Journal of wind engineering and industrial aerodynamics vol:92 issue:12 pages:1025-1037
The Multivariate Generalized Burr-Gamma distribution is used to model jointly the daily maximum wind data measured at four stations in The Netherlands. The main aim is to estimate high quantiles and tail probabilities based on a statistical model which includes covariate information such as wind direction. We discuss the advantages of such a multiple station approach in cases where the stations are not too distant and are therefore subject to a single meteorological mechanism for producing extreme winds. (C) 2004 Elsevier Ltd. All rights reserved.