Calcium isotope ratios of malformed foraminifera reveal biocalcification stress preceded Oceanic Anoxic Event 2

Pelagic calcifiers (foraminifera, pteropods, coccolithophores) account for up to 70% of modern marine carbonate export, and their calcification rates are tightly linked to seawater carbonate chemistry. Rapid increases in atmospheric CO2 drive ocean acidification (OA), lowering pH, [CO3 2−], and saturation states, and increasing DIC. Modern planktic foraminifera subjected to OA show reduced calcification and stress indicators (e.g., shell thinning). Reduced biocalcification can buffer OA by increasing surface ocean alkalinity, enhancing atmospheric CO2 drawdown, but potentially compounding biocalcification stress. Understanding the coupling between pelagic calcification and atmospheric CO2 is therefore critical for quantifying the marine carbon pump, carbon cycle feedbacks, and for assessing ocean alkalinity enhancement as a climate intervention. Geologic OA events offer insight into calcification responses under high-CO2 background climates. The Cretaceous greenhouse experienced episodic perturbations to the carbon cycle linked to large igneous province (LIP) volcanism and oceanic anoxic events (OAEs). OAE2 (~94 Ma) is marked by organic-rich deposits, a +3–4‰ δ13Corg excursion, and evidence for CO2 increases associated with Caribbean and/or High Arctic LIP activity. Biotic changes (nannofossil extinctions, size reductions, and foraminiferal morphometric stress) often precede OAEs, including OAE2 (the lower critical interval, LCI). Calcium isotope fractionation in carbonates is sensitive to precipitation rate; slower rates yield higher δ44/40Ca (smaller ΔCaCO3–H2O). Prior lab, modeling, and geologic records indicate δ44/40Ca can track OA-related biocalcification stress. In this study, we test for biocalcification stress preceding OAE2 using high-resolution, high-precision δ44/40Ca measurements paired with δ13C and δ18O in bulk carbonate and planktic foraminifera (Rotalipora cushmani) from the Bottaccione section (Gubbio, Italy) and bulk records from the Aristocrat Angus core (Denver Basin, USA). We examine their relation to morphometric indicators and osmium isotope (Osi) records of volcanism to evaluate timing and mechanisms of OA and feedbacks.

• OA impacts on modern pelagic calcifiers include decreased calcification rates and shell thinning, with reduced biocalcification acting as a rapid negative feedback that increases surface ocean alkalinity and CO2 uptake.
• The Cretaceous greenhouse climate saw multiple OAEs associated with LIP volcanism, carbon isotope excursions, and widespread anoxia. OAE2 exhibits a pronounced positive δ13Corg excursion and is linked to Caribbean/High Arctic LIP activity.
• Biotic turnover and morphometric stress (e.g., nannofossil size reductions, foraminiferal abnormalities and coiling changes) often precede OAEs; for OAE2 these occur within the LCI.
• Calcium isotopes: Inorganic precipitation experiments show δ44/40Ca increases with decreasing precipitation rate; models tie ΔCaCO3–H2O to carbonate system parameters. Culture and fossil studies indicate some foraminifera fractionate Ca isotopes similarly to inorganic calcite, and several geologic δ44/40Ca records show positive excursions during candidate OA events (including foraminifera, mollusks, bulk sediments). Multi-proxy Ca–Sr isotope work across OAE1a supports precipitation-rate control on δ44/40Ca.
• Osmium isotope stratigraphy captures volcanic inputs and has been used to time LIP activity relative to OAE2 onset. These studies motivate using δ44/40Ca as a proxy for biocalcification stress and OA timing relative to OAE2.

Study sites and materials:

• Bottaccione section, Gubbio, Italy (Scaglia Bianca Formation; 43°21′56.04″N, 12°34′57.56″E; 597 m a.s.l.): Outcrop sampling of bulk carbonate and planktic foraminifera (Rotalipora cushmani; when possible, Whiteinella archaeocretacea) across the lower critical interval (LCI) and into OAE2.
• Aristocrat Angus (AA) core, Denver Basin, Colorado, USA (40°14′32.97″N, 104°41′42.98″W; 1506 m a.s.l.): Bulk carbonate sampling across the interval spanning the onset of OAE2.

Bulk carbonate processing:

• Approximately 100 g of rock per interval collected. Samples crushed (shatter box or mortar/pestle with silica sand rinse), 27–100 mg powder reacted with 10 mL of 5% HNO3 on rocker (~12 h) to ensure complete dissolution. Solutions filtered (0.45 µm) and stored in acid-washed LDPE. Ca concentrations measured by Thermo Scientific iCAP 6500 ICP-OES at Northwestern University (NU).

Foraminiferal processing:

• ~100 g rock treated with concentrated acetic acid (Lirer, 2000), sieved at 63 µm, dried at 50 °C, then washed through 212 µm mesh. 15–25 R. cushmani (and available W. archaeocretacea) per interval picked from nannofossil-rich chalk; Feret diameters measured. In NU lab, 7 specimens subsampled, lightly crushed, homogenized, cleaned with MilliQ (per Barker et al.), dissolved in 5% HNO3. Ca concentrations measured by Thermo iCapQ ICP-MS. SEM (Hitachi S3400N-II) imaging performed on selected samples to assess taxonomy, texture, and potential authigenic features; calcite infilling observed across intervals.

Calcium isotope analysis (δ44/40Ca):

• High-precision 43Ca–42Ca double-spike technique on Thermo-Fisher Triton MC-TIMS (10^11 Ω amplifiers). Clean lab protocols; blanks negligible (~100× lower than sample Ca). Standards: OSIL Atlantic Seawater (ASW) and NIST SRM-915b analyzed every ≤30 samples. Bulk: 50 µg Ca; foraminifera: 25 µg Ca spiked, equilibrated (60 °C, 12 h), dried (90 °C). ASW and bulk passed through AG MP-50 cation exchange columns to isolate Ca; OM oxidized (35% H2O2), converted to nitrate (16 N HNO3), final load in 3 N HNO3; 10–16 µg Ca loaded on Ta filaments. Foraminifera not column-processed due to low [Ca]. Double-spike recalibrated using ASW each session. Results reported relative to ASW in δ-notation. Long-term external reproducibility ±0.05‰; ASW δ44/40Ca ≈ 0.000 ± 0.044‰ (2σ, n=706); 915b ≈ −1.136 ± 0.049‰ (2σ, n=276). Replicates/duplicates within ±0.05‰.

Carbon and oxygen isotopes (δ13C, δ18O):

• Thermo Delta V-Plus IRMS with Gas Bench II; bulk: ~200 µg in 12 mL vials, He purge 10 min, 200 µL 103% H3PO4 (ρ=1.92 g/cm3), dissolve at 30 °C for 12 h. Foraminifera optimized for ≤40 µg masses: weighed into Sn capsules, He purge 10 min, heated 90 °C ≥30 min to dissolve Sn; standards size series for linearity. Standards: CLMS (δ13C +2.31 ±0.05‰; δ18O −3.70 ±0.10‰), NBS18 (δ13C −5.01 ±0.05‰; δ18O −23.01 ±0.05‰), bracketing every 10 samples; values referenced to VPDB via NBS19. Precision: ±0.10‰ (2σ) for both δ13C and δ18O.

Statistics:

• Non-serial data (Bottaccione δ44/40Ca, δ18O, morphometrics): linear regression; Cook’s distance evaluated; alternate models removing influential points confirmed robustness. Assumptions for linear models met. Morphometrics from Coccioni et al. (n=4140) used to represent R. cushmani population. Time series with varied sampling densities analyzed using Astrochron surrogateCor with linear interpolation and Spearman’s rank correlation; tests repeated per δ44/40Ca dataset to ensure no single record drives correlations.

Data integration:

• Paired bulk and foraminiferal δ44/40Ca, δ13C, δ18O from Bottaccione; bulk δ44/40Ca and δ13C from AA; comparisons to previously published Portland core δ44/40Ca and Os isotope records. Stratigraphic alignment via integrated timescales as in supplementary materials.

• δ44/40Ca increased prior to OAE2 onset: • Bottaccione bulk carbonate (n=31): −1.56‰ to −1.45‰, increasing through the LCI to the last occurrence (LO) of R. cushmani. • Bottaccione R. cushmani (n=18): from ~−1.60‰ rising to a maximum −1.41‰ at 110.72 msl, then decreasing to −1.48‰ just below LO at 110.85 msl. • AA core bulk: −1.58‰ to −1.40‰, with maximum values well before bentonite A (≈ LO of R. cushmani). Positive excursions span multiple lithologies (calcareous mudstone and limestone). • Pre-CIE positive δ44/40Ca shifts begin ~60 kyr before the positive δ13C excursion defining OAE2.
• Correlation with morphometric stress indicators in foraminifera (Bottaccione): • δ44/40Ca vs Foraminiferal Abnormality Index (FAI): p=0.004, R²=0.50 (positive relationship). • δ44/40Ca vs Feret diameter: p=0.006, R²=0.45 (higher δ44/40Ca associated with smaller test size). • δ44/40Ca vs coiling direction: not significant (p=0.14, R²≈0.11).
• Multi-archive agreement and species differences: • At one horizon (110.43 msl), R. cushmani and W. archaeocretacea δ44/40Ca values are identical. • Bulk carbonate δ44/40Ca, δ13C, and δ18O values are on average higher than foraminiferal values; within lower LCI, bulk δ44/40Ca is ~0.18‰ heavier than R. cushmani, δ13C up to 0.64‰ higher, and δ18O ~1‰ higher, suggesting surface-ocean signal dominance (coccolithophores) relative to deeper-dwelling R. cushmani.
• Lack of temperature control on δ44/40Ca during LCI: no significant δ18O–δ44/40Ca correlation (p=0.16, R²=0.09) and δ18O varies by only ~0.17‰, supporting kinetic/precipitation-rate control.
• Diagenesis assessment: multiple lines indicate minimal fluid-buffered diagenesis; foraminifera show recrystallization/infilling but δ44/40Ca values remain low and distinct from bulk, implying preserved primary signals.
• Link to volcanism and OA timing: • δ44/40Ca positive excursions coincide with declines in seawater Os isotope values (Osi), a volcanism proxy; overall δ44/40Ca–Osi correlation significant (p<0.01; R²=0.38) with δ44/40Ca increases leading the δ13C excursion by ~60 kyr. • Suggests rapid CO2 input from LIP volcanism induced OA and biocalcification stress before the main CIE and anoxia.
• Evolution during OAE2 and feedbacks: • Decreasing δ44/40Ca during the main phase of OAE2 indicates increased precipitation rates consistent with ocean alkalinization following suppressed biocalcification. • Persistent OA (elevated δ44/40Ca) for ~150 kyr after OAE2 start (Portland and AA records) suggests episodic/continued CO2 input with biological compensation feedbacks. • The shift to lower δ44/40Ca coincides with the Plenus Cold Event, consistent with transient atmospheric CO2 drawdown via biologically mediated alkalinization.

The study demonstrates that δ44/40Ca increases began ~60 kyr before the OAE2 δ13C excursion and coincide with foraminiferal morphometric stress, indicating OA-induced reductions in carbonate precipitation rates (smaller ΔCaCO3–H2O) ahead of widespread anoxia. The timing and correlation with declining Osi implicate rapid CO2 release from LIP volcanism as the trigger. Absence of a temperature signal in δ18O and diagenetic overprint considerations support a kinetic/precipitation-rate control on δ44/40Ca. Bulk versus foraminiferal offsets suggest surface-ocean calcifiers (e.g., coccolithophores) experienced early biocalcification stress, with potential downward propagation of acidification. These results motivate redefining the onset of OAE2 to include the LCI, capturing the initial OA perturbation preceding the positive δ13C excursion. After OA onset, suppressed biocalcification increased surface ocean alkalinity (biological compensation), enhancing CO2 invasion and potentially sustaining OA under continued volcanic CO2 input. As CO2 inputs waned, recovery in δ44/40Ca and Osi indicates rapid biocalcification rebound and possible alkalinity overshoot, providing a mechanism for transient CO2 drawdown and the Plenus Cold Event during a greenhouse climate. The findings underscore dynamic feedbacks among CO2 forcing, biocalcification, and air–sea carbon partitioning, and they inform strategies for ocean alkalinity enhancement by illustrating the magnitude and persistence required to achieve CO2 removal while mitigating OA.

Foraminiferal and bulk δ44/40Ca records from the Bottaccione section (Gubbio, Italy) and bulk δ44/40Ca from the Aristocrat Angus core (Western Interior Seaway) show elevated values beginning ~60 kyr prior to OAE2. Significant correlations between δ44/40Ca and foraminiferal morphometric stress (test dwarfing, malformation) indicate reduced carbonate precipitation rates driven by OA. The δ44/40Ca records correlate with decreases in seawater Os isotopes, linking OA to CO2 inputs from emplacement of the Caribbean and/or High Arctic LIP. Elevated δ44/40Ca prior to the δ13C excursion shows that OA and biocalcification stress preceded anoxia, suggesting a need to redefine OAE2 onset to include the LCI. During OAE2, lower δ44/40Ca values point to surface ocean alkalinization following reduced carbonate production, facilitating additional atmospheric CO2 drawdown and potentially explaining the Plenus Cold Event. Overall, OAE2 provides a geologic analog for ocean alkalinity enhancement, offering insights into feedbacks relevant to anthropogenic CO2 removal and OA mitigation.

• Diagenesis: Foraminiferal specimens exhibit recrystallization and calcite infilling; while multiple lines of evidence argue against fluid-buffered alteration and suggest preserved primary δ44/40Ca signals, some degree of diagenetic overprint cannot be entirely excluded.
• Temperature constraints: High-resolution, independent temperature proxy records are lacking for the pre-OAE2 interval at both sites; although δ18O shows minimal variation and no correlation with δ44/40Ca, unrecognized temperature variability could influence fractionation.
• Sample representation: Bulk carbonate integrates multiple species and habitats, potentially introducing variability due to changing assemblages and making species-specific responses less resolvable.
• Spatial coverage and chronology: Records are from two regions (Tethys and Western Interior Seaway) with stratigraphic alignment relying on integrated timescales; global synchrony and exact temporal leads/lags carry uncertainties.
• Statistical power: Some regressions (e.g., δ44/40Ca–δ18O) may be limited by sample size and influence of individual points, though robustness tests were performed.