Predictive modeling of vulnerability level of French irrigated farms

Climate change has significantly increased the complexity of farm management. This work focuses on water-related risks and aims to highlight the major challenges that French farmers face. Using a mixed-methods approach, with data sourced from two different databases: the first one is an online survey with 111 responses launched in 2024, and the second one is the national database, Farm Accountancy Data Network (FADN), covering 631 irrigated farms between 2020 and 2022. For qualitative data, we used several methods such as graphical representations, response mapping, word clouds, and others. For quantitative data, we employed the machine learning algorithm « Random Forest ». Our findings indicate that farmers perceive climate risks, particularly those related to water, and feel that the risk of drought represents a real threat. Irrigation, as a management tool, appears to be a subject of debate and controversy, especially in areas with water-use conflicts. By examining irrigated farms, we found that each parameter associated with this technique impacts the farms' vulnerability to risk. On a national scale, farmers seem unaware of government initiatives for water resource management. Our goal is to provide valuable insights for policymakers and farmers and inspire them to improve or develop strategies to enhance the resilience of farming systems in the face of climate change.