Process-Oriented Diagnostics of Tropical Cyclones in Global Climate Models

Simulating tropical cyclone (TC) activity with global climate models (GCMs) remains a challenging problem. While some GCMs are able to simulate TC activity that is in good agreement with the observations, many other models exhibit strong biases. Decreasing horizontal grid spacing of the GCM simulations tends to improve the characteristics of simulated TCs, but this enhancement alone does not necessarily lead to greater skill in simulating TC activity.

This study uses process-based diagnostics to identify model characteristics that could explain why some GCM simulations are able to produce more realistic TC activity than others. The diagnostics examine how convection, moisture, clouds and related processes are coupled at individual grid points, which yields useful information into how convective parameterizations interact with resolved model dynamics. These diagnostics share similarities with those originally developed to examine the Madden-Julian Oscillations in climate models.

This study will examine TCs in seven different GCM simulations performed at NOAA/GFDL, NCAR, NASA, and CMCC that have different horizontal resolutions and ocean coupling. Preliminary results suggest that stronger TCs are closely associated with greater rainfall – thus greater diabatic heating – in the inner-core regions of the storms, which is consistent with previous theoretical studies. Other storm characteristics that can be used to infer why GCM simulations with comparable horizontal grid spacings produce different TC activity will be examined.