n Annual Proceedings of the South African Statistical Association Conference - Bivariate threshold excess models with application to extreme high temperatures in Limpopo province of South Africa

Volume 2018 Number Congress 1
  • ISSN :


A common focus of statistical analysis is on the average or centre of the data. In many situations, inference concerning the tails of the distributions is of interest. This leads to problems that are addressed by using extreme value theory (EVT). In the presence of multiple variables, there arises a need for multivariate extreme value theory (MEVT) which makes it possible to jointly model the variables and extract more information for inferential purposes. In this paper we use MEVT in modelling extreme high temperatures in South Africa. The data is on maximum daily temperatures provided by the South African Weather Services over the period January 2000 to December 2016. The aim of this paper is to model extremal dependence of temperatures at three meteorological stations in Limpopo province which are Polokwane, Thohoyandou and Lephalale. A penalised cubic smoothing spline is used for nonlinear detrending of the data prior to fitting the models. The residuals are extracted, but only the positive residuals above the threshold are used for modelling the dependence structure using the bivariate threshold excess models based on the Laplace margins.

Loading full text...

Full text loading...


Article metrics loading...


This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error