Impacts of Diabatic Initialization and Cumulus Parameterization on Numerical Typhoon Prediction

[摘要] Numerical experiments were conducted with a regional atmospheric model to investigate the suitability of the model for simulation/prediction of tropical cyclones (TCs). The model uses primitive equations and is similar in construction to a regional atmospheric model, called RegCM, which has been used extensively for regional climate modeling. The suitability of the model for TC study was tested in the mode of shortrange forecasts to find out how well the model can reproduce TC behaviors with a special attention on their genesis and development. Objective analysis data for typhoons Ed (9018) and Flo (9019) in September 1990 were used as the input data. In addition, “observed” precipitation (PR) rates estimated from hourly infrared satellite measurements were used.In the prediction experiments, we focused on two aspects: One was to study the impact of data initialization. The other was to examine the sensitivity of forecast depending on the choice of two cumulus parameterizations, Kuo and relaxed Arakawa-Schubert (RAS) schemes. The initialization to the input data was done through the combination of diabatic nonlinear normal mode initialization for dynamical fields and cumulus initialization for moisture. The diabatic heating rate was determined from radiative and condensation heating rates using the input data, including observed PR rates. The initialization of moisture was done through the inversion of cumulus scheme for a specified PR rate.Positive impacts of the initialization on reproducing the behaviors of both typhoons are clearly demonstrated in the use of both cumulus schemes. In particular, the initialization results in successful simulations of initial PR field without spinup problems. Thus, the positive impact is greatest in the case of TC genesis. The reproducibility of PR by the model with RAS scheme is superior, and the RAS scheme appears more suitable than the Kuo scheme for the intensity prediction of developing typhoons. However, the use of the Kuo scheme produces better scores in mean displacement errors.

[学科分类] 大气科学