Abstract: Based on Weather Research and Forecast (WRF)-Rapid Update Cycle (RUC) model system,three operational visibility schemes,including Stoelinga and Warner (SW) scheme,RUC scheme and Forecast Systems Laboratory (FSL) scheme,were used to make a forecast experiment for 25 fog events from October of 2015 to March of 2016 in Shandong province.The experiment included quantitative forecast of visibility and diagnostic forecast on weather fog occurs.The forecasting result is evaluated using routine meteorological observations.The causes of bad forecast on fog weather by different schemes were analyzed and then the RUC scheme was improved.The results indicated that the mean absolute error of visibility computed by the SW scheme ranges between 9 km and 12 km,while that computed by RUC scheme and FSL scheme varies within 2-4 km.For most of the fog events over Shandong province in the forecast experiment,the hit rate is larger than the false alarm rate based on the SW scheme and the FSL scheme,which means that the two schemes have a certain capability to forecast fog events.However,the hit rate and false alarm rate of fog prediction by the RUC scheme are both zero,indicating that this scheme has no ability to forecast fog.Combining the evaluating results of visibility and fog,the FSL scheme is more suitable to produce the operational product of fog in the WRF-RUC model.The major causes of bad predictive results include that the simulated visibility is very sensitive to the error of relative humidity,the influence of air pollutant on visibility is not considered,and lack of appropriate localized revision impact on the forecast effect of visibility scheme on fog.After improving the RUC scheme,the mean absolute error of the visibility deceases and the forecasting performance of fog is improved.

Key words: Fog, Visibility, WRF-RUC model, Forecast experiment