Entropy-weight-based spatiotemporal drought assessment using MODIS products and Sentinel-1A images in Urumqi, China

Drought is one of the most severe natural hazards influenced by many factors, which can in turn cause severe damage to agricultural, economic, social and ecological systems. For assessing drought intensity, early studies have typically used single-factor-based modeling approaches to delineate a specific aspect of drought. In this study, we developed an entropy weight method (named LNPS-EWM) for drought assessment based on MODIS products and Sentinel-1A images, considering four important factors, including land surface temperature (LST), normalized difference vegetation index (NDVI), potential evapotranspiration (PET), and soil moisture. The new LNPS-EWM method was applied to analyze the spatiotemporal drought patterns in Urumqi for 2018-2021. The results show that LST and PET were the dominant factors, which accounted for about 70% while NDVI and soil moisture only accounted for about 30%. A five-level drought classification shows that severe drought accounts for the largest portion and exceptional drought for the smallest portion. From 2018 to 2021, the Urumqi city center is the most drought-prone area, followed by the low-lying areas, while the southwestern and eastern mountainous areas are in a mild drought. In the central region in the north-south direction, the drought intensity in Urumqi was mitigated from 2018 to 2021. These results are useful for risk assessment, large-scale monitoring, and early warning of drought conditions. This study improves our understanding of drought intensity patterns in arid Northwest China and should help improve regulatory and regional policies to combat drought to maintain eco-friendly cities in other arid regions.