文章基本信息

标题：Uncertainties in estimates of mortality attributable to ambient PM2.5 in Europe
本地全文：下载
作者：Jonilda Kushta ; Andrea Pozzer ; Jos Lelieveld 等
期刊名称：Environmental Research Letters
印刷版ISSN：1748-9326
电子版ISSN：1748-9326
出版年度：2018
卷号：13
期号：6
页码：064029
DOI：10.1088/1748-9326/aabf29
语种：English
出版社：IOP Publishing Ltd
摘要：The assessment of health impacts associated with airborne particulate matter smaller than 2.5 μm in diameter (PM_2.5) relies on aerosol concentrations derived either from monitoring networks, satellite observations, numerical models, or a combination thereof. When global chemistry-transport models are used for estimating PM_2.5, their relatively coarse resolution has been implied to lead to underestimation of health impacts in densely populated and industrialized areas. In this study the role of spatial resolution and of vertical layering of a regional air quality model, used to compute PM_2.5 impacts on public health and mortality, is investigated. We utilize grid spacings of 100 km and 20 km to calculate annual mean PM_2.5 concentrations over Europe, which are in turn applied to the estimation of premature mortality by cardiovascular and respiratory diseases. Using model results at a 100 km grid resolution yields about 535 000 annual premature deaths over the extended European domain (242 000 within the EU-28), while numbers approximately 2.4% higher are derived by using the 20 km resolution. Using the surface (i.e. lowest) layer of the model for PM_2.5 yields about 0.6% higher mortality rates compared with PM_2.5 averaged over the first 200 m above ground. Further, the calculation of relative risks (RR) from PM_2.5, using 0.1 μg m⁻³ size resolution bins compared to the commonly used 1 μg m⁻³, is associated with ±0.8% uncertainty in estimated deaths. We conclude that model uncertainties contribute a small part of the overall uncertainty expressed by the 95% confidence intervals, which are of the order of ±30%, mostly related to the RR calculations based on epidemiological data.