|
Activity Number:
|
504
- Spatial Modeling on the Sphere and Over Large Domains
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 6, 2020 : 10:00 AM to 2:00 PM
|
|
Sponsor:
|
Section on Statistics and the Environment
|
|
Abstract #309870
|
|
|
Title:
|
Modeling a Tangential Vector Random Field on a Sphere: Application to Earth’s Magnetic Field
|
|
Author(s):
|
Amy Kim* and Debashis Paul and Thomas Lee and Tomoko Matsuo
|
|
Companies:
|
University of California, Davis and University of California, Davis and University of California, Davis and University of Colorado, Boulder
|
|
Keywords:
|
Spatial Statistics;
Random Field;
Geomagnetic Field;
Helmholtz–Hodge decomposition;
Linear Mixed Model;
Multivariate Analysis
|
|
Abstract:
|
We propose a parametric model for random tangential vector fields on the surface of a sphere through a vector spherical harmonics representation. It allows offering a unified framework for modeling both mean and residual fields by the linear mixed model approach and producing a meaningful decomposition of the vector field into curl-free and divergence-free fields. This random effects representation enables an efficient computation of the maximum likelihood estimate of the parameters. When the data are on an equiangular grid, the computational efficiency can be enhanced by making use of a discrete vector spherical harmonics transformation. We conduct extensive numerical studies to illustrate the estimation, model selection and prediction performance of the proposed method. We apply the proposed methodology to analyze Orsted satellite-based measurements on the earth's magnetic field.
|
Authors who are presenting talks have a * after their name.
