Air pollutant data are compositional in character because they describe quantitatively the parts of a whole (atmospheric composition). However, it is common to use air pollutant concentrations in statistical models without considering this characteristic of the data and, therefore, without control of common statistical problems, such as spurious correlations and subcompositional incoherence. This paper now proposes a daily multivariate spatio-temporal model with a compositional approach. The air pollution spatio-temporal model is based on a dynamic linear modelling framework with Bayesian inference. The novel modelling methodology was applied in an urban area for carbon monoxide (CO, mg·m−3), sulfur dioxide (SO2, μg·m−3), ozone (O3, μg·m−3), nitrogen dioxide (NO2, μg·m−3), and particulate matter less than 2.5 μm in aerodynamic diameter (PM2.5, μg·m−3). The proposal complemented and improved the conventional approach in air pollution modelling. The main improvements come from a fast multivariate data description, high spatial-correlation, and adequate modelling of air pollutants with high variability.

Wildfires are natural ecological processes that generate high levels of fine particulate matter (PM_2.5) that are dispersed into the atmosphere. PM_2.5 could be a potential health problem due to its size. Having adequate numerical models to predict the spatial and temporal distribution of PM_2.5 helps to mitigate the impact on human health. The compositional data approach is widely used in the environmental sciences and concentration analyses (parts of a whole). This numerical approach in the modelling process avoids one common statistical problem: the spurious correlation. PM_2.5 is a part of the atmospheric composition. In this way, this study developed an hourly spatio-temporal PM_2.5 model based on the dynamic linear modelling framework (DLM) with a compositional approach. The results of the model are extended using a Gaussian–Mattern field. The modelling of PM_2.5 using a compositional approach presented adequate quality model indices (NSE = 0.82, RMSE = 0.23, and a Pearson correlation coefficient of 0.91); however, the correlation range showed a slightly lower value than the conventional/traditional approach. The proposed method could be used in spatial prediction in places without monitoring stations.

Revista Tecnología y Ciencias del Agua (TyCA) del Instituto Mexicano de Tecnología del Agua (IMTA)