Apparent Forest‐Cover Stability Conceals Structural and Functional Landscape Reorganization in a Human‐Modified Atlantic Forest–Cerrado Ecotone

Aim To evaluate how long‐term forest turnover reshapes habitat structure and functional connectivity in a human‐modified Atlantic Forest–Cerrado ecotone landscape. Location Sarapuí River Basin, southeastern Brazil. Taxon Native forest vegetation and forest‐dependent taxa of the Atlantic Forest–Cerrado ecotone. Methods Land‐use and land‐cover (LULC) maps from 1986 to 2021 were analysed to quantify forest persistence, loss and regeneration dynamics. Forest‐cover transitions were integrated with structural landscape metrics describing patch area, shape complexity and isolation to evaluate habitat reorganization through time. Functional connectivity was assessed using the graph‐based Probability of Connectivity (PC) metric combined with network analysis, modularity‐based clustering and connectivity surface interpolation to identify connectivity gradients and hotspot subnetworks. Results Matrix intensification was observed throughout the historical series, with pasture areas declining by 33.8% due to the expansion of agriculture and silviculture between 1986 and 2021. Despite these land‐use transitions, total forest cover remained relatively stable at approximately 20% of the basin area. However, persistent forests declined by 28% relative to 1991, while regenerating forests represented approximately 35% of total forest cover by 2021, indicating substantial internal forest turnover and rejuvenation. Structural metrics revealed persistent fragmentation dominated by small patches, accompanied by gradual reductions in inter‐patch distances and increasing aggregation among forest remnants. Connectivity analyses indicated progressive network integration through increasing numbers of functional links and a reduction in subnetworks from 513 to 326. Main Conclusions Apparently stable forest‐cover percentages may conceal substantial internal landscape reorganization processes driven by simultaneous forest loss and regeneration. In human‐modified tropical landscapes, relatively subtle changes in habitat configuration may substantially alter connectivity structure and potential movement dynamics even when total habitat amount changes little. The integration of spatiotemporal forest dynamics with connectivity modelling provides a useful framework for identifying hidden connectivity dynamics and supporting landscape diagnosis and conservation planning in fragmented tropical landscapes.