A Copula-Based Multivariate Frequency Analysis for Hydrological Applications
Sander Vandenberghe1, Niko E.C. Verhoest1, Bernard De Baets2
1Laboratory of Hydrology and Water Management, Ghent University, Ghent, Belgium; 2Department of Applied Mathematics, Biometrics and Process Control, Ghent University, Ghent, Belgium

In the last decade, hydrologists are increasingly aware of the importance to account for the multidimensional character of hydrological processes when performing a frequency analysis. Recent research has shown the great potential of copulas to couple marginal distribution functions into a multivariate one, describing the association between the variables involved. In this presentation, the most important concepts of a copula-based multivariate frequency analysis will be discussed and illustrated by applying them in the context of rainfall and runoff modeling, paying attention to some pitfalls that might occur. A particular focus will be on multivariate return periods and their interpretation. Also an intuitively understandable way of constructing higher dimensional copulas will be explained and evaluated from a practical point of view.

Keywords: Copula; Rainfall-runoff modelling; Multivariate return period; Extremes

Biography: Niko Verhoest received the engineering and the Ph.D. degrees in applied biological sciences, both from Ghent University, Belgium, in 1994 and 2000, respectively. He has been Teaching Assistant (1998-2000) and Assistant Professor (2000-2002) with the Laboratory of Hydrology and Water Management, Ghent University. In 2002, he became Associate Professor of Hydrology and Water Management at the Faculty of Bioscience Engineering, Ghent University. His research interests include, amongst others, the analysis and modeling of rainfall.