The main purpose of this research is to provide the utility and applicability of GCM simulations for Korea water resources management. In order to utilize climate information with water resources management and planning, screening methodologies are necessary to determine the regions and conditions under which climate model simulations and predictions may be useful. We used a probabilistic methodology for measuring the comparison of model and observation climatology (means, standard deviations, sample frequency distributions) of large regions, which was recently proposed by Georgakakos (2003). Three different general circulation models (GCMs) with various spatial resolutions and model options operated by domestic and foreign agencies are used to clarify the usefulness of GCM for water resources applications in the study area. The formulation uses the significance probability of the Kolmogorov-Smirnov test for detecting differences between two variables. An estimator that accounts for climate model simulation and spatial association between the GCM data and observed data is used. Atmospheric general circulation model (AGCM) simulations done by ECMWF (European Centre for Medium-Range Weather Forecasts), and METRI (Meteorological Research Institute, Korea) were used for indicator variables, while observed mean areal precipitation (MAP) data, discharge data and mean areal temperature data on the seven major river basins in Korea were used for target variables. A Monte Carlo simulation was used to establish the significance of the estimator values. Monthly analyses by season and tercile discrimination condition were used for the analysis. The results show that GCM simulations are useful in discriminating the high from the low of the observed precipitation, discharge, and temperature values. Temperature especially can be useful regardless of model and season. In this study, we use the probabilistic utility index to evaluate the utility of climate-model simulations for seven watersheds on the Korean Peninsula under significant model and downscaling uncertainty. The important results obtained from this study are summarized as follows.