Air pollution epidemiology faces the problem of searching for small increases in risk against a background of great uncertainty and noise in the available environmental and health data. It is therefore essential both to be able to combine and compare results from local-scale studies, and to conduct large, population level studies over wide geographical areas. Both of these approaches are typically hampered by the lack of consistent data on air pollution, and the use of different methods and measures of exposure assessment. The APMoSPHERE (Air Pollution Modelling for Support to Policy on Health, Environment and Risk management in Europe) project is aimed at using GIS and remote sensing methods to develop a set of consistent air pollution maps (for PM10, black smoke, SO2, NO2, CO and O3), at a spatial resolution of ca. 1km, for the whole of the European Union, that can be used for both epidemiological and policy purposes. Different modelling approaches are being developed and assessed including: a) the compilation of high resolution emissions inventories by disaggregating national emission statistics to the local level, using land cover, socio-economic and other data; b) use of geostatistical and stochastic methods to model spatial variations in air pollution; c) development of a landscape stratification system for modelling air pollution; d) use of Bayesian hierarchical modelling techniques to model time-space variations in air pollutionl e) evaluation of the use of satellite data to monitor urban and regional variations in air pollution.
(1) Imperial College London
(2) Centre for International Climate and Environmental Research (CICERO), Oslo
(3) University of Utrecht
(4) National Obersavotory of Athens
(5) University of Bath
(6) AEA Technology, Didcot