30 40 50 60 70 80 30 40 50 60 70 80 Population-weighted PM 2.5 ( µ g m -3) China India China India 2000 2002 2004 2006 2008 2010 2012 2014 2016 30 40 50 60 70 80 30 40 50 60 70 80 Population-weighted PM 2.5 ( µ g m -3)
FIGURE1.14: Time-series of population-weighted PM2.5 concentra- tions in China and India from 1990-2016. Data taken from State of
Global Air 2018https://www.stateofglobalair.org/data/.
key polluting regions (e.g. Zheng et al.,2017).
Understanding how ambient PM2.5 concentrations will change into the future is complex and depends on a number of different factors such as changes in anthro- pogenic emissions, climate and meteorology, and natural emissions. However, con- sidering anthropogenic emissions alone, regional changes will be dependant on the emission pathway scenario assumed (Figure 1.15). That said, it is important to con- sider how changes or emission abatement in key polluting anthropogenic sectors might affect and improve ambient levels of PM2.5in the near-term (Chapter 5).
1.6 Trends in the global burden of disease attributable to PM
2.5In addition to providing disease burden estimates for the present-day, the GBD CRA also provided trends for the burden of disease attributable to individual known risk factors (1990 to present-day). In the most recent CRA (Gakidou et al.,2017; Feigin,
2016; Cohen et al.,2017), the global burden of disease attributable to long-term am- bient PM2.5exposure was found to increase 20% from 3.5 (3.0 to 4.0) million deaths in 1990 to 4.2 (3.7-4.8 ) million deaths in 2015 (Figure 1.16).
Figure 1.17 shows the change in ambient PM2.5attributable mortality over the same period (1990 to 2015) for ten populous countries together with the contribution of individual factors that influence mortality changes. The increase in the number of attributable deaths globally was largely caused by increases in PM2.5exposure and
FIGURE1.15: Regional annual mean population weighted PM2.5con- centrations (μg m-3) (left axis) in 2005 and 2050 under different emis- sion scenarios including ’Reference’ or Shared Socio-Economic Path- way (SSP) scenarios (blue color bars) and average of 3 RCP scenarios (grey bar). Green, orange and red colored markers indicate the frac- tion of the population exposed to<10,<25 and<35 μg m-3 respec- tively (right axis), and contribution of natural PM2.5 is represented
by the hatched area. Figure taken from Rao et al.,2017
the absolute numbers of deaths from non-communicable diseases in highly popu- lated countries such as India and China, where populations are growing and ageing rapidly. These changes were enough to counteract the reduction in attributable mor- tality experienced across high-income countries (e.g. US) where reductions in PM2.5 have been achieved since 1990, and where populations are not growing rapidly. In contrast, age-standardised death rates from ambient PM2.5 decreased by 12.3% from 65.6 per 105deaths (56.9-74.9) in 1990 to 57.5 per 105deaths (50.2-64.8) in 2015. This reduction was a result of improved air quality across high-income countries and overall global improvements in global healthcare, which resulted in overall declines in background disease rates that are causally associated with PM2.5exposure (Cohen
1.6. Trends in the global burden of disease attributable to PM2.5 33
FIGURE1.16: Total global deaths attributable to ambient PM2.5pollu- tion by year and cause. Figure taken from Cohen et al.,2017. et al.,2017).
FIGURE1.17: Changes in attributable deaths from ambient PM2.5ex- posure due to the changing contribution from population growth, ageing, background disease mortality, and exposure. Figure taken
from Cohen et al.,2017
Given the large risk factor to public health, understanding historical changes in am- bient PM2.5 disease burden is vital for informing future air quality policy design. However, current estimates are restricted to examining changes over the last 25 years only (Figure 1.16), when satellite and ground-based observations are typically available. Few studies have investigated changes in attributable disease burdens associated with changes in ambient PM2.5 over the last 50 years or so, a period of widespread implementation of air quality regulation and emission controls across North America and Western Europe coincided with extensive economic growth and
limited emission controls across developing Asia (Figures 1.10). Such changes would have undoubtedly resulted in regional contrasts in ambient PM2.5 and associated disease burden trends, and thus a focus of this thesis in Chapter 3.
Understanding how the PM2.5 disease burden will change in the future is complex and depends not only on changes in PM2.5 concentrations (previous section), but also on changes in demographics and background disease epidemiology. Given the large demographic contribution to changes in PM2.5disease burden over the last 25 years (Figure 1.17), future changes in demographics will likely play an important role in future attributable disease burdens. Regional studies have shown that future population growth and ageing will increase deaths from PM2.5in China and India, even when PM2.5levels have been substantially reduced relative to the present-day (GBD MAPS Working Group,2016; GBD MAPS Working Group,2018; Conibear et al.,2018b).
Studies that examine changes in PM2.5mortality under future emission pathway sce- narios, report global increases of between 50-335% by 2050 relative to the present- day (Lelieveld et al.,2015; Stohl et al.,2015; IEA,2016), while other studies predict reductions (Silva et al., 2016; IEA, 2016). These differences are attributable to the differences in assumed emissions and demographic pathways. Nevertheless, few global studies have examined the likely contribution of changes in demographics and background disease to future changes in PM2.5mortality, which is focus of this thesis in Chapter 5. Understanding these contributions may be important for craft- ing future air quality policy.
1.6.1 Options for mitigating particulate air pollution
Air pollution in the form of PM will persist as a major public health problem until governments take the necessary action to mitigate its effects. PM air pollution can be mitigated through control or end-of-pipe technologies that reduce emissions at the point of emission (e.g., vehicle catalytic converters and smokestack scrubbers), and more structural shifts that avoid the occurrence of emissions (e.g., fuel switching,
1.6. Trends in the global burden of disease attributable to PM2.5 35 energy efficiency and low or non-combustion technologies). However, with the evi- dence base for PM health effects growing at an ever increasing rate, the need to take action quickly using the range of mitigation options available should nevertheless be modulated by the need to find a long-term path that also does not compromise other policy goals (IEA,2016). The optimum policy path is one that takes decisive ac- tion in coordination with others, which should include both the setting of ambitious long-term air quality goals and a clean air strategy for the important polluting sec- tors, with effective monitoring, enforcement, evaluation and communication (IEA,
2016).
Reducing air pollution can also bring desirable co-benefits. Climate change co- benefits have recently gained attention through high profile political initiatives, such as the Climate and Clean Air Coalition (CCAC) (www.ccacoalition.org). However, decision makers need to be made aware that while action may provide benefits in one area (e.g., climate), they can worsen the situation in another (e.g., air quality). For example, climate policy favouring low emitting CO2 diesel vehicles replacing equivalent petrol vehicles can lead to detrimental effects on air quality (e.g. Jon- son et al.,2017). Another example includes the shift from light fossil fuel used for space heating to ’carbon-neutral’ biomass (wood), which may have greater adverse effects on PM air quality (e.g. Haluza et al., 2012). While this last example may be an growing problem facing wealthy regions (e.g., Europe), providing appropri- ate ’win-win’ solutions for the 3 billion users of residential (household) solid fuels (mainly solid biomass) in low and middle-income countries (Bonjour et al.,2013), which contribute both to household and ambient PM pollution, and climate change, (Smith et al.,2014a; Chafe et al.,2014; Lelieveld et al.,2015) is a complex but pressing issue.
The large scale adoption and implementation of clean burning solid fuel cookstoves has been suggested as a possible ’win-win’ solution (e.g. Anenberg et al., 2013). However, others argue that use such cookstoves are being driven by an interna- tional policy agenda focussing on climate goals (i.e., one promoting carbon-neutral solid biomass) rather than a promoting cleaner burning light fossil fuels (e.g., LPG), which would provide greater reductions in PM air pollution (both household and
ambient) and benefits for health (Goldemberg et al.,2018; Smith and Sagar,2014). It is thus important to consider how future changes in emissions (e.g., residential en- ergy) due to policy measures might impact on both air quality (and human health) and climate, which is a focus of this of this thesis in Chapter 5.