Abstract:Spatial data are often derived from multiple sources (e.g. satellites, in-situ sensors, survey samples) with different supports, but associated with the same properties of a spatial phenomenon of interest. It is common for predictors to also be measured on different spatial supports than the response variables. Although there is no standard way to work with spatial data with different supports, a prevalent approach used by practitioners has been to use downscaling or interpolation to project all the variables of analysis towards a common support, and then using standard spatial models. The main disadvantage with this approach is that simple interpolation can introduce biases and, more importantly, the uncertainty associated with the change of support is not taken into account in parameter estimation. In this article, we propose a Bayesian spatial latent Gaussian model that can handle data with different rectilinear supports in both the response variable and predictors. Our approach allows to handle changes of support more naturally according to the properties of the spatial stochastic process being used, and to take into account the uncertainty from the change of support in parameter estimation and prediction. We use spatial stochastic processes as linear combinations of basis functions where Gaussian Markov random fields define the weights. Our hierarchical modelling approach can be described by the following steps: (i) define a latent model where response variables and predictors are considered as latent stochastic processes with continuous support, (ii) link the continuous-index set stochastic processes with its projection to the support of the observed data, (iii) link the projected process with the observed data. We show the applicability of our approach by simulation studies and modelling land suitability for improved grassland in Rhondda Cynon Taf, a county borough in Wales.

Nonstationary Spatial Gaussian Markov Random Fields

Why not a thin plate spline for spatial models? A comparative study using Bayesian inference

Local Linear M‐estimation in Non‐parametric Spatial Regression

Flexible Basis Representations for Modeling Large Non-Gaussian Spatial Data

Bayesian Nonstationary Spatial Modeling for Very Large Datasets

Spatial Gaussian Markov Random Fields: Modelling, Applications and Efficient Computations

Bayesian inference for high-dimensional nonstationary Gaussian processes

Ideal Bayesian Spatial Adaptation

Mixed-Stationary Gaussian Process for Flexible Non-Stationary Modeling of Spatial Outcomes

Modeling massive highly-multivariate nonstationary spatial data with the basis graphical lasso

Direct Sampling of Bayesian Thin-Plate Splines for Spatial Smoothing

Gaussian Markov random field spatial models in GAMLSS

Modelling of Non-Stationary Spatial Structure Using Parametric Radial Basis Deformations

Spatial Latent Gaussian Modelling with Change of Support

Nearest-Neighbor Mixture Models for Non-Gaussian Spatial Processes

Bayesian Modeling for Nonstationary Spatial Point Process via Spatial Deformations

A Scalable Partitioned Approach to Model Massive Nonstationary Non-Gaussian Spatial Datasets

Non-Gaussian Matérn fields with an application to precipitation modeling

Nonstationary Spatial Process Models with Spatially Varying Covariance Kernels

Bayesian Modeling with Spatial Curvature Processes

Bayesian Function-on-Function Regression for Spatial Functional Data