A national-scale drought assessment in Uganda based on evapotranspiration deficits from the Bouchet hypothesis

Abstract

For effective management of drought risks, a process-specific indicator needs to be employed rather than depending only on precipitation anomalies. In this study, we performed an evapotranspiration-based (ET-based) assessment of agricultural droughts in Uganda where drought has usually been assessed with precipitation deficiency. The land-surface ET and evaporative demand were estimated with a state-of-the-art complementary principle across the country under poor ET data availability. Using the ET estimates, the Standardized Evapotranspiration Deficit Index (SEDI) was quantified to assess agricultural droughts for 1984–2017. The ET-based drought assessment was compared with conventional precipitation-based analyses using the Standardized Precipitation Index (SPI) in respect of severity, areal extent, and temporal trends. Results showed that the complementary principle provided acceptable performance in reproducing interpolated latent flux observations. SPI and SEDI were strongly correlated, and thus yielded identical temporal variations of mean drought severity and areal extent. The trend maps of SPI and SEDI also tended to agree, and indicate that central and northwestern Uganda would face increasing drought risks in the future. The declining precipitation in Uganda has been attributed to rising surface temperatures of the Indian Ocean, and is expected to persist up to the next decade. This study suggests that droughts in major croplands could be often accompanied by high heat-wave temperatures in Uganda due to strong land-atmosphere coupling, and thus mortality risks in ecosystems could increase together with water stress.