**1. Introduction**

The Turkish Strait System (TSS) consisting of the Marmara Sea, the Bosphorus, and the Dardanelles Strait is the only connection between two fundamentally different marginal seas: the Black Sea and the Mediterranean Sea. While the Black Sea has the dynamics of a typical estuarine circulation i.e., precipitation plus river input surpass evaporation (*E* − *P* − *R* < 0) [1,2], the Mediterranean Sea is an example of an inverse–estuary (*E* − *P* − *R* > 0) [3,4]. The TSS has unique dynamics since the flow is both density-driven between the salty (≈38 psu) Mediterranean Sea and the brackish (≈17 psu) Black Sea, and also barotropic-driven because of the permanent sea level difference between these two marginal seas. The sea level height in the Black Sea is approximately 0.3 m higher than the Marmara Sea due to its lower density. Climatological northeasterly winds weakly influence the general circulation in the Sea of Marmara [5].

The Marmara Sea is a relatively small inland sea which covers an area of 11,500 km<sup>2</sup> with three deep basins (>1000 m) and an extended shelf in the south. The Bosphorus and the Dardanelles Straits share common physical and geographical properties. They are both relatively narrow (Bosphorus 0.7–3.5 km, the Dardanelles 1.2–7 km), long (31 km for the Bosphorus and 61 km for the Dardanelles), and shallow (30–100 m for the Bosphorus and 50–120 m for the Dardanelles) channels [6–8]. Both Straits enable a water exchange between two different marginal seas with significant density differences.

**Citation:** Ilicak, M.; Federico, I.; Barletta, I.; Mutlu, S.; Karan, H.; Ciliberti, S.A.; Clementi, E.; Coppini, G.; Pinardi, N. Modeling of the Turkish Strait System Using a High Resolution Unstructured Grid Ocean Circulation Model. *J. Mar. Sci. Eng.* **2021**, *9*, 769. https://doi.org/ 10.3390/jmse9070769

Academic Editor: Matthew Lewis

Received: 16 June 2021 Accepted: 7 July 2021 Published: 12 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

The fresh water inflow from the Marmara Sea into the Aegean is crucial for the northern part of the Aegean Sea. Convection in the northern part of the Aegean is regulated by the brackish waters of the Black Sea coming from the Dardanelles. In the Black Sea, the river input is balanced by the salty and warm flow coming from the Mediterranean Basin through the Bosphorus Strait. This salty dense water (with the Cold Intermediate Water) is also important for the ventilation process of the deep, oxygen depleted Black Sea waters. Salty Mediterranean Sea waters are equilibrated at the depth of the suboxic and anoxic layers of the Black Sea and play an important role in the redox potential of the chemistry of the Black Sea [8].

Most ocean circulation models use structured grids with finite difference/volume discretizations. These types of grids are particularly challenging in modeling the TSS system because of the narrow and long Bosphorus and Dardanelles Straits, and the deep basins of the Marmara Sea. The complex topography and two-layer density structure require multiple hydraulic controls in both Straits [9]. These narrow Straits thus require a high resolution (less than 100 m) to resolve their dynamics. In addition, the TSS salinity and heat balances are controlled by the Black Sea and the Mediterranean Sea. Hence, the entire TSS region needs to be part of the modeling effort.

Some studies have modeled the transports of individual straits using either 2D idealized reduced gravity or non-hydrostatic models or 3D regional idealized models [10–14]. Some of these studies have attempted to model the Marmara Sea with or without the Turkish Straits using structured grid ocean models. Demyshev et al. conducted a finite difference numerical simulation without atmospheric forcing and reproduced the S-shaped circulation of the jet current exiting the Bosphorus and crossing the Marmara Sea with a basin scale anti-cyclonic circulation [15]. However, the model does not contain strait dynamics. Similarly, the Regional Ocean Modeling System (ROMS [16]) is used to model the Marmara Sea using realistic atmospheric forcing but with open boundaries at the Straits nudged to observation fields [17]. ROMS is a terrain following a structured grid model which needs bathymetry smoothing due to pressure gradient error. Without including the Straits in the model, the authors assumed that the surface circulation depends solely on the strength and directional pattern of the wind force in the Sea of Marmara. Sannino et al. used the structured grid MITgcm model [18] with curvilinear coordinates for a high resolution around the Bosphorus Strait [19]. Their model also did not use any atmospheric forcing. They investigated the circulation of the TSS changing the barotropic flow between the Black Sea and the Aegean Sea.

Few studies have used unstructured grid models to simulate the three-dimensional baroclinic circulation of the TSS. Stanev et al. used the SCHISM model for interconnected basins including the Black Sea, the TSS, and the northern Aegean Sea [20]. They used realistic atmospheric forcing with lateral open boundaries in the south. Their aim was to accurately represent the transport at the straits and the resulting circulation dynamics of the Black Sea. An implicit advection scheme was used in SCHISM for larger time steps. This leads to a coarser model resolution, with only 53 vertical levels at the deepest point of the Black Sea.

Similarly, the Finite Element Sea-Ice Ocean Model (FESOM), another unstructured grid model, was used by Aydogdu et al. to study the circulation of the TSS [21]. They analyzed the combined response of the Sea of Marmara with atmospheric forcing and strait dynamics. Although the FESOM model has a high resolution of up to 65 m in the horizontal and 110 vertical levels, the setup has a closed boundary, and a volume correction is needed to maintain the sea level difference between the Black Sea and the Aegean Sea sides.

Our aim in this study is to simulate the entire TSS region using an unstructured grid model with high resolution in both horizontal and vertical directions. The model has been forced by realistic atmospheric reanalysis and open boundaries with ocean analysis data sets. The main goal is to develop a regional model with an adequate representation of the mean and variability of the TSS circulation. We plan to use the output from the new model as lateral boundary conditions for future Copernicus Marine Service (CMEMS) Mediterranean Sea and Black Sea models.

This paper is organized as follows: the numerical model and details of the experiment are introduced in Section 2. The main results including mean circulation properties, validation of water mass structure, and volume fluxes across the straits are presented in Section 3. Finally, we summarize and conclude in Section 4.
