Reproducibility of Regional Climate in Central Japan Using the 4-km Resolution WRF Model

Abstract

The present study provides a preliminary evaluation of the WRF model at 4-km horizontal resolution in central Japan with a 5-year control simulation. The results show that the spatial distribution of the annual mean temperature is well reproduced with a bias of +0.6°C. However, the diurnal range is underestimated by 1.8°C, which is mainly caused by an overestimation of the daily minimum temperature from October to April. Additional analyses indicate that there is a statistically significant correlation between the monthly bias and the number of clear sky days in a particular month, which implies the incomplete reproduction of the surface inversion layer at night. The model overestimates the annual precipitation amount averaged over the analysis domain by 7%, compared to the observations. Such a positive bias results from the overestimation of weak precipitation. On the other hand, the model underestimates the frequency of the heavy rainfall. The performance of the WRF model shown in this study is similar to that of the NHRCM (by Sasaki et al. 2008). The results indicate that the WRF model can be used for high resolution, regional climate simulation/projection research for the studied region.