Abstract
Climate change and water scarcity have put food security and sustainable development in arid regions at risk. Irrigation based actual evapotranspiration (ETc) has recently been added as a new tool in the Decision Support System for Agrotechnology Transfer (DSSAT) models and might improve irrigation water management, thus more research is needed. For this purpose, three Wheat models (CERES, CROPSIM and N-Wheat) in the latest version of DSSAT (v. 4.7.5) were calibrated and evaluated using experimental field data across three growing seasons. Field data included irrigation by different fractions of ETc as 80%, 100% and 120%. The calibrated models were then employed to predict wheat grain yield (GY), biomass yield (BY), irrigation, evapotranspiration, water use efficiency-based evapotranspiration (WUE_ET), and water use efficiency-based irrigation (WUE_Irri) for 10 locations represent Nile delta in long term simulation (1991-2020). The models showed robust simulations of ETc compared to observed values under all corresponding treatments, demonstrating high calibration accuracy and the ability to predict yield and water for other locations in the long term. Simulation treatments included automatic irrigation with different fractions of 50%, 60%, 70%, 80%, 90% and 100% from ETc. Hereinafter, the simulated GY and WUE_ET were compared with those obtained by farmers in all locations to specify the recommended treatment achieving higher yield and water productivity. In all locations, simulated GY and BY ranged (4000-9000 kg ha-1), and (10,500-18,000 kg ha-1), respectively with associated uncertainty between treatments and locations. Averaged over ten locations, and 30 years, the simulated GY under full irrigation treatment (100% ETc), showed the superiority with an increase of 27.5%, 13.0%, 5.0%, 1.5%, and 0.4% relative to irrigation with 50%, 60%, 70%, 80%, and 90% ETc, respectively. Deficit irrigation-based ET decreased WUE_Irri, whilst increased WUE_ET, achieving the higher value (20.0 kg ha-1 mm-1) with irrigation based 90% ETc. However, deficit irrigation with 90% ETc (I5) produced higher WUE values than full irrigation (100% ETc), with increases of 0.08% and 10.6% for WUE_ET and WUE_irri, respectively. Comparing simulated GY and WUE_ET with farmers values in all locations, simulated values under irrigation based 90% ETc increased by 1.7% and 63%, respectively, confirming the importance of irrigation scheduling based 90% ETc in maximizing wheat yield and water productivity in arid regions.