Etude de l'utilisation d'une moyenne arithmétique des mesures d'un réseau de surveillance comme indicateur de niveau de pollution atmosphérique en milieu urbain
Indice de fumées noires, région parisienne, 1991-1992

Epidemiologists who study the relationship between air pollution and health, often compare time-series analysis of daily levels of air pollutants to daily incidence of health effects among the population recorded in the same area. Most of them simply use the arithmetic mean of the daily means of the levels of a pollutant measured at each station of the network to estimate the daily level of this pollutant over the study area. Nevertheless, few authors have been interested in the validation of this use. The purpose of this work was to validate it in the case of black smoke measurements at six stations located in Paris and his close suburbs. The analysis of the statistical distribution of the levels measured at each station, of the time series by the Box-Jenkins approach and, of the correlations between them, allowed us to verity the temporal and geographical homogeneity between data collected at the six stations. This work is a first step for the validation of ecological exposure assessment of an urban population.