Past and future weather- induced risk in crop...

22
A Center for Robust Decision Making on Climate and Energy Policy Past and future weather- induced risk in crop production Joshua Elliott University of Chicago and Argonne National Lab Computation Institute Columbia University Center for Climate Systems Research and NASA GISS

Transcript of Past and future weather- induced risk in crop...

Page 1: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

NSF Decision Making Under Uncertainty Collaborative Groups

A Center for

Robust Decision Making on Climate and Energy Policy

Past and future weather-induced risk in crop production

Joshua Elliott University of Chicago and Argonne National Lab Computation Institute

Columbia University Center for Climate Systems Research and NASA GISS

Page 2: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

2

Acknowledgements

Collaborators RDCEP (rdcep.org)

AgMIP (agmip.org)

ISI-MIP (isi-mip.org)

GGCMI (agmip.org/ag-grid/ggcmi/)

UK-US Taskforce on Extreme Weather and Global Food System Resilience

Support from USDA, UK-AID (DFID), DOE, USAID, NSF, …

Page 3: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

• Compare past & future distributions from ensembles of global crops models produced in AgMIP/ISI-MIP – Extreme (-) percentiles, variance & skewness of distributions

generally getting worse

– Global 1-in-100 year historical event occurs almost 1-in-30 years by mid-century

Non-stationary risk in agriculture

3

Reproduced from Extreme weather and resilience of the global food system

Prepared for the UK-US Taskforce on Extreme Weather and Global Food System Resilience

Page 4: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

• US maize, 1-in-200 year event ~25% worse in near-term

• China maize, no change at extreme tail, though skewness becomes much more negative

Regional effects sometimes very large

4

Page 5: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

• Resource availability – Water constraints will increase vulnerability in key regions

– Need: Improved reps of irrigation, collaboration with hydrological and groundwater modelers

• Management, technology and adaptation – Tech trends and adaptation interact w/ extremes and resource

demand to impact system-wide risk

– Need: Better data and models of technological change and detailed scenario analysis

• Better understand climate and extremes – Especially present, coincident, and non-stationary risk

– Need: Historical analogs for future extreme conditions and large fixed-forcing climate model ensembles (>1000 years)

Integrated analysis of climate and extremes

5

Page 6: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

• Resource availability – Water constraints will increase vulnerability in key regions

– Need: Improved reps of irrigation, collaboration with hydrological and groundwater modelers

Integrated analysis of climate and extremes

6

Page 7: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

Global water and climate

• With [CO2], direct losses of 400-1400 Pcal (8-24% of present)

• W/o [CO2] = 1400-2600 Pcal Elliott et al (2014)

PNAS, 111 (9): 3239-3244.

Compare ensembles of 11 global hydrological models (GHMs)

and 6 global gridded crop models (GGCMs) driven by climate

change scenarios from 5 GCMs under RCP8.5.

• Freshwater limits imply reversion of 20-60 Mha of irrigated cropland to rainfed by 2100, and a further loss of 600-2900 Pcal

Page 8: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

• Groundwater accounts for 60% of irrigation and 40% of drinking water in the US

• We find that warming, management, and adaptation pressures are expected to increase demand for irrigation in groundwater regions

• Working with groundwater modelers to create high-resolution scenarios for future irrigation availability

Irrigation and groundwater

8

Scanlon et al 2012.

Page 9: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

• Management, technology and adaptation – Tech trends and adaptation interact w/ extremes and resource

demand to impact system-wide risk

– Need: Better data and models of technological change and detailed scenario analysis

Integrated analysis of climate and extremes

9

Page 10: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

• Trends in planting date and season length expected to continue in absence of climate

– ~20% increase in maize season from 1980-2050

– Accounts for > 1/2 of yield trend from 1980-2012

Technology and development

10

Page 11: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

• W/o warming, yield growth continues through 2050

• W/ perfect adapt, most impact removed until 2040s – -13.9% loss in yield in 2040s relative to no-climate baseline

• Big % incr. in Canada and Northern states

• Declining yields across most of the southern states

11

Technology and adaptation (e.g. maize)

2040s yield, warming+adaptation

vs. no-climate baseline (%)

Page 12: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

Thirstier crops (technology and water)

• 1980-2012, yield incr. almost 70%, H2O only +8%

– W/o climate change, trend projected to continue

• But w/ warming and adaptation, water use accelerates

• Tech development and adaptation costs water

– +6.3% increase in irrigation demand without any warming

– +16% increase with climate change and adaptation

Page 13: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

• Better understand climate and extremes – Especially present, coincident, and non-stationary risk

– Need: Historical analogs for future extreme conditions and large fixed-forcing climate model ensembles (>1000 years)

Integrated analysis of climate and extremes

13

Page 14: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

• CMIP5 ensemble projects persistent severe drought conditions in US in the 21st century

Extreme drought risk 21st century

14

Reproduced from Cook et al 2015

• By mid-century, unprecedented in 1000 year record

• Use worst persistent drought in 20th century (1930s dustbowl) to understand potential future impacts

Page 15: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

The next dustbowl?

• Impact of dustbowl w/ 2012 and 2035 tech – 1936 drought today would be 50% worse than 2012

– 2035 earlier/longer season incr. avg. yield and slightly reduces drought impact

– Results confirmed with empirical model

• Higher temps makes much worse

Reproduced from Glotter and Elliott (2016). In review

Page 16: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

Temp/precip sensitivity

• T-P sensitivity plot for simulated crop yields using the median yield for the 2000s

• Temperature expected to increase significantly, precip largely same

• At extremes, damage from high T dominates (lines get more vertical)

• By ~mid-century, a typical year as bad as 1936 (w/o adaptation)

Page 17: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

Long-term LUC vulnerability

• Assume some threshold of loss (avg over 5 year period) will lead to long-term LUC

– Invest in irrigation, switch crops, CRP, or abandonment

• Best (1930s w/ 60% loss threshold): little LUC risk

• Worst (1930s+ΔT3 w/ 40% threshold): 85% of rainfed production at risk for long-term LUC

Page 18: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

• Extreme shocks to global and regional food systems likely to become more frequent and more severe – Must include climate, tech, and resource changes to capture risk

• Tech and management will push yields higher in coming years

– But not without more water

– CC slows yield growth, demands much more water

– Adaptation possible, but not without more water

– Some crops (e.g. sorghum) do better than others (e.g. cotton)

• Optimistic climate models give 10-15 years breathing room

• System-level resilience improves very slowly; still highly vulnerable to large scale and especially persistent drought events

• CC offers exploitable opportunities, used to offset losses

Conclusions

18

Page 19: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

• Large-scale drought and heat events accounted for 12% of all billion-dollar disaster events in the US from 1980-2011 but almost 25% of total monetary damages

• The 1988 US drought is estimated to have cost $79 billion (2013 USD), behind only Hurricane Katrina as the most costly weather-related disaster in US history.

19

• Warming temperatures and shifting precipitation patterns may exacerbate the problem, increasing the frequency and/or severity of large-scale droughts in sensitive agricultural regions

Vulnerability: Large-scale drought

Page 20: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

• Focus on extreme events

• Adding many sectors, including forestry

• Improving cross-sector interactions and land-use change and water.

• Adding a regional focus at the watershed level

• Adaptation to climate

GGCMI Phase 3 and ISI-MIP2.0 (2016)

20

Page 21: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

• Partial solution: AgMIP Linked global and regional assessments of climate, food systems, and extremes.

– Convening Aspen Global Change Institute workshop in September to discuss protocols for linked assessments

– Developing methodology and prototypes for North America and South Asia

Linked global & regional assessments

21

Page 22: Past and future weather- induced risk in crop …communications.ext.zalf.de/sites/crop-modelling/PDF...Past and future weather-induced risk in crop production Joshua Elliott University

• Three teams

– Climate extremes and ag impacts

– Government and market response

– Impacts to economies and health

• Study vulnerability of food system to extremes (1-in-200 year events)

– Opportunities for coordinated risk management

– Improve functioning of international markets

– Bolster national resilience to shocks

– Adapt agriculture for changing climate

UK-US Taskforce on Extreme Weather and Global Food System Resilience

22