Soil water balance in rice fields: modelling and irrigation management in a Piedmont area

Variations in rainfall and temperature due to climate change, combined with unsustainable water consumption patterns, are affecting water availability. According to the Global Environmental Outlook 2000 of the United Nations Environment Programme (UNEP), water scarcity can be considered as one of the most critical environmental issues of the 21st century and will negatively impact all sectors. Water scarcity and increasing competition for water resources pose critical challenges to agriculture, particularly for water-intensive crops such as rice. In this context, assessing irrigation requirements is essential for the development of effective water-related policies and for optimizing irrigation management, ensuring sustainable water use and preserving agricultural productivity. This thesis develops a Python-based soil water balance model specifically designed for paddy rice, extending an existing high-resolution framework for the assessment of irrigation requirements and adapting it to the conditions of the Piedmont area. In particular the pilot area where the model was applied is the Associazione d’Irrigazione Ovest Sesia (AIOS) district in Piemonte, a 100,000-hectare area where rice covers about 75% of the cultivated land. The model performs a daily soil water balance coupled to vegetation dynamics by integrating climatic, soil, and crop parameters to simulate soil moisture and saturated water level dynamics throughout the growing season based on FAO guidelines. Input data include: i) downscaled ERA5-based climatic data (2.2 km), ii) Soil physical and hydraulic properties from SoilGrids and HIhydrodata, iii) Regional Agricultural Land Use data from the Geoportale Piemonte (cadastral mosaic, SigmaTer project) and iv) local crop and irrigation information from district technicians and previous studies in nearby areas. The irrigation requirements obtained from the model, organized on a spatial grids covering the main rice-cropped area in North Italy, are coupled with GIS analysis to compute irrigation depths and volumes at the district scale. Three irrigation strategies were simulated—Continuous Flooding (CF), Alternate Wetting and Drying (AWD), and Dry Seeding with Delayed Flooding (DFL). Despite some limitations introduced in the water balance and in relation to groundwater, the model produced results consistent with previous studies. Both Alternate wetting-drying and Dry Seeding strategies reduce the district irrigation requirements compared to Continuous Flooding, with the second providing the largest overall water savings. However, Dry seeding increases water demand during critical summer months (June–July) and limits early-season groundwater recharge, which may otherwise support the soil water balance in downstream areas. These findings highlight the model’s reliability and potential for scenario analyses aimed at improving sustainable water management in rice-growing regions.