本期目录

  • 熊亚军,徐敬,孙兆彬,李梓铭,吴进,尹晓梅,乔林,赵秀娟.基于数据挖掘算法和数值模拟技术的大气污染减排效果评估[J].环境科学学报,2019,39(1):116-125

  • 基于数据挖掘算法和数值模拟技术的大气污染减排效果评估
  • Air pollution reduction effect evaluation based on data mining algorithm and numerical simulation technology
  • 基金项目:国家重点研发计划(No.2016YFC0202100);北京自然科学基金(No.8161004);国家自然科学基金(No.41575010);国家科技支撑计划(No.2014BAC16B04)
  • 作者
  • 单位
  • 熊亚军
  • 京津冀环境气象预报预警中心, 北京 100089
  • 徐敬
  • 中国气象局北京城市气象研究所, 北京 100089
  • 孙兆彬
  • 中国气象局北京城市气象研究所, 北京 100089
  • 李梓铭
  • 京津冀环境气象预报预警中心, 北京 100089
  • 吴进
  • 京津冀环境气象预报预警中心, 北京 100089
  • 尹晓梅
  • 京津冀环境气象预报预警中心, 北京 100089
  • 乔林
  • 京津冀环境气象预报预警中心, 北京 100089
  • 赵秀娟
  • 中国气象局北京城市气象研究所, 北京 100089
  • 摘要:近年来,京津冀地区采取了大量污染减排措施进行大气污染治理,如何客观评估减排效果是目前大气环境领域的研究难点.为准确评估大气污染过程的减排效果,本文利用北京地区常规气象资料、国控站PM2.5浓度资料,遴选了北京地区2018年3月11—14日和2013年3月14—17日两次空气污染过程,计算了大气容量系数、静稳指数,并利用KNN数据挖掘算法和WRF-Chem模式,对比分析了有无减排条件下的PM2.5日均浓度.结果表明:两次空气污染过程的天气形势和局地气象条件较相似,就大气热力和动力的垂直结构来看,2018年空气污染过程比2013年空气污染过程的大气稳定性更强、边界层高度更低、环境容量更小,但PM2.5峰值浓度却显著下降,平均浓度明显降低,PM2.5小时浓度的增长趋势相对平缓,重污染持续时间缩短.KNN数据挖掘算法减排评估结果显示,该方法能够较好地预测PM2.5日均浓度的变化趋势,2018年3月11—14日,在减排和不减排情景下PM2.5日均值分别为171和229 μg·m-3,减排使得污染过程PM2.5平均浓度下降了25.3%.数值模拟结果与KNN数据分析结论吻合,进一步验证了减排措施的有效性.综合看来,2018年空气污染过程中PM2.5浓度相比历史相似气象条件下的污染过程显著降低,这是长期大力度减排效果的体现.
  • Abstract:Air pollution reduction measures have been recently adopted in the Beijing-Tianjin-Hebei region. Objectively assessing atmospheric pollution reductions is difficult in the atmospheric environment field. This study used meteorological data, the PM2.5 concentration data in Beijing, and two air pollution processes (March 11—14, 2018 and March 14—17, 2013) in the Beijing area to calculate the atmospheric capacity coefficient and static stability index for assessing the effects of the air pollution reduction measures. Average daily PM2.5 concentrations were analyzed under an emission reduction condition by using the KNN data mining algorithm and WRF-Chem mode. The results showed that the weather conditions and local meteorological conditions were similar for the two air pollution processes from the perspective of atmospheric thermal and dynamic vertical structures. For the 2018 process, the atmosphere was more stable, the boundary layer height was lower, and the environmental capacity was smaller than that of the 2013 process. Peak and average PM2.5 concentrations dropped sharply and the PM2.5 hourly concentration grew relatively slowly. Moreover, the duration of heavy pollution was shortened. Evaluating emission reduction results based on the KNN data mining algorithm showed that this method can reliably predict the changing tendency of daily average PM2.5 concentrations. The average PM2.5 concentration during March 11—14, 2018 was 171 and 229 μg·m-3 under the emission reduction and nonemission reduction conditions, respectively. The PM2.5 concentration decreased by approximately 25.3% because of the emission reduction measures. Moreover, numerical simulation results were consistent with those from the KNN data analysis, which further verified the effectiveness of the emission reduction measures. In summary, the PM2.5 concentration for the 2018 pollution process was significantly lower than that in similar meteorological conditions in history, which is a reflection of the long-term, large-scale emission reduction efforts.

  • 摘要点击次数: 184 全文下载次数: 267