Comparative Analysis And Simulation Of Two Continuous Severe Haze Weather Processes In Beijing-Tianjin-Hebei

With the continuous expansion of the scale of cities in China,the continuous development of urbanization and rapid economic growth,the problem of air pollution is becoming more and more serious,and fog and haze weather occur frequently.The Beijing-Tianjin-Hebei region,the Yangtze River Delta,the Pearl River Delta and other regions have repeatedly experienced air pollution incidents,causing great harm and losses to people and society.On December 16,2016 solstice 21 and November 12,2018 solstice 15,China’s Beijing-Tianjin-Hebei region occurred persistent severe haze weather.In this paper,a variety of data are used to analyze the haze pollution in the Beijing-Tianjin-Hebei region.At the same time,the circulation background,meteorological conditions,PM_2.5and visibility of the two severe haze weather processes are comprehensively analyzed and simulated by using laser radar,wind profile radar,WRF-CAMx model and neural network method.The results show that the number of haze days in China and all over the country presents a trend of fluctuation and decline.PM_2.5and PM₁₀concentrations in the Beijing-Tianjin-Hebei region show consistent seasonal variations,with higher concentrations in winter and lower concentrations in summer.At the same time,the diurnal variation features of certain bimodal characteristics appear peak in the morning and night.Compared with the Beijing-Tianjin-Hebei region,Beijing had the lowest PM_2.5and PM₁₀concentrations overall,while Hebei had the highest.The two typical processes are controlled by the zonal circulation in front of the high pressure ridge in the upper air.The warm ridge extends northward in the lower air.The ground is located in the southeast of the high pressure,which is controlled by the uniform pressure field,and the pressure gradient is small,which is affected by the southerly wind.The existence of temperature inversion stratification and weak subsidence motion inhibit the diffusion of pollutants in the vertical direction.Small ground wind,high relative humidity,low mixing layer height,is not conducive to the horizontal and vertical diffusion of pollutants.These are the common reasons for the long duration and high intensity of the two haze weather processes.However,the differences in the intensity and duration of the two haze weather processes are caused by the difference in the maintenance time and adjustment period of atmospheric circulation and the intensity of meteorological conditions.According to the observations of laser radar and wind profile radar,it is found that the low and stable boundary layer height and the low-level westerly wind are conducive to the convergence and retention of pollutants.At the same time,there were foreign pollutants imported from Hebei and Shandong in these two pollution processes.In the later stage,as the cold air moves south,the temperature decreases and the wind speed increases,and the haze process quickly ends.However,in the haze process in 2016,the transmission rate of pollutants is lower than that in the haze process in 2018.Moreover,the vertical cutting edge of wind speed is large,the wind direction changes slowly,and the accumulation time of pollutants is long.Therefore,the duration is longer and the haze process is more serious.Compared with the meteorological elements in the past 30 years,most of the anomaly values of the haze process in 2016 are greater than those of 2018,and the pollution degree is more serious.The comprehensive influence of various meteorological elements leads to the severity and anomaly of the haze process.The PM_2.5concentration in the two haze processes simulated by WRF-CAMx model was in good agreement with the observation,but the simulation results of the 2016 process were underestimated to a certain degree,while the simulation results of the 2018 process were slightly higher.From the perspective of pollution sources,there are some differences in the sources of the two processes.During the course of 2016,the pollution sources of PM_2.5changed greatly.On December 16 and 20,local pollution in Beijing dominated.On December17 and 19,the contribution ratio of local pollution in Beijing decreased significantly,and the main pollution was affected by transmission.In 2018,the impact of local pollution sources was low and was mainly affected by transmission.In this study,the LSTM neural network method was used to forecast visibility,and it was found that the model modified by ECMWF forecast field had better forecasting ability than other single air quality models.The forecast results of the visibility corrected model based on LSTM neural network method show that the average error is the lowest about 6km,and the forecast performance is relatively stable under different ages.After correcting the forecast results of the numerical model,the correlation coefficient of LSTM model is increased to more than 0.6,and the level of visibility forecast is significantly improved.