Synoptic airflow and UK daily precipitation extremes: development and validation of a vector generalized linear model (Q549640)

A vector valued generalized linear regression model with generalized extreme value (GEV) distribution is proposed for probabilistic prediction of local precipitation extremal values. Dependence of scale and shift parameters of GEV from the airflow strength, vorticity and direction are modeled. The time of observations is used as a regressor to model the annual cycles. Maximum likelihood estimates are used for fitting. An application to real data from UK is discussed.