Integrating GRACE for Drought Monitoring and Decision Support

Integrating GRACE into Drought Monitoring, Flood Forecasting and Agricultural Markets related Decisions Richard Bernknopf, Yusuke Kuwayama, Matthew Rodell, and Benjamin Zaitchik October 30, 2019

Measuring the socioeconomic benefits of GRACE in the VALUABLES framework 1 satellite and 3 impact assessments: Probabilistic model (ensemble product): Spatiotemporal RSDI improve the reliability of existing forecast and monitoring methods of natural hazards Gravity and Recovery Climate Experiment (GRACE) Valuation (quantitative method market, nonmarket): Economic estimation of return on investment (expected benefits) to reduce societal risks USDA Drought Monitor for disaster assistance NOAA NWS river flow forecast for floods Soil moisture and yield forecasts for grain markets Gravity Recovery and Climate Experiment (GRACE) Source: NASA/JPL-Caltech (2013) Empirical analysis (VALUABLES framework): Integration of the probabilistic earth science model and values at risk for impact assessment 2 Integrating GRACE into Drought Monitoring, Flood Forecasting and Agricultural Markets Related Decisions

How does information yield benefits in application 1? INFORMATION USDM is a drought severity classification of an objective blend of data (PDSI, CPC Soil Moisture, USGS daily streamflow precipitation, remotely sensed vegetation health) and expert weighting. GRACE informs the US Drought Monitor (USDM) with 3 water storage indicators to monitor weekly drought conditions. DECISIONMAKER ACTIONS Managers implement regulations with total water storage data for loss estimation and disaster assistance decision. Managers implement USDM to determine eligibility for drought disaster assistance. OUTCOMES FOR PEOPLE AND THE ENVIRONMENT USDA drought assistance programs are administered with more information that reduce government costs. USDM is the basis for decisions associated with the application and allocation of drought disaster assistance and insurance. 3 Integrating GRACE into Drought Monitoring, Flood Forecasting and Agricultural Markets Related Decisions

1. Integrating GRACE into the U.S. Drought Monitor (USDM) R. Bernknopf, Y. Kuwayama, D. Brookshire, M. Macauley, B. Zaitchik, A. Thompson, and P. Vail Drivers: US Agricultural Act 2014 Payments to eligible livestock producers due to drought (Livestock Forage Disaster Program) USDA Farm Service Agency Emergency Designation / Declaration Emergency loans to eligible producers suffering losses due to drought - Secretarial Disaster Designation Process 7 CFR Part 759 New Information: Ensemble of GRACE data & a catchment land surface model of water storage information as a statistical distribution of wetness (percentiles) Production of 3 weekly drought indicators: surface soil moisture, root zone soil moisture, and groundwater 4 Integrating GRACE into Drought Monitoring, Flood Forecasting and Agricultural Markets Related Decisions

Case study results Outcome: Reduction in the magnitude of the prediction error for farm income with the addition of GRACE to the USDM for 2002 2013 Reduction of prediction error should increase the efficiency of management decisions Policy Analysis and Implication: USDA Livestock Assistance Grant Program - State block fund to recover forage losses in 2006 drought $50 million for eligible counties; eligibility if USDM classified county at D3 or D4 between 3/7 and 8/31 How would county eligibility have been affected if more GRACE-DA information were taken into account? About $16 million of the $50 million would have been allocated to different States. 5 Integrating GRACE into Drought Monitoring, Flood Forecasting and Agricultural Markets Related Decisions

How does information yield benefits in application 2? INFORMATION GRACE informs National Weather Service river flow models with 3-month forecasts of soil moisture and groundwater storage statistical distributions to improve flood forecasts. NWS produces expert-based probabilistic river flow forecasts for flood protection that includes a regional updating process to accommodate local conditions. DECISIONMAKER ACTIONS USACE flood hazard communications are based on 2 new wetness indicators to improve timing of hazard announcements. Informs USACE hazard assessments and warnings to stakeholders for short term flood protection, management decisions and response. OUTCOMES FOR PEOPLE AND THE ENVIRONMENT A hazard announcement helps communities minimize losses. Communication of information and warnings uses predictive information to protect people and property. NWS and USACE programs provide inputs to communities for more cost-effective short-term flood protection decisions. 6 Integrating GRACE into Drought Monitoring, Flood Forecasting and Agricultural Markets Related Decisions

2. Integrating GRACE into flood hazard communication R. Bernknopf, Y. Kuwayama, R. Gibson, M. Rodell, B. Zaitchik, R. Schueneman, M. DeWeese Drivers: National Weather Service (NWS) Policy Directive 90-2 September 30, 2016 Staffing and Organization NWS Mission and Organization Department of the Army U.S. Army Corps of Engineers Risk Assessment for Flood Risk Management Studies, Engineer Regulation 1105-2-101, July 17, 2017, Circular No. 1110-2-6074, 31 January 2018, Engineering and Design Guidance for Emergency Action Plans, Incident Management and Reporting, and Inundation Maps for Dams and Levee Systems New Information: GRACE provides a distribution of soil moisture storage to the Sacramento Soil Moisture model in the NWS Community Hydrologic Prediction System 3-month soil moisture wetness percentiles as a statistical distribution Valley City, ND 7 Integrating GRACE) into Drought Monitoring, Flood Forecasting and Agricultural Markets Related Decisions

GRACE 3-month soil moisture indicator GRACE DA provides an input to the Sacramento Soil Moisture Accounting model in a flood hazard forecast (3/7/2011) 8 Integrating GRACE into Drought Monitoring, Flood Forecasting and Agricultural Markets Related Decisions

Flood exceedance probabilities without and with GRACE (03/18/2011 06/16/2011) Source: NWS 2018 ESP + GRACE 03/11/ 2011 ESP Open 03/11/2011 Conditional simulation (CS) line indicates chances of the river to rise above current conditions. 9 Integrating GRACE into Drought Monitoring, Flood Forecasting and Agricultural Markets Related Decisions

2011 Emergency Levee Construction Cost (Source: R. Schueneman, Field Report, Sheyenne River, Valley City in Barnes County, ND, May 2011) Valley City mitigation had 3 phases using clay levees, sandbags, and Hesco barriers. Implemented Phase 1, then moved into Phase 2 as conditions changed, ultimately implemented Phase 3. NWS forecast and USACE risk information helped place protective measures for April 13- 15 high flows that would be released from Baldhill Dam. Event in progress by April 11 Emergency levees for risk management Option 1 Option 2 Option 3 Option 3 implemented Protection Effectiveness Protect City to Option 1 and if conditions warrant raise to 30% event (20.4-feet) Protect City to 30% event (20.4-feet) Protect City to 24-feet and construct contingency levees behind all sandbag and Hesco lines 50% event (19.1- feet) Effort 5-6 days, 12 hrs. per day 4-5 days, 24 hrs. per day 9-10 days, 24-hrs per day City priorities: 1. Downtown business district 2. Valley City State University 3. Remaining areas Total cost $1.3M $3.0M 10

Cost effectiveness of estimated protection cost savings from an improved hazard communication Estimate the savings with GRACE DA that could have been realized for the emergency measures Efficiency gained via better planning with a more accurate flood crest estimate by 3/11/11 to implement option 3. Waiting to choose the option caused a scramble for supplies due to a greater severity of the flood resulting in a higher price per yard. A savings would have been $6.11 per yard of clay placed. Assumptions: No change in structural damage to buildings from flood due to hazard communications 1 event determined by USACE as high flow event in April 2011 Option 3 update implemented: $1.3 million at $16.85 per cubic yard. Hypothetical example: Valley City, ND property at risk = $169M (2009) Actual mitigation cost = $3.0M (2011) Estimated mitigation cost with GRACE = $1.3M Cost savings = $1.7M 11 Integrating GRACE into Drought Monitoring, Flood Forecasting and Agricultural Markets Related Decisions

3. Integrating GRACE to Improve Price Forecasts for Corn to Reduce Market Risks R. Bernknopf, Y. Kuwayama, M. Rodell, and B. Zaitchik Driver: Markets such as the Chicago Board of Trade rely on weather related information to estimate the future value of the price of corn. The futures markets use data inputs to forecast the supply of corn currently and at other critical times of the crop year. Corn yield forecasts inform decisions to purchase security options by farmers and commodity market traders and by government program managers. Source: https://www.cannontrading.com/ tools/futures-market/grains/Corn-Futures 12 Integrating GRACE into Drought Monitoring, Flood Forecasting and Agricultural Markets Related Decisions

Applying Market-Based Methods for Monetizing Uncertainty Reduction: A Case Study (Cooke and Golub 2019) to Corn Futures Crop yield forecasts are based, in part, on field sampling of soil moisture (Place and Brown 1987) and simulations for estimating soil water budgets (Boorn and Zidek 2012) Quantification of corn production and price are based on: Local and regional agricultural production models (includes weather) Market factors e.g., international exchange rates, fuel prices New Information: GRACE provides input to the soil water transpiration rate in the production model 3-month soil moisture wetness percentiles represented as a statistical distribution Study status 13 Integrating GRACE into Drought Monitoring, Flood Forecasting and Agricultural Markets Related Decisions

Summary and Lessons Learned GRACE Reduces uncertainty of water storage estimates in a variety of decisions Adds a data input to improve estimation skill in monitoring and forecasting droughts and floods VOI benefits are the benefits that is due to a change in the information. The net benefits to the decision maker are the result of the information being available + the actions taken (e.g., disaster mitigation) to avoid a loss or promote a gain. Seek out, identify and engage the decision maker. Convince stakeholders a VOI study is not an evaluation of them or their work. Be mindful of agency mission and turf. 14 Integrating GRACE into Drought Monitoring, Flood Forecasting and Agricultural Markets Related Decisions

Thank you. Learn more about VALUABLES: rff.org/valuables Follow us on Twitter: @RFFvaluables Email with questions: rbern@unm.edu