Heteroscedastic Bayesian point source model for spatial data

We introduce a Bayesian point source model which may be useful for modelling spatial data. It may provide a simple explanatory model for some data, whilst in other cases it may give a parsimonious representation. The model assumes that there are point sources (or sinks), usually at unknown positions, and that the mean value at a site depends on the distance from these sources. We discuss the general form of the model, and the MCMC approach for estimating model parameters. We demonstrate the methodology applying the model to a real data set.