Integrating HYDRUS-2D and AquaCrop for simulating soil water–salt dynamics and crop growth under film-mulched drip irrigation

CONTEXT Sustainable agricultural production in arid regions is constrained by water scarcity and soil salinization. Effective irrigation and salinity management requires integrated understanding of soil water–salt dynamics and crop responses, yet existing modeling approaches often treat two-dimensional soil processes and dynamic crop growth separately, limiting their utility for field-scale management analysis. OBJECTIVE This study aims to develop and validate a coupled modeling framework that integrates two-dimensional soil water and solute transport with crop growth simulation, in order to improve the representation of soil–crop interactions under film-mulched drip irrigation and to provide a quantitative basis for irrigation and salinity management. METHODS A HYDRUS-AquaCrop-Coupled (HAC) model was constructed by externally linking the HYDRUS-2D model (water flow and solute transport) and the AquaCrop model (water-driven crop growth) via a central coupler written in Python. The two models operate on synchronized daily time steps with bidirectional data exchange: AquaCrop supplies daily potential transpiration, soil evaporation, and canopy cover to HYDRUS-2D, while HYDRUS-2D returns soil water content and salinity profiles for use as initial conditions in the subsequent AquaCrop time step. The coupled model was calibrated and validated using two-year field data, encompassing multiple irrigation and initial salinity treatments. Twenty-five scenarios combining five irrigation levels and five initial salinity levels were simulated to examine crop yield and biomass responses. RESULTS AND CONCLUSIONS The HAC model outperformed the standalone AquaCrop model in reproducing soil water content and salinity, with average RMSE improvements of 27.6% and 18.36%, respectively, while maintaining good agreement with observed canopy cover, aboveground biomass, and crop yield ( R 2 > 0.89, d > 0.81). Scenario simulations for the tested site indicated that the yield remained near its maximum when initial ECe ≤ 7.7 dS·m −1 and declined beyond this level. Model-derived, site-specific stage thresholds increased from 8.46 dS·m −1 at the seedling stage to 12.22 dS·m −1 at the flower boll stage. Under non-saline conditions, increasing irrigation amount within the tested range accelerated attainment of maximum yield but provided limited additional yield benefit. SIGNIFICANCE This study demonstrates a practical external coupling methodology that leverages the complementary strengths of a mechanistic soil model and a parsimonious crop model without requiring extensive source code modification. The model application here is restricted to the specific experimental site of the study region. To enable broader extrapolation, additional evaluations across varied soil types, climatic zones and cropping systems are necessary. Such follow-up work can quantify irrigation–salinity–yield trade-offs in water-limited salinized regions and enhance the generalizability of the proposed coupling framework.