Turbidity effects on the Aegean sea surface properties using numerical simulations

Abstract

This study examines the impact of different turbidity products on the Aegean Sea surface physical characteristics, by performing twin-experiment simulations using a high-resolution regional ocean model. The turbidity products used include an in-situ based diffuse attenuation coefficient dataset at 490 nm (k_d490, in m^− 1) and a satellite derived k_d490 product. Satellite turbidity products are broadly used in ocean simulations due to their spatiotemporal coverage and algorithm universality. Their validation and empirical components are trained mainly in phytoplankton driven regions and this may cause systematic differences in oligotrophic areas of variable optical properties’ composition. In the Aegean Sea, the in-situ based turbidity product accounts for the contribution of suspended particles in the solar heating profile, having further implications in the surface characteristics. The Aegean Sea upper-ocean thermohaline characteristics and general circulation patterns, reveal distinct differences between the twin-experiment simulations, showcasing mesoscale to locally induced impact of the turbidity variations. The turbidity impact on the air-sea interaction fluxes affects both thermodynamic processes i.e., solar radiation penetration and absorption in the water column, as well as dynamic processes i.e., momentum fluxes due to changes of the sea surface temperature and subsequently to the momentum drag coefficient. The Aegean Sea surface characteristics in the in-situ based turbidity product simulation, show a stronger decoupling between the North and the South Aegean Sea, when compared with the satellite derived turbidity product simulation. These results highlight the importance of incorporating more realistic turbidity products in ocean models, especially for optically complex regions such as the Aegean Sea.

Introduction

The Aegean Sea is considered an oligotrophic region (Siokou-Frangou et al. 2010), that receives turbidity inputs of terrestrial, riverine, and atmospheric origin (Karageorgis et al. 2022). Extensive research has documented its turbidity’s spatiotemporal distribution in relation to hydrographic conditions and the basin’s trophic regime (Durrie de Madron et al. 1992; Ignatiades 1998; Zervakis et al. 2005; Karageorgis et al. 2008, 2012, 2017, 2022; Banks et al. 2020; Chaikalis et al. 2021). However, the impact of these turbidity variations on the Aegean’s dynamics remains unexplored.

The ocean’s surface turbidity plays a crucial role in regulating the solar radiation distribution and absorption within the upper layer of the water column, affecting the surface heating processes. Consequently, turbidity is a parameter affecting the surface layer’s dynamics and the surface forcing in ocean general circulation models (OGCMs). Its accurate representation is vital in comprehending complex oceanic processes (Rochford et al. 2001; Murtugudde et al. 2002; Liu et al. 2021).

Previous modeling studies focusing on the solar absorption impact in the upper ocean, primarily use global and regional configurations where turbidity variations are dominated by chlorophyll-a (a proxy for phytoplankton biomass). These studies mainly explore the effect of solar radiation on the heat budget, the ocean stratification, and its implication on major circulation patterns. They noted that increasing turbidity limits light penetration into the water column, altering the surface temperature locally (Nakamoto et al. 2000; Strutton and Chavez 2004; Wu et al. 2007), as well as remotely through a circulation feedback (Nakamoto et al. 2001; Sweeney et al. 2005; Lin et al. 2007; Subrahmanyam et al. 2008; Zhou et al. 2015).

In regions like the Aegean Sea, where colored dissolved organic matter (CDOM), suspended particles, and other turbidity components (e.g., detritus, inorganic matter, dissolved minerals), aside from chlorophyll-a, significantly impact the optical properties (Drakopoulos et al. 2015; Karageorgis et al. 2017), the most representative turbidity metric is the diffuse attenuation coefficient at the blue part of the spectrum, k_d490 (Lee et al. 2005; Mallick et al. 2019). Although satellite turbidity products are broadly used in ocean simulations due to their spatiotemporal coverage and their algorithms’ universality, their empirical components and their validation performed in phytoplankton-driven regions could arise systematic differences in oligotrophic areas of variable optical composition (IOCCG 2020). In oligotrophic seas that receive particle inputs from terrestrial or aeolian sources, satellite turbidity products sometimes misattribute greener reflectance to chlorophyll-a concentration instead of accounting for optical variations caused by other suspended substances (Claustre et al. 2002; Morel et al. 2007; Morel and Gentili 2009; Organelli et al. 2017; IOCCG 2020). Consequently, the aforementioned differences pose a significant challenge in accurately representing these regions (Morel and Gentili 2009; Organelli et al. 2017; IOCCG 2020). One way to achieve their more realistic representation in ocean models would be the utilization of in-situ based turbidity products. Such a product has been developed for the Aegean Sea broader region (Metheniti et al. 2023), using a synthetic dataset of in-situ and satellite measurements. By incorporating a realistic turbidity field, and thereby solar absorption and scattering, modeling simulations can provide valuable insights into the mechanisms governing the ocean’s physical properties and dynamic processes (Kim et al. 2018).

In this paper, the impact of turbidity on the Aegean Seas’s surface circulation and physical characteristics is examined through twin-experiment simulations, using the in-situ based turbidity forcing dataset by (Metheniti et al. 2023) and a satellite-based product. Section 2 describes the regional settings and the model configuration, the initial and boundary conditions, the twin-experiment simulations, and the methods employed for result assessment. Section 3, presents the experiment results, focusing on turbidity effects on the surface physical and dynamical characteristics of the Aegean Sea. The discussion and conclusions are presented in Sect. 4.

Methods and materials

Regional settings

The Aegean Sea can be divided into three major sub-basins according to their hydrographic characteristics (Vervatis et al. 2011): (a) the North Aegean Sea, which includes the Sporades, Athos, and Limnos basins along with Thermaikos Gulf, Samothraki and Limnos shelf areas, (b) the Central Aegean, that includes the Skyros and Chios basins, and (c) the South Aegean, consisting of the Cyclades plateau, and the Myrtoan and Cretan Seas (Fig. 1). The general surface circulation of the Aegean is cyclonic in the largest part of the basin and anticyclonic in the northeastern part of the Aegean Sea, consisting of various basin-scale transient features and permanent gyres (Olson et al. 2007; Vervatis et al. 2013, 2014; Ruti et al. 2016; Mavropoulou et al. 2016, 2022; Velaoras et al. 2021). The main buoyancy contributor is the lower salinity Black Sea Water (BSW) inflow through the Dardanelle straits, along with fresher riverine waters (Fig. 1), creating a north-to-south buoyancy gradient when encountering the Levantine Surface Waters (LSW) (Zervakis and Georgopoulos 2002; Olson et al. 2007; Vervatis et al. 2011; Velaoras et al. 2021). Furthermore, the Aegean Sea is affected by the seasonally variable wind forcing, which along with the varying topography, contributes further to the general circulation patterns and the mesoscale activity of the basin (Poulos et al. 1997; Zervakis and Georgopoulos 2002; Kourafalou and Tsiaras 2007; Sylaios 2011).

Primary and secondary production, riverine discharge, and atmospheric and anthropogenic deposition of inorganic matter create a surface turbidity pattern with a noticeable north-to-south gradient (Karageorgis et al. 2022). The major turbidity sources are located mainly on the North Aegean, where primary production is locally induced by the BSW inflow, along with suspended particles and CDOM from river discharges (Drakopoulos et al. 2015; Karageorgis et al. 2017), while their dispersion is influenced by the hydrographic conditions of the domain (Durrie de Madron et al. 1992; Karageorgis et al. 2008). Saharan dust deposition on the basin is also a turbidity contributor, affecting seawater productivity (Tsagkaraki et al. 2020).

Model and input description

A high-resolution regional model was set up based on the MED36 configuration of (Mavropoulou et al. 2022), enclosing the Aegean Sea in the region 22.29–28.98 ^oE, 34.05–41.16 ^oN (Fig. 1). The ocean model is based on the NEMO v3.6 platform (Nucleus for European Modelling of the Ocean version 3.6; (Madec 2017), and it is 1:3 subset of the global operational system PSY4V3R1(Lellouche et al. 2018). The model is discretized in an Arakawa-C grid, with the horizontal resolution of the configuration at 1/36^o and 75 unevenly spaced vertical levels using a partial-z step, starting from 1 m resolution at the surface and reaching 250 m resolution at the bottom layer. Further details on the setup and physical parameters of the model are described in (Mavropoulou et al. 2022).

The European Center for Medium-Range Weather Forecasts (ECMWF) three-hourly reanalysis dataset (ERA5; https://www.ecmwf.int/en/forecasts/dataset/ecmwf-reanalysis-v5) was used to generate surface forcing with the CORE bulk formula (Large and Yeager 2004). The wind stress and turbulent heat fluxes (i.e., latent and sensible ) are parametrized in terms of near surface atmospheric and ocean state variables:

where the air density, the wind speed at 10 m height, the ocean surface current velocity, the latent heat of vaporization, the specific heat of air, and the air temperature and specific humidity at 2 m height, and the Sea Surface Temperature and the specific humidity of air assumed to be saturated at the ocean surface. The complete description of the air-sea fluxes parameterization includes the bulk coefficients for momentum drag , evaporation and sensible heat transfer . The transfer coefficients are estimated from Monin-Obukhov similarity theory and parametrized as functions of sea surface roughness and atmospheric stability i.e., and , where the von Kármán constant, the Obukhov length and the similarity function for momentum (e.g., in neutral conditions approaches zero ).

The model has four open boundaries, connecting to the adjacent basins, three on the south, west, and east borders, and one on the Dardanelle strait entrance to the Aegean, where the BSW inflow takes place. The west, south, and eastern monthly mean open boundary data, along with the initialization data (sea surface height, temperature, salinity, and current velocity components) were derived from the Copernicus Marine Environment Monitoring Service (CMEMS; (Simoncelli et al. 2019).

The Dardanelles Strait is explicitly introduced in the model as an open boundary, following Mavropoulou et al. (2022). The Dardanelles open boundary data contain monthly climatological values for the whole ocean state vector (i.e., sea surface height, temperature, salinity, current velocity components), obtained from the global operational system PSY4V3R1. The model uses tidal forcing at the open boundaries on the east, west and south provided by TPXO 7.1 global tide model(Egbert and Erofeeva 2002)as the sum of 11 constituents (M2, S2, K2, N2, K1, O1, P1, Q1, M4, Mf, Mm). We also use the same validated model set up as (Mavropoulou et al. 2022) for the river run-offs of Axios and Evros, adapted for the region by (Vervatis et al. 2013, 2014). The present NEMO configuration is further validated against satellite Sea Surface Temperature (SST data) from CMEMS database (confer supplementary material).

[See PDF for image]

Fig. 1

Aegean Sea bathymetric map (depth in m), corresponding to the configuration’s domain. Important topographical features are numbered as follows: Sporades basin (1), Athos basin (2), Limnos basin (3), Skyros basin (4), Chios basin (5). The rivers’ input are those of Axios and Evros as located in the Figure. The black lines denote the sections dividing the domain in North, Central and South Aegean. The section for the southward exchange is denoted by a red star

Solar penetration parameterization

The solar radiation penetration is parameterized through the downward irradiance (E_d) which is added as an additional term in the time evolution equation of temperature. This configuration uses a 2-waveband light penetration scheme (Paulson and Simpson 1977):,

Where z refers to the depth, Q_sr is the penetrative part of the solar radiation, R = 0.54, is the fraction of Q_sr for the non-penetrative wavebands longer than 700 nm and ξ₀ = 0.35 m, is an e-folding length scale, corresponding to Type I water (Jerlov 1978) and the second extinction length scale is associated with the shorter wavelengths, which is chosen to be spatially varying i.e., ξ₁ = 1/k_d490.

Twin-model sensitivity experiments

The first experiment was conducted using a reconstructed dataset created from in-situ and remotely sensed optical measurements, in the region of the Aegean Sea (Metheniti et al. 2023), from now on referenced as KDAEG experiment (Fig. 2a). The input of the second experiment was a multi-year mean field, generated using the k_d490 variable extracted from the ESA-CCI v5.0 database (Sathyendranath et al. 2021; https://climate.esa.int/en/projects/ocean-colour/), hereinafter referred as KDSAT experiment (Fig. 2b).

The two experiments start from the same initial conditions and the simulation is carried out for 5 years, i.e., 1997–2001, which corresponds to a period of continuous in-situ optical measurements, utilized for the creation of the KDAEG experiment’s dataset (Metheniti et al. 2023). The model produced daily outputs for the ocean state vector (SSH, T, S, U, V) and other diagnostic variables. The analysis is based on the mean annual results of the fifth (last) year of the model integration to avoid spin-up trends in the model’s physical and dynamical properties. The analysis is performed for the Aegean Sea domain, bounded by the Cretan arc straits, and a Lanczos filter with 20 grid step filtering is used for spatially smoothing the resulting maps. The filter was applied to eliminate noise while preserving the domain’s mesoscale features. The assessment of the different results is focused on the individual patterns located in the sub-basins, based on the surface physical characteristics, the mesoscale eddy activity, and the main circulation patterns.

The KDAEG experiment’s k_d490 field has a sharper north-south gradient than the KDSAT experiment’s field (Fig. 2c). Observing their surface k_d490 differences (Δk_d490) (Fig. 2d), the North Aegean exhibits greater variability and more turbid waters (positive median), while the inter-quartile range is smaller in the less turbid (negative median) in the Central and South Aegean. The higher turbidity areas of the North Aegean can be explained by the local primary production, the presence of the BSW in the area (Lykousis et al. 2002; Varkitzi et al. 2020), and the CDOM excess due to outflow transfer from rivers and BSW (Drakopoulos et al. 2015). Exceptions are the very nearshore regions (approximately depth < 40 m) and the eastern Aegean coasts where the KDSAT experiment field is more turbid due to the lack of spatial coverage in the in-situ dataset. Additionally, the region northeast of the Cretan island is more turbid in the KDAEG experiment, probably due to the timing of the measurement conduction (Karageorgis et al. 2008).

[See PDF for image]

Fig. 2

(a) Spatial distribution of the k_d490 used in the KDAEG and (b) the KDSAT experiments, (c) and their difference, and (d) boxplots of k_d490 difference’s corresponding distribution on the North, Central, and South Aegean. The box encompasses the interquartile range, where the mid-line represents the median, and the upper and lower vertical lines the third and fourth quartiles respectively. Outliers based on the interquartile range method, comprise 9% of the dataset and are excluded

Results

Sea surface temperature and salinity

[See PDF for image]

Fig. 3

(a) Boxplot for the distribution of the ΔSST (in ^oC) between the KDAEG and KDSAT experiments, for the North, Central and South Aegean. Outliers, comprising 3% of the dataset, are excluded. (b) The spatial distribution of ΔSST

[See PDF for image]

Fig. 4

Same as Fig. 3, for ΔSSS (in psu). Outliers, comprising 4% of the dataset, are excluded in (a)

The results of the KDAEG and KDSAT experiments are compared, according to their surface differences. The main pattern of the simulated physical properties and general circulation in both experiments agree with observational (Zervakis and Georgopoulos 2002; Olson et al. 2007; Vervatis et al. 2011) and numerical studies (Vervatis et al. 2013; Mavropoulou et al. 2016; Mamoutos et al. 2021). Briefly, the surface circulation is cyclonic, and a north-to-south temperature and salinity gradient is reproduced, with cooler and fresher waters in the North and warmer and saltier in the South. The differences mentioned hereinafter refer to the comparison of the KDAEG to the KDSAT experiment results.

The effect of the turbidity induced surface heating can be quantified by defining the Sea Surface Temperature (SST) difference (ΔSST) between the two experiments. ΔSST (Fig. 3a) and Δk_d490 (Fig. 2d), compared for each sub-basin, show similar variability range. Specifically, in the North Aegean, both Δk_d490 and ΔSST display a similar dispersion with a slightly positive median that corresponds to warmer and more turbid surface waters. In the Central and South Aegean, the range is smaller compared to the North, and the medians are negative indicating cooler and less turbid surface waters, a result that can be attributed to deeper light penetration in the water column. Overall, enhanced turbidity restricts the light penetration in a shallower surface layer, strengthening the stratification and, consequently, leading to higher SST (Nakamoto et al. 2000, 2001; Wu et al. 2007; Mallick et al. 2019). Conversely, reduced turbidity allows solar radiation to penetrate deeper, leading to cooler SST.

However, inconsistencies regarding this local mechanism arise when comparing the spatial distribution of ΔSST (Fig. 3b) to that of Δk_d490 (Fig. 2c). While there are regions like the coastal areas and the vicinity of the North Aegean continental shelf, where the cooling/warming corresponds to the decrease/enhancement of turbidity (local effect), other regions indicate the influence of three-dimensional dynamical mechanisms (dynamic effect). This effect will be discussed in the next subsection.

To further investigate the act of turbidity on the surface physical properties, the response of the Sea Surface Salinity (SSS) field is also explored. Comparing the SSS difference (ΔSSS) computed for each sub-basin (Fig. 4a, b) to the Δk_d490 (Fig. 2c, d), we observe a large negative (less saline) pattern in the North and Central Aegean, while the South Aegean appears overall the same for the two experiments with minor differences. This can be directly attributed to the BSW salinity distribution and its surface circulation and spreading (Vervatis et al. 2011; Androulidakis et al. 2012; Mavropoulou et al. 2016). Both SST and SSS distribution indicate that the two configurations lead to differences in surface circulation pattern and features that will be presented in the next subsection.

Surface circulation features

[See PDF for image]

Fig. 5

(a) Spatial distribution of the surface vorticity field of the KDAEG experiment and (b) the KDSAT experiment (hashed lines represent positive values). (c) Their difference (in s-1)

A rough comparison of the general circulation pattern of the Aegean Sea (Fig. 5a, b) shows a general similarity, which is expected since the only difference of the two experiments is related to the solar radiation absorption. But if we take a closer look in the differences of the various features (Fig. 5c) we can recognize important differences in their strength, location, and even the vorticity sign (cyclonic/anticyclonic).

Figure 5a-c shows a much stronger and more pronounced Samothraki anticyclone (around 40-40.5^ο N, 25-25.6^ο E) in KDAEG experiment. This is a permanent feature of the Aegean Sea, in agreement with previous surveys (Olson et al. 2007; Skliris et al. 2010) and numerical simulations (Sofianos et al. 2005; Androulidakis et al. 2012; Vervatis et al. 2014), and is linked to the northwestward intrusion of BSW in the North Aegean Sea. The strengthening of the Samothraki anticyclone appears to enhance further the northward propagation of the BSW towards the Limnos basin explaining also the lower values of SSS in the region (Fig. 4b).

Another circulation feature of the North Aegean, located over the Sporades basin, is the Sporades eddy (around 39–40^ο N, 23.5–24.5^ο E in Fig. 5). The vorticity sign changes between the two experiments, rotating in a cyclonic manner on the KDSAT experiment and anticyclonic on the KDAEG experiment. The alteration of the Sporades eddy rotation has been observed also in previous studies and is linked to the intensity of the westward propagation of the BSW (Kontoyiannis et al. 2003; Olson et al. 2007), where low salinity waters result in anticyclonic eddy rotation, as it is the case in the KDAEG experiment (Fig. 4b). The Sporades anticyclonic eddy in the KDAEG experiment appears to trap the spreading of BSW in the northwestern Aegean Sea and Thermaikos Gulf, and as a result the salinity of the gulf shows increased values. Other mesoscale features in the North Aegean e.g., the eddies over the Athos basin, are also enhanced (around 39.5–40^ο N, 24.5–25^ο E in Fig. 5).

In the Central Aegean, the most prominent feature in both simulations is the Chios gyre (around 37.5–385^ο N, 25–26^ο E in Fig. 5), which is a permanent cyclone in the Chios basin (Olson et al. 2007). In the KDAEG experiment, this formation is confined to the southern part of the Chios basin, whereas in the KDSAT experiment extends over the entire basin. The position and strength of the Chios gyre are affected by the southward propagation of the North Aegean surface waters along the Evoia coast, as well as the northward propagation of LSW along the Asia Minor coast (Olson et al. 2007). Changes in the location of the Chios cyclone between experiments, can explain the opposing patterns of ΔSST (Fig. 3b), and ΔSSS (Fig. 4b) in the region. The core of positive ΔSST values (negative ΔSSS) corresponds to the location of the Chios gyre in the KDSAT experiment, in which upwelling of cool and saline waters from the deep layers reach the surface.

In the South Aegean, the circulation differences between the two experiments are minor and mostly refer to the permanent anticyclone in the western Cretan Sea (around 35.5–36.5^ο N, 24.5–25.5^ο E in Fig. 5), being shifted eastward in the KDAEG experiment.

Overall, we have identified that the enhanced anticyclonic activity and the reduced SSS field between experiments are connected to each other in the North Aegean and affected by the atmospheric forcing. Sofianos et al. (2005) proposed that the anticyclonic circulation of the North Aegean is primarily regulated by the wind stress curl, pertaining to the fact that the intrusion of the BSW in the basin is subject to surface circulation driven by the atmosphere (Kontoyiannis et al. 2003; Mavropoulou et al. 2016). To examine the cause of these mesoscale circulation differences and their link to turbidity, the wind forcing of the two experiments was assessed in the following section.

The act of wind stress on the surface properties and its link to turbidity

[See PDF for image]

Fig. 6

Spatial distribution of the difference in wind stress (in Nm^-2) between KDAEG and KDSAT experiments, with positive values indicating increased wind stress and negative reduced wind stress

[See PDF for image]

Fig. 7

The KDAEG experiment’s surface wind stress vorticity spatial distribution field (in Nm^-3), provided as a reference for both experiments. Negative values indicate anticyclonic and positive cyclonic wind stress curl

The two experiments use the same input of atmospheric forcing, suggesting that the wind fields remain unchanged. However, changes are brought in the air-sea momentum fluxes through changes in the momentum drag coefficient applied in the bulk formula for the estimation of the wind stress. The drag coefficient depends on the SST model prognostic variable (Large and Yeager 2004) through the atmospheric stability criteria close to the sea surface (Sect. 2.2). Thus, differences in the SST field produce differences in the wind stress fields, as shown in Fig. 6.

The wind stress curl is negative (i.e., anticyclonic) in the North Aegean and positive (i.e., cyclonic) elsewhere (Fig. 7), which is in line with the findings from (Sofianos et al. 2005). In the northeastern Aegean, the wind stress field (Fig. 6) increases, leading to strengthening of the anticyclonic wind stress curl (i.e., more negative). This fact appears to reinforce the sea surface anticyclonic activity confirmed by the vorticity field in both experiments (Fig. 5), further affecting the surface hydrographic properties of the Aegean. In the KDAEG experiment the increased turbidity on the North Aegean explains the enhanced SST locally, which in turn, affects the stability of the air-sea interaction layer and the momentum drag coefficient, subsequently increasing the wind stress and its curl negativity. This appears to intensify and perhaps shift the location of the mesoscale anticyclonic features of the region, imposing changes in the Aegean’s hydrographic properties.

The impact of SST on surface heat fluxes

As detailed in Sect. 2.2, the SST has a direct impact on the air-sea fluxes themselves in the Large and Yeager (2004) bulk scheme, as well as on the momentum drag coefficient, the heat transfer coefficients and evaporation via stability functions. In general, the Aegean Sea experiences heat loss at the surface in both experiments (Fig. 8a-b), except for some regions in the southeastern coastal Aegean Sea. The differences between experiments show spatial variability in Fig. 8c with a similar pattern to the SST differences (Fig. 3b); negative (positive) values represent increased (reduced) heat loss in the KDAEG relative to the KDSAT.

[See PDF for image]

Fig. 8

(a) Spatial Distribution of the net heat flux, where negative values indicate heat loss, and positive values indicate heat gain, for (a) the KDAEG and (b) the KDSAT experiments. (c) Their difference (KDAEG-KDSAT), where negative values indicate increased heat loss, and positive reduced heat loss (except a coastal region in the southeastern Aegean, where positive values indicate increased heat gain)

Pertaining to the fact that, the shortwave radiation at the ocean surface is identical in both experiments, the following remark can be made to elucidate on the modifications of air-sea exchanges. The net ocean heat loss (Fig. 8a-b) and its differences between the two experiments (Fig. 8c), are mainly attributed to the latent heat flux, whilst the sensible and longwave radiation have a moderate contribution (Fig. 9a-c). The source of those differences is the modifications observed in SST between experiments (Fig. 3b), where in regions with increased (decreased) SST the evaporation and latent heat fluxes are also increased (decreased). The annual mean net heat flux is negative in both experiments, in agreement with findings from previous studies (Poulos et al. 1997; Vervatis et al. 2013, 2014), calculated at about − 31.6 Wm^− 2 and − 29.9 Wm^− 2 for the KDESA and KDAEG respectively.

[See PDF for image]

Fig. 9

The KDAEG to KDSAT difference, for a) the longwave radiation flux, (b) the sensible heat flux and (c) the latent heat flux. Negative values indicate increased heat loss and positive values indicate reduced heat loss

Figure 10 shows large seasonal variability for the turbulent fluxes, with net heat loss (gain) over the winter (summer) period. A seasonal signal is also observed when comparing the differences of the heat fluxes between the two experiments. The KDAEG simulation exhibits greater heat loss over winter in comparison to KDSAT, as a result of warmer surface waters and increased stratification. The opposite situation is observed during summer and the two models self-equilibrate to different states. The main mechanism regulating the seasonal variability of the Aegean Sea heat budget, is the modifications brought in SST by the different turbidity products (Fig. 10i, dashed blue line), in agreement also with results discussed for the annual maps in Figs. 8 and 9. Overall, SST increase (decrease) leads to increase (decrease) of evaporation and latent heat loss controlling the net heat flux in the Aegean. The temporal variability of the net heat flux difference is small (Fig. 8i) within the range of the bulk model errors themselves, perhaps considered as a secondary effect for the aforementioned mechanisms.

[See PDF for image]

Fig. 10

Timeseries of the turbulent and net heat fluxes (W/m2) for the twin experiments (a-c) KDAEG and (d-f) KDSAT, and (g-i) their differences KDAEG-KDSAT. (i) ΔSST (dashed blue line; ^oC)

Discussion and conclusions

The objective of this study is to assess the impact of a realistic turbidity field on the surface characteristics of the Aegean Sea, using a high-resolution regional ocean model and investigating the sensitivity to heating by solar irradiance. Twin-experiment simulations are conducted using two different, spatially varying turbidity datasets: (a) the so-called “KDAEG” simulation experiment, using an annual k_d490 field from in-situ measurements (in combination with remotely sensed optical products for the Aegean Sea; (Metheniti et al. 2023)) (Fig. 2a), and (b) the “KDSAT” experiment using a multi-year mean product for a k_d490 field derived from the ESA-CCI v5.0 database (Fig. 2b). Both simulations spanned a period of 5 years (1997–2001) and our analysis focus on the annual fields of the final year model outputs.

The use of the in-situ based turbidity product, as opposed to the simulation using the satellite product, reveals that the sea surface physical properties of SST and SSS, as well as the circulation patterns in the Aegean Sea, are sensitive to turbidity spatial variations. Specifically, two primary mechanisms are identified.

The first, is a direct feedback mechanism between the turbidity and the upper-ocean temperature field, in particular along the coastal areas and the North Aegean continental shelf, where the contribution of CDOM in turbidity is prominent. This relationship is consistent with previous studies (Nakamoto et al. 2000; Kara et al. 2005; Löptien and Meier 2011) and is attributed to changes in stratification due to light absorption in the upper layer. Higher (lower) turbidity (Fig. 2c-d) leads to increased (decreased) stratification and subsequently to higher (lower) SST (Fig. 3).

The second feedback mechanism is attributed to changes in the Aegean Sea circulation controlled by changes in SST and subsequently in momentum atmospheric fluxes. The KDAEG simulation experiment uses a more turbid product for the Northeastern Aegean Sea, compared to the KDSAT simulation experiment and therefore, shows elevated values of SST, that in turn affect the air-sea interaction momentum fluxes, the stability conditions and the momentum drag coefficient, increasing the strength of the wind stress (Fig. 6). As a result, the wind stress curl (Fig. 7) becomes more negative and strengthens the anticyclonic circulation of the North Aegean (Fig. 5). This finding is in agreement with (Sofianos et al. 2005), showcasing the impact of the negative wind stress curl to the region’s anticyclonic circulation. Furthermore, the anticyclonic circulation of the North Aegean Sea increases the residence time of the BSW, with the Samothraki anticyclone playing a pivotal role in recirculating the BSW in the basin (Zervakis and Georgopoulos 2002). The intensified anticyclonic features of the KDAEG experiment, assisted by changes in the North Aegean SSS patterns shown in Fig. 4b (i.e., negative in the North, positive in the South), indicate a prolonged circulation period for the fresh BSW in the North and Central Aegean basins. To quantify the intensity of the meridional exchange between the North-Central Aegean and the South Aegean Sea, the residence time for the southward exchange is estimated as:,

where V is the volume of the basin north of the border section in Central Aegean (red star in Fig. 1) estimated approximately at 28.4 10³km³, including the additional volume resulting from the sea surface height variations for each experiment, Q_out is the southward flux of water through the section, and E is the evaporation of water on the area above the section. T_R is estimated at approximately 895 days for the KDAEG experiment and 782 days for the KDSAT experiment, the order of magnitude being analogous to (Mamoutos et al. 2017), considering also differences in size between domains, the simulation period, and the inclusion of data assimilation and nature of forcing applied to the twin experiment. The prolonged period as indicated by the T_R values of the North-Central Aegean, confirms the decoupling mechanism between the Aegean sub-basins, resulting to more than 100 days (or approximately 12%) longer for a water parcel to exit the North-Central towards the South Aegean, in the KDAEG compared to the KDSAT experiment. This time period difference between experiments corresponds also to the seasonal cycle of the general circulation and physical properties of the Aegean Sea (Poulos et al. 1997; Olson et al. 2007; Vervatis et al. 2014), hindering possible exchanges that could impact Aegean stratification, mixing processes, deep water formation events, nutrient distribution, and regional weather patterns, to name a few.

The impact of SST is extended beyond momentum also to air-sea heat fluxes, consistent with stability functions incorporated in bulk schemes. The latter is of importance in this work, since an unstable boundary layer due to increased SST leads to larger evaporative heat loss and the system self-equilibrate to another state, depending on which turbidity product is used. In the Aegean Sea, the net heat flux exhibits strong seasonal variability, with heat loss (gain) over winter (summer), and an annual mean net heat loss, in agreement with previous studies (Poulos et al. 1997; Vervatis et al. 2013, 2014). In our twin experiments the evaporation and latent heat flux are the dominant sources controlling the Aegean heat budget, ensuing modifications in buoyancy production on vertical mixing and stratification, with possible climatic implications on the overturning circulation of the region.

This work underscores the critical need for a more realistic representation of turbidity forcing in ocean simulations. Minor discrepancies between turbidity products, driven by factors such as CDOM and other suspended particles derived by in-situ measurements, can have a substantial impact on regions such as the Aegean Sea, having a diverse optical regime. Therefore, it is crucial to thoroughly characterize turbidity products, because dissolved and particulate components interact differently with solar irradiance, leading to various modelling methods on heat redistribution in the water column. Future research could focus on investigating the aforementioned impacts over longer (decadal/climatic) simulations, by using updated turbidity products, especially given the increasing frequency of extreme weather events (Nastos and Zerefos 2007; Ganor et al. 2010; Pakalidou and Karacosta 2018; Cardell et al. 2020). In this context, there is a potential to alter the quantity and quality of optically significant constituents affecting the ocean forcing, with further implications on ecosystem dynamics, marine biodiversity, and other crucial environmental processes.

Acknowledgements

This work was supported by computational time granted from the National Infrastructures for Research and Technology S.A. (GRNET S.A.) in the National HPC facility – ARIS (https://hpc.grnet.gr/) - under project ID pa210603 (AEG36) and project ID pr013013 (BOFCOAM). We thank two anonymous reviewers for their constructive comments.

Author contributions

V.M: Sofware, Methodology, Conceptualization, Formal Analysis, Investigation, Visualizaiton, Writing - Original Draft, Writing - Original Draft, Writing - Review & Editing.V.V: Conceptualization, Software, Methodology, Formal Analysis, Writing - Review & EditingN.K: Resources, Formal AnalysisS.S: Supervision, Conceptualization, Methodology, Formal Analysis.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

Data will be made available upon request.

Declarations

Competing Interests

The authors have no relevant financial or non-financial interests to disclose.

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

Androulidakis, YS; Krestenitis, YN; Kourafalou, VH. Connectivity of North Aegean circulation to the Black Sea water budget. Cont Shelf Res; 2012; 48, pp. 8-26. [DOI: https://dx.doi.org/10.1016/j.csr.2012.08.019]

Banks AC, Drakopoulos PG, Chaikalis S et al (2020) An in situ optical dataset for working towards fiducial reference measurements based satellite ocean colour validation in the Eastern Mediterranean. 63. https://doi.org/10.1117/12.2571193

Cardell, MF; Amengual, A; Romero, R; Ramis, C. Future extremes of temperature and precipitation in Europe derived from a combination of dynamical and statistical approaches. Int J Climatol; 2020; 40, pp. 4800-4827. [DOI: https://dx.doi.org/10.1002/joc.6490]

Chaikalis S, Parinos C, Möbius J Optical Properties and Biochemical Indices of Marine Particles in the Open Mediterranean Sea: The R/V, Maria S, Merian, Cruise et al (2021) March 2018. Front Earth Sci (Lausanne) 9:695. https://doi.org/10.3389/feart.2021.614703

Claustre H, Morel A, Hooker SB et al (2002) Is desert dust making oligotrophic waters greener? Geophys Res Lett 29. https://doi.org/10.1029/2001gl014056. 107-1-107–4

Drakopoulos PG, Banks AC, Kakagiannis G et al (2015) Estimating chlorophyll concentrations in the optically complex waters of the North Aegean Sea from field and satellite ocean colour measurements. In: Hadjimitsis DG, Themistocleous K, Michaelides S, Papadavid G (eds) Third International Conference on Remote Sensing and Geoinformation of the Environment (RSCy2015). p 95351I

Durrie de Madron, X; Nyffeler, F; Balopoulos, ET; Chronis, G. Circulation and distribution of suspended matter in the Sporades basin (northwestern Aegean sea). J Mar Syst; 1992; 3, pp. 237-248. [DOI: https://dx.doi.org/10.1016/0924-7963(92)90003-Q]

Egbert, GD; Erofeeva, SY. Efficient Inverse Modeling of Barotropic Ocean Tides. J Atmos Ocean Technol; 2002; 19, pp. 183-204. [DOI: https://dx.doi.org/10.1175/1520-0426(2002)019]

Ganor, E; Osetinsky, I; Stupp, A; Alpert, P. Increasing trend of African dust, over 49 years, in the eastern Mediterranean. J Geophys Res; 2010; 115, D07201. [DOI: https://dx.doi.org/10.1029/2009JD012500]

Ignatiades, L. The productive and optical status of the oligotrophic waters of the Southern Aegean Sea (Cretan Sea), Eastern Mediterranean. J Plankton Res; 1998; 20, pp. 985-995. [DOI: https://dx.doi.org/10.1093/plankt/20.5.985]

IOCCG. Synergy between Ocean Colour and Biogeochemical/Ecosystem Models; 2020; Canada, IOCCG, Dartmouth:

Jerlov NG (1978) The optical classification of sea water in the euphotic zone. Københavns universitet, Institut for fysisk oceanografi

Kara, AB; Wallcraft, AJ; Hurlburt, HE. Sea surface temperature sensitivity to water turbidity from simulations of the turbid Black Sea using HYCOM. J Phys Oceanogr; 2005; 35, pp. 33-54. [DOI: https://dx.doi.org/10.1175/JPO-2656.1]

Karageorgis, AP; Gardner, WD; Georgopoulos, D et al. Particle dynamics in the Eastern Mediterranean Sea: A synthesis based on light transmission, PMC, and POC archives (1991–2001). Deep Sea Res Part I; 2008; 55, pp. 177-202. [DOI: https://dx.doi.org/10.1016/J.DSR.2007.11.002]

Karageorgis, AP; Georgopoulos, D; Kanellopoulos, TD et al. Spatial and seasonal variability of particulate matter optical and size properties in the Eastern Mediterranean Sea. J Mar Syst; 2012; 105–108, pp. 123-134. [DOI: https://dx.doi.org/10.1016/j.jmarsys.2012.07.003]

Karageorgis, AP; Drakopoulos, PG; Psarra, S et al. Particle characterization and composition in the NE Aegean Sea: Combining optical methods and biogeochemical parameters. Cont Shelf Res; 2017; 149, pp. 96-111. [DOI: https://dx.doi.org/10.1016/j.csr.2017.03.008]

Karageorgis, AP; Metheniti, VC; Chaikalis, SP et al. Particulate Matter Spatial, Temporal Distribution and Size Properties in the Aegean Sea; 2022; Berlin, Heidelberg, Springer: pp. 1-32. [DOI: https://dx.doi.org/10.1007/698_2022_899]

Kim, GE; Gnanadesikan, A; Del Castillo, CE; Pradal, M-A. Upper Ocean Cooling in a Coupled Climate Model Due to Light Attenuation by Yellowing Materials. Geophys Res Lett; 2018; 45, pp. 6134-6140. [DOI: https://dx.doi.org/10.1029/2018GL077297]

Kontoyiannis H, Kourafalou VH, Papadopoulos V (2003) Seasonal characteristics of the hydrology and circulation in the northwest Aegean Sea (eastern Mediterranean): Observations and modeling. J Geophys Res Oceans 108. https://doi.org/10.1029/2001jc001132

Kourafalou, V; Tsiaras, K. A nested circulation model for the North Aegean Sea. Ocean Sci; 2007; 3, pp. 1-16. [DOI: https://dx.doi.org/10.5194/OS-3-1-2007]

Large WG, Yeager SG (2004) Diurnal to decadal global forcing for ocean and sea-ice models: The data sets and flux climatologies. NCAR Tech Note TN–460 + ST:105pp. https://doi.org/10.5065/D6KK98Q6

Lee, ZP; Du, KP; Arnone, R et al. Penetration of solar radiation in the upper ocean: A numerical model for oceanic and coastal waters. J Geophys Res C: Oceans; 2005; 110, pp. 1-12. [DOI: https://dx.doi.org/10.1029/2004JC002780]

Lellouche, JM; Greiner, E; Le Galloudec, O et al. Recent updates to the Copernicus Marine Service global ocean monitoring and forecasting real-time 1 g 12° high-resolution system. Ocean Sci; 2018; 14, pp. 1093-1126. [DOI: https://dx.doi.org/10.5194/OS-14-1093-2018]

Lin, P; Liu, H; Zhang, X. Sensitivity of the upper ocean temperature and circulation in the equatorial Pacific to solar radiation penetration due to phytoplankton. Adv Atmos Sci; 2007; 24, pp. 765-780. [DOI: https://dx.doi.org/10.1007/s00376-007-0765-7]

Liu Y, He R, Lee Z (2021) Effects of Ocean Optical Properties and Solar Attenuation on the Northwestern Atlantic Ocean Heat Content and Hurricane Intensity. Geophys Res Lett 48. https://doi.org/10.1029/2021GL094171. e2021GL094171

Löptien, U; Meier, HEM. The influence of increasing water turbidity on the sea surface temperature in the Baltic Sea: A model sensitivity study. J Mar Syst; 2011; 88, pp. 323-331. [DOI: https://dx.doi.org/10.1016/j.jmarsys.2011.06.001]

Lykousis, V; Chronis, G; Tselepides, A et al. Major outputs of the recent multidisciplinary biogeochemical researches undertaken in the Aegean Sea. J Mar Syst; 2002; 33–34, pp. 313-334. [DOI: https://dx.doi.org/10.1016/S0924-7963(02)00064-7]

Madec G and team (2017) NEMO ocean engine. In Notes du Pôle de modélisation de l’Institut Pierre-Simon Laplace (IPSL) (v3.6-patch, Number 27). Zenodo. https://doi.org/10.5281/zenodo.3248739

Mallick, SK; Agarwal, N; Sharma, R et al. Impact of Satellite-Derived Diffuse Attenuation Coefficient on Upper Ocean Simulation Using High-Resolution Numerical Ocean Model: Case Study for the Bay of Bengal. Mar Geodesy; 2019; 42, pp. 535-557. [DOI: https://dx.doi.org/10.1080/01490419.2019.1664677]

Mamoutos, I; Zervakis, V; Tragou, E et al. The role of wind-forced coastal upwelling on the thermohaline functioning of the North Aegean Sea. Cont Shelf Res; 2017; 149, pp. 52-68. [DOI: https://dx.doi.org/10.1016/j.csr.2017.05.009]

Mamoutos IG, Potiris E, Tragou E et al (2021) A high-resolution numerical model of the north aegean sea aimed at climatological studies. J Mar Sci Eng 9. https://doi.org/10.3390/jmse9121463

Mavropoulou, AM; Mantziafou, A; Jarosz, E; Sofianos, S. The influence of Black Sea Water inflow and its synoptic time-scale variability in the North Aegean Sea hydrodynamics. Ocean Dyn; 2016; 66, pp. 195-206. [DOI: https://dx.doi.org/10.1007/s10236-016-0923-5]

Mavropoulou AM, Vervatis V, Sofianos S (2022) The Mediterranean Sea overturning circulation: A hindcast simulation (1958–2015) with an eddy-resolving (1/36°) model. Deep Sea Res 1. https://doi.org/10.1016/j.dsr.2022.103846. Oceanogr Res Pap 187:

Metheniti, V; Karageorgis, AP; Drakopoulos, P et al. Deriving the diffuse attenuation coefficient in the Eastern Mediterranean Sea, using observational optical measurements and a multi-layer perceptron regression model. Deep Sea Res Part I; 2023; 199, 104105. [DOI: https://dx.doi.org/10.1016/j.dsr.2023.104105]

Morel, A; Gentili, B. The dissolved yellow substance and the shades of blue in the Mediterranean Sea. Biogeosciences; 2009; 6, pp. 2625-2636. [DOI: https://dx.doi.org/10.5194/BG-6-2625-2009]

Morel, A; Huot, Y; Gentili, B et al. Examining the consistency of products derived from various ocean color sensors in open ocean (Case 1) waters in the perspective of a multi-sensor approach. Remote Sens Environ; 2007; 111, pp. 69-88. [DOI: https://dx.doi.org/10.1016/j.rse.2007.03.012]

Murtugudde, R; Beauchamp, J; McClain, CR et al. Effects of Penetrative Radiation on the Upper Tropical Ocean Circulation. J Clim; 2002; 15, pp. 470-486. [DOI: https://dx.doi.org/10.1175/1520-0442(2002)015]

Nakamoto, S; Kumar, SP; Oberhuber, JM et al. Chlorophyll modulation of sea surface temperature in the Arabian Sea in a mixed-layer isopycnal general circulation model. Geophys Res Lett; 2000; 27, pp. 747-750. [DOI: https://dx.doi.org/10.1029/1999GL002371]

Nakamoto, S; Kumar, SP; Oberhuber, JM et al. Response of the equatorial Pacific to chlorophyll pigment in a mixed layer isopycnal ocean general circulation model. Geophys Res Lett; 2001; 28, pp. 2021-2024. [DOI: https://dx.doi.org/10.1029/2000GL012494]

Nastos, PT; Zerefos, CS. On extreme daily precipitation totals at Athens, Greece. Adv Geosci; 2007; 10, pp. 59-66. [DOI: https://dx.doi.org/10.5194/adgeo-10-59-2007]

Olson, DB; Kourafalou, VH; Johns, WE et al. Aegean surface circulation from a satellite-tracked drifter array. J Phys Oceanogr; 2007; 37, pp. 1898-1917. [DOI: https://dx.doi.org/10.1175/JPO3028.1]

Organelli, E; Claustre, H; Bricaud, A et al. Bio-optical anomalies in the world’s oceans: An investigation on the diffuse attenuation coefficients for downward irradiance derived from Biogeochemical Argo float measurements. J Geophys Res Oceans; 2017; 122, pp. 3543-3564. [DOI: https://dx.doi.org/10.1002/2016JC012629]

Pakalidou, N; Karacosta, P. Study of very long-period extreme precipitation records in Thessaloniki, Greece. Atmos Res; 2018; 208, pp. 106-115. [DOI: https://dx.doi.org/10.1016/j.atmosres.2017.07.029]

Paulson, CA; Simpson, JJ. Irradiance Measurements in the Upper Ocean. J Phys Oceanogr; 1977; 7, pp. 952-956. [DOI: https://dx.doi.org/10.1175/1520-0485(1977)007]

Poulos, SE; Drakopoulos, PG; Collins, MB. Seasonal variability in sea surface oceanographic conditions in the Aegean Sea (Eastern Mediterranean): an overview. J Mar Syst; 1997; 13, pp. 225-244. [DOI: https://dx.doi.org/10.1016/S0924-7963(96)00113-3]

Rochford, PA; Kara, AB; Wallcraft, AJ; Arnone, RA. Importance of solar subsurface heating in ocean general circulation models. J Geophys Res Oceans; 2001; 106, pp. 30923-30938. [DOI: https://dx.doi.org/10.1029/2000jc000355]

Ruti, PM; Somot, S; Giorgi, F et al. Med-CORDEX initiative for Mediterranean climate studies. Bull Am Meteorol Soc; 2016; 97, pp. 1187-1208. [DOI: https://dx.doi.org/10.1175/BAMS-D-14-00176.1]

Sathyendranath S, Jackson T, Brockmann C et al (2021) ESA Ocean Colour Climate Change Initiative (Ocean_Colour_cci): Version 5.0 Data. In: NERC EDS Centre for Environmental Data Analysis [dataset]. https://catalogue.ceda.ac.uk/uuid/1dbe7a109c0244aaad713e078fd3059a. Accessed 1 Apr 2022

Simoncelli S, Fratianni C, Pinardi N et al (2019) Mediterranean Sea Physical Reanalysis (CMEMS MED-Physics) (Version 1) [Data set]

Siokou-Frangou, I; Christaki, U; Mazzocchi, MG et al. Plankton in the open Mediterranean Sea: a review. Biogeosciences; 2010; 7, pp. 1543-1586. [DOI: https://dx.doi.org/10.5194/bg-7-1543-2010]

Skliris, N; Mantziafou, A; Sofianos, S; Gkanasos, A. Satellite-derived variability of the Aegean Sea ecohydrodynamics. Cont Shelf Res; 2010; 30, pp. 403-418. [DOI: https://dx.doi.org/10.1016/j.csr.2009.12.012]

Sofianos S, Skliris N, Vervatis V et al (2005) On the Forcing Mechanisms of the Aegean Sea Surface Circulation. Geophys Res Abstracts 7

Strutton, PG; Chavez, FP. Biological Heating in the Equatorial Pacific: Observed Variability and Potential for Real-Time Calculation. J Clim; 2004; 17, pp. 1097-1109. [DOI: https://dx.doi.org/10.1175/1520-0442(2004)017]

Subrahmanyam, B; Ueyoshi, K; Morrison, JM. Sensitivity of the Indian Ocean circulation to phytoplankton forcing using an ocean model. Remote Sens Environ; 2008; 112, pp. 1488-1496. [DOI: https://dx.doi.org/10.1016/j.rse.2007.05.021]

Sweeney, C; Gnanadesikan, A; Griffies, SM et al. Impacts of Shortwave Penetration Depth on Large-Scale Ocean Circulation and Heat Transport. J Phys Oceanogr; 2005; 35, pp. 1103-1119. [DOI: https://dx.doi.org/10.1175/JPO2740.1]

Sylaios, G. Meteorological influences on the surface hydrographic patterns of the North Aegean Sea. Oceanologia; 2011; 53, pp. 57-80. [DOI: https://dx.doi.org/10.5697/OC.53-1.057]

Tsagkaraki, M; Theodosi, C; Kanakidou, M; Mihalopoulos, N. Atmospheric Deposition over the Aegean Sea and Its Impact on the Seawater Productivity; 2020; Berlin, Heidelberg, Springer: pp. 1-21.

Varkitzi I, Psarra S, Assimakopoulou G et al (2020) Phytoplankton dynamics and bloom formation in the oligotrophic Eastern Mediterranean: Field studies in the Aegean, Levantine and Ionian seas. 171:104662. https://doi.org/10.1016/j.dsr2.2019.104662

Velaoras, D; Zervakis, V; Theocharis, A. The Physical Characteristics and Dynamics of the Aegean Water Masses. The Handbook of Environmental Chemistry; 2021; Berlin, Heidelberg, Springer: pp. 1-29.

Vervatis, VD; Sofianos, SS; Theocharis, A. Distribution of the thermohaline characteristics in the Aegean Sea related to water mass formation processes (2005–2006 winter surveys). J Geophys Res Oceans; 2011; 116, C09034. [DOI: https://dx.doi.org/10.1029/2010JC006868]

Vervatis, VD; Sofianos, SS; Skliris, N et al. Mechanisms controlling the thermohaline circulation pattern variability in the Aegean–Levantine region. A hindcast simulation (1960–2000) with an eddy resolving model. Deep Sea Res Part I; 2013; 74, pp. 82-97. [DOI: https://dx.doi.org/10.1016/j.dsr.2012.12.011]

Vervatis, VD; Skliris, N; Sofianos, SS. Inter-annual/decadal variability of north Aegean Sea hydrodynamics over 1960–2000. Mediterr Mar Sci; 2014; 15, pp. 696-705. [DOI: https://dx.doi.org/10.12681/mms.852]

Wu, Y; Tang, CCL; Sathyendranath, S; Platt, T. The impact of bio-optical heating on the properties of the upper ocean: A sensitivity study using a 3-D circulation model for the Labrador Sea. Deep Sea Res 2 Top Stud Oceanogr; 2007; 54, pp. 2630-2642. [DOI: https://dx.doi.org/10.1016/j.dsr2.2007.08.019]

Zervakis, V; Georgopoulos, D. Hydrology and circulation in the North Aegean (eastern Mediterranean) throughout 1997 and 1998. Mediterr Mar Sci; 2002; 3, pp. 5-19. [DOI: https://dx.doi.org/10.12681/mms.254]

Zervakis, V; Karageorgis, AP; Kontoyiannis, H et al. Hydrology, circulation and distribution of particulate matter in Thermaikos Gulf (NW Aegean Sea), during September 2001-October 2001 and February 2002. Cont Shelf Res; 2005; 25, pp. 2332-2349. [DOI: https://dx.doi.org/10.1016/j.csr.2005.08.010]

Zhou, X; Marsland, SJ; Fiedler, R et al. Impact of different solar penetration depths on climate simulations. Tellus A: Dynamic Meteorol Oceanogr; 2015; 67, 25313. [DOI: https://dx.doi.org/10.3402/tellusa.v67.25313]