珠三角空气质量模拟关键不确定性来源识别  

Identification of key uncertainties of air quality simulation in Pearl River Delta

作　　者：于凯阳 黄志炯 史博文 郑传增 白莉[2] 黄江荣 郑君瑜 YU Kaiyang;HUANG Zhijiong;SHI Bowen;ZHENG Chuanzeng;BAI Li;HUANG Jiangrong;ZHENG Junyu(Institute for Environmental and Climate Research,Jinan University,Guangzhou 511486;Guangdong Environmental Monitoring Center,Guangzhou 510220)

机构地区：[1]暨南大学环境与气候研究院,广州511486 [2]广东省环境监测中心,广州510220

出　　处：《环境科学学报》2020年第8期2952-2961,共10页Acta Scientiae Circumstantiae

基　　金：国家自然科学基金(No.41805068);广东省自然科学基金(No.2018A030310654)。

摘　　要：由于受到模型输入参数不确定性和模型结构不确定性的影响,利用大气化学传输模型模拟空气质量普遍存在偏差.对大气化学传输模型进行不确定性诊断分析、识别其关键不确定性来源是提高空气质量模拟的重要手段,本研究以珠三角为研究区域,利用HDDM-SRSM不确定性诊断方法量化了清单排放(SO2、NOx、VOCs和NH3)、边界条件浓度和气象(风速和温度)等模型输入参数不确定性对空气质量模拟的影响.结果表明:SO2、NO2和O3模拟受排放、边界条件和气象不确定性影响明显,其相对不确定性为15.19%~43.33%.在这些因素中,边界条件、风速和前体物(NOx和VOCs)排放是O3模拟的关键不确定性来源,但各因素不确定性贡献比例在昼夜存在明显差异.在夜间,风速不确定性对O3模拟影响增大,其平均贡献比例上升至29.6%,表明改进风速模拟有助于改善夜间O3模拟;在白天,NOx和VOCs排放不确定性对O3峰值浓度模拟影响增大,其平均贡献比例上升至32.26%,表明改进前体物排放模拟有助于提高白天O3模拟准确性.不同于O3,SO2、NO2模拟更容易受到排放不确定性的影响,尤其是垂直分配的不确定性.模拟与观测结果对比也表明,合理的烟囱参数设置可以降低源排放垂直分配不确定性,提高SO2和NO2的模拟效果.Affected by uncertainties of input parameters of model and model structures, atmospheric chemical transport models are generally biased when simulating air quality. Uncertainty diagnostic analysis of atmospheric chemical transport models and identification of their key uncertainty sources are important means to improve air quality simulations. In this study, taking Pearl River Delta as the target area, HDDM-SRSM uncertainty diagnostic method was applied to quantify the impact of uncertainties of input parameters of model on air quality simulation, such as inventory emissions(SO2、NOx、VOCs and NH3), O3 concentrations in lateral boundary conditions and meteorological parameters(wind speed and temperature). Results show that simulations of SO2, NO2 and O3 were significantly affected by the uncertainty of emissions, lateral boundary conditions, and meteorological, and their relative uncertainties were between 15.19% and 43.33%. Among these factors, lateral boundary conditions, wind speed, and precursor emissions were the key uncertainty sources of O3 simulation, but the proportion of uncertainty contributions of each factor varied significantly between day and night. At night, the impact of uncertainties of wind speed on simulation of O3 increased, and its average contribution ratio increased to 29.6%,implying that improving wind speed simulation can help improve O3 simulation at night;during the daytime, the impact of uncertainties of NOx and VOCs emissions on simulation of O3 peak concentrations increased, and its average contribution ratio increased to 32.26%, implying that improving precursor emissions simulation can help improve daytime O3 simulation accuracy. Unlike O3, simulations of SO2 and NO2 were more susceptible to the uncertainty of emissions, especially the uncertainty of vertical distribution. The comparison of simulation and observation results also showed that reasonable stack parameter settings can reduce the vertical distribution of point source emissions and improve simulations of SO2 and NO2.

关 键 词：空气质量模拟 不确定性分析 来源识别 垂直分配 

分 类 号：X51[环境科学与工程—环境工程]