A mixed approach for urban flood prediction using Machine Learning and GIS

Abstract

Extreme weather conditions, as one of many effects of climate change, is expected to increase the magnitude and frequency of environmental disasters. In parallel, urban centres are also expected to grow significantly in the next years, making necessary to implement the adequate mechanisms to tackle such threats, more specifically flooding. This project aims to develop a flood prediction system using a combination of Machine Learning classifiers along with GIS techniques to be used as an effective tool for urban management and resilience planning. This approach can establish sensible factors and risk indices for the occurrence of floods at the city level, which could be instrumental for outlining a long-term strategy for Smart Cities. The most performant Machine Learning model was a Random Forest, with a Matthew's Correlation Coefficient of 0.77 and an Accuracy of 0.96. To support and extend the capabilities of the Machine Learning model, a GIS model was developed to find areas with higher likelihood of being flooded under critical weather conditions. Therefore, hot spots were defined for the entire city given the observed flood history. The scores obtained from the Random Forest model and the Hot Spot analysis were then combined to create a flood risk index.