ABSTRACT: Statistical downscaling models evaluation: A regional case study for Quebec regions, Canada.

CGRG Bibliography of Canadian Geomorphology
Search Results

Author :

Date :

Title :

Publication :

Issue :

Page(s) :

Abstract

The evaluation of statistical downscaling tools constitutes a key step toward the development of a regional Climate Scenarios Set (CSS) for: 1) stakeholders implicated in the evaluation of the impacts of climate variability and change and 2) decision makers that regard information related to climate change as being an integral variable in planning resources, activities and programs for a new climate reality. Statistical Downscaling (SDS) models used for this purpose, namely Statistical Downscaling Model ½ USDSM (e.g. Wilby et al., 2001), and Long Ashton Research Station Weather Generator- LARS-WG (e.g. Semenov and Barrow, 2002), are rigorously and systematicallyinter-compared over a wide range of climatic conditions. These tools are evaluated for the period 1961-1990 in focusing on extreme events and climate variability. Daily precipitation and maximum, minimum, and mean surface temperature series are generated for western Quebec. The ability of the statistical downscaling models to simulate these variables are evaluated using (1) Percentile analysis, (2) variance analysis, (3) model bias and RMSE (root mean square error), (4) extreme indices using in part the STARDEX Diagnostic Extremes Indices Software (SDEIS-A, cf : http://www.cru.uea.ac.uk/cru/projects/stardex/). SDEIS-A is applied to calculate indices using statistically downscaled GCMs and observed daily time series. A comparison between indices derived from statistically downscaled GCM output and observed data enables estimation of uncertainties introduced by using GCM output and statistical downscaling. Trend assessment of the simulated extreme indices versus historical ones examines changes in the extreme indices and skill of the SDS models.

Bibliography of Canadian Geomorphology