Enhancing Databases for AI Assisted Diagnosis and Documentation of Acid-rain-affected Sandstone Facades

Abstract

Three historic sandstone buildings in El-Mina, Lebanon, the Dr. Yaacoub Labban Habitat, the Al Hayek Building 318/12, and the Mameluke commercial structure, show large façade deterioration that conservation specialists attribute to acid rain. We documented cracks penetrating the structural elements, and considerable material loss across all three buildings. Sandstone and rainwater samples were collected from these façades and the area nearby. Together, they provided the foundation for laboratory experiments. The experiments were designed to quantify the relationship between precipitation acidity and material degradation. This study investigates the effect of acid rain on sandstone structures through laboratory simulations, focusing on the relationship between acid rain and crack formation. Our experiments exposed sandstone samples to solutions ranging from pH 3.8 to 7.0. These levels match the acidity we measured in El-Mina’s urban precipitation and identified a critical threshold at pH 4.0; below this level, degradation accelerates. Samples exposed to pH 3.8 solutions lost 9-10% of their mass over the experimental period. Five distinct crack morphologies were identified from the analysis; each type correlated with specific pH exposure conditions and material loss percentages. We documented these patterns through macro photography and digital image segmentation. This created a relational database, which links environmental data with quantitative crack measurements and high-resolution images. This database helps conservation professionals move beyond reactive visual inspections toward predictive maintenance based on environmental monitoring data. The method is reproducible and can be adapted for other heritage sites facing similar environmental challenges.