LAUSR

The AquaCrop model is crucial for assessing innovative irrigation strategies to increase water productivity, yet its reliability over multiple seasons in semi‐arid Ethiopia is underexplored. Field experiments were conducted in 2020/21 and 2021/22 with four irrigation treatments, that is, 100%, 80%, 60% and 40% ETc, to evaluate the model's performance in simulating maize growth, yield, evapotranspiration and water productivity. The 2020/21 data were used for calibration, and the 2021/22 data were used for validation. The results indicated that the model underestimated canopy cover (NRMSE < 25%) but accurately simulated biomass (NRMSE < 10%). Very good simulations were achieved for grain yield and water productivity (NRMSE < 10 tons ha−1, NSE > 0.75), whereas crop evapotranspiration was underestimated (NRMSE < 15%). The model was more reliable for the 100% and 80% ETc treatments than for the 60% and 40% ETc treatments. The 60% ETc treatment yielded an optimal value of 6.1–7.0 tons ha−1, with a water productivity ranging from 2.4 to 2.5 kg m−3. These findings demonstrate that AquaCrop can effectively predict maize growth and yield, with 60% ETc as the optimal irrigation level for water‐scarce regions such as Arba Minch.