Efficient Statistical Approach to Multisite Downscaling of Extreme Temperature Series Using Singular-Value Decomposition Technique

Abstract

Downscaling techniques are required to describe the linkages between global climate model (GCM) outputs at coarse grid resolutions and surface variables at suitable finer scales for climate change impact and adaptation studies. The present paper proposes an improved statistical approach to downscaling of daily maximum (Tmax) and minimum (Tmin) temperature series located at many different sites concurrently. This new approach is based on a combination of a multiple-regression model and the modeling of its stochastic component by the singular-value decomposition (SVD) technique to represent more effectively and accurately the space-time variabilities of these extreme daily temperature series. Results of an illustrative application using data from a network of 1 weather stations located in the southwest region of Quebec and southeast region of Ontario in Canada and from the available National Centers for Environmental Prediction/National Centre for Atmospheric Research (NCEP/NCAR) reanalysis data set indicated the effectiveness and the accuracy of the proposed approach. In particular, this new approach was found to be able to reproduce accurately the basic statistical properties of the Tmax and Tmin time series, including their mean, standard deviation, Tmax 9th percentile, and Tmin 1th percentile. In addition, the at-site autocorrelations, interstation correlations, and intervariable correlations of the daily Tmax and Tmin series have been accurately reproduced. Furthermore, the proposed approach was able to adequately reproduce the interannual variability of the Tmax and Tmin.