Civil aviation, air pollution and human health

Sources related to transport of people and goods are amongst the largest emitters of locally-acting air pollutants, but are also often amongst the sources most amenable to control. Thus road traffic tail-pipe emissions in the developed world have declined over the past 20 years, despite a continuing growth in traffic volumes. Technology has delivered greatly reduced emissions of health-related pollutants per vehicle-kilometre, whilst simultaneously improving fuel efficiency and greenhouse gas emissions by a combination of improved engine technology and exhaust after-treatment. Until recently, emissions from shipping were largely unregulated and as a consequence were steadily increasing as a percentage of SO₂ and NO_x emissions (Eyring et al 2010). Recent international agreements have curbed emissions from shipping, especially in coastal waters and ports (Tao et al 2013), and the regulations are set to tighten further.

International aviation is generating a large interest within the scientific community because of its constant growth: in 2012, there were more than 79 million aircraft movements carrying 5.7 billion passengers between 1598 airports located in 159 countries (ACI 2013). Annual growth rates of ∼5% per year are typical and numbers are predicted to double between 2012 and 2030 (ICAO 2013).

While most of the gas phase content of typical airliner engine exhaust consists of N₂, O₂, CO₂, and H₂O, many residual products are also jointly released in the atmosphere, including NO_x, CO, SO₂, a large number of hydrocarbons and aerosol particles containing organic and inorganic components having non-volatile and semi-volatile properties (Masiol and Harrison 2014). As from other combustion sources, the list of such products may also include compounds with known or suspected adverse effects on human health. Furthermore, aviation accounts for around 1–2% of global greenhouse gas emissions per year.

While aviation has not been immune from the trend towards emissions reduction, it has thus far got away fairly lightly. For example, while automotive gasoline and diesel (with a volatility range similar to aviation fuels) have reduced to a sulphur content of <10 ppm, the limit for aviation fuel remains at 3000 ppm with actual concentrations reported to be within the range 300–1100 ppm. While automotive diesels sold in Europe now have to meet particle emission standards requiring the installation of a particle filter, jet engine emissions are still evaluated with a semi-quantitative 'smoke number' with little pressure to improve.

Amongst the transport sectors, aviation is more difficult to study and apportion as it encompasses both the horizontal and vertical dimensions. Nevertheless, while it is tempting to believe that the localization of ground and low altitude aircraft movements limits their impact on air quality, even the high altitude emissions can impact on air quality in the atmospheric boundary layer where people live and breathe (e.g., Barrett et al 2010). Therefore, impacts of aviation upon air quality and the assessment of related health effects are major questions currently addressed in scientific research and debated by policy-makers and stakeholders.

In this context, the paper by Yim et al (2015) is a valuable contribution toward the assessment of emissions and health effects of civil aviation. The study takes account of emissions of NO_x, hydrocarbons, SO₂, black carbon and organic carbon using dispersion and chemistry-transport models to calculate concentrations of fine particles (PM_2.5, primary plus secondary sulphate) and ozone. Its main strength is the multi-scale approach, which is capable to capture the impacts of civil aviation emissions at ground-level arising from dispersion at various scales, ranging from local (∼1 km), near airport (∼10 km), regional (∼1000 km) and even global (∼10 000 km). The study also estimates the premature mortality per year associated with population exposure to PM_2.5 and O₃ attributable to civil aviation emissions using concentration-response functions (CRF) reported in the literature. For PM_2.5 the CRF used in the WHO (2004) Global Burden of Disease study was used, as well as alternative CRF from more recent studies for sensitivity analysis. For O₃, a relative risk coefficient for respiratory mortality reported by Jerrett et al (2009) was applied. Furthermore, Yim et al monetised the estimated premature mortality impacts using country-specific values of statistical life (VSL) and compared these estimates with other societal costs of civil aviation, namely the global costs associated with aviation related accidents, climate and noise impacts. According to this methodology, a global total of ∼5000 people who live within 20 km of airports are estimated to die prematurely each year due to long-term exposure to aviation-attributable PM_2.5 and O₃ with associated costs estimated to be ∼$21 billion per year.

Despite the valuable contribution made by the paper by Yim et al it should be viewed as a starting point from which to focus forthcoming efforts to generate improved future estimates and to develop mitigation measures rather than a complete and comprehensive estimation. In fact, information on aviation impacts upon the environment and public health remain very sparse and many questions are still unresolved. In this context, the study of Yim et al is highly conservative and considers just direct aircraft emissions, omitting a large component of the ground-level aviation related emissions which may have potentially high impact on air quality at the surface. With this in mind, a list of future considerations that need to be addressed in future research should include:

• A comprehensive estimation of PM emissions, including both volatile and non-volatile phases and their partitioning and dilution effects at ambient temperatures. The study by Yim et al includes primary emissions of black carbon and particulate organic carbon calculated from fuel burn, and sulphate, estimated by a simple parameterization of sulphur dioxide oxidation. This is liable to underestimate PM_2.5 concentrations, as nitrate was omitted despite large emissions of precursor NO_x, and conversion of VOC to secondary organic aerosol, admittedly more difficult to model, is also left out. Work on automotive emissions has shown that the large semi-volatile hydrocarbon component of the emitted particles will tend to desorb with advection downwind of source and form a greater mass of secondary aerosol through oxidation (Robinson et al 2010).
• Aircraft non-exhaust and other airport-related emissions. While the work of Yim et al is restricted to engine emissions from aircraft, there is a strong case to extend the analysis to include other emissions and impacts associated with air traffic. Aircraft themselves also cause emissions from tyre, brake and runway wear. Within the airport, auxiliary and ground power units are widely in use to power stationary aircraft, and ground service equipment which often uses engine technologies and fuel of quality far inferior to modern road vehicles inevitably accompanies aircraft movements. Finally, and perhaps most importantly, airports both operate and attract large volumes of road traffic in the form of shuttle buses, taxis, private cars, service vehicles etc, which together generate a large pollutant load (Masiol and Harrison 2014) as well as greenhouse gas emissions and noise.
• An extended parameterization and assessment of health impacts. More recent CRFs are now available in the literature, including coefficients used in the most recent Global Burden of Disease (2010) study (Lim et al 2012). Moreover, similar global burden studies have relied on Disability Adjusted Life Years (Lim et al 2012) or Years of Life Lost (Anenberg et al 2010), which are generally more appropriate metrics as they take into account the number of years lost due to ill-health. Additionally, the substantial burden of morbidity associated with air pollutant exposure should be included in further studies. Finally, although monetary valuation provides a common basis for comparison, there are ethical challenges around the valuation of non-market goods, such as ecosystem services, and the difference between VSL in high and low income countries (Hallegatte et al 2008).

Increasingly, cost-benefit analyses, taking account of the human health-related costs of industry and transport are being used to evaluate mitigation measures. It seems inevitable that civil aviation will continue to expand at a substantial rate and it will be essential to make sound decisions relating to its public health impacts. In this context, the work of Yim et al (2015) makes an important contribution and we look forward to further studies extending the analyses in the ways suggested above and being put to use in justifying proportionate mitigation measures.