|
|
Publications
Lucas, C., Muraleedharan, G. and Guedes Soares, C. (2015), “Outliers identification in a wave hindcast dataset used for Regional Frequency Analysis”, Maritime Technology and Engineering, Guedes Soares, C. & Santos T.A. (Eds.), Taylor & Francis Group, London, UK, pp. 1317-1328
Presence of extreme outliers in a data sample will adversely affect the estimation of extreme quantiles of designated return periods from the proposed probabilistic models. Hence an outlier detection approach is applied and the effect of outliers on extreme value predictions is assessed. The hindcast dataset used in this work is the daily maximum significant wave heights from the HIPOCAS database. The outlier detection based on the classical boxplot was applied to a dataset of 15 sites from an offshore region off Portugal extended over a period of 21 years (1958 - 1978). The outlier analysis was performed to the data samples for execution of regional frequency analysis based on L-moments for estimation of extreme significant wave heights. Boxplot analysis revealed that the configuration of the sub-regions (group of sites of similar L-moment ratios) formed by execution of the algorithms of RFA altered in certain cases and 100 year regional as well as the at-site extreme quantiles of the member sites of the regions were adversely affected by outliers.
If you did not manage to obtain a copy of this paper: Request a copy of this article
For information about all CENTEC publications you can download: Download the Complete List of CENTEC Publications
|