Assessing the reliability of AquaCrop as a decision-support tool for sustainable crop production

The AquaCrop model has a wide applicability for crop growth simulation under climate change. Nevertheless, its reliability for designing sustainable production system, particularly under environmental stresses, requires to be analyzed in advance. Hence, we carried out a 4-year field-modeling approach to address this issue. This is the first field-modeling study with AquaCrop which assesses the influence of crop type, irrigation water quality, and irrigation strategy and management on the model performance simultaneously. A total of 11 water-saving irrigation treatments were defined by combining two crop types (maize and sunflower), two irrigation water resources (fresh and saline water resources), and three irrigation strategies: full irrigation, deficit irrigation (DI), and partial root-zone drying (PRD). AquaCrop model was calibrated and validated based on the field-collected data during 2010–2011 maize growing seasons and 2014–2015 sunflower growing seasons. The model performance was evaluated based on RMSE, MBE, and RE indices. AquaCrop model provided a reliable estimation of both crop’s yield (RMSE = 0.31–0.57 ton ha−1, MAE = 0.29–0.44 ton ha−1, and RE = 1.08–13.24%), total biomass at harvest (RMSE = 0.87–2.40 t ha−1, MAE = 0.77–2.25 t ha−1, and RE = 1.46–11.75%), and the rooting-zone salinity (RMSE = 0.06–0.98dS m−1, MAE = 0.06–0.70 dS m−1, and RE = 0.03–19.33%). The model was also significantly capable in capturing the temporal variations of the canopy cover (R2 ≥ 0.93) and the soil salinity of the different soil layers (R2 ≥ 0.85). The highest range of uncertainty (RU) was obtained for estimating the rooting-zone salinity (RU = ± 11.58%), which induced the largest uncertainty in simulated crop’s growth parameters (RU = ± 16.56%). Based on the results, the AquaCrop model could be more reliably used for simulating crop growth under freshwater application rather than under saline water application.