Research

Spatio-temporal modeling of particulate matter concentration through the SPDE approach


Reference:

Cameletti, M., Lindgren, F., Simpson, D. and Rue, H., 2012. Spatio-temporal modeling of particulate matter concentration through the SPDE approach. AStA Advances in Statistical Analysis, 97 (2), pp. 109-131.

Related documents:

[img]
Preview
PDF (Author's accepted version) - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (3671kB) | Preview

    Official URL:

    http://dx.doi.org/10.1007/s10182-012-0196-3

    Abstract

    In this work, we consider a hierarchical spatio-temporal model for particulate matter (PM) concentration in the North-Italian region Piemonte. The model involves a Gaussian Field (GF), affected by a measurement error, and a state process characterized by a first order autoregressive dynamic model and spatially correlated innovations. This kind of model is well discussed and widely used in the air quality literature thanks to its flexibility in modelling the effect of relevant covariates (i.e. meteorological and geographical variables) as well as time and space dependence. However, Bayesian inference-through Markov chain Monte Carlo (MCMC) techniques-can be a challenge due to convergence problems and heavy computational loads. In particular, the computational issue refers to the infeasibility of linear algebra operations involving the big dense covariance matrices which occur when large spatio-temporal datasets are present. The main goal of this work is to present an effective estimating and spatial prediction strategy for the considered spatio-temporal model. This proposal consists in representing a GF with Matérn covariance function as a Gaussian Markov Random Field (GMRF) through the Stochastic Partial Differential Equations (SPDE) approach. The main advantage of moving from a GF to a GMRF stems from the good computational properties that the latter enjoys. In fact, GMRFs are defined by sparse matrices that allow for computationally effective numerical methods. Moreover, when dealing with Bayesian inference for GMRFs, it is possible to adopt the Integrated Nested Laplace Approximation (INLA) algorithm as an alternative to MCMC methods giving rise to additional computational advantages. The implementation of the SPDE approach through the R-library INLA (www.r-inla.org) is illustrated with reference to the Piemonte PM data. In particular, providing the step-by-step R-code, we show how it is easy to get prediction and probability of exceedance maps in a reasonable computing time. © 2012 Springer-Verlag.

    Details

    Item Type Articles
    CreatorsCameletti, M., Lindgren, F., Simpson, D. and Rue, H.
    DOI10.1007/s10182-012-0196-3
    DepartmentsFaculty of Science > Mathematical Sciences
    Publisher StatementCameletti_et_al_2012.pdf: The original publication is available at www.springerlink.com
    RefereedYes
    StatusPublished
    ID Code32278

    Export

    Actions (login required)

    View Item

    Document Downloads

    More statistics for this item...