Master's lecture in Environmental Engineering - Bryndís Tryggvadóttir

Master's student: Bryndís Tryggvadóttir

Title: Adaption of multivariate extreme value modeling for extreme coastal flood events

 ___________________________________________

Faculty: Faculty of Civil and Environmental Engineering

Advisors:  
Hrund Ólöf Andradóttir, Professor at the Faculty of Civil and Environmental Engineering .
Sigurður Sigurðarson, Engineer at The Icelandic Road and Coastal Administration.
Birgir Hrafnkelsson, Professor at the Faculty of Physical Sciences.

Examiner: Helgi Gunnarsson, Engineer at The Icelandic Road and Coastal Administration.

Abstract

Population residing near the ocean lives with the constant threat of coastal flooding, which intensifies with the rising sea level. To mitigate and adapt to this threat requires a reliable coastal flood risk prediction tool. The goal of this project was to adapt a multivariate (joint) probability model to predict extreme coastal flooding events in Southwest Iceland. A Monte Carlo sampling procedure was used together with multivariate analysis to generate a large sample of extreme ocean events, representing approx. 10,000 years, based on local 35-year time series on wave, wind and tidal data. This procedure preserves the dependence structure in the extremes, which has been considered a limitation of previously used multivariate extreme value methods. Due to the computationally and time consuming process of using the wave transformation model MIKE 21 SW to transform the offshore events to nearshore wave points, a limited number of events were run through MIKE 21 SW and then used to fit a meta-model. The meta model was then used to transform the large sample of offshore events to nearshore wave points. The resulting large sample of nearshore events was used to estimate coastal flooding on the basis of wave overtopping discharge for six coastal structures, both for current conditions and with increased sea level, in the capital area and Akranes on the southwest coast of Iceland. The results show that the method is promising and provides a platform for expanding to other coastal regions in Iceland.