Geostatistical Modelling Using Non-Gaussian Matern Fields
Journal article, 2015

This work provides a class of non-Gaussian spatial Matern fields which are useful for analysing geostatistical data. The models are constructed as solutions to stochastic partial differential equations driven by generalized hyperbolic noise and are incorporated in a standard geostatistical setting with irregularly spaced observations, measurement errors and covariates. A maximum likelihood estimation technique based on the Monte Carlo expectation-maximization algorithm is presented, and a Monte Carlo method for spatial prediction is derived. Finally, an application to precipitation data is presented, and the performance of the non-Gaussian models is compared with standard Gaussian and transformed Gaussian models through cross-validation.

Matern covariances

Markov random fields

variance Gamma

normal inverse Gaussian

Laplace

MCEM algorithm

SPDE

Author

Jonas Wallin

University of Gothenburg

Chalmers, Mathematical Sciences, Mathematical Statistics

David Bolin

Chalmers, Mathematical Sciences, Mathematical Statistics

University of Gothenburg

Scandinavian Journal of Statistics

0303-6898 (ISSN) 1467-9469 (eISSN)

Vol. 42 3 872-890

Subject Categories

Probability Theory and Statistics

DOI

10.1111/sjos.12141

More information

Created

10/7/2017