Spatial And Temporal Distribution Characteristics Of Air Quality And Prediction Of PM2.5 Concentration In Beijing-Tianjin-Hebei Region

In recent years,air pollution has become more and more serious.Among them,PM2.5 contributes the most to air quality pollution.The situation of preventing and controlling air pollution is getting more and more serious for local governments and environmental protection departments,especially for the beijing-tianjin-hebei region.How to prevent and control air quality in Beijing Tianjin Hebei region has become the focus of attention of the whole nation.We will control air pollution in the beijing-tianjin-hebei region and improve air quality.Understand and master all kinds of pollutants in atmospheric environment changes in time and space distribution characteristics and distribution rule,analyze the spatial and temporal variation rules of PM2.5 concentration in the beijing-tianjin-hebei region,and study a reasonable prediction model of PM2.5 concentration.In this paper,the method of using GIS spatial interpolation analysis,comparative analysis of the studied area from 2015 to 2018,time and space distribution characteristics of PM2.5 concentrations,and the air monitoring station of Beijing shunyi new town is taken as an example and the PM2.5 concentration is predicted for 0-72 hours with the combination of long and short time memory neural network model.In this paper,the main research work is as follows:(1)Base on the analysis of the seasonal spatial and temporal variation characteristics of PM2.5 concentration in the study area from 2015 to 2018,this paper concludes that the maximum PM2.5 concentration in the beijing-tianjin-hebei region decreases with the increase of time,and the spatial impact range of pollution shrinks.Pollution degree and pollution range in different seasons:winter>autumn>spring>summer;There is no significant difference in the distribution range of PM2.5 high values in the same season.In terms of spatial distribution,PM2.5 concentration in the study area is from low to high:North>middle>South(2)This paper analyze the correlation between the concentrations of various pollutants and the correlation between meteorological conditions and the concentrations of various pollutants in the study area from 2015 to 2018.The analysis show that the correlation coefficients of AQI with PM2.5 and PM10 pass the significance level test ofP<0.01,showing a strong positive correlation with PM2.5 and PM10 concentrations,and the correlation coefficients are 0.877 and 0.947,respectively.In other words,PM2.5 and PM10 have the largest impact on air quality(AQI)and were decisive pollutants.(3)The correlation between each pollutant concentration and the correlation between each meteorological factor and each pollutant concentration from 2015 to 2018 is analyze,and the correlation coefficients of AQI,PM10,SO2,CO,NO2,O3,air pressure,wind speed,wind direction,temperature,relative humidity and precipitation with PM2.5 were obtained.It can be found that among the five pollutants,PM10,CO,NO2 and PM2.5 concentration have the highest correlation coefficient,that is,the correlation degree is closer.Among the six surface meteorological factors,the correlation coefficient between surface pressure,temperature,relative humidity,2-minute average wind speed and PM2.5 concentration is the highest,that is,the correlation degree is closer,compared with the 2-minute average wind direction and precipitation.(4)The input data of the model are the hourly concentrations of various pollutants at the air quality monitoring station of shunyi new town(station no.1008A)in Beijing from 2015 to 2018 and the average hourly meteorological factors at the nearest meteorological monitoring station in terms of spatial distance.A prediction model of PM2.5 concentration with multiple input factors and a single output was established base on the long and short time memory neural network model,and the PM2.5concentration of the air quality monitoring station in shunyi new city of Beijing is predicted for 0-72 hours,and the accuracy of the prediction results and the actual values is verified.Research results show that the model can by higher precision of single site of PM2.5 concentrations.