Spatio-Temporal Evolution and Driving Mechanisms of Linear Heritage Landscape Fragmentation Based on Heritage-Corrected Datasets: A Case Study of Nested Cumulative Buffer Zones of the Yangzhou Ancient Grand Canal

Urbanization-driven landscape fragmentation undermines the spatial continuity and cultural integrity of linear canal heritage. Traditional land-cover datasets often group heritage patches with general land-use classes, leading to systematic bias in fragmentation evaluation. Furthermore, analyses using a single buffer boundary fail to capture scale-dependent gradient changes along heritage corridors. Taking the Yangzhou Ancient Grand Canal, a representative living segment of the Grand Canal World Heritage Site, this study developed a heritage-corrected seven-class land-use raster dataset by defining heritage areas as an independent land-use category. Using multi-temporal datasets covering 2000–2025, we created three nested cumulative buffer zones (0–200 m, 0–1000 m, and 0–2000 m). Six landscape-fragmentation metrics were calculated in Fragstats 4.2. We constructed the comprehensive heritage-landscape fragmentation index (CHFI) via min–max normalization and the entropy-weight method. The Geodetector model was then applied to quantify both individual-factor explanatory power and interactive driving effects across multiple buffer scales. Our results identify 2010 as a critical turning point for fragmentation dynamics: fragmentation was partly reduced before 2010 but increased steadily afterwards. Fragmentation risks concentrate within the 0–200 m near-shore core zone; the composite index is diluted as the buffer expands from 0–200 m to 0–1000 m but stabilises thereafter. Fragmentation follows a dual-core driving pattern shaped by vegetation and impervious-surface conditions; fractional vegetation cover represents the most consistent dominant factor across all scales and time periods. Near-shore zones are jointly influenced by elevation and distance-to-river gradients, while outer buffer zones are mainly controlled by vegetation conditions. All factor interactions produce enhanced explanatory power. After 2010, near-shore areas underwent a phase shift toward strong multi-factor coupling. Fragmentation hotspots gradually became decoupled from heritage-site locations, suggesting heritage-related pressures act largely through combined factor effects rather than the direct influence of heritage sites themselves. This research offers a refined analytical framework for urban canal heritage and supports evidence-informed conservation and living-heritage revitalization.