Statistical Diagnostic Tests for Multivariate/MultiSite Hyrological Time Series
Ian McLeod, Esam Mahdi
Statistical and Actuarial Sciences, University of Western Ontario, London, ON, Canada

A new type of diagnostic test based on the generalized variance is discussed. The purpose of this diagnostic check is to test the assumption of statistical independence of the error terms or innovations in these models. Most of the theoretical component of this work has been accepted for publication in the Journal of Time Series Analysis. We will focus on hydrological extensions and applications to multisite and multivariate riverflow. Simulation of daily flow using wavelet methods will also be discussed.

Keywords: Multivariate; Diagnostic check; Monte Carlo test; Wavelets

Biography: A. Ian McLeod received his Ph.D. from the University of Waterloo in 1977. As of October 2005, Dr. McLeod has published 60 articles in refereed journals as well as one textbook and numerous other articles. Dr. McLeod has held visiting research positions at CSIRO, Canberra; COPPE/UFRJ, Brasil; Institute of Statistical Mathematics, Tokyo; University of Kyoto; University of Montreal; Wolfram Research Inc. and was worked on statistical consulting projects for Environment Canada, Statistics Canada, Traffic Injury Research Foundation, Southern California Consolidated Edison Company and sales forecasting for a management consulting company. Dr. McLeod served as Department Chair, Department of Statistical and Actuarial Sciences, University of Western Ontario 2000-2004. As of Fall 2005, 10 Ph.D. theses have been completed under Dr. McLeod's supervision as well as 1 M.Sc. Thesis and 23 M.Sc. Projects.