Australia: The Land Where Time Began

A biography of the Australian continent 

LIPs - Siberian Traps

The volcanics

The Siberian Trap province has become a part of geological folklore because of their immense size and they are often depicted to cover vast areas (e.g. Officer & Page, 1996, p. 169). Flood basalts cover, in fact, only 3.4 x 10⁵ km² of northwest Siberia, though pyroclastics and intrusives, particularly sills, increase greatly the local area of the total province to 1.5 x 10⁶ km² (Zolotukhin & Al’Mukhamedov, 1988). There is no doubt the original area of the flood basalts was greater, though proposed figures of 5 x10⁶ km² have been proposed (Czamanske et al., 1998; Kozur, 1998), are probably exaggerated. According to Wignall estimates of their original volume are even more speculative. The volume of the Siberian Traps that have been assumed by many authors is 4 x 10⁶ km³, though only 4 x 10⁵ km³ remains at present. A thick pile of basalts from the Early Triassic was discovered in a borehole from the western Siberian depression which suggests the province may extend to the west beneath the sediments from the Jurassic-Cretaceous of this major basin (Westphal et al., 1998). Within the province itself, however, it is suggested by the distribution of lavas that they do not constitute a single continuous province, rather, the amalgamation of several “subprovinces” (Mitchell et al., 1994).

In western Siberia volcanism began with the formation of the Tuffaceous Series (Sadovnikov & Orlova, (1993, 1998). These are predominantly basaltic tuffs that underlie the flood basalts in most of the province and they dominate the entire succession in the south of the region. It has proved to be difficult to date the tuffs, but the discovery recently of conchostracan faunas within interbedded strata, as well as their comparison with similar faunas in Chinese sections that are well dated, indicates that they span the middle part of the Late Permian (Kozur, 1998). Over wide areas flood basalt flows overlie the tuffs over wide areas and it is suggested by recent radiometric dating efforts that the onset of eruptions was coincident with the Permian-Triassic (P-Tr) boundary. U-Pb dates were obtained from the ash band immediately below the P-Tr boundary in South China (Claoué-Long et al., 1991) of 251.2 ± 3.4 Ma. Sanidine from the same ash band was analysed (Renne et al., 1995) which obtained an ⁴⁰Ar-³⁹Ar age of 249.91 ± 1.52 Ma, which included external error. When the same ash band was redated (Bowring et al., 1998) they obtained a U-Pb age of 241.4 ± 0.3 Ma. It is suggested by these dates that the P-Tr boundary occurred between 250 and 251 Ma, with a greater likelihood for the older end of this age range. Contrasting with this, initial ⁴⁰Ar-³⁹Ar obtained from the Siberian Traps were substantially younger (Baksi & Farrar, 1991), though the date (Renne & Basu’s, 1991) of 248.4 ± 0.3 Ma (internal error only) ⁴⁰Ar-³⁹Ar from a basalt flow near the base of the lava pile at Noril’sk was the first to indicate that the onset of the eruption was close in age to the P-Tr boundary. Redating of the standard (Fish Canyon sanidine) that was used to calculate this age, produced an age of 250.0 ± 1.6 Ma and therefore an even closer correspondence (Renne et al., 1995), including all uncertainties this date becomes 250.0 ± 2.3 Ma (Renne et al., 1998). Within error this is comparable to a U-Pb age of 251.2 ± 0.3 Ma that was obtained (Kamo et al., 1996) from an intrusion in the lower third of the Noril’sk lava pile. Samples were analysed (Venkatesan et al., 1997) from throughout the Noril’sk succession and obtained ages of 247.1 ± 1.9 Ma for the base and 247.6 ± 2.5 Ma for the top of the lava pile which indicates a rapid rate of eruption, possibly 1 Myr or less. A different standard was used (Venkatesan et al., 1997) to calculate their ⁴⁰Ar-³⁹Ar ages, which therefore cannot be compared with the date of Renne et al. though it is clear for the Noril’sk succession, at least, the eruptions began around the P-Tr boundary, persisting possibly for 1 million years or less. This conclusion is, essentially, in accord with the biostratigraphic dating that is available from the region, which suggests that the onset of fissure eruptions occurred in the later part of the Dorashamian Stage, which is the last stage of the Permian, also known as the Changxingian Stage, slightly below the P-Tr boundary as defined by palynological data (Sadovnikov & Orlova, 1993, 1998). A similar age assignment (Kozur, 1998) is suggested by conchostracan evidence. This age that is slightly older is below the resolution that is available from radiometric dating.

Though the timing of eruptions in the Noril’sk region is constrained reasonably well, it is important to remember that this area constitutes only 7% of the total volume of the province (Venkatesan et al., 1997), and so may not be representative of the entire province. Up to 3 km of lavas infill a pre-existing graben from the Late Permian in the Maimecha-Kotui area in the NE of the province. The Maimecha-Kotui volcanics consists of a suite of alkali-ultra basics that are chemically distinct, which appear to be older than the Noril’sk succession, unlike the remainder of the flows in the province, which are tholeiitic basalts. Thus an ⁴⁰Ar-³⁹Ag plateau age was obtained (Basu et al., 1995) of 253.0 ± 2.6 Ma from an olivine nephelinite in the region. The same Fish Canyon sanidine standard was used to calibrate this date, and the recalibrated age (Renne & Basu, 1991) was used to recalculate the ages they reported (Renne & Basu, 1991). The age of onset of Noril’sk fissure eruptions and Maimecha-Kotui eruptions can be compared; they indicate that eruptions in the latter region occurred 0.7-5.0 Myr earlier at the 95% confidence level (Venkatesan et al., 1997). It is clear that much radiometric dating is required from elsewhere in the province.

Noril’sk provides much of the knowledge of the nature of the Siberian Traps. In this area individual flows are small by the standards of CFBPs, rarely exceeding a few 10s of metres in thickness and an extent of a few 10s of km (Sharma, 1997). The eruption history therefore consisted of many small flows of small volume, typically a flow of 10,000 m³ every 10,000 years. An unusually high proportion (10%) of basaltic pyroclastics is contained in Noril’sk lava pile. In total, there are approximately 30 beds, which range from a few 10s of centimetres to more than 100 m thick (Venkatesan et al., 1997). An important component of the Maimecha-Kotui volcanics as well as in the succession to the south of Putorana is tuffs. Their intrinsic cratonic setting is also noteworthy. With the exception of the small Columbia River Province, all other provinces occur at locations that are marked by subsequent continental rifting, or in the case of the case of the Emeishan and Panjal Volcanics, former sites. This relationship is ascribed by many authors to a plume-related mechanism of continental breakup, as noted above (Courtillot et al., 1999). Some of the predictions of the plume model, however, are not present in the Siberian traps, notably the absence of pre-eruption doming (Czamanske et al., 1998). It is suggested by plume modelling of the size required to generate the Siberian traps that up to 4 km of uplift should mark the arrival of the plume at the base of the lithosphere (Farnetani and Richards, 1994). In western Siberia, and in both the Noril’sk and Maimecha-Kotui areas, a pre-existing graben topography in infilled by lava, which indicates that rifting, and not uplift, predated the eruptions (Zorn & Vladimirov, 1989; Czamanske et al., 1998). Nonetheless, the preferred option in most studies of the Siberian Traps remains the plume model (Renne & Basu, 191; Veevers et al., 1994; Conaghan et al., 1994; Coffin & Eldholm, 1994; Sharma, 1997).

Extinction mechanisms

The significance of the Siberian Traps, whatever their origin, is well established due to the central role played by them in the majority of current models for the end-Permian mass extinction event. Attention has been focused by the higher proportion of tuffs (relative to many other CFBPs) in the Siberian Traps on the likely cooling effects of volcanic dust and sulphate aerosols (Campbell et al., 1992; Conaghan et al., 1994; Renne et al., 1992; Kamo et al., 1996; Kozur, 1998). It was therefore proposed (Campbell et al., 1992) that cooling of sufficient to cause an intense glaciation of the duration of the Siberian Traps eruption – a period of 600 ka in their estimation. This, in turn, is suggested to have been the cause of the major eustatic sea level fall at the end of the Permian that has been widely reported (e.g. Holser & Magaritz, 1987). It was suggested (Erwin, 1993), alternatively, that broad uplift that centred on western Siberia, immediately before the onset of volcanicity in the region, may have been recorded by the fall. As noted above, however, this alternative is made untenable by the lack of evidence for uplift in the region (Kamo et al., 1996; Czamanske et al., 1998). There is an equal lack of evidence for end Permian glaciation, though it was suggested (Campbell et al., 1992) that it may have been too brief to have left physical evidence. Reassessment of the latest Permian and earliest Triassic changes in sea level that were recently carried out suggests that the absence of such evidence may not be so puzzling as rapid sea level rise, not fall, is seen in most sections at this time (Hallam & Wignall, 1999). Hitherto, much of the evidence for a fall of sea level in the latest Permian has been based on the absence of diagnostic biostratigraphic markers from the latest Permian, though it was revealed by sequence stratigraphic analysis that the latest Permian-earliest Triassic interval was marked by a phase of rapid coastal onlap (Haq et al., 1987; Wignall & Hallam, 1993). It has been speculated (Hallam, 1999) that a major pulse of intracratonic volcanism may be reflected by the sea level rise. The same effect has been proposed for some changes in sea level that occurred in the Cretaceous (see below), though this proposition has been made difficult to judge  for the P-Tr sea level changes by the absence of any pre-Jurassic oceanic crust.

If glaciation can be discounted as a factor in the mass extinction event in the end-Permian, the same cannot be said for global cooling. The mechanism was favoured as the main cause of the mass extinction (in particular Kozur, 1998). The principal evidence he based this on was from the changing species of conodont thst were involved in distribution during the crisis. In many sections of the Tethyan P-Tr boundary, the appearance of Clarkina carinata, a species that has been interpreted to have a preference for cool water, marks the extinction interval (Kozur, 1998). The temperature preferences of Clarkina species are, however, by no means clear and it noteworthy that other species of Clarkina, such as the C. subcarinata which is closely related, have been inferred to prefer warm water (Kozur, 1998). Wignall suggests it needs to be considered that an alternative possibility, that C. carinata was a deep benthic form that expanded its range during the deepening that was caused by the sea level rise of the end-Permian. Contrasting with the conodont evidence, contemporaneous changes in the terrestrial flora suggested that forms that were cold adapted faired particularly badly during the end-Permian crisis. In the high southern palaeolatitudes the eradication of the Glossopteris flora is a particularly noteworthy feature of the event (Retallack, 1995). Extinctions among the Boreal marine invertebrates in Spitzbergen were also at least as severe as those from the lower latitude sections of Tethys (Wignall et al., 1998).

Acid rain, the other climatic consequence of SO₂ eruptions, has also been postulated as a cause of one of the more intriguing aspects of the extinction interval: the proliferation of fungi. It was revealed by palynological samples from the P-Tr sections around the world an increase in fungal spores in the latest Permian times, in particular in low palaeolatitudes where assemblages are dominated by them (Eshet, 1992). This has been attributed to this degradation, by volcanogenic acid rain, of floral ecosystems resulting in proliferation of fungi on the ample decaying vegetation. An alternative origin was proposed (Hallam & Wignall, 1997) linked to the extinction of many insect orders at the end of the Permian (Labandeira & Sepkoski, 1993). At present insects destroy a large proportion of terrestrial flora, as they probably did in the Permian, but fungi may respond to the increased availability of decaying vegetation in the absence of insects. According to Wignall a factor that requires consideration (Wignall et al., 1996; Kozur, 1998), is the possibility that the fungal spores event records the proliferation of marine rather than terrestrial fungi, and this weakens further the link between the fungal spike and acid rain.

The most obvious climatic signal that is associated with the mass extinction event of the end-Permian is a major phase of global warming. The evidence includes calcareous algae migrating to the Boreal latitudes (Wignall et al., 1998), the preferential loss of high latitude floras noted above (Retallack, 1995), and palaeosols developing that are typical of latitudes of 50^o or lower at palaeolatitudes as high as 80^o (Retallack, 1999; Retallack & Krull, 1999). Equatorial temperatures are suggested by oxygen isotope data to possibly have risen by as much as 6^oC at the P-Tr boundary, though such inferences need to assume the fluctuations of salinity can be disregarded (Holser et al., 1989). Strontium isotope ratios in marine carbonates and phosphates also indicate an increase of CO₂ around this time. After the low point at the end of the Middle Permian, the ratios of ³⁷Ar/³⁶Ar began to rise at an increasingly rapid rate, reaching a high point at the end to the Early Triassic (Martin & McDougal, 1995). Increased continental weathering in an atmosphere that was increasingly rich in CO₂ may be reflected by this trend (Martin & Macdougall, 1995) and/or enhanced leaching of the volcanogenic acid rain (Conaghan et al., 1994). Alternatively, the enhanced continental erosion during major sea level fall (Holser & Magaritz, 1987), does not account for the lack of correlation between the eustatic sea level curve and the isotopic trends (Hallam & Wignall, 1997) allam & Wignall, 1997). Hallam Hallam

as reflected by the Sr isotope curve.

Wignall suggests that global warming may also have been the cause of the rapid development of marine anoxia as seen in many shelf sections from the latest Permian (Wignall & Twitchett, 1996). In the deep water pelagic chert sections from the accreted terranes of Japan at the end of the Permian this “superanoxic” event first developed (Isozaki, 1997), though it is the expansion of oxygen-poor conditions into shallow waters in the latest Permian which coincides with the marine mass extinction event (Wignall & Hallam, 1992, 1996; Wignall et al., 1995). The decline of the equator-to-pole temperature gradient and the consequent increase in oceanic circulation, as well as the lower solubility of oxygen in warmer waters, are the 2 effects of warming that may have been responsible for marine anoxia (Wignall & Twitchett, 1996). Though the ultimate cause of the elevated atmospheric concentrations of CO₂ required to generate the warming is not known. The preserved source of many is volcanic CO₂ that was derived from the Siberian Traps (e.g. Campbell et al., 1992; Conaghan et al., 1994; Veevers et al., 1994; Wignall & Twitchett, 1996), though contributions from gas hydrates (Erwin, 1993; Morante, 1996), and oxidation of coals from Gondwana (Faure et al., 1995), may also have been important.

Carbon isotopic trends

According to Wignall in marine as well as terrestrial sections during the P-Tr interval dramatic δ¹³C fluctuations may hold a clue to the cause of changes of atmospheric CO₂. Values of δ¹³C gradually decreased in marine carbonates through the Dorashamian Stage prior to a rapid negative excursion, of 4-5‰ magnitude, began just below the P-Tr boundary (Holser et al., 1989, 1991; Xu & Yan, 1993; Morante et al., 1994; Morante, 1996). A brief influx of isotopically light carbon is indicated by this negative spike into the ocean-atmosphere system around the time of the end-Permian mass extinction event. A potential source is volcanogenic CO₂, though its δ¹³C value of -5‰ (McLean, 1995) is probably not sufficiently light to achieve the swing that is observed. This becomes readily appreciated when the amount of C in surficial reservoirs, which is estimated to be around 5 x 10⁹ g C (Berner, 1999), is compared with the volumes of CO₂ that has been estimated during the formation of CFBPs. It has been calculated, based on measurements from Hawaiian eruptions (McCartney et al., 1990), that 5 x 10¹² g of C is emitted by 1 km³ of basalt. Therefore, if it is assumed that a high value of 2 x 10⁶ km³ of basalt was present originally in the Siberian Traps, and then 1 x 10¹⁹ g of CO₂ C may have been released during the eruption. This is enough to cause only roughly 20% of the isotopic swing that is observed in the latest Permian. It is clear that there are many assumptions implicit in this reasoning, not least of which is the -5% isotopic composition of volcanic CO₂. The debate on the ultimate origin of CFBPs, in this context, is apposite to this assumption. It was suggested (Anderson, 1994, 1999) that rather than being sourced from deep mantle plumes, CFBPs may result from shallow mantle enriched in recycled lithosphere melting. If this layer, that is termed the perisphere, is enriched in isotopically light, subducted organic C then the δ¹³C of the volcanogenic CO₂ that was emitted from the source may be considerably lighter than the -5‰ value. The isotope excursion that occurred at the end-Permian could have been caused by the eruption being about 1 x 10¹⁹ g C with a δ¹³C value of -20‰. Before these speculations are accepted it would be salient, however, to recall that much, if not most of the Siberian traps province was erupted in the early Triassic after the δ¹³C excursion. At the latest Permian the sudden influx of light C into the atmosphere and oceans has been attributed to a near-total collapse of primary productivity (Holser & Margaritz, 1992; Margaritz et al, 1992; Wang et al., 1994). Once again, however, consideration of the masses involved highlights that are not likely to cause more than a small negative shift. Therefore, the modern biomass, which amounts to 8.3 x 10¹⁷g C, and which has an isotopic composition of -20‰, if it was added to the 5 x 10¹⁹ g C that is present in the inorganic C component of the ocean-atmosphere system, would clearly have little impact on values of δ¹³C.

It has been appreciated in recent years that one of the major repositories of C is isotopically light C which occurs in the form of methane hydrates (clathrates) that are buried at shallow depths beneath cold and/deep seas (dickens et al., 1997). Isotopically, hydrates are very light (δ¹³C = -65‰), and there may be as much as 1 x 10¹⁹g of C present in the hydrates beneath modern seas. The release of only 10% of this material into the atmosphere as methane would be enough to cause the δ¹³C shift that is observed across the P-Tr boundary (Erwin, 1993; Browning et al., 1998). The magnitude of the end-Permian δ¹³C excursion clearly renders its origin problematic. The eruption of vast volumes of CO₂ and the shutdown of productivity may have occurred at this time, though these events can account for only a fraction of the C isotope changes that are observed. The warming effect of the release of CO₂ may, however, have triggered the disassociation of huge amounts of methane hydrates, thereby producing the δ¹³C excursion, exacerbating the warming trend. There is substantial geological and palaeontological evidence for conditions that were globally warm in the Early Triassic, as noted above. There have been proposals for similar “runaway greenhouse” scenarios for the Cenomanian-Turonian interval of the Cretaceous as well as in the Palaeocene (see below) with flood basalt volcanism implicated in both cases.

Sources & Further reading

Wignall, P. B. (2001). "Large igneous provinces and mass extinctions." Earth-Sci. Rev. 53: 1-33.