| South Central Hebei,an important part of Beijing-Tianjin-Hebei Region,is one of the regions where the air pollution is the most serious.There,heavy pollution frequently happens in autumn and winter,which has greatly threatened the health of local residents.What's worse,it adversely affects its investment environment and economic competitiveness.Under this background,this paper,depending on the concentration of air pollutants and meteorological data in 2013~2015 of South Central Hebei,studies the spatial-temporal distribution of the air pollution;then,this paper sorts out the weather resulting in heavy air pollution in South Central Hebei.This paper also analyzes the weather feature reflected by the two typical heavy pollutions,as well as the impact on pollutants by meteorological factors;at last,by analyzing the stepwise regression analysis on meteorological factors and pollutants,this paper selects out optimal factors and sets up an air quality forecast model for South Central Hebei based on BP neural network.Therefore,this article can provide technological supports for implementing local air pollution forecast and improve the forecast level.Here below are the major research results:(1)In South Central Hebei,the time changing feature of air pollution is: from 2013 to 2015,AQI in South Central Hebei kept decreasing,which meant that the air quality was being improved gradually;by analyzing the monthly change of AQI,we can see that the heavy pollution was mainly occurred from Dec.to next Jan..After analyzing the annual change of the six pollutants,it was found that the change of five pollutants(PM2.5,PM10,SO2,NO2 and CO)were similar to each other,the change was obvious in winter half year while the change was slight in summer half year;however,the annual change of O3 is totally opposite to the other pollutants---the change was slight in winter half year while obvious in summer half year.After analyzing the quarterly change of the six pollutants,it was found that the concentration of PM2.5,PM10,SO2,CO and NO2 was the highest in winter while the lowest in summer;the quarterly change of PM2.5,SO2 and CO was relatively obvious,whose concentration in winter was much higher than that in summer;the quarterly change of concentration of NO2 was not obvious;the concentration of O3 was the highest in summer while lowest in winter,which meant that there was an obvious quarterly change.(2)In South Central Hebei,the spatial distribution of air pollution was: from the spatial distribution of AQI,we can see that South Central Hebei mainly suffered a middle level pollution,which was AQI in north > AQI in south > AQI in central;from the view of quarterly distribution,South Central Hebei mainly suffered a mild contamination in spring,where the distribution was AQI in northwest > AQI in southwest > AQI in northeast > AQI in southeast;air quality suffered a mild contamination in summer,and the air quality in summer was a little worse than that in spring,pollution in north Shijiazhuang,southeast Hengshui,northeast Xingtai,north Handan and south Handan was severe;in autumn,air quality suffered a middle level pollution,pollution in north Shijiazhuang,southwest Hengshui,central Xingtai and east Handan was severe;in winter,air quality suffered a heavy pollution,pollution in most Shijiazhuang,northwest Hengshui,southwest Xingtai,north Handan and south Handan was severe.(3)By classifying the weather at the time of heavy pollution(19 times)occurred from 2013 to 2015 in South Central Hebei,we can see that the weather situation would be affected by ground system,which mainly included: low pressure,high pressure bottom;uniform pressure field and high pressure rear.The low pressure type was controlled by the low pressure belt,and the pressure field was weak,Such polluted weather was 28% of the total;for high pressure bottom,the cold air was from east of Baikal Lake;such polluted weather was 28% of the total,such a weather was easy to result in heavy fog in the morning;for uniform pressure field,the field of pressure was controlled by uniform pressure field,there was no obvious weather system,such polluted weather occurred was 24% of the total;for high pressure rear was 20% of the total,such polluted weather was easy to get southwest warm and wet stream flow to north,which would result in heavy pollution.(4)For the aforementioned heavy pollution occurred from 2013 to 2015,there were two typical heavy pollution(one occurred from Jan.29,to Feb.2 2014,while the other one occurred from Feb.19 to 27,2014).After an analysis on these two typical heavy pollutions,it introduced the daily change of concentration of AQI,PM2.5 and PM10.It was found that the higher average wind speed was,the lower concentration of the pollutants would be,which reflected that wind speed was a key factor affecting the air pollution.For the cause and support mechanism of the two heavy pollution,it analyzed the high altitude trend fields 500 hPa,700 hPa,850 hPa and ground trend field.It was found that the field trend of these two times of heavy pollution were low-uniform pressure and high pressure bottom respectively(for the second one,high pressure bottom appeared lately).(5)It introduces the process to set up BP neural network air quality forecast model based on good original sample mechanism,by taking Shijiazhuang and Xingtai as an example and on the basis of calculating the forecast average daily difference of air pollution,a forecast model for the concentration of air pollution is established,whose weather forecast factor is the pollutant concentration of the previous day and the daily average value of the meteorological factors on that day.Then,a forecast on the air quality by combining both the meteorological factors can be done.Take the winter air quality model as an example,the correctness on air quality forecast shall be verified.From the result,we can know that the forecast rating accuracy on SO2 and O3 in Shijiazhuang and Xingtai were 90% correct or more,forecast rating accuracy on PM2.5 and PM10 was 80% correct or more.Generally speaking,correctness of air quality forecast in Shijiazhuang was better than that in Xingtai.From the forecast,we can know that for the air quality forecast in winter,the weather classification for heavy pollution should be taken into account.That is when there is a kind of weather which will result in heavy pollution(just as it is mentioned in Chapter 4),it is mandatory to do a proper man-made correction on the weather forecast according to the air pollution status which has been occurred in this history.Besides,the first forecast model mentioned in Chapter 6 shall be compared with the forecast model whose factor is the daily average difference of the meteorological factors occurred in two adjacent days(the meteorological factors in the two adjacent days have a good correlation).After the comparison,we know that the first model whose meteorological factors are the pollutant concentration and daily average value of meteorological factors of last day has a better forecast effect than the second model whose meteorological factor is the daily average difference of the meteorological factors in two adjacent days.Compared with the second model,the first model is more accurate in forecasting the air quality rating and primary pollutants.Therefore,the first model can be used as an optimized proposal for forecast. |
PDF Full Text Request