Spatial-Temporal modelization of the NO2 concentration data in mainland Portugal

In this work, we propose characterizing the evolution of the NO2 levels in Portugal, by using geostatistical approaches that deal with both the space and time coordinates and by establishing different factors affecting the concentrations detected. We also discuss appropriate tools for modelling the presence of a trend or a seasonal effect, together with specific approaches for estimation of the space-time variability. The correlation structure of the NO2 levels can be approximated, enabling to make use of the kriging techniques for prediction, without requiring data from a dense monitoring network. Proceeding in this way, the spatio-temporal patterns of the NO2 data can be derived, as well as the corresponding deviation errors. Among the potentialities of the information provided, compliance with the regulations about the NO2 concentrations can be checked.