Clean air in the Anthropocene

In atmospheric chemistry, interactions between air pollution, the biosphere and human health, often through reaction mixtures from both natural and anthropogenic sources, are of growing interest. Massive pollution emissions in the Anthropocene have transformed atmospheric composition to the extent that biogeochemical cycles, air quality and climate have changed globally and partly profoundly. It is estimated that mortality attributable to outdoor air pollution amounts to 4.33 million individuals per year, associated with 123 million years of life lost. Worldwide, air pollution is the major environmental risk factor to human health, and strict air quality standards have the potential to strongly reduce morbidity and mortality. Preserving clean air should be considered a human right, and is fundamental to many sustainable development goals of the United Nations, such as good health, climate action, sustainable cities, clean energy, and protecting life on land and in the water. It would be appropriate to adopt “ clean air ” as a sustainable development goal.


Introduction
In the past century, the inuence of mankind on the environment has grown immensely, and has essentially developed into a geological force, hence the Earth system has entered the Anthropocene epoch. 1 During the second half of the 20 th century this inuence accelerated sharply, with a doubling of the human population from 3 to 6 billion (and probably 9 billion in 2050), and with leaps in agricultural, economic and industrial activity. 2 Globally, more than 50% of the population lives in the urban environment, typically associated with degraded and poor air quality. However, air pollution emissions and their impacts extend well beyond the city boundaries, also transforming biogeochemical cycles and contributing to regional and global climate change. 3 The World Health Organization (WHO) has identied air pollution as a major environmental health risk, and notes that its abatement will also reduce emissions of carbon dioxide (CO 2 ) and short-lived climate pollutants. Nearly 90% of humanity lives in areas where the WHO air quality guideline of 10 mg m À3 annual average PM 2.5 is exceeded 4 (PM 2.5 is ne particulate mass with a diameter less than 2.5 mm). It has been estimated that worldwide about 3.3 million deaths were attributable to PM 2.5 and ozone (O 3 ) pollution in the year 2010, 5,6 and it will be argued here that this number needs to be revised upward.
Building on the Millennium Development Goals by the United Nations (UN), articulated for the period 2000 until 2015, the General Assembly of the UN has adopted 17 Sustainable Development Goals (SDGs) to be pursued for the subsequent phase up to 2030 (Fig. 1). The SDG Knowledge Platform on the topic "atmosphere" focusses on the Agenda for Sustainable Development paragraphs related to climate change, e.g. by calling for the reduction of CO 2 and other greenhouse gas emissions. It also acknowledges the need to enhance the implementation of strategies to monitor the Earth's atmosphere, and mentions international agreements such as the Montreal Protocol on Substances that Deplete the Ozone Layer. While the prominence of atmospheric composition is undisputed with respect to ozone loss in the stratosphere and climate change, it is crucial that the UN also recognise the central importance for air quality and public health. In terms of the value of lives lost and ill health, air pollution is costing OECD societies plus China and India an estimated 3.5 trillion US dollars per year, with a rising trend. 7 It is proposed here to adopt an additional SDG on "clean air", at the same level as good health, clean water, sustainable industrialization and cities, responsible production, climate action, marine and terrestrial life (SDGs 3, 6, 9, 11, 12, 13, 14, 15, respectively), not only because ambient air pollution is the leading environmental risk factor of the global burden of disease, 8 but also because many other SDGs cannot be achieved without considering air quality. The next Section 2 focusses on results from the Faraday Discussion meeting Atmospheric chemistry in the Anthropocene, held in York, UK, from 22-24 May 2017, which presented scientic results that underpin this statement. Section 3 presents an update of the Fig. 1 The 17 SDGs of the UN, aimed at ending poverty, protecting the planet, and ensuring prosperity for all as part of the sustainable development agenda. Each goal has specific targets to be achieved over the period 2015-2030 (http://www.un.org/ sustainabledevelopment/sustainable-development-goals/). An 18 th SDG aimed at clean air should be added. contribution of outdoor air pollution sources to the global burden of disease by Lelieveld et al., 6 taking into account recent health data and statistics by the WHO, and revised integrated exposure-response (IER) functions to compute mortality attributable to ambient air pollution. 8 Section 4 discusses the link between air quality and climate change, and the concluding Section 5 reiterates the "clean air" proposal in view of recent developments regarding air quality in the Anthropocene.

Atmospheric chemistry in the Anthropocene
An important subject of the Faraday Discussion meeting was the link between atmospheric chemistry and the biosphere, starting with a dedicated session, and revisited in subsequent sessions that focussed on chemical processes and air quality aspects. The discussion reected a central quality of the Anthropocene, namely that human inuence, through atmospheric composition change, profoundly interferes with the biological state of the environment, associated with adverse effects on ecosystems and agricultural crops, changing biogeochemical cycles and public health risks. It also manifested a growing interest in the complex interactions between the physical-chemical and biological workings of the Earth system. Many studies presented in York addressed processes across and at interfaces, e.g. on multiphase and heterogeneous chemistry, including air pollution impacts on bioaerosols and human tissue, being novel developments in the eld. In the following paragraphs the presentations are briey summarized, and details can be found in the referenced articles. Brüggemann et al. (DOI: 10.1039/C7FD00022G) reported their discovery that organic surfactants of biogenic origin (biolms) lead to volatile organic compound (VOC) production through sunlight-induced radical chemistry, a potential major marine source of VOCs and reaction products that can inuence secondary organic aerosols (SOA) and oxidants in the background atmosphere. A modelling study of natural and anthropogenic halogen chemistry indicated signicant effects on oxidant formation. It found that iodine and bromine chemistry may actually help improve air quality by reducing ozone, while chlorine chemistry might have a moderate opposite effect (Sherwen et al., DOI: 10.1039/ C7FD00026J). Anthropogenic land use change was shown to reduce biogenic VOC emissions and consequent SOA formation, which is expected to inuence air quality and climate change (Scott et al., DOI: 10.1039/C7FD00028F). While aerosols decrease direct and increase diffuse sunlight, thus cooling and drying the climate system, the net primary productivity (net carbon uptake) of forest expanses e.g. in the Amazon and the boreal region may change in converse directions (Unger et al., DOI: 10.1039/C7FD00033B). Because of the geographically counterbalancing outcomes, the net effect on global atmospheric CO 2 may thus be small.
The interaction of some noxious air pollutants, such as polycyclic aromatic hydrocarbons (PAHs), with SOA appears to remarkably strongly affect the properties of aerosols, e.g. by increasing the SOA mass, particle number concentration and viscosity, which may inuence the long-range transport potential of PAHs sulfate), but anthropogenic SOA can actually affect mixed ice-liquid water clouds and hence precipitation by enhancing glaciation, depending on the particle phase state. Long-term aircra measurements taken from commercial airliners, presented by Petzold et al. (DOI: 10.1039/C7FD00006E), demonstrate that the relative humidity distribution over ice in cirrus clouds, RH ice , strongly correlates with ice particle number concentrations. Measurements showed that within thin cirrus clouds in the subtropics and tropics RH ice z 100% and at mid-latitudes RH ice z 105% (only occasionally higher). A dedicated session of the Faraday meeting addressed new research tools (not elaborated here †), which was also reected in discussions of atmospheric chemistry processes, showing that recent instrument development is opening new territory for research. For example, Tong et al. (DOI: 10.1039/C7FD00023E) studied interactions of mineral dust with SOA particles, leading to formation of reactive oxygen species (ROS) in aqueous solution, e.g. from the decomposition of organic hydroperoxides. This could be a main mechanism of oxidant formation in clouds and fogs, and possibly also in the human respiratory tract, causing oxidative stress with health outcomes. Liu et al. (DOI: 10.1039/C7FD00005G) uncovered that air pollutants such as O 3 and NO 2 can enhance the allergenic potential of proteins in bioaerosols, e.g. through their nitration and oligomerization. The extent of these processes depends on the RH due to moisture-induced phase transition of proteins, also highlighting the importance of cloud processing of aerosols. Increasing tropospheric O 3 in the Anthropocene can promote proallergic protein modications and may have contributed to the observed increase of allergies.
New particle formation (NPF) in the atmosphere can be induced by a number of processes that involve anthropogenic and natural (including biogenic) precursor gases, and current theory explains NPF in clean and moderately polluted environments, however, not in heavily polluted megacities. NPF is nevertheless observed in megacities, e.g. in China, which indicates that molecular cluster scavenging may be over-, or their growth under-estimated (Kulmala et al., DOI: 10.1039/C6FD00257A). The recently constructed atmospheric reaction chamber HELIOS was shown to simulate realistic atmospheric conditions, and used to study the chemistry of isoprene, reaction intermediates, oxidant recycling and SOA yields (Ren et al., DOI: 10.1039/C7FD00014F). An upcoming research topic is the atmospheric chemistry of Criegee intermediates (CIs), i.e. biradicals formed from the reaction of ozone with olenic VOCs of natural and anthropogenic origin, which can react with SO 2 and NO 2 . Caravan et al. (DOI: 10.1039/ C7FD00007C) performed laboratory studies on the CI reaction with NO 2 , and found that NO 3 is not a likely product, while a species with the exact mass of the Criegee-NO 2 adduct was observed, possibly the major reaction product.
Another emerging topic is that of ne particulate (PM 1 ) chemistry and thermodynamics, e.g. based on aerosol mass spectrometer measurements in eld campaigns. PM 1 measurements from a tower and a zeppelin indicated excess † Additional papers were presented at the meeting that provided insight into new instrumentation,  ammonium with respect to inorganic ions (Schlag et al., DOI: 10.1039/ C7FD00027H). It was shown that heterogeneous reactions with organic compounds, notably (di)carboxylic acids, i.e. oxidation products of atmospheric VOCs, likely explain the NH 3 uptake. This suggests that NH 3 emission controls will have benets for air quality beyond the reduction of inorganic ammonium salts, i.e. from controlling SO 2 and NO 2 emissions. Murphy et al. (DOI: 10.1039/ C7FD00086C) used a thermodynamic gas-particle partitioning model to study the impact of the ammonium/(bi)sulfate ratio on the pH of aerosols. It was found that an increase of this ratio may lead to a steep rise of particle pH, which inuences the chemistry and environmental impacts of aerosols.

Public health impact of air pollution
The session on "The air we breathe: past, present and future", provided compelling support for dedicating an SDG to clean air. To contribute to the Faraday Discussion on this topic, our group used a global atmospheric chemistrygeneral circulation model, 9 applying emission estimates from the Emission Database for Global Atmospheric Research (EDGAR) as boundary conditions. 10 The model results have been combined with IER functions, following Lelieveld et al., 6 updated according to Cohen et al., 8 to compute the relative risk of mortality by acute lower respiratory infections (ALRI) in childhood, chronic obstructive pulmonary disease (COPD), cerebrovascular disease (CEV), leading to mortality by ischaemic and haemorrhagic stroke (i.e. CEVI and CEVH, respectively), ischaemic heart disease (IHD) and lung cancer (LC) (Fig. 2). These diseases are related to PM 2.5 , while COPD is additionally related to O 3 . All data sets used as input for the calculations, such as country level baseline mortality rates for these diseases and population data, have been adopted from the WHO Global Health Observatory (http://www.who.int/gho/database/en/), being representative of the year 2015.
The minimum risk exposure level distribution for annual mean PM 2.5 , adopted in the IER functions, is 2.4-5.9 mg m À3 , which is lower than the previously assumed range of 5.8-8.8 mg m À3 , 11 and that of O 3 is 33.3-41.9 ppbv. 8 This recent downward adjustment of the lower bound for PM 2.5 , i.e. from 5.8 to 2.4 mg m À3 , is a main reason for the higher mortality estimates presented here compared to earlier work, since a larger fraction of the population is exposed to hazardous levels of air pollution. Further, the mortality rates for COPD are signicantly higher than reported previously. Thus, the state-of-the-science is that only the very low PM 2.5 concentrations <2.4 mg m À3 do not contribute to morbidity and mortality. The results presented here agree closely with those of Cohen et al. 8 for the year 2015, on average within a few percent. Our global estimates for PM 2.5 compare well, and for O 3 they are slightly higher, largely related to the different methodologies in calculating the global concentration distributions of ozone.
Our updated results indicate a total mortality attributable to outdoor air pollution of 4.33 million persons per year, which is about 30% higher than estimated previously by Lelieveld et al. and by the global burden of disease (GBD) for the year 2010. 5,6 The statistical uncertainty range is estimated at AE50% (95% condence interval). Note that the IERs for very high levels of PM 2.5 are associated with the largest uncertainty, as epidemiological data for such conditions are scarce. About 739 500 per year are due to ALRI, 743 000 per year to COPD, 377 000 per year to CEVI, 491 000 per year to CEVH (hence 868 000 per year in total due to CEV), 1.4 million per year to IHD, 282 000 to LC and 298 000 per year due to O 3 related COPD (hence 1.04 million in total due to COPD) (Fig. 2). The estimated mortality attributable to air pollution in the EU-28 is 274 000 per year, and for Europe as a whole 332 000 per year (not including Russia and Ukraine). The global total number of years of life lost (YLL) is 123 million per year, which implies that each individual who dies prematurely due to air pollution loses about 28 years of his/her life. The life expectancy reduction is largest for ALRI as it affects infants, especially in countries where nutrition and medical care are decient, being on average 54 years, and 20-25 years for the other disease categories.
Considering that the diseases attributable to air pollution are analogous to those from tobacco smoking, the problem of degraded air quality may be characterized as "passive outdoor smoking". Among the leading GBD risk factors, ambient particulate matter pollution ranks h, aer high blood pressure, smoking, diabetes and high cholesterol, 8 and it is reiterated that it is the largest environmental risk factor in the GBD. To put the problem in context, the WHO estimates that the number of people who die annually from malaria and HIV/AIDS together is about 1.5-1.6 million, which is roughly one third compared to air pollution. It should be mentioned that the loss of life from malaria and HIV/AIDS has been brought down signicantly in the past decade as a result of active research, and energetic prevention and control programmes, which are also needed for global air pollution. Table 1 ranks the countries where the annual mortality exceeds 30 000 individuals per year. China and India lead, partly related to their large population size. On the other hand, the population weighted, per capita mortality attributable to air pollution is highest in the Ukraine and Russia, with about 1.7 Â 10 À3 and 1.6 Â 10 À3 persons per year, respectively, while in China it is $1.0 Â 10 À3 persons per year (i.e. one per 1000 person-years) (Fig. 3). The attribution to source sectors in Table 1 is from Lelieveld et al., 6 calculated by omitting the respective emissions in the model. On a global scale, residential energy use for heating and cooking is the largest source category (31%), notably because it is the major one in Asia. In Eastern Asia and Europe agricultural emissions are important, notably NH 3 from animal husbandry and fertilizer use, because this gas is an effective precursor of PM 2.5 . 12 In Germany, for example, 24 000 deaths per year may be   functions, focussed on the impacts of air pollution on public health in Africa, where population, urbanisation and industrialization are increasing rapidly. Based on improved Africa-specic emission estimates, residential energy use was found to contribute most to anthropogenic air pollution related mortality. In future, impacts related to fossil energy use are expected to become more prominent. Further, Archibald et al. (DOI: 10.1039/C7FD00004A) applied a chemistryclimate model to study the reduction of mortality due to air pollution emission controls in the last decades, focussing on the USA and Western Europe. An alternative time trajectory of emissions was considered as a "world avoided" scenario. A business-as-usual scenario since 1970 was compared to one that accounts for the legislation that has been implemented between 1970 and 2010. It was found that more than 520 000 mortalities have been mitigated owing to extensive reduction in sulphate aerosol and improvements in O 3 and NO 2 pollution. This sends the important message that air pollution controls in the USA and EU have actually achieved considerable public health benets, even though additional efforts will be essential to come down from the high numbers presented in Table 1 and Fig. 3.

Short-lived climate pollutants
While air pollutants have a typical lifetime in the atmosphere of a few days to weeks, greenhouse gases (GHGs) such as CO 2 , methane (CH 4 ), nitrous oxide (N 2 O) and halocarbons remain airborne for years up to more than a century. Since this is longer than the mixing time scale of the whole atmosphere (about a year), the radiative forcing of climate by the increase of these gases is inevitably a global phenomenon. Tropospheric O 3 , which also acts as a GHG, has an intermediate lifetime of about a month and can disperse on a hemispheric scale. The humaninduced increase of CO 2 contributes about 55-60% to the radiative forcing of climate by GHGs, and all non-CO 2 gases together about 40-45%. 3 The recent total radiative forcing of climate by increasing GHGs is about 3 W m À2 . In addition, anthropogenic ne aerosol particles (i.e. PM 2.5 ) cause regional radiative forcings that can be either positive (warming) due to the absorption of sunlight by black carbon (BC) or negative by the backscattering of sunlight by water-soluble acids, salts and organics. The net aerosol radiative forcing of climate is determined by the specic pollutant mixture in the particles and is generally negative (cooling). Furthermore, the aerosols can affect cloudiness (indirect effects) because they act as cloud condensation nuclei, also leading to a negative radiative forcing. The overall radiative forcing by aerosols (direct and indirect), averaged globally, is estimated at À0.9 W m À2 and may thus mask approximately 30% of the climate warming by GHGs. 3 Because the control of CO 2 emissions has shown to be challenging, short-lived climate pollutants (SLCPs), including CH 4 , O 3 , some hydrouorocarbons (HFCs) and BC, have come into the fore. SLCPs offer the advantage of responding relatively rapidly to control measures, which is particularly helpful if climate change needs to be averted in the near term. Considering that SLCPs also affect public health and ecosystems, policies on air quality and climate intersect, and it is useful to consider win-win strategies. While the latter may be expected for mitigation measures that address BC and O 3 , a general reduction of PM 2.5 to improve public health will likely moderate the overall negative radiative forcing of aerosols. On the other hand, air pollution controls are oen achieved through technology improvements, which typically also improve energy efficiency and reduce GHG emissions. It should be stressed that advocating SLCP control, e.

Concluding remarks
The Faraday Discussion meeting presented compelling evidence for a signicant Anthropocene footprint discernible in the composition and chemistry of the atmosphere. The eld has developed powerful tools to measure and to model processes that control atmospheric composition, and determine the environmental impacts, e.g. on air quality and climate. Considerable progress has been achieved regarding the climate forcing of reactive GHGs and aerosols, offering knowledge that can be used to formulate policies on SLCPs such as HFCs, CH 4 and O 3 . Much uncertainty remains on the chemistry and radiative properties of aerosols. Direct and indirect effects of aerosols on radiative forcing are important to estimate the climate sensitivity to GHG emissions. If the net cooling effect of aerosols would be smaller than assumed, for example due to light absorption by BC, the estimated rate of warming due to increasing CO 2 based on observations would be proportionally smaller. An emerging topic in atmospheric chemistry is the impact of aerosols and ROS on public health. Results from our group, based on updated air pollution, health and population data sets, suggest that the mortality attributable to outdoor air pollution amounts to 4.33 million individuals per year, associated with 123 million years of life lost. This refers to deaths that could have been avoided when annual average PM 2.5 would be below 2.4 mg m À3 and O 3 below 33.3 ppbv. More than 70% of the mortality occurs in Asia, and more than 50% in China and India (together 2.2 million per year). While the latter countries lead the ranking of total mortality, the per capita mortality is highest in Eastern Europe with 1.6 persons per 1000 inhabitants per year, being more than twice that in India and four times compared to North America (Fig. 3). The steepest recent increase in air pollution related mortality occurs in South Asia, especially India. It should be noted that globally about 18% (up to 780 000 deaths per year) cannot be easily managed, and are thus difficult to avoid, being attributable to natural PM 2.5 , mostly aeolian desert dust.
Recently, Zhang et al. investigated the relationship between exports of goods and services and the exposure to poor air quality. 13 In effect, by exporting goods and services, many countries in Asia and Eastern Europe "import" air pollution related mortality. These countries exchange monetary revenue against health expenditure through emissions embodied in exports. It is conceivable that some polluting industries have moved, e.g. from the EU and the USA, to regions with more permissive environmental regulations. To prevent such practices, international norms and agreements are needed, and an indispensable step will be that the UN adopt an SDG on "clean air". In view of the recently updated IER functions, with the minimum risk exposure level lowered from 5.9 to 2.4 mg m À3 (annual average PM 2.5 ), air quality standards will need to be revised, currently being 12 mg m À3 in the USA and 25 mg m À3 in the EU. The air quality guideline by the WHO of 10 mg m À3 PM 2.5 should also be reconsidered, even though its global implementation could already reduce mortality attributable to air pollution by more than 50%. 14