Back to Friday's Agenda | Back to Abstracts Page | Next >>> (D102)
On the statistical approach to Lagrangian data assimilation
Leonid Piterbarg
University of Southern California
piter@math.usc.edu(Abstract received 04/14/2005 for session D)
ABSTRACT
The Lagrangian data assimilation problem is treated as a statistical problem of optimal estimating a random velocity field given noisy observations of particles released in the flow. One of the most serious obstacles on this way is the non-linear relation between Eulerian and Lagrangian velocities. We, first, discuss how to overcome this difficulty. Then, a new justification is presented for the simplest optimal interpolation procedure which already was used in our twin experiments with OGCM's. Finally, we investigate the dependence of the assimilation skill on the observation time in the framework of Markov LSM's.
Back to Friday's Agenda | Back to Abstracts Page | Next >>> (D102)
2005 LAPCOD Meeting, Lerici, Italy, June 13-17, 2005