The Levantine Sea
Coupled Modeling and Forecasting Experience
Emin Özsoy*, Ali Aydoğdu, Hazem Nagy, Adil Sözer and Murat Gündüz
Institute of Marine Sciences, Middle East Technical University, Mersin, Turkey
* e-mail address:
Abstract. The complex, multi-scale dynamics of the Mediterranean Sea gives fascination to oceanographers trying to understand and predict its short or long-term behavior. Eddies, jets, meanders, wind driven currents, topographic / continental shelf waves, inertial / internal / gravity waves and sea-level oscillations add significant time and space dependence to the basic circulation exemplified by the satellite SST and chlorophyll fields. Confirmation of these features of the Levantine Basin circulation elaborated by long-term experiments and modeling is used to make short-term operational simulations of the Mediterranean Forecasting System.
Keywords: Eddies, jets, currents, Levantine basin, Mediterranean.
INTRODUCTION
In the Levantine Basin (area east of the Cretan Passage) of Mediterranean Sea (Fig.1), the only wide continental shelf areas to speak of are those adjoining the Nile river and the Mersin and İskenderun Bays of the Cilician Basin (area between Cyprus and Turkey) in the northeast. Perennial rivers Göksu, Lamas, Tarsus, Seyhan, Ceyhan and Asi, in this region plus some smaller rivers account for a total fresh water flux of 27km3/yr, much greater than the present discharge of the Nile (presently estimated to be about 16 km3/yr [1]), constituting one of the rare regions of freshwater influence (ROFI) supporting active shelf ecosystems in contrast with the oligotrophic deep sea. A sea-breeze system is superposed on westerly winds in summer. Storms steered by steep mountain ranges but intercepted by valleys often lead to local gale force winds during the rainy and mild winters [2].
Figure 1. Major current systems of the Mediterranean Sea [4].
OBSERVATIONS
Contrary to the mean cyclonic circulation proposed earlier in the Levantine Sea, the combined data of the international program Physical Oceanography of the Eastern Mediterranean (POEM) in the 1985-95 period showed for the first time a Mid-Mediterranean Jet that bifurcated several times, forming the Asia-Minor current following the Turkish coast, the complex anticyclonic circulations of the Mersa Matruh and Shikmona gyres to its south, interspersed with cyclonic and quasi-permanent anticyclonic long-lived eddies circulating around the basin with the mean currents and transporting water masses such as AW and LIW in deep eddy structures (Fig. 2) [4-7].
Figure 2. (a) The surface circulation in the first POEM experiment 1986, (b) a west-east cross-section approximately along the 36.5°N latitude in June 1987 showing water masses transported by eddies [4,7].
Modeling AND FORECASTING
Following the demands of an earlier generation models of the circulation dynamics of the Mediterranean Sea and its sub-basins, integration on European / international scale was achieved by the Mediterranean Forecasting System (MFS) providing data at open boundaries and common hydro-climatic forcing for the whole basin with nested models of many sub-domains. The Cilician Basin (CIL) and Northern Levantine (NLEV) operational forecast models have been developed, initially based on the Princeton Ocean Model (POM) [8] and recently on the Regional Ocean Modeling System (ROMS), published at http://linux-server.ims.metu.edu.tr/. The NLEV forecast model domain currently covers the Turkish Mediterranean coast (35.12-36.93ºN, 28.15-36.25ºE) with fine scale horizontal grid resolution of Dx = Dy = 1.35 km and 30 vertical s-levels, nested in the MFS Aegean Levantine Regional Model (ALERMO) and the Athens University SKIRON system providing atmospheric fields to determine interactive surface fluxes of mass, heat and momentum. The fine scale model bathymetry was generated from UNESCO bathymetric data, smoothed with a selective filter to make r=ΔH/(2H)<0.2 (where H is the depth) between adjacent grids and 10 rows of grids at open boundaries imitating the bathymetry of the coarse grid. The Generic Length-Scale (GLS) turbulence scheme is used in the model, with MPDATA advection for scalars, third-order upstream horizontal and fourth-order vertical advection of momentum, and selectable Smagorinsky lateral diffusivity and viscosity on geopotential surfaces. The 2d Flather (momentum) and Chapman (free surface) boundary conditions with 2d and 3d radiation and 3d nudging of momentum, temperature, salinity are applied at the open boundaries. The external and internal integration time steps typically are Δte=3 s and Δti=60 s respectively, and other model constants are: horizontal mixing coefficients Am= Ah=200 m2/s, background vertical mixing coefficients Km=Kh=Ks=2x10-4 m2/s respectively for momentum, heat, salt and bottom roughness parameters z0=0.01, Cb,min=0.0025.
In continuing work, coupled simulations with a nitrogen based ecosystem model [9] have been added in the ROMS system based on the original NLEV model configuration, including the riverine water and nutrient fluxes to the coastal ocean (Fig.3). The coupled model has been developed with nesting in the OPABFM Mediterranean coupled model [10] and has been used to test climate scenarios comparing two 5-year slices 1996-2001 and 2030-2035.
A coastal observation network concurrent with marine and atmospheric modeling (http://linux-server.ims.metu.edu.tr/) supports the system for calibration / validation as illustrated in Fig. 4 (further details can be found at http://gnoo.bo.ingv.it/myocean/calval/).
One of the successful predictions of the forecast model is the dense water formation on the shallow continental shelf are adjacent to İskenderun Bay in winter, verified by observations (Fig. 5).
Figure 3. (a) Currents and temperature at 10 m depth in the POM version of the NLEV model [6], sea surface height in (b) March and (c) November 2009 based on the ROMS version of the NLEV model, (d) sea surface salinity and (e) surface phytoplankton concentration (bloom conditions) simulated in April 1998, based on the NLEV model.
Figure 4. Comparison of observed daily mean sea level data at Taşucu (36°16'53"N, 33°50'09"E) on the Turkish coast with the MFS and NLEV model results at the same location.
Figure 5. Dense water formation in the Gulf of İskenderun based on observations and the NLEV model.
Acknowledgments
This work owes much to observational studies carried out over the last 30 years, including those under the POEM research programme, and modeling studies under the research project MOMA (105G029) sponsored by the Turkish Scientific and Technical Research Council (TÜBİTAK), the European projects MFSTEP (FP5), ECOOP, IASON, SESAME, MyOcean (FP6) and MyOcean2 (FP7) carried out at the IMS-METU.
References
[1] Reiter, E. R. (1979). Handbook for Forecasters in the Mediterranean; Weather Phenomena of the Mediterranean Basin; Part 1: General Description of the Meteorological Processes, Tech. Pap. 5-75, 344 pp., Environmental Prediction Research Facility, Naval Postgraduate School, Monterey, California
[2] Özsoy, E. (1981). On the Atmospheric Factors Affecting the Levantine Sea, European Center for Medium Range Weather Forecasts, Reading, U.K., Technical Report No. 25,30p.
[3] POEM Group (1992), General circulation of the Eastern Mediterranean, Earth Sci. Rev. 32, 285–309.
[4] Özsoy, E., Hecht, A. and Ü. cnlüata, (1989). Circulation and Hydrography of the Levantine Basin, Results of POEM Coordinated Experiments 1985/1986, Prog. Oceanogr., 22, 125-170.
[5] Özsoy, E., Hecht, A., Ünlüata, Ü., Brenner, S., Oğuz, T., Bishop, J., Latif, M. A., and Z. Rosentroub (1991). A Review of the Levantine Basin Circulation and Its Variability During 1985-88, Dyn. Atmos. Oceans, 15, 421-456.
[6] Özsoy, E. and H. Güngör (1993). The Northern Levantine Sea Circulation Based on Combined Analysis of CTD and ADCP Data, In: P. Brasseur (editor), Data Assimilation: Tools for Modelling the Ocean in a Global Change Perspective, NATO ASI Seeries, Springer-Verlag, Berlin.
[7] Özsoy, E., A. Hecht, Ü. Ünlüata, S. Brenner, H. İ. Sur, J. Bishop, M. A. Latif, Z. Rozentraub and T. Oğuz (1993). A synthesis of the Levantine Basin circulation and hydrography, 1985-1990. Deep-Sea Research, 40, No. 6, pp. 1075-1119.
[8] Özsoy, E. and A. Sözer (2006). Forecasting circulation in the Cilician Basin of the Levantine Sea, Ocean Sci. Discuss., 3, 1481-1514.
[9] Fennel, K., J. Wilkin, J. Levin, J. Moisan, J. O’Reilly, and D. Haidvogel (2006). Nitrogen cycling in the Middle Atlantic Bight: Results from a three-dimensional model and implications for the North Atlantic nitrogen budget, Global Biogeochem. Cycles, 20, GB3007, doi:10.1029/2005GB002456.
[10] Lazzari, P., A. Teruzzi, S. Salon, S. Campagna, C. Calonaci, S. Coltella, M. Tonani, and A. Crise (2010). Pre-operational short-term forecast for Mediterranean Sea biogeochemistry, Ocean Sci., 6, 25-39.