Abstract
The main purpose of this paper is to investigate the multivariate depencence relationship between meteorological variables and thermal energy demand through copula modelling. In order to keep into account both the multivariate dependence structure and the diverse bivariate relationships observed in the data, mainly two approches can be followed. One is to adopt a mixture of copulas that allows us to combine different copula families and to generate complex dependence structures not captured by the existing models. The other way is to use vine copulas that make it possible to analyse multivariate probability distributions through bivariate and conditional bivariate copulas organised in a suitable tree. A comparison of the two approaches is discussed.