Understanding the behaviour of environmental extreme events is of utmost importance for a number of sectors, for example the government, insurance companies and oil firms. Interest lies in determining the level at which to protect infrastructure against events that are potentially larger than those that have been previously observed. In this presentation, we illustrate the development of novel statistical models for univariate and multivariate extreme values through two applications, firstly changes in the extreme wave climate of the North Sea and secondly through gaining a better understanding of the risk of widespread flooding.
For safe offshore operations, accurate knowledge of the extreme oceanographic conditions is required for the next 50 years. We address this problem using a multi-step statistical downscaling algorithm using data from low-resolution global climate model (GCM) and local-scale hindcast data. The GCM is unable to produce wave data accurately so instead we downscale wind speed and exploit the relationships between wind characteristics and wave heights.
As a result of storms Desmond, Eva and Frank, a National Flood Resilience Review (NFRR) was formed by the UK government to understand the risk of widespread flooding. Some of the questions raised by this committee include ‘What is the chance of an extreme river flow occurring at one or more gauges, somewhere within the national river gauge network in any one year?’. In order to address questions of this nature it is vital to understand the dependence of large values of river flow and determine which combinations of locations are likely to simultaneously observe high flows. We use a multivariate extreme value model to summarise the changing behaviour of floods across the UK.
About the Speaker: Ross Towe (Senior Research Associate, Lancaster University)
Ross Towe is a statistician with a background in developing novel statistical models that characterise the natural environment. He is currently a Senior Research Associate at Lancaster University on the EPSRC funded project: The Role of Digital Technology in Understanding, Mitigating and Adapting to Environmental Change. He was previously a Knowledge Transfer Partnership (KTP) Associate, who focussed on improving statistical models for flood risk assessment. During this post, he also provided scientific analysis for the Scientific Advisory Group of National Flood Resilience Review. His PhD was a collaborative project with Shell, which developed statistical methods to predict the future extreme wave climate of the North Sea through utilizing information from global climate models.