{"id":{"repo_id":"exeter","oai_identifier":"oai:figshare.com:article/32446578"},"canonical_url":"https://search.dev.ndltd.org/etd/exeter/oai:figshare.com:article/32446578","repository":{"repo_id":"exeter","name":"University of Exeter","base_url":"https://api.figshare.com/v2/oai"},"display":{"title":"Collective risk modelling of multivariate compound events: correlation of aggregate wind and precipitation in European extratropical cyclones","abstract":"Extratropical cyclones cause both extreme wind and precipitation in Europe, which can result in substantial aggregate insurance losses. Understanding the relationship between these aggregates can potentially aid insurers in their portfolio management. This thesis uses annually aggregated wind and precipitation severity indices as proxies for yearly insured losses from windstorm and flood. The compound nature of cyclones means each event can contribute to yearly losses for both hazards. As such, aggregation over the calendar year was chosen to align with the duration of reinsurance contracts, so risk managers can understand the relationship between the hazards that cause losses to their portfolios. The proxy aggregate severity indices are derived by tracking cyclones in reanalysis from 1980-2020. The sample correlation between these indices is found to decrease with increasing severity threshold. Positive correlation occurs at low thresholds and decreases towards small negative correlations over land for the most extreme events. Six random sum collective risk frameworks of varying complexity have been explored for modelling the correlation between the aggregate indices. Four of these frameworks can capture the sample correlation across different thresholds at both individual locations and over larger national regions. The frameworks are used to decompose which factors drive the sample correlation. The clustering of events in time induces positive correlation while the between year wind-precipitation association plays a dominant role in the correlation decreasing with threshold. Annual mean storm duration is identified as a potential inter-annual modulator of the relationship between yearly mean wind and precipitation severity. Mean storm duration has a positive correlation with annual mean precipitation, but a negative correlation with annual mean windspeeds. Storm duration is found to be related to large scale modes of climate variability: negative correlation with the North Atlantic Oscillation over the UK, and positive correlation with the Scandinavian pattern over central Europe.<p></p>","abstract_html":"Extratropical cyclones cause both extreme wind and precipitation in Europe, which can result in substantial aggregate insurance losses. Understanding the relationship between these aggregates can potentially aid insurers in their portfolio management. This thesis uses annually aggregated wind and precipitation severity indices as proxies for yearly insured losses from windstorm and flood. The compound nature of cyclones means each event can contribute to yearly losses for both hazards. As such, aggregation over the calendar year was chosen to align with the duration of reinsurance contracts, so risk managers can understand the relationship between the hazards that cause losses to their portfolios. The proxy aggregate severity indices are derived by tracking cyclones in reanalysis from 1980-2020. The sample correlation between these indices is found to decrease with increasing severity threshold. Positive correlation occurs at low thresholds and decreases towards small negative correlations over land for the most extreme events. Six random sum collective risk frameworks of varying complexity have been explored for modelling the correlation between the aggregate indices. Four of these frameworks can capture the sample correlation across different thresholds at both individual locations and over larger national regions. The frameworks are used to decompose which factors drive the sample correlation. The clustering of events in time induces positive correlation while the between year wind-precipitation association plays a dominant role in the correlation decreasing with threshold. Annual mean storm duration is identified as a potential inter-annual modulator of the relationship between yearly mean wind and precipitation severity. Mean storm duration has a positive correlation with annual mean precipitation, but a negative correlation with annual mean windspeeds. Storm duration is found to be related to large scale modes of climate variability: negative correlation with the North Atlantic Oscillation over the UK, and positive correlation with the Scandinavian pattern over central Europe.&lt;p&gt;&lt;/p&gt;","abstract_has_math":false,"creators":["Toby Jones (21041453)"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2026,"date_issued":"2026-05-27T00:00:00Z","date_published":"2026-05-27T00:00:00Z","updated_at":"2026-07-27T19:32:54Z","subjects":["extratropical cyclones","windstorms","statistics","random sums","correlations","aggregate risk","collective risk"],"languages":[],"rights":["All rights reserved"],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier","label":"Identifier","values":["10779/exe.32446578.v1"],"render_values":[{"text":"10779/exe.32446578.v1","href":null,"code":true}]}]},"links":{"outbound_url":null,"outbound_label":null,"outbound_source":null},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:creator","label":"Author","values":["Toby Jones (21041453)"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2026-05-27T00:00:00Z"]},{"key":"dc:relation","label":"Dc Relation","values":["https://figshare.com/articles/thesis/Collective_risk_modelling_of_multivariate_compound_events_correlation_of_aggregate_wind_and_precipitation_in_European_extratropical_cyclones/32446578"]},{"key":"dc:type","label":"Dc Type","values":["Text","Thesis"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["extratropical cyclones","windstorms","statistics","random sums","correlations","aggregate risk","collective risk"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:rights","label":"Dc Rights","values":["All rights reserved"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["10779/exe.32446578.v1"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Extratropical cyclones cause both extreme wind and precipitation in Europe, which can result in substantial aggregate insurance losses. Understanding the relationship between these aggregates can potentially aid insurers in their portfolio management. This thesis uses annually aggregated wind and precipitation severity indices as proxies for yearly insured losses from windstorm and flood. The compound nature of cyclones means each event can contribute to yearly losses for both hazards. As such, aggregation over the calendar year was chosen to align with the duration of reinsurance contracts, so risk managers can understand the relationship between the hazards that cause losses to their portfolios. The proxy aggregate severity indices are derived by tracking cyclones in reanalysis from 1980-2020. The sample correlation between these indices is found to decrease with increasing severity threshold. Positive correlation occurs at low thresholds and decreases towards small negative correlations over land for the most extreme events. Six random sum collective risk frameworks of varying complexity have been explored for modelling the correlation between the aggregate indices. Four of these frameworks can capture the sample correlation across different thresholds at both individual locations and over larger national regions. The frameworks are used to decompose which factors drive the sample correlation. The clustering of events in time induces positive correlation while the between year wind-precipitation association plays a dominant role in the correlation decreasing with threshold. Annual mean storm duration is identified as a potential inter-annual modulator of the relationship between yearly mean wind and precipitation severity. Mean storm duration has a positive correlation with annual mean precipitation, but a negative correlation with annual mean windspeeds. Storm duration is found to be related to large scale modes of climate variability: negative correlation with the North Atlantic Oscillation over the UK, and positive correlation with the Scandinavian pattern over central Europe.<p></p>"]},{"key":"dc:title","label":"Title","values":["Collective risk modelling of multivariate compound events: correlation of aggregate wind and precipitation in European extratropical cyclones"]}]}],"canonical_facts":{"dc:creator":["Toby Jones (21041453)"],"dc:date":["2026-05-27T00:00:00Z"],"dc:description":["Extratropical cyclones cause both extreme wind and precipitation in Europe, which can result in substantial aggregate insurance losses. Understanding the relationship between these aggregates can potentially aid insurers in their portfolio management. This thesis uses annually aggregated wind and precipitation severity indices as proxies for yearly insured losses from windstorm and flood. The compound nature of cyclones means each event can contribute to yearly losses for both hazards. As such, aggregation over the calendar year was chosen to align with the duration of reinsurance contracts, so risk managers can understand the relationship between the hazards that cause losses to their portfolios. The proxy aggregate severity indices are derived by tracking cyclones in reanalysis from 1980-2020. The sample correlation between these indices is found to decrease with increasing severity threshold. Positive correlation occurs at low thresholds and decreases towards small negative correlations over land for the most extreme events. Six random sum collective risk frameworks of varying complexity have been explored for modelling the correlation between the aggregate indices. Four of these frameworks can capture the sample correlation across different thresholds at both individual locations and over larger national regions. The frameworks are used to decompose which factors drive the sample correlation. The clustering of events in time induces positive correlation while the between year wind-precipitation association plays a dominant role in the correlation decreasing with threshold. Annual mean storm duration is identified as a potential inter-annual modulator of the relationship between yearly mean wind and precipitation severity. Mean storm duration has a positive correlation with annual mean precipitation, but a negative correlation with annual mean windspeeds. Storm duration is found to be related to large scale modes of climate variability: negative correlation with the North Atlantic Oscillation over the UK, and positive correlation with the Scandinavian pattern over central Europe.<p></p>"],"dc:identifier":["10779/exe.32446578.v1"],"dc:relation":["https://figshare.com/articles/thesis/Collective_risk_modelling_of_multivariate_compound_events_correlation_of_aggregate_wind_and_precipitation_in_European_extratropical_cyclones/32446578"],"dc:rights":["All rights reserved"],"dc:subject":["extratropical cyclones","windstorms","statistics","random sums","correlations","aggregate risk","collective risk"],"dc:title":["Collective risk modelling of multivariate compound events: correlation of aggregate wind and precipitation in European extratropical cyclones"],"dc:type":["Text","Thesis"]},"updated_at":"2026-07-27T19:32:54Z"}