Development of a gis supported automatic road closure model for buildings at risk of collapse and its integration with network analysis

In this thesis study, a Geographic Information Systems (GIS)-supported Automatic Road Closure Detection Model was developed to predict road closures caused by building collapses after earthquakes, and this model was integrated with network analysis applications. The model identifies roads that are directly closed due to debris spread from collapsed buildings; it can also consider roads with a passage width below 3,5 meters as closed, depending on user preference, and Monte Carlo Simulation (MCS) can be applied to reduce uncertainties. The model can evaluate directly closed roads and can use the 3,5 m condition and/or MCS options individually or in combination. Accuracy performance was tested under three scenarios: (i) directly closed roads and MCS, (ii) directly closed roads and the 3,5 m condition, (iii) directly closed roads, the 3,5 m condition, and MCS. Eight formulas based on building height were tested in calculating debris spread. These tests were conducted considering the roads closed in Elbistan during the February 6, 2023 Kahramanmaraş earthquake. The analyses revealed that the H/2 formula yielded the best results, with F1 scores determined as 0,77, 0,76, and 0,77, respectively. The model was integrated into the network analysis in the Tepebaşı district of Eskişehir. Within this scope, the roads that would be closed due to the collapse of 71 risky buildings were determined. It was observed that these closures would affect the travel time to hospitals and emergency health stations, approximately doubling the access time.