Establishment and Application of Air Quality Statistical Forecasting Model - Taking Air Quality Data from City A as an Example

Abstract

By using the atmosphere automatic monitoring data and meteorological data of January and February 2017 in city A, 19 forecasting factors are selected and the statistical prediction model of winter air quality in city A is established by using the stepwise regression method. Forecast items include fine particulate matter (PM 2.5 ), inhalable particle (PM 10 ), sulfur dioxide (SO 2 ), nitrogen dioxide (NO 2 ), carbon monoxide (CO) average daily concentration and ozone (O 3 ) maximum 8h average concentration daily. From November 2017 to January 2018, the model was applied and revised in combination with human experience to carry out the environmental air quality forecast in city A. The comparison between the forecast results and the measured results showed that the level accuracy rate of the environmental air forecast results was 79.1%, and the accuracy rate of the primary pollutant was 73.6%.