*5.1. Factors Controlling Planktonic Fauna Distribution in the Aegean Sea*

Of the oceanographic factors typically considered, SST shows the highest explanatory power for the distribution of the planktonic fauna during the Late Quaternary. This is in accordance with previously published studies showing temperature as the dominant factor controlling the biogeography of planktonic foraminifera and pteropods at both global and local scales [61,67,97]. However, the 2 remaining factors (PCA-2, PCA-3) exhibit a bipolar character and could be considered as indicators of the annual stability of the water column. For the planktonic foraminifera fauna, they show that the faunal composition in the south Aegean Sea was not only controlled by SST, but also seems to be affected by the degree of development and location of a permanent or seasonal thermocline/pycnocline. Its vertical placement in the water column is a direct consequence of changes in sea surface salinity (SSS) and productivity (SSP), which ultimately reflected the seasonal fluctuations of the periods of vertical mixing of water in the periods of intense stratification. The interpretation of the second axis focused on the appearance depths of pycnocline and deep chlorophyll maximum (DCM) and the thickness of mixed layer. The interpretation of the third axis focused on upwelling currents and/or river inputs (e.g., *G. bulloides*), parameters which primarily control the food availability and reproductive cycles of foraminifera [73] and are directly correlated to the seasonal fluctuations they present [94,98–100]. Therefore, a useful additional dimension of planktonic foraminifera ecology that is underlined by the PCA conducted in this study is the degree of vertical stratification of the water column and the way it is recorded (seasonal presence/absence pycnocline and DCM and upwellings and runoff), which are inextricably linked with the factors of primary productivity and seasonality. Similarly, the second and third factor (PCA2, PCA3) of pteropods are focused on the hydrological conditions and the overall oxygenation of the water column. Particularly, the down-core scores of the second factor coincide with the δ <sup>13</sup>C and E-index values, indicating that variations in primary productivity have an impact on pteropod abundances. Even though nutrient concentrations are not a limited factor for their distribution [67], our data suggest that fluctuations in nutrients and salinity due to the increased freshwater inputs during the sapropel deposition favor the flourishment of some species (*Cavolinia* spp., *B. chierchiae*; Figure 4). Additionally, the third factor suggests that oxygen concentration, and thus the intensity of the oxygen minimum zone (OMZ), are parameters that affect pteropod distribution and particularly the mesopelagic species [66].

#### *5.2. Paleoceanographic Reconstruction*

The results of the multivariate statistical analyses, in combination with paleoceanographic indices and isotopic data (Figure 6), reveal a succession of Late glacial to Holocene paleoclimatic and paleoceanographic changes. The evidence of these changes is interpreted and discussed in terms of the events that mainly accompanied the transition out of the late glacial period and the deposition of sapropel S1 during the Holocene Climatic Optimum (HCO).

δ δ **Figure 6.** Comparison between down-core score plots of the factors revealed by PCA analysis, micropaleontological and geochemical results of core KIM-2A: (**a**) planktonic paleoclimatic curve (PPC); (**b**) oxygen isotope record (δ <sup>18</sup>O*G. ruber*); (**c**) factor 1 of planktonic foraminifera (PCA1; temperature factor); (**d**) factor 1 of pteropods (PCA1; temperature indicator); (**e**) eutrophication index (E-index); (**f**) carbon isotope record (δ <sup>13</sup>C*G. ruber*); (**g**) factor 2 of pteropods (PCA2; productivity factor); (**h**) factor 2 of planktonic foraminifera (PCA2; stratification factor); (**i**) factor 3 of pteropods (PCA3; stratification factor); (**j**) *G. bulloides*/*G. ruber* ratio (S-index); (**k**) factor 3 of planktonic foraminifera (PCA3; seasonality factor); and (**l**) *G. glutinata* %.

#### 5.2.1. Late Glacial

δ Ο − During the late glacial period (21.1–15.7 ka BP), the heaviest δ <sup>18</sup>O values (2.49–3.26‰), accompanied by relatively low PPC values (−32% to +4%), suggest a cold upper water column (Figure 6a,b). Particularly, this interval was characterized by high percentages of the cold water foraminifera species *T. quinqueloba* (~30%), accompanied by *G. glutinata* (9%), and *G. scitula* (8%), and significant percentages of the warm-water *G. ruber* f. alba (~34%), that are suggestive of milder

climate after the Last Glacial Maximum (LGM), which is in accordance with other records in the Mediterranean [101,102]. Pteropod fauna is composed mainly by the cold water species *L. retroversa* (98%) and the cold-tolerant mesopelagic *C. pyramidata* in very low percentages (3%) which is consistent with relevant late glacial Mediterranean records [33,61,103]. Eutrophic species are also abundant in this interval (*N. pahyderma, N. dutertrei, T. quinqueloba* and *G. bulloides*) and are associated with the high values of E-index (Figure 6e). Notably, the high abundance of *N. pachyderma* (26%) indicates shallowing of the pycnocline and the formation of a DCM layer. In addition, δ <sup>13</sup>C values, around +1‰, and the trend to higher S-index values (Figure 6f,j), also support the development of eutrophicated waters. The moderate abundance of *G. bulloides* (5−17%; Figure 3) suggests little to no upwelling and/or runoff contribution in primary productivity. Therefore, the injection of nutrients into the euphotic zone can be attributed to the intensification and southward shift of westerly winds, as indicated by atmospheric circulation models for this time interval [104].

#### 5.2.2. Deglaciation

At 15.7 ka BP an abrupt shift of PPC to positive values, accompanied by lighter δ <sup>18</sup>O values (+2.2‰) (Figure 6a,b), are indicative of the climatic amelioration that occurred during the last deglaciation. These warmer conditions are also supported by the occurrence of the warm water species *H. inflatus* and *D. trispinosa* (up to 60% and 50% respectively), and the decrease in *L. retroversa* percentages (between 33% and 85%). This warming trend is in agreement with relevant paleoclimatic records from the eastern Mediterranean, and is attributed to the Bølling–Allerød (B-A) interstadial [7,13,14,105,106]. The increased SST and humidity are also recorded by the higher abundance of the terrestrial biomarkers in the south Aegean [17], and by a change in the benthic faunas from oxic to dysoxic indicator species [13]. In the beginning of this interval, Neogloboquadrinids were temporarily replaced by *G. bulloides* (26%), suggesting local upwelling. Though, later on the eutrophic *N. pachyderma* and *N. dutertrei* present their highest abundance (42% and 14% respectively). Additional components of this interval are *G. ruber* (both variants), *G. bulloides, T. quinqueloba*, and *G. inflata*, suggesting temperate and meso- to eutrophic waters, with strong seasonal mixing and local upwelling.

This state persisted until the onset of the Younger Dryas (YD) at about 12.9 ka BP, which is depicted in the abrupt decrease in PPC (from +28% to −10%) and in heavier values of δ <sup>18</sup>O (1.0–2.5‰) (Figure 6a,b). The planktonic foraminiferal fauna shows an increase in cold water species (*N. pachyderma, T. quinqueloba, G. inflata* and *G. glutinata*). Pteropod fauna is composed mainly of the cold-water species epipelagic *L. retroversa* and the mesopelagic temperate to warm-water species *C. pyramidata* and *D. trispinosa*, while the warm-water *H. inflatus* presents a decreasing trend (Figure 3). This climate response of south Aegean depression to the YD event (12.9–11.7 ka BP) seems to be in accordance with relevant signals from other Aegean sub-basins [6,7]. Towards the end of YD (12.6–12.2 ka BP) the species *G. rubescens, C. acicula*, and *B. chierchiae* are added to the fauna suggesting that mild climatic conditions prevailed for a brief time interval of about 400 years within the YD event. This brief climatic amelioration in the mid YD, which has been attributed to the displacement of the polar front by a few degrees north [26], has been also observed in north-central Aegean marine records [7,106] and coincides with the pattern of GRIP and NGRIP ice-core records [107,108], pollen-based reconstructions from the Jura [109] and the Balkans [110], as well as chironomid-based reconstructions from North Italy [111]. The increases in the S-index and E-index at around 12.2 ka BP (Figure 6e,j) coincide with this amelioration, reflected by the improved ventilation and eutrophication of the water column.

#### 5.2.3. Holocene

With the ending of the YD, a general climatic amelioration is seen in the records (increase in PPC, lighter values in δ <sup>18</sup>O, decline of temperature factors PCA1; Figure 6a–d) marking the beginning of the Holocene. Planktonic foraminifera fauna consists mainly by *G. ruber* f. alba that increases in abundance towards the onset of sapropel deposition, along with *G. bulloides* and *N. pachyderma*. The two latter species present an opposite trend, decreasing towards the onset of sapropel deposition. This trend

suggests the gradual development of stratified and oligotrophic surface waters. Pteropod fauna follows a similar pattern with the decreasing abundances of the species *D. trispinosa* and *C. pyramidata*, which are indicative of a well ventilated water column [112], and the increase in the epipelagic *B. chierchiae* and *C. acicula* (Figure 4). The reduction of E-index along with the values of δ <sup>13</sup>C and the S-index (Figure 6e–j) are further evidence for the development of these conditions. The replacement of *G. inflata* by *G. glutinata* (Figure 3) may be tentatively explained in terms of increased seasonality, and of increased freshwater input, which reduced surface buoyancy loss and hence suppressed mixing during the beginning of the Holocene [7]. This was also enhanced by the presence of the epipelagic pteropods (*Creseis* spp.) that proliferate in low-salinity waters [113], and the seasonality factor (PCA3 of planktonic foraminifera; Figure 6k).

At around 9.4 ka BP, the deposition of sapropel S1 as witnessed by their high organic carbon content (Corg: 1.8–3.3%, Figure 2a), coincided with the start of the overall δ <sup>18</sup>O depletion (+0.6‰ to −0.9‰; Figure 6b). The changes of the planktonic foraminifera fauna are characterized by an increase in *G. ruber* f. rosea, *G. ruber*f. alba, *G. siphonifera* and *O. universa* that are suggestive of extremely warm and stratified conditions (PPC ~100%; PCA1 low values; Figure 6b–d). The lower values of *G. bulloides*/*G. ruber* ratio in this interval are also indicative of strongly stratified water column and are in agreement with the stratification factor for both faunas (PCA2 of planktonic foraminifera and PCA3 of pteropod; Figure 6h–j). An increase in temperature and humidity around this time has been documented in all marine and terrestrial pollen records in the eastern Mediterranean region [6,7,14–16,114]. This paleoclimate change coincides with the Holocene summer precession-related insolation maximum in the Northern Hemisphere [115], and the monsoon intensification that resulted in a widespread increase in humidity over the Mediterranean region and concomitant increase of freshwater input to the Mediterranean Sea [116,117]. Pteropod fauna is characterized by the dominance of the warm oligotrophic *H. inflatus*, and the warm epipelagic *B. chierchiae, C. acicula*, and *Cavolinia* spp. (Figure 4). Mesopelagic species (*C. pyramidata* and *D. trispinosa*) are decreasing dramatically due to the enhanced stratification of the entire water column. Mesopelagic pteropods are affected by the OMZ alterations, which are climatically controlled [63,112]. In the humid and warmer conditions that persisted during the formation of S1, the subsequent stratification of the water column favored a strong and well developed OMZ that probably led to the reduction of mesopelagic species. The presence of the mesopelagic *H. inflatus* into the sapropel sublayers can be explained by its habitat. More explicitly, this species adopts a variable depth habitat during its growth stages, and it is more susceptible to the low oxygen concentration in the OMZ [63,118,119]. The presence of *L. bulimoides* in the upper part of S1a, with peaks at 8.4 ka BP, 8.1 ka BP, and 7.9 ka BP (~6%) and its absence in the S1b, suggest that during the end of S1a (8.6–7.7 ka BP) the conditions were more arid than during the onset of S1a (9.4–8.6 ka BP) and the interval of S1b (6.9–6.4 ka BP). In these two phases of S1a, δ <sup>13</sup>C present a decreasing trend with an average value of +0.8‰ in the first phase and +0.6‰ in the second (Figure 6f).

The warm and stratified conditions favorable for the sapropel deposition were interrupted between 7.7 ka BP and 6.8 ka BP. This interval (S1i) is marked by the decrease in PPC (from 90% to 60%) and the heavier values of δ <sup>18</sup>O (+0.5‰) and δ <sup>13</sup>C (+0.8‰) as shown in Figure 6a,b,f. The subsequent cooling is also reflected in significant faunal changes, such as the increase in abundance of *G. inflata*, *T. quinqueloba, G. bulloides*, and *N. pachyderma* (Figure 3). These species are associated with relatively cold temperatures and increased food supply, suggesting high primary production and stronger mixing of the water column [46,52]. In addition, the pteropod *L. trochiformis*, which is related to the mixed layer of the water column and thrives in upwelling conditions [120–123], presents a peak at the beginning of S1i (Figure 4).

From 6.4 ka BP to the top of the core (~5.0 ka BP), a trend to heavier δ <sup>18</sup>O values (from 0‰ to +1.3‰) and the drop of PPC (~60%) are recorded (Figure 6a,b). In planktonic foraminifera fauna an increase in *G. inflata, G. truncatulinoides* and *N. pachyderma* along with the reduction of *G. ruber* f. rosea*, G. siphonifera* gr*. O. universa* indicate lower SST, and stronger seafloor oxygenation due to vertical mixing. This latter is also suggested by the increase in the mesopelagic pteropod *D. trispinosa*. The peak

of *L. trochiformis* at 5.7 ka BP (Figure 4) is suggestive of upwelling conditions. At this point, and up to 5.0 ka BP, increased percentages of *G. bulloides* and *G. sacculifer* are indicative of increased productivity. This is also indicated by the high values of E-index, the heavy δ <sup>13</sup>C (up to +1.7‰) and the PCA2 factor of pteropods (Figure 6e–g).
