研究报告

  • 王辉,吴其重,周任君,钟方潜,易明建.结合GIS数据对合肥市大气污染状况的初步数值模拟研究[J].环境科学学报,2016,36(11):3902-3912

  • 结合GIS数据对合肥市大气污染状况的初步数值模拟研究
  • Preliminary numerical simulation of air pollution in Hefei with GIS data
  • 基金项目:国家自然科学基金项目(No.41305121);中央高校基本科研业务费专项资金资助
  • 作者
  • 单位
  • 王辉
  • 全球变化与地球系统科学研究院, 北京师范大学, 北京 100875
  • 吴其重
  • 全球变化与地球系统科学研究院, 北京师范大学, 北京 100875
  • 周任君
  • 地球和空间科学学院, 中国科学技术大学, 合肥 230026
  • 钟方潜
  • 地球和空间科学学院, 中国科学技术大学, 合肥 230026
  • 易明建
  • 1. 地球和空间科学学院, 中国科学技术大学, 合肥 230026;2. 安徽省环境科学研究院, 合肥 230022
  • 摘要:应用WRF-SMOKE-CMAQ模式结合地理信息系统(GIS),数值模拟研究了合肥市大气污染状况.在排放处理过程中,安徽省统计类面源排放采用GIS工具进行空间分配,并收集重点企业排放数据,以点源形式导入数值模型中,驱动多尺度空气质量模式CMAQ模拟合肥2014年12月大气污染现状.结合合肥市环境空气质量站点PM2.5、NO2日均浓度和小时浓度对比验证表明:①采用GIS数据空间优化获得的统计类排放面源较合理地呈现相关联排放的地理特征.②当前模式系统可以较好地模拟出2014年12月合肥市PM2.5污染物变化特征,尤其是12月20日-26日PM2.5污染累积和消散过程,站点的模拟实测两倍因子在63%~77%之间,合理反映出当月合肥市区PM2.5污染状况.③模式对于NO2模拟,在部分区域具有较好的模拟效果,但在部分区域模式只能模拟出NO2大致变化趋势;所有有效站点的比对结果平均偏差为21.92 μg·m-3,整体存在偏高现象;有40%站点FAC2在73%~88%之间,除个别观测数据异常较多站点外,其他站点FAC2在50%左右,这种差异是由于模式网格分辨率较低、排放源分配及站点选取引起的.
  • Abstract:WRF-SMOKE-CMAQ model system and GIS data are applied to study the air pollution in Hefei, Anhui province. The statistical area emissions in Anhui are spatially distributed by GIS tools, and the key industrial emissions are collected as point sources and processed into the SMOKE model. The SMOKE emissions are used to drive CMAQ for simulating air pollution in Hefei during December 2014. Comparing hourly and daily simulated concentration of PM2.5 and NO2 with the observational data, it is found that the area sources spatially distributed by GIS data are more reasonable in terms of their geographic characteristics. The model system shows good performance on simulating PM2.5 concentration in Hefei during December in 2014, especially the accumulation and dispersal processes from Dec 20th to Dec 26th with the FAC2 ranging from 63% to 76%. This illustrates that the model system can reasonably represent the situation of air pollution in the urban of Hefei city. The model system performs well on simulating NO2 in some regions, although it only catches the general tendency in other regions. The mean bias of simulated NO2 is 21.92 μg·m-3, generally higher than the observational concentration of all valid stations. The FAC2 ranges from 73% to 88% for 40% stations, and about 50% for the other stations, except for the stations with significant amount of invalid data. The differences are caused by the low resolution of model grid,distribution of emission sources and selection of stations.

  • 摘要点击次数: 1432 全文下载次数: 1316