Advancements in the Temperature-Soil Moisture Dryness Index (TMDI) for Drought Monitoring in Southwestern Taiwan

Drought, a destructive natural disaster, poses a significant threat to vulnerable areas worldwide. Its occurrence in Taiwan brings up concerns, particularly for vital sectors, such as the high-end semiconductor chip industry. Many satellite-based indexes have been developed to monitor the drought. The temperature-vegetation dryness index (TVDI), a commonly used drought index, uses an empirical simplification of land surface temperature (LST) and fractional vegetation cover (FVC). The newly developed temperature-soil moisture dryness index (TMDI) using the LST-normalized difference latent heat index (NDLI) space is regarded as an alternative to TVDI due to its improved ability in vegetation-sparse areas. This article presents advancements in the TMDI using the novel fractional surface water availability (FSWA) derived from the NDLI, with an emphasis on enhanced edge selection in the LST-FSWA trapezoidal space for observing drought states. An effective method has been used to select the dry and wet edges within this trapezoid. The ability of the indexes was evaluated using indicators, including the surface energy balance algorithm for land (SEBAL)-based crop water stress index (CWSI) and evapotranspiration (ET), gross primary productivity (GPP), and in situ precipitation. The results show high correlations (r) between the TVDI and both CWSI and ET, with r values of 0.85 and -0.83, respectively. The TMDI reveals even stronger relationships with CWSI (r = 0.93) and ET (r = -0.94) and is more sensitive than individual variables (FVC, FSWA, and LST) and TVDI. It also indicates a high correlation between the TVDI and GPP (r = -0.69), while the TMDI displays a higher correlation with GPP (r = -0.75). Based on the spatiotemporal analysis, the TMDI was spatially well-matched with CWSI and GPP across most of the study area. Compared to other indexes, the TMDI exhibits the highest sensitivity to precipitation (r = -0.60). By leveraging the CWSI classification, a new TMDI threshold is proposed to assess drought status in southwestern Taiwan during the fourth quarter of the years 2014-2021. Overall, the TMDI accurately captures spatiotemporal variations in drought status, providing valuable insights for irrigation managers to effectively manage limited water resources.