Ali Gorjizade, Laleh Divband Hafshejani
Drought monitoring in Iran’s semi-arid regions remains constrained by data scarcity, reanalysis bias, and inconsistent performance of drought indices across temporal scales. This study evaluates the combined effects of bias-corrected reanalysis data and index selection on event-based drought characterization at the Dez Dam station (Khuzestan Province). ERA5-Land precipitation and temperature data were bias-corrected and integrated with station observations. Drought conditions were quantified using SPI, SPEI, and RDI at 3, 6, 9 and 12-month scales. Bias correction substantially improved ERA5-Land precipitation accuracy, reducing mean bias from +0.49 to +0.01 and RMSE from 4.76 to 4.54, while maintaining correlation with observations (r ≈ 0.72). Results reveal pronounced scale dependency among indices. At the 3-month scale, SPI and RDI captured the highest cumulative drought severity and peak intensity, whereas SPEI underestimated drought conditions. At longer scales, RDI showed the greatest cumulative severity (13.2 at 12 months), followed by SPI and SPEI. Run theory identified major drought episodes, notably a 14-month event during 2021-2022 and a significant period spanning 2010-2015, indicating increased hydrological stress, while Mann-Kendall analysis revealed mostly non-significant long-term trends due to high interannual variability. However, event-based analysis confirmed pronounced drought intensification during 2019-2023. Overall, integrating bias-corrected reanalysis data with a multi-index, multi-scale framework enhances drought characterization. The proposed approach provides actionable insights for reservoir-scale drought monitoring and water resources management in data-scarce, semi-arid environments. The primary novelty lies in integrating bias-corrected high-resolution reanalysis data with Run Theory for station-scale, event-based drought characterization tailored to local reservoir management.