Statistical comparison of scales and spatial distributions of convection in observational data and simulated hurricanes

Much is known about the driving forces behind hurricane convection, the spatial scale of convection and overall storm structure. However, the symmetric and asymmetric storm structure in a simulated hurricane does not always match that of an observed hurricane. That is especially true at the initialization time in the simulation, and in the earliest hours of the simulation when the hurricane is responding to the often non-physical initial state. Incorrect distribution, driving forces and scales of convection can have an effect on the storm size and intensity, and an incorrect storm size can lead to an incorrect track.

In this study, we simulate several Atlantic tropical cyclones using HWRF run at operational resolution, and run higher than operational resolution. We analyze the convection and storm structure throughout the simulation, with a special focus on the initial state and model response to the initial state. We compare with observational data to analyze the role of storm structure and driving forces behind convection in producing inaccuracies in the forecast.