**Contents**


## **About the Special Issue Editors**

**Miroslav Gačić** is a physical oceanographer at the National Institute of Oceanography and Applied Geophysics, Trieste (Italy), and was, until 2015, Head of the Oceanography Group for about 20 years. His main research interests and activities have been focused on the Adriatic and Mediterranean Seas, and have been related to: a) deep-water formation processes and their influence on the nutrient input into the euphotic zone, b) thermohaline circulation inside the basin and its interannual variability, c) the response of the basin to atmospheric forcing on various time scales, d) coastal circulation and response to local forcing and remote forcing and e) water exchange between semi-enclosed bays and seas and adjacent basins. Within coastal studies, he has been working on the analysis and interpretation of surface current measurements in the Northern Adriatic, in the Venice lagoon inlets. Special attention has been paid to understanding the water exchange dynamics between the lagoon and the open sea, and to quantify the relative contribution of tides and meteorological forcing. Within open-sea studies, he has worked on the winter convection and water exchange through the Strait of Otranto, and longitudinal water fluxes in the Adriatic Sea. An important aspect of his activity has been oriented towards the interannual and decadal variability of the Adriatic thermohaline properties and interaction with the Ionian Sea. One of the most important scientific results he has achieved is his contribution to the understanding of the decadal variability of the Ionian circulation and the feedback with the Adriatic Sea (Adriatic–Ionian bi-modal oscillating system—BiOS). He is the author of about 150 scientific papers in international journals and chapters in books.

**Manuel Bensi**, Ph.D., a researcher at the National Institute of Oceanography and Applied Geophysics, Trieste (Italy), is a marine scientist in the field of physical oceanography, with extensive experience in ocean data acquisition and analysis. His scientific activity focuses on thermohaline deep sea and large-scale circulation processes, dense water formation, and climate change in the Mediterranean Sea and Polar regions. He is the author and co-author of more than 80 scientific publications, including articles in journals, book chapters, and conference proceedings. Since 2003, he has participated in more than 25 oceanographic cruises (in the Mediterranean Sea, Atlantic and Arctic Ocean). He is the coordinator and task leader of several national and international projects. He is also a supervisor of doctoral students, and he constantly participates in scientific dissemination activities. He serves as a reviewer and guest editor for several ISI journals. He speaks three languages (Italian, English, Portuguese).

### *Editorial* **Ocean Exchange and Circulation**

#### **Miroslav Gaˇci´c and Manuel Bensi \***

National Institute of Oceanography and Experimental Geophysics, Borgo Grotta Gigante 42-C, I-34010 Trieste, Italy; mgacic@inogs.it

**\*** Correspondence: mbensi@inogs.it

Received: 12 March 2020; Accepted: 18 March 2020; Published: 20 March 2020

**Abstract:** The great spatial and temporal variability, which characterizes the marine environment, requires a huge effort to be observed and studied properly since changes in circulation and mixing processes directly influence the variability of the physical and biogeochemical properties. A multi-platform approach and a collaborative effort, in addition to optimizing both data collection and quality, is needed to bring the scientific community to more efficient monitoring and predicting of the world ocean processes. This Special Issue consists of nine original scientific articles that address oceanic circulation and water mass exchange. Most of them deal with mean circulation, basin and sub-basin-scale flows, mesoscale eddies, and internal processes (e.g., mixing and internal waves) that contribute to the redistribution of oceanic properties and energy within the ocean. One paper deals with numerical modelling application finalized to evaluate the capacity of coastal vegetated areas to mitigate the impact of a tsunami. The study areas in which these topics are developed include both oceanic areas and semi-enclosed seas such as the Mediterranean Sea, the Norwegian Sea and the Fram Strait, the South China Sea, and the Northwest Pacific. Scientific findings presented in this Special Issue highlight how a combination of various modern observation techniques can improve our understanding of the complex physical and biogeochemical processes in the ocean.

**Keywords:** mesoscale eddies; deep-sea thermohaline variability; dense-water formation; picoplankton distribution; glider; floats; high-frequency radar; moorings; tsunami; Kuroshio bifurcation; internal waves

#### **1. Introduction**

The oceanic circulation shows great spatial and temporal variability that requires a huge effort to be observed and studied properly [1]. Satellites, autonomous platforms, fixed point observatories located in key regions of the world ocean, and repeated hydrographic surveys are a means to collect new data. However, when the spatial and temporal resolution is not adequately resolved, numerical models can complement or compensate for the lack of information needed for the observation of the most complex oceanic dynamics and processes.

This Special Issue collects a series of scientific articles (Table 1) that address oceanic circulation and water mass exchange. The study areas in which these topics are developed are the Mediterranean Sea, the Norwegian Sea and the Fram Strait, the South China Sea, and the Northwest Pacific. The papers deal with the oceanic circulation generated by wind and/or density gradients, which contributes to the exchange of water between different parts of the ocean or between marginal seas and adjacent ocean areas. These processes develop at different spatial and temporal scales. Other than mean circulation, basin and sub-basin-scale flows, mesoscale eddies and internal processes (e.g., mixing induced by bottom roughness, internal waves) also contribute to the redistribution of oceanic properties and energy within the ocean, but they are not always correctly taken into consideration. Thermohaline ocean circulation is driven by winter convection and dense water formation processes, which are directly influenced by weather conditions. Subsequently, long-term and climatic changes in circulation and vertical mixing processes directly influence the variability of the ocean biogeochemical properties. A special role in trapping and/or transporting the biogeochemical properties of seawater is played by travelling eddies; however, this is yet to be quantified. The scientific papers included in this Special Issue try to provide new evidence related to the processes associated with the circulation and mixing in both oceanic areas and semi-enclosed seas.



#### **2. Oceanic Circulation and Mesoscale Features: Experimental and Modelling Approaches**

One paper in this Special Issue represents an example of numerical modelling application, displayed by Zhang et al. [2], where the authors study the interaction of the ocean with the coastal area: in fact, coastal areas are the most densely populated and urbanized, but they are also complex and delicate environments. In coastal areas, waves have a strong impact that can be devastating and produce great damage as well as the loss of human lives. In this case, prevention, through the application of numerical models for case studies, e.g., the impact of tsunami waves, is very important. A depth-integrated numerical model was established to simulate wave propagation in a coastal region

with and without forest cover, which could be considered buffer zones able to mitigate the impacts from tsunami waves over coastal areas. Overall, these findings provide cost-effective natural strategies to improve the effectiveness of vegetated bioshields against tsunami hazards.

Other papers in this Special Issue reflect the large interest and importance of mesoscale eddies in a number of semi-enclosed seas of the world ocean. A detailed census of mesoscale eddies and their generation and migration within the South China Sea is presented by Wang et al. [3]. A combined use of satellite measurement data and Hybrid Coordinate Ocean Model (HYCOM) reanalysis data, together with the self-organizing map (SOM) method, was used to investigate the east branch of the Kuroshio bifurcation where four coherent patterns associated with mesoscale eddies in the Pacific Ocean were found [4]. On the other hand, generation mechanisms like wind curl and dense water flow of the mesoscale eddies in the Ionian Sea (central Mediterranean) were addressed and some comparisons between the two forcing mechanisms were discussed in the paper by Menna et al. [5]. The Sicily Channel, in between the eastern and western Mediterranean basins, is another area characterized by large mesoscale variability and water mass exchanges that was analyzed in the paper by Reyez–Suarez et al. [6]. In this paper, data obtained by the High Frequency (HF) radar enabled authors to identify semi-permanent gyres, which were clearly evident from altimetry data also. By using satellite chlorophyll data, authors showed the signal present in the phytoplankton coinciding with altimetry or HF circulation data. Geostrophic flow obtained from the altimetry data fits very well with the HF radar data averaged over a few weeks. The paper presented by Menna et al. [5] deals with such variability more or less in the same area using a combination of long-term, time-series altimetry data, Sea Surface Temperature (SST), and remotely sensed wind data in order to study the seasonal and multiannual variability of mesoscale structures. Wind data allowed the authors to discuss in detail the generation mechanism of mesoscale eddies and the importance of wind stress and wind stress vorticity fields on their seasonal variability.

The role of mesoscale eddies in enhancing the primary production both at the open ocean [7,8] and in semi-enclosed seas (e.g., [9,10]) has been extensively addressed in oceanographic literature ever since the mid-1990s. There is also an important interaction between mesoscale eddies and plankton spatial patterns, and thus the biological features in the ocean are, to a large degree, conditioned by these circulation features [11]. In this issue, picoplankton biomass, distribution, and activity were addressed in the paper by Šanti´c et al. [12], where the authors used data collected during two Eurofleets2 cruises gathered in the Adriatic Sea (eastern Mediterranean) to compare pre- and post-winter convection phases. The Adriatic Sea is, in fact, one of the main dense water formation sites in the Mediterranean Sea [13], where long-term variability and climatic shifts can produce important effects on the overall circulation of this marginal sea, with possible repercussions on the global thermohaline circulation also [14]. Indeed, despite being a marginal sea, the large-scale dynamical effects linked to the interaction between the Mediterranean water outflow from the Strait of Gibraltar and the Atlantic Ocean has always been the subject of study [15]. The so-called "Meddies", for example, are lenses of warm and salty Mediterranean Water that travel through the North Atlantic at depths between ∼500 and 1500 m. They are typically 40–100 km in diameter with their core being 500–1000 m thick [16]. Signatures of the timing of Mediterranean outflow water activity are recorded in the depositional sediment, revealing how the addition of the warm saline water to the cooler less-salty waters of the Atlantic was related to climate change, deep ocean circulation, and plate tectonics [17].

Furthermore, the deep flow variability and the dynamics related to internal waves are analyzed in the paper published by Bensi et al. [18], where the authors take into account the bottom current flows and thermohaline variability in the eastern part of the Fram Strait, which is a crossroad of waters between the North Atlantic and the Arctic Ocean. Starting from a detailed analysis of time series data collected through deep-sea oceanographic moorings, the authors discuss the ocean–atmosphere interactions in a region where they are particularly intense, and where they lead to multiple oceanographic processes, like shelf-slope dynamics, deep water variability through the mixing of Polar and Atlantic waters, as well as sea ice and dense water formation. A strong eddy activity was observed at 1000 m depth

in the Fram Strait, associated with the passage of topographically trapped waves enhanced by the atmospheric events particularly intense in the winter season.

A multiplatform approach is an efficient experimental tool to study oceanic features and their evolution in time and space, and this is illustrated successfully in a series of papers in this issue. Based on the analyses of drifters, floats, altimetry, and glider data, the paper by Mauri et al. [19] focuses on quasi-permanent features, i.e., meandering structures and eddies/gyres, in the easternmost part of the Mediterranean Sea [19], which is also a region where the Levantine intermediate water originates from and starts its travel westwards, contributing to the generation of the necessary pre-conditions for the dense water formation in the Aegean, Adriatic, and West Mediterranean basins [20]. In particular, by the visual inspection of more than 800 vertical profiles of temperature, salinity, and potential density gathered from ARGO floats, Kubin et al. [20] detected the timing and evolution, of the Levantine intermediate and deep waters, revealing that, in some conditions, the algorithm used for the detection of the mixed layer depth can underestimate it during winter convection events.

#### **3. Conclusions**

This Special Issue contributes to highlight and discuss topics related to the oceanography of semi-enclosed seas and the open ocean. An important number of articles address the theme of mesoscale eddies and their generation/variability by external or internal forcing. In this Special Issue, the effects of mesoscale circulation on biological processes are also covered. The approach to these studies used a broad spectrum of modern measurement techniques that allow for a high temporal and spatial resolution; in this context, fixed-point observations, remote sensing techniques, data from gliders, floats, and drifters are used to a large extent. The characteristics of the circulation on larger spatial scales were also addressed and some important aspects, such as the thermohaline variability of deep layers or the formation of dense water, have been discussed. Future works should be more oriented towards an increasingly enhanced combination of numerical modelling data assimilation techniques and experimental data. Particularly promising results could be obtained from the assimilation of high-frequency radar data, which can also be used very efficiently in submesoscale vortex studies. A multi-platform approach, in addition to optimizing data collection—in terms of quality, also—will bring the scientific community to more efficient monitoring and better predictions of the world ocean processes. This must be a collaborative effort, useful to optimize and integrate ocean observing systems, sensor deployment, and usage [21]. Only with this approach can the scientific community take a step forward in the understanding of climate variability and the anthropogenic effect.

**Author Contributions:** the two authors contributed equally to the preparation of the article. Conceptualization, M.G. and M.B.; writing—original draft preparation, M.G. and M.B.; writing—review and editing, M.G. and M.B; All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** Thanks are due to the editors of the journal, as well as to the authors who contributed with their articles to the Special Issue. Finally, special thanks go to the anonymous reviewers, who have contributed efficiently to improve the quality of the articles.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 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 (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Levantine Intermediate and Levantine Deep Water Formation: An Argo Float Study from 2001 to 2017**

**Elisabeth Kubin 1,\*, Pierre-Marie Poulain 1,2, Elena Mauri 1, Milena Menna <sup>1</sup> and Giulio Notarstefano <sup>1</sup>**


Received: 12 July 2019; Accepted: 23 August 2019; Published: 27 August 2019

**Abstract:** Levantine intermediate water (LIW) is formed in the Levantine Sea (Eastern Mediterranean) and spreads throughout the Mediterranean at intermediate depths, following the general circulation. The LIW, characterized by high salinity and relatively high temperatures, is one of the main contributors of the Mediterranean Overturning Circulation and influences the mechanisms of deep water formation in the Western and Eastern Mediterranean sub-basins. In this study, the LIW and Levantine deep water (LDW) formation processes are investigated using Argo float data from 2001 to 2017 in the Northwestern Levantine Sea (NWLS), the larger area around Rhodes Gyre (RG). To find pronounced events of LIW and LDW formation, more than 800 Argo profiles were analyzed visually. Events of LIW and LDW formation captured by the Argo float data are compared to buoyancy, heat and freshwater fluxes, sea surface height (SSH), and sea surface temperature (SST). All pronounced events (with a mixed layer depth (MLD) deeper than 250 m) of dense water formation were characterized by low surface temperatures and strongly negative SSH. The formation of intermediate water with typical LIW characteristics (potential temperature > 15 ◦C, salinity > 39 psu) occurred mainly along the Northern coastline, while LDW formation (13.7 ◦C < potential temperature < 14.5 ◦C, 38.8 psu < salinity < 38.9 psu) occurred during strong convection events within temporary and strongly depressed mesoscale eddies in the center of RG. This study reveals and confirms the important contribution of boundary currents in ventilating the interior ocean and therefore underlines the need to rethink the drivers and contributors of the thermohaline circulation of the Mediterranean Sea.

**Keywords:** Levantine intermediate water formation; Levantine deep water formation; Rhodes Gyre; boundary currents; heat fluxes; Argo floats; dense water formation

#### **1. Introduction**

The Mediterranean Sea (Figure 1) is composed of two basins of nearly equal size, the Western and the Eastern Mediterranean Sea, connected by the Sicily Channel. The general circulation of the Mediterranean Sea can be divided into three dominant scales of motion: the basin scale including the thermohaline circulation, the sub-basin scale including permanent and quasipermanent cyclonic and anticyclonic gyres, and the mesoscale with small but energetic temporary eddies [1,2]. All these scales are interacting.

Through the Strait of Gibraltar, the relatively fresh Atlantic water (AW) enters the Western Mediterranean Sea within the upper 100 to 200 m. It is modified flowing eastward, passes the Sicily Channel and the Ionian Sea and enters the easternmost part of the Mediterranean, the Levantine Sea. The salinity of the AW in the Levantine Sea depends on the circulation patterns during its path, mainly influenced by the variability of the circulation of the North Ionian Gyre (NIG) which varies significantly at seasonal and decadal scales ([3–5]; Figure 1a).

Due to evaporation and air–sea exchanges, the AW is becoming saltier and warmer when reaching the Levantine Sea. The AW that enters the Levantine Sea is identified by a subsurface minimum of S < 38.6 psu and by a temperature of 14–15 ◦C. The Levantine intermediate water's (LIW) properties are defined by salinity values greater than 39 psu, potential temperature values greater than 15 ◦C and potential density values between 29 and 29.05 kgm−<sup>3</sup> while typical ranges for Levantine deep water (LDW) are 13.7–14.5 ◦C and 38.8–38.9 psu [6].

The strong advective surface salinity preconditioning [7] and buoyancy losses due to heat and freshwater fluxes in winter lead to dense water formation, sinking takes place, and the LIW is formed (Figure 1a). The LIW is characterized by a subsurface salinity maximum and occupies and moves in the intermediate layers between 200 and 600 m throughout the Mediterranean Sea until it reaches the Atlantic Ocean through the Strait of Gibraltar. Therefore, the LIW contributes as an important driver to the thermohaline circulation of the Mediterranean Sea. The specific pathways of the thermohaline circulation depend strongly on where and when the LIW is formed.

According to the prevailing view, the LIW formation takes place within the cyclonic Rhodes Gyre (RG) during the winter months ([1,8], Figure 1a). However, experimental studies showed that the RG is also a place of LDW formation and that the Levantine basin is a site of multiple and different water mass formation processes [6,9,10]. Furthermore, recent theoretical models revealed that no net mean sinking takes place within Mediterranean convection sites such as the RG, while boundary currents undergo net intense sinking ([11–16]; mean from 1980 to 2013 for all seasons; Figure 1c). This is due to vorticity dynamics: Only dissipation at the boundary (and bottom friction) can balance the vortex stretching that arises from vertical motions induced by net sinking.

The focus of this study is on the Northwestern Levantine Sea (NWLS; latitude: 33◦–37◦, longitude: 26◦–32◦; red rectangle in Figure 1b), the larger area around the RG, including also the area along the Northern coast. The NWLS is characterized by a general cyclonic surface circulation and the presence of the permanent RG, cyclonic and anticyclonic structures (such as Ierapetra (IG) and Mersa-Matruh Eddy (MME)), intense jets (such as the Mid Mediterranean Jet (MMJ); with minimum subsurface salinity values), a strong coastal current along the Northern coastline (the Asia Minor Current (AMC)), and the passage from the Cretan to the Levantine Sea ([17], Figure 1b).

This study aims to describe where and when LIW and LDW formation take place within the NWLS, using Lagrangian Argo float data over a period of 16 years. The results were compared to heat and freshwater fluxes and satellite data (SSH, SST).

Dense water formation processes occur on daily scales and are linked to temperature and salinity extrema which are not represented using climatological data sets or mixed layer depth (MLD) detection algorithms. Therefore, more than 800 Argo float profiles from 2001 to 2017 were analyzed visually to investigate LIW and LDW formation during winter months.

The paper is organized as follows: In Section 2, the data sets and analysis methods are presented. In Section 3, the time evolution of the heat and freshwater fluxes from 2001 to 2017 for the center of the RG are shown to identify periods of extreme heat losses due to outbreaks of strong cold and dry winds leading to dense water formation. Section 3 gives two examples of the LDW formation within the RG and one example of the typical LIW formation along the Northern coastline and characterizes the newly formed water masses. Section 4 gives an overall discussion and summarizes the major results of this work.

**Figure 1.** (**a**) General concept of the thermohaline circulation which, according to the prevailing view, is driven by a few convection sites within the Mediterranean Sea, adapted from [1]. The red arrow shows the entering Atlantic water (AW) while the orange arrow shows the Levantine intermediate water (LIW), which travels throughout the Mediterranean and flows into the Atlantic Ocean. The orange and yellow frames highlight the area of the Levantine Sea and of the Northwestern Levantine Sea (NWLS), respectively. The blue circle indicates the position of the North Ionian Gyre (NIG) the circulation of which is important for the advective salinity preconditioning. (**b**) Mean surface geostrophic circulation in the Levantine Sea from 1992 to 2010 derived from drifter data. The yellow rectangle indicates the area of study, the NWLS (latitude: 33–37◦, longitude: 26–32◦). Adapted from [17]. AMC—Asia Minor Current; CC—Cilician Current; CE—Cyprus Eddy; EE—Egyptian Eddies; IE—Ierapetra Eddy; LEC—Libyo-Egyptian Current; LTE—Latakia Eddy; MME—Mersa-Matruh Eddy; MMJ—Mid Mediterranean Jet; ShE—Shikmona Eddy. (**c**) A model run from 1980 to 2013 [15] showed that little to no net sinking takes place at convection sites (blue arrows; from left to right: Gulf of Lion, South Adriatic, Aegean Sea, Rhodes Gyre (RG)) while boundary layer currents undergo net intense sinking (brown arrows). The yellow rectangle indicates the area of study. Adapted from [15].

#### **2. Datasets and Methods**

The datasets used for this study are Argo floats vertical profiles of temperature and salinity (T/S) collected in the NWLS during winter months (January, February, March—JFM) between 2001 and 2017. In total 879 T/S profiles from 20 floats were analyzed visually. Figure 2 shows the position of the Argo float profiles and the annual distribution of profiles for JFM between 2001 and 2017 for the area of study.

**Figure 2.** (**a**) Position of the 879 analyzed float profiles for January, February, and March (JFM) 2001–2017; Orange rectangle defines the area of study (the NWLS), the black ellipse describes the center of RG. (**b**) Annual distribution of float profiles in the NWLS: JFM 2001–2017.

With the help of an external bladder, the Lagrangian floats descend to a programmed parking depth (350 or 1000 m) where they stay for a specified period (1–10 days, mainly 5 days [18]). Then, they descend to greater depths (up to 2000 m). During their ascent back to the surface, they measure temperature and salinity throughout the water column. At the surface they transmit the data to satellites and descend again. The transmitted data are stored at data assembly centers (DAC) which apply a quality control and provide open access to real time and delayed mode quality controlled data.

Quality controlled Argo float data were downloaded from the Ifremer Data Assembly Center (DAC; ftp://ftp.ifremer.fr/ifremer/argo/dac/coriolis/). For this study, only data with the best quality control (qc=1) were taken into account. Downloaded parameters included float number, position, time, pressure, temperature, and salinity.

Hydrographic properties were expressed as potential temperature, potential density, and salinity according to the practical salinity scale (PSU).

The visual inspection of the Argo float profiles is important due to the fact that the Argo floats may pass an area not exactly during the event of mixing or convection. They can instead sample days or weeks later when the recapping (i.e., a newly formed shallow MLD) already occurred. In such a case, MLD detection algorithms indicate a shallow MLD, but do not give any information about mixing or convection events before the recapping. While at the top, there can be already a newly formed MLD and the convection event can still be visible deeper in the water column. MLD detection statistics rarely give information about deep mixing events while the visual inspection of the form of the profile (potential temperature, salinity, and potential density) reveals clearly such events. Figure 3a shows the climatology of the winter maximum MLD derived from Argo float data and downloaded from [19] in the period 2000 to 2018. The maximum MLD is 225 m along the coastline of the NWLS (Figure 3a). The visual inspection of the Argo profiles in the same region reveals deeper dense water formation events. For example, in winter 2007, the float WMO 6900098 (Figure 3b), moving along the northern coastline of the NWLS, shows the deepening of the winter MLD from 100 m (Figure 3c) to 200 m (Figure 3d) in January. In March, when the maximum MLD is about 550 m, the recapping occurred (due to surface warming) with a newly formed MLD of about 50 m (Figure 3e). In this case, the MLD detention algorithm can fail, indicating the depth of 50 m as maximum MLD.

**Figure 3.** (**a**) Climatology of the winter maximum mixed layer depth (colors) and location of float profiles (grey dots) from 2000 to 2018. The white rectangle indicates the area of study, the NWLS. (**b**) Trajectory of the float WMO 6900098 during JFM 2007; the numbers along the trajectory show the locations of float profiles. Float potential density profiles at cycle 2 (**c**), 5 (**d**), and 13 (**e**).

The SST and sea surface height (SSH) data were downloaded from Copernicus (marine.copernicus.eu). The interpolated SST product (SST\_MED\_SST\_L4\_NRT\_OBSERVATIONS\_010\_004\_c\_V2) has a daily temporal resolution and a spatial resolution of 0.04◦ × 0.04◦. The interpolated SSH product

(SST\_MED\_SST\_L4\_REP\_OBSERVATIONS\_010\_021) has a daily temporal resolution and a spatial resolution of 0.125◦ × 0.125◦. Monthly means of SST superimposed with the geostrophic currents from SSH were used to describe the negative slope of eddies within the RG during intense convection events.

The freshwater fluxes were derived from ERA-INTERIM (daily) data. The downloaded parameters are evaporation (E) and total precipitation (P). The downloaded data have a time step of 12 hours, i.e., daily data at 00:00:00 and at 12:00:00 and a spatial resolution of 0.25◦ × 0.25◦. The daily freshwater fluxes (FWF) were calculated as the subtraction of the daily means of E and P: FWF=E−P.

The air–sea heat fluxes were derived from ERA-INTERIM (daily) data. The downloaded parameters are: Surface net solar radiation (Qsw), surface net thermal radiation (Qlw), surface sensible heat flux (Qs), and surface latent heat flux (Ql). The heat budget can be expressed as the difference between the net shortwave solar radiation (incoming minus reflected) absorbed by the sea surface, the sum of the longwave back radiation, the sensible, the latent, and the advective heat flux. The advective heat flux (Qadv) was not available at ERA-INTERIM and therefore not considered. The downloaded data have a time step of 12 hours, i.e., daily data at 00:00:00 and at 12:00:00 and a spatial resolution of 0.25◦ × 0.25◦. The daily mean of each parameter as well as the daily net heat fluxes were calculated as the sum of the daily means of each parameter: Qnet = Qsw + Qlw + Ql + Qs. The surface buoyancy flux B, composed by thermal (BT) and haline (BS) components, was calculated according to [20]:

$$\mathbf{B} = \mathbf{a} \times \mathbf{g} \times (\mathbf{C}\mathbf{p} \times \rho\_0)^{-1} \times \mathbf{Q} \mathbf{net} - \boldsymbol{\upbeta} \times \mathbf{S}\_0 \times \mathbf{g} \times (\rho\_0)^{-1} \times (\mathbf{E} - \mathbf{P})^2$$

where α is the thermal expansion coefficient, g = 9.8 ms−<sup>2</sup> is the gravity acceleration, Cp <sup>=</sup> 3.9715 <sup>×</sup> <sup>10</sup>−<sup>3</sup> Jkg<sup>−</sup>1K−<sup>1</sup> is the specific heat capacity of sea water, <sup>ρ</sup><sup>0</sup> = 1029 kgm−<sup>3</sup> is a reference sea water density, β is the haline contraction coefficient and S0 = 38.9 is a reference salinity. α and β were calculated at surface pressure, using monthly mean surface salinity and monthly mean surface temperature, downloaded from Copernicus (MEDSEA\_REANALYSIS\_PHYS\_006\_004). B is positive when surface water gets lighter and negative when surface water becomes denser (river inputs as well as horizontal and vertical advection also contribute to density changes, but were not considered due to lack of data).

The Turner angle (Tu) was computed to evaluate the relative roles of temperature and salinity gradients on the density gradients. Tu is defined as the four-quadrant arctangent [21], which units are degrees of rotation and was calculated with the Gibbs-SeaWater (GSW) Oceanographic Toolbox [22]. Argo float salinity and temperature were converted to absolute salinity and to conservative temperature, respectively. The conservative temperature represents more accurately the heat content [22].

Tu = 45◦ indicates that temperature is the only contributor, while Tu = −45◦ indicates that salinity is the only contributor to density changes; |Tu|< 45◦ indicates stable stratification and in this condition both temperature and salinity contribute to the density change ; 45◦ < Tu < 90◦ shows that salinity is working against temperature and is also called the 'salt finger regime' with the strongest activity near 90◦; −90◦ < Tu < −45◦ is called the 'diffusive regime' and shows that temperature is working against salinity, reaching the strongest activity near −90◦; |Tu|> 90◦ characterizes a statically unstable water column (where the Brunt–Vaisalaa frequency N2 < 0).

#### **3. Results**

#### *3.1. Heat and Freshwater Fluxes within the Northwestern Levantine Sea*

The intensity of the mixing and convection events depends mainly on the surface buoyancy fluxes B, which in turn depend on the heat fluxes through the air–sea interface, scaled by the thermal expansion coefficient α, and the freshwater fluxes, scaled by the haline contraction coefficient β. Monthly surface buoyancy fluxes and their thermal and haline (freshwater) components, integrated over the center of RG (longitude: 28–31◦E, latitude: 34–36◦N), are shown in Figure 4.


**Table 1.** Pronounced (mixed layer depth (MLD) >250 m) dense water formation events within the center of RG and along the Northern coastline: Area of formation, float WMO, time period, watermass characteristics, and maximum depth during the dense water formation events.

<sup>1</sup> T/S plots show the formation of dense water with a potential density that corresponds to the density of the upper deep boundary layer which is found at approximately 1000 m depth [23].

**Figure 4. Upper panel:** Time series of the monthly thermal component (BT) from 2001 to 2017, integrated over the center of RG (longitude: 28–31◦E, latitude: 34–36◦N). Blue and yellow circles indicate events of pronounced dense water formation detected by the Argo floats within RG and along the coastline, respectively (Table 1). **Lower panel:** Time series of the monthly haline component (BS; magenta line) and buoyancy fluxes (B; black dotted line).

The haline components (BS) dominate the surface buoyancy fluxes (Figure4, Lower Panel), i.e., that intense evaporation, especially during the preconditioning phase (e.g., Figure 8 for winter 2006), controls the surface buoyancy loss. Detected events of pronounced (i.e., with a MLD deeper than 250 m) dense water formation by the Argo float data, are indicated with blue (RG) and yellow (coastline) circles (Table 1).

The climatology of monthly heat fluxes Qnet for the center of RG from 2001 to 2017 shows that the largest heat losses which induce the preconditioning phase occurred mainly in November and in December (Figure 5). The subsequent heat losses in JFM induce the formation of dense water and therefore lead to convection and mixing.

**Figure 5.** Climatology of monthly integrated heat fluxes for the center of RG from 2001 to 2017. Main heat losses generally occur in November and December (turquoise bars) and initiate the preconditioning phase. Subsequent heat losses in January, February, and March induce dense water formation.

#### *3.2. LIW and LDW Formation within the Northwestern Levantine Sea*

Intermediate and deep water formation events in the NWLS (latitude: 33–37◦N, longitude: 26–32◦E) were analyzed during the winter months (JFM) from 2001 to 2017 (879 T/S profiles from 20 floats). Most of the float profiles within the NWLS showed 'regular' winter MLDs, i.e., MLDs between 100 and 200 m. Pronounced dense water formation, i.e., with a MLD deeper than 250 m, occurred only within the center of RG, along the Northern coastline and along the Cretan Arch passage. Events of pronounced LDW (13.7 ◦C < potential temperature < 14.5 ◦C, 38.8 psu < salinity < 38.9 psu) and 'lower range' LIW (potential temperature around 15 ◦C and salinity around 39 psu) formation were detected within the center of RG in winter 2004, 2005, 2006, and 2008 and events of pronounced LIW formation (potential temperature > 15 ◦C and salinity > 39 psu) were detected along the Northern coastline in winter 2007, 2012, 2015. and 2016 (Table 1). More than 800 profiles of 20 floats were analyzed, but only four floats (Tables 1 and 2) captured pronounced dense water formation, being at the right place at the right time. To document the dense water formation events the float had to be either inside or pass later through the area of dense water formation. Float WMO 6900098 had an exceptionally long lifetime of nearly 6 years and therefore it was able to capture one event of pronounced LIW formation along the Northern coastline and four events of pronounced DWF. Unfortunately, it stopped measuring at 600 m depth. The WMO numbers of the Argo floats that found pronounced dense water formation events within the center of RG and along the coastline are listed in Table 2.


**Table 2.** Argo floats capturing pronounced (MLD >250 m) dense water formation events.

#### 3.2.1. LDW formation within the Rhodes Gyre

Two examples of LDW formation within RG are given in this subsection.

(1) In JFM 2006, float WMO 6900098 was entrapped in the center of RG (Figure 6a). Hoevmueller plots of salinity, potential temperature, and potential density describe two pronounced events of mixing and convection during this winter (Figure 6b–d). The first event occurred by the end of January until mid-February and led to LDW formation (13.7 ◦C < potential temperature < 14.5 ◦C, 38.8 psu < salinity < 38.9 psu) while the second event around mid-March led to LDW and 'lower range' LIW (temperature about 15◦C and salinity about 39 psu) formation.

**Figure 6.** (**a**) Mean sea surface height (SSH, m) and float trajectory of float WMO 6900098 for JFM 2006. (**b**) Salinity (PSU), (**c**) potential temperature (◦C) and (**d**) potential density (kg/m3) from December 2005 to April 2006.

In December, very high surface salinity values (S > 39.15 psu) were detected in the upper 50 m (Figure 6b). However, mixing and convection is still prevented by relatively high surface temperatures during December. The surface temperature has a decreasing trend from 17.5 ◦C by the beginning of December to 15.5 ◦C in early January and reached a minimum of about 14 ◦C from the end of January to the end of February. From the beginning of March, the surface temperature gradually increased, reaching 15.5 ◦C during March with a successive increase to 17.5 ◦C by the end of April.

The MLD deepens from 50 m within December to about 100 m in the beginning of January and the high surface salinity is mixed to intermediate layers.

By the end of January, when lowest surface temperatures (T = 14–14.5 ◦C) are reached, dense water formation starts to occur. Potential density reaches its highest values (29.1 kg/m3) by the end of January until mid-February and the examination of single profiles shows that deep convection takes place down to at least 600 m during this period.

The Hoevmueller plot of the Turner angle (Figure 7) reveals statically unstable conditions (|Tu|> 90◦; dark blue and dark red points) from mid-January to the end of March and indicates the deep dense water formation events down to at least 600 m in February and March 2006. The deep dense water formation events are characterized by a stronger contribution of temperature (−45◦ < Tu < −90◦; blue points), while the main contributor to the stable stratification in December and April is mainly the salinity (45◦ < Tu < 90◦; yellow and light orange points).

**Figure 7.** The Turner angle (◦) of float WMO 6900098 describes the contribution of salinity and temperature gradients to the density gradient.

The heat and freshwater fluxes integrated over the center of RG show an intense preconditioning phase during December 2005, due to strong dry and cold winter winds which led to heat losses (Figure 8a) and evaporation (Figure 8b) and consequently to high surface salinity values. Additional heat losses by the end of January and the beginning of February coincide with the LDW formation event within the RG described above. The heat losses in mid-March coincide with the second dense water formation event within RG which led to a mixture of LDW and 'lower range' LIW formation.

**Figure 8. Upper panel:** Time series of integrated daily surface heat fluxes from December 2005 to April 2006 for the center of Rhodes Gyre. The heat losses by the end of January and mid-February induced deep convection and formation of Levantine deep water (LDW) while the heat losses in March induced mixing and formation of LDW and 'lower range' LIW (see also Figure 6). **Lower panel:** Time series of integrated daily freshwater fluxes for the same time period as the above panel. The freshwater fluxes in December show a strong evaporation which led to increased surface salinity as shown by the Argo float data (Figure 6b).

This deep convection event from the end of January until mid-February coincides with a strong depression of SSH within the RG area during that time, overlapping the exact position of the float (longitude: 28.5–29◦E, latitude: 35–35.5◦N; Figure 9). Figure 9a shows the float trajectory and mean SSH of January, February, and March 2006 while Figure 9b,c,d show the negative daily SSH and geostrophic currents for three specific days during the period of deep convection event from the end of January to mid-February. The eddy in which the float was trapped, represents the strongest depression (SSH < −0.3 m; about 20 cm below the seasonal mean (Figure 9a)) during winter months reaching a negative maximum during the days of deep water formation (Figure 6b–d).

The mesoscale eddy during that time shows a diameter of about 60 km which is within the typical mesoscale eddy diameter within the Levantine Sea (40–80 km).

**Figure 9.** (**a**) Trajectory of float WMO 6900098 depicted within white rectangle overlaid on mean SSH (m) for JFM 2006 for the NWLS. (**b–d**) daily SSH (m) during the LDW convection events (from 25 January to 7 February 2006). The mesoscale eddy within the white box has a diameter of about 60 km.

Figure 10 shows the monthly means of satellite SST superimposed on the geostrophic currents derived from SSH. The deep convection event occurred by the end of January until mid-February 2006 when the sea surface temperature was lowest. The lowest surface temperatures measured by the Argo float, evidenced within the Hoevmueller plots (Figure 6c), coincide with lowest temperatures by daily satellite SST (Figure 10e,f) and with the strongest depression of SSH (Figure 9b–d) by the end of January until mid-February.

Figure 11 shows the T/S plots for the two events of dense water formation during JFM 2006. Water masses above 100 m were not taken into account for the T/S plot to exclude shallow MLDs and recapping and to capture the events of pronounced intermediate and deep water formation. Water masses from 100 to 500 m are plotted with a green dot while water masses under 500 m are plotted with blue dots.

**Figure 10.** Monthly means of sea surface temperature (SST, ◦C) and geostrophic currents for (**a**) December 2005, (**b**) January, (**c**) February, and (**d**) March 2006. Daily SST (◦C) of (**e**) 2 February 2006 and (**f**) 6 February 2006. Dense water formation and the deep convection event occurred by the end of January until mid-February when SST was lowest.

**Figure 11.** (**a**) The temperature and salinity (T/S) plot for float WMO 6900098 from 20 January to 20 February 2006 indicates LDW formation. The additional density line with a potential density value of 29.17 kgm−<sup>3</sup> shows the upper deep-water boundary density which corresponds to approximately 1000 m depth for the NWLS [23]. (**b**) T/S plot for float WMO 6900098 for March 2006 indicates a mixture of LDW and 'lower range' LIW formation. Green dots represent depths from 100 to 500 m while blue dots represent depths from 500 to 600 m.

Figure 11a shows the T/S plot for the first dense water formation event: The potential temperature exhibits values smaller than 14.5 ◦C, the salinity shows values smaller than 39 psu, and the potential density shows a constant value of about 29.17 kg/m3. The typical ranges for LDW for potential temperature are between 13.7 ◦C and 14.5 ◦C and for salinity between 38.8 to 38.9 psu [6,9,10]. The potential density line of 29.17 kgm−<sup>3</sup> represents the upper deep-water boundary density for the NWLS, corresponding to approximately 1000 m depth [23]. All potential temperatures and salinity values lay on the line of constant potential density of 29.17 kgm<sup>−</sup>3, i.e., that the formed water masses sank to at least 1000 m, until the same potential density was reached. Therefore, the T/S plot confirms that LDW took place during the first event by late January until mid-February (Figure 11a).

For the dense water formation event in March 2006, the T/S plot shows potential temperatures smaller than 15 ◦C, salinities smaller than 39.1 psu, and potential densities between 29.125 kg/m<sup>3</sup> and 29.17 kg/m3. By nearly reaching 15 ◦C and with some salinity values above 39 psu, a part of the water mass reaches the lower range of LIW water mass characteristics (Figure 11b). This indicates a mixture of LDW and 'lower range' LIW formation during the second dense water formation event within RG.

(2) LDW formation took also place from the end of February until mid-March 2004 within another cyclonic mesoscale eddy located in the western part of the RG. Hoevmueller plots for DJFMA for salinity, potential temperature, and potential density are shown in Figure 12. During January and February, the MLD deepens constantly and the event of LDW formation occurs by the end of February until mid of March when the minimum surface temperature was reached. The examination of single profiles shows convection down to at least 600 m.

**Figure 12.** (**a**) Mean SSH (m) of the Northwestern Levantine Sea and float trajectory of float WMO 6900098 for JFM 2004. (**b**) Salinity (PSU), (**c**) potential temperature (◦C), and (**d**) potential density (kgm<sup>−</sup>3) from December 2003 to April 2004.

The Hoevmueller plot of the Turner angle (Figure 13) reveals statically unstable conditions (|Tu| > 90◦; dark blue and dark red points) and therefore a continuous deepening of the MLD from mid-January to mid-March 2004 and indicates the deep dense water formation events down to 400 m by the end of February and mid-March. The main contributor to the stable stratification in December and April is the salinity (45◦ < Tu < 90◦; yellow and light orange points), while the deep dense water formation

events are also characterized by a stronger contribution of temperature gradients (−45◦ < Tu < −90◦; blue points). 'Salt-fingering' (45◦ < Tu < 90◦; yellow points) can be noticed at a depth of about 350 m.

**Figure 13.** The Turner angle (◦) of float WMO 6900098 shows the contribution of salinity and temperature gradients to the density gradient.

The T/S plot of JFM2004 shows mainly LDW formation (Figure 14). Green dots represent water masses between 100 and 500 m while blue dots represent water masses between 500 and 600 m. The water mass characteristics show LDW and 'lower range' LIW (potential temperature around 15 ◦C and 39 psu < salinity < 39.1 psu) formation.

Figure 15 shows: (a) the mean SSH for JFM 2004 and the float trajectory (longitude: 27.5◦ E–28.5◦ E, latitude: 34.5◦ N–35◦ N) during this period and daily SSH (b) before; (c) during; and (d) after the convection event. A strong powerful structure develops by mid-February with a negative SSH of more than −0.3 m. Analysis of single profiles shows that on 8 March 2004, a MLD down to about 450 m was formed while on the surface recapping already took place (newly formed MLD of about 80 m).

**Figure 14.** The T/S plot for float WMO 6900098 for JFM 2004 shows LDW and 'lower range' LIW formation. Green dots represent depths from 100 to 500 m while blue dots represent depths from 500 to 600 m.

**Figure 15.** (**a**) mean SSH (m) for JFM 2004 and float trajectory of float WMO 6900098 (white rectangle). (**b**) SSH (m) for the 14 January 2004; float profiles show 'regular' winter MLDs around 150 m. (**c**) SSH (m) for the 16 February 2004; float profiles reveal dense water formation to about 350 m; (**d**) SSH (m) for 8 March 2004; recapping with a newly formed MLD of about 80 m occurred, while the dense water formation event down to 450 m is still captured by the float profile.

Monthly means of satellite SST are shown in Figure 16. The satellite SST reached a minimum between mid-February and mid-March, i.e., during the event of deep water formation, and coincides with the surface potential temperature measured by the Argo float (Figure 12).

**Figure 16.** Monthly means of the satellite SST (◦C) and absolute geostrophic currents from December 2003 to March 2004. The deep convection event occurred by the end of February and beginning of March 2004 when SST was lowest. The white rectangle shows the position of float WMO 6900098 during JFM 2004.

#### 3.2.2. LIW Formation along the Coastline

One example of LIW formation along the coastline is given in this subsection.

While LDW was formed inside RG, typical LIW was instead formed along the Northern Turkish coastline, i.e., along the AMC. The salinity, potential temperature, and potential density values characterize a typical LIW formation event by the end of March (Figure 17).

In December and January, very high surface salinity values (S > 39.3 psu) can be seen in the upper 50 to 100 m (Figure 17b). However, deep mixing and convection is still prevented by relatively high surface temperatures (T > 17 ◦C) during this period. Surface temperature has a decreasing trend from about 20 ◦C by the very beginning of December to about 17 ◦C in March when deep convection occurs. By the end of April, the surface temperature gradually increases reaching 17.5 ◦C. The observed surface water temperature along the coastline is about 1 ◦C to 3 ◦C warmer than in the open sea (Figures 6 and 12).

**Figure 17.** (**a**) Bathymetry of the Northwestern Levantine Sea (m) and float trajectory for float WMO 6900096 during JFM 2007. The red circle indicates the position of the float during March when the deep convection event occurred. Hoevmueller plots of (**b**) salinity (PSU), (**c**) potential temperature (◦C), and (**d**) potential density (kgm<sup>−</sup>3).

The MLD deepens gradually from about 50 m within December to about 250 m during February; therefore, the high surface salinity is mixed throughout intermediate layers (Figure 17b).

By the end of March, when lowest surface temperatures are reached (Figure 17c), dense water formation starts to occur. Surface potential density reaches values between 29 and 29.1 kg/m<sup>3</sup> during this period. The examination of single profiles shows that the mixing event takes place down to about 550 m.

The Hoevmueller plot of the Turner angle (Figure 18) reveals statically unstable conditions (|Tu|> 90◦; dark blue and dark red points) and therefore a continuous deepening of the MLD in January and February 2007. The main contributors to the stratification in December are both temperature and salinity (−45◦ < Tu < 45◦), while in April the main contributor is the salinity (45◦ < Tu < 90◦; yellow and light orange points). The Turner angle indicates the deep LIW formation event down to 550 m in March, which is also characterized by a stronger contribution of temperature gradients (−45◦ < Tu < −90◦; blue points).

The evolution of monthly mean SST from December 2007 to March 2007 is shown in Figure 19a–d. This deep convection event that was detected by the Argo float by the end of March 2007 coincides with the strongest depression of SSH and lowest surface temperatures along the coastline during winter 2007 (Figure 19a–c).

**Figure 18.** The Turner angle (◦) of float WMO 6900098 shows the contribution of salinity and temperature to density change.

**Figure 19.** (**a–d**) Monthly mean of satellite SST (◦C) from December 2006 to March 2007. (**e**) SSH (m) and (**f**) SST (◦C) on 30 March 2007, exhibiting strongest depression of SSH and lowest SST within this area during the whole winter period, coinciding with the deep dense water formation event. The white circle indicates the float position during March 2007.

The T/S plot for March 2007 reveals the formation of typical LIW, i.e., potential temperature above 15 ◦C and salinities higher than 39 psu (Figure 20). The potential density lines of 29–29.06 <sup>×</sup> kgm−<sup>3</sup> represent the potential density range of typical LIW ([6]; Figure 20).

**Figure 20.** T/S plot for float WMO 6900098 for March 2007. Green points indicate depths from 100 to 500 m while blue points indicate depths from 500 to 600 m. The additional density line with a potential density value of 29.17 kgm−<sup>3</sup> shows the upper deep-water boundary density while the potential density lines of 29–29.06 <sup>×</sup> kgm−<sup>3</sup> represent the potential density range of typical LIW [6].

#### *3.3. Climatology of Winter Mean MLD from 2000 to 2018*

The climatology of the winter (JFM) mean MLD from 2000 to 2018 for the Levantine Sea is shown in Figure 21. Within the cyclonic RG, the mean winter MLD is quite shallow (around 70 m). Deeper mean winter MLDs are found within anticyclonic eddies (IE, MME, CE, ShE; see Figure 1b for position of eddies) and along the coastlines, indicating dense water formation along boundary currents.

**Figure 21.** Climatology of the winter (JFM) mean MLD from 2000 to 2018 for the Levantine Sea.

#### **4. Discussion and Conclusions**

The present study is focused on the LIW and LDW formation events in the NWLS (Figure 22a) as detected by Argo float data from 2001 to 2017. The new and most interesting result is that the typical LIW (potential temperature > 15 ◦C and salinity > 39 psu) formation mainly occurred along the Northern coastline (Figure 18), while 'lower range' LIW (potential temperature about 15 ◦C and salinity about 39 psu) and LDW (13.7 ◦C < potential temperature < 14.5 ◦C, 38.8 psu < salinity < 38.9 psu) formation took place within mesoscale eddies located within the center of RG (Figure 22a; Figure 11a, Table 1).

**Figure 22.** (**a**) The white rectangle confines the area of study, the NWLS. Typical LIW formation was found along the Northern coastline (red arrow), while 'lower range' LIW and LDW formation was found within submesoscale eddies within the center of RG (black ellipse). (**b**) This figure, adapted from [15], summarizes the obtained results for winter seasons in the NWLS (yellow rectangle): Dense water formation along the Northern coastline reached intermediate (200–500 m; red line) and deeper layers (500–600 m; dashed red line), while LDW formation within the center of RG reached intermediate and deep layers (at least 1000 m; orange lines).

The schematic summary of the results for the winter seasons from 2001 to 2017 is evident in Figure 22b. Blue and brown arrows describe the convection and net sinking areas of the Mediterranean Sea as derived from theoretical models by [15,16]. Red and orange arrows are derived from the results of the present work. Red solid and dashed lines represent the formation of intermediate (200–500 m) and deeper waters (500–600 m) along the Northern coastline. Orange arrows represent the formation of intermediate (200–500 m) and deep waters (about 1000 m) within the RG, respectively. The examination of individual profiles showed dense water formation events reaching down to 550 m depth along the coastline and down to 1000 m depth within the center of RG (Figure 22b).

The specific event of LIW formation, captured by the Argo float data in March 2007 along the coastline, reached a depth of about 550 m. The T/S plot (Figure 20) showed typical LIW formation during this event and the observed surface water temperature along the coastline was about 1◦C to 3◦C warmer than in the open sea (Figures 6, 12 and 17), in agreement with the results of [24,25]. In January–February 2006, the Argo float data detected the LDW formation in the core of a cyclonic mesoscale structure located in the center of RG. This structure (diameter of about 60 km) shows the typical horizontal scale of convective chimneys. The T/S plot in Figure 11a reveals the LDW characteristics of the convection event. All potential temperatures and salinity values lay on the line of constant potential density of 29.17 kgm<sup>−</sup>3, revealing the sinking of the formed water masses to at least 1000 m as previously observed [6,9,10].

The intensity of the mixing and convection events depends mainly on the surface buoyancy flux B, which in turn depends on its thermal (BT) and haline (BS) components. The calculation and plot of the surface buoyancy flux and its components revealed that the haline component dominates over the thermal component (Figure 4, lower panel), i.e., intense evaporation (BS < 0) controls the surface buoyancy loss, especially during the preconditioning phase (e.g., Figure 8 for winter 2006). Therefore, the area of the RG is an area of net buoyancy loss, driven by the haline component, as shown by [26].

The influence of salinity and temperature gradients to the density gradients are described in the Hoevmueller diagrams of the Turner angle (Figures 7, 13 and 18). During the preconditioning phase (November, December) and the constant deepening of the MLD in the beginning of January, the influence of the salinity gradient was dominant, while during strong unstable conditions and consequently dense water formation, also the temperature gradient was influential. The Turner angle also approximately indicated the depth of the convection events.

The deep dense water formation events within the area of the RG can be described by the following phases: The whole process is influenced by the cyclonic rotation of the RG leading to the upwelling of cooler waters to the surface. In November and December, the preconditioning phase starts (Figures 4 and 5): the heat losses due to cold and dry winds lead to increased surface salinity (Figures 6, 12 and 17) through the evaporation and to a steady deepening of the MLD. Additional, temporarily outbreaks of strong winds during January, February, or March cause strong heat losses (Figures 5 and 8), which cause further cooling of surface waters. When lowest surface temperatures are reached (Figures 10, 16 and 19), dense water formation starts to occur. Within hours, the newly formed dense water sinks down rapidly to a depth of equal density where it spreads horizontally, forming an anticyclonic circulation due to the influence of the Coriolis force.The convection event also implicates a stretching of the water column leading to a change in vorticity, an increased geostrophic velocity, and a depression of the SSH. In fact, all pronounced dense water formation events documented by the Argo floats were indicated by a strong depression of satellite SSH (Figures 9, 15 and 19) and by lowest SSTs (Figures 10, 16 and 19).

The Argo float data revealed that LDW formation took place within the RG during winter months and showed the key role of the boundary currents for the LIW formation. The climatology of the mean MLDs of the Levantine Sea (Figure 21) reveals that, despite the deep convection events, little to net mean sinking takes place within the center of RG, while the deeper MLDs along the coastline indicate dense water formation occurring along boundary currents. Therefore, the drivers, sources, and main contributors of the Mediterranean thermohaline circulation have to be rethought not only within the Levantine, but also within the Mediterranean Sea. Deployments of additional Argo floats

to survey boundary currents and deployment of deep Argo floats within the main Mediterranean convection sites, i.e., the Gulf of Lion, the South Adriatic, and the RG area, will contribute to further understanding of dense water formation processes.

A better understanding of the Mediterranean thermohaline circulation is needed not only for a wider knowledge of the effects of climate change, but also for the impact on the ecology. Newly formed intermediate or deep waters can be polluted (with oil, microplastics, nutrients from extensive agriculture, heat due to global warming etc.), e.g., the Northern Levantine coastline, where pronounced dense water formation occurs, has the highest coastline plastic pollution within the Mediterranean Sea [27]. The newly formed water masses with the above-mentioned water properties and pollutants are transported throughout the Mediterranean to finally reach the Atlantic Ocean. The full impact (in terms of pollution and different water mass characteristics) will only be seen by future generations when these water masses emerge after decades or even centuries at different places within the Mediterranean Sea.

Therefore, it is obvious and evident that scientists and policy makers are obliged to join forces now to support and make commitments towards a real sustainable world that is not threatening, but protecting our ecosystems and lives.

**Author Contributions:** Conceptualization, E.K., P.-M.P. and E.M.; methodology, E.K., G.N. and M.M.; writing—original draft preparation, E.K.; writing—review and editing, M.M., E.M., P.-M.P. and G.N.; investigation, E.K.; data curation, E.K. and G.N.; supervision, P.-M.P. and E.M.; funding acquisition, P.-M.P.

**Funding:** The float data were collected and made freely available by the International Argo Program and the national programs that contribute to it (http://argo.jcommops.org). The Argo Program is part of the Global Ocean Observing System.

**Acknowledgments:** The authors would like to thank the two anonymous reviewers for their constructive comments and all the people who have deployed Argo floats in the Mediterranean Sea. We acknowledge Antonio Bussani and Massimo Pacciaroni for their technical support with the dataset and floats programming. In addition, we thank Vedrana Kovacevic for her suggestions and fruitful discussions.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


*Water* **2019**, *11*, 1781

27. WWF Mediterranean Marine Initiative Report. Stop the Flood of Plastic: How Mediterranean Countries Can Save Their Sea. 2019. Available online: http://awsassets.panda.org/downloads/a4\_plastics\_reg\_low.pdf (accessed on 25 June 2019).

© 2019 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 (http://creativecommons.org/licenses/by/4.0/).

#### *Article*

## **On the Variability of the Circulation and Water Mass Properties in the Eastern Levantine Sea between September 2016–August 2017**

#### **Elena Mauri 1,\*, Lina Sitz 1, Riccardo Gerin 1, Pierre-Marie Poulain 1,2, Daniel Hayes <sup>3</sup> and Hezi Gildor <sup>4</sup>**


Received: 18 July 2019; Accepted: 15 August 2019; Published: 21 August 2019

**Abstract:** The surface circulation and the thermohaline properties of the water masses of the eastern Levantine Sea (Mediterranean Sea) were monitored with mobile autonomous systems (surface drifters and gliders) during the period September 2016–August 2017. The drifters provided data for more than a year and revealed complex circulation features at scales ranging from the basin scale to the sub-mesoscale. Three drifters were captured in a semi-permanent gyre (Cyprus Eddy) allowing a quantitative study of its kinematics. During the experiment, three gliders were operated, in two different periods: September to December 2016 and February to March 2017. The autonomous instruments crossed the prevailing sub-basin structures several times. The collected in-situ observations were analyzed and interpreted in concert with remote sensing products (sea surface temperature and altimetry). The evolution of some of the prevailing features confirmed the complexity of the circulation of the basin. The Cyprus Eddy is the most persistent anticyclone, moving its geographical position and sometimes merging with the North Shikmona Eddy in a bigger structure. The gliders sampled this wide anticyclonic feature revealing its vertical structure in the two different periods. In fall, in stratified conditions, a high salinity core is evident below the thermocline. The isopycnals are characterized by an upward bending over the high salinity lens and a downward bending below it, typical of an anticyclonic modewater eddy. In winter, the core disappears following the vertical mixing that, homogenizes the upper Cyprus Eddy water down to 300 m.

**Keywords:** Mediterranean Sea; drifters; sub-basin-scale eddies; gliders

#### **1. Introduction**

The Levantine Sea (LS) is a complex multiscale system [1–3]. The basin-scale mean circulation, mesoscale and sub-basin scale eddies interact in a non-linear way producing a highly variable current field [4]. Despite many studies focusing on the LS, the basin has not been extensively sampled due to its high complexity and variability and to logistical or political issues.

Studies of the LS surface current started over a hundred years ago based on hydrographic data [5–11], revealing a basin-wide cyclonic circulation and the most persistent sub-basin scale features. A more detailed and complex circulation became evident only after the analysis of longtime series of satellite measurements, such as the Sea Surface Temperature (SST) [12,13] and the Absolute Dynamic Topography (ADT) [4,14–16] and numerical simulations [2,17–19]. Starting in 2005, as part of the Eddies and GYres Paths Tracking (EGITTO) [20] and North East MEDiterranean (NEMED) [4] projects,

numerous drifters were deployed in the region, allowing to calculate pseudo-Eulerian velocity statistics for different time periods. In particular, the use of the ADT for the period 1993–2010 allowed to describe the inter-annual variability of the Eddy Kinetic Energy (EKE) of the prevailing mesoscale features in the LS. The basin surface circulation map resulting from this study shows an along-slope cyclonic coastal circulation named the Libyo-Egyptian Current [12,20,21] extending as a northward current along the Middle-East coast. The other dominant feature is a central eastward cross-basin meandering current named Mid-Mediterranean Jet (MMJ) between 24◦ E and the longitude of Cyprus [22–24], and a series of mesoscale features, including some eddies such as the Cyprus (CE), the Shikmona (ShE) and the Latakia Eddy (LE). The CE is described as a persistent anticyclonic eddy characterized by seasonal variability in shape, dimension and position with an average diameter of 250–300 km [4,25]. The ShE instead, represents a complex system, composed of several cyclonic and anticyclonic eddies off the Israeli coast, in which the positions, sizes and intensities vary markedly [4,26–28]. The LE can be present as a cyclone or an anticyclone and the change in rotation is induced by the interaction of the MMJ with the northward meandering coastal current [4].

The thermohaline structure of the eastern LS is well-defined in the warm season and it is characterized at surface by the Levantine Surface Water (LSW), with temperature values between 22 and 28 ◦C and salinity of 39 to 39.6 PSU. The Atlantic Water (AW) with temperature values of 18 to 22 ◦C and salinity between 38.6 and 39.2 PSU, is positioned below the LSW and it is advected from the western Mediterranean Sea. The Levantine Intermediate Water (LIW) formed when LSW cools down and sinks along isopycnals to intermediate depths (ca. 130 m < *z* < 350 m), it presents typical values of 15 to 17.5 ◦C and 38.95 to 39.3 PSU [29]. Finally, the Levantine Deep Water (LDW) with its nearly constant values of temperature and salinity (13.8 ◦C and 38.7 PSU) is found below 750 m [30].

In the framework of the CINEL (CIrculation and water mass properties in the North Eastern Levantine) project drifters and gliders were operated in the eastern part of the LS for more than a year, starting in September 2016, to gain more insights on the variability of the physical and biochemical properties in the region and in particular, to study the major sub-basin scale and mesoscale eddies governing the dynamics of the eastern LS. The in situ observations provided by the drifters and gliders were used in concert with satellite products of SST and ADT to describe the spatial structure and the temporal evolution of the main eddies.

The use of satellite images in past studies to track mesoscale and sub-mesoscale features has been widely used. In the Mediterranean Sea, SST, chlorophyll and altimetry imagery were utilized in different studies [4,16,31–34], some of them the eddies were sampled by gliders [35–38].

In this study, we concentrate on the eastern LS, located between 30◦ E and 36◦ E, and 31◦ N and 37◦ N in the period between September 2016 and August 2017. The main focus of the analysis is on the detection and monitoring of strong sub-basin scale features and on their motion and evolution during the period of the experiment.

The paper is organized as follows. Information on the in-situ platforms (drifters and gliders) and the data they provided, as well as on the remotely sensed data (SST and ADT), is provided in Section 2. The methods applied to process all the data are also explained. In Section 3, the features highlighted by the SST anomaly and ADT images are interpreted in concert with the drifter tracks and the surface geostrophic currents computed from the satellite altimetry data. The vertical description of some sub-basin scale features using glider observations is also included in this section. Discussion and conclusions are found in Section 4.

#### **2. Data and Methods**

#### *2.1. Glider Data*

Three Seagliders (sg149 and sg150 of University of Cyprus—UCY; sg554 of National institute of Oceanography and Applied Geophysics—OGS) were operated in the eastern LS between 1 September 2016 and 16 March 2017 (See Table 1). The glider dataset includes five glider campaigns (Figure 1), OC-UCY (sg149)

4 November–

which covered two different seasons: fall 2016 and winter 2016–2017. During the first period, three glider campaigns were organized; the first one was performed in September 2016 (C1) and the other two (C2 and C3) were almost concomitant and covered the month of November and the beginning of December 2016. The winter sampling comprised two simultaneous campaigns (C4 and C5) starting at the beginning of February and ending in mid-March 2017.


6 December C3 143 10 February–

12 March C4 168

**Table 1.** Dates of glider campaigns in the two periods, conventional name of the missions used in the paper and number of recorded profiles.

The OGS sg554 glider is equipped with a pumped CTD (GPCTD) while the UCY gliders have a regular un-pumped CTD. The un-pumped data were corrected for the thermal lag using Kongsberg routines. Temperature and salinity data were collected in the top 950 m of the water column in all the five campaigns. Oxygen concentration was recorded by an Aanderaa optode 4330 and 3835 (for the OGS and UCY glider, respectively) and set to record data down to 600 m depth. All the gliders were also fitted with a Wetlab FLNTU sensor to collect chlorophyll and turbidity data at two wavelengths (470 and 700 nm) down to 300, 400 m or 600 m while crossing specific structures. The glider traveled along a ~26◦ inclined path with respect to the water surface. During the up casts, the glider collected high vertical resolution data (about 0.1 Hz, corresponding to about one sample every 2 m). The top 20 m CTD data of the OGS campaigns are missing. This behavior is common in Seabird GPCTD and is due to the incomplete download of the GPCTD buffer when the glider reaches the surface (SeaBird, 2013). The horizontal speed of the instrument spanned between 0.7 and 1.4 km/h depending on the sea currents. A glider took about 4–5 h to conclude a 950 m dive, therefore, the horizontal resolution was about 4 km. The data were transmitted via satellite Iridium links at each surfacing. In total, 974 casts were collected during the five missions.

During fall 2016 (campaigns C2 and C3), the OGS and UCY gliders were operated simultaneously. On 3 December, they were close by and the two profiles (one per glider) recorded at that particular moment (separated by about 1.5 km and 1 h) were used to intercompare the CTD data. Temperature and salinity profiles showed a good agreement (offset of 0.02 ◦C and 0.01 PSU, RMSD of 0.07 ◦C and 0.01 PSU). Oxygen data were not inter-compared because the oxygen factory calibration for the UCY gliders resulted too old and a sensor degradation (shift over times) was suspected. Therefore, only the relative values of the oxygen concentration are used in this paper. Chlorophyll and turbidity data were corrected using the dark counts computed from deep dives performed at the beginning and at the end of each glider campaigns. After a first quality control to eliminate obvious spikes [39], the data were averaged in 2-m non-overlapped bins. In this paper, the temperature, salinity, density and dissolved oxygen data are described while other parameters are only occasionally considered.

**Figure 1.** (**a**) Glider tracks during the CINEL experiment, (**b**) C1 mission performed by units SG150 glider during fall; (**c**) C2 by unit SG554 in fall, (**d**) C3 by unit SG149 in fall; (**e**) C4 by SG149 in winter; (**f**) C5 by SG554 in winter. Colors represent the evolution in time of each trajectory, with blue colors indicating the initial measurements and red colors showing the final observations. The actual dates for each mission are reported in Table 1.

#### *2.2. Drifter Data*

The CINEL drifter dataset includes the tracks provided by 16 drifters launched in 3 episodes: 4 drifters on 20–21 October 2016 along a meridional transect south of Cyprus and 8 drifters off the central coast of Israel on 7 February 2017 and 4 drifters again south of Cyprus (same positions as in October 2016) on 25 February 2017. One drifter stranded near Larnaca in Cyprus on 14 March 2017. It was recovered and redeployed south of Cyprus on 29 March. The drifters used during the experiment were the Surface Velocity Programme SVP drifter design [40] with a drogue centered at 15 m depth, equipped with a sensor for the SST and a tension sensor to monitor the drogue presence. They were manufactured by METOCEAN. Each drifter provides its location through the global positioning system (GPS) and transmits the data on land via Iridium satellite link. The drifter data were first edited from spikes and outliers [39], then the data of position, temperature, voltage and drogue presence were interpolated at 30-min uniform intervals using a kriging optimal interpolation method [41,42]. The velocities were then calculated as finite differences of the interpolated position. The interpolated positions were also subsampled at 2-h regular interval and low-pass filtered using a Hamming filter with a cut-off period at 36 h in order to remove high frequency current components (tidal and inertial currents) and then sub-sampled every 6 h. A composite diagram with all the low-pass filtered drifter trajectories between 20 October 2016 and 31 August 2017 is shown in Figure 2.

Drifter trajectory segments were superimposed on SST anomaly and ADT images to describe specific snapshots of surface circulation. A quantitative analysis of the drifter trajectories was also performed to characterize the CE, when some drifters were caught in its core for several months. To estimate the motion of the eddy center, loops in the drifter tracks were identified by considering the drifter positions bounded by two successive longitude maxima [43]. Then, the center of each loop was computed by averaging the longitudes and latitudes corresponding to each loop. A regression line was fitted through the center estimates to estimate the mean displacement of the eddy. The size and strength of the eddy were determined by considering the drifter speeds as a function of distance with respect to the eddy center.

**Figure 2.** Tracks of drifters operated during the CINEL experiment from September 2016 to August 2017.

#### *2.3. Sea Surface Temperature*

The satellite SST products considered in this study are daily gap-free maps (L4) at high (HR 0.0625◦) spatial resolution over the Mediterranean Sea. These products are based on night-time images collected by the infrared sensors mounted on different satellite platforms and cover the Southern European Seas. Remotely-sensed L4 SST datasets are operationally produced by the Consiglio Nazionale delle Ricerche-Gruppo di Oceanografia da Satellite (CNR-GOS). The basic design and the main algorithms used are described in [44]. The products are distributed by the Copernicus programme (http://marine.copernicus.eu). In this study, we generated SST anomaly maps for the east LS (30–36◦ E and 31–37◦ N) with a constant colorscale (between −2 and +2 ◦C) in order to enhance the features present in the area. In particular, each image was generated after subtracting the mean SST of the area and dividing the result by the standard deviation.

#### *2.4. Absolute Dynamic Topography*

The altimetry data are processed by the DUACS multi-mission altimeter data processing system and provided by Copernicus. Satellite gridded Sea Level Anomaly (SLA) are computed with respect to a twenty-year mean and is estimated by Optimal Interpolation [45], merging the measurement from all the available altimeter missions (Jason-3, Sentinel-3A, HY-2A, Saral/AltiKa, Cryosat-2, Jason-2, Jason-1, T/P, ENVISAT, GFO, ERS1/2) (see QUID document or http://duacs.cls.fr pages for processing details). The SLA data resolution is 0.125◦ by 0.125◦; to derive the ADT, the SLA is added to the Synthetic Mean Dynamic Topography (SMDT). The geostrophic currents are then derived by finite differencing and normalizing the ADT.

#### **3. Results**

#### *3.1. Surface Circulation and Sub-Basin Features*

The analysis of the drifter trajectories (Figure 2), and the SST anomaly/ADT daily images confirm the eastern LS as a very dynamical area, characterized by a number of permanent eddies and other features that are intermittent, eventually disappearing or merging with other eddies, creating wider structures. The ADT averaged throughout the whole period of study (September 2016–August 2017; Figure 3) and the drifter tracks (Figure 2) show the CE as the most persistent feature present in the area, covering a broad area south of Cyprus. Other relevant anticyclones are the North Shikmona Eddy (NSE), which is the easternmost feature off the Lebanese coast and another eddy located more to the south in front of the Israeli coast, hereafter called E1. In between these two anticyclones, a cyclone is evident. It is referred to as the South Shikmona Eddy (SSE) hereafter. Finally, another two cyclones, evidenced in the mean ADT, can be seen: one south of the CE, named E2 and the second one east of Cyprus island called the Latakia Eddy (LE). The above-mentioned features with the exception of the CE, are not permanent throughout the year, they can disappear for short periods or move slightly their geographical position or sometimes change their shape merging with other eddies. A qualitative analysis of the most salient features during a 1-year period starting from September 2016 was performed using SST anomaly and satellite altimetry maps. To analyze the mesoscale circulation, geostrophic currents computed from the altimetry data of the same day were superimposed on the daily SST anomaly images. Drifter tracks were also integrated to study the dynamics of the area.

**Figure 3.** Mean Absolute Dynamic Topography (ADT) for the period September 2016 to August 2017, evidencing the Cyprus eddy (CE), the north Shikmona Eddy (NSE), the south Shikmona Eddy (SSE), the anticyclone E1 in front of the Israeli coast, a cyclone south of CE called E2 and the Latakia Eddy (LE). The colorscale unit is m.

#### 3.1.1. Qualitative Description

Mid Mediterranean Jet and Thermal Front between Cyprus and Syria

From the beginning of September 2016 a strong jet (Figure 4a), crossing from west to east the eastern LS south of Cyprus, was evidenced by low temperature and by strong geostrophic currents. The wind over the whole basin blew from the west starting from June 2016 and becoming stronger and more stable at the beginning of August and lasting until 25 September 2016 (https://winds.jpl.nasa.gov/missions/quikscat/). The jet, called from now on the Mid Mediterranean Jet (MMJ), bended toward Cyprus skimming the CE to the north, and proceeding toward the Syrian coast. Before reaching, the coast, it flowed to the north between two eddies: the LE to the north and the NSE to the south (Figure 4b). The cold water of the MMJ (1.5–2 ◦C lower than the surrounding water), intruded from the west, joining the coastal upwelled water along the southern coast of Cyprus. In early October, the cold MMJ faded away (Figure 4c) and the warmer waters of the CE, intruded to the north, confining the cold upwelled water to the southeastern part of Cyprus (Figure 4c–i). The cold water, together with the general cooling of the area northeast of Cyprus, generated a zonal thermal front (Figure 4e–i). From September to November 2016 (Figure 4a–f) the strong geostrophic currents flowing from south of Cyprus toward the Syrian coast, were mainly located south of the front (in warmer waters). Three of the four drifters deployed along the meridional section of the CE in October 2016, clearly highlighted the zonal jet associated with the front. The northern drifter was immediately captured by the jet (Figure 4d) and reached the Syrian coast after about a week, then it veered to the north (not shown). The drifter deployed in the CE more to the south, after being involved for 10 days in a small eddy southwest of CE, was first caught by the CE, then it moved to the north and entered in the jet (Figure 4e). On 21 November, it reached the Syrian coast. Another drifter reached this area on 25 November and then both instruments moved southward along the Lebanese coast (Figure 4f). The geostrophic current in front of the Lebanon/Syrian coast, computed by the altimetry, show for this period a current flowing against the drifter tracks (not shown). One drifter was caught in the SSE, while the other one proceeded to the south along the coast until 11 December when it changed direction and moved northward to reach the SSE, in agreement with the geostrophic currents (not shown).

The zonal thermal front northeast of Cyprus persisted until March 2017 changing its shape and orientation. After a first phase in September 2016 during which the front appears jagged and not well defined in the SST anomaly (Figure 4c), in October and November it became sharper with a surface temperature gradient of 1.5–2 ◦C (Figure 4f). Two gliders were operated in November and both crossed the front simultaneously a few km apart. During this period, the zonally oriented front persisted until 15 December when a cyclonic eddy located around 34◦ E, intruded in the warmer side bringing colder water to the south (not shown). The wide cold bulge expanded more than 1 degree of latitude to the south in 1 December (Figure 4g) and the intrusion became narrow due to the weakening of the cyclone by 11 January 2017 (Figure 4h). A small feature of cold water, which detached from the above-mentioned bulge, remained at 33.5◦ E 34.3◦ N in mid-January (Figure 4i) and was included in a small cyclonic eddy (probably the SSE). On 23 February, two cold intrusions of 10 km size propagated to the east along the front moving more than 40 km in 4 days (Figure 4j,k). In early March the jet decreased in intensity and more meanders and filaments developed. Some of them were also identified by the drifter tracks (Figure 4l).

**Figure 4.** Geostrophic currents overlaid on the Sea Surface Temperature (SST) anomaly images, near the bottom the right corner of the sub-plots, mean, and standard deviation of the image are reported as well as the speed scale of the currents. Black (gray) lines indicate the glider track 4 days before (4 days after) the date reported on top of the image (black dot). White lines represent the drifter tracks 4 days before the reported date (white dot). Mid-Mediterranean Jet (MMJ) and the upwelling south of Cyprus, marked by cold water (**a**); the fading of MMJ with instabilities along the edge (**b**); cold front south east of Cyprus (**c**); cold front highlighted by a drifter (**d**); a second drifter evidences the fast geostrophic current north of Cyprus (**e**); drifters proceeding south along the Lebanon coast (**f**); north intrusion in the front (**g**); evolution of the intrusion (**h**); creation of a separate cyclonic eddy (**i**); intrusions (**j**); evolution of the intrusions (**k**) and cold front weakening (**l**).

Cyprus Eddy and North Shikmona Eddy

Even though the CE prevailed throughout the period of study from September 2016 to the end of August 2017, its shape and position changed significantly with time as revealed by the analysis of the ADT maps. At the beginning of September, the CE and NSE were present as two separate anticyclones (Figure 5a), then, on 4 October, three small eddies developed (Figure 5b) and in a few days, a wide zonally-elongated anticyclonic structure formed (Figure 5c). This broad feature persisted (Figure 5d) until the beginning of December, when the two eddies eventually detached from each other (Figure 5e) and the CE moved back to its original location, while the NSE completely disappeared (Figure 5f).

**Figure 5.** Selected maps of ADT with overlaid geostrophic currents with black, gray and white lines and dots as in Figure 4. CE and NSE evolution; anticyclones as separated eddies (**a**); presence of 3 small eddies in the same area (**b**); wide zonally-elongated anticyclonic structure (**c**); persistence of the CE-NSE anticyclone (**d**); weakening of the elongated anticyclone eddies (**e**) and disappearance of NSE (**f**).

To better show the time evolution of the two eddies and their geographical position, two Hovmoeller diagrams using as reference the altimetry data at 33.81◦ N (Figure 6a) and 34.2◦ E (Figure 6b) were produced. The latitude and longitude of the CE were selected based on the mean ADT for the period September 2016 to August 2017 (see Figure 3). As mentioned before, from October to December 2016 a large zonally-expanded anticyclone structure characterized the area. In November, this feature covered zonally the broadest area from 32◦ E to 35◦ E (Figure 6a) and in the meridional direction from 33.4◦ N to 34.7◦ N (Figure 6b). Starting from December, the CE was evident again as a single anticyclone with a lower signature in ADT and a smaller extension of about one degree in both longitude and latitude.

**Figure 6.** Hovmoeller diagram of ADT (**a**) along latitude 33.81◦ N and (**b**) longitude 33.19◦ E. Position of the CE determined from the annual ADT image. The black line shows the eddy displacement derived from the drifter data.

In September the gliders sampled the water column in the CE while in October-November 2016 they captured the CE-NSE merging event (refer to paragraph 3.2.).

In stratified conditions, from May until August 2017, the anticyclone was evidenced by a core of cold water (Figure 7). The area south of Cyprus was characterized by a strong upwelling with cold waters advected southward by the CE (Figure 7a–c). Starting from May, the eastern part of the basin becomes increasingly cooler. In July, cold filaments were advected toward the north, tracing the external edge of the CE (Figure 7b). By mid of August 2017, the cold western water was not entering as MMJ as in the previous year (Figure 7c). From June until August the NSE moved more than <sup>1</sup> <sup>2</sup> degree north, despairing for 2 weeks in mid-July.

**Figure 7.** SST anomaly images with overlaid geostrophic current with white lines and dots as in Figure 4. CE marked by a cold core from May to August 2017; cold core and upwelled water south of Cyprus advected by the CE (**a**); CE cold core and a westward filament tracing the external edge of the CE (**b**); CE extending to the Cyprus coast, upwelling south of Cyprus, absence of the MMJ (**c**).

Upwelling off Israel and South Shikmona Eddy

In late November 2016, an event of upwelling took place along the Israeli coast and lasted for about 7 days. The phenomenon, induced by eastern wind blowing for 10 consecutive days over 3 m/sec, was visible from 26 November to 2 December 2016. The upwelled water showed a temperature decrease of 2 ◦C close to the coast. The SSE and other smaller eddies to the south probably advected the cold water in the open sea and the offshore flowing filaments maintained a difference in temperature while moving away from the coast (Figure 8). The two gliders sampled the area during the same period, but they did not capture the upwelling signature. The cold layer, advected away from the coast, probably affected only the top surface layer, which was not sampled by the gliders.

**Figure 8.** SST anomaly image with overlaid geostrophic currents with black, gray and white lines and dots as in Figure 4. Event of upwelling in front of the Israeli coast at the end of November 2017.

During fall and winter, the ADT maps show a distinct motion of the SSE. In order to have a better description of the 1-year eddy evolution Hovmoeller diagrams of the ADT were constructed (Figure 9). The reference position (latitude 33.06◦ N–longitude 34.31◦ E) of the eddy was identified from the mean ADT map (Figure 3). In September during the first glider campaign the eddy was weakly detected in ADT, then from October to the end of November during the second and the third glider campaigns the eddy became well defined (Figure 5c) and moved southwestward ( <sup>1</sup> <sup>2</sup> degree south and 1 degree west; Figure 9a,b). In early December, the SSE moved farther west and a new SSE appeared at the usual position. From mid-December to the beginning of May, the SSE alternated periods when the eddy was well defined and periods in which it merged with another cyclone intruded to the south; sometimes the south tongue was close to the Lebanon coast (Figure 10a), other times the presence of NSE pushed the intrusion farther offshore (Figure 10b). The SSE finally splits in two cyclones at the end of May 2017 as shown in Figure 10c. During the summer months the eddy was weak or disappeared for weeks while in August it moved <sup>1</sup> <sup>2</sup> degree north (Figure 7c).

**Figure 9.** Hovmoeller diagram of ADT (**a**) along latitude 33.06◦ N and (**b**) longitude 34.31◦ E.

**Figure 10.** ADT with overlaid geostrophic currents with white lines and dots as in Figure 4. SSE merged with a cyclone intruded from the north: sometimes the intrusion was close to the Lebanon coast (**a**); the presence of NSE pushed the northern intrusion father offshore (**b**); SSE splits in two cyclones (**c**).

#### 3.1.2. Quantitative Description of CE Using Drifter Data

Three drifters moved in and around the CE from February to May 2017, while two of them remained trapped in the eddy until August 2017. From the analysis of the data of the 3 drifters it was found that the eddy can be considered to be in quasi-solid body rotation up to a radius of 40–50 km, where the maximum drifter speed of 50 cm/s occurred [43]. In the 4 months period, the radius ranged between <10 km to about 100 km with some oscillations, due to the fat that the eddy might not be perfectly circular and also due to possible variations in rotation strength. From the estimation of the eddy center position, it was found that the anticyclone moved toward the southeast at a speed of about 150 m/day (see more details in [43]). This displacement is in good agreement with the ADT maps. The black line depicted using the coefficients computed by the drifter data follows the motion of the CE displayed by Hovmoeller diagrams (Figure 6).

#### *3.2. Qualitative Vertical Description of Some Sub-Basin Features Using Glider Surveys*

In Fall 2016, the three glider campaigns were designed to sample the vertical structure of some sub-basin scale features between September and the beginning of December. The data of the second campaign (see track in Figure 1c) were plotted as a function of time (Figure 11). The eastern LS area appears strongly influenced by the presence of sub-basin eddies and mesoscale features. The thermohaline structure throughout the mission, shows a strong vertical stratification, with the presence of warm and salty LSW near the surface down to 40 m (Figure 11a), which gradually deepens, cooling and freshening (Figure 11b). The temperature (salinity) parameter between 20 m to 40 m reaches a maximum value of 26.24 ◦C (39.57 PSU). By the end of the mission, the mixed layer has deepened to 90 m and the maximum temperature has decreased to 20.38 ◦C (39.36 PSU). The AW is found at depths just below the LSW, under the thermocline (at about 50 m in October and 100 m depth at the end of November), revealed as a local minimum in salinity (about 38.8 PSU; Figure 11b), which slowly disappeared as the season comes to the end. Furthermore, a 70 m thick layer with relative salinity maximum is observed between 100 and 400 m corresponding to the thermohaline characteristics of the LIW. Below that depth in the LDW, the salinity gradually decreases with depth (not shown).

The contour plots of the second campaign (Figure 11) reveal the downward bending of the isolines of the water mass properties in correspondence to the CE anticyclone at the beginning and at the end of the mission. In between these anticyclones, a sequence of 4 transects, marked by individual upward domings, depicts different parts of the SSE area. The data after the second SSE transect (10–11 November) corresponds to the CE only partially sampled. After the fourth transect (22–25 November) data describe an anticyclone south of CE. The anticyclonic and cyclonic features are identified by acronyms and vertical black lines in Figure 11. The isopycnal curves overlaid on the temperature field show a homogeneous top layer and a highly stratified layer in correspondence to the AW below the thermocline.

The contour lines, representing the oxygen concentration (overlaid in Figure 11b) show a subsurface maximum between 65–100 m, diminishing gradually along the water column, while the top layer is quite homogeneous. The oxygen stratification is evident throughout the mission with the exception of the CE, where the isolines between 150 m depth and the LIW show a less stratified field. Some mesoscale structures visible in temperature, salinity and oxygen are present inside the CE (see description in the following section). The chlorophyll concentration (Figure 11c) displays a Deep Chlorophyll Maximum (DCM) at around 50 and 150 m depth. Its spatial structure (thickness and patchiness) is strongly influenced by the presence of the eddies. In the anticyclone, the DCM is deeper than in the cyclone, where the chlorophyll concentration is slightly higher. Maximum values are around 0.8 μg/L while the minimum is around 0.2 μg/L. Some filaments detach from the DCM in particular along the edges of the anticyclone, reaching 200–250 m depth. The backscattering (Figure 11d) exhibits different vertical distributions in the anticyclones and cyclones. In the CE backscattering is relatively high between 50 and 150 m, while inside the eddy below 150 m it is extremely low. This feature is present not only in the two CE transects but also in the area where two anticyclones were partially sampled (10–11 and 22–25 November). The extremely low backscattering values in the CE (the absolute minimum of the mission) also corresponds to the region of lower stratification in oxygen concentration. In the deeper layer down to 600 m (not shown, because sampled in mission C3) the backscattering is low, but the recorded absolute minimum is still between 150 and 300 m depth. The cyclone exhibits higher values of backscattering in the upper layer in the first 2 crossings of the SSE, then when the winter mixing starts, the values are lower even in the upper part of the SSE. There are relatively higher backscattering values below 200 m in the SSE area.

**Figure 11.** Glider data collected during the second campaign as a function of time and depth (20–450 m). Temperature field (◦C) with isopycnal (intervals of 0.1 kg/m3) overlaid (**a**), salinity field with oxygen (intervals of 0.2 mL/L) overlaid (**b**), fluorescence field (μg/L) with isopycnal overlaid (**c**), backscattering at 700 (m<sup>−</sup>1) with oxygen overlaid (**d**).

#### *The Cyprus Eddy and North Shikmona Eddy*

During fall 2016, the area of the CE formation was sampled five times (see Table 2 for timing and other details). Since the gliders crossed the anticyclones along almost meridional transects (in both directions), the data were plotted in Figure 12 as a function of latitude for a better description and comparison. Each crossing transect was completed in 5 to 10 days from the beginning of October to the beginning of December 2016.

**Figure 12.** Fall salinity fields of the anticyclonic structures in the region of the CE (**left**). Transects are ordered chronologically and dates are reported in Table 2. Isopycnal (intervals of 0.1 kg/m3) and isolines of oxygen concentration (intervals of 0.2 mL/L) are respectively shown with black and white lines. On the right, the ADT field and geostrophic currents. Black, grey, white lines and dots as in Figure 4. Salinity sections depicted in panels (**e**,**g**,**i**) correspond to the northern, western and eastern glider tracks in panels (**f**,**h**,**j**), respectively.


**Table 2.** Season (F—fall and W—winter), numbers assigned to the transect crossing the CE, mission name, period and approximate direction of sampling.

All five salinity fields collected during fall are displayed in chronological order in Figure 12. The ADT fields, with the associated geostrophic currents and glider/drifter tracks overlaid, (Figure 12, right column), give additional information of the surface dynamics during the sampling interval. The dates of the selected ADT maps correspond to the times when the gliders sampled the center of the isolines deepening. The black dot on the glider track corresponds to its position on that date. From the analysis of the ADT the first transect (Figure 12b) crosses a zonally elongated anticyclone corresponding to the merging of the CE and NSE, described in paragraph 3.1.1. Transect 2 (Figure 12d), one week later, crosses the wide formed structure close to the NSE usual position. As time passes the anticyclone becomes larger with strong geostrophic currents as observed during the third glider crossing (Figure 12f). Approximately one week later (Figure 12g–j), the same structure is captured almost simultaneously along transects 4 and 5. The transect 4 evidences the eddy in its central part, while the transect 5 is relative to its easternmost portion.

On the left side of Figure 12 the salinity vertical structure with density (black curves) and oxygen (white curves) overlaid, describes the five sampled transects of the anticyclone in the CE area between October and the beginning of December 2016. The water masses of the CE are typical of the eastern LS but the thermohaline vertical structure is strongly influenced by the deepening of the isolines extending down to 950 m. It is noticeable that all the transects exhibit an asymmetry in all parameters that does not seem to be related to the sampling method, since it is present both in the transects sampled from north to south (Figure 12c,e) and in those sampled in the opposite direction (Figure 12a,g,i).

A high salinity subsurface core of different shape and size is present in all transects. The homogeneous salinity core or lens is around 39.1 PSU and mainly enclosed within an isohaline of 39.05 PSU. The vertical extension of the core in transects 2, 3, 4 is between 50 and 250 m and laterally from 33.4 to 34◦ E (more specifically 33.4–33.6◦ E in transect 2, 33.5–33.8◦ E in transect 3 and 33.9–34◦ E in transect 4). This core is marked also in temperature (not shown): the 18 and 18.5 ◦C isotherms are bended downward, surrounding the lower part of the lens (e.g., Figure 11a). The core in its upper part evidences an upward doming in salinity, density and oxygen concentration.

The isopycnals highlight a strong vertical stratification in density in the upper part of the core while the lower part is rather homogeneous. A minimum of oxygen is clearly visible in the lower part of the lens (white lines in Figure 12c,e,g). The high salinity core evident in transects 1 and 5 (Figure 12a,i) have smaller vertical extension spanning between 250 and 300 m and laterally from 33.3 to 33.5◦ E in transect 5, while it is barely visible in transect 1.

In the attempt to have a better perception of the high salinity core geographical position, the five glider pathways, crossing the CE, were overlaid on the bathymetry of the area (Figure 13). The dots correspond to the salinity at 150 m and they are color-coded according to the salinity values. The transects are numbered following a chronological order. During the first and fifth transect a small high salinity core is barely visible because the lens is located deeper (Figure 12a), while in transect 2, 3, 4 the core is well captured. Figure 13 shows that the salty core was sampled first to the west during transect 2, then it was captured <sup>1</sup> <sup>2</sup> degree to the east in transect 3 and again <sup>1</sup> <sup>2</sup> degree to east during transect 4. Transect 5 was almost concomitant to the fourth transect but did not capture the salty core

at 150 m. High salinity is also evident in the northern part of the transects, where the bathymetry is shallower than in the southern part, but this is another feature different from the signature of the high salinity core.

**Figure 13.** The fall 2016 glider pathways corresponding to the transects shown in Figure 12. The color-coded dots are the salinity sampled at 150 m depth. Transects are numbered chronologically according to Table 2.

In winter, the gliders crossed the CE twice (see transects 6 and 7 in Table 2) and only transect 6 is described as an example. The deepening of the water mass properties is evident. The salinity field is displayed in Figure 14a, while its geographical position is shown on the ADT image in Figure 14b. The CE appears like a body of homogeneous salty water (39.2 PSU) down to 400 m depth, included in the downward doming of the LIW. The temperature in the surface layer (not shown) is homogeneous only to 100 m depth in the northern edge and 200 m in the southern edge, confirmed by the isopycnals (black lines in Figure 14a). The AW is no longer visible as a local minimum in salinity, having been mixed with the layers below and above during the winter in the previous weeks. The LIW is also no longer identifiable as a relative maximum in salinity but it still conserves the thermohaline properties found during fall (Figure 15). In the winter structures, both chlorophyll and turbidity fields (not shown) are vertically homogeneous reaching 300 m depth.

**Figure 14.** Winter salinity field of transect 6 (dates in Table 2) isopycnals (black lines, intervals of 0.1 kg/m3) overlaid (**a**); ADT image of the day when the glider crossed the center of the eddy (**b**). The transect is relative to the glider track depicted to the left, the dot is the position of the glider on the same day of the image, the black line is relative to 4 days before and the gray line is relative to 4 days after. White curves are 4-day trajectory segments of drifters (shown with white dots).

**Figure 15.** Theta-S diagram in the CE region obtained from the fall (winter) glider data in dark gray (light gray), the dashed lines are isopycnals. Turquoise, light blue, dark blue and magenta dots correspond to the high salinity core found in the transects 2, 3, 4 and 5 respectively. Red dots are relative to values between 20–200 m in winter transect, while orange dots correspond to the salinity maximum associated to the LIW, both for fall and winter missions.

To analyze further the thermohaline properties of the high salinity core a Theta-Salinity (TS) diagram of the CE region is depicted in Figure 15. All the glider temperature and salinity data of transects 1 to 6 (see Table 2 and Figures 12 and 14) are represented during fall (dark grey) and winter (in light grey). The high salinity core data, found in transects 2, 3, 4 and 5 are respectively plotted in

turquoise, light blue, dark blue and magenta. The red is used to distinguish the homogeneous salty water mass present in the upper layer (20–200 m) inside the CE during winter (Figure 14a), while orange corresponds to the salinity maximum associated with the LIW, in both seasons. All the high salinity cores present the salinity of the LIW, with exception of a few values of the core in transect 5. The density of each core ranges between 28 and 28.34 kg/m<sup>3</sup> and the vertical structure inside the core is mainly driven by temperature (as visible from the isopycnals in Figure 12). The core in transect 5 has a lower temperature in comparison with the others resulting in a higher density (28.36–28.4 kg/m3).

#### **4. Discussion and Conclusions**

The use of satellite altimetry and SST anomaly in concert with drifter data allowed us to analyze the surface evolution of some of the major sub-basin and mesoscale circulation structures in the eastern LS in two different seasons between September 2016 and August 2017. The eastern LS appears very variable over time: the geostrophic position, the size and the shape of most eddies change in a few weeks due to their interaction or their complete disappearance.

The anticyclonic CE is the most persistent feature in the above-mentioned period covering a broad geographical region south of Cyprus. It evolves while changing in shape, merging with other eddies like the NSE during fall 2016. Drifters were captured by it between February and May 2017 allowing a quantitative study of its kinematics. The mean radius of the eddy is about 40–50 km and the maximum speed of about 50 cm/s is reached at 40 km from the eddy center. The CE moves 150 m/day to the southeast. The NSE is present mainly during fall until the beginning of December, when it becomes partially merged with the CE. Then, it disappears until July when it reappears and it moves <sup>1</sup> 2 degree more to the north. The cyclonic SSE in some periods moves far from the coast to the southwest, merging with other cyclones and disappearing from its most probable position for weeks. Other eddies are present in the area like E1, E2 and LE (Figure 3), but their presence and position during the year of the study seems to be very variable and difficult to easily describe.

Besides eddies, the eastern LS is also characterized by a strong upwelling south of Cyprus in the summer months (September 2016 and May–August 2017), and other features occasionally present like the MMJ and the cold front east of Cyprus. The MMJ is present during the whole month of September 2017, crossing the basin from west to east, clearly traced by the cold water in the SST anomaly images, by the drifter tracks and by the geostrophic currents computed from the ADT fields. The jet skims the southern coast of Cyprus and flows toward the Syria/Lebanon coast, where it decreases in strength and it turns north along the coast. The disappearance of the jet corresponds to a wider extension of the CE to the north. The thermal zonal front, spanning from the southeast coast of Cyprus to the Syria/Lebanon coast represents a persistent feature from October to the beginning of March, concomitant with the cooling of the northern part of the basin. The front with a prevailing zonal orientation is first jagged and undefined, then it becomes sharper characterized by a strong current in the warmer side and finally, when the currents weaken, the north intrusions start to degrade the front. The upwelling in front of Israel at the end of November 2016 represents another sporadic event.

The glider campaigns sampled the vertical structures present in the area allowing a general description of a few circulation features in fall 2016 and in winter 2016–2017. The CE is the most sampled eddy during the fall season; it shows the usual thermohaline structure of the area, modulated below the mixed layer (50–100 m), by the downward bending of all the physical parameters measured. However above it, the isopycnals bend upward (Figure 12c,e,g) and in between the up and downward isopycnals, a homogeneous volume of water is present. This kind of eddies has been described as anticyclonic modewater eddies or submesoscale coherent vortices in different basins of the world [46–50]. The CE has a strong signature in ADT, but usually these eddies are also characterized by a negative SST anomaly, that was not seen at least during fall and winter. In summer, instead, the signature in SST anomaly was evident for three full months, but it was not supported by any glider measurement.

The existence of a high salinity core in the CE area was previously reported in studies implying CTD and XCTS surveys [3,25,30,51–53]. More recently, this lens was also described in glider missions, which took place in late fall and at the beginning of winter during the years 2009, 2010 and 2011–2012 [54]. However, in these earlier glider observations, the maximum of the salinity (around 39.3 PSU) showed higher values and lower temperature (around 17 ◦C) with respect to the LIW. In the present study, the core has the same salinity and a higher temperature (around 18.5 ◦C) than the LIW. A minimum in oxygen is evident in the lower part of the core; it is probably related to biological processes as those indicated in the Tropical North Atlantic by [48,55] and in the Mediterranean Sea by [56].

The CE shows a north-south asymmetry with steeper isolines in the north, where the bathymetry is shallower (Figure 13). This can be due to an interaction between the CE and the bathymetry as already shown in previous studies [13,30,36,57]. The shallow bathymetry is probably responsible for the strengthening of the geostrophic currents in the northern part of the CE, also underlined by the ADT maps analyzed in this study and the results described in [4].

Based on the obtained results, it is difficult to identify a specific process involved in the generation of the high salinity core found in the CE region. We can only speculate on the possible generation mechanisms. The shallower bathymetry in the northern part of the CE region, leads to a rising of the high salinity maximum of the LIW during summer or fall. The interaction of this layer with the bathymetry could generate instabilities that induces the creation of a nucleus of high salinity. Another origin of the core thermohaline properties could be related to the water mass created during winter mixing, as mentioned in [25,56]. This is consistent with the results of [58,59], who showed that the CE is able to maintain its core temperature and salinity characteristics for periods of up to two years. However further understanding of the local mechanisms of water mass exchanges and mixing processes is needed. This will be the subject of future studies.

**Author Contributions:** Conceptualization, E.M. and P.-M.P.; Data curation, L.S., R.G. and D.H.; Formal analysis, E.M. and L.S.; Funding acquisition, P.-M.P.; Investigation, D.H. and H.G.; Validation, L.S. and R.G.; Writing—original draft, E.M., L.S.; Writing—review & editing, L.S., R.G. and P.-M.P.

**Funding:** This study is part of the CINEL project sponsored by the U.S. Office of Naval Research (ONR) under grant N000141512459.

**Acknowledgments:** We thank the following persons who have helped with the sea-going operations and the processing of the data: Antonio Bussani, Milena Menna, Piero Zupelli, Massimo Pacciaroni, Stefano Kuchler and Aya Hozumi.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 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 (http://creativecommons.org/licenses/by/4.0/).
