Global ocean redox changes before and during the Toarcian Oceanic Anoxic Event

The ca. 183 Ma Toarcian Oceanic Anoxic Event (T-OAE) was a major perturbation to the Earth system coincident with mass extinction and linked to emplacement of the Karoo-Ferrar Large Igneous Province. Environmental changes included climate and hydrological intensification, enhanced continental weathering, increased marine productivity, and expansion of seafloor anoxia and euxinia. Despite the OAE designation, the global extent of total anoxia ([O2]aq = 0) and the partitioning between non-euxinic anoxia (including ferruginous) and euxinia remain poorly constrained. Prior work using Mo isotopes in organic-rich mudrocks inferred expansion of euxinia (2–10% of seafloor) during the T-OAE, but potential basinal restriction in classic European sections challenges global interpretations, and Mo isotopes cannot robustly quantify non-euxinic anoxia. Thallium isotopes from western Canada indicated two expansions of low-oxygen seafloor (~500 kyr before and at the onset of the negative carbon isotope excursion, N-CIE) but cannot distinguish redox subtypes. This study addresses these gaps by using temporally calibrated rhenium (Re) and molybdenum (Mo) elemental mass balance models applied to organic-rich mudrocks from the Gordondale Member (Fernie Formation, British Columbia) to quantify global total anoxic and euxinic seafloor areas before and during the T-OAE, leveraging the shorter oceanic residence time of Re to detect higher-temporal-resolution changes including potential contraction before the end of the event.

• Mo isotope records from European T-OAE sections suggest expanded global euxinia and contracted oxic seafloor at the event’s onset, with estimated euxinic area of ~2–10% (much larger than modern ~0.1–0.2%). Concerns over basinal restriction in European sections complicate global extrapolation, and Mo isotopes do not constrain non-euxinic anoxia due to similar fractionations across suboxic to anoxic settings.
• Thallium isotopes from western Canada record two shifts to higher (less negative) δ205Tl prior to and at the onset of the N-CIE, interpreted as two expansions of global low-oxygen seafloor and at least a 50% decrease in Mn-oxide burial; however, Tl cannot resolve the proportions of dysoxic/suboxidizing versus anoxic redox states and can be sensitive to local restriction due to very short residence time.
• Redox-sensitive elemental mass balances (U, Cr, Re for total anoxia; Mo for euxinia) have been successfully applied to Proterozoic oceans but had not been calibrated and applied to shorter Phanerozoic events like the T-OAE, although drawdown of trace-metal reservoirs during the T-OAE has been inferred. Previous Mo and S isotope studies indicate contraction of euxinia after the T-OAE, consistent with slower response times of longer-residence-time elements.

Study site and stratigraphy: New elemental and δ13Corg data were obtained from drill core c-B6-A/94-B-8 (Gordondale Member, Fernie Formation, British Columbia, Canada), which exhibits a Toarcian N-CIE (−2.0 to −2.4‰ over 7.1 m). Correlations with nearby sections and isotopic/geochronologic constraints indicate the presence of T-OAE strata. Local redox classification: To ensure ORM trace-metal enrichments reflect global seawater reservoirs, local bottom-water redox was constrained using Al-normalized trace metal ratios and authigenic ratios: V/Al, U/Al, Mo/Al, Re/Al, Re/Moauth, and Mo/Uauth. Thresholds were derived from a compiled modern dataset spanning oxic margins, OMZs, and silled euxinic basins. Key thresholds include: oxic characterized by low Re/Al (<1.3 ppb/%) and Mo/Al (<0.4 ppm/%); dysoxic below P-OMZs with V/Al <23 ppm/% and U/Al 1–5 ppm/% and high Re/Moauth (>21 ppb/ppm) and low Mo/Uauth (<0.3 ppm/ppm); suboxidizing within P-OMZs (O2 <5 µM) indicated by Re/Moauth 2–10 ppb/ppm (conservative split at 10 ppb/ppm from oxic/dysoxic); fully anoxic settings (P-OMZ and silled euxinic basins) show Mo/Al >5 ppm/% and Re/Moauth <2 ppb/ppm, further differentiated by V/Al and U/Al and Mo/Uauth (P-OMZ <2.4 ppm/ppm < silled euxinic). Intervals with low Mo but elevated Re and U indicate non-euxinic anoxia/suboxidizing conditions; elevated enrichments of all four metals indicate euxinia. Hydrography check: Mo–U and Cd/Mo–Co*Mn cross-plots indicate vigorous exchange with the open ocean and nutrient upwelling, minimizing local restriction effects. Interval subdivision: Based on δ13Corg and redox proxies, the section was divided into Pre-N-CIE and N-CIE, each split into lower/upper portions by redox to define Model Intervals (MI): MI-1 (lower Pre-N-CIE, dominantly euxinic), MI-2 (upper Pre-N-CIE, suboxidizing/non-euxinic anoxic), MI-3 (lower N-CIE, suboxidizing/non-euxinic anoxic), MI-4 (upper N-CIE, euxinic). Post-N-CIE (Poker Chip Shale) is dominantly oxic/dysoxic and excluded from modeling. Mass balance modeling: Authigenic concentrations of Re (for total anoxic seafloor area, Aanoxic) and Mo (for euxinic seafloor area, Aeuxinic) were related to global sink areas using an iterative ocean mass balance: Xauth = (1/BMAR) * [(Mb ac / b ac) * (Fin / Ax)], where Xauth is the authigenic metal concentration in locally anoxic (Re) or euxinic (Mo) sediments, BMAR is local bulk mass accumulation rate, Fin is riverine input flux, Ax is the seafloor area of each sink, and b values are environment-specific burial rates (modern-calibrated and updated iteratively with assumed depth of anoxia and associated organic carbon burial). Re burial is most efficient in anoxic sediments and insensitive to H2S availability; Mo burial is strongly enhanced only under euxinic conditions. Hydrothermal inputs are minor and excluded after sensitivity tests. Parameters and inputs: BMAR was calculated using the median T-OAE duration (400 kyr). Modern riverine fluxes assumed for Pre-N-CIE are 4.29 × 10^5 mol Re yr−1 and 3.00 × 10^8 mol Mo yr−1; at the onset of the N-CIE, riverine fluxes were increased threefold to reflect enhanced continental weathering, maintained through MI-3 and MI-4. Dysoxic samples identified by Re/Moauth were excluded from Re modeling. Mo modeling was restricted to locally euxinic intervals (MI-1, MI-4); it was not applied to MI-2 and MI-3 (non-euxinic conditions). Thermal maturity, dataset filtering, and BMAR sensitivity were assessed; full derivations, constants, and code are provided in Supplementary Information. Sample selection and local conditions: Only suboxidizing or anoxic samples were used for Re modeling, and only fully anoxic with evidence for euxinia for Mo modeling, to ensure that local enrichments track changes in global dissolved metal reservoirs.

• The lower Pre-N-CIE interval (MI-1; locally euxinic) shows high authigenic Re and Mo (Reauth = 215 ± 88 ng g−1; Moauth = 168 ± 91 µg g−1), indicating small anoxic/euxinic global areas. Model outputs: total anoxic seafloor 0.14–2.5%; euxinic 0.18–1.9% of total seafloor.
• Upper Pre-N-CIE (MI-2; non-euxinic anoxic/suboxidizing): Reauth = 125 ± 68 ng g−1 yields total anoxic seafloor 1.1–8.1% (Mo model not applicable).
• Lower N-CIE (MI-3; non-euxinic anoxic/suboxidizing): Reauth = 192 ± 95 ng g−1 yields total anoxic seafloor 3.9–100%, with large uncertainty due to model asymptote beyond ~10% anoxic area (Mo model not applicable).
• Upper N-CIE (MI-4; euxinic): Reauth = 277 ± 74 ng g−1 and Moauth = 108 ± 22 µg g−1 correspond to total anoxic seafloor 2.9–6.4% and euxinic seafloor 3.9–6.2%.
• Mean global area estimates based on interval means: total anoxic = 0.84% (MI-1), 2.6% (MI-2), 6.9% (MI-3), 4.1% (MI-4); euxinic = 0.47% (MI-1) and 4.8% (MI-4).
• Overall, models indicate an expansion to as much as ~7% total anoxic seafloor at the onset of the N-CIE, dominated by euxinia, followed by contraction before the end of the event.
• Local δ13Corg in the Gordondale Member records a −2.0 to −2.4‰ N-CIE across 7.1 m; the Re-based redox maximum aligns with the onset of this excursion. Re data suggest an earlier expansion of anoxia before the N-CIE, consistent with Tl isotope records, and a contraction during the N-CIE’s later phase.
• Comparison with biotic records shows the reconstructed redox evolution mirrors global ammonite and foraminiferal diversity collapse at or near T-OAE onset and recovery before the event’s end.
• Spatial implications: The magnitude of anoxic expansion likely remained focused on continental margins and epicontinental seas, with the open deep ocean largely oxygenated (though potentially with lower O2 than today).

By jointly applying Re and Mo mass balances to an unrestricted, marginally connected setting, the study quantitatively constrains both total anoxic and strictly euxinic seafloor areas through the T-OAE. Findings address the long-standing uncertainty over the global extent and character of deoxygenation: at peak, total anoxia covered up to ~7% of seafloor and was largely euxinic, but this expansion was transient and contracted prior to the end of the N-CIE. The results reconcile proxy discrepancies: shorter-residence-time proxies (Re, Tl) capture rapid contractions within the N-CIE, whereas longer-residence-time proxies (Mo, S) register post-event trends. The inferred margin-focused anoxia helps explain extensive organic-rich mudrock deposition and aligns with ecological selectivity during the mass extinction, while suggesting deep-ocean waters remained mostly oxygenated. The temporal redox pattern mirrors marine biodiversity collapse and recovery, supporting a causal link between transient, widespread low-oxygen conditions (especially euxinia) and biotic turnover. The approach demonstrates that integrating redox-sensitive elemental mass balances with rigorous local redox screening can separate global signals from local depositional noise and provide first-order, quantitative bounds on seafloor redox during Phanerozoic OAEs.

This study calibrates and applies Re- and Mo-based oceanic mass balance models to a Phanerozoic anoxic event to quantify, for the first time, both total anoxic and euxinic seafloor areas during the T-OAE. Results indicate a rapid expansion to as much as ~7% anoxic seafloor dominated by euxinia at the onset of the N-CIE, followed by contraction before the event’s end, with anoxia concentrated on continental margins and epicontinental seas and the deep ocean remaining largely oxygenated. The Re-derived redox trajectory aligns with documented patterns of marine biodiversity collapse and recovery, highlighting the ecological significance of transient but widespread euxinia. Future work should expand this framework to additional well-connected sections globally, refine constraints on riverine flux changes and BMAR, integrate complementary proxies (U, Cr, Tl, Fe speciation), and obtain tighter age control to resolve intra-event redox dynamics and the relative timing of euxinia versus general anoxia within the N-CIE.

• Proxy and modeling sensitivity: The Re model becomes asymptotic when anoxia exceeds ~10% of seafloor, inflating uncertainty (notably MI-3). Variance in authigenic enrichments within intervals likely reflects local depositional fluctuations (redox variability, sedimentation rates, host phase changes) rather than global redox, so interval means are more robust than individual samples.
• Local redox classification: Inexact agreement among redox proxies necessitated conservative interval delineation; some subintervals likely experienced transient re-oxygenation not fully captured.
• Euxinic constraints: Mo modeling is only applicable under locally euxinic conditions; hence Aeuxinic could not be estimated for MI-2 and MI-3, limiting resolution of euxinia within the N-CIE.
• Flux assumptions: Riverine flux during the N-CIE is scaled to 3× modern based on weathering proxies; Os-based weathering estimates from a nearby, mildly restricted site may overestimate global weathering, introducing uncertainty in flux scaling.
• Age control and stratigraphic completeness: Direct age constraints are lacking in the study core; correlations rely on δ13Corg and gamma-ray ties to dated sections. Differences in δ13Corg morphology across sections may reflect local sedimentation, sea-level change, or minor hiatuses.
• Spatial representativeness: Results derive from a single, open-margin site; while hydrographic analyses suggest good oceanic exchange, broader spatial coverage would better constrain global heterogeneity.
• Residence times: Longer-residence-time proxies (e.g., Mo isotopes) respond more slowly to redox changes, complicating direct temporal alignment across proxies; conversely, very short residence time proxies (Tl) can be sensitive to local restriction.