Discussing windstorms at the Royal Statistical Society international conference

JBA Marginal standardisation - Modelling spatial extent - Paul Sharkey report
Marginal standardisation – modelling spatial extent

Paul Sharkey, Statistician, is perfecting his talk for the 2018 Royal Statistical Society international conference which is taking place from 3-6 September.

Following a paper that Paul wrote with peers from Lancaster University and the Met Office, he will be discussing the impact of windstorms on the United Kingdom and northern Europe climate.

“It is an exciting prospect to share the outputs of our research with peers from academia and industry at the RSS conference, as we believe these tools could be useful for practitioners for assessing the risk associated with extreme windstorm events” commented Paul.

He added, “Working in a team of researchers from statistics and meteorology was an incredibly valuable experience that prepared me well for the multi-disciplinary collaborations I’ve been involved with in my time at JBA so far.”

Modelling the spatial extent and severity of European windstorms

Windstorms are a primary natural hazard affecting Europe that are commonly linked to substantial property and infrastructural damage. Extreme winds are typically generated by extratropical cyclone systems originating in the North Atlantic, which are often characterised by a track of local vorticity maxima.

While there have been numerous statistical studies on modelling extreme winds, little has been done to model the influence of the extratropical cyclone on the wind speeds that they generate. By modelling the development of windstorms in a Lagrangian frame of reference, we can assess the joint risk of severe events occurring at multiple sites.

Modelling windstorms

In this talk, Paul will present a novel approach to modelling windstorms that preserves the physical characteristics linking the windstorm and the cyclone track. This was done by exploring the dependence structure of these characteristics in a Lagrangian frame of reference.

The team explored a copula-based filtering approach to identify and extract the spatial footprint of extreme windstorm events. Then they used a Markov process to propagate the characteristics of the footprint in time relative to the cyclone track.

The overall model allows simulation of synthetic windstorm events, which can be used to quantify the risk associated with previously unobserved events at different sites, thus representing a useful tool for practitioners with regard to risk assessment.

In particular, for case studies in the northwest of England and eastern Germany, the team show that the spatial extent of windstorms become more localised as its magnitude increases, while the model captures the varying degrees of spatial dependence at different sites.

Curious how future windstorms will affect our climate?

Click here to read the full report.

Want to know more?

Email Paul Sharkey for more information on both his paper and his attendance at the Royal Statistical Society international conference. You can also find out more on our climate resilience web page.



Leave a Reply