(D106) <<< Previous | Back to Abstracts Page | Back to Friday's Agenda
The impact of temperature and salinity data from profiling floats in the Mediterranean Forecasting System.
F. Raicich, A. Griffa, A. Molcard, V. Rupolo
CNR, Istituto di Scienze Marine, viale Romolo Gessi 2, 34123 Trieste, Italy
fabio.raicich@ts.ismar.cnr.it(Abstract received 04/18/2005 for session D)
ABSTRACT
The impact of temperature (T) and salinity (S) from profiling floats (MedARGO) in the Mediterranean Forecasting System (MFS) is assessed by means of numerical twin experiments. These consist of a control run, in which ‘true’ T and S data are extracted from a Mediterranean General Circulation Model (MOM-1, 1/8°x1/8°, 31 vertical levels), a free run, with different initial conditions relative to the control run, and an assimilation run. The assimilation T and S data is performed by means of SOFA bivariate optimal interpolation scheme. The convergence of the assimilation run towards the control run is assessed by means of standard deviations of differences between the two runs. The convergence of the free run towards the control run, assessed in the same way, is used for reference, since it shows the ability of the model to converge towards the control run without data assimilation. The data impact is quantified by the error reduction achieved in the assimilation run relative to the free run. Both idealized, although realistic, and real MedARGO float distributions are studied. Twin experiments are performed in summer and winter circulation conditions. In the idealized case 47 floats are released along Volunteer Observing Ship (VOS) tracks and advected by the velocity field computed by 1/8° OPA model, according to the real operative scheme. The maximum error reduction is about 20-30%, depending on season, geographic area and depth. The most effective float trajectories sample dynamically active regions, such as the frontal areas in the western Mediterranean and the northwestern Ionian Sea. The MedARGO data impact is compared with that of T and S observations collected along VOS tracks. Real data distributions are also studied, using the profile positions during the MFS operational activity. In this case the sampling does not provide enough data to achieve notable error reductions when only MedARGO data are assimilated, but the data impact is improved when MedARGO and VOS data are associated.
(D106) <<< Previous | Back to Abstracts Page | Back to Friday's Agenda
2005 LAPCOD Meeting, Lerici, Italy, June 13-17, 2005