Global water scarcity poses a significant challenge to agricultural sustainability, particularly in semi-arid regions like Morocco. This study evaluates the benefits of coupling AquaCrop-OSPy with optimization algorithms to improve irrigation scheduling for wheat, aiming to reduce water use while maintaining yield and enhancing water productivity. The optimization framework integrates two methods within AquaCrop-OSPy: (i) a unidimensional optimization based on the minimizeₛcalar bounded algorithm and (ii) a multidimensional optimization using the L -BFGS-B algorithm. This coupling enables the automated adjustment of irrigation thresholds within the model, allowing efficient identification of water-saving yet yield-preserving irrigation schedules. Experiments were conducted over two growing seasons (2016/2017 and 2017/2018) in two wheat fields in the Haouz Plain, Morocco: Field 1 (stressed conditions) and Field 2 (normal conditions). Model calibration using canopy cover, biomass, and actual evapotranspiration (ETa) achieved R² values between 74% and 93%, RMSE values of 0. 09–0. 10 for CC, 0. 15–0. 44 t/ha for biomass, and 0. 52–0. 61 mm/day for ETa. The NSE ranged from 0. 71 to 0. 95, and the Index of agreement (d) ranged from 0. 86 to 0. 98. Validation in the second season confirmed similar performance, demonstrating the reliability of AquaCrop-OSPy for wheat simulation under semi-arid conditions. Three irrigation strategies were evaluated: Real Irrigation (farmers’ practice), NET irrigation, and Soil Moisture Target (SMT). NET irrigation applies a constant soil-water threshold throughout the crop cycle, whereas SMT uses stage-specific thresholds aligned with key phenological phases (emergence, canopy expansion, maximum canopy, and senescence). Optimization produced substantial improvements. In Field 1, optimized SMT reduced irrigation by 26–31% relative to farmers’ practices and 23–31% relative to the pre-optimized SMT baseline, while sustaining yields of 6. 43 t/ha in S1 and 6. 24 t/ha in S2 and increasing WP Irrigation to 2. 08–2. 33 kg/m³. In Field 2, optimized SMT reduced irrigation by 42–45% compared to farmers (23–31% relative to pre-optimization), while maintaining yields of 6. 43 t/ha in S1 and 6. 24 t/ha in S2, with similarly high water productivity. Optimized NET achieved 4–34% savings relative to farmers (10–11% relative to pre-optimization), with moderate gains in WP Irrigation. These results demonstrate the clear advantage of SMT over NET in achieving a more favorable water–yield balance. This study confirms that coupling AquaCrop-OSPy with optimization algorithms can substantially improve irrigation scheduling and WP, providing a practical tool for sustainable crop management in water-scarce regions. • AquaCrop-OSPy was coupled with gradient-based optimizers to automatically optimize irrigation for drip-irrigated wheat. • The dynamic SMT irrigation strategy substantially reduced irrigation water use while preserving yield. • Optimization irrigation water productivity by aligning irrigation events with crop water demand across key growth stages.
Kaissi et al. (Tue,) studied this question.