Integration of Satellite-Derived Geophysical Information into the Monitoring Framework of Monumental Buildings

Understanding the influence of environmental and soil-related factors on structural response is a key challenge in long-term structural health monitoring (SHM) of heritage structures. Among these factors, soil moisture can play a significant role in soil-structure interactions (SSIs), although on-site measurements are often unavailable. This paper proposes a framework that integrates heterogeneous satellite-derived and on-site data for modelling complex, non-linear relationships between environmental variables and structural response. The methodology is based on relevance vector machines (RVMs) within the sparse Bayesian learning (SBL) paradigm and incorporates latent parameters governing the expansion order of basis functions, enabling automatic selection of informative predictors and promoting model sparsity and interpretability. Satellite-derived soil moisture products are integrated with structural, environmental, and soil measurements to reconstruct monitoring time histories and infer unobserved states influencing structural behaviour. The methodology is validated using long-term monitoring data from an extensively monitored major historic dome. Results demonstrate that satellite-derived information captures soil-related environmental effects and explains variations in structural behaviour. These findings, supported by the large-scale availability of geophysical satellite data, highlight their potential to support SHM ofheritage structures, particularly where direct measurements are limited, improving the interpretation of environmental influences and enabling more efficient and informed monitoring strategies.