Skip to content

Simulation of extreme functionals in meteoceanic data: application to surge evolution over tidal cycles

Home / Papers / Simulation of extreme functionals in meteoceanic data: application to surge evolution over tidal cycles

Simulation of extreme functionals in meteoceanic data: application to surge evolution over tidal cycles

Authors: Nathan Gorse, Olivier Roustant, Jérémy Rohmer, Déborah Idier Date: 2026-04-28 Paper ID: openalex:2508.13687

Summary

This paper addresses the challenge of simulating extreme meteoceanic events, specifically coastal flooding, which often violate standard functional extreme value theory assumptions due to temporal dependence and short-tailed distributions. The authors propose a two-stage approach that first uses an autoregressive model to remove dependence between tidal cycles, followed by an adaptation of Pareto processes to simulate extreme functional scenarios. The framework is validated against observational surge data from Gâvres, France, using robust statistical and machine learning evaluation metrics.

Key Contributions

  • Introduces a two-stage simulation framework for rare meteoceanic events that accounts for temporal dependence and non-standard tail behavior in tidal surge data.
  • Develops an autoregressive pre-processing step to decorrelate tidal cycles, enabling the application of Pareto-process-based extreme value theory.
  • Demonstrates superior simulation of extreme surge scenarios at the Gâvres site using a multi-criteria validation approach involving PCA and extreme value analysis.

Metadata & Links

url
https://arxiv.org/abs/2508.13687
paper_id
2508.13687
paper_source
openalex
domain
time-series
tags
time-seriesforecastinganomaly-detection
architectures
datasets
concept_slugs
dataset_slugs
skill
TimeSeriesSkill
created_at
2026-05-01T06:04:41Z