Control of a free-surface barotropic model of the Bay of Biscay by assimilation of sea-level data in presence of atmospheric forcing errors
Julien Lamouroux(1) , Pierre De Mey(2) , Florent Lyard(2) , Frédérique Ponchaut(3) , and Eric Jeansou(3)
31401 Toulouse Cedex 09,
(2) POC/LEGOS, POC/LEGOS, 31401 Toulouse Cedex 09, France
(3) NOVELTIS, Parc Technologique du Canal, 31520 RAMONVILLE SAINT AGNE, France
A data assimilation method is set up in the barotropic, free-surface,
finite element MOG2D model, implemented in the Bay of Biscay and nested
in a North East Atlantic domain. The model is forced by the ARPEGE
In a first step, we explore the model error subspace in presence of
coherent atmospheric forcing errors. This is done via an ensemble
modelling approach in which the atmospheric fields are perturbed in a
multivariate and coherent way : by generating an a priori ensemble of
perturbed atmospheric forcing fields, and calculating the corresponding
a posteriori ensemble of model simulations, one can approximate the
forecast errors of the model by ensemble spread statistics, such as
background error Ensemble EOFs.
These approximated model error covariances, in form of 6D-EOFs (Sea
Level Anomaly, barotropic velocities, atmospheric pressure and
wind-stress), are shown to be neither homogeneous over the domain, nor
stationary, since they are very dependent on the meteorological forcing.
Such statistics are then used in a Reduced-Order Optimal Interpolation
sequential scheme (SEQUOIA, developed at LEGOS/POC) to constrain the
model forecast via sea level data assimilation. Twin experiments are
conducted in the last quarter of 1999. Results show that the use of
time-independent error statistics allows to control the model but that
time-dependent statistics often lead to better results, advocating for a
more advanced scheme in a future step.
Finally, several Observing System Experiments (OSEs) and
Observing-System Simulation Experiments (OSSEs), are carried out in
order to test the sensitivity of the results to the altimetric
configuration, with or without tide gauges acting in a complementary
manner. The impact of velocity measurements near the coast (HF radars)
is also illustrated.