Search

Drought is one of the most extreme climatic events in South Asia (SA) and has affected 1.44 billion people in last 68 years. The agriculture in many areas of this region is highly dependent on rainfall, which increases the vulnerability to drought. To mitigate the impact of drought on agriculture and food security, this study aims to develop a state-of-the-art system for monitoring agricultural drought over SA at a high spatial resolution (0.25 ^∘ ) in near real-time. This study currently focuses on the rain-fed area, and the impact of irrigation is not incorporated. This open and interactive tool can assist in monitoring the near-present soil moisture conditions, as well as assessing the historical drought conditions for better management. The South Asia Drought Monitor (SADM) runs the mesoscale hydrologic model to simulate the soil moisture using observation-based meteorological forcing (at near real-time), morphological variables, and land cover data. The soil moisture index (SMI) has been calculated by estimating the percentile of the simulated soil moisture. The drought monitor displays the SMI in five classes based on severity: abnormally dry, moderate drought, severe drought, extreme drought and exceptional drought. The main functions of this open interactive system include the provisioning of up-to-date and historical drought maps, displaying long-term drought conditions and downloading soil moisture data. Comparison of the SMI with the standardized precipitation evapotranspiration index (SPEI) shows that the SMI and SPEI depict similar temporal distribution patterns. However, the SPEI (for 4, 6, 9 and 12 months) differs in the representation of the dry conditions in 1992, 2009, and 2015 and the wet condition in 1983, 1988, and 1990. We evaluated the implications of using different precipitation forcings in a hydrological simulation. A comparison of major drought characteristics such as areal extent, duration, and intensity, using different precipitation datasets show that uncertainty in precipitation forcings can significantly influence model output and drought characteristics. For example, the areal extent of one of the most severe droughts from 1986 to 1988 differs by 9% between ERA5 and CHIRPSv2.