Introduction

Dramatic middle and late Mesozoic climate events are often associated with significant spread of reducing conditions and linked to increases in extinction rates1,2. These discrete episodes are characterised by intense perturbations in the global carbon cycle that typically record low O2 conditions and widespread changes in ocean chemistry, called Oceanic Anoxic Events (OAEs)2,3,4,5. The OAEs are conventionally described as being characterized by black and dark shales with fine laminations and absence of bioturbation, increased sedimentary organic matter and presence of pyrite crystals that attest to low-oxygen conditions2. However, the black shale record is sometimes limited in some OAEs or even absent, as occurred for the Paleocene-Eocene Thermal Maximum (PETM)5,6. In addition to these sedimentological markers, concentrations and isotopes of several chemical species (e.g., Mo, V, U) are largely applied as redox indicators to identify the OAEs1,5. Most OAEs also are characterized by an initial negative δ13C excursion (CIE) due to input of light carbon (12C) into the ocean–atmosphere system, followed by a positive CIE due to burial of excess 12C in organic-rich sediments1,5,7. Notably, the position of the CIEs related to OAEs can vary locally when compared to the deposition of the dark to black shales, indicating that the oceanic oxygen depletion events are diachronous across sedimentary basins 1. Furthermore, OAEs are generally related to sea water warming indicated by low δ18O values, which is however commonly interrupted by cooling phases likely caused by the increase of organic matter burial and pCO2 decrease1,5.

Phanerozoic OAEs are defined by intervals of widespread oxygen depletion in the ocean, recognized by sedimentological and/or geochemical records, yet they are significantly dissimilar from each other5. Jurassic and Cretaceous OAEs are particularly well-studied, such as the early Toarcian (Posidonienschiefer event, T‐OAE, ~ 183 Ma), early Aptian (Selli event, OAE 1a, ~ 120 Ma), early Albian (Paquier event, OAE 1b, ~ 111 Ma) and Cenomanian–Turonian (Bonarelli event, OAE 2, ~ 94 Ma) events1,8. Potential forcing mechanisms for OAEs include massive releases of methane (CH4) or volcanogenic carbon dioxide (CO2) outgassing into the atmosphere during the emplacement of Large Igneous Provinces (LIPs) as continental flood basalts or oceanic plateaus1,9. For example, OAE 1a correlates to the Ontong Java Plateau, OAE 2 is associated with the Caribbean-Colombian Plateau1 and the early Toarcian OAE is linked to the Karoo–Ferrar subaerial flood basalt extrusions10. Similar scenarios occurred during the late Sevatian (late Norian, Late Triassic) and across the Norian-Rhaetian boundary (NRB), during which significant biotic extinctions of important terrestrial and marine taxa occurred11. Notably, cosmopolitan bivalves (e.g., monotids, halobiids) are one of the major fossil groups that suffered a global and severe extinction, starting from the Sevatian to the lowermost Rhaetian11,12,13, while ammonoids suffered the largest decline among invertebrates, with a reduction in genera by more than an order of magnitude14,15. Other faunal extinctions and diversity loss also are documented for radiolarians and conodonts16,17,18, marine vertebrates (e.g., actinopterygian fishes) and reptiles (particularly among ichthyosaurs)19. In addition, reef communities declined significantly during the Late Triassic20. This interval of strong faunal extinctions at the NRB is associated with large carbon isotope excursions recorded on land21,22 and in marine environments11, which are linked to oscillations of biological pump efficiency and/or ocean stratification11.

We investigated and compared redox conditions recorded in several NRB sections deposited in different sedimentary environments (from shallow to deep water) on opposite sides of Pangea, spanning both hemispheres and across latitudes (Fig. 1). The Pignola-Abriola and Sasso di Castalda sections were deposited in the deep pelagic Lagonegro Basin (Italy), which was part of the western Tethys23,24. The Lagonegro Basin records the change from carbonate to biosiliceous sedimentation below the carbonate compensation depth (CCD) across the NRB25,26,27. Both these sections were previously investigated for bio- and chemostratigraphy11,23,24,25,27,28,29,30,31,32,33,34, while magnetostratigraphy was performed only in Pignola-Abriola section35. The southern Tethys Wombat Basin (NW Australia) was a neritic continental shelf environment located in the southern Tethys on the margin of Pangea, and was recently studied for biostratigraphy36 and chemostratigraphy11. The Kiritehere section (New Zealand) was deposited from mid-shelf to deeper water basin on the southernmost margin of Gondwana, close to the South Pole, and it has been recently investigated for organic carbon37. The Holberg section (British Columbia, Canada) is illustrated here for the first time and it represents a shallow to moderate depth depositional environment off an oceanic plateau in eastern Panthalassa, and it has been studied for organic carbon stable isotopes (δ13Corg) and chemostratigraphy (this study; Supplementary material). The Katsuyama section (Japan) was deposited below the CCD in the north Equatorial belt of the Panthalassa Ocean has been studied for bio- and chemostratigraphy and δ13Corg (this study; Supplementary material). All these sections spanned the uppermost Sevatian to lower Rhaetian, which is usually characterized by laminated layers and dark to black shales, except for the Katsuyama section that consists of red and dark red shales and cherts(11 and Supplementary material). Notably, the Kiritehere section mostly consists of dark volcanoclastic rocks without any particularly darker portions of the succession. The NRB is identified with the first occurrence of the conodont Misikella posthernsteini or (when necessary) is approximated by the negative δ13Corg peak just below the base of the Rhaetian11,24. The Wombat and Holberg sections and the Calcari con Selce Formation in the Pignola-Abriola and Sasso di Castalda sections are cherty limestones.

Figure 1
figure 1

δ13Corg, Ba, and redox-sensitive elements (Mn, Cr, V, U, and Mo) show enrichments in the study sections. The stratigraphic position of the NRB is discussed in Supplementary Information. Blue dots represent raw data, and red lines show moving average on three samples.

Full size image

Geochemical evidence of redox conditions at the NRB

Geochemical proxies are used to reconstruct oceanic redox conditions (e.g., V, U, and Mo enrichment)8,38,39,40,41,42,43. Enrichment factors (expressed as XEF, see Methods in Supplementary material) for Cr, V, U, and Mo have been widely used to characterize marine redox conditions [e.g38,44,45,46], which generally can be classified as oxic, suboxic, anoxic, or euxinic (presence of free H2S)47. In this study, these redox-sensitive elements were used to investigate redox conditions across different regions and depths.

Higher concentrations of elements that indicate suboxic to anoxic environments (Cr, V, U, and Mo) were observed in the NRB of the Tethys and eastern Panthalassa oceans (Fig. 1). A slight enrichment of Cr in the Wombat section suggests that at least suboxic conditions existed in the southeastern Tethys Ocean during the NRB interval. Across the NRB, a significant increase in VEF is recorded in sediments deposited in intermediate to deep water environments, represented by the Holberg, Pignola-Abriola and Sasso di Castalda sections (Fig. 1). Vanadium reduction occurs under lower oxygen concentrations45. In oxic seawater, V is present as soluble V(V) in the quasi-conservative form of vanadate oxyanions (HVO42– and H2VO42–). When conditions change from suboxic to weakly anoxic, V(V) converts to V(IV) and forms the vanadyl ion (VO2+), related hydroxyl species (VO(OH)3–), and insoluble hydroxides (VO(OH)2)48,49. The highest concentration of V is recorded in the NRB of the Pignola-Abriola section, where VEF values are relatively constant and close to the value of 1 in the Sevatian and abruptly increase across the NRB (Fig. 1). Similar V enrichments are found in the NRB of the Holberg and Sasso di Castalda sections. In the Wombat section, V increases across the NRB, indicating that the shallow marine environment in the southeastern Tethys has changed to a suboxic–anoxic environment (Fig. 1).

On the other hand, the lack of V and U enrichments in the shallow marine sections of Kiritehere apparently does not indicate that these shallow depositional environments might have been oxygen-poor too. However, the Kiritehere section recorded the global extinction of pelagic and benthonic bivalves that belonged to the monitid group, so other mechanisms related to the spread of oxygen-deficient waters might be thus indicated (e.g., ocean acidification, competition, food shortage, eutrophication). In addition, no stratigraphic changes in concentrations of redox-sensitive elements were observed in the deep-sea pelagic deposits of Panthalassa, represented by the Katsuyama section. This suggests that the development of oxygen-poor conditions may not have extended to the deep seafloor of the mid-Panthalassic Ocean (Fig. 1).

Another common redox proxy for reducing conditions is the MoEF/UEF ratio, which shows enrichment trends similar to those observed for VEF in both the Pignola-Abriola and Holberg sections (Fig. 1). Covariations of Mo and U enrichment patterns (Fig. 2) and increased MoEF/UEF ratios suggest the development of anoxic bottom water and scavenging of Mo by a Fe–Mn (oxyhydr)oxide particulate shuttle during the upper Sevatian to Rhaetian38,49. The occurrence of the particulate shuttle requires an oxic–anoxic redox boundary in the water column and rapid water replacement to maintain the predicted Fe–Mn redox behavior 38,50. Thus, the MoEF/UEF increase in the particulate shuttle field in the Holberg and Pignola-Abriola sections may indicate the development of intermediate-depth O2 minima with redox boundaries on both sides of Pangea. However, Mo enrichment due to Mn- and Fe- particulate shuttles was less significant compared to modern sites with anoxic oxygen minimum zones (OMZs)50. Consequently, V enrichment and UEF and MoEF trends suggest that anoxic conditions occurred in mid- to deep-water environments over a large area around Pangaea (Fig. 1).

Figure 2
figure 2

UEF vs. MoEF diagram for samples in the study sections. Base figure modified from Algeo and Tribovillard38.

Full size image

Mo is a conservative hexavalent oxyanion (MoO42--) in oxygenated water. The largest sink of Mo in the modern ocean (50%) is adsorption to manganese oxides, which preferentially retains light isotopes. This also occurs in MnO2 formed on land. Reporting Mo isotope compositions and correcting published data sets to the NIST 3134 reference solution (δ98Mo = ((98Mo/95Mo)sample/(98Mo/95Mo)3134–1) × 1000)51, river water δ98Mo averages + 0.54‰52 and seawater + 2.09‰ ± 0.1051, higher than the average continental crust based on molybdenite (0.04‰ ± 1.04) 53 or silicate rocks (+ 0.15‰)54. As the size of the marine Mn oxide sink of Mo changes over time, depending on the extent of anoxic basins, the δ98Mo composition of seawater is expected to also change over time. For example, prior to OAE-2 at 94 Ma, seawater δ98Mo was lower (+ 1.65‰)55. In conditions where free dissolved sulfide exists at concentrations exceeding 10 mM, Mo is quantitatively removed, and underlying sediments can preserve the isotopic composition of Mo in seawater. Today, euxinic basins are restricted to isolated systems like the Black Sea and fjords where exchange with the open ocean is poor, but in the past they may have been more common. In anoxic but non-euxinic basins, the isotope composition of sedimentary Mo depends on the presence or absence of Mn-oxides and bottom water oxygen. If porewater sulfide is sufficiently close to the sediment–water interface, while bottom water may not be euxinic, seawater Mo can diffuse into the sediment to give them a similar δ98Mo to seawater, but the Mo concentration will be lower than in truly euxinic basins57. With slightly higher O2 in bottom water, but with anoxic pore water, Mn-oxides with low δ98Mo that formed within seawater can dissolve in the anoxic sediments, releasing their Mo with low δ98Mo. This Mo can diffuse up towards the seawater-sediment interface, or down where it can be trapped permanently in deeper, sulfidic sediments. Today, sediment δ98Mo can range from ~ -1‰ in fully oxic systems to + 2.0‰ in euxinic systems, with intermediate values corresponding to redox state and the presence of sulfide. While modern ooids have a composition close to the seawater they precipitate from, use of skeletal carbonates (δ98Mocarb) as a proxy of seawater can be affected by non-equilibrium (biological) fractionation. Furthermore, δ98Mo of impure carbonates will also reflect contributions from detritus and Fe–Mn oxides, which is much lower than the seawater value.

Samples analysed for Mo isotopes were either carbonate-absent (CaO < 1%) or carbonate-bearing. The Mo in carbonate-absent sediments is derived from detritus or authigenic phases, namely Fe–Mn hydroxide and sulfides, while he carbonate-bearing samples also contain significant amounts of potentially isotopically distinct carbonate. Analyses were made on bulk samples, but to address the isotope composition of seawater derived phases, correction for detrital material is made. This is done by assuming all Ti is detrital, and using the upper continental crust (UCC) Mo/Ti composition to determine the amount of detritus-bound Mo. The isotope composition of the UCC is assumed to represent the detrital material, and for this we use UCC Mo concentrations and isotope composition of 1.1 ppm and + 0.1‰58,59 to subtract the detrital contribution to the whole rock data by isotope mass balance, leaving a δ98Mo value for authigenic (Fe–Mn hydroxide or sulfide) plus carbonate phases. The highest δ98Mo in carbonate will be close to seawater, and our highest δ98Mocarb of + 1.65 is similar to that of Cretaceous seawater suggesting that Late Triassic seawater was somewhat similar55. In the Pignola-Abriola section, the δ98Mo in carbonate-absent samples increases from -3.5‰ at 30 m to just 0.1‰ at 42.9 m, while the δ98Mo in carbonate-bearing samples are more variable (-2.7‰ to + 1.7‰) (Fig. 3).

Figure 3
figure 3

Geochemical profile of Pignola-Abriola section. Enrichment factors (EF) were calculated relative to UCC58,60. δ98Moauth for carbonate-absent is of Fe–Mn hydroxides and for carbonate-bearing is Fe–Mn hydroxides plus carbonate. Strontium curve modified after64. A description of the taxonomic range of conodonts is as follows: 1. Mockina bidentata; 2. Parvigondolella andrusovi; 3. Misikella hernsteini; 4. Mi. hernsteini/posthernsteini; 5. Mi. posthernsteini; 6. Mi. ultima.

Full size image

In an oxic water column, Fe–Mn hydroxides are 3‰ lower in δ98Mo than seawater56, but the lowest values of δ98Mo we measure are more than 3‰ lower than our estimate of seawater at the time of formation. To increase this difference beyond 3‰, dissolution of Fe–Mn hydroxides (bearing low δ98Mo signatures) followed by their re-oxidation is indicated, suggesting that the oxygen content of water close to the seawater sediment interface fluctuated in response to physical restriction, which is unlikely given the paleo-reconstruction of the Pignola-Abriola area, or to changes in the delivery of organic matter, supported by the correspondence between the lowest δ98Mo and highest TOC%. The increase in δ98Moauth from values consistent with a mostly, albeit variable, oxic water column to values consistent with anoxic to euxinic sediments underlying an expanding area of high productivity with a persistent oxygen minimum zone61. The increase in δ98Moauth towards that of contemporaneous seawater indicated by the highest δ98Mocarb as MoEF/UEF also increases supports the concept of intensification of an OMZ driven by increasing productivity, and the progression of a porewater sulfide front towards the sediment–seawater interface.

Possible cause(s)

These environmental perturbations can be explained by various mechanisms, any of which may create a global record, such as dissociation of clathrate hydrate and/or enhanced magmatic activity and outgassing1,62,63. Likely, a large volume of 13C-depleted CO2 entered the ocean–atmosphere system prior to the NRB, possibly as a product of volcanic outgassing by the emplacement of the Angayucham LIP11,21,24,32,64. This hypothesis is supported by the comparison of C-isotope records from marine11 and terrestrial settings21,22 with the seawater Sr-isotope65 and Os-isotope curves66,67,68. The 87Sr/86Sr and 187Os ⁄188Os values started to decline in the upper Norian, indicating a relative increase in the supply of mantle-derived unradiogenic Sr and Os from marine hydrothermal or other mafic igneous sources. Furthermore, the mechanism invoked to explain the spread of reducing conditions also implies significant input of CO2 into the ocean–atmosphere system that would enhance chemical weathering via acceleration of the hydrological cycle during radiative-forced warming and cause increased nutrient discharge (e.g., nitrates and phosphates) to the global ocean1,69,70. The coincidence of these three concatenated steps has been clearly documented worldwide. Our weathering proxy data (Chemical Index of Alteration—CIA, RbEF, and KEF values) have revealed that chemical weathering intensified during the development of anoxic conditions (Figs. 3, 4).

Figure 4
figure 4

Stratigraphic profiles of KEF, RbEF, and CIA in the Pignola-Abriola section. Enrichment factors (EF) were calculated relative to UCC60. KEF and RbEF decrease with increasing CIA values across the NRB.

Full size image

Values of the CIA from the shale samples71 indicate the extent of decomposition of feldspar minerals, which are the most abundant mineral group in the UCC. Because of the high carbonate content of the study section, we used a modified form of the CIA equation by33 as:

$${\text{CIA}}^* = {\text{Al}}_{2} {\text{O}}_{3} /\left( {{\text{Al}}_{2} {\text{O}}_{3} + {\text{Na}}_{2} {\text{O}} + {\text{K}}_{2} {\text{O}}} \right)/100,$$

CIA values are relatively constant below the FO level of M. posthernsteini s.l. sensu34, and then rapidly increase in the S3 interval (sensu32, Fig. 3). These stratigraphic variations in CIA values indicate intensified chemical weathering on continents during the S3 interval, implying that the climate changed to relatively warm and humid conditions. K and Rb were released from primary minerals, possibly micas and feldspars in felsic rocks, during continental chemical weathering. The stratigraphic variations in KEF and RbEF decrease across the NRB, associated with a rapid increase in CIA values (Figs. 3, 4), which suggest the intensity of aluminosilicate chemical weathering increased during the anoxic event.

In theory, enhanced chemical weathering via acceleration of the hydrological cycle would cause increased nutrient discharge to the global ocean, with a consequent increase in biological productivity1,70,72. Barium is commonly used as a tracer of paleo-productivity73,74,75,76, because barite precipitation occurs in decaying particulate organic matter while it sinks to the seafloor74,77,78, which is supported by enhanced BaEF values below high-productivity areas79. Our data show that BaEF increased in the NRB, except in the Panthalassic Katsuyama section, deposited below the CCD (Fig. 1). Hence, our geochemical proxies show an apparent link between the volcanic event and oceanic anoxia via increased chemical weathering and paleo-productivity during the S3 interval.

High CIA continues after the boundary, indicating the persistence on land of the conditions that triggered the OAE even as the ocean began to recover, first exhibited by MoEF/UEF50. δ98Moauth also begins to record increasing oxic conditions well after the boundary, showing the control that Mn has on δ98Mo. The prolonged chemical weathering could be the result of a million-year timescale required for silicate weathering to consume the massive amounts of CO2 emitted during volcanism. Long-term increases in CIA similar to the NRB have also been reported from the early Late Triassic Carnian Pluvial Episode (CPE; 232–234 Ma) and from the Triassic-Jurassic boundary80,81. In the case of the Triassic-Jurassic boundary, it is estimated that 1–2 myr were necessary for chemical weathering80 to consume the large amount of CO2 released by 600 kyr of volcanic activity82. CO2 emissions from voluminous, pulsed volcanic activity across the NRB may have thus caused prolonged continental weathering in the Rhaetian.

In summary, our data document a previously unknown OAE of global extent, similar in nature to other OAEs. The OAE that spanned the NRB extended across the Panthalassa Ocean to both sides of the Pangea supercontinent and it is recorded in both the Northern and Southern Hemispheres, at different latitudes. As is the case for most OAEs, we propose that this Norian-Rhaetian oceanic oxygen depletion event resulted from large-volume emissions of volcanogenic greenhouse gases that accelerated the hydrological cycle, thereby increasing weathering rates and increasing nutrient delivery to the oceans. The increased nutrient supply enhanced primary productivity and export of organic carbon, resulting in oxygen depletion and expansion of the OMZ, which in turn increased the burial efficiency of organic carbon in sediments. Fossil extinctions and reducing conditions were documented in basins around the world (Italy, Canada, NW Australia, New Zealand), leaving an unmistakable record of the NRB OAE. The onset of the stepwise Late Triassic extinctions coincided with the NRB OAE, indicating that the combined climate and environmental changes impacted biota at this time. The trigger of this event is attributed to a volcanic event pre-dating the NRB or an alternative source of volcanogenic gas emissions, likely from the Angayucham LIP.