On the Use of Satellite Data for Spatial Health Risk Assessment of Urban Air Pollutants

Background/Aims: The state of the art in air quality assessmentcomprises information and data processing tools using only data fromground-based measurement and atmospheric modelling. Groundmeasurements are not taken from dense enough monitoring networksaround the world to permit a satisfactory analysis of the influence of airpollution on the health of vulnerable population groups. Attempts toimprove our estimation of atmospheric pollutant concentrations at theurban and regional scale from combining ground data with numericalmodeling are hampered by the need for high quality and up-to-dateemissions inventories, as well as accurate estimates of initial andboundary conditions of the models. Information derived from earthobservation satellites can bridge the gap between models, simulating thetransport and chemical transformation of atmospheric pollutants andanalytical observations.Methods: A data fusion methodology was developed to integrate satellitedata with ground-based information and atmospheric modeling to deriveparticulate matter and ozone loading at the ground level. Physicalproperties of tropospheric aerosol and ozone are linked with the atmospheric physical– chemical processes that determine the total massconcentration and size distribution of particulate matter and theconcentration of ozone. Coupling these with spatially explicitlyexposure–response functions and population data, it results in refinedmaps of health risk attributable to air pollution.Results: The methodology was implemented in Athens, Greece andRome, Italy, 2 capitals characterized by intense photochemical pollutionand long-range transport of dust. Maps of health risk were produced. Thespatially scalar nature of the approach allowed us to evaluate the impactof risk modifiers such as the existence of urban vegetation and populationsusceptibility.Conclusion: Satellite data can be used efficiently to improve the spatiallink between environmental pollution and human health. The data fusionmethod proposed in the present study opens the way toward the enhanceduse of this valuable information in spatial epidemiology andenvironmental health science