Comparison of momentum roughness lengths of the WRF-SWAN online coupling and WRF model in simulation of tropical cyclone Gonu

Abstract

The surface enthalpy fluxes (latent and sensible heat fluxes) provide the necessary energy to intensify tropical cyclones (TCs). The surface momentum fluxes modify the intensity of TCs. Various parameters affect the surface fluxes. Drag and enthalpy exchange coefficients are known as parameters that lead to ambiguity in the surface fluxes. Thus, for simulating a TC, various drag and enthalpy exchange coefficients are tested in numerical simulations. Exchange coefficients directly relate to roughness lengths. The main goal of this study was comparing derived roughness lengths of the COAWST (WRF-SWAN online coupling) and WRF models with one another. The TC Gonu formed over the Arabian Sea in 2007 was selected for this study. Employing self-reliant WRF, it is found that track and intensity of the TC Gonu can be best simulated when Donelan parameterization is applied for momentum roughness length and Large-Pond parameterization for enthalpy roughness length. In the second section, simulations were conducted in the coupled mode using the COAWST model with the Large-Pond parameterization (for the enthalpy roughness length). Results cleared that the use of Oost parameterization leads to a high overestimation of simulated momentum roughness length compared with other parameterizations while both the Drennan and Taylor parameterizations were somewhat close to reality. Differences of simulated maximum 10-m wind speed between the Oost parameterization and other parameterizations were more than differences of their simulated minimum central pressure. These results proved that the momentum roughness length affects the maximum 10-m wind speed more significantly than the minimum central pressure. In the last section, a simulation was performed using the COAWST model with the Donelan–Large-Pond parameterization. In this simulation, only 10-m wind speed from the WRF model transferred to the SWAN model. The results showed that simulated wave heights in the open ocean were in good agreement with the results of other researchers. In general, the SWAN model performance is evaluated satisfactorily while the parameterizations using the wave information need to be more investigated.