The maximal diameter of the specimens (MDS) and the number of chambers, both related to test size, are weakly positively correlated (Supplementary Figure S1; R ² = 0.17, p < 0.0001). Because this study focuses on the entire test measure, the MDS was rather used than the number of chambers. The MDS varies between 149 and 300 µm (Figure 4). The calcite volume (varying from 2.9 E⁺⁰⁵ to 2.8 E⁺⁰⁶ μm³) and the calcite surface area (varying from 1.4 E⁺⁰⁵ to 9.1 E⁺⁰⁵ μm²) increase sharply with the MDS (Figure 4A; R ² = 0.76, p < 0.0001, and Figure 4B; R ² = 0.65, p < 0.0001, respectively). Well-fitted polynomial functions are also observed between the calcite volume (Figure 4A) and the calcite surface area (Figure 4B) with the MDS. The number of pores is scattered (varying from 356 to 14,559), but increases significantly with the MDS (Figure 4C; R ² = 0.16, p < 0.0001). Most of the specimens have <10,000 pores, except four specimens from the early 21st century showing higher values (Figure 4C; Supplementary Table S2). The average thickness indicates scattered values (varying from 2.71 to 11.71 µm), besides a weak but significant increasing correlation with the MDS is observed (Figure 4D; R ² = 0.05, p = 0.01).

The calcite volume indicates no significant correlation with the average thickness (Figure 5A; R ² = 0.02, p = 0.17). The calcite surface area displays a significant decreasing linear correlation with the average thickness (Figure 5B; R ² = 0.36, p < 0.0001). The calcite surface area and calcite volume show a significantly increasing correlation (Figure 5C; R ² = 0.21, p < 0.0001). Then, the calcite SV ratio displays a significant decreasing linear correlation with the average thickness (Figure 5D; R ² = 0.54, p < 0.0001). Well-fitted polynomial functions are noted between the calcite surface area (Figure 5B; R ² = 0.49, p < 0.0001) and the calcite SV ratio (Figure 5D; R ² = 0.67, p < 0.0001) with the average thickness. Interestingly, an increasing correlation is found between the average thickness calculated from the BoneJ plugin and the calcite VS ratio used as an indicator of thickness by Zarkogiannis et al. (2020) (Supplementary Figure S2; R ² = 0.52, p < 0.0001).

The number of pores in the entire test shows a significantly increasing linear correlation with the calcite surface area (Figure 6A; R ² = 0.74, p < 0.0001) and a well-fitted polynomial function (Figure 6A; R ² = 0.78, p < 0.0001). The pore density indicates a significant decreasing linear correlation with the average thickness (Figure 6B; R ² = 0.48, p < 0.0001) and also a well-fitted polynomial function (Figure 6B; R ² = 0.65, p < 0.0001). Weak but significant decreasing correlations are found between the pore density and the calcite volume (Figure 6C; R ² = 0.04, p = 0.03, and polynomial function; R ² = 0.06, p = 0.03), and a highly significant increasing correlation between the calcite SV ratio and the pore density (Figure 6D; R ² = 0.69, p < 0.0001).

Even if the adult specimens come from the same size fraction, the MDS affects the morphology of the entire test but not to the same extent (Figure 4). The calcite volume (Figure 4A) and calcite surface area (Figure 4B) are highly correlated with the size of the specimens, which was already demonstrated in previous studies (Belanger, 2022 and references therein). Because the average thickness is weakly affected by the MDS (Figure 4D), this parameter could be mainly influenced by environmental factors such as varying salinity, pH, or temperature. Conversely, a non-negligible correlation between the number of pores and the MDS was demonstrated (Figure 4C). Comparisons with previous studies are difficult since the pore pattern is species-specific and has only been performed on small parts of the tests from 2D images (Petersen et al., 2016 and references therein). Interestingly, four specimens from the early 21st century may be considered outliers regarding their very high number of pores (Figure 4C). Several hypotheses may explain these outliers; 1) a threshold value (14,559 pores, Supplementary Table S2) because the number of pores can be limited by the robustness of the test and the metabolic demands of the cell (Richirt et al., 2019), 2) an over-estimated number of pores linked to traces of dissolution in some damaged specimens that may generate additional holes, and 3) taphonomic effects that may increase test porosity (Oakes et al., 2019). This contribution illustrates that 3D reconstructions allow quantifying the morphological parameters of tests that have calcified under different environmental conditions, and highlights the need to standardize specimens by the same MDS for more accurate comparisons.

A wide range of morphological patterns in Elphidium clavatum was observed with two distinct patterns; thinner tests have a higher calcite SV ratio (Figure 5D) i.e., a larger surface area (Figure 5B), and a higher pore density (Figure 6B). Conversely, thicker tests have a lower calcite SV ratio and a lower pore density. The well-fitted polynomial functions found between the average thickness and surface calcite area (Figure 5B), calcite SV ratio (Figure 5D), and pore density (Figure 6B), may suggest a compromise between the robustness of the test and the metabolic needs of the cell. This hypothesis can be compared to the scaling laws driving pore patterns described by Richirt et al. (2019). Here, the thickness of the test can have a major role in the pore pattern and the overall shape of the test.

Under natural conditions, it is difficult to associate the variation of morphological patterns with a single environmental factor. However, some morphological traits of foraminifera such as the thickness, SV ratio, and pore density, were previously associated with environmental stressors, allowing us to extrapolate some broad conclusions to Elphidium clavatum for palaeoecological interpretations. We expect that E. clavatum would decrease calcification (i.e., thickness loss) in response to ocean acidification (OA). Indeed, thinner parts of small benthic foraminiferal tests are commonly observed in culture experiments at lower pH values (Allison et al., 2010; Dissard et al., 2010; Haynert et al., 2011). Moreover, the thinning of the entire test (i.e., outer and inner walls) related to OA is also demonstrated in 3D imaging foraminiferal studies (Supplementary Table S1). Ocean acidification is not the only factor that can lead to test thinning. The combined impact of OA and lower salinity may induce a synergistic effect on the calcification process, decreasing resistance to dissolution (Charrieau et al., 2018a; 2018b). Moreover, thinner walls intensify gas exchange in low-oxygen environments (Bernhard, 1986; Sen Gupta and Machain-Castillo, 1993; Kaiho, 1994). The combined effects of warming and OA may have an antagonist effect on calcification due to the positive effect of increasing temperature on calcification and growth (Haynert and Schönfeld, 2014). Consequently, warmer temperatures may also increase the variability in test thickness.

The assumption would be that E. clavatum would have more flattened tests, i.e., a higher calcite SV ratio, in response to deoxygenation and pollution. Some benthic foraminiferal species adapt their tests with flattened shapes to maximize more surface area per unit volume in low-oxygen environments (Bernhard, 1986; Sen Gupta and Machain-Castillo, 1993; Kaiho, 1994). A higher SV ratio can be also associated with a decreasing roundness or an increasing test asymmetry, previously interpreted as an adaptive response toward environmental stress or more variable environmental conditions (Leung et al., 2000; Weinkauf et al., 2014). A decrease in roundness may be also associated with morphological abnormalities. Deformed tests are reported in areas subject to different types of pollution e.g., heavy metals (Alve, 1991), and hydrocarbons (Morvan et al., 2004) but also from areas with a large gradient of salinity such as brackish conditions (Charrieau et al., 2018b). However, lower SV ratios are also observed in benthic foraminifera from high-latitude regions due to the increased volume and size of specimens, probably related to the availability of organic matter even in low-oxygen environments (Belanger, 2022). Therefore, food availability may also increase the calcite SV ratio variability because of a larger calcite volume.

Elphidium clavatum would increase its pore density in response to deoxygenation. In previous studies, correlations are observed between the increase in pore density with lower dissolved oxygen concentrations in the surrounding water (Kuhnt et al., 2013, 2014). Some studies describe that a flattened and thin test facilitates gas exchange by diffusion through the pores by minimizing oxygen consumption and increasing oxygen uptake efficiency (Bradshaw, 1961; Corliss, 1985; Sen Gupta and Machain-Castillo, 1993; Glock et al., 2019). In some benthic foraminifera species from oxygen minimum zones, positive relationships between pore density and temperature, and between pore density and bottom water [NO₃ ⁻] are demonstrated (Glock et al., 2011; Kuhnt et al., 2013). However, these relationships are species-specific and require further investigation in Elphidium clavatum. In this study, the thinnest tests have a higher pore density (Figure 6B), thus OA may have a synergistic effect with low-oxygen conditions on the pore density. Some authors argue that deoxygenation could lead to higher porosity, i.e., the percentage of the test surface covered by pores (Richirt et al., 2019). Achieving test porosity with 3D imaging remains a challenge, in this contribution, the pore area is visually highly variable between specimens and cannot be studied without robust statistical methods taking into account the variability in test thickness.

Although the morphological variability is large, significant long-term trends in morphological changes over the last 200 years can be noted, especially in the average thickness, number of pores, calcite volume, and calcite surface area (Figure 7). We computed the decrease in average thickness, and the increase in the number of pores, calcite volume, and surface area from the modern foraminifera in ∼2013 compared to their historical counterparts in ∼1807 (details in Supplementary Table S4). The modern specimens reveal a thickness loss of 28 ± 14% (n = 18), an increase of 35 ± 11% in calcite surface area, an increase of 15 ± 4% in calcite volume, and an increase of 91 ± 67% in the number of pores. These long-term trends can be interpreted as the result of gradual environmental changes in the Baltic Sea entrance. Fox et al. (2020) demonstrate a larger reduction in shell thickness of up to 76% in the planktonic foraminifera Neogloboquadrina dutertrei over the last ∼140 years in the Pacific ocean. These authors also find a thickness loss of ∼20% in Globigerinoides ruber (Fox et al., 2020), corresponding to a similar result for Elphidium clavatum. Globigerinoides ruber is known to display a mechanism of resistance to OA linked to photosynthetic algal symbionts (Fox et al., 2020 and references therein). The same mechanism of resistance for Elphidium clavatum cannot be applied, as they are living in the aphotic zone. Putative mechanisms of resistance to OA in non-photosynthetically benthic foraminifera from high-latitude regions need to be further investigated.

During the early industrial period referring to the period from ∼1807 to 1939 in our historical context, the BFAR of Elphidium clavatum remained stable and low (<44 specimens cm⁻² yr⁻¹) (Figure 7, Charrieau et al., 2019). From the total foraminiferal assemblage in Charrieau et al. (2019), two subzones were described ∼1807–1873 and ∼1873–1923, with associated environmental conditions characterized by low oxygen conditions, a salinity of ∼30, and the onset of human-induced impacts with various types of pollution (Zillén et al., 2008; Charrieau et al., 2019; Ljung et al., 2022). No significant difference was found in the morphology of the Elphidium clavatum specimens between the two subzones, however, the variations in both BFAR and the morphological parameters are not synchronized. Although the BFAR remained unchanged, a large variability in test thickness, calcite SV ratio, and pore density can be observed in ∼1873. The relative stability of the morphological patterns excepted in ∼1873 indicates a pivotal period, already noted by Charrieau et al. (2019), suggesting wider variability in pH and [O₂] values. In summary, during the early industrial period, although the BFAR of Elphidium clavatum remained unchanged, the morphological variations instead reveal the natural variability of environmental conditions in the region.

The ∼1939–2002 period corresponds to the intensification of human activities in the region, such as the massive increase of carbonaceous pollution from petroleum products for energy use (Ljung et al., 2022), and excess nutrient loading from terrestrial to marine environments (Gustafsson et al., 2012). This period was marked by favorable growth conditions of Elphidium clavatum with two successive sharp increases in the BFAR up to 826 and 1,247 specimens cm⁻² yr⁻¹ in ∼1939–1978 and ∼1986–2002 respectively (Charrieau et al., 2019). These BFAR variations were previously associated with the increase in organic matter as a food source for foraminifera, and major changes in the current and sediment pattern (Charrieau et al., 2019 and references therein). Interestingly, there is a synchronicity between both increases in BFAR and morphological variability, especially for the average thickness, calcite SV ratio, and pore density. Further, the decrease in the BFAR from the short period ∼1978–1986 (276 specimens cm⁻² yr⁻¹) was associated with the improved environmental conditions to reduce eutrophication in the region (Carstensen et al., 2006; Conley et al., 2007; Charrieau et al., 2019). A decrease in variability for all the morphological parameters is observed in ∼1986, suggesting that the specimens calcified in less contrasted environmental conditions. In summary, the historical record of Elphidium clavatum reveals the intensification of anthropogenic activities through the synchrony between high reproductive success and broad morphological diversification and conversely reduced variability in test morphology that may be associated with a short event of improved environmental conditions.

The early 21st century foraminiferal record reveals sharp contrasts between the BFAR and morphological variations in Elphidium clavatum. The associated environmental conditions during this period were characterized by low oxygen conditions, high organic matter content, and open ocean salinity (Charrieau et al., 2019). Particularly in ∼2002, E. clavatum dominated the fauna (Charrieau et al., 2019), but the specimens are the most negatively affected over the last 200 years: the average thickness compared to those from ∼1807 has decreased by 36 ± 17% (n = 18), the calcite surface area has increased by 63 ± 21%, and the number of pores has increased by 151 ± 120% (Supplementary Table S4). Moreover, in ∼2002 the largest variability in the calcite SV ratio and pore density is observed for the whole record (Figure 7). These results can be attributed to the larger seasonal hypoxia event recorded in the Danish Straits in 2002, explained by the combination of bottom water transport, nutrient supply from land, and rising temperature (Conley et al., 2007). By contrast in ∼2005, the morphological variability is lower and the specimens significantly differ from the general patterns observed around the 21st century (Figure 7). Especially, the specimens are thicker 11 ± 6% (n= 15), with a larger calcite volume of 77 ± 15% and a lower number of pores of 54 ± 26% compared to those from the ∼1807 (Supplementary Table S4). These unexpected results may be related to a massive inflow of highly saline, cold, and extremely oxygen-rich water from the North Sea, called Major Baltic Inflows, affecting occasionally the deep basins of the Baltic Sea and reported in 2003 (Lehmann et al., 2004; Feistel et al., 2006). Then, from the 2010s, a desynchronization is notable between the lower and stable BFAR (∼225 specimens cm⁻² yr⁻¹) and the large morphological variations of all parameters (Figure 7). The persisting environmental stressors i.e., warming, hypoxia, and OA since the 1940s, in addition to the possible inter-specific competition with opportunistic species such as Nonionella sp. T1 and Nonionoides turgidus (Charrieau et al., 2019), would not allow Elphidium clavatum to combine high reproductive success with a wide diversification of its morphological patterns. Recently, Bernhard et al. (2021) demonstrated in a triple-stressors experiment with propagules that Elphidium cf E. excavatum indicates high abundance under pre-industrial and cold acidified conditions, low abundance in present-day and cool + OA + hypoxic conditions, and absence in warm + OA + hypoxic conditions, indicating that Elphidium clavatum and probably other species from high-latitude regions will be challenged in the next decades.