A Process-Cognizant Remote Sensing Model for Subtropical Vegetation Drought in the Hunan-Jiangxi Region, China

In the context of global climate warming, the rising temperatures have triggered a surge in both the frequency and severity of drought in subtropical regions. Consequently, extensive vegetation mortality has emerged, posing a substantial threat to vegetation ecosystems. Accurate quantification and understanding of vegetation drought are imperative for regulating vegetation mortality rates of drought events. However, controversy persists surrounding the precise quantification of vegetation drought. Therefore, the development of timely and effective methods for the accurate monitoring of widespread vegetation drought is of utmost importance. Despite the studies revealing vegetation drought at different temporal and spatial scales, the precise quantification of drought variations among various vegetation covers, as well as within the same vegetation cover, remains unexplored.

In this study, for the precise quantification of "same vegetation cover with different drought degrees" and "same drought degree with different vegetation covers", a vegetation drought response (VDR) module is developed. This module accurately characterizes the spatiotemporal response of vegetation to soil moisture over time, based on spatio-temporal features constructed using the multispectral-based modified vegetation index and land surface temperature. To scientifically define drought boundaries, the study leverages knowledge involving "decreasing soil moisture leading to withering vegetation" and "increasing soil moisture resulting in flourishing vegetation" to identify the time intervals during the vegetation drought process (VDP). Within the VDP intervals, the sensitivity of vegetation response to soil moisture determines the characteristics of VDR in the beginning of drought, which is then utilized to establish the vegetation drought threshold (VDT). By applying the VDT to VDR, the study constructs a process-cognizant vegetation drought model (PCVDM) to achieve a quantitative inversion of vegetation drought. Using the Hunan-Jiangxi region in the central subtropical zone of China as a case study, this research employs remote sensing techniques to quantitatively retrieve the spatiotemporal changes in vegetation drought from 2000 to 2023. Furthermore, it conducts a spatiotemporal differentiation analysis and causation discrimination by integrating altitude and lithology conditions.

The findings of this study highlight the valuable insights on the spatiotemporal dynamics of vegetation drought supported with the PCVDM. The PCVDM can be utilized for remote sensing monitoring of vegetation drought in subtropical regions, enabling the identification of spatial differentiation in vegetation drought in the Hunan-Jiangxi region based on altitude and geological lithology. This study reveals the overall trends in vegetation changes in the Hunan-Jiangxi region since the beginning of this century: areas at higher altitudes (>800m) exhibit increased greenness due to rising temperatures, while lower altitude areas (<200m) experience intensified vegetation drought due to increased evapotranspiration. Meanwhile, moderate altitude areas (~400m) are influenced by the spatial differences in geological lithology, where increased greenness coexists with vegetation drought phenomena.