Title: Rainfall interpolation in Canada by a smooth copula-based generalized extreme value approach
Authors: Fatima Palacios Rodriguez - Universidad de Sevilla (Spain) [presenting]
Elena Di Bernardino - LJAD Laboratoire J.A. Dieudonné (France)
Melina Mailhot - Concordia (Canada)
Abstract: A statistical methodology is proposed based on a hybrid hierarchical smooth GEV copula model to map and predict extreme rainfall in Central Eastern Canada. The rainfall dataset contains a large portion of missing values. In addition, several non-concomitant record periods at different stations are observed in the considered dataset. The proposed approach combines GEV parameters' smooth functions in space through the use of spatial covariates and a flexible hierarchical copula-based model to take into account the dependence between the recording stations. The hierarchical copula structure is detected via a clustering algorithm implemented with a dissimilarity measure designed to handle missing data. Finally, the classical interpolation techniques are compared with our proposed approach.