An Improved TVPDI for Spatiotemporal Drought Dynamics Analysis in Xinjiang, China

The Temperature-Vegetation-Precipitation Drought Index (TVPDI) performs poorly in complex terrain due to Normalized Difference Vegetation Index (NDVI) saturation and land surface temperature (LST) retrieval inaccuracies. To address this, we adopted an improved TVPDI (ITVPDI) by incorporating Leaf Area Index (LAI) and the land surface–air temperature difference (LST−T). By using multi-source data from 2000 to 2022 in Xinjiang, China, we validated ITVPDI and analyzed drought dynamics. Results show: (1) ITVPDI correlates better with solar-induced chlorophyll fluorescence (SIF) (r = 0.17) and the moisture index (MI) (r = 0.22) than the traditional TVPDI, demonstrating superior performance in densely vegetated and topographically complex areas. (2) Drought frequency ranked as follows: severe (31.55%) > moderate (29.04%) > extreme (23.44%) > mild (15.94%). Mild and moderate droughts occurred in Northern Xinjiang and the Tianshan Mountains, while severe and extreme droughts clustered around the Tarim Basin and Eastern Xinjiang desert margins. As drought intensity increases, its center of gravity shifts “from north to south” and “from mountains to basins.” (3) ITVPDI showed a slight upward trend over the 23-year period, with autumn experiencing the most severe drought (mean ITVPDI = 0.293). (4) A mean Hurst index of 0.468 indicates weak anti-persistence, suggesting the current wetting trend may reverse, and increasing future drought risk. The ITVPDI proves to be a robust tool for drought monitoring in arid and semi-arid regions with complex terrain. This study provides crucial scientific support for regional water resource allocation, precision irrigation, and collaborative drought resistance and disaster mitigation in Northwest China.