Inverses of Matern Covariances on Grids
This work identifies limitations in a widely used approximation for spatial statistics, which is incremental but important for researchers in geostatistics and Gaussian process modeling.
The study analyzed the aliased spectral densities of Matérn covariance functions on regular grids, revealing that a popular SPDE approximation overestimates high-frequency power and fails to improve inverse accuracy with finer grids, except in 1D exponential cases, leading to overestimation of spatial range parameters in simulations.
We conduct a study of the aliased spectral densities of Matérn covariance functions on a regular grid of points, providing clarity on the properties of a popular approximation based on stochastic partial differential equations; while others have shown that it can approximate the covariance function well, we find that it assigns too much power at high frequencies and does not provide increasingly accurate approximations to the inverse as the grid spacing goes to zero, except in the one-dimensional exponential covariance case. We provide numerical results to support our theory, and in a simulation study, we investigate the implications for parameter estimation, finding that the SPDE approximation tends to overestimate spatial range parameters.