Assessment and Monitoring of Drought Crisis Impacts Using Spatial Data Mining with an Emphasis on the Lake Urmia Basin

Iran, situated in the arid geographical belt and desert strip of the northern hemisphere between latitudes 20° to 40°, experiences an average annual precipitation of 250 mm. Combined with the uneven distribution of rainfall, this places the country among the arid and semi-arid regions of the world. Consequently, addressing the threat of drought and implementing effective management strategies for this potential crisis is unavoidable. The present study aims to identify the most suitable combination of remote sensing drought indices compared to the Standardized Precipitation Index (SPI) using spatial data mining methods and association rules. The Lake Urmia basin was selected as the study area for evaluation. SPI was calculated by interpolating precipitation data from rain gauge and synoptic stations between 2011 and 2022 (1390 to 1401 in the Persian calendar), resulting in raster layers for 1, 3, 6, 9, and 12-month intervals. In parallel, remote sensing indices, including VHI, TCI, VCI, EVI, LST, and NDVI, were derived from MODIS sensor data for the same period. The results indicate a high correlation between remote sensing indices (LST, NDVI, and VCI) and SPI, with reliability exceeding 70%. Furthermore, drought zoning maps derived from the extracted rules reveal that the most severe droughts occurred in the years 2013, 2020, and 2021 (1392, 1399, and 1400).This study underscores the potential of integrating spatial data mining and remote sensing indices for more effective drought assessment and monitoring, particularly in regions prone to severe water stress.