Implicit particle filtering for models with partial noise, and an application to geomagnetic data assimilation
Authors:
Open Access
- 19 June 2012
- journal article
- Published by Copernicus GmbH in Nonlinear Processes in Geophysics
- Vol. 19 (3) , 365-382
Abstract
Implicit particle filtering is a sequential Monte Carlo method for data assimilation, designed to keep the number of particles manageable by focussing attention on regions of large probability. These regions are found by minimizing, for each particle, a scalar function F of the state variables. Some previous implementations of the implicit filter rely on finding the Hessians of these functions. The calculation of the Hessians can be cumbersome if the state dimension is large or if the underlying physics are such that derivatives of F are difficult to calculate, as happens in many geophysical applications, in particular in models with partial noise, i.e. with a singular state covariance matrix. Examples of models with partial noise include models where uncertain dynamic equations are supplemented by conservation laws with zero uncertainty, or with higher order (in time) stochastic partial differential equations (PDE) or with PDEs driven by spatially smooth noise processes. We make the implicit particle filter applicable to such situations by combining gradient descent minimization with random maps and show that the filter is efficient, accurate and reliable because it operates in a subspace of the state space. As an example, we consider a system of nonlinear stochastic PDEs that is of importance in geomagnetic data assimilation.Keywords
All Related Versions
This publication has 42 references indexed in Scilit:
- Implicit Particle Methods and Their Connection with Variational Data AssimilationMonthly Weather Review, 2013
- Inferring internal properties of Earth's core dynamics and their evolution from surface observations and a numerical geodynamo modelNonlinear Processes in Geophysics, 2011
- Implicit particle filters for data assimilationCommunications in Applied Mathematics and Computational Science, 2010
- Beyond Gaussian Statistical Modeling in Geophysical Data AssimilationMonthly Weather Review, 2010
- Implicit sampling for particle filtersProceedings of the National Academy of Sciences, 2009
- Stochastic Tools in Mathematics and Science2009
- Sharp failure rates for the bootstrap particle filter in high dimensions2008
- Spectral Methods2006
- A tutorial on particle filters for online nonlinear/non-Gaussian Bayesian trackingIEEE Transactions on Signal Processing, 2002
- Tropical Cyclone Prediction Using a Barotropic Model Initialized by a Generalized Inverse MethodMonthly Weather Review, 1993