中文摘要：

大气污染物排放清单是了解区域污染物排放特征、准确模拟空气质量的重要资料,而工业点源是大气污染的重点排放源.通过收集相关活动水平信息和合理的排放因子,采用＂自下而上＂的方法建立了南昌市2014年点源大气污染物排放清单.结果表明,SO2、NOx、CO、PM10、PM2.5和VOC排放总量分别为29576.2、17115.1、25946.6、4689.4、922.9和1190.4 t,其中,金属炼制行业对SO2、CO和VOC的贡献最高,分别占37.75%、30.59%和38.45%;火电行业是NOx的主要来源,其贡献率为47%;水泥等建材制造行业对PM10和PM2.5排放贡献最高,分别为26%和25%.根据排放源污染物排放量及地理坐标信息,建立了0.4 km×0.4 km的污染物排放量空间分布特征图,结果表明,南昌市大气污染物排放较为集中,青山湖区北部和新建区北部是SO2、NOx、CO和VOC的主要排放区,而PM10和PM2.5的排放量相对分散,并在安义县出现排放高值区.通过将计算结果与统计数据结果进行对比,了解所估算清单的准确程度.对SO2和NOx的计算值和统计值进行统计分析,结果显示,NMB（标准化平均偏差）和NME（标准化平均误差）值均小于50%,清单计算精度较高.同时,为了解清单数据质量,对清单的不确定性进行定量分析,结果显示,SO2和VOC不确定性较低而PM10和PM2.5的不确定性相对较高,清单整体不确定性与其他研究结果相差不大.建议后期研究可以从提升基础数据质量和建立具有区域代表性的排放因子数据库着手,从而减小排放量的不确定性,获得精准可靠的大气污染物清单并应用于空气质量模型预报等更深入的研究.

英文摘要：

Atmospheric pollutant emission inventory is important for understanding regional emission characteristics and predicting air quality. The point sources from industries are the major sources of atmospheric pollution. Based on collected activity data and reasonable emission factors, an atmospheric pollutant emission inventory of the point source for Nanchang City in 2014 was established by the bottom-up approach. The results showed that the estimations for SO2, NOx, CO, PM10, PM2.5 and VOC emissions were 29576.2, 17115.1, 25946.6, 4689.4, 922.9 and 1190.4 t, respectively. Metal refining industry was the highest contributor to SO2, CO and VOC, which accounted for about 37.75%, 30.59% and 38.45% of the total emission of point source, respectively. About 47% emissions of NOx were from thermal power industry. Building materials manufacturing industry accounted for 26% and 25% of total PM10 and PM2.5 emissions, respectively. The established pollutants spatial distribution map with a resolution of 0.4 km×0.4 km showed that the emissions of SO2, NOx, CO and VOC were mainly distributed in the northern Qingshanhu and Xinjian districts. The high values of PM10 and PM2.5 emissions were located in the Anyi County. Comparison between calculated and statistical values of SO2 and NOx revealed that NMB （standard deviation） and NME （ normalized mean error） values were less than 50%, indicating a high accuracy of emission inventory. In order to understand the quality of inventory and improve the credibility, quantitative analysis of emissions inventory was carried out. There was relatively low uncertainty in SO2 and VOC emissions, and the values of both PMl0 and PMz5 were relatively high. The overall uncertainty of the inventory was not quite different from other research results. It is recommended to improve the quality of basic data and establish a regional representative of the emission factor database in future studies so as to obtain accurate and reliable inventory to be used in air quality forecasting model.