Study on the In-Field Water Balance of Direct-Seeded Rice with Various Irrigation Regimes under Arid Climatic Conditions in Egypt Using the AquaCrop Model

Crop growth models are cost-effective and user-friendly tools for decision-makers to develop efficient in-field management strategies. These models are particularly important in countries such as Egypt, where the risk of water scarcity is inevitable. The present study aimed to examine the in-field water balance of direct-seeded rice (Giza 178) under various irrigation regimes and arid conditions during two growing seasons (2019 and 2020). Four irrigation regimes, namely, continuous flood irrigation with a fixed water depth of 5 cm, and 3-, 6-, and 10-day irrigation frequencies (FI, 3IF, 6IF, and 10IF, respectively), were arranged in a randomized complete block design with three replicates. Then, the feasibility of using AquaCrop in simulating direct-seeded rice development and in-field water balance was assessed. Five statistical indicators, including normalized root-mean-squared error (NRMSE), index of agreement (d), coefficient of determination (R-2), Nash-Sutcliffe efficiency coefficient (EF), and percent deviation (P-d), were used to evaluate the performance of AquaCrop. The field trial results demonstrated that both the 3IF and 6IF irrigation regimes were the best for achieving the highest biomass (21.0 t center dot ha(-1), under 3IF), yield (9.8 t center dot ha(-1), under 3IF), and saving irrigation water (18.3-22.4%, under 6IF), making them the best to apply in Egypt. Moreover, the AquaCrop simulation results showed a good correlation between the observed and simulated rice yield (Y) in both seasons (R-2 = 0.99 and 0.98 in 2019 and 2020, respectively). AquaCrop showed excellent performance in simulating canopy cover (CC) and biomass (B) during both growing seasons (5.0 <= NRMSE <= 15.0, 0.97 <= d <= 0.99, 0.92 <= R-2 <= 0.99, and 0.92 <= EF <= 0.99). In addition, the model showed acceptable performance in simulating in-field water balance components. Reasonably good model efficiency was recorded in simulating crop actual evapotranspiration (ETact). Meanwhile, the average P-d for percolation (P) was between -15.3% and 5.4% during both growing seasons. Overall, AquaCrop showed adequate accuracy in simulating CC, B, Y, ETact, and P but relatively low efficiency in simulating ETact and P under severe water scarcity. Therefore, AquaCrop may serve as a valuable tool for irrigation management and crop yield prediction even in arid regions, such as Egypt.