Investigating the Uncertainty Effect of GCMs Output on the Prediction of Meteorological Parameters in Golestan Province

Abstract

One of the important issues in assessing climate change using the output of General Circulation Models (GCMs) is their uncertainty so that the outputs of a model in a region may vary with another model in the same region. Disregarding the uncertainty of these models reduces the accuracy of the final outputs (Ashofteh and Massah, 2012). Various methods have been developed to analyze and reduce the amount of uncertainty. Among the methods used to investigate the uncertainty of the output of GCMs, one can mentioned the weighted means of observation, Wilcoxon Signed Rank test, Bootstrap confidence-interval estimation technique, Box Plot method, and the cumulative frequency distribution function. Accordingly, the present study, while predicting the temperature, precipitation and drought variables in Golestan province for the future 30 years via two general circulation models including ECHO-G and HadCM3, examined the uncertainty of these models by weighted means of observation and Box Plot methods. Also, statistical analysis of data by analysis of variance and mean comparison tests are among other goals of this research.

Keywords

• echo-g
• hadcm3
• lars-wg
• precipitation
• spi
• temperature