Flood Risk Models

Objective

Implements multiple data-driven candidate models to predict the occurrence and severity of flood incidents along the Red River of the North, a historical trade route between MN and ND prone to major flash floods. The criteria used to assess model's optimality is the maximization of the true positive rate (TPR) of flood detection. All scripts are adapted to take in data on site conditions sourced from the United States Geological Survey and spans 2007-2019. Each model sets Gage Height (water surface elevation) as the target variable.