ASkew–Gaussian Spatio–Temporal Process with Non–Stationary Correlation Structure

Abstract.this paper develops a new class of spatio-temporal process models that cansimultaneously capture skewness and non-stationarity. the proposed approach whichis based on using the closed skew-normal distribution in the low-rank representation ofstochastic processes, has several favorable properties. in particular, it greatly reducesthe dimension of the spatio-temporal latent variables and induces flexible correlationstructures. bayesian analysis of the model is implemented through a gibbs mcmcalgorithm which incorporates a version of the kalman filtering algorithm. all fullyconditional posterior distributions have closed forms which show another advanta-geous property of the proposed model. we demonstrate the eciency of our modelthrough an extensive simulation study and an application to a real data set comprisedof precipitation measurements.