Algorithms for space, time and content aggregation of air quality statistics

Official statistics on air quality are usually published either as simple averages of pollutant concentrations at single monitoring stations over long time intervals, or as simple averages of concentration values collected at the set of stations which belong to a monitoring network. Rules are needed to aggregate the elementary air quality data over space, time, and content of pollution, for the intelligence of cause-and-effect relations of pollution with health, demographic, ecological, and developmental statistics. In this paper, a model for time, space, and content aggregation of elementary pollution data, which is statistically optimal for air quality estimation, is put forward. Systems of weights for the estimation of the pollution level over a region, during an observational period, are deduced from the model. The expected value and the estimate variance of an air quality index which combines together normalised elementary indicators of atmospheric pollution are analytically determined. An index for the mediation of stratum estimates over an area is suggested, too.