Evaluation of the performance of the latest WRF model (V4.5) in simulating surface elements and its sensitivity to the diurnal SST cycle in the South China Sea

The performances of six surface layer parameterizations in the latest Weather Research and Forecasting Model (WRF Version 4.5) are examined via observations at Yongxing Island, South China Sea. Three short-term simulations and two long-term simulations were carried out, and four statistical indicators were used to quantify the performance of different schemes. The statistical indicators suggested that all the surface schemes had a better accuracy in simulating surface wind than in simulating surface temperature and moisture. Meanwhile, the unresolved precipitation in the WRF model made significant increases in the model biases for all the schemes. Overall, the Global Forecast System surface scheme performed better in simulating surface variables than other schemes on nonrainy days in the study area. The simulation results also demonstrated that the diurnal SST cycle input could lead to remarkable improvements in the accuracy of the simulated surface heat fluxes and the temperature for all the schemes. The daily averaged determination coefficients between the simulation and observation results almost doubled, and the daily averaged root mean square error (RMSE), as well as the relative RMSE, were reduced by approximately half. However, the surface air specific humidity was not sensitive to the diurnal SST cycle, and only small improvements were identified in these simulations with the diurnal SST cycle. Therefore, the calculation module for the surface humidity in the surface schemes needs to be further optimized.