Use of multivariate time series techniques to estimate the impact of particulate matter on the perceived annoyance

As well known, Particulate matter (PM) is an air pollutant that causes damage to the health of humans, other animals, plants, affects the climate and is a potential cause of annoyance through deposition on various surfaces. The perceived annoyance caused by particulate matter is related mainly to the increase of settled dust in urban and residential environments. PM can originate from many sources, i.e., paved and unpaved roads, buildings, agricultural operations and wind erosion represent the largest contributions beyond the relatively minor vehicular and industrial sources emissions. The aim of this paper is to quantify the relationship between perceived annoyance and particulate matter concentration and to estimate the relative risk (RR). The data was collected in the Metropolitan Region of Vitoria (MRV), Brazil. For this purpose, the variables of interest were modelled using vector time series model (VAR), principal component analysis (PCA), and logistic regression (LOG). The combination of these techniques resulted in a hybrid model denoted as LOG-PCA-VAR which allows to estimate RR by handling multipollutant effects. This study shows that there is a strong association between the perceived annoyance and different sizes of PM. The estimates of RR indicate that an increase in air pollutant concentrations significantly contributes in increasing the probability of being annoyed. [Display omitted] •Particulate matter is an air pollutant that causes damage to the health of humans.•Association between air pollutants and annoyance is interest in many studies.•The combination of statistical tools is a new contribution in this methodology.•The relative risk (RR) is computed for all methods considered.•Even low particles deposition induces high levels of nuisance reported in Vitória.