Forecasting Water Stress Across the Sahel and East Africa
NASA Harvest Consortium Researcher Kimberly Slinski of the University of Maryland Earth System Science Interdisciplinary Center and NASA Godard Space Flight Center (GSFC) is the Principal Investigator on “Earth Observation-Based Monitoring and Forecasting of Rangeland Water Resources”, a newly funded project that aims to develop novel capabilities for monitoring and forecasting water availability in African rangeland ponds. The Project Team includes Shrad Shukla and Chris Funk of the Climate Hazards Center; Mike Jasinski of NASA GSFC; Gabriel Senay, Jim Rowland, and Mike Budde of the United States Geological Survey (USGS); Amy McNally of the Famine Early Warning Systems Network, Inbal Becker-Reshef of Harvest, Evan Thomas of the University of Colorado, as well as Action Against Hunger, the Regional Centre for Mapping of Resources for Development, Comité permanent Inter-État de Lutte contre la Sécheresse au Sahel/AGRHYMET, and the the Alliance of Bioversity International and the International Center for Tropical Agriculture (CIAT).
This project partners closely with the United States Agency for International Development’s Famine Early Warning Systems Network (FEWS NET), a leading provider of early warning and analysis on acute food insecurity around the world. FEWS NET partner USGS maintains the Water Point Viewer, an interactive map that monitors the depth and area of 338 water points across arid and semi-arid regions of the Sahel and East Africa, from Senegal to Somalia.
In its existing form, the tool helps stakeholders understand the current availability of water for livestock and human use, which in turn informs food security analysis, humanitarian assistance planning, and a range of other activities. The Rangelands Monitoring and Forecasting System project will significantly expand and improve the existing Water Point Viewer, increasing the locations monitored, developing new time series of water point surface area using high-resolution satellite data, and improving overall model physics. These developments will also lead to one of the most exciting contributions of this work: new predictive capabilities for forecasting water point stress.
Currently, climate services in East and West Africa have limited capacity to predict water availability. The proposed system will fill this gap by forecasting water point stress at weather (~2 weeks), sub-seasonal (~4 weeks), and seasonal time scales (~6 months). The advanced data streams will allow disaster response agencies to better plan for and implement mitigating actions, including ensuring access to alternate water supplies through water trucking or increased maintenance and repair of groundwater pumping stations. Mitigating actions also include limiting herd size to a number that can be supported by rangeland conditions (de-stocking). These actions are costly to implement. Therefore, good data on current and forecasted conditions are essential to the decision makers responsible for implementing drought-response actions.
The proposed work consists of four tasks:
Task 1: Employ machine learning methods using high-resolution synthetic aperture radar and multispectral data (such as from Landsat, Sentinel 1 & 2, NISAR, ICESat-2) to identify rangeland water points and generate an observation-based time series of water point area.
Task 2: Simulate water availability at each water point with a satellite data-driven water balance model. The model will be forced using satellite- and Earth systems model-based meteorological data, and calibrated using the observation-based time series.
Task 3: Forecast water availability at each water point using the calibrated model driven by meteorological forecasts at the weather, sub-seasonal, and seasonal time scales.
Task 4: Engage stakeholders at critical points in the project cycle, including needs assessment, verification of new water points, system pilot testing, and during the development of the web interface. This task also includes stakeholder training to ensure their full integration into stakeholder water stress and food security assessment and prediction workflows.
This project is one of the 30 chosen from NASA’s recent solicitation for proposals to support the research, development, and deployment of applications using Earth observations for water resources management. The total combined funding for these investigations is approximately $22 million over 3 years.