INTERPRETABLE AI-BASED FRAMEWORK FOR CLIMATE CHANGEDRIVEN AGRICULTURAL LAND SUITABILITY MAPPING IN EURASIA

Abstract

The increasing severity of climate change poses a major threat to global food security by altering agricultural productivity and land suitability patterns. Traditional modeling approaches often fail to capture the complex, nonlinear interactions between climatic variables and soil characteristics, limiting their ability to provide actionable insights for sustainable land-use planning. To address this challenge, this research proposes an interpretable AI-based framework for assessing and mapping climate changedriven agricultural land suitability across the Eurasian region. The proposed model integrates machine learning algorithms with explainable artificial intelligence (XAI) techniques to predict future land suitability under multiple climate change scenarios. Key environmental parameters such as temperature, precipitation, soil type, topography, and vegetation indices are incorporated to develop a comprehensive dataset representing both current and projected agro-climatic conditions. Models such as Random Forest, XGBoost, and SHAP (SHapley Additive exPlanations) are employed to ensure both high predictive accuracy and interpretability, enabling researchers and policymakers to understand how individual climatic and soil variables contribute to land suitability outcomes. Experimental results demonstrate that the interpretable machine learning framework outperforms conventional statistical methods in predictive performance and transparency, providing spatially explicit insights into potential shifts in suitable agricultural zones across Eurasia. The study highlights regions most vulnerable to climate-induced degradation and those with emerging agricultural potential under future climate projections. This work contributes to the advancement of climate-smart agriculture by offering a transparent, datadriven decision-support system that enhances understanding of climate–agriculture interactions and supports resilient land management strategies for sustainable food production in a changing environment.