Spatio-Temporal Evaluation of Agricultural Drought Indices Using Remote Sensing and Soil Moisture Data in the Shiriya River Basin

Abstract

The current study presents an exploratory analysis in the Shiriya River Basin of the relationship between agricultural drought and LULC dynamics using remote sensing data over the period from 2002 to 2026. We used multi-sensor time series data to estimate rainfall, land surface temperature, and drought. The analysis was carried out using the NDVI, LST, VCI, SMI, TCI, VHI, and TVDI indicators. Rainfall showed a certain increase (trend line slope = 40.78mm/year; Sen’s slope) over the period, although with significant year-to-year variability. The LST was characterised by a slight decrease (trend line slope = -0.0988℃/year; Sen’s slope). There were extreme temperature fluctuations, particularly in 2006 and 2018. Based on the vegetation indicator data (NDVI, VCI, VHI), it was found that the study basin was dominated by vegetation with signs of significant improvement for the period 2002–2022; the share of vegetation in the basin was about 27% and increased to 88%. An increase in the level of soil moisture and a decrease in thermal stress confirmed this trend. SMI, TCI, and TVDI data indicate positive dynamics of the study area and the localised occurrence of drought throughout the period. However, the analysis of the last three years reveals signs of fragmentation of vegetation and significant fluctuations in surface temperature. The LULC data for 2026 indicate a decrease in the share of agricultural land to 18% with the transformation of the forest into barren land, an increase in the specific weight of built-up areas both fragmented and contiguous, and a decrease in the share of fallow land. The analysis showed that the basin has been experiencing improvement during the period under review and has been relatively free of drought. At the same time, these positive trends are gradually being replaced by signs of stress for the period of study in 2026.