Ecological Constraints of Plankton Bio-Indicators for Water Column Stratiﬁcation and Productivity: A Case Study of the Holocene North Aegean Sedimentary Record

: This study presents novel ﬁndings on the drivers of the calcitic planktonic foraminiferal and aragonitic pteropod Holocene assemblages of the North Aegean Trough (northeastern Mediter-ranean), an area recording the interaction between dynamic water masses as they exchange between the northern and southern Mediterranean sub-basins. Both of these groups of microorganisms are the major producers of calcium carbonate in the ocean, and are particularly sensitive to climate and oceanographic changes over the late Quaternary. Downcore micropaleontological data from the gravity core AEX-15, supplemented with multivariate statistical Q-mode cluster and principal component analyses (PCA) results, provide signiﬁcant insights on the water column dynamics during the Holocene. Focusing on the last ~10 calibrated thousands of years before the present day (ka cal BP), our integrated study reveals that primary productivity is the dominant factor controlling the planktonic foraminifera distribution in the North Aegean Sea, whereas water column stratiﬁcation, and particularly the intensity of the oxygen minimum zone, seems to be the major driver on the pteropod distribution. Besides productivity and thermal stratiﬁcation, which show the highest explanatory power for planktonic foraminifera and pteropod communities, respectively, though they affect both groups to a different extent, upwelling seems to further affect both faunal groups. Overall, our ﬁndings are consistent with those derived by published late Quaternary eastern Mediterranean records, highlighting in parallel a useful additional dimension on planktonic foraminiferal and pteropod ecology, which is inextricably linked with the factors of primary productivity and vertical stratiﬁcation of the warm Holocene water column.


Oceanographic and Geologic Setting
The Aegean Sea is in the northern sector of the eastern Mediterranean, between the Turkish coastline to the east, the Greek mainland to the north and west, and bounded on the south by the island of Crete and the Cretan Arc. It is connected to the Black Sea through the Straits of Bosphorus and Dardanelles, and to the Levantine Sea through several larger and deeper straits between the Peloponnese, the islands of Crete and Rhodes, and south-western Turkey (Figure 1a). It is separated into two major sub-basins with different climatic conditions: the "north" and the "south" Aegean Sea. The north is more humid than the semiarid south basin [60]. The studied core comes from the North Aegean Trough (NAT; Figure 1a), a northeast-southwest (NE-SW) elongated depression with a depth range between 800 and 1590 m that includes several interconnected sub-basin depressions, separated from each other by 100-to 350-m-deep intervening shoals and associated islands [5,61]. The hydrography of the Aegean Sea is characterized, in general, by a cyclonic surface water circulation, although the most active dynamic features are the mesoscale permanent and/or recurrent cyclonic and anticyclonic eddies [60], which form as a consequence of the Etesian winds [62]. In the study area, the water column structure comprises three major water masses: the superficial nutrient-rich and less saline Black Sea Water (BSW; 9-22 °C and 22-23 psu), the warmer and saline Levantine Intermediate Water (LIW; 16-25 °C and 39.2-39.5 psu), and the North Aegean Deep Water (NADW; 13-14 °C and 39.1-39.2 psu) [63,64]. The surface and intermediate waters follow the general counterclockwise circulation of the Aegean Sea and progressively mix as they flow southward along the east coast of Greece [65]. The main characteristic of the north Aegean is the fluvial freshwater inputs that discharge from the Black Sea and the river runoff from the Greek and Turkish mainland, which dilute the highly saline waters from the Levantine and south-central Aegean. The lateral and diapycnal mixing of warm subsurface intermediate waters with low salinity surficial waters leads to strong near-surface thermal stratification, which is a dominant feature of the north Aegean water column during the summer. During the winter, upwelling nutrient-rich waters seem to create favorable conditions for primary production [66]. Such seasonal upwelling pathways are very important for the low-chlorophyll and phosphorus-controlled oceanic system of the eastern Mediterranean basin [67][68][69], highlighting the importance of small-scale hydrographic dynamics in controlling primary productivity of this mesotrophic to oligotrophic setting. The hydrography of the Aegean Sea is characterized, in general, by a cyclonic surface water circulation, although the most active dynamic features are the mesoscale permanent and/or recurrent cyclonic and anticyclonic eddies [60], which form as a consequence of the Etesian winds [62]. In the study area, the water column structure comprises three major water masses: the superficial nutrient-rich and less saline Black Sea Water (BSW; 9-22 • C and 22-23 psu), the warmer and saline Levantine Intermediate Water (LIW; 16-25 • C and 39.2-39.5 psu), and the North Aegean Deep Water (NADW; 13-14 • C and 39.1-39.2 psu) [63,64]. The surface and intermediate waters follow the general counterclockwise circulation of the Aegean Sea and progressively mix as they flow southward along the east coast of Greece [65]. The main characteristic of the north Aegean is the fluvial freshwater inputs that discharge from the Black Sea and the river runoff from the Greek and Turkish mainland, which dilute the highly saline waters from the Levantine and south-central Aegean. The lateral and diapycnal mixing of warm subsurface intermediate waters with low salinity surficial waters leads to strong near-surface thermal stratification, which is a dominant feature of the north Aegean water column during the summer. During the winter, upwelling nutrient-rich waters seem to create favorable conditions for primary production [66]. Such seasonal upwelling pathways are very important for the low-chlorophyll and phosphorus-controlled oceanic system of the eastern Mediterranean basin [67][68][69], highlighting the importance of small-scale hydrographic dynamics in controlling primary productivity of this mesotrophic to oligotrophic setting.

Material
The study core was recovered from the NAT (39 • 39.900 N, 24 • 15.369 E; 1242 m water depth, length 178 cm), with R/V Aegaeo in October 2013 (Figure 1a,b). The planktonic foraminiferal and pteropod assemblages derived from 115 downcore samples, were picked from 10 cm 3 of wet sediment after washing through a 125 µm mesh sieve and cleaning using the HyPerCal protocol of [70]. All shells (300 individuals for each group) were handpicked, identified, and counted in each sample and then converted into percentages (Table S1; Supplementary Material). Detailed micropaleontological procedures, along with the time stratigraphic framework of the analyzed core are described in [30]. Based on the published age model of the study core, which is based on a combination of accelerator mass spectrometry ( 14 C AMS) radiocarbon date measurements, additional control points, and bioevents of planktonic foraminifera, its sediments cover the last 10.1 ka BP with an average sedimentation rate of 17.32 cm/ka.

Multivariate Statistical Analyses
The number of planktonic foraminifera and pteropods (Euthecosomata) counted is statistically reliable for both paleoceanographic and paleoclimatic reconstructions [71]. To determine the overall statistical similarity between samples, Q-mode cluster analysis was used, following the algorithms of [72] using the correlation coefficient matrix. In the case of planktonic foraminifera, the totality of the samples was used, whereas for pteropods, cluster analysis was carried out in 105 out of 115 samples. This was due to the lack of pteropod fauna in samples 1.5, 47, 51, 57, 59, 107, 122.5, 141.5, 142.5, and 143.5 cm. For paleoenvironmental reconstructions, multivariate statistical analyses have been performed on the data set, after exclusion of rare species (<3%) and grouping of species that have a discontinuous, scattered distribution at generic level. The results of cluster analysis were reported as Morisita similarity and arranged in two-dimensional hierarchical dendrograms, wherein locations were presented along the Y-axis while similarity level was plotted on the X-axis. Q-mode cluster analysis was performed in both faunal groups (planktonic foraminifera and pteropods) in order to investigate the differences or correlations among the biotopes identified within plankton communities.
Furthermore, principal component analysis (PCA) was applied to reduce the dimensionality of a multivariate data set to a few principal factors that determine the distributions of species. Raw data were processed using PAST (2.17) multivariate statistical software package of [73]. The factors obtained were rotated using a varimax-normalized algorithm, which allows more straightforward interpretation of the loadings of the principal components and maximization of the variances explained by the factors extracted. The resulting factor scores show the contribution of each factor in every sample and, therefore, the downcore contribution of each factor. The total number of factors was defined by minimizing the remaining "random" variability, and by the possibility to relate the factors to modern hydrographic conditions, and planktonic foraminiferal and pteropod ecology.

Results
Both faunal groups are abundant and well preserved in the samples of the studied core. In particular, 17 species of planktonic foraminifera lumped in 12 groups and 11 species of pteropods were identified (Table S1; Supplementary Materials). The most abundant planktonic foraminiferal species are Globigerina bulloides, Turborotalita quinqueloba, and Globigerinoides ruber alba, whereas the species Globorotalia inflata and Neogloboquadrina pachyderma present sporadic peaks. Pteropod species in abundance with a continuous distributional pattern are Heliconoides inflatus and Boasia chierchiae, whereas the rest of them present a scattered distributional pattern. More details on the distributional patterns of both planktonic groups are given in [30].

Q-Mode Cluster Analysis
The Q-mode hierarchical cluster analysis represents the grouping of samples based on the downcore abundance of the species higher than at least 3% in one sample. Three distinct assemblages of planktonic foraminifera and pteropods were identified by Q-mode cluster analysis in the studied core, reflecting different biotopes during the Holocene. Each assemblage is characterized by the dominant species and named accordingly after it.

Planktonic Foraminifera
Cluster I (T. quinqueloba assemblage): it contains 47 samples (samples 61 cm, 97 cm, and samples between 120 and 166 cm; Figure 2). Apart from samples 61 and 97, the remaining 45 samples correspond to the most recent sapropel S1. This assemblage is dominated by T. quinqueloba (up to 65%), an eutrophic species preferring low salinities [74]. In this assemblage, G. bulloides also occurs with high percentages (up to 50%), as well as the warm oligotrophic Globigerinoides ruber rosea (25%) and SPRUDTS group.

Q-Mode Cluster Analysis
The Q-mode hierarchical cluster analysis represents the grouping of samples based on the downcore abundance of the species higher than at least 3% in one sample. Three distinct assemblages of planktonic foraminifera and pteropods were identified by Q-mode cluster analysis in the studied core, reflecting different biotopes during the Holocene. Each assemblage is characterized by the dominant species and named accordingly after it.

Planktonic Foraminifera
Cluster I (T. quinqueloba assemblage): it contains 47 samples (samples 61 cm, 97 cm, and samples between 120 and 166 cm; Figure 2). Apart from samples 61 and 97, the remaining 45 samples correspond to the most recent sapropel S1. This assemblage is dominated by T. quinqueloba (up to 65%), an eutrophic species preferring low salinities [74]. In this assemblage, G. bulloides also occurs with high percentages (up to 50%), as well as the warm oligotrophic Globigerinoides ruber rosea (25%) and SPRUDTS group. Cluster II (G. inflata assemblage): this assemblage groups three samples (115 cm, 117 cm, and 119 cm; Figure 2), which belong to the oxidized part of sapropel S1. The domi- Cluster II (G. inflata assemblage): this assemblage groups three samples (115 cm, 117 cm, and 119 cm; Figure 2), which belong to the oxidized part of sapropel S1. The dominant component of this assemblage is G. inflata (up to 32%), which is a temperate species indicative of a homogenous water column that prefers to live at the bottom of thermocline depths [47]. Additional components of this assemblage are the species N. pachyderma, G. ruber alba, G. ruber rosea, T. quinqueloba, and G. bulloides. Neogloboquadrina pachyderma thrives in cool water close to or below the thermocline [75]. Its abundance is related to the development of the deep chlorophyll maximum (DCM) [75].

Pteropods
Cluster I (Diacria trispinosa assemblage): it contains six samples (samples 6.25 cm, 165.5 cm, and from 168 cm to 173.5 cm; Figure 3) that mainly correspond to the presapropel interval. This assemblage is characterized by the high relative abundance of the mesopelagic D. trispinosa (up to 60%) along with the epipelagic Creseis acicula (up to 50%). Additional components are the species B. chierchiae (25%), Clio pyramidata (20%), and H. inflatus (15%). nant component of this assemblage is G. inflata (up to 32%), which is a temperate species indicative of a homogenous water column that prefers to live at the bottom of thermocline depths [47]. Additional components of this assemblage are the species N. pachyderma, G. ruber alba, G. ruber rosea, T. quinqueloba, and G. bulloides. Neogloboquadrina pachyderma thrives in cool water close to or below the thermocline [75]. Its abundance is related to the development of the deep chlorophyll maximum (DCM) [75].

Principal Component Analysis
Principal component analysis is a commonly used method for multivariate statistical analysis, mainly because of its simple algebra and direct interpretation. The application of this statistical analysis yielded a three-factor model for both planktonic foraminiferal and pteropod communities. The interpretation of the three components in each case was based on the screen plots of eigen values, and the factor loadings of the planktonic foraminiferal and pteropod species, respectively. The distinguished factors account for 90.94% and 79.46%, of the total variance for the planktonic foraminiferal and pteropod, respectively (Tables 1 and 2), with their factor loadings showing the contribution of each factor in every sample and, therefore, the downcore contribution of each factor (Tables 3 and 4, Figure 4). In the case of a bipolar factor, which has extremes of positive and negative loadings, high positive factor scores are related to the positive pole and high negative scores to the negative pole, respectively.

Planktonic Foraminifera
The first varimax factor (PCA1) accounts for 67.95% of the total variance (Table 1) and exhibits a bipolar character with the negative pole to be dominated by the eutrophic species T. quinqueloba and the positive pole by the oligotrophic G. ruber alba (Table 3) [76][77][78]. Therefore, this factor is interpreted as a productivity indicator (Figure 4a). The second factor (PCA2) explains 15.58% of the total variance (Table 1). It is characterized by positive values of the surface dwellers G. ruber alba and T. quinqueloba [78]. Negative loadings are dominated mainly by the species G. bulloides, which is highly dependent on enhanced food levels, upwelling, strong seasonal mixing, or freshwater inputs [79][80][81][82]. Thus, the PCA2 factor is referred to as the stratification factor (Figure 4b). The third varimax factor (PCA3) accounts for the 7.41% of the total variance (Table 1) and is interpreted as an upwelling indicator (Figure 4c) since its main positive representative species, G. bulloides (Table 3), is strongly associated with the seasonal upwelling [83][84][85][86]. On the other hand, the negative loadings are represented mainly by the species G. ruber rosea, a warm subtropical species, whose abundance is mainly controlled by the thermal stratification of the upper water column [21,22,78]. (b) PCA2 of planktonic foraminifera was interpreted as a stratification factor; (c) PCA3 of planktonic foraminifera was interpreted as an upwelling factor; (d) PCA1 of pteropods was interpreted as a stratification factor; (e) PCA2 of pteropods was interpreted as a productivity factor; (f) PCA3 of pteropods was interpreted as an upwelling factor. Gray and light gray bands correspond to the sapropel S1 and its oxidized part, respectively.

Planktonic Foraminifera
The first varimax factor (PCA1) accounts for 67.95% of the total variance (Table 1) and exhibits a bipolar character with the negative pole to be dominated by the eutrophic species T. quinqueloba and the positive pole by the oligotrophic G. ruber alba (Table 3) [76][77][78]. Therefore, this factor is interpreted as a productivity indicator (Figure 4a). The second factor (PCA2) explains 15.58% of the total variance (Table 1). It is characterized by positive values of the surface dwellers G. ruber alba and T. quinqueloba [78]. Negative loadings are dominated mainly by the species G. bulloides, which is highly dependent on enhanced food levels, upwelling, strong seasonal mixing, or freshwater inputs [79][80][81][82]. Thus, the PCA2 factor is referred to as the stratification factor (Figure 4b). The third varimax factor (PCA3) accounts for the 7.41% of the total variance (Table 1) and is interpreted as an upwelling indicator (Figure 4c) since its main positive representative species, G. bulloides (Table 3), is strongly associated with the seasonal upwelling [83][84][85][86]. On . Environmental factors controlling planktonic foraminifera and pteropod distribution resulting from the PCA. (a) PCA1 of planktonic foraminifera was interpreted as a productivity factor; (b) PCA2 of planktonic foraminifera was interpreted as a stratification factor; (c) PCA3 of planktonic foraminifera was interpreted as an upwelling factor; (d) PCA1 of pteropods was interpreted as a stratification factor; (e) PCA2 of pteropods was interpreted as a productivity factor; (f) PCA3 of pteropods was interpreted as an upwelling factor. Gray and light gray bands correspond to the sapropel S1 and its oxidized part, respectively.

Pteropods
The first varimax factor (PCA1) explains 37.76% of the total variance (Table 2), with the positive loadings expressed by the mesopelagic H. inflatus, tolerant to a low oxygen concentration of the oxygen minimum zone (OMZ), [57] and the epipelagic B. chierchiae (Table 4). In contrast, negative loadings are expressed mainly by species indicative of a well-ventilated water column (L. trochiformis and C. pyramidata) [57,87]. Thus, PCA1 can be explained as a stratification factor (Figure 4d). The second factor (PCA2) explains the 27.66% of the total variance (Table 2). Its positive pole is represented mainly by the mesopelagic oligotrophic H. inflatus and the negative pole by the epipelagic B. chierchiae, which proliferates in nutrient-enriched waters (Table 4) [29,57]. Therefore, PCA2 is interpreted as a productivity factor (Figure 4e). The third factor (PCA3) describes 14.04% of the total variance ( Table 2). The main representative of the positive pole, L. trochiformis (Table 4), is characterized as a typical upwelling species [87,88]; thus, PCA3 can be interpreted as an upwelling factor (Figure 4f).

Biotopes of the North Aegean
The main assemblages identified in core AEX-15 through the cluster analysis, represent the sapropel S1 or the pre-and/or post-sapropel interval. Based on the ecological characteristics of the species participating in each assemblage, a unification of some of them into biotopes was made. Thus, each biotope is indicative of paleoenvironmental processes that affect both faunal groups.
More explicitly, the assemblage of T. quinqueloba, which corresponds to the sapropel S1 interval in the north Aegean, represents a biotope with low salinity and increased food availability in the surface waters [6,74,89,90]. This is explained by the larger river inflows draining the surrounding land [5,89] and supplying nutrients that fuel the productivity. In addition, the warm oligotrophic G. ruber rosea, and the SPRUDTS group that are also abundant in this assemblage, suggest enhanced thermal stratification of the upper water column [47,79,82]. This is further reinforced by the increased percentages of H. inflatus (H. inflatus assemblage) and the reduction/absence of mesopelagic species typical of a well-ventilated water column (e.g., Clio cuspidata, C. pyramidata, D. trispinosa) in pteropod's fauna. Although Heliconoides inflatus is a mesopelagic species, its presence in this biotope is explained by its habitat. This species adopts a variable depth habitat during its growth stages and is also tolerant to low oxygen concentration levels [57]. The presence of low salinity surface waters, as inferred from the planktonic foraminifera, is also enhanced by the high relative abundance of the epipelagic pteropod B. chierchiae. This species is known to proliferate in low, as well as high, salinity shallow waters [29,91,92]. Thus, both plankton assemblages (T. quinqueloba and H. inflatus) that correspond to the sapropel S1 represent a biotope with a strongly stratified water column characterized by intense OMZ, with low salinity and high fertility of the surficial waters.
The next biotope includes the G. ruber alba and L. trochiformis assemblages that correspond to the post-sapropel interval (Northgrippian and Meghalayan stages). In this biotope, oligotrophic (G. ruber alba, G. ruber rosea, and G. rubescens) and eutrophic (G. bulloides, T. quinqueloba, and N. pachyderma) species simultaneously occur, indicating a seasonal contrast in primary productivity probably related to the food availability. Seasonality is also reinforced by the continuous presence of Globigerinita glutinata since this species can survive, both in the oligotrophic surface and in more eutrophic waters, by changing its diet from diatoms to chrysophytes [47]. Moreover, the findings suggest a seasonal contrast in temperature as warm-characteristic species (G. ruber alba, G. ruber rosea, and G. rubescens) coexist with cold-indicative species (T. quinqueloba and N. pachyderma). On the other hand, the occurrence of mesopelagic pteropods along with epipelagic, suggests a well-ventilated water column [93]. The high relative abundance of L. trochiformis is indicative of local upwellings [87], whereas its reduction at~3.0 ka BP, along with the increased relative abundance of S. subula, is suggestive of an increase in sea surface salinity [56,57,94]. Thus, our findings suggest a high seasonal biotope indicative of a well-ventilated and high saline water column as a result of the reduction in freshwater inputs. This interpretation is in accordance with clay mineral data derived from north Aegean sediments [5,95] that point to particularly dry conditions in the north Aegean catchment as well as the reduction in the surface water Black Sea outflow at~4.5 ka BP [10,96].
The biotope characteristic of the G. inflata assemblage corresponds to the oxidized part of S1. The occurrence of the deep mixing species G. inflata and N. pachyderma in this assemblage reflects the prevalence of horizontally and/or vertically homogenized conditions in the water column [78,79,97]. Moreover, the biotope characteristic of the D. trispinosa assemblage corresponds to the pre-sapropel interval. Diacria trispinosa is a species typical of Atlantic-influenced waters [94,98], reflecting the influence of Atlanticmodified waters into the eastern Mediterranean basin prior to the S1 deposition. The epipelagic euryhaline species C. acicula is also precent in this assemblage. Its distribution pattern in the Holocene record of the Aegean sea was described for the first time in this gravity core (AEX-15; [30]), and is negatively affected by the water column stratification. Similar observations are recorded from the western continental shelf of India [57], where it appears to be related to the conspicuous differences in water column stratification linked with the vertical salinity gradient. Thus, our findings reflect a biozone with homogenized conditions in the water column influenced by Atlantic waters.

Factors Controlling Planktonic Fauna Distribution
Both faunal groups seem to be reliable indicators for past paleoenvironmental changes in this setting. The differences relay the extent to which each factor affects the fauna. In particular, the productivity factor is expressed as being more intensified in planktonic foraminifera fauna (68%) than in pteropods (28%) (Tables 1 and 2). In the case of planktonic foraminifera, this can be explained by the species T. quinqueloba and G. bulloides that render the role of productivity as the main factor, especially in the more humid and eutrophicated north Aegean region [6,60,89]. In contrast, the pteropod distribution pattern does not respond in the same way as planktonic foraminifera to changes in productivity. It is known that nutrient concentration is not a limited factor for their distribution [99], though our data suggest that fluctuations in nutrients and salinity due to the increased freshwater inputs, especially during the sapropel S1 deposition, favor the flourishment of the species B. chierchiae [29]. The stratification factor, on the contrary, affects the pteropod fauna (38%) more than the planktonic foraminifera (16%) (Tables 1 and 2). Our data suggest that the pteropod distribution pattern is controlled primarily by the oxygen concentration of the water column and, thus, the extent and intensity of the OMZ [29,56], which is related to the stratification of the entire water column. OMZ alterations are climatically controlled, with humid and milder climates favoring a strong and well developed OMZ, whereas cooler and/or arid periods are characterized by a more aerated OMZ [93,100]. In particular, mesopelagic species that are affected negatively by low oxygen concentration and intensified OMZ, are absent or present with very low percentages during the stratified conditions of the sapropel S1 layer. In contrast, such intense changes in species abundances of planktonic foraminifera, are not observed.
Finally, the third factor affecting equally both faunal groups' distribution in the north Aegean Sea seems to be upwelling. Upwellings control food availability and foraminifera reproductive cycles [47,101]. The participation of this factor in Holocene records of the north Aegean is explained by the presence of cyclonic and anticyclonic eddies and, thus, to the development of down and upwellings of the organic matter [60,89].

Conclusions
Plankton foraminiferal and pteropod faunas have been studied in core AEX-15, from the North Aegean Trough (northeastern Mediterranean), to evaluate their relationship with surface water dynamics, and to decipher the control of the Holocene productivity and stratification changes coupled with the impact of the regional eastern Mediterranean climate regime. The ecological interpretation of both of these planktonic groups, integrated with the multivariate statistical analysis data, has provided an accurate interpretation of the changes in surface water properties prior, during, and after the deposition of the most recent sapropel S1. The Q-mode cluster and principal component analyses performed in the planktonic fauna of core AEX-15 revealed three distinct assemblages in the north Aegean Sea that can be summarized into relevant biotopes, each one representative of different hydrological and climatic conditions. These biotopes correspond to the sapropel S1 and its oxidized part, pre-and, post-sapropel intervals. The PCA performed in both fauna groups, suggested that primary productivity, stratification of the water column, and upwelling are the main drivers of the planktonic foraminifera distribution. The same factors are recognized for the pteropod fauna but with a different order, indicating stratification as the primary factor controlling their distribution. Overall, the observed high-resolution water column variability allows a detailed palaeoecological reconstruction for the Holocene archive in the Aegean Sea and, furthermore, provides a notable contribution to paleoclimatic studies, facilitating intercorrelations between similar oceanographic basins and offering a better comprehension of the paleoceanographic history of these basins.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/jmse9111249/s1. Table S1: Raw data of the percentages of planktonic foraminifera and pteropod species in downcore samples.
Funding: This research received no external funding.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available in Supplementary material [ Table S1].