HR: 1300h
AN: OS52D-06 [PDF]
TI: Sensitivity of a Tropical Pacific Data Assimilation Scheme to Model Parameterizations and Wind
Errors
AU: * Perez, R C
EM: rperez@coas.oregonstate.edu
AF: College of Oceanic and Atmospheric Sciences, 104 Ocean Administration Building, Corvallis, OR 97331
United States
AU: Miller, R N
EM: miller@coas.oregonstate.edu
AF: College of Oceanic and Atmospheric Sciences, 104 Ocean Administration Building, Corvallis, OR 97331
United States
AB:
We have developed an efficient four-dimensional multivariate
data assimilation scheme for the tropical Pacific
Ocean.
The scheme consists of the Gent-Cane (1989) general circulation model
(GCM)
and assimilation of tide gauge data via a reduced state space
Kalman filter. The error models are constructed
in spaces spanned by enough Empirical Orthogonal Functions to capture
up to 99\% of the variance.
We will analyze changes in the model error statistics that result
from improving the quality of the wind
stress forcing and increasing
the horizontal and vertical resolution of our
GCM.
Statistical hypothesis testing and time series analysis
will be employed to evaluate the results of our experiments.
In addition, we will
study the consequences of the linearity assumption for the evolution of
the error covariances and the meridional structure of the resulting
error covariances.
DE: 3337 Numerical modeling and data assimilation
DE: 4231 Equatorial oceanography
DE: 4255 Numerical modeling
SC: Ocean Sciences [OS]
MN: 2004 Ocean Sciences Meeting