A Comparison Study On The Performance Between MM5and WRF In Simulation Of Winter PBL And Fog In East China

To compare the performance of MM5and WRF in simulating fog in the planetary boundary layer (PBL), MM5and WRF were run daily for December of2006and December of2007. The model meteorological parameters in the planetary boundary layer (PBL) were validated objectively using the ground-level observations of the China Meteorology Agency routine meteorological network, and the high resolution sounding data at observatories of Nanjing and Anqing.The evaluation results of meteorological parameters in the planetary boundary layer show that:1) Both MM5and WRF at the ground level, the simulated temperature and relative humidity were reliable, but the simulated wind speed was a little worse. Frequency distributions of the biases between simulated and measured temperature, relative humidity and wind speed showed near-normal shapes, the MM5performs better during daytime than during nighttime, but the WRF performs better during nighttime than during daytime for temperature and wind speed. Compare with MM5and WRF, the WRF performs better than MM5in simulating the surface temperature, relative humidity and wind speed, that the WRF has lower MAE and RMSE, but has the higher IA.2) Take Nanjing and Anqing for examples, the WRF and MM5modeled sounding in PBL at both08BT and20BT were acceptable. The results at20BT were better than those at08BT, and both models perform better with increasing height. As for temperature and relative humidity, MM5performed better in Nanjing than in Anqin for both model. By comparing two models, WRF performed better near the surface.3) Take Nanjing as an example, comparing with normal days, MM5and WRF both performed worse in simulating temperature and relative humidity during the fog and before fog, but no significant differences was found for wind speed.4) The near surface temperature inversion were more than40%in Nanjing and around30%in Anqing, frequently with multi-level temperature inversion. MM5and WRF both tended to over-predict the near surface temperature inversion, while miss the temperature inversion in the middle-to-upper boundary layer. In simulating the ground inversion layer, one side MM5performs better than WRF, on the other side both models performs better at Nanjing than at Anqing.The ground-level observations of the China Meteorology Agency routine meteorological network were also used to assess the performances of MM5and WRF in simulating fog. Some common used statistics measures (e.g. FAR, FB, TS, ETS) were calculated for fog simulation. The result of different techniques in fog predictions were compared and analysed.1) The MM5and the WRF fog region were roughly in line with actual situation, but there simulated fog regions have southerliness compared with actual situation. About two models, MM5performs better than WRF.2) By assessing and comparing with four different fog prediction techniques, the performances of fog simulation were a litter worse if LWC was only considered, but results were improved after relative humidity was considered. Compared with WRF, MM5has higher HR and FAR, in all WRF performs than MM5.3) Ensemble fog prediction from Ensemble prediction system enhanced the prediction performances. The reason is that ensemble system can effectively capture various fog types with multi-models and multi-physics schemes.4) Both model correctly simulate the onset and disappear of fog roughly. Compared with observations, simulated fog is occurred faster and is disperse quickly.