mehta nifa talk-final
TRANSCRIPT
Predictability of Impacts of Decadal Climate Variability on Water and Crop Yields in the Missouri River Basin, and Its Use in Agricultural Adaptation
Vikram M. MehtaThe Center for Research on the Changing Earth System, Catonsville,
Maryland
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Supported by the U.S. Department of Agriculture – National Institute of Food and Agriculture Grant 2011-67003-30213 under the NSF-USDA-DOE Earth System Modeling Program.
⦿ Objectives of this project
⦿ Importance of the Missouri River Basin (MRB)
⦿ Decadal climate variability (DCV) impacts on MRB water and crops
⦿ Data and models
⦿ Major accomplishments and other outcomes
⦿ The future?
The Team
Vikram Mehta, Norman Rosenberg, Katherin Mendoza, and Hui WangThe Center for Research on the Changing Earth System, Catonsville, Maryland
Cody Knutson, Nicole Wall, Tonya Haigh, and Tonya BernadtNational Drought Mitigation Center, Univ. of Nebraska – Lincoln, Nebraska
Bruce McCarl, Mario Fernandez, Pei Huang, Jinxiu Ding, Theepakorn Jithitikulchai
Raghavan Srinivasan, Prasad Daggupati, and Deb DebjaniTexas A & M University, College Station, Texas
Amita Mehta and Saikumar PopuriNASA-UMBC Joint Center for Earth System Technology, Catonsville,
Maryland
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Using the Missouri River Basin as a case study:
☞ To develop an ‘end-to-end’, decadal climate and impacts prediction system
☞ To develop an adaptive water and agriculture management system
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Overarching Objectives
Importance of the Missouri River BasinMissouri River Basin
Largest river basin in the US
Covers 500,000sq. miles,10 States, many Native American reservations, parts of Alberta and Saskatchewan
Value of crops and livestock over$100 billion per year
117 million acres cropland, only 12 million acres irrigated
Produces 46% of wheat, 22% of grain corn, 34% of cattle in the United States
Dependence on the Missouri River for drinking water, irrigation and industrial needs, hydro-electricity, recreation, navigation, and fish and wildlife habitat
Over 2000 urban centers of various sizes
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Urban Areas in the Missouri River Basin
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Over 2000 urban areas
Increasing competition for water among various sectors
Droughts and Water in the Main Stem Reservoirs
Water in the Main Stem Reservoirs
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Severe to Extreme Drought Area
(Lower figure, courtesy Kevin Grody, USACE)
Pacific Decadal Oscillation and wheat production in the MRB
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Average production6.6 million tons
Average production13.8 million tons
Correlation coefficients between the PDO and wheat production time series0.5 without smoothing and 0.65 after smoothing all time series.
Decadal Climate and Impacts Information for Decision Support in the Missouri River Basin
DCVphenomena
Influences on Basin hydro-
meteorology
Influenceson
AgricultureUrban water
IndustriesNavigationRecreation
Others
Rural andurban
economies; Local,
regional, national,
international economies
Adaptation strategies
viaunderstanding, prediction, and
scenario development
Data, information,
and decision- support systems
Active involvement of stakeholders
and policymakers
Applications In various
sectors
A system adaptableto other river basins also
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Data and Models
◉ Observed streamflow data from U.S. Geological Survey: Many gauge locations, 1961 - 2010
◉ Crop yield estimates from the USDA – National Agricultural Statistical Service (NASS): County totals, 1961 - 2010
◉Observations-based precipitation, daily max. and min. temperatures, surface winds, surface air humidity: 12 km x 12 km, 1961 - 2010
◉Decadal sea-surface temperature and hydro-meteorological predictions by 4 Earth System Models (NCAR-CCSM4, GFDL CM2.1, UKMO-HadCM3, and MIROC5): Monthly, various spatial resolutions, 1961 to 2010
◉Calibrated and validated 12 km x 12 km version of the Soil and Water Assessment Tool (SWAT) for the entire MRB
◉ Water and crop choices model RIVERSIM, optimized for the MRBVikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Hybrid Dynamical-Statistical Prediction System for Decadal Climate and
Hydro-meteorology
1961 – 2010ocean-
atmosphere- land hindcasts
Earth System Model (ESM)
1961-2010=me history
of greenhouse
gases, volcanic and
other aerosol op=cal
depths, solar radia=on; Projected values
aFer 2010 Ensemble ini=aliza=on system; 10-year experiments
ini=alized in 1960, 1970, …, 2000
2011 – 2036ocean-
atmosphere- land forecasts
Dynamical System for SST Prediction(CMIP5)
Models used in this study
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
GFDL – CM2.1 UKMO – HadCM3 MIROC5NCAR – CCSM4
SWAT Setup and Calibration - Validation
�Characterization of ~ 14,000 watersheds�Sub-watersheds and streams�Land use – land cover at30 m resolution, crop rotation and irrigation�Irrigated land and soil data�Precipitation, temperature, winds, solar radiation data at 12 km x 12 km�Crop yield calibration; Winter and spring wheat, corn (dryland and irrigated), soybean (dryland and irrigated)�Water yield (total surface and base flow) calibration�Water abstractions and other man- made changes not captured
Calibration and validation in each of these 11 land use classes
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Types of SWAT Experiments
SWATObserved Hydro- meteorological data; 1961 - 2010
Idealized Hydro- meteorological data from DCV scenarios
Water yield, stream flow, crop yields
Comparison with observed data; 1961 - 2010
SWAT Water yield, stream flow, crop yields
Inter-comparison of impacts of scenarios
SWATHindcast Hydro- meteorologicaldata; 1961 - 2010
Water yield, stream flow, crop yields
Comparison with observed data; 1961 - 2010
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Observed and SWAT-simulated streamflow anomalies (cu. m/s) in wet (1982-86) and dry (1987-90) Epochs
Dry: 1987 - 90
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
USGS
SWAT
Wet: 1982 - 86
USGS
SWAT
Decreased flows in western Montana and northern Kansas, and increased flows elsewhere
Increased flows in western Montana and northern Kansas, and decreased flows elsewhere
Observed and SWAT-simulated winter wheat yield anomalies (t/ha) in wet (1982-86) and dry (1987-90) Epochs
Wet: 1982 - 86 Dry: 1987 - 90
Decreased yields in western Montana and southeast MRB, and increased yields elsewhere
Increased yields in western Montana and southeast MRB, and decreased yields elsewhere
NASS
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
SWAT
NASS
SWAT
Selected sub-basins for development of predictability and adaptation methodologies
Platte
Lower Grand
James
Marias
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Decadal prediction of the Pacific Decadal Oscillation index by the MIROC5 Earth System Model: 1981-2020
Vertical dashed lines – forecast start times
Shading:±1 std. dev. of ensemble
Impacts prediction periods indicatedby
PDO+ 1982-842003-06
PDO- 1988-902007-13
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
ehta NIFA PDs’ Meeting, San Francisco, CA
Winter wheat prediction in the Platte River sub-basin
Winter wheat yield anomaly prediction with 50-year NASS data statistics
PDO+
Winter wheat yield anomaly prediction with SWAT
Actual 1982-84
Actual yield anomalies
PDO-
Actual 1988-90
Predicted PDO+ 1982-84 Predicted PDO- 1988-90
Vikram M 17 December 2016
ehta NIFA PDs’ Meeting, San Francisco, CA
Barley prediction in the Marias sub-basin
Barley yield anomaly prediction with 50-year NASS data statistics
Barley yield anomaly prediction with SWAT
Actual 1982-84
Actual yield anomalies
Predicted PDO+ 1982-84 Predicted PDO- 1988-90
Actual 1988-90
PDO+ PDO-
Vikram M 17 December 2016
Corn prediction in the James River sub-basin
Corn yield anomaly prediction with 50-year NASS data statistics
Actual yield anomalies
Predicted PDO+ 2003-06 Predicted PDO- 2007-13
Actual 2003-06
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Actual 2007-13
Why is there value?�Decision makers would likely make adjustments for revised expectations of crop yields and water supply.v If farmers knew water demands would be higher and supplies smaller, they might plant a smaller area.
v If they knew some crop yields would be lower and others higher, they might shift crop mix.
v If water planners knew water would be short, they might encourage conservation, reduce use, and negotiate options to buy out some agricultural water rights.
�These actions compared to actions based on the historical climatic record would generate value.
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Value of decadal climate information to MRB agriculture
Difference
82 29 53
• Modeling Scope– 16 crops, 411 counties in MRB; irrigated and
dryland cropping; municipal and industrial water use– DCV impacts on 5 crops: wheat (spring and winter),
corn, sorghum, soybeans; hydrological flow balance; reservoir storage
• Uncertainty about which combination of phases of major DCV phenomena will occur in the next one year
• Includes crop insurance
Value of DCV Information (in million $)
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Perfect information on next year’s DCV phase relative to
historical frequency
Reduced uncertainty on next year’s DCV phase (based on this
year’s phase) relative to historical frequency
Water and Agriculture Choices RIVERSIM model
NIFA PDs’ Meeting, San Francisco, CA
Decadal climate variability based crop adaptation options for farmersExample: Corn – soybean – wheat – hay planting choices in
the Platte River sub-basin, given DCV phase combination prediction
PDO+, TAG- PDO-, TAG+
Vikram Mehta 17 December 2016
Corn Corn
Hay Hay
Soybeans Soybeans
Winter wheat
Winter wheat
Summary of accomplishments
• Simulations of DCV phenomena and their decadal predictability assessed in four Earth System Models (ESMs)
• Statistical hydro-meteorological prediction system, using predicted DCV indices as predictors, developed
• Statistical downscaling scheme developed for ESM data
• Very high resolution land use-hydrology-crop model Soil and Water Assessment Tool (SWAT) calibrated and validated for the MRB; water and crop yield simulation and prediction experiments conducted
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Summary of accomplishments
•Skillful, multiyear to decadal prediction of indices of the PDO and other DCV phenomena possible in some cases, especially since the 1980s.
•These predicted DCV indices are shown to be useful for multiyear to decadal prediction of water yields, streamflow, and crop yields in the Missouri River Basin.
•The predicted crop yields are shown to be useful for, among other applications, adaptation of choice of crops to plant.
•Value of DCV information to MRB agricultural economy estimated at $30 – 80 million per year.
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Research Capacity-building
Project people: 6 senior scientists, 2 research scientists, 1 post-doctoral scientist,1 research associate, 5 Ph.D. students, 16 undergrad. students, 3 outreach specialists, 1 web programmer, 1 administrative officer, 1 information technologist
Minorities and Under-represented groups: 6 women, 1
Hispanic Institution network: 1 non-profit
organization and 3 Universities Project website:
Missouri.crces.org
Stakeholder Advisory Team (SAT): Representatives from water, agriculture, and natural resources management sectors in the MRB; academics; state and federal officials
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Other outcomes
Completed Ph.D.s: 4 (Texas A & M University)
Ph.D. in progress: 1 (University of Maryland – Baltimore County (UMBC))
Undergraduate student interns: 16 in UMBC under the NSF Research Experience for Undergraduates (REU) Program
Papers published: 7; Papers in preparation: 4
Conference/Workshop and other talks and posters: 10
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
The Future?
●Decadal prediction of water and crop yields at the county level and development of fine-scale adaptation options in the MRB.
●County-level assessment of value of DCV information to agricultural economy in the MRB.
● Simulation and prediction of coupled food-energy-water securities in the MRB and the Mississippi River Basin.
● Adaptation of models and methodologies to the Mississippi and Ohio River Basins to develop prediction and adaptation systems for water, crops, and water-borne transportation of agricultural and other materials/products.
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016
Thank you!
Vikram Mehta NIFA PDs’ Meeting, San Francisco, CA
17 December 2016