Multiple air pollutants and their health impacts for both present-day and future scenarios

The adverse health impacts of air pollution, both short-term and long-term, have been widely studied in recent years; however there are a number of uncertainties to consider when carrying out health impact assessments. Health effects attributable to exposure to air pollutants are typically estimated using measured or modelled pollutant concentrations which vary both temporally and spatially. The goal of this thesis is to perform health impact assessments using modelled pollutant concentrations for present-day and future. The specific aims are: (i) to study the influence of model horizontal resolution on simulated concentrations of ozone (O3) and particulate matter less than 2.5 μm in diameter (PM2.5) for Europe and the implications for health impact assessments associated with long-term exposure (ii) to model air pollutant concentrations during two air pollution episodes in July 2006 together with the corresponding short-term health impact in the UK (iii) to estimate potential future health burdens associated with long-term pollutant exposure under future UK emission changes for 2050 in the UK. First, the impact of model horizontal resolution on simulated concentrations of O3 and PM2.5, and on the associated long-term health impacts over Europe is examined, using the HadGEM3–UKCA (UK Chemistry and Aerosol) chemistry– climate model to simulate pollutant concentrations at a coarse (~140 km) and a finer (~50 km) horizontal resolution. The attributable fraction (AF) of total mortality due to long-term exposure to warm season daily maximum 8-hr running mean (MDA8) O3 and annual-mean PM2.5 concentrations is then estimated for each European country using pollutant concentrations simulated at each resolution. Results highlight seasonal variations in simulated O3 and PM2.5 differences between the two model resolutions in Europe. Simulated O3 concentrations averaged for Europe at the coarse resolution are higher in winter and spring (~10 and ~6 %, respectively) but lower in summer and autumn (~-1 and ~-4 %, respectively) compared to the finer resolution results. These differences may be partly explained by differences in nitrogen dioxide (NO2) concentrations simulated at the two resolutions. Compared to O3, the opposite seasonality in simulated PM2.5 differences between the two resolutions is found. In winter and spring, simulated PM2.5 concentrations are lower at the coarse compared to the finer resolution (~-8 and ~-6 % averaged for Europe, respectively) but higher in summer and autumn (~29 and ~8 %, respectively). Differences in simulated PM2.5 levels are largely related to differences in convective rainfall and boundary layer height between the two resolutions for all seasons. These differences between the two resolutions exhibit clear spatial patterns for both pollutants that vary by season, and exert a strong influence on country to country variations in the estimated AF of mortality for the two resolutions. Results demonstrate that health impact assessments calculated using modelled pollutant concentrations, are sensitive to a change in model resolution with differences in AF of mortality between the countries ranging between ~-5% and ~+3%. Under climate change, the risk of extreme weather events, such as heatwaves, is likely to increase. Thus the UK health burden associated with short-term exposure to MDA8 O3 and daily mean PM2.5 is examined during two five-day air pollution episodes during a well-known heatwave period in July 2006 (1st - 5th July and 18th – 22nd July) using the UK Met Office air quality model (AQUM) at 12 km horizontal resolution. Both episodes are found to be driven by anticyclonic conditions (mean sea-level pressures ~1020hPa over the UK) with light easterly and south easterly winds and high temperatures that aided pollution build up in the UK. The estimated total mortality burden associated with short-term exposure to O3 is similar during the each episode with about 70 daily deaths brought forward summed across the UK. The estimated health burden associated with short-term exposure to daily mean PM2.5 concentrations differs between the first and second episode resulting in about 43 and 36 daily deaths brought forward, respectively. The attributable fraction of all-cause (excluding external) mortality for both pollutants differs between UK regions and ranges between 1.6% to 5.2% depending on the pollution levels in each episode; the overall total estimated health burdens are highest in regions with higher population totals. Results show that during these episodes, short-term exposure to MDA8 O3 and daily mean PM2.5 is between 36- 38% and 39-56% higher, respectively, than if the pollution levels represented typical seasonal-mean concentrations. Finally, emission scenarios for the UK following three Intergovernmental Panel on Climate Change (IPCC) Representative Concentration Pathways (RCPs); RCP2.6, RCP6.0 and RCP8.5 are used to simulate future concentrations of O3, NO2 and PM2.5 for 2050 relative to 2000 using the AQUM air quality model at 12km resolution. The present-day and future AF of mortality associated with long-term exposure to annual mean MDA8 O3, NO2 and PM2.5 and the corresponding mortality burdens are estimated for each region in the UK. For all three RCPs, simulated annual mean MDA8 O3 concentrations in 2050 are estimated to increase compared to 2000, due to decreases in nitrogen oxides (NOx) emissions reducing titration of O3 by NO, and to increases in methane (CH4) levels across all of the UK. In contrast, annual mean NO2 concentrations decrease everywhere. This highlights that the whole of the UK is simulated to be in a NOx-saturated chemical environment. PM2.5 concentrations decrease under all RCPs for the 2050s mostly driven by decreases in NOx and sulphur dioxide (SO2) emissions affecting secondary inorganic aerosols concentrations. For all pollutants the largest changes are estimated under RCP8.5 while the smallest changes are estimated for RCP6.0 in 2050 as compared to present-day. Consequently, these two RCPs represent the high and low end of the AF and mortality burden difference range relative to present-day for all three pollutants. For all UK regions and all three RCPs, the AF of mortality associated with long-term exposure to O3 is estimated to increase in 2050 while the AF associated with long-term exposure to NO2 and PM2.5 is estimated to decrease as a result of higher and lower projected pollutant concentrations, respectively. Differences in the UK-wide mortality burden attributable to long-term exposure to annual mean MDA8 O3 across the RCPs range from +2,529 to +5,396 additional attributable deaths in 2050 compared to 2000. Long-term exposure to annual mean NO2 and PM2.5 differences in health burdens are between - 9,418 and -15,782 and from - 4,524 to -9,481 avoided attributable deaths in 2050 relative to present-day, respectively. These mortality burdens are also sensitive to future population projections. These results demonstrate that long-term health impact assessments estimated using modelled pollutant concentrations, are sensitive to a change in model resolution across Europe, especially in southern and eastern Europe. In addition, air pollution episodes are shown to have the potential to cause substantial short-term impacts on public health in the UK. Finally the sensitivity of future MDA8 O3-, NO2- and PM2.5-attributable health burdens in the UK to future emission scenarios as well as population projections is highlighted with implications for informing future emissions control strategies for the UK.