A0688
Title: Bayesian spatial modeling for data fusion adjusting for preferential sampling
Authors: Paula Moraga - King Abdullah University of Science and Technology (KAUST) (Saudi Arabia) [presenting]
Abstract: Spatially misaligned data are becoming increasingly common due to data collection and management advances. A Bayesian geostatistical model for combining data obtained at different spatial resolutions is presented. The flexible model can be applied in preferential sampling and spatiotemporal settings. The model assumes that underlying all observations, a spatially continuous variable can be modelled using a Gaussian random field process. The fast inference is performed via the integrated nested Laplace approximation (INLA) and the stochastic partial differential equation (SPDE) approaches. A new SPDE projection matrix for mapping the Gaussian Markov random field from the observations to the triangulation nodes is proposed to allow spatial data fusion. The performance of the new approach by means of simulation and air pollution applications is shown. The approach presented provides a useful tool in a wide range of situations where information at different spatial scales needs to be combined.
