Evapotranspiration estimation using thematic mapper spectral satellite data in the Ethiopian rift and adjacent highlands

Quantification of actual evapotranspiration from spectral satellite data requires determination of the surface energy balance components with the support of field observations. In this paper an approach called Surface Energy Balance Algorithm for Land (SEBAL) is presented to calculate daily evaporation of some of the Ethiopian rift lakes and the surrounding landscape from the Thematic Mapper spectral satellite data on the basis of the energy balance approach. The result has been used in part to support long-term evapotranspiration estimation made using conventional methods. Conventional open-water evaporation estimation methods usually assume that the surface temperatures of lakes remain uniform and that evaporation proceeds more or less at the same rate over the total lake surface. However, this study clearly demonstrated that evaporation from a single lake varies depending upon surface temperature, albedo and on the incoming fluxes from groundwater and surface waters. The daily evaporation of the rift lakes is highly variable, ranging from 4.9 to 5.9 mm/day. On the land surface, actual evapotranspiration increases with altitude following the availability of moisture in the soil instead of the temperature gradient. On average it ranges from 0.2 mm/day in rift lacustrine soils and salty lake shores to around 4.5 mm/day in the highland mountain peaks covered with afro-alpine vegetation. The relative spatial differences of the SEBAL result from the land surface is in good agreement with the conventional evapotranspiration estimates. (C) 2003 Elsevier B.V. All rights reserved.