Extreme value statistics for analysing simulated environmental extremes

Henry Elsom; Matthew Pawley

doi:10.1007/s10687-024-00502-3

Extreme value statistics for analysing simulated environmental extremes

Research output: Contribution to journal › Article › peer-review

Abstract

We present the methods employed by team ‘Uniofbathtopia’ as part of a competition organised for the 13th International Conference on Extreme Value Analysis (EVA2023), including our winning entry for the third sub-challenge. Our approaches unite ideas from extreme value theory, which provides a statistical framework for the estimation of probabilities/return levels associated with rare events, with techniques from unsupervised statistical learning, such as clustering and support identification. The methods are demonstrated on the data provided for the EVA (2023) Conference Data Challenge – environmental data sampled from the fantasy country of ‘Utopia’ – but the underlying assumptions and frameworks should apply in more general settings and applications.

Original language	English
Pages (from-to)	47-73
Journal	Extremes
Volume	28
Early online date	9 Nov 2024
DOIs	https://doi.org/10.1007/s10687-024-00502-3
Publication status	Published - 31 Mar 2025

Data Availability Statement

The data for the EVA (2023) Conference Data Challenge has been made publiclyavailable by Rohrbeck et al. (2024+). The code used for Challenges 3 and 4 is available at https://github.com/pawleymatthew/eva-2023-data-challenge.

Acknowledgements

The authors are thankful to Christian Rohrbeck, Emma Simpson and Jonathan Tawn of the University of Bath, University College London, and Lancaster University, respectively, for organising the challenges that inspired this work.

Funding

Matthew Pawley is supported by a scholarship from the EPSRC Centre for Doctoral Training in Statistical Applied Mathematics at Bath (SAMBa), under the project EP/S022945/1. Henry Elsom is supported by the University Research Studentship Award (URSA) from the University of Bath.

Funders	Funder number
EPSRC Centre for Doctoral Training in Statistical	EP/S022945/1

Keywords

Bootstrapping
Extreme value analysis
Max-linear model
Sparse projections
Tail pairwise dependence matrix

ASJC Scopus subject areas

Statistics and Probability
Engineering (miscellaneous)
Economics, Econometrics and Finance (miscellaneous)

Access to Document

10.1007/s10687-024-00502-3Licence: CC BY

Cite this

@article{552dff6f85bb4c67850e46dadeb6ce7b,

title = "Extreme value statistics for analysing simulated environmental extremes",

abstract = "We present the methods employed by team {\textquoteleft}Uniofbathtopia{\textquoteright} as part of a competition organised for the 13th International Conference on Extreme Value Analysis (EVA2023), including our winning entry for the third sub-challenge. Our approaches unite ideas from extreme value theory, which provides a statistical framework for the estimation of probabilities/return levels associated with rare events, with techniques from unsupervised statistical learning, such as clustering and support identification. The methods are demonstrated on the data provided for the EVA (2023) Conference Data Challenge – environmental data sampled from the fantasy country of {\textquoteleft}Utopia{\textquoteright} – but the underlying assumptions and frameworks should apply in more general settings and applications.",

keywords = "Bootstrapping, Extreme value analysis, Max-linear model, Sparse projections, Tail pairwise dependence matrix",

author = "Henry Elsom and Matthew Pawley",

year = "2025",

month = mar,

day = "31",

doi = "10.1007/s10687-024-00502-3",

language = "English",

volume = "28",

pages = "47--73",

journal = "Extremes",

issn = "1386-1999",

publisher = "Springer Netherlands",

}

TY - JOUR

T1 - Extreme value statistics for analysing simulated environmental extremes

AU - Elsom, Henry

AU - Pawley, Matthew

PY - 2025/3/31

Y1 - 2025/3/31

N2 - We present the methods employed by team ‘Uniofbathtopia’ as part of a competition organised for the 13th International Conference on Extreme Value Analysis (EVA2023), including our winning entry for the third sub-challenge. Our approaches unite ideas from extreme value theory, which provides a statistical framework for the estimation of probabilities/return levels associated with rare events, with techniques from unsupervised statistical learning, such as clustering and support identification. The methods are demonstrated on the data provided for the EVA (2023) Conference Data Challenge – environmental data sampled from the fantasy country of ‘Utopia’ – but the underlying assumptions and frameworks should apply in more general settings and applications.

AB - We present the methods employed by team ‘Uniofbathtopia’ as part of a competition organised for the 13th International Conference on Extreme Value Analysis (EVA2023), including our winning entry for the third sub-challenge. Our approaches unite ideas from extreme value theory, which provides a statistical framework for the estimation of probabilities/return levels associated with rare events, with techniques from unsupervised statistical learning, such as clustering and support identification. The methods are demonstrated on the data provided for the EVA (2023) Conference Data Challenge – environmental data sampled from the fantasy country of ‘Utopia’ – but the underlying assumptions and frameworks should apply in more general settings and applications.

KW - Bootstrapping

KW - Extreme value analysis

KW - Max-linear model

KW - Sparse projections

KW - Tail pairwise dependence matrix

UR - http://www.scopus.com/inward/record.url?scp=85208782614&partnerID=8YFLogxK

U2 - 10.1007/s10687-024-00502-3

DO - 10.1007/s10687-024-00502-3

M3 - Article

SN - 1386-1999

VL - 28

SP - 47

EP - 73

JO - Extremes

JF - Extremes

ER -

Extreme value statistics for analysing simulated environmental extremes

Abstract

Data Availability Statement

Acknowledgements

Funding

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this