Characteristics Analysis Of Air Quality Monitoring Data In Shanghai City And Its Application In Data Screening

The air quality has a direct influence on the people's living quality and the ecological living environment. The 45 air quality monitoring stations in the districts of Shanghai form the air quality monitoring network system in Shanghai which makes great contribution to a quick, accurate and complete picture of the air quality in Shanghai.This topic analyzes the statistical results of the monitoring data of three routine pollutants and six regular meteorological factors in the city of Shanghai in the year of 2008-2009. According to the national air quality standard, a comprehensive evaluation in the city of Shanghai has been conducted by four different methods. Based on the assessment of the meteorological conditions in Shanghai, the correlation analyses have done between the routine pollutants and meteorological factors. Then the establishment of some concentration models on the pollutant by adopting principal component analysis and polynomial regression has been done. Furthermore, by relying on the correlation characteristics of the pollutants, some screening methods on the regular air pollutants have been formulated which fit the actual situation of Shanghai and to offer some theoretical basis for the automatic data mining in the future.By analyzing the monitoring data of the routine pollutants and meteorological factors, some conclusions have drawn as following: the air quality of Shanghai is in good condition on the whole and meets the national standards for Level II. The comparison of four air quality evaluations indicates the changing laws of the air pollution in Shanghai: the air pollution is more serious in winter than in summer and is more serious in day than at night, and is more serious in downtown areas than in the rural areas. The meteorological factors and the routine pollutants are close in relevance, besides there are negatively correlated between the routine pollutants and air temperature, dew point, humidity, wind speed.Because of serious collinearity between factors, the forecasting models are adopt principal component regression to establish .Then the models are test by the monitoring data in the year of 2008-2009 which shows the advantage of high rate and good stability.Moreover, on the basis of the correlation among the routine monitoring pollutants, the screening method is established on abnormal monitoring data of pollutants by drawing a scatter diagram which is based on the monitoring data of PM₁₀ collected from 45 stations in 2008 and on the abnormal data located on the boundary curve and outside the boundary curve. In the end, we intend to adopt a joint screening system which depends primarily on the single index with one spot and on the single index with multiple spots as a supplementary.