Applicability of Site-Level Operational and Environmental Information for Long-Term Heritage Timber Monitoring: Vibration Response and Tilt-Monitoring Baselines at Xuankong Temple

Operational and environmental variability complicate long-term monitoring of publicly accessible heritage timber buildings, while local visitor activity and member-level hygrothermal conditions are difficult to observe continuously. Using field data from Xuankong Temple, this study develops a task-specific framework to evaluate three site-level information sources: opening and closure records for organizing vibration references, entrance counts for estimating open-period vibration responses, and ambient environmental measurements for assessing cross-period tilt baselines. Robustness tests, incremental generalized additive models with date substitution, and time-ordered frozen-model evaluation were used to assess the three tasks. Open-to-closure ratios of three-axis resultant root-mean-square (RMS) acceleration were 13.331, 17.872, and 1.195 at the North Building, Boardwalk, and South Building, respectively; this spatial contrast persisted across robustness tests. Entrance counts improved follow-up log-RMS estimation at all three nodes (ΔR2 = 0.132–0.364), with the clearest original-date advantage at the North Building. Temperature history improved development-period tilt estimation; frozen environmental–tilt models showed no consistent advantage over a fixed mean-tilt reference based on the final seven development days. This framework links site-level information to monitoring decisions, providing an evidence-based approach to organizing vibration references, selecting auxiliary variables, and identifying when finer local observations are needed.