The Features And Forecast Model Of Air Pollution Index Of Urumqi City

The problem of air pollution is always one of the closely watched environmentalissues, especially the main pollutants, such as PM10and SO₂, have seriously harmedto people’s health. The air pollution in Urumqi is very serious, the city had been listedas one of the world’s10most serious polluted cities, its SO₂, NO₂and PM10pollutionwere ranked the top three in China in2010, to research the air pollution and itsforecast has very important significance to improve the air quality in the region. Thearticle researched the air pollution in Urumqi from the view of air pollution index （theAPI）, selected the API data and ground meteorological elements data form2001to2010in Urumqi, made the year, quarter and month as cycle to analysis the change ruleof the air pollution in Urumqi in recent ten years detailedly. To use the DanielInspection method to do quantitative analysis to the API trends, to use correlationanalysis method to discuss how meteorological elements influenced the API indifferent season, based on extracted principal component form meteorologicalelements by using principal component analysis method, to use multiple regressionanalysis to establish the API prediction model. The result of the air pollution statusand law in Urumqi can provide theoretical basis for testing the effectiveness of themeasures of air pollution control in Urumqi and for the future governance of thepollution, at the same time, the API prediction model can provide theoretical referencefor the air pollution forecast in the area in the future.The results show that:⑴In recent decades, the downward trend of API in Urumqi wasnot significant, the air pollution has not been effectively controlled, the days of LevelII were most, and significantly increased year by year, the percent of unpolluted daysmaintained about70%, every year,81.7%of the pollutions were caused by PM10. Themonthly change trend of API was like parabolic, it monthly declined before July andmonthly rised after July, the air quality was worst in January and was best in July,every month,84.7%of the pollutions were caused by PM10. The changes of API havesignificant seasonal variations, it was maximum in winter and was minimum insummer, pollutional weather mostly occured in winter, the interannual fluctuations of API was mainly from the fluctuations in winter.⑵The meteorological elements thathave a significant negative correlation with API and are arranged according to the sizeof the order of the correlation coefficient are that: Daily minimum temperature>dailymaximum temperature>average vapor pressure>small evaporation>sunshinehours>average temperature>average wind speed, the bigger these meteorologicalelements are,the better the air quality is. The meteorological elements that have asignificant positive correlation with API and are arranged according to the size of theorder of the correlation coefficient are that: large evaporation>minimum relativehumidity>average relative humidity>Daily maximum site Pressure>Daily minimumsite Pressure>average site Pressure>precipitation. Wind speed, air pressure, airtemperature, relative humidity have obviously different influence to API in differentseasons.⑶The regression prediction model of API is: API=-18.023Z1-8.770Z2+103.918（R²=0.306），Z1and Z2are the main ingredient whichare extracted from the meteorological elements, The model’s prediction relative errorof API is40%, and its level prediction accuracy is63%, so the predicted result issatisfactory, and it has practical value.