15.1.1 Indicator Species

Regulations are set either for a specific air pollutant (for example, lead or benzene) or for a group of air pollutants. The latter include, for example, photochemical oxidants, nitrogen oxides (NO_x), sulfur oxides (SO_x), and particulate matter (PM). In the case of a group of pollutants, a representative pollutant must be selected. For photochemical oxidants, ozone (O₃) is used because it is the chemical species of that group that is present in the highest concentrations. For NO_x, nitrogen dioxide (NO₂) is used because it shows well-documented adverse effects on the human respiratory system. For SO_x, sulfur dioxide (SO₂) is used because it also shows well-documented adverse effects on the human respiratory system. Sulfur trioxide (SO₃) has a very short atmospheric lifetime, because it is rapidly hydrolyzed to sulfuric acid (H₂SO₄). H₂SO₄ is present in the atmosphere in the particulate phase because of its very low saturation vapor pressure. Particles lead to adverse respiratory and cardio-vascular effects; therefore, particles are regulated separately. Thus, it is appropriate to select SO₂ as the indicator species for SO_x. In the case of particles, the regulations have targeted their size rather than their chemical composition up to now. The reason is that there is a large body of health studies that associate adverse health effects with inhalable particles (PM₁₀) and fine particles (PM_2.5), whereas there is currently insufficient epidemiological evidence to identify the adverse health effects of individual constituents of PM (one exception could be diesel particles).

15.1.2 Exposure Duration

Health effects may result from acute exposure (ranging from a few minutes to a few hours) and/or chronic exposure (ranging from a few months to a few years). Therefore, the regulation must correspond to the exposure duration that is considered representative of the health effects. For some air pollutants, regulations may be appropriate for both short (acute exposure) and long (chronic exposure) durations. Table 12.2 summarizes the exposure durations relevant to health effects due to air pollutants regulated in North America and in Europe. In some cases, a pollutant may be regulated for an exposure duration in the absence of conclusive evidence of adverse health effects (see Tables 12.2, 15.1, and 15.2). Such cases (e.g., the annual standard for NO₂ in the United States and in Europe and the annual standard for PM₁₀ in Europe) generally result from historical reasons (earlier health data led to the regulation, which is then kept by default). The regulation that corresponds to the exposure duration with the best evidence of adverse health effects is typically the most constraining. On the other hand, it may happen that a regulation is not set for both acute and chronic exposures in spite of epidemiological evidence for both. This is the case, for example, for PM_2.5 in Europe, where only a long-term standard is used (see Tables 12.2 and 15.2). This choice may, however, be appropriate because the long-term exposure is considered to correspond to the most important adverse health effects. In the United States where both short- and long-term exposures to PM_2.5 are regulated, the long-term regulation (annual concentration averaged over three years) is the most constraining and it is rare that a region would be in attainment of the annual standard, but would not be in attainment of the short-term (daily) standard. It is also the case for ozone (O₃), which is only regulated for short-term exposure (8-hour average concentration), although there is some evidence of likely adverse health effects for chronic exposure to ozone (see Table 12.2). It is considered that an 8-hour average regulatory value also protects against chronic exposure, based on air quality data analyses relating ozone concentrations averaged over short and long periods.

15.1.3 Regulatory Value

The regulatory value (the value of the national ambient air quality standard in the U.S., the limit or target value in Europe) is the best-known aspect of the air pollutant regulation. However, it cannot be dissociated from the exposure (or sampling) duration, the statistical form of the regulation (see Section 15.1.4), or the location of the monitoring stations (see Section 15.1.5). The regulatory value is determined from toxicological and/or epidemiological studies. For standards corresponding to durations of 24 h or more, epidemiological studies are typically used, since human toxicological studies are not conducted for exposure durations exceeding a few hours. For shorter exposure durations, toxicological studies may be favored. However, epidemiological studies add some useful information concerning sensitive populations, which may not be available from toxicological studies. In addition, experimental air quality data may be used to extrapolate ambient pollutant concentrations temporally or spatially, in order to relate results available from toxicological or epidemiological studies with some characteristics of the regulation (e.g., sampling duration or location; see examples later in this section). Studies used to set up regulatory values in the United States are summarized in Table 15.3.

Toxicological studies (clinical studies conducted under controlled exposure) have been used for O₃ (concentration averaged over eight hours) and CO (concentrations averaged over one and eight hours). On the other hand, epidemiological studies have been used for the NO₂ and PM_2.5 annual values. Epidemiological studies were used in combination with a model representing the relationships between atmospheric concentrations, exposure, blood concentrations, and adverse health effects for Pb (3-month moving-average concentration) in order to better account for various possible exposure pathways. Epidemiological studies have also been used for the 24-hour average values of PM_2.5 and PM₁₀.

In some cases, toxicological and epidemiological studies are used in combination. In the case of the hourly NO₂ value, toxicological studies provided the low value of the range of concentrations corresponding to adverse health symptoms for individuals with moderate asthma. Epidemiological studies were then used to better account for more sensitive individuals (e.g., severe asthma). The regulatory value was to be set for near-source situations. Since the epidemiological studies were mostly based on existing air quality monitoring networks, which only included urban background locations, the analysis was complemented with an analysis of air quality data to extrapolate the urban background data to near-roadway locations. In the case of the hourly SO₂ regulatory value, toxicological studies have also provided the low value of the range of concentrations leading to adverse health effects for individuals with moderate asthma exposed for short periods (5 to 10 minutes). Since the regulatory value was set for a one-hour average concentration to minimize temporal fluctuations in the measured concentrations, ambient concentration data were used to relate the 5-minute average toxicological data and the 1-hour ambient concentrations. In the case of O₃, epidemiological studies have been used only to complement toxicological studies to confirm that significant adverse health effects do not appear below the low value of the range of concentrations proposed for the standard.

In summary, the U.S. regulations are based on epidemiological studies for concentrations averaged over periods ranging from 24 hours to one year. They are mostly based on toxicological studies for concentrations averaged over periods ranging from one to eight hours. However, those latter regulatory values are in some cases also based on epidemiological studies that provide useful information concerning sensitive individuals, who cannot participate in toxicological studies. In addition, air quality data have been used to extrapolate the results of health studies spatially (in the case of NO₂) or temporally (in the case of SO₂).

15.1.4 Statistical Form of the Regulation

A regulation may allow a number of exceedances of the regulatory value. This aspect may be stated either explicitly in terms of the number of allowed exceedances (for SO₂, NO₂, O₃, and PM₁₀ in the European regulations and for CO, O₃, and PM₁₀ in the U.S regulations) or in the form of a percentile that corresponds to the concentration not to be exceeded (for SO₂, NO₂, and PM_2.5 in the U.S. regulations). Since a number of allowed exceedances may be converted into a percentile and vice-versa, both are provided in Tables 15.1 and 15.2 to facilitate comparisons between the U.S. and European standards. The ambient concentrations of air pollutants depend strongly on meteorological conditions. Thus, extreme meteorological events may lead to large fluctuations in the highest measured pollutant concentrations from year to year. Therefore, allowing a limited number of exceedances of a regulatory value eliminates those highest measured concentrations from consideration and leads to comparisons of the measured concentrations to the regulatory value that are statistically more robust; i.e., they do not fluctuate as much from one year to the next compared to the highest concentrations. In addition, the meteorological variability may be taken into account for long periods by averaging results over several years. For example, some U.S. regulations, such as PM_2.5, apply to three-year periods. Averaging over several years may also be applied for short-term regulations. For example, the 99^th percentile of the hourly SO₂ concentration is averaged over a three-year period in the U.S. regulation.

15.1.5 Locations of Monitoring Stations

The location of monitoring stations that measure the ambient concentrations of the regulated air pollutants is important for the definition of the air quality standard, particularly in the case of primary pollutants. Urban background stations measure concentrations of primary pollutants that are lower than those measured near the source of those pollutants. For example, an urban background station will measure lower concentrations of CO, NO₂, and PM than those measured near a major roadway, with a difference of a factor of 1.5 to 2, for example, in Paris. In Europe, the air pollutant concentrations are measured at urban background stations as well as at near-source stations (mostly near-roadway stations). In the United States, air pollutant concentrations in urban areas were measured only at background stations until 2009. However, near-source monitoring stations are now required for NO₂ (since 2009) and PM_2.5 (since 2013). For a primary air pollutant, a given regulatory value (including averaging time and number of allowed exceedances) will correspond to a more constraining standard if a near-source monitoring location is used instead of an urban-background location.

15.1.6 Measurement Method

Some experimental uncertainty is associated with any measurement method. In some cases, there may also be a bias due to artifacts. As a result, different sampling instruments may give different values for a given concentration at a given time and location. Therefore, it is essential to specify which method and sampling instrument must be used to monitor air pollutant concentrations. This measurement method, which is specified in the regulation, is called the reference method. Other methods may be used, if they are shown to give results equivalent to the reference method. For example, the reference method in the United States for PM is a gravimetric method (a filter is weighed under controlled conditions for temperature and relative humidity before and after sampling). Continuous measurements with an instrument such as a tapered element oscillating microbalance (TEOM) are considered equivalent to the reference method. In early TEOM instruments, there were some significant artifacts due to the volatilization of semi-volatile particulate compounds such as ammonium nitrate and some organic compounds. A correction is now performed with a filter dynamics measurement system (FDMS), which minimizes the effect of those artifacts.

15.1.7 Regulations in the United States and in France

Regulations of air pollutants in use in 2018 in the United States and in France are presented in Tables 15.1 and 15.2, respectively. These regulations are based on toxicological and epidemiological studies conducted to identify and quantify the cause-effect relationships for acute and chronic exposure to those pollutants (see Table 15.3). There are several differences. First, some pollutants are regulated in both countries (NO₂, SO₂, O₃, PM, CO, and Pb), whereas others are regulated only in one country. Benzene is regulated in France, but not in the United States and, until 2008, fine particles (PM_2.5) were regulated in the United States since 1997, but not yet in Europe. Benzene is a carcinogenic pollutant (category 1 according to the IARC) and the European regulation corresponds to an excess cancer risk of about 10⁻⁵ (see Chapter 12). Benzene is not the only carcinogenic air pollutant and the U.S. Environmental Protection Agency favors treating all carcinogenic air pollutants together in terms of their excess cancer risk, rather than targeting a single pollutant. The sampling times are generally identical, although there are a few differences (one year in Europe for lead and three months in the United States). Statistical forms of the regulatory values are rarely identical and it is difficult to compare two regulations with different statistical forms and regulatory values. For example, the regulatory value for ozone is 120 μg m⁻³ in France (it is a target value), but this value may be exceeded 25 days per year. The regulatory value is greater in the United States since it is 137 μg m⁻³, however, the number of allowed exceedances is much lower (three days per year only).

15.2.1 Public Policy in the United States

In the United States, the Clean Air Act (CAA) and its amendments govern the air quality regulations. In the CAA, the major air pollutants are regulated through National Ambient Air Quality Standards (NAAQS, see Table 15.1). These NAAQS are regularly updated as follows. The setup of a standard starts with a review of the scientific literature, targeting primarily the health effects of the pollutant(s) considered. There are six pollutants or pollutant categories that undergo this process of standard setting: lead, carbon monoxide, nitrogen oxides, sulfur oxides, ozone, and particulate matter (PM). These pollutants are called “criteria pollutants” because the product of the scientific literature review was originally called the “criteria document” (it is now called the integrated scientific assessment or ISA). The next step is the preparation of an evaluation of the population exposure and associated health effects. This work is presented in a report titled “Risk and exposure assessment” (REA). The REA documents quantitatively the exposure and health risks of the population for different regulatory scenarios. These two documents constitute the scientific basis for the development of the NAAQS. Both documents are reviewed by a scientific committee, called the Clean Air Scientific Advisory Committee Review Panel (CASAC Review Panel). NAAQS are then proposed by the U.S. EPA and subsequently promulgated by the federal government. Primary NAAQS concern the protection of public health. Secondary NAAQS may be added to protect the environment. At each step of this public process, the stakeholders concerned by the regulation (industry, ecologists, citizens …) may comment.

Monitoring air quality is essential to ensure that an area of the United States is in attainment or not of the NAAQS. The U.S. EPA sets up a protocol for the air quality monitoring networks, including monitoring sites, measurement instruments, sampling chain procedure (quality assurance/quality control), etc. The air quality monitoring networks are typically operated by the states and measured concentrations are transferred by the state agencies to EPA and incorporated into a database. Based on these air quality data, EPA determines which areas are in attainment of the NAAQS and which ones are not (i.e., non-attainment). In case of non-attainment in a state, the state agency must develop an action plan, called a State Implementation Plan (SIP). A SIP lists all the emission control measures that will be implemented by the state agency to attain the NAAQS within the period allowed by EPA (the length of the period depends on how severe the non-attainment conditions are). The efficiency of these measures is typically evaluated quantitatively with an air quality numerical model (also called chemical-transport model, see Chapter 6). The models that are the most widely used in the United States for this task are CMAQ (Appel et al., 2017) and CAMx (Vijayaraghavan et al., 2012). EPA may ask for revisions if a SIP seems inappropriate (for example, poor performance of the air quality model against available data).

In most states, the state agency is responsible for stationary sources, whereas mobile sources are under the federal government authority. California is an exception because it regulates its own mobile source emissions. Then, the regulation of stationary sources in California may be delegated to districts. If a state does not attain the objectives listed in its SIP, EPA may implement some sanctions, such as the cancellation of funding for road networks, and/or take control of the SIP management by temporarily cancelling the delegation of the process to that state.

In some cases, it is not possible for a state to attain the NAAQS because long-range air pollution imported from other regions may be a major cause for the exceedance of some NAAQS. This has been the case, for example, in some small states of the northeastern U.S., which are downwind of states that are large emitters of primary air pollutants (in particular NO_x), and, therefore, could not meet the ozone NAAQS. The states with large NO_x emissions experienced little ozone formation, because this secondary pollutant was being formed farther downwind. EPA decided to set up a multi-state approach over the northeastern U.S. to tackle this issue on a regional basis and control NO_x emissions in the states that were the large emitters (Godowitch et al., 2008). This process was called the “SIP Call.” This example illustrates the fact that the existing law (here the SIP, which is limited to a state) was not sufficient to attain the regulatory objectives (the NAAQS), but that the federal agency, EPA, did not hesitate to modify the regulatory process to attain the desired objectives. In other words, reaching the objective prevailed over the simple application of the regulatory process, once the latter was shown to be inadequate.

Benzene, which is carcinogenic, is not regulated in the United States, because all carcinogenic pollutants are regulated as a whole as “air toxics.” They are subject to a regulatory process that differs from the NAAQS. Initially, their emissions were regulated via National Emission Standards for Hazardous Air Pollutants (NESHAP). This approach implied the concept of zero risk, which in theory could be attained if the emissions complied with the NESHAP values. Although this approach is feasible for non-carcinogenic risks (which involve a health threshold), it is inappropriate for carcinogenic pollutants, which typically do not have such a threshold (see Chapter 12). In the 1990s, a different approach was introduced that is based on the calculation of health risks to evaluate the carcinogenic risks due to air pollutants. If, after implementation of emission control technology required by the regulations (which are defined by source categories), individual excess cancer risks exceed a value in the range of 10⁻⁶ to 10⁻⁵, additional emission controls may be required for specific sources to decrease the risk (Ohshita and Seigneur, 1993).

The temporal trends of the maximum concentrations of three pollutants regulated in the U.S. are shown in Figure 15.1 for the Los Angeles Basin in California. These three pollutants, O₃, NO₂, and PM_2.5, exceed the NAAQS. The inter-annual variability is mainly due to the meteorological variability. In addition, the NO₂ monitoring network was modified starting in 2014 in response to a revision of the NO₂ NAAQS to include near-roadway locations in addition to the urban background stations (see Section 15.1.5). This change led to an increase in the maximum NO₂ concentration. There are significant decreases of the concentrations of these three air pollutants over the long term. On average, the decrease is 4 to 5 ppb per year for O₃ over the past 40 years, 9 to 10 ppb per year for NO₂, and about 1 μg m⁻³ per year for PM_2.5 over the past 15 years. However, these annual decreases may become less important with time, because the sources that are the easiest to control or to eliminate have typically been addressed upfront and, consequently, it becomes harder to control the emissions of primary pollutants and those of precursors of secondary pollutants. In addition, the relative contribution of the long-range transport of air pollution increases when the local emission contribution decreases. In the Los Angeles Basin, this imported pollution originates mostly from Asia. In summary, these figures illustrate the efficiency of regulatory measures that have been implemented to improve air quality in the Los Angeles Basin, while highlighting the efforts that are still pending to attain the NAAQS.

Figure 15.1. Timelines of the concentrations of ozone (O₃), nitrogen dioxide (NO₂), and fine particles (PM_2.5) in the Los Angeles basin, California. Top figure: maximum 8-hour averaged O₃ concentration from 1976 to 2015; middle figure: maximum hourly NO₂ concentration from 2000 to 2015; bottom figure: maximum PM_2.5 annual concentration from 2000 to 2015. Source: SCAQMD (2017).

15.2.2 Public Policy in France

In France, an air pollution regulation is based on directives issued by the European Union. The European directive is then converted into French law. A European directive for a given air pollutant takes into account the health effects of that air pollutant. However, there is a major difference between the U.S. approach and the European one in the fact that the European directive also takes into account the time needed to bring the air quality toward the health-based objective. As a result, the limit values (equivalent to the regulatory value of the U.S. NAAQS) may vary over time, starting with a high value for the first few years and decreasing to lower values as time goes on, tending toward a value corresponding to public health protection. It is currently the case for fine particles, PM_2.5, with an annual limit value of 25 μg m⁻³ in 2015 and a target value of 20 μg m⁻³ for 2020 (for comparison, the U.S. NAAQS annual value for PM_2.5 is 12 μg m⁻³).

The European approach uses an ensemble of values that present various levels of constraints:

• 
  

  Air quality objective: A concentration level of air pollutants to be attained over the long term in order to achieve an efficient protection of human health and the environment, except when it is unachievable through reasonable measures.

  
  
• 
  

  Target value: A concentration level of air pollutants defined in order to prevent, warn or reduce adverse effects on human health and the environment, to be attained to the extent possible by a given deadline.

  
  
• 
  

  Limit value: A concentration level of air pollutants based on scientific knowledge, not to be exceeded, in order to prevent, warn or reduce adverse effects on human health and the environment. (Note, however, that the regulation may allow a limited number of exceedances.)

  
  
• 
  

  Information and recommendation threshold: A concentration level of air pollutants above which a short-term exposure presents a health risk for sensitive individuals, requiring timely and relevant communication to the public.

  
  
• 
  

  Alert threshold: A concentration level of air pollutants above which short-term exposure presents a health risk for the whole population or environmental degradation and justifies emergency measures.

In France, air quality monitoring is conducted by non-governmental organizations (Associations agréées de surveillance de la qualité de l’air, i.e., AASQA), which transfer the air quality monitoring data into a national air quality database (Base de données sur la qualité de l’air, BDQA). Air quality monitoring must follow standard procedures defined at the European level.

In the U.S., the same entity (the state or, in California, the district) monitors air quality and is responsible for taking actions to meet the NAAQS. In France, the AASQA have no authority on air pollutant emission control. Instead, French government agencies are responsible for regulating stationary sources. Those are the Regional agencies for environment, planning, and housing (Directions régionales de l’environnement, de l’aménagement et du logement, DREAL), except in Paris, where it is the Regional agency for environment and energy (Direction régionale et interdépartementale de l’environnement et de l’énergie, DRIEE). Each country belonging to the European Union must then demonstrate that it is in attainment of the limit values of the European regulations. If not, an action plan must be developed to reach attainment within a reasonable time. In addition, exceedances of the alert and information thresholds must be communicated to the public and reported to the European Agency. An exceedance of limit values may lead to important fines by the European Union Court of Law (several tens of million euros per year). For example, as of May 2018, six European countries have been referred to the Court of Justice of the European Union because of exceedances of air pollutant limit values: France, Germany, and the United Kingdom for non-attainment of NO₂ limit values and Hungary, Italy, and Romania for non-attainment of PM₁₀ limit values.

There are three main programs in France that pertain to air quality protection. They are part of the 1996 Air Law (Loi sur l’air et l’utilisation rationnelle de l’énergie, LAURE).

1. 
   

   – Urban transportation planning (Plans de déplacements urbains, PDU), which are developed by 59 urbanized areas with more than 100,000 inhabitants. They have actually limited effect on air quality, because they mostly impact the spatio-temporal distribution of on-road traffic emissions, without significant impact on the area-wide total mobile source emissions.

   
   
2. 
   

   – Regional planning for climate, air, and energy (Schémas régionaux climat air énergie, SRCAE), which have replaced the earlier regional air quality plans (Plans régionaux de la qualité de l’air, PRQA). SRCAE were put into place following the so-called Grenelle de l’Environnement meeting in 2007 and were implemented through the Grenelle I (2009) and Grenelle II (2010) laws. They are developed by the administrative regions. They are open for public consultation and are revised every five years. They present recommendations to reduce emissions in order to reach objectives for air quality, while minimizing greenhouse gas emissions and energy consumption. However, they do not lead to specific regulations to reduce air pollutant emissions. Therefore, they are tools for providing recommendations in terms of public policy, rather than decision tools for emission controls.

   
   
3. 
   

   – The Atmosphere protection plans (Plans de protection de l’atmosphère, PPA), which are developed by 24 urbanized areas with more than 250,000 inhabitants. They are developed under the authority of the government representative for the region (“préfet”). The Paris PPA is developed by the DRIEE. PPA are developed with the objective of reaching attainment of the European limit values. Therefore, they may include regulations to reduce stationary source emissions. Thus, a PPA is a regulatory tool (actually, it is the only regulatory tool at the regional level). Its development typically involves the government agency (DRIEE or DREAL), technical support from the regional air quality organization (AASQA, e.g., Airparif in the Paris region), as well as stakeholders and experts. In the case of the Paris PPA, the emission control measures have been selected according to a process where all stakeholders participated in the PPA development. A PPA does not always demonstrate conclusively that the measures listed will lead to attainment of the limit values. Nevertheless, a PPA documents, at least partially, which impacts those measures will have on air quality improvement. The list of long-term measures identified to control air pollutant emissions may be complemented with emergency measures in the event of exceedance of alert thresholds.

It is important that these three plans (PDU, SRCAE, and PPA) be consistent among each other. Note that the French PPA have the same limitation as the U.S. SIP, i.e., they are limited to a geographical region and, therefore, may not be sufficient in the case of imported long-range air pollution. A national or European approach may be needed to treat not only regulated air pollutants with long atmospheric lifetimes (O₃ and PM_2.5), but also their precursors (VOC, NO_x, SO₂, and NH₃). Emission control scenario simulations are conducted to evaluate the impact of emission control measures on future air quality. Numerical air quality models used to that end include, for example, Chimere (Menut et al., 2014) and Polyphemus/Polair3D (Sartelet et al., 2012).

The reduction of stationary source emissions and the regulation of fuels are under the authority of the government. Emission standards for on-road vehicles are set at the European level (see Chapter 2).

Temporal trends of maximum concentrations at urban background stations for three regulated air pollutants are presented in Figure 15.2 for the Paris region. The maximum 8-hour average O₃ concentrations and the maximum annual NO₂ concentrations are shown for the 1994–2015 period. The maximum annual PM_2.5 concentrations are shown only for the 2007–2015 period, because of a change in monitoring technique in 2007. The inter-annual variability is mostly due to meteorological variability. For example, there is an increase in the maximum O₃ concentration in 2003 due to a heat wave in August 2003. Nevertheless, there are overall decreasing trends over the long term. O₃ concentrations, despite large year-to-year fluctuations, decreased by 23 μg m⁻³ (11.5 ppb), if one compares the averages of the maximum concentrations over the last two decades, i.e., 1996–2005 and 2006–2015). The number of exceedances of the target value remains within the authorized amount (<25 per year); however, the air quality objective is exceeded (8-hour average concentration of 120 μg m⁻³; Airparif, 2016). The maximum urban background NO₂ concentration decreases annually by about 1.4 μg m⁻³ on average over the period analyzed. The NO₂ concentrations are currently below the annual limit value at urban background stations. However, they exceed this limit value at stations located near roadways (Airparif, 2016). The maximum annual PM_2.5 concentration has decreased by about 1 μg m⁻³ per year on average over the past eight years. It is likely that the use of diesel particle filters has contributed to the decrease of fine particle concentrations in Paris. The annual PM_2.5 concentration is below the target value of 20 μg m⁻³ at urban background stations. At near-roadway stations, the limit value of 25 μg m⁻³ was attained in 2015 (Airparif, 2016). In the case of O₃ and the secondary fraction of PM_2.5, the contribution of long-range atmospheric transport is important and, therefore, limits the impact of local emission control strategies on the concentration levels of these secondary pollutants in the Paris region.

Figure 15.2. Timelines of concentrations of ozone (O₃), nitrogen dioxide (NO₂), and fine particles (PM_2.5) in the Paris region. Top figure: maximum 8-hour averaged O₃ concentration from 1994 to 2015; middle figure: maximum annual NO₂ concentration at urban background stations from 1994 to 2015; bottom figure: maximum annual PM_2.5 concentration at urban background stations from 2007 to 2015.

Source: Airparif (2016).