基于B-P神经网络的城市大气SO_2浓度预测(英文)

Base on the principle and method of B-P neural network,the prediction model of SO2 concentration in urban atmosphere was established by using the statistical data of a city in southwest China from 1991 to 2009,so as to forecast atmospheric SO2 concentration in a city of southwest China.The results showed that B-P neural network applied in the prediction of SO2 concentration in urban atmosphere was reasonable and efficient with high accuracy and wide adaptability,so it was worthy to be popularized.

Effect of traffic restriction on atmospheric particle concentrations and their size distributions in urban Lanzhou, Northwestern China

During the 2012 Lanzhou International Marathon, the local government made a significant effort to improve traffic conditions and air quality by implementing traffic restriction measures. To evaluate the direct effect of these measures on urban air quality, especially particle concentrations and their size distributions, atmospheric particle size distributions （0.5-20μm） obtained using an aerodynamic particle sizer （model 3321, TSI, USA） in June 2012 were analyzed. It was found that the particle number, surface area and volume concentrations for size range 0.5-10 μm were （15.0±2.1） cm-3, （11.8±2.6） μm2/cm3 and （1.9±0.6） μm2/cm3, respectively, on the traffic-restricted day （Sunday）, which is 63.2%, 53.0% and 47.2% lower than those on a normal Sunday. For number and surface area concentrations, the most affected size range was 0.5-0.7 and 0.5-0.8 μm, respectively, while for volume concentration, the most affected size ranges were 0.5-0.8, 1.7-2.0 and 5.0-5.4 μm. Number and volume concentrations of particles in size range 0.5-1.0μm correlated well with the number of non-CNG （Compressed Natural Gas） powered vehicles, while their correlation with the number of CNG-powered vehicles was very low, suggesting that reasonable urban traffic controls along with vehicle technology improvements could play an important role in improving urban air quality.

Pollutant concentrations and pollution loads in stormwater runoff from different land uses in Chongqing

To investigate the distribution of pollutant concentrations and pollution loads in stormwater runoff in Chongqing,six typical land use types were selected and studied from August 2009 to September 2011.Statistical analysis on the distribution of pollutant concentrations in all water samples shows that pollutant concentrations fluctuate greatly in rainfall-runoff,and the concentrations of the same pollutant also vary greatly in different rainfall events.In addition,it indicates that the event mean concentrations （EMCs） of total suspended solids （TSS） and chemical oxygen demand （COD） from urban traffic roads （UTR） are significantly higher than those from residential roads （RR）,commercial areas （CA）,concrete roofs （CR）,tile roofs （TRoof）,and campus catchment areas （CCA）;and the EMCs of total phosphorus （TP） and NH3-N from UTR and CA are 2.35-5 and 3 times of the class-III standard values specified in the Environmental Quality Standards for Surface Water （GB 3838-2002）.The EMCs of Fe,Pb and Cd are also much higher than the class-III standard values.The analysis of pollution load producing coefficients （PLPC） reveals that the main pollution source of TSS,COD and TP is UTR.The analysis of correlations between rainfall factors and EMCs/PLPC indicates that rainfall duration is correlated with EMCs/PLPC of TSS for TRoof and TP for UTR,while rainfall intensity is correlated with EMCs/PLPC of TP for both CR and CCA.The results of this study provide a reference for better management of non-point source pollution in urban regions.