A Retrospective View of the Application of Global Gridded Reference Evapotranspiration (Invited Presentation)

The availability of gridded precipitation estimates combining satellite and in-situ data began in the 1990s, and provided a valuable new means of drought monitoring for famine early warning, alongside the vegetation index imagery that had been in use since the 1980s. Nonetheless, it soon became clear that these data only spoke to the supply side of the crop water balance question. Consequently, scientists supporting the Famine Early Warning System (FEWS) began extracting, from the Global Data Assimilation System (GDAS), the variable fields necessary to solve the Penman-Monteith equation for reference evapotranspiration (ET). This made possible a first-order reconciliation of rainfed crop water supply with demand over a growing season, so that the timing of any within-season rainfall deficits were properly weighted. Later, these time-series grids of reference ET were used in conjunction with estimates of land surface temperature (LST) to estimate actual ET. This proved useful for monitoring both rainfed and irrigated croplands, and rangelands too. Continued experience with the data over time revealed shortcomings attributable to the non-homogeneity of the GDAS time series. To overcome this, a more rigorous formulation of reference ET was implemented using NASA’s MERRA reanalysis. This put previous applications, like crop water balance calculations and actual ET estimation by remote sensing of LST, on a much firmer basis. It also made possible the development of the Evaporative Demand Drought Index (EDDI), whose sensitivity can detect the rapid onset of a “flash drought” which is not well resolved by other indices. Today, global gridded estimates of reference ET are incorporated in a variety of products for monitoring meteorological and agricultural drought, and famine early warning is significantly strengthened as a consequence.