Integration of Geospatial Correlation for Rapid Highway Damage Identification Following Flooding

Transportation infrastructure systems provide operations to support the mobility of goods and people. Following flooding, the blocked road access would prevent people from accessing essential services; thus, it requires efficient restoration. Efficient infrastructure restoration cannot be achieved without rapid damage identification. However, due to the difficulties of road access, only a limited number of damage inspection data are available immediately following flooding. Aimed at reliably predicting infrastructure flood damage status in a timely manner, this research integrates geospatial correlation (i.e., closer things are more alike than things that are farther apart) to address the issues with incomplete damage inspection data. To achieve this, a Bayesian generalized linear geostatistical model was adopted. The model has been demonstrated to provide reliable prediction of highway flood damage with limited damage assessment data. In practice, the specialized model is expected to alleviate the effort and cost associated with infrastructure damage identification and facilitate the rapid execution of restoration operations.