Crop production in arid and semi-arid regions faces increasing threats from severe weather conditions during the growing season. Irrigation, while critical for mitigating the negative impacts of these conditions, requires datadriven scheduling approaches to optimize productivity and curb excessive withdrawal of limited groundwater resources. This study utilizes the Decision Support System for Agrotechnology Transfer (DSSAT) - Crop Environment Resource Synthesis (CERES) Maize model to evaluate maize yield, water savings, and irrigation water productivity under two irrigation scheduling methods: a conventional soil-moisture-based method and a crop evapotranspiration-based (ETc-based) method in Finney County, Southwestern Kansas, United States. Heat and water stress scenarios were induced by increasing the growing season maximum temperatures by 1 degrees C, 2 degrees C, and 4 degrees C, and extending the duration of dry periods by one day during critical growth stages. Twelve ETc-based treatments, comprising of varying ETc requirement thresholds needed to trigger irrigation (15 mm, 20 mm, 25 mm, and 30 mm), and ETc replacement levels (50 %, 75 %, and 100 %), were simulated over 30 years (1991-2020). The ETc-based treatments were compared to a soil moisture-based strategy which automatically applied a fixed irrigation amount whenever the plant available water in the soil dropped to 50 %. Results indicate that applying 75 % of the required ETc at a 30 mm threshold is the most effective strategy. This approach limited yield losses to 10 %, enhanced water savings by 25 %, and irrigation water productivity by 12 % (mean growing season ETc = 736 mm) when compared to the conventional soil-moisture-based method (mean-growing season ETc = 838 mm) under normal and extreme weather conditions. These findings highlight the potential of ETcbased irrigation to conserve water resources, ensuring more sustainable, climate-resilient agriculture.