Elsevier

Atmospheric Environment

Volume 251, 15 April 2021, 118275
Atmospheric Environment

Sensitivity analysis of O3 formation to its precursors-Multifractal approach

https://doi.org/10.1016/j.atmosenv.2021.118275Get rights and content

Highlights

  • The O3-NMHCs-NOx sensitivity is identified by coupling detrended fluctuation analysis.

  • A novel index is proposed to quantify the response of O3 formation to its precursors.

  • Time dependences of O3-NOx-NMHC sensitivity in different regions are revealed.

  • Researching results are test by the comparative analysis of classic empirical kinetic modeling.

  • Multifractal method can provide potential supplement to the methodology of O3-NOx-NMHC sensitivity analysis.

Abstract

As a secondary pollutant with strong oxidation, surface O3 is harmful to human health and ecosystem. Accurate identification in the response of ozone formation to its precursors (O3-NOx-NMHCsensitivity) is significant for policy-making of O3 pollution control. Great progress has been made in O3-NOx-NMHC sensitivity analysis based on model-based methods in previous studies. However, large uncertainties from meteorological factors, emission inventory and photochemical mechanism may lead to the deviation between the simulation results and the observations. Properties of nonlinear coupling detrended fluctuation analysis (CDFA) is the ability to reveal the coupling correlations between O3 and its precursors at different time scales, helping us to quantity the contribution degree of NOx and NMHC to O3 formation in our previous work. Here, we further set up a new index to determine directly the O3-NOx-NMHC sensitivity based on CDFA method, and its availability is test by the comparative analysis of classic Empirical Kinetic Modeling Approach (EKMA). Based on this new evaluation index, seasonal change, time dependence and diurnal variation of O3-NOx-NMHC sensitivity for two traffic sites and two general sites located in the north and south of Taiwan respectively are analyzed. The results confirmed that the new index is able to identify the regional difference in temporal variation of O3-NOx-NMHC sensitivity. Thus, the features of quantification, time-saving and easy-operation make it possible for multifractal methods to become an important potential supplement to the methodology of O3-NOx-NMHC sensitivity analysis in the future.

Introduction

As one main components of atmospheric photochemical smog, surface ozone (O3) is a secondary pollutant produced by photochemical reactions of nitrogen oxides (NOx) and non-methane hydrocarbons (NMHC) under solar radiation (Haagen-Smit, 1952; Derwent et al., 1996). Due to the strong oxidation and corrosion, high O3 level poses a serious threat to human health and ecosystem (Lippmann, 1993; Gryaris et al., 2004; Agathokleous et al., 2020). In order to control O3 pollution, many cities have developed some strategies including vehicle emission controls to reduce the emission of O3 precursors. But in some regions, O3 level keeps increasing instead of decreasing due to the lower O3 titration by NO (Sicard, 2021). It indicates that O3 formation is not a simple linear process in relation to its precursors, but a highly nonlinear process (Chou et al., 2006; Wang et al., 2017). In previous researches, surface O3 formation was divided into “NOx-sensitive”, “NMHC-sensitive” and “mixed-sensitive” regime (Sillman, 1999; Lu et al., 2019). In NOx-sensitive regime, O3 formation rate mainly depends on NOx concentration, but has little response to the change of NMHC concentration. On the contrary, in NMHC-sensitive regime, the formation rate of O3 increases with the increase of NMHC concentration, and does not increase with increasing NOx. In this case, reducing NOx concentration blindly may result in O3 accumulation due to the weakening of NO titration, which can lead to more serious O3 pollution. Therefore, accurate identification of O3-NOx-NMHC sensitivity in specific regions is of great significance for policy-making in O3 pollution control.

At present, the study of O3-NOx-NMHC sensitivity is generally based on model-based methods (Zhou et al., 2013; Jin et al., 2017; Oikonomakis et al., 2018; Wang et al., 2019; Pfister et al., 2019). In order to describe the occurrence and evolution of O3 pollution, various mathematical models have been established based on atmospheric chemistry, physics and meteorology, with considering the interaction among various air pollutants as much as possible. From the microscopic point of view, these models include meteorological factors, emission inventory and photochemical mechanism in detail, possessing clear physical and chemical meaning. However, the incomplete understanding of photochemical mechanism as well as the great uncertainty in emission inventory and meteorological factors, may lead to large errors in the simulation results. For example, using eight different air quality models, simulations of O3-NOx-NMHC sensitivity in UK and other regions of Europe in July 2006 were conducted by Richard et al. (2014). It had been found that results obtained from each model were inconsistent, some even contradictory. Moreover, for regions lacking emission inventories, the application of global emission inventory in the simulation of O3 formation can promote the uncertainty, due to the spatial difference (Abdallah et al., 2016). With regard to photochemical mechanisms, Kitayama et al. embedded four different photochemical mechanisms (Carbon Bond (CB05TUCL), Regional Atmospheric Chemistry Mechanism (RACM2), and State Air Pollution Research Center (SAPRC07TC and SAPRC99)) into the same air quality model (Community Multiscale Air Quality (CMAQ)) to simulate O3 formation in Japan, and significant differences had been found in the four simulation results (Kitayama et al., 2019). For meteorology, wind speed and the height of convective mixed layer were the main sources of uncertainty in modeling ozone concentrations (Sillman, 1999; Marsik et al., 1995). Some evidences showed that wind speed can lead to much uncertainty in the simulation of O3 formation by affecting pollution dispersion (Sillman et al., 1998). Thus, the great uncertainty in the estimation of numerous parameters in numerical models may lead to large errors in simulation results of O3-NOx-NMHC sensitivity. Especially under combined air pollution, the heterogeneous reactions on the surface of particles will promote the complexity of photochemical reactions. Therefore, in order to minimize the uncertainty in simulation process of O3 formation, it is necessary to find new ways to study the nonlinear response of O3 formation to its precursors.

In recent decades, based on the theory of complexity science, some new nonlinear methods (such as Detrended Fluctuation Analysis (DFA)) have been developed to study the multi-scale characteristics of O3 series, with the objective of revealing the intrinsic dynamics of high O3 concentration level (Windsor and Toumi, 2001; Varotsos et al., 2005; Varotsos and Kirk-Davidoff, 2006; Thomas et al., 2017; Jan et al., 2018). In general, nonlinear methods are performed based on time series of field observations without any assumptions, so the uncertainties in the results have been reduced. Numerous studies have shown that the auto-correlation function of O3 concentration fluctuation decays in the form of power law rather than the exponential decay (Chelani, 2009, 2012, 2013). It indicates that concentration distribution of O3 is not random, but has the characteristics of long-term persistence and memory (Chelani, 2016). Moreover, some evidence had been observed that long-term persistence in concentration distribution of air pollutants were associated with the prediction precision (Dong et al., 2017; Yuval and David, 2010).As a secondary pollutant, atmospheric O3 formation and evolution is a reflection of photochemical reactions of precursors (mainly NOx and NMHC). Thus, only analyzing the scale characteristics of single O3 series is not sufficient, and the relations between O3 and its precursors must be considered. Coupling correlations among multiple non-stationary time series can be studied using coupling detrended fluctuation analysis (CDFA) developed by Hedayatifar et al. (2011). Based on CDFA, we had demonstrated the long-term persistence in coupling correlations among O3, NOx and NMHC in our previous work (Liu et al., 2018). The existence of long-term persistence indicates that the concentration fluctuations of precursors in the past have a long persistent effect on O3 concentration in the future. Meanwhile, the heterogeneity can be observed in the long-term persistence at different time scales, which can be described by multifractal structure. Consequently, in our previous work, a new multifractal index had been proposed to measure the coupling sensitivity degree of NOx or NMHC in atmospheric O3 formation system. However, the index in our previous work can only quantify the contribution of each precursor to O3 formation, but it is unable to estimate the relative contribution of the two precursors for determining O3-NOx-NMHC sensitivity. So it failed to identify whether O3 formation in a region is sensitive to NOx or NMHC, which has little application value in the decision-making of ozone pollution control.

In this paper, based on CDFA method and our previous work(Liu et al., 2018), a new index R is proposed to determine quantitatively the O3-NOx-NMHC sensitivity at different time scales, by calculating the proportion of coupling sensitivity degree of each precursor in O3 formation system. The advantage of this index is that it can directly identify “NOx-sensitive”, “NMHC-sensitive” and “mixed-sensitive” regime of O3 formation for a specific region. Further, in order to test the availability of the new index, the results obtained by the CDFA method are compared with those obtained by the classical Empirical Kinetic Modeling Approach (EKMA). Moreover, for testing the regional universality of this new index, this work has made a comparative analysis of different function areas of air quality in southern and northern Taiwan respectively.

Section snippets

Materials

Taiwan is located at the southeast of East Asia, and its north and south regions belong to two different climate types. The northern and central regions of Taiwan are controlled by subtropical monsoon climate, while the southern region is controlled by tropical monsoon climate. Therefore, in order to reveal the spatial differences of surface O3 pollution, materials from traffic monitoring sites and general monitoring sites in the northern and southern regions have been selected for comparative

Seasonal change of O3-NOx-NMHC sensitivity based on CDFA method

In order to study the seasonal change of O3-NOx-NMHC sensitivity for the four sites, the research period from March 1, 2018 to February 28, 2019 is classified into four seasons: spring (March to May), summer (June to August), autumn (September to November) and winter (December to February in 2019). Then the CSD and R values in each season for the four sites are calculated, as shown in Fig. 3.

As shown in Fig. 3, CSD values of NMHC and NOx at four sites are almost higher in summer than that in

Conclusion

Although air pollution seems to be a stochastic-type process, the concentration fluctuation of air pollutants always exhibit complex statistical correlations at different time scales. For nonlinear and non-stationary atmospheric system, nonlinear methods are able to reveal the scaling behavior of intrinsic fluctuations in air quality variables. In our previous work, based on nonlinear CDFA method, we have quantified the contribution degree of NOx and NMHC to O3 formation by analyzing the

CRediT authorship contribution statement

Chunqiong Liu: Writing – original draft, preparation, Methodology. Li Zhang: Data curation, Software. Ye Wen: Writing – review & editing, Visualization. Kai Shi: Conceptualization, Methodology, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

The work is supported by Hunan Provincial Natural Science Foundation of China (No. 2020JJ4504).

References (45)

  • C.H. Lin et al.

    Experimental investigation of ozone accumulation over night during a wintertime ozone episode in south Taiwan

    Atmos. Environ.

    (2004)
  • C.Q. Liu et al.

    Coupling detrended fluctuation analysis of the relationship between O3 and its precursors –a case study in Taiwan

    Atmos. Environ.

    (2018)
  • Y.P. Peng et al.

    Concentrations of H2O2 and HNO3 and O3-VOC-NOx sensitivity in ambient air in southern Taiwan

    Atmos. Environ.

    (2006)
  • Min Shao et al.

    Ground-level ozone in the pearl river delta and the roles of VOC and NOx in its production

    J. Environ. Manag.

    (2009)
  • P. Sicard et al.

    Ozone weekend effect in cities: deep insights for urban air pollution control

    Environ. Res.

    (2020)
  • P. Sicard

    Ground-level ozone over time: an observation-based global overview

    Environmental Science & Health

    (2021)
  • S. Sillman

    The relation between ozone, NOx and hydrocarbons in urban and polluted rural environments

    Atmos. Environ.

    (1999)
  • C. Varotsos et al.

    Scaling properties of air pollution in Athens, Greece and Baltimore

    Md. Atmos. Environ

    (2005)
  • X.S. Wang et al.

    Decoupled direct sensitivity analysis of regional ozone pollution over the Pearl River Delta during the PRIDE-PRD2004 campaign

    Atmos. Environ.

    (2011)
  • T. Wang et al.

    Ozone pollution in China: a review of concentrations, meteorological influences, chemical precursors, and effects

    Sci. Total Environ.

    (2017)
  • H.L. Windsor et al.

    Scaling and persistence of UK pollution

    Atmos. Environ.

    (2001)
  • W. Zhou et al.

    Reconciling NOx emissions reductions and ozone trends in the U.S., 2002-2006

    Atmos. Environ.

    (2013)
  • Cited by (22)

    • Fractal analysis of impact of PM<inf>2.5</inf> on surface O<inf>3</inf> sensitivity regime based on field observations

      2023, Science of the Total Environment
      Citation Excerpt :

      The quantitative index system used to measure O3 sensitivity regime is established based on CDFA technique, as shown in Fig. 2(a). To describe the influence of PM2.5 on O3 sensitivity regime more accurately, fractal analysis of the temporal evolution of R values on hourly scale is necessary, which can be achieved by sliding window method (Liu et al., 2021). Considering the weekend effect of surface O3 pollution, we set the time series of O3, NMHC and NOx in 7-day scale (1–168 data) as the initial window and calculate the first R value.

    View all citing articles on Scopus
    View full text