New Flexible Compact Covariance Model on a Sphere

TL;DR

This paper introduces a new flexible covariance model for spherical data using kernel convolution, enabling accurate non-stationary spatial prediction and simulation for large datasets.

Contribution

It presents a novel covariance approximation method on a sphere based on kernel convolution, suitable for large, non-stationary spatial datasets.

Findings

Provides detailed derivation of the covariance approximation formulas

Enables efficient spatial prediction and simulation on large spherical datasets

Improves accuracy of covariance modeling on spherical domains

Abstract

We discuss how the kernel convolution approach can be used to accurately approximate the spatial covariance model on a sphere using spherical distances between points. A detailed derivation of the required formulas is provided. The proposed covariance model approximation can be used for non-stationary spatial prediction and simulation in the case when the dataset is large and the covariance model can be estimated separately in the data subsets.