随机辐射参数化过程扰动方案对温度预报的影响

THE IMPACT OF STOCHASTICALLY PERTURBED RADIATION PARAMETERIZATION TENDENCIES SCHEME ON TEMPERATURE FORECASTING

关键词:  温度  辐射参数化  集合预报  随机物理过程扰动

Key words:  temperature  radiation parameterization  ensemble prediction  stochastically perturbed parameterization tendency

摘要:

数值模式中辐射参数化过程的不确定性是导致温度预报不准确的原因之一，为了在WRF集合预报系统中提高温度预报的效果，提出一种针对辐射参数化倾向的随机扰动方案（Stochastically Perturbed Radiation Parameterization Tendencies, SPRPT）。并将这种方法与多辐射参数化物理过程方案、多参数扰动方案及传统的随机物理过程扰动方案（SPPT）方法对比。针对2014年7月的温度模拟过程中，多辐射参数化物理过程方案虽然在700 hPa以下及地面2 m温度的预报上有较大离散度，但会增大温度预报的误差，但综合效果没有SPRPT方案好；扰动散射调谐参数可在一定程度上提高温度预报，但不显著；在同样扰动参数下，SPPT方案对温度的预报改进不明显。而SPRPT方案能显著提高集合预报系统的离散度，降低地面2 m温度的暖偏差。评分指出该方案对集合预报系统，尤其是在模式底层及近地面的温度预报上，改善明显。

Abstract:

The uncertainty of the radiation parameterization process in numerical models is one of the reasons for inaccurate temperature prediction. To improve the performance of the WRF ensemble system in temperature prediction, a scheme called stochastically perturbed radiation parameterization tendencies, or SPRPT, is proposed in the present study. This scheme is compared with the multi-radiation parametrization process scheme, the multi-parameter perturbation scheme, and the traditional stochastically perturbed parameterization tendencies scheme (SPPT). In the simulation of the temperature in July 2014, although the multi-radiation parameterized physical process scheme has a large dispersion in the prediction of the temperature below 700 hPa and the temperature of two meters on the ground, it increases the error of temperature prediction, and in contrast, SPRPT has a better performance in temperature prediction. Perturbing the scattering tuning parameter can improve temperature prediction, but the improvement is not significant; under the same disturbance parameters, the SPPT scheme does not improve temperature prediction significantly. The SPRPT scheme can significantly improve the dispersion of the ensemble forecasting system as well as reduce the warm deviation of the two-meter temperature. Scores indicate that the scheme has improved the ensemble forecasting system significantly, especially in the prediction of temperature at the bottom layer and the surface of the model.