Simulation Of TC Climatic Characteristics Over The Western North Pacific Based On CMIP5 Models

Based on 30 global climate models from the Coupled Model Inter-comparison Project 5(CMIP5), TC activity over Western North Pacific (WNP) has been successfully simulated. To investigate simulated TC activity based upon the climate models, a downscaling TC system is adopted. Its input mainly includes three variables:sea surface temperature, large-scale steering flows and vertical wind shear. These three variables are provided by the projected large-scale environments derived from a selection of CMIP5 based on the Bayesian model averaging (BMA).The downscaling technique driven by the large-scale environment from NCEP/NOAA reanalysis data can reasonably reproduce TC track and intensity features over the WNP, but the simulation of TC intensity’s spatial distribution and interannual variability still remains to improve. In consideration of difference in large-scale environments among climate models, Bayesian model averaging and simple multi-models mean are applied to reduce the uncertainty of large-scale environments. A comparison of BMA and ensemble mean is also represented, on large-scale environments and simulated TC characters by the downscaling system, which shows that BMA have an advantage over the multi ensemble mean method. Further, two BMA methods are compared, one is 30 CMIP5 models’ multi-ensemble based on the BMA, and the other is the effective models’ multi-ensemble based on the BMA selected according to the evaluation of large-scale environments derived from 30 CMIP5 models. Both of two can obtain the comparatively precise large-scale environments and TC activities, the distinction between them is negligible. After the verification, BMA can obtain consistent large-scale environments from multiple CMIP5 models without evaluation and selection. Finally, some sensitive tests are conducted for investigating the role of sea surface temperature, large-scale steering flows and vertical wind shear on the TC activity over the WNP basin. The sensitive tests are designed under the premise of maintaining the balance between environmental factors in models, with BMA and the downscaling technique. It’s found that TC track is more sensitive to SST and large-scale steering flows; TC intensity is sensitive to all three environmental factors, while the impact of parameter SST and vertical wind shear exhibit more significant in TC lifetime maximum intensity and the number of intense TCs.