Author(s): Agnieszka Indiana Olbert; Sogol Moradian

Linked Author(s):

Keywords: No Keywords

Abstract: Accurate flood modelling plays a pivotal role in averting flood-related damages. In different studies, the existing guidelines for assessing compound flood risks arising from the interplay between coastal and fluvial factors often fall short in considering their intricate interactions. This omission can lead to an underestimation of the flood risk. This paper introduces a comprehensive approach to compound coastal-fluvial flood mapping, even in scenarios with restricted hydraulic and hydrological data. This approach leverages the power of seven distinct Machine Learning (ML) models. The core of this study revolves around the examination of oceanfluvial floods, taking into account upstream hydrographs in the fluvial section and downstream ocean water level timeseries in coastal areas as pivotal boundary conditions. These ML algorithms were applied to a specific study area in the southwestern part of Ireland, with a focus on generating high-resolution flood maps. The study particularly zoomed in on a significant historical compound ocean-fluvial flood event that occurred on November 19-20,2009. The results obtained in this study substantiate the reliability of ML models in estimating the extent of compound flood inundation. To gauge model performance, a set of skill scores were calculated. This research underscores the ability of ML models to generate coastal-fluvial flood inundation.

DOI:

Year: 2024