环境科学  2023, Vol. 44 Issue (7): 3669-3675   PDF    
基于响应曲面法的臭氧生成敏感性分析
朱禹寰1, 陈冰2, 张雅铷1, 刘晓1, 李光耀1, 舍静1, 陈强1     
1. 兰州大学大气科学学院, 半干旱气候变化教育部重点实验室, 兰州 730000;
2. 聚光科技(杭州)股份有限公司, 杭州 310000
摘要: 准确判断臭氧(O3)生成敏感性对O3污染成因分析和防控对策的制定至关重要.首次利用响应曲面方法设计最优试验方案, 基于盒子模式模拟结果, 快速量化O3对其前体物变化的响应.结果表明, CO对O3有正贡献, NOx和VOCs与O3呈现显著非线性关系, 当φ(VOCs)与[φ(NOx)-13.75]比值大于4.17时, 为NOx控制区, 小于4.17时, 为VOCs控制区; 烯烃为影响O3生成的关键VOCs组分, 当φ(烯烃)与[φ(NOx)-15]比值小于1.10且φ(烯烃) < 35×10-9时, 烯烃有利于O3的生成.响应曲面法在多因素和其交互作用对O3生成影响的研究中取得了良好效果, 为高效判断O3敏感性提供了新的思路和方法.
关键词: 臭氧(O3)      响应曲面      O3敏感性      盒子模式      一氧化碳(CO)      挥发性有机物(VOCs)     
Sensitivity Analysis of Ozone Formation Using Response Surface Methodology
ZHU Yu-huan1 , CHEN Bing2 , ZHANG Ya-ru1 , LIU Xiao1 , LI Guang-yao1 , SHE Jing1 , CHEN Qiang1     
1. Key Laboratory for Semi-Arid Climate Change, Ministry of Education, College of Atmospheric Sciences, Lanzhou University, Lanzhou 730000, China;
2. Focused Photonics(Hangzhou) Incorporated Company, Hangzhou 310000, China
Abstract: Identifying the nonlinear relationship between O3 and its precursors accurately plays an important role for the policy-making of O3 pollution control. In this study, the response surface methodology based on the box model simulation was used to quickly and efficiently quantify the O3 response to their precursors with the optimal experimental design. The results showed that CO had a positive contribution to ozone generation, whereas NOx and VOCs had a significant nonlinear relationship with O3. When the ratio of φ(VOCs) to [φ(NOx)-13.75] was greater than 4.17, the ozone formation regime was NOx-limited and became VOCs-limited when the ratio was less than 4.17. Olefin was the key VOCs' component to affect the formation of O3; when the radio of φ(olefin) to [φ(NOx)-15] was less than 1.10 and the value of the φ(olefin) was less than 35×10-9, olefin went far towards generating O3. Response surface methodology demonstrated that it can be well used to explore the influence of multiple factors and their interactions on O3 formation and provides a new approach for efficient O3 sensitivity analysis.
Key words: ozone(O3)      response surface      O3 sensitivity      box model      carbon monoxide(CO)      volatile organic compounds(VOCs)     

对流层臭氧(O3)是环境空气中主要的污染物, 对人体健康、植物生长和气候均有不利影响[1~3]. O3是氮氧化物(NOx)、挥发性有机化合物(VOCs)和CO等前体物发生复杂光化学反应生成的二次污染物[4], 其和前体物之间存在高度非线性关系[5], 同时颗粒物[6~8]和气象条件[9]等因素通过改变前体物的浓度、物种组成和化学反应速率等也对O3的生成和消耗产生影响.自我国实施《大气污染防治行动计划》等严格大气污染防治政策以来, SO2、NO2和CO等污染物浓度明显下降, PM2.5污染问题得到了较大的改善, 但是我国大多数地区O3浓度却呈现波动上升趋势[10~12]; 2020年初COVID-19病毒大流行, 疫情管控期间人为活动受到限制使得一次污染物的排放量大幅减少, 但在部分地区O3仍保持较高浓度[13], 可见对前体物的不当控制可能会导致O3浓度的升高, 所以准确判断O3生成对前体物的敏感性是制定有效控制O3污染决策的关键.目前, 常见研究O3生成敏感性的方法有O3等值线(O3 isopleths)、相对增量反应活性(relative incremental reactivity, RIR)和光化学指示剂(photochemical indicator, PI)等[14], 其中PI阈值的选取受O3污染程度的影响且有较大地区差异[15], 所以其应用存在局限性.O3等值线是表征O3与前体物之间关系的曲线[16, 17], 绘制研究区域O3等值线的方法主要分为基于现场观测数据和数值模式模拟两种.基于观测数据的方法是利用反映各因素综合影响的现场观测数据构建O3与NOx和VOCs的回归模型[18]、卷积神经网络(CNN)模型[19]或随机森林(RF)模型[20]等得到O3等值线.基于数值模式的方法是利用盒子模式(box models)[21, 22]或三维大气化学传输模式(chemical transport models, CTMs)[23~25], 模拟NOx和VOCs在不同排放或浓度水平下O3的生成, 从而绘制等值线.相较于盒子模式来说, 多次运行CTMs需要非常大的计算资源, 高阶解耦直接法(high-order decoupled direct method, HDDM)[26]和响应曲面法(response surface methodolog, RSM)[27~30]等被用于减少CTMs的模拟次数, 以提高构建等值线的效率.目前在利用O3等值线判断O3生成敏感性的多数研究中, 通常考察NOx和VOCs两类前体物整体变化的影响[31~33], 忽略了不同VOCs物种之间的活性差异和其他前体物对O3生成的影响.RIR被定义为O3生成潜势的变化与前体物浓度变化的比值[34], 根据各前体物RIR值的正负判断O3生成对其的敏感性[35, 36].RIR法可以判断具体前体物对O3生成的影响, 但其属于单因素分析, 未考虑不同VOCs物种及和NOx之间交互作用对O3生成的影响.

响应曲面法是通过设计合理的试验方案, 利用统计学方法对试验结果进行拟合, 构建多因素和其交互作用与响应值间函数关系的方法[37, 38], 目前在减少CTMs模拟次数从而高效绘制O3等值线相关研究中有较为广泛的应用, 但尚未发现将其用于探究多因素和其交互作用对O3生成的影响.针对O3敏感性研究存在的问题, 本研究通过利用响应曲面法分析NOx、VOCs和CO多因素和其交互作用对O3生成的影响, 实现O3对主要前体物生成敏感性的快速判断; 进一步考虑VOCs不同物种间活性的差异, 探究NOx、烷烃、烯烃、芳香烃、醛酮和其交互作用对O3生成的影响, 识别O3生成的关键VOCs组分并通过·OH收支分析解释其影响显著的原因, 以期为判断O3生成敏感性提供新的思路和方法.

1 材料与方法 1.1 模式简介和参数设置

KPP(kinetic preprocessor)是由Damian等[39]开发出的模拟大气化学反应动力学过程的软件, 是目前空气质量模式的核心模块之一.采用KPP-3.0.0版本, 化学机制选用包含33个物种和81个化学反应的CBM-Ⅳ[40], 输入的初始数据包括O3、NO、NO2、CO和VOCs等污染物以及温度和湿度等, 模拟时长为12 h, 分辨率为60 s.

1.2 模式输入数据来源

选用乌海超级观测站(观测站具体介绍见参考文献[41])某日10:00的NO、NO2、CO和VOCs(包括29种烷烃、9种烯烃、34种卤代烃、17种芳香烃、9种醛类、4种酮类、3种醚类、3种酯类以及异丙醇、二硫化碳和乙炔)监测数据为例, 作为基准情景下的污染物组成, 各污染物和VOCs模式物种的体积分数见表 1所示.

表 1 基准情景下的污染物组成×10-9 Table 1 Initial pollutant composition used in baseline scenario×10-9

1.3 响应曲面法(RSM)

响应曲面法(RSM)是通过合理的试验设计, 采用多元回归方程拟合因素和响应值之间函数关系的统计方法.中心复合设计(central composite design, CCD)是常用的响应曲面试验设计方法之一, 探究k个因素对响应值影响的CCD方案由中心点、2k个顶点和2k个定义参数α的轴向点组成, 其中中心点和顶点用于估计模型的一次项和交互项, 轴向点用于估计平方项.利用Design-expert 13软件(试用版), 选择CCD方法分别设计两组多因素试验情景, 即: ①NOx、VOCs和CO; ②NOx、烷烃、烯烃、芳香烃和醛酮作为影响因素, 以O3日最大体积分数(O3max)为响应值进行建模.

2 结果与讨论 2.1 NOx、VOCs和CO这3因素对O3生成的影响

利用CCD方法设计得到18个试验情景(包含3个重复试验), 探究NOx、VOCs和CO对O3生成的影响, 试验方案中各因素中心点、顶点和轴向点对应的取值如表 2所示.

表 2 中心复合设计3因素布局水平×10-9 Table 2 Level of three-factor layout for CCD×10-9

通过RSM建立A(NOx)、B(VOCs)和C(CO)与O3max的回归方程, 如式(1)所示:

(1)

模型的方差分析结果显示, F值为32.30, P值< 0.000 1, 表示模型极显著, 具有统计学意义; 调整决定系数Radj2=0.946 3, 说明模型能够解释94.63%响应值的变化; 模型的拟合信噪比为15.05, 大于4, 综上该模型可用于分析前体物NOx、VOCs和CO对O3max的影响.根据模型各项的P值可得, NOx二次项、VOCs一次项、VOCs二次项、NOx和VOCs的交互作用和CO一次项均对O3max有显著影响, 已有研究也表明NOx、VOCs和CO均为O3生成的重要前体物[42~45].由于CO与NOx和VOCs的交互作用均不显著, 仅CO一次项显著且回归方程中CO项系数为正, 即可得CO对O3max有正贡献, 对O3-NOx-VOCs敏感性影响较小.而NOx和VOCs存在交互作用且其二次项均显著, 说明O3max与NOx和VOCs存在非线性关系, 由回归方程得到O3等值线[图 1(a)]反映O3-NOx-VOCs敏感性, 脊线φ(NOx)=0.24×φ(VOCs)+13.75下方区域为NOx控制区, φ(NOx)增加会导致O3max升高, 但φ(VOCs)对O3max影响较小; 脊线上方区域为VOCs控制区, 随着φ(VOCs)的增加, O3max显著升高, φ(NOx) < 50×10-9时O3max对NOx不敏感, 但φ(NOx)>50×10-9时NOx存在滴定作用.图 1(b)是保持NOx和VOCs取值范围相同, 对121个等浓度梯度情景的模拟结果进行插值得到的O3等值线, 其脊线为φ(NOx)=0.28×φ(VOCs)+13.75, 两种方法得到的NOx和VOCs控制区域面积相差小于5%, 即验证了RSM所得结果的可靠性.

(a)RSM拟合得到; (b)121个KPP模拟结果插值得到 图 1 对NOx和VOCs变化响应的O3等值线 Fig. 1 Isopleths of the O3 response to NOx and VOCs' changes

2.2 NOx、烷烃、烯烃、芳香烃和醛酮这5因素对O3生成的影响

从2.1节3因素分析结果可得, CO对O3max影响相对较小且其与其他因素间不存在显著交互作用, 而NOx和VOCs与O3max均呈现显著的非线性关系, 同时不同VOCs反应活性差异较大, 因此除了浓度, VOCs的物种组成对O3生成敏感性也有重要影响.所以进一步将VOCs分为烷烃、烯烃、芳香烃和醛酮, 利用CCD方法设计得到46个试验情景(包含3个重复试验), 探究其和NOx对O3生成的影响, 试验方案中各因素的中心点、顶点和轴向点对应的取值如表 3所示.

表 3 中心复合设计5因素布局水平×10-9 Table 3 Level of five-factor layout for CCD×10-9

通过RSM建立A(NOx)、B(烷烃)、C(烯烃)、D(芳香烃)和E(醛酮)与O3max的回归方程如式(2)所示:

(2)

模型的方差分析结果显示, F值为14.36, P < 0.000 1表示模型极显著, 具有统计学意义; 调整决定系数Radj2=0.858 6, 说明模型能够解释85.86%响应值的变化; 模型的拟合信噪比为14.41, 大于4, 综上该模型可用于分析NOx、烷烃、烯烃、芳香烃和醛酮对O3max的影响.通过F值大小得到各显著项对O3max影响的排序为:NOx二次项>烯烃>NOx>NOx和烯烃的交互作用>烯烃二次项>烷烃.在4类VOCs中, 只有烯烃和烷烃对O3max有显著影响, 其中烷烃项系数为正, 对O3生成有正贡献; 烯烃一次项和二次项对O3max影响都大于烷烃项, 且烯烃与NOx存在交互作用, 综上烯烃可识别为O3生成的关键VOCs组分.图 2为NOx-烯烃-O3等值线, 反映O3max与NOx和烯烃间的非线性关系, 脊线φ(NOx)=0.91×φ(烯烃)+15下方区域为NOx控制区, NOx对O3max有正贡献, 烯烃对O3max影响较小; 脊线上方区域为烯烃控制区, O3max随烯烃的增加而明显升高, 但当φ(烯烃)>35×10-9时O3max呈现略微降低的趋势, 说明高浓度烯烃对O3存在消耗作用, φ(NOx) < 50×10-9时NOx对O3max影响较小, 但φ(NOx)>50×10-9时NOx对O3max存在负贡献.

图 2 对NOx和烯烃变化响应的O3等值线 Fig. 2 Isopleth of the O3 response to NOx and changes in olefin

为了进一步探究烷烃和烯烃影响O3生成存在较大差异的原因, 对CCD方案中37~40号试验情景下(表 4) ·OH的生成和消耗速率进行分析.

表 4 37~40号试验情景下各因素值和响应值×10-9 Table 4 Value of factors and responses in 37-40 scenario×10-9

当其他因素均保持中心点值时, 烷烃由轴向点低值[图 3(a)]增加到轴向点高值[图 3(b)], 对应·OH平均总生成速率由(1.1±0.5)×107 molecular·(cm3·s)-1升高至(1.2±0.8)×107 molecular·(cm3·s)-1, 平均总消耗速率由(8.2±4.7)×107 molecular·(cm3·s)-1升高至(9.5±6.6)×107 molecular·(cm3·s)-1, O3max升高了40%; 烯烃由轴向点低值[图 3(c)]增加到轴向点高值[图 3(d)], 对应·OH平均总生成速率由(5.2±2.1)×106 molecular·(cm3·s)-1升高至(1.5±1.0)×107 molecular·(cm3·s)-1, 平均总消耗速率由(3.8±2.0)×107 molecular·(cm3·s)-1升高至(1.1±0.81)×108 molecular·(cm3·s)-1, O3max升高了313%, 可见烷烃和烯烃增加比例相同时, 烯烃能更有效地促进·OH循环, 从而使O3max明显升高.

图 3 不同烷烃和烯烃情景下的·OH生成和消耗速率 Fig. 3 Formation and loss rates of ·OH in scenarios with different alkanes and olefins

表 5为37~40号试验情景下·OH主要生成和消耗途径贡献率, 对比分析各情景下·OH主要生成途径的占比, φ(烯烃)的增加能有效促进HO2将NO氧化为NO2; 由于O3的光解是O(1D)的来源, 所以在O3max较高的情景38和40中O(1D)+H2O的贡献较高.对比各情景下·OH主要消耗途径的贡献率可得, 相对于烷烃来说, φ(烯烃)的增加更显著地加快了VOCs+·OH的反应速率; 同时使NOx+·OH的反应速率降低幅度更大, 烯烃在与NOx争夺·OH时表现出更强的竞争力, 所以其和NOx之间的交互作用对O3生成有重要影响.

表 5 37~40号试验情景下, ·OH主要生成和消耗途径贡献率/% Table 5 Contributions of the main reaction pathways for radical formation and loss in 37-40 scenarios/%

3 结论

(1) 基于结合CBM-Ⅳ化学机制的盒子模式, 设计合理的试验情景, 利用响应曲面法建立反映O3和前体物非线性关系的方程, 在模拟次数较少的情况下实现多因素和其交互作用对O3生成影响的定量研究, 显著提高了O3生成敏感性的判断效率.

(2) CO对O3max有正贡献, 其与NOx和VOCs的交互作用对O3max影响不显著.NOx和VOCs与O3max存在显著非线性关系, 当φ(VOCs)与[φ(NOx)-13.75]比值大于4.17时, 为NOx控制区; 当比值小于4.17时, 为VOCs控制区.

(3) 烷烃和烯烃对O3max生成影响显著, 其中烷烃对O3max有正贡献, 烯烃和NOx与O3max存在显著非线性关系, 当φ(烯烃)与[φ(NOx)-15]比值大于1.10时, 为NOx控制区; 当φ(烯烃)与[φ(NOx)-15]比值小于1.10且φ(烯烃) < 35×10-9时, 为烯烃控制区, 但φ(烯烃)增加至35×10-9以上时, 对O3表现出消耗作用.

参考文献
[1] U.S. EPA. Air quality criteria for ozone and related photochemical oxidants (Final Report, 2006)[R]. EPA/600/R-05/004aF-cF, Washington: U.S. Environmental Protection Agency, 2006.
[2] Turner M C, Jerrett M, Pope Ⅲ C A, et al. Long-term ozone exposure and mortality in a large prospective study[J]. American Journal of Respiratory and Critical Care Medicine, 2016, 193(10): 1134-1142. DOI:10.1164/rccm.201508-1633OC
[3] Monks P S, Archibald A T, Colette A, et al. Tropospheric ozone and its precursors from the urban to the global scale from air quality to short-lived climate forcer[J]. Atmospheric Chemistry and Physics, 2015, 15(15): 8889-8973. DOI:10.5194/acp-15-8889-2015
[4] Haagen-Smit A J. Chemistry and physiology of Los Angeles smog[J]. Industrial & Engineering Chemistry, 1952, 44(6): 1342-1346.
[5] Sillman S. The relation between ozone, NOx and hydrocarbons in urban and polluted rural environments[J]. Atmospheric Environment, 1999, 33(12): 1821-1845. DOI:10.1016/S1352-2310(98)00345-8
[6] Chi X Y, Liu C, Xie Z Q, et al. Observations of ozone vertical profiles and corresponding precursors in the low troposphere in Beijing, China[J]. Atmospheric Research, 2018, 213: 224-235. DOI:10.1016/j.atmosres.2018.06.012
[7] Li K, Jacob D J, Liao H, et al. Anthropogenic drivers of 2013-2017 trends in summer surface ozone in China[J]. Proceedings of the National Academy of Sciences of the United States of America, 2019, 116(2): 422-427. DOI:10.1073/pnas.1812168116
[8] Li K, Jacob D J, Liao H, et al. A two-pollutant strategy for improving ozone and particulate air quality in China[J]. Nature Geoscience, 2019, 12(11): 906-910. DOI:10.1038/s41561-019-0464-x
[9] Lu H X, Lyu X, Cheng H R, et al. Overview on the spatial-temporal characteristics of the ozone formation regime in China[J]. Environmental Science: Processes & Impacts, 2019, 21(6): 916-929.
[10] Wang Y H, Gao W K, Wang S, et al. Contrasting trends of PM2.5 and surface-ozone concentrations in China from 2013 to 2017[J]. National Science Review, 2020, 7(8): 1331-1339. DOI:10.1093/nsr/nwaa032
[11] Li K, Jacob D J, Shen L, et al. Increases in surface ozone pollution in China from 2013 to 2019: anthropogenic and meteorological influences[J]. Atmospheric Chemistry and Physics, 2020, 20(19): 11423-11433. DOI:10.5194/acp-20-11423-2020
[12] Liu Y M, Wang T. Worsening urban ozone pollution in China from 2013 to 2017-Part 1: the complex and varying roles of meteorology[J]. Atmospheric Chemistry and Physics, 2020, 20(11): 6305-6321. DOI:10.5194/acp-20-6305-2020
[13] Cazorla M, Herrera E, Palomeque E, et al. What the COVID-19 lockdown revealed about photochemistry and ozone production in Quito, Ecuador[J]. Atmospheric Pollution Research, 2021, 12(1): 124-133. DOI:10.1016/j.apr.2020.08.028
[14] Liu C Q, Shi K. A review on methodology in O3-NOx-VOC sensitivity study[J]. Environmental Pollution, 2021, 291. DOI:10.1016/j.envpol.2021.118249
[15] Liu J W, Li X, Tan Z F, et al. Assessing the ratios of formaldehyde and glyoxal to NO2 as Indicators of O3-NOx-VOC sensitivity[J]. Environmental Science & Technology, 2021, 55(16): 10935-10945.
[16] Chameides W L, Fehsenfeld F, Rodgers M O, et al. Ozone precursor relationships in the ambient atmosphere[J]. Journal of Geophysical Research, 1992, 97(D5): 6037-6055. DOI:10.1029/91JD03014
[17] Kinosian J R. Ozone-precursor relationships from EKMA diagrams[J]. Environmental Science & Technology, 1982, 16(12): 880-883.
[18] Qian Y, Henneman L R F, Mulholland J A, et al. Empirical development of ozone isopleths: applications to Los Angeles[J]. Environmental Science & Technology Letters, 2019, 6(5): 294-299.
[19] Santos F M, Gómez-Losada A, Pires J C M. Empirical ozone isopleths at urban and suburban sites through evolutionary procedure-based models[J]. Journal of Hazardous Materials, 2021, 419. DOI:10.1016/j.jhazmat.2021.126386
[20] Zhan J L, Liu Y C, Ma W, et al. Ozone formation sensitivity study using machine learning coupled with the reactivity of volatile organic compound species[J]. Atmospheric Measurement Techniques, 2022, 15(5): 1511-1520. DOI:10.5194/amt-15-1511-2022
[21] Xie Y T, Cheng C L, Wang Z H, et al. Exploration of O3-precursor relationship and observation-oriented O3 control strategies in a non-provincial capital city, southwestern China[J]. Science of the Total Environment, 2021, 800. DOI:10.1016/j.scitotenv.2021.149422
[22] Hui L R, Ma T, Gao Z J, et al. Characteristics and sources of volatile organic compounds during high ozone episodes: a case study at a site in the eastern Guanzhong Plain, China[J]. Chemosphere, 2021, 265. DOI:10.1016/j.chemosphere.2020.129072
[23] Luo H H, Zhao K H, Yuan Z B, et al. Emission source-based ozone isopleth and isosurface diagrams and their significance in ozone pollution control strategies[J]. Journal of Environmental Sciences, 2021, 105: 138-149. DOI:10.1016/j.jes.2020.12.033
[24] Sierra A, Vanoye A Y, Mendoza A. Ozone sensitivity to its precursor emissions in northeastern Mexico for a summer air pollution episode[J]. Journal of the Air & Waste Management Association, 2013, 63(10): 1221-1233.
[25] Cui M, An X Q, Xing L, et al. Simulated sensitivity of ozone generation to precursors in Beijing during a high O3 episode[J]. Advances in Atmospheric Sciences, 2021, 38(7): 1223-1237. DOI:10.1007/s00376-021-0270-4
[26] Shen H Z, Sun Z, Chen Y L, et al. Novel method for ozone isopleth construction and diagnosis for the ozone control strategy of Chinese cities[J]. Environmental Science & Technology, 2021, 55(23): 15625-15636.
[27] Xing J, Wang S X, Jang C, et al. Nonlinear response of ozone to precursor emission changes in China: a modeling study using response surface methodology[J]. Atmospheric Chemistry and Physics, 2011, 11(10): 5027-5044. DOI:10.5194/acp-11-5027-2011
[28] Xing J, Ding D, Wang S X, et al. Quantification of the enhanced effectiveness of NOx control from simultaneous reductions of VOC and NH3 for reducing air pollution in the Beijing-Tianjin-Hebei region, China[J]. Atmospheric Chemistry and Physics, 2018, 18(11): 7799-7814. DOI:10.5194/acp-18-7799-2018
[29] Xing J, Zheng S X, Ding D, et al. Deep learning for prediction of the air quality response to emission changes[J]. Environmental Science & Technology, 2020, 54(14): 8589-8600.
[30] Li J Y, Dai Y Z, Zhu Y, et al. Improvements of response surface modeling with self-adaptive machine learning method for PM2.5 and O3 predictions[J]. Journal of Environmental Management, 2022, 303. DOI:10.1016/j.jenvman.2021.114210
[31] 钱骏, 徐晨曦, 陈军辉, 等. 2020年成都市典型臭氧污染过程特征及敏感性[J]. 环境科学, 2021, 42(12): 5736-5746.
Qian J, Xu C X, Chen J H, et al. Chemical characteristics and contaminant sensitivity during the typical ozone pollution processes of Chengdu in 2020[J]. Environmental Science, 2021, 42(12): 5736-5746. DOI:10.3969/j.issn.1000-6923.2021.12.030
[32] 李凯, 刘敏, 梅如波. 泰安市大气臭氧污染特征及敏感性分析[J]. 环境科学, 2020, 41(8): 3539-3546.
Li K, Liu M, Mei R B. Pollution characteristics and sensitivity analysis of atmospheric ozone in Taian City[J]. Environmental Science, 2020, 41(8): 3539-3546. DOI:10.3969/j.issn.1000-6923.2020.08.034
[33] Liu Y F, Qiu P P, Li C L, et al. Evolution and variations of atmospheric VOCs and O3 photochemistry during a summer O3 event in a county-level city, Southern China[J]. Atmospheric Environment, 2022, 272. DOI:10.1016/j.atmosenv.2022.118942
[34] Cardelino C A, Chameides W L. An observation-based model for analyzing ozone precursor relationships in the urban atmosphere[J]. Journal of the Air & Waste Management Association, 1995, 45(3): 161-180.
[35] Xue L K, Wang T, Gao J, et al. Ground-level ozone in four Chinese cities: precursors, regional transport and heterogeneous processes[J]. Atmospheric Chemistry and Physics, 2014, 14(23): 13175-13188. DOI:10.5194/acp-14-13175-2014
[36] An J L, Zou J N, Wang J X, et al. Differences in ozone photochemical characteristics between the megacity Nanjing and its suburban surroundings, Yangtze River Delta, China[J]. Environmental Science and Pollution Research, 2015, 22(24): 19607-19617. DOI:10.1007/s11356-015-5177-0
[37] Box G E P, Wilson K B. On the experimental attainment of optimum conditions[J]. Journal of the Royal Statistical Society: Series B (Methodological), 1951, 13(1): 1-38. DOI:10.1111/j.2517-6161.1951.tb00067.x
[38] 刘祥, 王婧瑶, 吴娟娟, 等. 微藻固定化条件优化及其污水氨氮去除潜力分析[J]. 环境科学, 2019, 40(7): 3126-3134.
Liu X, Wang J Y, Wu J J, et al. Optimization of the parameters for microalgae immobilization and analysis of its recovery potential for ammonia nitrogen in wastewater[J]. Environmental Science, 2019, 40(7): 3126-3134.
[39] Damian V, Sandu A, Damian M, et al. The kinetic preprocessor KPP-a software environment for solving chemical kinetics[J]. Computers & Chemical Engineering, 2002, 26(11): 1567-1579.
[40] Gery M W, Whitten G Z, Killus J P, et al. A photochemical kinetics mechanism for urban and regional scale computer modeling[J]. Journal of Geophysical Research, 1989, 94(D10): 12925-12956. DOI:10.1029/JD094iD10p12925
[41] 朱玉凡, 陈强, 刘晓, 等. 基于气团老化程度对挥发性有机物分类改善PMF源解析效果[J]. 环境科学, 2022, 43(2): 707-713.
Zhu Y F, Chen Q, Liu X, et al. Improved performance of PMF source apportionment for volatile organic compounds based on classification of VOCs' aging degree in air mass[J]. Environmental Science, 2022, 43(2): 707-713.
[42] Wang T, Xue L K, Brimblecombe P, et al. Ozone pollution in China: a review of concentrations, meteorological influences, chemical precursors, and effects[J]. Science of the Total Environment, 2017, 575: 1582-1596. DOI:10.1016/j.scitotenv.2016.10.081
[43] Atkinson R. Atmospheric chemistry of VOCs and NOx[J]. Atmospheric Environment, 2000, 34(12-14): 2063-2101. DOI:10.1016/S1352-2310(99)00460-4
[44] Sillman S, Logan J A, Wofsy S C. The sensitivity of ozone to nitrogen oxides and hydrocarbons in regional ozone episodes[J]. Journal of Geophysical Research, 1990, 95(D2): 1837-1851. DOI:10.1029/JD095iD02p01837
[45] Christian K E, Brune W H, Mao J Q, et al. Global sensitivity analysis of GEOS-Chem modeled ozone and hydrogen oxides during the INTEX campaigns[J]. Atmospheric Chemistry and Physics, 2018, 18(4): 2443-2460. DOI:10.5194/acp-18-2443-2018