Assimilating Ferry Box data into the Aegean Sea model

作者：Korres G^*; Ntoumas M; Potiris M; Petihakis G

来源：Journal of Marine Systems, 2014, 140: 59-72.

DOI：10.1016/j.jmarsys.2014.03.013

• Aegean Sea
• Data assimilation
• SEEK filter
• FerryBox
• POSEIDON system

Operational monitoring and forecasting of marine environmental conditions is a necessary tool for the effective management and protection of the marine ecosystem. It requires the use of multi-variable real-time measurements combined with advanced physical and ecological numerical models. Towards this, a FerryBox system was originally installed and operated in the route Piraeus-Heraldion in 2003 for one year. Early 2012 the system was upgraded and moved to a new high-speed ferry traveling daily in the same route as before. This route is by large traversing the Cretan Sea being the largest and deepest basin (2500 m) in the south Aegean Sea. The HCMR Ferry Box is today the only one in the Mediterranean and thus it can be considered as a pilot case. The analysis of FerryBox SST and SSS in situ data revealed the presence of important regional and sub-basin scale physical phenomena, such as wind-driven coastal upwelling and the presence of a mesoscale cyclone to the north of Crete. In order to assess the impact of the FerryBox SST data in constraining the Aegean Sea hydrodynamic model which is part of the POSEIDON forecasting system, the in situ data were assimilated using an advanced multivariate assimilation scheme based on the Singular Evolutive Extended Kalman (SEEK) filter, a simplified square-root extended Kalman filter that operates with low-rank error covariance matrices as a way to reduce the computational burden. Thus during the period mid-August 2012-mid January 2013 in addition to the standard assimilating parameters, daily SST data along the ferryboat route from Piraeus to Heraldion were assimilated into the model. Inter-comparisons between the control run of the system (model run that uses only the standard data set of observations) and the experiment where the observational data set is augmented with the FerryBox SST data produce interesting results. Apart from the improvement of the SST error, the additional assimilation of daily of FerryBox SST observations is found to have a significant impact on the correct representation of the dynamical dipole in the central Cretan Sea and other dynamic features of the South Aegean Sea, which is then depicted in the decrease of the basin wide SSH RMS error.

• 出版日期2014-12