THEORETICAL CLIMATE DYNAMICS GROUP
UCLA's IEM 4 Abstract

Data Assimilation

TCD
home		 DA DAMEE IEM3 IEM4
Model

An Expert-System Approach to Efficient State and Parameter Estimation for Ocean and Coupled General Circulation Models

Kayo Ide, Michael Ghil and Achim Wirth

Department of Atmospheric Sciences and
Institute of Geophysics and Planetary Physics,
University of California, Los Angeles, CA 90095-1565, U.S.A.

Accurate parametrization of mass, momentum and energy fluxes across the ocean-atmosphere interface is essential in modeling the oceans' and coupled ocean-atmosphere general circulation. Erroneous values of the surface-forcing parameters for an ocean general circulation model (GCM), as well as of coupling coefficients for a coupled GCM, may lead to highly distorted model dynamics and have a negative impact on both long-term simulations and short-term prediction skills. To improve these crucial parametrizations, data assimilation can be used to estimate model parameters, along and simultaneously with the state (Hao, 1994; Hao and Ghil, 1995; Navon, 1997, and further references therein). However, the fluxes across the interface are difficult to observe directly and require therefore an innovative observing-system design. The huge computational burden of ocean and coupled GCM runs alone has forced us to use fairly simple-minded data-assimilation methods even for state estimation alone; relatively little work has been done so far on parameter estimation.

Our goal here is to develop a comprehensive data-assimilation system to estimate concurrently both the state and a few key parameters for the ocean and coupled GCMs. In the present study, we propose an intelligent, data-adaptive sequential-estimation approach to the development of such a system. The system automatically chooses the best suited methods for each of the forecast and analysis steps in an assimilation cycle from several -- simple to sophisticated -- sequential methods. The selection is made based on the expected number and nature of the observations available at the analysis stage, as well as on the model dynamics valid during the forecast interval. This approach also allows us to estimate the error covariance in regions where advective nonlinearities play an important role, such as the western boundary currents and mid-ocean jets.

References
Hao, Z., 1994: Data Assimilation for Interannual Climate-Change Prediction, Ph.D. Thesis, UCLA, 224pp.
Hao, Z., and M. Ghil, 1995: Sequential parameter estimation for a coupled ocean-atmosphere model. In Proc. 2nd WMO Int'l Symp. Assim. Obs. Meteor. & Oceanogr., Tokyo, March 1995, WMO/TD-No. 651, PWPR Report Series No. 5, WMO Geneva, Switzerland, Vol. I, pp. 181-186.
Navon, I.M., 1997: Parameter estimation and identifiability in meteorology and oceanography, Dyn. Atmos. Oceans, in press.

[ TCD home | DA | DAMEE | IEM3 | IEM4 Model ]
April 1, 1997