Additive functional gauge modelling with multivariate copulas for the assessment of High-Risk Scenarios

Abstract

This paper addresses the analysis of multivariate extreme value data using the geometric approach based on gauge functions, with an additive model proposed by Wadsworth et Campbell (2024). Building on wave data from Newlyn, we revisit this framework using the new gauge function we previously introduced in Bagr\'e \textit{et al.} (2025). A hybrid approach combining additive formulation and Gaussian elements shows a better fit than the pure Gaussian model, highlighting the relevance of our approach given the limitations of Gaussian models for non-normal data as in Nodle (2010)