Multivariate extreme storm surge flooding events on the UK’s east coast
thesisposted on 2018-12-10, 09:38 authored by Cristina Caballero-Megiddo
In the United Kingdom (UK), floods, and specifically coastal flooding, are a hazard that is commonly thought likely to increase due to the impacts of climate change and the results of development in areas at risk. East coast storm surges have been extremely devastating in the recent past, such as the events of 1953 or the winter of 2013/14. The challenge is to analysis the risk of widespread, concurrent and clustered coastal flooding in a regional scale. It is widely accepted that extreme value analysis (EVA) is an important tool for studying coastal flood risk, but it requires the estimation of a threshold to define extreme events and has to cope with the problems of missing values within the dataset. The main areas of research discussed in this thesis involve making improvements to the way that extreme thresholds are selected and providing an alternative approach for multivariate missing values. By applying an automated threshold selection method to the data, more plausible and less subjective results can be yielded over the traditional manual approach. The alternative multivariate analysis at regional scale considers the statistical dependences between locations and which possible combination of events to take into account in order to handle missing values within time series dataset. Both areas of research provide developments to existing extreme value methodologies, hence enhancing the predicted future storm surge coastal flood modelling. An application of this research is to analysis the potential impacts of proposed nuclear power stations considering the increase likelihood of occurrence of extreme storm surge events. This research undertakes EVA with the statistical programming language R. However, R provides a range of functions embedded in different R packages, it was necessary to create new functions, scripts and commands to improve the analysis of extremes in order to undertake the threshold selection and cope with missing values. This research selects, as a case study, fourteen tide gauges along the East Coast of the UK from Lerwick to Dover. The main measure is skew surge due to be an independent and identically distributed variable and all phase differences in the calculations are removed. The multivariate model provides the likelihood of future significant storm surge flooding events along the East Coast of the UK. Results show that return levels for 50, 100 and 250 years estimates higher impact of ≈1m in Felixstowe, Sheerness, Immingham, Cromer and Lowestoft, while the northern gauges show an increment of ≈0.5m. Moreover, due to the overdispersion of the dataset, high predicted values are estimated in Lowestoft, Felixstowe and Dover where currently nuclear power sites are generating energy and new sites will be built in the future. In summary, the main aim of this research is to undertake a multivariate extreme model to analysis the potential impacts of future storm surge coastal flooding at a regional scale. By analysing extreme skew surge events at a regional level, a more complex storm surge coastal flooding model can be elaborated, and therefore, better results can be obtained. The multivariate extreme model requires how to select extreme events and how to handle missing values within the dataset. Hence, the proposed Automated Graphic Threshold Selection (AGTS) method provides a mathematical and computational tool to select extreme threshold, and moreover, the Multivariate Extreme Missing Value Approach (MEMVA) handles the missing values in time series dataset. The multivariate extreme model has the potential to improve the regional risk assessment of widespread, concurrent and clustered coastal flooding events.
- Architecture, Building and Civil Engineering
Publisher© Cristina Caballero
Publisher statementThis work is made available according to the conditions of the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) licence. Full details of this licence are available at: https://creativecommons.org/licenses/by-nc-nd/4.0/
NotesA Doctoral Thesis. Submitted in partial fulfilment of the requirements for the award of Doctor of Philosophy of Loughborough University.