A nonstationary framework for hydrological drought assessment in Iran

Study region: The Halil-Rud Basin in Iran, a semi-arid watershed, has been increasingly affected by climate change over the past decades, impacting hydrological processes and water resource Study focus: This study introduces a Nonstationary Standardized Runoff Index (NSRI) to enhance hydrological drought assessment in the Halil-Rud Basin for the period 1980–2019. Using the GAMLSS framework with a time-varying gamma distribution, NSRI models incorporate linear relationships between runoff and hydroclimatic covariates, including precipitation, temperature, potential evapotranspiration, and antecedent runoff. Four models—one stationary (M0) and three nonstationary (M1–M3)—were evaluated across 14 covariate combinations using monthly runoff data from three stations (Pole-Baft, Meidan, Kenaroyeh), comparing drought severity, duration, and intensity. New hydrological insights for the region: Results indicate that nonstationary models consistently outperform the stationary baseline, with M2 (temperature and antecedent runoff) providing the best fit. Analysis of S/NS drought indices along the Halil-Rud Basin reveals that NSRI more accurately captures spatiotemporal drought variability, especially in downstream regions affected by anthropogenic influences. Compared to SSRI, NSRI moderates extreme drought estimates, highlighting the risk of overestimation when using stationary assumptions. These findings demonstrate the value of nonstationary modeling for robust drought monitoring and adaptive water resource management in semi-arid regions.