際際滷

際際滷Share a Scribd company logo

Ncar global econ g nelson

Download as PPTX, PDF
Gerald Nelson
Gerald Nelson

This document discusses scenario-based assessments of climate change impacts on agriculture. It summarizes a modeling intercomparison exercise that harmonized population, GDP, agricultural yield growth, and climate change across models to compare results. The models showed reductions in agricultural prices and production by 2050 under climate change. However, the models do not fully capture impacts on pests, diseases, extreme events, and nutrition. The document calls for improved collaboration across modeling communities to develop standardized data, share code and modeling approaches, and build a 21st century modeling environment to better address climate change impacts on food security.

1 of 27
Download to read offline
Scenario-based assessment of climate change impacts on agriculture Gerald C. Nelson Professor Emeritus, University of Illinois, Urbana-Champaign Presentation at the AAAS session New Scenarios for Assessing Future Climate Change, 26 February 2014
or how we used the SSPs (and an RCP) in a model intercomparison exercise
or should we do any more integrated assessments until we fix the Lamppost problem
Policy questions that need scenario answers What is the future of agricultural prices? How will agricultural production evolve? How will climate change alter Prices Land use Trade Undernourishment Do we have the tools to answer these questions?
Alternate perspectives on future prices with no climate change, 2000-2050 in 2011 IMPACT big price increases (e.g. 80% increase in coarse grains price)
Why do the results differ? Differing perspectives on Todays unknowns that we should know Future unknowns Economic development Population growth Climate change Natural resource availability Technological advance Differing economic modeling approaches CGE models more flexible (?) Functional forms determine outcomes (e.g., Armington assumption, demand parameters; ref Bennetts Law) 6
We had no idea which is most important 7
Scenario harmonization: Common values for key drivers Harmonized four key drivers Population from SSP2 and 3 GDP from SSP2 and 3 Exogenous component of agricultural yield growth Climate change effects on yield growth No harmonization on other important drivers Three orthogonal comparisons in 2050 Socioeconomics SSP2 versus SSP3 Bioenergy policies not covered here Climate change no climate change versus RCP 8.5 with no CO2 fertilization
SSPs: What did we use and why? What? SSP2 and SSP3, version 0.5 GDP, OECD version Population What not? Storylines Population makeup Urbanization Why?
SSP per capita income Per capita incomes in 2010; SSP2 and SSP3 in 2050 0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 World Developing East Asia South Asia Europe & Central Asia WB definitions Middle East & North Africa Sub-Saharan Africa Latin America & Caribbean High Income 2010 SSP2 SSP3
The modeling chain: From biophysical to socioeconomic Nelson, et al., PNAS 2013
Results: price changes between 2005 and 2050 are smaller, but
Maximum price changes between 2005 and 2050 reduced. Large differences remain. General equilibrium Partial equilibrium
2050 price effects from the socioeconomic scenarios, with no climate change 14
Model differ dramatically in their responses to the negative future of SSP3 Price declines Price increases Price not affected
Climate change and economic responses Quantity Price Demand Supply Climate change shifts supply to the left Initial shortage caused by climate change Initial shortage = big price increase Final quantity Finalprice Model choices - Supply response? - Area - Yield - Demand response? - Where? - How much trade?
Results from the climate change scenarios Comparisons are differences in 2050 outcomes
Mean -17% -11% +11% -2% - 1% -3% +20% Climate change reduces 2050 yields. Producers, consumers, & trade partially compensate Nelson, et al., PNAS 2013 Final yield Biophysical effect
How do the model distribute their 2050 responses to climate change? Mostly area Mostly yield Demand Area Yield Demand/ Supply Model Type AIM 0.12 0.90 -0.02 0.14 CGE ENVISAGE 0.17 0.32 0.51 0.2 CGE FARM 0.04 0.67 0.29 0.04 CGE GTEM 0.06 0.29 0.65 0.06 CGE MAGNET 0.1 1.39 -0.49 0.11 CGE GCAM 0.22 0.78 0 0.28 PE GLOBIOM 0.49 0.12 0.38 0.98 PE IMPACT 0.38 0.53 0.09 0.61 PE MAgPIE -0.01 -0.08 1.09 -0.01 PE AVERAGE 0.17 0.53 0.3 0.21 Large demand Nelson, et al., Agricultural Economics 2014
How plausible are the ensemble results? Two views They are too pessimistic GHG concentration pathway with the greatest forcing (RCP 8.5) Crop models assumed constant CO2 concentrations throughout the period They are not too bad Actual GHG concentrations similar to RCP 8.5 so far Field results from FACE experiments suggest CO2 fertilization effect in the field is less than in the lab
And then there is the Lamppost Problem
What is missing in our climate change results? The models dont include effects of Increasing pest and disease pressure Increasing extreme events Increasing ozone Effects on nutrition Models for most crops dont exist These could swamp the negative effects already quantified
What to do about the Lamppost Problem? 23
Merge the silos! 24 Its not enough for each community to work together. We need a community of communities, old and new. Standard data protocols, developed together Commonly agreed aggregation methods Centrally managed data storage
Open the black boxes so they arent reinvented 25 Share code within each community Identify best versions Centrally manage code storage and dissemination Identify data needs and develop new sources
Develop 21st century modeling environment Approach depends on topic Crop modeling Code hierarchy with modular construction Plant-level functions (e.g. photosynthesis) Species-level functions Variety-level functions Design with modularity in mind Identify critical parameters and data needs at each level Talk to the computational biology folks Facilitates the needed bulk development of models for fruits and vegetables 26
Conclusions Substance RCP8.5 results in lower yields Adaptation reduces some of those effects across the supply and demand side Economic models allocate response differently between supply (area and yield) and demand Process Existing models/methods are underestimating the effects of climate change on food security. We urgently need to address the Lamppost Problem before we do more assessments. 27

More Related Content

Ncar global econ g nelson

  • 1. Scenario-based assessment of climate change impacts on agriculture Gerald C. Nelson Professor Emeritus, University of Illinois, Urbana-Champaign Presentation at the AAAS session New Scenarios for Assessing Future Climate Change, 26 February 2014
  • 2. or how we used the SSPs (and an RCP) in a model intercomparison exercise
  • 3. or should we do any more integrated assessments until we fix the Lamppost problem
  • 4. Policy questions that need scenario answers What is the future of agricultural prices? How will agricultural production evolve? How will climate change alter Prices Land use Trade Undernourishment Do we have the tools to answer these questions?
  • 5. Alternate perspectives on future prices with no climate change, 2000-2050 in 2011 IMPACT big price increases (e.g. 80% increase in coarse grains price)
  • 6. Why do the results differ? Differing perspectives on Todays unknowns that we should know Future unknowns Economic development Population growth Climate change Natural resource availability Technological advance Differing economic modeling approaches CGE models more flexible (?) Functional forms determine outcomes (e.g., Armington assumption, demand parameters; ref Bennetts Law) 6
  • 7. We had no idea which is most important 7
  • 8. Scenario harmonization: Common values for key drivers Harmonized four key drivers Population from SSP2 and 3 GDP from SSP2 and 3 Exogenous component of agricultural yield growth Climate change effects on yield growth No harmonization on other important drivers Three orthogonal comparisons in 2050 Socioeconomics SSP2 versus SSP3 Bioenergy policies not covered here Climate change no climate change versus RCP 8.5 with no CO2 fertilization
  • 9. SSPs: What did we use and why? What? SSP2 and SSP3, version 0.5 GDP, OECD version Population What not? Storylines Population makeup Urbanization Why?
  • 10. SSP per capita income Per capita incomes in 2010; SSP2 and SSP3 in 2050 0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 World Developing East Asia South Asia Europe & Central Asia WB definitions Middle East & North Africa Sub-Saharan Africa Latin America & Caribbean High Income 2010 SSP2 SSP3
  • 11. The modeling chain: From biophysical to socioeconomic Nelson, et al., PNAS 2013
  • 12. Results: price changes between 2005 and 2050 are smaller, but
  • 13. Maximum price changes between 2005 and 2050 reduced. Large differences remain. General equilibrium Partial equilibrium
  • 14. 2050 price effects from the socioeconomic scenarios, with no climate change 14
  • 15. Model differ dramatically in their responses to the negative future of SSP3 Price declines Price increases Price not affected
  • 16. Climate change and economic responses Quantity Price Demand Supply Climate change shifts supply to the left Initial shortage caused by climate change Initial shortage = big price increase Final quantity Finalprice Model choices - Supply response? - Area - Yield - Demand response? - Where? - How much trade?
  • 17. Results from the climate change scenarios Comparisons are differences in 2050 outcomes
  • 18. Mean -17% -11% +11% -2% - 1% -3% +20% Climate change reduces 2050 yields. Producers, consumers, & trade partially compensate Nelson, et al., PNAS 2013 Final yield Biophysical effect
  • 19. How do the model distribute their 2050 responses to climate change? Mostly area Mostly yield Demand Area Yield Demand/ Supply Model Type AIM 0.12 0.90 -0.02 0.14 CGE ENVISAGE 0.17 0.32 0.51 0.2 CGE FARM 0.04 0.67 0.29 0.04 CGE GTEM 0.06 0.29 0.65 0.06 CGE MAGNET 0.1 1.39 -0.49 0.11 CGE GCAM 0.22 0.78 0 0.28 PE GLOBIOM 0.49 0.12 0.38 0.98 PE IMPACT 0.38 0.53 0.09 0.61 PE MAgPIE -0.01 -0.08 1.09 -0.01 PE AVERAGE 0.17 0.53 0.3 0.21 Large demand Nelson, et al., Agricultural Economics 2014
  • 20. How plausible are the ensemble results? Two views They are too pessimistic GHG concentration pathway with the greatest forcing (RCP 8.5) Crop models assumed constant CO2 concentrations throughout the period They are not too bad Actual GHG concentrations similar to RCP 8.5 so far Field results from FACE experiments suggest CO2 fertilization effect in the field is less than in the lab
  • 21. And then there is the Lamppost Problem
  • 22. What is missing in our climate change results? The models dont include effects of Increasing pest and disease pressure Increasing extreme events Increasing ozone Effects on nutrition Models for most crops dont exist These could swamp the negative effects already quantified
  • 23. What to do about the Lamppost Problem? 23
  • 24. Merge the silos! 24 Its not enough for each community to work together. We need a community of communities, old and new. Standard data protocols, developed together Commonly agreed aggregation methods Centrally managed data storage
  • 25. Open the black boxes so they arent reinvented 25 Share code within each community Identify best versions Centrally manage code storage and dissemination Identify data needs and develop new sources
  • 26. Develop 21st century modeling environment Approach depends on topic Crop modeling Code hierarchy with modular construction Plant-level functions (e.g. photosynthesis) Species-level functions Variety-level functions Design with modularity in mind Identify critical parameters and data needs at each level Talk to the computational biology folks Facilitates the needed bulk development of models for fruits and vegetables 26
  • 27. Conclusions Substance RCP8.5 results in lower yields Adaptation reduces some of those effects across the supply and demand side Economic models allocate response differently between supply (area and yield) and demand Process Existing models/methods are underestimating the effects of climate change on food security. We urgently need to address the Lamppost Problem before we do more assessments. 27

Editor's Notes

  • #6: OECD comparison
  • #23: Image source: http://www.flickr.com/photos/morville/4273477501/.