The Improvement Of Meteorological Data Assimilation Of WRF-3DVAR For The Simulation Of Near-surface Wind Speed And Heavy Atmospheric Pollution

Due to the combined effects of human activities and meteorological elements,air pollution incidents occur frequently in the Beijing-Tianjin-Hebei region in the North China Plain,where PM_2.5 is one of wintertime primary air pollutants in this region.The structure of the atmospheric boundary layer and the near-surface wind field are important driving factors for the transport and diffusion of air pollutants.Therefore,the simulation accuracy of atmospheric boundary layer and near-surface wind speed largely affects the air quality simulation.This paper selects January 13 to February 13,2020 as the study period of PM_2.5 pollution in winter.Based on the optimal simulation of PM_2.5 in the Beijing-Tianjin-Hebei region during the study period,the atmospheric boundary layer parameterization scheme is selected,then by using the meteorological observation data,WRF-Chem with WRF-3DVAR（three-dimensional variation）is conducted to t apply the meteorological data assimilation in air quality simulation,evaluating the effects of meteorological observation data assimilation on near-surface wind speed and PM_2.5 in the process of air pollution.（1）Evaluation of the atmospheric boundary layer parameterization scheme on the simulation of near-surface wind field and PM_2.5 in Beijing-Tianjin-Hebei regionIn order to study the influence of the boundary layer parameterization scheme on the modeling near-surface wind speeds and PM_2.5concentrations,three boundary layer parameterization schemes YSU,MYJ and ACM2 were selected to simulate the same pollution process in the Beijing-Tianjin-Hebei region.The evaluations show that the boundary layer scheme had obvious differences in the simulation of near-surface wind speed.Beijing,Tianjin,Cangzhou,Shijiazhuang and Xingtai had the smallest deviation between the simulated wind speed of the YSU scheme and observations.This scheme also had the best simulation effect on the near-surface wind speed in Beijing-Tianjin-Hebei region.The simulations of PM_2.5concentrations show that the three boundary layer schemes could reflect the daily changes of PM_2.5 well.Using statistics to evaluate the impact of three boundary layer schemes on PM_2.5simulation,it was found that the simulation effect of YSU scheme was the best of three schemes regardless of whether clean days or pollution days,followed by ACM2 scheme.The statistical analysis of the daily variation of PM_2.5 concentrations,wind speeds and boundary layer heights in Beijing-Tianjin-Hebei region found that PM_2.5changes was inversely related to wind speed and boundary layer height;further analysis of hourly changes in PM_2.5 concentrations,wind speeds,boundary layer heights and air temperature in pollution process（January 17-19）indicated that the increase of wind speed,temperature rise,and boundary layer height could reduce PM_2.5 concentration,the accuracy of simulation on meteorological elements would significantly affect PM_2.5 simulation results.Therefore,selecting the optimal atmospheric boundary layer parameterization scheme is one of the key issues in improving air quality simulation.（2）Improvement of WRF-3DVAR meteorological data assimilation on near-surface wind speed and PM_2.5simulations in winter in Beijing-Tianjin-Hebei regionBased on 12h forecast products simulated by the WRF model in Beijing-Tianjin-Hebei region,a statistic calculation of the background error covariance（B）is conducted using the NMC method to obtain the eigenvectors,eigenvalues and length-scales of B and compare the statistical structure features of B in each triple-nested model domain.The result indicated that background error covariance was closely related to resolution and geographic characteristics of simulated region.Calculating background error covariance of specific region has an important role in the research of variational assimilation.Based on the WRF-3DVAR and WRF-Chem models,using conventional meteorological observation data for cyclic assimilation and simulation forecasts,evaluating the assimilation of conventional observation data on near-surface wind speed in the study period from January 13to February 13,2020.The results show that the assimilation and non-assimilation control experiments had reasonable simulation for the daily wind speed changes in Beijing-Tianjin-Hebei region,but the simulation is higher than observation data.Using statistical evaluations,it was found that the assimilation of meteorological observation data had an improved effect on the simulation of near-surface wind speed.Selected the pollution process from January 17 to19,2020 for further detailed analysis.During this pollution period,the improvement of near-surface wind speed simulation in Beijing,Tianjin and Baoding was more obvious than other cities.Assimilation of conventional observational data had significant improvement within 36h simulation.The control experiment without data assimilation had overestimated the simulation of PM_2.5 in polluted area of Beijing-Tianjin-Hebei region.The assimilation experiment weakened the convergence intensity of the wind field simulated by the control experiment,and increased the boundary layer height at the same time,which caused the simulated PM_2.5concentration reduced.The improvement of assimilating conventional observational data was still limit the PM_2.5 simulation,and the improvement effect is most obvious in 12-24 hours.