Flood mapping based on machine learning: a lexicometric analysis

Environmental changes occurring on the planet contribute to disasters that result in loss of life and severe socioeconomic problems. In urban areas, the increasing imperviousness of the soil leads to a notable rise in stormwater runoff and the frequency of floods and inundations, causing numerous physical, social, and financial issues. Studies have aimed to map flood and inundation susceptibility using various methodologies. Here, we identify that the use of Machine Learning (ML) methodologies in flood studies requires a large variety of input factors to drive its processes. Thus, this study reviews the reference literature to organize, classify, and qualify selected articles, and then perform a lexicometric analyses. The collected peer-reviewed articles address the production of susceptibility maps for flooding in urban areas using ML methodologies. Through this process, it was possible to identify current theoretical and methodological approaches, as well as to understand the state of the art in flood and urban flooding studies.