Flooding on a yearly basis represents a significant share of all economic and insured losses in
Slovenia and globally. To assist in the estimation of their sizes, catastrophe modelling emerged
as an appropriate tool in the last decades. These models consist of four main components:
hazard module, exposure module, vulnerability module and financial engine, which performs
the calculation of the losses while taking into account the financial conditions of the policies.
The hazard module forms the essence of these models as it simulates possible physical
phenomena, typical for the peril in question. When analysing flooding, it is necessary to define
the flood event set which describes the events with their exact geographical location and local
intensity. Due to the multi-site nature of flooding, classical univariate probability analyses are
less appropriate when analysing joint flood risk. Copula functions have started being used in
hydrology as an alternative to the classical approach.
The usability of copula function to develop a flood event set for the Republic of Slovenia has
been explored in the graduation thesis. Three Student t copula based and one Gumbel copula
based event set have been developed and compared. Bivariate analyses of observed and
simulated flows at pairs of stations showed that the Student t copula with 0,483 degrees of
freedom and Gumbel copula with the parameter
|
