A novel integrated modelling framework to uncover spatial and temporal evolutionary patterns and influence mechanisms...

Rapid and disorderly urban expansion leads to productivity loss, habitat fragmentation, and reduced land marginal returns, hindering sustainable urban development. Scientific identification of land use conflicts (LUCs) and understanding their driving mechanisms are therefore essential for optimizing national territorial spatial patterns. In this study, we focused on Xiong'an New Area, a state-level new area, and identified four levels of conflict zones and sixteen types of LUCs from a "production-living-ecology" spatial perspective. By combining spatial heterogeneity analysis with interpretable machine learning, a novel integrated framework was developed to reveal the spatio-temporal evolutionary patterns of land use conflict intensity (LUCI) and its response to socio-economic and natural factors. Results indicated that the share of medium- and high-conflict zones in Xiong'an New Area increased from 46 % to 74 % between 2010 and 2020, displaying clustered spatial distribution. LUC types were dominated by medium conflict among agricultural-construction-ecological land and high conflict between agricultural and ecological land. LUCI continuously increased, showing a spatial pattern of "high in the north and low in the south." Moreover, the relationships between LUCI and its factors were non-linear and exhibited threshold effects. Gross domestic product, normalized difference vegetation index, and distance from roads were the three most influential factors. Interactions among factors had a non-smooth impact on LUCI. This study offers decision support for optimizing land use patterns and mitigating LUCs in rapidly urbanizing areas.