Bioclimatic constraints and edaphic preferences of wheat: implications for environmental suitability forecasting under climate change

Background. Understanding how environmental factors influence the spatial suita­bility of wheat is critical for sustaining productivity under climate change. In regions like Ukrainian Polissia and the Forest-Steppe, where climatic and soil gradients are strong, changes in agroecological conditions may substantially affect cultivation potential. While global studies exist, regional assessments that integrate both climate and soil data remain limited. Identifying key environmental drivers and their response patterns supports targeted adaptation and land use planning, helping ensure food security in a changing climate. Materials and Methods. The spatial suitability of wheat cultivation in the Polissia and Forest-Steppe regions of Ukraine was assessed using agroecological modelling. We compiled a dataset of observed wheat cover from official agricultural statistics. The environmental predictors included 19 bioclimatic variables (WorldClim), soil properties (texture, pH, and organic matter content), and topographic factors. Multicollinearity was reduced via principal component analysis and correlation filtering. Four modelling approaches: ordinary least squares (OLS), ridge regression, generalised additive models (GAM), and random forest (RF), were applied to identify key predictors and response patterns. Results and Discussion. Among the tested models, random forest provided the highest accuracy, followed by GAM and ridge regression, while OLS lagged behind. Key predictors of wheat suitability included warm-quarter temperature (bio10), growing seasonal precipitation, and soil factors, such as pH, clay content, and bulk density. Wheat showed clear sensitivity to high summer temperatures, with response curves revealing nonlinear, bell-shaped patterns indicative of ecological optima. Climate projections suggest a northward shift and fragmentation of suitable areas, especially under SSP3-7.0 and SSP5-8.5 scenarios. While marginal gains are possible short-term, long-term suita­bility is likely to decline in the southern and central zones. These findings underscore the need to integrate climatic and soil data in regional planning and to support adaptation through targeted crop relocation and variety selection. Conclusion. This study demonstrates that the spatial suitability of wheat in Ukraine’s Polissia and Forest-Steppe regions is strongly influenced by both bioclimatic and edaphic factors. Random forest modelling proved the most effective for capturing complex environmental responses. Climate change projections indicate a northward shift and reduction of suitable areas, emphasising the need for adaptive land-use strate­gies. Integrating climate and soil data into agroecological assessments is critical for anticipating risks, guiding crop management decisions, and ensuring long-term food security in vulnerable agricultural landscapes.