Impact of Model Resolution on Secondary Eyewall Formation and Eyewall Replacement Cycle

Depicting Secondary Eyewall Formation (SEF) and Eyewall Replacement Cycle (ERC) in a numerical model is important for tropical cyclone (TC) forecasting. However, there is no consensus about what resolutions are appropriate to describe SEF/ERC within a full-physics mesoscale model. In this study, numerical experiments are conducted to examine the impact of the horizontal and vertical resolutions on SEF/ERC. The mesoscale model is configured through nesting to the horizontal grid spacings of 6, 4, 2, 1.33, 0.67-km, and with 27- and 54-levels on an f-plane in a quiescent environment. In addition, there are more levels below 1.5-km to better describe the TC boundary layer (TCBL). The simulations with 6 and 4-km grid spacings show no obvious SEF/ERC regardless of the number of vertical levels. When the horizontal grid spacings decrease to 2-km or smaller, the simulations manifest SEF/ERC. These results are supported by a few simulations with the ARW model using similar configurations. Furthermore, the spectra of kinetic energy and vertical velocity from various resolutions confirm that the grid spacings should be smaller than 4-km to resolve SEF/ERC. The impact of doubling vertical levels on the SEF/ERC is not as significant as doubling the horizontal resolutions. Finally, we discuss the coupling between the balanced/unbalanced flows (above/in the TCBL), and their effect on SEF. It is proposed that the coupled balanced/unbalanced processes that generate the quasi-steady cooling zone in the primary eyewall and two warming regions inside and beyond the cooling zone are essential for SEF.

Automated summary

Depicting Secondary Eyewall Formation (SEF) and Eyewall Replacement Cycle (ERC) in a numerical model is important for tropical cyclone (TC) forecasting. The impact of horizontal and vertical resolutions on SEF/ERC is examined through numerical experiments. Results show that SEF/ERC is manifested when the horizontal grid spacing is 2-km or smaller, while no obvious SEF/ERC is observed with grid spacings of 6-km and 4-km.