Early detection and monitoring of the extent and severity of agricultural drought is an important component of food security assessment in the developing world. Because livelihood systems are climate-sensitive, and conventional surface instrument networks are sparse and report with delays, satellite remote sensing and modeling play a vital role in early warning systems. Even in the U.S., where station networks are good, there are gaps that need to be filled in by these methods to get the complete picture of drought events. Consequently, there have been significant investments made to integrate NASA Earth science data and models into drought monitoring and related early warning activities.
MODIS vegetation index images are of high value for the detection of anomalies in plant cover and phenology that are indicative of drought. They can be readily integrated into harvest assessments to reveal drought impacts. The recently implemented NASA LANCE and USGS eMODIS systems have significantly reduced product latencies, enhancing early warning activities both in the U.S. and overseas.
MODIS land surface temperature (LST) data are also important. Day time LST data are used in an energy balance approach to estimate evapotranspiration by crops and native vegetation. Such estimates help identify drought-induced anomalies in crop production and rangeland conditions. They also provide essential boundary conditions for regional aquifer modeling studies. Night time LST data have been adopted as a key forcing variable in a malaria vectorial capacity model developed at the IRI and implemented at USGS as part of routine processing for famine early warning. Malaria control agencies in Africa use the model output in their GIS to plan deployment of their resources.
Crop water balance models provide an assessment of drought impact that is independent and complementary to the use of vegetation index imagery. Satellite rainfall estimates are a key forcing variable. Researchers at UC Santa Barbara have made important strides in the removal of bias in TRMM rainfall estimates by using them in conjunction with gridded station climatologies.
Livestock form the basis of food security of pastoralists in East Africa, whose livelihoods follow the seasonal changes in the availability of forage and drinking water. Techniques for tracking the status of watering holes with ASTER imagery have added a new dimension to monitoring by the USAID Livestock Early Warning System.
Perhaps the best way to take advantage of satellite observations is by assimilating them into land surface models that provide gridded estimates of essential variables like soil moisture, runoff, snow pack, and evapotranspiration. The Land Information System (LIS) is now being implemented by NOAA to provide snow monitoring over a broad swath of Central Asia for early detection of agricultural drought. An instance of LIS is also being developed for Africa to support famine early warning. These land data assimilation systems will not only make possible enhanced application of NASA data already used for drought monitoring, but will also set the stage for exploitation of GRACE gravimetric data and soil moisture products from the upcoming SMAP mission.