Ensemble Riemannian Data Assimilation over the Wasserstein Space
Methodology
2021-10-11 v3
Abstract
In this paper, we present an ensemble data assimilation paradigm over a Riemannian manifold equipped with the Wasserstein metric. Unlike the Eulerian penalization of error in the Euclidean space, the Wasserstein metric can capture translation and difference between the shapes of square-integrable probability distributions of the background state and observations -- enabling to formally penalize geophysical biases in state-space with non-Gaussian distributions. The new approach is applied to dissipative and chaotic evolutionary dynamics and its potential advantages and limitations are highlighted compared to the classic variational and filtering data assimilation approaches under systematic and random errors.
Cite
@article{arxiv.2009.03443,
title = {Ensemble Riemannian Data Assimilation over the Wasserstein Space},
author = {Sagar K. Tamang and Ardeshir Ebtehaj and Peter J. Van Leeuwen and Dongmian Zou and Gilad Lerman},
journal= {arXiv preprint arXiv:2009.03443},
year = {2021}
}