Tropical Cyclone Intensity Change in the Western North Pacific: Downscaling from IPCC AR4 Experiments

The influence of global warming on tropical cyclone (TC) intensity has been discussed extensively over the past decade and downscaling studies have been conducted by embedding high-resolution regional models in the large-scale environment derived from global warming experiments in previous studies. With theoretical models for TC movement and intensity, this downscaling study examines the impact of global warming on TC intensity in the western North Pacific basin with attention to the selection of climate models for the Intergovernmental Panel on Climate Change (IPCC) fourth assessment report. Individual contributions of SST, vertical wind shear and TC track change to TC intensity change are also discussed. The synthetic TC activity in the study is simulated with motion and intensity models respectively, while little change is assumed in TC formation due to uncertain predictions for regional changes in future TC formation. The downscaling models are first evaluated against the observations during 1965-1999, and then ten IPCC models are selected based on TC intensity downscaled with large-scale environments from their control runs. With an increase of 1.6-3.5 K in SST over the period 2065-2099, the TC peak intensity and number of intense TCs are projected to increase by 14% and 66% respectively, in agreement with previous studies. That is to say, it is likely that intense TC numbers in this region will increase in a warmer world unless the future TC frequency decreases considerably in the western North Pacific basin. Although SST warming plays a dominant role in affecting TC intensity in most of the climate models, this study suggests that changes in vertical wind shear and prevailing tracks can have significant influence on TC intensity change, sometimes comparable to or even larger than the SST change. However, changes in TC track and vertical wind shear do not have the same systematic effect on TC intensity as SST does, increasing the projection uncertainty.