1.4 Evaluation of NHRCM

As global warming projection involves systematic bias due to the climate model limitations described in the previous section, it is important to examine how well climate models reproduce observation results before climate projection data are used. For GWPV8, model climatology was first calculated from a historical run of the non-hydrostatic regional climate model (NHRCM) for grid points corresponding to observation stations. The model outputs were then compared with observation results, and the NHRCM's ability was evaluated.

Despite systematic biases with temporally and spatially small scales, the NHRCM is capable of reproducing the overall characteristics of Japan's climate. This indicates that the model provides outstanding sample data for the analysis of future changes in temporally or spatially averaged quantities. However, for some variables (e.g., numbers of high-temperature days, percentiles, 20-year return values of precipitation and relative humidity), the results may be highly sensitive to these small systematic biases. Accordingly, the NHRCM's outputs are statistically corrected using observational data before these variables are analyzed.

The ability of the NHRCM to reproduce observation is outlined in Appendix A along with the method of bias correction.