Uncertainty Analysis of SWAT‐Predicted Stream Flow in Dinevar Watershed Under CMIP6 Climate Scenarios

ABSTRACT This study evaluated the impacts of climate change on surface flow in the Dinevar Basin, Iran, via the SWAT model. It was calibrated (1986–2004) and validated (2005–2012) via monthly streamflow data from two hydrometric stations and exhibited satisfactory performance (NSE > 0.6). The model was forced with bias‐corrected data from five CMIP6 climate models under the SSP1‐2.6 and SSP5‐8.5 scenarios for the near‐future (2020–2060) and far‐future (2061–2100) periods. Sensitivity analysis revealed that 17 key parameters were the most influential. Future climate projections revealed high variability in precipitation patterns among different GCMs, whereas all the models consistently projected increased maximum temperatures, particularly under SSP5‐8.5. Runoff simulations revealed substantial uncertainty among GCMs, with multimodel annual averages ranging from 0.96 to 43.8 m 3 /s. Most models predicted an overall reduction in annual runoff, with significant decreases from March to May. Uncertainty assessment using a performance‐weighted hybrid multimodel ensemble indicated that increased average monthly flow is projected only in November for both scenarios and future periods, whereas decreases are expected in all other months. This highlights a critical shift in the hydrological regime towards concentrated flow in late autumn, presenting substantial challenges for sustainable water resource management under changing climatic conditions.