The Study Of Air Pollution Prediction In Tianjin

Statistically analyzing by using of the Tianjin pollutant data from automatic monitoring and synchronized meteorological data during 1997-1999, and coordinating with the former Sulfur dioxide, Nitrogen oxide compound, and TSP, the environmental meteorological characteristic and the trait of the meteorological element which influences pollutant dilution and diffusion were elicited in Tianjin. The category and the source of the main pollutant in the Tianjin air, the pollution and the time-spatial distribution of the main pollutant, and the variation of the pollution in recent years were analyzed. Employing Tianjin tower data the emergence and disappearance of inversion under city boundary layer, and the variation trait of the wind field gradient were analyzed. The state of land and sea breezes was analyzed by the use of observed data.Air pollution rests greatly with the category of the pollution source and the emission characteristic in the forecast region, the integrated and systemic database of the pollution source emission is founded based on the investigation of the pollution source of 1995 and 1998 in the study area, which processed and saved all sorts of pollution source, and deals with pollution source information adapt to the numerical model. Through the analysis of the trait of the pollutant emission at different height, the contribution rate of pollution source at different height for the different pollutant was obtained.Two numerical systems were introduced. One is the high-resolution chemical model system (EMH) with the source of pollution input carried out by the Institute of Atmospheric Physics of the Chinese Academy of Sciences, and the other is City Air pollution prediction system (CAPPS) with anti-inverse pollution source intensity developed by the Chinese Academy of Meteorological Sciences. In the EMH models the meteorological field forecast used the ETA model and the M-βweather forecast model, and in the CAPPS model system used MM5 meso-scale model. After improvement and debugging, we used the EMH to conduct experimential forcast of the air pollution from 17 February to 13 March 2000. The results indicated the output of the meteorological model can reflect the changement of the weather pattern; Among all the six monitoring stations, the prediction accuracy rate of the SO₂ concentration is 55.3-74.8%; the prediction accuracy rate of the NO₂ concentration exceeded 70%; and that of the NO_X concentration is 59.1-84.5%; that of the TSP concentration is 62.1-87.2%.The daily mean PSI of SO₂,PM10 and NO₂ at Taifeng and Yongming road station in New Area near to seashore of Tianjin was calculated by CAPPS from December 2003 to February 2004,and the polluted potential prediction experiment in urban area during 14-20 September 2006 was conducted, the output of CAPPS is consistent with the development tendency of the polluted potential. The result of polluted potential prediction simulated by CAPPS was satisfactory. The predicted pollutant concentration accuracy rate of SO₂ is 70.4%; the predicted pollutant concentration accuracy rate of PM10 is 65.8% and that of NO₂ is 76.8%.An empirical prediction equation was founded by analyzing the correlation of pollution concentration and meteorological element. A prediction experiment was conducted by using the empirical prediction equation from 17 February to 18 March 2000,we obtained the predict accuracy rate of the pollution concentration: SO₂ is 93%; NO₂ is 80% and TSP is 73% .To better validate the forecasting results of the various methods of predicting air quality in daily operation, the prediction results of EMH(chemical model system), CAPPS, MRA (multivariate regression statistical forecasting methods) and ECF (empirical prediction method is consensus forecast) were confirmed by applying three sample records pollutants (SO₂, NO₂ and PM10) concentration in Tianjin urban areas including New Area near to seashore at 12 automatic monitoring stations in 2002.The best forecasting method was ECF. But results simulated by EMH were obviously inferior to that by using other several methods, except the forecasting accuracy rate of NO₂ corresponding with other methods. The colligation score of CAPPS and that of MRA was nearly the same, but the forecast pollutant rank accurate rate of CAPPS was obviously higher than MRA. In addition to PM10 prediction accuracy rate just was under MRA, the index prediction accuracy rate of NO₂ and SO₂ of CAPPS was also higher than that of MRA methods.