Outlier Identification of Concentrations of Pollutants in Environmental Data Using Modern Statistical Methods

Abstract

The article is focused on identification of outlier measurements in environmental data which may significantly affect the future results of the analysis and interpretation of results. For this reason, their identification forms an integral part of data analysis. The aim of this article is to perform statistical analysis that automatically identifies segments of outlier measurements. The results were demonstrated on real concentration data. The methodological procedure was used to evaluate particulate matter of the PM10 fraction size from two monitoring stations located in Brno, Czech Republic.