Modeling and Optimizing Wheat Yield under Climate Variability Using Artificial Intelligent and CropWat: A Comparative Study in Nubariya, Egypt

In addition to CropWat simulations, machine learning (ML) as tools of artificial intelligent (AI) algorithms were employed to enhance predictive accuracy by analyzing non-linear patterns across irrigation systems, genotypes, and planting dates. This study was conducted at the Experimental Research Farm of the National Research Centre (NRC) in El-Nubaria, Beheira Governorate, Egypt, to evaluate the impact of irrigation systems, planting dates (as a climate change proxy), and wheat genotypes on wheat performance. The field experiment involved two irrigation systems (sprinkler and drip), three Egyptian wheat varieties (Misr 1, Sakha 95, and Giza 171), and four sowing dates: the Normal Date of Planting (NDP), and 10, 20, and 30 days after NDP (DAND). The CropWat model was used to simulate biological yield, straw yield, grain yield, harvest index (HI), and water productivity (WP), with results compared to observed field data. Findings indicated that CropWat generally underestimated yields under sprinkler irrigation and overestimated them under drip, especially for Misr 1. Giza 171 showed the closest alignment between simulated and observed results, while Sakha 95 displayed high variability. Delayed planting negatively affected all yield parameters, a trend captured in simulations but with less intensity. HI values were frequently overestimated under stress conditions, and water productivity was inconsistently simulated, especially under later planting dates. Statistical analysis (LSD 0.05) confirmed that many observed-simulated differences were significant, indicating the need for improved model calibration. Conclusion include: (1) prioritizing Giza 171 for its stable performance under climate variability, (2) optimizing drip irrigation for water-use efficiency, and (3) enhancing CropWat’s climate sensitivity and varietal calibration to improve predictive accuracy under changing environmental conditions.