P2.1 ENSEMBLE DATA ASSIMILATION: EXPERIMENTS USING NASAS GEOS COLUMN PRECIPITATION MODEL

1
P2.1 ENSEMBLE DATA ASSIMILATION EXPERIMENTS
USING NASAS GEOS COLUMN PRECIPITATION MODEL
D. Zupanski1, A. Y. Hou2, S. Zhang2, M.
Zupanski1, C. D. Kummerow1, and S. H. Cheung3
1Colorado State University, Fort Collins, CO
2NASA Goddard Space Flight Center, Greenbelt, MD
2NASA Ames Research Center, Moffett Field, CA
Goals Develop a unified probabilistic data as
similation, model error estimation, and ensemble
forecasting method Examine the method in applica
tion to NASAs GEOS column models
Assimilate satellite precipitation observations
(SSM/I, TMI, GPM) Estimate atmospheric state, mod
el errors, and models empirical parameters
Determine uncertainty of all estimates
Evaluate capability of the method to provide new
knowledge about atmospheric processes
Methodology Maximum Likelihood Ensemble Filte
r (MLEF, Zupanski 2005 Zupanski and Zupanski 20
05) Developed using ideas from Variational data
assimilation (3DVAR, 4DVAR) Iterated Kalman Filte
rs Ensemble Transform Kalman Filter (ETKF, Bishop
et al. 2001)
Initial tuning of the MLEF algorithm in
application to NASA/GEOS-4 column model
Single column version of the GOES-4 GCM 55 leve
l model, two state variables T and Q
10 ensembles, 110 observations of T and Q
10 data assimilation cycles 6-h data assimilation
interval PSAS analyses used as observations an
d forcing
Further experiments employing NASA/GEOS-5 column
model Single column version of the GOES-5 GCM
GOES-5 includes a finite-volume dynamical core
and full physics package The model is driven by
external data (ARM observations)
40 level model, two control variables T and Q
10 ensembles, 80 observations of T and Q
40 data assimilation cycles 6-h data assimilation
interval Model simulated observations with ran
dom noise
Analysis error covariance
Forecast error covariance
R1/2 ?
Observability matrix
R1/2 2 ?
Degrees of freedom (DOF) for signal
(Rodgers 2000)
Minimize cost function J
eigenvalues of C
Acknowledgements This research is partially funde
d by NASA grants 621-15-45-78 and NAG5-12105)
References Bishop, C. H., B. J. Etherton, and S.
Majumjar, 2001 Adaptive sampling with the
ensemble transform Kalman filter. Part 1
Theoretical aspects. Mon. Wea. Rev., 129,
420436. Rodgers, C. D., 2000 Inverse Methods fo
r Atmospheric Sounding Theory and Practice.
World Scientific, 238 pp. Zupanski, D., and M. Zu
panski, 2005 Model error estimation employing
ensemble data assimilation approach.
Submitted to Mon. Wea. Rev. Available at
ftp//ftp.cira.colostate.edu/Zupanski/manuscripts/
MLEF_model_err.revised2.pdf Zupanski, M., 2005
Maximum likelihood ensemble filter Theoretical
aspects. Accepted to Mon. Wea. Rev. Available
at ftp//ftp.cira.colostate.edu/milija/papers/MLEF
_MWR.pdf.
Change of variable
Experiments Examine sensitivity of innovation s
tatistics to observation errors
Evaluate information content of the observations
(DOF)
augmented control variable of dim Nstate Nens
(includes initial conditions, model error,
empirical
parameters)
control variable in ensemble space of dim Nens