How well do Global Climate Models detect tropical cyclones considering cyclone phase?

The center of a tropical cyclone (TC) is searched and tracked by determining whether the model data or the gridded parameters show static or thermodynamic characteristics. Similar to existing studies on searching and tracking TCs using gridded simulation data, the searching and tracking of the TCs in this study was divided into tracking, which searches for the center of the cyclone, and structure, which explores the vertical structure at the center. Moreover, to identify cyclones such as extratropical cyclones based on climate simulation data, the cyclone phase (CP) was also considered. To verify the TC tracking method, GFS analysis field data were collected to define the TCs that occurred in the Western North Pacific (WNP) during 2009–2014. The verified tracking method was used in the current climate simulation, which was generated using a AGCM GME, which are a high-resolution model for application and evaluation of current climate predictions. The experiments using GFS data for verifying the existing tracking method showed that 146 TCs occurred in the WNP over 6 years, which is 23 more than the observed number, and they showed monthly/seasonal characteristics similar to those of the observation. However, when the cyclonic phase was considered for the identified TCs, 14 other cyclones were found. Therefore, even when the other cyclones were excluded, the correlation coefficient was maintained at 0.9 or higher. Furthermore, this method was applied to the simulations of current climate from 1980 to 2009. As in the analysis field research, the results showed that the correlation coefficient of the seasonal variation between the model and observation increased from 0.6 to 0.8, particularly when the cyclone phase was considered.