Events

Maxime is a risk scientist at Risk Frontiers and helps develop and manage Risk Frontiers’ suite of models. He holds a PhD in quantitative Marine Science from the University of Tasmania. His research investigated the drivers and impacts of marine heatwaves.
Maxime’s role within Risk Frontiers focuses on Risk Frontiers’ Cyclone model CyclAUS and implementing a new machine learning based storm surge hazard component. He also uses his strong background in extreme and climate science to improve the development of other models including Flood, Hail and assist the ClimateGLOBE project.

In the context of global sea level rise, coastal flooding losses due to storm surge events are a growing threat for Australians. Such events are commonly observed during tropical cyclones (TCs). Although Australia is baring down a hot and dry El Nino summer, Australian coastlines are still at risk. The loss associated with storm surge during a TC is often ignored in traditional TC Catastrophe-Loss models due to the complexity of modelling extreme sea level events in Monte-Carlo simulation infrastructures. Here, we present a machine learning modelling approach to include storm surge risk within Risk Frontiers’ TC loss model – CyclAUS. Storm surge water levels are modelled using Convolutional Neural Networks trained on a coastal sea level reanalysis dataset, spanning the Australian coastline at a 25 km resolution. Outputs from CyclAUS, simulating 50000 years of current climatology TCs around Australia, are then fed to point-based individual storm surge models to extract a maximum sea level height. Storm surge model validation shows that our machine learning approach has comparable skills to hydrodynamic solutions, while demanding far fewer computational resources, thus enabling its implementation into stochastic CAT-loss models.