Rapid Response Of Atmospheric Fine Particle Concentration To Ambient Emission Change In The Yangtze River Delta In Winter

Urban areas in China are still prone to frequent regional heavy particulate pollution events in autumn and winter.A precise and scientific control of fine particulate matter(PM_2.5)pollution depends on the establishment of quantitative relationship between source pollution and pollution concentration.At present,the traditional scenario analysis method based on chemical transmission model is difficult to meet the effect evaluation of schemes for mass emission control.It is necessary to explore a new method to establish a fast response relationship between source pollution and pollution concentration.This study takes the Yangtze River Delta region as the research object,and selects January 2015 as the representative heavy pollution period.With eight provinces in the east of China,five pollutant types,and five emission reductions,a large number of different source emission scenarios are constructed.The sensitivity analysis technique of chemical transport model is used to obtain the response data set of large quantities of source emission and receptor concentration.Aiming at the massive response data set,a convolutional neural network is used to establish a deep learning model of rapid response of receptor concentration and source emission,which is applied to the long-term joint prevention and control work in the Yangtze River Delta region.The research results show that:（1）The prediction results of the convolutional neural network model constructed are basically consistent with the original results of the chemical transmission model simulation.The overall average overestimate is about 6.7%,and R² is above 0.99.In terms of response speed,the predicted time of the trained source-receiver model for different emission inputs is about 65ms,which is four orders of magnitude lower than the calculation time of traditional chemical transmission models,achieving quasi-real-time PM_2.5 concentration prediction.（2）The application results of the model in the Yangtze River Delta show that PM_2.5 pollution in the Yangtze River Delta region in winter is most affected by primary PM_2.5 emissions in four provincial regions within the region,and outside the region,it is mainly affected by ammonia and ammonia in Henan and Shandong provinces.Once the impact of PM_2.5emissions,the gaseous pollutants in Hebei and Shanxi also have a certain impact on the Yangtze River Delta.If the local PM_2.5 emissions are reduced by 60%each time,the average PM_2.5 mass concentrations in Anhui,Jiangsu,Shanghai and Zhejiang are expected to decrease by 23.24,18.32,15.17 and9.07μg/m³,respectively.This paper is expected to provide new technical means for establishing a rapid relationship between the concentration of atmospheric fine particles and pollution emissions.The results of its application can also provide technical support for the formulation of scientific,accurate,and effective heavy pollution reduction programs in the Yangtze River Delta region.