INTRATEST TRACE ELEMENT VARIABILITY IN POLAR AND SUBPOLAR PLANKTIC FORAMINIFERA: INSIGHTS INTO VITAL EFFECTS, ONTOGENY, AND BIOMINERALIZATION PROCESSES

The trace element geochemistry (TE/Ca) of foraminifera varies as a function of seawater composition, the physical conditions of the water in which they are calcifying, and the biological conditions during calcification (e.g., Lea et al., 1999; Erez, 2003; Russell et al., 2004; Katz et al., 2010; Allen et al., 2016; de Nooijer et al., 2017; Evans et al., 2018). Thus, great efforts have been made to quantitatively constrain relationships between foraminiferal test TE/Ca geochemistry and the conditions of the foraminifer’s living environment as a means for paleoenvironmental reconstruction. While some foraminiferal TE/Ca paleoproxies have been rigorously studied, such as Mg/Ca as a paleotemperature proxy (Nürnberg, 1995; Nürnberg et al., 1996; Lea et al., 1999; Anand et al., 2003; Russell et al., 2004; Martínez-Botí et al., 2011; Hönisch et al., 2013; Holland et al., 2020), incorporation of other TEs into foraminiferal calcite is less well-understood and represents new pathways of needed exploration. In addition to understanding the environmental controls on TE incorporation, it is equally important to constrain how biological processes modulate TE geochemistry. Biological processes that influence test geochemistry, broadly referred to as “vital effects” (Niebler et al., 1999; de Nooijer et al., 2014; Schiebel & Hemleben, 2017), can distort primary environmental signals, thus interfering with paleoreconstructions reliant upon test geochemistry. For example, trace element geochemistry may differ from chamber-to-chamber or when examining profiles of TEs across a chamber wall (Eggins et al., 2003; Steinhardt et al., 2014, Steinhardt et al., 2015; Davis et al. 2020; Hupp & Fehrenbacher, 2023). Furthermore, ontogenetic processes, such as test thickening observed in several species of planktic foraminifera syn- and post-gametogenesis, have been shown to impart different geochemical signatures in different portions of the test wall (see table 1 of Hupp & Fehrenbacher, 2023, and references therein). Recognizing intratest variability is informative to identifying potential vital effects as well as providing insight into ontogeny and biomineralization processes (Eggins et al., 2003; Hathorne et al., 2003; Davis et al., 2020). Fossilized test TE/Ca variability can be further confounded by the addition of diagenetic calcite, which varies even more from primary foraminiferal calcite compositions (e.g., Boussetta et al., 2011; Kozdon et al., 2013; Edgar et al., 2015).

Recent advances in in situ technologies, such as laser-ablation inductively coupled plasma mass spectrometry (LA-ICP-MS), provide a unique opportunity to investigate intratest trace element variability. LA-ICP-MS analysis allows for quantifying TEs in individual foraminifera in a population of specimens (i.e., intertest variability). By analyzing micron-scale intratest variability using LA-ICP-MS, one can also identify potential vital effects and investigate other mechanisms responsible for trace element incorporation in foraminifera (Eggins et al., 2003; Hathorne et al., 2003; Sadekov et al., 2010; Davis et al., 2020). This is necessary to identify intertest and intratest TE variability that potentially confound “averaged” TE measurements used in paleoreconstructions, as is the case in more traditional, pooled, whole-test ‘bulk’ analytical techniques, such as solution analysis.

LA-ICP-MS is commonly used to investigate TE geochemistry in modern planktic foraminiferal species (e.g., Hathorne et al., 2009; Steinhardt et al., 2015; Davis et al., 2020), however these investigations predominantly focus on tropical, subtropical, and transitional species. Here, we explore the intratest variability of polar and subpolar modern planktic foraminiferal species, specifically Neogloboquadrina incompta (Cifelli, 1961), Neogloboquadrina pachyderma (Ehrenberg, 1861), and Turborotalita quinqueloba (Natland, 1938), to bolster their usage as geochemical paleoproxy substrates. Specimens investigated in this study were collected from plankton tow samples throughout the Northern California Current region under a wide range of environmental conditions between 2011 and 2022. All specimens were dead upon collection, as identified by their thickly crusted test walls, and in recently collected samples, the absence of remnant cytoplasm. Only encrusted specimens were examined, as the presence of crust indicates the end or near-end of an individual’s life cycle, and thus are likely more representative of fossilized specimens found on the seafloor (Bé & Lott, 1964; Bé, 1980; Caron et al., 1990; Davis et al., 2020). By characterizing the intratest variability of crusted (i.e., dead) specimens collected across a wide range of conditions in the uppermost surface ocean, we aim to provide a foundation from which future trace element studies focused on polar and subpolar planktic foraminifera collected from sediment traps and the seafloor can be compared. Furthermore, this study illustrates the wide range of approaches that can be taken to investigate foraminiferal geochemistry when using LA-ICP-MS data. This study provides insight into vital effects and trace element incorporation into the tests of polar and subpolar planktic foraminifera taxa independent of potential trace element heterogeneity associated with travel through the water column (e.g., scavenging; Davis et al., 2020) or early diagenesis upon the seafloor (e.g., Branson et al., 2015).

Study specimens were collected via surface plankton tow from offshore sites throughout the Northern California Current (NCC) region, ranging from northern California to the border between Washington (USA) and Canada (Fig. 1). This region of the Northeast Pacific Ocean is characterized by a relatively narrow, deep continental shelf (0 to 200 m) and a steep continental slope that transitions into deep (>2500 m) water 25–50 km offshore. A well-developed seasonal oxygen minimum zone (OMZ, 600–1000 m) associated with North Pacific Intermediate Water exists offshore (e.g., Paulmier & Ruiz-Pino, 2009; Pierce et al., 2012). As in many other eastern boundary ocean margins, upwelling along the margin is a highly seasonal phenomenon driven by shifts in wind strength and direction (Barth & Wheeler, 2005). During spring and summer, this region experiences strong upwelling caused by prevailing northerly winds that induce Ekman transport of deeper water from the shelf break onto the continental shelf (e.g., Kirincich et al., 2005; Perlin et al., 2005). Fall and winter display starkly different conditions, with southerly, downwelling-favorable winds, large swells, high wave energies nearshore, and the arrival of moisture-laden storms (Kniskern et al., 2011). Environmental variability due to seasonal upwelling is further overprinted by longer-term oscillations such as El-Niño Southern Oscillation and the Pacific Decadal Oscillation. Collectively, these features create a dynamic region with diverse conditions under which regional planktic foraminifera calcify (Ortiz & Mix, 1992; Taylor et al., 2018; Lane et al., 2023).

Plankton tows were collected from several sites (Fig. 1) during a series of cruises conducted between 2011 and 2022. Vertical tows were conducted from 100 m depth to the surface using a 150-μm mesh net. Most tows were obtained during NOAA-led fisheries survey cruises, where samples were preserved using NaOH-buffered formalin immediately upon collection and stored in a sample archive (Feinberg & Peterson, 2003). Planktic foraminiferal tests were removed from archived tows, rinsed with DI water upon removal, and stored on micropaleontological slides until further cleaning was conducted in preparation for geochemical analysis. In addition to tests from archived tows, additional tows were collected by the authors where tests were immediately removed from tows, rinsed with DI water, and stored on micropaleontological slides, without any exposure to preservative chemicals. In total, specimens examined herein originate from 39 discrete tows collected across 29 different sites over a period of 11 years. Metadata for each test examined, including tow location (site name, latitude, longitude) and archived (preserved in buffered formalin) versus recently collected (never exposed to preservatives) status is detailed in Appendix 1.

Specimens of N. incompta (N = 92), N. pachyderma (N = 55), and T. quinqueloba (N = 42) were selected for individual foraminifer geochemical analysis. All three study taxa are asymbiotic; however, T. quinqueloba is a spinose species whereas the neogloboquadrinids are non-spinose (Schiebel & Hemleben, 2017). Specimens selected for analysis were limited to heavily-calcified tests (i.e., clearly apparent crusts) where it is inferred or confirmed that the specimens were dead upon collection. For archive specimens, dead specimens were inferred based upon evidence of a clearly apparent gametogenic crust (see fig 1. in Hupp & Fehrenbacher, 2023, for further description of crusted versus uncrusted individuals). For the recently collected specimens, dead specimens were heavily-calcified and lacked cytoplasm.

Study specimens underwent an oxidative cleaning process to remove remnant organic matter and detritus (Mashiotta et al., 1999) that consisted of: 1) rinsing the tests with methanol, 2) rinsing the tests with ultrapure milliQ water three times, 3) submerging tests in a 1:1 solution of 30% H₂O₂ buffered with 0.1 M NaOH and placing vials containing the tests and solution in a warm (∼65°C) water bath for 10 minutes, and lastly, 4) removing the H₂O₂-NaOH solution and rinsing the tests with milliQ water three times. While previous studies submerge the tests for 30 minutes, the water bath time was decreased to 10 minutes to prevent test dissolution (Bonnin et al., 2019).

Trace element geochemistry was determined via laser-ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). Tests were analyzed with a Thermo Scientific iCAP RQ quadrupole ICP-MS coupled to an Applied Spectra RESOlution (193 nm) laser-ablation system housed in the Keck Collaboratory for Plasma Spectrometry at Oregon State University. While ten isotopes were measured during analysis, seven isotopes are examined herein: ²⁴Mg,²³Na, ⁸⁸Sr, ¹³⁸Ba, ⁵⁵Mn, ⁶⁶Zn, and ⁴³Ca. Isotopes were measured using a rapid peak-hopping method where Mn, Zn, and Ba were assigned dwell times of 0.05 s, Na a dwell time of 0.04 s, Sr and Mg dwell times of 0.02 s, and Ca a dwell time of 0.01 s (total sweep time: 0.29 s). Analysis spots were ablated from the outer chamber wall (the test surface) through the chamber wall. For the neogloboquadrinids, laser energy was set to 4.0 mJ attenuated by 87.5% whereas for T. quinqueloba the laser energy was reduced to 3.0 or 3.5 mJ, also attenuated by 87.5%. Circular spot sizes range from 24 to 50 μm and were adjusted for each specimen to maximize the ablated area. Repetition rates were adjusted between 5 and 6 Hz to optimize laser profiles. Individual laser analyses were conducted on the F0, F1, F2, F3, and F4 chambers of T. quinqueloba and the F0, F1, F2, and F3 chambers for N. incompta and N. pachyderma. In rare cases, due to test orientation, the oldest chamber in the outer whorl (F4 and F3 for T. quinqueloba and the neogloboquadrinids, respectively) was not accessible to be analyzed. Repeat analyses were conducted on at least one chamber per specimen to assess reproducibility (Appendix 1).

Three standard reference materials were measured bracketing every 10 to 12 specimens (i.e., every 50 to 72 test ablation spots across 10 to 12 continuous profiles), including the glasses NIST 610 and NIST 612, and the compressed-powder USGS standard MACS-3. Each standard was ablated for 60 s using a repetition rate of 6 Hz on a spot size of 50 μm with a laser energy of 5 mJ attenuated by 50%. All TE/Ca were calibrated using the NIST 610 and 612 glasses, aside from Na/Ca which was calibrated using NIST 610 and MACS-3 due to a consistent concentration of Na within NIST glass standards. Elemental calibrations of standard reference materials measured throughout the analysis sessions yielded r² ≥ 0.99.

Data was processed with the Python package LAtools (Branson et al., 2019) following established data reduction protocols (Longerich et al., 1996). Data reduction involves despiking and signal smoothing to remove outliers in the laser spectra, evaluation and correction for any drift observed in the standard measurements throughout the analysis session, and removal of average background counts from each data point. Data output from this process produces individual specimen files, where each spot measurement composes a distinct section of the full spectrum, as well as mean TE/Ca for each spot analysis, normalized to known TE/Ca concentrations of the standards (Fig. S1; Jochum et al., 2011). Whole-test averages and individual chamber averages of all TEs for each specimen are reported in Appendix 2. For chambers where multiple measurements were taken, the replicate measurements were averaged to produce the chamber values reported in Appendix 2.

The range of whole-test average TE/Ca compositions for the study specimens provides context for further investigation of the intratest geochemistry in these species (Fig. 2; Table 1). Whole-test averages were determined by taking the mean of the average chamber measurements from each chamber in the outer whorl (Fig. 2A). Note that where multiple measurements were taken within a single chamber, those measurements were averaged to determine a chamber mean value before being used to calculate the whole-test average. It is recognized that the whole-test average as described here is technically an interpolated average, as we are calculating the mean of several discrete measurements taken from chambers in the outer whorl as opposed to dissolving and analyzing the entire test, as done in solution analyses. However, comparing the “whole test” value as measured by solution-based ICP-MS and interpolated by laser-based ICP-MS result in negligible differences (Fehrenbacher et al., 2020). We also recognize that while much of the interspecies variability is directly tied to environmental variability during specimen collection/period of calcification, the whole test TE/Ca geochemistry provides context for the subsequent discussion of intratest TE/Ca variability. Furthermore, examination of whole-test TE/Ca distributions provides insight into elements that are potentially modulated by environmental conditions. For example, TEs that do not vary significantly may suggest that these TEs are not strongly impacted by calcification conditions (e.g., Ba/Ca in T. quinqueloba; Fig. 2E).

Here we describe the whole-test average distributions by examining the distribution mean (Table 1) as well as the distribution range, as defined by the maximum whisker value minus the minimum whisker value of the boxplots shown in Figure 2. Neogloboquadrina incompta generally records the lowest average Mg/Ca, Na/Ca, and Sr/Ca, followed by N. pachyderma, and T. quinqueloba (Figs. 2B–D). The range of the distributions is fairly consistent between all taxa for Mg/Ca (∼3 mmol/mol) and Sr/Ca (∼0.25 mmol/mol), but the Na/Ca range increases from: ∼6 mmol/mol in N. incompta to 13 mmol/mol in N. pachyderma and to 22 mmol/mol in T. quinqueloba, respectively. Interspecies TE/Ca differences are more irregular when comparing Ba, Mn, and Zn (Figs. 2E–G). While all taxa exhibit similar Ba/Ca minima, N. incompta exhibits the largest range of Ba/Ca values (41 μmol/mol), followed by N. pachyderma (16 μmol/mol), and lastly with the smallest range, T. quinqueloba (5 μmol/mol; Fig. 2E). Neogloboquadrina incompta also exhibits a larger range of Mn/Ca values (25 μmol/mol) compared to N. pachyderma (7 μmol/mol) and T. quinqueloba (11 μmol/mol; Fig. 2F). The inner quartile ranges are comparable among all three taxa for Zn/Ca, but the full range of whole test values is largest in N. pachyderma (977 μmol/mol), followed by T. quinqueloba (728 μmol/mol), and lastly, N. incompta (586 μmol/mol). Neogloboquadrina pachyderma also records higher Zn/Ca values on average than T. quinqueloba and N. incompta (Fig. 2G).

To explore how TE/Ca may change through ontogeny, chamber-to-chamber TE/Ca averages are compared to one another by mean centering the TE/Ca data (i.e., whole test average TE/Ca is subtracted from each chamber average; Fig. 3A). By normalizing each chamber average by the test average, the range of chamber-to-chamber variability can be examined while removing variability that could be more directly attributed to calcification under different environmental conditions.

For all three taxa, the Mg/Ca composition decreases in each successively added chamber in the outer whorl, with the youngest chamber (F0) exhibiting the lowest values found in the outer whorl (Figs. 3B, E, H). The F4 chamber of T. quinqueloba does not seem to follow the intratest pattern (Fig. 3H), however this could be an artifact, as not all F4 chambers could be analyzed in this taxon. Ratios of Na/Ca and Sr/Ca show minimal chamber-to-chamber variability in N. pachyderma (Figs. 3F, G) and T. quinqueloba (Figs. 3I, J). However, in N. incompta, the F0 chamber records consistently higher Na/Ca than all other chambers in the outer whorl (Fig. 3C). A consistent increase in Sr/Ca is observed with each successively younger chamber in the outer whorl of N. incompta (Fig. 3D). The ratio of Ba/Ca differs little between chambers in N. incompta (Fig. 4A), however the older chambers (i.e., F2, F3) tend to record slightly higher Ba/Ca values than the ultimate and penultimate chambers of N. pachyderma and T. quinqueloba (Figs. 4D, G). A similar pattern of decreasing TE/Ca content in each sequentially added chamber is observed with Mn/Ca in N. pachyderma (Fig. 4E) and T. quinqueloba (Fig. 4H), whereas N. incompta shows only a relative Mn/Ca depletion in the F0 chamber (Fig. 4B). The neogloboquadrinids exhibit an opposite pattern with Zn/Ca where Zn incorporation increases with each progressively added chamber (Figs. 4C, F). Conversely, T. quinqueloba shows a progressive decrease in Zn incorporation with each progressively added chamber, aside from the ultimate F0 chamber (Fig. 4I).

In addition to characterizing the chamber-to-chamber variability, the range of values measured from individual chambers within a single test is also explored (Fig. 5). The intratest range is calculated by subtracting the lowest chamber average value of those measured in the outer whorl from the highest chamber average value of outer whorl chambers in a single specimen (e.g., Fig. 5A). Here, the medians of the boxplot distributions (Fig. 5) will be discussed, as some TE/Ca exhibit several outliers influencing the distribution mean. The median intratest range of Mg/Ca values is ∼0.8 mmol/mol in the neogloboquadrinids, however the distribution of ranges exhibits a greater spread for N. pachyderma than for N. incompta (Fig. 5B). Turborotalita quinqueloba exhibits a median intratest Mg/Ca range of ∼1.6 mmol/mol. The median of the intratest Na/Ca range increases from 2.0 mmol/mol in N. incompta to 3.4 mmol/mol in N. pachyderma to 8.8 mmol/mol in T. quinqueloba, and the distribution spread of intratest ranges increase in the same taxa order (Fig. 5C). Distributions of intratest Sr/Ca ranges are very similar among the three study taxa, with a median intratest range of ∼0.1 mmol/mol (Fig. 5D). Intratest Ba/Ca ranges show an opposite pattern among taxa as observed in Na/Ca (Fig. 5E). The median of intratest Ba/Ca ratios was 6.4 μmol/mol in N. incompta, 4.7 μmol/mol in N. pachyderma, and 2.4 μmol/mol in T. quinqueloba. The boxplot distribution of intratest Ba/Ca ranges shows the highest maximum value in N. incompta (36.6 μmol/mol), followed by N. pachyderma (18.1 μmol/mol), and lastly, T. quinqueloba (9.9 μmol/mol). For Mn/Ca, T. quinqueloba shows the highest median internal variability (6.6 μmol/mol), followed by N. incompta (6.3 μmol/mol), and N. pachyderma (3.0 μmol/mol), however the distribution of Mn/Ca intratest ranges recorded by N. incompta shows the greatest spread (Fig. 5F). Conversely, the highest median range of intratest Zn/Ca is recorded by N. pachyderma (357 μmol/mol), followed by T. quinqueloba (194 μmol/mol), and N. incompta (111 μmol/mol), with N. pachyderma exhibiting the greatest variability of intratest ranges (Fig. 5G).

Intratest TE/Ca variability, characterized by the intratest range and standard deviation of average chamber measurements in the outer whorl, can be compared to the whole-test TE/Ca value of an individual specimen to determine if intratest variability fluctuates consistently with increasing or decreasing TE/Ca content (Table 2). Intratest Mg/Ca variability shows a strong (range r² = 0.67, and standard deviation or “SD” r² = 0.68) and moderately strong (range and SD r² = 0.58) positive relationship with whole-test Mg/Ca values in N. incompta and N. pachyderma, respectively. No correlation is found between intratest Mg/Ca variability and the whole-test average value in T. quinqueloba. Neogloboquadrina incompta exhibits a moderate positive correlation (range r² = 0.53, SD r² = 0.54) between intratest Na/Ca variability and the whole-test value, whereas N. pachyderma (range and SD r² = 0.36) and T. quinqueloba (range r² = 0.32, SD r² = 0.33) exhibit only weak positive correlations. No relationship is found between Sr/Ca intratest variability and the whole-test Sr/Ca value in any of the three study taxa. Whole-test Ba/Ca shows a very strong positive correlation with intratest variability in N. pachyderma (range r² = 0.98, SD r² = 0.99), and a strong positive relationship in N. incompta (range and SD r² = 0.64) and T. quinqueloba (range r² = 0.75, SD r² = 0.76). A strong positive relationship is found in N. incompta (range r² = 0.70, SD r² = 0.72) and T. quinqueloba (range r² = 0.77, SD r² = 0.78) when comparing intratest Mn/Ca variability to whole-test values, and a moderate to weak positive correlation is found in N. pachyderma (range r² = 0.41, SD r² = 0.39). Moderate positive correlations are also found between intratest Zn/Ca variability and whole-test values in T. quinqueloba (range r² = 0.62, SD r² = 0.61) and N. pachyderma (range r² = 0.51, SD r² = 0.49), with a weak correlation observed in N. incompta (range r² = 0.37, SD r² = 0.38).

Laser ablation was conducted from the outside of the test through to the inner chamber wall allowing for individual depth profiles to be used to assess how the geochemistry of the test changes through the test wall (e.g., Fig. 6). All specimens examined for this study were chosen because they appeared to be thickly calcified, and all three species are known to exhibit a gametogenic crust (Schiebel & Hemleben, 2017; Pearson & Kucera, 2018). The inner versus outer calcite TE/Ca data are compared to evaluate if trace elements vary between the inner chamber wall and outer gametogenic crust (Figs. 6, 7). For simplicity, we assume the first half of a given depth profile (the outer calcite) is dominantly composed of crust calcite; the second half of the same depth profile dominantly represents the inner chamber calcite or “lamellar” calcite (cf., Hupp & Fehrenbacher, 2023; Fig. 6B). If there is little to no difference in the average of these two values (e.g., data points sit upon a 1:1 line, Figs. 6C, 7), we assume that TEs do not vary greatly between the outer crust and the inner chamber wall. Conversely, if a chamber exhibits higher TEs in the crust or inner chamber wall, data points will either plot above or below the line of equality, respectively. While there is some variability in the amount of gametogenic calcite precipitated at the end of ontogeny, the 50% rule is an appropriate approximation (Bé, 1980; Ariwaka, 1983; Mulitza et al., 1997; Bauch et al., 2002; Simstich et al., 2003; Bolton & Marr, 2013; Hupp & Fehrenbacher, 2023). Definitively quantifying the amount of early ontogenetic versus gametogenic crust calcite in our study populations (N = 189) would require laborious analyses (e.g., thin sectioning specimens for EMPA and SEM imaging), which is outside of the scope of this study.

For all trace elements, aside from Ba, all three taxa show greater enrichment in TEs in the inner 50% of the spectrum (i.e., the lamellar calcite) compared to the outer 50% of the same spectrum (i.e., the outer crust; Fig. 7). Deviation from the line of equality is greatest in Mg/Ca, Na/Ca, and Sr/Ca, whereas the inner chamber wall shows only minor enrichments in Mn/Ca and Zn/Ca compared to the crust, aside from large deviations observed in the Zn/Ca of T. quinqueloba (Fig. 7R). Ratios of Ba/Ca are also heavily enriched (in some cases by >200%) in the inner chamber walls of the neogloboquadrinids, yet T. quinqueloba shows a minor enrichment of Ba in the crust relative to the lamellar calcite of the inner test wall due to a few high-Ba specimens (Fig. 7P).

As potential controls on TE incorporation are considered, it is helpful to evaluate patterns in TE covariance. Figures 8A–C show covariance matrices for the whole-test TE/Ca values of our study specimens, separated by taxon. This results section will focus upon discussing moderate and strong relationships, as several elements covary at r² <0.40. For N. incompta, Mg/Ca moderately positively covaries with Na/Ca (r² = 0.47) and Sr/Ca (r² = 0.45) on the whole-test average scale (Fig. 8A). Moderate positive relationships are also found between Ba/Ca and Mn/Ca (r² = 0.52) as well as Ba/Ca and Zn/Ca (r² = 0.43). For N. pachyderma, a strong positive relationship is observed between Na/Ca and Sr/Ca (r² = 0.64; Fig. 8B). Ratios of Na/Ca also moderately positively covary with Mg/Ca (r² = 0.50), Ba/Ca (r² = 0.57), and Mn/Ca (r² = 0.47). Moderate positive covariations are found between Mg/Ca and Sr/Ca (r² = 0.45), Mg/Ca and Ba/Ca (r² = 0.47), Ba/Ca and Mn/Ca (r² = 0.48), and Mn/Ca and Zn/Ca (r² = 0.40) in this taxon. Turborotalita quinqueloba exhibits the least number of correlations between trace elements, with Mg/Ca and Na/Ca exhibiting a moderate positive correlation (r² = 0.49), and Sr/Ca and Ba/Ca exhibiting a strong negative correlation (r² = -0.80; Fig. 8C).

In addition to investigating covariance in TEs on the whole-test scale, the TE covariation in individual chambers is also explored (Figs. 8D–F). At first glance, there are many more significant (p-value ≤0.05) relationships at the chamber level. For example, some level of significant positive covariance is found among all elements studied in the neogloboquadrinids (Figs. 8D, E). Of note, N. incompta exhibits moderate positive covariance between Mg/Ca, Na/Ca, Sr/Ca, and Ba/Ca with r² ranging from 0.50 and 0.66 for each of the combinations of TE comparisons (Fig. 8D). Moderate positive relationships also exist between Mn/Ca and Na/Ca (r² = 0.42), as well as Mn/Ca and Ba/Ca (r² = 0.43). Neogloboquadrina pachyderma also shows moderate positive correlations between Mg/Ca, Na/Ca, and Sr/Ca (0.44 ≤ r² ≤ 0.51; Fig. 8E). However, the strongest positive relationships in this taxon are found between Ba and Mn (r² = 0.73) and Mn and Zn (r² = 0.71), with Ba and Zn exhibiting a moderate positive relationship (r² = 0.47). Comparisons of chamber TEs for T. quinqueloba result in more significant covariations than found on the whole-test scale, the strongest being between Mg and Na (r² = 0.57) and Na and Ba (r² = 0.52; Fig. 8F).

Interesting differences in whole-test trace element geochemistry are identified among the three study species (Fig. 2; Table 1). For example, whole-test Mg/Ca, Na/Ca, and Sr/Ca systematically increase when comparing N. incompta to N. pachyderma, and values increase furthermore when comparing N. pachyderma to T. quinqueloba (Fig. 2). These interspecies differences are likely reflective of differences in living depth habitat in the study region where N. incompta prefers the deepest depth habitat of the three study taxa, and T. quinqueloba prefers the shallowest. Magnesium is known to be incorporated at greater concentrations into the foraminiferal crystal lattice when the test is grown under higher temperatures (Nürnberg et al., 1996; Lea et al., 1999; Anand et al., 2003; Russell et al., 2004; Martínez-Botí et al., 2011; Hönisch et al., 2013; Holland et al., 2020). The highest whole test Mg/Ca values are found in T. quinqueloba, suggesting that they live closest to the relatively warmer sea surface, and the lowest Mg/Ca values are found in N. incompta, suggesting that they prefer a deeper, cooler depth habitat (Fig. 2B). If we are to assume that the Mg/Ca composition of the three study taxa represents their preferred depth habitat, then this general pattern in relative depth habitat among the three study taxa are consistent with those observed in much of the Nordic Seas (Simstich et al., 2003) but differ from those characterized for the subtropical eastern North Atlantic (Rebotim et al., 2017).

Foraminiferal Na/Ca has been identified as a potential paleosalinity proxy, where test Na/Ca compositions positively covary with salinity (Wit et al., 2013; Mezger et al., 2016; Bertlich et al., 2018, 2021; Watkins et al., 2021) or as a proxy for the major ion chemistry of seawater on long timescales (Zhou et al., 2021; Gray et al., 2023). Among the data presented herein, there is a systematic increase in Na/Ca where not only the whole-test value but also the range of whole test Na/Ca values progressively increase from N. incompta to N. pachyderma to T. quinqueloba (Fig. 2C). This observation contradicts the current understanding of foraminiferal Na/Ca and salinity. In the Northern California Current region, salinity is found to generally increase with depth (Ortiz et al., 1996; Lane et al., 2023), suggesting that one should expect to find a decrease in Na/Ca values when comparing N. incompta to N. pachyderma and furthermore when comparing N. pachyderma to T. quinqueloba However, existing studies on the Na/Ca composition of several planktic foraminiferal taxa have found distinct differences in species-specific Na/Ca-salinity calibrations (e.g., see fig. 7 in Bertlich et al., 2021). For example, tropical species Trilobatus sacculifer incorporates Na at lower rates than polar taxon N. pachyderma (Allen et al., 2016; Bertlich et al., 2018, 2021; Mezger et al., 2018). Deviation from predicted interspecies differences in Na/Ca composition with inferred depth habitat is most likely due to interspecies differences in Na partitioning. Despite the three study species not following the predicted pattern of decreased Na/Ca content with shallower depth habitats, T. quinqueloba does exhibit the largest range of Na/Ca values, supporting the interpretation of a shallower depth habitat where salinity conditions can be highly variable. All three taxa show moderate positive relationships between Mg and Na in the whole-test and individual chamber scale, providing support for the interpretation of temperature and salinity changes with depth habitat being the dominant controls on the modulation of these elements, or, more generally, that Mg and Na incorporation mechanisms are potentially linked (Fig. 8).

Measurements of Sr/Ca show the same systematic increase in whole test values among the three taxa, in which the lowest values are found in N. incompta, and the highest values are found in T. quinqueloba (Fig. 2D). Distributions of whole-test values of Sr/Ca appear to have an abrupt upper limit and a skewed distribution toward lower Sr/Ca values. This distribution architecture is likely reflective of maximum Sr content in the uppermost surface ocean with greater variability of Sr/Ca values found below the surface maximum. Recent compilations of Sr/Ca ratios in seawater have found that Sr/Ca decreases with depth in the Pacific surface ocean (Lebrato et al., 2020), although at much higher values than found in foraminiferal calcite (Allen et al., 2016), supporting the linkage between depth habitat and foraminiferal Sr/Ca in the study taxa.

Barium and manganese incorporation may be reflective of habitat, more specifically potential growth in deeper, oxygen-poor waters or the influence of calcification in microhabitats during periods of higher productivity (Fritz‐Endres et al., 2022). Non-spinose taxa have been found living and calcifying in marine snow particles, a habitat favorable for the incorporation of Ba and Mn (Fehrenbacher et al., 2018; Greco et al., 2019; Fritz‐Endres et al., 2022; Richey et al., 2022). It is also noteworthy that upon picking and removing tests for this study, it was not uncommon to find neogloboquadrinid specimens embedded in marine snow. Barite forms within marine snow particles, suggesting that the Ba in solution within particulate organic matter is highly elevated compared to ambient seawater (Dymond & Collier, 1996; Griffith & Paytan, 2012; Martinez-Ruiz et al., 2019, 2020; Carter et al., 2020). During periods of high export productivity, more tests of non-spinose taxa can be found living in the abundant marine snow, leading to increased barium incorporation into the calcite crystal lattice (Fritz‐Endres et al., 2022). Marine snow also acts as an oxygen-depleted microhabitat (Alldredge & Cohen, 1987), where redox-sensitive elements such as Mn could be more readily incorporated into the foraminiferal crystal lattice. However, Mn/Ca composition of some species may also be controlled by upwelling intensity (Davis et al., 2020). The Northern California Current is known for its exceptional seasonal fluctuations in upwelling intensity, in which the upwelling of nutrient-rich deep waters often lead to increased productivity in the surface ocean. Periods of higher productivity associated with upwelling or other factors are also associated with expanded oxygen minimum zones, thus potentially creating oxygen-depleted habitats on both the micro-(i.e., marine snow) and macro-scales. While the exact control on Ba and Mn incorporation in these planktic foraminiferal taxa is not well-constrained, there appears to be a similar control on the whole-test Ba/Ca and Mn/Ca of the neogloboquadrinids that helps to explain the covariation of whole-test Ba/Ca and Mn/Ca (N. incompta r² = 0.52; N. pachyderma r² = 0.48; Figs. 8A, B). This positive relationship between Ba and Mn is either similar (i.e., N. incompta, r² = 0.43; Fig. 8D) or stronger (i.e., N. pachyderma, r² = 0.73; Fig. 8E) when examining TE covariance within individual chambers. Ratios of Ba/Ca and Mn/Ca are not correlated on either the whole test or individual-chamber scale for T. quinqueloba (Figs. 8C, F). Whole-test Mn/Ca ratios of T. quinqueloba are in a similar range to the neogloboquadrinids, but the Ba/Ca ratios are considerably lower (Fig. 2). However, the Ba/Ca ratios recorded by T. quinqueloba are elevated compared to most other spinose taxa (cf. Hönisch et al., 2011). As a spinose foraminifera, T. quinqueloba is unlikely to live in a marine snow microhabitat, however the elevated Ba/Ca and Mn/Ca ratios of this species may also be influenced by upwelling (Davis et al., 2020).

Zinc incorporation has been found to be potentially reflective of productivity, oxygen concentration, and/or carbonate ion concentration in benthic foraminifera (Marchitto et al., 2000; Bryan & Marchitto, 2010; van Dijk et al., 2017) but is relatively understudied in planktic foraminifera (Marr et al., 2013; Davis et al., 2023; Hupp & Fehrenbacher, 2023), and thus incorporation mechanisms are poorly understood. Interspecies comparison of whole-test Zn/Ca values in this study furthers this conundrum, as whole-test Zn/Ca does not appear to follow any interspecies patterns similar to those described above for the other trace elements. However, Neogloboquadrina pachyderma exhibits distinctly higher Zn/Ca values than the two other study taxa, and comparisons of average TEs in individual chambers yields a strong positive covariance between Mn/Ca and Zn/Ca incorporation (r² = 0.71; Fig. 8E). Zinc is also a redox-sensitive element that can become soluble under low oxygen conditions during microbial sulfate reduction and thus could also be reflective of calcification in a marine snow microhabitat (Davis et al., 2023). Interestingly, we do not find a similar strength of relationship in our N. incompta population (Fig. 8). It is possible that N. pachyderma has a greater tolerance for calcifying under low-oxygen conditions in which more time spent in such an environment could allow for growth under different stages of the redox ladder. Dissolved manganese can be produced in hypoxic environments under manganese reduction, whereas anoxic conditions are needed for sulfate reduction to occur in which dissolved zinc could become available for incorporation in the calcite crystal lattice.

Interspecies differences in trace element incorporation provide useful insight into the general ecology of the study taxa in this region. Although N. incompta and N. pachyderma are of the same genus, differences in preferred depth habitat and potentially differences in marine microhabitat growth regimes create distinct geochemical signatures among these two species. Turborotalita quinqueloba also has a clearly different geochemistry than the neogloboquadrinids. All three study taxa are asymbiotic; however, T. quinqueloba is a spinose species, whereas the neogloboquadrinids are non-spinose. Despite lacking algal photosymbionts, T. quinqueloba appears to have a shallower depth habitat (inferred from Mg/Ca ratios) compared to N. incompta or N. pachyderma (Simstich et al., 2003). Turborotalita quinqueloba also exhibits very low Ba/Ca whole test ratios (Fig. 2) and lacks significant relationships between Ba, Mn, and Zn (Fig. 8), suggesting that even though it shares the asymbiotic ecology of the neogloboquadrinids, its TE/Ca incorporation mechanisms may be more similar to symbiont-bearing spinose species.

Investigation of intratest variability can provide insight into ontogeny. For example, chamber-to-chamber variability may be reflective of changing depth habitat or oceanographic conditions during an individual’s life cycle. While some chamber-to-chamber TE comparisons show little to no difference between adjacent chambers, other chamber comparisons yield interesting results (Figs. 3, 4; Table 3). Here, chamber-to-chamber trends demonstrate that Mg content progressively decreases in sequentially younger chambers in all three study taxa (Figs. 3B, E, H). A similar chamber-to-chamber trend has been observed in N. dutertrei (Steinhardt et al., 2014; Fehrenbacher et al., 2018) but has not been previously recorded in other studies of N. incompta (Bolton, 2011; Davis et al., 2017) or N. pachyderma (Davis et al., 2017; Hupp & Fehrenbacher, 2023). Decreasing Mg could be reflective of an increasing depth of living habitat over the course of an individual’s life (e.g., Meilland et al., 2021).

An enrichment in Na in the F0 chamber of N. incompta is evident (Fig. 3C). Recent studies have shown that addition of a gametogenic crust at the end of a planktic foraminifers’ life can lead to an overall lower concentration in the whole-test trace element geochemistry (Eggins et al., 2003; Sadekov et al., 2005; Bolton & Marr, 2013; Davis et al., 2017; Hupp & Fehrenbacher, 2023). Examination of intrachamber TE composition supports this observation, where the inner lamellar calcite records higher TEs than the outer crust calcite (Fig. 7). The gametogenic crust tends to be relatively thin in the F0 chamber compared to other chambers in the outer whorl. Therefore, the Na geochemistry of the F0 chamber of crusted specimens can seem to be relatively enriched. Similar observations have been made for Na and Sr in populations of N. pachyderma collected from plankton tows in the Northern California Current (Hupp & Fehrenbacher, 2023).

Chamber-to-chamber trends in Sr/Ca vary markedly between species, with N. incompta exhibiting a decreasing trend with each additional chamber, N. pachyderma exhibiting an increasing trend, and T. quinqueloba showing no systematic trend (Fig. 3; Table 3), furthering the conundrum of which environmental parameter controls Sr incorporation (Lea et al., 1999; Martin et al., 1999; Stoll et al., 1999; Billups et al., 2004; Allen et al., 2016). Compilations of systematic changes in Sr among adjacent chambers have revealed a wide range of intratest patterns (see table 1 in Davis et al., 2020). Since the study of Davis et al. (2020), Hupp & Fehrenbacher (2023) found little change in chamber-to-chamber Sr/Ca variability in a population of crusted N. pachyderma, aside from an enrichment in Sr in the F0 chamber. Conversely, prior study of the non-spinose taxa Globorotaloides hexagonus, known to inhabit oxygen minimum zones, has shown a distinct increase in Sr incorporation with the addition of each new chamber (Davis et al., 2023). Ratios of Sr/Ca have been shown to increase with increased temperature, salinity, and carbonate ion concentration in spinose tropical to subtropical species (Allen et al., 2016). However, systematic chamber-to-chamber fluctuations in Sr/Ca could directly mimic changes in water Sr/Ca, reflective of dynamic changes in depth habitat through ontogeny.

Chamber Ba/Ca and Mn/Ca comparisons in N. pachyderma and T. quinqueloba show distinct decreasing trends in TE incorporation with ontogeny, whereas N. incompta exhibits relatively consistent Ba and Mn composition among chambers in the final whorl (Fig. 4). Differences in Ba and Mn content among chambers in N. pachyderma may reflect movement into and out of a marine snow calcifying environment during the ontogeny of an individual, in which chambers recording higher Ba/Ca and Mn/Ca values signal calcification in a marine snow microenvironment, where conditions are conducive for elevated TEs (Alldredge & Cohen, 1987; Fritz‐Endres et al., 2022; Richey et al., 2022); however, this explanation would not extend to explain similar trends found in T. quinqueloba. While we do not find any chamber-to-chamber gradient in the Ba/Ca and Mn/Ca composition of N. incompta, the larger range of Ba and Mn whole-test values suggests that they too prefer a marine snow microhabitat. However, the lack of chamber variability may imply that individuals either spend nearly all of their adult life calcifying in marine snow when it is present or all of their life outside of marine snow.

Interspecies differences in Zn/Ca chamber-to-chamber trends inspire more questions; Zn/Ca shows a clear increasing trend in progressively younger chambers of the neogloboquadrinids but exhibits a decreasing trend, aside from the F0 chamber, in T. quinqueloba. These chamber-to-chamber observations could reflect changes in the carbonate ion concentration of the calcifying microhabitat (Marchitto et al., 2000; van Dijk et al., 2017), with potential differences between spinose and non-spinose taxa. However, the mechanism for producing such differences in [CO₃²⁻] among spinose and non-spinose taxa is unclear, particularly because all taxa examined in this study are asymbiotic, and thus changing pH conditions in the local microenvironment due to the photosynthesis of algal symbionts is not an explanation. Ratios of Zn/Ca have been suggested to be sensitive to oxygen concentrations (Davis et al., 2023), whereby incorporation of redox-sensitive Zn would increase when calcifying under low-oxygen conditions. This relationship has been observed in Globorotaloides hexagonus, a species commonly found living in oxygen minimum zones (Davis et al., 2023). However, oxygen conditions are unlikely to change systematically throughout the ontogeny, and thus is unlikely to be the cause of chamber-to-chamber Zn variability in our study taxa. Zinc incorporation into planktic foraminiferal tests continues to be an enigma.

While the causes of ontogenetic trends in TE incorporation are difficult to constrain, it is interesting to consider the interspecies differences in such trends (Table 3). For example, one would expect to potentially find similarities in ontogenetic TE incorporation among taxa of the same genus, same spinose versus non-spinose affinity, or same symbiotic versus asymbiotic status. However, we find little consistencies or natural groupings in chamber-to-chamber TE incorporation among our study taxa aside from a consistent decrease in Mg incorporation through ontogeny in all study taxa and similarities in systematic Zn incorporation among the neogloboquadrinids.

The range of intratest chamber measurements (e.g., Fig. 5A) can be used to identify more sporadic changes in environmental conditions throughout an individual’s lifetime. For example, the intratest Ba/Ca range is highest in the neogloboquadrinids (Fig. 5D), where either the intensity of export production may change during an individual’s life or they may choose to only live a portion of their life calcifying in a marine snow microenvironment. Generally, our data show that greater intratest Mg/Ca and Na/Ca variability is likely reflective of existence in a shallower depth habitat where temperature and salinity feature stronger fluctuations during a foraminifers’ lifetime. This expected relationship tied to depth habitat is well-supported by the moderate positive correlations (0.47 ≤ r² ≤ 0.53) found between whole-test and individual chamber Mg/Ca and Na/Ca in our study taxa (Fig. 8). However, while T. quinqueloba records overall higher whole-test Sr/Ca values than the other two study taxa, we find that their intratest ranges are comparable to the neogloboquadrinids, likely due to 1) the environmental condition(s) that regulate Sr incorporation not changing much in their depth habitat over the short timescale of an individual foraminifer’s lifetime or 2) a general insensitivity to changing Sr incorporation. Greater intratest Ba/Ca and Mn/Ca in the neogloboquadrinids and, in the case of N. pachyderma, Zn/Ca variability is reflective of ontogeny lived in a deeper depth habitat, likely intersecting with the oxygen minimum zone, where fluctuations in export production and oxygen saturation modulate trace element incorporation into the foraminiferal crystal lattice over an individual’s lifetime. Further characterization of the preferred depth habitat of these taxa in the Northern California Current will strengthen our understanding of these geochemical-environmental relationships.

In addition to examining chamber-to-chamber variability to delineate ontogenetic controls on trace element incorporation, we can investigate how trace elements vary with depth through the chamber wall. Previous studies have investigated either a few individuals or a single population of tests to examine how test trace element geochemistry may change with the addition of gametogenic crust (e.g., table 1 in Hupp & Fehrenbacher, 2023). However, the dataset presented herein allows for investigation into the potential for consistent patterns in the trace element differences between the gametogenic crust and early ontogenetic calcite in individuals grown under a wide range of environmental conditions. Here we find that for all three study taxa and nearly all trace elements examined herein, the inner chamber wall (i.e., dominantly lamellar calcite) has consistently higher trace element compositions than the outer chamber wall, dominated by the gametogenic crust (Fig. 7). This observation indicates that the biomineralization pathway for precipitating the crust post- or syn-reproduction differs from the “normal” biomineralization mechanism used for growing a test throughout the vast majority of a foraminifer’s life. While the gametogenic crusting process has been found to occur quickly (on the scale of hours to days; Bé, 1980), it is possible that rapid calcification rates could be linked to lower TE incorporation. However, studies of symbiont-bearing planktic foraminifers have found that calcification rate does not typically exert a dominant control on trace element partitioning (Allen et al., 2016). Furthermore, if calcification rate were to exert a control on TE incorporation, one might expect to find higher TE compositions in calcite precipitated at higher rates, as commonly found with trace elements incorporated during inorganic calcite precipitation (e.g., Lorens, 1981; Mucci, 1988; Tesoriero & Pankow, 1996; Nehrke et al., 2007). Yet, the rapidly-precipitated gametogenic calcite exhibits consistently lower TE concentrations than the lamellar calcite among the study taxa, supporting the idea that calcification rate is unlikely to be the cause of such TE differences throughout the chamber wall. More research is needed to better understand differences between and the motivation for these two modes of biomineralization.

Constraining controls on interspecies, interspecimen, and intratest trace element variability is critical to refining our understanding and utilization of foraminifera as paleoproxy substrates. The data explored herein provide insight into several considerations for foraminiferal-based paleoreconstructions. Firstly, interspecies Ba/Ca differences and variability in the range of intratest Ba/Ca chamber values lend support for the recently proposed Ba/Ca-productivity proxy in non-spinose foraminifera (Fehrenbacher et al., 2018; Fritz‐Endres et al., 2022; Richey et al., 2022). In both neogloboquadrinid species, we find elevated whole-test Ba/Ca and a wide range of Ba/Ca intratest values, suggesting that some of our study specimens grew under higher productivity regimes. Thus far, positive relationships between Ba/Ca and export productivity have been documented in nonspinose taxa Neogloboquadrina dutertrei, Pulleniatina obliquiloculata, and Globorotalia truncatulinoides (Fehrenbacher et al., 2018; Fritz‐Endres et al., 2022; Richey et al., 2022). This study suggests that a similar relationship may exist in N. incompta and N. pachyderma. To identify periods of high productivity in the fossil record, Ba/Ca ratios should be measured in individual specimens from each time slice of interest to examine both the interspecimen and intraspecimen (e.g., in N. pachyderma) variability in a population. A similar approach has been taken with examining changes in populations of individual δ¹⁸O foraminiferal analyses to characterize changes in El Niño intensity over time (Thirumalai et al., 2013).

Correlations between Ba/Ca and Mn/Ca (Fig. 8) are supportive of not only Ba as a productivity proxy, but also Mn as a potential dissolved oxygen proxy. Although research has explored dissolved oxygen concentrations as the primary driver of Mn incorporation in benthic foraminifera, much less research has been done to identify such a relationship in planktic foraminifera (e.g., Marr et al., 2013; Davis et al., 2023). Here we show that Mn, and possibly Zn in the case of N. pachyderma, incorporation may be a direct result of redox conditions during calcification, and after further development could serve as critical proxies for reconstructing changes in oxygen minimum zone behavior and, when paired with Ba/Ca, particulate organic carbon flux to the seafloor.

The Na/Ca-paleosalinity proxy was recently proposed, and while field-based and culturing studies have produced promising results (Mezger et al., 2016; Bertlich et al., 2018; Dämmer et al., 2020; Bertlich et al., 2021), others are skeptical of the applicability of this ratio as a salinity proxy in deep time (Zhou et al., 2021; Gray et al., 2023). Although the data provided herein cannot settle this debate, these data contribute to fill the large data-gap in quantifying the Na/Ca composition of different planktic foraminiferal species collected from their natural living environment. Whole-test Na/Ca averages of our three study taxa ranged from approximately 10.7 to 41.7 mmol/mol (mean = 15.0 mmol/mol) for N. incompta, 11.3 to 25.1 mmol/mol (mean = 17.3 mmol/mol) for N. pachyderma, and 14.5 to 93.1 mmol/mol (mean = 24.3 mmol/mol) for T. quinqueloba (Table 1). These values are consistent with those that have been previously published on N. pachyderma from the Northern California Current region (Hupp & Fehrenbacher, 2023) but are higher than values reported from N. pachyderma collected from the Weddell Sea that were subsequently grown in culture (Bertlich et al., 2021). To our knowledge, these are the first Na/Ca ratios acquired from N. incompta and T. quinqueloba collected from plankton tows. In general, all published values of N. pachyderma are higher than those found in tropical spinose species (e.g., T. sacculifer, G. ruber; Mezger et al., 2016; Bertlich et al., 2018), suggesting that interspecies Na/Ca differences and Na incorporation mechanisms are far from being understood. It is possible that elevated Na/Ca is associated with storage in NaOH-buffered preservatives (i.e., formalin) or exposure to NaOH-buffered solutions during the pre-analysis chemical cleaning procedure. However, Na/Ca data shown herein is consistent with that previously published by Hupp & Fehrenbacher (2023), in which all specimens were collected from “fresh” tows and were not exposed to preservatives. A constrained cleaning study is needed to determine the potential impact of using NaOH-buffered cleaning solutions and preservatives on individual specimens analyzed by laser ablation or solution-based ICP-MS. Regardless of this potential source of bias, we are confident in the relative differences found between species, between specimens, and within individual tests.

Contrasts in interspecies TE composition described herein appear to reflect differences in preferred habitat and therefore variation in calcification conditions. However, it is worth noting that interspecies differences are also in part controlled by the relative number of specimens collected from variable environmental regimes. Furthermore, it is recognized that individual species of foraminifera often require species-specific TE-environmental calibrations when developing paleoproxies; this is because each taxon may have a different sensitivity to environmental conditions or variable approaches to calcification that produce taxon-specific partition coefficients. We can already identify some evidence of different sensitivities to trace element incorporation from the data presented herein. For example, T. quinqueloba does not exhibit covariance between many of the trace elements examined on the whole-test or individual chamber scale (Fig. 8C, F), suggesting that either 1) their trace element incorporation is less sensitive to calcifying under high productivity, low oxygen, or variable carbonate chemistry conditions and/or 2) their preferred habitat is relatively homogenous in these environmental conditions (aside from changes in temperature and salinity). We also find that N. pachyderma shows a distinct Zn enrichment and strong positive correlation between Mn and Zn that is not found in either of the other study taxa (Figs. 2, 8), showing further taxon-specific patterns in trace element incorporation.

Herein we characterize the amount of chamber-to-chamber and general intratest variability with the aim of identifying vital effects that could influence the appropriateness of these polar to subpolar taxa for paleoreconstruction purposes. Chamber-to-chamber measurements show that on average individual chamber Mg/Ca only varies from the test average by <0.5 mmol/mol across the three study taxa (Fig. 3) with intratest ranges being <1 mmol/mol for the neogloboquadrinids and <2 mmol/mol for T. quinqueloba (Fig. 5B). Published calibrations for N. pachyderma and N. incompta (Davis et al., 2017) suggest that intratest variability of 1 mmol/mol would equate to a few degrees C depending on the location along the logarithmic Mg/Ca-temperature calibration curve. A Mg/Ca-temperature relationship has not been determined for T. quinqueloba. Measurements of Na/Ca indicate chamber measurements vary by <1 mmol/mol across the three study taxa (Fig. 3) when compared to the test average, and intratest ranges are found to be <4 mmol/mol for the neogloboquadrinids and <10 mmol/mol for T. quinqueloba (Fig. 5C). Published Na/Ca-salinity calibrations are relatively limited, and thus a difference of 1 to 10 mmol/mol could equate to a change of 2 to >10 psu. Strong species-specific (e.g., Bertlich et al., 2021) and instrument-specific (e.g., Gray et al., 2023) differences in Na/Ca make it difficult to constrain the impact of intratest differences in Na/Ca on paleoreconstructions at this time; however, further work illuminating the caveats of the proposed Na/Ca-salinity proxy will help to place these intratest differences into context in regards to their potential effect on salinity reconstructions. In general, intratest variability in Mg/Ca and Na/Ca is found to be greatest in T. quinqueloba of our three study taxa. Minimal intratest variability is found in Sr/Ca (<0.05 mmol/mol comparing chambers to test average; ∼1 mmol/mol intratest range) among the three study taxa (Figs. 3, 5D), and with the dominant control on Sr/Ca remaining elusive, we do not pinpoint a need for concern of ontogenetic influences on paleoreconstruction at this time.

When comparing chamber values to the test average for Ba/Ca, we find variability of <2 μmol/mol (Fig. 4) and intratest ranges among all chambers in the outer whorl of <5 μmol/mol (Fig. 5E). While Ba/Ca-export productivity calibrations have not been published for the study taxa, published calibrations of Neogloboquadrina dutertrei and Pulleniatina obliquiloculata suggest that differences of 2 μmol/mol would result in a change of interpreted export productivity by ∼0.5 mol C/m²/yr, and a difference of 5 μmol/mol could further offset export productivity estimates by up to approximately 2 mol C/m²/yr (Fritz‐Endres et al., 2022). Individual chamber comparisons to test averages for Mn/Ca show offsets of <3 μmol/mol (Fig. 4) and intratest chamber ranges of <7 μmol/mol (Fig. 5F). Similar to Ba/Ca, calibrations between Mn/Ca and dissolved oxygen have not been published for these taxa. However, published calibrations between Mn/Ca and dissolved oxygen measured from the oxygen minimum zone-dwelling foraminifer Globorotaloides hexagonus indicate that intratest differences of this magnitude would likely have very little impact on dissolved oxygen reconstructions (Davis et al., 2023). Furthermore, studies linking Mn/Ca to upwelling variability found differences in Mn/Ca for specimens of Globigerina ruber (pink), Orbulina universa, and Globorotalia menardii grown during non-upwelling and upwelling times to be >10 μmol/mol on average (Davis et al., 2020). This further suggests that intratest Mn/Ca found in the three study taxa herein are unlikely to obscure primary environmental signals interpreted from whole-test values. As discussed, the primary control on Zn/Ca is ambiguous. Individual chamber values of Zn/Ca vary by <100 μmol/mol when compared to the test average (Fig. 4). However, the intratest range of Zn/Ca values can reach concentrations up to nearly 400 μmol/mol. As our understanding of the primary controls of Zn/Ca evolves, so will the implications of these intratest differences in the three study taxa.

The magnitude of intratest variability described above could have impacts on Mg/Ca-temperature reconstructions, particularly for T. quinqueloba. Vital effects in the form of intratest differences could also impact Na/Ca-salinity reconstructions and Ba/Ca-export productivity reconstructions, though much more work is needed to refine our understanding of these potential proxies, particularly in the three taxa examined in this study. While the degree of intratest variability may be alarming, it is important to recognize that the depth habitat of an individual is not stagnant through ontogeny. Observations of population dynamics as a means to investigate reproductive strategies of planktic foraminifera living in the Red Sea have shown that individuals often deviate from a consistent depth habitat and that the depth habitat of a foraminifer changes throughout its lifetime (Meilland et al., 2021). However, statistically robust investigations of test populations show that the population, and thus the chemistry of their tests, will largely represent a consistent change in depth habitat during ontogeny. Therefore, intratest variability, such as that characterized in this study, is likely largely reflective of changes in habitat conditions or depth during an individual’s lifetime.

The data provided herein also lead to suggestions on how to approach LA-ICP-MS trace element analyses for the study taxa. Here we find that intratest trace element variability increases in tests with higher whole-test trace element values (Table 2). This pattern is observed as a moderate to strong positive relationship in all three taxa for most elements, aside from Sr/Ca. Therefore, when analyzing tests via LA-ICP-MS, it is imperative to conduct individual spot analyses in all chambers in the final whorl to determine your test average signal and not restrict your analyses to the ultimate or penultimate chambers. Measuring all chambers in the final whorl will also allow for more direct comparison to solution-based data (Fehrenbacher et al., 2020).

Lastly, we have shown that for our three study taxa, there are predictable intrachamber differences where early ontogenetic calcite records higher trace element compositions than the gametogenic crust for nearly all trace elements (Fig. 7). Differences in trace element geochemistry between lamellar calcite and crust calcite are most apparent in the Mg/Ca and Na/Ca of all three taxa, as well as in the Ba/Ca composition of the neogloboquadrinids and the Zn/Ca composition of T. quinqueloba (Fig. 7). This observation has implications for how TE-environmental calibrations are constructed and how we analyze fossil specimens. For example, TE-environmental calibrations are often established from controlled culture experiments. However, tests grown in culture do not often develop a thick gametogenic crust (Davis et al., 2017; Fehrenbacher et al., 2017). We therefore echo the importance of following the recommendations and considerations brought forth by Davis et al. (2017) and Hupp & Fehrenbacher (2023), where either 1) calibrations linking TE composition to environmental variables must be established from only crusted specimens, to allow for a more direct comparison to the fossil record which is dominated by crusted specimens (Caron et al., 1990; Johnstone et al., 2010); or 2) test measurements for paleoreconstructions should use LA-ICP-MS to measure and isolate the trace element signature of the early ontogenetic calcite, which more accurately records signals from the foraminifers living environment.

Here we explore interspecies, chamber-to-chamber, and intrachamber trace element variability in three subpolar to polar planktic foraminiferal taxa: Neogloboquadrina incompta, N. pachyderma, and Turborotalita quinqueloba. By examining large populations of thickly-calcified tests (i.e., dead upon collection) collected under a wide range of environmental conditions throughout the dynamic Northern California Current region, we investigate environmental versus biological controls on trace element incorporation. The three study taxa show distinctly different whole-test trace element values that are largely reflective of preferred living habitat. The whole-test Mg/Ca and Sr/Ca ratios are highest in T. quinqueloba and lowest in N. incompta, representing the shallowest and deepest depth habitats among the study taxa, respectively (Fig. 2). Turborotalita quinqueloba also exhibits wide distributions of whole-test averages and intratest ranges for Mg/Ca and Na/Ca, reflective of a shallow living depth with higher temperature and salinity variability over an individual’s ontogeny (Figs. 2, 5). Whole-test and individual chamber Mg/Ca and Na/Ca compositions positively covary in all three study taxa, further supporting their incorporation mechanisms being tied to living depth habitat. The neogloboquadrinids exhibit elevated Ba/Ca and Mn/Ca and strong positive covariance between these TEs compared to T. quinqueloba, indicating that some individuals calcified in an oxygen-depleted marine snow microhabitat, likely during periods of elevated export production (Figs. 2, 5, 8). These observations provide support for using the Ba/Ca of these non-spinose foraminifers as a paleoproductivity proxy and potentially Mn as a paleoredox proxy. Furthermore, strong correlations are found between Mn/Ca and Zn/Ca in N. pachyderma, suggesting that this species may be able to calcify more readily under variable redox conditions than the other study taxa in which hypoxic versus anoxic conditions allow for greater incorporation of Mn and Zn, respectively (Fig. 8E). Tests with higher whole-test trace element values tend to have greater intratest trace element variability (Table 2) and the outer chamber walls composed of dominantly gametogenic crust, consistently record lower trace element compositions than the lamellar calcite of the inner chamber wall formed throughout most of a foraminifers’ ontogeny; this pattern was found for nearly all trace elements in all three study taxa (Fig. 7). Intratest chamber-to-chamber variability can be quite high (Figs. 3–5), and thus laser-ablation ICP-MS studies of planktic foraminifera should always measure spot analyses of every chamber in the outer whorl. Lastly, results provide insight into the intratest trace element variability in subpolar and polar planktic foraminifera, and contribute vast amounts of data that can be used to better understand controls on trace element incorporation in these understudied high-latitude taxa.

This work was supported by an NSF OCE grant 2222365 to B.H. and J.F. and NSF OCE 2049143 to J.F. This research was further supported by the NOAA Climate and Global Change Postdoctoral Fellowship Program to B.H., administered by UCAR's Cooperative Programs for the Advancement of Earth System Science (CPAESS) under award #NA21OAR4310383. Special thanks to J. Fisher for assistance in accessing archived tow samples and the staff at the Oregon State University Keck Collaboratory for Plasma Spectrometry. Figure S1, Table S1, and the Appendices can be found linked to the online version of this article.

Appendix 1. Specimen metadata including site collected and if the specimen was from an archived or freshly-collected (i.e., “recent”) tow. The latitude and longitude for each tow site also included.

Appendix 2. Chamber (i.e., F0, F1, F2, F3, F4) and whole-test average TE/Ca values for each individual study specimen. Mg/Ca, Na/Ca, Sr/Ca values are reported in mmol/mol, whereas Ba/Ca, Mn/Ca, and Zn/Ca values are reported in μmol/mol.

Table S1. Average reproducibility for each TE/Ca ratio as determined by the average difference between repeat measurements in the same chamber of each study specimen.

Figure S1. Example of the continuous laser ablation spectrum for Neogloboquadrina incompta specimen #52. A) Raw laser ablation spectrum output from LAtools (Branson et al., 2019). B) Calibrated laser ablation spectrum in which the background (i.e., non-ablation) portions of the spectrum have been removed and ratios have been calculated and calibrated to bracketing NIST standards (see Materials and Methods: LA-ICP-MS). Note that for this specimen, replicate spot analyses were conducted on the F0 and F1 chambers.