On the Use of Satellite Data to Estimate Spatially Referenced Health Risk of Air Pollution

Background/Aims: The state of the art in air quality assessmentcomprises information and data processing tools using only data fromground-based measurement and atmospheric modeling. Groundmeasurements of air pollutants are not taken from dense enoughmonitoring networks around the world to permit a satisfactory analysis ofthe actual influence of fine urban aerosol and ozone on the health ofvulnerable population groups, such as the elderly, children under the ageof 15, asthmatics, people with cardiovascular problems. Introduction ofinformation derived from Earth Observation satellite data can be used tobridge the gap between models simulating the transport and chemicaltransformation of ambient air pollutants, and analytical observations.Methods: A data and model fusion methodology has been developed tointegrate the 3 information data sources (i.e., Earth Observation EO,ground-based information and atmospheric modeling) to derive PM10,PM2.5 and ozone loading at the ground level. The resulting pollution mapsare coupled to epidemiologically derived exposure-response functions andpopulation data, resulting in high resolution morbidity and mortalityindicator maps. Comparison of these maps with actual health outcomestatistics reveals new insight into the spatial link between air pollutionexposure and public health risk.Results: The data assimilation methodology was applied in Athens,Greece and Rome, Italy, 2 of the largest capitals in Southern Europe,characterized by increased photochemical pollution and long-rangetransport of PM. Results showed that the proposed methodology improvedsignificantly the spatial accuracy of health risk estimates. Given the scalarnature of the approach, refined risk estimates can be made in areaspopulated by susceptible sub-groups taking into account risk modifierssuch as the existence of urban vegetation and socioeconomic condition.Conclusion: Satellite-based atmosphere observation can be a keycontributor to the determination of the spatial relationship between airpollution and public health risk. Efficient data and model fusion is theoptimal way to achieving this.