney of Stratus Consulting started the terrestrial ecosystems and forestry discussion by
presenting the impact of climate change on terrestrial ecosystems. She noted that her presentation was
not meant to be comprehensive, instead aimed at highlighting some key impacts and related tools.

Dr. Carney described how terrestrial ecosystems provide numerous economically important services:
the provisioning of food, water, and raw materials (e.g., timber, non-timber forest products); regulation
of air quality, storm protection, and waste assimilation; and cultural services such as recreation and
passive use value. She noted that climate change will fundamentally and potentially dramatically affect
the location and character of today's ecosystems. She noted key changes including changes in species

-------
locations, ecosystem productivity, rates of ecosystem processes, and disturbance regimes (e.g., drought,
fire, pest outbreaks).

Next, Dr. Carney discussed three major ecosystem impacts—changes in vegetation distribution and
dynamics, wildfire dynamics, and species extinction risks—that have the potential to be included in
lAMs. She selected these impacts as they best met the following criteria: ecological importance,
economic importance, and being well understood. For each of the three impacts, Dr. Carney discussed
why the impact is likely to occur, the tools available to estimate the impact, what research has shown,
key uncertainties or other shortcomings with projecting future impacts, and what key services are likely
to be affected.

Dr. Carney noted that changes in vegetation distribution and dynamics, which will be affected across the
globe, are most commonly examined using dynamic global vegetation models (DGVMs). She noted that
there are many DGVMs available that can examine multiple scales (e.g., countries, regions, globe). Most
DGVMs consist of interacting biogeography, biogeochemistry, and fire modules. She highlighted a
couple of studies using DGVMs, one which examined vegetation changes in the United States and a
second which examined changes in global tree cover. She emphasized that both studies predict
fundamental and large-scale changes. Dr. Carney explained the limitations of DGVMs, including that
there is a significant amount of variability across models for the same region and climate scenario, with
results highly dependent on the GCM used. She noted additional limitations, including an absence of
most other anthropogenic factors, the assumption of no barriers to plant dispersal, and an absence of
pest and pathogenic influence. She noted that there are some general areas of agreement between
models and that scientists should look for these areas, perhaps averaging DGVM results, when possible.

Next, Dr. Carney explained that climate change will affect wildfire dynamics through direct (e.g., higher
temperatures, dryer fuels) and indirect (e.g., changes in vegetation type) mechanisms. She noted that
wildfire dynamics can be modeled using statistical models based on historic fire behavior, as well as
using the fire module of DGVMs. Dr. Carney presented the results of one study that predicts decreased
fire in northern Canada and Russia, and increased fire in the United States, central South America,
southern Africa, western China, and Australia. She explained that wildfire models can only roughly
approximate both historic and future wildfire dynamics, and that they are unable to predict the timing
and location of specific fires.

Finally, Dr. Carney discussed species extinctions, which are most commonly modeled using climate
envelope models. These models use current distributions of a species to construct climatic requirements
and then determine where species could live under future climate conditions. She noted that extinctions
are likely to occur, but that the results of these studies vary widely, with predicted extinctions ranging
from relatively low levels up to 60% of species. She noted several key uncertainties in climate envelope
models, including: that species may be flexible and able to survive in a wider range of climate conditions
than is predicted by their current range, that biotic interactions may be more important than climate in
determining species range, that dispersal is likely limited by habitat fragmentation, and that land use
change may amplify climate change impacts. She further noted that is difficult to value global

-------
biodiversity and that economic value is often tied to specific species or locations rather than global
extinctions.

Dr. Carney concluded with recommendations for future research needs. She suggested that methods
need to be developed to integrate results across studies and tools (e.g., meta-analyses, ensemble
means). She suggested a major need to develop large-scale, long term projections for changes in pest
outbreaks and interior wetland change and loss. She also noted the importance of understanding
changes in snow pack, particularly as related to ecosystems and recreation.

Estimating the Economic Impact of Climate Change on Forestry

After Dr. Carney's presentation, Dr. Brent Sohngen of the Department of Agricultural, Environmental,
and Development Economics at Ohio State University and a University Fellow for Resources for the
Future, presented on estimating the economic impact of climate change on forestry.

First Dr. Sohngen described the general process of measuring damages, which starts with future climate
scenarios and concludes with economic impacts. He noted that feedbacks and interactions between
different steps of the analysis are important and require additional research. Dr. Sohngen then
explained that both models and observations indicate increases in productivity due to: C02 fertilization,
warming in colder climates, and precipitation gains where water is limited. He noted that DGVMs
indicate limits to productivity gains and suggest ecosystems will change from a carbon sink to a carbon
source within the next several decades.

Dr. Sohngen presented results in the literature that predict a reduction in total U.S. ecosystem carbon,
with losses greatest in the eastern United States and under more recent climate scenarios. Without
accounting for adaptation, these ecosystem effects could result in emissions of up to 500 million t C per
year and a total loss over the century of 10-20 billion t C. He then presented regional estimates from the
literature on timber market results. He showed that timber output and consumer surplus is expected to
increase in almost all regions, but that producer returns only increase in about half of the regions.

Dr. Sohngen then presented preliminary results from an analysis that is currently underway. That study
incorporates several key factors into the economic analysis, including yield change, stock losses, and
area suitable for trees. It also incorporates adaptation options, including existing stock management by
changing rotations and salvage; replanting of new species if growing and economic conditions warrant
it; and future stock management by changing rotations, management, and investments. He showed that
that global output is expected to increase by 5-15% while global prices are expected to decrease by 5-
15%.

He explained that regional results suggest that there will be winners and losers, but that the allocation
of benefits and losses depend on the climate scenarios. He noted that Brazil, Canada, Russia, and
Oceania are likely to experience net benefits. Finally, he emphasized that the management of forest
stocks will be complicated by disturbance. He noted that large-scale disturbances are already influencing
outputs in many regions (e.g., mountain pine beetle outbreaks in Canada, forest fires in Russia) and that
disturbance patterns are expected to change with climate change. He noted that increases in
productivity are not expected to be able to counter falling global prices.

-------
Dr. Sohngen concluded by describing some of the study's limitations. He noted that timber markets may
not be most important demand on forestland in the future, that models are deterministic, and that
ecosystem models are calibrated without human influences. After the conclusion of his presentation, Dr.
Carney asked if crop shifting is incorporated into his model. She noted that if timber prices drop too low,
people may decide to use the land in other ways. Dr. Sohngen explained that this type of crop shifting is
partially incorporated.

Valuing Climate-associated Changes in Terrestrial Ecosystems and Ecosystem Services
Dr. Alan Krupnick provided the third and last presentation for the terrestrial ecosystems and forestry
impact category, on valuing the impacts of climate change on terrestrial ecosystem services. Dr.

Krupnick focused his comments on non-use values and stated preference studies. He noted that even a
low WTP per person can amount to significant totals.

Dr. Krupnick discussed the transition from natural science assessment to economic assessment, where
biophysical endpoints estimated by natural scientists are used as the starting points in valuation studies.
He explained a need for natural scientists to provide biophysical impacts assessment endpoints that
correspond to the items assessed in valuation exercises (valuation starting points), that people value
and care about, and that have functional relationships with climate drivers. He explained a parallel need
for economists to develop a consensus approach to classify endpoints to be used as valuation starting
points. He noted that natural scientists have identified large numbers of climate change impacts, from
which endpoints need to be identified. He further noted that economists have not been able to easily
define the things that matter from an economic perspective.

Dr. Krupnick explained that, when conducting stated preference studies, it is crucial to ask the right
questions. He noted that survey respondents should be asked to value biophysical outputs (e.g., number
of eagles), rather than biophysical inputs (e.g., number of acres of eagle habitat). He explained that
natural scientists should identify the production function that defines the relationship between inputs
and outputs. He also noted that it may be better to not mention climate change, particularly in U.S.
studies, as climate skeptics might provide biased answers. He questioned how best to admit
uncertainties in surveys without inducing protest bids.

Dr. Krupnick presented several examples of stated preference surveys where survey respondents are
given a set of options to choose from with a suite of associated conditions. He noted one study that
suggests the household monthly mean WTP for a 30% greenhouse gas reduction is $22 in Sweden, $17
in the United States, and $5 in China.

Dr. Krupnick classified starting points for climate change into four categories: use values; "standard"
non-use values; combinations associated with events or broad scale changes; and novel changes. He
then classified valuation studies into four categories: studies valuing relevant commodities in a non-
climate context; studies transferring non-climate values to a climate change context; studies directly
valuing relevant commodities in a climate change context; and stated preference top-down studies.

Dr. Krupnick went on to summarize and classify the literature using his set of starting points and survey
types. Dr. Krupnick noted that there is a broad range of existing studies falling into almost every

-------
combination of startpoint and survey type. He suggested these studies provide a lot of material for
meta-analyses and benefit-transfer. Dr. Krupnick noted the studies range widely in their spatial scale,
but that spatial specificity enhances credibility. He highlighted that scope sensitivity tests ensure WTP is
greater for avoiding larger damages or gaining larger benefits and that marginal returns decrease. He
noted that existing studies suggest timing of benefits is not significant, implying low or zero discount
rates. He explained that most studies assume certainty and very few vary uncertainty.

Dr. Krupnick noted that existing "non-climate" studies are useful but limited, that benefits transfers
studies are artificial and assumption-based, and that climate-driven studies are useful and growing in
number, but that they will always be location-specific and thus patchy. He noted that top-down studies
are tempting as they provide a broad coverage of endpoints and locations, but that they involve highly
imprecise commodity definitions and scenarios. He highlighted a need for holistic valuation estimates.

Discussion: Terrestrial Ecosystems and Forestry

After Dr. Krupnick's presentation, the terrestrial ecosystems and forestry discussion session
commenced. One participant noted the finding highlighted by Dr. Sohngen that forest productivity
would increase due to climate change. Since forests provide an important low-cost mitigation option,
she asked how this trend could be incorporated into mitigation costs in the SCC. Dr. Sohngen noted that
initial unpublished models suggest lower costs of carbon sequestration, but that it is a broad, uncertain
result.

Several participants and speakers discussed the usefulness of the concept of ecosystem services. Dr.
Krupnick expressed satisfaction that the concept had gained traction, as it does provide a bridge into the
economic sphere by using the term 'service'. However, he suggested it was only a starting point that
only partially overlaps with important endpoints lying underneath the services. Another participant
suggested that the literature does not provide good information on how climate change will impact
ecosystem services. He agreed with Dr. Krupnick that the concept has potential and begins to provide a
useful bridge. However, he suggested the concept had not gotten a lot of traction in policy making. He
suggested that the concept should continue to be pursued in a sensible way. Dr. Sohngen agreed with
the previous assessments. He added that the economic drivers for management and adaptation of
timber markers seem to be decreasing, suggesting it is more compelling to consider their ecosystem
services. Dr. Carney suggested that the concept of ecosystem services, while perhaps imperfect, is still
useful. She explained that ecosystem services provide a way to translate ecological effects into changes
that are important to individuals in a policy context.

Another series of comments focused on the language of stated preference surveys. Dr. Krupnick
explained that a tax is frequently used as a vehicle in surveys but that the standard practice is to try to
present a hypothetical real choice that has real costs. He emphasized that stated preference studies are
not attitude surveys, and that responses should be limited by income and choices should be binding. He
noted that surveys are aimed at estimating the individual willingness to pay. He added that studies are
constructed to eliminate the possibility of "free riding" and to incorporate the effect that one individual
paying in the absence of other contributions would have no effect.

-------
In response to a final question, Dr. Sohngen explained that models do, at least partially, incorporate
country variables (e.g., poverty) as timber production shifts across political borders. He explained that
models incorporate different production costs (e.g., labor costs), management structures, species uses,
and prices. He suggested the extent of incorporation may not be sufficient or perfect.

Energy Production and Consumption

The eighth session covered the impacts and damages to energy production and consumption. The
session was moderated by Dr. Stephanie Waldhoff of EPA and included presentations by Dr. Howard
Gruenspecht, U.S. Energy Information Administration; and Dr. Jayant Sathaye, Lawrence Berkeley
National Laboratory (LBNL).

U.S. Energy Production and Consumption Impacts of Climate Change

As the first speaker for the Energy Production and Consumption Impact Category, Dr. Howard
Gruenspecht of the U.S. Energy Information Administration discussed the energy system impacts of
climate change. He noted that climate change impacts on energy systems have received considerable
attention, despite high-profile reports finding that the impacts will be modest.

First, Dr. Gruenspecht presented climate change impacts on energy demand for space heating and
cooling. He noted that the United States is a relatively cold country, where the amount of energy used
for heating is three to four times as great as the amount used for cooling. He noted that this gap is even
greater in other industrialized countries. He further noted that energy use for space conditioning is
highly tied to development. Dr. Gruenspecht explained that the details of warming are very important in
considering energy impacts. This includes the latitudinal, diurnal, and seasonal gradients. He explained
that space conditioning is subject to thresholds and that measures of comfort produce very different
impact estimates than measures of energy expenditures. Finally, Dr. Gruenspecht noted the importance
of incorporating technology changes over relevant time horizons. Historic increases in cooling efficiency
had significant impacts, and new technologies such as smart grid will likely have similar impacts.

Dr. Gruenspecht noted that the literature has focused on energy demands for space conditioning but
that other areas of energy demand merit additional attention. He highlighted the energy-water nexus,
since climate change stresses traditional water sources. He showed that non-traditional sources such as
desalinized water require significant amounts of energy.

Next, Dr. Gruenspecht presented climate change impacts on energy supply. He noted impacts on access
to traditional resources, including hydroelectricity's sensitivity to melting glaciers and arctic oil
infrastructure's sensitivity to melting permafrost. He further noted the need for cool water and air to
maintain power plant operation. However, Dr. Gruenspecht emphasized his feeling that too much
attention has been placed on energy issues, which may not be quantitatively important in overall
effects, particularly after mitigation and adaptation are considered.

Finally, Dr. Gruenspecht discussed the impacts of climate change on non-traditional energy sources. He
noted the very significant effects of cloud cover and aerosols on solar power, the unclear changes in
wind patterns that will affect wind power, and the agricultural effects on biomass.

-------
Dr. Gruenspecht concluded that energy impacts may be beneficial for small to modest climate change,
but dominated by negative impacts in the long-run. He emphasized that details are crucial in modeling
impacts and that changes must be considered in the context of adaptation and technology change. He
suggested the importance of distinguishing between energy system impacts, which are important to
energy planners, and energy-system-related welfare impacts, which are important for cost-benefit
analysis of climate change policies.

Impacts of Climate Change on Global Energy Production and Consumption

Following Dr. Gruenspecht, Dr. Jayant Sathaye of LBNL presented the impacts of climate change on
global energy production and consumption. He started by presenting a list of over a dozen hydro-
meteorological and climate parameters that each have numerous effects on energy demand and supply.

Dr. Sathaye then presented a selected review of international impact analyses in the literature. He noted
that most of the literature focuses on energy demand, as opposed to energy supply. The literature
indicates that global reductions in energy demand for heating will be greater than global increases in
energy demand for cooling. For example, the POLES model estimates 200-300 million tons of oil
equivalent (Mtoe)reductions in heating demand compared to 60-130 Mtoe increases in cooling demand.
The literature indicates that global nuclear generation will decline, while hydroelectricity generation
may increase or decrease depending on the scenario (more likely increase). Dr. Sathaye also presented
examples of international studies at the national and regional scale.

Next, Dr. Sathaye presented an example of a study conducted in California to demonstrate the data and
information needed to conduct an energy impact analysis. He explained that the study, funded by the
California Energy Commission, focuses primarily on three impacts: increased temperature impacts on
electricity capacity and demand; sea level rise impacts on energy infrastructure; and wildfire impacts on
energy infrastructure. He presented the intricate flow chart of analysis stages, commencing with
AOGCM emission scenarios and culminating in a summary of damages.

Dr. Sathaye then presented results from the study. He explained that warming temperatures may lead
to both losses of up to 4,000 megawatts (4%) of available natural gas-fired power plant capacity, as well
as increases in peak load cooling demand of 20%. He noted that the combined effect of changes in
demand and supply result in a 24% gap between energy supply and demand that needs to be addressed.

Dr. Sathaye presented the maps of the wildfire analysis, which involved identifying the climate factors
affecting wildfires, overlaying transmission lines on near-term and long-term spatial models of wildfire
probability, and quantifying the length of transmission lines exposed to wildfires under modeled future
climate scenarios. Dr. Sathaye explained a similar analysis for sea level rise, which concluded that a 1.4
meter projected rise in sea levels would affect 25 power plants and approximately 90 substations.

Dr. Sathaye concluded that there is a general lack of quantitatively-based impacts information for the
energy sector, but that the base of international literature is growing. He reiterated global projections of
larger decreases in heating demand compared to increases in cooling demand. He noted that the
temperature impact on demand is much higher than on supply infrastructure and that the impact of
wildfires could potentially be significantly high. Finally, he suggested that more data and research are

-------
needed to evaluate wildfire and sea level rise impacts on power sector infrastructure and temperature
impacts on electricity transmission and distribution.

Discussion: Energy Production and Consumption

During the question and answer session, one participant again raised the need to incorporate
interactions and double-counting across sectors, highlighting the intersection of health impacts driven
by temperature with impacts on cooling demand. Another participant noted that an impact in one
sector might be an adaptation in another. Dr. Gruenspecht added that there are significant impacts from
adaptation, technology, and efficiency that must be considered. Dr. Sathaye agreed, noting the need to
develop a long-term scenario of future infrastructure possibilities and combine that scenario with
climate data.

Another participant asked how cooling penetrates lower socio-economic classes, noting that middle
class and poor country adoption of cooling greatly determines international impact. Dr. Sathaye agreed
with the importance of these effects. He noted that the air conditioning load in India has been
increasing annually by 25%. He suggested that similar changes are occurring elsewhere in developing
countries.

A third participant asked about distinguishing between costs of damages and costs of reducing risks,
noting that the costs of reducing risks are often significantly lower than costs of damages. Dr. Sathaye
agreed that this distinction is critical and should be reflected in the cost analysis. Dr. Gruenspecht also
agreed, emphasizing that the future must be considered in the context of technology change. He
acknowledged the extreme difficulty in attempting to predict the 100 year future, but emphasized its
necessity.

During the discussion session, both speakers emphasized a need for more and better climate data,
noting the need for information on things like cloud cover. One participant suggested that economists
need to move forward with the data available now, since some aspects of physical climate change are
going to be difficult to estimate more accurately anytime soon. Dr. Gruenspecht acknowledged the
validity of her point but suggested that there is a middle ground where climate scientists might be able
to provide more than what is provided now, but not everything desired by economists. For example, he
suggested it would be helpful to have information on cloud cover on a global average scale. Dr. Sathaye
agreed, noting that global average numbers provide a sense of the underlying information. Another
participant argued that global average numbers are enormously insufficient and could do more harm
than good when considering spatially specific investments and activities related to cloud cover and wind
patterns. Yet another participant challenged the community to do better. The first participant suggested
that economists need to lower their expectations. She explained that global average cloud cover is the
greatest uncertainty in models. She suggested a need to make decisions under uncertainty. Another
participant suggested it would be helpful to put bounds on the uncertainty with factors such as this.

Finally, one participant asked if heat waves and blackouts are incorporated in models. One of Dr.
Sathaye's colleagues explained that the California study did incorporate the effect of heat waves, but did
not include the costs of blackouts. The participant suggested that this would affect the overall

-------
conclusion related to heating and cooling demand. Dr. Gruenspecht reemphasized the distinction
between energy impacts and welfare impacts.

Socio-economic and Geopolitical Impacts

The ninth session covered the socio-economic and geopolitical impacts and damages. The session was
moderated by Dr. Alex Marten of EPA and included presentations by Dr. Nils Petter Gleditsch, Peace
Research Institute Oslo; and Dr. Robert McLeman, University of Ottawa.

Regional Conflict and Climate Change

Dr. Nils Petter Gleditsch of the Centre for the Study of Civil War, the Peace Research Institute Oslo, and
the Department of Sociology and Political Science at the Norwegian University of Science and
Technology commenced the last impact session with his presentation on regional conflict and climate
change. Dr. Gleditsch is an expert on conflict. During his presentation and through his abstract, Dr.
Gleditsch indicated that the policy debate is running well ahead of its academic foundation, and
sometimes even contrary to the best evidence.

First, Dr. Gleditsch presented current trends in armed conflicts and number of deaths. He explained that
the world is moving towards a liberal peace - as democracy and trade increase worldwide, conflict
becomes less likely. This movement includes increases in the number of international governmental
organizations (IGOs), in democracy, in wealth, and in trade. He noted four possible threats to the liberal
peace: shifting patterns of power, the financial crisis, fundamentalist religion, and climate change. He
noted that climate change is arguably the most serious threat, highlighting numerous statements from
non-governmental organizations, politicians, and some academics indicating climate change is a major
issue that will greatly impact conflict. Despite the rhetoric, however, there is little systematic evidence
to date that long-term climate change or short-term climate variability has had any observable effects
on the pattern of conflict at any level. Dr. Gleditsch then presented a flowchart from the World Bank
that presents numerous possible pathways that lead from climate change to conflict. He showed that
natural disasters, migration caused by sea level rise or other climate factors, and increasing resource
scarcity may all promote conflict.

Next, Dr. Gleditsch presented numerous, sometimes contradictory, findings from the literature
regarding the influence of climate factors on conflict. To date there is little published systematic
research on the security implications of climate change. The few studies that do exist are inconclusive,
most often finding no effect or only a low effect of climate variability and climate change. The scenarios
summarized by the Inter-Governmental Panel on Climate Change (IPCC) are much less certain in terms
of the social implications than the conclusions about the physical implications of climate change, and the
few statements on the security implications found in the IPCC reports are largely based on outdated or
irrelevant sources.

Dr. Gleditsch presented evidence regarding the effects of precipitation, temperature, sea level change,
and natural disasters. He noted that millions of people may become refugees due to sea level rise. He
also noted that natural disasters may reduce conflict as people tend to unite in the face of adversity. Dr.
Gleditsch discussed the economic effects of climate change, noting that economic factors are important

-------
in conflict. He explained that economic interdependence and economic development limit inter- and
intra-state conflict, respectively; but that economic decline could reverse this.

Dr. Gleditsch presented arguments and counterarguments for several climate change impacts on
interstate conflict. He suggested increased scarcity may or may not lead to interstate conflict. He also
explained that climate change will open up new trade routes and new ocean territories. He noted that
uncertainty about ownership and competition for exploiting these resources may or may not promote
conflict. He suggested that climate change may affect where nations fight, rather than whether or when.

Dr. Gleditsch described methods analyzing the scarcity theorem, highlighting several criticisms of past
studies. He highlighted the interactions of climate change with other factors, such as poverty, poor
governance, and ethnic dominance, suggesting that climate change may act as a threat multiplier and
destabilize conflict-prone regions. He suggested that, from a policy perspective, it is useful to examine
whether it is easier to reduce climate change or other factors in the interaction. Dr. Gleditsch presented
a map of the distribution of armed conflict, highlighting Africa, East Asia, and Central and South Asia as
particularly vulnerable regions.

Finally, Dr. Gleditsch presented a list of research priorities. He suggested that future research needs to
look at interactions between climate change and political and economic factors, to focus on countries
with low adaptive capacity, to examine a broader set of conflicts, to conduct disaggregated studies of
geo-referenced data, to balance negative and positive effects of climate change, and possibly to couple
models of climate change to models of conflict. Dr. Gleditsch suggested that if climate change has
negligible impacts on conflict, it matters significantly for the credibility of climate change research, very
little for mitigation, and possible a lot for adaptation.

After the conclusion of his presentation, Dr. Gleditsch agreed with one participant's concern that studies
of historic conflict may not inform the effects of unprecedented changes in climate. Another participant
asked if there was evidence for conflict in small islands, which are particularly vulnerable to sea level
rise. Dr. Gleditsch explained that there is not a lot of conflict in those areas, and that migration and
security concerns will more likely result from climate change, than conflict.

Migration Impacts of Climate Change

Following Dr. Gleditsch, Dr. Robert McLeman of the University of Ottawa's Department of Geography
presented the migration impacts of climate change. Dr. McLeman began with an overview of climate
change-caused migration. He noted that the media has already identified the first climate change
refugees, including those from Shishmaref, Alaska; the Cataret Islands, and the Lake Chad region.

Dr. McLeman provided a range of estimates for the number of future environmental refugees, ranging
from 50 million refugees by 2100 to 1 billion refugees by 2050. He noted that predictions are based on
identifying areas and populations exposed to negative climate change impacts. However, he noted that
exposure does not equate to migration, climate-migration does not result from a simple stimulus-
response, and there are numerous intervening socio-economic, cultural, and institutional factors. All of
these caveats affect the accuracy of the estimates.

-------
Dr. McLeman explained that migration may be caused by sudden onset events (e.g., hurricanes),
persistent conditions (e.g., drought), or other stimuli. He noted that one of the earliest groups of climate
change migrants will be oil workers migrating to the arctic. Dr. McLeman explained that climate change
will generate migration stimuli nearly everywhere people live, including the arctic, high latitudes, wet
tropics, mid-to low-latitudes, dry tropics, coastal plains, deltas, and small islands.

Dr. McLeman explained that climate events and conditions do not always stimulate migration and that
multiple migration outcomes can be generated by a single climate event (e.g., brief evacuation,
extended leave, permanent migration, new arrivals). He presented data from Hurricanes Katrina and
Mitch that inform ensuing migration patterns. He noted one study that shows a 10% decrease in
agricultural production in Mexico due to drought is associated with a 2% rise in Mexican-U.S. migration.
Dr. McLeman explained that migration is one of a range of potential adaptive responses to
environmental stress. Migration is used in many parts of world, is typically initiated by households, is not
available to everyone, is not always used, and, in the worst case, could be the only adaptation option.

Dr. McLeman explained that vulnerability is a function of exposure, sensitivity, and adaptive capacity. He
noted that migration changes the composition of the population left behind, which in turn changes the
area's adaptive capacity. He further noted that migration is motivated by numerous non-climate factors
(e.g., opportunity-seeking, cultural norms, lifestyle, love, persecution), with which climate interacts. He
explained that most observed climate-related migration is not conflict-related, is internal or intra-
regional, and generally follows established routes or transnational communities when international.

Dr. McLeman described numerous climate-migration models, including examples of each. Models
include: historical climate-migration models, spatial vulnerability models, multi-level hazard analysis
models, multi-stage regression models, and agent-based models. As part of one of the examples, he
explained that migrants tend to be young, healthy, skilled, educated members of the middle class with
uncertain land tenure and family ties elsewhere. Meanwhile, those less likely to migrate include
wealthier classes, landowners, owners of fixed assets, those with strong local social networks, the poor
and destitute, the elderly, the infirm, or those with broken families.

Dr. McLeman concluded with a list of challenges and opportunities. He noted many challenges related to
a lack of data availability and reliability, including the lack of a single global database, fragmented data,
and data missing reasons for migration. He noted other challenges including understanding system
linkages and the role of intervening variables, as well as uncertainty about future climatic stimuli. Dr.
McLeman listed three opportunities: to develop monitoring and data collection protocols, to enhance
empirical research into environment and migration linkages, and to develop and improve migration
models as climate change models improve.

Discussion: Socio-economic and Geopolitical Impacts

During the question and answer session, one participant highlighted the work of Robert Bates, which
uses a different approach than described by Dr. Gleditsch to examine conflict. Dr. Gleditsch commented
that he thought adding climate variables to Dr. Bates work would produce similar results to those he
discussed.

-------
Another participant asked whether climate change detection and attribution would affect the result that
people unite in the face of natural disasters. He asked whether the existence of human cause or blame
would affect the potential for conflict. Dr. Gleditsch clarified that the observation that people unite in
the face of adversity does not only apply to natural disasters, but includes human-induced disasters such
as bombings. He suggested that results may be different if a population's own government was
responsible for the climate change. Dr. McLeman added that climate change adaptation planning was
actually a fairly effective way to get otherwise quarreling parties to collaborate.

Dr. Gleditsch agreed with a third participant that climate conflict models should be focused on multiple
stressors rather than climate as a solitary force. He noted that there has been some work in this area
and reemphasized the notion of analyzing whether it is easier to address the issue by changing the
climate variable or the other variables.

In response to another participant, Dr. McLeman acknowledged that he overlooked the effects of
climate change on amenity migrants during his presentation. He agreed that climate change would
affect the places to which affluent and retired people migrate.

A final participant asked whether the literature has examined the interaction between climate change,
energy markets, and conflict and migration. Dr. Gleditsch reiterated the importance of resource scarcity
in climate change. He suggested that there could be a benefit from a reduction in oil dependence and oil
prices. Dr. McLeman noted that there may be an effect on energy markets from predicted rural-to-urban
migration. He explained that rural residents tend to have a smaller energy footprint than urban
residents, so that increased urbanization will lead to increased energy demand.

Panel Discussion: Incorporating Research on Climate Change Impacts into
Integrated Assessment Modeling

Following the impact-specific sessions, a five-member panel discussed the incorporation of research on
climate change impacts into integrated assessment modeling. The panel discussion was moderated by
Dr. Elizabeth Kopits of EPA and included Dr. David Anthoff, University of California, Berkeley; Dr. Tony
Janetos, Joint Global Change Research Institute, Pacific Northwest National Laboratory (PNNL); Dr.
Robert Mendelsohn, Yale University; Dr. Cynthia Rosenzweig, NASA Goddard Institute for Space Studies;
and Dr. Gary Yohe, Wesleyan University. The panel discussion started with comments from each of the
panelist members and concluded with questions from the audience. Dr. Kopits framed the discussion by
asking the panelists whether there was any hope in improving lAMs or whether it was only possible to
outline a long-term research agenda.

David Anthoff, University of California, Berkeley

Dr. David Anthoff of the University of California at Berkeley, who works on the FUND model with Richard
Tol, commenced the discussion. Dr. Anthoff reflected on each of the nine impact categories as
presented by the workshop speakers and reflected on how well the state of the literature is
incorporated into lAMs (specifically FUND). He noted that his comments would merely reflect how well
the literature is reflected in FUND, without assessing the state of the primary research itself. He further

-------
qualified his comments by noting that they simply reflect his impressions from listening to the two days
of presentations.

Dr. Anthoff suggested that FUND does a decent job incorporating the research for storms, water, sea
level rise, forestry, and energy demand. He noted that Dr. Cropper's suggestion regarding the income
elasticity for health impacts could be investigated fairly simply in the short-term. He suggested that the
primary literature for agriculture seemed contradictory and does not provide the aggregated numbers
necessary for 1AM incorporation. He noted the difficulties associated with translating research on
individual crops into the models. Dr. Anthoff noted that ocean acidification is not incorporated in any of
the three models. He suggested that progress could be made to incorporate ocean acidification in the
mid-term. Dr. Anthoff noted that FUND incorporates biodiversity loss, but that the primary research is
rough. He noted that while energy demand is incorporated in lAMs, energy supply is not. He suggested
the possibility of incorporating conflict is very far off. He noted that FUND incorporates a very simple
migration model for sea level rise, but that other causes of migration are not incorporated.

Dr. Anthoff then suggested that primary researchers need to evaluate the lAMs to assess how well the
data sources, damage functions, and outputs reflect the primary literature for each of the impact
categories. He noted that uncertainty and extreme impacts are critical in lAMs, but were not discussed
much during the workshop sessions. Lastly, Dr. Anthoff remarked that lAMs are severely understaffed
and underfunded, particularly as compared to GCMs.

Tony Janetos, Joint Global Change Research Institute, Pacific Northwest National Laboratory

Next, Dr. Anthony C. Janetos of the Joint Global Change Research Institute suggested that there are
many possibilities for improving lAMs based on the workshop presentations, noting that the physical
impacts research seems to have advanced more than the valuation research. However, he suggested
that very few of the advancements are readily incorporated into lAMs. He noted a need for additional
understanding of thresholds, non-linear behavior, and process-level understanding. He further indicated
a need to model interactions between sectors with an explicit representation of the sectors themselves,
as well as the economic and physical factors (e.g. competition for water and land) that connect them.

Dr. Janetos identified several reasons that limit the generation of good central estimates of physical and
economic parameters, which he noted are necessary for SCC development. First he cited the non-
linearity and thresholds that pervade physical systems. He noted that some thresholds are not
necessarily attributable to anthropogenic changes (e.g., climate changes driving pine beetle
infestations). Dr. Janetos suggested a need to improve knowledge of the reference case, noting that the
major drivers of big changes over the past half-century are human-driven (e.g., land-cover changes).

Dr. Janetos emphasized the importance of interactions among sectors, which he emphasized is a first-
order problem. He explained that competition for water among various human uses and ecosystem uses
is just the tip of the iceberg and not particularly well understood. He noted that aggregation and
disaggregation issues are extremely important, which is a challenge for the response-surface approach.

Dr. Janetos then enumerated well-known deficiencies in the ecological models. For example, in the
Vegetation/Ecosystem Modeling and Analysis Program (VEMAP), when all major ecosystem models

-------
were driven by same factors, they diverged. He noted that there has not been a subsequent
reconciliation of that divergence. Dr. Janetos noted other deficiencies: ecological models typical do not
include threshold responses; they underplay or omit biotic interactions like pests and pathogens; and
DGVMs are largely unverified and potentially unverifiable.

Dr. Janetos suggested that it is useful and important, while difficult, to infer or develop statistically- or
model-based response functions for use in reduced form lAMs. He noted that current damage functions
are not robust beyond the ranges for which they were originally designed, and suggested that a process-
based approach might be useful. He emphasized that uncertainty and error bars must be well
characterized, noting that lAMs are better at doing this than the impacts community.

Robert Mendelsohn, Yale University

Dr. Robert Mendelsohn of Yale University shared brief remarks following Dr. Janetos. He noted that
lAMs are not able to capture the level of detail available from climate modeling, ecological impact
assessment, and damage assessment. He suggested some concern regarding the lack of connection
between detailed impact studies and lAMs, however, he noted this lack of connection does not
necessarily mean the lAMs are biased.

Dr. Mendelsohn emphasized the absolute necessity for studies to include adaptation. He highlighted
that lAMs are interested in the actual damages of climate change, not the potential damages. He
explained that significant adaptation will be implemented and models must acknowledge it. Next, Dr.
Mendelsohn noted that the workshop seemed to be missing any discussion of catastrophic events and
tipping points.

Next, Dr. Mendelsohn emphasized that the community should not be disheartened about lAMs or
damage estimates. He emphasized that lAMs do a good job, in general, and that a lot of progress has
been made over the last 20 years. He noted that natural science, ecosystem models, and economic
models are all improving steadily, especially for short-term predictions. He acknowledged that long-term
predictions are more difficult. He suggested a need for a third generation 1AM to address spatial detail.

Finally, Dr. Mendelsohn suggested that the near-term agenda should be focused on capturing damage
assessment work within impact models, so that lAMs can incorporate all current knowledge.

Cynthia Rosenzweig, National Aeronautics and Space Administration

Next, Dr. Cynthia Rosenzweig of NASA discussed three points and proposed a way forward.

First, Dr. Rosenzweig discussed the impacts, adaptation, and vulnerability (IAV) component of impacts
assessment and valuation. She noted that adaptation has been severely underfunded but has been
getting increased attention recently. She acknowledged a need to improve the biological, physical, and
social science of impacts, as impacts research is much less advanced than climate science and has real
effects on society. She highlighted an eagerness to work with and improve lAMs, but noted the great
difficulty in doing so.

-------
Second, Dr. Rosenzweig discussed the economic components of impacts assessment and valuation. She
suggested that current work (e.g., SCC) is focused on justifying mitigation action. She suggested the
need for an adaptation lens in economics work, and even analysis of the balance of resource allocation
between adaptation and mitigation. She questioned whether lAMs are capable of addressing all three
questions. She suggested a need to understand the economic underpinnings of adaptation, to better
understand state changes arising from incremental and marginal changes, and to better address equity
and environmental justice issues.

Third, Dr. Rosenzweig discussed integration of scales, of mitigation and adaptation, and of sectors. She
noted that urban areas are where all sectors are integrated. She suggested climate change assessment
in cities be conducted.

Finally, Dr. Rosenzweig suggested a need for on-going trans-disciplinary groups to work to improve basic
research and translation. She highlighted EMF-24, OCMIP, and VEMAP as examples of trans-disciplinary
efforts aimed at creating processes and structures for progress. She suggested collaboration with the
National Climate Assessment and with international impacts efforts. She highlighted a movement to
coordinate IAV scientists behind research questions. She noted the United Nations Environment
Programme (UNEP) Programme of Research on Climate Change Vulnerability, Impacts and Adaptation
(PRO-VIA), a new organization aimed at setting research questions and directions.

Gary Yohe, Wesleyan University

Finally, Dr. Gary Yohe of Wesleyan University shared his comments. He noted that his comments serve
as an outline of the more complete paper he wrote to address the charge questions.

First, Dr. Yohe suggested a need for humility regarding our confidence estimates of the research. He
emphasized a need to identify uncertainty issues.

Second, Dr. Yohe suggested a possible Type 3 Error in assessing economic impacts from climate change
to build the SCC, cautioning scientists and economists not to spend time addressing the wrong issues,
with little value added. He discussed his use of PAGE with Chris Hope to do a Monte Carlo analysis of
probabilities with a range of different parameter assumptions. The analysis concluded that, in PAGE,
differences in damage estimates were not as important as other variables such as time preference, risk
aversion, etc.

Third, Dr. Yohe instead suggested an alternative approach for estimating benefits of marginal reductions
in emissions, with higher value added. He suggested that an iterative process be built to set a target and
work towards a shadow price. First, he suggested using an assessment of climate risk to determine the
long-term objective and medium-term climate budget. Second, he suggested the U.S. contribution to
this budget could be determined, working within the political process. Third, the results from this
analysis could be used to price carbon for non-climate policy needs. Within this process, lAMs would be
used to check the reasonableness of the assessment, to design cost-minimizing approaches (including
net economic damages), and to highlight areas where adaptation in economic sectors will be most
productive.

-------
Panel Discussion

Following remarks from the five panelists, the panel discussed questions from the audience. One
participant asked what detail is needed, what uncertainty is important to characterize, and what factors
most influence the results in lAMs. She noted the orders of magnitude difference resulting from
carefully conducted impact analyses. Dr. Janetos agreed that modelers must identify which complexity is
important to include. He noted structures arising to address this question, including validation studies
and a process-level understanding of the individual sectors. Dr. Anthoff noted that exploring relative
importance is a key strength of lAMs. He noted that lAMs can use ranges and limits from the impact
community as inputs, to determine how much the SCC reacts to a full range of inputs from a single
sector.

Another participant questioned the interaction of high non-use and non-market values with the imposed
limit that damages cannot be more than GDP. A third participant underscored a couple of Dr. Anthoff s
comments. He emphasized the importance of the tails of impacts (as opposed to means, medians) to
policy makers. He highlighted the need for impact studies beyond 2100. He emphasized the small size of
the IA community. He noted that a lot of the community's time is spent on discussing their work at
meetings like this workshop, which limits time available to do the work. Dr. Rosenzweig expressed her
hope that by expanding the community that is working on rigorously comparing models, they will be
able to work with and help integrated assessment modelers by providing more rigorous estimates.

Closing Remarks

The workshop concluded with closing remarks from Dr. Rick Duke, Deputy Assistant Secretary for
Climate Policy at DOE and Dr. Al McGartland, Director of the National Center for Environmental
Economics at EPA.

Summary Comments by U.S. Department of Energy

First, Dr. Rick Duke, the DOE Deputy Assistant Secretary for Climate Policy, thanked the participants for
attending, particularly those that braved the weather on Day 1. He expressed his appreciation of the
great conversation between natural scientists and economists, noting that he was struck by Dr.

Anthoff s desire to engage natural scientists to review economists' work on impacts.

Dr. Duke again noted that the workshop grew out of the interagency SCC work, which has since been
used in rulemaking. He acknowledged that the SCC values have numerous limitations that need to be
addressed, some beyond the scope of these workshops. He outlined a challenge to the community on
two timescales: to help to make better regulatory decisions in the near-term and to promote research to
improve assessment and valuation in the long-term.

Dr. Duke highlighted the need to evaluate the impacts of higher temperature outcomes, as well as
median outcomes. He noted that climate policy is much like insurance policy - a primary goal is to
reduce the consequences of particularly unfavorable states of the world (e.g., high climate sensitivities)
as well as to reduce expected losses. He emphasized the importance of evaluating the more significant
outcomes given the major challenges in achieving planned mitigation.

He closed by thanking the presenters, the broader research community, and DOE's partner, EPA.

-------
Summary Comments by U.S. Environmental Protection Agency

Finally, Dr. Al McGartland, Office Director for EPA's NCEE, extended both personal and EPA thanks to the
participants for attending despite the inclement weather. He said that he intended to finish the
conference with a ray of hope.

First, he noted that due to the field's interdisciplinary nature, everyone in the community must stretch
to accommodate other groups. He emphasized that policy institutions have to stretch as well. Dr.
McGartland noted that the SCC process was aimed at developing a set of numbers and asked if the right
questions are being asked. He noted that the process is not aimed at legislation or the next Kyoto
Protocol. Rather, the process seeks a shadow price so that EPA and DOE can incorporate the benefits of
carbon reduction in any rule affecting carbon emissions.

Next, Dr. McGartland highlighted the significant progress that has been made in risk assessment since
the work on particulate matter, lead, and pesticides in the 1980s. He suggested that simply duplicating
the historic rate of progress in this area would be great. He noted that EPA's long-term strategy is
dominated by regulatory work in areas where there are large net benefits.

Looking forward, Dr. McGartland highlighted a number of good points from the workshop. He
emphasized the need to address interactions among sectors. Finally, he highlighted his commitment to
move forward with the SCC using a transparent process.

52

-------
Improving the Assessment and Valuation of
Climate Change Impacts for Policy
arid Regulatory Analysis

Sponsored by

A J\ United Siaitis
I™^¦nviranminr,,:'1

Protection Afloncy

U.S. DEPARTMENT Of

IP ENERGY

January 27-28, 2011 Capital Hilton, Washington, DC

Workshop Agenda

Federal Meeting Room

RESEARCH ON CLIMATE CHANGE IMPACTS
AND ASSOCIATED ECONOMIC DAMAGES

DAY 1

Workshop Introduction

8:55-9:00 Welcome

Elizabeth Kopits, U.S. Environmental Protection Agency

9:00-9:20 Opening Remarks - Progress in estimating climate change
impacts

Michael Oppenheimer, Princeton University

9:20 - 9:40 Opening Remarks - Progress in valuing climate damages
William Cline, Peterson Institute for International Economics

9:40 - 9:45 Questions

Sessions covering research on various impact categories:

Storms and Other Extreme Weather Events

Moderator: Alex Marten, U.S. Environmental Protection Agency

9:45 - 10:05 Impact of Climate Change on Storms and Other Extreme
Weather Events

Tom Knutson, National Oceanic and Atmospheric Administration

10:10 - 10:30 Global Damages from Storms and Other Extreme Weather
Events

Robert Mendelsohn, Yale University
10:35 - 10:55 Open Facilitated Discussion

10:55-11:05 Break

-------
Water Resources

Moderator: Robert Kopp, U.S. Department of Energy

11:05 - 11:25 Hydrological/Water Resource Impacts of Climate Change

Ken Strzepek, University of Colorado, Boulder, and
Massachusetts Institute of Technology

11:25 - 11:45 Estimating the Economic Impact of Changes in Water
Availability

Brian Hurd, New Mexico State University
11:45 - 12:10 Open Facilitated Discussion

12:10-1:00 Lunch

Agriculture

Moderator: Charles Griffiths, U.S. Environmental Protection Agency

1:00- 1:20 Biophysical Responses of Agro-ecosystems to Climate
Change

Cynthia Rosenzweig, NASA Goddard Institute for Space Studies

1:20-1:40 Estimating the Economic Impact of Climate Change in the
Agricultural Sector

Wolfram Schlenker, Columbia University

1:40 - 2:20 Open Facilitated Discussion
Human Health

Moderator: Charles Griffiths, U.S. Environmental Protection Agency

2:20 - 2:40 Climate-Associated Changes in Health Conditions/Diseases
and Air Pollution

Kristie Ebi, Carnegie Institution for Science

2:40 - 3:00 Estimating the Economic Value of Health Impacts of Climate
Change

Maureen Cropper, Resources for the Future and University of
Maryland, College Park

3:00 - 3:40 Open Facilitated Discussion

3:40-3:50 Break

2

-------
Sea Level Rise

Moderator: Robert Kopp, U.S. Department of Energy

3:50-4:10 Sea Level Impacts of Climate Change

Robert Nicholls, University of Southampton

4:10-4:30 Estimating the Economic Impact of Sea Level Rise

Robert Tol, Economic and Social Research Institute

4:30 - 5:10 Open Facilitated Discussion

3

-------
• DAY 2

Day 2 Introduction

8:30-8:40	Welcome; Recap of Day 1; Overview of Day 2

Elizabeth Kopits, U.S. Environmental Protection Agency

Impacts Sessions Continued:

Marine Ecosystems and Resources

Moderator: Chris Moore, U.S. Environmental Protection Agency

8:40 - 9:00 Modeling Climate and Ocean Acidification Impacts on Ocean
Biogeoch em istry

Sarah Cooley, Woods Hole Oceanographic Institute

9:00-9:20 Modeling Climate and Acidification Impacts on Fisheries
and Aquaculture

Paul McElhany, National Oceanic and Atmospheric Administration

9:20-9:40 Economic Impact of Climate Change and Ocean Acidification
on Fisheries

David Finnoff, University of Wyoming

9:40 - 10:00 Non-market Valuation of Climate and Acidification Impacts
on Marine Resources

John Whitehead, Appalachian State University
10:00-10:10 Break
10:10 - 10:50 Open Facilitated Discussion
Terrestrial Ecosystems and Forestry

Moderator: Steve Newbold, U.S. Environmental Protection Agency

10:50 - 11:10 Biological Responses of Terrestrial Ecosystems to Climate
Change

Karen Carney, Stratus Consulting

11:10 - 11:30 Estimating the Economic Impact of Climate Change on
Forestry

Brent Sohngen, Ohio State University

11:30 - 11:50 Valuing Climate-associated Changes in Terrestrial
Ecosystems and Ecosystem Services

Alan Krupnick, Resources for the Future

11:50 - 12:30 Open Facilitated Discussion

4

-------
12:30- 1:30 Lunch

Energy Production and Consumption

Moderator: Stephanie Waldhoff, U.S. Environmental Protection Agency

1:30- 1:50 U.S. Energy Production and Consumption Impacts of Climate
Change

Howard Gruenspecht, U.S. Energy Information Administration

1:50 - 2:10 Impacts of Climate Change on Global Energy Production and
Consumption

Jayant Sathaye, Lawrence Berkeley National Laboratory
2:10-2:50 Open Facilitated Discussion

Socio-economic and Geopolitical Impacts

Moderator: Alex Marten, U.S. Environmental Protection Agency

2:50-3:10 Regional Conflict and Climate Change

Nils Petter Gleditsch, Peace Research Institute Oslo

3:10 - 3:30 Migration Impacts of Climate Change

Robert McLeman, University of Ottawa

3:30-3:40 Break

3:40 - 4:20 Open Facilitated Discussion

Panel Discussion: Incorporating Research on Climate Change Impacts into
Integrated Assessment Modeling

4:20-5:20	Moderator: Elizabeth Kopits, U.S. Environmental Protection

Agency

Panelists:

•	David Anthoff, University of California, Berkeley

•	Anthony Janetos, Joint Global Change Research Institute,
Pacific Northwest National Laboratory

•	Robert Mendelsohn, Yale University

5

-------
•	Cynthia Rosenzweig, NASA Goddard Institute for Space
Studies

•	Gary Yohe, Wesleyan University

Closing Remarks

5:20-5:25 Summary Comments by U.S. Department of Energy

Rick Duke, Deputy Assistant Secretary for Climate Policy

5:25-5:30 Summary Comments by U.S. Environmental Protection Agency

A1 McGartland, Director of the National Center for Environmental
Economics

6

-------
Progress in Estimating Climate Change Impacts
Michael Oppenheimer
Program in Science, Technology, and Environmental Policy
Princeton University

ABSTRACT

The assessment of potential impacts of climate change has progressed over time from taxonomies and
enumeration of the magnitude of potential direct effects of climate change on individuals, societies,
species, and ecosystems according to a limited number of metrics toward a more integrated approach
that encompasses the vast range of human response to risk, perceived risk, and experience. Recent
advances are both conceptual and methodological, and focus on analysis of some consequences of
climate change that were viewed heretofore as intractable. This presentation will review a selection of
these developments and represent them through a handful of illustrative cases. A key characteristic of
the emerging areas of interest is a focus on understanding human responses to impacts and developing
integrated approaches which assess impacts in an evolving socioeconomic and policy context.

1. Dynamic vulnerability

While climate impact analysis in some sectors, notably agriculture, has attempted to integrate human
responses by accounting in part for the potential to adapt, such approaches have been marginal and
particular, and unable to estimate the full interaction among humans, socioeconomic systems, and the
climate. Ideally, impacts would be assessed in the context of development scenarios which capture
vulnerability as an evolving feature rather than a static set of capacities and limits. Responses would
also be dynamic, described as resulting from experienced-based perception of risk as well as "objective"
risk. The latter is particularly important in situations where impacts are dominated by extreme and/or
rare events, where uncertainty is large, and where learning is a critical determinant of response.

The SRES1 represented a potential step in this direction. However, they were mainly used in impact
analysis to determine a range of climate futures rather than the range of human responses. The
emerging Shared Socioeconomic Pathways2 may provide an improved basis for integrated analysis of
impacts.

2. Mapping human responses and evaluating their indirect consequences

To date, impact studies have naturally tended to focus on the direct effect of changes in the physical
climate system (including sea level). But some of the key impacts are indirect, arising from decisions
stimulated by the initial physical changes, or expectation thereof. For example, it is well known that
people migrate, sometimes temporarily, sometimes permanently, in response to unfavorable
environmental changes, including climate.3 These movements have the potential for large scale effects
on natural resources, ecosystems services, and species. They are second order in the sense of being
indirect but not necessarily in the sense of their magnitude4.

For example, one recent study suggests that a potential relative shift in agricultural productivity in
western South Africa compared to the eastern part of that country could encourage cultivation in
regions now designated for protection for species conservation purposes5. Such indirect impacts may in
some cases be larger than the direct consequences of the changing climate for the species at risk. While

-------
ability to reliably quantify such responses runs into the limits imposed by regional modeling and
downscaling, the same is true of the direct responses. On the other hand, modeling of some responses,
like human migration6and the potential for large associated indirect impacts, is in its infancy, and this
presents a key obstacle. But at the present time, it is at least possible to investigate which areas may
become vulnerable because their relative attractiveness for cultivation or other economic activities is
projected to increase. Other shifts in human settlement such as those driven by sea level rise could
likewise bring about large scale indirect impacts.

Such impacts might fall under the category of action-at-a-distance, in the sense that a climate change in
one region stimulates responses which have impacts on people and resources in another region(s). The
potential number of such reverberations is large, including, for example via the interlinked global market
system (discussed further in section 3, below).

3.	Integration of impacts, adaptation, and mitigation (biofuels, geo-engineering)

It has long been known that adaptation actions bear consequences for mitigation strategies, e.g.,
projected increases in cooling and decreases in heating requirements bear implications for strategies to
mitigated carbon dioxide emissions. A new focus is developing on the implications which mitigation
actions bear for impacts and adaptation. For example, there is evidence that the conversation of
unmanaged forest and cultivated land for the purposes of growing crops intended for bio-fuel feed
stocks (encouraged in some instances by energy- and climate-related policy initiatives) could bring along
substantial consequences for biodiversity and world grain prices, respectively7. A complex set of
subsequent human responses would also result from the latter. These in turn would affect the initial
mitigation actions by raising their cost and potentially undercutting political support for them.

A second emerging area of interest is geo-engineering, particularly short-wave radiation management,
which is projected to produce significant, potentially harmful climate impacts far removed for the
location of initiation of the mitigation actions8. Such impacts would not only result in various human
impacts (via the water and agricultural sectors) and responses, but have the potential to feed back
through the political system and affect judgments about the viability of this mitigation approach. Both
of these examples illustrate the tightly couple nature of the mitigation-impact-adaptation system and
the unavoidable necessity of understanding both political and economic consequences to adequately
project future outcomes.

4.	Interacting systems and stressors

Consideration of interacting stressors and systems8 are not new to climate impact studies but just as
with the topics above, a new emphasis is emerging which examines such interactions through the lens of
human responses to general socioeconomic conditions as well as climate-related circumstances. Impacts
of climate change evolve in the context of multiple additional environmental stressors including air
pollution, water pollution, and the massive consequences resulting from urbanization and other
concentrations of human population such as occur in deltaic and estuarine regions. In urban
agglomerations, we see the potential increase in efficiency of use of some resources (energy),
accompanied by the shifting of environmental natural resource exploitation to outlying regions (water
withdrawal, food production). Interacting stresses include 1) the squeezing of an increasing population
within a potentially shrinking land area (due to sea level rise), 2) the increasing health risk of a growing
population subject to an increasing urban heat island effect, and 3) the increasing problems associated
with water and solid waste disposal under conditions of increased heat and population density.

-------
5. Extremes and disasters

Consideration of extremes and disasters provides an additional framework for understanding potential
impacts, adaptation, and socioeconomic ripple effects. Much of the past climate change impacts
research has focused on outcomes of changes in mean values of climate parameters. The difficulties
entailed in attempting to account for changes in extremes include among others, the difficulty of
projecting changes in many extremes, and the social and geographic specificity of conditions of
vulnerability and exposure which combine with extreme physical events to produce extreme impacts
and disasters.

The upcoming IPCC Special Report, Managing the Risks of Extreme Events and Disasters to Advance
Climate Change Adaptation (SREX), is expected to provide new insights which should help define an
emerging research agenda on such impacts. One noteworthy feature is the importance of the timing of
events and their interactions, which can amplify the effects of both underlying trends in the mean
climate as well as the effects of individual extreme events.

6. Methodological advances

As series of developments are gradually improving the ability to understand causation, to project future
impacts and responses, and to permit a fuller risk management approach to impact assessment. Among
these are advances in detection and attribution, further exploitation of methods commonly used in
econometrics9,10, and probabilistic and multi-metric frameworks for evaluating risk.

References

1. N. Nakicenovic and Rob Swart (Eds.) Emissions Scenarios, IPCC Special Report, Cambridge University
Press, UK (2000).

2. E. Kriegler et al, Socioeconomic Scenario Development for Climate Change Analysis (2010),

At http://www.ipcc-wg3.de/meetings/expert-meetings-and-workshops/files/Kriegler-et-al-2010-
Scenarios-for-Climate-Change-Analvsis-Working-Paper-2010-10-18.pdf

3. Care, In Search if Shelter: Mapping the Effects of Climate Change on Human Migration and
Displacement (2009)

4. W. Turner et al., Climate change: helping nature survive the human response, Cons. Lett., doi:
10.llll/j.l755-263X.2010.00128.x (2010)

5. B. Bradley et al., Predicting how human adaptation to climate change will affect ecological
conservation: a case study from South Africa, submitted (2010)

6. S. Feng, A. Krueger, and M. Oppenheimer, Linkages among climate change, crop yields and
Mexico-US cross-border migration, PNAS www.pnas.org/cgi/doi/10.1073/pnas.10026321Q7 (2010)

7. NRC, Liquid Transportation Fuels from Coal and Biomass: Technological Status, Costs, and

-------
Environmental Impacts, National Academic Press (2009)

8.	R. Warren, The role of interactions in a world implementing adaptation and mitigation solutions to
climate change, Phil. Trans. R. Soc. A (2010) 368, 1-25 doi:10.1098/rsta.2010.0271

9.	W. Schlenker and M. Roberts, Nonlinear temperature effects indicate severe damages to U.S. crop
yields under climate change, PNAS 106, 15594-15598 (2009).

10.	M.B. Burke et al., Warming increases the risk of civil war in Africa, PNAS 106, 20670-20674 (2009)

-------
Progress in Estimating Climate

Change Impacts

Michael Oppenheimer
Program in Science, Technology, and Environmental

Policy
Princeton University

At

Climate Damages Workshop
USEPA/DoE
27 January 2011

-------
Overview

•	Systematic assessment of potential impacts of climate
change and valuation of damages goes back to 1970s*

•	Recent advances in process-based and statistical
modeling

•	Limited progress is accounting for adaptation capacity
and human responses in general

•	Emerging issue: need an integrated approach to
impact/adaptation/development

*for example, Williams, J. (ed.): 1978, Proceedings of an IIASA Workshop on Carbon Dioxide,

Climate, and Society, Pergamon Press, Oxford, February 21-24

-------
Progress in physical exposure
and impact modeling

•	GCM resolution improves, downscaling, RCMs

(e.g., watershed-scale, coral reef studies)

•	Statistical modeling of responses (to variability):
agriculture, migration, conflict

•	Deployment of GIS data (coastal impacts)

-------
Where progress has been slow

•	Incorporating adaptation capacity into impact modeling

(arises in social and natural systems): obscure

•	Understanding the gap between capacity and
implementation of adaptation

•	Assessing indirect effects

•	Developing a comprehensive approach:

development paths plus top-down/bottom-up

-------
Emerging areas

Human Responses to Climate Change and
their

Indirect and Remote Consequences

»>Migration of human population affects resources
and people at a distance (Leman abstract)

»>Shift in regions exploited for agriculture

threatens or benefits unique ecosystems/species

-------
Overlap of areas losing crop suitability and
conservation land in Cape region, year 2050

Turner et al, Cons. Letters

-------
Integration of Impacts, Adaptation with
Mitigation

»>Bio-fuel feedstock production impacts on
land use for biodiversity, food production
and prices (Schlenker abstract) and various
reverberations, including political

»>Impacts of geo-engineering

-------
Interacting Systems and Stressors

»>Urbanization with urban heat islands and climate

change (McCarthy et al 2010): affects energy and
resource use and human health

»>Upstream water diversion causing deltaic subsidence
with exposure to sea level rise (Ericson et al)

-------
Deltas and Upstream Reservoirs

Worldwide

Fig. 1. Global distribution of the 40 deltas analyzed m this study, the potentially contributing drainage basin area of each delta (blue) and the large
reservoirs (>0.5 km' maximum capacity) in each basm.

Ericson et al, Global and Planetary Change 50 (2006) 63-82

-------
Climate Extremes and Disasters

»> Local specificity of exposure and vulnerability

(Knutson abstract)

»> How might learning occur as history becomes a
poor

guide to the future (SREX: lessons from disaster
response)?

-------
Dynamic Vulnerability

»> Impacts depend on vulnerability, which evolves

with

defense)

development (e.g., affordability of coastal

»> Both increases and decreases occur to
vulnerability

as learning competes with mal-adaptation
and risk-shifting behavior (withdrawal vs.
hardening in some cases of flood-plain defense)

»> How to integrate the contextual aspect associated
with diverse potential development pathways

mtn

-------
Not just a developing country issue...

-welcome to Atlantic City

V* r * 1

.

Courtesy Norm Psuty

-------
1

Valuation of Damages from Climate Change1

William R. Cline
Peterson Institute for International Economics
January, 2011

Introduction

It is an honor to be invited to speak at this workshop. I look forward to hearing the latest views
of the prominent experts assembled by the organizers. I believe the EPA and other agencies made a
good start on estimating the social cost of carbon in the February 2010 report of the Interagency
Working Group (2010). Strengthening those estimates has become all the more important with the
delay of US climate legislation and the de facto recourse at present to Plan B, in which EPA enforcement
and action by the three Regional Climate Initiatives at the state level constitute the interim delivery
mechanism for internationally promised US action. I will stress the importance of strengthening the
damage estimates in two dimensions: treatment of catastrophic damage and choice of the central
discount rate.

Brief Retrospective2

Let me first provide a brief retrospective on cost-benefit analysis of climate change. My 1992
book (Cline, 1992) used estimates by the EPA and other sources to estimate that 2.5°C warming from a
doubling of carbon dioxide by late this century would impose damages of 1 percent of GDP on the US
economy. In order of importance, the damages were in agriculture, electricity requirements for
increased cooling in excess of reduced heating; water supply; sea-level rise; loss of human life;
tropospheric ozone pollution; species loss, and forest loss. I note that these are broadly the same
categories on the agenda of this conference. However, I emphasized that the analysis should cover 300

1 Remarks at the conference on Improving the Assessment and Valuation of Climate Change Impacts for Policy and
Regulatory Analysis, Environmental Protection Agency and US Department of Energy, Washington DC, January 27-
28, 2011.

2 For a recent overview, see Cline (2010a).

-------
2

years, the horizon before major re-absorption into the deep ocean. Using the scientific relationships
reported in the first IPCC review, I estimated that over that horizon warming could reach 10°C,
increasing damage to 6 percent of GDP in the central case and three times as high in a higher-damage
variant. I invoked the Ramsey (1928) discounting method that imposes zero pure time preference, or
discounting for impatience, for intergenerational comparisons. With my discount rate of 1.5 percent for
per capita income rising at 1 percent, I estimated that - with modest risk-weighting, a cut in greenhouse
gas emissions by one-half at an annual abatement cost of around 3 percent of GDP was warranted on
social cost-benefit grounds. Inclusion of catastrophic damages would have reinforced the conclusion.
Using his DICE model and a considerably higher discount rate, William Nordhaus (1993) concluded that
much less abatement was warranted. In the 1995 IPCC survey of economic modeling results, social cost
of carbon by 2010-20 was placed in a range of about $5-$7 (1990 dollars) per ton of C02 in estimates by
Nordhaus as well as some other modelers, but reached $18 (or $30 at 2010 prices) in my alternative
runs of the DICE model using my discounting (Pearce et al, p. 215; Cline, 1997, pp. 110-17).

Even after an important revision of the DICE model in 2000 that tried to incorporate catastrophic
damages based on surveys of expert opinion, by 2008 Nordhaus (2008) continued to estimate low
optimal carbon dioxide taxes ($11 per ton in 2015 and still only $24 by 2050) and high optimal emissions
paths (rising from 30 GtC02 now to 44 GtC02 in 2050) and high optimal atmospheric concentrations (480
ppm by 2050 and 660 ppm C02 by 2200). In sharp contrast, in his 2007 review for the UK Treasury,
Nicholas Stern and his team found that social benefits of greatly exceeded abatement costs of limiting
atmospheric concentrations to 500-550 ppm C02-equivalent, requiring emissions about one-third lower
than the 2000 levels by 2050 and even lower thereafter (Stern, 2007). Stern used the PAGE model with
a damage function quite similar to that used in Nordhaus' DICE model, and found that by 2200 global
damages under business as usual would amount to 5-20 percent of world product. Using Ramsey's zero
pure time preference and considering an infinite horizon thereafter, the Stern Review also placed the

-------
3

equivalent "now and forever" value of unrestrained damages at 5 to 20 percent of world GDP. He
placed the abatement cost for the 500-550 ppm ceiling at -1% to +3.5% of world product by 2050, and
the average cost at about $50 per ton of C02 in 2015, falling to about $30 by 2025. (p. 260) Essentially
the same two central analytical features of my 1992 book, Ramsey-type discounting with zero pure time
preference and the adoption of a long horizon, led Stern to the same conclusion that much more
aggressive abatement was warranted on social cost-benefit grounds than identified by Nordhaus and
some other modelers.

At this point Martin Weitzman (2007) entered the debate with a new emphasis on the implications
of uncertainty about catastrophic effects. He judged that Stern was probably right for the wrong reason.
The pure time preference rate should not be set at zero, but future catastrophes from climate change
could be severe enough to drive consumption levels below those of the present and hence discounting
for consumption would turn negative. The "fat tail" of the probability distributions of warming and
damage are at the heart of this risk, and they introduce uncertainty about the discount rate that should
be used. However, Weitzman's mathematics involve a singularity in which the present value of future
loss is infinite, so his analysis is difficult to make operational. Sterner and Persson (2007) also arrive at a
favorable evaluation of Stern-like aggressive action but argue that this conclusion could be reached
"even with Nordhaus' conventional assumptions of a fairly high rate of discount... [if] the escalation of
prices for scarce environmental services were taken into account."

Catastrophe Update and Super-Contingent Valuation

Scientific work in recent years has increased the concern we should have about catastrophic
effects of climate change. The three catastrophes usually considered are: collapse of the ocean
conveyor belt that causes the Gulf Stream and keeps Northern Europe warm; melting of the Greenland
ice sheet or collapse of the West Antarctic ice sheet, either of which would raise sea levels by 7 meters ;

-------
4

and a runaway greenhouse effect as methane is released from clathrates on continental shelves and
from permafrost. With respect to the conveyor belt, a 2005 study found that "the Atlantic meridional
overturning circulation has slowed by about 30 percent between 1957 and 2004" (Bryden, Longworth
and Cunningham, 2005). With respect to the Greenland ice sheet, in a 2005 study Meinshausen (2005)
found that " the loss of the Greenland ice-sheet may be triggered by a local temperature increase of
approximately 2.7°C, which could correspond to a global mean temperature increase of less than 2°C."

Perhaps the most disturbing new evidence on catastrophic risks concerns massive extinctions as
a consequence of an eventual loss of oxygen in the oceans, a buildup in anaerobic bacteria, and the
release of hydrogen sulfide from the oceans in amounts toxic for plants and animals. A 2005 study by
Kump, Pavlov, and Arthur (2005) found that "fluxes of H2S to the atmosphere ... would likely have led to
toxic levels ...[that served] as a kill mechanism during the end-Permian, late Devonian, and Cenomanian-
Turonian extinctions" (p. 397). In the first of these, the Permian-Triassic extinction event 251 million
years ago, some 90 percent of species on land and in the oceans became extinct. Volcanic eruptions in
the Siberian "traps" (lava-flows) are likely to have caused sharp increases in atmospheric concentrations
of C02, methane releases from clathrates, and an increase in global temperatures by levels 6°C (Benton,
2003). "The evidence at hand links the mass extinctions with a changeover in the ocean from
oxygenated to anoxic bottom waters" (Ward, 2010, p. 189). A shut-down in the ocean conveyor belt
would have caused this changeover, setting the stage for the buildup of anaerobic bacteria and eventual
release of hydrogen sulfide. Similarly, a 2007 study found that over the past 520 million years,
extinctions were relatively high during warm "greenhouse" phases; four of the five worst mass
extinctions were associated with such phases (Mayhew and Benton, 2007).

The time scale for such a phenomenon is unknown, but is probably on the order of thousands of
years.3 Eventually a world free of ice sheets would mean sea levels 60 to 80 meters higher than today.4

3 Lee Kump, personal communication, November 1, 2007.

-------
5

If the H2S hypothesis is correct, humans could probably survive using gas masks out of doors and living in
atmospheric-controlled chambers, or at least those who could afford to do so would. However, food
supply would be challenging, because of the likely die-off of livestock animals.

These stakes pose an acute problem for cost-benefit analysis. Suppose the time horizon is 2,000
years. Suppose world product stabilizes at $500 trillion (compared to $340 trillion in the EMF-22
scenarios for 2100, and $50 trillion at present), and world population, at 9 billion. The Interagency
report's lowest discount factor of 2.5 percent expands $1 over 2,000 years to $2.8 x 1021 dollars, or $2.8
billion trillion. The policy maker would have to conclude it is not worth spending even a single cent
today to avoid the complete elimination of one year's worth of world product 2,000 years from now.

Hopefully, policy makers do not make calculations about such large but long-term stakes in this
fashion. It may be helpful to resort to a sort of "super-contingent valuation" thought experiment.
Instead of conducting a survey of how much the typical household would be willing to pay to save the
polar bear, one could think of how policymakers seem to be expressing revealed contingent valuation of
catastrophic damage. Consider the pledges at Copenhagen. The industrial countries have stated that
they will provide $100 billion annually by 2020 to help developing countries curb greenhouse gas
emissions. Business as usual emissions of developing countries are likely to be 21 GtC02 by then (Cline,
2010b). The pledges so far from Copenhagen amount to reducing that amount by only 0.7 GtC02, or by
less than 4 percent. Suppose the policymakers believed that by pledging resources, they could induce
the developing countries to more than double that effort, attaining a 10 percent reduction. That would
amount to a cutback of 2.1 billion tons at $100 billion, implying an average abatement cost of $50 per
ton of carbon dioxide. That is twice the central Interagency estimate for 2020. So why not think of the
Copenhagen pledges as revealed contingent evaluation by industrial country leaders placing the value of

4 The lower figure is from Hansen et al (2008) as interpreted in Cline (2010a); the higher figure, from Ward (2010,
p. 39).

-------
6

avoiding catastrophe at about equal to the value of the other global warming damages that have been
counted in the models.

Discount Rate Once Again

Interestingly enough, this exercise yields a price that is much closer to the Interagency's low-
discount case ($42) and lower than the 95th percentile high-damage case ($81). This comparison brings
one right back to the two central issues that have challenged the economics of global warming from the
start: the discount rate and proper valuation of catastrophic risk. I have just discussed one important
catastrophic risk. Let me say three specific things about the discount rate.

First, returning to proper discounting for a time scale of one or two centuries rather than
millennia, I would emphasize that the particular value chosen for one specific parameter makes an
immense difference: the so-called elasticity of marginal utility, or the percent decline in marginal utility
for a percent increase in per capita consumption. In the Ramsey equation, the discount rate equals pure
time preference, which many would agree should be set at zero for intergenerational comparisons, plus
the elasticity of marginal utility multiplied by the growth rate of per capita income. Stern's use of unity
for the elasticity of marginal utility, or a logarithmic utility function, probably understates how rapidly
marginal utility falls off as consumption rises. But the value of 2 used for this elasticity by both
Nordhaus and Weitzman probably overstates it. The evidence I would cite is the structure of
progressive tax regimes in industrial countries. A parameter of unity would lead to a strictly
proportional tax, in which it is considered fair that the poor man pays the same percent of income as the
rich man. We observe more progressive structures than that. But a parameter of 2 would mean, for
example, that the average (not marginal) income tax on an income of $650,000 would be 79 percent if

-------
7

the tax on an income of $20,000 is 10 percent.5 That is far more progressive than we observe. The
value of 1.5 that I used in 1992 still seems about right to me; in this example it would generate an
average tax rate of 42 percent for the rich household, much closer to what we observe.

Second, I urge the Interagency working group to use the long-term Treasury Inflation Protected
(TIP) bond as the best measure of the pre-tax risk-free real rate for discounting consumption. Using
instead the long-term nominal rate and deflating by actual inflation gives an understatement during the
high-inflation 1970s and early 1980s, but an overstatement for the following decades because markets
consistently lagged behind the actuality of falling inflation in adjusting inflation expectations. Using the
available 20- and 30-year TIP rates since 2004, the real rate has averaged 2.1 percent (Federal Reserve,
2011). When the Interagency's translation to after-tax return at 73 percent of the pre-tax rate is
applied, that yields 1.5 percent as the discount rate for consumption. So I would argue that the
"descriptive" approach using the observed consumption discount rate should place it at 1.5 percent,
more than a full percentage point below the rate of 2.7 percent used in the Interagency report for the
same concept.

Third, per capita growth is the other component of the discount rate. The Interagency group
expects global per capita income to rise at 2 percent annually through 2100. Actually the EMF-22
projection for 2100 amounts to an annual per capita growth of 1.77 percent for 2010-2100 (Interagency
Working Group, 2010, table 2). Moreover, that is at market exchange rates. The growth rate will be
lower at purchasing power parity, at about 0.8 times as much based on the Balassa-Samuelson
relationship (Subramanian, 2011). The consumption discount rate would then be 1.5 for the elasticity of
marginal utility, multiplied by 1.4 percent for ppp growth in per capita income, or 2.1 percent. That

5 In the constant relative risk aversion (CRRA) utility function, utility from consumption level C is: U = c'^'/tl-ri),
where r| is the absolute value of the elasticity of marginal utility. For a given average tax rate for the poor family,
the socially optimal average tax rate for the rich family is the level that just equates the reduction in utility for each
of the two families as a consequence of the tax.

-------
8

would only be for the 21st century. The 22nd century should be discounted at a lower rate because per
capita growth would decelerate.

An insurance approach

Even with refinements in discounting, the ultimate difficulty of placing a value on catastrophic
effects raises doubts about the use of integrated assessment models to arrive at optimal paths of
abatement and carbon dioxide shadow prices. That is why both Stern and Weitzman adopt essentially
an insurance approach to global warming policy, even though they disagree on the discount rate. Stern
suggests a ceiling of 500-550 ppm for carbon-dioxide-equivalent concentrations. At Copenhagen in
December 2009, heads of state set a ceiling of 2°C for eventual warming. Once such targets are set, the
social cost problem becomes one of identifying the least-cost way to achieve the target. The discount
rate chosen affects the timing of the cutbacks, but their cumulative magnitude is determined
exogenously given the climate target rather than endogenously as a function of damage avoided and
abatement cost. Even in this approach it would be important to calculate the best estimate of
quantifiable non-catastrophic damages avoided, as they would likely cover a considerable portion of
abatement costs if not the full amount. Given marginal abatement cost along the least-cost path, the
proper price to use for the social cost of carbon is by definition the marginal abatement cost identified
for that path.

It turns out that any extra cost paid for this insurance approach may be quite small even when
compared to a supposed optimal path using more conventional discounting. Thus, in Nordhaus' (2008)
results using the latest version of the DICE model, the difference between the future path of per capita
consumption in his optimal path and in a path adhering to a 2°C ceiling (p. 209) is, as Tom Schelling
tends to say, no wider than the lead of the pencil being used to draw the graph. The present value of
abatement cost in his preferred optimal path that allows eventual warming to reach 3.5°C is a tiny 0.11

-------
9

percent of the present value of future world product. If instead the 2°C limit is observed, the present
value of abatement cost rises to 0.57 percent of world product (Cline, 2010a). The additional insurance
costs 0.46 percent of the present value of world product over the next two centuries. That ought to be
a bargain if one gives much credence at all to the various catastrophe scenarios. Similarly, in the EMF-22
projections reported in the Interagency review, limiting atmospheric concentrations to 550 ppm C02-
equivalent would involve abatement costs amounting to only 0.66 percent of world product in 2030 and
1.3 percent in 2100 (p. 16). The insurance approach would thus seem to recommend that the
Interagency group include as at least one variant a social cost of carbon path set equal to the marginal
abatement cost along either the 550 ppm path or a 2°C ceiling path.

Workshop Issues

I look forward to the discussions in this workshop. Many questions seem relevant for an update
of damage valuation. What do the experts now say about storm damage given the experience of
Katrina? Was the Fourth Assessment Report of the IPCC understating the pace of likely sea-level rise in
light of new evidence? How does the FUND model's finding of initial benefits rather than damages for
up to 3°C warming square with Meinhausen's eventual loss of the Greenland ice sheet with only 2°C
warming? Where do the agricultural estimates now stand? My own take in my 2007 book was that by
the 2080s the losses in agricultural potential would reach about 5 to 15 percent globally, 30-40 percent
in South Asia, and 20-25 percent in Africa and Latin America, depending on whether carbon fertilization
is included (Cline, 2007). There is also a new category that I hope will be discussed in the session on
health impacts: the adverse effect of warming on labor productivity in outdoor sectors in warm climates
(Kjellstrom et al, 2008). World Bank modeling of climate policy applies large damage effects in this
category (van der Mensbrugghe and Roson, 2010). I would be interested in whether participants in this
workshop agree.

-------
10

Bottom Line

Let me conclude by returning to where I began: I welcome the February report of the
Interagency Working Group as a good start. I take some comfort from the fact that for the first two
decades, its path for the social cost of carbon is broadly consistent with the Congressional Budget Office
(CBO, 2009) estimates of the allowance price for carbon dioxide - or marginal abatement cost - along
the abatement path in the Waxman-Markey bill passed by the House of Representatives in 2009. That
bill would have cut US emissions by 83 percent below 2005 levels by 2050, arguably enough to be
consistent with global abatement close to what is needed for limiting warming to a range of 2 to 3
degrees C. Thus, the central Interagency estimate of the social cost of carbon dioxide is $26 per ton in
2020 and $33 in 2030; the CBO allowance price for Waxman-Markey is $25 in 2020 and $40 in 2030 (p.
11). However, in later periods the Interagency estimate falls increasingly short of would be needed
under Waxman-Markey: $39 versus $70 per ton in 2040 and $45 versus $120 in 2050. I would suggest
that given this growing discrepancy, EPA enforcement should take special care when applying the
Interagency social cost estimates in decisions affecting new plant equipment designed to be in operation
longer than 20 years.

This being said, it does seem to me that more attention to catastrophic considerations and a
revisiting of the discounting issue, including the use of TIPs as a guide to the pre-tax consumption
discount rate, are likely to lead to a higher path of the social cost of carbon than estimated by the
Interagency group in its February report. It is also the case, however, that sooner rather than later it will
be necessary to adopt comprehensive legislation on greenhouse gas abatement. When that is done, the
American public will no longer have to rely so heavily on the EPA to sort out the right social price of
carbon, because their elected representatives will implicitly have made that decision for them by setting
the terms of the climate legislation. Super-contingent evaluation will have taken place through the

-------
11

democratic process. In the meantime, political economy could plausibly counsel against any massive
shocks in the Interagency Group's revisions of the social cost of carbon. The EPA will need to walk a
tightrope between placing too low an estimate that risks the environment, on the one hand, and on the
other, placing so high an estimate that it provokes congressional blocking measures (such as threats to
block public debt bills unless they include a clause removing the agency's authority to enforce
greenhouse gas abatement). Continuing to build on the professional and rigorous approach already
begun seems likely to help assure that this narrow path can be successfully followed.

-------
12

References

Bryden, Harry L., Hannah R. Longworth, and Stuart A. Cunningham, 2005. "Slowing of the Atlantic
Meridional Overturning Circulation at 25°N," Nature, vol. 438, no. 7068, December 1, pp. 655-57.

CBO, 2009. Congressional Budget Office, The Economic Effects of Legislation to Reduce Greenhouse-Gas
Emissions (Washington: CBO, September)

Cline, William R., 1992. The Economics of Global Warming (Washington: Institute for International
Economics)

Cline, William R., 1997. "Modelling Economically Efficient Abatement of Greenhouse Gases." In Yoichi
Kaya and Keiichi Yokobori, eds., Environment, Energy, and Economy (Tokyo: United Nations University
Press, pp. 99-122).

Cline, William R., 2007. Global Warming and Agriculture (Washington: Peterson Institute for
International Economics)

Cline, William R., 2010a. "Economic Analysis and Climate Change Policy: An Editorial Comment".
Climatic Change, Vol. 101, Issue 3, pp. 387-94. Doi: 10.1007/sl0584-009-9766-0, January 13.

Cline, William R., 2010b. "Carbon Abatement Costs and Climate Change Finance." (Washington:
Peterson Institute for International Economics, November). Processed.

Federal Reserve, 2011. "Selected Interest Rates: Historical Data." (Washington: Federal Reserve).
Available at: http://www.federalreserve.gov/releases/hl5/data.htm

Hansen, J., Mki. Sato, P. Kharecha, D. Beerling, R. Berner, V. Masson-Delmotte, M. Pagani, M. Raymo,
D.L. Royer, and J.C. Zachos, 2008. "Target Atmospheric C02: Where Should Humanity Aim? Open Atmos.
Sci.J., 2, 217-231, doi:10.2174/1874282300802010217.

Interagency Working Group, 2010. Social Cost of Carbon for Regulatory Impact Analysis Under Executive
Order 12866. (Washington: Interagency Working Group on Social Cost of Carbon, US Government,
February). Available at: http://epa.gov/otaq/climate/regulations/scc-tsd.pdf

Kjellstrom, T., R.S. Kovats, S.J. Lloyd, T. Holt, R.S.J. Tol, 2008. The Direct Impact of Climate Change on
Regional Labor Productivity. ENSEMBLES Deliverable D7.8, European Commission, Sixth Framework
Programme. (London: UK Met Office).

Kump, Lee R., Alexander Pavlov and Michael A. Arthur, 2005. "Massive Release of Hydrogen Sulfide to
the Surface Ocean and Atmosphere During Intervals of Oceanic Anoxia." Geology, vol. 33, pp. 397-400.
doi: 10.1130/G21295.1

-------
13

Mayhew, Peter J., Gareth B. Jenkins, and Timothy G. Benton (2007). "A long-term Association between
Global Temperature and Biodiversity, Origination and Extinction in the Fossil Record. Proceedings of the
Royal Society B, No. 275,pp. 47-53. doi: 10.1098/rspb.2007.1302.

Meinshausen, Malte, 2005. "On the Risk of Overshooting 2°C." Paper presented at Scientific Symposium
"Avoiding Dangerous Climate Change," MetOffice, Exeter, 1-3 February, 2005.

Nordhaus, William, 1992. "Optimal Greenhouse-Gas Reductions and Tax Policy in the 'DICE' Model,"
American Economic Review, Vol. 83, No. 2, May, pp. 313-17.

Nordhaus, William, 2008. A Question of Balance: Weighing the Options on Global Warming Policies
(New Haven: Yale University Press)

Pearce, D. W., W. R. Cline, A. N. Anchanta, S. Fankhauser, R. K. Pachauri, R.S.J. Tol, and P. Vellinga, 1995.
"The Social Costs of Climate Change: Greenhouse Damage and the Benefits of Control." In James P.
Bruce, Hoesung Lee, and Erik F. Haites, eds., Climate Change 1995: Economic and Social Dimensions of
Climate Change (Cambridge, UK: Cambridge University Press), pp. 179-224. Contribution of Working
Group III to the Second Assessment Report of the Intergovernmental Panel on Climate Change.

Ramsey, F. P., 1928. "A Mathematical Theory of Saving." Economic Journal, vol. 138, no. 152, pp. 543-
59.

Stern, Nicholas, 2007. The Economics of Climate Change: the Stern Review (Cambridge: Cambridge
University Press)

Sterner, Thomas, and U. Martin Persson, 2007. An Even Sterner Review. RFF Discussion Paper 07-37
(Washington: Resources for the Future, July)

Subramanian, Arvind, 2011. "Projecting PPP-based GDP Growth." (Washington: Peterson Institute for
International Economics). Processed.

Van de Mensbrugghe, Dominque, and Roberto Roson, 2010. Climate, Trade and Development
(Washington: World Bank). Draft, June, processed.

Ward, Peter D., 2010. The Flooded Earth: Our Future in a World Without Ice Caps (New York: Basic
Books)

Weitzman, Martin, 2007. "A Review of The Stern Review on the Economics of Climate Change," Journal
of Economic Literature, vol. 45, no. 3, September, pp. 703-24.

-------
Tropical Cyclones and Climate Change

Thomas R. Knutson
GFDL/NOAA
January 2011

1. Introduction

This extended abstract addresses the question of climate change impacts on tropical
cyclones, with a focus on: 1) the detection or attribution of past anthropogenic changes in
tropical cyclone activity and 2) projected changes by the late 21st century under the IPCC
A1B scenario. A greater emphasis is placed on Atlantic hurricanes than other basins.

In February 2010, a World Meteorological Organization (WMO) Expert Team on
Climate Change Impacts on Tropical Cyclones published an assessment of "Tropical
Cyclones and Climate Change" in Nature Geoscience1. The WMO assessment forms the
basis for the "consensus" or "best estimate" views in this abstract, which are presented in
sections 2-3. Speakers at the workshop were also asked to address the range of possible
outcomes. The ranges of future projections presented in the WMO assessment were not
intended be interpreted as the range of possible future changes. Therefore, in sections 4-
5,1 expand on some issues which were not explicitly covered by the WMO team report,
particularly in section 5 with some speculations concerning a wider range of possible
tropical cyclone changes. These comments on the wider range of possible impacts and
on statistical vs. dynamical models (section 4) represent my personal views and not
necessarily those of the WMO team.

2. Detection of a climate change in tropical cyclones?

The term climate change detection as used in this abstract refers to a change which is
anthropogenic in origin and is sufficiently large that the signal clearly rises above the
background "noise" of natural climate variability (with the "noise" produced by internal
climate variability, volcanic forcing, solar variability, and other natural forcings). As
noted in IPCC AR42, the rise of global mean temperatures over the past half century is an
example of a detectable climate change; in that case IPCC concluded that most the
change was very likely attributable to human-caused increases in greenhouse gas
concentrations in the atmosphere.

In the case of tropical cyclones, the WMO team concluded1 that it was uncertain whether
any changes in past tropical cyclone activity have exceeded the levels due to natural
climate variability. While some long (century scale) records of both Atlantic hurricane
and tropical storm counts show significant rising trends, further studies have pointed to
potential problems (e.g., likely missing storms) in these data sets due to the limited
density of ship traffic in the pre-satellite era. After adjusting for such changes in
observing capabilities for non-landfalling storms, one study3 found that the rising trend in
tropical storm counts was no longer statistically significant. Another study4 noted that
almost the entire trend in tropical storm counts was due to a trend in short-duration (less

-------
than two days) storms, a feature of the data which those authors interpreted as likely due
in large part to changes in observing capabilities.

A global analysis of tropical cyclone intensity trends over 1981-2006 found increases in
the intensities of the strongest tropical cyclones, with the most significant changes in the
Atlantic basin5. However, the short time period of this dataset, together with the lack of
"Control run" estimates of internal climate variability of TC intensities, precludes a
climate change detection at this point. The intensity data also have uncertainties,
particularly in the Indian Ocean where the satellite record is less consistent over time.

3. Tropical Cyclone Projections for the Late 21st Century

Based on available studies, the WMO team concluded1 the following regarding tropical
cyclone projections for the late 21st century, assuming that the large-scale climate
changes are as projected by the IPCC AR4 A1B scenario (quoted from Box 1 of the
Nature Geoscience report):

"Frequency. It is likely that the global frequency of tropical cyclones will either
decrease or remain essentially unchanged owing to greenhouse warming. We
have very low confidence in projected changes in individual basins. Current
models project changes ranging from -6 to -34% globally, and up to +/-50% or
more in individual basins by the late twenty-first century.

Intensity. Some increase in the mean maximum wind speed of tropical cyclones
is likely (+2 to +11% globally) with projected twenty-first century warming,
although increases may not occur in all tropical regions. The frequency of the
most intense (rare, high-impact) storms will more than not increase by a
substantially larger percentage in some basins.

Rainfall. Rainfall rates are likely to increase. The projected magnitude is on the
order of +20% within 100 km of the tropical cyclone centre.

Genesis, tracks, duration, and surge flooding. We have low confidence in
projected changes in tropical cyclone genesis-location, tracks, duration, and areas
of impact. Existing model projections do not show dramatic large-scale changes
in these features. The vulnerability of coastal regions to storm-surge flooding is
expected to increase with future sea-level rise and coastal development, although
this vulnerability will also depend on future storm characteristics."

While the WMO team judged that a substantial increase in the frequency of the most
intense storms over the 21st century is more likely than not globally, their confidence in
this finding was limited, since the model-projected change results from a competition
between the influence of increasing storm intensity and decreasing overall storm
frequency. An example of such a change projected for the Atlantic basin is found in a
recent downscaling study6 by Bender et al. (GFDL) using an operational (9 km grid)

-------
hurricane prediction model. This downscaling framework projects a doubling in the
frequency of Atlantic category 4-5 hurricanes over the 21st century (A1B scenario) using
an 18-model average climate change signal. However, when four of the 18 individual
models were downscaled, three showed an increase and one a decrease in category 4-5
frequency. Differences in regional SST projections in the various climate models
appeared to be important for producing this large range of projections, implying that
uncertainties in future regional SST pattern changes must be narrowed to reduce the
uncertainty in Atlantic hurricane projections. The study also presented preliminary
estimates of the climate-induced change in hurricane damage potential for the Atlantic
basin (+28% by 2100 for the 18-model average, with a range of -54% to +71% for the 4
individual models runs). These damage potential projections do not include important
influences such as sea level rise, coastal development, and societal adaptation.

4. Methodologies for projecting Tropical Cyclone changes:
statistical vs. dynamical models

The projections in the previous section rely heavily on dynamical models including
global climate models, higher resolution global atmospheric models forced by SSTs from
global climate models, or even higher resolution regional downscaling models. In
addition, some studies employed either statistical/dynamical hybrid models or theoretical
intensity models. The WMO report also discussed an example of using purely statistical
(correlation) methods to project late 21st century Atlantic hurricane power dissipation. In
that case, two alternative statistical models of hurricane activity vs SST, both of which
perform comparably during the historical period, give dramatically different projections
of late 21st century activity, with the projection based on local tropical Atlantic SST
showing a dramatic increase of about 300% in power dissipation by 2100. The second
statistical approach (relative SST) projects much smaller changes in Atlantic power
dissipation by 2100—a scenario strongly favored by current dynamical models. The
differences between various dynamical model projections seem to be explained7 in large
part by differences in tropical Atlantic warming relative to the rest of the tropics as
projected by the parent climate model used to drive the downscaling model.

The example for Atlantic power dissipation illustrates how dynamical and statistical
downscaling techniques, or different statistical approaches, can differ substantially in
their projections of the tropical cyclone response to a given climate change scenario. In
terms of general modeling approaches, both dynamical modeling and statistical modeling
techniques can provide complementary approaches and are worthy of pursuit, although
each has its limitations, and results using either approach should be interpreted with due
caution. Dynamical modeling attempts to use fundamental physical laws such as the
equations of motion and the first law of thermodynamics, integrating systems of these
equations forward in time using computer models. One reason this approach is often
favored in the case of climate change is that one assumes that the fundamental laws are
more likely to be applicable in a changed climate than empirical relations derived by
training a statistical model on past climate data alone.

-------
5. Some speculations on the range of possible outcomes

Here I regard the projections in section 3 from the WMO report as consensus statements
based on available studies. However, it is possible that more dramatic future changes
could occur over the 21st century. While, in my opinion, these more dramatic changes
remain speculative, they are at least plausible enough to merit discussion here.

First, it is possible that 21st century changes in tropical cyclones will be less potentially
damaging than the scenarios outlined in the projections section. For example, some
studies suggest that TC activity in some basins, such as the NW Pacific and North
Atlantic, could shift eastward away from current landfalling regions and thus perhaps
reduce the percentage of storms that make landfall in major population regions. Global
climate transient sensitivity or sea level rise could be at the low end, or even lower than,
the range shown in IPCC AR4. Future greenhouse gas concentrations could be toward
the lower end or lower than IPCC AR4 scenarios. Alternatively, it is also possible that
the reverse could be true in these cases, i.e., that transient climate sensitivity, future
greenhouse gas concentrations, sea level rise, and so forth could be higher than expected,
or even that storm tracks could shift systematically more toward major landfalling
regions, in contrast to a number of current projections.

In addition to these contributors to uncertainty, for the remainder of this section, I will
focus on other more novel mechanisms under which future changes could imply
substantially greater damage potential than the projections of the WMO report.

Vertical profile of temperature change. A common characteristic of climate model
projections of greenhouse warmed climates is an increase in the temperature change with
height, such that the upper troposphere about 4 miles above the earth's surface warms
more than near the surface. This enhanced warming with height is one of the key factors
leading to relatively modest changes in hurricane activity in future climate projections. If
the warming were instead uniform with height through the troposphere, the atmosphere
would become more unstable and much more conducive to hurricane activity over time,
and the resulting increases in intensity could be several times larger than those currently
projected. Interestingly, observed vertical profiles of air temperature changes since about
1980 using radiosondes and some satellite records actually show a relatively uniform
warming with height through the troposphere. However, as argued by Santer et al.8, such
a change is not only inconsistent with climate models and with the notion that the tropical
atmospheric remain close to a moist adiabatic profile, but such as uniform change also
differs from the vertical profile of year-to-year fluctuations in temperature, where climate
models and observations agree that an such temperature variations have an amplified
signal with height in the troposphere. Further, Allen and Sherwood9 argue that the
observed destabilizing temperature trends are inconsistent with temperature trends
inferred from wind fields. Therefore I consider it more likely that data problems with the
radiosonde and satellite temperature datasets have led to unreliable observed temperature
trend profiles that falsely indicate a substantial destabilization of the tropical atmosphere
since 1980. Of course it remains important to confirm this assertion with further studies

-------
and to maintain a vigilant observing network to monitor the vertical profile of tropical
temperatures and TC activity as the planet continues to warm.

Lower stratospheric temperatures. A variant on the theme of vertical profile of
temperature changes is the recent study of Emanuel10, who reports that a cooling trend in
the lower stratospheric temperatures in recent decades implies an increase in potential
intensity of hurricanes in the Atlantic. According to his statistical/dynamical model, this
has further caused an increase in Atlantic tropical storm numbers. While the lower
stratospheric temperature decrease remains a subject of further investigation as to its
veracity and cause, preliminary results with another (dynamical) model from GFDL (G.
Vecchi, personal communication) suggest that lower stratospheric temperatures do not
affect tropical storm counts substantially in that model. Further work is needed to better
constrain lower stratospheric and upper tropospheric temperature changes, their causes,
and their impact on tropical cyclones in general. For example, one can speculate that
ozone changes and related atmospheric effects could have affected tropical upper
tropospheric temperatures enough to change tropical cyclone activity substantially. If so,
this mechanism would have implications for past and future (projected) changes in
Atlantic hurricane activity. For example, if it turns out that ozone depletion contributed
substantially to the increased Atlantic hurricane activity in recent decades, then the higher
activity levels since 1995 could be more persistent than expected on the basis of typical
internal variability (Atlantic Multidecadal Oscillation) arguments. Those internal
variability arguments typically suggest that hurricane activity will likely return toward
pre-1995 levels sometime in the next few decades. In any case, the potential links
between lower stratospheric and/or upper tropospheric temperatures, climate forcings,
and hurricane activity mentioned here remain speculative.

Impact of Tropical Cyclones on Ocean Heat Transport. Previous work by Emanuel
had suggested that tropical cyclones could influence the climate system through changes
in the rate of vertical oceanic mixing, leading to changes in the global oceanic heat
transport. More recent studies have estimated that this influence on heat transport is
confined mainly to the tropics. For example, Jansen et al.11 estimate that TC cause less
than 10% of the global poleward heat transport.

From a paleoclimate perspective, changes in tropical cyclone activity have been
proposed as a key mechanism for maintaining the 'equable' climates of the early Pliocene
(3 to 5 million years ago), when some geologic proxy indicators suggest that the warm
tropical SST region was markedly expanded poleward and the eastern equatorial Pacific
cold tongue was absent. Federov et al.12 simulated large increases in tropical cyclone
activity during this time, and suggested that the very different temperatures of that time
were linked to tropical cyclone feedbacks on climate. Enhanced tropical cyclone activity
in their downscaling in the eastern Pacific eventually leads, in their climate model
simulations, to permanent El Nino conditions. While the simulated changes in TC activity
and in sea surface temperatures in their study are dramatic, the implications of their
simulations for climate changes over the next century or so remain speculative.

6. Key research needs going forward

-------
Among the key research needs going forward is the urgent need to have consistent,
homogeneous long-term records of hurricane statistics (for trend and climate change
assessment) and the need to narrow uncertainties in future sea level rise and regional SST
pattern changes that drive regional tropical cyclone changes. Improved quantification or
reduction of uncertainty in SST pattern projections will likely depend on reducing
uncertainties in cloud feedback, aerosol forcing, and possibly in coupled ocean-
atmosphere interaction, which remain central problems in climate change research.
Continued monitoring of tropical cyclone activity globally for emergence of trends, as
well as further research concerning the vertical structure of the atmospheric temperature
changes and ocean mixing effects by tropical cyclones are all prudent measures for
earlier detection and/or anticipation of future "surprises" in the hurricane/climate realm.

7. References

1.	Knutson, T., J. McBride, J. Chan, K. A. Emanuel, G. Holland, C. Landsea, I. Held, J. Kossin,

A. K. Srivastava, and M. Sugi, 2010: Tropical cyclones and climate change. Nature
Geoscience, 3, doi:doi: 10.1038/ngeo779.

2.	IPCC, 2007: Climate Change 2007: The Physical Science Basis. Contribution of Working

Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate
Change. S. Solomon, et al. (eds.). Cambridge University Press, Cambridge, United
Kingdom and New York, NY, USA, 996 pp. (2007).

3.	Vecchi, G. A., & T. R. Knutson, 2008. On estimates of historical North Atlantic tropical

cyclone activity. J. Clim., 21, 3580-3600.

4.	Landsea, C, Gabriel A Vecchi, L Bengtsson, and Thomas R Knutson, 2010: Impact of duration

thresholds on Atlantic tropical cyclone counts. lournal of Climate, 23(10),
doi: 10.1175/2009ICLI3034.1.

5.	Eisner, I. B., Kossin, I.P., and lagger, T.H., 2008: The increasing intensity of the strongest

tropical cyclones. Nature, 455, 92-95, doi: 10.1038/nature07234.

6.	Bender, Morris A., Thomas R Knutson, Robert E Tuleya, loseph I Sirutis, Gabriel A Vecchi,

Stephen T Garner, and Isaac Held, 2010: Modeled impact of anthropogenic warming on
the frequency of intense Atlantic hurricanes. Science, 327(5964),
doi: 10.1126/science.l 180568.

7.	Villarini, G., G. A. Vecchi, T. R. Knutson, M. Zhao, and I. A. Smith, 2011: North Atlantic

tropical storm frequency response to anthropogenic forcing: projections and sources of
uncertainty. J. Climate, in press.

8.	Santer, B. D., and Coauthors, 2005: Amplification of surface temperature trends and

variability in the tropical atmosphere. Science, 309, 1551-1556.

9.	Allen, R. I. and S. C. Sherwood, 2009: Warming maximum in the tropical upper troposphere

deduced from thermal winds. Nature Geoscience, 1, 399-403.

10.	Emanuel, K., 2010: Stratospheric cooling and tropical cyclones. 29th Conference on

Hurricanes and Tropical Meteorology, 4A.4. American Meteorological Society, Boston.
http://ams.confex.com/ams/29Hurricanes/techprogram/paper 168302.htm

11.	lansen, M.F., R. Ferrari, and T.A. Mooring, 2010: Seasonal versus permanent thermocline

warming by tropical cyclones. Geophys. Res. Letters, 37, doi: 10.1029/2009GL041808.

12.	Federov, A., C. M. Brierley, and K. Emanuel, 2010: Tropical cyclones and permanent El

Nino in the early Pliocene epoch. Nature, 463, 1066-1070, doi:10.1038/nature08831.

-------
Tropical Cyclones and Climate Change

Tom Knutson

Geophysical Fluid Dynamics Lab/NOAA
Princeton, New Jersey

http://www.gfdl.noaa.gov/~tk

Hurricane Katrina (2005), damage estimate: ~$US125 Billion

-------
nature .

REVIEW ARTICLE

PUBLISHED ONLINE: 21 FEBRUARY 2010 |D0I:10.1038/NGE0779

Tropical cyclones and climate change

WMO Expert Team on Climate Change Impacts on Tropical Cyclones

World Meteorological Organization
Weather Research Programme

Tom Knutson, Co-Chair
John McBride, Co-Chair
Johnny Chan
Kerry Emanuel
Isaac Held
Greg Holland
Jim Kossin
Chris Landsea
A.K. Srivastava
Masato Sugi

Geophysical Fluid Dynamics Laboratory/NOAA, Princeton, USA

Center for Australian Weather and Climate Research, Melbourne, Australia

University of Hong Kong, Hong Kong, China

Massachusetts Institute of Technology, Cambridge, USA

Geophysical Fluid Dynamics Laboratory/NOAA, USA

National Center for Atmospheric Research, Boulder, USA

National Climatic Data Center/NOAA, Madison, USA

National Hurricane Center/NOAA, Miami, USA

India Meteorological Department, Pune, India

Research Institute for Global Change/JAMSTEC, Yokohama, Japan

-------
Overview of Assessments

Climate Change Detection and Attribution:

•	It remains uncertain whether past changes in tropical cyclone activity

exceed natural variability levels.

Projections for iate 21st century:

•	Likely fewer tropical storms globally (~no change to -34%), with even

greater uncertainty in individual basins (e.g., the Atlantic).

•	Likely increase in average hurricane wind speeds globally (+2 to11 %),

though not necessarily in all basins

•	More likely than not (>50% chance) that the frequency very intense

hurricanes will increase by a substantial fraction in some basins

•	Likely higher rainfall rates in hurricanes (roughly +20% within 100 km

of storm)

•	Sea level rise is expected to exacerbate storm surge impacts even

assuming storms themselves do not change.

-------
HadCRUT3 global mean temperature anomalies (1850-2010)

Year

Fri Jan 21 13:18:44 2011

Source: Climatic Research Unit, http://www.cru.uea.ac.uk/cru/data/temperature/

-------
There is some recent evidence that overall Atlantic hurricane
activity may have increased since in the 1950s and 60s in
association with increasing sea surface temperatures...

27.8

o

o

0)

CO

1-

ffl
O.

CD

m
o
«
t
i

m

m
©
co

26.8
26.6

Increasing data uncertainty

1950 1960 1970 1980

Year

Source: Kerry Emanuel, J. Climate (2007).

1990

2000

2010

PDI is proportional to the time
integral of the cube of the surface
wind speeds accumulated across all
storms over their entire life cycles.

-------
The frequency of tropical storms (low-pass filtered) in the Atlantic basin 6
since 1870 has some correlation with tropical Atlantic SSTs

16

14

12

10

8

Aug-Oct HADISST 6-18N, 20-60W
Annual Atlantic storm count

r2 =0.74 since 1970

But is the
storm record
reliable
enough for
this?

1860 1880 1900 1920 1940 1960 1980 2000 2020

Year

Source: Emanuel (2006); Mann and Emanuel (2006) EOS. See also Holland and Webster (2007) Phil. Trans. R. Soc. A

-------
=> Significant at
p=0.05

Adjustments to storm counts based on
ship/storm track locations and density

Sources:

Vecchi and Knutson (2008)
Laridsea et al. (2009)

Vecchi and Knutson (in press)

w
c

D
O

O

15

T3
c
ro

Normalized Tropical Atlantic Indices
tttti i i i i i I i i i i i I i i i i i I i i i i i I i i i i i rrr

_l A V	_













111II

111II 11111

111II

II1 1 1

II 1 11

Global Mean Temperature

MDRSST: HadlSST

Tropical Storms: Unadjusted *
Trop. Storms: > 2-day: Unadjusted
Hurricanes: Unadjusted *

Trop. Storms: Adjusted
Trop. Storms: > 2-day: Adjusted
Hurricanes: Adjusted

U.S. Landfall. Trop. Storms (Unadj.)
U.S. Landfall. Hurricanes (Unadj.)

1880 1900 1920 1940 1960 1980 2000

a)I966-2006 Storms

|b|1878-l,)l.

)1946-1965|

Year

-------
1	1 J-

66 72 99 74 79 70 7S 76 S8 72 64 S1 90
76 96 74 79 61 93 92 95 76 64 70 44 74

1981~

1966 1991 1996~
Year

T

T

2001 2006

Maximum wind speed (m s 1)
29	36	41	49

_J	I	I	L_

0.4-

0.3-



"0.2-



V 0,1-



0.0-

-0.1 -

—I	1	1—

0.2	0.4	0.6

Quantile

0.8

8

Global Tropical Cyclone
Intensity Trends

There is some statistical
evidence that the strongest
hurricanes are getting stronger.
This signal is most pronounced in
the Atlantic. However, the
satellite-based data for the global
analysis are only available for
1981-2006.

Quantile regression
computes linear trends for
particular parts of the
distribution. The largest
increases of intensity are
found in the upper quantiles
(upper extremes) of the
distribution.

Source: Eisner et al., Nature, 2008.g

-------
IPCC Projections of Future Changes in Climate

Global Mean Temperature Change

IPCC best estimates
(with likely ranges):

Low scenario (B1):

1.8 C (1.1 -2.9 C)

High scenario (A 1FI):

4.0 C (2.4 - 6.4 C)

ro 3.0
<1)

o 2.0

A2

A1E5

B1

Constant composition
commitment

20th century

Source: IIPCC 4th Assessment Report. Used with permission.

-------
Zetac Regional Model reproduces the interannual variability	10

and trend of Atlantic hurricane counts (1980-2006)

18-km grid model nudged toward large-scale (wave 0-2) NCEP Reanalyses

Atlantic Hurricanes (1980-2006): Simulated vs. Observed

Correlation = 0.84; Linear trends: +0.21 storms/yr (model) and +0.15 storms/yr (observed).

u

O

CO

14
12
10
8
6
4
2
0

Model Ensemble
Observed

1980	1985	1990	1995	2000	2005

Source: Knutson etal., 2007, Bull. Amer. Meteor. Soc.

Year

-------
Simulating past variability i

-(a)

Model Ensemble
Observed

—i—i—i—r-

1985

1990

1995

2000

2005

Year

Atlantic tropical cyclone activity

Progress has been made in developing
dynamical and statistical/dynamical
models for seasonal tropical cyclone
frequency.

Left: examples for the Atlantic basin,
using high resolution atmospheric
models; regional dynamical
downscaling models; and
statistical/dynamical techniques,
(a) and (b) use NCEP Reanalysis.

(c) and (d) use only SSTs.

Current question: Is the cooling of
tropopause transition layer (TTL)
temperatures crucial for simulating the
Atlantic trend in TCs over this period?

Source: Knutson et al., Nature Geoscience (2010)

-------
Projected Changes in Regional Hurricane Activity

GFDL 50-km HIRAM, using four projections of late 21st Century SSTs.

GFDL CM2.1

I

12

18-model CMIP3 Ensemble

HadCM3



"V W ^ cfe

0 -nn "cn 300 350

ECHAM5

^ ' <<
UrS

7

X ' W \ dt
» 'JJm j

v ^ ¦ r
" - (

hv \y

150	200

longitude

300	350

150	200

longitude

250	300

I

I

0

i:

Red I = increase
Blue/ = decrease

Unit: Number per year

•	Regional increases/decreases much larger than global-mean.

•	Pattern depends on details of SST change.

Source: Courtesy Mi rig Zhao, GFDL. Adapted from Zhao, et al. (J, Climate, 2009)

-------
Global Model Tropical Cyclone Climate Change

Experiments: Use A1B Scenario late 21st century projected
SST changes from several CMIP3 models

GFDLCM2.1	HadCM3

Source: Zhao, Held, Lin, and Vecchi (J. Climate, 2009)

Unit: Deg C

ECHAM5

100 T40

Urgtuh

CMIP3 18-model Ensemble

-------
TABLE SL
TC Frequency
Projection:

Reference

Modd/type

Experiment

Global

NH

SH

N Atl.

NW
Pac.

NE
Pac

N
Ind

Tropical Stonn
Frequency Changes
(«*)	

Sua et al. 2002
(ref 36)	

JMA
Timedice

T106L21
(-120km)

lOv

lxC02,2sC02

-34

-28

-39

-61

-66

-67

+9

McDonald et al.
2005 (ief 53)

HadAM3
Timeslice

N144L30

(~100km)

15vIS95a
1979-1994
2082-2097

-10

-30

-30

+80

+42

Hasegawa ami
Emori 2005 (ref 54)

CCSR^TESFRC
GCttmeslice

T106L56

(-120km)

5x20yatlxCO2
7x20v at 2xC02

Yoslumura et al.
2006 (ref 55)

JMA
Timeslice

T106L21
(-120km)

lOv

lxC02,2xC02

-15

Ooiichi et al. 2006
(ref 10)

MRTJMA
Timediee

TL959160
(-20km)

lOv A13

1982-1993

2080-2099

-30

-28

-32

-34

-38

-34

Chauvin et al. 2006
(ref 11)

ARPEGE Ounat
Timedice

Downscale
CNRMB2
Downscale
Hadley A2

-18

Stowasser et al.

2007

(ref 56)

IPRC Reaonal

Downscale

NCARCCSMl

6kC02

-19

Benetsson et al.
2007 (ref 23)

ECKAM5
timeslice

T213 (-60
km)

2071-2100. A1B

-20

-26

Benetsson et al.

2007

(ref 23)

ECHAM5
timeslice

T319 (-40
km)

2071-2100, A IB

-13

-28

-51

Emanuel et al. 2008
(ref 21)

Stansticai-
determimstic

Downscale 7
QvHP3 mods.:
A1B, 2180-2200
Average over 7

-4

-14

-5

Kmitson et al. 2008
(ref 22}

(SDL Zetac
regional

IS km

Downscale
CMIP3 ens. A1B,
20SQ-2100

-27

Leslie et al. 2007
(ref 57)

OU-CGCM with
hist-ies. window

Up to 50

2000 to 2050
conii'ol and IS92a
(6 members)

Gualdi et al. 2008
(ref 34)

SINTEX-G
coupled model

T106(-120
km)

30 yr IriCOl

2xC02„

4xC02

-16 (2x)
-44

(4x)

-14

-20

-3

-13

Semmler et al. 2008
(refSB)	

Smsby Centre
regional mocel

16 yr conirol ami

A2,20S5-2100

-13

Zhao et al. 2009
(ref 12)

GFDL HIRAM
time dice

50 km

Downstage A1B:
CMIP3 E=1S ens.
GEDL CM2.1
HadCM3
ECHAM5

-20
-20
-11
-20

-14
-14

+5
-17

-32
-33
-42

-39

-5

-62

-1

-29
-5
-12
-52

+15
-23
+61
+35

_2
-43

-2
-25

Sugi et al. 2009
(ref 59)

JMAMRI global
AGCM fameslrce

20 km
20 km
20 km
20 km
60 km
60 km
60 km
60 km

Downscale A1B:

MEICGCM2.3

MSICGCM1J

MIROC-H

CMIP3 le=1S ens.

MKICGCM2.3

MIROC-H

O.HP3 n=lS ens.

C3IFLO

-29

-25

-27

-20

-20

-6

-21

-31

_25

-15
-21
-21
0

-19
-29

-27
-25
-42
-19
-17
-16
-25
-11

+22

+23

-18

+5

+58

+6

+4

-37

-36
-29
+28
-26
-36
-64
-14
+13

-39
-30
-50
-25
-31
-42
-33
-49

-39

-29
+32
-15
-12
+79
+33
-7

Tropical Cyclones Frequency
Projections (Late 21st century) -
Summary

Blue =
decrease

Red =
increase

14

Source: Knutson et al., Nature Geoscience 2010.

-------
Statistical/Dynamical Downscaling Projections: Emanuel et al. (2008)

C)

50
40
10
20
10
0
•10
¦20
-30
-40
¦SO

Chanqe in Power Dissipation

Change in Frequency

Change in Intensity

7.4 0.2 10 8

I n J ill ¦ lli J

3.7 25

J. jil

U IV

^¦|CCSU3
^¦CMRU
1 CSIRO
1 ECHAM
1 OFDL
¦¦MIROC
MRI

1" IB

50
40
30
20

K

U 10
m

< o

|

-10

EastPac	VUsstPac Northlnd SouthHcm

Change in Duration

s

Atl E.Pac W.Pac N. Ind S.H.

-20
-30
-40
-50

| CCSM3
ICNRM
] CSIRO
ECHAM
GFDL
I MIROC
IMRI

JJl

¦=~n

0.7

2 5

2.7

-1.3

2.3

Atl E.Pac W.Pac N. Ind S.H.

Source: Emanual et al. (2008) Bull. Amer. Meteor. Soc

-------
Table S2. Intensity
Projections:

Metric/

Reference

Technique
Model

Resolution
Metric type

Climate Change
scenario

Global

NH

SH

NA4

nw

Pac,
NE
Pat

N
AH.

NW
Pac.

Pat.

NinA

s.

Ind.

S.TV
Pac.

Potential in ten dry or
stat dynamical
projections Change)















Avg
(low,
high)











Vecdbi and Soden 2007
(adapted from lef 60)

Emanuel PL
reversible

XV? diss. b»3fing

Mas Wind
speed (%)

CMIP3 lS-model
A1B (lOOyr trend)

2.6

2.7

2.4

2.1

0.05

(-8 0,
4 .Si

2 9

(-3.1,

12.6)

3.5

(-6,4
16.2)

4.4

(-3.3,
16.0)

3.7

(-7.6.
17.1)

0.99
(-86,
8.6)

Kmrtson and Tuleya
2004

(adapted from lef 9)

Potential Intensity

Emanuel

reversible

Pressure fall

<%)

CMIP2+
+l%'yr C02
SO-yr trend







5.0

2.6

(-5.6,

12.6)

7.0

(-L0.
19.6)

5.4

(-5.0,
21.9)







Krartson and Tuleya

2004

(ref 9)

Potential density,

Emanuel

pseudoadiabatic

Pies sure fall

(%)

CMIP2+
+l%*yr C02
SO-yr trend







7.6

6.0

(1.6,
13.2)

8.5

(28.
252)

8.2
(-3.3,
28.0)







Knutson and Tuleya

:>4

(ref 9)

Potential Intensity.
Holland

Pressme fall
(%)

CMIP2+
+l%^yr C02
SO-yr trend







15.2

12.4
(4.0.
28.9)

17.3

(94.
30.6)

15.8

(3.4,
42.5)







Emanuel et al., 200S
(ref21)

Stat.'Dyii. Model

Max Wind
speed (%)

Q.HP3 7-rnodel
A1B (2181-2200
minus. 1981-2000)

4.5



2.5

6.1

7.4

10.8

0.2

3.7

































Dynamical Model
Projections (Max •wind
speed 0 6 change)



























Kmrtson and Tuleya

2004

(rcf9)

GFDL Hurricane
Model

9km grid

inner nest

CMIP2+
+lWyrC02
SO-yr trend







5.9

5.5
(1-5,
S.l)

5.4

(3.3,6.7)

6.6
(1.1,
10.1)







Kimtson and Tuleya
2004 {Pressure fall)
(ref9)

GFDL Hum cane
Model

9 km grid
inner nest;
Pressure Ml
(•-.)

CMIP2+
+l%'yrC02
SO-yr trend







13.S

13.0

(3 2,
21.6)

13.6

(8.0,

i«i)

14.8

(3.6,
25.0)







Kmtsonet aL 2001
(ref 61)

GFDL Hunicane
Model

IS km gnd
wJ ocean
coupling

GEDLR30
downs cale,
+l%fyrC02yr
71-120 ave

6



















Knutson et al. 20QS
(ref 22)

GFDL Zetac
regional

18 km

Etownscale
OrflP3 ens. A1B,
2080-2100









2.9











Oouchi et aL 2006
(ref 10)

(Average intensity)

MRI'JMA
Tnneslice

TL959 L60
(-20km)

lOy A IB

1982-1993

2080-2099

10.7

8.5

14.1



11.2

4.2

0.6

-12.8

17.3

-2.0

Oouchi et al 2006
(ref 10)

(Average annual
Tn.mminn intensity)

MRI'JMA
Timeslice

IL959 L60
(-20km)

IOvAIB

1982-1993

2080-2099

13.7

15.5

6.9



20.1

-2.0

-5.0

-16.7

8.2

-22 5

Sernmkr et ai 2008
(ref 58)

Rossby Centre
regional model

28 km

16 yr control and
A2, 20S5-2100









44











Walsh etaL 2004
(ref 59)

CSIRODARLAM
regional model

30 km

3xC02; 2061-
2090 minus 1961-
1990



















4-26%
F<970
mb

Bengtsson et al. 2007
(ref 23)

ECHAMS
timeslice

T319(~40
km)

2071-2100, A IB



442%.
#>50

lll'S

















Chauvm et al. 2006
(ref 11)

ARPEGE Climat
Tnne-.Hce

-50 km

Etownscale
-CNRMB2
- Hadlev A2









-0
~0











Stowasser et al. 2007
(ref 56)

IPRC Regional

-50 km

Downscale
NCAR CCSM2,
6xC02











PDI:

4-50%









Leslie et aL 2007
(ref 57)

OU-CGCM with
high-res. window

Up to 50 km

2000 to 2050
control and IS92a
(6 members)



















4-100
%

#>30

m/s

Tropical
Cyclone
Intensity
Projections

Blue = decrease
Red = increase

16

Source: Knutson et al.,
Nature Geoscience 2010.

-------
2) Regional model projects
change ir| hurricane counts
from climate model output.

1)GI
i large
k chan
Hi aero

1) Global climate model projects
large-scale climate changes from
changes in greenhouse gases and
aerosols.

Example of a "double-downscaling" method used to explore frequencies
and intensities of Atlantic hurricanes at high resolution

Geophysical Fluid Dynamics Laboratory/NO A A

3) Hurricane model projects
change in most intense hurricanes
from regional model output.

-------
Late 21st Century Climate Warming Projection- Average of 18 CMIP3 Models

Modeled Category 4 & 5 Hurricane Tracks

Present Climate

Wormed Climate

Atlantic Ocean

North America

Tropical Storm - Category 2
Category 3

Category 4 - Category 5

Atlantic Ocean

Tropical Storms (1981-2005)
histograms of max wind

40°W

0°

100°W 80°W 60°W
(27 Simulated Hurricane Seasons)

20°W	0°	max wind speed (m/s)

Source: Bender et al., Science, 2010

| CONTROL
I WARMED.CM 2
t WARMED.MRI
» VVARMED.MPI
VVARMF.D.HADLEY

The Cat 4-5 increase
is not projected for
all of the 18
individual models:

Tropical Storms (1980-2006)
histograms of max wind









# > CONTROL











# < WARMED.ENSEMBLE





















































































































































max wind speed (m/s)

-------
SUMMARY OF PROJECTED CHANGE

Projected Changes in Atlantic Hurricane Frequency over 21 st Century

125

*->
c

CD
U
+->

LO

fN

i—

CD
>

O

CD
CD

£ -25

_c

u

vp
o^

75

25

-75

Trop. Storm+
Cat. 1 Hurr.

Cat. 2+3
Hurricane

Cat. 4+5
Hurricane

Cat 4+5 frequency:
81% increase, or
10% per decade

Estimated net impact
of these changes on
damage potential:

+28%

• Colored bars show changes for the18 model CMIP3 ensemble (27 seasons); dots
show range of changes across 4 individual CMIP models (13 seasons).

Source: Bender et al., Science, 2010.

-------
Tropical Cyclone Precipitation Rate Projections (Late 21st Century)











Table Si,
TC Precipitation
Projections









1 1 1

Blue = decrease; Red = increase





















Reference

Model/type



Eipemnenr

Basin;

Radius around
storm center

Percent Change















Hasegawa and Eincai

2005"(ref 54)

CCSRTSHESFRC
GC timeslice

T106L56
(~120km)

5x20yat lxC02
7x20v at 2xC'02

NW Pacific

1000 km

-8,4

Yosiimiira et al- 2006
(ref55)

IMA GSM8911
Tune slice

T106L21
(-120km)

lOy

lxC02,2xC02

Global

300 km

—10 (Arakawa-
Schubert)
—1-15 (Kuo)

QuminetaL 2006
(ref 11)

ARPEGE Climat
Timer'i.-g

-50 km

Dtromcale CNRM B2
Downscaie Hadley A2

Atlantic

a'a

Substantial increase

BengCsonet al. 2007
(ref 23)

ECHAM5

tune slice

T213 (-60 km)

2071-2100 . A IB

Not hem
Hemisphere

550 km
Accum. Along
Datt

+21 (all TC s)
+30 (TC > 33 m/s)

KnufcoE et al 2098

(ref 22)

GEDL Zetac
regional

18km

Dcramscaie CM3P3 en;.
A IB. 20SO-2100

Atiaitfic

50 km
100 km
400 km

+37
+23
+10

Knrtson ani Tuleva 200S
(ref 62)

GFDL Hunicane
Model (idealized)

9 Vm inner nest

CMIP2+
+l°VyrC02
SO-yr trend

Atlantic. NE
Pacific. NW
Pacific

-100 km

+22

Gualdi et 3l 20OS
(ref 34)

SENTEX-G
coupled mccel

T106(-120 km)

30 yi laC02,2xCOZ
4xC02

Global

100 km
400 km

+6.1
+2.8

precip CONTROL ASO

	

I I I I I I I I I I I I I I I I I I I
-6-4-3-2-10 1 S 3 4 5



5 -

[US

4 -

3&S

a -

IBS

2 -

94

1 -

»

0 -



-1 -

ft

-3 -

4

—3 -

i

-4 -

a

-5 -

precip TfAEMING ASO

						

B

1W34

ais
EM
ISA

m

32
16

e

d
£

I I I I I I I I I I I I I I I I I I I
5 -4 -a -K -1 {] 1 2 3 4 5

Knutson et al. (2008)
Avq. Rainfall Rate Increases:
50 km radius: +37%
100 km radius: +23%
150 km radius: +17%
400 km radius: +10%

Average Warming: 1.72°C

-------
SUMMARY ASSESSMENT (other storm
characteristics/impacts):

Tropical Cyclone Projections: Genesis.
Tracks, and Duration

We have low confidence in projected changes in
genesis location, tracks, duration, or areas of impact.
Existing model projections do not show dramatic
large-scale changes in these features.

-------
'Possible Rang©' of Projections?

Or, speculations on what could make things worse than projected?

Atlantic Hurricane Acitivity vs. Sea Surface Temperature

500

Based on Absolute SST (1946-2007)

	Annual Observed PDl

Five-year Observed PDl

Five-year PDl based on observed absolute SST;r = 0.79
Statistical Five-year PDl downscaling of global climate models (1946-2100)

	Individual model

Average of 24 models

High-resolution /
model projections (see caption)

11111111H11 111 111 I 11 I 1111111 111 I 111111 III I 111 11II111 111 I 111 1111 111 111)11 I 111T111111111 11111 111 111 I 111111 I III111 111T111 111 I 111 1111111111 I 11 I 1111 11111

1960 1980 2000 2020 2040 2060 2080 2100

500-

400-

300

Based on Relative SST (1946-2007)

—	Annual Observed PDl
Five-year Observed PDl

^ Five-year PDl based on observed relative SST; r = 0.79
Statistical Five-year PDl downscaling of global climate models (1946-2100)

—	Individual model
Average of 24 models

(b)

High-resolution
model projections
(see caption)

A significant statistical
correlation exists between
Atlantic TC power dissipation
and SST since 1950 (top).

A comparable correlation
exists between the power
dissipation and the tropical
Atlantic SST relative to mean
tropical SST (bottom).

These two statistical
relations lead to dramatically
different 'projections' of late
21st century Atlantic TC
activity, ranging from a
dramatic ~300% increase to
little change. The large
(~300%) increase scenario is
not supported by existing
downscaling models
(symbols).

11111111111111111111111111111111111111111111111111111111111111111 H 111 11111111111111111 TTirn [TIT 111111111111111111 11 Ml 111111 III III11 III 1111II1111111 111

1960 1980 2000 2020 2040 2060 2080 2100

-------
'Possible Range' of Projections?

Or. speculations on what could make things worse than projected?

Vertical profile of tropospheric warming:

• Models and theory predict that the vertical profile of tropical tropospheric warming
will amplify with height, while radiosonde-based and some satellite-based observations
suggest that the troposphere has warmed uniformly with height. A uniform warming
with height would be 'de-stabilizing', and would imply future hurricane activity
increases much larger than currently projected (by - 3-4x). Modeling studies and
critical reanalysis of observations (e.g., using winds to infer temperature trends)
suggest that the observed of 'destabilization' of tropical temperatures from
radiosondes and satellites are likely unreliable.

Interanriual Variability
vertical profile

1	2	3

' Scaling ratio Rs(z): s{T(z)} / s{Ts)

Trend (1979-1999)
vertical profile

*7 ¦ i »

1 ),

! i r

\XB

m

\

j \

1



w

J\ /

/ /

/

/

/I

/ /

/ \



/ ' t

iffl/Jh/



/ ' /

mf/\



/





0	1	2 13

Scaling ratio RJz): Trend in T(z) / nend in Ts

Trend (1970-2005) derived
from Radiosonde winds

Trend (1970-2005)
from climate models

15D

Jp
%
e 20

BSC

n



•••



too

\

r







ISO
200











300











500











850

Radiosondes,
models and theory

Radiosondes

Models and theory

0 30
Lartftuc)? i )

60 90 -60 -30

23

Sources: Sariter et al. 2005; Alien and Sherwood 2008

-------
'Possible Range' of Projections?

Or. speculations on what could make things worse than projected?

Seasonal Wind Shear Anomaly	Potential Intensity Anomaly

100°W 80°W 60"W 40°W 20°W 0"	100°W 80°W 60"W 40°W 20°W 0°

SST Anomaly (Tropical-mean SSTA)

Wind shear anomaly (m/s)

1 1.4 1.8 2.2 2.6 3
SST Anomaly (K)

The range of possible
projections could be even
broader than inferred
from the AR4 models
(sample of 4 models
shown at left):

• IPCC AR4 models favor a
weak El Nino-like signature to
the pattern of 21st century
warming, and strongly favor
enhanced vertical wind shear
over the Caribbean and
tropical Atlantic. However,
some models project little
change in wind shear and
some researchers (Cane et
al.) argue that the Pacific
warming signal will be
distinctly La Nina-like, which
could substantially impact
Atlantic hurricane projection^.

-------
'Possible Range' of Projections?

Or, speculations on what could make things worse than projected?

Lower stratospheric temperatures:

• Can lower stratospheric or tropopause transition layer (TTL) temperatures
(apparently cooling) affect tropical storm frequency or hurricane intensity? Emanuel
statistical/dynamical downscaling: yes for both. Current GFDL dynamical models: no
for tropical storm frequency, not clear for intensity (upper tropospheric temperatures
affect hurricane intensity in the GFDL models). Also, are NCEP potential intensity
trends since 1980 reliable or do they suffer from inhomogeneity problems?

Potential Intensity trends since 1980
from NCEP Reanalysis

iih

1S4S

1«0

IMS



2M5

Statistical/Dynamical Downscaling of Atlantic
Tropical Storm Frequency (1870-2005)

J0<

25



15

1&

S«s;l Track
EtHflM
PJOAWCIRES
mCAFfJICEF

1 iS6 ISflG 1306 *313 I-94G IflSfl 19id	215 C-

Vesr

Source: K. Emanuel. AMS Hurricanes and Tropical Meteorology Conference abstract, 2010.

-------
'Possible Range' of Projections?

Or, speculations on what could make things worse than projected?

Tropical cyclone-induced changes in ocean heat transport:

• Possible role of tropical cyclones in 'equable' climates of 3-5 million years
ago being investigated, but implications for this mechanism on climate for
next century or so remain highly speculative. Tropical cyclones cause less
than 10% of global poleward heat transport in the current climate,
according to the latest studies.

Tracks in modem climate

b Increasing CO,, and subtropical mixing

strength

Tracks in early Pliocene ctimate

Longitude (degrees)

Figure 2! The tracks of tropical cyclones simulated by the SDSM. a. In the

modem climate, and b, in the early Pliocene climate. The colours indicate
hurricane strength—from tropical depression (TD; Hue) to tropical storm

20

120 W

90 H

Longitude

4 ->20

2 C Increasing GO* only

20

20 S

150 W
? jctegrees;.

(TSjcyan) to category-5 hurricanes (red). The tracks shown in each p
a two-year sub sample of 10,000 simulated tropical cyclones.

Figure 4 i SST changes in the tropical Pacific simulated by the coupled

-------
Overview of Assessments

Climate Change Detection and Attribution:

•	It remains uncertain whether past changes in tropical cyclone activity

exceed natural variability levels.

Projections for iate 21st century:

•	Likely fewer tropical storms globally (~no change to -34%), with even

greater uncertainty in individual basins (e.g., the Atlantic).

•	Likely increase in average hurricane wind speeds globally (+2 to11 %),

though not necessarily in all basins

•	More frequent very intense storms (> 50% chance these will increase

by a substantial percentage in some basins).

•	Likely higher rainfall rates in hurricanes (roughly +20% within 100 km

of storm)

•	Sea level rise is expected to exacerbate storm surge impacts even

assuming storms themselves do not change.

-------
Emergence Time Scale: If the observed Cat 4+5 data
since 1944 represents the noise (e.g. through bootstrap
resampling), how long would it take for a trend of -10%
per decade in Cat 4+5 frequency to emerge from noise?
Answer: ~60 yr (by then 95% of cases are positive)

28

Number of Cat 4+5 Atlantic Hurricanes

With Emanuel Adjustment for Early Storm Intensities

1940

1950

1960

1970 1980
Year

1990 2000 2010

Instead, assume
residuals from a
4th order

polynomial: 55 yr

Instead, resample
chunks of length
3-7 yr: 65-70 yr

U.S. Landfalling Cat 4-5 hurricanes (1851-2008)

Line: 20-yr running mean; source: HURDAT

1X60 1880 1900 1920 1940 I960 1980 2000
Year

Source: Bender et al., Science, 2010.

-------
The Impact of Climate Change on Tropical Cyclone Damages

Robert Mendelsohn
Yale University

Kerry Emanuel

MIT

Shun Chonabayashi
Cornell University

Abstract

This paper constructs an integrated assessment model of tropical cyclones in order to quantify
the additional damage that climate change might cause by 2100 around the world. The paper
begins with the Alb SRES emission trajectory from the Intergovernmental Panel on Climate
Change (IPCC). The trajectory approaches a stable concentration of greenhouse gases of about
720ppm by 2100. This emission trajectory is used in four general circulation climate models:
CNRM, ECHAM, GFDL, and MIROC. The climate models are used to predict hurricanes in the
1980-2000 climate and the 2180-2200 climate. The models predict a range of future global
temperature changes from 2.9°C to 4.5°C. The climate outcomes from these models are then fed
into a regional climate model that is capable of predicting hurricane behavior in each ocean
basin.

A tropical cyclone generator creates potential hurricanes randomly in each basin for both the
current climate and the future climate. The model follows these storms across each ocean and
determines which storms develop into hurricanes and which do not. A total of 17,000 tropical
cyclones are generated in each of the 8 climate scenarios (current and future climate with each of
4 climate models). The model does a reasonable job of predicting the frequency, intensity, and
location of the hurricane distribution observed in the current climate.

We detect the influence of climate change by comparing the results of current predicted
hurricanes versus future predicted hurricanes. Except in the GFDL scenario which predicts a

-------
doubling of hurricanes, the frequency of hurricanes is not predicted to change because of
warming. However, in the western North Atlantic and the western North Pacific, hurricane
intensity consistently increases in all four climate scenarios. In the other ocean basins, the change
in hurricane intensity is inconsistent, sometimes increasing and sometimes decreasing across the
climate scenarios. Whether climate change has an effect on hurricane intensity in the other basins
is therefore highly uncertain.

This paper advances on the underlying science of hurricanes by examining the damages that
hurricanes might cause. Beginning with observed hurricanes and observed levels of damages, the
paper calculates a damage function of tropical cyclones. Using US data, the study calculate a
relationship between storm damages and storm intensity. The analysis confirms earlier results
suggesting a highly nonlinear relationship between storm intensity and storm damages. The
study finds that the minimum pressure of a storm is a better predictor of damages than maximum
wind speed (the measure used in the damage literature). Using international data, the study then
examines the link between storm damages and the characteristics of the affected areas. The study
finds that damages increase with both income and population density as expected. However,
damages increase less than proportionally with both variables contrary to assumptions made by
previous authors.

Current predicted tropical cyclone damages are then calculated using the current income and
population density of each country and the distribution of tropical cyclones predicted in the
current climate. Damages are calculated for each storm that is predicted to strike each country or
come close enough to cause significant damage. The expected value of damages is adjusted to
equal the observed damages in each country over the last 20 years. The global damages from
tropical cyclones are currently $26 billion (0.043 percent of Gross World Product (GWP)).

-------
The analysis then calculates what damages would occur if income and population density were to
increase as projected by 2100. Given the projected growth rates for each country, damages are
calculated again using the current distribution of tropical cyclones. Global damages are projected
to double to $55 billion just because of the increase in income and population in each country.
Damages increase faster in Asia because of the projected faster growth rate of income in that
region. Tropical cyclones damages as a fraction of GWP are expected to fall by 2100 to 0.01
percent of GWP because damages increase less than proportionally with income.

The final step in the analysis is to compute the effect of climate change. The damages from the
future economy with storms from the current climate are compared to the damages from the
future economy with storms from the future climate. Warming doubles the global damage caused
by tropical cyclones. Warming causes an additional $54 billion of damage per year (0.01 percent
of GWP). Looking across the different climate models, warming increases damages between
$28 and $68 billion/yr.

The effects are not uniformly felt across the world. The increase in storm intensity in the North
Atlantic and North Pacific lead to substantial damages in the northeastern edge of the Western
Hemisphere and in the eastern edge of Asia. The United States, China, and Japan account for 88
percent of the expected global damages. The countries with the biggest damages as a percent of
GDP are predominantly small islands in the Caribbean.

Warming also changes the distribution of damages. With current climate, the top 10% worst
storms (measured by damage) cause 90% of the total damage and the top 1% worst storms cause
58%) of the damage. With the future climate, the top 10%> worst storms cause 93% of the damage
and the top 1% of storms cause 64% of all the damage.

-------
Further work is required to fully understand how to adapt to tropical cyclones. There are three
mechanisms that cause damage: storm surge, high winds, and flash flooding. Further research is
needed to predict the consequences of each of these mechanisms and how different adaptation
strategies might change the distribution of damages. The fact that so much of the damages are
concentrated in rare (once in a century or millennium) but very powerful events makes
adaptation difficult.

-------
Capsule

Climate science and economics are combined to estimate the future tropical cyclone damages
from economic growth and from climate change.

-------
Main Text

I. Introduction

Tropical cyclones (hurricanes, typhoons) have become an icon of climate change.
Scientists report an increase in tropical cyclone intensity over the last 30 years1'2 and a dramatic
increase in tropical cyclone damages over time3'4. And yet despite these findings, the link
between climate change and tropical cyclone damage remains controversial. Tropical cyclones
are rare events and appear to be subject to long term variability so it is difficult to detect changes
in underlying frequencies and severity5. The people and assets in harm's way is also increasing
over time, which may explain the trend in tropical cyclone damage6'7'8. The historic record may
simply not be long enough and clear enough to detect how climate may be affecting tropical
cyclones.

The average current global damage from tropical cyclones is $26 billion per year9. A
tropical cyclone model predicts there will be an increase in tropical cyclone intensity in the
Atlantic Ocean10. Using this average change in intensity, several authors predict that damage will
double11'12'13.

This paper takes a different approach that captures the full range and distribution of
tropical cyclones to estimate the impact of climate change on the damages caused by tropical
cyclones. The paper relies on an integrated assessment that combines the insights of a hurricane
generator with the consequences of a damage model. Beginning with an emissions trajectory,
four climate models predict future climate scenarios. A tropical cyclone generator is then used to
seed potential storms in each ocean basin10. The storms are then permitted to develop given the

-------
conditions predicted by each climate model. A total of 17,000 storms are generated with and
without climate change. The model is able to capture the different outcomes in each ocean basin
and measure how storm intensity and location changes. For each storm, a damage model is then
used to predict the damages upon landfall.

The analysis begins by forecasting how current baseline damages from tropical cyclones
would change because of future increases in what is in harm's way. From this future baseline, we
then evaluate the effects of climate change. We predict how the change in tropical cyclones
generated by the current versus future climate affect damages. The analysis captures changes in
the frequency of storms, the landfall of storms, and the intensity of storms. The analysis carefully
controls for what is in harm's way before estimating the impact of climate change. The results
provide the first geographically detailed estimates of how storm damages change around the
world.

The next section of the paper describes the methodology in more detail. The empirical
findings of the paper are then reviewed in Section III. The paper concludes with a review of the
major findings and some policy observations.

II. Theoretical Methodology

The economic damage (D) from each tropical cyclone is the sum of all the losses caused by
it. In this analysis, we focus on economic damages primarily from lost buildings and
infrastructure. In order to model tropical cyclones, it is critical to recognize that they are rare
events and depend on the frequency or probability (n) of each storm in each place. The

-------
characteristics (X) of each storm are also important. Damages also depend upon where the
tropical cyclone strikes (/). Atmospheric science can help predict the probability a tropical
cyclone (j) with particular characteristics (X) will strike each place (J) given the climate (C):

7T,

= *(Xt,,C)

(1)

The actual damage associated with any given tropical cyclone (j) also depends on the
vulnerability (Z) of each place (/):

A =D(X1,Z1)

(2)

The expected value of tropical cyclone damages is:

E[D] = Y^X,.CWX,.Z,)	(3)

Because the damage function is highly nonlinear, the expected damage is the sum of the damages
caused by every storm. It is very important to model the entire distribution of damages in order
to capture the true effect of tropical cyclones.

-------
The damage caused by moving from the current climate CO to a future climate CI is the
change in the expected value of the damages:

W = E[D{C\)\ - E[D(C0)]	(4)

For any given time period, climate change could change damages because the frequency,
intensity, or the location of storms change. In this study, we compute tropical cyclone damages
in each country of the world and then aggregate the results to larger regions. Country specific
results are reported in Appendix A.

Equation 4 calculates the expected welfare loss from climate change. We also calculate
the return rate for storms causing each level of damage. This is a relationship between the
average years between tropical cyclones that cause specific amounts of damage:

return = 1 / prob(D) = g(D(X))	(5)

Policy makers may be interested in the return rate because it provides information about the
distribution of damages. Insurers would also be interested in the return rate because it provides
information about how much catastrophic insurance to buy.

III. Methodology

-------
The integrated assessment predicts tropical cyclone damages given different climates.
The analysis relies on the A1B SRES emission scenario generated by the Intergovernmental
Panel on Climate Change14. The emission scenario assumes that mitigation is tightened gradually
over time so that greenhouse gas concentrations finally peak and stabilize at 720 ppm.

We rely on four climate models: CNRM15, ECHAM16, GFDL17, and MIROC18. Each
climate model predicts both the current climate and the climate in 2100. CNRM predicts a global
warming of 2.9 C, ECHAM predicts 3.4 C, GFDL predicts 2.7 C, and MIROC predicts 4.5 C.
These changes in climate raise sea surface temperatures which in turn fuel the tropical cyclones.
However, there are other changes such as wind shear and humidity that can affect tropical
cyclone intensity. In addition, changing atmospheric winds can alter the tracks of tropical
cyclones.

Using a tropical cyclone generator in each ocean basin, the climate data is used to project
tropical cyclone tracks10. A total of 17,000 tropical cyclone tracks are generated across the five
oceans with and without climate change for each climate model (8 sets of 17,000 tropical
cyclones). For each track, we follow where the tropical cyclone makes landfall or passes close
enough to land to create damage. The minimum barometric pressure and the maximum wind
speed at landfall of each storm are recorded. The hurricane generator also predicts the expected
frequency of tropical cyclones in each ocean basin.

Figure 1 presents a sample of the tracks generated in each ocean basin. The figure reveals
that there is a zone just north and south of the equator where the storms are the most intense. As
storms veer off to middle and high latitudes, they tend to lose power.

-------
Figure 2 shows the changes in power by ocean basin attributable to climate change.

Power consistently climbs in the North Atlantic and the North Western Pacific ocean basins
across all four climate models. These predicted changes in tropical cyclone power will especially
influence damages in North America and eastern Asia respectively. Changes in the other ocean
basins are not consistent across the climate models.

A damage function is then estimated to predict the damages that each storm will cause.
The coefficient for storm intensity was estimated using aggregate damages per storm and storm
characteristics at landfall from US storms since I96019. This historic data was matched with
coastal population density and income20 near landfall. The log-log regressions in Table 1 reveal
the elasticities of each variable with respect to damage. The first two columns using US data
reveal that damages are a highly nonlinear function of wind speed and minimum barometric
pressure. The regressions also reveal that minimum pressure provides more accurate estimates of
damages than maximum wind speed. It is likely that minimum pressure is a better predictor of
storm intensity because it is difficult to measure maximum wind speed accurately. The literature

11 12 13 21

relies solely on wind speed to measure damages ' ' ' .

The third column of Table 1 presents a damage regression using international data9.
Damages increase with income but fall with population density. The elasticities of these
variables are significantly less than 1 (contrary to assumptions in the literature6'8'11'12). Storm
damages consequently do not rise proportionally with income or population.

Given the empirical results above, the preferred damage function has the following form:

D = Ad *MP-S6Y0 06Pop-°2

(6)

-------
The expected damages for each country were calculated by summing the product of the
probability of each storm times the damage it causes for each country. Storm damages are
truncated so that they cannot exceed the complete destruction of all the capital in the five
counties near landfall. The parameter AD is calibrated for each country so that the damage from
the current predicted set of storms striking each country equal the observed damage in recent
history.

IV. Results

The annual observed global damage from tropical cyclones is $26 billion (0.043 percent
of GWP)9. Our first task is to project how these damages would increase with future economic
growth, holding climate constant. Both population and income are projected to 2100. The
population in each country is assumed to follow projections that lead to a global population of 9
billion22. GDP is assumed to grow an annual rate of 2 percent in developed countries, 2.7 percent
in developing countries, and 3.3 percent in emerging countries. Dividing GDP by population
yields a future prediction of income per capita for each country in 2100. The damages from the
set of storms given current climate are then recalculated using the damage function and future
levels of population density and income. With future baseline conditions in 2100, the global
expected damage more than doubles to $55 billion per year (0.01 percent of GWP). Damage
grows more slowly than GDP because the coefficients on income and population in the damage
function are less than 1.

-------
In order to calculate the impact of climate change, a new set of tropical cyclones is
generated given the 2100 climate predicted by each of the climate models. The impact of climate
change is the difference in damages caused by the new set of cyclones versus the original set of
cyclones. Both measures of damage use future economic conditions. By evaluating the impact of
climate change using future conditions rather than current conditions, the impacts are larger
because more is in harm's way in the future. Damages are computed from all 17,000 storms
before and after climate changes.

The results reveal that climate change by 2100 is expected to cause tropical cyclone
damages to increase $54 billion/yr (a 100% increase above the future baseline). This additional
damage is equal to 0.01 percent of GWP. Looking across the different climate models, damages
rise between $28 and $68 billion/yr. These aggregate global results are very consistent with most
of the findings in the literature that climate change would double tropical cyclone damages.

However, the new results reveal that the distribution of damages is not even across the
world. Figure 3 displays the damages caused by climate change in each continent. Asia and
North America are the two continents that are consistently predicted to be damaged by warming
across all four climate models. The increased intensity of North Atlantic and Western Pacific
storms are causing these effects. The additional damage in North America is equal to $30
billion/yr and the additional damage in Asia is equal to $21 billion. The additional storm damage
in the rest of the world is just $3 billion. For some regions and models, the damages from
tropical storms actually fall with warming.

Figure 4 displays climate change tropical cyclone damages as a fraction of GDP in 2100.
The figure illustrates how burdensome the change in tropical storm damage will be to the

-------
economies in each region. The global average damage per unit of GDP is 0.01 percent. North
America (0.03 percent of GDP) and Asia (0.01 percent of GDP) have the largest additional
impacts per unit of GDP. The tropical cyclone damages per unit of GDP caused by climate
change are low in the remaining continents.

The continental averages, however, hide disproportionate effects in individual countries.
Damages to all affected countries and each model are shown in Appendix A. The countries with
the largest average impacts from climate change are the United States ($30 billion), Japan ($9
billion), and China ($8 billion). The damages from these three countries account for 88 percent
of the global damages. The impacts are above 0.2 percent of GDP only in Antigua-Barbados,
Cayman Islands, Dominica, Grenada, Honduras, Montserrat, St. Kitts-Nevis, Turks-Caicos, and
the US Virgin islands. All but Honduras is an island in the Caribbean.

Although expected damages reveal long term damages, they hide changes in the
distribution of damages. Figure 5 displays the relationship between damage and return rates for
the GFDL climate scenario. The results for the other climate scenarios are similar. The figure
reveals that common small storms are not different before and after climate change. Climate
change increases the intensity of large storms. With the nonlinear damage function, this
increased intensity translates into a significant increase in damages. The return period for the
most powerful storms becomes shorter.

A surprisingly large fraction of the expected damages of tropical cyclones is caused by
the most harmful storms. With current climate, the top 10% worst storms (measured by damage)
cause 90% of the total damage. The top 1% worst storms cause 58% of the damage. With the

-------
future climate, the top 10% worst storms cause 93% of the damage and the top 1% of storms
cause 64% of all the damage.

V. Conclusion

This study constructs an integrated assessment model to predict the tropical cyclone
damages caused by climate change. The paper relies on a tropical cyclone generator and four
climate models to predict thousands of tropical cyclones with and without climate change. The
results indicate that tropical cyclone intensity will consistently increase in both the North
Atlantic and North West Pacific ocean basins but not in the other ocean basins. The study then
estimates a damage function associated with tropical storms from United States and international
data. The analysis suggests that minimum pressure provides a more accurate measure of storm
intensity than maximum wind speed and that damages are a highly nonlinear function of storm
intensity. The results also suggest that damages increase with income but less than
proportionally.

Increasing future income and population is predicted to increase annual tropical cyclone
damages from $26 billion to $55 billion even with the current climate. However, damages as a
fraction of GWP are expected to fall from their current rate of 0.04 percent in 2010 to 0.01
percent in 2100.

The impact of climate change is expected to double the damages from tropical cyclones
by 2100 by $54 billion. This is equal to 0.01 percent of GWP. The estimated impact of climate

-------
change ranges from $28 to $68 billion depending on the climate model. The findings confirm the
rough results of earlier tropical cyclone studies that relied on cruder methods.

The damages, however, are not evenly spread across the planet. Because tropical
cyclones in the North Atlantic and North West Pacific Oceans consistently increase in intensity
with warming, North America and eastern Asia have the largest and most consistent impacts.
The average impact in Asia is an additional damage of $21 billion and the average impact in
North America is an additional damage of $30 billion. Damages to the United States, Japan and
China account for 88% of global damage. Climate change causes small damages in the rest of the
world because the remaining continents see both small harmful and beneficial impacts depending
on the climate model. Even controlling for GDP, North America and eastern Asia bear the
highest damages per unit of GDP. However, the most vulnerable countries are relatively small
Caribbean islands.

The results reveal that the damages from tropical cyclones are quite skewed. Even with
the current climate, the 10 percent worst storms (measured by damage) account for 90 percent of
the total damage. With warming, these powerful storms get even more harmful and the 10
percent worst storms account for 93 percent of the total damages. These especially large storms
explain most of the damages caused by climate change and yet they occur very rarely. It may
well take several centuries of observations to see whether the changes predicted in this paper
actually have occurred.

There are many uncertainties associated with the forecasts made in this study. The
emission path of greenhouse gases is highly uncertain because it depends upon the long term
growth of the economy, the long term relationship between GDP and energy, and mitigation

-------
policies that may be adopted over the next century. The relationship between climate change and
greenhouse gas concentrations is also quite uncertain as revealed by the results from the four
climate models. Exactly how tropical cyclones will react to climate change is also uncertain as it
depends upon many factors that are difficult to predict. The magnitude of the damages that future
tropical storms will cause is uncertain. The damages with respect to storm intensity are very
sensitive to minimum pressure and to the elasticity of population and especially income. Better
international records of storm tracks and intensities and storm damages would help increase the
accuracy of these estimates. How damages might change if there is both a change in tropical
cyclones and sea level rise is uncertain (although they may be just additive23).

Finally, how society will adapt to tropical cyclones in the future is not yet clear.

Currently, many countries have mal-adaptation policies that make matters worse by encouraging
assets to remain or be placed in harm's way. For example, subsidizing flood insurance and
capping the cost of catastrophe insurance makes it cheaper to live in risky locations. Even
providing emergency disaster relief reduces the overall cost of developing a risky location.
Reducing the implicit subsidies in these policies and actively discouraging development in risky
locations could reduce damages significantly. In contrast, physical protection strategies such as
building sea walls may be ineffective as protection against tropical cyclones. Most of the damage
is caused by rare but very powerful storms. Walls would have to be very high to prevent
inundation. These would be hard to justify if powerful storms are very infrequent at each location
(once in a thousand years). Developing effective adaptation strategies to tropical cyclones is an
important policy and research topic.

-------
Acknowledgement

This paper was commissioned by the Joint World Bank - UN Project on the Economics of
Disaster Risk Reduction and funded by the Global Facility for Disaster Reduction and Recovery.
The findings, interpretations, and conclusions expressed in this paper are entirely those of the
authors. We are grateful to William Nordhaus, Apurva Sanghi, Michael Toman and seminar
participants at the World Bank, Yale University, and United Nations for valuable comments and
suggestions.

-------
References

1.	Emanuel, K. 2005. "Increasing destructiveness of tropical cyclones over the past 30 years,"

Nature, 436: 686 - 688.

2.	Intergovernmental Panel on Climate Change. 2007. The Physical Science Basis, Cambridge

University Press, Cambridge UK.

3.	Intergovernmental Panel on Climate Change. 2001. Climate Change 2001: Impacts,

Adaptation and Vulnerability, Cambridge University Press, Cambridge UK.

4.	Swiss Re, 2006. 'The effect of climate change: storm damage in Europe on the rise', Focus

report.

5.	Landsea, C. W., B. A. Harper, K. Hoarau, and J. A. Knaff 2006. "Can we detect trends in

extreme tropical cyclones?" Science 313: 452 - 454.

6.	Pielke, R. A. Jr. and C. W. Landsea. 1998. "Normalized tropical cyclone damages in the

United States: 1925-1995," Weather and Forecasting 13: 621-631

7.	Pielke, R. A. Jr. 2005. "Are there trends in tropical cyclone destruction?" Nature 438: El 1

8.	Pielke, R. A. Jr. J. Gratz, C. W. Landsea, D. Collins; M. A. Saunders; and R. Musulin. 2008.

"Normalized tropical cyclone damage in the United States: 1900-2005. Natural Hazards
Review, 9: 1-29.

9.	EMDAT. 2009 "The OFDA/CRED international disaster database" www.emdat.be Universite

Catholique de Louvain, Brussels, Belgium

10.	Emanuel, K, R. Sundararajan, and J. William, 2008. "Tropical cyclones and global warming:

Results from downscaling IPCC AR4 simulations" Bulletin American Meteorological
Society 89: 347-367.

11.	Nordhaus, W. 2010. "The economics of hurricanes and implications of global warming"

Climate Change Economics 1: 1-24.

-------
12.	Pielke, R.A. Jr. 2007. "Future economic damage from tropical cyclones: sensitivities to

societal and climate changes" Philosophical Transactions Royal Society 365: 1-13.

13.	Narita, D., R.S.J., Tol, and D. Anthoff, 2008. Damage costs of climate change through

intensification of tropical cyclone activities: An application of FUND, Climate
Research 39: 87-97.

14.	IPCC (Intergovernmental Panel on Climate Change). 2000. Special Report on Emissions

Scenarios, Cambridge University Press, Cambridge, UK.

15.	Gueremy, J.F., M. Deque, A Braun, J.P. Evre. 2005. "Actual and potential skill of seasonal

predictions using the CNRM contribution to DEMETER: coupled versus uncoupled
model" Telhts 57: 308-319

16.	Cubasch, U, R. Voss, G. Hegerl, J. Waskiewitz, and T. Crowley. 1997. Simulation of the

influence of solar radiation variations on the global climate with an ocean-atmosphere
general circulation model" Climate Dynamics 13: 757-767.

17.	Manabe, S., J. Stouffer, M.J. Spelman, and K. Bryan. 1991. "Transient responses of a

coupled ocean-atmosphere model to gradual changes of atmospheric C02. Part I: mean
annual response" Journal of Climate 4: 785-818.

18.	Hasumi, H. and S. Emori. 2004. K-l Coupled GCM (MIROC) Description, Center for

Climate System Research, University of Tokyo, Tokyo.

19.	National Tropical Cyclone Center. 2009. "Tropical cyclone reports" National Oceanic

Atmospheric Administration, http://www.nhc.noaa.gOv/pastall.shtml#tcr

20.	United States Census of Population, 1960-2000, Department of Commerce, Washington D.C.

21.	Hallegate, S. 2007. "The use of synthetic hurricane tracks in risk analysis and climate change

damage assessment", Journal of Applied Meteorology and Climatology, 46: 1956-1966.

22.	United Nations. 2004. World Population in 2300 Department of Economic and Social

Affairs. New York.

-------
23. Nicholls, R.J. et al. 2008. "Ranking port cities with high exposure and vulnerability to
climate change" OECD Environment Working Paper No 1, Paris, France.

-------
Figure and Table Captions

Table 1: Regressions of Tropical Cyclone Damages
Figure 1: Storm tracks by minimum pressure (mb)

Figure 2: Change in Tropical Storm Power by Ocean Caused by Climate Change
Figure 3: Climate Change Impacts on Tropical Cyclone Damages by Region by 2100
Figure 4: Tropical cyclone damage in 2100 as a fraction of GDP
Figure 5: Return period in 2100 by US damage for ECHAM

-------
Table 1: Regressions of Tropical Cyclone Damages



US

US

International

Constant

12.19

607.5

15.17



(1.42)

(10.39)

(22.77)

Log (Wind Speed)

4.95
(7.83)





Log(Minimum



-86.3



Pressure)



(9.96)



Log(income)

0.903

0.370

0.415



(0.96)

(0.45)

(6.44)

Log(Population

0.458

0.488

-0.210

Density)

(1.28)

(1.53)

(3.04)

Adj Rsq

0.371

0.501

0.158

F Statistic

22.61

35.76

103.2

Observations

111

111

807

Note: T-statistics in parentheses. The functional form of the regression is log log. Source of US
data is NOAA 2009 and the source of the international data is EMDAT 2009.

-------
Figure I Storm tracks by minimum pressure (mb)

90N
70N
50 N
30N
10N
10S
30S
50S
70S

50E 70E 90E 110E130E150E170E170WI50W130W110W90W 70W 50W 30W 10W 10E 30E

860
880
900
920
940
960
980
1000

-------
Figure 2: Change in Tropical Storm Power by Ocean Caused by Climate Change



80



70

k.

0>

60

5



o



CL

50

"ro



TS
C

40

ro



-1



c

30

0)



O)



1=



ra

20

O



A-*



c

10

o



-------
Figure 3: Climate Change Impacts on Tropical Cyclone Damages by Region by 2100

Q







Q>

<<>

run

J-&	/b	. r

//







d

O



w

~	CNRM
¦ECHAM

~	GFDL

~	MIROC

Note: Calculated using minimum pressure damage model with future baseline.

-------
Figure 4: Tropical cyclone damage in 2100 as a fraction of GDP

Q.
O

0

c

a)
o

a)
a.

0.0700
0.0600
0.0500
0.0400
0.0300
0.0200
0.0100
0.0000
-0.0100





o

9>

:<&¦

Jd







j=U=H





&













&
&



0

U-1

&

cr

r?>

&

~	CNRM
¦ ECHAM

~	GFDL

~	MIROC

-------
Figure 5: Return period in 2100 by US damage for ECHAM

ECHAM Model

Return Period (Years)

Note: Damage calculated in 2100.

-------
The Impact of Climate Change

on Extreme Events

Robert Mendelsohn

Kerry Emanuel
Shun Chonabayashi
Gokay Saher

-------
Acknowledgements

•	Funding by World Bank-United Nations

•	Global Facility for Disaster Reduction &
Recovery

•	Report:

-	Natural Hazards, UnNatural Disasters: The
Economics of Effective Prevention

-	Apurva Sanghi

-------
Goal

•	Measure how climate change affects
future extreme events

•	Reflect any underlying changes in
vulnerability in future periods

•	Estimate damage functions for each type
of extreme event

•	Measure future extreme events caused by
climate change

-------
Extreme Events Examined

•	Cold Events

•	Drought

•	Flood

•	Hail

•	Heatwaves

•	Tornadoes

•	Thunderstorms

•	Tropical cyclones

•	Severe Storms
(extratropical)

-------
Forecast Future Baseline

Impacts

•	Estimate damage function with EM DAT
data

•	Use forecasts of future income and
population

•	Forecast how damages and deaths will
change as income and population
increase

-------
Current and 2100 Baseline Impacts of
Extreme Weather Events

Cold

Drought

Flood

Heat
Wave

Local
Storm

Severe
Storm

Tropical
Cyclone

*2010
~ 27 00

10

16

19
49

31

8

16

26
55

-------
Extreme Event Damages by

Region



Africa

Asia

Europe

Latin
America

North
America

Oceania

V2010

0

21

10

4

22

1

U2100

2

74

49

11

46

2

-------
Current and Future Deaths by

Extreme Event

~(/>

C3
CD

D

20000 ^
18000
16000
14000
12000
10000
8000
6000
4000
2000
0

Cold

Drought

Flood

Heat
Wave

Local
Storm

Severe
Storm

Tropical
Cyclone

2010 716

234

7508

4598

295

451

19486

U2100 215

25

3898

6123

239

287

7180

-------
Past Climate Results

•	IPCC 1996 guesses CC increases US tropical
cyclone damages by 0.02% of GDP and world
damages of 0.002% of GWP

•	Nordhaus 2006 estimates CC doubles US
tropical cyclone damages (0.06% of GDP)

•	Narita et al 2007 estimate CC doubles world
tropical cyclone damages by 0.006% GWP

•	Stern guesses CC increases all extreme event
damages by 5% of GWP

-------
Integrated Assessment Model

Emissions Trajectory

i

Climate Scenario

1

Event Risks

Vulnerabi

ity Projection

Damage Function

i

Damage Estimate

-------
IPCC Emissions Scenarios

SRES C02 CONCENTRATIONS : ILLUSTRATIVE SCENARIOS AND FULL RANGE



1100



1050



1000



950

£

900

s.



850

z

S

800

K

§

750

H



Z

til

700

U

650

O

600

u

Osl



o

550

u



500



450



400



350

1990 2000 2010 2020 2030 2040 2050 2060 2070 2000 2090 2100

YEAR

This

^study

-------
o

o

cn

4.0 -

3.0 H

¦

E

| 2.0 H

-------
Using Climate Models to Estimate Changing

Incidence of Extreme Events

•	Some events can be inferred directly from
climate model output

-	Heat waves and cold snaps

-	Droughts and certain kinds of floods

-	Large-scale non-tropical wind storms

•	Some must be inferred indirectly, by using
sub-models (e.g. "downscaling")

-Tropical cyclones, thunderstorms, tornadoes,
hail storms

-------
Climate Models

•	CNRM

•	ECHAM

•	GFDL

•	Ml ROC (tropical cyclones only)

-------
Tropical Cyclone Generator

•	Step 1: Seed each ocean basin with a very large
number of weak, randomly located cyclones

•	Step 2: Cyclones are assumed to move with the
large scale atmospheric flow in which they are
embedded, taken from the global climate model

•	Step 3: Run a detailed cyclone intensity model
for each event, and note how many achieve at
least tropical storm strength

•	Step 4: Using the small fraction of surviving
events, determine storm statistics.

Details: Emanuel et al., Bull. Amer. Meteor. Soc., 2008

-------
Tropical Cyclone Power by

Ocean Basin



80



70



60

§;

O



o_

50

15



£Z

40

TO



—I



¦E

30

-------
US Tropical Cyclone Damage

Function



Constant

Minimum
Pressure

Income

Populat.
Density

Damage
Model

607.5
(10.39)

-86.3
(9.96)

0.370
(0.45)

0.488
(1.53)

Fatality
Model

247.5
(4.10)

-33.3
(3.69)

-2.35
(1.74)

1.28
(2.78)

-------
International Tropical Cyclone
Damage Function



Constant

Income

Populat.
Density

Damage

15.17
(22.77)

0.415
(6.44)

-0.21
(3.04)

Fatality

6.25
(18.20)

-0.477
(14.01)

0.07
(1.86)

-------
Baseline Tropical Cyclone
Global Damages

•	Current Damages: $26 billion/yr

•	Future Damages: $55 billion/yr

•	Current Global Deaths: 19,500/yr

•	Future Global Deaths: 7,200/yr

•	Change in 2100 because of higher
population and income

•	Current climate for baseline estimates

-------
Climate Change Impacts on
Tropical Cyclones in 2100



CNRM

ECHAM

GFDL

Ml ROC

Damages
(billions)

$19.9

$54.7

$70.2

$69.7

Deaths

760

-1500

-3300

1300

-------
Impact of Climate Change on
Regional Cyclone Damage

G
w
D

W
£
O

m

70
60
50
40
30
20
10
0

Z

^=7

1

m

- # vj -

Africa

Asia

Europe

Latin America

North America

Oceania

m CNRM

0

10

0

3

15

0

~ ECHAM

0

25

0

-1

32

0

a GFDL

0

4

0

4

61

0

&MROC

0

46

0

5

13

0

-------
Climate Change Cyclone
Impacts as a Percent of GDP

0.0350

0.0300

0.0250

Q 0.0200
O

o 0.0150

./
./

./

2 0.0100 f

IL

0.0050
0.0000
-0.0050

Africa

Asia

Europe

Latin
America

North
America

Oceania

CNRM

0.0000

0.0051

0.0000

0.0032

0.0150

0.0006

~ ECHAM

-0.0001

0.0122

0.0000

-0.0012

0.0328

0.0002

GFDL

0.0009

0.0226

0.0000

0.0063

0.0136

-0.0011

P Ml ROC

0.0002

0.0104

0.0000

0.0032

0.0311

-0.0009

-------
Climate Change Impact by 2100 on

All Extreme Events



CNRM

ECHAM

GFDL

Damages

($billion/yr)
(%GDP)

$47.0
(0.008%)

$85.6
(0.015%)

$102.5
(0.018%)

Deaths
(per year)

1750

-500

-2277

-------
Climate Change Damages by

Event

Q

CO
D

c
©

m

80
70
60
50
40
30
20
10
0



-lU

Cold

Drough
t

Flood

Heat

Local
Storms

Severe
Storms

Tropica
1

Cyclon
e

MCNMR

3

0

0

3

6

16

20

DECHAM

3

0

0

6

6

16

55

BGFDL

6

0

1

5

6

16

70

-------
Climate Change Extreme Event

Damages by Region

-------
Climate Change Extreme Event
Damages in % GDP

0.080

0.070

Q.

Q

0

c


a.

-------
Limitations

•	Non hurricane impacts uncertain

•	Country scale may be too coarse- need
finer scale

•	Need better data about damages and
extreme events

•	Ecosystem effects are not measured

•	Adaptation is not explicitly modeled

-------
Summary

•	Predicted climate change impacts from all
extreme events (including tropical cyclones)
range from $47 to $100 billion/yr by 2100

•	Equivalent to 0.008 to 0.018 percent of GWP by
2100

•	Climate change has mixed effect on fatalities
because tropical cyclone deaths may fall more
than other deaths increase

-------
Volume 1 • Number 1 • April 2010

3

g

Editor-in-Chief:

Robert Mendelsohn

Yale University, USA

-------
Incorporating Water Resources

Impacts
sessment Models

Kenneth Strzepek

Joint Program on the Science
and Policy	of G

-------
-------
Elements of the Water System

•	The Hydrologic System

-	Climate and Land Use

•	The Managed Water Supply System

•	Water Demand

-	Aquatic Ecosystem

-	Market Activities

-	Human Health

-	Non-Market Activities

•	Excess Water

•	Role in Economic Development

-------
km3

3 200

2 800
2 400
2 000
1 600
1 200
800
400
0

Agricultural











iJL



Forecast

Evolution of Global Water Use

Withdrawal and Consumption by Sector

Domestic











































Industrial

































. 11



Reservoirs



























1900 1925 1950 1975 2000 2025 1900 1925 1950 1975 2000 2025 1900 1925 1950 1975 2000 2025 1900 1925 1950 1975 2000 2025

Withdrawal

Consumption

Waste

Withdrawal

Consumption

Vteste

Withdrawal

Consumption

Waste

Evaporation

Note: Domestic water consumption in developed countries (500-800 litres per person per day)
is about six times greater than in developing countries (60-150 litres per person per day).

PMIUPPE REKACEWICZ
FEBF*MR <2W»

Source: Igor A Shiklomanov, State Hydrological Institute (SHI St Petersburg) and United Nations Educational Scientific and Cultural Organisation (UNESCO Paris). 1999

-------
United States
Water Use
2005

Public supply. 11 percent

Puhbc supply maiar intaka. Say County. Florida

Irrigation. 34 percent

Galad-pia*flood imqntion, Mofaorr Courtly Wyoming

Aquacutture. less than 1 percent

World's l*»gD8i lr>.r'win, Buhl Idaho

Mining, less than 1 percent

Spofanre p4.jinii» iTine. Kjoq! Moun-«in.NomGiri>»f«

Domestic, less than 1 percent

5

i

Domestic woi. Lady County, Ginrpu

livestock, less than 1 percent

livestock wKonng. Rio Atriba County, Maw Mmlcn

Industrial. 5 percent

H.)[)«< mil Sirrannati.

Thermoelectric power. 48 percent

Cocfing luwr's. Boko Cuunly. Geoifta

-------
Water Use in Europe

Socieol write* um pc*
*ot> regie* 4p*c#ot)

Imt» ffi WJr

-------
Modeling Water Resources Impacts in

IAMS

•	We know how to model key water related at the River Basin Level

Hydrology, Crops, Energy, M&l, ....

Combined	Use of Optimization	and Simula

Henry D. Jacoby & d. p. Loucks	water resourc

•	What is the appropriate Spatial and Temporal Scale to accurately model
climate change impacts for the questions being asked by lAMs or
sectoral level analysis at what scale.

•	lAMs

-	Spatial Scale: 10 to 20 regions: National lowest Scale

-	Temporal Scale: 1 to 5 year time steps

•	Global Crop and Hydrologic Modeling at 0.1 to 0.5 degree dail

•	There are over 10,000 level 4 River Basin "~20,000 km2"

•	Water Mgt Models : "River Basin Scale" and Monthly

-------
Spatial Scale Economic Components of Selected lAMs

MiniCAM:	14 Regions

-	the United States, US, Canada, W. Europe, Australia & New Zealand, Japan, Eastern
Europe, The Former Soviet Union, China, Mid-East, Africa, Latin America, Korea,
Southeast Asia, and India. In addition, three others are underdevelopment: Mexico,

Argentina, and Brazil.

MERGE: 9 Regions

-	Canada, Australia and New Zealand (CANZ); China; eastern Europe and the former
Soviet Union (EEFSU); India; Japan; Mexico, and OPEC (MOPEC); western Europe
(WEUR); the United States of America (USA); and the rest of the world (ROW).

IGSM/EPPA: 16 Regions

-	United States (USA) European Union (EUR) Eastern Europe (EET) Japan (JPN) Former
Soviet Union (FSU) Australia & New Zealand (ANZ) Canada (CAN China (CHN) India
(IND) Higher Income East Asiad(ASI) Middle East (MES) Indonesia (IDZ) Mexico (MEX)
Central & South America (LAM) Africa (AFR) Rest of World (ROW)

Fund

National Level

-------
Global Water and AG IMPACTS
Regional and National Scale

Hp wpectsrtcy ; Esp«sn» IUt»

71.0 - 74 0
070- 70S
040-00)

62.0 * 03 9
MO -01 S
SS.0 - SS J
150 54)

JtO - 44 5
2S S- J4S

Low	;C»ptr»re« Mti* i M)> mptrentt

I	I Ho OlU i Pst 04 donnoot f HO tvay o»to»

uuiiii	n (-Wt l*i	lih-f rrtn

-------
Europe Region and
18 FPU

-------
Southern SSA Region
and 21 FPU

-------
Irrigation Water Demand

Europe-15 PEF/PET for months when crops are grown in 2001











i	i2

001-FPU

20

01-36 Reg

-







—















r























-



--



—



—



—













—



—



—







—



—

























-





















































U H—^—i— —i— —i— —i— —i— —i— —i— —i— —i— —i— —i— —i— —i— —i— —i— —i——i—

^ ^ ^	<#" 

Southern SSA PEF/PET
for months when crops are growi i n 2001









n n

DRY

f f f f f f	¦?

r ,r r 
-------
Irrigation Water Supply for Agriculture

Irrigation Water Demand in Central Asia
(kmA3)

20
15
10
5
0

Central Asia FPU
Central Asia 36

T	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1-

<5^	<# <#	<#

Irrigation Water Demand In Southern SSA

(kmA3)

10
8
6
4
2
0

Southern SSA-FPU
Southern SSA-36

i	1	i—

—i	1	1	1	1	1	1	i—

—i	1	i—

^ ^ ^ ^ ^ ^ ~N> ^

20
15
10
5
0

Irrigation Water Supply for
Agriculture in Central Asia (kmA3)

\ A

k AA k.





V V

Y \Mf

-^-Central Asia-FPU

Central Asia-36

C? C?' C?5 C?3 'O' ^ ^ qV
cV cV cV cV cV cv cv cv cv cv

Irrigation Water Supply for Agriculture
in Southern SSA (kmA3)

Southern SSA-FPU
Southern SSA-36

$ qCS* ^	^ ^ ^	^

VV^'v'v'v'v'v'v'v'v'v'v

-------
WATER MANAGEMENT

•	Bring Water to Where it is needed when it
is needed

•	Great Environmental Costs

•	Market Benefit

•	Social Costs and Benefits

-------






mm







¦m*









wm







































Annual
mean

Yield

Storage Yield Curves

Storage capacity

-------
Gtptpa|l Wet|anc|l Ipry

Change in average annual precipitation, 2000 - 2050
CSIRO (DRY)	NCAR (WET)

A2 SCENARIO

Two extreme GCMs used to estimate range of costs

16

-------
800

q 600

C/3
£

" 400

8

a>

¦ i. 200

¦a

a

U

(a) CSIRO

; —



——S=—



— r







*=-
f\

o



" i

o

«r>

o



o



ri

ri

rn

m

3;



yr\

o

O

O

o

o

©

©

©



O



r i

o

o

©

1

s

CI

ci

ci

ci

CI



o

IT)

o



ci
o

m
o

r^.
O





ci

ci

ci

CI

ci

| 300

£>

Q

w 200

(a) CSIRO

(b) NCAR

-------
J

\

"What is the
of spatial scale and
management
basin modeling for global
food producti

Pathfinder Res

North.Xjte

SPIatte Irrig (1) KS M

Fort Peck

Garrison

garrison. Mun. (1)
3arrison.Irrig (1)

North.Platte River

3he

WY.Irrig (1)

Big Beri,

Alcova ResQ[enc|q' ]am NPIatte Irrig (1)J

3uqhisey. Res * FtRandaft

'Seminoe Res

„ Sutherland Re
Riversida.Res _

Empire Res
M ilton. Resi »^ackson, Lak

Prewitt Res

Antero Res/	South, Platte

Lake CSeeseman
Eleven. Mile Canyon.Res	CO Irrig (1)

i/e'AnMi"" Osage Rfes
K§ Thermal. (1) Osage River

-------
4 Spatial Scale Representations of

River

Lower.Mis Mun.(1)

Missouri. River

Garrison

Missouri

Ft Peck.Mun.f1i Ft Peck. Irrig (1)

I

garrison. Mun. (1)
Barrison. Irrig (1)

WY. Irrig (1)

Detailed

North.Platte River

Pathfinder. Res
rSeminoe Res

Big Bei.

Alcova Res Glendgbam NPIatte Irrig (1)

uermsey Res A Ft Randal

iki "•

/lin.Rel.Req_GavPt (1)

PpintA Mun1 <1>

Sutherland Re:

Riverside_Res
rTkJzmpire Res
Milton. Resi »slackson. Laki

Antero Res,

Lake CXeeseman
Eleven. Mile Canyon.Res

Lake M^lpney j P|atte Irrig (1)

PrewittR^.VJ,- o ^ Platte Mun.(jj;

North/Starling Res
South. Platte

SPIatte Irrig (1^KS	OsagWes

CO Irrig (1)	K&Vrfermal. (1) Osage River

Basin

Missouri.River

1 Basin

Min. Flow. Req (1)



uri. River

is Irrig (1)

3 Basins

North. Platte River North Platte Irrig (1)

in. Rel. Req_GavPt (1)

1 (1)

1)

all (1)

al2 (1

S Platte Irrig (1)

-------
Average Monthly Hydropower Generation in
each of the Missouri River Spatial Representations

Average Monthly Hydropower Generation

0)
c
o
O

%
o
a.
o

>
X

50000000 n
45000000
40000000
35000000
30000000
25000000
20000000
15000000
10000000
5000000
0

^ ^







s$K ^ #



^	VQ® vqO

„
-------
Spatial Resolution Impacts on Estimated Crop Water Stress

(Farmer et al, 2010)

Ten Years of Yield Factors for FPU 44

~r

-------
Spatial Unit of Crop Modeling (Farmer et al, 2010)

No Statistical Difference between 0.5 & 2 degree
A p-vaiue less than 0.05 will reject the null hypothesis.

•0.5 Degree 3000 cells
•2 degree 180 cells

-------
River Basin Spatial Scale for USA
30 to 991 basins

1

Source: Subbasin Assessment Regions 1978 Water Resources Council, 2nd National Assessment

i^>omisfKj^3rJK.auiy

|Pnc i ticjNoi" 11 nvesr

^Dssouii

LlyppepAIississlppi!

.em lessee]

Loweij >
Mississippi,

20 \

i S<:£Ttii Arl!iutic-(jult

-------
Mean of Differences in Number of Drought-Months
Relative to 20th Century Baseline for the 99 U.S. Subbasins

I I Regional Watershed Boundaries (2-digit HUCs)
Mean of differences in number of drought-months
(Out of a total of 360 months, or over 30 years)

H -67 - -62

-25 --19

I - I

1 -10



40-44

-61 - -54

-18--14

I	I

11 -17



45-53

¦¦ -53 - -46

¦¦ -13--11

I I

18-25



54-64

H -45 - -37

HI -10--6

¦¦

26-34

HI

65-75

-36 - -26

I I -5-0

I I

35-39



76-89

Drought Index: SPI-12
GCMs: All (Mean Values)

Time Period: Mid-21 st Century
SRES Scenario: A1B

-------
UNCERTAINTY DISTRIBUTIONS ACROSS GCMS FOR THE SPI-12 DROUGHT INDEX
DIFFERENCE IN NUMBER OF DROUGHT MONTHS FROM 20TH CENTURY BASELINE FOR THE 99
SUBBASINS UNDER THE A1B SRES SCENARIO, MID-21ST CENTURY

ll

« ai

2 >

Q 3

..

2	S

S o

£ ¦=
E S

3	CM

Z c

.E £

O £

£ 2

CJ LL

o

o
o

CO

o

LO
CM

o
o
CM

o

LO

o
o

o

LO

o
LO

o
o

o
LO

o
o

Subbasin

9997959391 8987 858381 79 77757371 6967 656361 5957555351 4947454341 3937353331 292725 2321 1917151311 9 7 5 3 1

HUC
Region

o o



<
CO

5

-------
(.Oi

Mean of Differences in Number of Drought-Months
Relative to 20th Century Baseline for the 99 U.S. Subbasins

j

J

I I Regional Watershed Boundaries (2-digit HUCs)
Mean of differences in number of drought-months
(Out of a total of 360 months, or over 30 years)

¦H

-8--4

6-7



13



21 - 23

¦¦

-3-0

r~\ 8



14



24-27

H

1 -2

I I 9

WM

15-16



28-32

~

3-4

H 10-11



17-18



33-41

wm

5

H 12



19-20



42-51

Drought Index: PDSI Severe
GCMs: All (Mean Values)

Time Period: Mid-21 st Century
SRES Scenario: A1B

-------
¦=	£

a)	ra

3	CQ

8	>

Q	5

2	S

5	o

¦2	¦=

E	S

3	CM

Z	O

.E	£

0)	£

£	2

QJ	LL

0)

o
^r

o
CM

o
o

o
oo

o
CD

o

o
CM

o
CM

o

o
CD

UNCERTAINTY DISTRIBUTIONS ACROSS GCMS FOR THE PDSI EXTREME DROUGHT INDEX
DIFFERENCE IN NUMBER OF DROUGHT MONTHS FROM 20TH CENTURY BASELINE FOR THE 99
SUBBASINS UNDER THE A1B SRES SCENARIO, MID-21 ST CENTURY

Subbasin

9997959391 8987858381 7977757371 6967656361 5957555351 4947 454341 3937353331 2927252321 1917151311 9 7 5 3 1

-------
Mean Changes in Drought Index Values from Baseline

B1, A1B, and A2 SRES Scenarios n Late 21st Century (Strzepek et al 2010)

PDSI
Extreme

SPI-12

INDUSTRIAL ECONOMICS, INC.

-------
Observation 4b: Emissions have a more pronounced effect on droughts
when both temperature and precipitation are considered.

C02 Concentration vs. Mean Change in Drought Months from
Baseline: PDSI-Extreme Drought Index

+ Dallas (67)

•	Central California (95)
Iowa (41)

¦ Atlanta (18)

X Los Angeles (98)

~	Boston (3)

X Seattle (91)

750

C02 Concentration

INDUSTRIAL ECONOMICS, INC.

-------
Boston Design Storm 2050

Current	Uncertainty

much greater	than

Changes

Boston Absolute Yearly Daily Precipitation of Models: Comparrison of Seasons at 2050 by LPTIII

	[	[	[	[	I	I	I	[	[	[	

Winter Spring Summer Autumn Yearly Winter Spring Summer Autumn Yearly

-------
CC Impacts on Roadway Bridges

USEPA by Stratus

Table 13. Number of currently deficient bridges per 2-digit HUC
vulnerable to climate change for tlie historical 100-year, 24-hour
storm. Tins includes currently deficient bridges with a projected
increase in modeled flow of more than 20% for three fiinire emissions
scenarios (A2, A1B. Bl) and two time periods (2055, 2090).

HUC

A2
2055

A IB
2055

Bl

2055

A2
2090

A1B

2090

Bl

2090

1

2,430

3,026

1,428

2,904

2,717

2,726

2

7,395

3,871

3,409

8,734

8,100

4,749

3

15,106

7 772

6.860

23,072

19,552

3.987

4

4,556

3,675

4.482

4,975

5,198

4.629

5

11.996

8,458

6.350

12,654

12,870

10,156

6

2,875

3,105

3,677

4,306

4,306

2,980

7

7,246

6,030

5,591

8,129

8,405

4,354

8

6.114

541

1.404

11,761

4,669

1,325

9

351

221

343

351

356

325

10

11,722

8,205

6,539

10,329

12,453

9,979

11

5,774

1,648

925

6,516

5,451

4.612

12

6,254

1,783

1.440

7,433

5,300

9.617

13

453

426

291

52S

515

538

14

502

274

250

400

405

486

15

773

285

123

752

493

301

16

545

360

463

515

545

534

17

4,665

2,675

1.666

5,545

3,688

3,011

IS

2,357

459

2,087

2,530

2,551

496

Total

91,114

52,814

47,328

111,434

97,604

64.805

-------
Impacts on Roadway Bridges

Stratus Consulting	(Final, 8*12 '20100

Projected Number and Percent d H»gfiway Bridges At Risk trom Increased
Peak Hows Due to Climate Change Cry Hydrologic Region, 2046-2065
Emission Scenario; A IB, lOO-year 24-hour Precipitation Event

Nomtw 3i Bridges Vulnerasie
to Climate Change*

0

1	so
St'500
jOI-ICOO
1.001 2,500

| 2.501 5,00*

>S 0W

!«*M indkaat*

lh* pacH Mt« 113*. omu

uur rji arc Mjnaratta
lo dlnoM chargfl

I 'laxtmiiix

	 ?-Oig)1 UBOS 4JC

eased on age or gcms cccma oocm 3.1, cnhm CM3 o, ana qhx 2 0

'inciuMe oonvWf deficient Dndpn « ciimwtlty *-r*t*ahi* hrktoM IacmmI on rwv«»r\rty Ml

with a projected «vrease in modeled flow ot Dealer than SO percent, plus currently acteptaDte bridges located on
sandy so* with a protected reread in modeled Kr« ol o>»awr mar 100 percent trom tine pared 1981 20M to 2046 2C65

Figure 10. 2046-2065,100-year. 24-hout storm. Scenario A1B.

-------
Threats to Existing Ag Water (Strzepek & Boehlert, 2010)

Legend

No Data

map_data

NoCC_M_l_E

I -100 000000--86 075412
-86 875411 - -46 355323
-46 355322 - -28.036716
-28 036715--9 255413
-9255412 - 0 000000

.ST"



INDUSTRIAL ECONOMICS, INC.



-------
Threats to Existing Ag Water

Foresight Region

World
Europe

European Union
Northwestern Europe
United Kingdom
Former Soviet Union

Africa

Sub-Saharan Africa
Nile River Basin

North America

Asia

China
India

Latin America and the Caribbean

Brazil



2000
Agricultural
Withdrawals
(billion m3)

2,946

263

95
16
0.6
186

246

50
146

255

2,060

558
866

182

21

No Climate Change

2050 M&I

7.3%

2.5%

0.7%
4.5%
0.0%
3.2%

9.8%

11.9%
9.1%

-0.1%

8.8%

2.7%
13.5%

3.8%

0.0%

EFRs

9.4%

7.7%

12.8%
11.7%
0.0%
10.0%

5.8%

7.2%
0.2%

15.2%

8.9%

7.3%
12.1%

12.3%

0.0%

501 0.2% | 14.3%

2050 M&I
and EFRs

17.7%

14.4%

18.7%
8.2%
0.0%
19.7%

15.8%

16.4%
9.2%

14.9%

18.6%

10.1%
27.7%

16.1%

0.0%

Kim

14.5% |

-------
Climate Change Threats to Ag Water



Historic

WET

DRY

World

17.7

16.5

16.9

Europe

14.4

12.9

20.4

Africa

15.8

16.9

17.1

N.America

14.9

13.6

12

Asia

18.6

16.7

16.8

Latin Amer

16.1

19.9

16.8

Oceania

14.5

14.5

14.5

INDUSTRIAL ECONOMICS, INC.

-------
Water For Environment versus AG - Happening

•	Australian farmers are furious about a government
concession to nature Australia's water war

•	AFTER a ten-year drought, farmers along the
Murrumbidgee River now face ruin from a devastating
flood. But it is the government that riles them as much
as any caprice of nature. Last month officials called for
a cut of nearly 40% in the volume of river water they
take for irrigation. At a rowdy meeting in Narrandera, a
river town, John Bonetti, a third-generation farmer,
drew cheers from about 900 farmers when he told
visiting bureaucrats and scientists, "If you think this is
the end of the fight, I can assure you it's only the
bloody start."

INDUSTRIAL ECONOMICS, INC.

-------
Summary

•	SCALE MATTERS

•	Cannot sum water impacts across sectors
for Impacts and especially Adaptation
must model Basin Scale Water Mgt
Systems (Smith, Hurd next talk)

•	FLOODING VERY IMPORTANT

- Need "additional" information from GCMs

•	CLIMATE CHANGE IN THE CONTEXT
OF GLOBAL CHANGE

-------
-------
-------
-------
CC Impacts on Roadway Bridges

^tTfltus Consulting	(FiuaL, S'12 '2010)

$2DC.:

,1 J15C.3

c
e>

J3
W

in $1CD.D

¦i—'

in
O
u

n3

: 5C. ~

SC.3

B1 A1E A2
2D10—2D55

bi aib a:

2055-2C3C

~	Dencient bridges repaired

~	Dencient bridges repaired as
adaptat on to clinia:e change

Figure 16. C
scenario.

osts for adapting deficient bridges to climate change by time period and

-------
IPCC AR4 Precipitation

(3

30
20
10
0
-10
-20

-30 u	

1950

30
20
10
0

0
Id

1	-10

Pj

ir

-20

-30 u	

1950

30
20
10
0

a>
c

8

sz

0

Id

1	-10

to
IE

-20

-30

1950

West Africa





1975

2000

2025
Year

East Africa

2050

2075



1975

2000

2025
Year

South Africa

2050

2075







a\

















1975

2000

2025
Year

2050

2075

-------
Boston Design Storm 2050

Current	Uncertainty

much greater	than

Changes

Boston Absolute Yearly Daily Precipitation of Models: Comparrison of Seasons at 2050 by LPTIII

	[	[	[	[	I	I	I	[	[	[	

Winter Spring Summer Autumn Yearly Winter Spring Summer Autumn Yearly

-------
Change in Precipitation Event Interval ( %/°C)

DJF

JJA

MAM

SON

-------
Brief Summary of Studies that Estimate the Economic Impact of
Changes in Climate and Water Availability

Brian H. Hurd and Mani Rouhi-Rad (New Mexico State University, bhurd@nmsu.edu)
Jan 24, 2011

1. Overview

As is well known, water management infrastructures and institutions have evolved to
help communities cope with a moderate range of water supply variability and
uncertainty. With few exceptions, U.S. communities, industries and water users
regardless of their location have evolved capacities to sustain successfully within the
context of their local water supply fluctuations and climate variability that is with a
variability that is within their 'norm'. Mild fluctuations, moderate variability, even
occasional extremes are typically within the normal realm of expectation. Resilience is a
characteristic of communities, industries, organizations, and residents that are
moderately to well prepared and capable of responding well to 'occasional' extremes.
(Figure 1 from the USGS shows water use patterns across the U.S.) However, as the
accumulation of science indicates the climate forcing of anthropogenic greenhouse gas
emissions is highly likely to contribute to climate uncertainty (e.g., Parry et al., IPCC
FAR, 2007 and others). And if this uncertainty can be expected to lead to more frequent,
persistent and intense deviations beyond the prevailing capacities of water users to
cope, then the risks and consequences facing water users within communities,
industries, and organizations become an economic concern.

Both human and natural systems are vulnerable to such changes, and to their conflation
with other significant stressors like population growth, loss of habitat and biodiversity,
resource scarcity etc. Water systems are particularly sensitive to climatic changes. Both
surface- and ground-water supplies can be affected by extreme or persistent changes in
the amount and timing of precipitation, temperature driven processes such as
evaporation and vegetative evapotranspiration, snowmelt, vegetation cover, and
streamflow patterns. Water users are also directly and indirectly affected by extreme and
or persistent changes in climate, for example, as rising temperatures increase
consumptive irrigation requirements for many crops including irrigated lawns and
gardens. And in municipal water systems an increase in seasonal temperatures would
likely be experienced as an increase in the effective length of 'summer' and its inherently
higher water demands as well as greater 'peak' demands and the associated strain on
water delivery capacity and infrastructure.

In this brief abstract, we survey the literature on economic impacts to water systems
and resources, with a focus on national and region-wide estimates and on the most

1

-------
recent studies where they have been conducted. There are far more studies linking
climate change and hydrology than those considering economic endpoints. In fact there
are surprisingly few studies that complete the linkages between climate change, water
and economic consequences. We have tried to access and include as many as we
could identify.

2. National Scale Estimates

Examining climate change impacts on water resources on a national scale is quite
daunting and there are few examples to draw upon. This is really not at all surprising.
There is tremendous variation in water resources and water systems across the U.S. Not
only variation across regions but tremendous complexity within regions, and within
particular watersheds. Such variation and complexity hinders the development of a
comprehensive and consistent assessment of economic impacts on a national basis.
Estimation approaches such as large-scale statistical studies that have been used in
other sectors such as agriculture (e.g., Mendelsohn et al., 1994) are not well suited with
so much uncontrolled variation. Enumerative or aggregation approaches to measuring
economic impacts that build a national level estimate by aggregating impacts from each
of the nation's watersheds is conceivable in concept but very difficult and costly to
execute. Perhaps the closest example of this approach is Hurd et al. (1999a, 2004)
where national-level estimates were derived on the basis of only a few large-scale
regional estimates and some rather heroic assumptions about the comparability and
conformability of some very different regions. Finally, a third way has recently been
used to take aim at this difficult problem. Researchers at Sandia National Laboratories
have used REMI (Regional Economic Impact, Inc.) in conjunction with a system-
dynamics hydrology model and estimated precipitation changes based on the SRES
A1B scenario to estimate state-level economic impacts from reduced precipitation.

Research on climate change and its potential economic impacts has steadily evolved
from static models based on fixed marginal values to models that capture market
dynamics. Early studies by Cline (1992), Fankhauser (1995), and Titus (1992)
associated fixed economic values with projections of physical changes (e.g., runoff), with
no attempt to account for changes in the marginal value of water or the response of
water use to changes in marginal value. Both Cline's (1992) estimated cost of $7 billion
and Fankhauser's (1995) estimated cost of $13.7 billion to consumptive water users in
the United States are driven by an assumed 10% decrease in water availability. Titus
(1992) estimated costs ranging from $21 to $60 billion, including impacts to
nonconsumptive users (primarily hydropower and water quality losses), which he
observed would most likely exceed the magnitude of impacts to consumptive users.

Hurd, Callaway, Smith and Kirshen (1999a, 2004) approached the problem from a
region-specific perspective. Using models of the hydro-economy for four major water
resource regions (shown in Figure 2a) and driven by simulated streamflow changes for a
set of 15 incremental climate change scenarios, and an extrapolation model based on
comparable regions they developed national level estimates of economic damages

2

-------
related to water resources and climate change. They estimated total annual damages to
consumptive and non-consumptive water users by as much as $43.1 billion (1994$)
under an incremental level of climate change where temperatures rose by 5 deg C and 0
change in precipitation (estimates shown in Figure 2b).

In assessing the potential for climate change to affect water availability on a national
scale - specifically the impacts of reduced precipitation, Sandia National Laboratory
(Backus et al., 2010) estimates there is a 50-50 chance that cumulative direct and
indirect macro-economic losses in GDP through 2050 will exceed nearly $ 1.1 trillion
(2008$), not including flood risks. That is approximately 0.2% of the cumulative GDP
projected between 2010 and 2050. They estimate a 50-50 chance of non-discounted
annual losses of $60 billion (2008$) by 2050. Their estimation process uses the
MIROC3.2 (medium resolution) and the A1B emissions scenario as a motif to guide the
assignment of state-level precipitation changes and then uses results from the remaining
GCM runs to characterize and assess uncertainty. Water availability changes are
assessed at the county-level using Sandia Water Hydrology model. State-level impacts
on economic activity changes are analyzed using REM I. [REM I and other input-output
type model the changes in economic flows into and out from a region. They do not
measure or estimate economic costs and benefits in a theoretically consistent manner.
For example, these models do not estimate changes in willingness-to-pay associated
with changes in water availability but rather they simulate the economic consequences
that are entailed by such changes. For example, in the same way that a disaster can
stimulate regional economies as recovery and rebuilding efforts create jobs and raise
incomes. In a similar fashion, persistent and severe water shortages can lead to
adaptive responses, like building dams and power plants to replace storage and
hydropower generation, thus stimulating employment and incomes.

Figure 3 shows the estimated cumulative state-level economic impacts from 2010
through 2050 (green areas show net GDP increases - particularly in California and
PNW).

Although there are considerable differences in the above approaches used to estimate
national-level annual economic impacts of climate change on water resources, there is a
remarkable consistency in the order of magnitude and share of GDP as shown here:

Cline (1992)	$7 billion (~ 0.1% of 1992 US-GDP $6.3 trillion)

Titus (1992)	$21 - 60 billion (~ 0.3 - 0.9% of 1992 US-GDP $6.3 trillion)

Fankhauser (1995)	$13.7 billion (~ 0.2% of 1995 US-GDP $7.4 trillion)

Hurd et al. (1999a, 2004)	$9.4 - 43.1 billion (~ 0.13 - 0.58% of 1995 US-GDP $7.4 trillion)

Backus et al. (SANDIA, 2010)	$ 60 billion (~ 0.4% of 2009 US-GDP $14.1 trillion)

3

-------
3. Regional Estimates

There are several region-scale estimates. These include the regions underlying the national-
estimates of Hurd et al., namely the Colorado River, Missouri Basin, Delaware basin, and
Appalachicola-Flint-Chattahoochie in the Southeast, and the state-level assessments
provided in the Sandia report (Backus, 2010), as shown in Figure 3. Additional economic
studies include California (Lund etal, 2003; Medellin etal., 2006), the Pacific Northwest
(Climate Impacts Group, 2009), and the Upper Rio Grande (Hurd and Coonrod, 2007).

California.

Medellin et al. (2006) perform a comprehensive assessment of climate change impacts on
California water users. An example of their findings uses the relatively dry GFDL-A2 to
estimate a 27% decrease in water availability and with modeled adaptive responses they
find "an average annual scarcity of 17%". Water deliveries to agriculture fall by 24% and
urban deliveries fall by 1%. They break down the impacts across three categories: scarcity
costs, operating costs, and additional policy costs if interregional water transfers are
limited. "Of the $360 million/year in average water scarcity costs for 2050 with dry climate
warming, $302 million/year results from lost agricultural production and $59 million/year
is from urban water shortages. ... Dry climate warming imposes an additional increase of
$384 million/year in system operating costs. ... With the climate warming, the costs of
policies limiting interregional water transfers increases to $250 million/year." All together,
these costs amount to $994 million per year, or less than 0.1% of California's $1.5 trillion/yr
economy.

Columbia River & Pacific Northwest.

The Climate Impacts Group at University of Washington assessed the impacts of climate
change on the Pacific Northwest and the state of Washington, averaging across 20 GCMs
under both SRES B1 and A1B (Climate Impacts Group, 2009). Snowpack reductions were
significant, with snow water equivalent falling by as much as 65%. Although annual runoff
shows an increase of 6% there is a reduction of 43% in runoff during the summer irrigation
season by the 2080s. Without adaptation water delivery shortages to agriculture in the
Yakima River basin, for example, could be significant. Estimated deliveries fall by as much
as 77% by the 2080s. In the 2020s, regional hydropower production increases by 0.5-4% in
winter, decreases by 9-11% in summer, with annual reductions of 1-4%.Economic losses of
between $23 million and $70 million are estimated, with significantly greater probabilities
of annual net operating losses for junior water rights holders.

Rio Grande.

Hurd and Coonrod (2007) estimate economic impacts of climate change on water resources
in the Upper Rio Grande (primarily New Mexico, El Paso, Tx, and the San Luis Valley of
Southern Colorado). Under the relatively dry scenario (GFDL), runoff change was estimated

4

-------
to fall by 28% (using WATBAL) and annual direct economic damages in 2080 were
estimated at $100 million using a hydro-economic model of the watershed. This loss is
approximately 0.2% of the estimated GSP of $60 billion.

Colorado River.

Christensen and Lettenmaier (2007), did a similar research on the Colorado River
hydrology with the average of 11 GCM ensembles and two SRES emission scenarios: A2 and
B1 (reference). Annual runoff reduction was between 0.0 (2020 Bl) and 11.0 (2080 A2)
percent Average annual delivery shortage was estimated to be between 0.22 BCM/Yr
(115.8%) and 1.2 BCM/Yr (631.5). Energy Production is estimated to increase during 2020s
by the maximum of 120.5 GWh/Yr (1.4%) and experience a reduction during the rest of the
century which will result in a maximum of 1573.6 GWh/Yr (18.5%) of negative production
during 2080s.

Hurd et al. (1999a), following the work of Booker and Young, modeled the hydro-economy
of the Colorado River basin and the impacts of climate change using incremental climate
change scenarios and the VIC hydrology model. From an annual baseline of $7.7 billion
(1994$) they estimated economic losses for a 5 deg C rise with no change in precipitation of
nearly $1.2 billion when runoff was estimated to fall by 35%. Under a 2.5 deg C rise and a
10% reduction in precipitation the losses approached nearly $1.4 billion (1994$).

Other Regions.

A few other regional studies of economic climate change impacts have been documented.
Our survey is neither exhaustive nor comprehensive, though literature searches do not find
many that are geographically broad. This does not indicate that there are not likely to be
significant impacts in other regions. Exhibit 4 shows some of the other basins modeled by
Hurd et al. (1999a) and the estimated changes in runoff and economic impact

5

-------
Exhibits

Exhibit 1. Estimated Use of Water in the United States in 2000

Including Surface Water and Groundwater Withdrawals
source: USGS (2000)

si

C/) (£

5 11

M

b <
2

60,000

WEST

EAST

40,000 -

20,000

Surface-water withdrawals

Groundwater withdrawals

EXPLANATION

Water withdrawals, in million gallons per day

~ 0 to 2,000
I I 2,000 to 5,000
I I 5,000 to 10,000

10,000 to 20,000
20,000 to 52,000

Source: USGS (2000).

6

-------
Exhibit 2. a. U.S. Water Resource Regions and the Regional

Associations and b. Estimated National Level Impacts of
Climate Change on Water Resources from Hurd et al. (1999a,
2004)

a.

Maine

b.

Estimated Total Economic Welfare Impacts on U.S. Water Resource Users

(billions of 1994$)

Climate
Scenario

Consumptive
Use

Nonconsumptive Use

Total

Hydropower

Other
Nonconsumptive
Sectors*

Baseline

88.5

14.7

28.7

132.00

+1.5=C +15%P

0.085

0.69

8.98

9.76

+2.5^C +7%P

-0.98

-2.75

-5.68

-9.41

+5.0^€

-4.29

-7.42

-31.4

-43.11

* Not including damages from thermal heat pollution.

7

-------
Exhibit 3. Excerpted from Backus et al. (2010), Cumulative GDP
climate-change risk (40 years from 2010-2050) from
reduced precipitation for the ensemble of A1B climate
scenarios (in billions of dollars at a 0% discount rate).

K „

J26.6B

OR

S19.4B	ID

S4.0B

NV
-S38.7B

CA
$25.1 B

UT

-S10.5B

MT

S0.9B

vw

-$3.0B

CO
$1.2B

-$69.0B	NM

-$26.1B

NE
-$1.4B

KS
-S6.3B

MM
-$8.3B

OK
-$38.0B

IA

-$2.8B

IL

-$10.1B

MO

-$3.88

ME

-$0.3B

VT
-$0.7B i

NY
-$122.9B

-$1.8B
MA -S9.0B
CT Rl
-$9.5B-$0.7B

PA
-$64.6B

OH

-$26.7B

WV
-$45.9B

KY
-S40.6B

MS
-$7.3B

TN

•$58. SB

AL

-$29.2B

VA
-S45.4B

NC
463.4B

GA
-$10Z9B

TX
-$137.8B

LA

-$14.3B

FL
-S146.3B

Percent Change





-1.11--0.20





-0.20--0.10





-0.10- -0.03





-0.03 - 0.00





0.00-0.14

8

-------
Exhibit 4. Estimated Regional Changes in Runoff and Economic Welfare
under Selected Incremental Climate Changes



Watershed



Colorado

Missouri

Appalachicola-

Flint-
Chattahoochie

Delaware

Baseline

Runoff (kaf/yr)
Welfare (million 1994$)

17,058
$7,744

56,651
$10,804

24,363
$2,225

13,660
$6,565

Climate Change Scenario and Changes from Baseline

+2.5 deg C, +7% P

% Runoff chg (kaf/yr)
Welfare chg (M1994$)

-	4.2%

-	$102

-	9.1%

-	$519

-	0.3%

-	$15 (1)

-4.1%
- $22

+2.5 deg C, -10% P

% Runoff Chg (kaf/yr)
Welfare chg (M1994$)

-	37.9%

-	$1,372

-	42.5%

-	$2,041

-	27.5%

-	$12 (1)

-	33.2%

-	$187

+5 deg C, 0% P

% Runoff Chg (kaf/yr)
Welfare chg (M1994$)

-	34.7%

-	$1,193

-	42.4%

-	$2,239

-	23.5%

-	$31 (1)

-	33.9%

-	$207











(1)

The estimated changes in welfare for the AFC basin show a mixture of effects including changes in flooding and water quality
which confound simple comparison across scenarios. For example, a possible consequence of warmer and drier mean climate
might be an expected reduction in average annual flood damages as represented in the above results. However, this analysis
does not take into account possible changes in climate variability i.e., greater frequency and intensity of extreme events.

Source: Adapted from: Hurd, B. H., J. M. Callaway, J. B. Smith, and P. Kirshen. 1999a. "Economic Effects of Climate Change
on U.S. Water Resources." In The Impact of Climate Change on the United States Economy, ed. Robert
Mendelsohn and James NeumannCambride, UK: Cambridge University Press, 133-177.

9

-------
Cited References and Additional Relevant Literature

I.	Backus, George, Thomas Lowry, Drake Warren, Mark Ehlen, Geoffrey Klise, Verne

Loose, Len Malczynski, Rhonda Reinert, Kevin Stamber, Vince Tidwell etal.

Assessing the Near-Term Risk of Climate Uncertainty: Interdependencies Among the U.S.
States. SAND2010-2052, 1-259. 2010. Albuquerque, New Mexico, Sandia National
Laboratories.

2.	Barnett, T., R. Malone, W. Pennell, D. Stammer, B. Semtner, and A.

Washington.2004. "The effects of climate change on water resources in the west:
Introduction and overview." Climatic Change, 62(1-3): 1-11.

3.	Barnett, Tim, R. Malone, W. Pennell, D. Stammer, B. Semtner, and W. Washington.

THE EFFECTS OF CLIMATE CHANGE ONWATER RESOURCES IN THEWEST:
INTRODUCTION AND OVERVIEW. Climatic Change 62, 1-11. 2004.

4.	Christensen, N. S. and D. P. Lettenmaier. A multimodel ensemble approach to
assessment of climate change impacts on the hydrology and water resources of the
Colorado River Basin. Hydrol.Earth Syst.Sci. 11(4), 1417-1434. 9-7-2007.

5.	Climate Impacts Group. The Washington Climate Change Impacts Assessment. McGuire
Eisner, M., Littell, J., and Whitely Binder, L. 2009. Center for Science in the Earth System,
Joint Institute for the Study of the Atmosphere and Oceans, University of Washington, Seattle,
Washington.

6.	Cline, William R. 1992. The Economics of Global Warming. Washington D.C.: Institute for
International Economics

7.	Fankhauser, S. 1995. Valuing Climate Change: The Economics of the Greenhouse.
London: Earthscan Publications Ltd.

8.	Hamlet, A. F. Assessing water resources adaptive capacity to climate change impacts in
the Pacific Northwest Region of North America. Hydro 7(4), 4437-4471. 7-8-2010.

9.	Harou, J. J., J. Medell+jn-Azuara, T. J. Zhu, S. K. Tanaka, J. R. Lund, S. Stine, M. A.
Olivares, and M. W. Jenkins.2010. "Economic consequences of optimized water
management for a prolonged, severe drought in California." Water Resources Research,
46(5): W05522.

10. Hurd, B. H., J. M. Callaway, J. B. Smith, and P. Kirshen. 1999a. "Economic Effects of
Climate Change on U.S. Water Resources." In The Impact of Climate Change on the
United States Economy, ed. Robert Mendelsohn and James NeumannCambride, UK:
Cambridge University Press, 133-177.

II.	Hurd, B., N. Leary, R. Jones, and J. Smith. 1999b. "Relative regional vulnerability of
water resources to climate change." Journal of the American Water Resources
Association, 35(6): 1399-1409.

12. Hurd, B. H., M. Callaway, J. Smith, and P. Kirshen.2004. "Climatic change and US
water resources: From modeled watershed impacts to national estimates." Journal of the
American Water Resources Association, 40(1): 129-148.

10

-------
13.	Hurd, B. H. and J. Coonrod. Climate Change and Its Implications for New Mexico's
Water Resources and Economic Opportunities. 1-28. 2007. Las Cruces, NM, Agricultural
Experiment Station, New Mexico State University. Technical Report #45.

14.	Hurd, B. H. and M. Harrod. 2001. "Water Resources: Economic Analysis." In Global
Warming and the American Economy: A Regional Assessment of Climate Change
Impacts, ed. Robert MendelsohnNorthhampton, MA: Edward Elgar Publishing, 106-131.

15.	Lund, J. R., R. E. Howitt, M. W. Jenkins, T. Zhu, S. K. Tanaka, M. Pulido, M. Tauber,
R. Ritzema, and I. Ferriera. Climate Warming and California's Water Future. 03-1. 2003.
Center for Environmental and Water Resources Engineering, University of California, Davis,
California.

16.	Medellin, J, J. J. Harou, M. A. Olivares, J. R. Lund, R. E. Howitt, S. K. Tanaka, M. W.
Jenkins, Kaveh Madani, and T. Zhu. Climate warming and water supply management in
California. California Climate Change Center. 2006. San Diego, California, Scripps
Institution of Oceanography.

17.	Medellin, J, Julien J. Harou, Marcelo A. Olivares, Kaveh Madani, Jay R. Lund,
Richard E. Howitt, Stacy K. Tanaka, Marion W. Jenkins, and Zhu Tingju.2008.

"Adaptability and adaptations of California's water supply system to dry climate warming."
Climatic Change, 87(s1): S75-S90.

18.	Melillo, Jerry M. Climate change impacts on the United States; the potential
consequences of climate variability and change. Janetos, Anthony C., Karl, Thomas R.,
Corell, Robert, Barron, Eric J., Burkett, Virginia R., Cecich, Thomas F., Jacobs, Katherine,
Miller, Barbara, Morgan, M. Granger, Parson, Edward A., Richels, Richard G., Schimel,
David S., Carter, Lynne, Easterling, David, Felzer, Benjamin, Field, John, Grabhorn, Paul,
Hassol, Susan Joy, MacCracken, Michael, Smith, Joel, Taylor, Melissa, and Wlbanks,
Thomas. 1-1-2000. United States.

19.	Mendelsohn, R. 2001. Global Warming and the American Economy. Northhampton, MA:
Edward Elgar Publishing, Inc.

20.	Mendelsohn, R. and J. E. Neumann. 1999. The Impact of Climate Change on the United
States Economy. Cambridge, UK: Cambridge University Press

21.	Mendelsohn, R., W. D. NORDHAUS, and D. Shaw.1994. "THE IMPACT OF GLOBAL
WARMING ON AGRICULTURE - A RICARDIAN ANALYSIS." American Economic
Review, 84(4): 753-771.

22.	Miles, Edward, Marketa Eisner, Jeremy Littell, Lara Binder, and Dennis
Lettenmaier.2010. "Assessing regional impacts and adaptation strategies for climate
change: the Washington Climate Change Impacts Assessment." Climatic Change, 702(1):
9-27.

23.	Milly, P. C. D., K. A. Dunne, and A. V. Vecchia.2005. "Global pattern of trends in
streamflow and water availability in a changing climate." Nature, 438(7066): 347-350.

11

-------
24.	Neff, R., H. J. Chang, C. G. Knight, R. G. Najjar, B. Yarnal, and H. A. Walker.2000.
"Impact of climate variation and change on Mid-Atlantic Region hydrology and water
resources." Climate Research, 74(3): 207-218.

25.	Parry, M. L, O. F. Canziani, J. P. Palutikof, P. J. Linden, and C. E. Hanson. Climate
change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to
the Fourth Assessment Report of the Intergovernmental Panel on Climate Change.
Summary for Policymakers. Parry, M. L., Canziani, O. F., Palutikof, J. P., Linden, P. J.,
and Hanson, C. E. Climate change 2007: Impacts, Adaptation and Vulnerability.
Contribution of Working Group II to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change. Summary for Policymakers. 2007. Geneva;
Switzerland, Intergovernmental Panel on Climate Change (IPCC).

26.	Tanaka, Stacy K., Zhu Tingju, Jay R. Lund, Richard E. Howitt, Marion W. Jenkins,
Manuel A. Pulido, M. + Tauber, Randall S. Ritzema, and ln+ s C. Ferreira.2006.

"CLIMATE WARMING AND WATER MANAGEMENT ADAPTATION FOR CALIFORNIA."
Climatic Change, 76(3/4): 361-387.

27.	Titus, J. G. 1992. "The Costs of Climate Change to the United States." In Global Climate
Change: Implications, Challenges and Mitigation Measures, ed. S. K. Majumdar, and B.
Kalkstein, and L. S. Yarnal, and E. W. Miller, L. M. RosenfeldPhiladelphia, PA:
Pennsylvania Academy of Science.

28.	Xie, H., J. W. Eheart, and H. An.2008. "Hydrologic and economic implications of climate
change for typical River Basins of the agricultural Midwestern United States." Journal of
Water Resources Planning and Management-Asce, 134(3): 205-213.

12

-------
January 27-28, 2011

Research on Climate Change Impacts and
Associated Economic Damages:

Estimates of the Economic Impact of Changes

in Climate and Water Availability

Brian H. Hurd

Assoc. Prof of Agricultural Economics &

Agricultural Business

New Mexico State University

-------
Overview

Less Snowpack

Concepts and Complexities

National Estimates

Regional Estimates

Issues, Gaps, and Next
Steps

Reduced Streamflows

Earlier Snowmelt

4S3

Floods

Droughts

-------
Water, Climate & Communities
Form Complex Systems

•	Estimating water resource impacts is tough

-	Lots of variability: spatial, temporal, uses, infrastructure, vulnerability

•	What to measure?

-	Economic damages/benefits?

-	Changes in jobs, income & production?

•	How to measure?

-	Statistical models?

-	Simulation models?

-	Optimization models?

•	Adaptation & behavior

-------
Climate and Rivers

Rio Grande at Del Norte - Climate Change Simulation

Model assumptions
temperature \ 4°C
Precipitation 10%

source: Al Rango (usda/ars)
using Snow melt Runoff Model (SRM)

What does it mean for?

•	Water storage and
distribution systems?

•	Urban and rural water
users?

•	Water quality?

•	Hydropower?

•	Recreational and cultural
functions?

•	Riparian ecosystems and
migratory patterns?

-------
Spatial Heterogeneity:

Climate, Vegetation, Environment

Upper Rio Grande

RIO Spatial Variability

Upper Rio
Grande LTHO

Del Morte

Ft.

,/>\Raifi Gauuos
+ Radar Sites
[T"|Snow Course

Jemez

Source: Enrigue Vivoni, AZ State Univ.

-------
Water Use Patterns

60,000

WEST

EAST

I

ec

UJ

CL

Z

o

40,000 -

20,000 -

Ql=.

nnfln

D

l~lnnnn 1II In

in

a

On

nnl innn

.nn„n.









W:







5>T



V8

/95c

Surface-water withdrawals

Groundwater withdrawals



EXPLANATION

Water withdrawals, in million gallons per day
~ 0 to 2,000	CZI 10,000 to 20,000

r I 2,000 to 5,000 H 20,000 to 52,000
I I 5,000 to 10,000

Source: USGS (2000).

-------
Relative Regional Vulnerability of Water Resources

) 100 200 300400 500
Miles

o 100 200 300 400 500 600 700	.^stratus Consulting Inc.

Environment aid Energy Research^

Kilometers

Water Resource lnde>
d:/ginwater/gis/projects/finalindxupmap.aml April 28,1999

Hydrologic Unit Boundary
State Boundary

Average vulnerability <1.7

Average vulnerability >=1.7 and <=2

Average vulnerability >2

Note: Vulnerability values calculated as the average of the
vulnerability classification of the following subindicators:

(1)Water	Supply Distribution and Consumptive Use; and

(2)lnstream	Use, Water Quality, and Ecosystem Support.

Overall index

Source: Hurd, B.H., M. Leary, R. Jones, and J.B. Smith. 1999. "Relative Regional Vulnerability of Water Resources to Climate Change."

journal of the American Water Resources Association, December, 35(6): 1399-1410.

-------
National Estimates: Summary

Cline (1992)	$7 billion (~ 0.1% of 1992 US-GDP $6.3 trillion)

Titus (1992)	$21 - 60 billion (~ 0.3 - 0.9% of 1992 US-GDP $6.3 trillion)

Fankhauser (1995)	$13.7 billion (~ 0.2% of 1995 US-GDP $7.4 trillion)

Hurd et al. (1999a, 2004)	$9.4 - 43.1 billion (~ 0.13 - 0.58% of 1995 US-GDP $7.4 trillion)

Backus et al. (SANDIA, 2010)	$ 60 billion (~ 0.4% of 2009 US-GDP $14.1 trillion)

-------
National Estimates: Aggregating Benefits and Costs
Hydro-economic Model Approach

Estimated Total Economic Welfare Impacts on U.S. Water Resource Users

(billions of 1994S)

Climate
Scenario

Consumptive

Use

Nonconsumptive Use

Total

Hydropower

Other
Nonconsumptive
Sectors*

Baseline

88.5

14.7

28.7

132.00

+1.5°C +15%P

0.085

0.69

8.98

9.76

+2.5^€ +7%P

-0.98

-2.75

-5.68

-9.41

+5.0^€

-4.29

-7.42

-31.4

-43.11

* Not including damages from thermal heat pollution.

Source: Hurd, B. H., J. M. Callaway, J. B. Smith, and P. Kirshen. 1999.

"Economic Effects of Climate Change on U.S. Water Resources."

In The Impact of Climate Change on the United States Economy, ed. Robert Mendelsohn and James Neumann
Cambride, UK: Cambridge University Press, 133-177. .

-------
National Estimates: Jobs, Income & GDP Approach

Assemble climate

futures for
precipitation and

temperature
conditions with
uncertainty
2010 to 2050

(Via IPCC Ensembles
from Climate Models)

i

~

Map temperature
and precipitation
data to determine
water availability
and agricultural
production

(Via Sandia
Hydrological Model)

Calculate state and
national economic
consequences
2010 to 2050

(Via REMI
Macroeconomic
Model)

Sandia National Laboratory (Backus et al., 2010) estimates there is a 50-50 chance
that cumulative direct and indirect macro-economic losses in GDP through 2050 will
exceed nearly $1.1 trillion (2008$), not including flood risks. That is approximately
0.2% of the cumulative GDP projected between 2010 and 2050.

On an annual basis: a 50-50 chance of non-discounted losses of $60 billion (2008$) by
2050.

Source: Backus, G. et al. Assessing the Near-Term Risk of Climate Uncertainty: Interdependencies Among the U.S. States
SAND2010-2052, 1-259. 2010. Albuquerque, New Mexico, Sandia National Laboratories.

-------
Regional Estimates: Hydro-economic Model Approach

Estimated Regional Changes in Runoff and Economic Welfare under
Selected Incremental Climate Changes

Watershed



Colorado

Missouri

Appalachicola-

Delaware







Flint-









Chattahoochie



Baseline









Runoff (kaf/yr)

17,058

56,651

24,363

13,660

Welfare (million 1994$)

$7,744

$10,804

$2,225

$6,565

Climate Change Scenario and Changes from Baseline

+2.5 deg C, +7% P









% Runoff Chg (kaf/yr)

- 4.2%

-9.1%

- 0.3%

-4.1%

Welfare chg (M1994$)

-$102

-$519

- $15 (1)

-$22

+2.5 deg C, -10% P









% Runoff Chg (kaf/yr)

- 37.9%

- 42.5%

- 27.5%

- 33.2%

Welfare chg (M1994$)

-$1,372

- $2,041

- $12 (1)

-$187

+5 deg C, 0% P









% Runoff Chg (kaf/yr)

- 34.7%

- 42.4%

- 23.5%

- 33.9%

Welfare chg (M1994$)

-$1,193

- $2,239

- $31 (1)

-$207

Source: Hurd, B. H., J. M. Callaway, J. B. Smith, and P. Kirshen. 1999.

-------
Other Regional Estimates

Region

Study

Economic Impacts

California

Medellin et al. (2006)

Pacific Northwest Climate Impacts Group (2009)

Rio Grande	Hurd and Coonrod (2007)

Colorado River	Christensen and Lettenmaier(2007)

$302 M/yr agricultural scarcity cost, $59 M/yr urban scarcity cost, $384 M/yr
operating cost, $250 M/yr the costs of policies limiting interregional water transfers,
which is $994 M/yr totally (less than 0.1% California's economy)

Economic losses of between $23 million and $70 million are estimated, with
significantly greater probabilities of annual net operating losses for junior water
rights holders.

direct economic damages in 2080 were estimated to be $100 million/year

Energy Production is estimated to increase during 2020s by the maximum of 120.5
GWh/Yr (1.4%) and experience a reduction during the rest of the century which will
result in a maximum of 1573.6 GWh/Yr (18.5%) of negative production during
2080s.

-------
State-Level Estimates: SANDIA/REMI Approach

i

B| WA
$26.6B

MT

S0.9B

OR
S19.4B

ID
S4.0B

CA
$25.1 B

WY
-$3.0B

NV

-$38.7B	yy

-$10.5B

CO
$1.2B

NM

-$26.1B

ND
-$0.9B

NE
-$1.4B

KS
-S6.3B

MN

-$8.3B

OK
-$38. OB

Wl
-$6.2B

IL

-$10.1B

MO

•$3.8B

AR
-$11.9B

ME

-$0.3B

VT

-$0.7B ,.jh
-$1.8B

NY
-$122.9B

CT Rl

-$9.5B-$0.7B

MA -$9.0B

PA

-$64.6B

NJ
-$38.9B

OH
-$26.7B

WV
-S45.9B

KY
-$40.6B

NC
-$63.4B

MS

-$7.3B

AL

-S29.2B

GA
-$102-9B

TX
-$137.8B

LA
-$14.3B

FL
-$146.3B

Percent Change



¦

-1.11--0.20





-0.20- -0.10





-0.10--0.03





-0.03 - 0.00



y

0.00-0.14

Source: Backus, G. et al. Assessing the Near-Term Risk of Climate Uncertainty: Interdependencies Among the U.S. States
SAND2010-2052, 1-259. 2010. Albuquerque, New Mexico, Sandia National Laboratories.

-------
Issues, Gaps, and Next Steps

Understanding changes in extreme events

-	Severe, sustained drought risk

-	Flood risk changes are not well understood and are often locally sensitive

Water rights, federal & state regulation, and administration constraints
confound assessment of impacts and adaptation

Projecting market prices and trade flows of agricultural and other water-
intensive products is difficult

Groundwater. Measuring, monitoring, modeling.

Water security and food security are conflated and stir deep emotions

Water quality and environmental quality hard to assess and measure
economic outcomes

Coupling of hydro-economic and dynamic system simulation approaches
could bridge some gaps

-------
More nformation can be found at:
http://agecon.nmsu.edu/bhurd

-------
Biophysical Climate Change

Effects on
Ag ro-ecosystems

U.S. EPA/DOE Workshop

Research on Climate Change Impacts and
Associated Economic Damages

January 27-28, 2011

Cynthia Rosenzweig
NASA/Goddard Institute for Space Studies

1

-------
Outline

•	Estimates of current and likely impact of
climate change on biophysical response of
agricultural crops

•	Data and models used to make projections

•	Modulation of biophysical impacts via
adaptation

•	Gaps and uncertainties

2

-------
Current and Future
Impacts

• Estimates of the current and likely future
impact of climate change on biophysical
response of agricultural crops.

-What crops, (livestock), soil, and pests will be most
affected?

- Describe the best central estimates, the wider range
of possible outcomes, and the relative likelihood of
those outcomes.

3

-------
Observed Impacts on Agriculture

Yields

Phenology



Management
practices,
forest fires,
earlier pests
and
diseases

1973-2002 Annual temperature trends
<-1.2C to >1.2C

/

Temperature change °C
1970-2004

Livestock

Over the last 50 years:

• Very likely

less frequent cold days, cold
nights, and frosts
more frequent hot days and hot
nights

-	more frequent heat waves

-	more frequent heavy
precipitation events

-	increased incidence of extreme
high sea level

-	increased drought in some
regions

High temperature effect on rice yield; Earlier planting of
spring crops; Increased forest fires, pests in N America
and Mediterranean; Decline in livestock productivity

IPCC V\tall AR4

-------
Earlier Emergence of Insects

-1			1 ' I	'	1	'	!			f-

a Aphis mellifera

b Pieris rapae

* ;~



» * • •: * ;	I

* 4 i * iH

i ' i —> ¦ >

c Leptinotarsa decemlineata

d Bactrcpera oleae

In a six-decade long
study at a biological
research station in
Spain, increasing
earlier time of first
appearance for the
honey bee, cabbage
white butterfly, potato
beetle and olive fly were
found.

1940 1950 1960 1970 1980 1990 2000

1940 1950 1960 1970 1980 1990 2000

year

5

Gordon and Sanz, 2005; Gutierrez et al., 2010

-------
Photosynthesis Response to C02

60

50

40

30

20

10

0

C4 plant

—







C3 plant









I /^^C02 compensation points



i i i

C3 Plants

Wheat
Rice

Soybean
Barley

C4 Plants

Corn

Sorghum

Sugarcane

20	40	60	80

Intercellular C02 partial pressure, C\ (Pa)

100

PLANT PHYSIOLOGY, Fourth Edition, Figure 9.22 © 2006 Sinauer Associates, Inc.

-------
CO Yi eld Responses

I &

T—i—r

-I	L

—"—I—I	I—I	1—I—

FACE, wheat & barley

FACE, soybean

FACE, non-hybrid rice
FACE, hybrid rice

Kimball (1986) enclosures, wheat
Kimball (1986) enclosures, soybean
Long et al. (2006) FACE; Enclosures
Wheat, soybean, rice grain

Tubiello et al. (2007) from Amthor (2001)
Wheat grain in enclosures

Ainsworth et al. (2006) FACE; Enclosures
Wheat, soybean, rice grain

Ziska & Bunce (2007) wheat
FACE

	1 Open-top chambers

Tunnels

	I	Glasshouses

Growth chambers
Ziska & Bunce (2007) soybean

	1	FACE

Open-top chambers
SPAR units
Growth chambers

	1	Glasshouses

Ziska & Bunce (2007) rice
FACE

Open-top chambers
SPAR units
Tunnels
Glasshouses

-10

10

20

30

40

50

60

70

80

Relative C3 crop yield changes due to elevated C02 (%)

•	Biomass and yield with +200ppm were
increased by FACE in C3 species, but not in
C4 except under water stressed conditions.
Average C3 yield increase is -16% in FACE.

•	Low soil N often reduces these gains.

•	It appears unlikely that there is a significant
difference in the response of C3 grain crops to
elevated C02 between FACE and enclosure
experiments when the whole population of
enclosure experiments is included and their
variability is accounted for.

•	Important for simulation.

7

Kimball 2010

-------
Elevated C02 can also favor weeds

Crop

Weed

Increasing

c

[ CO2 ] favors Environment

Reference

A.	C4 Crops/C4 Weeds

So rg h u m Am a ran thus retroflexus

B.	C4Crops/C3 Weeds
Somhum Xanthium strumarium

CP

Sorghum Al but Hon theophrasti

C.	C3Crops/C3 Weeds
Soybean Chenopodium album
Lucerne Taraxacum officinale
Pasture Taraxacum and Plantago
Pasture Plantago lanceolate

D.	C3 Crops/C4 Weeds
Fescue Sorghum halapense
Soybean Sorghum halapense
Rice Echinoch 11oa glabrescens
Soybean A. retroflexus

Weed

Weed
Weed

Weed
Weed
Weed
Weed

Crop
Crop
Crop
Crop

Field

Glasshouse
Field

Field
Field
Field
Chamber

Glasshouse
Chamber
Glasshouse
Field

Ziska (2003)

Ziska (2001)

Ziska (2003)

Ziska (2000)

Bunce (1995)

Potvin and Vasseur (1997)
Newton et al. (1996)

Carter and Peterson (1983)

Patterson et al. (1984)

Alberto et al. (1996)

Ziska (2000)
	8

Ziska 2010

-------
Crop Response to Temperature

700

600

500

400

300

200

100













favored







D

\ /









\ /
/ \









/ \

\

C4 favored

i







•	Can shift photosynthesis
curve positively

•	Speed-up of phenology is
a negative pressure on yield

•	High-temperature stress
during critical growth
periods

•	T-FACE experiments now
underway.

10 20 30 40
Daytime growing-season temperature (°C)

PLANT PHYSIOLOGY, Fourth Edition, Figure 9.23 © 2006 Slnauer Associates, Inc.

-------
Yield Response to Water
Extreme events - Drought

Grain
yield 900-

(9 m-2)

600H

5 kPa

5 kPa

Har. Index = 0.5

i	1	r

100 200 300

400

i	1

500 600

•	Crops need water -
through precipitation or
irrigation

•	Drought stress affects
yield during critical growth
periods

•	Excess water can be
damaging as well

Transpirable water (mm)

Maximum grain yield plotted as a function of the amount of
transpirable soil water available through the growing season.
Two vapor pressure deficit environments are presented. C4
crops favored at both higher and lower water stress.

10

Sinclair 2010

-------
Extreme Events - Floods

~	Baseline

¦	HCGS 2030

~	CCGS 2030

~	HCGS 2090

¦	CCGS 2090

Damage to yield (%)

Number of events causing damage to maize yields due to excess soil moisture
conditions, averaged over all study sites, under current baseline (1951-1998) and
climate change conditions. Events causing a 20% simulated yield damage are

comparable to the 1993 US Midwest floods.	11

Rosenzweig et al. 2001

-------
AOGCM Projections of Surface Temperatures

/L 2020-2029



ifl



1



2090-2099



i liy\\



J]

B1

i j



_Q
03
_Q
O

CD
>

CC

-1 01 2345678
Global Average Surface Temperature Change (°C)

B1: 2020-2029

A1B: 2020-2029

A1B: 2090-2099

3

o

NJ

o
—*

§

I I I I I M

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5

Warm ing is Expected to be Greatest over Land
and at Most Hugh Northern Latitudes.
Hot Extremes and Heat Waves wrll
Continue to Become More Frequent

12

IPCC WGII AR4
Fig SPM-6

-------
Projected Patterns of Precipitation Changes

multi-model



DJF multi-model

A1B

JJA

%

©IPCC 2007: WG1-AR4

2090s

-20 -10 -5 5 10 20

Increases in Precipitation are Very Likely in the High-
Latitudes, whfle Decreases are (Likely n Most Subtropical

(Land Regions

Heavy Precipitation Events will Continue to

Become More Frequent

Droughts more frequent in some regions mcc AR4

13
Gl >

Figure SPM-7

-------
Projected Yield Changes 2050s

2050s	 2050s	

-2 5 d 2 5

Parry et al., 2004

Percent Change in Yield

Potential changes (%) in national cereal yields for the 2050s (compared with 1990)
under the HadCMS SRES A2a scenario with and without C02 effects (DSSAT)

Yield Effects with C02, rainfed wheat
CSIROA1B (DSSAT)

~	2000 old area lost
Yield loss > 25% of 2000'

~	.Yield loss 5-25%

Yield change within 5%

~	Yield gain 5-25%

Yield gain > 25%

~	2050 new area gained

IFPRI 2011

Parry et al.

-30% to +20%

IFPRI

-25% to +25%

GAEZ

-32% to +19%

GAEZ NASA 2009 rain-fed cereals Hadley A2
North America -7 to -1%; Europe -4 to 3;

Central Asia 14-19%; Southern Africa -32 to -29

Schlenker & Lobel Africa multi GCMs
-22 to -2% statistical approach ^

w/o adaptation

-------
Global Effects of Climate Change are Positive in
Short Term and Negative in Long Term

Percent Change in Food Production Potential

Inflection	WORLD

Points

m m m

80

0 123456789 10
0-10 = Severity of climate change (~time)

PRODUCTION potential with low crop response to C02
PRODUCTION potential with high crop response to C02
AREA EXTENT with low crop response to C02

AREA EXTENT with high crop response to C02	15

11 AS A

-------
Discuss the data and models used to make these

projections.

Are some modeling methods superior to others?

What are the main data requirements, spatial
resolution, and level of uncertainty in the outputs?

How are impacts expected to differ across
temperate and tropical regions?

-------
Statistical Approach

•	Uses historical data to estimate statistical relationships between
observed crop yields as a function of observed climate variables.

•	Uses these relationships to project the yield impact of changes in climate.
Advantages

Relationships should integrate biophysical responses to climate
variables; based on observations; data availability is improving.

Disadvantages

The approach does not explain process-based changes; does not represent
out-of-sample conditions; does not incorporate the effects of C02.

Data: yearly yield/aggregated 1° 4-hourly reanalysis, monthly growing season,

degree days climate; Spatial resolution: crop reporting districts; country level

	^

-------
Expert System Approach

•	Uses soil capability, climate, crop calendar, and simple productivity
relationships to estimate production potential of agricultural systems.

•	Use calculator to project effect of changes in climate on production
potential.

Advantages

Projects changes in both
production potential and
spatial extent of cropping
systems; global extent.

Disadvantages

Results not easily validated
in current climate.

Processes are represented
by simplified relationships.

Spatial distribution of land use

GAEZ Data: yearly yield/monthly climate; soils; crop calendars; ag systems;'

Spatial resolution 5'x5' lat/long

8

-------
Dynamic Process Crop Models

Advantages

•	Explicit simulation of processes affected by
climate, including C02 effects on growth and
water use.

•	Management practices included.

•	Cultivar characteristics can be tested for
'design' of adapted varieties.

•	Testable with experimental field data.

Data: daily T, P, SR; cultivar characteristics;
soils, management; yearly yield
Spatial resolution: Site-based; aggregated to
regions, countries

Disadvantages

•	Not all biophysical processes
included.

•	Aggregation from sites to regions
challenging.

•	Data availability varied.

Main Program

Land Unit Module

Primary Modules



Weather

Management

Secondary Modules

Soil

Organic Matter
Application

Fertilizer
Application

Soil - Plant -
Atmosphere

Soil Dynamics
-C Soil Water

Soil Inorganic N |
—| Soil Inorganic P |

Plant

{ Soil Temperature |

Potential
evapotransplration

Soil, flood, mulch
evaporation

Plant
Transpiration

-[ CROPGRO crops
—[ CERES Maize
—[ CERES Rice

CERES Wheat

—| SUBSTOR Potato
^Other crop models

Jones 2010^ g

-------
Cereal Yield Response

to Warming
Temperate vs. Tropical
Regions

With and Without
Simulated Adaptation

Temperate yields tend to
thrive until +3°C

Red = without adaptation
Green = with adaptation
	= reference line

a) Maize, Temperate

b) Maize. Tropical

1

Tropical yields tend to
decline immediately

Mean Local Temperature Change (°C)

c) Wheat, Temperate

Mean Local Temperature Change (°C)

d) Wheat, Tropical

6
¦£
3>
>-

60

40

20

-20

-40

-60

•

60
40
20



¦ H. *

¦ i • 4V f .

:		

!. ni „.
~ ~ ~ **•—- • ~ .

~

- a a

^ • }j i

* Si f

t

& i* rb
o o o c

*

% iTn1

r. . . . * jSi.



	]

0 12 3 4 5

Mean Local Temperature Change (°C)
e) Rice, Temperate

0 1 2 3 4 5 6
Mean Local Temperature Change (*C)

f) Rice, Tropical

IPCC, 2007

60
40

a>

1 20
6 „



60
40
20







¦

		

				

«C?-—



1 J ; ; :

Simple adaptations extend
temperate crops to +4-5 °C
but tropical yields only to
+2-3°C

	*—L	5	

-20
-40

	• 1 I J;.

|

~ ~





2 3 4 5 1
smperature Change (°C)

-60

5

J 1 2 3 4 5 20 e
Mean Local Temperature Change (°C)

-------
Projected Changes n Aggregate Cereal
Production in Sub Saharan Africa from
CI mate Change n 2046-2065

5 Percentile

Mean

95 Percentile

The benefits of adaptation are
uncertain.

-	A portfolio of strategies are
recommended

-	(e.g.) creating crops for both
drought and heat tolerance

There is a need to reduce the
uncertainty in how effective different
interventions are

-	It is recommended to accelerate
efforts to monitor and evaluate
current activities toward
adaptation.

Schlenker Lobell 2010

21

-------
Temperature

Precipitation

Carbon dioxide

Integrated impacts

River floods

Change in rice production (%)
(applies to all panels)

Median percentage changes in average pre-monsoon rice production in
sub-regions of Bangladesh based on 2040-2069 future climate simulations (as
compared to a 1970-1999 baseline). The impacts of changes in (clockwise
from bottom left) sea level rise, river floods, temperature, precipitation, and
carbon dioxide are presented absent other changes, along with a larger figure

showing the integrated production changes when all impacts are considered.
Ruane et al.. 2009; World Bank, 2009, Forthcoming	

Projected effects
of climate change

factors on
Bangladesh rice
production in the
2050s

22

-------
To what extent are changes in agricultural
practices and technologies capable of
modulating biophysical impacts?

23

-------
Progressive Levels of Adaptation

Challenges and Opportunities

E

p

c

CD

00

c

o

cS

Q-

"D

(0

Transformation from landuse or
distribution change

New products such as
ecosystem services

Production chain approaches
Climate change-ready germplasm
Diversification and risk management

Varieties, planting times, spacing

Stubble, water, nutrient and canopy
management etc

Climate change

Howden 20$®

-------
Adaptation is Not Always
Possible or Complete

Two examples for the CCGS 2030s Scenario

Spring wheat

Strategy: Early planting

Results: Successful heat stress
avoidance

10 r

(D 0
O)

£-10

sz

o

^ -20

cd

>-

sp -30

-40

Fargo ND Glasgow MT

Boise ID

¦

No adaptation

¦

Adaptation

Winter wheat

Strategy: Change of cultivar

Results: Unable to reverse damage
due to low precipitation

10 r

« °
O)

S -io

sz

o

^ -20

cd

>-

sp -30

-40

Dodge City KS North Platte NE Goodland KS

¦

No adaptation

¦

Adaptation

CC = Canadian Climate Centre GCM

U.S. National Assessment; Tubiello et al., 2d$2

-------
What are the most important gaps or
uncertainties in our knowledge regarding
biophysical responses of agro-ecosystems

to climate change?

What additional research would be most

valuable?

26

-------
Gaps and Uncertainties

a) 1970-1999 Baseline

12
10
8
6
4
2

"O
£

45

C
*03

i_

C

o
(/)
CD
0)

o

O)
c
03
-------
Ag

HSMIP

r

The Agricultural
Model Intercomparison
and Improvement Project

AgMIP Kickoff Workshop
October, 2010 Long Beach, CA

Information
Technologies:

On line Project
Guidance, Archive,
and Clearinghouse

Intercomparisons

Improve Crop and
Ag Trade Models
Gauge

.Uncertainties

Capacity Building

Vulnerability Assessment
Adaptation and Mitigation
Trade Policy Instruments
Technological Exchange

Extended
Applications

Water Resources
Pests and Diseases
Livestock and Pasture

i

AgMIP components and expected outcomes

Aggregation,	Uncertainty,

28

-------
-------
Research on Climate Change Impacts and Associated Economic Damages

Wolfram Schlenker
Columbia University and NBER

For most of human history, agriculture accounted for the dominant share of GDP and
employed most labor. Johnson (1997) estimates that in 1800 about 75-80% of the labor
force in developed nations was engaged in farming. Before 1930, production increases
were mainly driven by an expansion of the farming area while yields (output per area)
remained flat. The picture flipped around 1930, when production increases switched
from the extensive to the intensive margin: increases in output mainly came from
increases in yields, while the total farming area remained rather constant. Yields of most
commodities increased roughly threefold in the second half of the 19th century in the
United States as well as other developed countries. The large increase in yields has lead
to a general downward trend in agricultural prices over the 19th century. As a result,
agriculture now constitutes a small share of GDP in developed countries (2-3% in the
United States).

1)	Why impacts on US agriculture might be economically meaningful

While agriculture is a small share of GDP, it is arguably responsible for a large amount of
consumer surplus. GDP is simply the value of all produced goods and services in a
country. As far back as Adam Smith, researchers have examined the paradox of "value"
and asked why an essential good (water or food) can have a much lower value or price
than a nonessential good (diamonds). The reason is that the price of a product is
determined by its scarcity: food is currently abundant and therefore the price is low in
real terms. This, however, does not mean that changes in food production have small
impacts on welfare.

Demand for basic food is highly inelastic. The four basic commodities - corn,
soybeans, rice, and wheat - account for roughly 75% of the calories humans consume. A
demand elasticity of 0.05 for calories from these commodities implies that a 1% shortfall
in production increase prices by 20%. The recent tripling of commodity prices for the
basic four commodities has hardly impacted the amount of food consumed in developed
countries, yet reduced global consumer surplus by roughly 1.25 trillion dollars annually
(Roberts and Schlenker, 2010). Any shortfall in the production of basic food
commodities has the potential for large changes in welfare.

The U.S. is by far the largest producer of basic food calories and responsible for
23% of world caloric production of the four basic commodities. Its share of basic caloric
production is roughly three times as large as Saudi Arabia's share in oil production. Any
impact in the United States would have repercussions on world food markets simply due
to the dominating share of US production.

2)	Potential climate change impacts on US agriculture

Schlenker and Roberts (2009) use a new fine-scale weather dataset that incorporates the
whole distribution of temperatures within each day and across all days in the growing
season to estimate the influence of various temperatures on crop growth in a county-level

-------
panel analysis in the United States. Yields increase with temperature up to 29°C (84°F)
for corn and 30°C (86°F) for soybeans. If farmers could freely choose their growing
conditions, a temperature of, respectively, 84°F or 86°F every day all year long would be
ideal. Both lower and higher temperatures result in suboptimal yield growth. The
troublesome fact though is that the slope of the decline above the optimum is about ten
times steeper than the incline below it. In other words, being 1°F above the optimum
reduces yields ten times as much as being 1°F below it, or, equivalently, being 1°F above
the optimum reduces yields as much as being 10°F below it. The strong relationship
between temperatures above the optimum and yields implies that roughly half of the
year-to-year variation in crop yields can be explained by one single measure: how often
and by how much temperatures exceed the crop-specific optimum. The concept of
degree days simply adds all temperatures above the optimum for each day. One day that
is 10 degrees above the optimum is as harmful as 10 days that are 1 degree above the
optimum. Corn futures markets confirm this highly significant relationship: futures
prices for deliveries at the end of the growing season are highly sensitive to extreme heat
events during the growing season, but not average temperature.

Climate change is predicted to increase the daily minimum and maximum
temperatures. During the summer months, the minimum is usually below 84°F in the
Midwest, the major agricultural growing area in the United States. At the same time,
there are many days when the maximum temperature is above 86°F. Warming therefore
has countervailing effects: shifting minimum temperatures upward closer towards the
optimal growing temperature is beneficial for yields, however, shifting maximum
temperatures that already exceed the optimal levels further upward decreases yields.
Since the slope of the decline above the optimum is much steeper than the incline below
it, the latter effect dominates, resulting in sharp net yield losses for most climate
scenarios. Holding current growing regions fixed, area-weighted average yields are
predicted to decrease by -40% before the end of the century under the slowest (Bl)
warming scenario and decrease by -75% under the most rapid warming scenario (A1FI)
under the Hadley III model. Predicted temperature changes have larger effects that
predicted precipitation changes.

Year-to-year weather fluctuations are arguably different from permanent shifts in
climate. While the former are unknown at the time of planting, farmers can adapt to the
latter. To examine how farmers respond to changes in average condition, one can also
link average yields to average temperatures. A priori, one would have expected that
areas in the Southern United States that experience temperatures above 84-86°F more
frequently had an incentive to adapt to these temperatures and are hence less sensitive to
extreme heat. However, the same nonlinear and asymmetric relationship is found in the
time-series and cross-section. This suggests limited historical adaptation of seed varieties
or management practices to warmer temperatures because the cross-section includes
farmers' adaptations to warmer climates and the time-series does not. A model using
farmland values instead of crop yields finds similar predicted declines if one controls for
the damaging effects of extreme heat (Schlenker, Hanemann, and Fisher, 2006).
Moreover, the negative coefficient on extreme heat is highly robust to various
specification changes.

Similar relative sensitivities are found using a panel of yields in Africa (Schlenker
and Lobell, 2010). While countries in Africa are already hotter and hence more

-------
susceptible to further temperature increases, predicted temperature increases are lower
than in higher latitudes. Confidence bands on estimated yield-weather relationships are
larger in Africa where both yield and weather data are measured with less precision.

3)	Adaptation to climate change: evolution of heat tolerance

Given the large damaging effect of extreme heat on yields for at least two basic food
commodities (corn and soybeans), the big question becomes whether technological
innovation can reduce the sensitivity to these extreme temperatures. If changes in
climatic conditions reduce yields, prices would rise, giving seed companies a strong
incentive to innovate and make seeds more heat resistant. On the other hand, one might
wonder how difficult it is from a breeding standpoint to reduce heat tolerance.

The recent past might give us some guidance: while average corn yields increased
continuously in the second half of the 19th century by a total factor of three, the evolution
of heat sensitivity is highly nonlinear, growing with the adoption of double-cross hybrid
corn in the 1940's, peaking around 1960, and then declining sharply as single-cross
hybrids come online. Corn in Indiana, the state with the longest detailed daily weather
record, is most sensitive to extreme temperatures at the end of the sample. Since climate
change models predict an increase in extreme temperatures, the big question is whether
the next breeding cycles can increase both average yields and heat tolerance
simultaneously as in the period 1940-1960, or whether continued increases in average
yields can only be achieved at the expense of heat tolerance as in the period from 1960
onwards. Important areas for future research are to better understand how such
innovations could happen.

Genetically modified crops are the biggest hope to usher in a new era of
innovation that limits a plant's sensitivity to extreme heat. To date most commercially
successful genetically modified crops resist pests or herbicides. But more ambitious
efforts exist to develop plants that manufacture their own nitrogen fertilizer and possess
more nutrients. While public funding of basic research has diminished, private donations
from charities like the Gates Foundation or by profit-driven companies like Monsanto
might replace these funds. However, given public good attributes of research, there
remain important questions about the extent to which private incentives to fund basic
research align with potential social welfare.

4)	Biofuels as mitigation option: the US ethanol mandate and food prices

Previous sections highlighted the effect of changing climatic conditions on agricultural
yields. The reverse link has also received considerable attention: how does agriculture,
and more specifically agricultural policies, impact climate change? Forests store a large
amount of carbon, and most deforestation is done to convert forests to agricultural land.
Houghton et al. (1999) estimate that 10-30 percent of fossil fuel emissions in the United
States were offset by land use changes that lead to reforestation in the 1980s. By the
same token, biofuel policies, especially the US ethanol mandate, have received a lot of
attention as a tool to reduce CO2 emissions and limit climate change.

Roberts and Schlenker (2010) develop a new methodology to estimate both
demand and supply elasticities of agricultural commodities (maize, rice, soybeans, and
wheat). While current weather shocks have been used to estimate demand elasticities

-------
ever since P.J. Wright introduced the concept of instrumental variables, past weather
shocks can be used to estimate supply elasticities.

Since the estimated supply elasticity is roughly twice as large as the demand
elasticity, one third of the caloric input used in biofuel production comes from reduction
in food consumption while two thirds come from increases in food production. The US
ethanol mandate is predicted to decrease food consumption by 1% and increase
commodity prices by 20% assuming that one third of the calories used in ethanol
production are recycled as feedstock for animals. Future research should examine how
changes in the variance and correlation of weather shocks will impact food price spikes.

Lastly, the predicted increase in food prices due to biofuel mandates might lead to
expansion of agricultural areas, which, dependent on where they occur, might result in
significant increases in CO2 emissions (Searchinger et al., 2008). This is an ongoing
research area to correctly assess the effect of various mandates, e.g., the low carbon fuel
standard in California.

5) References

Houghton, R. A., J. L. Hackler, and K. T. Lawrence. 1999. "The U.S. Carbon Budget:
Contributions from Land-Use Change." Science, 285(5427):574-578.

Johnson, D. Gale, "Agriculture and the Wealth of Nations," American Economic Review,
May 1997,57(2), 1-12.

Roberts, M. J. and W. Schlenker, "Identifying Supply and Demand Elasticities of
Agricultural Commodities: Implications for the US Ethanol Mandate" NBER
Working Paper 15921, 2010.

Roberts, M. J. and W. Schlenker, "The Evolution of Heat Tolerance of Corn:
Implications for Climate Change," NBER Conference Volume: The Economics of
Climate Change - Adaptations Past and Present, 2011.

Searchinger, Timothy, Ralph Heimlich, R. A. Houghton, Fengxia Dong, Amani Elobeid,
Jacinto Fabiosa, Simla Tokgoz, Dermot Hayes, and Tun-Hsiang Yu. 2008. "Use
of U.S. Croplands for Biofuels Increases Greenhouse Gases Through Emissions
from Land-Use Change." Science, 319(5867): 1238-1240.

Schlenker, W., W. M. Hanemann, and A. C. Fisher, "The Impact of Global Warming on
U.S. Agriculture: An Econometric Analysis of Optimal Growing Conditions,"
Review of Economics and Statistics, 2006, 88(1), p. 113-125.

Schlenker, W. and M. J. Roberts, "Non-linear Temperature Effects Indicate Severe
Damages to U.S. Crop Yields under Climate Change," Proceedings of the
National Academy of Sciences, 2009, 106(37), p. 15594-15598.

Schlenker, W. and D. B. Lobell, "Robust Negative Impacts of Climate Change on
African Agriculture," Environmental Research Letters, 2010, 5(1), p. 1-8.

-------
Estimating the Economic Impact of Climate
Change in the Agricultural Sector

Wolfram Schlenker1

1Columbia University and NBER

Washington DC, January 27 2011

-------
Agriculture in US

Impacts

Technological Progress

Ethanol

Conclusions

OO

O

O

O

OO

Outline









Q Why Impacts on US Agriculture Might be Economically Meaningful
0 Potential Impacts of Climate Change on US Agriculture
Q Technological Progress: Evolution of Heat Tolerance
0 Agriculture's Role in Fighting Global Warming
o Conclusions

-------
Agriculture in US
OO

Impacts
O

Ethanol
O

Conclusions
OO

Outline

Q Why Impacts on US Agriculture Might be Economically Meaningful

Q Potential Impacts of Climate Change on US Agriculture
Q Technological Progress: Evolution of Heat Tolerance
Q Agriculture's Role in Fighting Global Warming
Q Conclusions

-------
Agriculture in US Impacts

Technological Progress

Ethanol

Conclusions

Role of Agriculture







Consumer Surplus







e Agriculture accounts for small share of US GDP

® 2-3% of GDP

a Does that mean impacts are negligible?

~ MS

5 O

-------
Agriculture in US	Impacts	Technological Progress	Ethanol	Conclusions

•O	O	O	O	OO

Role of Agriculture

Consumer Surplus

e Agriculture accounts for small share of US GDP

® 2-3% of GDP

a Does that mean impacts are negligible?

» Adam Smith: Paradox of value / price

» Why is the price of diamonds (nonessential good) so high, while the
price of water (essential good) is low
o Price of a good depends on scarcity!

-------
Agriculture in US Impacts

SO o

Technological Progress

Q

Ethanol

Q

Conclusions
OO

Role of Agriculture







Consumer Surplus







p

/N

Good 1
#

Good 2

	>



-------
Agriculture in US Impacts

SO o

Technological Progress

Q

Ethanol

Q

Conclusions
OO

Role of Agriculture







Consumer Surplus







p

/N

Good 1

Good 2

	> 4



-------
Agriculture in US Impacts

SO o

Technological Progress

Q

Ethanol

Q

Conclusions
OO

Role of Agriculture







Consumer Surplus







p

A









Good 1







Good 2

	^=»-



-------
Agriculture in US Impacts
SO O

Technological Progress

Q

Ethanol

Q

Conclusions
OO

Role of Agriculture







Consumer Surplus







1 >0^0

-------
Agriculture in US Impacts

SO o

Technological Progress

Q

Ethanol

Q

Conclusions
OO

Role of Agriculture







Consumer Surplus







-------
Agriculture in US	Impacts	Technological Progress	Ethanol	Conclusions

•O	O	O	O	OO

Role of Agriculture

Consumer Surplus

e Agriculture accounts for small share of US GDP

® 2-3% of GDP

a Does that mean impacts are negligible?

» Adam Smith: Paradox of value / price

» Why is the price of diamonds (nonessential good) so high, while the
price of water (essential good) is low
a Price of a good depends on scarcity!

» GDP is not a welfare measure

a Agricultural demand is highly inelastic (^0.05)
a Low price, but large consumer surplus

~ ~ ~ < 5 ~ < | ~ l

-------
Agriculture in US	Impacts	Technological Progress	Ethanol	Conclusions

•O	O	O	O	OO

Role of Agriculture

Consumer Surplus

e Agriculture accounts for small share of US GDP

® 2-3% of GDP

a Does that mean impacts are negligible?

» Adam Smith: Paradox of value / price

» Why is the price of diamonds (nonessential good) so high, while the
price of water (essential good) is low
a Price of a good depends on scarcity!

» GDP is not a welfare measure

a Agricultural demand is highly inelastic (^0.05)
a Low price, but large consumer surplus

» Climate impacts

a Small reduction in production result in
a Large price changes (inelastic demand)
a Potential for large welfare losses (consumer surplus)

0^0

-------
Agriculture in US Impacts
0» O

Technological Progress
O

Ethanol
O

Conclusions
OO

Role of Agriculture







Agriculture in US







» Four basic staple commodities

a Corn, rice, soybeans, and wheat

a 75% of calories humans consume worldwide

a Recent tripling of prices: 1.25 trillion dollar surplus loss per year

-------
Agriculture in US Impacts
0» O

Technological Progress
O

Ethanol
O

Conclusions
OO

Role of Agriculture







Agriculture in US







» Four basic staple commodities

a Corn, rice, soybeans, and wheat

a 75% of calories humans consume worldwide

a Recent tripling of prices: 1.25 trillion dollar surplus loss per year

® World caloric production

a Has been trending upward

» before 1940: Mainly area expansion
a after 1940: Mainly yield increases

a Real price has fallen over 20th century

~ MS

5 O

-------
Agriculture in US	Impacts	Technological Progress	Ethanol	Conclusions

0»	O	O	O	OO

Role of Agriculture

Agriculture in US

» Four basic staple commodities

a Corn, rice, soybeans, and wheat

a 75% of calories humans consume worldwide

a Recent tripling of prices: 1.25 trillion dollar surplus loss per year

® World caloric production

a Has been trending upward

» before 1940: Mainly area expansion
a after 1940: Mainly yield increases

a Real price has fallen over 20th century

» US share of caloric production

» Roughly constant around 23% (last 50 years)
a Larger than Saudi Arabia's share of oil production
a Impacts on US yields have potential to influence world markets

5 O

-------
Agriculture in US Impacts
0» O

Technological Progress
O

Ethanol
O

Conclusions
OO

Role of Agriculture







Agriculture in US







Year

-------
Agriculture in US

Impacts

Technological Progress



Ethanol

Conclusions

OO

O

O



O

OO

Outline











Q Why Impacts on US Agriculture Might be Economically Meaningful
0 Potential Impacts of Climate Change on US Agriculture

Q Technological Progress: Evolution of Heat Tolerance
Q Agriculture's Role in Fighting Global Warming
Q Conclusions

-------
Agriculture in US	Impacts	Technological Progress	Ethanol	Conclusions

OO	•	O	O	OO

Impact on Yields / Farmland Values

Link between Temperature and Yields

® Statistical Analysis

a Panel of county-level yields in Eastern United States
a Corn and Soybeans (two biggest staple commodities in US)
a Fine-scale weather (daily temperature / precip on 2.5mile grid)
« Years: 1950-2005

5 O

-------
Agriculture in US	Impacts	Technological Progress^M

OO	•	O

Impact on Yields / Farmland Values

Link between Temperature and Yields

® Statistical Analysis

a Panel of county-level yields in Eastern United States
a Corn and Soybeans (two biggest staple commodities in US)
a Fine-scale weather (daily temperature / precip on 2.5mile grid)
« Years: 1950-2005

» Model accounts for

a Amount of time spent in each 1UC interval
a Quadratic in total precipitation
a State-specific quadratic time trends
a County fixed effects

~ MS

5 O

-------
Agriculture in US
OO

Impact on Yields / Farmland Values

Technological Progress

Link between Temperature and Yields

Conclusions
OO

Panel of Corn and Soybean Yields

	Step Function

	Polynomial (8th-order)

	Piecewise Linear

15 20 25 30 35
Temperature (Celsius)

10 15 20 25
Temperature (Celsius)

~ MS

-------
Agriculture in US
OO

Impact on Yields / Farmland Values

Technological Progress

Conclusions
OO

Link between Temperature and Yields

Corn and Soybean Yields - Various Source of Identification

-

-	Cress sc:ti:n

-	CDS! iccto" l C:ntro ;¦

-	""i-ii:

— Panel

—Cross-section
-Cross-section (Soil Controls)
—Time Series

10

15

20

10

15

20

Temperature (Celsius)

Temperature (Celsius)

~ MS

5 O

-------
Agriculture in US
OO

Impact on Yields / Farmland Values

Technological Progress

Conclusions
OO

Link between Temperature and Yields

Climate Impacts - Hadley III model (Significant Warming)

-20

-40

-60

-80

-IOC-

Corn

B1 B2

A2 A1

}H i i :

1 f : i ( r 11

	Panel



	Cross-section



	Cross-section (Soil Controls)



	Time Series



~ MS

5 O

-------
Agriculture in US
OO

Impact on Yields / Farmland Values

Technological Progress

Conclusions
OO

Link between Temperature and Yields

Climate Impacts - Hadley III model (Significant Warming)

-20

-40

	Corn	

B1	B2	A2

A1

-60

-80-

-100L

~ MS

5 O

-------
Agriculture in US	Impacts	Technological Progress	Ethanol	Conclusions

OO	•	O	O	OO

Impact on Yields / Farmland Values

Link between Temperature and Yields

o Other parts of globe

a Tropics: hotter baseline but lower temperature increases

» Can crop switching save the day?

» Cross-sectional analysis of farmland values
® Accounting for extreme heat
» Limit to Eastern United States

® Similar results

o Large negative effect of extreme heat
® Robust to myriad of specification checks
® Different census years
® Permutations of other control variables

~ ~ ~ < 5 ~ < | ~ l

-------
Agriculture in US

Impacts

Technological Progress

Ethanol

Conclusions

OO

O

O

O

OO

Outline









Q Why Impacts on US Agriculture Might be Economically Meaningful
Q Potential Impacts of Climate Change on US Agriculture
Q Technological Progress: Evolution of Heat Tolerance
Q Agriculture's Role in Fighting Global Warming
Q Conclusions

-------
Agriculture in US Impacts

oo o

Technological Progress
•

Ethanol
O

Conclusions
OO

Long-term Study of Indiana







Evolution over Time







Weather in Indiana

~ MS

5 O

-------
Agriculture in US Impacts
OO O

Technological Progress
•

Ethanol
O

Conclusions
OO

Long-term Study of Indiana







Evolution over Time







Average Yields

200 	T	T	T	T	

180

1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
Year



-------
Agriculture in US Impacts

oo o

Technological Progress
•

Ethanol
O

Conclusions
OO

Long-term Study of Indiana







Evolution over Time







Corn: Weather-Yield Relationship

Corn

~ MS

5 O

-------
Agriculture in US Impacts

oo o

Technological Progress
•

Ethanol
O

Conclusions
OO

Long-term Study of Indiana







Evolution over Time







Heat Tolerance

-------
Agriculture in US	Impacts	Technological Progress	Ethanol	Conclusions

oo	o	o	o	oo

Outline

Q Why Impacts on US Agriculture Might be Economically Meaningful
Q Potential Impacts of Climate Change on US Agriculture
Q Technological Progress: Evolution of Heat Tolerance
Q Agriculture's Role in Fighting Global Warming
Conclusions

-------
Agriculture in US	Impacts	Technological Progress	Ethanol	Conclusions

OO	O	O	•	OO

Agriculture's Impact Climate Change

Agriculture and C02 Emissions

® Land use change responsible for ~20% of CO2 emissions
a Deforestation releases a lot of carbon

~ MS

-------
Agriculture in US
OO

Agriculture's Impact Climate Change

Technological Progress

Agriculture and C02 Emissions

® Land use change responsible for ^20% of CO2 emissions
a Deforestation releases a lot of carbon

® Use of biofuels to reduce CO2 emissions

» Long history of biofuels: Model T ran on ethanol
o Renewed interest due to climate change

~ MS

-------
Agriculture in US	Impacts	Technological Progress	Ethanol	Conclusions

OO	O	O	•	OO

Agriculture's Impact Climate Change

Agriculture and C02 Emissions

® Land use change responsible for ~20% of CO2 emissions
a Deforestation releases a lot of carbon

® Use of biofuels to reduce CO2 emissions

a Long history of biofuels: Model T ran on ethanol
a Renewed interest due to climate change

® Concern: Indirect land use change / food prices
a 2009 Renewable Energy Standard

a 5% of world caloric production of 4 basic staples go into biofuels
a Predicted commodity price increase 20% (one third recycling ratio)

a Reduces consumer surplus by 150 bi I lion annually
a Supply mainly from extensive margin

o Indirect land use change impacts CO2 balance

~ ~ ~ < 5 ~ < | ~ l >0^0

-------
Agriculture in US	Impacts	Technological Progress	Ethanol	Conclusions

OO	O	O	•	OO

Agriculture's Impact Climate Change

Agriculture and C02 Emissions

® Price increase / indirect land use change

a Depends on demand and supply elasticity for calories

o Study identifying demand / supply elasticities
a Demand: global yield shocks
a Supply: lagged yield shocks

» Data

a Yearly country level yield shocks

a Combine 4 basic staples (corn, rice, soybeans, and wheat)
a Futures price data (Chicago Board of Trade)

~ ~ ~ < 5 ~ < | ~ l

-------
Agriculture in US	Impacts	Technological Progress	Ethanol	Conclusions

OO	O	O	•	OO

Agriculture's Impact Climate Change

Agriculture and C02 Emissions







Model







2SLS

3SLS

2SLS

3SLS

2SLS

3SLS





Demand and Supply Elasticities





&cl

-0.0505***

-0.0554***

-0.0641**

-0.0797***

-0.0668***

-0.0634***

(se)

(0.0190)

(0.0167)

(0.0243)

(0.0215)

(0.0241)

(0.0226)

/3S

0.1165***

0.1337***

0.0826***

0.0951***

0.0957***

0.0979***

(se)

(0.0286)

(0.0241)

(0.0217)

(0.0189)

(0.0208)

(0.0189)

A p

31.41

27.01

36.10

29.31

32.14

32.16

(95%)

(21.32,50.14)

(20.69,36.62)

(23.75,60.31)

(22.01,40.80)

(22.23,50.00)

(22.79,48.40)

N

42

42

42

42

41

41

1

2

2

3

3

3

3

K

1

1

1

1

2

2

@d

Demand elasticity











P s

Supply elasticity











A p

Predicted price increase (0% recycling

as feed stock)







-------
Agriculture in US	Impacts	Technological Progress	Ethanol	Conclusions

oo	o	o	o	oo

Outline

Q Why Impacts on US Agriculture Might be Economically Meaningful
Q Potential Impacts of Climate Change on US Agriculture
Q Technological Progress: Evolution of Heat Tolerance
Q Agriculture's Role in Fighting Global Warming
o Conclusions

-------
Agriculture in US	Impacts	Technological Progress	Ethanol	Conclusions

OO	O	O	O	»o

Major Findings

Conclusions

a Agriculture responsible for large consumer surplus
a Food is essential good

» US produces 23% of calories (four basic staples)

5 O

-------
Agriculture in US	Impacts	Technological Progress	Ethanol	Conclusions

OO	O	O	O	»o

Major Findings

Conclusions

a Agriculture responsible for large consumer surplus
a Food is essential good

» US produces 23% of calories (four basic staples)

a Potential impact of climate change
a Driving force: extreme heat

a Large yield decline if maximum temperature rises a lot
a Impact depends on baseline and predicted increase

~ MS

5 O

-------
Agriculture in US	Impacts	Technological Progress	Ethanol	Conclusions

OO	O	O	O	»o

Major Findings

Conclusions

a Agriculture responsible for large consumer surplus
a Food is essential good

» US produces 23% of calories (four basic staples)

a Potential impact of climate change
a Driving force: extreme heat

a Large yield decline if maximum temperature rises a lot
a Impact depends on baseline and predicted increase

o Technological progress

a Average yields have risen steadily since 1940
a Not true for heat tolerance

~ MS

5 O

-------
Agriculture in US	Impacts	Technological Progress	Ethanol	Conclusions

OO	O	O	O	»o

Major Findings

Conclusions

a Agriculture responsible for large consumer surplus
a Food is essential good

» US produces 23% of calories (four basic staples)

a Potential impact of climate change
a Driving force: extreme heat

a Large yield decline if maximum temperature rises a lot
a Impact depends on baseline and predicted increase

o Technological progress

a Average yields have risen steadily since 1940
a Not true for heat tolerance

» Ethanol mandate

a Requires 5% of caloric production (4 basic staples)
a Predicted to increase commodity prices by 20%
a Indirect land use change

~ MS

5 O

-------
Agriculture in US

Impacts

Technological Progress

Ethanol

Conclusions

OO

O

O

O

O#

Major Findings









References









Papers with more detail available:
www.columbia.edu/^ws2162

~ MS

5 O

-------
Climate-Associated Changes in Health Outcomes

Kristie L. Ebi, Ph.D., MPH
Carnegie Institution for Science

Introduction

Climate change has the potential to affect any health outcome that is seasonal or that is
associated with weather and climate. In addition, many key determinants of human
health, such as food and freshwater availability, are strongly influenced by weather and
climate. Climate-sensitive health outcomes include injuries, illnesses, and deaths
associated with extreme weather events, and the effects of changing weather patterns
mediated through ecological systems, such as water- and food-borne diseases,
vectorborne and zoonotic diseases, respiratory diseases associated with ground-level
ozone and aeroallergens, and undernutrition. Climate change also may result in resource
depletion and other processes that could lead to large-scale migration, with associated
health impacts. While negative health effects are projected for all countries, the largest
impacts are expected in lower-income populations, primarily those living in tropical and
subtropical countries.

Health Risks of Climate Change

Infectious Diseases

Climate is a primary determinant of whether a particular location has environmental
conditions suitable for the transmission of a range of infectious diseases. Increasing
temperatures could affect vector and rodent borne diseases, in terms of the density of
insects and rodents in a particular area (and therefore the likelihood of infection) and by
changing the geographic range of the vector and pathogen. Expansion in range can
expose new populations who have little or no immunity to new infections, which could
result in large disease outbreaks. Although understanding of the potential impacts of
climate change on infectious diseases is still in its relatively early stages, expert
assessments have concluded that climate change is expected to be among the most
important drivers of infectious disease in the future.1 A UK review considered scenarios
for the next 10-25 years of infectious diseases in humans, animals, and plants for the UK
and sub-Saharan Africa, and aimed to produce a vision of new systems needed for
disease detection, identification and monitoring. The key driver in the UK was expected
to be increasing ambient temperature. In Africa, where people, animals and crops live in
conditions of much greater moisture stress, rising temperature were still considered to be
important but less so than changes to rainfall patterns and the frequency of droughts.
Climate-change mediated spread of infectious diseases was expected to cause direct
human suffering, especially in Africa, and increasingly challenge current production
systems of livestock and crops in the UK and Africa.

Malaria is the most important vectorborne disease in the world; it is also a preventable
disease. About 40% of the world's population is at risk of contracting malaria, and
roughly 75% of cases occur in Africa, with the remainder occurring in Southeast Asia,
the western Pacific, and the Americas.11 In sub-Saharan Africa, malaria remains the most
common parasitic disease and is the main cause of morbidity and mortality among

1

-------
children less than five years of age and among pregnant women.111 The 1990 Global
Burden of Disease study estimated that malaria accounted for approximately 10.8% of
years of life lost across sub-Saharan Africa.1V

There has been a great deal of interest in modeling how the incidence and geographic
range of malaria could change under different climate change projections. Results from
several models suggest that climate change could alter the season of transmission and
geographic range of malaria in Africa, particularly sub-Saharan Africa/ The results
suggest that climate change will be associated with geographic expansions of the areas
suitable for stable falciparum malaria in some regions and with contractions in others; the
projected areas of expansion are larger than the projected areas of contraction. For
instance, Ethiopia, Zimbabwe, and South Africa are projected to show increases of more
than 100% in person-months of exposure later in this century, changes that could
dramatically increase the burden of those suffering with malaria."

Studies have shown that some areas in Asia are projected to be at increased risk of
malaria, while reductions have been projected for some areas in Central America and
around the Amazon, due to decreases in rainfall/11 An assessment in Australia based on
climatic suitability for the main Anopheline vectors projected a likely southward
expansion of habitat, although the future risk of endemicity would remain low due to the
capacity to respond/111

Climate change could affect the incidence and geographic range of a large number of
vectorborne and zoonotic diseases of concern include dengue fever, Lyme disease, plague,
Chagas disease, Rift valley fever, and leishmaniasis; expansions and contractions of
ranges are possible as ecosystems and transmission pathways change with changing
weather patterns.1X

Several food- and waterborne diseases that cause significant numbers of cases of illness
are climate sensitive, suggesting that climate change may affect their incidence and
distribution. For example, an approximately linear association between temperature and
common forms of food-borne diseases such as salmonellosis suggests increasing cases
with increasing temperature/ Limited projections suggest these risks could increase with
climate change/1

Air Pollution

In some regions, climate change may increase concentrations of selected air pollutants,
particularly ozone, and could decrease concentration of other pollutants, such as
particulate matter (due to increasing heavy precipitation events). There is extensive
literature documenting the adverse health impacts of exposure to elevated concentrations
of air pollutants. In 2000, there were 800,000 deaths from respiratory problems, lung
disease, and cancer that were attributed to urban air pollution, with the largest burden in
low-income countries in the Western Pacific and South East Asia/11 In addition, there
were 1.6 million deaths attributed to indoor air pollution caused by burning biomass fuels,
such as wood and dung.

More is known about the potential impacts of climate change on ground-level ozone than
on other air pollutants. Acute exposure to elevated concentrations of ozone is associated

2

-------
with increased hospital admissions for pneumonia, chronic obstructive pulmonary disease,
asthma, allergic rhinitis and other respiratory diseases, and with premature mortality.xm

Changes in concentrations of ground-level ozone driven by scenarios of future emissions
and /or weather patterns have been projected for Europe and North America, with most
projections suggesting increasing concentrations.X1V xv Higher ozone concentrations will
likely increase a range of health problems and increase premature mortality in susceptible
individuals.™ Despite the heavier pollution burdens, no studies have been conducted for
cities in low- or middle-income countries.

Malnutrition

Climate change threatens human health through its effect on under-nutrition and food
insecurity. More than 800 million people are undernourished, causing over 15% of the
total global disease burden, and over three billion people are micronutrient deficient.™1
The prevalence of undernourishment has fallen over recent decades, with reductions in
Asia and Latin America partly offset by increases in Africa and the Middle East. Almost
60% of the world's undernourished people live in South Asia, while the highest incidence
of undernourishment is in Sub-Saharan Africa, where more than one-third of the
population is underfed.

Recent projections suggest that half of the world's population could face severe food
shortages by the end of the century as rising temperatures take their toll on farmers'
crops; a greater proportion of this will be in Africa.xvm Harvests of staple food crops
such as rice and maize could fall by between 20% and 40% as a result of higher
temperatures during the growing season in the tropics and sub-tropics. Although data are
limited, malnutrition associated with drought and flooding may be one of the most
important consequences of climate change due to the large number of people that may be
affected.xix

Extreme Weather Events

The adverse health consequences of flooding and windstorms often are complex and far-
reaching, and include the physical health effects experienced during the event or clean-up
process, effects brought about by damage to infrastructure related to water supply,
sanitation, and drainage, and population displacement.xx Extreme weather events are also
associated with mental health effects, such as post-traumatic stress disorder, resulting
from the experience of the event or from the recovery process. These psychological
effects tend to be much longer lasting and may be worse than the direct physical
effects.XX1 More than 90% of the disasters that occurred in 2007 were the result of
extreme weather- or climate-related events, together accounting for 95% of the reported
fatalities and 80% of the total USD82 billion economic losses. The health impacts of
extreme events in low- and middle-income countries are substantially larger.

Heat waves affect human health via heat stress, heatstroke, and death,xxu as well as
exacerbating underlying conditions that can lead to an increase in mortality from all
causes of death.xxm Older adults, children, city-dwellers, the poor, and people taking
certain medications are at the highest risk during a heat wave. The numbers of heat-
related deaths are projected to increase with climate change.XX1V

3

-------
Projections suggest that regions affected by moderate droughts are set to double by the
end of the century, with areas affected by extreme droughts increasing from 1% today to
30% in 2100. The most striking impact is expected in parts of southern Europe, North,
West and Southern Africa, western Eurasia, and the US. The loss of livelihoods due to
drought is a major trigger for population movements that may cause additional adverse
health burdens. The effects of drought on health include malnutrition (protein-energy
malnutrition and/or micronutrient deficiencies), infectious and diarrheal diseases, and
respiratory diseases.xxv Droughts, especially in rural areas, have a tendency to influence
migration into cities, increasing urbanization and stressing the socio-economic conditions
already affected by high levels of city population growth.

Prolonged droughts fuel fires, releasing respiratory pollutants, while floods can create
mosquito breeding sites, foster fungal growth, and flush microbes, nutrients and
chemicals into bays and estuaries, causing water-borne disease outbreaks from organisms
likeis. coli and Cryptosporidium.XXV1

Global Assessments of the Health Impacts of Climate Change

The most comprehensive evaluation of the health burden due to climate change used a
comparative risk assessment approach to estimate total health burdens from climate
change in 2000 and 2030, and to project how much of this burden might be avoided by
stabilizing greenhouse gas (GHG) emission.xxv" The health outcomes (diarrhoea, malaria,
malnutrition, heat-related mortality, and injury from floods and landslides) were chosen
based on sensitivity to climate variations, likely future importance, and availability of
quantitative global models (or the feasibility of constructing them). The projected
relative risks attributable to climate change in 2030 vary by health outcome and region,
and are largely negative, with the majority of the projected health burden due to increases
in diarrheal disease and malnutrition, primarily in low-income populations already
experiencing a large burden of disease. The study is described in more detail in the
Annex.

These results are consistent with a review that concluded that health risks are likely to
increase with increasing global mean surface temperature, particularly in low latitude
countries.xxvm Actual health burdens depend on assumptions of population growth,
future baseline disease incidence, and the extent of adaptation.

Research Needs

A recent cross-agency working group in the U.S. summarized the research needs for
better understanding of the linkages between climate change and health.XX1X Overarching
themes include focusing on systems and complexity, enhancing risk communication and
public health education, co-benefits of mitigation and adaptation strategies, and urgency
and scope

•	Improve characterization of exposure- response relationships, particularly at
regional and local levels, including identifying thresholds and particularly
vulnerable groups. This needs to be done within the context of complex systems.

•	Collect data on the early effects of changing weather patterns on climate-sensitive
health outcomes.

•	Collect and enhance long-term surveillance data on health issues of potential

4

-------
concern, including vectorborne and zoonotic diseases, air quality, pollen and mold
counts, reporting of food- and water-borne diseases, morbidity due to temperature
extremes, and mental health impacts from extreme weather events.

•	Develop quantitative models of possible health impacts of climate change that can
be used to explore the consequences of a range of socioeconomic and climate
scenarios.

•	Understand local- and regional-scale vulnerability and adaptive capacity to
characterize the potential risks and the time horizon over which climate risks might
arise.

•	Develop downscaled climate projections at the local and regional scale in order to
conduct the types of vulnerability and adaptation assessments that will enable
adequate response to climate change, and to determine the potential for interactions
between climate and other risk factors, including societal, environmental, and
economic.

•	Improve understanding of designing, implementing, and monitoring effective and
efficient adaptation options.

•	Understand the co-benefits of mitigation and adaptation strategies.

•	Enhance risk communication and public health education.

5

-------
ANNEX

Estimating Current and Future Population Health Burdens Attributable to Climate
Change: the WHO Global Burden of Disease Study

The first global estimate of current and possible future population health burdens
attributable to climate change was conducted as part of the World Health Organization's
Global Burden of Disease (2000) project (McMichael et al. 2004). This study remains
the most comprehensive projection of the health impacts of climate change.

The Global Burden of Disease study used published information on climate-health
(exposure-effect) relationships to estimate the proportion of the actually observed cases
of a specified disease (e.g., malaria or child diarrhea) that could be reasonably
attributable to climate change. The steps to estimate the current attributable burden of
disease and premature death were:

(i)	Determine or estimate changes in temperature (and other climate variables) over the
recent past.

(ii)	Determine (to the extent possible given data limitations), for each disease of interest,
the current rates of incidence or premature death, by geographic region.

(iii)	Determine from the published scientific literature, for each disease of interest, the
increase in disease risk per unit increase in temperature or other climate variable (i.e.
the relative risk).

(iv)	Apply the relative risk to the existing rates of disease or death to estimate the
'population attributable fraction' (assuming all persons are equally exposed to the
change in climate).

Estimation of the attributable burden of disease and premature death was limited to
malaria, malnutrition, diarrheal disease, and floods, plus, as a minor contribution, the
impacts of heatwaves. It is important to note that this study was conservative because it
was limited to those health outcomes for which the baseline climate-health relationship
had already been reasonably well characterized in the literature. Also, cautious
assumptions were made that the health risks would be significantly reduced with
economic development.

The same method was used to estimate the future burden of disease and premature death
attributable to climate change for the year 2030. This requires, for a specified future
time:

•	A modeled scenario of global climate change, geospatially differentiated at the
appropriate scale.

•	Estimations, by region/country, of population size, and age structure.

•	Estimations, by region/country, of the future baseline (counter-factual) rates of
disease incidence or premature death.

•	Assumptions about the applicable relative risk (e.g. does it stay constant, increase,
or decrease over time, given that there will be changes in the target population,
including changes due to adaptive actions).

6

-------
2030 was chosen as the time horizon because, among other reasons, beyond a few
decades into the future, there is increasing uncertainty about trends in social, economic,
and political circumstances, population living conditions, and the background population
health profile.

Study Details

The World Health Organization (WHO) Global Burden of Disease study began in 1992
with the objective of quantifying the burden of disease and injury in human populations
(Murray and Lopez 1996). The burden of disease refers to the total amount of premature
death and morbidity within a population. The goals of the study were to produce the best
possible evidence-based description of population health, the causes of lost health, and
likely future trends in health in order to inform policy-making. The WHO Global Burden
of Disease 2000 project (GBD) updated the earlier study (Murray et al. 2002). It drew on
a wide variety of data sources to develop internally consistent estimates of incidence,
prevalence, and mortality, and severity and duration, for over 130 major health outcomes,
for the year 2000 and beyond.

To the extent possible, the GBD synthesized all relevant epidemiologic evidence on
population health within a consistent and comprehensive framework, the comparative
risk assessment. Twenty-six risk factors were assessed, including major environmental,
occupational, behavioral, and lifestyle risk factors. Climate change was one of the
environmental risk factors assessed (McMichael et al. 2004).

The GBD used two summary measures of population health, mortality and the Disability
Adjusted Life Years lost (DALYs) (Murray and Lopez 1996). DALYs provide a better
measure than mortality of the population health impacts of diarrheal diseases,
malnutrition, and malaria. The attributable burden of DALYs for a specific risk factor
was determined by estimation of the burden of specific diseases related to the risk factor;
estimation of the increase in risk for each disease per unit increase in exposure to the risk
factor; and estimation of the current population distribution of exposure, or future
distribution as estimated by modeling exposure scenarios.

For climate change, the questions addressed were what would be the total health impact
caused by climate change between 2000 and 2030, and how much of this burden could be
avoided by stabilizing greenhouse gas emissions (McMichael et al. 2004). The
alternative exposure scenarios were:

•	Unmitigated emission trends (UE) (i.e. approximately following the IPCC IS92a
scenario);

•	Emissions reductions resulting in stabilization at 750 ppm C02-equivalent by 2210
(s750); and

•	Emissions reductions resulting in stabilization at 550 ppm C02-equivalent by 2170
(s550).

Climate change projections were generated using the HadCM2 global climate model
(Johns et al. 2001). The health outcomes included were chosen based on sensitivity to
climate variation, predicted future importance, and availability of quantitative global
models (or feasibility of constructing them); these were:

7

-------
•	the direct health impacts of heat and cold,

•	episodes of diarrheal disease,

•	cases of Plasmodium falciparum malaria,

•	fatal unintentional injuries in coastal floods and inland floods/landslides, and

•	estimated prevalence of malnutrition (indicated by non-availability of recommended
daily calorie intake).

Both global and WHO region-specific estimates were generated.

Results From the WHO Global Burden of Disease Project

Table 1 summarizes the health outcomes included, as well as the assumed mechanism by
which climate change induces each of the specified health outcomes.

Table 1: Health Outcomes Included in the WHO Global Burden of Disease Project

Class

Mechanism

Outcome

Direct impacts of heat and
cold:

Water-washed, waterborne,
and foodborne disease:
Vector-borne disease:

Natural disasters*

Risk of malnutrition

Thermal stress due to higher temperatures Cardiovascular disease deaths

Higher temperatures encourage proliferation
of bacterial pathogens
Rainfall and temperature affect vector
abundance. Temperature affects incubation
period of parasite in mosquito
Temperature affects incubation period of
virus in mosquito

Increased floods and landslides due to sea
level rise and extreme rainfall.

Changes in food production and per capita
food availability.	

Diarrhoea episodes
Malaria cases

Dengue cases

deaths due to unintentional
injuries

other unintentional injuries (non-
fatal)

non-availability of recommended
daily calorie intake	

Source: McMichael et al. 2004

For the year 2000, the mortality attributable to climate change was estimated to be
154,000 (0.3%) deaths, and the attributable burden was 5.5 million (0.4%) DALYs, with
approximately 50% of the burden due to malnutrition (McMichael et al. 2004). These
estimates are for the year 2000, by which time the amount of climate change since the
selected baseline year (1990) was small (approximately 0.2°C). Therefore, future disease
burdens would be expected to increase with increasing climate change, unless
(implausibly) fully effective adaptation measures were implemented.

Approximately 46% of the DALYs attributable to climate change were estimated to have
occurred in the WHO South-East Asia Region (which includes South Asia), 23% in
countries in the Africa region with high child mortality and very high adult male
mortality, and 14% in countries in the Eastern Mediterranean region with high child and
adult male mortality.

Table 2 summarizes the estimated numbers of deaths occurring in 2000 as a result of the
impacts of climate change on the occurrence of the five specified health outcomes

8

-------
amenable to quantitative modeling (see Annex 2 for WHO regions). Figure 3 maps these
results by WHO region.

Table 2: Estimated mortality (000s) attributable to climate change in the year 2000,
by cause and WHO region

Table 20.16 Estimated mortality (000s) attributable to climate change in
the year 2000, by cause and subregion

Subregion

Malnutrition

Diarrhoea

Malaria

Floods

CVD

All causes

Total deathsfmillion
population

AFR-D

8

5

5

0

1

19

66.83

AFR-E

9

8

IS

0

1

36

109.40

AMR-A

0

0

0

0

0

0

0.15

AMR-B

0

0

0

1

1

2

3.74

AMR-D

0

1

0

0

0

1

10.28

EMR-B

0

0

0

0

0

1

5.65

EMR-D

9

s

3

1

1

21

61.30

EUR-A

0

0

0

0

0

0

0.07

EUR-B

0

0

0

0

0

0

1.04

EUR-C

0

0

0

0

0

0

0.29

SEAR-B

0

1

0

0

1

2

7.91

SEAR-D

52

22

0

0

7

80

65.79

WPR-A

0

0

0

0

0

0

0.09

WPR-B

0

2

1

0

0

3

2.16

World

77

47

27

2

12

166

27.82

CVD Cardiovascular disease. As described in section 3,6, the estimated cardiovascular deaths represent
temperature-related mortality displacement. Therefore no disease burden is estimated for deaths from this
cause in Table 20.1 7.

9

-------
Figure 3: Burden of Premature Deaths Attributable to Climate Change, for Year 2000

Deaths from climate change

Estimates by WHO sub-region for 2000 (World Health Report, Geneva, WHO, 2002)
No GBlB	® world Health Organization 2005. All rights reserved.

v 	 	O			 'J —	—	'J	7

Malaria in 2030

The WHO GBD study used the calculated relative risks to estimate the excess number of
incident cases of diarrheal diseases, malnutrition, and malaria in 2030 for the three
scenarios (unmitigated emissions (UE) and stabilization scenarios at 550 and 750 ppm
C02-equivalent).

Diarrheal Diseases

For the estimations for diarrheal diseases, developing countries were defined as those
with per capita incomes less than US$6,000/year in 1990 US dollars. For such countries,
the exposure-response relationship used was a 5% increase in diarrheal incidence per °C
increase in temperature; this estimate was based on two studies (Checkley et al. 2000;
Singh et al. 2001). The study assumed that the climate sensitivity of diarrhea would
decrease with increasing GDP; once a country was projected to reach per capita incomes
of UD$6,000/year, then overall diarrhea incidence was assumed to not respond to
changes in temperature. The study assumed that diarrheal incidence in richer countries is
insensiti ve to climate change.

The relative risks for each region are a population-weighted average of the countries
within the region. The model output was used to generate mid-range estimates; the high
relative risks were calculated as a doubling of the mid-range estimate.

Malnutrition

Estimates of national food availability were based on the effects of temperature and
precipitation, and the beneficial effects of higher C02 levels, projected using the

10

-------
IBSNAT-ICASA dynamic crop growth models (IBSNAT 1989). Principal characteristics
of this model include:

•	No major changes in the political or economic context of world food trade or in food
production technology;

•	Demographic change follows the World Bank mid-range estimate (i.e. 10.7 billion
by the 2080s);

•	GDP to accumulate as projected by EMF14 (Energy Modeling Forum 1995); and

•	A 50% trade liberalization in agriculture is introduced gradually by 2020.

Note that malnutrition has multiple causes. Access to a range of affordable quality foods
is required for adequate nutrition. There may be sufficient food production within a
country, but families may not have access because the food is not culturally desirable, it
is too expensive, or there is inadequate transportation. Subsistence farmers and the urban
poor are particularly at risk. Therefore, economic and political factors can be as
important as climate in determining food availability. However, this model focused only
on the association between climatic factors (including CO2) and national food availability

Analyses suggested that the model output was positively related to more direct measures
of malnutrition, including incidence of underweight, stunting, and wasting in children <5
years of age. The relative risks of malnutrition were interpreted as being directly
proportional to the incidence of underweight. Again, the model output was used to
generate mid-range estimates; the high relative risks were calculated as a doubling of the
mid-range estimate.

Malaria

Estimates for the projected populations at risk of Plasmodium falciparum malaria were
based on the MARA/ARMA model (MARA/ARMA 1998). As for other health
outcomes, the model output was used to generate mid-range estimates; the high relative
risks were calculated as a doubling of the mid-range estimate.

The total estimated excess numbers of cases are shown in Tables 1-3 in Annex 3.

Summary

The projected relative risks attributable to climate change in 2030 vary by health outcome
and region, and are largely negative, with the majority of the projected disease burden
due to increases in diarrheal disease and malnutrition, primarily in low-income
populations already experiencing a large burden of disease (McMichael et al. 2004).
Absolute disease burdens depend on assumptions of demographic change, future baseline
disease incidence, and the extent of adaptation. Table 3 summarizes the current number
of cases of the three health outcomes, the projected number of cases under the
unmitigated emissions scenario, and the percentage increase (Ebi 2008).

Table 3: Comparison of current diarrheal disease, malnutrition, and malaria cases
with estimated climate change impacts in 2030 assuming the 750 ppm of CO2
scenario (thousands of cases)

Diarrheal diseases Malnutrition Malaria

11

-------
Total

4,513,981

46,352

408,227

Climate change impacts

131,980

4,673

21,787

% increase

3%

10%

5%

Climate change alone, without considering other factors that could increase or decrease
incidence, is projected to increase the burden of diarrheal diseases, malnutrition, and
malaria by several percentage points worldwide. Although there is high uncertainty in
the regional estimates, as would be expected, those regions with high current burdens of
these health outcomes are projected to experience the largest increase in 2030. For
example, unmitigated emissions are projected to more than double the number of incident
cases of diarrheal disease in Africa and parts of Southeast Asia. The largest increase in
malnutrition is projected to occur in the parts of Southeast Asia where malnutrition is
currently severe. The largest increase in incident cases of malaria is projected to occur in
Africa and parts of the Eastern Mediterranean region.

I	E.g. Brownlie, J., Peckham, C., Waage, J., Woolhouse, M., Lyall, C., Meagher, L., Tait, J.,

Baylis, M. and Nicoll, A. (2006): Foresight. Infectious Diseases: preparing for the future.
Future Threats. Office of Science and Innovation, London; Confalonieri U., Menne B.,
Akhtar R., Ebi K.L., Hauengue M., Kovats R.S., Revich B., and A. Woodward (2007):

Human Health. Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution
of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on
Climate Change. Parry M.L., Canziani O.F., Palutikof J.P., van der Linden P.J., Hansson C.E.
(Eds). Cambridge University Press, Cambridge, U.K.; Semenza, J.C., and B. Menne, 2009:
"Climate change and infectious diseases in Europ q," Lancet Infect Dis 9:365-75.

II	Snow, R. W., Guerra, C. A., Noor, A. N., Myint, H. Y. and S. I. Hay (2005): The global

distribution of clinical episodes of Plasmodium falciparum malaria. Nature 434, 214-217.

III	WHO (2004): Malaria epidemics: forecasting, prevention, early warning and control - from

policy to practice. Report of an informal consultation, Leysin, Switzerland, 8-10 December
2003.World Health Organization, Geneva, 52 pp.

IV	Murray, C. and A. Lopez (1996): Global Burden of Disease. Harvard University Press, Boston,

MA,

v Ebi K.L., Hartman, J. McConnell, J. K. Chan, N. and J. Weyant (2005): "Climate suitability for
stable malaria transmission in Zimbabwe under different climate change scenarios." Climatic
Change 73:375-393; Tanser F.C., Sharp, B. and and D. le Sueur (2003): "Potential effect of
climate change on malaria transmission in Africa." Lancet 362:1792-1798; Thomas C.J.,
Davies, G. and C.E. Dunn, (2004): "Mixed picture for changes in stable malaria distribution
with future climate in Africa." Trends Parasitol 20:216-220; Van Lieshout M., Kovats, R. S.,
Livermore, M.T.J, and P. Martens, (2004): "Climate change and malaria: analysis of the
SRES climate and socio-economic scenarios." Global Environmental Change 14:87-99;
Peterson, A.T., 2009: "Shifting suitability for malaria vectors across Africa with warming
climates," BMCInf Dis 9:59, doi: 10.1186/1471-2334-9-59.

12

-------
V1 Van Lieshout, M., Kovats, R. S., Livermore, M., et al. 2004. Climate change and malaria:
Analysis of the SRES climate and socio-economic scenarios. Global Environmental Change
14, 87-99; Tanser, F. C., Sharp, B. & LeSueur, D. 2003. Potential effect of climate change on
malaria transmission in Africa. Lancet 362, 1792-1798.

™ Van Lieshout et al., Op.Cit

vm McMichael,A.J., Woodruff, R. Whetton, P. Hennessy, K. Nicholls, N. Hales, S. Woodward A.
and T. Kjellstrom (2003): Human Health and Climate Change in Oceania: Risk Assessment
2002. Department of Health and Ageing, Canberra, 128 pp.

1X Gonzalez, C., Wang, O., Strutz, S.E., Gonzalez-Salazar, C., Sanchez-Cordero, V., and S. Sarkar,
2010: "Climate change and risk of leishmaniasis in North America: predictions from
ecological niche models of vector and reservoir species," PLOS Neglected Tropical Diseases
4:e585; Gould, E.A., and S. Higgs, 2009: "Impact of climate change and other factors on
emerging arbovirus diseases," Trans R Soc Trop Med Hyg 103:109-121; Holt, A.C., Salkeld,
D.J., Fritz, C.L., Tucker, J.R., and Gong P., 2009: "Spatial analysis of plague in California:
niche modeling predictions of the current distribution and potential response to climate
change," Int J Health Geographies 8:38, doi:10.1186/1476-072X-8-38; Click-Lambert, R.,
Kolivras, K.N., Resler, L.M., Brewster, C.C., and S.L. Paulson, 2008: "The potential for
emergence of Chagas disease in the United States," Geospatial Health 2:227-239; Mangal,
T.D., Paterson, S., and A. Fenton, 2008; "Predicting the impact of long-term temperature
changes on the epidemiology and control of schistosomiasis: a mechanistic model," PLoS
ONE 3: el438. doi: 10.1371/journal.pone.0001438; Ogden, N.H., St-Onge, L, Barker, I.K.,
Brazeau, S., Bigras-Poulin, M, Charron, D.F., Francis, C.M., Heagy, A. Lindsay, L.R.,
Maarouf, A., Michel, P., Milord, F., O'Callaghan, C.J., Trudel, L., and R.A. Thomson, 2008:
"Risk maps for range expansion of the Lyme disease vector, Ixodes scapularis, in Canada
now and with climate change," Int J Health Geographies 7:24, doi: 10.1186/1476-072X-7-24;
Zhou, X.N., Yang, G.J., Yang, K., Wang, X.H., Hong, Q.B., Sun, L.P., Malone, J.B.,
Kristensen, T.K., Bergquist, N.R., and J. Utzinger, 2008: "Potential impact of climate change
on schistosomiasis transmission in China," Am J Trop Med Hyg 78:188-194; Confalonieri et
al., Op Cit.

x E.g. D'Souza. R.M., N.G. Becker, G. Hall, and K.B. Moodie, 2004: "Does ambient temperature
affect foodborne disease?" Epidemiology 15:86-92; Kovats, R.S., S.J. Edwards, S. Hajat,
B.G. Armstrong, K.L. Ebi, and B. Menne, 2004: "The effect of temperature on food
poisoning: a time-series analysis of salmonellosis in ten European countries." Epidemiol
Infect 132, 443-453; Fleury, M., D. Charron, J. Holt, O. Allen, A. Maarouf, 2006: "A time
series analysis of the relationship of ambient temperature and common bacterial enteric
infections in to Canadian provinces,"/^ JBiometeorol 50:385-91.

X1 E.g. Patz, P.A., Vavrus, S.J., Uejio, C.K., and S.L. McLellan, 2008: "Climate change and
vectorborne disease risk in the Great Lakes region of the U.S," Am J Prev Med 35:451-8.

xu WHO, 2002: World Health Report 2002: reducing risks, promoting healthy li fe (Geneva:

World Health Organization).

xm See Mudway, I.S. and F.J. Kelly, 2000: Ozone and the lung: a sensitive issue. Mol. Aspects
Med., 21, 1-48.; Gryparis, A., B. Forsberg, K. Katsouyanni, A. Analitis, G. Touloumi, J.
Schwartz, E. Samoli, S. Medina, H.R. Anderson, E.M. Niciu, H.E.Wichmann, B. Kriz, M.
Kosnik, J. Skorkovsky, J.M.Vonk and Z.Dortbudak, 2004:Acute effects of ozone on mortality
fromthe "Air Pollution and HealthAEuropeanApproach" project. Am. J. Respir. Crit. Care
Med., 170, 1080-1087.; Bell,M.L., F. Dominici and J.M. Samet, 2005:Ameta-analysis of
time-series studies of ozone and mortality with comparison to the national morbidity,
mortality, and air pollution study. Epidemiology, 16, 436-445.; Ito, K, S.F. De Leon and M.

13

-------
Lippmann, 2005: Associations between ozone and daily mortality: analysis and meta-
analysis. Epidemiology, 16, 446-457. Levy, J.I., S.M. Chemerynski and J.A. Sarnat, 2005:
Ozone exposure and mortality: an empiric bayes metaregression analysis. Epidemiology, 16,
458-468.; Bell, M.L., R.D. Peng and F. Dominici, 2006: The exposure-response curve for
ozone and risk of mortality and the adequacy of current ozone regulations. Environ.Health
Persp., 114, 532-536.).

X1V Stevenson, D.S., C.E. Johnson, W.J. Collins, R.G. Derwent, and J.M. Edwards, 2000: "Future
estimates of tropospheric ozone radiative forcing and methane turnover - the impact of
climate change" Geophys Res Lett 27: 2073-2076; Derwent, R.G., W.J. Collins, C.E. Johnson,
and D.S. Stevenson, 2001: "Transient behaviour of tropospheric ozone precursors in a global
3-D CTM and their indirect greenhouse effects," Climatic Change 49: 463-487; Johnson,
C.S., D.S. Stevenson, W. Collins, R. Derwent, 2001: "Role of climate feedback on methane
and ozone studied with a coupled ocean-atmosphere-chemistry model," Geophys Res Lett 28:
1723-1726; Taha, H., 2001: Potential impacts of climate change on troposphereic ozone in
California: a preliminary episodic modeling assessment of the Los Angeles Basin and the
Sacramento Valley (Berkeley, CA: Lawrence Berkeley National Laboratories); Hogrefe, C., J.
Biswas, B. Lynn, K. Civerolo, J.Y. Ku, J. Rosenthal, C. Rosenzweig, R. Goldberg, and P.L.
Kinney, 2004: "Simulating regional-scale ozone climatology over the eastern United States:
model evaluation results," Atmos Environ 38: 2627; Doherty, R.M., Heal, M.R., Wilkinson,
P., Pattenden, S., Vieno, M., Armstrong, B., Atkinson, R., Chalabi, Z., Kovats, S., Milojevic,

A.,	and D.S., Stevenson, 2009: "Current and future climate- and air pollution-mediated
impacts on human health," Environmental Health 8(Suppl):S8, doi:10.1186/1476-069X-8-S1-
S8.

xv Future emissions are, of course, uncertain, and depend on assumptions of population growth,
economic development, and energy use.

XV1 Knowlton, K., J.E. Rosenthal, C. Hogrefe, B. Lynn, S. Gaffin, R. Goldberg, C. Rosenzweig, K.
Civerolo, J.Y. Ku, and P.L. Kinney, 2004: "Assessing ozone-related health impacts under a
changing climate," Environ Health Perspect 112: 1557-1563; Chang, H.H., Zhou, J., and M.
Fuentes, 2010: "Impact of climate change on ambient ozone level and mortality in
Southeastern United States" Int J Environ Res Public Health 7: 2866-2880;
doi: 10.3390/ijerph7072866.

xvu Herran H, Wakhungu J, (Co-chairs). (2008). International Assessment of Agricultural
Knowledge, Science, and Technology for Development. Executive Summary of the
Synthesis Report.

http://www.agassessment.org/docs/IAASTD_EXEC_SUMMARY_JAN_2008.pdf

xvm Battisti DS and Naylor RL. (2009). Historical Warnings of Future Food Insecurity with
Unprecedented Seasonal Heat, Science, 323 (5911), pp. 240 -244.

X1X Confalonieri U., Menne B., Akhtar R., Ebi K.L., Hauengue M., Kovats R.S., Revich B., and A.
Woodward (2007): Human Health. Climate Change 2007: Impacts, Adaptation and
Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change. Parry M.L., Canziani O.F., Palutikof J.P., van
der Linden P.J., Hansson C.E. (Eds). Cambridge University Press, Cambridge, U.K.

xx Ahern, M.J., R.S. Kovats, P. Wilkinson, R. Few, and F. Matthies, 2005: "Global health impacts
of floods: epidemiological evidence," Epidemiol Rev 27:36-45; Hajat S., K. Ebi, S. Kovats,

B.	Menne, S. Edwards, and A. Haines, 2003: "The human health consequences of flooding in
Europe and the implications for public health: A review of the evidence," Applied
Environmental Science and Public Health 1:13-21.

XXI Ibid.

14

-------
xxu Kilbourne, E.M., 1997: "Heat waves and hot environments," in The public health

consequences of disasters, edited by E.K. Noji (New York: Oxford University Press) 245-269.

xxm Kovats, R.S. and C. Koppe, 2005: "Heat waves: past and future impacts, Integration of Public
Health with Adaptation to Climate Change: Lessons Learned and New Directions, edited by
K.L. Ebi, J.B. Smith, and I. Burton (London: Taylor & Francis) pp. 136-160.

XX1V Keatinge, W.R., G.C. Donaldson, R.S. Kovats, and A. McMichael A, 2002: "Heat and cold
related mortality morbidity and climate change," in Health Effects of Climate Change in the
UK (London: Department of Health); Dessai, S., 2003: "Heat stress and mortality in Lisbon
Part II. An assessment of the potential impacts of climate change," Int J Biometeorol 48:37-
44; McMichael, A., R. Woodruff, P. Whetton, K. Hennessy, N. Nicholls, S. Hales, A.
Woodward, and T. Kjellstrom, 2003: Human Health and Climate Change in Oceania: Risk
Assessment 2002 (Canberra: Commonwealth of Australia); Hayhoe, K., D. Cayan, C.B. Field,
P.C. Frumhoff, E.P. Maurer, N.L. Miller, S.C. Moser, S.H. Schneider, K. Nicholas Cahill,
E.E. Cleland, L. Dale, R. Drapek, R.M. Hanemann, L.S. Kalkstein, J. Lenihan, C.K. Lunch,
R.P. Neilson, S.C. Sheridan, and J.H. Verville, 2004: "Emission pathways, climate change,
and impacts on California" PNAS 101:12422-427; Gosling, S.N., McGregor, G.R., and J.A.
Lowe, 2009: "Climate change and heat-related mortality in six cities Part 2: climate model
evaluation and projected impacts from changes in the mean and variability of temperature
with climate change," Int J Biometeorol 53( 1):31-51; Doyon, B., Belanger, D., and P.
Gosselin, 2008: "The potential impact of climate change on annual and seasonal mortality for
three cities in Quebec, Canada," Int J Health Geographies 7:23, doi:10.1186/1476-072X-7-
23; Muthers, S., Matzarakis, A., and E. Koch, 2010: "Climate change and mortality in Vienna
- a human biometeorological analysis based on regional climate modeling," Int J Environ Res
Public Health 7:2965-2977, doi: 10.3390/ijerph7072965.

xxv Menne B; Bertollini R. 2000. The health impacts of desertification and drought. The
Newsletter of the Convention to Combat Desertification,

XXV1 See Kovats, R. S., Bouma, M. J. & Haines, A. El Nino and Health. (World Health

Organization, WHO/SDE/PHE/99.4,Geneva, 1999); Dearborn, D. G., Yike, I., Sorenson, W.
G., Miller, M. J. & Etzel, R. A. Overview of investigation into pulmonary hemorrhage among
infants in Cleveland, Ohio. Environmental Health Perspectives, 495-499 (1999); Mackenzie,
W. R., Hoxie, N. J., Proctor, M. E. & Gradus, M. S. A massive outbreak in Milwaukee of
Cryptosporidium infection transmitted through public water supply. New England Journal of
Medicine 331, 161-167 (1994).

xxvn McMichael A.J., D. Campbell-Lendrum, S. Kovats, S. Edwards, P. Wilkinson, T. Wilson, R.
Nicholls, S. Hales, F. Tanser, D. LeSueur, M. Schlesinger, andN. Andronova, 2004: "Global
Climate Change." In: Comparative Quantification of Health Risks: Global and Regional
Burden of Disease due to Selected Major Risk Factors. Eds. Ezzati M, Lopez A, Rodgers A,
Murray C. World Health Organization, Geneva, pp 1543-1649.

xxvm Hitz S., and J. Smith, 2004: "Estimating global impacts from climate change." Global
Environmental Change 14:201-218.

XX1X Portier ,C.J., Thigpen Tart, K, Carter, S.R., Dilworth, C.H., Grambsch, A.E., Gohlke, J., Hess,
J., Howard, S.N., Luber, G., Lutz, J.T., et al., 2010: A Human Health Perspective On Climate
Change: A Report Outlining the Research Needs on the Human Health Effects of Climate
Change. Research Triangle Park, NC: Environmental Health Perspectives/National Institute
of Environmental Health Sciences, doi: 10.1289/ehp. 1002272 Available:
www.niehs.nih.gov/climatereport.

15

-------
I Climate-Associated Changes

in Health Outcomes

Kristie L. Ebi, Ph.D., MPH
Carnegie Institution for Science

27 January 201 I

-------
IPCC AR4 Health Impacts of Climate
Change

~ Emerg ng evidence of climate change impacts:

~	Altered distribution of some vectors

~	Altered seasonal distribution of some pollen species

~	Increased risk of heatwave deaths

Deaths from climate change

Estimates by WHO sub-region for 2000 (World Health Report. Geneva. WHO, 2002)
NO Udtd	® World Health Organization 2005. All rights reserved.

-------
Direction and Magnitude of CI mate Change
Health Impacts

Negative Impact

Positive Impact

Very High Confidence

Malaria: Contraction and expansion, „
changes in transmission season

High Confidence

increase in malnutrition

Increase in the number of people suffering
from deaths, disease and injuries
from extreme weather events

Increase in the frequency of cardio-respiratory
diseases from changes in air quality

Change in the range of infectious disease vectors
Reduction of cold-related deaths

Medium Confidence

Increase in the burden of diarrheal diseases

*

IPCC 2007

-------
Sum of Years of Life Lost and Years of Lfe
L ived with Disability

77.5
67.5
57.5
47.5
37.5
27 5
17.5

flj

-D 7.5

3

¦H -2 5

n

~1 -12.5
-22.5
-32.5
-42.5
-52.5
-62.5
-72.5
-82.5







I I

[
i

r

J Non-communicable diseases
1 Communicable diseases
Injuries











Annual c
Diarrhe;
million d
Malaria

Jeaths:
il diseases =
eaths
= 1 million



" T—=

— — - - i
•







i





i _____





i







	1 _











1





i





.



p



i





underlying

10.5
dhood deat







rition is an
f 50% ofthi
annual chil



¦¦¦¦¦¦

Malnut

cause o
million

0	100	200	300	400	500	600

Rate per thousand	Pitcher et Hi. 2008

-------
Prevalence Ch ildhood Diarrhea

www.Worldmapper.org

-------
Malar a Cases





4

www.Worldmapper.org

-------
Global Burden of Disease Undernutrition

~	21% disability-adjusted life-years (DALYs) for
children younger than 5 years

~	35% child deaths - 11% of total global Burden of
Disease

~	When all the effects of malnutrition are
considered (including loss of cognitive function,
poor school performance, and loss of future
earning potential), the total estimated costs of
environmental risk factors could be as high as 8-
9% of a typical developing country's GDP in South
Asia or Sub-Saharan Africa

Black et al. 2008

-------
Prevalence of Stunting in Children
Under 5 years (2005)

Black et al. 2008

-------
Prevalence of Stunting Children
Under 5 years in India (2005)

n
~
n
n
~

No data
20-29%
30-39%
40-49%
50-59%
s>6o%

India has more
than 61 million
stunted
children, 51%
of the national
population and
34% of the
global total.
However,
stunting
prevalence
varies

substantially by
state.

Black et al. 2003

-------
Interactions of Infectious Diseases and

Undernutrition

~	Poor nutritional status, especially in infants
and young children, makes infections more
severe and prolonged, and often more
frequent

In low-income countries, 27% of children under the age of 5
are chronically undernourished or stunted, and 23% are
underweight

~	Almost all infections influence a child's
nutritional status through loss of appetite,
changes in intestinal absorption, metabolism,
and excretion of specific nutrients

The effects of infections appear to be directly proportional to
the severity of the infection

-------
Projected Changes n Ozone and Related
Deaths, New York Metro Area

2020s

2080s

Baseline Daily Mortality Rate per 100,000

Percent Increase in 03-related Deaths
2050s

1990

2050s

Kinney et al. 2006

I 11.397- 1.769
I	11.77-2.141

¦	2.142-2.513

¦	2.514-2.885

¦	> 2.835

-------
Climate Change Impacts in 2030 under 750 ppm

CO2 Scenario (thousands of cases)

Estimated costs to treat the climate change-related
cases = $3,992 to $12,603 million



Diarrhea

Malnutrition

Malaria

Total

4,513,981

46,352

408,227

Climate

change

impacts

131,980

4,673

21,787

% increase

3%

10%

5%

Ebi 2008

-------
Vibrio	parahaemolyticusInfections by

Harvest Date and Mean Daily Water
Temperature

10-,

v
E

K

m

0

1

*

o

i	1	1	1	1	1	r

10 13 16 19 22 25 28

1G 19 22 25 2S 31

June

July

McLaughlin et al. 2005

-------
Research Needs

Improve characterization of exposure- response relationships, particularly at regional
and local levels, including identifying thresholds and particularly vulnerable groups

Collect data on the early effects of changing weather patterns on climate-sensitive
health outcomes

Collect and enhance long-term surveillance data on health issues of potential concern,
including vectorborne and zoonotic diseases, air quality, pollen and mold counts,
reporting of food- and water-borne diseases, morbidity due to temperature extremes,
and mental health impacts from extreme weather events

Develop quantitative models of possible health impacts of climate change that can be
used to explore the consequences of a range of socioeconomic and climate scenarios

Understand local- and regional-scale vulnerability and adaptive capacity to
characterize the potential risks and the time horizon over which risks might arise

Develop downscaled climate projections at the local and regional scale in order to
conduct the types of vulnerability and adaptation assessments that will enable
adequate response to climate change, and to determine the potential for interactions
between climate and other risk factors, including societal, environmental, and
economic

Improve understanding of the design, implementation, and monitoring of adaptation
options

Understand the co-benefits of mitigation and adaptation strategies
Enhance risk communication and public health education

-------
Estimating the Economic Value of Health Impacts of Climate Change

Maureen L. Cropper
University of Maryland and Resources for the Future

How should we value the health impacts of climate change? The answer is, in principle, simple:
we should value them by what people are willing to pay to avoid them. This includes the costs
of averting behavior—the costs of the energy used to mitigate the effects of temperature
extremes on health—as well as the cost of the illnesses themselves. Obtaining empirical
estimates of WTP for health—for adults and children—in countries at all income levels is
challenging. The purpose of this presentation is to discuss in more detail what empirical
estimates are needed and how they might be obtained, in the short run, through benefits transfer.

Nature of Health Impacts to Be Valued

The number of deaths and illnesses associated with climate change are likely to be greatest in
developing countries, at least over the rest of this century. Mc Michaels et al. (2004) estimate
that in 2000, climate change was associated worldwide with 166,000 deaths—77,000 due to
malnutrition, 47,000 associated with diarrhea, and 27,000 with malaria (see Figure 1). The
highest number of deaths (per 100,000 persons) was predicted to occur in Africa, South Asia and
the Middle East. It should also be noted that the majority of these deaths are children, and that
deaths among children account for the bulk of the 5.5 million Disability-Adjusted Life Years
(DALYs) that Mc Michael et al. (2004) estimate were lost due to climate change in 2000.

This implies that we must value the lives of children (and adults) in developing countries. The
illnesses that these individuals suffer are also important and must be valued. These include non-
fatal cases of diarrhea and malaria, respiratory illnesses and cardiovascular disease. Adults and
children in higher income countries will also be affected by climate change. The same valuation
concepts should be applied in all cases, as discussed in the next section.

Valuating Mortality

To value risk of death among adults, the appropriate valuation concept is what adults would pay
to reduce their own risk of dying. For children, it is what parents would pay to reduce their
children's risk of dying. Willingness to pay is constrained by ability to pay, and should increase
with income, assuming life extension is a normal good. This implies that WTP will generally be
lower in low-income than in high-income countries. It is often suggested that lives should be
valued equally in all countries—that the same WTP amount should be used regardless of income.
The problem with this suggestion is that it forces people in developing countries to spend more
on risk reduction than they would choose, based on their own preferences. The correct valuation
concept is what a person would pay for a small reduction in his risk of death.

By convention, the sum of WTPs for small risk changes is expressed as the Value per Statistical
Life (VSL)—the sum of WTPs for risk reductions that sum to one statistical life saved. For
example, if each of 10,000 people would pay $100 to reduce their risk of dying over the coming

1

-------
year, the VSL would be $1,000,000 (10,000*$100). The risk reduction (1 in 10,000) summed
over 10,000 people would result in one statistical life saved.

Empirical estimates of the VSL for adults most frequently come from hedonic wage studies,
which estimate compensating wage differentials in the labor market, or from contingent
valuation (stated preference) surveys in which people are asked directly what they would pay for
a reduction in their risk of dying. The empirical literature on the VSL in high income countries is
large.1 There are approximately 4 dozen compensating wage studies in high income countries
(see, for example, Viscusi and Aldy (2003)) and over 4 dozen stated preference studies (Braathen
et al., 2009). Several recent meta-analyses have summarized the results of these studies
(Cropper, Hammitt and Robinson, 2011). The literature in middle income countries is much
smaller.2 Robinson and Hammitt (2009) review 8 wage-risk and 9 stated-preference studies
conducted in 9 middle-income countries. Braathen et al. (2009) citel4 stated preference studies
conducted in middle-income countries, but only one in a low income country (Bangladesh).

VSL Benefits Transfer

What is clear is that the developing country literature at this point is not sufficiently mature to
provide estimates of the VSL for individual countries. This suggests transferring estimates from
countries where better studies exist to countries for which there are no empirical estimates of the
VSL. Most transfers are based on income differences between countries. The most common
approach to benefits transfer assumes that the ratio of the VSL to per capita income is constant
among countries. (This is equivalent to assuming an income elasticity of the VSL =1.)
Transferring values from the US, where this ratio is approximately 140 to 1, implies that the ratio
of the VSL to income is 140 to 1 for all countries.

Recent analyses, however, suggest that an income elasticity of 1 may be inappropriate for low-
income countries. This based partly on a comparison of the ratio of the VSL to income in high
income countries with the corresponding ratio based on studies in middle income countries.
Preliminary analyses (Cropper and Sahin, 2009) suggest the ratio is closer to 80 to 1 for middle
income countries v. 140 to 1 for high income countries. This suggests that the income elasticity
of the VSL is > 1. Analyses of the income elasticity of the VSL in the US (Costa and Kahn,
2004; Kniesner et al. 2011) and Taiwan (Hammitt, Liu and Liu, 2000) also suggest that the
income elasticity of the VSL is larger at low incomes than at high incomes. Pending additional
studies, Hammitt and Robinson (2010) suggest using an income elasticity of the VSL of 1.5 in
addition to an income elasticity of 1.0 to provide a range of values of for the VSL in middle and
low income countries.

Cropper, Hammitt and Robinson (2011) summarize this literature, including recent meta-analyses.

2

I follow the World Bank's definition, based on market exchange rates. The groups are: low income, $995 or
less; lower middle income, $996 - $3,945; upper middle income, $3,946 - $12,195; and high income, $12,196 or
more.

2

-------
Estimating the VSL for Children

There is a small but growing literature on parents' WTP to reduce health risks to their children,
including mortality risks. In the US and Europe, revealed preference studies have used
information on the purchase of car seats and bicycle helmets to infer WTP for reduced death and
injury. Other studies are based on parents' WTP for chelation therapy for children with body
lead burdens. Some of the literature relies on stated preference studies. As stated in a recent
OECD volume on children's health (OECD, 2010) only 15 studies directly compare parents'
willingness to pay for improvements in their own health with WTP for improvements in their
children's health. Many of these studies value reductions in acute illness, and only one study
was conducted outside of the US and Europe (Liu et al., 2000).

The consensus from studies conducted in high income countries is that parents are willing to pay
more to reduce health risks to young children than to themselves—generally about twice as
much—but that this effect decreases with child age. The result is also not universal: Jenkins et
al. (2001) and Mount et al. (2001) find that parents are willing to pay more to reduce mortality
risks to themselves than to their children. The USEPA uses the same VSL for children as for
adults.

The question is whether the VSL is higher for children than for adults in low income countries;
in particular, in countries with substantial under-5 child mortality and high fertility rates. There
no studies of which I know that directly address this issue. In studies conducted for the World
Bank (Larsen, 2011), VSLs used for children are often less than those used for adults. This is a
topic clearly requiring more research. The literature on the allocation of food and health care
resources within the household may shed some light on this issue.

Valuing Morbidity

Willingness to pay for avoided illness should capture the value of the pain and suffering avoided,
as well as the value of time lost due to illness (both leisure and work time) and the costs of
medical treatment. If some of these costs are not borne by the individual, and are therefore not
reflected in his willingness to pay, the value of the avoided costs must be added to WTP to
measure the total benefits of reduced illness. The Cost of Illness (COI) approach, which captures
medical costs and lost productivity, is often used as a lower bound to the more comprehensive
valuation concept.

In high income countries, WTP estimates for avoided morbidity are available for some illnesses,
but COI estimates are often used to measure the value of avoided illness. Due to the
heterogeneous nature of illness, providing WTP (or even COI) estimates for a variety of diseases
is a huge task. The most sensible approach would be to determine the diseases that are likely to
lead to the largest number of DALYs lost due to illness and to focus on obtaining COI estimates
for these diseases.

Morbidity, especially in Sub-Saharan Africa, is likely to have impacts on the economy beyond
traditional illness costs. Child morbidityis likely to affect human capital formation. (See for
example, Alderman, Hoddinott and Kinsey(2006) on the impacts of malnutrition on human

3

-------
capital formation.) Malaria may have impacts on economic growth through land use, crop
choice and other mechanisms (Gallup and Sachs, 2001; Tol, 2008). These effects are certainly
worth exploring.

Mortality per
million population

_2"4

¦ 70 120
No data

* ?

: Change in climate compared to baseline 1961-1990 climate

figure 1: Estimated Deaths due to Climate Change in 2000, by WHO Sub-Region

Source: Map created by Center for Sustainability and the Global Environment (SAGE), University of
Wisconsin using data from McMichael et al. (2004).

References

Alderman, H., Hoddinott, J.and B. Kinsey.2006. "Long-Term Consequences of Early Childhood
Malnutrition ''OxfordEconomic Papers 58(3):450-474.

Braathen, N.A., H.Lindhjem and S. Navrud. 2009. Valuing Lives Saved from Environmental
Transport, and Health Policies: A Meta-Analysis of Stated Preference Studies. Prepared for the
Organization for Economic Co-operation and Development.

Costa, Dora L., and Matthew E. Kahn. 2004. "Changes in the Value of Life, 1940-1980." The
Journal of Risk and Uncertainty 29(2): 159-80.

Cropper, M.L. and S. Sahin. 2009."Valuing Mortality and Morbidity in the Context of
Disaster Risks." World Bank Policy Research Working Paper 4832, February 2009.

Cropper, M. L., J.K. Hammittand L. A. Robinson. 2011. "Valuing Mortality Risk Reductions:
Progress and Challenges," forthcoming in the Annual Review of Resource Economics.

4

-------
Gallup, J.L. and J.D Sachs. 2001. "The Economic Burden of Malaria." American Journal of
Tropical Medical and Hygiene. 64(l):85-96.

Hammitt, J.K., J-T.Liu, and J-L. Liu.2000. "Survival is a Luxury Good: Thelncreasing Value of
a Statistical Life." Prepared for the NBER Summerlnstitute Workshop on Public Policy and the
Environment.

Hammitt, J.K. and L.A. Robinson.2011. "The Income Elasticity of the Value per Statistical Life:
Transferring Estimates between High and Low Income Populations." Journal of Benefit-Cost
A nalysis. 2(1), Art. 1.

Jenkins, R.R., N. Owens, and L.B. Wiggins. 2001. "Valuing Reduced Risks to Children:
The Case of Bicycle Safety Helmets."Contemporary Economic Policy. 19(4):397-408.

Kniesner, T.J., W.K. Viscusi, C. Woock, and J.P. Ziliak. 2011. "The Value of Statistical Life:
Evidence from Panel Data." Vanderbilt University Law School, Law and Economics,

Working Paper Number 11-02.

Larsen, Bjorn. 2011. Personal Communication, January 23.

Liu, J.-T., J.K. Hammitt, J.-D.Wang, and J.-L. Liu. 2000. "Mother's Willingness to Pay
for Her Own and Her Child's Health: A Contingent Valuation Study in Taiwan." Health
Economics. 9:319-326.

McMichael A.J., D. Campbell-Lendrum, S. Kovats, S. Edwards, P. Wilkinson, T. Wilson, et al.
2004. "Global climate change." In Comparative Quantification of Health Risks: Global and
Regional Burden of Disease Due to SelectedMajor Risk Factors, ed. M.J. Ezzati, et al. 1543-
1649. Geneva: World Health Organization.

Mount, T., W. Weng, W. Schulze, and L. Chestnut. 2001. "Automobile Safety and the Value of
Statistical Life for Children, Adults, and the Elderly: Results from New Data On Automobile
Usage," Paper presented at the Association of Environmental and Resource Economists 2001
Summer Workshop, "Assessing and Managing Environmental and Public Health Risks." Bar
Harbor, Maine, June 13-15, 2001.

Robinson, L.A. and J.K. Hammitt. 2009. The Value of Reducing Air Pollution

Risks in Sub-Saharan Africa. Prepared for the World Bank under sub contract to ICF International.

Tol, Richard. 2008. "Climate, Development and Malaria: An Application of FUND." Climatic
Change. 88(l):21-34.

Viscusi, W. Kip, and Joseph E. Aldy. 2003. "The Value of Statistical Life: A Critical Review of
Market Estimates Throughout the World." Journal of Risk and Uncertainty 27(l):5-76.

5

-------
Estimating the Economic
Value of Health Impacts
of Climate Change

Maureen L. Cropper

University of Maryland
Resources for the Future

January 27th, 2011

-------
The Task

¦	Given estimates of health impacts of climate change by
region and time period, monetize value of health
damages

¦	Should value damages after adaptation, plus costs of
adaptation; presentation will focus on valuing health
impacts per se

¦	Value changes in mortality risks

~	For children and adults

~	As a function of per capita income

¦	Value changes in morbidity

-------
Presentation

~	Main health impacts to be valued and countries
in which they are likely to occur

~	Valuation concepts

~	Estimating the value of mortality risk reductions
for adults in low income countries

~	Estimating the value of mortality risk reductions
for children in low income countries

~	Valuing morbidity

-------
Which Health Effects to Value?

¦	Possible health endpoints include:

~	Malnutrition

~	Diarrheal disease

~	Vector-borne diseases (malaria, dengue fever)

~	Deaths associated with temperature extremes, air pollution

~	Deaths associated with climate-related disasters

¦	According to McMichael et al. (2004) most DALYs lost
due to:

~	Malnutrition

~	Diarrhea

~	Vector-borne disease

-------
Estimated Deaths due to Climate
Change* in 2000, by WHO subregion

Source: Map created by SAGE using data from McMichael et al. (2004

-------
Overview of Approaches to Valuing
Death and Injury

~	Human Capital - Cost of Illness (COI)

~	Values a life by the PDV of forgone earnings

~	Values an injury by medical costs and lost
productivity

~	Value of Statistical Life - Willingness to Pay

~	Values a life by sum of what people will pay for
reductions in risk of death

~	For injuries, adds WTP for pain and discomfort to
COI

~	VSL - WTP approach is theoretically correct

-------
Valuing Reductions in Risk of Death

~	Goal is to estimate what an individual is willing and able
to pay for a	small reduction in his risk of death

•	It does NOT measure the amount an individual
would pay to avoid death with certainty

~	Suppose a person is willing to pay $500 to reduce his
risk of dying by 1 in 10,000 over the coming year:

•	If 10,000 people will each pay $500 for a 1 in 10,000
risk reduction, together they will pay $5,000,000 for
risk reductions that sum to 1 statistical life saved

•	We say that $5,000,000 is the Value of a Statistical
Life.

-------
Approaches to Valuing Mortality
Risk Reductions

~	Revealed Preference Studies

•	Use compensating wage (CW) differentials to value
risk of death (most common approach)

•	Use data on purchase of safer vehicles or safety
equipment (e.g., bicycle helmets)

~	Stated Preference Studies

•	Ask people directly what they would pay for a change
in risk of death (e.g., Contingent valuation (CV)
studies)

-------
Overview of VSL Estimates in the
Literature

High-income OECD countries

•	Approximately 4 dozen CW studies (30 in USA)

•	Over 4 dozen CV studies

•	6 published meta-analyses of these studies since 2000

Middle-income countries

•	Fewer than a dozen CW studies

•	About two dozen stated preference studies

Low-income countries

•	1 study for Bangladesh; none for Africa

-------
How Is VSL Transferred from One
Country to Another?

¦	Most common approach is:

VSI-india = VSLUSA*(Yjndja/ Yusa)£

where f is the income elasticity of the VSL. Usual
assumption is that £ = 1.

¦	This implies:

VSLusa/ Yusa ~ VSLlndja/Ylndja

10

-------
Is the Conventional Approach
Correct?

¦	In High Income Countries VSL/Y ratio ~ 140

~	Ratio of VSL/Y is about 140 in Miller (2000) based
studies in 13 high income countries

¦	In Middle Income Countries VSL/Y ratio ~ 80

~	Review of 17 VSL studies in middle income
developing countries by Robinson and Hammitt
(2009) implies a ratio of 80 (using better studies)

¦	This suggests that £ > 1.

¦	US labor market studies suggest that £
increases as incomes fall

-------
How to Estimate the VSL for
Developing Countries?

¦	Hammitt and Robinson (2010) suggest using an
income elasticity of 1.5

~	Supported by studies by Costa and Kahn (2004) and
Hammit, Liu and Liu (2000)

¦	Cropper and Sahin (2009) also suggest £ = 1.5
based on a life-cycle consumption model

¦	Using a US VSL of $6.3 million (2007 USD) and
YUS = $46,000 implies:

~	VSLlndia = (Ylndia)A1.5 * (.64)

-------
How to Estimate the VSL for
Children?

¦	Studies of parents' willingness to pay to reduce
risks to children used to estimate the VSL

¦	Studies in high income countries suggest child
VSL «2x adult VSL

¦	However....

~	Parents' WTP may be different in countries where 1
out of 5 children die before age 5

~	USEPA uses same value for adults and children

~	Many World Bank studies have used Human Capital
approach for children

-------
Valuing Morbidity

Want to capture:

•	Value of lost productivity

•	Cost of medical treatment

•	Value of discomfort, inconvenience and pain

Cost of Illness (COI) = Value of lost work time +
Cost of medical treatment

Could add value of Quality-Adjusted Life Years
(QALYs) lost to COI to capture pain and
suffering since few direct WTP estimate
available

-------
Valuing Morbidity

¦	In US studies of health effects of air pollution,
value of avoided morbidity is small relative to
premature mortality

~	Case of chronic bronchitis « .05 VSL

¦	Back-of-the-envelope calculations should be
done before refining estimates

¦	Other impacts that may be relevant are:

~	Macroeconomic impacts of malaria (Gallup and
Sachs, 2001; Tol, 2008)

~	Impacts of malnutrition on human capital formation
(Alderman, Hoddinott and Kinsey, 2003)

-------
Conclusions

¦	Greatest disease burden from climate change likely to be
in Sub-Saharan Africa, South Asia and the Middle East

¦	Much of the disease burden will fall on children

¦	Value associated with health impacts depends crucially
on:

~	How value of morality risks varies with income

~	How risks to children are valued v. risks to adults

¦	Most of the disease burden likely to come from mortality

~	But, link between diseases and economic growth could be
important

16

-------
Improving the Assessment and Valuation of Climate Change Impacts for Policy and Regulatory Analysis: Research on
Climate Change Impacts and Associated Economic Damages
January 27-28 2011, Washington, DC

Sea Level Impacts of Climate Change

Robert J. Nicholls

School of Civil Engineering and the Environment, and the Tyndall Centre for Climate Change Research

University of Southampton
Southampton UK SO 17 IB J, UK

Email: r.i.nicholls@soton.ac.uk

INTRODUCTION

Sea-level rise has been seen as a major threat to low-lying coastal areas around the globe since the issue of
human-induced global wanning emerged in the 1980s. What is often less appreciated is that more than 200
million people are already vulnerable to flooding by extreme sea levels around the globe. This population
could grow fourfold to the 2080s just due to rising population/coastward migration. These people generally
depend on natural and/or artificial flood defences and drainage to manage the risks, with the most
developed and extensive artificial systems in Europe (especially around the southern North Sea) and East
Asia. Most threatened are the significant populations (at least 20 million people today) already living below
normal high tides in many countries such as the Netherlands and the USA. Hurricane Katrina's impacts on
New Orleans in 2005 remind us of what happens if such defences fail. Increasing mean sea level and more
intense storms will exacerbate these risks. Despite these threats, the actual consequences of sea-level rise
remain uncertain and contested. This reflects far more than the uncertainty in the magnitude of sea-level
rise and climate change, with the uncertainties about our ability to adapt to these challenges being a major
uncertainty (Nicholls and Tol, 2006; Nicholls et al., 2007a).

CLIMATE CHANGE AND GLOBAL/RELATIVE SEA-LEVEL RISE

Human-induced climate change is expected to cause a profound series of changes including rising sea level,
higher sea-surface temperatures, and changing storm, wave and run-off characteristics. Although higher sea
level only directly impact coastal areas, these are the most densely-populated and economically active land
areas on Earth, and they also support important and productive ecosystems that are sensitive to sea level
and other change. Rising global sea level due to thermal expansion and the melting of land-based ice is
already being observed and this rise is likely to accelerate through the 21st century. From 1990 to the last
decade of the 21st century, a total rise in the range 18-59 cm has been forecast by the Intergovernmental
Panel on Climate Change (IPCC) Fourth Assessment Report (AR4) (Meehl et al, 2007). It is worth noting
that the current satellite observations of global sea-level rise are at the high end of the predicted SRES
scenarios (Rahmstorf et al., 2007), and if recent ice sheet discharge continues through the 21st Century at
current rates, the maximum projected rise increases to 79 cm1. Even this scenario excludes uncertainties
due to collapse of the large ice sheets, and as noted in the IPCC Synthesis Report (2007), the quantitative
AR4 scenarios do not provide an upper bound on sea-level rise during the 21st Century. A global rise of sea

1 Allowing for ice-melt uncertainties

1

-------
Improving the Assessment and Valuation of Climate Change Impacts for Policy and Regulatory Analysis: Research on
Climate Change Impacts and Associated Economic Damages
January 27-28 2011, Washington, DC

level exceeding one metre remains a low probability, but physically-plausible scenario for the 21st Century
due to large uncertainties concerning ice sheet dynamics and their response to global warming. While these
high end scenarios may be relatively unlikely, their large potential impacts makes them highly significant in
terms of climate risk. There is also increasing concern about higher extreme sea levels due to more intense
storms superimposed on these mean rises, especially for areas affected by tropical storms. This would
exacerbate the impacts of global-mean sea-level rise, particularly the risk of more damaging floods and
storms.

When analysing sea-level rise impacts and responses, it is (Nicholls, 2010) fundamental that impacts are a
product of relative (or local) sea-level rise rather than global changes alone. Relative sea-level change takes
into account the sum of global, regional and local components of sea-level change: the underlying drivers
of these components are (1) climate change such as melting of land-based ice, thermal expansion of ocean
waters, and changing ocean dynamics, and (2) non-climate uplift/subsidence processes such as tectonics,
glacial isostatic adjustment, and natural and human-induced subsidence. Hence relative sea-level rise is
only partly a response to climate change and varies from place to place. Relative sea level is presently
falling due to ongoing glacial isostatic adjustment (rebound) in some high-latitude locations that were
formerly sites of large (kilometre-thick) glaciers, such as the northern Baltic and Hudson Bay, while RSLR
is more rapid than global-mean trends on subsiding coasts, including many populous deltas. Most
dramatically, human-induced subsidence of susceptible areas due to drainage and withdrawal of
groundwater can produce dramatic RSLR, especially cities built on deltaic deposits. Over the 20th century,
parts of Tokyo and Osaka subsided up to 5 m and 3 m, respectively, a large part of Shanghai subsided up to
3 m, and most of Bangkok subsided up to 2 m2. Such human-induced subsidence can be managed by
stopping shallow sub-surface fluid withdrawals, but natural "background" rates of subsidence will continue.
The four example cities have all seen a combination of such policies combined with the provision of flood
defences and pumped drainage to avoid submergence and/or frequent flooding. In contrast, Jakarta and
Metro Manila are subsiding cities where little systematic action to manage and reduce the subsidence are in
place as yet.

SEA-LEVEL RISE AND RESULTING IMPACTS

Relative sea-level rise has a wide range of effects on the natural system, with the five main effects being
summarized in Table 1. Flooding/submergence, ecosystem change and erosion have received significantly
more attention than salinisation and rising water tables. Along with rising sea level, there are changes to all
the processes that operate around the coast. The immediate effect is submergence and increased flooding of
coastal land, as well as saltwater intrusion into surface waters. Longer term effects also occur as the coast
adjusts to the new environmental conditions, including wetland loss and change in response to higher water
tables and increasing salinity, erosion of beaches and soft cliffs, and saltwater intrusion into groundwater.

2 The maximum subsidence is reported as data on average subsidence is not available.

2

-------
Improving the Assessment and Valuation of Climate Change Impacts for Policy and Regulatory Analysis: Research on
Climate Change Impacts and Associated Economic Damages
January 27-28 2011, Washington, DC

These lagged changes interact with the immediate effects of sea-level rise and generally exacerbate them.
For instance, coastal erosion will tend to degrade or remove natural protective features (e.g. saltmarshes,
mangroves and sand dunes) so increasing the impact of extreme water levels and hence the risk of coastal
flooding.

A mean rise in sea level also raises extreme water levels, as shown by Zhang et al (2000) on the US East
Coast, and this is widely applied in impact studies for future conditions. Changes in storm characteristics
could also influence extreme water levels both positively and negatively. For example, the widely debated
increase in the intensity of tropical cyclones would increase in general terms extreme water levels in the
areas affected.

Changes in natural systems as a result of sea-level rise have many important direct socio-economic impacts
on a range of sectors with the effect being overwhelmingly negative. For instance, flooding can damage the
extensive coastal infrastructure, ports and industry, the built environment, and agricultural areas, and in the
worst case lead to significant mortality (e.g., Cyclone Nargis (2008), Myanamar). Erosion can lead to losses
of the built environment and related infrastructure and have adverse consequences for sectors such as
tourism and recreation. In addition to these direct impacts, there are indirect impacts such as adverse effects
on human health: for example, mental health problems increase after a flood, or the release of toxins from
eroded land fills and waste sites which are commonly located in low-lying coastal areas, especially around
major cities. Economically, sea-level rise will also have direct and indirect effects (see Tol, 2011, these
abstracts). Thus, sea-level rise has the potential to trigger a cascade of direct and indirect human impacts.

RECENT IMPACTS OF SEA-LEVEL RISE

Over the 20th century global sea level rose about 18 cm. While this change may seem small, it will have
had many significant effects, most particularly in terms of the return periods of extreme water levels (e.g.,
Zhang et al., 2000; Menendez and Woodworth, 2010), and promoting a widespread erosive tendency for
coasts. However, linking sea-level rise quantitatively to impacts is quite difficult as the coastal zone has
been subjected to multiple drivers of change over the 20th Century (Rosenzweig et al., 2007). Good data on
rising sea levels has only been measured in a few locations, and defences and other coastal infrastructure
have often been upgraded substantially through the 20th Century, especially in those (wealthy) places where
there are sea-level measurements. Most of this defence upgrade reflects expanding populations and wealth
in the coastal flood plain and changing attitudes to risk, and relative sea-level rise may not have even been
considered in the design. Equally, erosion can be promoted by processes other than sea-level rise (Table 1),
and human reduction in sediment supply to the coast must contribute to the observed changes. Decline in
intertidal habitats such as saltmarshes, mudflats and mangroves is often linked to sea-level rise, but these
systems are also subject to multiple drivers of change, including direct destruction (Nicholls, 2004). Hence,

3

-------
Improving the Assessment and Valuation of Climate Change Impacts for Policy and Regulatory Analysis: Research on
Climate Change Impacts and Associated Economic Damages
January 27-28 2011, Washington, DC

while global sea-level rise is a pervasive process, it is difficult to unambiguously link it to impacts, except
in some special cases - most recent coastal change was a response to multiple drivers of change.

On the US east coast, relative sea levels have risen at variable rates between 2 and 4 mm/yr over the 20th
century, reflecting a combination of global rise and subsidence. Both sea level and coastal change has been
measured during the 20th century, providing a laboratory for exploring shoreline response to sea-level rise.
Comparing the rate of shoreline retreat and the long-term rate of relative sea-level rise away from inlets and
engineered shores, supports the concept of the 'BruunRule' where the shoreline retreat rate is 50 to 100
times the rate of sea-level rise (Zhang et al., 2004), although this relationship remains controversial. Near
inlets, the indirect effects of sea-level rise which cause the associated estuary/lagoon to trap beach-sized
sediment can have much larger erosional effects on the neighbouring open coasts than predicted by the
Bruun Rule (Stive, 2004). Hence, more general relationships are required to understand coastal change
taking account of sea-level change, sediment supply and coastal physiography. Human responses to sea-
level rise are even more difficult to document. Human abandonment of low-lying islands in Chesapeake
Bay, USA during the late 19th/early 20th century does seem to have been triggered by a small acceleration
of sea-level rise and the resulting land loss (Gibbons and Nicholls, 2006).

There have certainly been impacts from the relative sea-level rise resulting from large rates of subsidence,
such as the Mississippi delta where relative sea-level rise is 5 to 10 mm/yr. Between 1978 and 2000, 1565
km2 of intertidal coastal marshes and adjacent lands were converted to open water, due to sediment
starvation and increases in the salinity and water levels of coastal marshes due to human development and
wider changes (Barras et al., 2003). By 2050, about 1300 km2 of additional coastal land loss is projected if
current global, regional and local processes continue at the same rate. There have also been significant
impacts of relative sea-level rise in deltas and in and around subsiding coastal cities, in terms of increased
waterlogging, flooding and submergence, and the resulting need for management responses (Nicholls et al.,
2007b). The flooding in New Orleans during Katrina in 2005 was significantly exacerbated by subsidence
compared to earlier flood events such as Hurricane Betsy in 1965 (Grossi and Muir Wood, 2006). In terms
of response, all the major developed areas that were impacted by relative sea-level rise have been defended,
even when the change in relative sea-level rise was several metres over several decades. In New Orleans,
the pre-existing dike system before Katrina have been rebuilt and substantially upgraded at a cost of $15
billion over 6 years. In less developed areas, coastal retreat has occurred such as south of Bangkok where
subsidence has led to a shoreline retreat of more than a kilometre.

Hence observations through the 20th Century reinforce the importance of understanding the impacts of sea-
level rise in the context of multiple drivers of change - this will remain true under more rapid rises in sea
level. Of these multiple drivers of change, human-induced subsidence is of particular interest, but this
remains relatively unstudied in a systematic sense. Observations also emphasize the ability to protect

4

-------
Improving the Assessment and Valuation of Climate Change Impacts for Policy and Regulatory Analysis: Research on
Climate Change Impacts and Associated Economic Damages
January 27-28 2011, Washington, DC

against RSLR, especially for more densely-populated areas such as the subsiding Asian cities already
discussed, or around the southern North Sea, including London and Hamburg.

FUTURE IMPACTS OF SEA-LEVEL RISE

The future impacts of sea-level rise will depend on a range of factors, including the degree to which sea-
level rise accelerates, the level and manner of coastal development and the success (or failure) of adaptation
(Nicholls, 2010). Assessments of the future impacts of sea-level rise have taken place on a range of scales
from local to global. They all confirm potentially large impacts following Table 1, although comprehensive
studies are limited and most available assessments only consider a subset of possible impacts. Taking
account of population exposure, sensitivity and adaptive capacity, South and South-East Asia and Africa
appear to be most vulnerable in absolute terms due to storm-induced flooding combined with sea-level rise.
Small island regions in the Pacific, Indian Ocean and Caribbean stand out as being especially vulnerable to
flooding (Mimura et al., 2007), even though relatively few people are affected in global terms. The
populations of low-lying islands such as the Maldives or Tuvalu face the real prospect of increased
flooding, submergence and forced abandonment: this perception may trigger a collapse in investment and
general confidence blighting these areas and triggering abandonment long before it is physically inevitable
(Barnet and Adger, 2003). An important lobby group for small islands and sea-level rise is the Alliance of
Small Island States (AOSIS), which contains 37 UN votes.

However, adaptation can greatly reduce the impacts. Benefit-cost models that compare protection with
retreat generally suggest that it is worth investing in widespread protection as populated coastal areas are
often of high economic value (Fankhauser, 1995; Tol, 2007; Sugiyama et al., 2008). (It is worth noting that
if no economic growth is assumed, protection is much harder to justify and hence the impacts of sea-level
rise depend on both climate and socio-economic scenarios (Nicholls, 2004; Anthoff et al., 2010)). With or
without protection, small island and deltaic areas stand out as relatively more vulnerable in most analyses
and the impacts fall disproportionately on poorer countries (Anthoff et al., 2010; Sugiyama et al., 2008).

Regional and global scale assessments

Compared to national assessments, regional and global assessments provide a more consistent basis to
assess the broad-scale impacts of sea-level rise.

Coastal Flooding

Globally, it was estimated that about 200 million people lived in the coastal flood plain (below the 1 in
1,000 year surge-flood elevation) in 1990, or about 4% of the world's population (Nicholls et al., 1999).
Based on estimates of defence standards, on average 10 million people/year experienced coastal flooding in
1990. These numbers will change due to the competing influences of relative sea-level rise (due to local
subsidence and global changes), changes in coastal population and improving defence standards as people

5

-------
Improving the Assessment and Valuation of Climate Change Impacts for Policy and Regulatory Analysis: Research on
Climate Change Impacts and Associated Economic Damages
January 27-28 2011, Washington, DC

become more wealthy (Nicholls et al., 1999). Relative sea-level rise is assumed to displace extreme water
levels upwards (assuming constant storm characteristics). The analysis is designed to explore the impacts of
global-mean sea-level rise if it is largely ignored. Therefore, the increasing protection standards only
consider existing climate variability (i.e. surges in 1990) and the analysis is considering a world that is
completely ignoring the issue of global-mean (and relative) sea-level rise. (This follows recent behaviour
globally). Outputs include:

people in the hazard zone (PHZ) - the population living below the 1 in 1,000 year flood plain (or
the exposed population);

people at risk (PAR) - the average number of people who experience flooding per year (a measure
of risk that takes account of flood protection);

Table 2 illustrates the impacts of no global-mean sea-level rise and the IS92a global-mean sea-level rise
scenarios on flooding (for a global-mean rise in the range 19 to 80 cm from 1990 to the 2080s). Generic
results include:

Even without sea-level rise, the number of people flooded each year first increases significantly
due to increasing coastal populations (i.e., exposure), and then diminishes as increasing protection
standards due to rising GDP/capita become the most important factor.

Significant impacts of sea-level rise are not apparent until the 2050s or later so sea-level rise is a
slow onset hazard.

The uncertainty about impacts is large with relatively minor impacts for the low rise scenario in
the 2080s, a 10-fold increase in PAR under the mid rise scenario and a 27-fold increase in PAR under the
high rise scenario for the 2080s.

Looking at 20 world regions, they all see an increase in the incidence of flooding compared to the baseline,
most especially under the higher sea-level rise scenarios. The most vulnerable regions in relative terms are
the small island regions of the Caribbean, Indian Ocean and Pacific Ocean. However, absolute increases in
the incidence of flooding are largest in the southern Mediterranean (largely due to the Nile delta), West
Africa, East Africa, South Asia and South-East Asia - these five regions contain about 90% of the people
flooded in all cases for the 2080s. This reflects the large populations of low-lying deltas in parts of Asia,
and projections of rapid population growth around Africa's coastal areas. While developed country regions
have relatively low impacts, sea-level rise still produces a significant increase in the number of people who
would be flooded assuming no adaptation for sea-level rise. These results show that sea-level rise could
have a profound impact on the incidence of flooding - the higher the total rise, the greater the increase in
flood risk, all other factors being equal. Any increase in storminess would further exacerbate the predicted
increase in coastal flooding.

6

-------
Improving the Assessment and Valuation of Climate Change Impacts for Policy and Regulatory Analysis: Research on
Climate Change Impacts and Associated Economic Damages
January 27-28 2011, Washington, DC

Using the DIVA model, we can examine flood impacts with and without dike upgrade (Nicholls, 2010). No
upgrade leads to results that are qualitatively similar to those in Table 2. The behaviour assuming dike
upgrade is quite different and independent of the magnitude of the sea-level rise scenario, the number of
people flooded is projected to decline through the 21st Century. This reflects that the dikes are raised more
than the magnitude of sea-level rise as people adapt to sea-level rise and become more risk adverse as they
become more wealthy. This illustrates that the success or failure of adaptation is fundamental to
understanding impacts as discussed later.

Environmental Refugees

Sea-level rise is often associated with a large potential for environmental refugees forcibly displaced from
their homes (Myers, 2002). Potentially, many tens or even hundreds of millions of peoples could be so
displaced, especially given that coastal populations are growing significantly worldwide. However, if we
can successful adapt to these challenges, this is a much smaller problem than is often assumed. Adaptation
could include flood defences for urban areas, and land use planning for new developments to avoid the
more risky areas. As a reference, Tol (2002a; 2002b) suggests that most coastal areas are worth protecting
in a benefit-cost sense (protection costs are less than damage costs). This formulation suggests that <75,000
people/year will be displaced by a 1-m sea-level through the 21st Century, after allowing for protection:
incrementally this is of order 1% of the potentially displaced population. This result has a large
uncertainty, but it illustrates again that the success or failure of adaptation is a key element to understanding
the scale of the problem.

Global Costs of Sea-Level Rise

Global estimates of the incremental costs of upgrading defence infrastructure3 suggest the costs are much
lower than the expected damage (Tol, 2007). IPCC CZMS (1990) estimated the costs of defending against a
1-m sea-level rise at $500 billion. Hoozemans et al (1993) doubled these costs to $1000 billion. Looking at
the total costs of sea-level rise including dryland and wetland loss and incremental defence investment,
Fankhauser (1995) estimated annual global costs of $47 billion using the IPCC CZMS (1990) data. Tol
(2002a; 2002b) made similar estimates using the Hoozemans et al. (1993) data, supplemented by other data
sources and adding the costs of forced migration. The protection was optimised and it was estimated that
the annual costs of sea-level rise are only $13 billion/year for a 1-m global rise in sea level: much lower
than estimated by Fankhauser, and much lower than widely assumed. However, any failure in protection
will lead to much higher costs. Sugiyama et al (2008) noted that that the spatial distribution of
infrastructure and wealth along the coast influences costs: the more wealth is concentrated the smaller the
protection costs.

3 These incremental costs should not be compared directly with projects such as the post-Katrina defence of
New Orleans as they only reflect the sea-level component of these needs.

7

-------
Improving the Assessment and Valuation of Climate Change Impacts for Policy and Regulatory Analysis: Research on
Climate Change Impacts and Associated Economic Damages
January 27-28 2011, Washington, DC

RESPONDING TO SEA-LEVEL RISE

The two potential responses to sea-level rise are mitigation and adaptation: only the latter is considered
here. Adaptation to sea-level rise involves responding to both mean sea-level rise and extreme sea-level rise
(Hallegatte, 2009). Planned adaptation options to sea-level rise are usually presented as one of three generic
approaches (Klein et al., 2001) with examples in Table 1:

•	(Planned) Retreat - all natural system effects are allowed to occur and human impacts are minimised
by pulling back from the coast via land use planning, development control, etc.;

•	Accommodation - all natural system effects are allowed to occur and human impacts are minimised by
adjusting human use of the coastal zone via flood resilience, warning systems, insurance, etc.;

•	Protection - natural system effects are controlled by soft or hard engineering (e.g., nourished beaches
and dunes or seawalls), reducing human impacts in the zone that would be impacted without
protection.

Given the large and rapidly growing concentration of people and activity in the coastal zone, autonomous
(or spontaneous) adaptation processes alone will not be able to cope with sea-level rise. Further, adaptation
in the coastal context is widely seen as a public responsibility. Therefore, all levels of government have a
key role in developing and facilitating appropriate adaptation measures. The required adaptation costs
remain uncertain, but as large amounts are already invested in managing coastal floods, erosion and other
coastal hazards, the incremental costs of including global sea-level rise does not appear infeasible at a
global scale over the coming decades (World Bank, 2010). However, in certain settings such as small
islands, these costs could overwhelm local economies (Fankhauser and Tol, 2005; Nicholls and Tol, 2006).
Another key issues are the adaptation deficit (Parry et al., 2009), and observed behaviour does not agree
with the implicit model in many of the benefit-cost analysis. Lastly, maintenance can substantially raise
costs compared to just capital costs, and this needs to be considered and poorly maintained flood defences
are worst than no defences as the engender a false sense of security.

DISCUSSION/CONCLUDING REMARKS

The abstract illustrates that understanding the impacts of sea-level rise crosses many disciplines and
embraces natural, social, and engineering sciences, and major gaps in that understanding remain. The
success or failure of adaptation in general, and protection in particular, is an important issue which deserves
more attention and has lead to what Nicholls and Tol (2006) termed the 'optimistic' and 'pessimistic' view
of the importance of sea-level rise. The pessimists tend to focus on high rises in sea level, extreme events
like Katrina, and view our ability to adapt as being rather limited resulting in alarming impacts, including
widespread human displacement from coastal areas. The optimists tend to focus on lower rises in sea level
and stress a high ability to protect and high benefit-cost ratios in developed areas and wonder what all the
fuss is about.

8

-------
Improving the Assessment and Valuation of Climate Change Impacts for Policy and Regulatory Analysis: Research on
Climate Change Impacts and Associated Economic Damages
January 27-28 2011, Washington, DC

The optimists have empirical evidence to support their views that sea-level rise is not a big problem in
terms of the subsiding megacities that are also thriving. Importantly, these analyses suggest that improved
protection under rising sea levels is more likely and rational than is widely assumed. Hence the common
assumption of a widespread retreat from the shore is not inevitable, and coastal societies will have more
choice in their response to rising sea level than is often assumed. However, the pessimists also have
evidence to support their view. First, the socio-economic scenarios used in climate impact assessments
assume substantial future economic growth and its more equitable distribution: lower growth and greater
concentration of wealth in parts of the world may mean less damage in monetary terms, but it will also lead
to a lower ability to protect in those poorer areas. Secondly, the benefit-cost approach implies a proactive
attitude to protection with extensive management in place for the hazards of climate variability. However,
experience suggests a widespread adaptation deficit in many parts of world and shows that most protection
has been built as a reaction to actual or near disaster. The cost of addressing the adaptation deficit will often
be significant in itself, although this has not been quantitatively assessed. If combined with high rates of
sea-level rise this will probably lead to more frequent coastal disasters, even if the ultimate response is
better protection. Thirdly, disasters (or adaptation failures) such as Hurricane 'Katrina' could trigger coastal
abandonment, a process that has not been analysed to date. This could have a profound influence on
society's future choices concerning coastal protection as the pattern of coastal occupancy might change
radically. A cycle of decline in some coastal areas is not inconceivable, especially in future world scenarios
where capital is highly mobile and collective action is weaker. As the issue of sea-level rise is so widely
known, disinvestment from coastal areas may be triggered even without disasters actually occurring: for
example, the economies of small islands may be highly vulnerable if investors become cautious (Barnett
and Adger, 2003). Lastly, retreat and accommodation have long lead times - benefits are greatest if
implementation occurs soon - but this is not happening widely as yet. For these reasons, adaptation may
not be as successful as some assume, especially if rises in sea level are at the higher end of the range of
predictions.

Thus the optimists and the pessimists both have arguments in their favour. Sea-level rise is clearly a threat,
which demands a response. Scientists need to better understand this threat, including the implications of
different mixtures of adaptation and mitigation, as well as the need to engage with the coastal and climate
policy process so that these scientific perspectives are heard. Importantly, it has been recognised that a
combination of mitigation (to reduce the risks of a large rise in sea level) and adaptation (to the inevitable
rise) appears to be the most appropriate course of action, as these two policies are more effective when
combined than when followed independently, and together they address both immediate and longer term
concerns (Nicholls et al., 2007a).

References

Anthoff, D., Nicholls, R.J., & Tol, R.S.J. (2010) The economic impact of substantial sea-level rise.

Mitigation and Adaptation Strategies for Global Change, 15, 321-335

9

-------
Improving the Assessment and Valuation of Climate Change Impacts for Policy and Regulatory Analysis: Research on
Climate Change Impacts and Associated Economic Damages
January 27-28 2011, Washington, DC

Barnett, J. & Adger, W. N. (2003) Climate dangers and atoll countries. Climatic Change, 61, 321-337.

Barras, J., Beville, S., Britsch, D., Hartley, S., Hawes, S., Johnston, J., Kemp, P., Kinler, Q., Martucci, A.,
Porthouse, J., Reed, D., Roy, K., Sapkota, S. & Suhayda, J. (2003) Historical and projected coastal
Louisiana land changes: 1978-2050, U.S. Geological Survey.Open File Report 03-334 39

Dasgupta S, Laplante B, Meisner C., Wheeler, D. & J. Yan (2009) The impact of sea level rise on
developing countries: a comparative analysis. Climatic Change 93, 379-388.

Fankhauser, S. (1995) Protection versus retreat: estimating the costs of sea-level rise. Environment and
Planning A, 27, 299-319.

Fankhauser, S. & Tol, R. S. J. (2005) On climate change and economic growth. Resource and Energy
Economics, 27, 1-17.

Grossi, P. & Muir-Wood, R. (2006) Flood risk in New Orleans: Implications for future management and
insurability. London, UK, Risk Management Solutions (RMS).

Hallegatte, S. (2009) Strategies to adapt to an uncertain climate change. Global Environmental Change, 19,
240-247.

Hoozemans, F. M. J., Marchand, M. & Pennekamp, H. A. (1993) A Global Vulnerability Analysis:

Vulnerability Assessment for population, coastal wetlands and rice production on a global scale. 2nd
edition. Delft, Delft Hydraulics.

IPCC (2007) Summary for Policymakers. IN Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M.,
Averyt, K. B., Tignor, M. & Miller, H. L. (Eds.) Climate Change 2007: The Physical Science Basis.
Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on
Climate Change. Cambridge, UK and New York, USA, Cambridge University Press.

IPCC CZMS (1990) Strategies for adaptation to sea level rise. Report of the Coastal Zone Management
Subgroup, Response Strategies Working Group of the Intergovernmental Panel on Climate Change.
The Hague, Netherlands, Ministry of Transport, Public Works and Water Management.

Klein, R. J. T., Nicholls, R. J., Ragoonaden, S., Capobianco, M., Aston, J. & Buckley, E. N. (2001)

Technological options for adaptation to climate change in coastal zones. Journal of Coastal Research,
17, 531-543.

Meehl, G. A., Stacker, T. F., Collins, W., Friedlingstein, P., Gaye, A., Gregory, J., Kitoh, A., Knutti, R.,
Murphy, J., Noda, A., Raper, S., Watterson, I., Weaver, A. & Zhao, Z.-C. (2007) Global climate
projections IN Solomon, S., Qin, D. & Manning, M. (Eds.) Climate Change 2007: The Physical
Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change. Cambridge, UK, Cambridge University Press.

Menendez, M. & P. L. Woodworth (2010). Changes in extreme high water levels based on a quasi-global
tide-gauge dataset, Journal of Geophysical Research, 115, C10011.

Mimura, N., Nurse, L., McLean, R., Agard, J., Briguglio, L., Lefale, P., Payet, R., Sem, G. (2007) Small
islands. In M. Parry, O. Canziani, J. Palutikof, P. van der Linden and C. Hanson (Eds.) Climate Change
2007:Impacts, Adaptation and Vulnerability. Cambridge University Press, Cambridge, UK., pp. 687-
716.

Myers, N. (2002) Environmental Refugees: A growing phenomenon of the 21st Century. Philosophical
Transactions of the Royal Society, B, 357, 609-613.

Nicholls, R. J. (2002) Analysis of global impacts of sea-level rise: A case study of flooding. Physics and
Chemistry of the Earth, 27, 1455-1466.

Nicholls, R. J. (2004) Coastal flooding and wetland loss in the 21st century: changes under the SRES
climate and socio-economic scenarios. Global Environmental Change - Human and Policy
Dimensions, 14, 69-86.

Nicholls, R. J. (2010). Impacts of and Responses to Sea-Level Rise. In: Understanding sea-level rise and
variability. (Church, J. A., Woodworth, P. L. Aarup T. & Wilson, S. (eds)). Wiley-Blackwell, 2010,
pp. 17-51.

Nicholls, R. J., Wong, P. P., Burkett, V. R., Codignotto, J. O., Hay, J. E., McLean, R. F., Ragoonaden, S. &
Woodroffe, C. D. (2007a) Coastal systems and low-lying areas. IN Parry, M. L., Canziani, O. F.,
Palutikof, J. P., Van Der Linden, P. & Hanson, C. E. (Eds.) Climate Change 2007: Impacts, Adaptation
and Vulnerability. Cambridge, UK, Cambridge University Press

Nicholls, R. J., Hanson, S., Herweijer, C., Patmore, N., Hallegatte, S., Corfee-Morlot, J., Chateau, J. &
Muir-Wood, R. (2007b) Ranking port cities with high exposure and vulnerability to climate

10

-------
Improving the Assessment and Valuation of Climate Change Impacts for Policy and Regulatory Analysis: Research on
Climate Change Impacts and Associated Economic Damages
January 27-28 2011, Washington, DC

extremes—exposure estimates. Environmental Working Paper No. 1. Paris Organisation for Economic
Co-operation and Development (OECD).

Nicholls, R. J. & Tol, R. S. J. (2006) Impacts and responses to sea-level rise: a global analysis of the SRES
scenarios over the twenty-first century. Philosophical Transactions of the Royal Society A
Mathematical Physical and Engineering Sciences, 364, 1073-1095..

Parry M.L., Arnell N.W., Berry P.M., DodmanD, Fankhauser S, Hope C, Kovats S, Nicholls R. J.,

Satterthwaite D, Tiffin R, Wheeler T, (2009). Assessing the costs of adaptation to climate change: A
review of the UNFCCC and other recent estimates. International Institute for Environment and
Development and Grantham Institute for Climate Change, London. 111pp.

Rahmstorf, S., Cazenave, A., Church, J. A., Hansen, J. E., Keeling, R. F., Parker, D. E. & Somerville, R. C.

J. (2007) Recent climate observations compared to projections. Science, 316, 709.

Rosenzweig, C., Casassa, G., Karoly, D. J., Imeson, A., Liu, C., Menzel, A., Rawlins, S., Root, T., Seguin,
B. & Tryjanowski, P. (2007) Assessment of observed changes and responses in natural and managed
systems. In: Parry, M., Canziani, O. F., Palutikof, J. P., Van Der Linden, P. & Hanson, C. E. (Eds.)
Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the
Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, UK,
Cambridge University Press.

Stive, M. J. F. (2004) How important is global warming for coastal erosion? An editorial comment.

Climatic Change, 64, 27-39.

Sugiyama, M., Nicholls, R. J. & Vafeidis, A. (2008) Estimating the economic cost of sea-level rise. Report

No. 156. Boston, USA, MIT Joint Program on the Science and policy of Global Change.

Tol, R. S. J. (2002a) Estimates of the damage costs of climate change. Part I: Benchmark estimates.

Environmental and Resource Economics, 21, 47-73.

Tol, R. S. J. (2002b) Estimates of the damage costs of climate change. Part II: Dynamic estimates.

Environmental and Resource Economics, 21, 135-160.

Tol, R. S. J. (2007) The double trade-off between adaptation and mitigation for sea level rise: An
application of FUND. Mitigation and Adaptation Strategies for Global Change, 12, 741-753.

World Bank (2010) The economics of adaptation to climate change. A synthesis report, final consultation
draft. Washington DC: World Bank

http://siteresources.worldbank.org/EXTCC/Resources/EACC_FinalSynthesisReport0803_2010.pdf
Zhang, K. Q., Douglas, B. C. & Leatherman, S. P. (2000) Twentieth-century storm activity along the US

east coast. Journal of Climate, 13, 1748-1761.

Zhang, K. Q., Douglas, B. C. & Leatherman, S. P. (2004) Global warming and coastal erosion. Climatic
Change, 64, 41-58

11

-------
Improving the Assessment and Valuation of Climate Change Impacts for Policy and Regulatory Analysis: Research on
Climate Change Impacts and Associated Economic Damages
January 27-28 2011, Washington, DC

Table 1. The main natural system effects of relative sea-level rise, including climate and non-climate
interacting factors and examples of adaptation to these effects. Some interacting factors (e.g., sediment
supply) appear twice as they can be influenced both by climate and non-climate factors. Adaptations are
coded: P - Protection; A - Accommodation; R - Retreat, (adapted from Nicholls and Tol, 2006; Nicholls
2010).			

NATURAL SYSTEM
EFFECT

INTERACTING FACTORS

ADAPTATION
RESPONSES

CLIMATE

NON-CLIMATE

1.

Inundation,
flood and
storm
damage

a. Surge
(flooding
from the
sea)

Wave/storm

climate,

Erosion,

Sediment supply.

Sediment supply,
Flood management,
Erosion,

Land reclamation

Dikes/surge
barriers [P],
Building

codes/floodwise
buildings [A],
Land use
planning/hazard
delineation [A/R],

b.

Backwater
effect
(flooding
from rivers)

Run-off.

Catchment
management and
land use.

2. Wetland loss (and
change)

C02 fertilisation of
biomass
production,
Sediment supply,
Migration space

Sediment supply,
Migration space,
Land reclamation
(i.e., direct
destruction).

Land use planning

[A/R],

Managed

realignment/ forbid
hard defences [R],
Nourishment/sedim
ent management
[PI.

3. Erosion (of 'soft'
morphology)

Sediment supply,

Wave/storm

climate.

Sediment supply.

Coast defences [P],
Nourishment [P],
Building setbacks
[R],

4.

Saltwater
Intrusion

a. Surface
Waters

Run-off.

Catchment
management
(overextreaction),
Land use.

Saltwater intrusion
barriers [P],
Change water
abstraction [A/R],

b. Ground-
water

Rainfall.

Land use,
Aquifer use
(overpumping).

Freshwater
injection [P],
Change water
abstraction [A/Rl.

5. Rising water tables/
impeded drainage

Rainfall,
Run-off.

Land use,
Aquifer use,
Catchment
management.

Upgrade drainage
systems [P],
Polders [P],
Change land use
[A],

Land use
planning/hazard
delineation [A/R],

12

-------
Improving the Assessment and Valuation of Climate Change Impacts for Policy and Regulatory Analysis: Research on
Climate Change Impacts and Associated Economic Damages
January 27-28 2011, Washington, DC

Table 2. Sea-level rise and coastal flooding of people under the IS92a sea-level rise scenarios for low, mid
and high climate sensitivities - see text for definitions of People in the Hazard Zone (PHZ) and People at
Risk (PAR). The population scenario assumes that population change within the coastal flood plain is
double national trends. Defences are upgraded with rising GDP/capita, but do not address sea-level rise
(adapted from Nicholls, 2002).	

Time
(years)

Sea-Level
Scenario

People in the
Hazard Zone
(PHZ)

People at Risk
(PAR)

1990

N/A

197

10

2020s

No Rise

399

22

Low

403

23

Mid

411

24

High

423

30

2050s

No Rise

511

27

Low

525

28

Mid

550

64

High

581

176

2080s

No Rise

575

13

Low

605

17

Mid

647

133

High

702

353

13

-------
^	. UNIVERSITY OF

Southampton

TyndairCentre

School of Civil Engineering

and the Environment

for Climate Change Research

Sea-Level Impacts of
Climate Change

Prof. Robert J. Nicholls

School of Civil Engineering and the Environment and
the Tyndall Centre for Climate Change Research

University of Southampton

Southampton S017 1BJ

United Kingdom

r.i.nicholls@soton.ac.uk

Improving the Assessment and Valuation of Climate Change Impacts for Policy and Regulatory Analysis:
Research on Climate Change Impacts and Associated Economic Damages
January 27-28 2011, Washington, DC

-------
Plan

•	Introduction

•	What is sea-level rise';

•	Impacts of sea-level rise

•	Responses to sea-level rise

•	Concluding thoughts

-------
Coasts and People

Population and economic density in the coastal zone is greater than

other areas of the earth's surface.

0	20	40	60	80	100

Elevation (m)

Source: Nicholls and Small, 1993, Journal of Coastal Research

-------
Current Exposure by Elevation

based on today's conditions n 84 developing countries

15 -r

•Land area (%)

as 10

QJ

1—

3

(/)
o

Q.

X

5

0

Population (%)

GDP (USD) (%)

Urban area (%)

0

Height (m)

rel. to mean sea level

¦Agricultural area (%)

Source: Dagsputa et al (2007) World Bank Report (2009) Climatic Change

-------
What is Sea-Level Rise?

-------
Climate-induced Sea-Level Rise

Rising temperatures lead to:

•	Thermal expansion of seawater;

•	Melting of land-based ice

-	Small glaciers (e.g., Rockies, Alaska)
-Greenland ce sheet

-	West Antarctic ice sheet	¦

-------
Global Sea-Level Rise

(Source: IPCC, 2007, AR4 WG1) 9 /

• /

/

/

/

/

/

1800 1850 1900	1950 2000 2050 2100

Year

-------
Subsiding Coastal Megacities

(maximum subsidence during the 20th Century)

Tokyo (5 m)
Shanghai (3 m)

Manila (1

2 o

ahtfkok

Madras Jakarta (<$5 m)
Calcutta ? - ' - °

Istanbul

Tianiin (2 m)

Dhaka ? Seou/

Osaka

(3

Los Angeles

Lima

Buenos Aires

Lagos

Karachi
Rio de Janeiro

Source: Nicholls (1995) GeoJournal

-------
What Are The Impacts
of Sea-Level Rise?

-------
Physical Impacts of Sea-Level Rise

NATURAL SYSTEM EFFECT	INTERACTING FACTORS

CLIMATE	NON-CLIMATE

1. Inundation,
flood and
storm damage

a. Surge

(flooding from the
sea)

Wave/storm climate,
Erosion,

Sediment supply.

Sediment supply,
Flood management,
Erosion,

Land reclamation

b. Backwater
effect (flooding
from rivers)

Run-off.

Catchment management and land
use.

2. Wetland loss (and change)

C02 fertilisation of
biomass production,
Sediment supply,
Migration space

Sediment supply,

Migration space,

Land reclamation (i.e., direct
destruction).

3. Erosion (of 'soft' morphology)

Sediment supply,
Wave/storm climate.

Sediment supply.

4. Saltwater
Intrusion

a. Surface
Waters

Run-off.

Catchment management (over-

extraction),

Land use.

b. Ground-water

Rainfall.

Land use,

Aquifer use (over-pumping).

5. Higher water tables/ impeded
drainage

Rainfall,
Run-off.

Land use,

Aquifer use,

Catchment management.

Source: Nicholls (2010) Book on "Understanding Sea-Level Rise and Variability"

-------
Socio-Economic Impacts of SLR

Coastal Socio-
economic
Sector

Freshwater
Resources
Agriculture and

Fisheries and
Aquaculture
Health
Recreation and
tourism
Biodiversity

Settlements/
infrastructure

Sea-level rise physical impact

Inundation,
etc.

Wetland loss

Erosion

Saltwater
intrusion

Higher water

tables/

etc.











v





v

v

X

X



X

X

X

X

X

X

-

X

X

-

X

X

X

X

X

-

-

X

X

X

X

X

X

-

X

X

X

X = strong; x= weak; - = negligible or not established.

Source: Nicholls (2010) Book on "Understanding Sea-Level Rise and Variability''

-------
loods: December Northeaster 1992

New York City - FDR Drive

-------
Submergence Due to Subsidence

Bangkok Area

Legend

	Raattf

(a) 1981

(b) 2002

» -*»







Source: Phien-Wej et al (2006) Engineering Geology

-------
" hreatened Coastal Areas

to 40-cm of SLR by the 2080s

Caribbean

Indian Ocean
Small Islands

Pacific Ocean
Small Islands

People at Risk
(millions per region)

¦	(A) > 50 million

¦	(B) 10 -50 million
(C) < 10 million

Vulnerable island region
Regional boundary

Source: Nicholls et al. (1999); see also Nicholls (2004) Global Environmental Change

-------
Exposed Population 2005

Top 20 Cities - based on 100 year flood plain

Alexandria

Kolkata.

Mumbai

A BangkoJ

Abidjan

Exposed population - Scenario C
(000s)

•	< 1000

#	1000-2000
^ 2000-3000

5,000 10,000 Kilometers
	I

Source: Nicholls et al., 2008, OECD Report

-------
Exposed Assets 2005

Top 20 Cities - based on 100 year flood plain

Amsterdam

Vancouver

New Yc

Mumbai

Bangkok

Exposed assets - Scenario C
US$ million

o <10,000
O 10,000 - 200,000
A >200,000

10,000 Kilometers

Source: Nicholls et al., 2008, OECD Report

-------
What Can We Do About
Sea-Level Rise?

Mitigation - source control
Adaptation - change behaviour

-------
Mitigation Scenarios

Hadley Coupled Ocean-Atmosphere Model 2

Time (years)

Source: Nicholls and Lowe (2004) Global Environmental Change

-------
Planned Adaptation to SLR

Accommodate

	

/V

I

'SOft	^^^1

Source: Nicholls (2010) Book on "Understanding Sea-Level Rise and Variability"

-------
Many Adaptation Options are Available

P - Protection; A - Accommodation; R - Retreat.

NATURAL SYSTEM EFFECT

POSSIBLE ADAPTATION RESPONSES

1. Inundation,
flood and storm
damage

a. Surge

Dikes/surge barriers [P],

Building codes/floodwise buildings [A],

Land use planning/hazard delineation [A/R].

b. Backwater
effect

2. Wetland loss (and change)

Land use planning [A/R],

Managed realignment/ forbid hard defences [R],

Nourishment/sediment management [P].

3. Erosion (of 'soft' morphology)

Coast defences [P],
Nourishment [P],
Building setbacks [R].

4. Saltwater
Intrusion

a. Surface Waters

Saltwater intrusion barriers [P],
Change water abstraction [A/R],

b. Ground-water

Freshwater injection [P],

Change water abstraction [A/R],

5. Rising water tables/ impeded
drainage

Upgrade drainage systems [P],

Polders [P],

Change land use [A],

Land use planning/hazard delineation [A/R],

Source: Nicholls (2010) Book on "Understanding Sea-Level Rise and Variability"

-------
Fraction of Coast Protected

Sensitivity Analysis on Protection Costs

FUND analysis (for the ATLANTIS Project)

a
o

*+->
o

<4H

0.9

0.8

0.7

0.6

0.5

0.4

0.3







2000 2025 2050 2075 2100 2125 2150 2175 2200 2225

Source: Nicholls et al (2008) Climatic Change

-------
Optimists vs. Pessimists

Optimists

Pessimists

Possible small rise in sea level (< 0.5 m by

Possible large rise in sea level (> 1 m by

2100)

2100)

High benefit-cost ratios

Extreme events and disasters

Adaptation will work

Adaptation will fail or is unaffordable



Thriving subsiding megacities

Optimistic socio-economic scenarios



Observed protection tends to be reactive



rather than proactive - the adaptation



deficit



Disasters could trigger coastal



abandonment, undermining the case for



protection

Retreat and accommodation have long lead
times and need to start now

-------
Concluding Remarks (1)

•	Climate-induced sea-level rise is inevitable -
the uncertainty is its magnitude.

•	This will be compounded by subsidence in
many densely-populated coastal areas.

•	Risks are already rising, and this will continue.

•	The worst-case (do nothing) impacts are
dramatic.

•	There are widely differing views concerning
the success or failure of adaptation.

-------
Concluding Remarks (2)

•	Mitigation of climate and subsidence is
needed to make the problem more
manageable.

•	To adapt to dynamic coastal risks, proactive
assessment is required including:

-	defining the relevant drivers,

-	the potential impacts,

-	the potential adaptation responses,

-	selection of sustainable adaptation pathways.

-------
^	. UNIVERSITY OF

Southampton

TyndairCentre

School of Civil Engineering

and the Environment

for Climate Change Research

Sea-Level Impacts of
Climate Change

Prof. Robert J. Nicholls

School of Civil Engineering and the Environment and
the Tyndall Centre for Climate Change Research

University of Southampton

Southampton S017 1BJ

United Kingdom

r.i.nicholls@soton.ac.uk

Improving the Assessment and Valuation of Climate Change Impacts for Policy and Regulatory Analysis:
Research on Climate Change Impacts and Associated Economic Damages
January 27-28 2011, Washington, DC

-------
rill ECONOMIC IMPACT OF SEA LEVEL RISE

Richard S.J. Tol

Economic and Social Research Institute, Dublin, Ireland

Institute for Environmental Studies, Vrije Universiteit, Amsterdam, The Netherlands
Department of Spatial Economics, Vrije Universiteit, Amsterdam, The Netherlands
Department of Economics, Trinity College, Dublin, Ireland

11 February 2011

Sea level rise has a range of impacts on the coast, including permanent inundation, increased
flood risk, wetland loss, and saltwater intrusion. Enhanced protection of the coast would alleviate
some of these impacts (e.g., flood risk), but may ameliorate others (e.g., wetland loss).

The bulk of the literature on the economic impact of sea level rise has used the so-called direct
cost method to estimate the total welfare loss. This method is conceptually straightforward. One
starts with estimates of the physical impacts, estimates the price, multiplies the two, and adds the
results across impacts, space and time.

While conceptually straightforward, there are practical difficulties. The price of permanent
inundation, for instance, is the average value of land. Although beach front property is
considerably more expensive than property further inland, sea level rise would shift the coastline.
Beach front property would get lost, but adjacent property would become beach front and thus
appreciate in value. The appropriate value of land is therefore the average value of land. But
where would one get an estimate of the average value? Some countries have a well-developed
market for land and a robust administration that collects and reports such data. Most countries,
however, lack either or both.

Figure 1 shows one attempt to fill the data gap. It assumes that land value is a function of income
density ($/yr/ha) - the product of per capita income ($/p/yr) and population density (p/ha). The
income density elasticity of land value is estimated using data for the states of the USA. The US
average land value is used as the basis for extrapolation to the rest of the world. Figure 1
contrasts this estimate to two other, equally crude attempts which agree on the broad picture but
disagree on the details.

-------
There are different issues with the cost of coastal protection. Dikes, seawalls, groins, etc are
often built in the same way around the world, and often by the same small group of multinational
companies. While estimates are available for the cost of raising a kilometer of dike by one meter,
say, the analysis is complicated by the fact that different places would opt for different types of
coastal protection.

Wetlands impose yet another challenge. There is a market price for land and for coastal
protection. There is no market for wetlands. One therefore has to rely on non-market valuation
techniques. Brander et al. conduct a meta-analysis of wetland values. Figure 2 reproduces some
of their results, which confirm expectations. Wetlands are more valuable in places where there
are many people and where there are rich people; larger wetlands are less valuable, per hectare,
than smaller wetlands. At the same time, Figure 2 reveals a large range of wetland values. This is
partly because wetlands are very heterogeneous, and partly because non-market valuation is
difficult and prone to measurement error.

One cannot study the impacts of sea level rise (or any other aspect of climate change for that
matter) without adaptation. Some forms of adaptation are trivial. Sunbathers are unlikely to
return to a beach, or the beach that their grandfather used to frequent, if it would be washed
away. There is no risk that sea level rise would drown them. Coastal protection, on the other
hand, is typically regarded as a collective or public good.

One could take one of two approaches to model and protect coastal protection. One could study
the type and design standard of coastal protection as it is. This is hampered by poor data.
Attempts to gather data on the design standard of dikes and seawalls have led nowhere, even for
data-rich and well-organized places like the European Union. Instead, one could study the
frequency of floods. Data are available - cf. Figure 3 - but while multiple regression analysis
reveals certain patterns - richer, more egalitarian, more authoritarian countries are less
vulnerable to natural disasters - a substantial part of the variance cannot be explained.

The second approach to modeling coastal protection is to consider optimal adaptation. This
approach circumvents the problem of collecting data on how people adapt, but it creates a
counterfactual set of data on how people should adapt. There are a few studies that compare
actual and optimal coastal protection. These studies suggest that decisions about coastal
protection are typically not based on a cost-benefit analysis. Nonetheless, optimal adaptation is
the method most prevalent in the literature.

Figure 4 shows some results for direct cost estimates. Figure 4 displays the fifty most vulnerable
countries in 2100 - that is, the countries with the highest total cost relatively to their gross
domestic product. While sea level rise would cost more than 0.5% of GDP in a handful of
countries, the relative cost is much smaller than that in the vast majority of countries. The main
reason for this result is that the absolute cost of coastal protection is stable over time, and
therefore falls relative to the value of land and the size of the economy. As a result, a greater

-------
share of the coastline is protected and the relative costs of sea level rise fall. Exceptionally
vulnerable are countries with a coast that is long relative to the hinterland - that is, small islands
- and poor countries in river deltas.

Direct costs are conceptually straightforward albeit uncertain in practice. Direct costs, however,
are only an approximation of the true impact of sea level rise on welfare. Particularly, a loss of
land would reduce production in agriculture, which would drive up food prices and leave less
money for other consumption. Coastal protection would increase the demand for construction
and for investment funds. In order to fully appreciate the economic implications of sea level rise,
one would need to use a computable general equilibrium model.

Figure 5 shows the results of such an analysis. In the scenario, it is assumed that there is no
additional coastal protection. The analysis is done for assumptions that may reflect the economy
of 2050, and sea level is assumed to rise by 25 cm. Two shocks are considered. First, only land is
lost. Second, both land and the capital on that land are lost. In the first shock, people anticipate
sea level rise and fully depreciate their houses, factories, roads etc before they are washed away
by the waves. In the second shock, there is no anticipation of sea level rise. Economic activity
falls if productive assets are lost to the sea. Developed economies respond little to a reduction in
the availability of land but more strongly to a loss of capital; less developed economies respond
in the opposite way. This reflects the relative land- and capital-intensity of production.

Figure 6 shows results from the same model and analysis, now assuming that all vulnerable and
inhabited coasts are fully protected. Two mechanisms explain the pattern in Figure 6. First,
coastal protection stimulates the economic activity through an increased demand for
construction. (This also illustrates that GDP is a poor indicator for welfare.) Second, the increase
in the demand for investment and hence savings suppresses consumption. Therefore, the impact
of coastal protection is net positive in regions that have a lot of coast to protect (Australia,
Canada, Russia) and it is net negative in regions that finance a lot of international investment
(European Union, Japan - the model is calibrated to data from the mid-1990s).

Figure 7 compares direct cost estimates to the true welfare impact (or rather, the Hicksian
equivalent variation), considering a scenario without additional coastal protection. Figure 7
reveals that, globally, the direct cost estimate underestimates the true welfare loss, but only by
15% or so. Direct costs are necessarily lower than welfare, because a loss of a productive asset
deflates the entire economy and raises production costs everywhere. The direct costs only
include the direct implications. Figure 7 further shows that the regional pattern of impacts is
different. In some regions, the true welfare loss may be lower than the direct cost estimate. In
this case, that is because relatively land-abundant regions (Brazil, Ukraine) can take advantage of
land loss elsewhere and increase their agricultural production and export.

Although sea level rise is one of the better understood impacts of climate change, the above
review suggests that current impact estimates leave much to be desired. There is a paucity of

-------
high-quality data. Partly, this is because not much of an effort has been made (e.g., land values).
Partly, this is because good data is expensive to collect (e.g., wetland values). Partly, this is
because most of the impact will take place in the future and studies necessarily rely on
extrapolation. Two big uncertainties are the value of wetlands and the nature and intensity of
coastal protection. Two unquantified unknowns the impact of saltwater intrusion and the effect
of change in the frequency, direction, and intensity of storms.

-------
100,000,000 n

10,000,000 -

1,000,000 -

~ This study
Fankhauser
¦ Darwin

100,000 -

10,000

Figure 1. Three alternative estimates of the national average value of agricultural land.

-------
Js I nti
¦

i

m tiiiiJi

£ «-
I ~

* -

liiii >

llii'i

LS11

r ¦ isr«4|*
if - L^uti



-t—?

¦^XL_
1.- ;

k

« a

V*"

G »

*

>•¦ +

Wttfjmi srei i ts; kf mk >

l!> I UP

itjii.:
mmc

iBI
attlfl
IM
Ifl

*

¦T

"f

¦ t

4. -	¦ •

I :A-

i'tMtta

t

» i

¦T t ¦:

»• ; i

* 11* ¦

—«¦——¦«
¦^flr

>*»;	3l.fl-

liDP JMf CdflU I 199! USI'

n

4>>n

lllrl

jj	nit; *

1

s

J»	"!I£IZ h

%,	Riea

J	iaa

13
Oil

* * » . i

• f" :* ~

« . 1 ¦. t» ~

4

mi

P:f*ilirii:
-------
100

10

10

0.1

*!#••

••

„ *



• • ^ •—

• iQ00^ *. toooo • iooooo

• ••• * ,*
—*—•	* •	

• • <

-g 0.01

#•

0.001

Income ($/p/yr)

Figure 3. The number of people affected by floods as a function of per capita income.

-------
Land loss
Wetland loss
Protection cost





l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l l

a =j v) vj

IJ

O si V)

2	u =J

a	=j L.

V)	cr =J

V)	— O

a	o

w	n

c	o

a v) w d
w u n-q

.E >— >
zj ~n ^ ^

sS

a £ a

V)

«Z °

A- ><0
_c

CJ

0,n* w Q_~a _*£_C3C3C5V)C3C3j_C3=3

c ^ u fi. l. a c - cn w~a t; o - -t-
°3j*§^|£§§"^.§§.3

_J	V) ^ r o ~ Oh 12 - Jl- ° C

g-s-oww®!*6^ s

r- t-	/ ) I>

aowdOiOdvid-aind

a =
n £ '-

w o e-

: d C L d



Q-

a

Q_

oi°o

-i2 E
en w
.En-^

U H	O O ^ „

l£>
=3

>

,E=-.<£ v) 5
*¦-> 1-1 "a

ES
°2

o
V)

y) CI/)

o v)~a o v) c
0~a C (JlW'r

x a-5 S a §
o-2 (U-g °00

S^htn
^ -"S

w

a .1=
~o -E

d^

~a

>

i/)

Figure 4. The fifty countries most vulnerable to sea level rise in 2100, and the composition of the
annual cost.

-------
0.05

0.00

-0.05

u
£_
0)
O-

-0.10

-0.15

-0.20

¦	Land / Capital loss

¦	GDP (land)

~	GDP (L+C)

¦	CO2 (land)

~	COZ (L+C)

Figure 5. The impact of sea level rise (without additional coastal protection) in 2050 on GDP and
C02 emissions.

-------
¦	Protection costs
~ GDP

¦	C02

L

EEx

CHIND

RoW

Figure 6. The impact of additional coastal protection to cope with sea level rise in 2050 on GDP
and CO2 emissions.

-------
World

Rest of the World
Southeast Asia
East Asia
Australiaand New Zealand

Japan
EU
Canada
USA

0 2000 4000 6000
Million dollar per year

Figure 7. The direct costs and total welfare impacts of sea level rise in 2050.

-------
Sea Level Rise' Economic Impact

Eimear Leahy, Sean Lyons, Richard S.J. Tol

Economic and Social Research Institute, Dublin

Trinity College, Dublin
Vrije Universiteit, Amsterdam

-------
Introduction

•	Prof Nicholls discussed the impact of sea
level rise on the coast

•	I will discuss the economic implications,
focussing on

-	direct costs

-	adaptation

-	general equilibrium effects

-------
Direct costs

•	Sea level rise has a number of impacts

-	Inundation / land loss

-	Flood frequency

-	Wetland loss

-	Coastal protection

-	Saltwater intrusion
• • •

•	For each of each, you can estimate a unit
cost and multiply that with the impact
estimate of the previous presentation

•	The sum is the direct cost

-------
100,000,000

10,000,000

1,000,000

~ This study
Fankhauser
¦ Darwin

Land value depends on income density
calibrated to US data





Use average values rather than beach front values as
property markets will adjust to coastal realignment

-------
>XHXILO i

_= I BHHd 1
3

YHXLG 1

imaa

u •"

= rL

P **

it
* «



IDMl
Hill
Lai
ai

y = 11711%:

R*

it

* f

1

IWOB i : a

imnu
tunc

¦.'DOC
iniC

iao
io

OJ

u

m t

: i	¦

T

¦i
" E

D

iOCffii

tookd:

J: mi ,

s

laxQi]
ICKC
1KC

w

BLQ'
OJ

~E '£

if.

J T

"L."

, ¦ -* " . v.1
_ + ¦'*¦ , ¦ t

— -1—Jl • 
-------
Gruenspecht's Law

V

t.

Utii	***

. a - i t ¦



One cannot study impacts without studying adaptation

-------
£_

\
o
o
o

V)
"D
O
O

_Q
"D

¦J—

a

fc)

o

4)
Q_

100

10

1

10

0.1

0.01

•• * Observed adaptation

* >' » _ .
• 4 • >

* .O *

•• • •

10Q) « 10$6£#*# ?0000#*# .100000

*,• • I ,*

	1 	#	'	1 •	

#•

0.001

Income ($/p/yr)

-------
3.0 n

2.5

2.0

Q_

a

C 1.5

o
£_

(1)

Q_

1.0

0.5

0.0

2100, A IB, annual impact
50 most vulnerable countries
Optimal adaptation

Land loss
Wetland loss
Protection cost

CJ ZJ V) V)

£li

OOEt
ca. a_2

O -C V)

j- at—i
U 00 —

^ q

a zj

V) O"
V)-

CQ"c!
§ N

II

zj a vj
z'sig

a



a

ZJ
CL

a
o_

ago

.b! a 'vj a^a

00 *--£$£¦
a_>^S'r c

p£i§"^E&§§a= §:0_o

£ ^q|_~0"P	ii— a P- n £-- a • — ^ -

o
o

Q_



_ _J|	 vj

O b •' h H U O C —

5aa|s>|w<5

lu	~a

N

o



£
o

._ a

>"6
o

s_

o

°3

a
a

ZJ
C CT
O-

Jl

I/)

_a

s_
O



vj a —
TJ'rtj

c o a
a-2 j-

V) ft) 00

M a

£ s

o^-
i/)

CJ V) CJ
(S) O t;

c s_ a-

•= ZJ-E

I £
o

-I-

V)

O

O V)

-------
General equilibrium effects

What are the wider economic implications?

Land loss would affect agriculture, and
hence all other markets

Coastal protection would affect
construction and capital

Static CGE model

Dryland lost is a loss of the endowment
land; we also include a case in which a
proportional amount of capital is lost

Coastal protection is a defensive
investment, financed by a forced mcrzasz
in savings

-------
No protection; 2050; 25 cm SLR

0.05

0.00

-0.05

PL

- .i

USA	EU EEFSU

c

o
o

S_

o

Q_

-0.10

-0.15

-0.20

¦	Land / Capital loss

¦	GDP (land)

~	GDP (L+C)

¦	C02 (land)

~	CO 2 (L+C)

Jtpjkl

¦

RoAl

EE

I

IHIND

RbW

Bosello, Roson, Tol, 2007, Environmental and Resource Economics

-------
Full protection; 2050; 25 cm SLR

¦ Protection costs
~ GDP

mcoz

-0.4

CHIND

-0.6

Bosello, Roson, Tol, 2007, Environmental and Resource Economics

-------
World

Rest of the World
Southeast Asia |
East Asia
Australiaand New Zealand

Canada
USA

0

HEV
Direct

2000 4000 6000

Million dollar per year

Darwin, Tol, 2001, Environmental and Resource Economics

-------
Conclusions

•	Sea level rise is one of the better
understood impacts

•	Estimates are uncertain, however, partly
because the current data are not very
good, and partly because the impact is in a
remote future

•	Two big uncertainties are wetland value and
adaptation

•	Unknowns include saltwater instrusion and
storms

-------
I 6 H



.j- &

-"V



_l	I	l_

-AIM
-,V2
-111
-B2

¦ -» *—* 1

-Ain
-A2
-SI
-B2

t ~ I t ¦

_i	1	1	1—

—i	1	1	1

-^Ain
-*-A2
-*-BI
-¦-112



Figiirc 30. PicrUacliin casta s& &	of current jtcm, ckim<*iic pffodufit fcr the 13 most

iJfcctad countrte; under the ti.a SFiES fcti-fcb (AlPT, A2, Hi, ft;;. JIQh (a), ai'xB (BJ ind
aw:« {4-

-------
Vulnerability to Natural Disasters and Per Capita Income

c

0	^

1	°

§•3

a. ^
*

II

a

s- «4i
^ a

2
0
-2
-4
-6
-8
¦10
¦12

~ ~ #»~ ~—~

~ ~

y = -1.0196X + 4.7624
R2 = 0.2105

Ln(Income per capita in US dollar)

Yohe and Tol, 2002, GEC

-------
Vulnerability to Natural Disasters and Income Distribution

c

0	^

1	°

8- 3

a. ^
*

II

a <^>
£- «4i

^ a

0
-2
-4
-6
-8
-10
-12

£ *	^ 1	* *

~ ** 3*% ~ S.7 * ~ l> * *«*£

. .	~	a *	—*Z	

~ ~

y = 2.5528x^- 12.716
R2 = 0.0753

Ln(6ini coefficient)

Yohe and Tol, 2002, GEC

-------
Australasia (7)
AfricidS)
Asia (46)
l Uinipc (23)
South America. (12)
North America (11 ])

I

WiMMlLind {52)
Fresh '*aLcr marsh (124)
Siltfhradkuh marsh l'57'i
(Jnvegetalsd .%dimcnL CIS)
Mangrove (£0)

Wetland *aJut by ciinliiKnl

~1	1—1

TTTTl	

ID

TTT1	

ICO

rmi	

DM

ID.OCO

100,000

Wi-tiand lulu* by m-Lhuid !>pt-

X



IH

10

100

1,000

10.000

H30.CO0

Btodtverelty ; I J-|
Amenity (43)
rtuiwixfi (18)
Malereils (32)
KccicillimaJ l-fifilnr (56)
llecrtaliafli, Musiim (50)
ItalHui Mi Nursery i67)
Wak* Quality (251
Vr'auf Supply |1E|
ETood (26)

1

M Ad »alut ht viii&jtdst-nta

IH

ill

IH

1U0

i-H

IH
Zh

H

IJKO

I0,C«I

11 ii i

IBDjSDD

Oppomiily Cos! 111)
Marias IIIom |"5L)
l^odiKUcn I 'unclian (19)
>tel E'-KrUsr locnme f22)
Rjpl jc^iTK :i Cost (28 ;<
Tcife) Cost riffi
IfedcvUc Pricing (3)
CVM |1B)

I

MMli nl yjIih hy valuation nuttvid

H

X

X

I

IH

	I 	!-¦

ID	ICQ	UDOO	10,000

Wetland vali».:L995 US Shi %rLog scale)

-i—	ii

laUKB

-------
Modeling changes in ocean biogoechemistry due to climate change and ocean acidification

Sarah Cooley (WHOI; scoolev@whoi.edu)

December 31,2010

As rising anthropogenic carbon dioxide emissions have contributed to climate change by altering
the Earth's radiative balance, about one-third of this carbon dioxide (Sabine et al. 2004;

Intergovernmental Panel on Climate Change 2007) has also dissolved in the ocean to cause ocean
acidification (OA), a much less well-publicized phenomenon. The physical chemistry of OA is very well
understood, and these changes have been observed at many locations worldwide. Observational data have
contributed to the development and testing of coupled ocean models used to examine climate change and
ocean acidification. However, we do not have enough information yet to predict the biological responses
to ocean acidification for more than a handful of organisms. As a result, forecasts of the ecological
responses to ocean acidification still contain great uncertainty. Predicting OA's socioeconomic effects is
also therefore in its infancy. Determining the end-to-end effects of ocean acidification will require a
combination of data collection and synthesis, model and method development in multiple disciplines, and
intercomparison and linking of earth system, ecological, and socioeconomic models.

Chemistry & observations

Ocean acidification refers to the suite of chemical changes that occur when excess atmospheric
carbon dioxide (C02) from human activities reacts with water molecules to form carbonic acid, a weak
acid that partially dissociates into hydrogen ions and bicarbonate. Some of the carbon dioxide molecules
also react with dissolved carbonate ions that are already present, forming more bicarbonate. The net
chemical consequences of these reactions are an increase in hydrogen ions, a decrease in carbonate ions,
and an overall increase in the content of dissolved C02 species in water. The increase in hydrogen ions
increases solution acidity, which also decreases measured pH.

The total quantity of dissolved C02 and carbonate system species in seawater, or the inorganic carbon
system, can be measured directly or calculated from other observed parameters. Any two of the four
parameters including total dissolved inorganic carbon (DIC; the total amount of dissolved C02,
bicarbonate, and carbonate ions), total alkalinity (TA; the excess base in seawater), pH, and the partial
pressure of C02 (pC02), can be derived from two other measured parameters. Other difficult-to-measure
parameters, such as the concentration of carbonate ions and the saturation state of calcium carbonate
minerals (Q), can also be derived similarly. When state-of-the-art methods, standards, and tightly
controlled laboratory conditions are used, measured DIC, TA, pH, and pC02 have uncertainties ranging
from ~0.03%~0.2% (depending on parameter; Dickson, 2009, personal communication). Uncertainties
double if analyses are done in less tightly controlled conditions (e.g., at sea). Historically, observational
campaigns have usually measured seawater DIC and TA, then calculated pH and pC02. Using this
method, measurement uncertainties plus error in equilibrium constants yield a combined resultant error of
0.6%-6.3% in derived parameters (Dickson and Riley 1978). The carbonate ion concentration derived by
this method has an error of 3.1 %; hydrogen ion concentration has an error of 5.6% (Dickson and Riley
1978). Consequently, uncertainty around values of Q calculated from derived carbonate ion
concentrations are of a similar magnitude as the annual rate of change in Q (-0.09 year"1) observed at
time-series stations like ALOHA (Figure 1 in Feely et al. 2009a), which underscores the necessity of
long-term ocean acidification monitoring with high-quality measurements.

Seawater chemistry measurements from time-series stations and repeat hydrography cruises show the
global extent and progress of ocean acidification. The inorganic carbon chemistry of upper-ocean
seawater has been tracked at monthly monitoring locations including ALOHA near Hawaii, BATS near
Bermuda, station PAPA in the North Pacific, and ESTOC near the Canary Islands, and records show a
progressive decrease in upper-ocean pH, Q, and/or carbonate ion concentration as seawater C02 rises
(Dore et al., 2009, updated in Doney 2010; Gruber et al. 2002; Gonzalez-Davila et al. 2010). Comparison
of datasets from repeat hydrography programs has shown that changes in ocean carbonate chemistry due
to the invasion of anthropogenic C02 penetrate thousands of meters in each ocean basin (Sabine et al.

1

-------
2004). Variabilities of pH and pC02 are naturally greater in coastal regions because of respiration,
photosynthesis, and runoff, so ocean acidification research must also determine what conditions may be
damaging there (National Research Council 2010). Numerous programs now focus on establishing
baseline conditions in many more locations, such as the Arctic Ocean (Azetsu-Scott et al. 2010; Cai et al.
2010), but infrastructure development is still required to collect enough data to determine baseline
conditions and indicate future changes in all regions (Feely et al. 2009b).

Ocean acidification and climate change are also expected to alter other nutrient cycles. Decreases in
pH and carbonate ions will affect the solubility, adsorption, toxicity, and rates of redox reactions for
metals in seawater (Millero et al. 2009). The biological availability of many metals could change, with
varying outcomes: increased copper could kill more phytoplankton, whereas increased iron could support
more phytoplankton growth. These changes could be especially important in estuarine biogeochemical
cycling, where redox reactions tightly control the behavior of metals and gaseous components like C02,
which in turn control phytoplankton community composition (Millero et al. 2009). Throughout the
oceans, ocean acidification and climate change may also alter nitrogen cycling. Bacterial nitrification
could slow as pH decreases and cyanobacterial nitrogen fixation could increase as temperature and C02
levels rise, promoting an overall shift towards a larger reduced nitrogen pool dominated by ammonia
(Hutchins et al. 2009). At the same time, increasing temperature could slow vertical mixing and thereby
reduce upwelling of nutrients from deep water, enhancing nutrient limitation. These consequences of
ocean acidification are somewhat less well quantified than the expected changes in pH and carbonate ion
concentration, so present OA forecasts primarily focus on inorganic carbon cycle-related changes in the
oceans.

Earth system models

Coupled ocean models used to study climate change often include carbon cycles that interact with
meteorological variables, oceanographic variables, and biogeochemical processes; therefore, these models
simulate ocean acidification as well as other anthropogenically forced changes in Earth systems. Model-
data comparisons are used to judge the models' skill at creating hindcasts, and the models that reproduce
major features of circulation and tracer transport are believed to provide credible estimates of future
climate change at large scales (Intergovernmental Panel on Climate Change 2007). Intercomparison
exercises, such as the international Ocean Carbon Model Intercomparison Project (OCMIP), then
compare forecasts from multiple skillful models to develop estimates of the range of future conditions.
Atmospheric C02 levels of -780 ppm by 2100 (IS92A scenario, Leggett et al. 1992) yielded a median
response for OCMIP's thirteen models in which ocean pH decreased by 0.3-0.4 and carbonate ion
concentrations dropped globally. Overall, there will be an equatorward contraction and shallowing of
high-carbonate waters suitable for animals that make hard shells and skeletons (Orr et al. 2005).
Meanwhile, temperature increases due to climate change that decrease C02 solubility will counteract less
than 10% of the chemical changes associated with ocean acidification (Orr et al. 2005). The subsequent
Coupled Climate-Carbon Cycle Intercomparison Project (C4MIP) found that in 11 climate models with
land and ocean carbon cycles, feedbacks between climate change and the carbon cycle would occur and
increase atmospheric C02 by an additional -50-100 ppm, causing additional warming and allied changes
(Intergovernmental Panel on Climate Change 2007).

The biogeochemical consequences of future ocean carbon cycle changes are less clear at present.
Intercomparison has shown that accurate physics are a very strong determinant of whether modeled
biogeochemical processes replicate observed conditions (Doney et al. 2004; Najjar et al. 2007). Detailed
intercomparisons of biogeochemical parameterizations in coupled models, though, are often limited for
several reasons. First, the level of biogeochemical complexity in the OCMIP/C4MIP models varies
greatly, and appropriate evaluation methods and criteria for each model depend on their specific
biogeochemical parameterizations. Second, spatially or temporally sufficient data is often lacking for
evaluating many modeled biogeochemical processes in detail. Third, modeled biogeochemical parameters
(e.g., growth of a generalized pool of zooplankton) are often not directly comparable to observational data
(Doney et al. 2009). Most biogeochemical models do undergo qualitative or quantitative model-data

2

-------
comparison (68%), but there is no standardized approach, and few assessments use in-depth statistical
techniques for model-data comparison (Stow et al. 2009). Some models involved in OCMIP and
subsequent intercomparison studies have undergone extensive model-data comparison individually. For
example, Community Climate System Model (CCSM3) output was evaluated using satellite-derived
surface ocean chlorophyll and primary productivity, climatologies of nutrients and pC02, and time-series
data from observational programs like JGOFS (Doney et al. 2009). This model had the strongest model-
data correlations for SST and nutrients, moderate correlations for surface pC02 and C02 air-sea flux, and
weaker correlations for ecosystem variables including chlorophyll, primary production, phytoplankton
growth rate, etc. (Doney et al. 2009). Further improvements to the biogeochemical model and its
feedbacks may bring the ecosystem variables, which are the most dependent on its parameterizations, into
better agreement with observations. To that end, a new model that builds upon CCSM, called the
Community Earth System Model (CESM), is being developed and tested (University Corporation for
Atmospheric Research 2010). Similar work is underway with other coupled models.

The primary uncertainty in ocean acidification chemistry forecasts comes not from the carbon
chemistry itself, measurements, or from coupled models' abilities to predict ocean carbon inventories, but
rather from the uncertainty in anthropogenic C02 emission trajectories (Intergovernmental Panel on
Climate Change 2007). The rate of change of ocean acidification in offshore seawater appears to mirror
the rate of atmospheric C02 rise (Figure 1 in Feely et al. 2009a). Even if the atmospheric C02 trajectory
levels off today, ocean pH, surface Q, and deep-ocean Q will continue to be depressed compared to
preindustrial conditions in the next five centuries (Frolicher and Joos 2010). Long-term forecasts suggest
that oceanic uptake of C02 will slow as the chemical changes from OA accumulate (Sabine and Tanhua
2010) and as ocean circulation slows from climate change (Intergovernmental Panel on Climate Change
2007), but these factors will not reverse ocean acidification either. Over shorter periods, a moderate
amount of uncertainty about ocean acidification's progress in coastal zones is associated with the
possibility of changes in freshwater cycling and deposition of other, acid-generating pollutants near shore
(Doney et al. 2007; Doney 2010).

Biological responses & models

Most of the uncertainty about OA's effects on marine ecosystems arises from our present incomplete
knowledge about the individual and population-level responses to OA. Any of the chemical changes due
to ocean acidification may be biologically relevant (reviewed in National Research Council 2010). The
decline in carbonate ions decreases the amount of carbonate building blocks available for marine animals
that create calcium carbonate shells and skeletons. These organisms include primary producers such as
coccolithophores and coralline algae, zooplankton such as pteropods, mollusks such as clams, oysters,
and mussels, crustaceans such as crabs and lobsters, and reef-forming corals. Most of the calcifying
organisms studied show negative responses to ocean acidification such as decreasing calcification rates,
delayed larval development, and smaller shells (Kroeker et al. 2010). The cellular and organismal
mechanisms behind these responses are not yet clear. In other organisms, the decrease in seawater pH, the
increase in C02, or both may affect concentration gradients of hydrogen ions across cell membranes or
change oxygen-C02 respiratory balances (National Research Council 2010). Finally, increasing C02
concentrations from ocean acidification may benefit photosynthetic organisms like phytoplankton,
macroalgae, and seagrasses.

Scaling ocean acidification's effects on individual organisms to populations and ecosystems
remains a challenge. Not only will OA's effects on individuals alter their individual performance, but it
may also alter their behavior in ways that will generate population-wide consequences. For example,
ecosystem-scale observations in natural environments have reported declines in calcifier populations and
increases in seagrass populations with increased proximity to C02 vents (Hall-Spencer et al. 2008), and
shifts from calcifier-dominated communities to photosynthesizer- and invertebrate-dominated
communities with long-term pH decreases (Wootton et al. 2008). Until ocean acidification's effects on
both life functions and behaviors of susceptible species are known, modeling studies are limited to using

3

-------
statistical responses generalized from individual studies and falling within trends observed in ecosystem-
based studies such as these.

In the absence of mechanistic knowledge of marine organisms' responses to OA, applying
statistical relationships describing general responses can still be instructive for providing first estimates of
OA's total impacts. Brander et al. (2009) estimated the economic impacts of OA on coral reefs by relating
loss of coral cover to ocean acidification using a general relationship summarizing multiple coral studies
and an economic value meta-analysis. They concluded that losses were approximately one order of
magnitude smaller than those from climate change. However, only broad insight is available from this
study because the analysis depends heavily on an assumed linear biological response (loss of coral cover)
that may in nature be nonlinear, stepped, or otherwise episodic (Kleypas and Yates 2009). In another
analysis, Cooley and Doney (2009) determined the potential losses from ocean acidification to United
States commercial mollusk harvests by assuming that calcification rate decreases in mollusks directly
correlated to population decreases. They concluded that annual losses in ex-vessel revenue could range
from the tens to hundreds of millions of dollars. Their linear damage function approximated trends
comparable to those in individual and ecosystem studies, but it did not explicitly include interspecies
interactions, adaptability, or long-lasting damage to juveniles, all of which could affect populations over
time periods relevant to the analysis. Until more biological data is available, initial studies like these must
necessarily use statistical fits instead of mechanistic responses, but these types of studies must be
interpreted with care to avoid drawing conclusions broader than the ingoing biological information
permits.

Conclusions & future work

The chemical reactions and equilibria governing carbon dioxide's behavior in seawater have been
well understood for decades to centuries, and worldwide observational datasets (e.g., Figure 1 in Feely et
al. 2009a) that show decreasing ocean pH and carbonate ion concentration with rising atmospheric C02
levels agree with scientific theory. Ocean acidfication's effects on ocean chemistry can be forecast when
this well-understood physical chemistry is included in the ocean carbon cycle of coupled ocean models.
Intercomparisons of multiple skillful models suggest that ocean acidification will progress globally, and
its progression depends greatly on atmospheric C02 emissions trajectories. C02 emissions, however,
depend fundamentally on human behavior, which is far more uncertain. Local factors and climate change
will exert secondary control on OA in nearshore regions.

Our ability to forecast ocean acidification's total effects on ecosystems and human economies
under different C02 emissions scenarios is limited because our knowledge about biological responses to
OA is incomplete. We also do not know how ocean acidification will affect coastal ecosystems where
marine organisms thrive, although C02 concentration and pH variabilities there are already high. We do
not know whether ecosystems will undergo stepwise responses or cross tipping points, and we do not
understand how best to scale individual effects to population-wide responses.

Until some of these chemical and biological questions are resolved, we will be limited to making
broad assessments of potential socioeconomic losses from ocean acidification using observed biological
trends. Until then, modeling work must continue to ensure that biogeochemical model skill continues to
improve. At the same time, ecosystem models of marine communities like EcoPath and Atlantis must be
built, tested, and tuned to permit the extrapolation of biological responses to OA to ecosystems. Finally,
socioeconomic studies must find improved ways to value the range of market and nonmarket services that
marine ecosystems provide. These parallel efforts will permit skillful biogeochemical models to be linked
to ecosystem models and to socioeconomic models.

References

Azetsu-Scott, K., A. Clarke, K. Falkner, J. Hamilton, E. P. Jones, C. Lee, B. Petrie, S. Prinsenberg, M.

Starr, and P. Yeats, 2010: Calcium carbonate saturation states in the waters of the Canadian

Arctic Archipelago and the Labrador Sea. J. Geophys. Res., 115, 18 PP.

Brander, L. M., K. Rehdanz, R. S. Tol, and P. J. van Beukering, 2009: The Economic Impact of Ocean

4

-------
Acidification on Coral Reefs. ESRI, (Accessed December 29, 2010).

Cai, W., L. Chen, B. Chen, Z. Gao, S. H. Lee, J. Chen, D. Pierrot, K. Sullivan, Y. Wang, X. Hu, W.

Huang, Y. Zhang, S. Xu, A. Murata, J. M. Grebmeier, E. P. Jones, and H. Zhang, 2010: Decrease
in the C02 Uptake Capacity in an Ice-Free Arctic Ocean Basin. Science, 329, 556-559.

Cooley, S. R., and S. C. Doney, 2009: Anticipating ocean acidification's economic consequences for
commercial fisheries. Environ. Res. Lett., 4, 024007.

Dickson, A., and J. Riley, 1978: The effect of analytical error on the evaluation of the components of the
aquatic carbon dioxide system .Marine Chemistry, 6, 77-85.

Doney, S. C., K. Lindsay, K. Caldeira, J. Campin, H. Drange, J. Dutay, M. Follows, Y. Gao, A.

Gnanadesikan, N. Gruber, A. Ishida, F. Joos, G. Madec, E. Maier-Reimer, J. C. Marshall, R. J.
Matear, P. Monfray, A. Mouchet, R. Najjar, J. C. Orr, G. Plattner, J. Sarmiento, R. Schlitzer, R.
Slater, I. J. Totterdell, M. Weirig, Y. Yamanaka, and A. Yool, 2004: Evaluating global ocean
carbon models: The importance of realistic physics. Global Biogeochem. Cycles, 18, 22 PP.

Doney, S. C., 2010: The Growing Human Footprint on Coastal and Open-Ocean Biogeochemistry.
Science, 328, 1512-1516.

Doney, S. C., I. Lima, J. K. Moore, K. Lindsay, M. J. Behrenfeld, T. K. Westberry, N. Mahowald, D. M.
Glover, and T. Takahashi, 2009: Skill metrics for confronting global upper ocean ecosystem-
biogeochemistry models against field and remote sensing data. Journal of Marine Systems, 76,
95-112.

Doney, S. C., N. Mahowald, I. Lima, R. A. Feely, F. T. Mackenzie, J. Lamarque, and P. J. Rasch, 2007:
Impact of anthropogenic atmospheric nitrogen and sulfur deposition on ocean acidification and
the inorganic carbon system. Proceedings of the National Academy of Sciences, 104, 14580-
14585.

Feely, R. A., S. C. Doney, and S. R. Cooley, 2009a: Ocean Acidification: Present Conditions and Future
Changes in a High-C02 World. Oceanography, 22, 36-47.

Feely, R. A., V. J. Fabry, A. G. Dickson, J. Gattuso, J. Bijma, U. Riebesell, S. C. Doney, C. Turley, T.
Saino, K. Lee, K. Anthony, and J. Kleypas, 2009b: An international observational network for
ocean acidification. Venice, Italy, http://www.oceanobs09.net/ (Accessed December 29, 2010).

Frolicher, T., and F. Joos, 2010: Reversible and irreversible impacts of greenhouse gas emissions in

multi-century projections with the NCAR global coupled carbon cycle-climate model. Climate
Dynamics, 35, 1439-1459.

Gonzalez-Davila, M., J. M. Santana-Casiano, M. J. Rueda, and O. Llinas, 2010: The water column

distribution of carbonate system variables at the ESTOC site from 1995 to 2004. Biogeosciences,
7, 3067-3081.

Gruber, N., C. D. Keeling, and N. R. Bates, 2002: Interannual Variability in the North Atlantic Ocean
Carbon Sink. Science, 298, 2374-2378.

Hall-Spencer, J. M., R. Rodolfo-Metalpa, S. Martin, E. Ransome, M. Fine, S. M. Turner, S. J. Rowley, D.
Tedesco, and M. Buia, 2008: Volcanic carbon dioxide vents show ecosystem effects of ocean
acidification. Nature, 454, 96-99.

Hutchins, D. A., M. R. Mulholland, and F. Fu, 2009: Nutrient Cycles and Marine Microbes in a C02-
Enriched Ocean. Oceanography, 22, 128-145.

Intergovernmental Panel on Climate Change, 2007: Climate Change 2007: The Physical Science Basis.
Intergovernmental Panel on Climate Change,

Kleypas, J. A., and K. K. Yates, 2009: Coral Reefs and Ocean Acidification. Oceanography, 22, 108-117.

Kroeker, K., R. Kordas, R. Crim, and G. Singh, 2010: Meta-analysis reveals negative yet variable effects
of ocean acidification on marine organisms. Ecology Letters, 13, 1419-1434.

Leggett, J., W. Pepper, R. Swart, J. Edmonds, L. Meira Filho, I. Mintzer, M. Wang, and J. Watson, 1992:
Emissions Scenarios for the IPCC: An Update. Cambridge University Press,
http://sedac.ciesin.columbia.edu/ddc/is92/ (Accessed December 29, 2010).

Millero, F. J., R. Woosley, B. Ditrolio, and J. Waters, 2009: Effect of Ocean Acidification on the
Speciation of Metals in Seawater. Oceanography, 22, 72-85.

5

-------
Najjar, R. G., X. Jin, F. Louanchi, O. Aumont, K. Caldeira, S. C. Doney, J. Dutay, M. Follows, N.

Gruber, F. Joos, K. Lindsay, E. Maier-Reimer, R. J. Matear, K. Matsumoto, P. Monfray, A.
Mouchet, J. C. Orr, G. Plattner, J. L. Sarmiento, R. Schlitzer, R. D. Slater, M. Weirig, Y.
Yamanaka, and A. Yool, 2007: Impact of circulation on export production, dissolved organic
matter, and dissolved oxygen in the ocean: Results from Phase II of the Ocean Carbon-cycle
Model Intercomparison Project (OCMIP-2). Global Biogeochem. Cycles, 21, 22 PP.

National Research Council, 2010: Ocean AcidificatiomA National Strategy to Meet the Challenges of a
Changing Ocean. The National Academies Press, Washington, D.C.,
http://www.nap.edu/catalog.php?record_id=12904#toc.

Orr, J. C., V. J. Fabry, O. Aumont, L. Bopp, S. C. Doney, R. A. Feely, A. Gnanadesikan, N. Gruber, A.

Ishida, F. Joos, R. M. Key, K. Lindsay, E. Maier-Reimer, R. Matear, P. Monfray, A. Mouchet, R.
G. Najjar, G. Plattner, K. B. Rodgers, C. L. Sabine, J. L. Sarmiento, R. Schlitzer, R. D. Slater, I.
J. Totterdell, M. Weirig, Y. Yamanaka, and A. Yool, 2005: Anthropogenic ocean acidification
over the twenty-first century and its impact on calcifying organisms. Nature, 437, 681-686.
Sabine, C., and T. Tanhua, 2010: Estimation of Anthropogenic C02 Inventories in the Ocean. Annual

Review of Marine Science, 2, 175-198.

Sabine, C. L., R. A. Feely, N. Gruber, R. M. Key, K. Lee, J. L. Bullister, R. Wanninkhof, C. S. Wong, D.
W. R. Wallace, B. Tilbrook, F. J. Millero, T. Peng, A. Kozyr, T. Ono, and A. F. Rios, 2004: The
Oceanic Sink for Anthropogenic C02. Science, 305, 367-371.

Stow, C. A., J. Jolliff, D. J. McGillicuddy Jr., S. C. Doney, J. I. Allen, M. A. Friedrichs, K. A. Rose, and
P. Wallhead, 2009: Skill assessment for coupled biological/physical models of marine systems.
Journal of Marine Systems, 76, 4-15.

University Corporation for Atmospheric Research, 2010: New computer model advances climate change
research. Press Release, August 18 http://www2.ucar.edu/news/2366/new-computer-model-
advances-climate-change-research (Accessed December 29, 2010).

Wootton, J. T., C. A. Pfister, and J. D. Forester, 2008: Dynamic patterns and ecological impacts of

declining ocean pH in a high-resolution multi-year dataset. Proceedings of the National Academy
of Sciences, 105, 18848-18853.

-------
Modeling changes in
ocean biogoechemistry
due to ocean acidification
and climate change

Sarah Cooley, scooley@whoi.edu
January 28, 2011

Improving the Assessment and Valuation of
Climate Change Impacts for Policy
and Regulatory Analysis

Sponsored by

United Slates
Environmental
Protoction Agency

v>EPA

U.S. DEPARTMENT 0»

WENERGY

January 27-28, 2011	Capital Hilton, Washington, DC

-------
Today's talk

Chemistry & observations

•	What is OA? How well can we detect it?

Earth system models

•	Ability to forecast future ocean conditions?

Biological responses & models

•	How well can we forecast the future?

Key knowledge gaps & needs

-------
Rising C02 causes ocean acidification

Increasing C02
•Lowers pH
•Lowers [C032 ]
saturation state

"Q"

Present change is
faster than rock
weathering & other
compensatory
mechanisms

-------
Observations show OA advancing

400-

37S

(atmC02)y = 1.811x- 3252.4
R2 = 0.95, st err = 0.028

350-

LEGEND

-m- Mauna Loa atmospheric C02 (ppmv)
Aloha seawater pC02 (patm)

Aloha seawater pH

O
u

325-

300-

275-



(pC02)y=1.90x-3453.96
R2 = 0.3431, st err = 0.20

(pH)y = -0.00188x +11.842
R2 = 0.289, st err= 0.00022

--8.05

8.00

1960

1970

1980

1990

2000

2010

4.9-

4.7-

4.5-

£ 4.3-
a;

fO

^ 4.1-
.S

&

2 3.9-
$

3.7-
35-
33-

y =-0.012x +29.615
R2 = 0.159, st err = 0.0021

LEGEND

~ Surface nc3 saturation
Surface 0„ saturation

23°N
22-N
21-N
20-N
19"N

160TV 1S8-W

• Station Aloha

Station Mauna Loa A

-r-

y = -0.00758x+ 18.86
R2 = 0.125, st err = 0.0015

1960

1970

1980
Year

1990

2000

2010

Now:

t Atmos. pC02
| Ocean pC02
| Ocean pH
| Calcite sat. st.

I Aragonite sat. st.
t Coastal variability

Anthropogenic C02 in upper ocean worldwide

60° 50° 40° 30° 20° 10°
Latitude

10° 20°

Doney et al. Oceanography 2009, Sabine et al. Science 2004

-------
Other effects & synergies

anammox

assimilation t Organic
degi'adation Nitrogen

Hutchins et al., Oceanography, 2010

Marine nitrogen pool shifts towards
ammonia as N2 fixers thrive in a high-

C02 ocean

+5

NO,

-------
Other effects & synergies

Metal ion speciation changes
from changing pH and/or

C02:

• Copper (Cu2+)
increases: toxic!

Iron (Fe2+)
increases: fertilizer?

Speciation Changes with Time

•o so

&

LA

•b 40
C

a

S 30

a.

20
10,

—o—

Cu3*

0

CuCO



3



	o

xP

o-

-o-

_0-*

.-O

2000

2050 2100

2150
Year

2200 2250

2300

-o-

Fe3"



FeC03

2300

Millero et al., Oceanography, 2010

-------
Other effects & synergies

Reactive
nitrogen

Deforestation Fossil fuels Agriculture
and industry

Coastal eutrophication
and hypoxia ^

Altered primary
production

Ocean uptake
and acidification

Other anthropogenic changes could
be antagonistic or synergistic

Low O, upwelling

Oeoxygenation

Doney, Science, 201

Climate change
(winds, temperature, sea
ice, precipitation, runoff
and ocean circulation)

Carbon dioxide

Organic pollutants and trace metals

Nutrients

-------
Earth system model simulations

"Skill" evaluated with model-data
comparisons of hindcasts

• Simple & not-so-simple statistics

Correct physics is key!

BGC parameterizations are
under continuous improvement

Mean Surface Chlorophyll (Sep 1997 - Dec 2004)

mg Chi m 3

I
i

3
1

0.5

0.3

0.2

0.175

0.15

0.125

0.1

0.07

0,05

0.03

0.01

mg Chi m

!

i

3
1

0.5

0.3

0,2

0.17

0.13

0.12

0.1

0.07

0.05

0.03

0.01

mg Chi mJ

flj,

- 8

a

0L

i

Doney et al., J. Mar. Sys.,

-------
Model intercomparison used to
create, evaluate forecasts

Multi-model median of % saturation of carbonate ion from OCMIP-2 models:
broad agreement that ocean pH and carbonate ion levels will decline in

response to rising atmospheric C02.

2080-2099

2011-2030	2046-2065

BOH

8011



MTW



WOT

GOV

IOMCITu: E

mam

¦ rm

LONGITUDE



40*M

40*N

Key question: what will C02atm be?

40H

0*
40*S
10*$

ICW

i rmJTi i*r



wr

IPCC, 2007

-------
5

o

u

3
IS

o.

¦

o

¦

C

<

Biological Groups at Risk

Pteropods

Some plankton

Warm-water corals Cold-water corals

Many mollusks

Marine predators

Reef

communities

Businesses

Fishermen

Coastal environments

-------
Calcif cation
responses vary

With decreasing Q,

•	Crustaceans

•	Urchins, some

algae, corals 4

•	Mollusks ^

• Individual & population
implications
not yet understood

Ries et al., 2009, Geology

OS
p.

D.5 10 15 2.0 2.5
50 Q Haimeda	»

O 4/1

_ _ ¦

i*°

1?

£ 20'X PeiKil uirhri

.S 10

Jj.

^ -10
3 -20
aS

0.5 1.0 1,5 2 0

M, SgrpuW HKjrm

1000 A. Blueaab

0.5
¦MJ ~.

B Shrimp

8KI

2.0 2,S 0.5 10 1.5

25 r. Coralline red alyat1

10 H a;..(5 rnussei *
*

t

OS 1.0 1.5
2 K Barg elarr

1

0

-1

-2

OS

10 |vj. Pariwirkle

ID 1,5
n

aragantB

20
IS
10
5

0_

1,0 1.5 2.0 2,5
"i5 I. TBrrparafja coral t

0.5 1.0 15
10 O. Say scallop *

C. Whelk

1.D „ 1.5
n

Saturation state

L. Conch

-------
Ecological implications

Food web effects of OA are unknown, could be extensive

-------
Ecosystem changes

Photo, U. Washington

In a coastal lagoon,
noncalcifiers replaced
many calcifiers over an
8-y. pH decline
(8.41-7.99)

(Wootton et al. PNAS 2008)

Near a volcanic C02 vent,

•adult mollusks damaged

•juvenile mollusks absent

•corals, coralline algae
absent

Hall-Spencer et al., Nature 2008

-------
How to value ecosystem services?

Ease of attaching dollar value

Public goods

Private goods

-------
OA's economic impacts

U.S. mollusk harvests

Uninfluencec
1%

Mollusks = $748M of
U.S. ex-vessel
harvests in 2007

Assume a 0.1-0.2 unit pH decrease
by 2060 = 6-25% lower harvests

-	Annual losses of $75-187M

-	NPV losses through 2060 of $1.7-10B

Oystel^ & mussels

%

ther calcifiers
1%

Calcifiers'
predators
24%

Shrimp
10%

Crabs
11%

Lobsters
9%

CD

co

o
o

ro
O

"o
E

ro
o
4=
j_>
TO
O

O)

300 800 1300 1800 2300 2800
pC02 (ppmv)

0.75-j
0.50-
0.25-
0

-0.25-

r2 = 0.70, P < 0.0001

Cooley & Doney 2009 ERL

Gazeau et al , 2007; Green et al. L&O 2009

-------
OA's economic impacts

Coral reefs

Value coral reefs via
meta-regression

More information needed
on relation of coral
cover to OA

Results strongly driven by
importance of reefs for
tourism - nonmarket servict
underestimated?

1000 —

Annual damage due



(/>
c
o

CD

A1
B2
A2
B1

500

Brander et al. 2009, http://en.scientificcommons.org/41882916

-------
Knowledge gaps for OA 1AM

Schematic from C. Moore, EPA 2010

-------
Uncertainty builds

Rising

atmospheric C02

Changing
benefits

Socioeconomic
consequences

Decreasing

0C®^_P^'	Altered

physiology
population
changes

[C032-]

Food web
changes,
ecosystem
shifts

-------
Certainty scorecard



Certainty

Data
limited?

Methods
limited?

Atmospheric C02 rising

High





Ocean pH, carbonate decreasing

High

~



Marine organisms affected

Medium

~

~

Ecosystems change

Medium/L

ow

~

~

Ecosystem services change

Low

~

~

Socioeconomic consequences

Low

~ S

-------
Modeling Climate and Acidification
Impacts on Fisheries, Aquaculture,
and Other Marine Resources

^ W7TiTjTj^^4



Paul McElhany,
Research Ecologist
Northwest Fisheries
Science Center

Background Photo byJared Figurski

-------
Talk Outline

Hope

100

Despair

-100

The Modeling Moeling
problem approaches examples

Reality Where from
check here

-------
The Problem

Living Aquatic Resource Issues

•	Capture fisheries

•	Aquaculture

•	Endangered species

•	Tourism

•	Shoreline protection

•	Human Health

C02 Effects

•	Growth and Survival

•	Range shifts

•	Stratification/circulation

—	Nutrients

—	Oxygen

—	Dispersal

•	Sea level rise

•	Acidification

•	Storms

•	Increased UV

-------
Model Flavors

Fishery stock assessments
Population Viability Analyses
Food web/ecosystem models
NPZ models

Minimum realistic models
Maximum unrealistic models
Modeled range maps
Individually-based models
Life-cycle models
Bioenergetics
Expert systems

-------
Incorporating C02:
Down-scaling IPCC- class models

•	Model Scales

-	Space

•	IPCC: typically 1° x 1° (~110 km latitude) or coarser

•	IPCC: Very poor on the coasts/nearshore, fronts and eddies

•	Biological scales: Sometimes meters mater

-	Time

•	IPCC: Does not resolve decadal scale patterns

•	Biological scales: annual and seasonal variation mater

•	Key Features to down-scale

-	Temperature

-	Stratification/Circulation/Salinity

-	Storms

-	Sea level

-	Carbon Chemistry

-------
>

Age 4 River Adult

season: 0.64
pooled group: River Run



Age 5 River Adult

season: 0.65
pooled group: River Run

Example 1: Salmon life
cycle, Climate change
and stream flow

1920	1940	1960	1980

From Crozier et al. 2008

-------
Example 2: Acidification in Puget Sound with Ecopath/Ecosim

9.66%

Calcium carbonate mineralogy

250-j High

49.03%

CaC03
Vertebrate
No biomineral
Non-CaC03

30.16%

11.14%



tfl
QJ

0J

200-

lo 150-

cu

100-

50-

Solubility

Low

Mineralogy

J Mostly amorphous CaC03, some calcite
] High Mg calcite, some amorphous CaC03
High Mg calcite

Mostly aragonite, some high Mg calcite
Aragonite

Mostly aragonite, some calcite
Mostly low Mg calcite, some aragonite
Low Mg calcite

From Busch, Harvey and McElhany in prep

-------
Puget Sound Ecosystem

4.5

4.0

3.5

0)

3.0

Q.

2 2.5

2.0

1.5

1.0

Adult lingcod

Salmon carcass l|Detritus]|| Ala/plant matte [Overstory kelp HI Benth macroalg IMI Benth miCCOalg

Phytoplankton

Busch, Harvey and McElhany in prep

-------
"1 Because	of their en

chemical composition of the open
oceans, with the exception of lead, has
not been greatly affected by human
activities."

Kates and Parris. 2003. Long-term trends and a
sustainability transition. Proceedings of the National

Academy of Science

-------
-------
Impacts on Puget Sound Harvest?

10

35

O)

'>>

CO
CD

0)

0

-HI r

-10

>.

04

to E

C tf)

¦— c
0) o

O)

c o
(0 -c

(— -I—»

o ®

^ S -20

0

0

	1	

CD

CL

*

c

o

e

CO
CO



*

O)

c

0
X

"D

o
o

O)
c

£0

M—

o
o
C£


'**—

CO

Q.

I

JZ

CO



H—

0
JZ

CO

Decline in

some

Calcifiers

~	5%

~	15%

~	25%

Busch et al. in prep

-------
Impacts on Puget Sound Biomass?

From a 25% decline in some calcifiers

I

Forage fish



1 Selected crustaceans





1 1 Oysters and clams

Sea lion c



1 1 Echinoderms
^¦AN 3 groups

Harbor seal

i

Migratory diving birds 1

¦

1

Resident diving birds





¦

1

Herbivorous birds

=-

[

Gulls



Raptors 1—

|—l—l—l—1—|—l—l—l—l—|—1—1—l—I—|—1—1—1—l—

	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1

-20 -15 -10 -5 0	5 10 15 20

Percent change in biomass

Busch et al. in prep

-------
Example 3: California Current, Climate
and OA with EwE (Ainsworth et al.)

150°0'0"W 145°0'0"W 140°0'0"W 135°0-0"W 130'0'0"W 125°0'0"W 120°0'0"W

_|		1	1	1	1	L

65°0'0"N-

60°0'0"N'

55°0'0"N-

50°0'0"N-

45°0'0"NH

40"0'0"N-

35o0'0"NH

31

150°0'0"W 145°0'0"W 140°0'0"W 135'0'0"W 130°0'0"W 125"0'0"W 120°0'0"W

-------
California Current Climate Effects

•	Primary productivity (from GFDL ESM2.1)

•	Biogeographic range shifts (from Cheung et al.)

•	Zooplankton size structure (Moran 2009)

•	Ocean acidification (Busch et al. review)

•	De-oxygenation (Whitney 2007)

Result Summary

•	General decline in fisheries, especially with all
climate effects

•	Range shifts biggest impact

-------
Example 4: Atlantis

Fisheries
Submodel

Ecology
Submodel

Oceanography
Submodel

Climate and Oceanography

Fulton, E. A. 2004. Ecological MM&lUm, 173:371-406.

-------
Atlantis Applications

System

Complexity

Understanding

System-
MSE

Fisheries

Nutrients

Mining

Pollution

Climate

Catchments

Indicators

Management

Port Phillip Bay (Australia)

















South-east Australia



















We stern port (Australia)















Victoria (Australia)



















Eastern Tasmania (Australia)













New South Wales shelf (Australia)











Clarence River (Australia)





















Soutt-west Australia

















Northeast United States













Chesapeake Bay (USA)

















California Current (USA and
Canada)















Central California (USA)











Sea of Cortez (Gulf of California)
(Mexico)



















Strart of Georgia and Puget Sound



















North Sea















Arctic

















Gulf of Carpentaria (Australia)

















Derwent River and Storm Bay
(Australia)





















CERF Tasmania (Australia)















Benguela
Gulf of Mexico

Key























Main
purpose

Secondary
purpose

Preliminary
efforts

Planned















-------
Example 5:
Bioclimate
envelope

2000 Emission Scenario

Emission Scenario A1B

Chango in catch potential
(% relative to 2005)

H <-53
H -50 to -30
¦i -31 to -15
~ -16 to -5
I I -6 to 5
I I 6-15
S3 15-30
¦I 31-50
¦I 51-100
n > io3

From Cheung et al. 2009

-------
(C)

Projected Change in Catch

Emission

Scenario

I - \.'t	k

A1B

, JW ^QrLtW> $4 ¦ - •

Change In catch (t km-2)
from 2005 to 2055

i v	ii-—w .--s« < * >!	-

v§U

' i •. 7^'

W' " v

Sf "2i i' ^	VC^Br *	<~C| so

M 6/

is,

A

A' -¦
L. J|AjC Hf

/3r

T2 *

W	St,	>-0.50 to -0.05

-4 *



I >-0.05 to -0.005
I > -0 005 lo 0 006
I I >0 005-0.05
>0.05-0.50
¦¦ >0.50

jnp	u

Emission

2000
Emiss
Scenario

From Cheung et al. 2009

-------
Example 6: Extinction risk for 82
species of tropical corals

EXCEPTIONALLY | VERY

	 f

UNLIKELY , UNLIKELY (

<1% 1-10%

LESS LIKELY
THAU NOT

33-50%

MOSf LIKELY
THAN NOT

50-66%

LIKELY

66-90%

I VEftY
LIKELY

| VIRTUALLY
CERTAIN

90-99% >99%

10-33%

UNLIKELY

-------
Incorporating Uncertainty

5000

Uncertainty
changes mean,
not just the
range

Productivity

From McElhany et al. 2010

-------
Reality Check - Some big questions

Florida - yes or no?

Gulf stream - same

Increased stratification - how much, where, effect?
Upwelling - same

Decadal oscillations ( regime shifts")???

Adaptation to OA and temp?

Ice ecosystems?

Rainfall changes and freshwater systems - where, how
much

Where will fishing get better?

-------
Details Matter

•	Species differences

•	Species interactions (predator-prey mismatch)

•	Phenology

•	Synergistic effects

•	Short term variability

•	Local circulation

•	Lab studies don't scale to ecosystems

-------
Moving Forward:

Coarse scale impact assessment

•	Back Of Envelope (BOE) estimates

•	Three Approaches:

—	Bioclimate envelope as key first pass estimates

—	Minimum realistic models on high value fisheries

—	Ecosystem/foodweb to look for interactions

•	Resolution of big climate questions

-------
Some References

•	Cheung et al. 2009. Large-scale redistribution of
maximum fisheries catch potential in the global
ocean under climate change. Global Change
Biology.

•	Kevern et al. (ed). 2009. Climate change
implications for fisheries and aquaculture. FAO.

•	Stock et al. in press. On the use of IPCC-class
models to assess the impact of climate on Living
Marine Resources. Progress in Oceanography.

-------
Modeling economic impacts of climate change and ocean acidification to fisheries

David Finnoff (University of Wyoming)

Abstract

Ocean acidification appears to have potential to be a significant problem. Past declines in ocean surface
pH have been linked to mass extinction events (Guinotte and Fabry, 2008). While I am not an expert in
the science, the issue starts with declines in pH (increased acidity) causing a reduction in carbonate ion
concentration which in turn causes a reduction in calcium carbonate saturation. This has impacts on
marine organisms that are calcifiers and essentially requires marine calcifying organisms to use more
energy to form biogenic calcium carbonate (Guinotte and Fabry, 2008). The observable consequences
are thought to be hampered reef formation of corals, algaes and hampered shell formation of oysters,
clams and crabs (although there are varying consequences on species depending on studies as shown by
Dr. Cooley).

There has been little work assessing the economic consequences of ocean acidification. The one
notable paper is that of Cooley and Doney (2009). In this paper the authors calculated potential
revenue losses for the U.S.A. from decreased mollusk harvests. If reductions of 6%-25% from 2007 level
of harvests were to occur in 2009, the authors calculate $75-187 million in direct revenue would be lost
each year into the future, with a net NPV loss of $1.7-10 billion through 2060. However it needs to be
noted that these values were calculated using what are commonly termed as replacement cost or
engineering cost estimates. From an economic viewpoint, there is no direct connection between
replacement costs and a useful welfare measure.

From an economic viewpoint, if ocean acidification affects the provisioning of ecosystem services, it can
result in lost consumer surplus (which are the opportunity costs to consumers). Consumer surplus is the
benefit to consumers of a market outcome and accrue whenever consumers pay less than their
maximum willingness to pay for that unit of a good.

Market prices simply capture the relative rate at which the market is willing to exchange one good for
another. The method employed by Cooley and Doney (2009) is the product of market price and a change
in quantity, or engineering cost estimates. If the reduction in mollusk harvests are given by the
difference in harvests from Qo to Qi as shown in Figure 1 evaluated at the constant price P0:

-------
Price

per

unit

b	a

Qi	Qo	Quantity

Figure 1 Replacement cost estimates

The lost revenues from ocean acidification are calculated (area Qi Qoab, shaded area in blue). Values
calculated in this manner tend to be rejected as they have no relationship to the economically relevant
surplus measures. Figure 2 illustrates the lost consumer surplus (area P0Pica, shaded area in red)
associated with the same reduction in harvests if price increases from P0 to Pi with the harvest
reduction Qo to Qi:

Price

per

unit

Figure 2 Consumer surplus estimates

-------
As Figure 2 illustrates there is no direction relationship between the replacement cost estimate and the
loss in consumer surplus. The replacement cost estimates do not measure or even approximate
economic welfare (see Bockstael et al. 2000). In addition, they omit key interactions within the
economy and between the economy and nature (Finnoff & Tschirhart 2008). However, applying an
economic approach can be a challenge because it requires measuring these surplus measures, which
requires more information than just market prices and quantities.

To apply an economic approach to the problem, it helps to consider the problem as one of a class of One
of a class of "Materials Damages" problems studied in detail by Tom Crocker 25 years ago (see a review
of the research effort for the EPA report archived at

http://yosemite.epa.gov/ee/epa/eerm. nsf/vwAN/EE-0043.pdf/$file/EE-0043.pdf). In this work Crocker
and his colleagues made the salient point that human welfare is dependent on biological systems
(material environment) that provide critical inputs to human activity. If there are damages or
improvements in material environment then there will be welfare changes.

Adams and Crocker (1991) laid out three basic steps to assess materials damage from environmental
changes. The first step is to provide an understanding of how the environmental change perturbs
production and consumption opportunity sets. The second was then to determine the input and output
market prices changes in response to the perturbations in opportunity sets. The third was to the
document all the adaptations humans can engage in to minimize losses or maximize gains from changes
in opportunities and prices.

In general, changes in production opportunities from perturbations in provisioning of ecosystem
services (ES) change producers production possibilities by the availability and combinations of ES input
sets (i.e. species compositions and densities). In turn this also affects output sets as there may be fewer
of some economically relevant species and potentially more of others. If the environmental degradation
reduces production possibilities then there will be less choice, higher costs and lower profits. Regardless
Adams and Crocker (1991) point out that human objective functions and behavioral conditions remain
the same in that firms still choose cost minimizing input combinations.

Similarly in consumption, perturbations in provisioning of ES may change costs facing households
directly or indirectly with corresponding welfare consequences. Again the underlying economic problem
remains the same with households choosing utility maximizing combinations of goods and services given
their income given the perturbations in provisioning of ES.

-------
The implication is that standard economic models can be used if the environmental perturbations can
be reliably brought into economic analysis. This is a primary challenge facing research in this area. To
bring the environmental changes into economic analysis there is a basic choice in the representation of
the natural system. On the one hand the assessment could employ a reduced form representation of
the natural system, reducing the entire natural system into one or two indicators (i.e. species). These
approaches are commonly seen in the bioeconomic literature (see Massey et al 2006, Smith 2007). They
are easy to fit to limited data and are typically thought to give a good overview of general processes.
However, it has been shown that aggregation (into a reduced form) can cause errors in economic
estimates (Kopp and Smith, 1980). On the other hand the natural system can be represented by a
detailed, or structural model (see Finnoff and Tschirhart 2008). Structural representations can represent
critical details explicitly and capture the complex adaptive nature of natural systems. However, it has
been shown that there are rapidly declining marginal returns to the inclusion of additional natural
science information (Adams, Crocker and Katz, 1984). The question then becomes what is the
appropriate balance of reality and tractability in the analysis?

One organizing principle that has roots in Tom Crocker's work is the potential for non-convexities in
natural system phenomena (see for example Crocker and Forester, 1981 and Brown et al. 2010). If the
natural system is reasonably convex, then environmental perturbations will have monotonic effects that
can be well represented with a reduced from representation. But if there are pervasive non-convexities
then a high level of abstraction may lead to trouble and it may well be necessary for the assessor to
know the entire possibilities surface.

The point is rather obvious if one considers the standard way an economist might consider correcting a
materials damage problem (to correct the problem one has to understand the welfare consequences
making an economic assessment one part of a corrective policy). Figure 3 illustrates a hypothetical
setting relating (loosely) to the problem of ocean acidification and a simple adaptation of Crocker and
Forester (1981). In the top panel, marginal control costs and marginal damages of acidification are
presented as downward and upward sloping functions of pH (acidity increases to the right of the
horizontal axis and decreases to the left). Economic theory would dictate that as there are costs of
control and damages that there is a single point of balance between the two marginal effects - a point
at which the net benefits to society of a plan of action are maximized (bottom panel). To find the
optimal point all one needs is information on marginal damages and marginal control costs to determine
how to maximize social net benefits.

-------
$

Figure 3. Optimal acidification in the standard setting

However, in many cases (see Crocker and Forester 1981) marginal damages or marginal control costs
may not be monotonically related to the environmental state. Figure 4 demonstrates the case Crocker
and Forester found for terrestrial acid deposition. Here, there are serious non convexities in marginal
damages. The implications are then that there is the possibility for multiple equilibria and having to
differentiate between local and global optimal. For example, as shown in Figure 4, without a knowledge
of the entire damage and cost functions would the researcher be able to determine which of the
equilibrium points A, B, or C would be globally optimal. In addition, unlike the standard setting, how
exactly natural and economic adjustments are to be made to bring the system into equilibrium are not
as clear. For example, in the region between A and B the marginal damages of acidification exceed the
marginal control costs, signally that a reduction in pH is optimal, directing the situation towards point A.
However, to the right of point B the reverse is true, signally that an increase in pH is optimal. This would
direct the situation towards point B which would only be appropriate if it were a global maximum. If
only a local max this would be problematic (to say nothing of the highly acidic end state). It appears that
an expansion of the scope of analysis is necessary as marginal comparisons alone (of marginal damages
to marginal control costs) are insufficient to signal how to maximize social net benefits. In these settings
it is likely necessary to know the entire surface (across environmental change) to locate the global
optimum and understand the signals provided by marginal measures.

-------
Figure 4. Acidification with non-convexity

Of course then the question becomes is there the potential for non-convexities with ocean acidification?

Using an extension of a Bering Sea ecosystem model developed in Finnoff and Tschirhart (2008) in work
for the EPA and National marine fisheries service (illustrated by Figure 5) the consequences of ocean
acidification were simulated in a very ad-hoc fashion. Under the assumption that acidification only
influenced the commercially important crab stocks, the ad hoc assumption was made in the model that
acidification increases variable respiration requirements of crabs for any level of biomass consumption.
The process could be expected to directly affect more species but the point is just to illustrate the
potential ecosystem consequences.

Using 3 arbitrarily chosen severities (1 being the most severe and 3 the least) and assuming that the full
effect would take time to unfold the model was used to generate multi-species growth functions for
ecosystem species in the presence of acidification. Figure 6 presents the growth functions generated
for three commercially important species, crabs, pacific cod and arrow tooth flounder under a

-------
benchmark of no acidification, low acidification, moderate acidification and high acidification. The
growth functions simply document the "surplus" production available or growth (vertical axis) at any
level of stock (horizontal axis) that could be appropriated by humans and the system remain in
equilibrium (a multispecies interpretation of bioeconomic yields)

blue whale

Illustrative Example:
Bering Sea Food Web

arrowtooth flounder



he rring

zooplankton

Figure 5

What is striking about Figure 6 is that for crabs alone there are non-monotonic changes from ocean
acidification. For the low to moderate levels of acidification (levels 2 and 3) the multispecies carrying
capacity of crabs (where the growth curves cut the horizontal axes) increases. In the absence of human
harvests crab populations might increase at these low levels of acidification! This is due to the food web
repercussions of acidification which see differential effects on predators (cod) and prey (bethos) which
reverberate throughout the ecosystem. High levels of acidification (level 1) here would lead to
extinction of crabs.

-------
For other commercially exploited species that are directly related through a direct predator prey
relationship, such as cod, a low level of acidification finds the carrying capacity only slightly altered but
there are significant declines at moderate and high levels (where the moderate and high lines overlay
one another). Arrowtooth flounder (ATF) are also commercially exploited yet are more distantly related
in the food web. They only experience minor effects on their carrying capacity across the levels of
acidification. However, for each of these commercially exploited species there are significant declines in
surplus growth (sustainably harvestable biomass).

—Benchmark —Acidifcation 1 —Acidification 2 —Acidification 3

Figure 6 Selected growth curves for commercially exploited species

There are also effects on charismatic mammals that could be expected to have significant non-market
values (Finnoff and Tschirhart, 2008) yet are only indirectly related to crabs in the ecosystem. Figure 7
presents growth curves for stellar sea lions (SSL) and sperm whales (SW). Sperm whales are more
directly related to the effects on crabs than sea lions yet both have effects on their carrying capacities
and growth (the moderate and high acidification curves overlay one another).

1.00

0.80

| 0.60
o

3

o 0.40

0.20

0.00

0.00	1.00	2.00	3.00

Cod stock

—Benchmark —Acidifcation 1 —Acidification 2

4.00	5.00

—Acidification 3

-------
Figure 7 Selected growth curves for charismatic mammals

In sum, the consequences from acidification reverberate across system in varying degrees and
magnitudes. There definitely seems to be the potential for non-convexities. As shown in the above
figures, the negative shock of acidification on the crab optimization problem can result in higher carrying
capacities yet less surplus growth. The changes are not typically monotonic. The implications for
bioeconomic harvests of fish and crab is that they will likely be affected in varying degrees and
magnitudes depending on their location in the food web. There are also perturbations in non-harvested
stocks in varying degrees depending on their location in the foodweb.

Regardless of the accuracy of these results, they point to the complexity in assessing the changes in
opportunity sets posed by acidification. To assess these or similar consequences an evaluation
mechanism would need to be able to assess changes in flows (harvests of commercially exploited
species) and stocks (changes in charismatic mammals) simultaneously. There is much the same reality
versus tractability debate in the assessment mechanism as in the inclusion of ecological detail.

One organizing lens is whether a reduced form (partial equilibrium) representation is sufficient for
accurate assessment or whether a structural form (general equilibrium) representation is required.
Partial equilibrium approaches are the bioeconomic standard (for example see Smith, 2007) for small
scale policies and welfare changes, while general equilibrium approaches are the public finance standard
(for example see Carbone and Smith, 2008) for larger scale policies and welfare changes.

Partial equilibrium approaches are typically easy to implement as they hold all other economic activity
constant (taking other prices and incomes as exogenous). They allow an uncluttered view of the
economic activity directly affected by the acidification and a clear representation of optimal planning

-------
over long time horizons through the effect of environmental dynamics on choices. In addition they
typically require few parameters. However they only provide a narrow viewpoint, they omit all other
human adaptation and often omit a connection to welfare economics.

In contrast a general equilibrium representation allows the adaptations in the economic system to be
represented. Prices and incomes are endogenous, there is an inclusion of producer and consumer
behavior throughout an economic and allow a clear link to the principles of welfare economics.
However these methods require numerous parameters, they are exceedingly hard to dynamically
optimize, their broad viewpoint makes decomposing welfare effects impossible and can obscure the
influence of environmental dynamics by economic responses.

Both methodologies have pros and cons, the question boiling down to a determination of the the
appropriate balance. For the problem of ocean acidification this would tend to depends on the setting.
For example, when considering the consequences on aquaculture a partial equilibrium approach may
suffice, especially if the consequences are confined to the near shore and few other exploited (or non-
market) populations. Regardless the lack of scientific research into this issue from an economic
viewpoint is glaring. To say much more requires some hard scientific effort.

In conclusion, the point of my talk is that welfare measurement of materials damages has some well
known characteristics but for this problem a lot remains unresolved and work remains. There is a high
likelihood in my opinion that generating accurate assessments will be tricky and generalities seem to be
lacking. A necessary first step is a a clear understanding of how production and consumption
possibilities are affected by the problem in a consistent setting. While dose response relationships of
environmental change from the natural sciences are key, but how much detail is necessary for a good
understanding remains to be resolved in this context.

The implications from this brief review are obvious. If problems are convex or well behaved then
aggregate representations of the natural science may be sufficient for good economic assessments. But
if these problems have pervasive non-convexities then policy makers must expand the scope of their
analysis for good economic assessments. Marginal assessments on their own may lead to trouble.

-------
References:

Adams, R.M. andT.D. Crocker. 1991. Materials Damages. Chapter IX. In: Measuring the Demand for

Environmental Quality. C.D. Kolstad and J. Braden, eds. North-Holland Publishers, pp. 271-

302.

Adams, R.M., T.D. Crocker and R.W. Katz. 1984. Assessing the Adequacy of Natural Science Information:
A Bayesian Approach. The Review of Economics and Statistics. 66(4):568-75.

Bockstael NE, Freeman AM III, Kopp RJ, Portney PR, Smith VK. 2000. On measuring economic values for
nature. Environ. Sci. Technol. 34:1384-89.

Brown, G., T. Patterson and N. Cain. 2010. The devil in the details: Non-convexities in ecosystem service
provision. Resource and Energy Economics, forthcoming.

Carbone, J. and V.K. Smith. 2008. Evaluating Policy Interventions with General Equilibrium Externalities.
Journal of Public Economics, 92(5-6): 1254-1274.

Cooley, S.R. and S.C. Doney. 2009. Anticipating ocean acidification's economic consequences for
commercial fisheries. Environ. Res. Lett, 4: 024007 (8pp).

Crocker, T. D., and B. A. Forster. 1981. Decision Problems in the Control of Acid Precipitation:
Nonconvexities and Irreversibilities. J. Air Pollution Control Assoc. 31:31-37.

Finnoff D, Tschirhart J. 2008. Linking dynamic economic and ecological general equilibrium models.
Resour. Energy Econ. 30:91-114.

Guinotte, J. and V. Fabry. 2008. Ocean acidification and its potential effects on marine ecosystems.
Annals of the New York Academy of Sciences, 1134: 320-342.

Kopp, R.J. and V.K. Smith. 1980. Measuring Factor Substitution with Neoclassical Models: An
Experimental Evaluation. Bell Journal of Economics, 11(2): 631-655.

Massey M, Newbold SC, Gentner B. 2006. Valuing water quality changes using a bioeconomic model of a
coastal recreational fishery. Journal of Environmental Economics and Management, 52:482-500.

Smith, M.D. 2007. Generating Value in Habitat-dependent Fisheries: The Importance of Fishery
Management Institutions. Land Economics, 83(1): 59-73.

-------
Economic Impact of Climate
Change and Ocean Acidification

on Fisheries

David Finnoff

University of Wyoming

U.S. EPA/DOE Workshop, ''Research on Climate Change
Impacts and Associated Economic Damages", January 27-28,

2011, Washington D.C.

-------
Ocean Acidification

•	Past declines in ocean surface pH linked to
mass extinction events

•	Reduction in carbonate ion concentration

—	> reduction in calcium carbonate saturation

—	> impacts on marine calcifiers

—	>requires marine calcifying organisms to use more
energy to form biogenic calcium carbonate

•	Hampered reef formation of corals, algaes

•	Hampered shell formation of oysters, clams and crabs

-------
Economic consequences: S.R. Cooley and S.C. Doney. 2009. "Anticipating ocean
acidification's economic consequences for commercial fisheries/' Environ. Res. Lett.

Price

per

unit

•Calculated potential revenue losses from decreased mollusk harvests of 6%-25%
from 2007 level were to occur in 2009, $75-187 million in direct revenue would be
lost each year into the future, with a net NPV loss of $1.7-10 billion through 2060

•No direct connection between replacement costs and a useful welfare measure

-------
Economic Consequences

•	May disrupt provisioning of ecosystem services (ES)

•	One of a class of "Materials Damages" problems
studied in detail by Tom Crocker 25 years ago

•	Human welfare is dependent on biological systems
(material environment) that provide critical inputs to
human activity

•	Damages or improvements in material environment
implies welfare changes

-------
Assessment of Materials Damage Requires:

1.	Characterization of the differential changes
across time and space that environmental
change causes in production and consumption
opportunities

2.	Determination of the responses of input and
output market prices to these changes

3.	Identification of the adaptations that affected
agents can make to minimize losses or maximize
gains from changes in opportunities and prices

-------
Economic Effects

•	Perturbations in provisioning of ES change
producers production possibilities

•	Degradation may reduce production possibilities

•	Perturbations in provisioning of ES may change
costs facing households directly or indirectly
(access costs)

•	Objective functions and behavior towards new
sets of production and consumption possibilities
remains the same

•	First key question: how do to bring these changes
in possibilities into economic analysis?

-------
Bringing Environmental Changes into
the Economic Assessment

Reduced Form Representation:

•easy to fit to limited data

•gives good view of general processes

BUT

•aggregation can cause errors in
economic estimates
(Kopp and Smith, 1980 BJE)

Structural Representation:

•can represent critical details explicitly
•capture within system adaptations
BUT

•contribution of additional natural science

information declines rapidly

(Adams, Crocker and Katz, 1984 RESTAT)

Appropriate Balance?

• in one dimension balance depends on potential of non-convexities
•If problems are convex reduced form representation likely sufficient
•If there are pervasive non-convexities high level of abstraction may lead
to trouble - need to know the entire possibilities surface

-------
Marginal damage

PH

Greater acidity i>

Greater acidity ->

Standard setting

• single equilibrium

•marginal comparisons
alone sufficient to signal
how to max social net
benefits

Net benefits

0

-------
Marginal control costs

reater acidity i>

Marginal damage

Greater acidity ->

$

Non-convexities

0

$

Net be nefits

0

•	multiple equilibria

•	natural and economic
adjustments not as clear -
requires an expansion of
the scope of analysis

•marginal comparisons
alone insufficient to signal
how to max social net
benefits

•need know the entire
surface (across
environmental change) to
locate global optimum
and understand how the
marginal damages change

-------
Illustrative Example:
Bering Sea Food Web

blue whale

benthos

III

detritus

-------
GEEM: Developed to track ecological adjustment

1)	fitness net energy

i-1	m	i-1	i-l

Ri =YXej ~e^xu ~ £ eX}+t,eJy,i<^x!J) ~ /'(Sxy) ~ Pi

7=1	k—i+l	7=1	7=1

2)	biomass transfers (similar to market clearing)

rijXjjie) < nj/pcf?))

3) population updating



? MSS

Si vi

-------
Acidification

Ad Hoc: Acidification increases variable respiration requirements
of crabs for any level of biomass consumption

fcrab(•)	^crab^crab.benth

A ,=

a

i

bi+e

-At

i = $,3

J

A.

A.

A:

0 20 40 60 80 100

-------
—Benchmark —Acidifcation 1 —Acidification 2 —Acidification 3

Cod stock

—Benchmark —Acidifcation 1 —Acidification 2 —Acidification 3

0	50	100	150	200	250	300	350

ATF Stock

—Benchmark —Acidification 1 —Acidification 2 —Acidification 3

-------
0.020

0.015

.c

1

2	o.oio

13

_i

co
co

0.005

0.000

0.00 0.02 0.04
—Benchmark —Acidification 1

0.06 0.08 0.10 0.12
SSL stock

—Acidification 2 Acidification 3

0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80

SW stock

—Benchmark —Acidification 1 —Acidification 2 —Acidification 3

-------
Implications:

•	Consequences reverberate across system in
varying degrees and magnitudes

•	Seems to be a potential for non-convexities

—	Acidification a negative shock to crab optimization
problem yet can see higher stocks (although less
surplus growth)

—	Changes not always monotonic

—	Problems with reduced form aggregations

-------
Representation of Environmental Changes

Reduced Form
Representation

v's

Structural
Representation

7.00

-1.00

	Benchmark

	Acidification 3

Crab stock

—Acidifcation 1	—Acidification 2

Acid 1 single species

5.00

-------
Point:

•	Bioeconomic harvests offish and crab likely
affected to varying degrees and magnitudes
depending on location in food web

•	Non-harvested stocks may or may not have
cascading effects depending on location in
foodweb

•	To assess tradeoffs have to be able to access
changes in flows and stocks simultaneously

-------
Evaluating Environmental Changes: Do
changes in relative prices matter?

Reduced Form / Partial

Structural /General Equilibrium

Equilibrium Representation:

Representation:

• other prices and incomes exogenous

• prices and incomes endogenous

• allows clear representation of optimal

• system wide adaptation

planning over long time horizons

• clear link to principles of welfare

• allows clear focus on effect of

economics and inclusion of producer and

environmental dynamics on choices

consumer behavior

• requires few parameters

• Enough detail to include specific scientific

BUT

information

• narrow viewpoint, omits all other

BUT

adaptation

• requires numerous parameters

• typically omits a connection to welfare

• hard to dynamically optimize

economics

• broad viewpoint makes decomposing

• not clear how specific scientific

effects tricky

information be included into lumped

• influence of environmental dynamics

parameters

obscured by economic responses

What is the Appropriate Balance????

-------
Conclusions

Point: Welfare measurement of materials damages has some well
known characteristics but for this problem a lot remains
unresolved and work remains

1.	Accurate assessments tricky, generalities seem to be lacking

2.	Need a clear understanding of how production and consumption
possibilities are affected by the problem in a consistent setting

- dose response relationships of environmental change from the
natural sciences are key, but how much detail is necessary
for a good understanding remains to be resolved

3.	If problems are convex or well behaved then aggregate
representations of the natural science may be sufficient for good
economic assessments

4.	If problems have pervasive non-convexities then policy makers
must expand the scope of their analysis for good economic
assessments - marginal assessments on their own may lead to
trouble (G. Brown, REE, forthcoming)

-------
Commodity
Markets

(Z^^)-/	

^	 *	< J	X

~'	" "V^			 ' ' N

^	;	>'V

North, fur seal )	,•	' X... V /"¦	

;/

L

Households

; f* - ¦ + y- ,-"i





5$
o



Ci

o
o
<3

U



u K§ ^	7 \ \ ^— 	' \ /!•

^ ^	' (" Small flatfish -X; \ " A k	1 / . / :
	:	 L-:	W L—1 7 ^	 			Nv \ *•» a A \ | ' \ ~ 1 :

¦ | A i f \ 7 j,"" _ — — "V"" """"	^ v \ \\ '	' * / 1/

I	j\	\"(f"~G^T^)/ \ \ 'Cjvenile polbdT^/) ./[

II	;i x,	1

	~1 !	\ 	*""* / (^sTa urchiT^)

Regulator •	*****•.. ( Benthos 	v"/'"			

	2		:			^					 A 7nnnhnknn ^	a

Factor
Markets

^ bentnos ) 	v ^	

	 ^nsBS-s^	 \C Zooplankton J

\	A 		

^TT~-n' \ t

(^^etritus^^

(^Phytoplankton^) (^^Kelp

Jl it

Valuing Ecosystem Services - Bering Sea

Monetary flows 	~

Real good flows 	~

Biomass flows in pred-prey relations	~

Cultural ecosystem service	. —>

Provisioning ecosystem service

Ecosystem externality	

Sun

-------
Output





Consu

mers

Valuing Ecosystem Services - Neuse River

Agriculture

Blue Crab
Harvests

Shallow Channel (s)

Blue Crabs (7)

Mobile
Species

-------
U.S. EPA/DOE Workshop, "Research on Climate Change Impacts and Associated
Economic Damages," January 21-IS, 2011, Washington D.C.

Abstract: Nonmarket valuation of climate change and ocean acidification impacts to marine
resources

John Whitehead
Department of Economics
Appalachian State University

December 22, 2010

Introduction

The purpose of this abstract is to describe existing methods of estimating the economic
values for avoiding the climate change impacts to marine resources. In the first section I describe
the available methods. In the next section I review the literature focused on recreation values
associated with climate change. In the third section I consider a conceptual model for valuing
climate change impacts to marine resources. In the fourth section I consider future research
needs.

Methods

Estimating the nonmarket values of climate change impacts to marine resources first
requires consideration of the type of impacts. Market values are the changes in outputs and
inputs associated with a resource reallocation and are valued with market prices. Nonmarket
values are those that accrue above and beyond market values and are variously called consumer
surplus, compensating surplus, equivalent surplus, willingness to pay and willingness to accept.
The total economic value is the sum of all nonmarket values.

Estimation of the total economic value for marine resources is complex. Consider coral
reefs which can provide recreation and tourism values, amenity values, fishery habitat values and
biodiversity values (Figure 1). The main categories of nonmarket values include direct use
values, indirect use values and nonuse values. Direct use values are those that arise from on-site
enjoyment of a natural resource. Direct use values that are generated by marine resources are
primarily recreational and tourism values. In Figure 1, individuals can enjoy recreational diving
on the coral reef ecosystem and gain direct use values. Indirect use values are those that are
enjoyed on-site as a by-product of coral reefs. For example, fish stocks are enhanced by coral
reef protection and anglers enjoy coral reef protection indirectly through improved catch rates.
Nonuse values are those values that arise without on-site enjoyment. Nonuse values may be
motivated by altruism, bequests or an environmental ethic.

Both revealed and stated preference methods can be used to estimate direct and indirect
nonmarket use values. The most advantageous revealed preference nonmarket valuation method
for outdoor recreational modeling is the travel cost method. The travel cost method exploits the
empirical relationship between outdoor recreation trips and site selection and the travel cost

1

-------
required to reach recreation sites. The most basic finding is that the further the distance the less
likely the recreation site will be selected and the fewer the number of trips.

Stated preference methods include the contingent behavior, contingent valuation and
attribute-based choice experiment (i.e., conjoint analysis) methods. The contingent valuation
method could be used by asking survey respondents for their willingness to pay to prevent
climate change to recreation resources. The contingent behavior method could be used by asking
survey respondents for hypothetical changes in visitation behavior (i.e., trips) with changes in
climate related variables. Attribute-based choice experiments can be used by asking survey
respondents about changes in visitation behavior (i.e., site selection) with changes in climate
related variables.

Both revealed and stated preference methods have limitations when valuing the impacts
of long term climate change. Revealed preference methods are constrained by current spatial
variations in temperature and other measures of climate change impacts. Forecasts of the impacts
of temperature change beyond current experience are possible but the range and types of
behavior change are constrained by the model and existing behaviors. Stated preference methods
are limited in that the measured behavior is hypothetical and subject to potential biases. One
approach for resolving these weaknesses is the combination and joint estimation of revealed and
stated preference data. Joint estimation allows the behavior change to range beyond historical
experience with the stated preference data while grounding the hypothetical data in revealed
preferences.

Stated preference methods must be used to estimate nonuse values. The contingent
valuation method can be used to ask survey respondents about their willingness to pay for
climate change policy that would change the characteristics of marine resources. One problem
with the contingent valuation method in this context is that it is most effectively employed to
estimate total economic values. Willingness to pay for climate change policy could also capture
marine resource values, coastal values, terrestrial values and others. Attribute-based choice
experiments can also be used to estimate nonuse values. Respondents are typically led through a
series of policy choices with varying characteristics of the policy. In the case of coral reef
valuation, these characteristics could include changes in the ecosystem, fish stocks and other
impacts with and without opportunities for recreation. Simulation methods can then be used to
estimate nonuse values.

Literature on Outdoor Recreation and Climate Change

Past research on the impact of climate change on outdoor recreational activities is
relatively sparse. Early studies found that precipitation and temperature affects beach recreation
activities (McConnell 1977, Silberman and Klock 1988). Mendelsohn and Markowsi (1999)
considered the effects of changes in temperature and precipitation on a wide range of outdoor
recreational activities using state-level aggregate demand functions. Considering a range of
climate scenarios, the authors found that increased temperature and precipitation increase the
aggregate economic value of some activities and decreases the aggregate economic value of
others. Loomis and Crespi (1999) took an approach similar to Mendelsohn and Markowsi (1999)
but used microdata. They considered the effects of temperature, precipitation and other climate

2

-------
change impacts (e.g., beach length, wetland acres) on a wide range of outdoor recreational
activities. Overall, they found that climate change is likely to have positive impacts on the
aggregate economic value of outdoor recreation activities.

Several studies have focused on more narrow regions and outdoor recreational activities.
Pendleton and Mendelsohn (1998) related the effects of temperature and precipitation to catch
rates for trout and pan fish in the northeastern United States. Climate change is expected to
decrease trout catch rates and increase pan fish catch rates. Using microdata, the authors found
that fish catch rates influence fishing site location choice. Combining the effects of climate
change on catch rates the authors found that climate change would benefit freshwater fishing in
the northeastern United States. Ahn et al. (2000) focused on trout fishing in the Southern
Appalachian Mountain region of North Carolina. Using methods similar to Pendleton and
Mendelsohn (1998) the authors found contrasting results. Based on their results climate change
would reduce the economic value of trout fishing in this region. The contrast may be due to a
lack of species-substitution possibilities. More recently, Englin and Moeltner (2004) estimated
weekly skiing and snowboarding trip demand models and integrate weekly weather conditions as
a factor affecting demand. They find that temperature and precipitation affect the number of
skiing and snowboarding days in expected ways.

All of the previous studies used revealed preference methods. In contrast, Richardson and
Loomis (2004) employed a stated preference approach to estimate the impacts of climate change
on economic value for recreation at Rocky Mountain National Park. Richardson and Loomis'
hypothetical scenario explicitly considered the direct effects of climate, temperature and
precipitation, and the indirect effects of temperature and precipitation on other environmental
factors such as vegetation composition and wildlife populations. They found that climate change
would have positive impacts on visitation at Rocky Mountain National Park.

A Conceptual Model

There are a number of relationships that need to be modeled to estimate a marine rsources
damage function (Figure 2).1 First, a simple model of the effect of carbon dioxide emissions on
ocean acidification is needed. The simple model should be able to abstract away from the
biophysical complexities and allow focus on the endpoints that are important for anthropogenic
valuation. For example, a description of how carbon dioxide emissions affect seawater variables
and other weather-related variables important to recreation (e.g., ambient temperature,
precipitation) is needed. Considering the example of coral reef ecosystems, let this relationship
be expressed as equations (1) and (2):

(1)	5 = f(C02)

(2)	W = f(CO2)

where S represents seawater variables (e.g., temperature, chemistry), Wrepresents weather-
related climate change variables (e.g., ambient temperature, precipitation) and C02 represents
carbon dioxide. In Figure 2 this relationship is represented by the arrows labeled (1) and (2).

1 Note that this model is what is understood by an economist with no training in climate science.

3

-------
Next, a biophysical description of the effect of seawater and other climate variables on
coral reef ecosystems and fish stocks (e.g., range shifts, habitat loss, and prey availability) is
needed.

(3)	CR=f{S,W)

(4)	FS = f(S, W, CR)

where CR is coral reef ecosystem and FS is fish stocks. In Figure 2 these relationships are
represented by the arrows labeled (3) and (4). Note that fish stocks are affected by seawater
variables and other climate variables directly and indirectly through coral reef ecosystems.

Next, behavioral models could be estimated with revealed preference methods such as the
travel cost method:

(5)	RD=f(p,y,CR)

(6)	RF = f(p,y,FS)

where RD is recreational diving, RF is recreational fishing, p is the access cost of each activity
(e.g., travel cost) and .y is income. In Figure 2 these are illustrated by the arrows labeled (5) and
(6). The link between carbon dioxide emissions and recreational behavior can be found by
substituting equations (1) and (2) into (3) and substituting equations (1), (2) and (3) into equation
(4). Then equation (3) would be substituted into equation (5) and equation (4) would be
substituted into equation (6). The reduced form behavioral models are:

(5 ')RD =f(p,y,S,W,CR{S,W))

(6') RF = f(p, y, S, W, FS(S, W, CR))

To estimate recreational impacts from climate on marine recreational behavior in a
revealed preference study, one would follow the methods employed in previous studies.
Considering the conceptual framework developed by Shaw and Loomis (2008), one would
estimate the relationship between the direct effects of climate change (e.g., temperature,
precipitation, climate variability), the indirect effects (e.g., fish stocks and composition) and the
effects on outdoor recreational behavior and economic value. Data with spatial variation in the
climate change variables is required.

In particular, consider the random utility model version of the revealed preference travel
cost method. In this model, it is assumed that individuals choose recreation sites based on
tradeoffs among trip costs and site characteristics (e.g., temperature, precipitation, catch rates). If
anglers make fishing site selections based on these characteristics, then the existing relationship
between site characteristics and fishing site selection can be used to simulate the impact of
climate change. This model could then be linked to models of visitation frequency to estimate the
aggregate impacts of climate change on marine recreation behavior. Stated preference recreation
scenarios can be designed to elicit hypothetical behavior data to supplement the revealed
preference data in a joint estimation framework.

4

-------
The simple biophysical descriptions represented by equations (1), (2), (3) and (4) could
be used to design stated preference recreation scenarios for estimation of use values and policy
scenarios for estimation of nonuse values. First, nonuse values must be conceptually defined.
Economists use the utility function to conceptualize the relationship between consumption and
welfare (i.e., happiness). Considering the example above:

(7)	U = U(X, RD, RF, CR, FS, UaUb)

where U is the utility of an individual, Xrepresents market goods, If represents the utility of
individual a (i.e., altruism) and if represents the utility of individual b (i.e., bequests to future
generations). Changes in RD and RF that affect utility represent behavior that generates use
values. Changes in CR and FS generate nonuse values motivated by an environmental ethic.
Changes in If and if generate nonuse values motived by altruism and bequests to future
generations. These relationships are represented by the arrows labeled (7) in Figure 2.

Substitution of equation (7) for individuals a and b leads to a reduced form utility
function which can be maximized subject to a budget constraint to find the indirect utility
function:

(8)	v = v(p,y, CR, FS)

Use and nonuse values can be conceptually defined using equation (8). The total economic value
of a change in coral reef ecosystems from the baseline level to a degraded state (CR, FS') is:

(9)	v(p,y - TEV, CR, FS) = v(p,y, CR', FS')

where TEV is the total economic value, the amount of income that must be taken from the
individual in order to maintain utility at a level equal to that with full income but a degraded
environment. Total economic value is the sum of use value and nonuse value, TEV = UV +
NUV, where nonuse value is:

(10)	v(p*,y — NUV, CR, FS) = v(p*,y,CR',FS')

where p* is the price at which the quantity of recreation demanded is equal to zero. The residual
difference between TEV and NUV is equal to the sum of the use values from equations (5') and
(6'). Contingent valuation or attribute-based choice experiment scenarios can be described to
convey the information included in equations (1) - (10) and obtain estimates of total economic
values and nonuse values for the impacts of climate change on marine resources. Joint estimation
with recreation demand functions (5') and (6') can be used to decompose total economic value
into use value and nonuse value and further calibrate the model.

Future Research

Future research must be conducted to determine how the biophysical models could be
integrated with the economic models. To my knowledge there are no good examples in the

5

-------
literature.2 Important gaps and uncertainties in our knowledge regarding the economic impact of
changes in fisheries and coral reef ecosystems due to climate change are the lack of empirical
relationships described above. One of the next steps to improving how nonmarket impacts to
marine ecosystem service impacts are handled in an integrated assessment modeling framework
is to gather the necessary revealed preference and stated preference data and estimate the
relationships described above. The accuracy of transfers of these damage functions across
regions and over long periods of time is an open question, requiring validity studies. Research
examining these relationships would be most fruitful. In the interim, investigation of benefit
transfer methods with existing estimates of coral reef recreation and recreational fishing values
would allow preliminary estimation of these damage functions.

2 Note that I still have a stack of papers to read and have not yet exhausted my literature search abilities.

6

-------
Figure 1. Nonmarket Economic Values

Figure 2. Climate Change and
Nonmarket Economic Values

7

-------
References

Ahn, Soeun, Joseph E. DeSteiguer, Raymond B. Palmquist and Thomas P. Holmes, "Economic
Analysis of the Potential Impact of Climate Change on Recreational Trout Fishing in the
Southern Appalachian Mountains: An Application of a Nested Multinomial Logit Model,"
Climatic Change 45:493-509, 2000.

Englin, Jeffrey, and Klaus Moeltner, "The Value of Snowfall to Skiers and Boarders,"
Environmental and Resource Economics 29:123-136, 2004.

Loomis, John and John Crespi, "Estimated Effects of Climate Change on Selected Outdoor
Recreation Activities in the United States," Chapter 11 in The Impact of Climate Change on the
United States Economy, Robert Mendelsohn and James E. Neumann, eds., Cambridge University
Press, pp. 289-314, 1999.

McConnell, Kenneth E., "Congestion and Willingness to Pay: A Study of Beach Use," Land
Economics 53(2): 185-195, 1977.

Mendelsohn, Robert and Maria Markowsi, "The Impact of Climate Change on Outdoor
Recreation," Chapter 10 in The Impact of Climate Change on the United States Economy, Robert
Mendelsohn and James E. Neumann, eds., Cambridge University Press, pp. 267-288, 1999.

Mendelsohn, Robert and James E. Neumann, eds. The Impact of Climate Change on the United
States Economy, Cambridge University Press, pp. 267-288, 1999.

Pendleton, Linwood H., and Robert Mendelsohn, "Estimating the Economic Impact of Climate
Change on the Freshwater Sportsfisheries of the Northeastern U.S.," Land Economics 74(4):483-
496, 1998.

Richardson, Robert B., and John B. Loomis, "Adaptive Recreation Planning and Climate
Change: A Contingent Visitation Approach," Ecological Economics 50(83-99): 2004.

Shaw, W. Douglass, and John B. Loomis, "Frameworks for Analyzing the Economic Effects of
Climate Change on Outdoor Recreation," Climate Research 36:259-269, 2008.

Silberman, J. and Klock, M., "The Recreation Benefits of Beach Nourishment," Ocean and
Shoreline Management 11:73-90, 1988.

8

-------
Nonmarket Valuation of Climate
Change and Ocean Acidification
Impacts to Marine Resources

John C. Whitehead
Department of Economics
Appalachian State University
Boone, NC

2/16/2011

USEPA/DOE Workshop "Research on
Climate Change Impacts and Associated
Economic Damages"

1

-------
Nonmarket Values

2/16/2011

USEPA/DOE Workshop "Research ori
Climate Change Impacts and Associated
Economic Damages"

2

-------
Nonmarket Values for Coral Reefs

Coral Reef Ecosystem

2/16/2011

USEPA/DOE Workshop "Research on
Climate Change Impacts and Associated
Economic Damages"

-------
Use values

•	Willingness to pay to avoid climate change to
marine resources due to use of these
resources on-site

•	Direct use

—	Diving

—	Snorkeling

—	Viewing

•	Indirect use

—	Fishing (coral reef habitat and nursery functions)

2/16/2011

USEPA/DOE Workshop "Research on
Climate Change Impacts and Associated
Economic Damages"

4

-------
Nonuse (aka, passive use) values

•	Willingness to pay to avoid climate change to
marine resources without the intent to use
these resources on-site

•	Motives

-	Altruism (WTP today for Aq today)

-	Ecological ethic (WTP today for Aq today)

-	Bequests (WTP today for Aq in the future)

2/16/2011

USEPA/DOE Workshop "Research on
Climate Change Impacts and Associated
Economic Damages"

5

-------
Measurement of Total Economic Value

Types of Value

Valuation





Methods

Use

Nonuse

Revealed Preference

Yes

No

Stated Preference

Yes

Yes

2/16/2011

USEPA/DOE Workshop "Research on
Climate Change Impacts and Associated
Economic Damages"

6

-------
Revealed Preference Methods

• Types

—	Hedonic price method

•	Property values

—	Averting behavior method

•	Health values

—	Travel cost method

•	Recreation values

-	Single site TCM

-	Multiple site RUM

—	NFI, PF, GR (generally not appropriate)

2/16/2011

USEPA/DOE Workshop "Research on
Climate Change Impacts and Associated
Economic Damages"

7

-------
Stated Preference Methods

• Types

-	Contingent valuation

•	Used to estimate UV, NUV and TEV

•	difficult to avoid double counting in the case of climate
change

•	WTP to climate change policy = bequest values

-	Choice experiments

•	Similar values as CVM Use to estimate UV, NUV and TEV

•	can be used to separate marine values from total values of
climate change policy

-	Contingent behavior

•	Used to estimate recreation and other UVs

2/16/2011

USEPA/DOE Workshop "Research on
Climate Change Impacts and Associated
Economic Damages"

8

-------
RP-SP Methods

•	Problems with both RP and SP Methods

•	Joint estimation of RP-SP data can mitigate
some of these problems

•	TCM/RUM with SP methods is used to
estimate use and nonuse values

2/16/2011

USEPA/DOE Workshop "Research on
Climate Change Impacts and Associated
Economic Damages"

9

-------
Climate Change and Nonmarket Values

Coral Reef
Ecosystem

(4)

Fish Stocks

Seawater
variables

Other
climate
variables

Nonuse
Motives

(7)

Utility

2/16/2011

USEPA/DOE Workshop "Research on
Climate Change Impacts and Associated
Economic Damages"

10

-------
Literature

•	RP: Spatial variation in climate variables

—	Mendelsohn and Markowsi, 1999

—	Loomis and Crespi, 1999

—	Ahn, et al., 2000

—	Pendleton and Mendelsohn, 1998

•	RP: Temporal variation in climate variables

—	Englin and Moeltner, 2004

—	Carter and Letson, 2009

•	SP: Richardson and Loomis, 2004

2/16/2011

USEPA/DOE Workshop "Research on
Climate Change Impacts and Associated
Economic Damages"

11

-------
A reduced form damage function

•	Data

-	NSRE (1990, 2000)

—	NSFHWAR (every 5 years)

•	Recreation Days = f(X; temp, precip, etc)

2/16/2011

USEPA/DOE Workshop "Research on
Climate Change Impacts and Associated
Economic Damages"

12

-------
Saltwater Fishing Participation

Linear probability model

Variable

Estimate

t-value

3F

7F

Intercept

1.9467

27.42





income

-0.0009

-9.45





white

-0.0379

-4.20





male

-0.1035

-15.42





age

0.0013

5.56





educ

0.0046

3.11





hhnum

-0.0087

-3.13





under6

-0.0002

-0.03





metro

-0.0315

-3.82





jantemp

-0.0022

-5.90

-0.00666

-0.01554

jultemp

0.0017

1.76

0.00504

0.01176

janpcp

-0.0063

-3.16





julpcp

-0.0194

-7.62











-0.00162

-0.00378

2/16/2011

USEPA/DOE Workshop "Research on
Climate Change Impacts and Associated
Economic Damages"

13

-------
Saltwater Fishing Days

Negative Binomial Intensity Model



Variable

Estimate

t-value

3F

7F

Intercept

2.4058

3.83





income

0.0002

0.22





white

-0.0609

-0.71





male

0.24

3.93





age

0.0031

1.35





educ

-0.0686

-4.94





hhnum

-0.068

-3.00





under6

0.0419

0.83





metro

0.0358

0.47





jantemp

-0.0044

-1.22

-0.0132

-0.0308

jultemp

0.0128

1.47

0.0384

0.0896

janpcp

-0.0265

-1.46





julpcp

0.117

5.74











0.0252

0.0588

USEPA/DOE Workshop "Research on
2/16/2011	Climate Change Impacts and Associated	14

Economic Damages"

-------
A more structural damage function

•	MRFSS data

—	temporal
variation

-	Spatial
variation

•	Climate change
would affect
species
composition
and potential
fishing days

41

71'-''—

NOAA FISHERIES:

Office of Science & Technology

Directorate | Fisheries Statistics | Assessment & Monitoring | Economics & Social Analysis | Science Information | Marine EcoSystems

Recreational Fisheries

Home

Marine Recreational
Information Program

Program Overview

Survey Materials

Querying the Data

Local Fish names

Downloads

Meetings

Contact Us

Related links

Sign-up FishNews

Downloads

Documents and data are available for download after the data are declared "final". When
downloads are separated by coasts, combined Atlantic and Gulf of Mexico data are
annotated with "AG", Gulf-only data with "GU", and Pacific data with "PC". Some downloads
may have been produced in the past, but are not currently available.

¦ SAS Data

Home | About Us | Regions | Science Centers | Councils | Commissions | Advisory Committee | Search
Contact Us | Forms | Privacy Policy | Information Quality Guidelines | Disclaimer | People Locator

2/16/2011

USEPA/DOE Workshop "Research ori
Climate Change Impacts and Associated
Economic Damages"

15

-------
Marine recreational fishing and

climate change

•	Household production model

-	HCKR = f(X; cs, ts)

-	Changes in season length

-	Changes in species composition

•	Participation / Site selection model

-	Y = f(TC,HCKR; cs, ts)

•	Estimate WTP with simulated changes of
climate change

2/16/2011

USEPA/DOE Workshop "Research on
Climate Change Impacts and Associated
Economic Damages"

16

-------
Conclusions

•	No study to date explicitly addresses nonmarket
valuation of climate change and marine resources

-	WTP review finds no mention of marine values

• Is it insignificant or missing?

•	Meta-analyses could be used in a benefit transfer
study

-	Coral reef recreation values

-	Outdoor recreation values

-	Recreational catch values

•	But, behavioral response to climate change is
missing

2/16/2011

USEPA/DOE Workshop "Research on
Climate Change Impacts and Associated
Economic Damages"

17

-------
Future Research

•	All sorts of studies are needed: RP, SP; TEV, UV, NUV

•	Most promising with existing RP data

—	Reduced form

—	More structural

•	New studies

-	SP data

•	CVM - difficult to avoid double counting

•	CE - can differentiate between marine and other values

•	CB - behavioral response to climate change

-	RP-SP joint estimation

•	Can differentiate between UV and NUV

2/16/2011

USEPA/DOE Workshop "Research on
Climate Change Impacts and Associated
Economic Damages"

18

-------
The impacts of climate change on terrestrial ecosystems

Karen Carney, Stratus Consulting Inc.

Improving the Assessment and Valuation of Climate Change Impacts for Policy and Regulatory

Analyses

Washington, DC

January 27-28, 2011

1. Background

Concerns over the impacts of rising global atmospheric greenhouse gas (GHG) concentrations
are growing. Although human welfare and well-being will be directly affected by changes in
climate, many important impacts to human welfare will occur indirectly due to climate-induced
changes in ecosystems. Terrestrial ecosystems provide critical goods and services to humans. For
example, they provide raw materials (e.g., food, water, and timber), regulate air quality,
assimilate waste, and provide recreational opportunities. Terrestrial ecosystems are also valuable
simply for existing; that is, society puts value on the existence of species and habitats and invests
resources to protect them. To fully understand the benefits and costs of alternative climate
policies, increases or declines in these critical services resulting from climate change must be
evaluated.

It is well understood that climate is a key determinant of the structure and function of
ecosystems. Climate affects which species are able to reside in a given location, how productive
an ecosystem is, the rates of ecosystem processes (e.g., nitrification, methane production), and
the nature, frequency, and intensity of natural disturbances (e.g., wildfires, pest outbreaks). Thus,
as climate changes, it will fundamentally, and potentially dramatically, affect the location and
character of ecosystems.

Which ecological impacts should be examined?

The Intergovernmental Climate Change reports, U.S. Climate Change Science Program synthesis
reports, and a variety of other synthetic reports issued by states, governments, and NGOs provide
a long list of potential impacts to terrestrial ecosystems. The impacts range across geographic
scales (i.e., some are sub-national, country-specific, or global) and across different levels of
biological organization (i.e., some address individual species, ecosystems, or global
biodiversity). Which of these myriad of impacts should be addressed in integrated assessment
models? I suggest that the focus be on impacts that are:

~ Ecologically important - the impact is large and relatively widespread geographically

-------
~	Economically important - the impact will affect ecosystem services with high economic
values

~	Well understood - one needs to be able to project the magnitude of the impact in a
scientifically robust manner.

2. Key terrestrial ecosystem impacts

Neither this abstract, nor the oral presentation that accompanied it, is intended to provide a
comprehensive overview of terrestrial climate change impacts. Rather, I intend to highlight some
of the key impacts and related tools that either have been or could be incorporated into integrated
assessment models. Specifically, I discuss three large-scale terrestrial ecosystem impacts: (1)
changes in vegetation distribution and dynamics, (2) changes in wildfire dynamics, and (3)
potential increases in species extinction risks.

2.1 Changes in vegetation distribution and dynamics

Why climate change will affect vegetation dynamics. Climate is a fundamental driver of key
ecological properties and processes. Temperature, precipitation, and relative humidity (and other
climatic variables) affect where species can persist, ecosystem productivity, rates of ecosystem
processes (e.g., organic matter decomposition), and frequency and intensity of disturbance events
(e.g., wildfire, droughts, and pest outbreaks). Changing climate will thus fundamentally affect
our environment, changing where grasslands and forests are located, the productivity of
ecosystems, and kinds of disturbance regimes ecosystems experience.

Tools used to project changes in vegetation. Dynamic global vegetation models (DGVMs) are
the most commonly used tool to project future changes in vegetation. Although many different
models exist, typically all of them can address three specific issues: (1) how soil and climate
affect the distribution of vegetation types, (2) how nutrients move within a given geographic
area, and (3) wildfire dynamics. For a specific time period and climate change scenario, DGVMs
can provide information about the potential distribution of vegetation types, the biomass of
different types of vegetation (e.g., tree, shrub, grass), terrestrial carbon storage, and the
frequency and extent of wildfires.

Examples of recent research. There has been a large volume of DGVM-related research over the
past decade, and it is beyond the scope of this abstract to capture the history and evolution of that
research. However, I will note that recent studies have used DGVMs to examine the impact of
climate change on vegetation dynamics in specific countries, regions, and globally. For example,
Lenihan et al. (2008) used the MCI DGVM to demonstrate that there may be rather significant
shifts in the distribution of vegetation in the United States by 2100 under climate change
scenarios (based on the SRES A2 and B2 emissions scenarios). This study also showed that the
extent of change in vegetation type and carbon storage is heavily influenced by fire suppression.
Sitch et al. (2008) examined the potential impact of climate change on global vegetation and
compared results across five DGVMs and four different SRES emissions scenarios. Their results

-------
showed substantial differences across models in vegetation responses to drought in the tropics
and warming temperatures in boreal ecosystems. In this study, for all but the most extreme SRES
scenarios (A1FI), the DGVMs suggested that the terrestrial biosphere would continue to be a
sink throughout the 21st century. The stimulative effect of elevated C02 compensated for the
direct suppressive effects of climate change on terrestrial carbon uptake. Galbraith et al. (2010)
assessed the extent of projected Amazon forest die-back under future climate change using three
DGVMs and the Hadley climate model (HadCM3). All DGVMs showed some degree of die-
back, but the extent and intensity of the die-back varied significantly across DGVMs.
Importantly, the models varied in their sensitivities to changes in rainfall and temperature; one
model was equally sensitive to both changes in precipitation and temperature, but the other two
were strongly affected by changes in temperature and insensitive to changes in precipitation.

Key uncertainties and shortcomings in projections of vegetation change. DGVMs can provide
insights into the nature and magnitude of potential climate change impacts on terrestrial
ecosystems, but there are important sources of uncertainty that should be kept in mind. First,
these models provide information about "potential" vegetation only, ignoring the very real and
critical impact that human intervention can have on the composition and productivity of
vegetation. One can address fire suppression in DGVMs, and one can also screen out current and
future urban and agricultural areas, but other effects (e.g., direct plantings, fertilization, invasive
species) will be ignored. Second, many DGVMs assume there are no barriers to plant dispersal.
This is clearly not the case, particularly in highly fragmented urban or agricultural landscapes.
Third, the impacts of pests and pathogens are ignored, despite how critical these disturbance
agents can be to shaping ecosystems. Finally, results across DGVMs can vary substantially for
the same region and the same climate change scenario.

Affected ecosystem services. Changes in vegetation will affect a multitude of ecosystem
services. Among the most important will be the provisioning of timber and non-timber forest
products, grazing, and carbon storage and sequestration, which are critical to understanding
potential terrestrial feedbacks to anthropogenic climate change.

2.2 Changes in wildfire frequency

Why climate change will affect wildfire frequency. Fires are likely to increase in many areas due
to both the direct and indirect effects of climate change. Higher temperatures will directly
increase the likelihood of fires - a spark is more likely to turn into a fire when temperatures are
hotter. Higher temperatures also desiccate vegetation and forest floor, which provide the fuel for
fires. Indirect effects on fire can be brought about via changes in vegetation. For example,
grasslands burn more readily than forests. And changes in productivity affect fuel load.

Tools used to project changes in wildfire. In my review of the literature, I identified two main
approaches to projecting wildfire under climate change. The first is to use statistical modeling.
This involves examining past fire behavior and identifying the factors that best predict historical
fire outbreaks (e.g., via step-wise linear regression). Changes in these factors are then used to

-------
predict fire behavior in the future. The second approach involves utilizing the relatively
simplistic fire models embedded in DGVMs to project future fire dynamics.

Examples of recent research. As with the review of dynamic global vegetation modeling, this
section is not at all meant to be comprehensive, but rather to highlight studies that demonstrate
the kinds of analysis that can be (and have been) done in this field recently. Many studies have
been done at relatively small scales or at the country level, but here I'll mention two global
studies, as that is the scale most relevant for integrated assessment models. Gonzales et al. (2010)
used output from the fire model of MCI to develop estimates of changes in fire frequency
between 2000 and 2100 under SRES scenario A1B. They found the fire frequency decreased on
two-fifths of global land, just slightly more than the area experiencing increases. Areas with
potentially lower fire frequencies included the coterminous United States, and northern Eurasia;
higher fire was projected for sub-Saharan Africa and northern South America. Krawchuk et al.
(2009) used statistical modeling to estimate changes in fire probability under SRES scenarios A2
and B2. Some of their results agree with Gonzales et al. (2010) - decreases in fire probabilities
are projected in northern Eurasia and higher fire probabilities are projected for South America.
However, Krawchuck et al. (2009) found that fire probabilities would be higher in the United
States and Europe and lower probabilities are estimated in sub-Saharan Africa.

Key uncertainties and shortcomings in projections of wildfire. For both statistical modeling and
DGVM approaches to projecting future fire, the models only roughly approximate historical
patterns of fire. They thus can only be expected to provide relatively rough indices of what might
happen with fire in the future. Like DGVM studies, the results of fire modeling studies can vary
significantly for the same region, and it is difficult for non-experts to assess which results are
more accurate. Finally, the timing and location of specific fires cannot be projected - only rough
approximations in overall changes in fire frequency and intensity for a given location can be
provided.

Affected ecosystem services. Like changes in vegetation, fire will affect the provisioning of
timber and non-timber forest products. Fire will also affect recreation, as people tend to stay
away for a period of time from areas that have recently burned before they return to hike, fish, or
camp. Although not an ecosystem service, changes in fire dynamics could also affect the amount
of money that is spent on fire suppression, an effect not likely addressed by other sectoral
analyses addressed in integrated assessment models. Finally, wildfire could have important
effects on air quality via the release of aerosols, which would have important health and visibility
implications.

2.3 Potential increases in species extinction risks

Why climate change will affect species extinctions. As noted earlier, climate is a critical driver
of where different species and ecosystems are found. As climate shifts, the areas providing the
climatic conditions that a species requires may move, sometimes into areas that don't have any
habitat that could support that species (e.g., into agricultural or urban areas). It is also possible

-------
that the climatic conditions a species requires may disappear altogether. This is more likely to
occur with species that live at high altitudes or latitudes.

Tools used to project changes in species extinction risks. There are a variety of approaches that
can be used to estimate the risk of future species extinctions. However, the most commonly used
approach involves the application of climate envelope models. These models use information
about the current distributions of species and the associated range of climate conditions to
construct their climate requirements. Under future climate scenarios, one estimates where that
species could live and how much area is available to it. Some studies then use species-area
relationships (species diversity is known to increase with size of geographic area) to determine
how many species can be supported in a future climate. Climate envelope models can be used
alone or in conjunction with expert opinion to estimate species extinction risks. Another, less
commonly used approach for examining future extinction risks involves utilizing vegetation
models to estimate habitat loss within a specific geographic area under different climate change
scenarios; such analyses often make the simplifying assumption that species ranges cannot shift
to accommodate changes in habitat (Pereira et al., 2010).

Examples of recent research. Estimates of species extinctions vary widely. For example, Thomas
et al. (2004) estimate that between 9 and 52% of species would be committed to extinction by
2050, depending on the assumptions made about dispersal ability and the specific climate change
scenario. The IPCC (2007a) estimated, using a combination of expert judgment and information
from climate envelop studies, that 20 to 30% of plant and animal species would be at risk of
extinction with an increase of 2 to 3°C in global temperature. A range of studies, using quite
different methodologies and examining different taxa, found that from 0 to 60% of species may
be at risk of extinction under future climate change (Pereira et al., 2010). Interestingly, Beale et
al. (2008) found that climate envelop models did no better than chance in explaining why species
reside where they do for approximately two-thirds of European bird species.

Key uncertainties and shortcomings in estimates of future extinction risks. Although climate
change poses a real, critical threat to species across the globe, there is a great deal of uncertainty
both within and among modeling studies about the magnitude of climate change impact on
species extinctions. Climate envelope models are known to have some specific technical issues;
these issues need to be specifically addressed because these types of models are used so often.
First, they may overestimate extinctions because species may be more flexible climatically than
their current distribution suggests. Envelope models may also underestimate extinctions because
climate change may interact with other factors, making things worse than predicted. For
example, some species may not be able to persist near human settlements even if the climate is
suitable.

Affected ecosystem services. Understanding the economic impact of global-level species
extinctions is another challenge. Economic studies have been done to estimate the existence
values of species, but these studies can be highly controversial. The impacts of extinctions on
other services could be explored, given the services that specific species or suites of species can
provide. For example, a tree species may provide valuable wood, and bird and wildlife viewing

-------
provide another type of value. However, such values are typically tied to species or geographic
locations rather than global extinctions, making this approach impractical.

3. Future research needs

3.1	Integrating across studies

Across all ecosystem impacts, there are a variety of methods available to project future
dynamics, and the methods typically give different answers regarding the magnitude of the
impact in question. The question is - how should all the different tools and studies be integrated?
Some ideas include:

~	Conducting meta analyses, which involves pooling data across many studies to detect
general patterns.

~	Developing ensemble means, as is done for climate models, across different impact
models. This approach would likely require ensembles to be developed across different
climate change scenarios and GCMs, making its feasibility questionable.

~	Soliciting expert opinions. Although imperfect and subjective, this is a cost-efficient
method for providing rough estimates of the potential direction and rough magnitude of
specific impacts.

3.2	Key knowledge gaps

There are also some critical knowledge gaps that need to be addressed when considering the
kinds of impacts that would appropriately be addressed in integrated assessment models. Three
key gaps include:

~	Pest outbreaks. There is a dire need for models that project the impact of climate change
on pest outbreaks. We know that pests are critical drivers of the productivity and
structure of ecosystems and that they will have significant impacts on the provisioning of
ecosystem services.

~	Freshwater wetlands. Large-scale, interior, freshwater wetlands provide critical
ecosystem services, and it is clear they will be affected by changes in precipitation and
temperature. Although some models have been developed to conduct sensitivity analyses
related to wetland impacts, projections of impacts for specific regions (e.g., the Prairie
Pothole region) are needed.

~	Snow pack dynamics. A lot of research has examined the impact of climate change on
snow pack dynamics, but this research has typically focused on water resource
implications. Changes in snow pack volume and the timing of snow pack melt can affect
freshwater and marine ecosystems as well as snow-related recreation.

-------
Literature cited

Beale, C.M., J.L. Lennon, and A. Gimona. 2008. Opening the climate envelope reveals no
macroscale associations with climate in European birds. Proceedings of the National Academy of
Sciences (105): 14908-14912.

Galbraith, D., P.E. Levy, S. Stitch, C. Huntingford, P. Cox, M. Williams, and P. Meir. 2010.
Multiple mechanisms of Amazonian forest biomass losses in three dynamic global vegetation
models under climate change. New Phytologist (187) 64-655.

Gonzales, P., R.P. Neilson, J.M. Lenihan, and R.J. Drapek. 2010. Global patterns in the
vulnerability of ecosystems to vegetation shifts due to climate change. Global Ecology and
Biogeography (19): 755-768.

IPCC. 2007a. Climate Change 2007: Synthesis Report. An Assessment of the Intergovernmental
Panel on Climate Change. Intergovernmental Panel on Climate Change.

IPCC. 2007b. Climate Change 2007: The Physical Science Basis. Summary for Policymakers.
Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change, S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt,
M. Tignor and H.L. Miller (eds.). Cambridge University Press, Cambridge, UK and New York.

Krawchuk, M.A., M.A. Moritz, M. Parisien, J. Van Dorn, and K. Hayhoe. 2009. Global
pyrogeography: the current and future distribution of wildfire. PloS ONE (4): e5102.

Lenihan, J.M., D. Bachelet, R.P. Neilson, and R. Drapek. 2008. Simulated response of
coterminous United States ecosystems to climate change at different levels of fire suppression,
CO2 emission rate, and growth response to CO2. Global and Planetary Change 64: 16-25.

Pereira, H.M., P.W. Leadley, V. Proenca, R. Alkemade, J.P.W. Scharlemann, J.F. Fernandez-
Manjarres, M.B. Araujo, P. Balvanera, R. Biggs, W.W.L. Cheung, L. Chini, H.D. Cooper, E.L.
Gilman, S. Guenette, G.C. Hurtt, H.P. Huntington, G.M. Mace, T. Oberdorff, C. Ravenga, P.
Rodrigues, R.J. Scholes, U.R. Sumaila, andM. Walpole. 2010. Scenarios for global biodiversity
in the 21st century. Science (330): 1496-1501.

Sitch, S., C. Huntingford, N. Gedney, P.E. Levy, M. Lomass, S.L. Piao, R. Betts, P. Ciais, P.
Cox, P. Friedlingstein, C.D. Jones, I.C. Prentice, and F.I. Woodward. 2008. Evaluation of the
terrestrial carbon cycle, future plant geography and climate-carbon cycle feedbacks using five
Dynamic Global Vegetation Models (DGVMs). Global Change Biology (14): 2015-2039.

Thomas, C.D., A. Cameron, R.E. Green, M. Bakkenes, L.J. Beaumont, Y.C. Collingham, B.F.N.
Erasmus, M.F. de Suiquiera, A. Grainger, L. Hannah, L. Hughes, B. Huntley, A.S. van Jaarsveld,
G.F. Midgley, L. Miles, M.A. Ortega-Huerta, A.T. Peterson, O.L. Phillips, and S.E. Williams.
2004. Extinction risk from climate change. Nature (427): 145-148.

-------
STRATUS CONSULTING

t of Climate Change on
Terrestrial Ecosystems

Climate Damages Workshop

Karen Carney, PhD
Stratus Consulting
Washington, DC
January 28, 2011

-------
Outline

n Background

n Descriptions of key ecosystem impacts

-	Vegetation distribution and dynamics

-	Wildfire dynamics

-	Species extinction risks
n Future research needs

STRATUS CONSULTING

-------
Background

n Why do ecosystems matter when assessing
economic impacts of climate change?

n Provide critical services to people

-	Provisioning (e.g., food, water, raw
materials)

-	Regulating (e.g., air quality, storm
protection, waste assimilation)

-	Cultural (e.g., recreation, passive use)

n These services have substantial economic
value

STRATUS CONSULTING

-------
Background (cont.)

° Climate change affects:

-	What species are where

-	How productive an ecosystem is

-	Rates of ecosystem processes
(e.g., decomposition, denitrification)

-	The disturbance regimes it experiences

•	Drought

•	Pest outbreaks

Photo credits: USFWS

STRATUS CONSULTING

n

-------
Background (cont.)

n Which ecological impacts?

n Given focus on use in integrated assessment

models, focus on impacts:

-	Ecologically important

•	Impact is large and relatively
widespread

-	Economically important

•	Impact will affect ecosystem services
with high values

-	Well understood

•	Need to quantify projected impacts in
scientifically robust way

STRATUS CONSULTING

-------
Outline

n Background

n Descriptions of key ecosystem impacts

-	Vegetation distribution and dynamics

-	Wildfire dynamics

-	Species extinction risks

n Future research needs

STRATUS CONSULTING

-------
Key Ecosystem Impacts

n For each impact, will discuss:

-	Why the impact is likely to occur

-	The tools available to estimate the impact

-	What research has shown

-	Key uncertainties or other shortcomings with
projecting future impacts

-	What key services are likely to be affected

STRATUS CONSULTING

-------
Outline

n Background

n Descriptions of key ecosystem impacts

-	Vegetation distribution and dynamics

-	Wildfire dynamics

-	Species extinction risks
n Future research needs

STRATUS CONSULTING

-------
Changes in Vegetation

n How will climate affect vegetation?

-	Changes in temperature, precipitation,
relative humidity affect:

•	What species can live where

•	Ecosystem productivity

•	Wildfire frequency and intensity, a key
disturbance agent

-	Will fundamentally alter our environment -
where grasslands and forests are, and what
kinds of animals we see in different areas
(not static)

STRATUS CONSULTING

-------
Changes in Vegetation (cont.)

n Projecting future vegetation dynamics

- Dynamic global vegetation models (DGVMs)

•	Large scale patterns of vegetation change

•	Typically have interacting modules:

-Biogeography model - potential
vegetation given climate and soil
parameters

-Biogeochemistry model, which
simulates the movement of nutrients

-Fire model - disturbance by wildfire

STRATUS CONSULTING

-------
Changes in Vegetation (cont.)

n Projecting future vegetation dynamics (cont.)

- For specified time period and climate
scenario, DGVMs can tell you:

•	Potential vegetation type (e.g., temperate
deciduous forest, temperate mixed forest)

•	Plant biomass (by life form - trees,
shrubs, grasses)

•	Carbon storage (above and below-
ground)

•	Burned area/wildfire frequency

STRATUS CONSULTING

-------
Changes in Vegetation (cont.)

n Projecting future vegetation dynamics (cont.)
- Many DGVMs are available; commonly used:

•	MC1 - United States

•	Lund-Potsdam-Jena (LPJ) -
Germany/Sweden

•	SDGVM - United Kingdom

•	Integrated Biosphere Simulator (IBIS) -
United States

STRATUS CONSULTING

-------
Changes in Vegetation (cont.)

° What research has shown

Historical

A. SF-HIST

B. SF-A

From: Lenihan et al , 2008. Global and Planetary Change 64:16-25.

Fig. 4. Model simulated vegetation type with suppressed fire(SF) for 1971-2000 historical period and 2CF70-2099 future period. A: SRES-A2, B: SRES-B2.

Alpine

Suoalpine Forest

Temperate Conifer Forest

Cool Mixed Forest

Temperature Decidous Forest

Warm Temperate Mixed Forest

Tropical Forest

Woodland/ Savanna

Shrubland

Grassland

Desert

STRATUS CONSULTING

-------
Changes in Vegetation (cont.)

° What research has shown (cont.)

-	% change in tree coverage, SRES
A1FI, 4 DGVMs, Hadley GCM

-	Significant variability across
models

-	Some areas of general agreement

•	Varying degrees of Amazon
forest dieback

•	Boreal forest expansion

From: Sitch et al., 2008. Global Change Biology
14:2015-2039.

STRATUS CONSULTING

-------
Changes in Vegetation (cont.)

n Key uncertainties

-	Potential vegetation only - most anthropogenic
factors ignored; some can be addressed

•	Fires suppression can be accounted for

•	Can screen out urban/agricultural lands

-	Assume no barriers to plant dispersal

-	Pests and pathogens are ignored

-	Significant differences across DGVMs for the
same region and climate scenario

STRATUS CONSULTING

-------
Changes in Vegetation (cont.)

° Affected ecosystem services

Forestry	I mm

•	Timber	Ififp

•	Non-timber forest products
Grazing

•	Forage productivity in grasslands,
shrublands, savannas, and forests

Carbon sequestration and storage

STRATUS CONSULTING

-------
Changes in Vegetation (cont.)

n Take home

-	Ecosystems across the globe will be
affected, so this is a key impact to consider

-	Can examine multiple scales - countries,
regions, the globe

-	Linked to critical ecosystem services

-	Good models, but difficult to know which
ones are most reliable

-	Highly dependent on the GCM used

-	Look for areas of agreement, perhaps
average DGVM results when possible

STRATUS CONSULTING

-------
Outline

n Background

n Descriptions of key ecosystem impacts

-	Vegetation distribution and dynamics

-	Wildfire dynamics

-	Species extinction risks
n Future research needs

STRATUS CONSULTING

-------
Wildfire Dynamics

] low will climate affect wildfire?

- Fires will likely increase in many areas
via various mechanisms	Photc

•	Direct

-Higher temperatures = more fires

-Higher temperatures (and decreased
precipitation) = desiccation of vegetation
and forest floor (fuel)

•	Indirect

-Changes in vegetation type
(grassland/forest)

-Changes in productivity (fuel load)

Photo credit: USFWS

STRATUS CONSULTING



-------
Wildfire Dynamics (cont.)

n Projecting future wildfire dynamics
- Statistical models

•	Examine past fire behavior

•	Identify factors (e.g., via stepwise linear
regression) that are key to predicting fire

•	Use equation to predict fires in future
(based on key variables)

-DGVMs

STRATUS CONSULTING

-------
Wildfire dynamics (cont.)

° What research has shown

-	Change wildfire freq. from 2000-2100, A1B

-	More fire: U.S., central South America, southern
Africa, western China, Australia

-	Less fire: northern Canada, northern Russia

From: Gonzales et al. 2010. Global Ecol. Biogeogr. 19: 755-768

STRATUS CONSULTING

-------
Wildfire Dynamics (cont.)

n Key uncertainties

- For both statistical model and DGVM
approaches

•	Methods only roughly approximate
historical fires

•	Thus, provide similarly rough estimates of
future wildfire dynamics

•	Timing/locations of specific fires cannot be
predicted

STRATUS CONSULTING

-------
Wildfire Dynamics (cont.)

n Affected ecosystem services

Timber/non-timber forest product
provisioning

Recreation

Fire suppression (not an ecosystem
service but a real cost)

Regulation of air quality - aerosols

(see Spracklen et al., 2009, Journal of Geophysical
Research)

Photo credit: USFWS

Photo credit: USFWS

STRATUS CONSULTING

-------
Outline

n Background

n Descriptions of key ecosystem impacts

-	Vegetation distribution and dynamics

-	Wildfire frequency/intensity

-	Species extinction risks
n Future research needs

STRATUS CONSULTING

-------
Species Extinctions

How will climate affect it?

Climate (temperature/precipitation^
is a key driver of species and
ecosystem distributions

As climate shifts, areas that support
specific species may move
(sometimes into areas inhabited by
humans)

Habitat may disappear (e.g., alpine,
cloud-forest dependent species)

These dynamics will likely increase
the risk of species extinctions

> 'V

Photo credits: USFWS

STRATUS CONSULTING

¦¦

-------
Species Extinctions (cont.)

n Projecting future species extinctions

- Most commonly involves application of
"climate envelope" models

•	Use current distributions of a species to
construct its climatic requirements

•	Under future climate change, then
determine where species could live

•	Use species-area relationships to project
extinctions

STRATUS CONSULTING

-------
Species Extinctions (cont.)

n What research has shown

-	Results vary

•	9-52% of species will be "committed" to
extinction by 2050 (Thomas et al., 2004)

•	20-30% of plant and animal species at
risk of extinction with increase of 2-3 C
(IPCC, 2007)

•	0-60% extinctions for different
taxa/methodologies (Pereira et al., 2010)

-	Envelope model did no better than "null"
models in predicting species occurrence (null
= species ranges are randomly placed in
region; Beal et al., 2010)

STRATUS CONSULTING

-------
Species Extinctions (cont.)

n Key uncertainties

-	Great deal of uncertainty within and across
studies and modeling methods

-	Climate envelope models

•	May overestimate extinctions

-Species may be flexible climatically

-Biotic interactions may be more
important than climate

•	May underestimate extinctions

-Dispersal may be limited by habitat
fragmentation

-Impacts of climate change may be
amplified by land use change

STRATUS CONSULTING

-------
Species Extinctions (cont.)

n Affected ecosystem services

-	Another key issue...

-	How do you value global biodiversity?

• Could query public

-Some species may matter more to the
public, and ecologically, than others

STRATUS CONSULTING

-------
Species Extinctions (cont.)

n Affected ecosystem services (cont.)

-	Values could be tied to specific species, or
suites of species

•	A given tree may provide highly valued
wood

•	Bird watching/wildlife viewing is valuable

-	But values not tied to global extinction risk -
linked to species, suites of species, and/or
specific locations

STRATUS CONSULTING

-------
Species Extinctions (cont.)

n Take home

-	Climate change is a threat to species, and
more extinctions are likely to occur

-	Range of estimates available for species
extinction risk

-	Robustness of estimates highly contested

-	Link to ecosystem services and values
difficult

-	Proceed with caution

STRATUS CONSULTING

-------
Outline

n Background

n Descriptions of key ecosystem impacts

-	Vegetation distribution and dynamics

-	Wildfire frequency/intensity

-	Species extinction risks
n Future research needs

STRATUS CONSULTING

-------
Photo credit: USFWS

Future research

° Integrating across approaches

-	Across all impacts, variety of methods available
that provide different estimates of impact

-	Need to think carefully about how to integrate
across studies/tools

•	Meta-analyses?

•	Ensemble means' of ecosystem impacts
with different models?

•	Need to be done with different climate
scenarios/GCMs

•	How can this be done practically?

STRATUS CONSULTING



-------
Future research (cont.)



Major Gaps

- Need to develop large-scale, long
term projections for changes in

•	Pest outbreaks

•	Interior wetland change/loss

Photo credit: USFWS

Changes in snow pack dynamics

•	Large-scale impacts on
freshwater/marine ecosystems

•	Implications for recreational
values

Photo credit: USFWS

STRATUS CONSULTING

-------
Thank you!

Photo credit: USFWS

STRATUS CONSULTING

El

-------
Wildfire Dynamics (cont.)

D What research has shown



- Fire risk for three



different time periods



over 21st century



- Higher fire risk:



• U.S.



• Amazon



• Western China



- Lower fire risk:



• Northern Canada



• Russia



• Australia (?)



From: Krawchuck et aL 2009. Plos One 4: e5102.

i —win— i

STRATUS CONSULTING

-------
Valuing the impacts of climate change on forestry
Brent Sohngen (Ohio State University)

1. Briefly review the existing estimates of the value of climate change impacts on forestry. In
addition to the best central estimates, also describe the wider range of possible outcomes—
including those that may arise from potential economic catastrophes—and the relative
likelihoods of these outcomes.

Current estimates suggest that forestry outputs are likely to increase globally over the century
(see Table 1). As a result, consumers will gain from increased timber output and lower timber
prices. Producers could gain if timber production increases due to climate change, although
lower prices could have negative impacts in some regions (for discussion of overall welfare
impacts, see Sohngen et al., 2001). The strongest gains are projected for subtropical regions
where producers are able to adapt more quickly with faster growing timber types.

Table 1: Estimates of impacts of climate change on timber outputs by region (reproduced from
Table 4.2 in Seppala et al., 2004).

Region

Output

Producer Returns



2000-2050

2050-2100



North America1

-4% to +10%

+ 12 to+16%

Decreases

Europe2

-4% to +5%

+2 to+13%

Decreases

Russia3

+2 to +6%

+7 to+18%

Decreases

South America4

+10 to +20%

+20 to +50%

Increases

Australia/New Zealand4

-3 to+12%

-10 to +30%

Decreases& Increases

Africa5

+5 to+14%

+ 17 to +31%

Increases

China5

+10 to+11%

+26 to +29%

Increases

South-east Asia5

+4 to+10%

+ 14 to +30%

Increases

1	Alig et al. (2002), Irland et al. (2001), Joyce et al. (1995, 2001), Perez-Garcia et al. (1997, 2002), Sohngen et al. (2001),
Sohngen and Mendelsohn (1998, 1999), Sohngen and Sedjo (2005)

2	Karjalainen et al. (2003), Nabuurs et al. (2002), Perez-Garcia et al. (2002), Sohngen et al. (2001)

3	Lelyakin et al. (1997), Sohngen et al. (2001)

3	Lelyakin et al. (1997), Sohngen et al. (2001)

4	Perez Garcia et al. (1997, 2002), Sohngen et al. (2001)

5	Sohngen et al. (2001)

Although the general results suggest higher output in forestry, there is large uncertainty about
these results. The ranges shown in Table 1 are not uncertainty bounds, but they are instead
ranges based on different studies in the literature. These do not reflect the full set of uncertainty
that would be expected to affect estimates of economic impacts, but they are illustrative of the
current state of knowledge.

One of the difficulties of measuring uncertainty in economic outcomes relates to method used to
conduct integrated assessment modeling of forestry impacts. Figure 1, for example, represents
the typical modeling steps that are undertaken to calculate the impacts of climate change on
forestry. Modelers start with the climate models, which are linked to ecosystem models, which
are in turn linked to economic models. All of the models have their own uncertainties, and

-------
researchers will handle these uncertainties in different ways, depending on the resources they
have to conduct a study.

For example, uncertainty in climate outcomes from the climate models can be incorporated, at
least tentatively, by utilizing several different models. There are a large number of climate
models, and if researchers have access to many of them, they can choose them in order to
represent the range of potential outcomes from the models. The ecosystem models, nowadays
the Dynamic Global Vegetation Models (DGVMs), also contain uncertainty. There are fewer
ecosystem models than climate models, but in the past, research teams have collaborated to
prepare results across different ecosystem models based on common climate inputs (e.g.,
VEMAP Members, 1995). In these cases, the research teams have represented at least some of
the uncertainty in ecosystem outcomes by using results from several models.

Figure 1: Flow of forestry integrated assessment models of climate change impacts (from
Seppala et al., 2009)

2

-------
2.	How do these estimates vary across regions? Characterize the uncertainty / robustness / level
of confidence in these estimates, on average globally and by region.

The results in table 1 suggest potential negative effects in the shorter-term in temperate regions
like the United States, Canada, Europe, Russia, etc. There are several reasons for this. The
ecological models utilized in the studies in Table 1 suggested that climate change could cause
relatively large disturbances in forests over the next several decades, and these disturbances
could negatively influence outputs. While the impacts are mitigated to some extent by
adaptation through salvage harvesting, the changes in disturbance patterns modeled by the
ecological models were large enough to have important impacts.

In contrast, most subtropical and tropical regions are projected to potentially benefit from climate
change according to the results in Table 1. These trends are anticipated to continue. Over the
past half century, there has been a continued increase in the area of fast-growing timber
plantations in subtropical regions world-wide. Current estimates suggest that there are 90-100
million hectares of fast-growing timber plantations globally with20-40 million of these hectares
located in subtropical regions (ABARE-Jaako-Poyry, 1999; Sohngen, 2010). They are
estimated to provide 15-25% of global timber supplies currently (Daigneault et al., 2008;
Sohngen, 2010), and are expected to provide much of the growth in output in the coming
decades. The fast-growing plantations have timber species that can be harvested in 10-25 year
rotations and produce 10-20 m3 per hectare per year in wood (Cubbage et al., 2010).

Economic studies suggest that managers are able to adapt to climate change relatively rapidly
with these fast-growing plantation species. As supplies in temperate zones are affected by
disturbances, supply of timber from plantations expands to limit any shortfalls globally. In fact,
managers of plantation forests appear to be able to take advantage of some of the impacts of
climate change in forests that have longer rotations. Furthermore, ecosystem models used in the
earlier economic studies did not suggest as large of disturbance patterns in subtropical regions as
in temperate regions, so plantations were exposed to less risk than their counterparts further
north.

As noted above, it is difficult to quantify the uncertainty in economic outcomes. Most of the
studies conducted so far are greater than 5 years old, and thus are reliant on climate and
ecosystem modeling that occurred in the early to mid-1990s. This constitutes an important
limitation to the robustness of the result described above. Utilizing more recent climate and
ecological modeling may lead to very different estimates of economic impacts.

3.	Briefly review the models and data used to estimate the value of climate change impacts on
forestry.

a. What types of natural science models and data are used to inform these estimates, and
what categories of values have been included?

A number of different ecosystem models have been used to date by economic modelers. For
example, the models in the study by VEMAP Members (1996) have been used by a number of
different modelers to examine economic effects of climate change. These earlier models have
been supplanted by more recent Dynamic Global Vegetation Models (e.g., Fischlin et al., 2007;

3

-------
Bachelet et al., 2003, Bachelet et al., 2004). These models project changes in ecosystem type,
changes forest productivity (net primary productivity, net ecosystem productivity, and net
biological productivity), and in some cases changes in carbon content due to fire or other
disturbances. The models can be implemented at a range of scales depending on the inputs.
Often, for example, they are implemented globally at the 0.5 degree grid cell basis; however,
they can be implemented at a finer scale for more specific regional analysis. For climate
analysis, however these models all rely on climate model inputs, which are often provided at a
much more aggregate level.

b.	What physical and economic factors make some regions more or less vulnerable to the
impacts of climate change on forestry than others?

The most vulnerable regions to climate change in forestry appear to be regions that currently
produced the greatest share of output. For instance, in the United States, the Southern US
produces the greatest share of output nationally and is also projected to be the most vulnerable in
analyses to date (e.g., Sohngen and Mendelsohn, 1998, 1999; Sohngen et al, 2001; Bachelet et
al., 2003; Bachelet et al., 2004). In a global context, the study by Sohngen et al. (2001) suggests
that North America, Europe, and Russia are more vulnerable to climate change than other
regions due to ecological and economic factors. Note that these regions currently constitute
over 65% of industrial timber outputs. Ecologically, the models suggested that these regions
would experience greater disturbance with climate change. Economically, these temperate
regions could adapt, but because other regions were able to adapt more rapidly, prices fell, and
the lower prices reduced welfare for landowners in temperate regions.

The physical factors that make a region more or less vulnerable relate mainly to the growth rate
of the timber stocks and the area of fast-growing plantations. Regions with faster growing
species appear to be more able to adapt to climate change, whereas regions with slower growing
species appear to be more susceptible to damages from forest fires and other impacts.

c.	How are the values of forestry impacts projected into the future, accounting for changes
in other economic and environmental conditions?

For the most part, process based economic models are utilized. These models used either
dynamic optimization approaches, or other static simulation approaches to projection timber
harvests. Most studies have made timber price endogenous so they are able to account for other
factors that influence timber demand, such as changes in population and income.

4. What are the most important gaps or uncertainties in our knowledge regarding the value of
forestry impacts? What additional research in this area would be most useful?

There are a number of important gaps. Three potential gaps and additional research topics are
listed below:

• The pace of change in ecosystem and climate models appears to be much more rapid than
the pace of change in economic modeling. For example, new scenarios of climate models

4

-------
and new scenarios of ecosystem models seem to appear about every 5 years, while new
economic analysis emerges much more slowly.

There may be a number of reasons for this. One possible explanation is that the effect of
humans on forested ecosystems is changing dramatically. The economic studies reviewed
above focus on timber demand as the driving human influence; however, timber demand
may not be the most important demand of forest resources in the future. For example,
future demand for natural ecosystems may be driven by non market vaues or recreational
values, or it may be driven more importantly by land-use change (e.g., conversion of
productive timberland to private recreational land). Alternatively demand may be driven
by agricultural uses in some regions of the world (e.g., tropical regions where agricultural
land is expanding). Thus, more complex models that account for different kinds of
demands for land may be necessary to fully assess the implications of climate change on
forestry.

•	The economic models have not fully reflected the uncertainty. The ecosystem models
suggest that disturbance patterns could change dramatically over time, but there has been
little use of this information by economic modelers to date. There are a number of
stochastic models of forest management under uncertain disturbance regimes (e.g.,
Daigneault et al., 2010), but few of these are linked to climate models .

•	Ecosystem models are calibrated without reference to models of human behavior. This
likely causes them to over-estimate the potential effects of climate change on ecosystems.
In many ecosystems, for instance, one would expect humans to adapt to damages, and
this adaptation is missing from the ecosystem models. There is substantial room for
modelers to conduct integrated economic and ecological analysis that would capture
these effects.

References

ABARE - Jaako Poyry. 1999. Global Outlook for Plantations. ABARE Research Report 99.9.
Canberra, Australia.

Alig, R., D. Adams, and B. McCarl. 2002. Projecting impacts of global climate change on the
U.S. forest and agriculture sectors and carbon budgets. Forest Ecology and Management, 169:3-
14.

Bachelet, D., R. P. Neilson, T. Hickler, R. J. Drapek, J. M. Lenihan, M. T. Sykes, B. Smith, S.
Sitch, and K. Thonicke. 2003. Simulating past and future dynamics of natural ecosystems in the
United States. Global Biogeochemistry Cycles 17(2): 14.1-14.21.

Bachelet, D., R.P. Neilson, J.M. Lenihan, R.J. Drapek. 2004. Regional differences in the carbon
source-sink potential of natural vegetation in the U.S.A. Ecological Management 3 3 (Supplement
1):S23-S43.

5

-------
Cubbage, Frederick, Sadharga Koesbanda, Patricio MacDonagh, Gustavo Balmelli, Virginia
Morales Olmos, Rafael Rubilar, Rafael de la Torre, Vitor Hoeflich, Mauro Murraro, Heynz
Kotze, Ronalds Gonzalez, Omar Carrerro, Gregory Frey, James Turner, Roger Lord, Jin Huang,
Charles Maclntyre, Kathleen McGinley, Robert Abt, and Richard Phillips. 2010. Global timber
investments, wood costs, regulation, and risk. Biomass and Bioenergy. In press.
Doi:10.1016/j.biombe.2010.05.008.

Daigneault, A., B. Sohngen, and R. Sedjo. 2008. " Exchange Rates and the Competitiveness of
the United States Timber Sector in a Global Economy." Forest Policy and Economics. 10(3):
108-116.

Fischlin, A., G.F. Midgley, J.T. Price, R. Leemans, B. Gopal, C. Turley, M.D.A. Rounsevell,
O.P. Dube, J. Tarazona, A.A. Velichko, 2007: Ecosystems, their properties, goods, and services.
Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II
to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, M.L. Parry,
O.F. Canziani, J.P. Palutikof, P.J. van der Linden and C.E. Hanson, Eds., Cambridge University
Press, Cambridge, 211-272.

Irland, L.; D. Adams, R. Alig, C.J. Betz, C. Chen, M. Hutchins, B. A. McCarl, K. Skog, and
Brent L. Sohngen. 2001. "Assessing Socioeconomic Impacts of Climate Change on U.S. Forest,
Wood-Product Markets, and Forest Recreation." Bioscience. 51(9): 753- 764.

Joyce, L.A., J.R. Mills, L.S. Heath, A.D. McGuire, R.W. Haynes, and R.A. Birdsey. 1995.
Forest sector impacts from changes in forest productivity under climate change. Journal of
Biogeogj'ciphy, 22: 703-713.

Joyce, L., J. Aber, S. McNulty, V. Dale, A. Hansen, L. Irland, R. Neilson, K. Skog. 2001.
Potential Consequences of Climate Variability and Change for the Forests of the United States.
Chapter 17 in Climate Change Impacts on the United States: The Potential Consequences of
Climate Variability and Change. National Assessment Synthesis Team, US Global Change
Research Program. (http://www.usgcrp.gov/usgcrp/Library/nationalassessment/)

Karjalainen, T., Pussinen, A., Liski, J., Nabuurs, G.J., Eggers, T., Lapvetelainen, T., Kaipainen,
T. 2003. Scenario analysis of the impacts of forest management and climate change on the
European forest sector carbon budget. Forest Policy and Economics 5: 141-155.

Lelyakin, A.L., A.O. Kokorin, I.M Nazarov. 1997. Vulnerability of Russian Forests to Climate
Changes, Model Estimation of CO2 Fluxes. Climatic Change. 36:123-133.

Nabuurs, G.J., Paivinen, R., Schanz, H. 2001. Sustainable management regimes for Europe's
forests — a projection with EFISCEN until 2050. Forest Policy and Economics 3: 155-173.

Nabuurs, G.J., Pussinen, A., Karjalainen, T., Erhard, M., Kramer, K. 2002. Stemwood volume
increment changes in European forests due to climate change-a simulation study with the
EFISCEN model. Global Change Biology 8: 304-316.

6

-------
Perez-Garcia, J., L.A. Joyce, A.D. McGuire, and C.S. Binkley. 1997. "Economic impact of
climatic change on the global forest sector" In Economics of Carbon Sequestration in Forestry,
eds. R.A. Sedjo; R.N. Sampson; and J. Wisniewski. Lewis Publishers, Boca Raton.

Perez-Garcia, J., L.A. Joyce, A.D. McGuire, and X. Xiao. 2002. "Impacts of Climate Change
on the Global Forest Sector." Climatic Change. 54: 439-461,

Seppala, R., A. Buck, P. Katila. 2009. "Adaptation of Forests and People to Climate Change -
A Global Assessment Report. IUFRO World Series Volume 22. Helsinki, Finland, 224 p.

Sohngen, B. and R. Mendelsohn. 1998. "Valuing the Market Impact of Large Scale Ecological
Change: The Effect of Climate Change on US Timber." American Economic Review, 88(4): 689
-710.

Sohngen, B. and R. Mendelsohn. 1999. "The US Timber Market Impacts of Climate Change."
Chapter 5 in The Market Impacts of Climate Change on the US Economy. Edited by R.
Mendelsohn and J. Neumann. Cambridge, U.K.: Cambridge University Press.

Sohngen, B., R. Mendelsohn, and R. Sedjo. 2001. " A Global Model of Climate Change Impacts
on Timber Markets." Journal of Agricultural and Resource Economics. 26(2): 326-343.

Sohngen, B. and R. Sedjo. 2005. "Impacts of Climate Change on Forest Product Markets:
Implications for North American Producers." Forestry Chronicle 81(5): 669-674

Sohngen, B. 2010. Outsourcing Timber: The Role of Emerging Region Plantations on Southern
US Production. Working Paper. AED Economics. Ohio State University.

VEMAP Members (J.M. Melillo, J. Borchers, J. Chaney, H. Fisher, S. Fox, A. Haxeltine, A.
Janetos, D.W. Kicklighter, T.G.F. Kittel, A.D. McGuire, R. McKeown, R. Neilson, R. Nemani,
D.S. Ojima, T. Painter, Y. Pan, W.J. Parton, L. Pierce, L. Pitelka, C. Prentice, B. Rizzo, N.A.
Rosenbloom, S. Running, D.S. Schimel, S. Sitch, T. Smith, and I. Woodward). 1995.
Vegetation/ecosystem modeling and analysis project: comparing biogeography and
biogeochemistry models in a continental-scale study of terrestrial ecosystem responses to climate
change and C02 doubling. Global Biogeochemical Cycles 9:407-437.

7

-------
he Eco
Chang*

Brent Sohngen

Department of Agricultural, Environmental &
Development Economics,

University Fellow, RFF

Sohngen.l@osu.edu

-------
Outline of Presentation

•	Methods for assessment

•	Ecosystem impacts important for economic
analysis

•	Some results from a recent assessment.

-------
Scenario of future

economicand
population growth

	

Important
feedbacks
between
management,

ecological
impacts, and
market
interactions has
not been fully
addressed to
date.

Economic impacts (prices,
timber production, welfare
effects)

How are impacts measured?

Scenario of future climate ~L
forcing (e.g., SRES
scenarios)

Translation of ecological
results into growth and
disturbance effects

Ecological scenarios from
DGVMs or other ecological
models

Climate Scenarios from

General Circulation Models

^^^

Adaptation of Forests and People to Climate Change. 2009. Alexander Buck, Pia Katila and Risto
Seppala. (eds.). IUFRO World Series Volume 22. Helsinki. 224 p.

-------
Ecosystem Impacts

•	Productivity changes (IPCC, 2007)

-	C02fertilization (e.g., Norby et al., 2006).

-	Warming in colder climates.

-	Precipitation gains where water is limited.

•	Some current evidence that historical climate
change and C02 change have increased
productivity to date (e.g., Myneni et al., 1997;
Boisvenue and Running, 2006; McMahon et al.,
2010).

•	Potential limits to productivity gains: Net impacts

-	Species composition, age structure, seasonal and daily
precipitation and temperature patterns, etc.

-	Drying and forest fire effects

-------
Global Ecosystem Impacts

Losses ultimately	weigh do Ecosystems

turn from carbon sink to source within the next
several decades, due to fire and other disturbance



8 -i

CD

—





b

ti -

cn



CL

4 -

0



cn



c

2 -



-

o

o -



£=



o

-2 -

_Q







o

-4 -

-t—'



—	LPJ HadCM3 A2

—	LPJECHAM5B1

—	LPJ CRU Climatology

1900

1950

2000

T

2050

1

2100

IPCC (2007) WG 2, Chapter 4, Figure 4.2

-------
US Ecosystem Impacts

• Reduction in total ecosystem carbon with climate change.

—	Losses greatest in eastern US

—	Losses greater with more recent climate scenarios

%CHANGE between HISTORICAL (1961-1990) and FUTURE (2070-2099) CONDITIONS

HADCM2SUL	HADCM3a	HADCM3b

CGCM1	CGCM2a	CGCM2b

Bachelet et al. (2008)

-------
US Ecosystem Impacts

• How big might the losses be?

—	Emissions	of up to

—	Total loss over century

































' 1 & -t

T



A _



r











f

^ iJ

rty

A













V









1890 1910 1930 1950 1970 1990 2010 2030 2050 2070 2090 2110
	CGCM2-a CGCM2-b	HAl)CM3-a HAI)CM3-b

Bachelet et al. (2008)

-------
Scenario of future

economicand
population growth

Sp^nario of future chrH^te
f forcing (e.g., SRES
scenarios)





Climate Scenarios from
General Circulation Models





Ecological scenarios from
DGVMs or other ecological
models





Translation of ecological
results into growth and
disturbance effects

Need to integrate...

Economic impacts (prices,
timber production, welfare
.	effects) W

Adaptation of Forests and People to Climatb ChaogR 2009. Alexander Buck, Pia Katila and Risto
Seppala. (eds.). IUFRO World Series Volume 22. Helsinki. 224 p.

	

Important
feedbacks
between
management,

ecological
impacts, and
market
interactions has
not been fully
addressed to
date.

-------
Summary: Timber market results to date

Region

Output

Producer Returns



2000-2050

2050-2100



North America

-4% to +10%

+12 to+16%

Decreases

Europe

-4% to +5%

+2 to+13%

Decreases

Russia

+2 to +6%

+7 to+18%

Decreases

South America

+10 to +20%

+20 to +50%

Increases

Aus./New Zealand

-3 to+12%

-10 to +30%

Deer. & Incr.

Africa

+5 to +14%

+17 to +31%

Increases

China

+10 to +11%

+26 to +29%

Increases

SE Asia

+4 to +10%

+14 to +30%

Increases

Alig et al. (2002), Irland et al. (2001), Joyce et al. (1995, 2001), Perez-Garcia et al. (1997, 2002), Sohngen et al. (2001), Sohngen and Mendelsohn (1998, 1999),
Sohngen and Sedjo (2005); 2 Karjalainen et al. (2003), Nabuurs et al. (2002), Perez-Garcia et al. (2002), Sohngen et al. (2001) ; Lelyakin et al. (1997),

Adaptation of Forests and People to Climate Change. 2009. Alexander Buck, Pia Katila and Risto
Seppala. (eds.). IUFRO World Series Volume 22. Helsinki. 224 p.

-------
Updated Analysis

•	Climate Change:

•	A2, Alb scenarios

•	CSIRO, Hadley, MIROC models

•	Ecological Analysis: DGVM

•	MCI model (MAPPS and Century Model)

•	Economic Analysis:

•	Global Land Use Model (Sohngen and
Mendelsohn, 2007)

-------
Change in tmax
2070-2099 vs 1961-1990

0-1

1-2

2-3

3-4

4-5

5-6

6-7
>7

A2

A1B

B1

-------
<-50
-50--40
-40 - -30
-30 - -20
-20--10
-10-0
0-11
11 -25
25-43
43-67
67-100
> 100

B1

% Change in precip.
2070-2099 vs 1961-1990

-------
Approach to Economic Analysis

•	Ecosystem Model (DGVM) provides
information on

>	Shift in range for timber species

>	Natural disturbance losses (% stock burned)

>	Net primary productivity, net ecosystem productivity, and
net biological productivity

•	Data provided by DGVM

>	0.5 degree grid cells for globe.

>	Annually to 2100.

-------
Approach to Economic Analysis

Incorporate several factors

•	Yield change is
proportional to
the change in NPP

•	Stock losses due to
burned area

• Yield changes captured

as: I	a	1

V =^SV

A J / i t a J
a=\

• Stock losses captured as

^a+lj+1 ^ Yt Jfa.t ^a,t Sa=Lt

Area suitable for
trees changes

Use maps of shifts in
ecosystem types.

-------
Adaptations Incorporated

•	Manage existing stock by

-	changing rotations

-	Salvage

•	Replant new species if growing and economic
conditions warrant

•	Manage future stock by

-	Changing rotations

-	Changing management & investments

-------
Some Results from Economic Analysis

• Climate Change strengthens current trends
towards shorter rotations and production in
subtropical regions.

— South/Central America, Oceania, South Africa



Age

m3/ha/yr

$/ha

US Southern Pine

30

4.8

$3,180

S. China mixed

50

1.8

$771

Canada Boreal SW

70

1.6

$288

Russia Boreal SW

100

1.0

$58

South Amer. Eucalypt

10

7.0

$8,453

Oceania SW

30

13.5

$7,937

Source: Sohngen, 2010

-------
Market Projections: No Climate

Change

2500

2000

1500

(TJ

>
s_

V
Q.

m

E

g 1000

500

0

ROW

South America

Most growth in output
Is outside North America

I960 1980 2000 2020 2040 2060 2080 2100

-------
Market Projections with Climate

Change

• South America gains some advantage under
A2 for example

(0

a)

a)
o.

to

E

c
o

700
600
500
400
300
200
100
0

•US

South America

1 US A2

South America A2

1960

1980 2000

2020

2040

2060 2080

2100

-------
Some Results from Economic Analysis

•	Climate Change strengthens current trends
towards shorter rotations and production in
subtropical regions.

— South/Central America, Oceania, South Africa

•	Global output rising and timber prices falling

-------
Global Output and Prices
fall by 5-15%

4800

4300

% 3800

OJ
Q.

m 3300
c

= 2800

2300

HADA2
HADA1B
MIROCA1B
¦Base
¦CSIROA2
CSIROA1B
¦MIROCA2

1800 	1	1	1	1	1	1	1	1	1	1	1

% ^ ^

TO
aj
>

QJ

a.

m
E
c
o

180
160
140
120
100
80
60
40
20
0

HADA2
HADA1B
MIROCA1B
¦Base
¦CSIROA2
CSIROA1B
¦MIROCA2

	1	1	1	1	1	1	1	1	1	r

^ ^ ^



OS

-------
Some Results from Economic Analysis

•	Climate Change strengthens current trends
towards shorter rotations and production in
subtropical regions.

— South/Central America, Oceania, South Africa

•	Global output rising and timber prices falling

•	Regional results suggest winners and losers, but
dependent on climate scenarios.

-------
Regional results variable

1.20
1.00
0.80
0.60
0.40
0.20
0.00
(0.20)
(0.40)
(0.60)
(0.80)

2050

w

i

I

X

r

-¦ I' *¦

C

Ln

n cp

X 73

^ I

n
>

>
o
>

73

c

oo
oo

C

>

oo
O
C

>
oo

n

73
oo
>

OO

oo
>

>
Ln

> ±

1.50
1.00
0.50
0.00
(0.50) -
(1.00) -
(1.50) -

TT

Al

am

\

2090

M

"*T

rV

"T

V



c

Ln

0	CP

1	s

n
>

>
o
>

;xj

c

oo
oo

C

>

oo
O
C

>
oo

n

LO	LO

ld lo	m

> £	>

^ ^	E/>

O

n

>
TJ
>

>
OO

-------
Some Results from Economic Analysis

•	Climate Change strengthens current trends towards shorter
rotations and production in subtropical regions.

-	South/Central America, Oceania, South Africa

•	Global output rising and timber prices falling

•	Regional results suggest winners and losers, but dependent
on climate scenarios.

•	Management of forest stocks complicated by disturbance.

-	Large scale disturbances already influencing outputs in many
regions (Mountain pine beetle in Canada, Forest fires in Russia,
etc.).

-	Disturbance patterns expected to change with climate change.

-------
Disturbance and Adaptation.

• US and Canada example...

800 n

700

_ 600

TO
O)

? 500
a;

Q.

m 400
E

o 300

200

100

0

US

Canada

US A2

1960 1980 2000 2020 2040 2060 2080 2100

-------
US: Ecosystem models projects a stock increase, but
economic model projects a decrease in output...

Aboveground C declines
from the beginning.



0.30



0.25

0J

0.20

two



c

(D

0.15

U

0.10

"ni
c
o

0.05

'¦E

0.00

o



Q.
O

(0.05) -

Q_

(0.10) -



(0.15) -



(0.20) -

US - Change in Aboveground C

(% change relative to 2005)

2005

2025

2045

2065

2085

Forest rises a bit over time

-------
Canada: Ecosystem models project that
stocks decline, but output increases

Aboveground C declines
from the beginning.



0.20



0.10

-------
Summary and Key Limitations of

Analysis

•	Newer analysis has different scale of effects (smaller)
and different regional implications.

•	Economic analysis is evolving relatively slowly.

•	Timber markets may not be most important demand on
forestland in the future.

•	Models are deterministic.

•	Ecosystem models are calibrated without human
influences.

-------
Valuing the impacts of climate change on terrestrial ecosystem services

Alan Krupnick and David McLaughlin
Resources for the Future

Improving the Assessment and Valuation of Climate Change Impacts for Policy

and Regulatory Analyses

Capital Hilton, Washington, DC
January 27-28, 2011

Climate change is already having impacts on terrestrial ecosystem services, according to the
IPCC, the Millennium Ecosystem Assessment and many other scientific reports, and such
impacts are only expected to broaden and worsen as greenhouse gas emissions (GHGs) continue
at their historic levels. To set appropriate policies for reducing GHG emissions, economists
recommend the use of cost-benefit analysis to help decide on the appropriate stringency of
policies, such as the size of a cap in a cap and trade system, the size of a carbon tax, or the
stringency of a carbon fuels standard. To perform such analyses, the predominant approach has
been to use integrated assessment models (IAMs), such as DICE. However, these models lack
geographic specificity, must make hugely simplifying assumptions to capture the myriad effects
caused by climate change and the welfare losses associated with them and not all components are
based on public preferences. As such, there is a need for more targeted valuation studies to serve
as further evidence about the willingness to pay (WTP) to reduce climate change.

The purpose of this brief paper is to sample and classify the literature valuing terrestrial
ecosystem services and make some judgments about its usefulness to benefits analysis associated
with climate change mitigation. As the valuation literature relevant to all types of terrestrial
ecosystem services is enormous, this review is limited to studies valuing ecosystems, primarily
nonuse values, which are likely to provide the largest aggregate values of any service one would
label as based on terrestrial ecosystems (defined as land, river, and lake-based systems,
excluding coastal and saltwater systems). With the emphasis on nonuse values, this paper
focuses on stated preference studies, but also gives some attention to use values, and so includes
revealed preference studies, such as those on recreation.

Classification of Ecological Endpoints Associated with Climate Change

Prior to the entrance of climate change into the valuation literature, this literature was mainly
focused on ecological endpoints related to acid rain, ozone, land use change from urbanization,
dam creation/removal, etc. This literature has relevance to climate change valuation, to be sure;
yet it is inadequate for several reasons. First, there are novel types of ecosystem effects
associated with climate change, such as shifts in the range of a species or subtle perturbations in

1

-------
ecosystem function related to incremental seasonal changes, such as from early alpine snowmelt.
Second, climate change may effect geographical locations that have not been previously studied
for valuation purposes. Third, climate change may produce larger scale effects (e.g., mass
extinctions rather than one at a time). Fourth, the geographic scale of effects related to climate
change, even if these effects are familiar, may be much larger, and the time phasing of these
effects may take longer to begin and longer to reach a new equilibrium.

Fortunately, there are a variety of studies (e.g., IPCC report) that classify the full range of
ecosystem damages associated with climate change. Unfortunately, these classifications involve
much double-counting and contain endpoints that the public would be unable to value (Boyd and
Krupnick, 2009) because they are inherently complex, require advanced scientific knowledge or
are too far from their experience. It is beyond the scope of this paper to develop a complete
classification system that includes only "valuation-relevant" endpoints and eliminates double
counting, meaning that inputs to the processes affecting valuation-relevant endpoints, as well as
the processes themselves, will not both be counted. Take for example an input and process such
as submerged aquatic vegetation (SAV) (an input) and its provision of shelter to fish eggs and
hatchlings (a process). Changes to the SAV affect its ability to provide shelter, which will
ultimately affect the fish populations, the endpoint in this example. In the paper, we argue that
people most easily and reliably value, and understand such endpoints and counting only them
avoids double-counting.

For this literature review of valuation studies, we nevertheless needed some classification
framework of relevant endpoints drawn from the scientific literature on climate change. These
endpoints are not meant to be comprehensive. They appear along the top row of table 1,
covering, first, recreation-related endpoints, such as fish populations, snow cover in ski areas,
tourism, etc. Another category is related to "standard" nonuse values associated with species,
which could include plant, bird, mammal and various aquatic species and cover changes in the
population size, whether they are endangered or facing extinction and, looking across many
species living interrelated in an ecosystem, measures of biodiversity. The term "standard" is
used to indicate that endpoints under these subcategories are quite familiar to economists
engaged in their valuation. The next set of categories is related to combinations of endpoints.

Three subcategories are highlighted that appear to be related to climate change, and other drivers
of disturbance (e.g., land use or management changes). These include the changes in endpoints
associated with wildfires and other events related to climate change, as well as aggregations of
endpoints associated with climate change, such as large changes in biodiversity or mass
extinctions. The last subcategory, "complete," is included to capture combinations of endpoints
that include all the major categories of endpoint changes identified in the scientific literature
(such as IPCC reports). The last category are endpoints that are unique to climate change (or at
least long lasting changes in weather patterns), such as changes in the range of a species or
ecosystem and perhaps some of the more subtle changes in an ecosystem associated with early

2

-------
snowmelt (e.g., changes in migratory patterns).1 How one classifies endpoints of the same type
that are affected over larger time periods or geographical area is arbitrary (are they "unique" or
are they "standard"?). Not noted above, but essential to be mindful of, is the large degree of
uncertainty associated with the magnitude and timing of climatic effects relative to the more
modest uncertainty associated with, say, effects related to acid rain or ozone.

Table 1. Terrestrial Ecosystem Studies by study design and commodity: Classification of Reviewed
Literature

Classification of the Valuation Literature

Table 1 contains a classification framework for the valuation literature that is somewhat
unorthodox, in that it does not distinguish these studies by whether they are stated or revealed
preference or meta-analysis, etc. The classification in the first column of table 1 relates to the
credibility of cost-benefit analyses that would use this literature, beyond that of the methodology
itself.

1 Note that there are positive effects of climate change predicted for terrestrial ecosystems, such as faster tree growth
(at least for a time). For simplicity, we ignore these in the discussion.

3

-------
The top category covers studies that are designed to elicit WTP to avoid most, if not all,
ecosystem-related changes expected from climate mitigation policy. The breadth of this
"commodity" is so wide that only stated preference approaches can be used. The second
category is studies valuing ecosystem endpoints in a climate change context. These are studies
that were designed with the idea of valuing the types of ecosystem changes thought to be
associated with climate change. Nevertheless, even if designed this way, they may not actually
mention climate change. They are applicable to specific locations and types of terrestrial
ecosystems, (e.g., the Murray River ecosystem or Colorado forests). The third category covers
studies that have applied findings from studies valuing changes in terrestrial ecosystem services
in a non-climate change context to estimated climate change impacts in a benefit transfer
exercise. Before writing this paper, we were unaware whether many such studies existed and
were surprised to learn that they do.

The final category (at the bottom of the first column) refers to studies valuing relevant endpoints
in a non-climate change context. In this category is basically the entire ecosystem valuation
literature that is motivated by non-climate change problems (e.g., acid rain). This literature
provides values for a large number of the types of effects associated with climate change, e.g.,
species extinction, but might lack the scale or magnitude of effects associated with policies to
mitigate climate change.

Turning to the interior of table 1, an "X" means that studies of one of the four types apply to the
ecosystem endpoints indicated in the columns. By definition, the top category has an X only
under "Complete" (even though their descriptions may be far from complete). And by
definition, there are blank cells for the bottom row under the Complete and Unique columns.

The literature search that supports these X's was conducted using standard Google and Google
Scholar searches, augmented by reference lists found in studies, as well as the Environmental
Resource Valuation initiative (EVRI) database. We make no claim that this search was
comprehensive. But, we feel we have a reasonable handle on the literature.

The surprise in the table is that there are X's in so many cells. Missing is benefit transfer studies
for disturbances, but this may be occurring because of the limitations of our literature review.

Results from examining this literature

The following tentative findings emerge.

Timing. Because of the long lead times associated with the onset of some types of climate
changes (or at least their most severe manifestations) and their potentially long duration, how
preferences are related to this type of timing is important. While there are few studies upon
which to make firm conclusions, in these it appears that the timing of the benefits of climate
change mitigation doesn't seems to matter. That is, the longer term and the very distant future
appear to be treated equally with respect to willingness to pay, implying very low or zero

4

-------
discount rates (Layton and Brown, 1999; Fleischer et al., 2006). These findings are consistent
with the general tenor of the literature on temporal preferences. However, in our recent
experience with focus groups comparing WTP for commodities offered in the near term (10
years) versus those offered further into the future (50 years), the latter timing creates scenario
credibility problems. Respondents tend to believe that ecological improvements are less likely
to occur the further into the future such changes are offered (Boyd and Krupnick, 2009). In a
field study, statistically distinguishing such behavior from normal discounting would be very
challenging. Some studies simply punt on the issue of communicating long-term changes and
bring such changes into the near-future, or within the lifetime of the respondent, thereby
overestimating WTP, assuming any positive discounting.

Scope sensitivity. With the profession moving more and more to choice experiment formats,
scope sensitivity is now generally limited to showing that there is a positive and statistically
significant coefficient on an attribute (i.e., that people are willing to pay statistically more for
larger reductions in the same commodity). Such tests are run with panel data, where each
respondent answers multiple choice questions, each with different levels of at least one attribute,
including an associated cost.2 This approach is less restrictive than the split sample set-up
recommended by the NOAA Panel for contingent valuation studies.

Nevertheless, with this set-up, the studies reviewed here indicate that scope sensitivity is
generally demonstrated, and further that there is decreasing marginal willingness to pay for
increased number of species protected or for other metrics of increasing ecosystem services.

Uncertainty. Science has limited ability to predict both the future status quo effects from climate
change and the ecosystem improvements arising off this baseline following a given GHG
reduction. Thus, characterizing this scientific uncertainty in stated preference studies is
important. Very few of the studies we reviewed explicitly vary the certainty with which
mitigating actions will improve ecosystem qualities or quantities. Indeed, most appear to treat
ecosystem improvements as if they would occur with certainty. In our focus group experience,
admitting to uncertainty in ecosystem improvements from an intervention scenario results in
respondents' questioning the science or the survey creator's understanding of the science, which
itself results in lower or zero bids from some people. Statistically separating this type of
"protest" bid from the normal behavior of being willing to pay less for a commodity that has a
non-zero probability of being realized (relative to the same commodity offered with certainty) is
another major challenge.

A Tempting Option. The studies classified as "top-down" (see row one of table 1) are a very
tempting alternative to the messy and almost impossible business of doing very detailed
valuation studies in many habitats and using benefit transfer to fill in the rest. The existing
literature in this category covers studies that ask for WTP for reducing greenhouse gas emissions

2 Occasionally, studies bundle several terrestrial ecosystem services to account for tradeoffs between valuing
different ecosystem services. (Riera et al., 2007)

5

-------
and avoiding the consequences of climate change. Some studies ask for the WTP to offset air
travel (Brouwer et al., 2008), or for taking mitigation actions (Akter and Bennet, 2008), or to
reduce dependence on foreign oil and carbon emissions (Li et al., 2009), to implement the Kyoto
Protocol (Berrens et al., 2004), or more explicitly (Berk and Fovell (1999)) to prevent significant
climate changes. Cameron (2005) used a convenience sample of college students and found that
respondents who are more certain about a given increase in average temperatures have higher
WTP to prevent such an increase. In line with these results, Viscusi and Zeckhauser (2005) and
Akter and Bennett (2008) also found that people who find global warming to be more likely also
have higher WTP. Hence, as might be expected, one important explanatory factor for how much
individuals are willing to pay for mi tigating climate change is if they believe in climate change.

Details on a Broad Climate Mitigation Valuation Study (Carlsson et al, 2010)

In a survey performed by Carlsson et al. (2010), respondents were told that the magnitude of
future temperature increases will depend on the amount of future global C02 emissions;
specifically, if C02 emissions are reduced from current emission levels by 30%, 60%, and 85%
respectively, then the temperature increase will be limited to 4°F, 3°F, or 2°F. If the world
instead does not reduce emissions but continues with "business as usual" (BAU), the temperature
is expected to increase by more than 4°F in 2050. The survey explained, based on information
from the IPCC, that this would most likely correspond to large changes in the global ecosystem
and most countries would be negatively affected. An information screen (figure 1) summarized
these effect of temperature increases on harvests, increased flooding and storms, and ecosystem
effects by the year 2050.

Global emissions
reduction



85% reduction



60% reduction



30% reduction

Temperature increase

2°F increase

3°F increase

4°F increase

Harvest

Harvests in countries
near the equator
decrease by 4-6%.
Harvests in countries in
the northern
hemisphere increase by
1-3%.

Harvests in countries
near the equator
decrease by 10-12%.
Harvests in countries in
the northern
hemisphere are
unaffected.

Harvests in countries
near the equator
decrease by 14-16%.
Harvests in the northern
hemisphere decrease
by 0-2%.

Increased flooding and
storms

Small tropical islands
and lowland countries
(for example,
Bangladesh) experience
increased flooding and
storms.

Additional low-lying
areas in the Americas.
Asia, and Africa
experience increased
flooding and storms.

Populous cities face
increased flood risks
from rivers and ocean
storms.

Existence of small
island countries is
threatened.

Threatened ecosystems

Sensitive ecosystems,
such as coral reefs and
the Arctic, are
threatened.

Most coral reefs die.
Additional sensitive
ecosystems and
species around the
world are threatened.

Sensitive and less-
sensitive ecosystems
and species around the
world are threatened.

6

-------
Figure 1. Global Emissions Reduction, Temperature Increase and Its Effects as Presented to
Survey Respondents in Carlsson et al. (2010)

This survey screen is a good example of responses to the above key issues (and others). Overall,
the "commodities" being valued encompass both ecosystem effects as well as harvests and
storms/flooding, with some specific species - coral reefs - and specific locations called out.
These choices are in line with the IPCC's "most likely" predictions. Timing of the effects of
global warming is set at 2050, a simplification and forward telescoping of the path of effects we
might see. Uncertainty is handled obliquely by providing ranges of likely effects (e.g., 1-3
percent reduction in harvests) in the table, and using words like "most likely" in the text.
However, in the information screen above (which also doubles — with slight modifications - as
the choice experiment screen) declarative phrases are used, e.g., "most coral reefs die", as
opposed to using qualifiers on the verb, such as "may die." These choices were guided by focus
group feedback. Probably the most subtle approach to uncertainty is with description (or lack of
description) of the future baseline, of which is said only that temperature change of greater than
4 degrees F ".. .would most likely correspond to large changes in the global ecosystem and most
countries would be negatively affected" (Carlsson et al., 2010). This decision to so vaguely
define the baseline also was made in response to focus group feedback and the difficulty of
concisely describing widespread and possibly dramatic effects. Finally, the survey passed scope
sensitivity, although as designed the test was of internal (rather than external) scope sensitivity.

In any event, the survey yielded many interesting results. Carlsson et al. (2010) found that a
large majority of the respondents in all three countries believe the mean global temperature has
increased over the last 100 years and that humans are responsible for the increase. Americans,
however, believe less in both aspects compared to Chinese and Swedes. A larger share of
Americans appears to be pessimistic; they believe that we cannot do anything to stop climate
change. Carlsson et al. (2010) also found that Sweden has the highest WTP for reduction of
C02 and China has the lowest. In going from a 30% reduction in GHGs to 60%, for instance,
the Swedes were willing to pay $20 per household month, while the U.S. sample was willing to
pay $10 and the Chinese $4. Interestingly, when the WTP is measured as the share of household
income, the willingness to pay is about the same for American and Chinese sample, but much
higher for the Swedes.

The findings from Carlsson et al. (2010) show that the U.S. population contains a far larger
percentage of climate skeptics (24 percent by one metric) than in Sweden or even China (both
around 5-6 percent). With such a large fraction of the population thinking this way, survey
researchers must be concerned that WTP for ecosystem improvements themselves, irrespective
of the cause, not be biased downwards simply because respondents discount the link between
climate change and the ecosystem changes being offered or biased upwards through double-
counting due to the respondent's inclusion of joint benefits (e.g., human health) in their WTP.

7

-------
This situation is not unlike that faced by researchers seeking values for reducing mortality risks
from air pollution. Best practice is to avoid conveying the cause of reducing such risks (e.g.,
reducing particulate emissions) because such reductions carry with them cognitive linkages to
other types of improvements (say in morbidity or visibility or materials damages), which tend to
inflate WTP, or feelings leading to scenario rejection or downward bias. The most common of
these feelings in our focus group work has been that respondents should not be responsible or
pay for reductions in air pollution because it is "industry's fault." Both of these problems -
linkage bias and scenario rejection - are likely to be present to an even greater extent in climate
change valuation studies relative to valuation studies for conventional air pollutants.

This situation could lead to more efforts, such as that from MacDonald et al. (2010) on the
Murray River region in Australia, where climate change is not even mentioned as a causal driver
of change, and mitigation of GHGs is not mentioned as a motivator for ecosystem improvements.
It remains to be determined whether plausible alternative stories can be constructed for
delivering such widespread and large improvements in ecosystems as are thought to be realizable
from large GHG reductions. If they can't, perhaps the best approach would be to include
questions to determine whether or to what degree a respondent is a climate skeptic and to adjust
for this statistically or through their exclusion. To do so, however, risks overvaluing the
improvements, as legitimate zeros or low values may be excluded.

Conclusion

Is this literature ready for prime time, i.e., to be used to help develop and justify a social cost of
carbon? A top level response is that, with the pervasiveness of the effects of global warming on
all types of natural and human systems, and given the interconnectedness of those systems, it
seems too reductionist to focus on valuation of changes to specific resources or systems, in this
case terrestrial ecosystems. That is, the value of slowing climate change needs to be estimated
from a holistic perspective. To do so, the only possible way to go is with the top-down studies
like those defined in the first row of table 1, recognizing that these studies can never provide the
detail and the preciseness of commodity definition that is desirable in, say, natural resource
damage assessments. However we must ask ourselves as a society if we are willing to trade off
precision for comprehensiveness/breadth.

What is the alternative? In our view, the vast literature simply valuing ecosystem services is not
largely motivated or directly applicable to climate change. And use of these studies in benefits
transfers therefore involves huge assumptions (e.g., about how much two extinctions are worth
avoiding relative to one) and, even then, there will be gaps in geographic coverage. On a more
hopeful note, there are an increasing number of ecosystem valuation studies motivated by
climate change that have the right scale and type of commodities being valued (see table 1, row
2). Yet, such studies are invariably place-based and draw relatively tight, rather than porous

8

-------
boundaries between the ecosystem of interest and its linkages to other systems. Thus, one will
not be able to easily aggregate such studies, properly account for overlaps and gaps and
eventually come out with a cost of carbon. However, examining such studies one at a time and
drawing insights out of them may both inform the design of information treatments for the
studies, like those found in the first row of our classification, and lead to a more
qualitative/judgmental basis for settling on a cost of carbon number.

References

Akter, S., and Bennet, J. 2009. "Household Perception of Climate Change and Preferences for
Mitigation Action: The Case of the Carbon Pollution Reduction Scheme in Australia,"
Research Report No. 19. Canberra: Australian National University, Crawford School of
Economics and Government. ISSN 1835-9728.

Berk, R.A., and R.G. Fovell. 1999. "Public Perceptions of Climate Change: A Willingness to Pay
Assessment," Climatic Change, 41: 413-446.

Berrens, R.P., A.K. Bohara, H.C. Jenkins-Smith, C.L. Silva, andD.L. Weimer. 2004.

"Information and Effort in Contingent Valuation Surveys: Application to Global Climate
Change Using National Internet Samples," Journal of Environmental Economics and
Management, 47(2): 331-336.

Boyd, J. and A. Krupnick. 2009. "The Definition and Choice of Environmental Commodities for
Nonmarket Valuation," RFF Discussion Paper 09-35, Resources for the Future,
Washington D.C.

Brouwer, R., L. Brander, and P. van Beukering. 2008. "A Convenient Truth: Air Travel

Passengers' Willingness to Pay to Offset Their C02 Emissions," Climatic Change 90(3):
299-313.

Cameron, T.A. 2005. "Individual Option Prices for Climate Change Mitigation," Journal of
Public Economics, 89(2-3): 283-301.

Carlsson, F., M. Kataria, A. Krupnick, E. Lampi, A. Lofgren, P. Qin, S. Chung, and T. Sterner.
2010. "Paying for Mitigation: A Multiple Country Study," Working Papers in
Economics, no. 447. Gothenburg, Sweden: Department of Economics, University of
Gothenburg.

Fleischer, A. and M. Sternberg. 2006. "The Economic Impact of Global Climate Change on
Mediterranean Rangeland Ecosystems: A Space-for-Time Approach," Ecological
Economics, 59:287-295.

9

-------
Layton, D., G. Brown, and M. Plummer. 1999. "Valuing Multiple Programs to Improve Fish
Populations," Washington Department of Ecology. Olympia, WA.

Li, H., H.C. Jenkins-Smith, C.L. Silva, R.P. Berrens, and K.G. Herron. 2009. "Public Support for
Reducing US Reliance on Fossil Fuels: Investigating Household Willingness-to-Pay for
Energy Research and Development," Ecological Economics, 68: 731-742.

MacDonald, D. H., M. Morrison, J. Rose, and K. Boyle. 2010. "Valuing a Multi-State River: The
Case of the River Murray," Forthcoming in the Australian Journal of Agricultural and
Resource Economics.

Riera P., J. Penuelas, V. Farreras, M. Estiarte. 2007. "Valuation of Climate Change Effects on
Mediterranean Shrublands," Ecological Applications, 17(1), 91-10.

Viscusi, W.K., and R.J. Zeckhauser. 2006. "The Perception and Valuation of the Risks of
Climate Change: A Rational and Behavioral Blend," Climatic Change, 77: 151-177.

10

-------
Valuing the impacts of climate
change on terrestrial ecosystem

services

Alan Krupnick
Resources for the Future

Improving the Assessment and Valuation of Climate Change
Impacts for Policy and Regulatory Analyses

Capital Hilton, Washington, DC
January 27-28, 2011

RESOURCES

FOR THE FUTURE

-------
Definitions and scope

•	Terrestrial: everything but coastal and ocean

•	Here, my focus is on the squishiest of ecosystem
services: non-use values

¦ Stated preference (survey-based) studies. A low WTP
per person goes a long way!

•	Endpoints: biophysical effects estimated by
natural scientists that are used as startpoints in
valuation studies

-------
The task

Based on climate drivers

Based on preferences ~ WTP A mil ion of them. Which to choose?

SCCT = Unit values * A Endpoint

RESOURCES

FOR THE FUTURE

-------
Issues

•	Does the natural science examine the
appropriate endpoints and build the
appropriate functional relationships to link
back to climate variables and interventions?

•	Are those endpoints valued? Credibly
valued? Are the valuation studies
comprehensive enough?

FOR THE FUTURE

-------
ENDPOINTS AT ISSUE:

IPCC

o

Global mean annual temperature change relative to 1980-1999 (°C)

12	3	4

Increased water availability in moist tropics and high latitudes

Decreasing water availability and increasing drought in mid-latitudes and semi-arid low latitudes ™ ™ ¦
Hundreds of millions of people exposed to increased water stress ——————————————

Up to 30% of species at	Significant extinctions,

increasing risk of extinction	around the globe

Increased coral bleaching Most corals bleached	Widespread coral mortality

Terrestrial biosphere tends toward a net carbon source as:
-15% ^-40% of ecosystems affected

Increasing species range shifts and wildfire risk

Ecosystem changes due to weakening of the meridional — a
overturning circulation

5 °C

WATER

ECOSYSTEMS

RESOURCES

FOR THE FUTURE

-------
ECOSYSTEM SERVICES

Provisioning

FOOD

FRESH WATER
WOOD AND FIBER
FUEL

Supporting

NUTRIENT CYCLING
SOIL FORMATION
PRIMARY PRODUCTION

Regulating

CLIMATE REGULATION
FLOOD REGULATION
DISEASE REGULATION
WATER PURIFICATION

Cultural

AESTHETIC
SPIRITUAL
EDUCATIONAL
RECREATIONAL

/

LIFE ON EARTH - BIODIVERSITY

ARROW'S COLOR
Potential for mediation by
socioeconomic factors

Low
Medium

ARROW'S WIDTH

Intensity of linkages between ecosystem
services and human well-being

i	Weak

~ Medium

CONSTITUENTS OF WELL-BEING

Security



PERSONAL SAFETY



SECURE RESOURCE ACCESS



SECURITY FROM DISASTERS



Basic material



for good life

Freedom

ADEQUATE LIVELIHOODS

of choice

SUFFICIENT NUTRITIOUS FOOD

and action

SHELTER

ACCESS TO GOODS

OPPORTUNITY TO BE

ABLE TO ACHIEVE
WHAT AN INDIVIDUAL



Health

STRENGTH

VALUES DOING
AND BEING

FEELING WELL



ACCESS TO CLEAN AIR



AND WATER



Good social relations



SOCIAL COHESION



MUTUAL RESPECT



ABILITY TO HELP OTHERS



Source: Millennium Ecosystem Assessment

-------
Ecological Production

Theory

Same thing

¦	Biophysical inputs

¦	Transformed via natural processes into

¦	Biophysical outputs
Qi =/(Iu, Ii2,...)

-------
Production Function Error

•	What is the value of "more acres of eagle
habitat?"

•	Need to know two things

(1)	The value you place on eagle abundance

(2)	The production function that translates eagle
habitat into eagles

Respondents will intuit a relationship
But won't know magnitude

RESOURCES

FOR THE FUTURE

-------
Startpoint Categories for
Climate Change

•	Use (e.g., fish populations)

•	"Standard" non-use (e.g., single species
population change, extinction)

•	Combinations associated with events (e.g.,
wildfires) or broad scale changes (e.g.,
desertification)

•	Novel changes (e.g., range shift, mass
extinctions)

£ RESOURCES

FOR THE FUTURE

-------
Valuation studies classification

•	Studies valuing relevant commodities in
non-climate context

•	Studies transferring these values to a
climate change context

•	Studies valuing relevant commodities in a
climate change context

•	Stated preference top-down studies

FOR THE FUTURE

-------
Standard Endpoints

Please vote:

The following vote offers a choice between No Program and an Improvement Program
oof.or?. The future conditions of the SAM Region by 2019 for each choice are
summarized below. What is your vote?



No Program

Improvement Program

Streams and fish

150,000 healthy streams: 150,000
streams of concern.

Up to 6 species of fish harmed in
streams of concern

20% of all streams (60,000 streams of
concern) will improve and be stocked
with these fish.

*

Bird populations

Three songbird species are 65% of
what they once were.

Three songbird species improve to
85% of what they once were

X

Forests

3% (780 000 acres) of SAM Region
has damage to red spruce and sugar
maple trees

1% of SAM Region (260,000 acres)
improves

Your additional state
tax

SO

S500 in total (S50 per year for 10
years)





Your Vote?



o



-------
Murray

River

Watershed

Boyle et aL
2010





Option A

Current
Condition

Option B

Option C

Native fish





10% of original
population

20% of original
population

40% of original
population

Healthy
Riverside
Vegetation and
Wetlands



f/f

pm ri

ij„, | i. jt

icq

50% of
original area

178,000 ha

60% of
original area

200,000 ha

70% of
original area

240.000 ha





























Frequency of

Waterbird

breeding





Every 10 years

Every S years

Every 3 years





























Coorong and
Lower Lakes





Coorong
declining rapidly



Coorong
declining rapidly



	'—

. . T*

Coorong healthy





Time until

improvement

occurs

CB





5 years

20 years

Household cost

per year for 10
years

$

SO

SI 00

S250

-------
Fleischer and
Sternberg, Ecol.
Econ, 2006

RESOURCES

FOR THE FUTURE

Program 1
No action

Program 2
Forcstation is
used to slow
down

erccnhousc
effect

Program 3
Reduction in
the use of
greenhouse
gases

Program 4
Forcstation and
greenhouse-gas
reduction

Program 5
Drastic
reduction in
greenhouse
eases

Landscape in
the Galilee3
will become
arid, also loss
of plant life
will occur

Landscape in
the Galilee3
will become
semiarid

Landscape in
the Galilee3
will become
semiarid

Landscape in
the Galilee3
will have less
plant life

Landscape in
the Galilee3
will not
chance

$0 per month
Mesic

Mediterranean





Arid

$7.5 per month
Mesic

Mediterranean

$7.5 per month
Mesic

Mediterranean

$ 15 per month
Mesic

Mediterranean

$20 per month

Mesic

Mediterranean

Semiarid

Semiarid

Mediterranean

-------
Table 1. Global Emission Reduction, Temperature Increase, and Its Effects as

Presented to Survey Respondents

Global emissions
reduction



85% reduction



60% reduction



30% reduction

Temperature increase

2°F increase

33F increase

4°F increase

Harvest

Harvests in countries
near the equator
decrease by 4-6%.

Harvests in countries in
the northern
hemisphere increase by
1-3%.

Harvests in countries
near the equator
decrease by 10-12%.

Harvests in countries in
the northern
hemisphere are
unaffected.

Harvests in countries
near the equator
decrease by 14-16%.

Harvests in the northern
hemisphere decrease
by 0-2%.

Increased flooding and
storms

Small tropical islands
and lowland countries
(for example,
Bangladesh) experience
increased flooding and
storms.

Additional low-lying
areas in the Americas,
Asia, and Africa
experience increased
flooding and storms.

Populous cities face
increased flood risks
from rivers and ocean
storms.

Existence of small
island countries is
threatened.

Threatened ecosystems

Sensitive ecosystems,
such as coral reefs and
the Arctic, are
threatened.

Most coral reefs die.

Additional sensitive
ecosystems and
species around the
world are threatened.

Sensitive and less-
sensitive ecosystems
and species around the
world are threatened.

-------
Figure 1, Distribution of WTP Responses for 30% Greenhouse Gas Reduction

-------
Usefulness of literature

•	Existing "non-climate" studies - useful but
limited

•	BT with above studies: artificial and
assumption-based

•	Climate-driven studies: useful, growing
literature, but will always be "patchy"

-------
Top-level studies as
tempting option

•	Broad coverage of endpoints and locations

•	But highly imprecise commodity definitions
and scenarios

•	What's the alternative?

¦ Perhaps benefits transfers from well-done
climate-based valuation studies.

-------
Classification of the Valuation

Literature

3.) Studies
transferring
values to a
climate

context
4.) Studies

relevant
endpoints in a
non-climate

Recreation

Species

Disturbances

Multiple
commodities

Complete

Events

e.g. Fishing,
skiing,
hunting,
beach

Population
change

Endangered
or facing
extinction

Decreased
biodiversity

e.g. Wildfires

e-g-

Biodiversity
and mass
extinctions

All/Most
relevant
commodities

Range or
ecosystem
shift

Early snow
melt impacts













X





X

X

X

X

X

X



X



X

X

X

X



X



X

X

X

X

X

X

X

X

X





-------
Results

• Most cells filled in ^ a lot of studies to
work with for meta-analyses and benefit-
transfer

-------
Spatial Scale

¦	Studies range widely in spatial scales

¦	Desire for specificity to enhance credibility:

> "tangible" commodities and convincing scenarios

-------
Scope Sensitivity

¦	WTP more for avoiding larger
damages/gaining larger benefits

¦	Decreasing marginal returns

-------
T iming

¦	Timing of benefits doesn't seem to matter
much

¦	Low discount rates

¦	Not addressed by many studies

-------
Uncertainty

Most assume certainty

Very few vary uncertainty

Admitting to uncertainty may induce protest
bids

¦	Rejection of science or survey

¦	Difficult to sort out from "legitimate"
responses

RESOURCES

FOR THE FUTURE

-------
What is needed

•	From Ecologists: Endpoints that match
valuation startpoints and have functional
relationships with climate drivers

•	From Economists: consensus approach to
classifying endpoints to be used as
valuation startpoints

-------
Final thoughts

•	Should surveys mention climate change?

¦ Climate skeptics

•	How to admit uncertainties in surveys?

•	Need holistic valuation estimates (more
than just terrestrial ecosystem effects) - no
presumption of additivity top down SP
studies? Or top down SP studies for non-
market ES only?

FOR THE FUTURE

-------
Energy System Impacts of Climate Change: An Overview

for

EPA/DOE Workshop on Improving the Assessment and Valuation of Climate Change Impacts

for Policy and Regulatory Analysis
Research on Climate Change Impacts and Associated Economic Damages

Washington DC
January 27-28, 2011

Howard K. Gruenspecht1
U.S. Department of Energy

Assessment and valuation of the impacts of climate change on energy systems, including both
effects on energy demand and effects on energy supply systems, have received considerable
attention over the past 25 years. While the literature encompasses a wide range of results,
recent assessments, including high-profile reports such as the Stern Review and the IPCC 4th
Assessment Report that have identified a high likelihood of significant adverse impacts in other
areas, have generally found modest impacts, both positive and negative, on energy systems.
There is nothing in the more recent literature that suggests any major change in that
assessment. Below, we review impacts on energy use for space heating and cooling which have
been considered in most analyses of energy demand impacts, as well as other potential effects
on energy demand. We also consider effects on energy supply systems, both new and existing.
We conclude with summary observations about the analysis of energy impacts to date and
identify factors that may be important in extending the literature.

Impacts on Space Conditioning Energy Demand: The most direct way in which climate change
potentially affects energy demand is through its effect on energy use for heating and cooling.
Some early studies of impacts on energy demand in the United States focused exclusively on
the demand for electricity for cooling in the summertime. Subsequently, several papers noted
that from the space conditioning perspective, the United States is a cold country, with
expenditures on winter heating fuel several times higher than expenditures on electricity for
cooling, and that some degree of warming would likely decrease overall demand and
expenditures for space conditioning energy. The traditional grouping of "industrialized
countries" ~ the OECD countries plus Russia and Eastern Europe have an even larger gap
between their baseline energy use and expenditures for heating and cooling, so initial
warming is likely to provide savings in energy use and expenditures for space conditioning in

1 The views expressed in this note, which draw on the author's past involvement with the literature on neregy
impacts of climate change, go beyond topics that fall within the purview of the Energy Information Administration,
where he now serves as Deputy Administrator. They should not be construed as reflecting the views of that
agency.

Energy System Impacts of Climate Change

Page 1

-------
those regions as well. The developing countries, which include both tropical and non-tropical
areas, present something of a mixed bag, in part because the use of cooling equipment is highly
sensitive to economic development as well as local climate conditions.

Any analysis of the impact of climate change on space conditioning energy use is likely to be
highly sensitive to both the magnitude of climate change considered, and its detailed
composition. The latitudinal, diurnal, and seasonal gradient of warming and changes in relative
humidity all play a crucial role in determining whether warming reduces or increases energy use
and/or expenditures for space conditioning at any particular location, or cumulatively across
any set of locations. In part, the spread of results across studies on space conditioning impacts
reflects different approaches to specifying the global warming scenarios that are considered.

Energy expenditure changes and measures of individual or aggregate comfort in buildings, a
welfare indicator, may diverge considerably. On the one hand, the change in capital and
energy expenditures for space cooling in a higher temperature and humidity scenario is likely to
overstate the cost of maintaining a constant indoor summertime comfort level for those who
acquire new space conditioning equipment in the face of climate change. Space cooling, unlike
space heating, is subject to very significant threshold effects, even in relatively rich countries.2
Once installed, cooling equipment is likely to be used to provide improved comfort relative to
that which householders might have accepted under baseline conditions before the threshold
was crossed. However, energy expenditure changes do not reflect the value of incremental
indoor discomfort for those who do not cross the cooling equipment threshold. In addition,
incremental summertime outdoor discomfort for the wider public, are not reflected at all in
changes in space conditioning costs.

Energy implications of changes in space conditioning energy demand, which are of great
interest to energy planners without regard to their value as welfare indicators, must be
assessed in the context of technology changes over relevant time horizons. Assessments of
energy impacts of climate change that are made without consideration of changes in energy
technologies and practices can badly miss their mark. For example, the efficiency of new air
conditioning units has nearly doubled since 1990, when the first studies claiming large impacts
on summer peak energy load due to warming were published. More recently, the prospect of
the smart grid and attendant opportunities to manage load in real time are likely to greatly
ameliorate the implications of higher peak space conditioning loads for the electricity supply
infrastructure, since other loads can now be more readily incentivized to "make room" for
cooling loads.

2 For example, many homes along the California coast, and in Europe, both relatively wealthy regions of the world,
do not have air conditioners. It is possible that climate change could result in the crossing of a comfort threshold
that leads households to install such equipment.

Energy System Impacts of Climate Change

Page 2

-------
Other Energy Demand Impacts: While space conditioning impacts have been the focus of
research on energy demand impacts, some other areas, including the energy-water nexus,
merit additional attention. Significant amounts of energy are used to supply water for
household, agricultural, and industrial purposes, and also to move and treat wastewater from
all categories of water use. The potential impact of climate change on water supply has been
considered elsewhere in this workshop. To the extent that climate change has adverse impacts
on water supply, the need to provide replacement water may have significant energy
implications. Many types of replacement supplies, such as desalinization plants, long-distance
pumping solutions, and cleaning of wastewater to a standard that allows for reuse, can use
significantly higher amounts of energy than is required to supply water under baseline
conditions. The issue can be important in both developed and developing country contexts.

Energy Supply Systems: Access to Traditional Energy Resources. It is well known that climate
change can have significant impacts on access to traditional energy resources. For example,
hydroelectricity, by far the most significant source of renewable electricity in both the United
States and the world today, is quite sensitive to patterns of precipitation and snowpack
accumulation, which are in turn likely to be affected by climate change. The pattern of
impacts is likely to vary across locations, and also to be dependent on both the passage of time
and the extent of climate change. A traditional energy resource where the initial impact on
supply is likely to be positive is the Arctic oil resource, as access would be significantly improved
by a reduction in Arctic ice cover. However, not all northern latitude resources will necessarily
benefit from climate change, as any change in permafrost conditions and the length of that
annual hard freeze period could limit the ability to build and maintain energy infrastructures
needed to access certain energy resources, both onshore and offshore, at high latitudes.

Energy Supply Systems: Impacts on Existing Energy Supply Infrastructure. Another category
of energy supply impacts that has been extensively examined involves existing energy supply
infrastructures that may be affected by changes in temperature or precipitation patterns. In
addition to hydroelectric dams that are directly dependent on water flows, nearly all existing
generating facilities require access to cooling water. As discussed above, climate change
impacts are likely to include changes in precipitation, snowpack and evaporation patterns that
will affect water availability and temperature in and around existing power plants. A change in
cooling water availability and temperature can affect power plant operation. Changes in
ambient air temperature can also affect the effective maximum capacity of existing units.
However, it is questionable whether or not any of these impacts are quantitatively important in
the overall context of climate change impacts after consideration of actions to mitigate and/or
adapt to them.

Energy System Impacts of Climate Change

Page 3

-------
Energy Supply Systems: Impacts on Non-Traditional Energy Sources. Given that energy-
related emissions from the combustion of fossil fuels account for at least three-fifths of
anthropogenic greenhouse gas emissions globally, and more than four-fifths of U.S. emissions,
strategies to mitigate emissions often focus on the replacement of fossil fuels with emissions-
free energy sources. Expanded use of wind, solar, and biomass energy for electricity
generation, and the use of biofuels in the transportation sector, are often cited as potential
alternatives to fossil fuels. Given this, it is important to consider the possible impacts of climate
change on these technologies.

With respect to solar, both photovoltaic (PV) and solar thermal technologies are sensitive to
changes in cloud cover. Pan et al. (2004) modeled changes in global solar radiation reaching
the surface through the 2040s based on the Hadley Center Circulation model and projected a
solar resource reduced by as much as 20 percent seasonally in key U.S. regions for solar energy,
presumably from increased cloud cover. The energy assessment published by the U.S. Climate
Change Science Program in February 2008 notes that aerosols can also play a role in cloud
cover, and that interactions between aerosols and greenhouse gases are complex.

Biomass already rivals hydropower as a renewable energy source in the United States, and
mandates for renewable fuel use in transportation first enacted in 2005 and then significantly
strengthened in the Energy Independence and Security Act of 2007 call for biofuels to
significantly grow in both absolute terms and as a share of the total liquid fuels used in
transportation. Biomass also has growth potential in the electric power sector, where it can be
co-fired with coal in existing power plants. Much attention in the recent literature and the
regulatory sphere has focused on the carbon cycle impacts of increased biomass energy use,
which depends on the sustainability of biomass cultivation and proper accounting practices.
The impacts of climate change on the economics of biomass energy are closely related to the
effects of climate change on agriculture, which are addressed in another part of this workshop.

Finally, with respect to wind, there is little information regarding the effects of climate change.
The siting of wind farms and the cost of wind generation are both very sensitive to the specific
location of the wind resource. One question that arises, in addition to the impact of climate
change on overall available wind resource, is if climate change will cause shifts in wind patterns
within the 20-to 30-year lifetime of wind projects.

Concluding Observations:

1. Energy is a sector that is likely to be impacted by climate change. As climate changes

considered grow ever larger, common sense suggests that negative impacts on energy use and
supply will dominate, but for small to modest climate change it is quite possible that net energy
"damages" will be negative.

Energy System Impacts of Climate Change

Page 4

-------
2.	For energy, as in some other sectors where impacts must be assessed, the devil is really in the
details, such as, but not limited to, the assumed latitudinal, seasonal, and diurnal gradient of
climate change, and its effects on humidity, cloud cover, and wind patterns as well as its effect
on temperature. Studies that make different assumptions in these areas can reach wildly
different conclusions even if they are both carefully executed.

3.	Future impacts of climate change on energy systems will occur in the context of future
opportunities for adaptation and responses. While it is hard to predict the future, it is important
to consider the implications of the past track record of technology improvements and the
impact of technologies now being deployed in assessing the cost of adaptation and response
strategies.

4.	It is useful to distinguish between energy system impacts, which are of greatest importance to
energy planners, and energy-system-related welfare impacts, which are of primary importance
to cost-benefit analysis of policies to address climate change.

5.	Both analysts and research funders can advance the utility and credibility of research on energy
system impacts of climate change through a commitment to carefully scope and prioritize
research needs in the area.

Energy System Impacts of Climate Change

Page 5

-------
Energy System Impacts of Climate Change:

Background Slides

for

EPA/DOE Workshop on Improving the Assessment and Valuation
of Climate Change Impacts for Policy and Regulatory Analysis

Washington DC
January 27-28, 2011

Howard Gruenspecht, Deputy Administrator

U.S. Energy Information Administration

Independent Statistics and Analysis

-------
Despite a continued shift of the U.S. population to warmer areas, much
more energy is used to heat buildings than to cool them through 2035

Projected Space Conditioning Energy Use in Buildings, Quadrillion Btu

Residential

2010

2015

2025

2035

Space Heating

4.33

4.27

4.18

4.10

Purchased Electricity

0.29

0.28

0.30

0.31

Natural Gas

3.29

3.27

3.28

3.27

Distillate Fuel Oil

0.50

0.48

0.39

0.33

Liquefied Petroleum Gases

0.26

0.23

0.21

0.19

Space Cooling

1.11

0.82

0.90

0.99

Purchased Electricity

1.11

0.82

0.90

0.99

Natural Gas

0.00

0.00

0.00

0.00

Ventilation

0.14

0.15

0.18

0.19

Commercial

Space Heating

1.93

2.01

2.04

2.04

Purchased Electricity

0.18

0.17

0.17

0.18

Natural Gas

1.61

1.70

1.75

1.76

Distillate Fuel Oil

0.14

0.13

0.11

0.10

Space Cooling

0.62

0.56

0.60

0.65

Purchased Electricity

0.58

0.53

0.56

0.61

Natural Gas

0.04

0.04

0.04

0.04

Ventilation

0.51

0.55

0.64

0.71

Source: EIAAEO2011 Reference case

Howard Gruenspecht, EPA/DOE Workshop on Assessment and Valuation of Climate Change Impacts, 1/28/11

-------
Despite a continued shift of the U.S. population away from cold areas,

projected energy expenditures to heat buildings exceed cooling
	expenditures by a wide margin through 2035	

Space Conditioning Energy Expenditures in Buildings, Billions of Year 2009 Dollars

Residential

2010

2015

2025

2035

Space Heating

63.47

59.90

65.74

69.86

Purchased Electricity

9.66

9.09

9.43

9.83

Natural Gas

36.65

33.74

39.08

44.21

Distillate Fuel Oil

10.33

10.14

10.12

9.03

Liquefied Petroleum Gases

6.82

6.93

7.11

6.80

Space Cooling

37.42

26.31

28.09

31.46

Purchased Electricity

37.42

26.31

28.09

31.46

Natural Gas

0.00

0.00

0.00

0.00

Ventilation

4.66

4.93

5.65

6.17

Commercial









Space Heating

22.12

21.75

24.56

26.94

Purchased Electricity

5.00

4.58

4.59

4.82

Natural Gas

14.59

14.64

17.26

19.52

Distillate Fuel Oil

2.53

2.54

2.70

2.60

Space Cooling

16.94

14.51

15.37

16.97

Purchased Electricity

16.54

14.20

15.02

16.58

Natural Gas

0.40

0.31

0.35

0.39

Ventilation

14.55

14.95

17.05

19.30

Source: EIAAEO2011 Reference case
Howard Gruenspecht, EPA/DOE Workshop on Assessment and Valuation of Climate Change Impacts, 1/28/11	3

eia

-------
Growing Use of Non-traditional
Water Resources

Power Requirements For Treating

"5
5

20

10

o

= £ 15
r S3

c a

S3

=> c

k. Q

0	=

is

T>

B

¦a

1

"o

k.

CL

Waste Water Reuse
Desali nation

/

/

CO
<

E

I
*









I



Today

The Future





/









.f

























































Conventional
Treatment

NF

RO

RO

2020

2000	2010

Year

(Modified from Water Reuse 2007, EPA 2004, Mickley 2003)

(Einfeld 2007)



ENERGY^ W0

cHcBuX

Desal growing at 10% per year, waste water reuse at 15% per year
Reuse not accounted for in USGS assessments
Non-traditional water use is energy intensive

|	-	j

Source: Mike Hightower, Sandia National Laboratory

-------
For more information

U.S. Energy Information Administration home page

www.eia.gov

Short-Term Energy Outlook

www, e i a. q o v/e m e u/steo/p u b/co ntents.html

Annual Energy Outlook

www.eia.gov/oiaf/aeo/index.html

International Energy Outlook

www.eia.gov/oiaf/ieo/index.html

Monthly Energy Review

www.eia.gov/emeu/mer/contents.html

EIA Information Center

(202) 586-8800
Live expert from 9:00 AM - 5:00 p.m. EST
Monday - Friday (excluding Federal holidays)

email: lnfoCtr@eia.gov

ceia; Howard Gruenspecht, EPA/DOE Workshop on Assessment and Valuation of Climate Change Impacts, 1/28/11

-------
Impacts of Climate Change on Global Electricity Production and Consumption:
Recent Literature and a Useful Case Study from California

Jayant Sathaye
Lawrence Berkeley National Laboratory, Berkeley

Abstract

Climate change affects both energy demand and energy supply through various parameters. These
parameters include warmer air and water caused by higher temperatures, changes in flow of rivers,
snowfall and ice accretion, coastal inundation, wildfires, soil conditions, cloudiness and wind speeds.
Increases in energy demand and supply loss create a combined problem for ensuring an adequate
supply of fuels and electricity. Projections of these parameters combined with those of energy demand
and supply over the next century are needed to improve our understanding of the increased
vulnerability of the energy sector. In addition, a detailed physical layout of the various facilities is
necessary to understand the exposure of energy infrastructure to the climate-related challenges.

Despite a potentially significant impact on energy demand and supply, the international literature base
on these topics is very limited particularly in the developing countries and on the supply component. As
a result, this presentation reports on selected international quantitative evaluations of energy demand,
qualitative evaluations of energy supply impacts, and related policy implications. Given the limited
amount of literature on this subject, we discuss an approach that we have used for evaluating the
impact of climate change on the California energy demand and supply systems. We believe this method
could provide insights and form the basis for "bottom-up" evaluations in other countries.

Table 1 shows the hydro-meteorological and climate parameters for selected energy uses. This table
indicates the various connections between the sets of parameters. For example, changes in air
temperature would affect electricity generation efficiency including that of solar PV panels and the
demand for cooling and heating. Robust evaluation of energy supply and demand impacts should
examine each of the listed parameters while also taking into consideration the interconnections
between them. Warmer temperatures may affect generation, transmission and transformer substations
leading to a compounded impact.

A number of papers discuss how cooling and heating energy use will be affected by projected changes in
temperature. Previous analyses of climate impacts on demand has shown that the overall impact of
higher temperatures is likely to reduce demand for heating more than the effect of increased cooling
load.

Adjusting for other variables such as income and energy price is also important in assessing the effect of
temperature increases. A recent publication (Petrick et al. 2010)1 evaluates residential data for 157

1 Petrick S., K. Rehdanz, and R. Tol (2010). The impact of temperature changes on residential energy consumption.
Kiel Institute for the World Economy, No. 1618.

-------
countries over three decades and shows that energy use declines due to rising temperatures indicating
that reduction in heating has played a more important role than the increase in air conditioning load.

An analysis using the POLES Model for Europe (EU27) also notes that only a limited literature develops
the discussion of these issues, and no definitive conclusions exist about quantified evaluations of these
impacts and their respective costs (Mima et al. 2010)2. Mina et al. (2010) This paper estimate that
European energy expenditures on supply-side resources will be $65 billion higher in 2100 - orO.OB
percent of GDP - in one climate change scenario. Conversely, energy expenditures on the demand side
are projected to decrease by $480 billion for heating and increase by $10 billion for cooling. Another
paper by Isaac and Van Vuuren (2009: estimates that global heating energy demand decreases by 800-
1000 Mtoe while cooling demand increases by 80-100 Mtoe by 2100.

Table 1: Hydro-meteorological and Climate Parameters for Select Energy Uses

Hydro-meteorological and/or

Select energy uses

climate parameter



Turbine production efficiency, air source generation potential and output,

Air temperature

demand (cooling/heating), demand simulation/modeling, solar PV panel
efficiency

Rainfall

Hydro-generation potential and efficiency, biomass production, demand.

demand simulation/modeling

Wind speed and/or direction

Wind generation potential and efficiency, demand, demand
simulation/modeling

Cloudiness

Solar generation potential, demand, demand simulation/modeling

Snowfall and ice accretion

Power line maintenance, demand, demand simulation/modeling

Humidity

Demand, demand simulation/modeling

Short-wave radiation

Solar generation potential and output, output modeling, demand, demand
simulation/modeling

River flow

Hydro-generation and potential, hydro-generation modeling (including dam

control), power station cooling water demands

Coastal wave height and frequency,

Wave generation potential and output, generation modeling, off-shore

and statistics

infrastructure protection and design

Sub-surface soil temperatures

Ground source generation potential and output

Flood statistics

Raw material production and delivery, infrastructure protection and design,
cooling water demands

Drought statistics

Hydro-generation output, demand

Storm statistics (includes strong

Infrastructure protection and design, demand surges

winds, heavy rain, hail, lightning)



Sea level

Offshore operations, coastal energy infrastructure

Formal analysis of impacts of climate change on energy supply infrastructure is extremely limited.
Studies exist for the UK, Brazil, and the US state of Alaska, but there may be other studies currently
being conducted elsewhere. Lawrence Berkeley National Laboratory (LBNL) is in the process of
completing a "bottom-up" study for California. The results of which are described below. This multi-year
research effort included participation by utility companies in a technical advisory role.

2 Mima S. and Criqui P. (2010). Analysis of Europe energy system in the POLES model A1B case under future climate
change. Draft Report, LEPil, Grenoble.

Isaac M. and D. Van Vuuren (2009). Modeling global residential sector energy use for heating and air conditioning
in the context of climate change. Energy Policy

-------
Our study examined the impact of climate change on California energy infrastructure, including the San
Francisco bay region. We estimated second-order impacts on power plant generation, transmission line
and substation capacity during heat spells, wildfires near transmission line corridors and a limited study
of sea level encroachment on power plants, substations and natural gas facilities.

We conclude that negative impacts on electricity infrastructure can be avoided, if climate change is
anticipated and sufficient adaptation measures are employed. These measures might include installing
new generation, substation, and transmission capacity, improving energy efficiency, and increasing
investments in cooling equipment and wildfire mitigation strategies.

More specifically, the study finds that higher temperatures will decrease the capacity of existing natural
gas fired power plants to generate electricity during particularly hot periods in the future. The estimated
decrease in capacity varies by region, emission scenario, climate model, and plant type. During the
hottest periods in August (at the end of the century) and under the high emission scenario (A2), our
models estimate a decrease in simple cycle natural gas power plants generating capacity of3%-6% in
California and 3%-4% in the San Francisco region. Under similar conditions, our models suggest
diminished transformer and substation capability—between 2 and 4% across California and between 2
and 3% in the San Francisco region with a small increase in transmission line carrying capacity.

Climate change and fire risk may pose a more difficult challenge to the electric utilities. Our work, building
on the results of existing fire studies, suggests that higher temperatures resulting from climate change will
increase fire risk to transmission lines in California, including the San Francisco region. For example, the
likelihood of fires occurring next to large transmission lines is expected to increase dramatically in parts of
California and San Francisco at the end of the century, under some climate scenarios. It should be noted
that fires do not always cause electricity outages—they more often increase electricity maintenance costs
and decrease transmission line efficiency. In addition, rising sea levels at the end of the century could
flood as many as 25 power plants, scores of electricity substations and numerous natural gas facilities
located along the coast of California and within the San Francisco region. Properly anticipated however,
flooding could be avoided by building dykes, moving plants to higher elevations and other preventative
actions. We also conducted site visits to several power plants and learned that the vertical resolution of
California coastal topography is of a coarse resolution, which makes estimating impacts at the local level
very difficult. We also learned that electricity infrastructure was occasionally not located at the latitude
and longitude reported in the database that was supplied to us.

We concluded that electric utilities can deal with anticipated climate change, but we also recognize that
the level of system capacity needed to do this may be difficult to quantify and finance. It is clear that
utility engineering practices traditionally used to determine generation or transmission capacity may
need to be revised. Similarly, utility tariff setting guidelines may need to be altered to finance the
necessary infrastructure to maintain system reliability. In short, uncertainty about climate change is
likely to pose both institutional and scientific challenges of a type that go beyond the scope of the
current study. These institutional challenges may present as large a problem to the electricity system of
California as the economic costs of anticipated climate change described in this study.

-------
Impacts of Climate Change on Global
Energy Production and Consumption:

Recent Literature and a Useful California Case Study

Dr. Jay ant
Lawrence Berkeley National Laboratory
Berkeley, California

Email: JASathave@lbl.gov

With assistance from
Dr. Peter Larsen and Dr. Larry Dale, LBNL

DOE/EPA Climate Damages Workshop II
Washington D.C.

-------
Presentation Outline

I.	Context

II.	Selected Review of International
Impact Analyses

III.	U.S. Case Study: California

IV.	Lessons Learned

-------
Presentation Context

•	Traditional focus has been on GHG mitigation policy
effects to this sector.

•	General lack of impacts information for the energy sector,
but base of international literature is growing.

•	Qualitative "scoping studies", global, and regional risk
assessments are underway.

•	Analysis methods carried out in our ongoing research into
California energy infrastructure at risk to climate change
could be replicated in other regions, especially
probabilistic and risk-based mapping.

3

-------
Presentation Context:

Parameter Impacts on Energy Demand and Supply



Hydro-meteorological and/or
climate parameter

Select energy uses

Air temperature

Turbine production efficiency, air source generation potential and output,
demand (cooling/heating), demand simulation/modeling, solar PV panel
efficiency

Rainfall

Hydro-generation potential and efficiency, biomass production, demand,
demand simulation/modeling

Wind speed and/or direction

Wind generation potential and efficiency, demand, demand
simulation/modeling

Cloudiness

Solar generation potential, demand, demand simulation/modeling

Snowfall and ice accretion

Power line maintenance, demand, demand simulation/modeling

Humidity

Demand, demand simulation/modeling

Short-wave radiation

Solar generation potential and output, output modeling, demand, demand
simulation/modeling

River flow

Hydro-generation and potential, hydro-generation modeling (including dam
control), power station cooling water demands

Coastal wave height and frequency,
and statistics

Wave generation potential and output, generation modeling, off-shore
infrastructure protection and design

Sub-surface soil temperatures

Ground source generation potential and output

Flood statistics

Raw material production and delivery, infrastructure protection and design,
cooling water demands

Drought statistics

Hydro-generation output, demand

Storm statistics (includes strong
winds, heavy rain, hail, lightning)

Infrastructure protection and design, demand surges

Sea level

Offshore operations, coastal energy infrastructure

4

-------
Selected Research List: Global, National and Local

Larsen et al
(2008)

Wilbanks et
aL/USCCRP (2007)

Sathaye et al.
(in progress)

Hulrne et al./ADAM

Mima et al./ClimateCost

ma et
al (2010);
Schaeffer
et al.
(2008)

sadoorian
et al. (2007)

Wang et
al. (2010)

GLOBAL:

Vergara et al./World Bank (in progress)
Petrick et al. (2010)

Wilbanks et al./IPCC-AR4 (2007) 5

-------
Selected Research: Global and Multi-national

Climate impact on energy demand:

•	Heating Demand:

•	Models typically show a decline in heating demand with rising
temperatures

•	e.g., Mina et al.(2010) using the A1B reference scenario in the POLES
model show a decline that ranges from 200-300 Mtoe (-38% to -62%) by
2100.

•	Cooling Demand:

•	Models show an increase in cooling demand with rising temperatures

•	e.g., Increase in cooling demand is typically lower than the increase in
heating demand - 60-130 Mtoe in the POLES model

6

-------
Selected Research: Global and Multi-national

Climate impact on energy supply:

•	Quantitative analysis of global supply options is limited to date

•	e.g., POLES model shows that hydroelectricity generation may increase or
decrease depending on the scenario, while nuclear and thermal generation
declines by 2100

Impact of climate change on World nuclear gene ration

-1000

-2 000

-3 000

-4 000
-5 000
-6 000
-7 000

¦	BCM2

¦	EGMAM1

¦	EGMAM2
EGMAM3

¦	IPCM4
MPEH5_1
MPEH5_3
DMIEH5
HADGEM

2020 2030 2040 2050 2060 2070 2080 2090 2100

Impact of climate change on World hydroelectricity
generation

2020 2030 2040 2050 2060 2070 2080 2090 2100

IBCM2
ICNCM3
I EGMAM1
I EGMAM2
EGMAM3
INGVSX
IPCM4
MPEH5_1
MPEH5_2
MPEH5_3
DMIEH5
HADGEM

7

-------
Selected Research: National

~Least-cost adaptation options for the Brazilian electric power
system (Lucena et al 2010)-

•	Researchers applied an integrated resource planning approach to
calculate least-cost adaptation measures to a set of projected climate
impacts in 2100 on the Brazilian power sector.

•	Used MAED (demand) and MESSAGE (supply) models, and A2
and B2 scenarios

•Focus is on impacts on electricity demand, hydropower capacity
factor, and natural gas efficiency

•	Electricity demand increases in residential and service sectors by
6% and 5%

•	Hydropower firm capacity factor declines by about 30%

•	Natural gas generation decreases by about 2%

•	Above impacts are offset by efficient adaptation technologies, and
increased use of renewable, nuclear and thermal plant use

8

-------
Selected Research: Local/Regional

~Alaska Infrastructure at Risk (Larsen et al. 2008)-

• Developed preliminary model to estimate quantitative risk to AK public
infrastructure, including energy systems. Model estimated additional costs
with and without adaptation scenarios and included probabilistic framework.
Researchers acknowledged shortcomings including the need to: 1) improve
count/value of infrastructure, 2) develop "ground-truthed" damage functions,
and 3) properly discount uncertain future risk to the present.

~California Energy Infrastructure at Risk (Sathaye et al; in progress)-

• Estimating risk to power plant, substation, and transmission line
performance to projected temperature maximums. Team is overlaying
reported energy infrastructure locations on top of sea-level rise and wildfire
projections and visiting sites to ground-truth modeled results.

9

-------
Case Study: Risk to CA Energy Infrastructure

California's Major Power Infrastructure

Change in August Mean Maximum Temperature from I 999: A2

BACKGROUND:

•	California Energy Commission funded study to
estimate power demand and explore physical risk
to CA energy supply system.

•	Technical advisory committee, including power
sector stakeholders, provide feedback on data
sources and methods.

•	Estimated risk for A2 and B1 scenarios for three
time periods up to 2100

BASIC METHOD:

•	Coupled downscaled AOGCM projections to
electrical system thermal equations to estimate
changes to system capacity and demand from
increased ambient temperature.

•	Overlaid sea-level rise estimates and
wildfire projections with known location of
CA energy infrastructure.	10

San
Francisco
Bay Araa
Insel

Transmission Lines (kv)
60-69
92-161

	220-282

	345

	500

•	Substations
Power Plants

CC
- CT

*	Other

A Temperaturf (C)
2 4

Los

Diego

-------
11

-------
Overview of Research: Assessing vulnerability of...

1. Electricity
to warming
temperatures.

• Literature review to determine
quantitative relationships
between ambient temperature
and power plant, substation, and
transmission capacity.

Without additional cooling
equipment, C A natural gas-fired
power plants typically lose
-0.7% to 1.0% of capacity for
every degree of ambient
temperature above 15C.

Estimated potential physical
impacts without

adaptation/growth scenarios and
reported results using mapping
and numerical simulation
software.

Temperature (Degrees Celsius)

Without additional cooling
equipment, CA substations
typically lose -1.0% of capacity
for every degree of ambient
temperature above 30C.

Combined-Cycle

Desert

Coast

Mountain

110%
105%
g 100%
95%
I 90%
I 85%

^ 80%

£

75%
70%

12

-------
End-of-Century Incremental Impact Distributions

Natural gas-fired Power Plants







6 40 "



o 35 -



£ 30 "

CM

£ 25 "



^ 20 ~



-S 15 "



§ 10 -



° c -
CO o



u



^ 40 -



o 35 -



I 30 "



£ 25 "

LL)

-1 20 "



¦S 15 "



§ 10 -



Bo

O CJ1

I I

A^g Cap. Change:	(	2.7"X)

Std Dev.	0.4*

ruin:	(	4.6*)

Max:	(	2.0*)

A/g Cap. Change: ( 1.7*)

Std Dev.
kin:

Ivkx:

0.3*
( 2.6-t)

C

II I I I I I I I I-

( 4.6%) ( 4.1%) ( 3.6%) ( 3.1%) ( 2.6%) ( 2.1%) ( 1.6%) ( 1.1%) ( 0.6%) ( 0.1%)

Average Change in Peak Capacity at Natural Gas Plants

Warming temperatures may lead to loss up to 4,000 megawatts (4%) of
available natural gas-fired power plant capacity.

Incremental losses are reported (i.e., losses above and beyond the losses
estimated for the base period: 1961-1990).

13

-------
End-of-Century Impact Mapping

A2 Scenario, Three AOGCMs
Average Peak Capacity Loss in August

CC Power Plants

Source: Scripps; CEC; LBNL

CC Plants
End of Century
Difference

2070-2099

. CC Power Plants

10 15 20 25 30 35%	3	4	5 6%~

CT Power Plants

-------
Peak demand load vs. peak temperature

1700

1530

1360

1190

RA2 = 0.8244

1020

850

21.0

weighted average temperature, 11 stations, °C

25.0

29.0

33.0

37.0

41.0

15

-------
Electricity Demand and Supply: Results Summary

California's Major Power Infrastructure

•	Peak Capacity Losses

•Natural gas-fired power plants

•	up to 4000 MW (4%)

•	Electricity supply sub-stations

•	1.6% to 2.7%

•	Transmission lines

•	Limited data on sizes, locations,
and usage capacity

•	-7%

•	Cooling demand

•	20% increase in peak load

•	Demand and supply combined effect

•24%

Source: CEC 2010

Other

	re

-------
Overview of Research: Assessing vulnerability of....

2. Electricity infrastructure to wildfires.

•	Discuss climate factors affecting wildfires

•	Overlay transmission lines on near-term spatial models of
wildfire probability

•	Overlay transmission lines on long-term spatial models of
wildfire (as influenced by climate projections)

•	Quantify transmission length of lines exposed to wildfires
under modeled future climate scenarios

17

-------
Length ot transmission Lines Potentially Impacted:
Increase in burned area within cells (1/8°)

Transmission Lines and Fire Risk: A2 Scenario

2020

2050

2085

Exposed
Ratio Line (Km)
001 to 1 2,009
1 to 2 35.287

, t f ,;. . 2 to 3 547

V isdtnL/ - >3.. o

GFDL

n28j|l\.

' > • Exposed

jf, g~. I Ratio Line (Km)
001 to 1 2.480
1to2 26,840

iX ThfcvJj 210 3 8457

v: -yfcbz • 66

- r>

Vsrif"r . Exposed
jf. Line (Km)
jfeyVt' 0-01 to t 5,600

1,02 10816

-X 2toJ 9873

VVV^T* >3.-11.755

im S

- i>

Exposed
Ratio Line (Km)
/ J 0.01 to 1 2.129
1 to 2 35,660

Vv-i; x 2to3 54

i®li^ >3"°

CNRM

f" >-6 j

Exposed

^ Ratio Line (Km)

M.X^r* 001 to' 8.134

1 to 2 28,696
2to3 2.776

/ Expoud
g. s.' tV. MS Line IKm)
XT') ' 0.01 to 1 2.514

rX -U&V> "°2 ,7'3,°

. 2to 3 10,332

yi -3

TtIVj ¦ ^ - \

<3 Hf:- , ,

Exposed
Ratio Line (Km)
001 to 1 6,865

X Tt1to2 30 978

V fi „ *. J<»3 0

PCM1

V

Exposed
Ratio Line (Km)
0.01 to 1 1.195
1 to 2 34,896

» 2,0 3 1739

>3... 13

M ' /.

V y. Exposed
Bafis Line (Km)
0* -p j 0.01 to 1 660
V i f "•> 1 to 2 27.040
-VafeT 2 to 3 7.716
NQK >3 - 2.427

-

A

Ratio of Expected Area Burned to Base Period (1975)

Coarse spatial
resolution of
fire projection
data limited our
impact analysis
to the length of
line in a fire-
prone area.

N

0 200

No Data 0.01 1.0 1.5 2.0 2.5 3.0 3.5 20.00

18

-------
Projected fire risk to transmission lines for the A2 scenario

Transmission Lines and Wildfire Risk

A2 Scenario

1

2020	2050	2085

Probability Line
Affected by Fire

w/in 30-yr Period

* Lines 220kv & above

	0% - 10%

- 10.1% - 20%
20.1% - 30%
30.1% - 40%
40.1% - 50%
50.1% - 60%

	60.1% - 70%

	70.1% - 80%

Source: Westerling; CEC; LBNL

19

-------
Overview of Research: Assessing vulnerability of....

3. Electricity,	natural	ga

infrastructure to sea level rise

•	Review current sea level trends

•	Incorporate data:

-	Land area affected by sea level rise (Pacific Institute, Knowles)

-	Power plant, substation, natural gas locations (CEC)

•	Mapping analysis:

-	Overlay infrastructure locations over sea level areas

-	CompareLBNL and Pacific Institute study results

20

-------
Sea Level Rise Impact Mapping & Comparisons

Power Plants Potentially at Risk from Sea Level Rise

Projected sea level rise - 1.4 meters

25 power plants and about 90
substations are vulnerable to sea
level rise

Humboldt Bay and Antioch Site
visits indicated that coarse vertical
resolution of CA topography may
have over- or under-stated impacts
in power plant locations.

At-Risk Power Plants

(MW)
©

Fuel



0- 10
11 - SO
51 - ISO
151 - 250
251 - 2000

San Francisco Bay Area

OIL/GAS
LANDFILL GAS
MSW

Predicted inundation of 100-year
flood with 1,4m Sea Level Rise

Source: Pac if tc Institue

HUMBOLDT BAY
"(137 MW)

. 10 K

SOLANO COGEN

2.92 MW)

NOVE
> ^v|3MW)

C & H SUGAR <9.5 MW)
CROCKETT COGEN
(247.4 MW)

HUNTER,
(215

OAKLAND (223-5 MW)
—ALAMEDA (49.9 MW)

UNITED COG€N?>	W

INC. (31 MW) V

(3.75

NEWBY IS. I

(2 MW)
NEWBY IS.2
(3.3 MW)

Santa Cruz Area

9-

- STREET JAIL (0.18 MW)

-SANTA CRUZWWTP (0.65 MW)

BUENA VISTA (3.2 MW)-
WATSONVILLE COGEN (31 MW)

Los Angeles Area

a SEGUNDO (1020 MW)
lOGBsl |

0*9

HARBOR. COGEN
(107.45 MW)

r- THUKS(47.8 MW)
QUEEN MARY (I MW)

HUNTINGTON BEACH
.10 km	(1013 MW)

<5

San Francisco
Bay Area Inset

SOUTHERN CALIFORNIA GAS -
UCSB (0.2 MW)

o

ORMOND BEACH
"(1612.8 MW)

Los Angeles
Area Inset



~2T

-------
Lessons Learned

•	General lack of quantitatively-based impacts information
for energy sector, but base of international literature is
growing.

•	Projected global heating demand reduction due to higher
temperatures is larger than the increase in cooling demand

•	Temperature impact on demand is much higher than on
supply infrastructure

• Impact on hydropower supply may increase or decrease
generation depending on water supply conditions

•	Impact of wildfires could potentially be significantly high

•	More data and research are needed to evaluate wildfire and
sea level rise impacts on the power sector infrastructure
and temperature impacts on electricity transmission and
distribution

22

-------
Acknowledgements for CA Research

List of Authors:

•	Jay ant Sathaye, Larry Dale, Peter Larsen, and Gary Fitts (LBNL)

•	Kevin Koy and Sarah Lewis (Geospatial Innovation Facility at
UC-Berkeley)

•	Andre Lucena (Federal University of Rio de Janeiro)

Funder:

•	Guido Franco (PIER Program at California Energy Commission)

23

-------
Regional Conflict and Climate Change

Nils Petter Gleditsch1

Centre for the Study of Civil War (CSCW), Peace Research Institute Oslo (PRIO),

&

Department of Sociology and Political Science, Norwegian University of Science

and Technology, Trondheim

Paper prepared for the workshop on Research on Climate Change Impacts and Associated
Economic Damages, Washington, DC, 27-28 January 2011; Session on Socio-economic and
Geopolitical Impacts, Friday 28 January 2011, 2:50 pm-3:10 pm

Charge questions

Briefly review existing studies of the impacts of climate change on intra- or inter-regional conflicts,
with special attention to any existing quantitative estimates of the effects of changes in
temperature, precipitation patterns, or sea level on conflict. Which regions are likely to be the
most vulnerable to these impacts?

Briefly review the models and data used to estimate these impacts. What factors are most
important to capture in such models when thinking about the conflict impacts of climate change
over a long time frame?

Characterize the uncertainty/robustness/level of confidence in these estimates, globally and by
region. What are the most important gaps or uncertainties in our knowledge regarding the conflict
impacts of climate change? What research in this area would be most useful in the near term?

Abstract

The world is generally becoming more peaceful, but the debate on climate change raises the
specter of a new source of instability and conflict. In this field, the policy debate is running well
ahead of its academic foundation - and sometimes even contrary to the best evidence. To date
there is little published systematic research on the security implications of climate change. The
few studies that do exist are inconclusive, most often finding no effect or only a low effect of
climate variability and climate change. The scenarios summarized by the Inter-Governmental
Panel on Climate Change (IPCC) are much less certain in terms of the social implications than
the conclusions about the physical implications of climate change, and the few statements on the
security implications found in the IPCC reports are largely based on outdated or irrelevant
sources. This paper reviews briefly the models and the uncertainties and outlines some priorities
for future research in this area.

* This paper builds on various publications from the Centre for the Study of Civil War at PRIO
including Buhaug (2010a), Buhaug, Gleditsch & Theisen (2008, 2010), Gleditsch & Nordas
(2009), Gleditsch, Nordas & Salehyan (2007), and Nordas & Gleditsch (2007b). I am grateful to
my colleagues Halvard Buhaug and Ole Magnus Theisen for comments and suggestions. Our
research is principally funded by the Research Council of Norway. Author address: Centre for the
Study of Civil War, PRIO, P. O. Box 9229, Gronland, 0134 Oslo, Norway; nilspgS'prio.no.

-------
2

Introduction

A liberal peace seems to be in the making (Gleditsch, 2008), with a decreasing
number of armed conflicts (Gleditsch et al., 2002; Harbom & Wallensteen,
2010) and lower severity of war as measured by annual battle-related deaths
(Lacina, Gleditsch & Russett, 2006; HSRP, 2010). At the same time, there has
been a strong in democracy, trade, international economic integration, and
memberships in international organizations, as well as in international peace-
keeping and mediation efforts. Figure 1 illustrates the trends in the frequency
and severity of armed conflict.

Figure 1. The frequency and severity of armed conflict, 1946-2009

60 i	r 700 000

e	I f ¦	r	ro

r I / 1	/	\ y b 400000 -g

M-

o

i-

S I' I	/ 'IV	h 300 000 Ja

E	I	/ a / \A /vvvA	i

3 20 \ rv |	/*-* /V v\ / ^\l\	|	|

200 000
100 000

1966	1976	1986	1996	2006

Armed conflicts	Battle deaths

Source: UCDP/PRIO Armed Conflict Dataset, v. 4-2006 (Gleditsch et al., 2002) and PRIO Battle
Deaths Dataset, v. 2.0 (Lacina & Gleditsch, 2005). Figure created by Halvard Buhaug. Data
available from www.prio.no/cscw/datasets and www.per.uu.se/research/UCDP/. The figure
includes all state-based conflicts with more than 25 battle deaths in a calendar year.

The financial crisis, fundamentalist religion, and other factors are widely
seen as obstacles on the road towards a more peaceful world. But the greatest
challenge to the global liberal peace, according to an increasingly widespread
view, is the threat of climate change. Fears on this score have been expressed
by the Norwegian Nobel Committee (Mjos, 2007), which awarded the Nobel
Peace Prize for 2007 to Al Gore and the Inter-Governmental Panel for Climate

2

-------
3

Change and by President Barack Obama (2009). The UN Security Council
discussed the security implications of climate change for the first time in April
2007 (UN, 2007).

Despite the rhetoric, there is little systematic evidence to date that long-
term climate change or short-term climate variability has had any observable
effects on the pattern of conflict at any level. The Intergovernmental Panel on
Climate Change (IPCC) is the main source of scientific information on the
causes and consequences of climate change and has had a strong influence on
the agenda of the public debate. However, so far the IPCC has not made the
security implications a priority issue. The Third and Fourth Assessment
Reports (IPCC, 2001, 2007) make scattered comments on climate change in the
reports from Working Group II on 'Impacts, Adaptation, and Vulnerability', but
these comments are very weakly founded in peer-reviewed research. There is no
thematic chapter for security or conflict, so the scattered comments turn up in
chapters on other topics such as freshwater management and in some of the
regional chapters (notably in the Africa chapter of AR4).

Had the IPCC systematically reviewed the conflict literature, it would
have discovered some relevant research relating to scarcity models of conflict.
And since 2007, more systematic research on the security implications effects
of climate change has emerged. In what follows, I will review this literature,
assess the level of uncertainty of this area of research (which is high), and
discuss priorities for future research. But first, a brief primer on conflict.

Defining conflict2

In our research, we distinguish between conflict, understood as an incompat-
ibility between actors over interests or values, and conflict behavior. Although
for convenience, the literature often refers just to 'conflict', we are interested in
armed conflict, defined by the Uppsala Conflict Data Program (UCDP) as a con-
tested incompatibility that concerns government or territory or both where the
use of armed force between two parties results in at least 25 battle-related
deaths in a calendar year. Of these two parties, at least one is the government
of a state. A war is defined as an armed conflict with more than 1,000 battle-
deaths in a calendar year. UCDP's Armed Conflict Dataset (ACD) has been
compiled for the time-period 1946-2009 (Harbom & Wallensteen, 2010) and is
updated annually. To distinguish them from other types of armed conflict, such
conflicts are now frequently referred to as state-based armed conflict. They can
be subdivided into interstate conflict (between two or more states), extra-state

2 Detailed definitions from the Uppsala Conflict Data Program are found at
www.pcr.uu.se/research/ucdp/definitions.

3

-------
4

conflict (between a state and a non-state group outside its own territory, e.g.
colonial war), intrastate conflict (between the government of a state and internal
opposition groups), and internationalized intrastate conflict (where troops from
another country supports one or both parties to the conflict). The term civil war
is used for intrastate armed conflict with more than 1,000 battle deaths.

Two additional forms of conflict, both with the same lower threshold of
25 battle deaths in a calendar year and covering the period 1989-2008, are now
regularly recorded by the UCDP, although not necessarily updated annually3:
One-sided violence is the use of armed force by the government or an organized
group against civilians. This dataset, which covers the period 1989-2009,
includes genocide and politicide. Non-state conflict is the use of armed force
between two organized armed groups, neither of which is the government. This
includes communal violence. A final form of violence, not coded as a separate
category by UCDP, is Riots, rural or urban, where the violence is not carried out
by an organized group, and where the target is mostly the government but
which can also be directed against private actors. A borderline case is violent
crime, which often accompanies riots and even organized violence and
sometimes can be hard to separate from violent conflict (Collier, 2000).

Of the different types of conflict, disregarding crime, interstate conflict
and one-sided violence claimed the greatest numbers of lives in the twentieth
century. Civil war follows next, while communal conflicts and riots are usually
smaller. Given the small number of interstate wars after the end of the Cold
War and the sparsity of major episodes of one-sided violence, civil war is now
the main killer.

The political rhetoric is unclear about the kinds of conflict expected to
result from climate change, but all these forms of violence have been mentioned
at times. The academic work on the topic needs to be more specific, and many
scholars expect climate change to have a greater impact on non-state violence
than on state-based conflict.

The term 'regional conflict' in the assigned title for this talk is interpreted
in the first charge question as 'intra- or inter-regional conflict'. The common
meaning of regional conflict is probably conflict within certain regions.4 In fact,
a large share of the emerging research focuses on Sub-Saharan Africa as the
most probable venue for climate-induced violence. The alternative interpre-
tation, conflict between regions, would potentially involve violence at a higher

3	The data can be downloaded from www. per. uu. se / research / ucdp / datasets /.

4	See, for instance, an early discussion of environmental quality (including climate change) and
regional conflict (Kennedy et al., 1996).

4

-------
5

level, possibly even "world war'. Most of the research discussed here is relevant
to the first interpretation, but I will also pay brief visits to interregional conflict.

Linking climate change to conflict

Figure 2 is a theoretical model linking climate change to intrastate armed
conflict. The model incorporates insights from case studies as well as statistical
studies of conflict. Three effects of climate change (natural disasters, sea-level
rise, and increasing resource scarcity) are posited to lead to loss of livelihood,
economic decline, and increased insecurity either directly or through forced
migration. Interacting with poor governance, societal inequalities, and a bad
neighborhood, these factors in turn may promote political and economic
instability, social fragmentation, migration, and inappropriate responses from
governments. Eventually this produces increased motivation for instigating
violence as well as improved opportunities for organizing it.

In the following we review the evidence for some of these links via the
three mechanisms mentioned in Charge question 1 (precipitation, temperature,
and rising sea level) as well as two others (natural disasters and arctic rivalry)
that are frequently mentioned in the literature.

Figure 2. Possible pathways from climate change to conflict

The diagram gives a synthesized account of proposed causal linkages between climate change and
armed conflict. For the sake of clarity, possible feedback loops, reciprocal effects, and contextual
determinants are kept at a minimum. Source: Buhaug, Gleditsch & Theisen (2008: 21).

5

-------
6

Evidence

Only a limited number of peer-reviewed studies deal with climate
change/variability and conflict. In the following, I include a few unpublished
papers in the discussion. These are generally papers that have been circulating
in the academic community for some time, have been revised, and are currently
under review at major journals or in press.

Precipitation

The scarcity (or neo-malthusian) model of conflict assumes that if climate
change results in a reduction in essential resources for livelihood, such as food
or water, those affected by the increasing scarcity may start fighting over the
remaining resources. Alternatively, people may be forced to leave the area, and
create new scarcities when they encroach on the territory of other people who
may also be resource-constrained. Barnett 85 Adger (2007) review a broad range
of studies of both of these effects, focusing particularly on countries where a
large majority of the population is still dependent on employment in the
primary sector. If climate change results in reduced rainfall and access to the
natural capital that sustains livelihoods, poverty will be more widespread and
the potential for conflict greater. Published statistical studies of conflicts
globally (Raleigh 85 Urdal, 2007) or in Africa (Hendrix 85 Glaser, 2007; Meier,
Bond 85 Bond, 2007) provide only limited support for these hypotheses. For
instance, Raleigh 85 Urdal concluded (p. 674) on the basis of local-level data,
that the effects of land degradation and water scarcity were "weak, negligible, or
insignificant'. Many of these early studies were inspired by a study by Miguel,
Satyanath 85 Sergenti (2004), which found a relationship between negative
rainfall deviation and increased risk of civil war in Africa. These authors were
not primarily interested in climate change, but used rainfall deviation as an
instrument for economic shocks. Jensen 85 Gleditsch (2009) have pointed out
that Miguel et al. misinterpreted the UCDP data and included countries that
intervene in civil war as countries at civil war. Correcting for this, their results
are weaker. And as Ciccone (2010) has remarked, Miguel et al. look only at
year-to-year rainfall deviations rather than deviations from a long-term mean.
Using this indicator, which better reflects abnormality in rainfall and conforms
more closely to the idea of climate change, their results evaporate. All of these
studies are conducted at the national level. But rainfall variations do not follow
national boundaries. Theisen, Holtermann 85 Buhaug (2010) used disaggregated
data on conflict and climatic variations and found no relationship at the local
level. Looking at a broader set of conflicts for the past two decades, Hendrix 85

6

-------
7

Salehyan (2010) found rainfall to be correlated with civil war and insurgency,
but it is wetter years that are more likely to suffer from violent events. Extreme
deviations in rainfall - particularly dry and wet years - are associated with all
types of political conflict.

Temperature

Two of the authors behind Miguel et al. (2004) were also involved in a more
recent study of temperature and conflict. In a widely publicized study, Burke et
al. (2009, 2010) claimed to find a link between temperature and civil war in
Sub-Saharan Africa for the period 1981-2002 and argued that over a 35-year
period climate change would produce a major increase in the incidence and
severity of civil war in the region, despite the expected conflict-dampening effect
of economic growth and continued democratization during this period.5
However, Buhaug (2010a,b) found that their results were not robust to
standard control variables, to variations in the model specification, to different
cut-offs for the severity of conflict, or to an extension of the time series to the
most recent years. Buhaug concluded that climate variability is not a good
predictor of civil war. Instead, civil war can be better accounted for by poverty,
ethno-political exclusion, and the influence of the Cold War. Figure 3 from
Buhaug's work indicates that using one of the models from Burke et al. (2009),
the climate variables (temperature and precipitation) add virtually nothing to
the explanatory power of the model.

Figure 3. Predicted values of civil war - does climate matter?

ID

-Q

CO

(0
>

Q)

TO IT) -

E
o
o

£

TO
_£Z

>- O -

LO

5 They also suggest (Burke et al., 2009: 20672) that 'earlier findings of increased conflict during
drier years' may have captured the effect of temperature and that 'the role of precipitation
remains empirically ambiguous'

7

-------
8

This figure plots predicted values of civil war for Model 2 of Burke et al. (2009) on the horizontal
axis and a similar model without climate parameters on the vertical axis (r=.999). The linear
models predict outside the range of possible values (0,1). The climate variables add 0.002 to a
total explained variance of 0.657. Source: Buhaug (2010b: E186-187).

A study that looked at long-term trends (a millennium) in climate and
war for China (Zhang et al., 2006) showed that China suffered more often from
war, population decline, and dynastic changes during cold periods. A follow-up
paper found more that cooling impeded agricultural production, in turn
resulting in price inflation, war, famine, and population decline (Zhang et al.,
2007) A study of Europe over the last millennium (Tol & Wagner, 2010) found
that violent conflict (data from www.warscholar. com /1 was more intense during
colder periods, but that this relationship disappears in the past three centuries
and is not robust to details of the climate reconstruction or to the sample
period.6 It makes sense that by and large a colder climate over some time would
lead to a drop in agricultural production and thus in food scarcity and also
makes sense that these Malthusian constraints are becoming less important
over time with increasing industrialization and long-distance trade But the
conflict data have not yet been frequently used in academic research and so far
these findings have not been tested by other scholars.

A recent study of Central Europe by Biintgen et al. (2010), while not
addressing armed conflict directly, links climate to the rise of fall of
civilizations. It confirms the link between warmer summers and improved
conditions for human settlements but also finds that climate variability has a
major impact. However, the authors concede that modern societies may be less
vulnerable to climatic fluctuations.

Several decades ago there was widespread concern in the scientific
community that the world might be facing a period of global cooling, possibly
even a new ice age. The CIA warned of an era of drought, famine, and political
unrest, and even a potential for international conflict. The agency's analysis
suggested that forecasting climate was vital to the planning and execution of US
policy and would occupy a major portion of US intelligence assets (CIA, 1974).

A long line of research links hot temperatures to individual aggression,
including violent crime and riots. Anderson (2001) suggests that therefore
global warming may increase violence. But the causal mechanism proposed in

6 The positive correlation between low temperature and conflict holds for most of Europe, but in
the Balkans it is reversed. However, they note that the Balkans is largely excluded from the
conflict database. They also report a positive correlation between precipitation and conflict for
most of Europe in the earlier centuries (which they attribute to a decline in agricultural output
due to waterlogging) and a negative correlation in the Balkans (which may be due to drought).
Again, this correlation is not found for the most recent three centuries.

8

-------
9

these studies (personal discomfort) is different from the scarcity thesis that is at
the core of the relationship proposed by Burke et al. (2009) and the kind of
violence is also different.

Sea-level change

IPCC (2007, WG II: 323) forecasts a global mean sea-level rise of between 0.28
and 0.43 meters within this century, depending on the scenario chosen.7
Projections for the size of coastal populations (residing below 100 m elevation
and less than 100 km from the coast) show that they may rise from 1.2 billion
(1990 estimate) to between 1.8 and 5.2 billion (Nicholls & Small, 2002). Sea-
level rise will threaten the livelihood of the populations on small island states in
the Indian Ocean, the Caribbean, and the Pacific. However, a much larger
number of people in low-lying areas, rural and urban, and particularly in South
Asia and West Africa, may become more exposed to soil erosion, seasonal
flooding, and extreme weather. Depending on the degree of protection that can
be offered, this may lead to 'climate migration', and conflict with the host
population is a possible consequence (Nicholls & Tol, 2006). However, this is
going to be a slow process and urbanization and industrialization may well
absorb a large fraction of the people who move.

In a global study covering the period 1951-2001, Salehyan & Gleditsch

(2006)	found that an influx of refugees increased the probability of civil war.
However, since a large proportion of these people have fled from conflict, they
are likely to bring with them the attitudes, the weapons, and the organization
that fuel a continuation of the conflict in the host location. It is not obvious that
economic migrants, including environmental migrants, will generate armed
conflict in the same way (Gleditsch, Nordas & Salehyan, 2007). However, this
has not been studied systematically, due to conceptual problems (what is the
definition of an environmental migrant?) and lack of systematic data. Reuveny

(2007)	examined 38 cases of environmental migration since the 1930s and
found that in half of them there was some kind of armed conflict, most
frequently when the migration cross international boundaries. While suggestive,
his study is unlikely to include all cases of environmental migration during this
period and the conflicts are of different types. Moreover, he did not have any
control variables.

7 Several more recent estimates are higher, cf. Grinsted, Moore & Jevrejeva (2009) who project
sea-level rise to the end of the twenty-first century from 0.9 to 1.3 m for the A1B scenario.

9

-------
10

Natural disasters

Global warming is predicted to increase the frequency and intensity of
natural disasters such as tropical storms, flash floods, landslides, and wild
fires, and substantially alter precipitation patterns in many parts of the world.
There has been a sharp increase in the number of disasters over the last sixty
years8, although it is not certain how much of this can be accounted for by
improved reporting, population growth, and shifting patterns of settlement. In
2009, 335 natural disasters were reported, killing more than 10,000 people
(Vos et al., 2010: 1). Asia is the region most heavily affected. Geological
disasters like volcanic eruptions, earthquakes, and tsunamis need not concern
us here, since they are unlikely to be influenced by climate change. The
temporal increase in disaster frequency is largely accounted for by hydrological
and meteorological disasters, particularly by floods, as shown in Figure 4.

The severity of disasters, measured as the number of casualties, shows
no evident time trend, presumably because of increasing coping capacity in
many countries. Future economic development is likely to further increase the
ability of many societies to absorb natural disasters without great loss of
human life, so an increase in extreme weather events need not be accompanied
by higher casualty figures. Geological events are slightly more deadly, but the
more numerous climate-related generate the highest overall death toll.

Figure 4. Frequency and severity of hydro-meteorological disasters
since 1946

Year

Hydro-met. disasters	Disaster deaths

8 Vos et al. (2010: 5) define a disaster as 'a situation or event which overwhelms local capacity,
necessitating a request to a national or international level for external assistance; an unforeseen
and often sudden event that causes great damage, destruction and human suffering'.

10

-------
11

Source of Figure: Buhaug, Gleditsch & Theisen (2008: 11). Data from EM-DAT, Centre for
Research on the Epidemiology of Disasters (CRED). An update from CRED (Vos et al., 2010) does
not show any time trend in the number of disasters for the most recent decade.

Natural disasters may exacerbate conflict risk primarily through
economic loss and a weakening of government authority. Some statistical
studies find the risk of conflict to be higher following natural disasters (Drury &
Olson, 1998; Brancati, 2007; Nel & Righarts, 2008).9 However, Slettebak & de
Soysa (2010), drawing on a long tradition in disaster sociology, argue that
disasters are just as likely to unite those who are adversely affected, at least in
the short run, implying that various forms of anti-social behavior, including
violence, should decline. Using a global sample from 1950 until today and a set
of standard control variables they find that countries affected by climate
disasters face a lower risk of civil war. Similarly, Bergholt & Lujala (2010) find
that climatic natural disasters such as floods and storms have a negative
impact on economic growth but have no effect on the onset of conflict, either
directly or as an instrument for economic shocks.

Arctic rivalry

The melting of the Arctic icecap has been predicted to lead to a scramble for
shipping lanes and natural resources in previously inaccessible territories
(Borgerson, 2008; Paskal, 2010). Since there is no established legal regime for
the region, some observers feel that this could lead to armed conflict. Several
major powers have interests in the region, so potentially this could lead to some
serious conflicts. On the other hand, the vast extension (from the early 1970s)
of national sovereignty through the establishment of Exclusive Economic Zones
(EEZs) points in a different direction. Despite legal action, unresolved bounda-
ries, and occasional confrontations, particularly over fisheries, the estab-
lishment of EEZs to 200 miles off the coastline has proceeded in overwhelming-
ly peaceful fashion. Although several countries (including the US) have not
ratified the UN Convention on the Law of the Sea (concluded in 1982, entered
into force in 1994), its provisions are generally respected. Most observers seem
to agree with Haftendorn (2010) that a mad race to the Pole is not very likely,
nor is a military conflict among the contenders. Historically, the role of disputed
territory is one the central issues of war (Holsti, 1991; Huth, 1996) but
interstate war, regardless of issue, has declined to the point where it is now
very rare (Harbom & Wallensteen, 2010).

9 Brancati (2007) studied only earthquakes and Nel & Righarts (2008) also found stronger results
for geological than for climatic disasters.

11

-------
12

Vulnerable regions

Which are the most vulnerable regions? Empirical studies of rainfall and
temperature (such as Miguel et al., 2004; Burke et al., 2009, Buhaug, 2010)
have largely focused on Africa South of Sahara. In part, this is because Africa is
more dependent on rain-fed agriculture and thus more severely affected by
major climate change or variability. But it is also because climate change is
expected to be associated with conflict in interaction with other conflict-
inducing factors, such as poverty, economic decline, ethnic exclusion etc.
(Buhaug, Gleditsch & Theisen, 2010), all of which also have been frequent in
Africa. Of the 58 countries included in the "bottom billion' (the countries that
are both poor and stagnating) close to two-thirds are found in Africa (Collier,
2009).

Africa is also one of the more conflict-prone regions, along with South
Asia and the Middle East. In the late 1990s, Africa accounted for more battle-
related deaths than all other regions together. However, since then, all regions -
and Africa in particular - have experienced a decline in battle deaths. Since
2005 most battle deaths have occurred in Central and South Asia, driven in
particular by the wars in Sri Lanka, Afghanistan, and Pakistan.

In the second half of the twentieth century, East Asia experienced the
three largest wars anywhere, the Chinese Civil War, the Korea War, and the
Vietnam War, However, since the Vietnamese invasion of Cambodia in 1978
and the Sino-Vietnamese War in 1979 (followed by some minor skirmishes in
the 1980s), this region has been largely free of war.10

Since the physical effects of climate change are so varied, it is hard to
compare regions in terms of the overall effects of climate change. IPCC (2007,
WG II: 435) characterizes Africa as 'one of the most vulnerable continents to
climate change and climate variability', but this judgment is made as much
because of Africa's low adaptive capacity as much as the absolute size of the
climate changes.

Unfortunately, the climate change projections for Africa are highly
uncertain (IPCC, 2007, WG I: 266ff.). Paradoxically, where accurate measure-
ment of historical climate variables is the most needed, the information is also
the most limited.

10 Cf. www, prio. no / cscw/ cross / battledeaths. The major exception is provided by the two insur-
rections in the Philippines, which have claimed some than 20,000 battle deaths over this thirty-
year period. By contrast, each of the three major East Asian wars claimed more than one million
battle deaths each over much shorter time periods.

12

-------
13

Major climate change challenges in Asia include possible increased
seasonal flooding and drought in the areas downstream from the shrinking
Himalayan glaciers, environmental refugees following sea-level rise, and threats
to major coastal cities such as Dhaka, Mumbai, and Hong Kong as a result of
increased tropical storms as well s sea-level rise (IPCC, 2007; Wischnath, 2010).
These challenges are particularly serious since the population of Asia makes up
more than half of the world total. On the other hand, economic growth has been
particularly rapid in large parts of Asia in the past two decades, so the adaptive
capacity is clearly larger than in Africa.

Models

The climate models used in studies of the effects on conflict are generally
derived from standard sources, such as those used by the IPCC. For instance,
Burke et al. (2009) use time series on precipitation and temperature from the
Climatic Research Unit at the University of East Anglia and climate projections
from general circulation models from the World Climate Research Program's
Coupled Model Intercomparison Project under the IPCC's A1B emissions
scenarios, with some alternative calculations under the A2 and B1 scenarios.
Although different scenarios yield somewhat different results, current
controversies about the effects of climate change on conflict do not seem to
depend on the choice of historical data or emissions scenarios.

There is no standard model of conflict which is universally accepted, but
the two most frequently used models of civil war are those used in Fearon &
Laitin (2003) and Collier & Hoeffler (2004), and Hegre & Sambanis (2006) have
conducted a sensitivity analysis to identify the most robust variables from a
large number of common explanatory schemes. Buhaug (2010a) employs some
of the variables from these studies as controls and alternative explanations.
Burke et al. (2010), however, insist that controlling for endogenous variables,
i.e. independent variables that can be influenced by conflict (or the anticipation
of it) will bias the analysis. In the early work of Miguel et al. (2004) the
endogeneity problem was tackled by using rainfall deviation as an instrument
for economic shocks, but it is not always possible to find suitable instruments
and in Burke et al. (2009) there are none.

As already shown in Figure 3 above, the climate variables add very little
to the explanatory power of the model used by Burke et al. (2009). The relatively
high explanatory power, with R2 as high as 0.66 in their Model 1, is driven by
the fixed country effects and the time trends. Standard opportunity models of
civil war, such as Fearon & Laitin (2003) and Collier & Hoeffler (2004) as well as
studies that place more emphasis on ethnic grievances, such as Cederman &

13

-------
14

Girardin (2007), explain more of the variance with explanatory variables and
control variables. However, as Ward, Greenhill 85 Bakke (2010) point out, such
models nevertheless do a very poor job of prediction. The Fearon 85 Laitin (2003)
model does not correctly predict a single onset of civil war, while the Collier 85
Hoeffler (2004) model correctly predicts 3, at the expense of predicting 5 false
positives.11 At the moment, social scientists are poorly equipped to predict rare
events like conflict but climate change is just one of many areas where policy
prescriptions are dependent on more successful efforts at prediction (Schneider,
Gleditsch 85 Carey, 2010).

Uncertainty

The IPCC assessment reports employ quantitative as well as qualitative
assessments of uncertainty. In the Fourth Assessment Report, each Working
Group used a different variation. Working Group I, which assessed the physical
science, relied primarily on a quantitative likelihood scale, with Virtually
certain' (>99% probability of occurrence) at the top.12 For instance, WG I
estimated it to be Very likely' (i.e. > 90%) that the frequency of heavy preci-
pitation events would increase in the future for most areas.13 WG2 relied mostly
on a quantitative confidence scale, where e.g. "high confidence' indicates an
80% or higher chance of being correct.14 WG III relied exclusively on a
qualitative level-of-understanding scale.

The uncertainties in the IPCC assessments are exacerbated by the
inclusion of non-peer reviewed material. The basic principle is that material
used by IPCC and included in the assessment reports should be peer-reviewed.
In WG I on the physical consequences of climate change, this provides the bulk
of the evidence. However, the IPCC has concluded that 'it is increasingly
apparent that materials relevant to IPCC Reports, in particular, information
about the experience and practice of the private sector in mitigation and
adaptation activities, are found in sources that have not been published or
peer-reviewed' (IPCC, 1999/2008: Annex 2). Each such source is to be critically
assessed by the authors of the IPCC assessment and will be archived and made
available to IPCC review authors who request them. An outsider cannot know
exactly how these guidelines have been used in the preparation of the Third and
Fourth Assessment Report, but it is obvious to a reader who knows the

11	When the threshold is set at p (onset) > 0.5. With a lower threshold, both models predict more
conflicts correctly, but they yield an even larger number of false positives (from two to four as
many as the correct predictions).

12	IPCC (2007, WG I: 23), IAC (2010: 29).
is IPCC (2007, WG I: 8).

14 IPCC (2007, WG I: 22), IAC (2010: 28).

14

-------
15

literature that a number of sources have in fact been used quite uncritically in
references to conflict.15

Following the discovery of an error in the Fourth Assessment Report16
that had cited a non-peer reviewed source to back up a an alarmist statement
that the Himalayan glaciers were likely to disappear within 35 years, the UN
and the IPCC itself asked the InterAcademy Council, an umbrella group of
national academies of science in fifteen countries, to review the IPCC's
organization and procedures. Although the evaluation report (IAC, 2010) was
generally favorable, there were critical comments that the review editors had
insufficient authority to ensure that the authors followed up their comments,
that Working Group II (which deals with the social consequences of climate
change) had overemphasized the negative aspects of climate change, that it had
reported high confidence in some statements for which there was little evidence
(p. 4), and that the selection of authors for regional chapters often excludes
some of the best experts because they don't live in the region (p. 18). The report
also noted that peer-reviewed journal articles comprised 84% of the references
in Working Group I, but only 59% in WG II and 36% in WG III (p. 19). An
implication, not stated explicitly by the IAC, is that the IPCC's statements on
the social implications of climate change are less reliable than assessments of
the physical basis.

In the Fifth Assessment Report (AR5), scheduled for 2013, there will be a
chapter on human security, which it is expected will also include a discussion
of violent conflict.17 This is a promising development. However, the expertise of
the group of authors responsible for this chapter leans heavily towards broader
aspects of human security rather than conflict. It seems likely that they will
produce a more balanced assessment of the literature on climate change and
conflict, as the authors have signaled a stronger emphasis on peer-reviewed
literature. But it remains to be seen whether this will prevent more extravagant
and empirically unsupported statements being made in other chapters of the
report and restrain the more dramatic interpretations by NGOs and politi-
cians. 18

15	For a detailed examination, see Nordas & Gleditsch (2009).

16	And, at about the same time, the leaking of thousands of documents and e-mails from the
Climate Research Unit at the University of East Anglia. For a balanced account of the
'Climategate' affair, see Pearce (2010).

17	However, the scoping document of the AR5, approved in October 2010, does not reveal the
contents at this level of detail, cf.

www.ipcc.ch/meetings/session32/svr final scoping document, pdfl.

18	The IPCC November 2010 announcement about the Table of Contents and the authors is found
at www.ipcc.ch/meetings/session32/inf07 p32 ipcc ar5 authors review editors.pdf. cf. Chapter
12.

15

-------
16

Research priorities

Research on the security effects of climate change should focus on interactions
between climatic variables and other conflict-inducing factors, to test the notion
that climate change can act as a 'threat multiplier' (CNA, 2007: 1).

Secondly, although data and models may be more readily available for
rich countries, research on conflict as a possible effect of climate change needs
to focus on the poorer parts of the world, where the adaptive capacity is lower
today. Of course, some countries in the third world now have high economic
growth and are likely to be in a position to absorb greater changes fifty years
from now. Therefore, particular attention needs to be paid to countries that are
not only poor but also stagnating.

Third, we need to go beyond the state-based violence considered in most
statistical studies to date. Much of the case study literature refers to non-state
or one-sided violence, but this has hardly been tested in large-n studies.
Unfortunately, the time series for these types of conflict data are still quite
short, so improved data collection will be a priority.

More work needs to be put into the geographical disaggregation of the
effects of climate change since these effects will not follow national boundaries.

Further, the study of climate change and conflict needs to balance the
negative and positive effects of climate change. While food production is likely to
decrease in some areas, it may increase in others. Although the global net effect
of climate change seems likely to be negative, the effects would vary
considerably both geographically and by sector.

Finally, if we are to go beyond the simple projection of past changes into
the future, we will need a tighter coupling of climate change models and the
conflict models. The development of more fine-grained data for the physical
effects of climate change, incorporating geographic variation, rates of change,
and adaptive measures, will facilitate the scientific interface. But for the
moment, it may be more realistic to concentrate on the past impact of climate
change. If such research indicates that the link to conflict is weak, efforts to
establish projections into the future probably should have lower priority.

Conclusions

Given the potential range and scope of consequences of climate change, it is not
surprising that there is widespread concern about its security implications. In
part, this concern has been directed at raising awareness about 'environmental
security' in a broad sense. Climate change will have many serious effects, parti-
cularly transition effects, on peoples and societies worldwide. The hardships of
climate change are particularly likely to add to the burden of poverty and

16

-------
17

human insecurity of already vulnerable societies and weak governments.19
However, the use of such wider concepts of security must not stand in the way
of a focused effort to analyze empirically the possible link between
environmental change and violent conflict. Assuming such a link without the
necessary evidence may lead peacemaking astray and can eventually also
undermine the credibility of the IPCC and the efforts to reach a consensus of
knowledge about human-made climate change and a concerted global effort at
mitigation and adaptation. The climate-conflict discourse is easily exploited by
cynical governments and ruthless rebels who would like to evade any direct
responsibility for atrocities and violence and prefer to put the blame on
developed countries and their greenhouse gas emissions (Salehyan, 2008).

Finally, what if the academic community were to conclude that climate
change has very little impact on armed conflict. Does it matter? It matters a
great deal for the credibility of climate change research. Extremely low-
probability hazards should not be promoted to major threats under the
precautionary principle. For adaptation to climate change, clarifying the conflict
effects may also be important. Preventing armed conflict is likely to require
countermeasures that are different than preventing biodiversity loss. For the
need to mitigate the effects of climate change, however, the effects of climate
probably matter very little. There are many other reasons to reduce the human
impact on the climate and to prevent global warming from getting out of hand.

References

Anderson, Craig A (2001) Heat and violence. Current Directions in Psychological Science
10(1): 33-38.

Barnett, Jon & W Neil Adger (2007) Climate change, human security and violent
conflict. Political Geography 26(6): 639-655.

Bergholt, Drago & Paivi Lujala (2010) Natural disasters, economic growth, and armed
civil conflict. Paper presented to the international conference on Climate Change
and Security, Trondheim, 21-24 June, www.dknvs.no/climsec.

Borgerson, Scott (2008) Arctic meltdown: The economic and security implications of
global warming. Foreign Affairs 87(2): 63-77.

Buhaug, Halvard (2010a) Climate not to blame for African civil wars. PNAS 107(38):
16477-16482.

Buhaug, Halvard (2010b) Reply to Burke et al.: Bias and climate war research. PNAS
107(51): E186-E187.

Buhaug, Halvard; Nils Petter Gleditsch & Ole Magnus Theisen, 2008. Implications of
Climate Change for Armed Conflict. Working paper, Social Dimensions of Climate
Change workshop. Washington, DC: World Bank, Social Development
Department,

19 However, Buhaug, Falch, Gleditsch & Wischnath (2010) found that climate change so far has
had no greater absolute impact on temperature and precipitation in 'bottom billion' countries in
Africa than in other African countries.

17

-------
18

http: / / siteresources.worldbarLk.org/INTRANETSQCIALDEVELQPMENT /Resourc
es/SDCCWorkingPaper Conflict.pdf.

Buhaug, Halvard; Ashild Falch, Nils Petter Gleditsch & Gerdis Wischnath (2010). The
bottom billion and climate change. Paper presented at the workshop on climate
change and security, Zurich, 3 December.

Buhaug, Halvard; Nils Petter Gleditsch & Ole Magnus Theisen (2010) Implications of
climate change for armed conflict. Ch. 3 in Robin Mearns & Andy Norton (eds),
Social Dimensions of Climate Change: Equity and Vulnerability. New Frontiers of
Social Policy. Washington, DC: World Bank, 75-101.

Biintgen, Ulf; Willy Tegel, Kurt Nicolussi, Michael McCormick, David Frank, Valerie
Trouet, Jed O Kaplan, Franz Herzig, Karl-Uwe Heussner, Heinz Wanner, Jiirg
Luterbacher, Jan Esper (2011) 2500 years of European climate variability and
human susceptibility, Scienceexpress, 13 January, 1-10.

Burke, Marshall B; Edward Miguel, Shanker Satyanath, John A Dykema & David B
Lobell (2009) Warming increases the risk of civil war in Africa. PNAS 106(49):
20670-20674.

Burke, Marshall B; Edward Miguel, Shanker Satyanath, John A Dykema & David B
Lobell (2010) Climate robustly linked to African civil war. PNAS 107(51): E185.

Cederman, Lars-Erik & Luc Girardin (2007) Beyond fractionalization: Mapping ethnicity
onto nationalist insurgencies. American Political Science Review 101(1): 595-715.

CIA (1974) A Study of Climatological Research As It Pertains to Intelligence Problems.
Working paper, Office of Research and Development, Central Intelligence
Agency, August. Available in the Library of Congress (control no. 76603473, call
no. QC981.8.C5 U513 1974) and at www, climatemonitor. it / wp- content /
uploads/2009 712/ 1974.pdf.

Ciccone, Antonio (2010) Transitory economic shocks and civil conflict,
www, antoniociccone. eu / wp- content / uploads /2010/02 / transitory-shocks-
februarv-201 Of.pdf.

CNA (2007) National Security and the Threat of Climate Change. Report from a panel of
retired senior US military officers. Alexandria, VA: CNA Corporation,
http: / / securitvandclimate.cna.org/.

Collier, Paul (2000) Rebellion as a quasi-criminal activity. Journal of Conflict Resolution
44(6): 839-853.

Collier, Paul (2009) Wars, Guns & Votes. Democracy in Dangerous Places. London:
Bodley Head.

Collier, Paul & Anke Hoeffler (2004) Greed and grievance in civil war. Oxford Economic
Papers 56(4): 563-595.

Drury, Cooper A. & Richard Stuart Olson (1998) Disasters and political unrest: An
empirical investigation. Journal of Contingencies and Crisis Management
6(3):153-161.

Fearon, James D. & David Laitin (2003) Ethnicity, insurgency, and civil war. American
Political Science Review 97(1): 75-90.

Gleditsch, Nils Petter (2008) The liberal moment fifteen years on. Presidential Address,
International Studies Association, International Studies Quarterly 52(4): 691-
712.

Gleditsch, Nils Petter & Ragnhild Nordas (2009) Climate change and conflict: A critical
overview. Die Friedens-Warte 84(2): 11-28.

Gleditsch, Nils Petter; Ragnhild Nordas & Idean Salehyan (2007) Climate Change and
Conflict: The Migration Link. Coping with Crisis Working Paper Series. New York:
International Peace Academy,

www.ipacademv.org/our-work/coping-with-crisis/working-papers.

Gleditsch, Nils Petter; Peter Wallensteen, Mikael Eriksson, Margareta Sollenberg &
Havard Strand (2002) Armed conflict 1946-2001: A new dataset. Journal of
Peace Research 39(5): 615-637.

18

-------
19

Grinsted, Aslak; John C. Moore & Svetlana Jevrejeva (2009) Reconstructing sea level
from paleo and projected temperatures 200 to 2100 ad. Climate Dynamics
34(4):461-472.

Haftendorn, Helga (2010) Coping with climate change in the Arctic: The applicability of
soft solutions to hard conflicts. Paper presented to the international conference
on Climate Change and Security, Trondheim, 21-24 June,
www, dknvs. no / climsec.

Harbom, Lotta & Peter Wallensteen (2010) Armed conflicts, 1946-2009. Journal of Peace
Research 47(4): 501-509.

Hegre, Havard & Nicholas Sambanis (2006) Sensitivity analysis of empirical results on
civil war onset. Journal of Conflict Resolution 50(4): 508-535.

Hendrix, Cullen S & Sarah M Glaser (2007) Trends and triggers: Climate, climate
change and civil conflict in Sub-Saharan Africa. Political Geography 26(6): 695-
715.

Hendrix, Cullen & Idean Salehyan (2010) After the rain: Rainfall variability, hydro-
meteorological disasters, and social conflict in Africa. Paper presented to the
international conference on Climate Change and Security, Trondheim, 21-24
June, www, dknvs.no / climsec.

Holsti, Kalevi J (1991) Peace and War: Armed Conflict and International Order, 1948-
1989. New York: Cambridge University Press.

HSRP (2010) Human Security Report 2009/2010. The Causes of Peace and the Shrinking
Costs of War. Vancouver, BC: Human Security Report Project, Simon Fraser
University.

Huth, Paul (1996) Standing Your Ground. Territorial Disputes and International Conflict.
Ann Arbor, MI: University of Michigan Press.

IAC (2010) Climate Change Assessments: Review of the Processes and Procedures of the
IPCC. Amsterdam: InterAcademy Council,

http: / / r eviewipcc. interacademycouncil. net / report / Climate%2 0Change%2 OAssessments
,%20Review%20of%20the%20Processes%20&%20Procedures%20of%20the%20I
PCC.pdf.

IPCC (1999/2008) Procedures for the Preparation, Review, Acceptance, Adoption,
Approval and Publication of IPCC Reports. Appendix A to the Principles Governing
IPCC Work. Adopted 1999 and amended 2003 and 2008,
http: / / ipcc. ch / pdf / ipcc-principles / ipcc-principles-appendix- a. pdf.

IPCC (2001) Third Assessment Report. Climate Change 2001. Geneva: Intergovernmental
Panel on Climate Change & Cambridge: Cambridge University Press. Available
online at www.ipcc.ch.

IPCC (2007) Fourth Assessment Report. Climate Change 2007. Geneva: Intergovern-
mental Panel on Climate Change & Cambridge: Cambridge University Press.
Available online at www.ipcc.ch.

Jensen, Peter Sandholt & Kristian Skrede Gleditsch (2009) Rain, growth, and civil war:
The importance of location. Defence and Peace Economics 20(5): 359-372.

Kennedy, Donald; David Holloway, Erika Weinthal, Walter Falcon, Paul Ehrlich, Roz
Naylor, Michael May, Steven Schneider, Stephen Fetter & Jor-San Choi (1998)
Environmental Quality and Regional Conflict. New York: Carnegie Commission on
Preventing Deadly Conflict,

www.wilsoncenter. org / subsites / ccpdc / pubs / ken / kenfr.htm.

Lacina, Bethany & Nils Petter Gleditsch (2006) Monitoring trends in global combat: A
new dataset of battle deaths. European Journal of Population 21(2-3): 145-166.

Lacina, Bethany; Nils Petter Gleditsch & Bruce Russett (2006) The declining risk of
death in battle. International Studies Quarterly 50(3): 673-680.

Meier, Patrick; Doug Bond & Joe Bond (2007) Environmental influences on pastoral
conflict in the Horn of Africa. Political Geography 26(6): 716-735.

19

-------
20

Miguel, Edward; Shanker Satyanath & Ernest Sergenti (2004) Economic shocks and
civil conflict: An instrumental variables approach. Journal of Political Economy
112(4): 725-753.

Mjos, Ole Danbolt (2007) Award ceremony speech by the Chair of the Norwegian Nobel
Committee for the IPCC and A1 Gore, Oslo, 10 December,
http:/ /nobelprize.org/nobel prizes/peace/laureates/2007/presentation-
speech, html.

Nel, Philip & Marjolein Righarts (2008) Natural disasters and the risk of violent civil
conflict. International Studies Quarterly 52(1): 159-185.

Nicholls, Robert J & Christopher Small (2002) Improved estimates of coastal population
and exposure to hazards. Eos. Transactions of the American Geophysical Union
83(28): 301-305.

Nicholls, Robert J & Richard S J Tol (2006) Impacts and responses to sea-level rise: A
global analysis of the SRES scenarios over the 21st century. Philosophical
Transaxtions of the Royal Society London A 361(1841): 1073-1095.

Nordas, Ragnhild, & Nils Petter Gleditsch, guest editors (2007a) Climate change and
conflict. Special issue of Political Geography 26(6), August.

Nordas, Ragnhild & Nils Petter Gleditsch (2007b) Climate change and conflict. Political
Geography 26(6): 627-638.

Nordas, Ragnhild & Nils Petter Gleditsch (2009) IPCC and the climate-conflict nexus.
Paper presented at the 50th Annual Convention of the International Studies
Association, New York, 15-18 February,
http: / / convention2. allacademic. com / one / isa / isa09 /.

Obama, Barack H (2009) Remarks by the president at United Nations Secretary General
Ban Ki-Moon's climate change summit 22 September. New York: United Nations,
www.whitehouse.gov/the press office/.

Paskal, Claire (2010) Global Warring. How Environmental, Economic, and Political Crises
Will Dedraw the World Map. London: Palgrave Macmillan.

Pearce, Fred (2010) The Climate Files. The Battle for the Truth about Global Warming.
London: Guardianbooks.

Raleigh, Clionadh & Henrik Urdal ( 2007) Climate change, environmental degradation
and armed conflict. Political Geography 26(6): 674-694.

Reuveny, Rafael (2007) Climate change-induced migration and violent conflict. Political
Geography 26(6): 656-673.

Salehyan, Idean (2008) From climate change to conflict? No consensus yet. Journal of
Peace Research 45(3): 315-326.

Salehyan, Idean & Kristian Skrede Gleditsch (2006) Refugees and the spread of civil
war. International Organization 60(2): 335-366.

Schneider, Gerald; Nils Petter Gleditsch & Sabine Carey (eds) (2010) Exploring the past,
anticipating the future. Presidential symposium, International Studies Review
12(1): 1-104.

Slettebak, Rune T and Indra de Soysa (2010) Don't blame the weather! Climate-related
natural disasters and civil conflict. Paper presented to the international
conference on Climate Change and Security, Trondheim, 21-24 June,
www, dknvs. no / climsec.

Theisen, Ole Magnus; Helge Holtermann & Halvard Buhaug (2010) Drought, political
exclusion, and civil war. Oslo: Centre for the Study of Civil War, PRIO, under
journal review.

Tol, Richard SJ & Sebastian Wagner (2010) Climate change and violent conflict in
Europe over the last millennium. Climatic Change 99(1-2): 65-79.

UN (2007) Security Council Holds First-Ever Debate on Impact of Climate Change, 5663rd
Meeting. New York; United Nations, Department of Public Information,
www.un.org/News/Press/docs/2007/ sc9000.doc. htm.

Vos, Femke; Jose Rodriguez, Regina Below & Debarati Guha-Sapir (2010). Annual
Disaster Statistical Review 2009: The Numbers and Trends. Brussels: Centre for

20

-------
21

Research on the Epidemiology of Disasters (CRED), Universite catholique de
Louvain, http:/ /cred.be/sites/default/files/ADSR 2009.pdf.

Ward, Michael D, Brian Greenhill & Kristin Bakke (2010) The perils of policy by p-value:
Predicting civil conflicts. Journal of Peace Research 46(4): 363-375.

Wischnath, Gerdis (2010) Climate strange and civil war? Variability of climate
parameters and internal armed conflict in Asia. Paper presented to a workshop
of the Working group on environmental factors in civil war. Centre for the Study
of Civil War, PRIO, 16-17 December,

www.prio.no/upload/cscw/wg3/workshop dec 2010/Climate%20Strange%20a
nd%20Civil%20War 1210.pdf.

Zhang, David D; CY Jim, George C-S Lin, Yuan-qing He, James J Wang & Harry F Lee
(2006) Climatic change, wars and dynastic cycles in China over the last
millennium. Climatic Change 76(3-4): 459-477.

Zhang, David D., Jane Zhang, Harry F Lee & Yuan-qing He) Source: Climate change
and war frequency in eastern China over the last millennium. Human
Ecology 35(4): 403-414.

Biographical information

NILS PETTER GLEDITSCH, b. 1942, Research professor at the Centre for the Study of
Civil War, PRIO and Professor of political science, Norwegian University of
Science and Technology, Trondheim. Associate editor, Journal of Peace Research
(Editor, 1983-2010). President of the International Studies Association 2008-09.

[g: /konferanser/2011/110127-110128 Washington/Gleditsch abstract revised]

21

-------
P RIO

Regional conflict and climate change

0NTNU

Regional Conflict and Climate change

EPA/DoE workshop on Climate Change Impacts and Associated Economic Damages

Capitol Hilton, Washington, DC, 27-28 January 2011

Nils Petter Gleditsch

Centre for the Study of Civil War (CSCW), Peace Research Institute Oslo (PRIO) &
Department of Sociology and Political Science, Norwegian University of Science and

Technology (NTNU)

-------
Regional conflict and climate change

0NTNU

Armed conflicts and battle deaths, 1946-2009

60

700 000

50

40

c

o

o

o 30

a;
.q

20

10

600 000
500 000

in
-C

fU

400 000 -g

M—

o

300 000 Ja
E

3

z

200 000
100 000

1946	1956	1966

^^"Armed conflicts

i i i i i i i i i i i i i i

1976	1986

1996
Battle deaths

2006

0

-------
P RIO

Regional conflict and climate change

0NTNU

:=-
_Qj

CTf

r--

cn

CD

¦U1

CD
CL

Towards a liberal peace?

CD —

GDP/pop
IGOs

Democracy
Trade/GDP

o

cn

CD
•ZD
(---I

1=1

CD

	1	

1 950

	1	

1 960

	1	

1 970

	1	

1 930

	1	

1 990

	1	

2000

Year

-------
P RIO

Regional conflict and climate change

0NTNU

Possible threats to the liberal peace

•	Shifting patterns of power

•	The financial crisis

•	Fundamentalist religion

•	Climate change

-------
Regional conflict and climate change

0NTNU

Enter climate change:

Are we heading towards disaster?

•	Darfur is the first of many climate wars (Ban Ki-Moon, 2007-08)

•	There is little scientific dispute that if we do nothing, we will face more drought,
more famine, more mass displacement - all of which will fuel more conflict for
decades (President Barack Obama's Nobel Peace Prize Lecture, 10 December
2009)

•	Evidence is fast accumulating that, within our children's lifetimes, severe
droughts, storms and heat waves caused by climate change could rip apart
societies from one side of the planet to the other. Climate stress may well
represent a challenge to international security just as dangerous — and more
intractable — than the arms race between the United States and the Soviet
Union during the cold war or the proliferation of nuclear weapons among rogue
states today. (Thomas Homer-Dixon, NYT} 24 April 2007)

-------
P RIO

Regional conflict and climate change

0NTNU

From climate change to conflict:
Possible pathways

-------
Regional conflict and climate change

0NTNU

Evidence: Precipitation (I)

•	*Miguel, Satyanath & Sergenti (2004): the probability of
conflict in sub-Saharan Africa increases the year after a year
with reduced rainfall (instrument for economic shock)

•	*Hendrix & Glaser (2007): the level of available freshwater is
positively linked to conflict, but negative deviations also yield
more conflict

•	*Jensen & Gleditsch (2009): the results in Miguel et al. (2004)
are weaker when removing countries that participate in civil
wars in other countries

-------
Regional conflict and climate change

0NTNU

Evidence: Precipitation (II)

•	*Hendrix & Salehyan (2010): (47 African countries, 1990-
2009) Wetter years are more likely to see civil wars. Rainfall
variability has a significant effect on other forms of political
unrest.

•	Theisen, Holtermann & Buhaug (2010): In a disaggregated
analysis, drought has no influence on civil conflict in Africa

•	Ciccone (2010): Miguel et al. look only at annual deviations
rather than deviations from the long-term mean

•	*Burke et al. (2009): Precipitation changes in Africa cannot be
predicted precisely from existing climate models

-------
Regional conflict and climate change

0NTNU

Evidence: Temperature (I)

•	*Burke et al. (2009, 2010): Higher temperatures in SS Africa yield
more conflict (impact on agriculture);

•	*Buhaug (2010a,b) Their results are not robust to standard
control variables, to variations in the model specification, or to an
extension of the time series to more recent years

-------
P RIO

Regional conflict and climate change

Civil war risk with/without climate variables

ID



0J

TO ID

E "

o

o

J

TO

_c

>~ o

ID

I

-------
Regional conflict and climate change

0NTNU

Evidence: Temperature (II)

•	*Zhang (2006, 2007) War, population decline, and dynastic
changes were more common in China in cold periods (1000-year
time frame)

•	*Tol & Wagner (2010) Violent conflict in Europe was more
common in cold periods, but the relationship disappears in the
most recent three centuries

•	*Buntgen et al. (2010) Warmer summers improve conditions for
human settlements and the rise of civilizations - but this may be
less relevant for modern civilizations

•	CIA (1974) Global cooling threatens to produce drought, famine,
and political unrest, particularly in the Sahel region. Climate
modification could lead to international conflict

-------
0NTNU

Evidence: Sea-level change

•	IPCC (2007, WG II: 323): Global mean sea-level rise to 2100:
0.28-0.43 cm

•	*Grinsted, Moore & Jevrejeva (2009): 0.9-1.3 m

•	Myers, IPCC, Stern: 150-200-250 mill 'climate refugees'

•	*Nicholls & Small (2002): 1.2 bill, live in coastal areas, rising to
5.2 bill, by the end of the century

•	*Salehyan & Gleditsch (2006): Countries with a high influx of
refugees have a greater risk of civil war

•	Gleditsch, Nordas & Salehyan (2007): Will this also apply to
climate refugees?

•	*Reuveny (2007): In 38 cases of environmental migration since
the 1930s, half experienced armed conflict of some kind - but is
this representative?

P RIO

Regional conflict and climate change

-------
0NTNU

Evidence: Natural disasters (I)

•	CRED data show that the number of natural disasters is
increasing, more people affected, fewer people die

•	Is the increase in numbers due to global warming, better
reporting, shifting settlements?

•	Increase in cost, but mainly due to more high-value objects
insured?

•	Analyses of disasters and conflict suggest a connection (*Drury &
Olson, 1998; *Brancati, 2007; *Nel & Righarts, 2008), but mostly
for geological disasters, and mechanisms unclear

P RIO

Regional conflict and climate change

-------
Regional conflict and climate change

0NTNU

Hydro-meteorological disasters

Year

— Hydro-met. disasters	— Disaster deaths

-------
Regional conflict and climate change

0NTNU

Evidence: Natural disasters (II)

•	Even for geological disasters, Aceh points in a different
direction (*Le Billon & Waizenegger, 2007; *Enia, 2008;
*Beardsley & McQuinn, 2009)

•	Slettebak & de Soysa (2010): Earlier studies fail to include
proper controls, particularly population size. Using the
Fearon & Laitin model, climate-related disasters, tend (if
anything) to lower the probability of conflict; consistent with
a long tradition in disaster sociology that people unite in the
face of adversity

•	Bergholt & Lujala (2010): Natural disasters lower economic
growth but do not increase conflict via this mechanism

-------
0NTNU

Economic effects of climate change

•	Economic factors important in conflict - economic interdependence
limits interstate conflict, economic development limits intrastate conflict

•	Economic decline could reverse this

•	Debate about the economic effects of climate change hinges on the
value of discounting future economic effects - Stern (2007) uses a low
value, while Nordhaus (2007) uses a high value

•	Few empirical studies: Bernauer et al. (2010) study effects of
precipitation and conflict, Bergholt & Lujala (2010) natural disasters
and conflict, neither study finds any effect on conflict via economic
growth, but Bernauer et al. find that political institutions modify the
relationship

P RIO

Regional conflict and climate change

-------
P RIO

Regional conflict and climate change

0NTNU

Climate change and interstate conflict

•	Argument 1: Increased scarcity interstate conflict

•	Counterargument: Scarcity model generally unpersuasive and less so today

•	Argument 2: Climate change will open up new trade routes and new ocean
territories for exploration, there will be uncertainty about ownership and
competition for exploiting these resources, danger of conflict

•	Counterargument: a) little systematic research, b) introduction of EEZs
proceeded largely peacefully

•	Tir & Stinnett (2010): Institutionalized cooperation in shared rivers is likely to
prevent distribution conflicts

•	Gartzke (2010): climate change may affect where nations fight, rather than
whether or when (militarized disputes move to higher latitudes in summer,
lower latitudes in winter)

-------
prio ^^hqvfm:

Methods

The neomalthusian theory of conflict has generally drawn on case studies for
support, notably those by Homer-Dixon and others

Large-n studies have found little support for the scarcity theories. So is it a
methodological divide?

Several recent case studies, by *Benjaminsen (2008) on Mali, *Witsenburg &
Adano (2009) on Northern Kenya, *Brown (2010) on Darfur, and others have
also questioned the scarcity perspective

The neomalthusian case studies in the scarcity tradition have been criticized for
selecting on the dependent variable, i.e. studying only the conflict cases

But they can also be criticized for relatively shallow case description and for
focusing too rapidly on scarcity factors

We may perhaps see a convergence of case studies and statistical work,

including time-series for single countries and disaggregated statistical studies

-------
BNTNU

Interactions

Critics of Homer-Dixon and others may have overlooked how scarcity
interacts with poverty, poor governance, ethnic dominance, etc.

Threat multiplier (CNA , 2007)

Double exposure (O'Brien), also Temesgen (2010)

'Unfortunately, pollution, population growth and climate change are not in
the distant future: they are occurring now and hitting the poorest and
most vulnerable hardest. Environmental degradation has the potential
to destabilize already conflict-prone regions, especially when
compounded by inequitable access or politicization of access to scarce
resources.' - Kofi Annan (2006)

Hard to test for interactions of four factors ...

From a policy perspective, easiest to reduce climate change or to change
other factors in the interaction?

P RIO

Regional conflict and climate change

-------
0NTNU

Vulnerable regions

•	Africa is high on conflict, low on development, low on governance;
includes two thirds of the 'bottom billion' countries. However,

Africa is experiencing a decline in conflict, increasing economic
growth, and improving governance

•	East Asia had the most severe wars in the second half of the
twentieth century; now largely peaceful

•	Most battle deaths currently occur in Central and South Asia.
Middle East also sees frequent conflict, but not currently very
severe

•	Empirical studies have focused on Africa (particular SSA) a)
because it is more vulnerable b) because of low adaptive capacity

P RIO

Regional conflict and climate change

-------
o

Regional conflict and climate change

Q

NTNU

The distribution of armed conflict

i



4t

t-

•X S ; Ca t Vb .

X- j M $2

iL

Intrastate conflict

No conflict
Conflict 1989-2007
I Conflict 2008

%

-------
0NTNU

Models

•	Disagreements about security effects do not appear to depend on
the choice of emissions scenarios

•	No standard conflict model, but *Fearon & Laitin (2003) and
*Collier & Hoeffler (2004) frequently used

•	Endogeneity problems?

•	*Ward, Greenhill & Bakke (2010): Standard conflict models do a
poor job of predicting new conflicts

•	If studies of historical data provide little evidence for a security
effect, projection is less urgent

P RIO

Regional conflict and climate change

-------
0NTNU

Uncertainty

•	IPCC WG I : quantitative likelihood scale: Virtually certain = 99%
probability of occurrence, etc.

•	IPCC WG II: quantitative confidence scale: Very high confidence
= 90% or higher chance of being correct

•	IPCC WG III: qualitative level-of-understanding scale, high to low
agreement on one axis, much to little evidence on the other

•	IAC (2010) criticizes WG II for reporting high confidence in
statements for which there was little evidence

•	Peer-reviewed sources: relatively fewer in WG II than in WG I and
even lower in WG III

P RIO

Regional conflict and climate change

-------
Regional conflict and climate change

0NTNU

Research priorities

•	Look at interactions between climate change and

political and economic factors

•	Focus on countries with low adaptive capacity

•	Look at a broader set of conflicts (one-sided, non-

state, riots)

•	Disaggregated studies of geo-referenced data

•	Balance negative and positive effects (e.g. food)

•	(possibly) Couple models of climate change to models

of conflict

-------
0NTNU

What if climate change has negligible
impact on conflict?

Does it matter?

•	For the credibility of climate change research -
very much

•	For mitigation - very little

•	For adaption - possibly a lot

p RIO

Regional conflict and climate change

-------
P RIO

Regional conflict and climate change

0NTNU

THANK YOU FOR YOUR ATTENTION

-------
Migration Impacts of Climate Change

Robert McLeman, Associate Professor, Department of Geography, University of Ottawa
rmcleman(5)uottawa.ca

Extended abstract for US EPA/DOE Workshop, "Research on Climate Change Impacts and Associated
Economic Damages", January 27-28, 2011, Washington, D.C.

1. Briefly review existing studies of the impacts of climate change on intra- or inter-regional
migration, with special attention to any existing quantitative estimates of the effects of changes in
temperature, precipitation patterns, or sea level on migration patterns. Which regions are likely to be
the most vulnerable to these impacts?

Scholars have long known that environmental conditions, including climatic variability and change, can
and do influence human migration (Hugo 1996, Hunter 2005). Contemporary discussions of climate-
related migration tend to be framed in terms of "environmental refugees" (a term coined by El-Hinnawi
1986), whereby people are involuntarily displaced in response to environmental conditions or events
such as floods, droughts and so forth. A range of climatic events and conditions known from past
experience to have stimulated distress migration are expected to increase in terms of frequency and
severity in many reasons regions as a result of climate change (Solomon et al 2007, Parry et al 2007)
(Table 1). However, distress migration represents only one end of a continuum of possible climate-
migration outcomes, the other end being environmental amenity migrants who voluntarily seek better
quality environmental conditions (e.g. "snowbird" migration of retirees from northern US to the
sunbelt). Many other possibilities exist between the extremes of environmental refugee and amenity-
seeker, and in many instances it may be difficult to distinguish environmental influences from political,
economic, social, and similar cultural factors that influence migration behavior (Hunter 2005, Massey et
al 2010, Suhrke 1994). For example, often overlooked in discussions of climate change-related migration
is the potential effect on labor migration patterns, as the impacts of climate change reduce income
possibilities in some regions or sectors and open up opportunities in new ones (e.g. economic
development in the warming Arctic creating new development and labour migration there (McLeman
and Hunter 2010)).

Table 1 Expected impacts of anthropogenic climate change reported by IPCC and potential associations
with future population displacements/migrations (adapted from McLeman 2011; McLeman & Hunter
2010)

Expected biophysical
changes (from Solomon et
al 2007, Parry et al 2007)

Regions at risk

Possible linkages to migration

Decreased snow and sea ice
cover

Arctic

Economic migrants arriving to take
advantage of newly accessible
resources

Higher average river runoff
and water availability; more
heavy precipitation events

High latitudes, some wet
tropical areas

Flood-related displacements

Lower average river runoff
and water availability; more

Mid-to low-latitudes and dry
tropics; drought-prone

Water scarcity, drought, & decreased
crop productivity leading to

1

-------
droughts in dryland areas

continental areas; areas
receiving mountain snowmelt

migration, especially higher rates of
rural-urban migration

Coastal erosion, extreme
storms, sea level rise

Low-lying coastal regions, deltas
small island states

Relocation of coastal settlements &
infrastructure; salinization of water
supplies

In the scientific community, human responses to the impacts of climate change are typically described in
terms of vulnerability, which is in turn seen as being a function of the sensitivity of a given population,
region or system to the types of climatic disturbances to which it may be exposed (often simply
described as exposure), and the capacity of the population to adapt (Adger 2006, Parry et al 2007).

Some types of settlement locations are more exposed to migration-inducing climate events than others,
such as low-lying coastal areas and small islands; river valleys and deltas; dryland areas; regions where
precipitation is highly seasonal; and, high latitudes and high altitudes (McLeman and Hunter 2010). In
this context, migration is a process by which exposed individuals or households may adapt to climatic
exposures (McLeman and Smit 2006, Perch-Nielsen et al 2008, Tacoli 2009, Bardsley and Hugo 2010).
There are past examples of state-organized population relocations in response to climate-related events
(e.g. resettlement after drought in East Africa in the 1980s and in Alberta/Saskatchewan, Canada in the
1930s (Ezra and Kiros 2001, Marchildon et al 2008)). However, most climate-related migration occurs as
the result of autonomous responses by households and individuals, and consequently takes on many
different shapes and forms. A single climate event may stimulate a variety of possible migration
responses, as was seen following Hurricane Katrina (Fussell et al 2010).

The greatest amount of climate-related migration presently occurs at intra-national or intra-regional
scales, and this is expected to continue to be the case in coming decades (Adamo and Izazola 2010,
Massey et al 2010, Nelson 2010). In developing regions, where economic systems and livelihoods are
closely tied to agriculture and natural resources, extreme climatic events and conditions are expected to
accelerate already growing levels of rural-to-urban migration (Hunter 2005, McLeman and Hunter 2010).
People at the lowest end of the socio-economic spectrum - particularly landless laborers and tenant
farmers - are the most mobile and most easily displaced (Massey et al 2010). Landowners, business
operators and others at the upper end of the socioeconomic spectrum will also experience economic
hardship, but are more likely to resist migration because their household capital is tied to land and other
assets that are not transportable (McLeman and Smit 2006). Cyclical intra-regional migration in response
to seasonal variability in precipitation and periodic droughts has already long been practiced in Sudano-
Sahelian Africa and rural South Asia and this is expected to continue and potentially grow (Deshingkar &
Start 2003, Hampshire 2002, Mortimore and Adams 2001, Nyong et al 2006).

International movements of people are also expected to increase in response to climate change,
particularly along established migration routes and making use of social networks and transnational
communities (McLeman and Hunter 2010). This belief is supported by evidence from recent climate-
related migration movements, including examples involving the US. For example, Feng et al (2010) have
observed that migration from Mexico to the US surges when drought conditions exist in rural Mexico.
Hurricane Mitch was followed by a pulse of Honduran migration into neighbouring countries and to the
US (Figure 1). Popular media have suggested that anthropogenic climate change has already begun
causing migration from small Pacific islands to Australia and New Zealand, but there currently exists no

2

-------
peer-reviewed research to support this suggestion (Mortreux and Barnett 2009). Case study findings
from the EU/UNU-led EACHFOR project on climate and migration,1 which was completed in 2009, may
provide additional insights into international migration prospects under future climate change, but the
results have yet to appear in scholarly journals.

Figure 1: Apprehensions of improperly documented Honduran migrants along southern US border pre-

and post-Hurricane Mitch (Oct-Nov 1998)

Data source: US Department of Homeland Security Office of Immigration Statistics

While there is increasing agreement on the regions and populations most at risk of experiencing climate
change-related migration, quantitative forecasts are few and vary considerably. The most widely-cited
prediction is one made by British ecologist Norman Myers, who suggested there may be 200 million
environmental refugees worldwide by mid- to late century, to be displaced by a variety of
environmental changes including climate change and sea level rise (Myers 2002). Similar predictions
have been made by CARE International (2009), while the relief organization Christian Aid (2007)
suggested as many as one billion people could be displaced from their homes by mid-century from the
combination of anthropogenic climate change and other global environmental changes. McGranahan et
al (2007) maintain a Low Elevation Coastal Zone database and have used it to estimate that 10% of the
world's population (15% of the global urban population) lives within ten metres of sea level, and is
potentially exposed to the impacts by sea level rise.

2. Briefly review the models and data used to estimate these impacts. What factors are most
important to capture in such models when thinking about the migration impacts of climate change over
a long time frame?

Data for estimating climate change-related migration

Lack of reliable data constitutes a severe and ongoing impediment to reliable forecasting of climate
change-related population movements. Data on global-scale population movements are generally
coarse in nature, and those pertaining to environmental stimuli are particularly unavailable (Brown
2008). The Population Division of the UN Department of Economic and Social Affairs estimates the

1 http://www.each-for.eu

3

-------
world's current annual migrant population at slightly more than 200 million (UN DESA 2010); it is not
indicated what proportion migrate for environmental reasons. The United Nations High Commission for
Refugees (UNHCR) reported an estimated 10.4 million refugees worldwide, and another 15.6 million
involuntarily displaced within their own borders at the end of 2009 (the last year for which figures were
reported at time of writing)(UNHCR 2010). Because environmental stimuli do not qualify as valid reasons
for seeking refugee protection, these statistics do not capture people who are involuntarily displaced for
climate-related reasons, and the UNHCR offers no estimates for such categories of people. The UN's
International Strategy for Disaster Reduction and the Centre for Research on the Epidemiology of
Disasters provide annual estimates of the number of people affected by natural disasters affecting 100
people or more per event, broken down by type of disaster (of which some, but not all, are climatic in
nature). These provide crude proxy figures from which to make estimates of involuntary climate change-
related migration. It is important to note, however, that not all of those affected by disasters become
migrants; many resume their former place of residence as soon as it is safe to do so. Furthermore, many
environmentally induced displacements and movements of people are driven by small-scale, frequent or
repetitive events that may not show up in disaster reporting (Gutmann and Field 2010).

Modeling of climate change-related migration

Modeling of climate change-related migration is still an emergent area of research. Much of the current
work to date can be loosely described as spatial vulnerability modeling, having been influenced by
techniques developed in natural hazards vulnerability research (e.g. Clark et al 1998, Cutter et al 2000,
Wilhelmi and Wilhite 2002). These types of models identify areas or populations vulnerable to particular
impacts of climate change by using geographic information systems (GIS) to combine modeled climate
data from general circulation models (GCMs) or regional climate models (RCMs) with various types of
population, agro-economic and/or resource data (e.g., Byravan et al 2010, Mcgranahan et al 2007,
O'Brien et al 2004, Polsky 2004, Vorosmarty et al 2000). From these, assumptions are then made about
the potential for population displacement and migration, as was done for example in the CARE
International 2009 report cited previously. These models can be extended to identify potential sites of
climate change-related conflicts (which would have feedback effects on migration), as is presently being
done at the University of Texas-Austin to identify sites of potential climate change-related conflict in
Africa2 and at Oregon State to identify potential sites of freshwater conflict.3

Migration estimates based on spatial models make an assumption that an increase in exposure to a
particular climatic stress stimulates a corresponding increase in migration (Piguet 2010). This
assumption is inherently unreliable, because climate-migration rarely unfolds in simple stimulus-
response fashion, but is instead heavily moderated by intervening economic, social and cultural
variables (McLeman and Smit 2006, Massey et al 2010). For example, McLeman et al (2010) combined
regional climate data and census information to create a GIS model of drought-related population
change known to have occurred in western Canada in the 1930s. While the model successfully captured
spatial associations between population change and drought for that particular decade at regional
scales, the model has not yet been able to reproduce drought migration patterns in subsequent decades
for the same region. This is because institutional and economic structures changed substantially over
subsequent decades, requiring incorporation of additional data and modification of the underlying

2	http://ccaps.strausscenter.org/about

3	http://www.transboundarvwaters.orst.edu/research/case studies/index.html

4

-------
assumptions of causality built into the model. Enhancing the predictive capacity of spatial vulnerability
models and "ground-truthing" them requires complementary qualitative field research to identify the
factors and interactions (macro-level and context-specific) that transform vulnerability to migration.

Identification of vulnerable areas and populations that might experience climate-related migration is
not, however, the same as being able to quantify the number of likely migrants. A second type of
modeling that may hold promise for climate change migration research is hazard analysis modeling,
which focuses on capturing the migration behavior of individuals or particular population groups (Barber
et al 2000). Somewhat confusingly, the use of the term "hazard" with respect to this modeling method
does not relate to environmental hazard stimuli but is simply a generic term denoting any potential
lifecourse event (e.g. having a child, changing jobs, migrating, etc) that is contingent upon other
variables, one of which could conceivably be changes in climatic or environmental conditions. In general
migration research, this type of modeling has been used to understand the timing of migration events in
response to particular stimuli (i.e. time-hazard modeling (e.g. Odland and Shumway 1993)) and in
identifying potential migration stimuli operating across multiple scales (i.e. multi-level hazard modeling
(e.g. Massey & Espinosa 1997)). The types and quantity of data necessary to apply this type of modelling
to climate change migration are not widely available at present, although it has been applied in studies
of other types of environmental migration, such as the impacts of land degradation on rural migration in
Nepal (e.g. Massey et al 2010). A research group at the University of Sussex, England, is currently
developing a multi-level hazard method described as agent-based modelling to develop forecasts of
climate change migration, a method which derives multiple hypotheses about migrant behaviour from
known migration data to create computer simulations (Kniveton et al 2008). The researchers have been
attempting to apply the method to drought migration in Burkina Faso; results have yet to appear in
scientific literature.

3. Characterize the uncertainty / robustness / level of confidence in these estimates, on average
globally and by region.

There is a great deal of convergence in the research in terms of global and regional scale identification of
areas and populations potentially at risk of experiencing population displacements and distress
migration due to climate change. This situation will likely improve in the short run due to improvements
in the availability of regional climate model data. Reliable local and sub-regional identification of
potential climate change-related distress migration hotspots is not yet widely available and requires
more research.

Existing estimates of future climate change migration numbers are inherently speculative and often
anecdotal, and are consequently viewed with considerable scepticism by many scholars (Massey et al
2010, McLeman and Hunter 2010). This is to be expected given the limited availability and quality of
regional climatic and population data and our weak understanding of the process linkages between
climatic stimulus, migration outcome and intervening socio-economic and cultural processes. Most
climate change-related migration is expected to occur within regions and borders, and is likely to include
not only distress migrants but large numbers of voluntary migrants as well.

No global monitoring program presently exists for capturing environmentally-related population
movements across international borders or internal movements. For particular regions and sub-regions,
researchers have developed detailed datasetsthat include linked environmental information and
population and migration data over particular time periods, with Burkina Faso, Nepal, and Amazonian
Brazil being just some examples (Kniveton et al 2008, Massey et al 2010, Parry et al 2010). These
disparate datasets are not necessarily linkable to create larger scale models, may not cover similar time

5

-------
periods and may not be maintained on an ongoing basis. In summary, reliable forecasts of climate
change migration numbers (as opposed to populations at risk) are many years off.

4. What are the most important gaps or uncertainties in our knowledge regarding the migration
impacts of climate change? What research in this area would be most useful in the near term?

One important area for additional research is in enhancing our understanding of the underlying
connections between climatic stimuli, intervening socio-economic factors and migration decision-
making outcomes. Evidence from known climate-related migration events shows that migration
responses to climatic stimuli are highly variable within and across populations (McLeman and Hunter
2010). Not all households exposed to a given climate event adapt through migration, and not all those
who might migrate do so (McLeman and Smit 2006). Understanding the underlying forces responsible
for differential migration responses is important for translating spatial vulnerability models into reliable
forecasting models. Massey et al (2010) have suggested that migrants may act on the perception of an
impending environmental risk rather than waiting for the actual occurrence of the environmental risk
itself; if so, this is an area that is greatly understudied. Social networks and social capital are also
believed to be significant influences on climate-related migration and therefore warrant further
research attention (Gilbert and McLeman 2010, Massey et al 2010). The potential effects of climate
change on intraregional and international labor migration patterns is virtually unexplored and warrants
close attention, particularly given recent empirical findings regarding the influence of climatic conditions
on labor migration within the Himalayan region and between Mexico and the US (Banerjee 2010, Feng
et al 2010).

A second area of uncertainty, and one where US and international policymakers have an opportunity to
play an important role, is in the creation of a protocol and mechanisms for generating global statistics
on internal and international migration undertaken for environmental reasons. As indicated above,
existing datasets relating to refugees and disaster displacements provide only rough and unreliable
proxies for measuring the effects of climate and other environmental events and conditions on
migration. A global environmental migration monitoring initiative would in principle be a relatively
straightforward undertaking, requiring a simple protocol that might be enacted through an existing
international agreement such as the UN Framework Convention on Climate Change. A range of existing
international institutions, including various UN agencies and the International Organization for
Migration, have the potential wherewithal for collecting and maintaining such statistics and would
require modest incremental resources to do so. Such an initiative would be a particularly useful step
forward in transforming discussion of climate change migration from informed speculation to evidence-
based policy-planning.

References cited

Adamo, S. B., & Izazola, H. (2010). Human Migration and the Environment. Population and
Environment2, 32(2-3), 105-108.

Adger, W. N. (2006). Vulnerability. Global Environmental Change, 16(3), 268-281.

Banerjee, S. (2010). Labour Migration in Response to Rapid Onset Water Hazards in the Hindu Kush-
Himalayas: Is Labour Migration an Adaptation to Impacts of Rapid Onset Water Hazards? Conference
presentation, European Science Foundation Conference, Environmental Change and Migration: From
Vulnerabilities to Capabilities, 5-9 December 2010, Bielefeld, Germany.

6

-------
Barber, J. S., Murphy, S. A., Axinn, W. G., & Maples, J. (2000). Discrete-Time Multilevel Hazard Analysis.
Sociological Methodology, 30(1), 201-235.

Bardsley, D. K., & Hugo, G. J. (2010). Migration and climate change: examining thresholds of change to
guide effective adaptation decision-making. Population and Environment, 32(2-3), 238-262.

Brown, O. (2008). The Numbers Game. Forced Migration Review, 31, 8-9.

Byravan, S., Rajan, S. C., & Bangarajan, R. (2010). Sea level rise: Impact on major infrastructure,
ecosystems and land along the Tamil Nadu coast (p. 44p). Madras: IMFR-CDF, Indian Institute of
Technology.

CARE International. (2009). In Search of Shelter: Mapping the Effects of Climate Change on Human
Migration and Displacement. Washington DC: CARE International.

Christian Aid. (2007). Human tide: the real migration crisis. London: Christian Aid.

Clark, G. E., Moser, S. C., Ratick, S. J., Dow, K., Meyer, W. B., Emani, S., et al. (1998). Assessing the
vulnerability of coastal communities to extreme storms: The case of Revere, MA., USA. Mitigation and
Adaptation Strategies for Global Change, 3(1), 59-82.

Cutter, S., Mitchell, J. T., & Scott, M. S. (2000). Revealing the Vulnerability of People and Places: A Case
Study of Georgetown County, South Carolina. Annals of the Association of American Geographers, 90(4),
713-737.

Deshingkar, P., & Start, D. (2003). Seasonal Migration for Livelihoods in India: Coping, Accumulation and
Exclusion. London: Overseas Development Institute.

El-Hinnawi, E. (1985). Environmental refugees. Nairobi: United Nations Environmental Program.

Ezra, M., & Kiros, G.-E. (2001). Rural out-migration in the drought prone areas of Ethiopia: a multilevel
analysis. International Migration Review, 35(3), 749-771.

Feng, S. F., Krueger, A. B., & Oppenheimer, M. (2010). Linkages among climate change, crop yields and
Mexico-US cross-border migration. Proceedings of the National Academy of Science, 107(32), 14257-
14262.

Fussell, E., Sastry, N., & VanLandingham, M. (2010). Race, socioeconomic status, and return migration to
New Orleans after Hurricane Katrina. Population and Environment, 31(1-3), 20-42.

Gilbert, G., & McLeman, R. (2010). Household access to capital and its effects on drought adaptation and
migration: A case study of rural Alberta in the 1930s. Population and Environment, 32(1), 3-26.

Gutmann, M. P., & Field, V. (2010). Katrina in historical context: environment and migration in the U.S..
Population and Environment, 31(1-3), 3-19.

Hampshire, K. (2002). Fulani on the move: Seasonal economic migration in the Sahel as a social process.
The Journal of Development Studies, 38(5), 15-36.

Hugo, G. (1996). Environmental concerns and international migration. International Migration Review,
30(1), 105-131.

7

-------
Hunter, L. M. (2005). Migration and Environmental Hazards. Population and Environment, 26(4), 273-
302.

Kniveton, D. R., Schmidt-Verkerk, K., Smith, C., & Black, R. (2008). Climate change and migration:
improving methodologies to estimate flows. Geneva: International Organization for Migration.

Marchildon, G. P., Kulshreshtha, S., Wheaton, E., & Sauchyn, D. (2008). Drought and Institutional
Adaptation in the Great Plains of Alberta and Saskatchewan, 1914-1939. Natural Hazards, 45(3), 391-
411.

Massey, D. S., Axinn, W. G., & Ghimire, D. J. (2010). Environmental change and out-migration: evidence
from Nepal. Population and Environment, 32(2-3), 109-136.

Massey, D. S., & Espinosa, K. E. (1997). What's driving Mexico-US migration? A theoretical, empirical and
policy analysis. American Journal of Sociology, 102(4), 939-999.

Mcgranahan, G., Balk, D., & Anderson, B. (2007). The rising tide: assessing the risks of climate change
and human settlements in low elevation coastal zones. Environment and Urbanization, 19(1), 17-37.

McLeman, R., Herold, S., Reljic, Z., Sawada, M., & McKenney, D. (2010). GIS-based modeling of drought
and historical population change on the Canadian Prairies. Journal of Historical Geography, 36(1), 43-56.

McLeman, R., & Hunter, L. M. (2010). Migration in the Context of Vulnerability and Adaptation to
Climate Change: Insights from Analogues. Wiley Interdisciplinary Reviews: Climate Change, 1(3), 450-
461.

McLeman, R., & Smit, B. (2006). Migration as an Adaptation to Climate Change. Climatic Change, 76(1-2),
31-53.

McLeman, R. (2011, in press). Climate change, migration, and critical international security
considerations. Geneva: International Organization for Migration.

Mortimore, M. J., & Adams, W. M. (2001). Farmer adaptation, change and "crisis" in the Sahel. Global
Environmental Change, 11(1), 49-57.

Mortreux, C., & Barnett, J. (2009). Climate change, migration and adaptation in Funafuti, Tuvalu . Global
Environmental Change, 19(1), 105-112.

Myers, N. (2002). Environmental refugees: a growing phenomenon of the 21st century. Philosophical
transactions of the Royal Society London: Biological sciences: Series B, 357(1420), 609-613.

Nelson, D. R. (2010). Climate change: understanding anthropogenic contributions and responses.
Population and Environment, 31(5), 283-285.

Nyong, A., Fiki, C., & McLeman, R. (2006). Drought-Related Conflicts, Management and Resolution in the
West African Sahel: Considerations for Climate Change Research. Die Erde (The Earth), 137(3), 223-248.

O'Brien, K., Leichenko, R., Kelkar, U., Venema, H., Aandahl, G., Tompkins, H., et al. (2004). Mapping
vulnerability to multiple stressors: climate change and globalization in India . Global Environmental
Change, 14(4), 303-313.

8

-------
Odland, J., & Shumway, J. M. (1993). Interdependencies in the timing of migration and mobility events.
Papers in Regional Science, 72(3), 221-237.

Parry, L., Day, B., Amaral, S., & Peres, C. A. (2010). Drivers of rural exodus from Amazonian headwaters.
Population and Environment, 32(2-3), 137-176.

Parry, M. L., Canziani, O. F., Palutikof, J. P., Der Linden, P. J. van, & Hanson, C. E. (2007). Contribution of
Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change,
2007. Cambridge: Cambridge University Press.

Perch-Nielsen, S., Battig, M., & Imboden, D. (2008). Exploring the link between climate change and
migration. Climatic Change, 91(3-4), 375-393.

Piguet, E. (2010). Linking climate change, environmental degradation, and migration: a methodological
overview. Wiley Interdisciplinary Reviews: Climate Change, 1(4), 517-524.

Polsky, C. (2004). Putting Space and Time in Ricardian Climate Change Impact Studies: Agriculture in the
U.S. Great Plains, 1969-1992. Annals of the Association of American Geographers, 94(3), 549-564.

Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K. B., et al. (2007). IPCC Fourth
Assessment Report: Climate Change 2007: The Physical Science Basis. Cambridge: Cambridge University
Press.

Suhrke, A. (1994). Environmental Degradation and Population Flows. Journal of International Affairs,
47(2), 473-496.

Tacoli, C. (2009). Crisis or adaptation? Migration and climate change in a context of high mobility.
Environment and Urbanization, 21(2), 513-525.

UN DESA (Department of Economic and Social Affairs), Population Division. (2010). International
Migration 2009. New York. Retrieved from

http://www.un.org/esa/population/publications/2009Migration Chart/2009lttMig chart.htm
UNHCR. (2010). UNHCR Statistical Yearbook 2009. Geneva.

Vorosmarty, C. J., Green, P., Salisbury, J., & Lammers, R. B. (2000). Global water resources: Vulnerability
from climate change and population growth. Science, 289(5477), 284-288.

Wilhelmi, O. V., & Wilhite, D. A. (2002). Assessing vulnerability to agricultural drought: A Nebraska case
study. Natural Hazards, 25(1), 37-58.

9

-------
Migration Impacts of Climate

Change



• r.Vl/i.

-------
Questions

•	What regions most vulnerable?

•	What models/data are available?

•	How confident are we in these?

•	Gaps & opportunities

-------
Predictions of a coming exodus

KITCHENER • CAMBRIDGE • WATERLOO A3

Friday, March 5,2004

Global climate
change could lead
to refugees: study

OTTAWA

Canada could see an influx of envi-
ronmental refugees from countries
rocked by hurricanes, droughts and
other disturbing effects of global cli-
mate change, says a study prepared for
the national spy agency

Others might be drawn to Canada
as icy regions of the vast North be-
come warmer and more hospitable to
marine traffic, posing possible new se-
curity challenges.

"Climate-related disruptions of hu-
man populations and consequent mi-

Last summer Europe's
• hottest in 500 years

PAGED14

Many scientists believe human ac-
tivity has prompted global warming
that will lead to an increase in average
temperatures. Other anticipated
changes include rising sea levels, en-
hanced risk of drought, more frequent
and intense storms, and other extreme
weather events.

The paper notes sea ice in Canada's
Northwest Passage has thinned to pre-

Before the Flood

By SUJATHA BYRAVAN and SUDHIR
CHELLA RAJAN
Published: May 9, 2005

Cambridge, Mass. —. One of the paradoxes of
global warming is that developing countries,
which were not responsible for most of the
greenhouse gas emissions that are changing the
climate and did not reap the benefits of
industrialization, will bear the brunt of the
consequences. One of these consequences will
be rising seas, which in turn will generate a
surge of "climate exiles" who have been

-------
Media identification of the first
climate change refugees

Shishmaref, Alaska

Cataret Islands

Lake Chad region

-------
Predictions of future environmental

refugees

•	Up to 1 billion by 2050 (Christian Aid)

•	200 million by 2050 or 2100 (Norman
Myers, CARE International press release)

•	50 million by 2010 (UNU 2005 press
release)

•	10% of world population lives within 10m
of sea level (Mcgranahan et al 2007)

-------
Context

cn
c
o

1000
900
800
700
600
500
400
300
200
100
0

UNHCR
refugees
actual

UNU

forecast
2010

Mye rs
forecast

world
migrant
actual

within 10m Christian Aid
of sea forecast

-------
Existing forecasts of climate
change migration

•	Identify areas/populations exposed to
negative CC impacts

•	Exposure f migration

•	Climate-migration not simple stimulus-
response

•	Intervening socio-economic, cultural &
institutional factors

-------
Climatic stimuli known to be
associated with migration

•	Sudden onset events (e.g. hurricanes,
tropical storms, extreme rainfall events)

•	Persistent conditions (e.g. drought,
changes in monsoons)

•	Climate change expected to exacerbate
existing stimuli, create new ones (e.g. sea
levels, Arctic ice)

-------
Hurricane Katrina

-------
New Orleans population

post-Katrina

g 500,000
CO

450,000

'3

3 400,000

"cij 350,000

£ 300,000

| 250,000
0)

£ 200,000

O 150,000
c

¦2 100,000
_co

g_ 50,000
o

Q-	0

2005	2006	2007	2008	2009

Data source: US Census bureau

http://www.census.gov/popest/counties/CO-EST2QQ9-Q1 .html

-------
-------
Undocumented Hondurans
arrested at US-Mexico border

Hurricane Mitch strikes Honduras Oct-Nov 1998

18000

16000

1996

1997

1998

1999

2000

2001

2002

-------
Drought & migration

• Feng et al (2010) find that a 10% decrease
in agricultural production in Mexico due to
drought is associated with a 2% rise in
Mexican migration to US

Feng SF, Krueger AB, Oppenheimer M. (2010) Linkages among climate change,
crop yields and Mexico-US cross-border migration. Proceedings of the National
Academy of Science. 107(32): 14257-14262.

-------
Where will climate change
generate migration stimuli?

•	Arctic (permafrost, sea & land ice melt)

•	High latitudes, wet tropics (heavy
precipitation events, floods)

•	Mid- to low-latitudes, dry tropics (drought,
water scarcity)

•	Coastal plains, deltas, small islands
(erosion, storm surges, salinization)

IPCC 2007 distilled by McLeman & Hunter 2010

-------
Differential outcomes

Climate events, conditions don't always
stimulate migration

Multiple migration outcomes can be
generated by single climate event

Why? What distinguishes migrants from
non-migrants?

New Orleans population post-Katrina

500,000 -j—

450,000
400,000
350,000
300,000
250,000
200,000
150,000
100,000
50,000

-------
Vulnerability (V)

• Potential to experience loss or harm

V = f( E,S,A)

E = exposure (i.e. climatic stimulus)
S = sensitivity of the exposed system
A = adaptive capacity

-------
V = f( E,S,A)

Adaptive capacity

• Options for adapting to drought not the
same in rural Nigeria as in rural
Saskatchewan

-------
Migration as adaptation

•	Migration is one of a range of potential
adaptive responses to environmental
stress

•	Is presently used in many parts of world

•	Is typically initiated at the household level

•	Is not available to everyone

•	Is not always used by all who might do so

•	In worst cases, could be the only
adaptation

-------
Climatic stress

Vulnerable
population

Adaptation

Other than
migration

Migration

feedback effect of population change

Simplified from McLeman R, Smit B. (2006) Migration as an Adaptation
to Climate Change. Climatic Change. 76(1-2):31-53.

-------
Climatic stress

Why might peopl

Vulnerable
population

Adaptation

feedback effect of population change

Simplified from McLeman R, Smit B. (2006) Migration as an Adaptation
to Climate Change. Climatic Change. 76(1-2):31-53.

-------
What else motivates people to

migrate?

•	Opportunity/benefit seeking (economic,
public services)

•	Household risk diversification

•	Macro-scale systems

•	Cultural norms

•	Lifestyle

•	Bright lights-big city

•	Love

•	Persecution, fear of violence

-------
What else motivates people to

migrate?

•	Opportunity/benefit seeking (economic,
public services)

•	Household risk diversification

•	Macro-scale systems

Cultural norms
Lifestyle

Bright lights-big city (e>
Love

Persecution, fear of violence

Clima

influence/interact
with any of these
(except maybe love)

-------
What do we tend to focus on?

GWYNNE DYER

the

COMING
ANARCHY

SHATTERING THE DREAMS f
OF THE POST-COLD WA«

ROBERT D.
KAPLAN' "

Author of BALKAN GHOSTS

CLEO

PASKAL

HOW ENVIRONMENTAL.
ECONOMIC AND

POLITICAL CRISES WILL
REDRAW THE WORLD MAP

-------
But most observed climate-
related migration...

•	Is not conflict-related

•	Is internal/intra-regional

•	When international, follows established
routes, transnational communities

•	Is shaped by other motivations as well

Mexican migration to US

Crop yields in Mexico

Feng et al 2010

-------
Climate-migration models

-------
Historical climate-migration

modeling

•	Use known climatic data and known
population change data from past events

•	Generates learning analogues

•	Can be ground-truthed

-------
Canadian drought refugees, 1930s

-------
Drought & rural population
loss.1931-36, Canadian prairies

McLeman et al. (2010). GlS-based modeling of drought and historical population change on the Canadian
Prairies. Journal of Historical Geography, 36, 43-56.

-------
What distinguished migrants
from non-migrants?

Hfl

Qualitative research

-------
Who migrates?

More likely

•	Young, healthy, skilled,
educated

•	Middle class

•	Uncertain land tenure

•	Family ties elsewhere

Less likely

•	Wealthier classes,
landowners (especially
good land), owners of
fixed assets

•	Those with strong local
social networks

•	Poor, destitute

•	Elderly, infirm, broken
families

-------
Other types of modeling

Spatial vulnerability models

•	GIS-based modeling to identify places/
populations at future risk (potential
hotspots)

•	Are silent on likelihood of migration
outcomes

-------
CIESIN models for "In Search of

Shelter" report

WindowsXP: T ake a scree...	Q) ciesin.Columbia.edu/binarie...

f- CO ciesln.columbia.edu.'binaries/web/global/news^Ooa/climmigr-rpt-juneOg.pdf

CD
%

§

O
CD

O

3
3

n"

c
3

CD
O

 /
3'

poor satisfactory good excellent

3.2 Mexico and Central America: Migration in response to drought and disasters

-------
o w

a Is

Byravan et al models of sea level rise
& coastal settlement, Tamil Nadu

MAP SHOWS LOCATIONS OF SEZs,
PORTS, POWER STATIONS, ECO-ZONES
OVER POSSIBLE FLOOD RISK AREAS.

PH

I "<

S£Z Location*

1

' f

i

4

o

o



73

CD

CL

6"

o

o

5*



w

p.

3

o

—h

3

b

Q_

w

Q.

-p

CD

to"

Q.

CD
-------
Other types of modeling

•	Multi-level hazard analysis models

•	Does not refer to natural hazards, but is
statistical tool to isolate the relative effect of
particular variables on migration outcomes

•	Used by Massey, Axinn, others to estimate
determinants of Mexico-US migration,
environmental drivers of migration in rural Nepal

Massey, D. S., Axinn, W. G., & Ghimire, D. J. (2010). Environmental change and
out-migration: evidence from Nepal. Population and Environment, 32(2-3), 109-136.
Massey, D. S., & Espinosa, K. E. (1997). What's driving Mexico-US migration? A
theoretical, empirical and policy analysis. American Journal of Sociology, 102(4), 939-999.

-------
Other types of modeling

•	Multi-stage regression model of known &
estimated migration + crop yield change

•	Then combined with crop simulation
models for forecasting

•	E.g. estimating potential Mexico-US
migration (Feng et al 2010)

Feng SF, Krueger AB, Oppenheimer M. Linkages among climate change, crop
yields and Mexico-US cross-border migration. Proceedings of the National
Academy of Science. 2010; 107(32): 14257-14262.

-------
Other types of modeling

•	Agent-based modeling

•	Simulation modeling to attempt to replicate
& then predict interactions (in this case
between climatic stimuli & migration
outcomes)

•	Being used by group at U of Sussex to
model drought migration in Burkina Faso

Kniveton, D. R., Schmidt-Verkerk, K., Smith, C., & Black, R. (2008). Climate
change and migration: improving methodologies to estimate flows. Geneva:
International Organization for Migration.

-------
Agent-based model by Smith for
Burkina Faso migration

http://www.informatics.sussex.ac.uk/users/cds21/abm/

CjWindowsXP: Takeascree... J US Burkina Faso - Simulation * NewTab

¦[_ f (51 x |

f CO www,informaflcs.sussex.ac.uk/users/cds21/abm/Burkina%20Faso.html

	~jj,

. ~ ' 1	1 M

500-699 mm
700-899 mm
> 900 mm





B"

B

-------
Challenges & opportunities

-------
Challenges

•	Data availability, reliability

•	No single global database

•	Fragmented data for various regions, time
periods

•	Even where you have census data for
population change/migration, reasons for
migration often missing

•	Proxy data: disaster displacements (not
the best)

-------
Challenges

•	Understanding system linkages

•	Role of intervening variables (e.g.
perception, social networks, labour
migration pressures/opportunities...)

•	Uncertainty about future
frequency/severity of migration-assocated
climatic stimuli

-------
Opportunities

To develop monitoring & data collection
protocols

To enhance empirical research into
environment & migration linkages

To develop & improve migration models
climate change models improve

-------
Thanks! Merci!

Robert McLeman

Associate Professor

nm

u Ottawa

Departement de geographie
Department of Geography

e-mail: rmcleman@uottawa.ca
web: http://www.qeoqraphv.uottawa.ca/prof/rmcleman.htm

-------
inal Panel Discussion: Food for
Thought

Anthony C. Janetos, Director
Joint Global Change Research Institute

28 January 2010

Pacific Northwest

NATIONAL LABORATORY
Proudly Ofitnltd Iry B
-------
Our Charge

~	How do we take all we have learned in the past two days to
improve reduced-form integrated assessment models (lAMs)?

~	In which sectors or categories has research on physical
impacts of climate change or methods for valuing the
associated damages developed beyond what is currently
represented in reduced-form lAMs? Which of these can most
readily be incorporated into modified versions of existing lAMs?
How could one approach modeling the interactions across
individual impact sectors?

~	From the perspective of your discipline/area of expertise (e.g.
economist, scientist, 1AM modeler), what are the most
important gaps or uncertainties in our knowledge regarding the
impacts of climate change and associated economic
damages? What research would be most useful in the near vs.

-------
The Short Version of My Answers

How do we take all we have learned in the oast two days to improve reduced-
form integrated assessment models (lAMs)?

~	Many different possibilities, but not clear to me that it's always a good idea.

In which sectors or categories has research on physical impacts of climate
change or methods for valuing the associated damages developed beyond what
is currently represented in reduced-form lAMs?

~	Pretty much all of them with respect to physical impacts. Methods for economic
valuation appear at first glance not to have advanced nearly as much. Methods
for incorporating valuation and physical impacts into reduced form models are
increasingly sophisticated, but we need to be careful about what either the data or
the models are capable of doing.

Which of these can most readily be incorporated into modified versions of existing
lAMs?

~	Relatively few, without some fairly extensive thought given to thresholds, non-
linear behavior, and process-level understanding.

How could one approach modeling the interactions across individual impact
sectors?

~ Need explicit representation of the sectors and both the economic and physical
factors (e.g. competition for water and land) that connect them.

-------
Background

~	The challenge to all the modelers in the workshop has
essentially been framed in a "social cost of carbon"
framework

~	Assumes that we have good central estimates of a large
number of parameters, both physical and economic, but is
this reasonable?

~	Many reasons in particular cases that we should be
humble about our ability to generate really good
estimates, so I will highlight only a few...

-------
Background

~	Ubiquity of "bad behavior" in physical systems

~	Thresholds are routine phenomena- we've looked at
much of the literature on ecological thresholds, and in
some ways the greater challenge is finding a system that
does not respond in this way

~	But our ability to model such changes is rudimentary -
yesterday, saw the example of the sensitivity of crop
productivity to temperature thresholds, many other
examples where there is an ecosystem threshold that is
not necessarily related to an extreme in climate
variability...

-------


i . 1: <¦ ^	• 1 w

*	;

¦• '^^SLL^>ir'	i

,!V^> • E i' *4>	'

. JM*,, t' ' V J<





SW-



n m , J
&



« N *

(it'

j§

* "JA

s

i/St'-

«

& 5

\m12



!%



K " V'



KM

IKkSi

Hi v

HW

' ,V'

• t.tw i

&•!

w

«J!SSP

m

B

w

m

m

;m

i'ffllv

h.inj

!W

W?I'

I'W

M



* »l

Ml

:M

* i •<

'./ ,

i

I

LU

i«SM

mlm

1

i

iViW

1

Iv!



f.i

-------
Background

~	The major drivers of big changes over the past half-
century in both managed and unmanaged ecosystems
are in fact human-driven

~	Land-cover changes as just one example

~	We need to be able to take these sort of changes into
account; heard this point made in a very interesting talk
on forests this morning

-------
Pacific Northwest

NATIONAL LABORATORY
Proudly Operated by BalKNie Since 1965

UNIVERSITY OF

'MARYLAND

-------
Background

~	Interaction among sectors is clearly a first-order problem,
not a second-order problem as we have typically treated it
in impact assessments

~	Competition for water among agricultural, energy,
industrial and other human uses - and ecosystem
uses/needs is just the tip of the iceberg

~	Competition for land among economically productive uses
(e.g. agriculture, forestry), provision of ecosystem
services that are not valued in markets, provision of
services that are not currently valued in markets, but
could be in different policy regimes

~ Aggregation/disaggregation issues turn out to be
extremely important, and this is a challenge for the
response-surface approach

-------
Background

Many of the ecological models that are being used have
well-known deficiencies that are not being taken into
account

¦	They do a quite poor job of parameterizing the C02-driven
increases in water-use efficiency, for example

¦	They typically do not include the type of threshold responses
mentioned before

¦	They underplay or don't include biotic interactions like pests and
pathogens

¦	Some, including the DGVM's, are essentially unverified, and how
they could be verified is not all that clear

Some of the potential ecological changes are still in the
category of being theoretically possible, but our
techniques for projecting them are very preliminary (e.g.
extinction risk, climate envelope modeling for range shifts]



Pacific Northwest

NATIONAL LABORATORY
Proudly Operated by BalKMie Since 1965

Sc UNIVERSITY OF

w MARYLAND

-------
Proudly Operated by BalKHie Since 1965

-------
Background

The technique of inferring or developing simple, statistically- or
model-based response functions for use in reduced form lAMs
faces some very difficult challenges

~	My personal conclusion is that these techniques have utility for
understanding some of the interactions of climate impacts and
economic concerns in today's world - AND THIS IS REALLY
IMPORTANT TO DO!

~	But their ability to do projections that are intrinsically far beyond
the range in which the original parameterizations and damage
functions have been developed is likely to be quite limited

~	My second conclusion is that a more process-based approach
to linking concerns about impacts with their economic
consequences and with the economic and technological
evolution of both the impact sectors and climate policy is more
likely to be helpful at the end of the day

-------
Some Thoughts on the Value Added from a New Round of
Climate Change Damage Estimates

Gary Yoheac and Chris Hopeb

EPA/DOE Workshop on Improving the Assessment and Valuation of Climate
Change Impacts for Policy and Regulatory Analysis:

Research on Climate Change Impacts and Associated Economic Damages

January 28, 2011

a Huffington Foundation Professor of Economics
and Environmental Studies
Wesleyan University
238 Church Street
Middletown, CT 06480, USA

b Judge Business School
University of Cambridge
Cambridge CB2 1AG, UK

c Contact author:
gyohe@wesleyan.edu

1

-------
The organizers of the workshop on "Research on Climate Change Impacts and Associated
Economic Damages" asked us (among others) to reflect briefly on three summary questions. The
first focused on improving reduced-form integrated assessment models. The second asked for an
assessment of recent progress with particular attention paid to interactions across sectors. The
third invited us to identify important gaps and uncertainties. We will not attempt to answer any of
these questions comprehensively. We will, though, offer some hopefully provocative thoughts that
address the content of each of them, taken in turn, from a value-added perspective. In doing so, we
hope to speak to the issues raised by the broader title of the two-day meeting: "Improving the
Assessment and Valuation of Climate Change Impacts for Policy and Regulatory Analysis".

Our first set of comments expresses some concern about the value of specific contributions
to integrated assessments and their products. To that end, Section 1 offers a warning to beware of
analyses that are so narrow that they miss good deal of the important economic ramifications of the
full suite of manifestations of climate change; i.e., they miss interactions in the climate system that
allow climate change, itself, to be a source of multiple stress even within one particular sector.
Section 1 also makes the point that the largest value added by updated economic analyses of
impacts may be found in using their results to identify where more careful consideration of site-
specific and path dependent adaptation might be most productive.

Our second set of comments focuses attention on one of the most visible products of
integrated assessment modeling - estimates of the social cost of carbon which we take as one
example of aggregate economic indicators that have been designed to summarize climate risk in
policy deliberations. Our point, argued in Section 2, will be that these estimates are so sensitive to a
wide range of parameters that improved understanding of economic damages across many (if not
all) climate sensitive sectors may offer only limited value added. Some of these parameters reflect
interactions across sectors. Others fall within the prerogative of decision-makers who use the
results of integrated assessment to judge the value of mitigation policy. Still others fall within the
prerogative of "Mother Nature"; and we must humbly admit that she is not being particularly
forthcoming in providing information from which we can glean reliable and timely estimates. We

2

-------
fear, in other words, that the very focus of this workshop may have been guilty of a "type-three
error" - that is, in the words of Richard Tol, "barking up the wrong tree".

Having cast some doubt on the ability of improved estimates of economic damages to
increase the value of economic damage estimates in integrated assessment modeling designed to
inform climate policy deliberations, we offer an alternative approach in Section 3. We begin with
the idea that climate policy can perhaps best be understood as a question of setting a carbon-
emissions budget for a period of decades rather than centuries - say limiting cumulative emission
from the United States to between 170 to 200 gigatons through 2050 as suggested in the report of
the "Limiting Panel" to America's Climate Choices [NAS (2010)]. Working from there to suggest
how to set a price on carbon, we end this brief note by describing implicitly a research agenda that
could (a) effectively inform mitigation decisions while, at the same time, (b) providing economic
estimates for aggregate indicators like the social cost of carbon. It is these estimates that can be
applied to considerations of the value (or harm) caused by the carbon-emission consequences of
non-climate regulations and other market interventions. We believe that working out the technical
and practical details of such an approach could pay the greatest dividends - an approach that would
use the results of integrated assessment models to characterize policy context and judge economic
tradeoffs.

Section 1: Beware of Spurious Precision and Incomplete Models.

The workshop offered glimpses into current work across a wide range of sectors and
contexts, but we are worried that any single paper could be taken as comprehensive coverage of
what is known and/or what needs to be known. Take, for example, the contribution by
Mendelsohn, Emanuel, and Chonabayashi on tropical cyclone damage. We do not mean to pick on
this paper, but it does speak to climate impacts in a sector with which we have some familiarity.
The authors used historical records to calibrate simulations of future cyclones with and without
climate change using a collection of 4 global circulation models along the Alb SRES storyline. Based
on statistical associations of storm intensity and observed damages, they conclude that "Increasing
future income and population is predicted to increase annual tropical cyclone damages from $26
billion to $55 billion even with the current climate. However, damages as a fraction of GWP are
expected to fall from their current rate of 0.04 percent in 2010 to 0.01 percent in 2100."

While the analysis is solid as far as it goes, we are afraid that it makes only a small
contribution to our understanding of vulnerability to coastal storms that could easily be

3

-------
misinterpreted for two reasons. First of all, while the analysis did use four alternative climate
models to simulate the future implications of 70,000 simulated cyclones, it did not provide any
insight into the true range of possible damage futures. It did not, for example, explore alternative
socio-economic futures (either within Alb with respect to geographical distribution of populations
and development or across alternative story-lines). Nor did it explore uncertainty boundaries
defined by its estimates of damage elasticities (with respect to income and population). It did not
even explore uncertainty boundaries defined by any portion of the reported range of equilibrium
climate sensitivity - an increasingly common feature of contemporary impacts analyses. It follows
that the $26 to $55 billion range must be understated; it is easy to envision not-implausible
economic futures for which $26 billion is too high, but it is equally easy to envision futures for
which $55 billion is way too low.

The analysis also falls well short of providing comprehensive estimates of the economic
damage of either tropical cyclones or coastal storms more generally. This is, in part, because it
completely ignores major components of potential damage. Loss of life comes to mind in this
regard; and while ignoring this risk avoids the controversy about international distributions of the
value of a statistical life, it does so at the expense of severely limiting the coverage of the reported
estimates.

In addition, because the analysis relies heavily on central tendencies in its statistical
representation of future damages, it misses entirely the enormous inter-annual variability in
cyclone damage about which insurance and re-insurance companies would be far more interested.
Katrina dominates any damage time series over the past few decades in a way that is not
reasonably reflected in the annual means (or medians, for that matter). Indeed, only researchers
who recognize that the sheer magnitude of a Katrina-like outlier cannot be excluded from any
year's potential exposure will be able to appreciate the enormous adaptation challenge that it
poses. Spreading annual risk geographically may not be enough for tropical cyclones. It may be
necessary to spread risk over time, as well; but to do so would require regulator reform of the sort
now being suggested by Kunreuther and Useem (2010).

Mendelsohn, et al. also ignore the contribution of even modest sea level rise to damages
associated with storms of all shapes and sizes. The authors are, in fact, completely wrong when
they assert on the basis of simple statistical analysis of damages (in the text that describes the
content of Figure 5) that "common small storms are not different before and after climate change."
Kirshen, et al (2008), Rosenzweig, et al. (2010), and others have argued convincingly that sea level

4

-------
rise elevates storm surges associated with any coastal storm and therefore amplifies any storm's
potential for causing economic damage. The mechanism is really quite simple. Elevated storm
surges driven by routine sea level rise can make what is now, for example, a 20-year storm look like
the current 50-year storm in terms of economic exposure. In other words, what is now the 50-year
storm in terms of economic consequence can turn into an every other decade (on average) event at
some point - and for some locations, some time in the relatively near-term future. Table 1 offers
some evidence of what this association could mean for what is currently the 100-year storm in
Boston and New York along two SRES emissions trajectories and central tendency sea level rise.

Figure 1 brings this simple process (for storms of all dimensions) into geographic focus by
plotting the frequency of threshold anomalies per year for 5 different locations along the north-
eastern coastline of the United States from 1920 through 2005; these are locations that have
experienced, on average between 2.6 cm and 2.8 cm of sea level rise per decade since 1920. The
various panels of Figure 2 show what this process understanding means for an urban coastal
community in Boston. Offered simply as an illustrative example, it shows damage profiles (without
adaptation) at 20-year increments that were drawn randomly from probabilistic representations of
historical weather patterns (without altering intensity or frequency in anticipation of climate
change). This historical pattern was then superimposed upon sea level trajectories that reach 100
cm and 60 cm by 2100.

Notice that damages from the worst 5% of the storms (including, perhaps, an occasional
representation of a hurricane or a severe winter nor-easter with hurricane force winds) are
expected to climb over the century by as much as 250% (along the 100 cm trajectory); this is
flooding analog to what Mendelsohn, et al. estimate as a function of storm intensity that is implied
by the first rows of Table 1. More importantly, notice that damages from the other 95% of the
storms are expected to increase similarly and persistently over time at rates that are determined by
the underlying sea level rise scenario.

Clearly, these risk profiles show that common storms can be quite different under climate
change when the local characteristics of climate change are more comprehensively represented; and
clearly, those differences can produce some relatively large economic consequences. These sorts of
risk profiles can also help decision-makers decide how and when to respond to a growing climate-
related risk. Table 2, for example, charts the increase in the estimated expected internal rate of
return for an investment in protective infrastructure that would (a) cost $390 million (in real
dollars) to implement, (b) commit the city to 10% maintenance expenses thereafter, and (c) not

5

-------
guarantee complete protection from the upper end of the damage distribution. These economic
estimates show that the need for adaptation could be urgent (or not), depending on the degree to
which this public investment would complement private investment [see, e.g., Ogura and Yohe
(1977)] and the speed with which sea level are seen to be rising rise.

Section 2: Value Added for Aggregate Economic Indicators like the Social Cost of Carbon.

Downing and Watkiss (2003) warned that economic analyses of climate change damages
failed to cover much of what might be in store for the planet (especially in terms of socially
contingent consequences and abrupt events). While little has changed to allay their concerns, this
section will not rehash their arguments. It will, instead, ask (and, to some degree, answer) a simple
question: "What difference would marginal contributions to economic damage estimates (for the
impacts and sectors that we can model) make on the major economic aggregates that some believe
most significantly inform climate policy deliberations?" We know that uncertainty compounds
through the climate system as we move from (a) economic activity to (b) greenhouse gas emissions
to (c) changes in their atmospheric concentrations to (d) changes in global mean temperature and
other climate variables to (e) impacts in physical and biological systems to (f) economic estimates of
associated damages with and without adaptation. Since new estimates of economic damages speak
only to the last (italicized) association, it would seem fool-hearty not to hypothesize that the
answer to this question is "Not much!"

To begin to explore the potential validity of this hypothesis, we used the latest version of
the PAGE integrated assessment model (PAGE 09) to track the implications of three possible
implications of a new round economic damage estimates (of the sort presented at the workshop) on
the distribution of estimates of the social cost of carbon.1 The baseline scenarios worked from a
representation of the SRES A1B storyline whose default settings produced the range of temperature
trajectories depicted in Figure 3. The three experimental changes from the default settings were
designed to reflect improved (or at least altered) understanding of economic damages across the
board. Results (calibrated in terms of the social cost of carbon) from the default-setting baseline
and three experiments are recorded in Table 3 and depicted graphically in Figure 4. In every case,
the summary statistics of Table 3 and the histograms of Figure 4 were produced from monte carlo
simulations that involved 100,000 distinct manifestations of the complete set of underlying random

1 See Hope (2006) for details of the structure of the PAGE models. For updates included in PAGE09, see
http: //climatecost.cc/images/Policy brief 4 PAGE09 Model vs 2 watermark.pdf.

6

-------
variables that PAGE 09 can accommodate.

In the first experiment (Case A in Table 3), new economic research was assumed to reduce
the range of the parameters that calibrates estimates for economic sectors and coastal zones by
50% without changing their means or the modes. In the second experiment (Case B), new research
was assumed to reduce the modes by 50%. Since the distributions of all parameters are triangular
in PAGE 09, reducing the mode by 50% reduces the mean by almost 9% and puts an additional 17%
of the probabilistic density below the old mean. This might not seem like much from a modeling
perspective, but we submit that it reflects what would be a gigantic change against conventional
wisdom that is anchored by the inertia of decades of earlier research. The third experiment (Case C
in Table 3) repeats Case B in the opposite direction; i.e., the mode is increased by 50%.

Given that these results are based on 100,000 runs, there is a 95% chance that another set
of 100,000 runs would produce means in every case that are within $2 of these reported values.
The summary statistics therefore strongly suggest that it would be unlikely that reducing the range
of economic damage estimates would change the mean estimate for the social cost of carbon even
though the 99th percentile estimate might fall by more than 10%. Cases B and C, where the mode
changed, did show significant changes in the mean and slight changes in the 5th to 95th percentile
ranges; but these changes are nothing to write home about in terms of making policy. Indeed, the
histograms portrayed in Figure 4 depict vivid portraits of robust insensitivity to new information
about economics. Estimates range from $0 through nearly $10,000 or more per ton in every case,
but the modal estimates all lie between $25 and $50 per, the median estimates all fall in the
neighborhood of $50 per ton, and the means (excluding the top 1% of the estimates) all hover
between $80 and $90 per ton (adding the top 1% of the estimates would add roughly $20 to these
values).

The relative insensitivity of these statistical values is supported by analysis of the marginal
contributions of uncertainty in the underlying random variables to the overall variability in
estimates of the social cost of carbon. Transient climate response dominated for every case,
followed (among sources reflecting human attitudes or activities) by the pure rate of time
preference (about 60% as influential and transient sensitivity), relative risk or inequity aversion
(about 50% as influential), indirect effects of sulfates (about 25% as influential), and non-economic
effects (also about 25% as influential). The influence of the exponent coefficient for economic
damages lies below all of these and some others - roughly one-eighth as influential in determining
the range of estimates in the social cost of carbon as transient climate sensitivity.

7

-------
The various panels of Figure 5 display the actual correlation estimates. They show, for
example, that increasing transient climate response parameter (TCR) by 1 standard deviation
above its mean in the default case would increase the social cost of carbon by $67 while doing the
same for the economic damages parameter (POW-1) would increase the social cost of carbon by
only $9. Similar disparity is clearly apparent for the other three cases. Put another way, any
change in economic estimates of damages that new literature might produce is easily undone by
small adjustments in other parameters and/or purposeful adjustments in judgmental parameters
(e.g., time preference or risk and inequity aversion).

The numerical results reported here are, to be sure, highly model-specific both with respect
to the sources of uncertainty that are represented explicitly in its structure and the way those
sources are depicted. Other models may suggest that dramatic change in the overall distributions
of economic damages might be more (or less) influential in determining the social cost of carbon,
but we do not think that the qualitative conclusion that they illustrate. We do not think, in other
words, that our hypothesis of minimal value added is right would be weakened substantially if
other models were similarly exercised.

Section 3: Barking up a Different Tree for Value Added.

To us, at least, it follows from the hypothesis that we raises and supported in Section 2 that
economic aggregates should not be the (sole) foundation upon which to build climate policy. They
can, at best, contribute to an understanding of context within which policy alternatives derived
from other sources should be evaluated. That is to say, they can contribute to analyses of whether
or not those alternatives can achieve their stated climate objectives at least cost and, in some cases,
whether or not they might be doing more harm than good. There is, after all, such a thing as
dangerous climate policy; see, for example Tol and Yohe (2007). In addition, the more detailed
modules from which these aggregates are constructed can help decision-makers and researchers
alike identify where careful consideration of an expanded set of adaptation options might be most
productive. Nonetheless, we fear that trying to devise a way to set the price of carbon (or the
economic value of emissions reductions or increases from a non-climate policy, for that matter)
equal to something like the social cost of carbon is probably a fruitless enterprise. Moreover,
justifying impacts analyses completely on the basis of improving the quality of their contributions
to estimates of the social cost of carbon is likely to be a misguided enterprise.

8

-------
So what should we be doing, instead? The authors of the report of the Limiting Panel to
America's Climate Choices [NAS (2010)] offered what we view to be a solid suggestion. They
recommended a multi-step process that would begin with assessing a wide range of climate risks
that will materialize over the medium to long-term. They recognized that these risks will be
calibrated in many monetary and non-monetary metrics and that it will be up to the political
process to determine a socially acceptable level of risk. Given that determination, it should be
possible to identify long-term mitigation targets in terms of temperature increases and associated
ranges of atmospheric concentrations; and from there, it should be possible (a) to deduce a
medium-term global carbon emissions budget that would put the planet on a path from which
iterative decisions based on new climate science and technological development could be designed
and implemented effectively and (b) infer the United States (and other country, for that matter)
contributions to that budget

Each of the steps noted above can, of course, be identified as a research topic, particularly
the iterative component of evolving long-term policy objectives and medium-term carbon budget
targets. Several researchable topics come to mind almost immediately. What should be monitored
to inform iterative decisions, for example? How should "mid-course" corrections be implemented,
and what types of institutions need to be created to make them maximally efficient? And how
frequently should they be undertaken?

More to the point of this workshop, though, how could a medium-term carbon budget target
be achieved? NAS (2010) concluded that it is necessary (but not sufficient by any means) to seta
price on carbon that increases predictably and persistently over the applicable time period. Since
even a medium-term emissions budget can be viewed as an inter-temporal exhaustible resource
problem, the first-order answer to how to price carbon comes straight from Hotelling (1931):
compute the scarcity rent for year one and let it increase over time at the rate of interest The
actual best trajectory will depend, of course, on the rate of growth in economic activity, the rate of
technological innovation in non-carbon intensive energy sources and carbon sequestration, and
other factors that cannot be predicted accurately for 40 year time periods; but these insight
highlights yet another set of researchable questions about quantification and short-term term
iteration processes. Perhaps the most practical approach would involve identifying technologies
that could contribute most to emissions reductions and evaluating the cost of carbon that would be
required to make them economically competitive with fossil-base alternatives at the time they
would become viable. As described in Yohe, et al (2007), the appropriate initial scarcity rent could,

9

-------
quite simply, be the level that would, if it were to climb at the rate of interest, reach the pricing
threshold at just the right time; but this, too, is a researchable issue.

And what role can damage estimates play in all of this? It seems to us, as noted above, that
they provide context in a very important sense. Ranges of aggregates like the social cost of carbon
offer fundamental access to the answers of questions like "What combinations of normative and
scientifically-based parameters produce discounted marginal damage estimates that are consistent
with carbon pricing proposals born of technological modeling and national carbon emissions
budgets? And are those combinations consistent with the normative view of how the world should
behave from which the long-term objectives and medium-term targets were derived?" Their
content, in other words, is not numerical; it is, instead interrogatory.

Section 4: A Concluding Thought.

Answers to the research questions identified in Section 3 that were informed directly by our
brief comments in Sections 1 and 2 would not be unique, of course, and that complication must be
acknowledged from the start. So, too, should the pervasive uncertainties that will not, in many
cases, be resolved in a timely fashion. We close, therefore, with a reference to a lesson articulated
almost two decades ago by Lester Lave - an economist of considerable note and wide experience in
climate-related issues who worked for decades at Carnegie Mellon University in Pittsburgh. He
once told the then fledgling Center for the Study of the Human Dimensions of Global Environmental
Change that "If it does not make a difference of a factor of two, then it is inside the noise. With that
fact of life we will simply have to learn to cope." Correcting for misrepresenting trends inside that
noise is, quite fundamentally, why iteration is so essential in all of this - it is the first order question
that must be confronted directly if we are to have any success in Improving the Assessment and
Valuation of Climate Change Impacts for Policy and Regulatory Analysis.

10

-------
References

Downing, T. & Watkiss, P., "The marginal social costs of carbon in policy making: Applications,
uncertainty and a possible risk based approach", 'DEFRA International Seminar on the Social Costs
of Carbon', 2003.

Hope C., "The marginal impact of C02 from PAGE2002: An integrated assessment model
incorporating the IPCC's five reasons for concern", Integrated Assessment 6: 19-56. It can be found
at http://iournals.sfu.ca/int assess/index.php/iai/article/view/227. 2006.

Hotelling H., "The economics of exhaustible resources", Journal of Political Economy 39: 137-175,
1931.

Kirshen, P., Watson, C., Douglas, E., Gontz, A., Lee, J., and Tian, Y., "Coastal Flooding in the
Northeastern United States due to Climate Change", Mitigation and Adaptation Strategies for Global
Change 13: 437-451, DOI: 10.1007/sll027-007-9130-5, 2008.

Kunreuther, H., and Useem, M., Learning from Catastrophes: Strategies for Reaction and Response,
Wharton School Publishing Philadelphia, 2010.

Mendelsohn, R., Emanuel, K, and Chonabayashi, S., "The impact of climate change on tropical
cyclone damages", Workshop on Climate Change Impacts and Associated Economic Damages,
January 27, 2011, Washington, DC, 2011.

National Academies of Science (NAS), Limiting the Magnitude of Future Climate Change, Report of
the Panel on Limiting the Magnitude of Future of Future Climate Change, America's Climate Choices,
National Research Council, National Academies Press, Washington, DC. http: //www.nas.edu. 2010.

Ogura, S. and Yohe, G., "The complementarity of public and private capital and the optimal rate of
return to government investment," Quarterly Journal of Economics, 91: 651-662,1977.

Rosenzweig, C., Solecki, W., Gornitz, V., Horton, R, Major, D., Yohe, G., Zimmerman, R., 2011,
"Developing Coastal Adaptation to Climate Change in the New York City Infrastructure-shed:
Process, Approach, Tools, and Strategies", Climatic Change, forthcoming.

Tol, R. and Yohe, G., "On Dangerous Climate Change and Dangerous Emission Reduction" (with
Richard Tol) in Avoiding Dangerous Climate Change (Schellnhuber, H.J., Cramer, W. Nakicenovic, N.
Wigley, T., and Yohe, G. eds.), Cambridge University Press, 2006.

Yohe, G., Tol, R. and Murphy, D., "On Setting Near-term Climate Policy while the Dust Begins to
Settle: The Legacy of the Stern ReviewEnergy and Environment 18: 621-633, 2007.

Yohe, G., Knee, K. and Kirshen, P., "On the Economics of Coastal Adaptation Solutions in an
Uncertain World", Climatic Change, DOI: 10.1007/sl0584-010-9997- , 2010

11

-------
Table 1. Estimated Storm Surge Elevations and Return Times of the Current 100-year Storm
Anomalies for Boston and New York. Estimates based on median sea level rise scenarios for the
B1 and A1FI SRES scenarios with historical pace of local sea level rise indicated in parentheses.
Source: Kirshen, etal. (2008).

Storm Surge Elevation
Location 2005	2050

Boston (2.65 mm/yr; 1921-2005)

B1	2.9 m	3.0 m

A1FI 2.9 m	3.2 m

New York (2.77 mm/yr; 1920-2005)
B1	2.8 m	2.9 m

A1FI 2.8 m	3.1m

Recurrence of 2005
100 yr Storm

2100	2050	2100

3.1m	15 yr	5 yr

3.8 m	3 yr	«2 yr

3.0 m	50 yr	30 yr

3.7 m	30 yr	3 yr

12

-------
Table 2. Estimated Internal Rates of Return for Investment in Protective Infrastructure over
Time: Estimates of the expected internal rates of return for investing in a $390 million (real terms)
protective infrastructure against the increasing economic risk driven by climate change and
portrayed in Figure 1 for an urban area in Boston. Source: Yohe, et al. (2010)

Year

1 meter SLR(2100)

0.6 meter SLR(2100)

2010

2.1%

-0.5%

2015

3.8%

0.2%

2020

4.3%

0.4%

2025

5.2%

0.8%

2030

6.4%

1.3%

2035

8.4%

1.8%

2040

12.4%

2.5%

2045



3.4%

2050



5.0%

13

-------
Table 3: Summary Results for the Social Cost of Carbon (per ton of C02): Summary results
from 100,000 runs for the default settings are compared with cases in which (Case A) the range of
economic damages in general and attributed to sea level rise shrinks by 50%, (Case B) the ranges of
both stay the same but the modes shrink by 50%, and (Case C) the ranges of both stay the same but
the modes increase by 50%. Schematics of the critical distributions are provided for each.

Mean of Contribution of

Case	Min 5th Mean 95th 99th Max Lower 99% Top 1% to Mean

Default	-$4 $12 $106 $259 $1191 $12215 $85	20%

Symmetric default settings for the economic damage and sea level rise calibrations

Case A	-$1 $12 $106 $258 $1168 $10084 $85	20%

A Ranges for the two economic damage parameters diminished by 50%

Case B	-$2 $10 $102 $248 $1108 $9131 $80	22%

Ranges preserved but distribution skewed with the mode 50% lower

CaseC	-$3 $13 $111 $272 $1218 $13166 $89	20%

Ranges preserved but distributions skewed with the mode 50% higher

14

-------
Figure 1: Observed Frequencies of "Over-threshold" Events in Select Locations along the
Northeastern Coastline of the United States since 1920: The number of "points-over-threshold
[POT] anomalies per year for each site; a strongly increasing trend in the number of POT anomalies
was detected at all sites. Source: Kirshen, et al. (2008).

60

X Atlantic City

o Boston	0

A New London	^

50

z

10

. ~ IVJYC

i. o Woods Hole	° a

» 40	O	D	O	a

CX„/»	Xaf1aa



V/	U W 7\ yVl, S

X wXX Jx >c

X O 	~

1915 1925 1935 1945 1955 1965 1975 1985 1995 2005

Year

15

-------
Figure 2: Damage Profiles from Coastal Storms over Time for Two Sea Level Rise
Trajectories: Distributions of economic damage across 100 runs for two sea level rise scenarios.
Panels A and B indicate economic damages from coastal flooding in selected years in the future for
an urban area in Boston along 1.0 and 0.6 m sea level rise scenarios, respectively. These estimates
do not include adaptation. Source: Yohe, etal (2010)

Distributions of Damages over 100 Runs - Urban 1 meter

Annual Damages ($miIIion)

—•—2010
—¦—2030
—*—2050
—«—2070
—¦—2090

Panel A

Panel B

16

-------
Figure 3: Global Mean Temperature (relative to pre-industrial levels): The thick middle line
represents the mean for an Alb-style story-line with default settings, 75th and 95th percentiles runs
for the 100,000 permutations run above the mean; 25th and 5th percentile trajectories run below.

Global mean temperature rise

DegC 10.00
9.00
8.00
7.00
6.00
5.00
4.00
3.00
2.00
1.00
0.00

2000

Year

2050	2100	2150	2200

17

-------
Figure 4: Histograms of the Social Cost of Carbon. Distributions of estimates of the social cost of
carbon from 100,000 randomly selected futures (excluding the upper 1% of the estimates so that
the shapes become clear). Panel A depicts the default baseline. Panel B depicts Case A - reduction
in the range of the parameters that calibrates estimates for economic sectors and coastal zones by
50% without changing their means or the modes. Panel C depicts Case B - 50% reductions in the
modes of those parameters without changing their ranges. Panel D depicts Case C - 50%
exaggeration of the modes of those parameters without changing their ranges.

SCC02

0.012

0.010

0.008

0.006

0.004

0.002

0.000

5.0%

SCC02

Minimum	-1.39

Maximum	1,168.49

Mean	84.70

Std Dev	110.97

Values	99000 / 100000

Filtered	1000

Panel A

SCC02

0.012-

0.010

0.008 -

0.006 ¦

0.004 ¦

0.002 ¦

0.000

5.0%

SCC02

Minimum	-1.39

Maximum	1,168.49

Mean	84.70

Std Dev	110.97

Values	99000 / 100000

Filtered	1000



Panel B
18

-------
SCC02

0.012-

0.010 -

0.008

0.006 -

0.004 -

0.002 ¦

0.000

5.0%

U scc°2

Minimum
Maximum
Mean
Std Dev

Values 99000
Filtered

-2.27
1,207.97
80.37
109.61
/ 100000
1000

UO	oo

Panel C

SCC02

0.012 n

0.010 ¦

0.008 ¦

0.006 ¦

0.004 ¦

0.002

0.000

SCC02

-2.83
1,217.83
88.89
114.33

Values 99000 / 100000
Filtered	1000

vD	00

Panel D

19

-------
Figure 5: Marginal Contributions of Various Parameters to Variability in Estimates of the
Social Cost of Carbon. The bars indicate the direction and strength of various parameters in
sustaining variability in estimates of the social cost of carbon; cases are as defined in Figure 4.2 The
value of 67 assigned to transient climate response (TCR) indicates, for example, that increasing TCR
by 1 standard deviation above its mean would increase the social cost of carbon by $67. Increasing
the economic damages parameter (POW-1) by 1 standard deviation would, by way of contrast,
increase the social cost of carbon by only $9.

SCC02

Regression - Mapped Values

SCC02

Panel A

SCC02

Regression - Mapped Values

o	o	o	o	o	o	o

ST	fM	fM	"3"	VO	CO

SCC02

Panel B

2

Partial Glossary: TCR - transient climate response; PTP - pure time preference rate; EMUC - (negative of
the) elasticity of the marginal utility of consumption; FRT - feedback response time; IND - indirect effect of
sulfates; POW-2 - exponent of the non-economic impact function; W_2 - non-economic impact at calibration
temperature; TCAL - calibration temperature; POW-1 - exponent of the economic impact function.

20

-------
SCC02

Regression - Mapped Values

SCC02

Panel C

SCC02

Regression - Mapped Values

TCR
PTP
EMUC
FRT
RAND_DIS
IND
POW_2
W_2
TCAL
POW_l
IA weights factor
WDIS
DISTAL)
RLO
CCF
AIR_C02

SCC02

Panel D

21

-------
Some Thoughts on the Value Added from a New Round of
Climate Change Damage Estimates

Gary Yoheac and Chris Hopeb

EPA/DOE Workshop on Improving the Assessment and Valuation of Climate

Change Impacts for Policy and Regulatory Analysis:

Research on Climate Change Impacts and Associated Economic Damages

January 28, 2011

-------
Outline of Brief Remarks

•	More complete paper available.

•	Section	1 - Issues with Coas

•	Section 2 - Type 3 Error - Barking up the wrong tree
means very little value added.

•	Section 3 - There is an alternative - the Limiting
Panel plus iteration - here is value added for an
aggressive research agenda.

•	Economic analyses of impacts help ID places where
adaptation would be important; "laugh test context
for the alternative.

-------
Experiment Results - SCC

Mean of Contribution of

Case	Min 5th Mean 95th 99th Max Lower 99% Top 1% to Mean

A

Default	-$4 $12 $106 $259 $1191 $12215 $85	20%

Symmetric default settings for the economic damage and sea level rise calibrations

Case A	-$1 $12 $106 $258 $1168 $10084 $85	20%

Ranges for the two economic damage parameters diminished by 50%

Case B	-$2 $10 $102 $248 $1108 $9131 $80	22%

Ranges preserved but distribution skewed with the mode 50% lower

Case C

-$3 $13 $111 $272 $1218 $13166 $89
Ranges preserved but distributions skewed with the mode 50% higher

20%

-------
Experiment Results -

SCC

SCC02

Panel A - Default

Panel B - Reduced Range

5CC02

I 5CCD2

Mmmum	-2.27

Maiimum	1,207.97

Mean	80.37

9d Dev	109.61

Values	WOO I 100000

filtered	1000

Sid Dev	114.33

V*km 99000 1 100000
fifcered	1000

Panel C - Mode 50% Lower

Panel D - Mode 50% Higher

-------
An Alternative Approach - A Different Tree for
Barking with higher Value Added

•	Use assessment of climate risk to determine long-term
objective and medium-term carbon budget - build the
iterative process

•	Work within the process to determine US contribution to the
budget

•	Compute scarcity rent trajectory for the budget (a la Hotelling)
and then add details of economic growth, technological
development, etc... build the iterative process.

•	Use the results to price carbon for non-climate policy needs

•	Use 1AM results to (1) check the "laugh test", (2) design cost-
minimizing approaches (including net economic damage) and
(3) highlight areas where adaptation in economic sectors will
be most productive.

-------
-------