Author(s): Diego Urrea Mendez; Dina Gomez Rave; Salvador Navas; Manuel Del Jesus
Linked Author(s): Manuel del Jesus
Keywords: No Keywords
Abstract: The integration of multivariate analysis into the study of compound hydrological events reveals its potential to better understand the complexity of extreme events and generate likely scenarios. Our methodology, divided into two parts, focuses on comparing different multivariate models that capture all the uncertainty and implementing machine learning techniques that enable the simulation of floods considering multivariate scenarios. The analysis underscores the importance of carefully selecting the appropriate multivariate model, as Gaussian models underestimated extreme events, while extreme vine copula models provided more accurate results. This approach contributes to the understanding and effective management of compound hydrological events, benefiting critical infrastructure planning and flood prediction, and opening new perspectives in climate risk management.
Year: 2024