Integration of SEBAL ‐Derived Evapotranspiration With Climate Change Projections to Assess Basin‐Scale Water Resources and Crops Yield

ABSTRACT Assessing the combined impacts of climate change on basin hydrology and crop production is essential for designing robust adaptation strategies in data‐scarce arid regions. The Maroon Basin in southwestern Iran is a major agricultural zone affected by climate change and groundwater depletion. In this study, remote‐sensing‐based evapotranspiration (ET) derived from the Surface Energy Balance Algorithm for Land (SEBAL) was integrated with a deep‐learning Long Short‐Term Memory (LSTM) model and the SWAT hydrological–crop model to assess future water and yield responses under CMIP6 climate scenarios. SEBAL was applied to Landsat imagery to estimate spatially distributed ET for the historical period, and these estimates were used to train an LSTM model driven by daily minimum/maximum air temperature and precipitation. Bias‐corrected CMIP6 projections under SSP1–2.6 and SSP5–8.5 were then used to generate daily ET scenarios for 2025–2070, which, together with the adjusted climate inputs, forced SWAT to simulate runoff and crop yields under future conditions. CMIP6 projections indicate substantial warming under both SSPs and seasonally varying changes in precipitation relative to the 1990–2014 baseline. SWAT simulations suggest mean annual runoff decreases by about 18%–22% and yields of three representative crops (alfalfa, wheat and apple) decline by up to ~40%–50%, with larger reductions under SSP5–8.5. Using SEBAL/LSTM‐based ET instead of the default SWAT PET preserves the negative direction of change but yields systematically smaller magnitudes of runoff and yield reductions, indicating that improved ET representation moderates—but does not eliminate—adverse climate‐change impacts. Yield results should be interpreted as scenario‐based, internally consistent model responses (crop parameters were not tuned to local yield statistics). Additional uncertainty stems from the single GCM selection, the bias‐correction approach, and the LSTM‐based ET emulation. Nevertheless, the proposed SEBAL–LSTM–SWAT framework provides a practical approach for assessing the direction and relative magnitude of climate‐change impacts on water availability and agricultural production in data‐scarce, groundwater‐stressed basins.