The State of AgMIP
Cynthia Rosenzweig, Jim Jones, Jerry Hatfield, John Antle,
Alex Ruane, Ken Boote, Peter Thorburn, Cheryl Porter, Sander Janssen,
Jerry Nelson, and Carolyn Mutter AgMIP Mid-Term Workshop, Kathmandu, Nepal, July 22nd, 2013
Worldwide Science Community
2nd Global Oct 2011
South America #1
1st Global Oct 2010
Rice
Wheat #2
Sub-Saharan Africa #1
South Asia #1
3rd Global Oct 2012
~500 listserve members
Maize
2
Track 1: Model Improvement and Intercomparison
Track 2: Climate Change Multi-Model Assessment
Cross-Cutting Themes
Uncertainty, Aggregation and Scaling,
Representative Agricultural Pathways Regional and Global Scales
AgMIP Sentinel Sites
Silver
Gold
Platinum
Two-Track Science Approach
Rosenzweig et al., 2013 AgForMet
3
Capacity Building
and Decision Making • Regional expertise
• Adaptation strategies
• Technology exchange
Climate Team
Crop Modeling Team
Economics Team
Information
Technology
Team
Improvements and
Intercomparisons • Crop models
• Agricultural economic models
• Scenario construction
• Aggregation methodologies
Cross-Cutting
Themes
• Uncertainty
• Aggregation and
Scaling
• Representative
Agricultural
Pathways
Assessments • Regional
• Global
• Crop-specific
Work
Groups
• Soils
• Water
Resources
• Livestock
and
Grasslands
• Pests and
Diseases
Teams, Linkages and Outcomes
Rosenzweig et al., 2013
4 Links to CCAFS, Global Yield Gap Atlas, Global Futures, et al.
• Regional projects awarded on competitive basis
• 15 countries in Sub-Saharan Africa
• 5 countries in South Asia
• 60 institutions and 120+ scientists.
• Methods for Integrated Regional Assessment (IRA) of climate change impacts on agriculture
• Link climate, crop, and economics models in protocol
approach to provide distribution of impacts across farm
households at regional scales.
• Create and disseminate handbook, climate scenarios
guidebook, and tools to enable research teams
• Train over 250 scientists
• Engage regional stakeholders and national media
• Handbook Series on Climate Change
and Agroecosystems (2012-2013)
• AgMIP regional project partners are chapter authors
• Citable IPCC references.
• Multiple crop model training
• 10 scientists ‘trained as trainers’ for Africa, Asia
• 50 scientists trained in multiple crop models and analytical methods
AgMIP Regional Projects
5
Ghana, 2012
Sri Lanka, 2012
Local Climate Change
Projections
6
Substantial
uncertainty in
projections,
particularly in
precipitation
changes
Mid-Century
(2040-2069)
RCP8.5
Compared to
1980-2009 period
Median of 20
CMIP5 GCMs
Temperature changes (˚C) Precipitation changes (%)
New Methods for
Regional Integrated Assessment
Antle et al., 2013 7
Regional climate projections
Calibrated crop model genetic
coefficients
On-farm economic survey data
Winners and losers
What AgMIP has learned so far
• Tremendous interest in agricultural research community in
interdisciplinary multi-model research and assessment
• Median of crop model ensembles reproduces observed yields
• Best Practices’ for model calibration essential for rigorous results • Crop responses to CO2, temperature, and water remain key sources
of uncertainty
• Regional integrated assessments extending methods for projecting
changes in farm systems
• Global crop yield impacts project greater vulnerability in lower
latitudes
• Limitations in fresh water available for irrigation may compound
climate impacts in many regions, while its abundance could help to
alleviate them in others.
• Agricultural prices projected to go up
• Crop and economic model uncertainty is large and generally greater
than climate model uncertainty
8
= Wheat
= Maize
= Rice
0˚
0˚ 90˚ -90˚
45˚
-45˚
= Sugarcane
Ames
Morogoro
Wongan Hills
Delhi
Ludhiana
Ayr
Los Baños
Piracicaba
Shizukuishi
Rio Verde
La Mercy
Haarweg
Lusignan
Balcarce
Nanjing
AgMIP Sentinel Sites
North America
South
America
Sub-Saharan
Africa
Europe
South
Asia
Asia*
Australia*
Rosenzweig et al., 2013
9
• Wheat (27 models), Maize (18), and Rice Model (14) Pilots underway
• Pilots under development for sugarcane, millet/sorghum, soybean, groundnut, and potato
• North America, South America, Europe, Sub-Saharan Africa, South Asia, Asia*, Australia*
Crop Model Intercomparison Pilot Studies and Integrated Assessment Regions
Accomplishments
•Implementation of “Global-Maize”
model run in DSSAT
•Collaboration among crop modelers
and experimentalists from Monsanto,
government agencies, and universities
•Completed development of
phenology and leaf area index
subroutines
•Preliminary testing of Global-Maize
using a number of datasets from
temperate and tropical Sentinel Sites
Crop Model Improvement
Groups
Maize Project goal
To develop and test new open-source maize model routines that build on advances in
crop modeling over the last 50 years
10
• Maize Model Improvement Work Group
– Leaders M.Tollenaar, S. Kumudini,
K. Boote, J. Jones, and participants worldwide
• Experimentalists and modelers interacting to use data on
CO2, temperature, water, and nitrogen responses
– Leaders J. Hatfield, K. Boote
-- First workshop held in Ames, Iowa, Sep 2012
-- BSSG Meeting April 2013
11
Global Gridded Crop Model
Results
median of 7 GGCMs and 5 GCMs
Lower latitudes are more vulnerable to climate change
Global Economics Models
Effects of climate change on agricultural prices
(S3-S6 results in 2050 relative to S1 results in 2050)
Source: AgMIP model runs, December 2012.
Nelson, Gerald C. et
al., “Agriculture and
Climate Change in
Global Scenarios:
Why Don’t the
Models Agree.”
Agricultural
Economics,
In review.
9 global
economic
models
There is potential for large price increases with climate
change, although uncertainty is also large
Phase 2
workshop
in Dublin
April 9,
2013
12
Coordinated Climate-Crop Model Pilot
(C3MP)
[CO2]
T
P
T
[CO2]
P
= baseline
Submitted results are used to fit emulators to estimate impacts
response surfaces. The response surfaces help to efficiently and
probabilistically assess crop responses to CMIP5 climate scenarios
Below: 135+ C3MP registrants come from 40 countries, analyzing at
least 14 crops with more than 10 crop models.
Change in Temperature [-1 ˚ C to +8˚C]
20 CMIP5 GCM
Projections over
Henry County, FL
%C
ha
nge in
Pre
cip
[-5
0%
to
50
%]
13 Cross Sections of Hypercube Emulator for percentage changes
in 30-year mean Peanut yield, from Ruane et al., in review.
Upcoming
– Sub-Saharan Africa: South Africa, Jul 2013
– South Asia: Nepal, Jul 2013
– Latin America: Brazil, Sep 2013
– East Asia: China, Oct 2013
– Fourth Global Workshop: New York, Oct 2013
– Climate Methods: ICTP, Trieste, early 2014
Next Steps and Upcoming Events
For protocols, up-to-date events and news,
and to join AgMIP listserve – www.agmip.org
Contacts
Cynthia Rosenzweig [email protected]
Alex Ruane [email protected] 14