A Computationally Efficient Flood Evacuation Planning Tool to Assess the Impacts of Flooding on Transportation Networks.

Floods are a serious natural hazard that may disrupt essential infrastructure, towns, and communities worldwide. As a result, hundreds of lives are lost every year, and floods cause massive economic damage to critical infrastructure (CI). Preparing in advance for possible evacuations can drastically reduce the likelihood of potential deaths and damage to the environment and CIs. However, the absence of a reliable cyber system to define evacuation routes creates considerable delays in the process of evacuation, which in turn slows down disaster preparation and response efforts. Recent advancements in data-driven and machine-learning approaches have given faster and more inventive ways to forecast flooding events, which serve as the primary causes and processes for a large number of issues associated with flood prediction. The purpose of this research is to devise a one-of-a-kind, computationally efficient surrogate model for defining evacuation routes in Low Country, South Carolina, USA. The tool incorporates the distinctive characteristics of machine learning (ML) modeling, transportation geospatial data, and hydraulics and drainage qualities to assess evacuation routes. The system was expertly designed to estimate flood stage levels using ML across USGS gaging stations, combine the findings with the results of Manning's equation, and traffic data, and then integrate all this information into a remote web application. The architecture of the tool is made up of several interchangeable components, such as modules for ML modeling, performance assessment, inundation mapping, and online visualization. The efficacy of  the Flood Evacuation Planning Tool, with a simple conceptual inundation model and a dynamic user interface, is shown by thorough testing and processing measurements. Enhancements to the system that are now being implemented and those that will be implemented in the near future include expanding coverage to areas that are more prone to flooding and boosting the capabilities and accuracy of the tool will be discussed.