Modelling spatio–temporal data from mobile monitoring stations

Environmental data are typically indexed in space and time. This work deals with modelling spatio–temporal air quality datasets, when multiple measurements are available for each space–time point. Typically this situation arises when different measurements referred to several response variables are observed in each space–time point, e.g. different pollutants or size resolved data on particular matter. Nonetheless, such a kind of data can also be obtained when using a mobile monitoring station moving along a path for a certain time period. High–frequency measurements are then obtained that refer to different space and time points. These data can be dealt with by discretising the space and time information on a grid, and then summarizing response variable(s) within each space–time point, so to have point–referenced data. Now, each spatio–temporal datum has different measurements referred to the response variable observed several times over several locations in a close neighbourhood of a specific space–time point. We model this type of data within a Bayesian hierarchical modelling framework, in which observed measurements are modelled in the first stage of the hierarchy, while the unobserved spatio–temporal process is considered in the following stages. The final model is very flexible and includes autoregressive terms in time, different structures for the variance–covariance matrix of the errors, and can manage covariates available at different space–time resolutions. This approach is applied and tailored to data coming from the PMetro project (http://www.pmetro.it), which studies urban pollution dynamics in the town of Perugia (Italy) since September, 2012: fast measure of gases and size resolved particulate matter is collected using an Optical Particle Counter located on a cabin of the Minimetro, a public conveyance that moves on a monorail on a line transect of the town. Urban microclimate information is also available for including covariates.