NEGLECTED SMALL FORAMINIFERS OF THE EASTERN MEDITERRANEAN AND THEIR PALEOECOLOGICAL IMPORTANCE Open Access

The Mediterranean Sea is a semi-restricted basin due to its limited connection with the open ocean. The Sicily Channel divides the Mediterranean Sea into eastern and western basins with different temperature and salinity profiles (Figs. 1A, B). The surface of the eastern Mediterranean Sea is characterized by high salinity, whereas the relatively low salinity inflow of Atlantic waters through the Strait of Gibraltar dominates the western Mediterranean Sea (Tanhua et al., 2013). The deep water of the eastern basin is significantly more saline and warmer compared to the deep water of the western basin, and as a result, the density of the eastern basin is higher than the western basin. The Mediterranean Sea surface layers are mainly characterized by low nutrient concentrations causing oligotrophic, or even ultra-oligotrophic features that increase towards the eastern Mediterranean (e.g., Kress et al., 2003; Pujo-Pay et al., 2011). Consequently, various foraminiferal assemblages developed and have adapted to the chemical and physical conditions of the distinct basins (e.g., Cimerman & Langer, 1991; Schmidt et al., 2015).

Most research on foraminifers worldwide and specifically in the Mediterranean Sea have focused on >125-µm (e.g., Sprovieri et al., 2003, 2006; Öğretmen et al., 2018; McNeill et al., 2019) and >150-µm size fractions (e.g., Violanti et al., 1991; Jorissen et al., 1993; Casford et al., 2002; Pérez-Folgado et al., 2003; Béjard et al., 2024). However, gaps in the fossil records of many species may result from using these larger mesh sizes (Kroon & Nederbragt, 1990; Boltovskoy, 1991). Using both smaller and larger mesh sizes allows the sampling of a wider species spectrum, which is of special interest for paleoceanographic reconstructions (Carstens et al., 1997) and makes possible an increased biostratigraphic resolution (Filipescu & Silye, 2008). Because of their rapid reaction to organic matter input, small-sized foraminifers are considered potential indicators of the seasonality of organic matter flux in marine environments (Smart et al., 1994; Thomas et al., 1995; Gooday, 2003; Duchemin et al., 2007). In addition, the small-sized species are more adaptively responsive due to their tendency to have shorter generation times (Arnold et al., 1995). During times of stress, this adaptive responsiveness may give them relatively enhanced resistance to extinction due to Cope’s Rule (Arnold et al., 1995). The Lilliput effect (dwarfing of taxa) marks morphologic and intraspecies size reductions associated with environmental stresses (e.g., restricted basins) and has been observed in many organisms including foraminifers (Keller & Abramovich, 2009). On the other hand, Por (1989) suggested that the body sizes of the specimens in the Laventin basin were smaller than their conspecifics in the Western Mediterranean, expressed as “Levantine nanism (dwarfism)”. He suggested that nanism might be the result of exceptional environmental factors (high salinity and temperature), low productivity, or a combination of all these aspects. Nanism had been suggested for various taxonomic groups: by Levi (1957) for sponges; by Stephen (1958) for sipunculids; by Ben-Eliahu & Fiege (1994) for polychaetes; by Spezzaferri et al. (1998) for planktic foraminifers; by Sonin et al. (2007) and Goren (2014) for fish; by Miebach et al. (2017) for pollen. Also, Zarkogiannis et al. (2020) suggested that the average size of planktic foraminifers decreased slightly from the North Aegean toward the Levantine Sea and may be mainly related to nutrient availability.

In the semi-restricted, ultra-oligotrophic eastern Mediterranean Sea, it is reasonable to expect relatively small foraminifers. In addition to the reduction in energy needed to maintain smaller body sizes in such ultra-oligotrophic regions, smaller test volumes provide negative buoyancy to descend into nutrient-rich depths (Zarkogiannis et al., 2019). Therefore, analyzing a smaller size fraction (>63-µm) might prevent overlooking some details (e.g., new and dominant species in the assemblage) and reveal the true paleoecological factor(s) in a marine setting. Accordingly, I analyzed >63-µm residue to obtain an accurate paleoecological model using Cluster Analysis and Principal Component Analysis. In addition, the method is checked by comparing results to those using the >125-µm size fraction. Since many of the species are extant, it is possible to project into the past environmental conditions using their current life preferences ( Appendix 1).

Sampling was carried out at two different locations, the Kargı section and the Finike Seamount (Figs. 1C, D, E). The Kargı Section is an outcrop located approximately 100 m above sea level in the Aksu Basin (Fig. 1D). In total, eight samples were collected from the Kargı section at one-meter intervals. The samples from the Finike Seamount were collected with a beam trawl from two stations at 1800 m and 2200 m water depth from the R/V Yunus S in September 2021 (Fig. 1E). All the samples were washed through a 63-μm sieve separately with tap water without any prior acid treatment, and residues were dried at either room temperature or in the furnace at 60°C. Foraminifers were handpicked with a brush using a stereomicroscope. Planktic foraminifer analyses were carried out by counting, when possible, up to 300 specimens from the >63-μm and >125-μm fractions of the dry residue. Although quantitative and qualitative analyses of planktic foraminifers were performed for the Kargı section, only a qualitative assessment was carried out for the Finike Seamount samples due to the small number of samples. Sample preparation and microscope studies were performed at 1) Süleyman Demirel University Department of Geological Engineering and 2) the Eastern Mediterranean Center for Oceanography and Limnology (EMCOL) and Eurasia Institute of Earth Sciences at Istanbul Technical University.

The taxonomic identification of planktic foraminifers was based on Parker (1962), Postuma (1971), Kennett & Srinivasan (1983), Loeblich & Tappan (1988), Iaccarino et al. (2007), and Schiebel & Hemleben (2017). Hansen & Lykke-Andersen (1976), Gregory & Bridge (1979), McNeil et al. (1982), Loeblich & Tappan (1988), and McDougall (1994) were used for the taxonomic identification of Cribroelphidum ustulatum. For further visual comparison, I used foraminifer databases www.microtax.org, WoRMS, and foraminifera.eu. Because Trilobatus sacculifer and T. trilobus have been shown to be conspecific in the modern plankton by André et al. (2013), in this study Trilobatus sacculifer includes both T. sacculifer and T. trilobus (i.e. Trilobatus sacculifer and T. trilobus are merged in Tables 1, 2, and 4).

The geographical distribution (in the Mediterranean Sea) of each species identified in these samples was determined using some keywords and filters in the database PANGAEA (https://www.pangaea.de/) ( Appendix 1). For instance, for Berggrenia pumilio, firstly “Quaternary Berggrenia pumilio” was searched from the database and then “Mediterranean Sea” was used as a filter to constrain the results. For further searches, www.microtax.org and common search engines were used.

Microfaunal analyses of the planktic foraminifers were carried out on the >63-μm and >125-µm size fractions of the samples to test the paleoecological differences (Tables 1 and 2). Microfaunal analyses of planktic foraminifers were performed considering only species with more than 5% relative abundance. (Fig. 2, Tables 3 and 4). To obtain a reasonable paleoecological model, species were grouped according to their genera based on their ecological requirements following Bé (1977) and Schiebel & Hemleben (2017). Neogloboquadrina pachyderma is associated with cold polar and subpolar waters in both hemispheres, whereas N. incompta is restricted to more temperate waters of global subpolar and transition zones (Darling et al., 2006). Therefore, these two species were not grouped and analyzed separately due to their different ecological preferences.

Two algorithms were used: the unconstrained Cluster Analysis (Chord distance measure and the un-weighted pair group method using arithmetic average-UPGMA) in Q-mode (column mode) and the Principal Component Analysis (PCA) using the PAST software, version 4.12 (Hammer et al., 2001). The Cluster Analysis was applied in Q-mode to find dominant associations throughout the sections. The Cluster Analysis dendrogram was tested through the cophenetic correlation coefficient (c) (Sokal & Rohlf, 1962; Mouchet et al., 2008). The highest cophenetic correlation coefficient points to the cluster that holds most of the information (Borcard et al., 2018).

The PCA was applied to quantify variables (components) within the multivariate dataset. To test the accuracy of the PCA, bootstrap was used to estimate the variances of factor loadings (Chatterjee, 1984). The bootstrapping was carried out with 999 bootstrap replicates for all analyses. In the scree plots, 95% bootstrapped confidence intervals were evaluated for each Eigenvalue and broken bar value.

The Cluster Analysis applied to >63-μm size fraction in Q-mode (c = 0.95) yields three main clusters that are discriminated in the Kargı section at a distance value equal to 0.40, including in total seven taxa showing more than 5% abundance (Figs. 2, 3A). The three clusters of the section were investigated in six groupings as Globigerinoides obliquus, Globoturborotalita rubescens, Orbulina universa, Turborotalita group (consists of Turborotalita clarkei and T. quinqueloba), Neogloboquadrina incompta, and Globigerinita glutinata. Accordingly, three clusters of the section reported in Figure 3A are the following: cluster A, including four samples, is dominated by Turborotalita assemblage; cluster B, including three samples, dominated by Turborotalita and Globigerinita glutinata assemblages; and cluster C including only sample 7 and dominated by Orbulina universa. In cluster A, Turborotalita (consisting of T. clarkei and T. quinqueloba) is the dominant taxa in samples S6, S3, S5, and S2 with relative abundances of ∼46, 42, 55 and 61%, respectively, and Globigerinita glutinata is the second most abundant species relative abundances of ∼17, 17, 16 and 14%, respectively, in the samples of this cluster. In cluster B, Turborotalita (consisting of T. clarkei and T. quinqueloba) is the dominant taxa in samples S8 and S1 with relative abundances of ∼34 and 37%, respectively. Only sample S4 is subordinated by Globigerinita glutinata with a relative abundance of ∼30%, but Turborotalita is the second most dominant taxa with a relative abundance of ∼26%. In cluster C, Orbulina universa is the dominant species in Sample S7 with a relative abundance of ∼29%.

Excluding the <125-μm size fraction results in two main clusters that are dominated by (cluster A) Orbulina universa and (cluster B) Orbulina universa and Globigerinoides (Fig. 4A). The Globigerinoides group consists of Globigerinoides elengatus and Globigerinoides obliquus. In cluster A, Orbulina universa is the dominant species in sample S4 with a relative abundance of ∼48%. In cluster B, Orbulina universa is the dominant species in samples S5, S8, S1, S6, S3, S2, S7 with relative abundances of ∼52, 54, 43, 28, 29, 32, 38%, respectively. Globigerinoides is the second most abundant taxon with relative abundances of ∼14, 11, 18, 23, 21, 23, 34%, respectively.

The Kargı section was deposited in a marine setting and hosted three planktic foraminifer assemblages (Turborotalita, Turborotalita-Globigerinita glutinata, and Orbulina universa; Fig. 3A). The PCA of the Kargı section reveals that only one Principal Component (PC) completely lies above the broken stick values, falling within the 95% confidence interval (Fig. 3B). In other words, the Kargı marine setting was deposited in an environment with a single controlling parameter. This PC explains 79% of the variance (Fig. 3B). The PC loadings >0.3 were considered for the PCA interpretation. Accordingly, PC is positively correlated with Turborotalita (loading 0.89), but negatively correlated with Orbulina universa (loading -0.33; Fig. 3C).

By excluding the <125-μm size fraction and based on the PCA assessment, the Kargı marine setting also appears controlled by only one PC (Fig. 4B), but the PC shows a positive correlation with Globigerinoides (consisting of Globigerinoides elengatus and Globigerinoides obliquus) and Orbulina universa, and no species is weighted negatively for PC (Fig. 4C).

In addition to abundant small-sized planktic species (such as Turborotalita clarkei, T. quinqueloba, Globigerinita glutinata, and other small-sized planktic species), Globoturborotalita rubescens, Turborotalita humilis, Berggrenia pumilio, Globorotalia bermudezi, and benthic species Saidovina karreriana were captured from the 63–125-µm residue. Turborotalita clarkei, Berggrenia pumilio, and Globorotalia bermudezi were not captured from the >125-µm size fraction (Table 2). Also, Globoturborotalita rubescens, Turborotalita quinqueloba, and Globigerinita glutinata were captured from the >125-µm residue with maximum relative abundances of ∼ 3, 2, and 6%, respectively (Tables 2 and 4). Cribroelphidium ustulatum was only found in the >63-µm residue from the Finike Seamount seafloor. Geographic locations of Turborotalita clarkei, Berggrenia pumilio, Globorotalia bermudezi, and Cribroelphidium ustulatum in the studied locations in the Mediterranean Sea are shown in Figures 1D, E, and the selected species are illustrated in Figure 5. Additionally, for visual comparison, C. ustulatum, identified by Hansen & Lykke-Andersen (1976) from Quaternary deposits in Lundergaard, Denmark, is presented in Figure 6.

By analyzing the small-sized fraction (>63-µm), new previously unreported species might be recorded from different biostratigraphic ranges in the Mediterranean Sea. In this way, Kanbur & Öğretmen (2022) reported for the first time T. clarkei and B. pumilio from Calabrian (Early Pleistocene)-aged onland sediments in the eastern Mediterranean Region. Up until their work, in the entire Mediterranean Region, Turborotalita clarkei was only mentioned in recent sediments of the eastern Mediterranean Sea and late Holocene deposits of the Marmara Sea by Alavi (1980, 1988, respectively). Furthermore, the same study (Kanbur & Öğretmen, 2022) was the first (Calabrian) record of Berggrenia pumilio in the Mediterranean. According to the related literature, first reported occurrence (the base) of Berggrenia pumilio was in the Holocene stage (Parker, 1962; Saito et al., 1981; Loeblich & Tappan, 1994; Rebotim et al., 2017; Schiebel & Hemleben, 2017; Siccha & Kucera, 2017; Brummer & Kucera, 2022; Meilland et al., 2022; Mikrotax database). Hence, in this study, the biostratigraphic importance of analyzing small-size fractions is highlighted. For more time-efficient research, it is recommended to keep frequently used databases such as PANGAEA, Microtax, and WoRMS up to date.

Likewise, Kanbur & Öğretmen (2022) presented G. bermudezi, which was the first record from the Calabrian stage (Early Pleistocene) of the Mediterranean region. Globorotalia bermudezi was regarded as a distinct species by Kennett & Srinivasan (1983) and Aze et al. (2011), but Brummer & Kucera (2022) consider it a variant of G. scitula. Additionally, as an ecological parameter, according to Rögl & Bolli (1973), G. bermudezi lives closer to the surface than Globorotalia scitula. Reiss et al. (1971) described Globorotalia sp., which is identical to G. bermudezi from the Mediterranean Late Pleistocene to Recent (Rögl & Bolli, 1973).

The first report of Cribroelphidium ustulatum was presented by Öğretmen et al. (2022) from the Mediterranean region. The presence of the species in the Mediterranean Region is puzzling. Cribroelphidium ustulatum, a cold water species in the Arctic region ranging from Pliocene to Pleistocene (Gregory & Bridge, 1979; Feyling-Hanssen, 1980; Knudsen & Asbjörnsdóttir, 1991; among others), was collected from 2200 m of water depth in the surface sediments of the eastern Mediterranean Sea. The reason for the occurrence of the species might relate to upwelling cold-water influx to the Mediterranean Basin from the North Atlantic pre-Holocene and displacement/reworking of the species to the depositional setting.

Although Turborotalita clarkei, Berggrenia pumilio, Globorotalia bermudezi and Cribroelphidium ustulatum were included in the dataset from the eastern Mediterranean by Kanbur & Öğretmen (2022) and Öğretmen et al. (2022), they have not been reported from the western Mediterranean so far. In particular, the absence of Turborotalita clarkei, whose relative abundance exceeds 25% at some levels in Kargı Section (S2, S5, S6) and provides the accurate paleoecological assessment as presented in this study, arouses curiosity about whether the reason for its absence is ecological factors or exclusion due to small size fractions (<125-μm) not sampled/represented in previous studies (e.g., Pérez-Folgado et al., 2003; Béjard et al., 2024 among others in the western Mediterranean). In another case, such as Amore et al. (2000, 2004), paleoecological conditions can be re-investigated. By analyzing the >106-μm residue, they detected high abundances of T. quinqueloba at some levels of the gravity logs but no T. clarkei.

Since many previous studies focused on the >125-µm size fraction, little attention has been given to T. clarkei, one of the smallest planktic foraminifers. Boltovskoy (1991) reported: “Although it (i.e. Turborotalita clarkei) is widely distributed throughout the world, even dominating the assemblages of many areas (at least in the Quaternary deposits), it is missing from numerous reports based on materials where it must have been present.” Because adult Turborotalita clarkei dominate the 63–125-μm size fraction (Chernihovsky et al., 2018; Kanbur & Öğretmen, 2022), excluding the taxa of the <125-μm size fraction might cause a significant knowledge gap for ecological studies. It is stated by Sen Gupta et al. (1987) that “in ecological investigations, a significant amount of information on species diversity and dominance may be lost when only the >125-μm size fraction is considered.” This suggestion is also in agreement with a study about small-sized pteropod distribution in the eastern Mediterranean by Beccari et al. (2023). They concluded that neglecting the small size fraction (i.e., >63-µm) may result in a remarkable (up to 50–60%) underestimation of the relative abundance and may significantly affect paleoenvironmental reconstructions.

Turborotalita clarkei is common in tropical, subtropical latitudes and transitional temperate zones (Schiebel & Hemleben, 2017), as well as in hypersaline environments like the Gulf of Aqaba (Levy et al., 2023). Turborotalita quinqueloba also lives in low latitudes (Oberhänsli et al., 1992; Schiebel & Hemleben, 2017). According to Chernihovsky et al. (2018), T. clarkei is an asymbiotic and omnivorous species. In term of living depth, as reported by Rebotim et al. (2017) from the eastern North Atlantic, T. clarkei is a subsurface dweller with an average depth of 218 m. It was considered by Chernihovsky et al. (2018) a deep-dwelling (>200 m depth) species. Levy et al. (2023) tested T. clarkei at depths of 120–570 m in the Gulf of Aqaba, meaning that T. clarkei lives at different depths in different seas. Turborotalita quinqueloba is an asymbiotic heterotrophic species (Takagi et al., 2019). It was considered a subsurface dweller with an average living depth of 144 m by Rebotim et al. (2017) and a surface/subsurface dweller in low latitudes by Oberhänsli et al. (1992) and Schiebel & Hemleben (2017). The highest fluxes of T. clarkei emerge between late fall and early spring with increased Chl-a concentrations (Chernihovsky et al., 2018). Turborotalita. quinqueloba is a temperate and cold-water species (Amore et al., 2004; Jonkers & Kucera, 2015) with ecological preferences mostly linked to cold and very productive surface waters (Rohling et al., 1993; Rasmussen & Thomsen, 2012).

Based on cluster analysis including the <125-μm size fraction reported in Figure 3A, three main clusters are dominated by (1) Turborotalita assemblage, (2) Turborotalita, and Globigerinita glutinata assemblage, and (3) Orbulina universa. Here, the Turborotalita assemblage consists of T. clarkei and T. quinqueloba. On the other hand, the PCA analysis has revealed that only one paleoecological factor (PC) shows a positive correlation with Turborotalita but that Orbulina universa is weighted negatively for PC (Figs. 3B, C). Both species of Turborotalita are asymbiotic deep dwellers, related to cold water and high productivity. On the other hand, Orbulina universa, with which PC is negatively correlated, is a tropical-subtropical species (Jonkers & Kucera, 2015) and a mixed layer dweller (Chaisson & Ravelo, 1997). O. universa is a symbiont-bearing species and has high potential for photosynthesis (i.e. productivity) (Spero and Parker 1985; Saraswati 2021). Orbulina universa is considered by Amore et al. (2004) as a warm water species. If we look at ecological preferences of Turborotalita and Orbulina universa as a whole: Turborotalita is asymbiotic deep dweller and related to cold water and high productivity, on the other hand Orbulina universa is symbiotic mixed layer dweller and related to warm water and high productivity. According to PCA analysis, PC is positively correlated with Turborotalita and negatively with Orbulina universa. In other words, Turborotalita and Orbulina universa are inversely correlated with each other. Meaning that, the paleoecological factor controlling the marine setting is primarily influenced by cold temperature rather than productivity.

In conclusion, the only paleoecological factor controlling the Kargı marine setting that is positively correlated with Turborotalita dominance is suggestive of cold temperature. Because both T. clarkei and T. quinqueloba are asymbiotic deep dwellers related to high productivity and show a positive correlation with PC, nutrient availability might be another paleoecological factor controlling the marine setting.

Although Turborotalita clarkei and T. quinqueloba are the dominant species and both have an average relative abundance of over 18% when the >63-μm size fraction is analyzed, Turborotalita clarkei is absent and the relative abundance of T. quinqueloba significantly decreases (<2%) if the 63–125-μm size fraction is excluded. Likewise the relative abundance of Globigerinita glutinata, which is another dominant species and has an average relative abundance of over 18% when the >63-μm size fraction is analyzed, significantly decreases if the 63–125-μm size fraction is excluded. Only in S8 does its relative abundance exceed 5%. The relative abundance of Globoturborotalita rubescens exceeds 5% in S3 when the >63-μm size fraction is analyzed, but its relative abundance decreases significantly if the 63–125-μm size fraction is excluded, never reaching 5% at any level (Tables 2–4).

Excluding the 63–125-μm size fraction (i.e., Turborotalita clarkei, T. quinqueloba, and Globoturborotalita rubescens) results in two main clusters that are dominated by (A) Orbulina universa, and (B) Orbulina universa and Globigerinoides (Fig. 4A). The Globigerinoides group consists of Globigerinoides elengatus and Globigerinoides obliquus. Based on PCA assessment, Kargı marine setting was controlled by only one PC (Fig. 4B). Accordingly, PC shows a positive correlation with Globigerinoides and Orbulina universa, but no species is weighted negatively for PC (Fig. 4C).

Globigerinoides elongatus is considered the best indicator of warm waters since its maximum abundance was recorded during interglacial periods (Bonfardeci et al., 2018). Globigerinoides obliquus is a tropical, mixed-layer, shallow-dwelling species preferring temperatures >24°C and low seasonal changes in sea surface temperature as well as vertical temperature gradients (Bali et al., 2020). Orbulina universa is a tropical and subtropical species (Jonkers & Kucera, 2015) and mixed layer dweller (Chaisson & Ravelo, 1997) that tolerates a wide range of salinity and temperature (Bijma et al., 1990; Schiebel & Hemleben, 2017). It is considered by Amore et al. (2004) to be a warm water species. After assessing the life preferences of Globigerinoides group and Orbulina universa, both are temperature-related mixed layer taxa. Accordingly, when 63–125-μm size fraction is excluded, the controlling paleoecological parameter for the Kargı section is suggestive of warm temperature.

It is assumed that the expanded PCA that includes the 63–125-μm size fraction, the paleoecological preferences of Turborotalita, and the paleoecological characteristics of the Early Pleistocene (Calabrian) stage of the eastern Mediterranean overlap. The fact that Turborotalita clarkei and T. quinqueloba are deep water and productivity-related species is consistent with the oligotrophic character of the eastern Mediterranean Sea. Because the surface layers are generally almost nutrient depleted and the deep water has higher nutrient values, it is reasonable that its dominant planktic dwellers are small-sized deep water species that can access the nutrients in deep water (Zarkogiannis et al., 2020). Moreover, the dominance of the asymbiotic species (i.e., both Turborotalita clarkei and T. quinqueloba-symbiont-barren and Globigerinita glutinata-facultative symbiont-bearing) in the oligotrophic Kargi marine setting supports the suggestion that the percentage of the annual flux of the symbiont-bearing species can be a possible indicator of the degree of annual upper ocean oligotrophy (Avnaim-Katav et al., 2020). Actually all these assumptions need more consideration of alternative hypotheses.

The species in the 63–125-μm size fraction with greater than 5% relative abundance (i.e., the sum of Turborotalita clarkei, T. quinqueloba, and Globoturborotalita rubescens) exceed 61%. When including Globigerinita glutinata, whose relative abundance is largely in <125-μm fraction, these species (i.e., Turborotalita clarkei, T. quinqueloba, Globoturborotalita rubescens, and Globigerinita glutinata) represent >75% of the total assemblage (Table 3). In brief, the small-sized species dominate the Kargı marine setting. According to Zarkogiannis et al. (2019, 2020), in ultra-oligotrophic regions smaller body sizes provide a reduction in energy needs and negative buoyancy to organisms to descend into nutrient-rich depths or deeper Deep Chlorophyll Maximum.

Turborotalita clarkei is common in tropical and subtropical latitudes and also in transitional temperate zones (Schiebel & Hemleben, 2017). Turborotalita clarkei is not a typical warm water species and lives in a wide range of latitudes. Siccha (2009) generated transfer functions to reconstruct the paleoclimate in the Red Sea by characterizing the relationship between planktic foraminiferal assemblages and chlorophyll-a concentration, which is a proxy for productivity. He took into account the >150-µm size fraction to determine the planktic foraminiferal assemblages in the basin. Chernihovsky et al. (2020) reported that Turborotalita clarkei is the most dominant species in the oligotrophic Gulf of Aqaba (Red Sea) and constitutes 60% of the total planktic foraminifer assemblages in the >63-µm size fraction. They also suggested that Turborotalita clarkei fluxes increase gradually throughout the winter, reaching an annual peak value in spring coeval with maximal chlorophyll-a concentrations. This means that Chernihovsky et al. (2020) verified that primary productivity determines the distinct gradient in the foraminifer community in the Red Sea. But Siccha’s (2009) transfer functions were established by excluding the <150-µm size fraction, meaning that he ignored the dominant species in the basin according to Chernihovsky et al.’s (2020) findings. It is recommended that paleoecological transfer functions, which were used effectively in extreme environments such as the Red Sea (Siccha, 2009), should be revised by including small-sized (<125-µm) individuals along with larger size fractions.

The most important reason for focusing on the >125-µm size fractions in many previous studies was to exclude the juvenile forms and avoid high percentages of indeterminate specimens. However, this approach for the ultra-oligotrophic environments might cause researchers to lose small-sized abundant species such as Turborotalita clarkei, T. quinqueloba, Globoturborotalita rubescens, and Globigerinita glutinata. Turborotalita clarkei was not included in the datasets of paleoenvironmental and paleoclimatological studies such as Violanti et al. (1991), Amore et al. (2000, 2004), and Öğretmen et al. (2018) from the western and eastern Mediterranean Sea, which did not use >63-µm residue in their quantitative analysis.

Moreover, Turborotalita quinqueloba, which is commonly underrepresented by analyzing the >125-μm size fraction, is an important proxy for surface productivity during glacial periods (Sprovieri et al., 2003; Capotondi et al., 2004; Di Donato et al., 2015) and exceeds 34% at one level (S2) and ∼20% in average throughout the Kargı Section (Calabrian Stage, Early Pleistocene).

Selected small-sized (<125-µm) foraminifers that were identified from the Kargı section including planktics Globoturborotalita rubescens, Turborotalita clarkei, T. humilis, T. quinqueloba, Berggrenia pumilio, and Globorotalia bermudezi along with the benthic foraminifer Saidovina karreriana are presented in Figure 5. Note that Globoturborotalita rubescens and Saidovina karreriana are biostratigraphic marker species of the Mediterranean region (Vaiani & Venezia, 1999; Lirer et al., 2019). Although the largest individuals are selected for the best visual representation in Figure 5, the species identified from Finike Seamount, such as Globigerinita glutinata and G. uvula, measure slightly over 125-μm, and the size of the benthic foraminifer Cribroelphidium ustulatum is markedly less than 150-μm.

In this study, multivariate statistical analyses were applied to compare the paleoenvironmental interpretations based on the inclusion of smaller forams to interpretations using only the bigger size fraction in the Kargı marine setting where small-sized species dominate. Two cases were investigated: 1) foraminifers >63-µm, and 2) foraminifers >125-µm. In both cases, statistical analyses yield only one PC controlling the Kargı marine setting. When the small size fraction is included, the PC shows a positive correlation with Turborotalita and a negative correlation with Orbulina universa. In another case (i.e., excluding the size fraction <125-μm), the PC is positively correlated with Globigerinoides and Orbulina universa, but no species is weighted negatively for the PC. If we compare both cases, the PCA yields opposite positive and negative correlations for the paleoecological approach. When the size fraction 63–125-μm is included, the paleoecological factor is suggestive of cold temperature (or nutrient availability), but when the small fraction (63–125-μm) is excluded the PC yields the opposite paleoecological factor (warm water) and does not accurately reflect the marine conditions, showing that small planktic foraminifer species play a determinant role in paleoecological assessments.

I am grateful to Dr. N. Öğretmen for her encouragement, enlightening discussions, and useful comments provided during the course of this manuscript. I would like to thank Dr. S. Akyürekli who assisted during the SEM photography of the samples. I am grateful to Dr. M. Robinson whose comments helped me improve the manuscript.

Appendix 1. Quaternary Planktic Foraminifers found in the eastern Mediterranean Sea, their ecological requirements, and the references for their distribution in the study area. The Appendix can be found linked to the online version of this article.