Organised session at Spatial statistics conference
Published: 10 August 2015
The invited session ‘Spatio-temporal modeling of exposure to air pollution’ was held on Wednesday 10th June in Avignon as part of the conference ‘Spatial Statistics 2015: Emerging Patterns’. The session consisted of contributions from three speakers and was delivered to a large audience with around 60 delegates in attendance.
The invited session ‘Spatio-temporal modeling of exposure to air pollution’ was held on Wednesday 10th June in Avignon as part of the conference ‘Spatial Statistics 2015: Emerging Patterns’. The session consisted of contributions from three speakers and was delivered to a large audience with around 60 delegates in attendance.
Prof. Alessandro Fasso, from the University of Bergamo was the first speaker who discussed new multivariate approaches for modeling the spatial distribution of seven pollutants across the whole of Europe between 2009 and 2011. Handling the non-Gaussianity and multi-resolution nature of the data were particularly emphasized.
The second speaker, Alastair Rushworth from the University of Strathclyde, presented an innovative spatial model for capturing complex heterogeneity in disease risk after accounting for exposure to a heterogeneous air pollution process and other relevant risk factors. The talk highlighted the challenges inherent in incorporating modelled air pollution estimates and uncertainty measures, and a two-stage approach to incorporating this information was proposed.
The final speaker in the session was Prof. Sujit Sahu, from the University of Southampton who proposed new Bayesian spatio-temporal hierarchical models for predicting the pattern of air pollution at a fine grid of locations. The models incorporated data from a combination of sources including atmospheric air quality model output and raw observations obtained at air pollution sensors. The resulting models were found to be highly accurate and performed better than competing approaches.
First published: 10 August 2015