(D103) <<< Previous | Back to Abstracts Page | Next >>> (D105)
Assimilation of Lagrangian Data Using Particle Filters
Keith Thompson, Kassiem Jacobs
Dalhousie University, Canada
keith.thompson@dal.ca(Abstract received 04/29/2005 for session D)
ABSTRACT
A straightforward scheme for assimilating Lagrangian data is described and illustrated using the idealized dynamical model of Kuznetsov, Ide and Jones (2003). This model includes a drifter that is passively advected by the flow fields of two self-advecting point vortices. The theory underlying the assimilation method is fully Bayesian and the method is implemented using a Particle Filter (PF). It is shown that the PF can accomodate the highly non-Gaussian probability density function of drifter position and can provide reliable estimates of vortex position from intermittently observed drifter positions over a wide range of parameter values. The number of ensemble members required by the PF is discussed and some simple schemes for reducing this number below O(100) are described and shown to work in this simple dynamical system. The prospect of using PF in realistic, fully nonlinear ocean model to assimilate Argo data is discussed.
Kuznetsov, L., K. Ide and C.K.R.T. Jones, 2003: A method for assimilation of Lagrangian data. Mon. Wea. Rev., 131(10), 2247-2260.
(D103) <<< Previous | Back to Abstracts Page | Next >>> (D105)
2005 LAPCOD Meeting, Lerici, Italy, June 13-17, 2005