A Quantitative Method to Evaluate Tropical Cyclone Tracks in Climate Models

The ability to simulate tropical cyclones (TCs) realistically is an important factor in the performance evaluation of climate models. In previous studies, indirect evaluation methods have been proposed that are based on the comparison of TC-related background circulation between model results and observations. Direct model evaluation methods, in most cases, are limited to the model skill in simulating the TC frequency, intensity, and track density. Here we propose a new method to quantitatively and directly evaluate the ability of climate models in simulating TC tracks. The method consists of two indicators that account for the model performance in simulating TC track density and the geographic properties of TC tracks, respectively. This method is applied to evaluate the skill of climate models in simulating TC tracks over the western North Pacific Ocean. The explicit models include seven from phase 5 of the Coupled Model Intercomparison Project and eight from the U.S. CLIVAR Hurricane Working Group (HWG), as well as four downscaled HWG models. Our results indicate the order of these 15 explicit models according to their ability to simulate TC tracks. In addition, we show that, for one of the models, the TC track simulation is greatly improved by using downscaling.