Monitoring and Modeling Long-term Environmental Influences on a Shallow Excavation Using Machine Learning: Case Study of Tomb TT95 in the Theban Necropolis, Luxor, Egypt

Exposure of rock in thermal cycling has been linked to reduction in intact rock sample strength at a laboratory scale and has been characterized as a potential triggering mechanism for rock falls. However, the long-term effect of climatic fluctuations on shallow rock excavations still remains unknown. This research investigates the long-term effects of climatic fluctuations on shallow rock excavations by monitoring a damaged pillar in TT95, an underground funerary chapel in Theban Necropolis, Luxor, Egypt. Using two orthogonally placed extensometers, relative displacements of the pillar were measured alongside temperature and humidity. Findings indicate that pillar displacements correlate with seasonal temperature changes, showing a 0.02 mm/year drift. Data-driven CNN models were developed for forecasting, with one extensometer model achieving high accuracy (R²: 0.98) and the other performing poorly (R²: -0.2). The study suggests that continued pillar drift could cause loose rock blocks on the chapel ceiling to detach over time.