In the Schandelah-1 record, the abundance of malformations of fern spores across the TJB varies in concert with negative CIEs and sedimentary Hg-enrichments in the Schandelah-1 record (Fig. 4). The record for Schandelah-1 is strikingly similar to the Danish Basin2, suggesting a synchronous widespread mutagenic event at the TJB with a single underlying cause. Multiple mechanisms have been proposed to explain malformation of pollen and spores associated with large igneous province emplacement and mass-extinction. Ozone layer depletion due to volcanic halocarbon emissions may have led to increased UV-B radiation and has been previously linked to malformations in gymnosperm pollen and permanent lycopsid tetrads across the Permian-Triassic boundary18–20. Other explanations include rising global temperature due to volcanic emission causing heat-stress, which can induce polyploidy, abnormal meiosis and cytokinesis in extant angiosperm pollen21. Although increased UV-B radiation and heat-stress could have played a contributing role, these scenarios fall short of explaining the prolonged and repeated mutagenesis in fern communities as seen during the Hettangian in the German Basin after extrusive CAMP activity had ceased (Fig. 4). U-Pb dates of CAMP basalts indicate two phases of extrusive flood volcanism that ended following the Spelae event1,44. This also concluded the main eruptive phase of CAMP emplacements, while only a few sill intrusions are observed within the Planorbis zone (early Hettangian)1. This post-ETME intrusive phase was likely restricted to passive and diffusive emissions unable to significantly affect the stratosphere, where the ozone layer resides19. The absence of eruptive volcanism would lead to a relatively quick recovery (decades to hundred years) of the ozone layer eliminating UV-B radiation as multi-million-year stressor. In contrast, Hg is continuously present in the Earth’s crust and can be mobilized though gaseous volatilization, transported in particulates (rivers/runoff) and organic matter degradation that involve several surface processes such as deforestation, weathering and wildfires49. The strong correlation between Hg-concentrations and elevated malformed spore abundances throughout the studied section (Fig. S3) suggests that the accumulation of mobilized mercury in terrestrial environments played a key role.
The enrichment of Hg in sediments is governed by several processes that drive burial and preservation on geological timescales. Most Hg is adsorbed to S-rich minerals or organic matter (OM), a process that takes place mostly in aquatic systems50. The size of the carbon sink normally dictates the adsorption potential and concentration of Hg in sediments27. This is clearly observed for the Marshi CIE anomaly, where the highest Hg-concentrations are correlated with high TOC levels. In NW Europe this event is linked to widespread marine transgression12 driving increased carbon burial. A decoupling of Hg and TOC is observed at the Spelae CIE/anomaly (Fig S2), suggesting carbon burial did not play a significant role. The pathway of volcanic Hg-enrichment in marine sediments is facilitated through atmospheric deposition of oxidized Hg2+ and subsequent scavenging by organic matter (OM). In addition, other minerals such as sulfides and clays can be dominant hosts of Hg31,51. For instance, during photic zone euxinia (PZE) in the upper water column, the amount of free H2S can result in in situ pyrite framboid formation. These conditions favor Hg-mitigation through S-drawdown in the case of excess Hg, overriding the OM-drawdown29. For the Schandelah-1 record, a short-lived peak in TOC at the onset of the Spelae CIE is indicative of a transgression with several sites showing evidence for euxinic conditions at this level15,16. Therefore, the initial rise in bulk Hg-concentrations is likely coupled to increased sulfur-binding and burial, while the subsequent decrease in TOC and increase in Hg/TOC during the Spelae CIE reflects excess burial of volcanogenic Hg. Although CAMP-eruptions spanned at least 800 kyr, many global sites only record a single prominent Hg-anomaly at the Spelae CIE2,25,27,28,52,53. Early activity in the CAMP was mostly restricted to intrusive volcanism leading to sill emplacement, while later phases were dominated by extrusive basalt eruptions that overlapped with the Spelae CIE45,46,54. Global atmospheric distribution of Hg requires atmospheric degassing which would be most efficient during extrusive phases.
Similar to the Spelae anomaly, the four Hettangian intervals with Hg-anomalies show a decoupling of Hg to TOC. The cyclic nature of these Hg-enrichments and the absence of eruptive CAMP activity suggests they are unlikely driven by volcanism. Instead, terrestrial reservoirs (bedrock, soil and vegetation) need to be considered for their ability to accumulate Hg which can be intermittently delivered to shallow marine depositional environments. In addition, the effects of redox conditions and early diagenesis need to be taken into account55. We utilize the Hydrogen Index (HI) and Oxygen Index (OI) of organic matter to assess the influence of post-depositional oxidation on our Hettangian Hg record. Hg/TOC show no significant correlation with either HI or OI (Fig. S2), although three samples with exceptionally high Hg/TOC values (> 200 ppb/wt%) exhibit lower HI in the upper Hettangian red clay intervals. This could indicate the presence of halted paleo-redox fronts which are characterized by lower HI and higher OI in open marine systems, expressing itself through sharp increases of Hg/TOC55. However, the organic matter from the Schandelah-1 section is clearly dominated by terrestrial input12 where intervals showing elevated Hg/TOC coincide with pulses of increased weathering as evident from reworked palynomorphs12. This indicates that particle-bound Hg was delivered via increased transport of terrestrial sediments/organic material.
Large-scale volcanic activity increases the global Hg budget, resulting in Hg accumulating in terrestrial environments38,56. Terrestrial reservoirs, such as plants and soils, can accumulate Hg and can act as a source to shallow marine/lacustrine environments when Hg is remobilized30,31,56. The residence time of Hg determines the redeposition potential, which is closely tied to soil-organics in terrestrial environments. About 50% of Hg deposited on the ocean’s surface is re-emitted to the atmosphere, while only a small fraction (10%) of Hg deposited in soils is recycled57. This results in a residence time of Hg in soils of about 1000 years with a total residence time of 3000 years in the atmospheric-ocean-terrestrial system57. However, bedrock reservoirs, such as coal beds and mineral sources, have accumulated Hg over longer geological timescales which can be similarly mobilized by erosion and runoff58. Part of this terrestrial Hg finds its way to the marine realm via rivers which may contribute up to 10% of the total oceanic Hg input59. Thus, the terrestrial Hg flux to shallow marine basins is significantly larger than the total oceanic input. Periods of increased runoff and erosion would ultimately displace large quantities of terrestrially-stored Hg from vast catchment areas and concentrate it in low-lying basins and deltaic/coastal fronts, such as the Central European Basin (CEB; Fig. 1). This can occur within relatively short geological intervals where it potentially imposes mutagenic consequences on the surrounding vegetation. In addition, a regional collapse of terrestrial ecosystems through the loss of dense forest vegetation and soil erosion would further impede Hg-mediation through organic-sequestering and subject the prolificating fern vegetation to relatively higher concentrations of Hg. However, the mutagenic potency of Hg on vegetation would be significantly magnified if shifted to its gaseous form (Hg0), circumventing root-protective systems and being absorbed through stomatal uptake22–24,37.
The volcanic origin of the Spelae anomaly in the Schandelah-1 record is supported by the sharp positive shift in MIF co-occurring with a negative shift in δ13CTOC (Fig. 4). An increase in atmospheric dispersal of volcanic Hg0 (assumed to be near-zero MIF) fits the observed pattern, causing the Δ199Hg values of in those beds enriched in Hg to shift towards 0‰ (Fig. 5A, black dashed arrow). In the Rhaetian interval, where background Hg-concentrations are low, Δ199Hg values show minimal variability (–0.50‰ to − 0.30‰), which is more consistent with a source from plant material. Negative MIF in vegetation is attributed to photochemical reduction (loss of Hg) in foliage37,41, whereby subsequent litterfall carries very low MIF values (Δ199Hg = − 0.6‰ to − 0.2‰)39. Mass-dependent fraction (δ202Hg, − 2.00‰ to − 1.00‰) shows considerably more variation during the Rhaetian (Fig. 5B). Positive shifts in MDF in the Triletes Beds likely driven by the reduction of Hg2+ via microbial or abiotic processes39. Although Hg within the Triletes Beds is derived from plant sources (highly negative MIF), the low abundance of vegetation impeded plant uptake of Hg causing a decreased flux to soils, which was subsequently subjected to degradation (causing positive MDF). In addition, the low Hg concentrations are indicative of the absence of volcanic Hg-input. Subsequent volcanic Hg-enrichment at the Spelae anomaly caused a notable shift towards more negative MDF, indicative of a combined atmospheric (volcanic) and plant influence.
During the Hettangian, higher MIF values (Fig. 4D) mark a shift in Hg-sourcing and/or change in photochemical reduction. Both Δ199Hg and δ202Hg values exhibit repeated positive shifts in tandem with Hg-anomalies, which suggest that increased input of Hg was either derived from a mixture of multiple terrestrial sources and/or isotopic alteration during transport42. For instance, modern soil reservoirs typically show higher MIF (Δ199Hg = − 0.2 to + 0.1‰) and likely represent a contribution of both of leaf foliage (litter fall, low MIF) and mineral sources (near-zero MIF)37,49,60. Modern foliage also tends to show lower δ202Hg values than those of mineral soils41,61. Other studies suggest significant shifts in MIF can be attributed to a higher proportion of reworking from abundant Triassic coal beds52,58. Fossilized terrestrial organic matter derived from coal deposits shows low MIF49. This suggests that the terrestrial Hg-isotope signature in the marine system is dependent on the composition of the transported material. While during the Rhaetian, the MIF composition of Schandelah-1 represents a vegetation endmember (low Hg/TOC and highly negative MIF), during the Hettangian MIF could represent a mixture of plant (litter fall) and soil/bedrock (mineral) derived Hg. The fraction of Hg-mineral sources is larger during intervals of high Hg/TOC indicative of soil-weathering and redeposition of older coal/bedrock-derived Hg. In contrast, during intervals that show background Hg-concentrations, the fraction of vegetation-derived Hg was likely higher, which resulted in relatively low Hg-input to shallow marine environments with lower MIF/MDF.
Hettangian variability in MIF corresponds with Hg-anomalies thus indicating swings in soil/bedrock-derived Hg-mobilization and vegetation-mediated Hg-sequestering (Fig. 5A, blue dashed arrow). We use the ratio of Δ199Hg/Δ201Hg to further examine MIF variability (Fig. 5C). A slope of ~ 1.00 is shown to be derived from photochemical reduction of Hg2+ and is consistent with rainfall, foliage, sediments and coals39. This confirms that most of the Hg in the Schandelah-1 section is derived from terrestrial sources. However, photochemical processes impose another important control on environmental Hg-mobilization by causing volatilization to gaseous elemental Hg0. Photochemical changes in Hg-speciation (degassing) can also lead to increases in both MDF and MIF39. During the Hettangian, the repeated positive shifts in the MDF and MIF signatures hint towards a combined effect of increased soil/bedrock-derived Hg and photochemical reduction. Photochemical reduction of Hg, which yields higher MIF (higher Δ199Hg values), shows correlation with water depth62, turbidity63 and shading from canopy64. Sediments from salt marshes also show highly positive MIF, which has been attributed to in situ photochemical reduction65. Hence, shifts towards open landscapes through the loss of high canopy vegetation affects the mobility of Hg, not only by making soils and the underlying bedrock more susceptible to erosion, but also by increasing exposure to sunlight causing volatilization/degassing of terrestrial Hg and allowing for increased photochemical reduction.
In Schandelah-1, Hg-enrichments driven by volcanic emissions coincide with a negative CIE (Spelae) reflecting gaseous Hg being distributed atmospherically as noted by a positive shift in MIF (Fig. 6A). On the other hand, shifts in landscape and vegetation composition during the Hettangian seem to have direct implications for the speciation of Hg and interactions with the environment. This pattern is also evident in sections in China that straddle the Triassic-Jurassic boundary, confirming a prominent role terrestrial Hg during the early Hettangian through increased weathering38. Previous examinations of the Schandelah-1 core have revealed the periodic nature of the Hettangian to be the result of increased wildfire activity and weathering due to major swings in the hydrological regime12. Shifts to open landscapes, likely facilitated by wildfire activity12 resulted in large inputs of terrestrial material and soil/bedrock-stored Hg, driving positive shifts in both MIF and MDF. In addition, open and flooded landscapes would provide the ideal setting for terrestrial (soil, bedrock and vegetation) Hg to be photochemically reduced and shift to a gaseous Hg0 species (Fig. 6B). This would additionally cause positive shifts in MIF/MDF. This further indicates that cyclic positive shifts of both MIF and MDF during the Hettangian were unlikely caused by volcanic inputs due to the opposite shifts for MIF and MDF values, as seen in the Spelae anomaly. A negative shift in MDF at the Spelae anomaly suggests higher influence of biological processes with minimal impact of photochemical reduction in the terrestrial realm during volcanic activity.
While the main extinction phase (Triletes Beds) is characterized by low Hg-concentrations, the Spelae anomaly shows evidence of increased extinction rates and increased stress/mutagenesis in fern communities (i.e., spore malformations), suggesting this to be an expression of atmospherically (gaseous) distributed Hg-toxicity derived from eruptive volcanic activity. Although other toxic metals are emitted from volcanic events and could potentially impose plant mutagenesis, the volatile nature of Hg results in uptake in plants via stomata 37 and increases its re-deposition potential57. During the Hettangian stage, the periodically increased volatilization of Hg0 as a result of photochemical reduction of terrestrially-sourced Hg could have similarly induced mutagenesis in local fern communities as seen by volcanic atmospheric deposition at the TJB (Fig. 5D). Coastal mires and hinterland areas acted as catchments and storage for Hg during times of increased terrestrial biomass production (Fig. 6B). This rhythmic storage and mobilization of Hg was facilitated through orbitally paced climate changes acting on the state of terrestrial vegetation12. Temperate conditions during long-eccentricity (405 kyr) minima promoted high canopy mire vegetation and Hg-sequestering. Increasing seasonal (precessional) contrast during long-eccentricity maxima caused a collapse of vegetation cover and promoted the proliferation of pioneering fern taxa that were subjected to Hg-mobilization and degassing. These local fern communities experienced stress as a consequence as recorded in malformed spores. However, this mutagenic effect seems to be mostly evident from several tree/ground fern families which include Dipteridaceae, Dicksoniaceae and Matoniaceae2, suggesting certain fern species were more susceptible to Hg-pollution or that malformations in less-abundant fern groups are not well recorded. Although Hg-environmental dynamics across timescales of > 100 kyr are not well understood, a collapse of terrestrial and marine biomass due to the ETME likely impeded the re-absorption of excess Hg66 and caused long-term consequences. These disturbed conditions appear to have ceased during the early Sinemurian when a conifer-dominated biome stabilized, and sea level rise and enhanced carbon burial led to mediation of Hg-pollution.
Based on a quantification of spore teratology and Hg concentrations, as well as Hg-isotope records, we establish a link between fern mutagenesis and Hg-pollution due to CAMP volcanism and subsequent mobilization from soil/bedrock reservoirs during times without large-scale volcanic eruptions. Hg-mobilization continued for at least 2 million years up to the Hettangian-Sinemurian boundary under extreme greenhouse climate conditions. It can be shown that Hg was mainly sourced through continental erosion during the Hettangian based on positive shifts in δ202Hg and Δ199Hg (MDF and MIF). In contrast, Hg at the Triassic-Jurassic boundary was volcanically sourced. Climate-driven collapse of vegetation through wildfire activity and increased weathering impeded the re-absorption of Hg, which continued to disturb and stress Hettangian coastal ecosystems. The strong correlation of spore malformations with high MIF/MDF values further indicates that the photo-reduction of terrestrial Hg was periodically enhanced during the Hettangian in open coastal/wetland areas, due to the loss of canopy cover. Our results further indicate that gaseous Hg played a significant role in mobility and toxicity, which directly impacted fern communities. Hg pollution may have been especially severe in coastal regions depending on the ability of such areas to absorb and store Hg for longer periods. Although environmental Hg-dynamics over hundreds of thousand years are still unclear, our results point to the long-term implications of large-scale volcanism on terrestrial vegetation following major extinction.