Non-parametric approximations for anisotropy estimation in two-dimensional differentiable Gaussian random fields

Abstract

Spatially referenced data often have autocovariance functions with elliptical isolevel contours, a property known as geometric anisotropy. The anisotropy parameters include the tilt of the ellipse (orientation angle) with respect to a reference axis and the aspect ratio of the principal correlation lengths. Since these parameters are unknown a priori, sample estimates are needed to define suitable spatial models for the interpolation of incomplete data. The distribution of the anisotropy statistics is determined by a non-Gaussian sampling joint probability density. By means of analytical calculations, we derive an explicit expression for the joint probability density function of the anisotropy statistics for Gaussian, stationary and differentiable random fields. Based on this expression, we obtain an approximate joint density which we use to formulate a statistical test for isotropy. The approximate joint density is independent of the autocovariance function and provides conservative probability and confidence regions for the * petrakis@mred.tuc.gr † Corresponding author; dionisi@mred.tuc.gr anisotropy parameters. We validate the theoretical analysis by means of simulations using synthetic data, and we illustrate the detection of anisotropy changes with a case study involving background radiation exposure data. The approximate joint density provides (i) a standalone approximate estimate of the anisotropy statistics distribution (ii) informed initial values for maximum likelihood estimation, and (iii) a useful prior for Bayesian anisotropy inference.

Boltzmann-Gibbs Random Fields with Mesh-free Precision Operators Based on Smoothed Particle Hydrodynamics

“…In this case, isotropic dependence can be restored by means of suitable rescaling and rotation transformations, e.g. [20,21].…”

On Some Characteristics of Gaussian Covariance Functions

“…However, in such cases, the Gaussian covariance function that is extremely smooth is not the most suitable candidate, while ‘rougher’ models, for example, Whittle–Matern functions that offer controlled smoothness and functions based on rational spectral densities (Hristopulos, 2003; Hristopulos & Elogne, 2007), are better candidates. Finally, the prevalence of the isotropic assumption is rarely due to the underlying physical reality (which is often anisotropic) but rather to the difficulty to accurately estimate the anisotropy parameters in the case of scattered data (Chorti & Hristopulos, 2008; Petrakis & Hristopulos, 2017). •When a small part of the Earth is considered, it is worth working with two‐dimensional Euclidean space of a planar projection of the data points.…”

The land–sea coastal border: a quantitative definition by considering the wind and wave conditions in a wave-dominated, micro-tidal environment

“…1). The metric of the geometric anisotropy then becomes their ratio and Hristopulos, 2008;Petrakis and Hristopulos, 2017). An R value close to unity means that u and v are isotropic, i.e.…”