Evaluation of a Fine Resolution Climate Model in the Variability of Potential Tropical Cyclone Genesis Frequency over the Western North Pacific
This study evaluated the performance of GFDL HiRAM, a fine resolution AGCM, in the simulation of GPI (Genesis Potential Index) of tropical cyclone and its temporal variation over the Western North Pacific (WNP). We analyzed the AMIP simulation by the AGCM for the 30-year (1979-2008) forced by observed sea surface temperatures as the lower boundary condition. Since GPI depends on the five large-scale environmental factors(850 hPa absolute vorticity, 700 hPa relative humidity, vertical wind shear, maximum potential intensity, and 500 hPa vertical velocity), the biases of the simulation are examined for these factors as well as GPI itself. The results are compared with the ECMWF Interim reanalysis (ERA-I), and the analyses show that both the mean spatial pattern and the seasonal cycle of GPI over the WNP are reasonably simulated by HiRAM. But the magnitude of GPI is significantly underestimated due to the combined contribution of negative biases in four factors excluding the low-level vorticity. It is demonstrated that the three leading modes of spatio-temporal variability of GPI in EOF analysis for ERA-I are associated with ENSO, climate change with long-term trends, and SST anomalies over the WNP. The response of GPI to ENSO is more or less captured by HiRAM, including the east-west shift of Typhoon genesis location. However, it is supposed that unrealistic response of GPI and its factors to La-Nina or eastern Pacific El-Nino is an important shortcoming of HiRAM. In addition, HiRAM fails to reproduce the characteristic spatiotemporal variation associated with the climate change mode of GPI. The key findings from this study provide helpful guidance for improvement of HiRAM.