Research on Climate Change Impacts in the United States James McFarland, Jeremy Martinich, Marcus Sarofim Climate Change Division, US EPA Stephanie Waldhoff, PNNL Latin American Modeling Project San José, Costa Rica 4 October 2012

Overview •  Goals and motivation •  Methodology •  Illustrative results •  Next steps

U.S. Environmental Protection Agency

2

Goals and Motivation for Impacts Research •  Our goal is to quantify where possible and communicate the benefits (i.e., avoided or reduced impacts) of mitigation & adaptation actions. •  The research explores how impacts and damages may change under a consistent set of scenarios, data, and assumptions. –  Existing impacts literature is largely based on inconsistent assumptions along the causal chain from socio-economics to emissions to climatic effects and impacts.

•  Initial focus is on: –  Risks and impacts within the U.S., without ignoring key global linkages or key regional components. •  Impacts and benefits across a range of sectors, e.g., water resources, human health, ecosystems, energy.

–  Potential benefits of mitigation scenarios (adaptation later). –  Analyzing key sources of uncertainty, including emissions pathway, climate sensitivity, climate models, etc.

3

Our ideal tools and results

Socio-economic, policy drivers Emissions

• 

Atmospheric concentrations

Integrated model(s) with internally consistent emissions drivers, impact sectors, and economic valuation –  Climate impacts feed back into the economy and climate

Radiative forcing

• 

Identify, quantify, and be transparent about key uncertainties along the causal chain

Historical and projected temperature, precipitation, sea level rise, etc.

• 

Multiple future scenarios, BAU and policies

• 

Outputs that communicate effectively to multiple audiences about how impacts and risks change from one scenario to another.

Potential risks and impacts, economic damages GHG mitigation and adaptation measures

U.S. Environmental Protection Agency

4

Methodology

U.S. Environmental Protection Agency

5

Analytical Goals •  Develop estimates of climate change impacts and damages in multiple sectors that can be synthesized –  Begin with integrated assessment (IA) models to develop three internally consistent socio-economic, emissions, and climate scenarios (BAU, RF 4.5, RF 3.7) –  All sectoral models use consistent population, GDP, and emissions data –  Climate inputs consistent with all socio-economic and emissions scenarios

•  Explore uncertainties around impacts estimates –  Scientific: Multiple climate sensitivities (2.0, 3.0, 4.5, and 6.0) –  Model: Use of multiple IA and sectoral models where possible –  Variability: Analysis of changing temperature and precipitation patterns

•  Understand what drives differences in model results –  Comparison of data inputs and outputs –  Discussions about model structures, methods, etc.,. U.S. Environmental Protection Agency

6

Methodology •  Begin with IA models (IGSM and GCAM) to develop three internally consistent socio-economic, emissions, and climate scenarios –  Reference: Business as usual •  GDP and population harmonized with US (EIA) data through 2035, EPPA projections through 2100

–  Policy scenarios: •  4.5 W/m2 and 3.7 W/m2, stabilization in 2100

•  Multiple climate sensitivities (2.0, 3.0, 4.5, and 6.0) •  Climate data from MIT’s 3D (CAM) component of IGSM •  Sectoral models develop estimates with these consistent socioeconomic and climate data U.S. Environmental Protection Agency

7

Impacts Research Operational Schematic

Yield Changes (Crops, forests)

Yield Changes (Crops, forests)

8

Data Flow • 

Inputs –  Reference GDP and population (EIA through 2035)

• 

IA Model Outputs –  Global GHG concentrations –  Global and domestic emissions • 

CO2, non-CO2 GHG, criteria pollutants

–  Sea Level Rise

–  Policy scenario, RF targets

–  Temperature change •  • 

Global annual average Gridded monthly, daily, hourly

–  Precipitation • 

Gridded monthly, daily, hourly

•  Changes in impact sectors (use IA outputs as inputs) •  •  •  •  •  •  •  •  • 

Temperature-related mortality SLR property damages and adaptation response costs Road and bridge infrastructure adaptation Inland flooding damages Water supply and demand Drought risk (not monetized) Electricity supply Energy demand Population

•  •  •  •  •  • 

Crop yields projections Vegetative carbon sequestration and provisioning of grazing lands Forest fire frequency/magnitude and suppression costs Coral reef cover and recreational/ existence values Freshwater fish habitat and recreational fishing impacts Air quality

U.S. Environmental Protection Agency

9

Examples of Data Needs for Sectoral Modeling Inputs Sector/Model COMBO   SLR/Coastal  Property  Model

Socio-­‐economic

Ecoservices/forest  fires

ΔLU  (Developed  Land)

Inland  flooding

Population  growth Pop  growth,  demographic   Max  and  min  daily  T changes,  V SL  =  f(GDP/cap) Daily  precip  to  calculate  2  y   and  100  y  24  hour  max   Land  cover  type precip Monthly  avg  temp  and   precip Value  of  fishing  day,   Monthly  avg  max  T  and  avg   Population  growth precip Monthly  avg,  max,  and  min   T,  total  monthly  precip,   Population  growth cloud  cover,  wind,  relative   humidity Yield  changes  due  to   Demand  (population,  GDP) climate  changes  (EPIC) ΔT  (daily/hourly) Population,  GDP

Heat  health Bridge  vulnerability Drought  risk Freshwater  fisheries

Water  supply-­‐demand

FASOM IPM

GDP  growth

Climate Other Global  avg  ΔT Global  avg  SLR Monthly  avg  T,  daily  max  T,   CO2  Concentrations,   monthly  mean  precip elevation Monthly  ΔPrecip

10

Illustrative Results

U.S. Environmental Protection Agency

11

CO2 Emissions U.S.  Annual  CO2 Emissions

MIT   Reference GCAM   Reference

8,000

MIT   Policy4.5

6,000

2095

2085

2075

2065

2055

GCAM   Policy3.7

2045

0 2035

MIT   Policy3.7 2025

2,000 2015

4,000

GCAM   Policy4.5

Global  Annual  CO2 Emissions

MIT   Reference

100,000 90,000

GCAM   Reference

80,000

MIT   Policy4.5

60,000 50,000

GCAM   Policy4.5

40,000 30,000

GCAM  GlobalFossilfueland  LU  CO2  Emissions  (Mt-­‐CO2)

90,000

80,000 70,000

60,000 50,000 40,000

30,000 20,000

10,000 0 2000

2020

2040

2060

2080

MIT   Policy3.7

20,000 10,000

2095

2085

2075

2065

2055

2045

2035

11/2/12

2025

2015

0 2005

Mt-­‐CO2/year

70,000

100,000 Annual  CO 2 Emiaaions  ( Mt-­‐CO2)

10,000

2005

Mt-­‐CO2/year

12,000

GCAM   Policy3.7

12

2100

CO2 Concentrations 3.7 Policy

Reference

ppm CO2

750 650

490 MIT-CS2

470

MIT-CS6

450

PNNL-CS2 ppm CO2

850

PNNL-CS6

550

430 MIT-CS2 410

MIT-CS6

390

PNNL-CS2

450 370 350 2000

2020

2040

2060

2080

2100

350 2000

PNNL-CS6 2020

U.S. Environmental Protection Agency

2040

2060

2080

2100

13

Forcing 3.7 Policy

Reference 4.5 Ref-MIT-2

10

Ref-MIT-3

8

Ref-MIT-6

4 W/m2 Since Preindustrial

W/m2 since Preindustrial

12

Ref-MIT-4.5 Ref-PNNL-2

6

Ref-PNNL-6

4 2 0 2000

3.5 3

3.7MIT-2

2.5

3.7MIT-3

2

3.7MIT-4. 5 3.7MIT-6

1.5 1

3.7PNNL

0.5 2020

2040

2060

2080

2100

0 2000

2020

U.S. Environmental Protection Agency

2040

2060

2080

2100

14

Temperature Reference

3.7 Policy 3.5

9

Degrees C Since 1990

7 6 5

PNNL-3.7CS2

PNNLRef,CS2 PNNLRef,CS6 MITRef,CS2 MITRef,CS6

3 Degrees C Since 1990

8

4 3

2.5

PNNL-3.7CS6 MIT-CS2 MIT-CS6

2 1.5 1

2 0.5

1 0 1990

2010

2030

2050

2070

2090

0 1990

2010

2030

U.S. Environmental Protection Agency

2050

2070

2090

15

Presentation of Results

Probability  

(Global Average ΔT from 1990, GCAM) 1.0

60%

0.8

50%

0.6

40% 30%

0.4

20%

0.2

10%

0.0

0

2

4

6

8

10

0%

Observed ΔT  in  2100   (above  1990) Reference

Policy  4 .5

Reference

0

Policy  3 .7

2

4

6

Observed ΔT  in  2100   (above  1990) Reference

Policy  4.5

8

10

Policy  3.7

Policy  3.7

Policy  4.5 0-­‐2

0-­‐2

0-­‐2

2-­‐3

2-­‐3

2-­‐3

3-­‐4

3-­‐4

3-­‐4

4-­‐5

4-­‐5

4-­‐5

5-­‐6

5-­‐6

5-­‐6

6-­‐8

6-­‐8

6-­‐8

>8

>8

>8

U.S. Environmental Protection Agency

16

Presentation of Results

Probability  

(Global Average ΔT from 1990, IGSM) 1.0

60%

0.8

50%

40%

0.6

30%

0.4

20%

0.2

10%

0.0

0

2

4

6

8

10

0% 0

Observed ΔT  in  2100   (above  1990) Reference

Policy  4 .5

Reference

2

4

Reference

Policy  3 .7

6

Observed ΔT  in  2100   (above  1990) Policy  4.5

8

10

Policy  3.7

Policy  3.7

Policy  4.5 0-­‐2

0-­‐2

0-­‐2

2-­‐3

2-­‐3

2-­‐3

3-­‐4

3-­‐4

3-­‐4

4-­‐6

4-­‐6

4-­‐6

6-­‐8

6-­‐8

6-­‐8

>8

>8

>8

U.S. Environmental Protection Agency

17

Sea Level Rise (meters) 0.8 0.7 0.6

Ref,CS2 Ref,CS3

0.5

Ref,CS4.5 0.4

Ref,CS6 MIT-2

0.3

MIT-3 0.2

MIT-4.5 MIT-6

0.1 0 1990

11/2/12

2010

2030

2050

2070

2090

U.S. Environmental Protection Agency

18

th percentile ChangeChanges in # of days above present day 95 in Temperature Extremes

Without mitigating GHGs, today’s hottest days become more frequent, and the number of frosts will decrease. Daily Max Temperature

Frost Frequency

Less Stringent Policy (RF4.5)

More Stringent Policy (RF3.7)

# of future frost days per year

Change in # of days above present day 95th percentile

Business As Usual

19

Changes in Extreme Precipitation Without mitigating GHGs, extreme precipitation will become more common. Business As Usual

Summer

Less Stringent Policy (RF4.5)

More Stringent Policy (RF3.7)

Change in # of days above present day 95th percentile

Change in # of days above present day 95th percentile

Winter

20

BAU Less stringent (RF 4.5) More stringent (RF 3.7)

Change in number of drought months within a 30yr window (difference between 1980-2009 and 2085-2115) Change in number of drought months within a 30yr window (diference between 1980-2009 and 2085-2115)

Changes in Drought Risk Through 2100

21

Estimated Decline in U.S. Coral Reefs Loss of Hawaiian Coral Cover Sum of Lost Annual Rec. Benefits in Hawaii S. Florida

•  GHG mitigation delays Hawaiian coral reef loss compared to BAU. –  The more stringent policy scenario (RF3.7) avoids ~$9B in lost recreational value for Hawaiian reefs, compared to the BAU.

•  GHG mitigation provides only minor benefit to coral cover in South Florida and Puerto Rico (not shown), as these reefs are already being affected by climate change, acidification, and other stressors.

22

Energy: National HDD & CDD % Change in Heating Degree Days Relative to 2005 2012

2020

2030

2040

2050

0.00% -10.00% -20.00% -30.00% -40.00% -50.00% -60.00%

Business as usual:

-70.00%

Policy (RF3.7):

70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% 0.00% 2012

2020

2030

% Change in Cooling Degree Days Relative to 2005

2040

2050

Next Steps

U.S. Environmental Protection Agency

24

Peer Review of Impacts Research •  Peer review of methods and results –  Special issue •  Overview of methodology and goals •  Individual papers on each component of the project

–  Individual papers for each topic/impacts sector •  Overall approach and scenario development •  Extreme events and assessing uncertainty of regional climate change •  Coastal development, infrastructure, and heat health •  Energy supply/demand and water resources (drought, flooding damages, water supply/ demand) •  Ag/forestry and ecosystems (coral reefs, freshwater fish, vegetation/wildfire)

–  Key methods and results assembled in a single paper •  Will require a significant amount of supplementary material.

U.S. Environmental Protection Agency

25

Communication of Results •  Estimating impacts and economic damages in an analytically rigorous and consistent way will enable clear communication of climate change impacts and risks to a variety of audiences –  Researchers •  Distribute findings through peer reviewed publication and conference presentations

–  Policy makers •  Schedule briefings with interested committees •  Incorporate results into legislative analyses

–  Public •  Share results through EPA's updated climate change website •  Summary report

U.S. Environmental Protection Agency

26

Other Potential/Future Impact Analyses •  Population –  Leverage EPA-ORD’s ICLUS model to examine climate change impacts on regional population growth –  Disaggregated data may be used in future iterations as inputs to other sectoral models (e.g. land use, energy)

•  Energy Supply –  NREL’s ReEDS model to look at climate change impacts on energy transmission, including extreme events

•  State-level impacts –  Penn State, Boston University developing a state-level impacts model using sectoral damage functions to examine impacts with interstate trade U.S. Environmental Protection Agency

27

Appendix: Supplementary Materials

11/2/12

U.S. Environmental Protection Agency

28

Coordination with Integrated Assessment (IA) Models •  EPA/OAP supports number of global climate/economic modeling groups:

The MIT IGSM Model

–  MIT Joint Program’s IGSM framework (IA) –  PNNL-JGCRI’s GCAM (IA)

•  Coordinate with these groups to: –  Harmonize key inputs (GDP, pop. growth, radiative forcing targets) –  Obtain climate projections for use in sectoral models –  Utilize sectoral components of these broader frameworks for impact and damage analyses

U.S. Environmental Protection Agency

29

Scenario Design •  Reference scenario: As discussed previously •  Policy: As discussed, target forcing in 2100 as a change from preindustrial –  MIT: Manual target –  PNNL: Automated

•  Climate Parameters: –  MIT: KV = 0.5 cm2/s, Aerosol in 80s = -0.25 to -0.95 W/m2, depending on CS parameter –  PNNL: Kv = 2.3 cm2/s, Aerosol in 1990= about -1.3 W/m2 11/2/12

U.S. Environmental Protection Agency

30

MIT: Climate Sensitivity and Aerosol Forcing

11/2/12

U.S. Environmental Protection Agency

31

Existing Data Sets for Integrated Analysis • 

IPCC Special Report on Emissions Scenarios (SRES, 2000) –  Insufficient regional disaggregation (only four world regions) –  Scenarios do not reflect explicit climate policies, but rather development paths with unclear costs –  Critiqued for unrealistic narrowing of incomes across regions

• 

IPCC Relative Concentration Pathways (RCPs) –  Break the link between socio-economics, emissions and atmospheric concentrations. –  This is a design feature to allow for socio-economic and climatic research to proceed in parallel.

32

Data Flow • 

Inputs –  Reference GDP and population (EIA through 2035)

• 

IA Model Outputs –  Global GHG concentrations –  Global and domestic emissions • 

CO2, non-CO2 GHG, criteria pollutants

–  Sea Level Rise

–  Policy scenario, RF targets

–  Temperature change •  • 

Global annual average Gridded monthly, daily, hourly

–  Precipitation • 

Gridded monthly, daily, hourly

•  Changes in impact sectors (use IA outputs as inputs) •  •  •  •  •  •  •  •  • 

Temperature-related mortality SLR property damages and adaptation response costs Road and bridge infrastructure adaptation Inland flooding damages Water supply and demand Drought risk (not monetized) Electricity supply Energy demand Population

•  •  •  •  •  • 

Crop yields projections Vegetative carbon sequestration and provisioning of grazing lands Forest fire frequency/magnitude and suppression costs Coral reef cover and recreational/ existence values Freshwater fish habitat and recreational fishing impacts Air quality

U.S. Environmental Protection Agency

33

What is being measured? • 

GDP estimates that include the cost of the policy (i.e. GDPPolicy < GDPReference)  GDP/cap is lower under policy scenarios

• 

Economic damages are calculated using two components: –  – 

Impacts that measure physical units (e.g. deaths) A measure of the economic value of those impacts

• 

Damages measure the economic value of those impacts

• 

When the economic value is correlated with GDP/cap (e.g. VSL or property values), the damages under a policy scenario will be lower for two reasons: –  Impacts are smaller –  Economic value is smaller

• 

Therefore, the benefits of the policy are larger than if GDP/cap was constant

• 

Is this a problem? – 

11/2/12

Only because the climate change impacts are not themselves included in the GDP measures—this work is intended to enable inclusion of these damages

U.S. Environmental Protection Agency

34

Example of limited benefits analysis to date: Probability of Observed Temperature Changes in 2100 (S. APA) Reference

8.1%

0-­‐2

10.1%

4.4%

° Fahrenheit

11.1%

0-­‐3.6

3.6-­‐5.4

20.9%

29.3%

2-­‐3

Full  Participation

4.0% 0.6%

4.7% 1.18%

° Celsius

3-­‐4

Developing  Country  Delay  

5.4-­‐7.2

18.9%

4-­‐5

7.2-­‐9.0

27.0%

5-­‐6

47.2%

6-­‐8

74.8%

9.0-­‐10.8 10.8-­‐14.4

37.8%

•  The pie charts show the approximate probability of observed global mean temperature changes in 2100, relative to pre-industrial, falling within specific temperature ranges under reference, developing country action delayed until 2050, and G8 international action scenarios. –  The figures were developed using MAGICC 5.3 and the truncated (at 10° C) Roe and Baker (2007) distribution over climate sensitivity. –  Observed temperature change is that resulting from the concentration levels in a specific year. –  See appendix 5 for equilibrium temperature results.

•  Observed temperature change does not equal the change in equilibrium temperature because –  CO2e concentrations rise after 2100: Equilibrium temperature change is not achieved until after CO2e concentrations are stabilized. In this analysis, CO2e concentrations will continue to rise after 2100. Therefore, changes in equilibrium temperature will differ from the observed temperature changes. –  Ocean temperature inertia: This inertia causes the equilibrium global mean surface temperature change to lag behind the observed global mean surface temperature change by as much as 500 years. Even if CO2e concentrations in 2100 were stabilized, observed temperatures would continue to rise for centuries before the equilibrium was reached.

•  Under the Reference scenario (1st chart), the probability of the observed temperature change in 2100 being below 2 degrees C is approximately 1%, while there is a nearly 75% probability associated with this under the Full Participation scenario (3rd chart). •  The probability of being above 4 degrees C is about 32% in the Reference case, while it is just under 15% in the Delayed Participation scenario (2nd chart) and zero under Full Participation (3rd chart).

35

GDP per capita U.S.  G DP per  capita $150,000 GCAM  A ll MIT   Reference MIT   Policy4.5 MIT   Policy3.7

$110,000 $90,000 $70,000 $50,000

$30,000

$40,000 $35,000

Gross  World  Product per  capita GCAM  A ll

$30,000

GDP/cap (2005$)

GDP/cap (2005$)

$130,000

$25,000 $20,000 $15,000 $10,000

MIT   Reference MIT   Policy4.5 MIT   Policy3.7

$5,000

$0

11/2/12

36

Example: Heat Deaths 2050 GCAM 80,258

GDP/cap-­‐Ref GDP/cap-­‐Policy VSL-­‐Ref 10.1 VSL-­‐Policy Value-­‐Ref  deaths          123,401 Value-­‐Pol  deaths              89,761 Policy  Benefit              33,641

IGSM 86,209 77,764 10.1 9.7          123,161              85,967              37,195

2100 GCAM IGSM 152,445 159,252 137,943 13.0 12.9 12.2              289,262    285,501              129,895    121,047              159,367    164,454

Using  Policy  V SL  times  change  i n  #  deaths:              32,219 11/2/12

U.S. Environmental Protection Agency

   148,512 37