Assessing the Impacts of Climate Change on U.S. Agriculture by Combining Agroecosystem and Economic...

Assessing the Impacts of Climate Change on U.S. Agriculture by Combining

Agroecosystem and Economic Models

Robert Beach, RTI International; Allison Thomson, JGCRI; Bruce McCarl, Texas A&M University

Presented atNorth American Carbon Project 4th All-Investigators Meeting

Albuquerque, NM, February 4-7, 2013

RTI International is a trade name of Research Triangle Institute

3040 Cornwallis Road ■ P.O. Box 12194 ■ Research Triangle Park, NC 27709Phone 919-485-5579 e-mail rbeach@rti.orgFax 919-541-6683

2

Introduction

Agricultural production is inherently risky, dependent upon weather and other factors

Climate change may lead to more rapid changes in mean yields as well as yield and price variability than in the past, making adjustment more difficult Extreme events Historical experience less informative Faster depreciation of research/knowledge stock

Complex set of interactions between impacts with heterogeneity across time and space Need for detailed characterization of potential effects on

yields for future scenarios as inputs for decision support tools

3

IPCC Scenario and GCMs Used

IPCC Scenario A1B was the scenario selected for all global circulation models Actual emissions have been above the A1B scenario so we felt it

was reasonable to focus on that scenario vs. B1 and B2 scenarios with lower emissions given that there has also been interest in non-fossil fuel energy (vs. A1FI or A2)

GCMs (daily simulation data available for 2045-2054 and 1991-2000) GFDL-CM2.0 and GFDL-CM2.1 models developed by the

Geophysical Fluid Dynamics Laboratory (GFDL), USA Both GFDL-CM2.0 and GFDL-CM2.1 are being included

because they have substantially different seasonal precipitation patterns

Coupled Global Climate Model (CGCM) 3.1 developed by the Canadian Centre for Climate Modelling and Analysis, Canada

Meteorological Research Institute (MRI) Coupled atmosphere-ocean General Circulation Model (CGCM) 2.2 developed by the Meteorological Research Institute, Japan Meteorological Agency, Japan

4

Comparison of Average Changes in U.S. Temperature and Precipitation Across GCMs

Model Season Change Max

Temp (°C) Change Min Temp (°C)

Change in Precipitation (%)

GFDL-CM2.0 MAM 2.78 2.41 -7.4

GFDL-CM2.0 JJA 4.34 3.44 -8.5

GFDL-CM2.1 MAM 1.66 1.72 0.6

GFDL-CM2.1 JJA 4.03 3.45 -16.5

CGCM3.1 MAM 2.45 2.41 2.1

CGCM3.1 JJA 2.27 2.17 0.7

MRI-CGCM2.2 MAM 1.23 1.37 9.5

MRI-CGCM2.2 JJA 1.28 1.57 8.7

5

EPIC Model

Process-level agro-ecosystem model originally developed at USDA, currently used by multiple research groups

Has been used previously to simulate regional productivity of corn, soybeans, wheat (winter and spring), cotton, hay, and switchgrass for the U.S. at the 8-digit hydrologic unit scale

Multiple soil types are represented within each of the 1,450 hydrologic units, resulting in 7,540 total runs

Modeling system modified to use 1991-2000 baseline climatology simulations and future climate change projections for daily simulations from IPCC SRES scenario A1B

Projections for this study incorporate GCM results from four different models for the time period 2045-2054

Additional cropping systems were added: sorghum, rice, barley, and potatoes for the appropriate regions

6

Change in Average Spring and Summer Temperature and Precipitation: GFDL-CM2.0, 2045-2054

Summer

Spring

7

Change in Average Spring and Summer Temperature and Precipitation: GFDL-CM2.1, 2045-2054

Summer

Spring

8

Change in Average Spring and Summer Temperature and Precipitation: CGCM3.1, 2045-2054

Summer

Spring

9

Change in Average Spring and Summer Temperature and Precipitation: MRI-CGCM2.2, 2045-2054

Summer

Spring

10

Examining Potential Shifts in Production Regions

Changes in yield distributions may alter production regions

Areas where crops were modeled in EPIC were expanded outside recent historical rangeFocused on suitable cropland areas in proximity to

historical rangeEPIC simulations of mean and variance of yield

Equilibrium production is being simulated based on stochastic version of FASOM

11

Current and Expanded Crop Production Ranges Modeled in EPIC

Barley Corn Cotton Potatoes

Rice WheatSoybeansSorghum

12

Percentage Change in Dryland and Irrigated Corn Yields Simulated Using EPIC, 2045-2054

Irrigated Yields

Dryland Yields

13

Percentage Change in Dryland and Irrigated Soybean Yields Simulated Using EPIC, 2045-2054

Irrigated Yields

Dryland Yields

14

Percentage Change in Dryland and Irrigated Wheat Yields Simulated Using EPIC, 2045-2054

Irrigated Yields

Dryland Yields

15

Sample Comparison of Simulated Corn Yield Distributions Across Climate Scenarios

16

FASOM Model Structure

Objective: Welfare Maximization Land is allocated between activities (and combined with other inputs) based on

relative rents (including GHG payments) and suitability to maximize intertemporal welfare

Sectoral and Land Coverage Forest – approximately 80 products from private timberland Agriculture – crops and pasture

Over 70 primary and about 60 processed commodities, 20 processed feeds Developed – Tracks conversion of forest, crop, and pastureland for development

3 GHGs — CO2, N2O, CH4

Stocks and flows of GHGs for more than 50 sources and sinks 63 US regions (11 market regions) and international trade with 37 major trading

partners Detailed Bioenergy Market

Forestry & agricultural dedicated and residue feedstocks Tracks production of starch- and sugar-based ethanol, cellulosic ethanol,

biodiesel, and bioelectricity

17

FASOM Regions

South Central

Northeast

Southeast

Lake States

PacificSouthwest

GreatPlains

Rocky Mountains

PacificNorthwest

West East

Corn Belt

SouthWest

18

Simulated Changes in Regional Acreage Relative to Baseline, Corn (Acres)

Note: CB=Corn Belt, GP=Great Plains, LS=Lake States, NE=New England, RM=Rocky Mountains, PSW=Pacific Southwest, PNWE=Pacific Northwest East Side, SC=Southcentral, SE=Southeast, SW=Southwest

19

Simulated Changes in Regional Acreage Relative to Baseline, Soybeans (Acres)

Note: CB=Corn Belt, GP=Great Plains, LS=Lake States, NE=New England, RM=Rocky Mountains, PSW=Pacific Southwest, PNWE=Pacific Northwest East Side, SC=Southcentral, SE=Southeast, SW=Southwest

20

Simulated Changes in Regional Acreage Relative to Baseline, Wheat (Acres)

Note: CB=Corn Belt, GP=Great Plains, LS=Lake States, NE=New England, RM=Rocky Mountains, PSW=Pacific Southwest, PNWE=Pacific Northwest East Side, SC=Southcentral, SE=Southeast, SW=Southwest

21

Equilibrium Changes in National Average Yield Relative to Baseline

22

Simulated Changes in Average Market Price Relative to the Baseline

23

Annual Welfare Change in US Agricultural Sector (2011$)

CGCM-3.1 MRICGCM-2.2 GFDL-2.0 GFDL-2.1

(5.0)

(4.0)

(3.0)

(2.0)

(1.0)

-

1.0

2.0

Producer SurplusConsumer Surplus

(bill

ion

$)

24

Summary

Crop yields are affected by changes in temperature, precipitation, and other aspects of climate in a very heterogeneous way

Important to consider changes in both mean and variance of yields

Combined results from EPIC crop process model with FASOM model of forest and agricultural production and markets

Potentially substantial shifts in crop production and commodity markets, though considerable uncertainty

25

Ongoing Research/Extensions

Updated climate scenarios and new EPIC runs

Incorporation of impacts on U.S. forest productivity

International climate impacts on forests and agriculture

Improved incorporation of variability and uncertainty

Integration of insurance/risk management into market modeling

Integration of climate impacts and adaptation with mitigation Energy prices Renewable fuels assumptions Technological progress assumptions Carbon price paths