Numerical Simulation Of Heavy Fog Events Based On Different Parameterization Schemes And An Assessment On Forecasting Performance

In order to compare different effects on simulations of heavy fog processes under various microphysics, planetary boundary layer and land surface schemes,21 heavy fog events selected from the observed data of weather monitoring system from 2011 to 2013 were simulated using WRF (Weather Research and Forecasting Model). Influence of different parameterization schemes on simulating these fog events was discussed and the forecasting indexes of fog formation based on the output physical variables of WRF were determined. The results are as follows:(1) The simulation of the whole fog occurring processes works best when selecting WDM6 Scheme as the microphysics scheme, QNSE Scheme as the planetary boundary layer scheme and the SLAB Scheme as the land surface scheme considering the effects of different parameterization schemes in simulating the near-surface meteorological factors, the air temperature in the upper-level atmosphere and the distribution of fog areas comprehensively. (2) Under the optimum configure of the parameterization schemes and giving consideration to the limited computing resources of the meteorological professional departments in China and the suitable vertical resolution of the model, the analysis on the numerical simulation results of atmospheric backgrounds during 21 heavy events demonstrated that the forecasting indexes of heavy fog on the expressway of Jiangsu province were the liquid water content at lowest model level (30-40 m) is more than 0.015 g/kg, or 2m level relative humidity is more than 95% and 10m level wind speed is less than 3m/s at the same time.Numerical forecasting was simulated day by day from October to November in 2014 using the determined optimum configure of the parameterization schemes and multi-factors forecasting indexes, and fog forecasting ability of WRF model was verified. One dimensional model PAFOG driven by the output variables of WRF model was used to forecast the same fog events day by day, and the forecasting performance of two models was analyzed comparatively. The results are as follows: (1) Either WRF model or WRF-PAFOG model shows some forecasting ability of regional fog, but both exist the shortage of high missing rate. (2) In general, forecasting performance of WRF model is slightly better than that of WRF-PAFOG model, and TS (Threat Score), FAR (False Alarm Rate), MR (Missing Rate) of WRF model on day-by-day forecast is 13.3%,84.9% and 46.8%, while that of WRF-PAFOG model is 11.5%,82.9% and 74.1%. (3) Two models have their own advantages on fog forecasting, WRF model is better on forecasting large area regional fog because of its lower MR, while WRF-PAFOG model is better on forecasting decentralized local fog because of its lower FAR.