Abstract
Unabated emissions of precursor gases from coal-fired power stations in South Africa greatly contribute to ambient PM2.5 concentration in the region, but estimates of the number of associated premature mortalities vary vastly. The application of the exposure-response function is one of the reasons for the differences in estimates of premature mortalities; since the relationship between mortality and PM2.5 exposure flattens out at higher PM2.5 concentrations, a smaller number of premature mortalities is calculated if a reduction in PM2.5 concentration occurs at the actual ambient concentration ("avoidable deaths"), while a greater number of mortalities is calculated if other sources are not considered (at low ambient concentrations) ("single-source deaths"). It is more appropriate to attribute deaths to a source in proportion to the contribution made by the source to ambient PM2.5 levels ("attributable deaths"). Total ambient PM2.5 concentrations are extrapolated from 38 ambient monitoring stations in the north-eastern region of South Africa using a regression model based on household socio-economic factors, and PM2.5 concentrations derived from power station emissions are estimated using a reduced-form intake fraction model. It is estimated that there are around 800 attributable premature mortalities per year due to PM2.5 from coal-fired power station emissions in South Africa. The number of premature mortalities is 44% higher when calculating single-source deaths and 53% lower when calculating avoidable deaths. Using the attributable death approach is particularly important in highly polluted areas such as low-income residential areas where households use solid and liquid fuels.