Title: Offline Test Criteria for Filtering Complex Dynamical Systems

Speaker: John Harlim, New York University

Date/Time: Monday, June 25, 2007, 9:00am – 10:00am

Location: CSRI Building, Room 90 (Sandia NM)

Brief Abstract: The current practical numerical weather forecasting involves General Circulation Model where the prognostic model is discretized in space and time while the unresolved processes are parameterized. Making accurate predictions with partial observation of this high dimensional complex dynamical system is a major challenge. Many ensemble Kalman filter approaches have been successfully developed for solving these extremely complex systems; however, new issues arise in the practical application of these filtering strategies. One such issue is that it requires many realizations that involve expensive calculations for propagating each ensemble member forward in time.
In this talk, I will discuss recent theoretical work of Majda and Grote (PNAS 2006, 2007): the possibility of using larger time steps that sometimes violates the classical CFL-stability condition for an explicit difference scheme. With this approach, one can reduce the computational time of propagating the large ensemble size required for filtering while still obtaining stable and statistically accurate filtering results. I will also show results from the ongoing joint work with Emilio Castronovo and Andy Majda in developing and implementing new test criteria for addressing these issues. Specifically, I will discuss results of filtering an Ornstein-Uhlenbeck process, an advection-diffusion equation, and a 40-dimensional chaotic Lorenz-96 model that mimics the midlatitude weather pattern.

CSRI POC: Danny Dunlavy, (505) 284-6092