资源环境科技发展态势分析平台

The sensitivity of simulated tropical cyclones (TCs) to resolution, convection scheme and ocean surface flux parameterization is investigated with a regional climate model (RegCM4) over the CORDEX Central America domain, including the Tropical North Atlantic (TNA) and Eastern Tropical Pacific (ETP) basins. Simulations for the TC seasons of the ten-year period (1989-1998) driven by ERA-Interim reanalysis fields are completed using 50 and 25 km grid spacing, two convection schemes (Emanuel, Em; and Kain-Fritsch, KF) and two ocean surface flux representations, a Monin-Obukhov scheme available in the BATS land surface package (Dickinson et al. 1993), and the scheme of Zeng et al. (J Clim 11(10):2628-2644, 1998). The model performance is assessed against observed TC characteristics for the simulation period. In general, different sensitivities are found over the two basins investigated. The simulations using the KF scheme show higher TC density, longer TC duration (up to 15 days) and stronger peak winds (> 50 ms(-1)) than those using Em (< 40 ms(-1)). All simulations show a better spatial representation of simulated TC density and interannual variability over the TNA than over the ETP. The 25 km resolution simulations show greater TC density, duration and intensity compared to the 50 km resolution ones, especially over the ETP basin, and generally more in line with observations. Simulated TCs show a strong sensitivity to ocean fluxes, especially over the TNA basin, with the Monin-Obukhov scheme leading to an overestimate of the TC number, and the Zeng scheme being closer to observations. All simulations capture the density of cyclones during active TC seasons over the TNA, however, without data assimilation, the tracks of individual events do not match closely the corresponding observed ones. Overall, the best model performance is obtained when using the KF and Zeng schemes at 25 km grid spacing.