Constructing and Evaluating Heritage Corridor Networks for Timber Arch Covered Bridges: A MaxEnt–XGBoost–SHAP Evaluation in Northeastern Fujian, China

Fragmented point-based conservation limits the integrated protection and adaptive reuse of timber arch covered bridges. This study develops a quantitative decision-support workflow for identifying and evaluating heritage corridor networks, using 99 timber arch covered bridges in Northeastern Fujian, China, as the regional sample and three representative bridges for bridge-scale interpretation. The workflow combines spatial pattern analysis, Maximum Entropy (MaxEnt) suitability modeling, XGBoost–SHAP-assisted resistance analysis, circuit theory corridor extraction, complex network analysis, and maturity assessment. The results show that the bridges form a significant clustered pattern with a Shouning primary core and a Pingnan secondary core. MaxEnt identified high-suitability areas of 1.857 × 105 ha, accounting for 3.65% of the study area, mainly in the Shouning–Pingnan–Zhenghe border zone. The XGBoost–SHAP procedure was used as a secondary modeling step to translate suitability patterns into a composite resistance surface; it indicated that road proximity, elevation, nighttime light intensity, and distance to points of interest (POIs) were the main variables shaping the model output. Based on 27 selected heritage source sites, 75 primary corridors totaling 4078.22 km were identified, together with 33 pinch points and 27 barrier points. A corridor protection width of 7 km was determined as optimal. Network analysis identified Luanfeng Bridge, Longtan Bridge, and Denglong Bridge as important structural nodes and divided all 99 bridges into five river basin-related communities. Maturity assessment revealed a pattern of “one core, two belts, and multiple nodes” but also considerable functional disparity. The three representative cases are reframed as preliminary interpretations of model results: Luanfeng Bridge emphasizes landscape-integrity and visitor-pressure assessment, Denglong Bridge emphasizes route continuity and low-intervention interpretation, and Qiancheng Bridge emphasizes livelihood coordination and incremental rural tourism. This study provides decision-making support for advancing the integrated conservation of timber arch covered bridges and the synergistic development of rural tourism.