The PT extinction was rapid, probably taking place in less than a million years. Although it was much more severe in the ocean, it affected terrestrial ecosystems too. A prolific swamp flora in the Southern Hemisphere had been producing enough organic debris to form coals in Australia, but the coal beds stop abruptly at the P-T boundary. No coal was laid down anywhere in the world for at least 6 m.y. afterward. A large change in carbon isotopes occurred across the PT boundary, which signifies an important and global drop in photosynthesis that lasted a long time.
There is some evidence for an impact at the PT boundary. The continental collisions that formed Pangea in the Permian would account for a major drop in diversity but not for a sudden, enormous mass extinction. Perhaps most important of all, the Permian extinction coincides with the largest known volcanic eruption in Earth history.
In addition to plate tectonics, the Earth also has plume tectonics. Occasionally, an event at the boundary between the Earth's core and mantle sets a giant pulse of heat rising toward the surface as a plume. As it approaches the surface, the plume melts the crust to develop a flat head of basalt magma that can be 1000 km across and 100 km thick. Penetrating the crust, the plume generates enormous volcanic eruptions that pour hundreds of thousands of cubic kilometers of basalt‹flood basalts‹out over the surface. If a plume erupts through a continent, it blasts material into the atmosphere as well. After the head of the plume has erupted, the much narrower tail will continue to erupt for 100 m.y. or more, but now its effects are more local, affecting only 100 km or so of terrain as it forms a long-lasting hot spot of volcanic activity.
Plume events are rare: there have been only eight enormous plume eruptions in the last 250 m.y. The most recent is the Yellowstone plume: at about 17 Ma it burned through the crust to form enormous lava fields that are now known as the Columbia Plateau basalts of Oregon and Washington, best seen in the Columbia River gorge. North America drifted westward over this "hot spot," which continued to erupt to form the volcanic rocks of the Snake River plain in Idaho (Valley of the Moon and so on), and it now sits under Yellowstone National Park. The hot spot is in a quiet period now, with geyser activity rather than active eruption, but it produced enormous volcanic exposions about 500,000 years ago that blasted ash over most of the mountain states and into Canada.
A massive plume eruption took place exactly at the PT boundary. A new plume burned through the crust in what is now western Siberia to form the "Siberian Traps," gigantic flood basalts that cover 2.5 million sq km in area and are perhaps 3 million cu km in volume. The eruptions coincided exactly with the PT boundary, at 253 Ma, and lasted at full intensity for only about a million years‹the largest known, most intense eruptions in the history of the Earth. They lie across the PT boundary and were formed in what was obviously a major event in Earth history.
There is a feeling, particularly among physical scientists, that if we can show that a physical catastrophe occurred at a boundary, we have an automatic explanation for an extinction. But this connection has to be demonstrated, not just assumed.
In 1993 Douglas Erwin felt obliged to suggest the "Murder on the Orient Express" hypothesis for the PT extinction: that is, many factors, all acting together, led to the extinction. This is not a particularly "clean" hypothesis to accept or to test. However, we have eight more years' research now, by Erwin and others, and we can do better. Any explanation of the PT event must account for the severity of the extinctions. The PT extinction is the largest in Earth history, but so was the Siberian Traps eruption. In 1995 Paul Renne and his colleagues suggested a scenario based on the eruption as a primary cause of the extinction.
The plume rises toward the crust and erupts. The tremendous amount of sulfate aerosols would cool the climate enough to form ice-caps, rather quickly, and this in turn would cause a rather rapid drop in sea level along with global cooling, early in the eruptive sequence. In the rock record, we would expect to see changes in carbon and sulfur isotopes, and we do. Furthermore, as the plume erupts, the crust would be raised by the buoyant magma, perhaps enough to form a land footing for the large continental ice sheets that would grow in these high latitudes. Finally, as the eruption dies off, the crust would subside and the aerosols would disperse, making for a rapid end to the volcanically induced glaciation and another rapid change in climate. It is possible, but not calculated yet, that the volcanic gases that had built up during the eruption could have had a greenhouse effect for some time after the eruption ended, taking the earth from a volcanic glaciation to a volcanic hothouse.
Add to this scenario the "usual" effects of a giant eruption, such as acid rain, ozone depletion, a massive dose of carbon dioxide into the atmosphere, or any combination of the above, and the ingredients are in place for a mass extinction.
Though it is easy to imagine that a giant eruption might have caused a catastrophe at the PT boundary, it is not certain that it would. We do not know how much dust, smoke, and aerosols would be produced, even though it is absolutely critical to calculations of their effects that we know those factors rather precisely. We do not know how far volcanic aerosols and stratospheric dust would be carried over the Earth, or in detail what effects they would have. Dust and aerosols in the air can help absorb solar heat rather than reflect it, for example, and some models suggest that parts of the Earth would warm, parts would cool, and parts would stay at about the same temperature.
The most persuasive scenarios of volcanic extinction are quickly summarized. Even a short-lived catastrophe among land plants and surface plankton at sea would drastically affect normal food chains. Large animals would have been vulnerable to food shortage, and their extinction after a catastrophe seems plausible. In the oceans, invertebrates living in shallow water would have suffered greatly from cold or frost, or perhaps from CO2-induced heating. High-latitude faunas and floras in particular were already adapted to winter darkness, though perhaps not to extreme cold. Thus, tropical reef communities could have been devastated, but high-latitude communities could have survived much better.
These general patterns are observed at the PT boundary, though high-latitude floras were affected worse than one would have predicted. Even so, the patterns do not prove that the eruptions caused the extinction: other factors could have been at work too. Some specific evidence shows that eruptions do not necessarily cause catastrophes. For example, the eruption of Krakatau in 1883 destroyed all life on the island and severely damaged ecosystems for hundreds of miles around. But those ecosystems have completely recovered 100 years later, in a geologically insignificant time. There's no biological trace of the much larger eruption of Toba, 75,000 years ago. No North American extinctions coincided with explosive eruptions from Long Valley caldera, California, from Crater Lake, Oregon, or from Yellowstone, all of which blew ash as far as Canada within the last million years.
Other major plume eruptions are not linked with extinctions: examples include the Jurassic Karroo Basalts of South Africa and the Miocene Columbia Plateau Basalts. However, one should beware of dismissing catastrophic explanations because small events do not trigger catastrophes. There may be a threshold effect: if the event is not big enough it will do nothing, but if it is big enough it will do everything. Perhaps only two eruptions in the last 500 m.y. were large enough to cause a mass extinction, at the P-T boundary and perhaps also the K-T boundary.
In short, we don't yet know whether an eruption would have catastrophic, severe, or only mild biological and ecological effects, or whether those effects would be local, regional, or global. In any scenario, however, the killing agent is transient: it would have operated for only a short time geologically. Clearly, if such events occur, they are rare. That does not make them impossible, only unlikely. And that means they have to be very persuasive indeed before we accept them!
A gigantic eruption might have caused the PT extinction by inducing climatic changes near and at the boundary. But climatic changes can be caused by more normal agents such as geographic change, and perhaps these could have caused the extinctions. In particular, the Earth's climate is driven dominantly by oceanic factors. An eruption at the PT boundary could have added to a climatic change that was already happening. We should look at some of the data.
In the Late Permian, the continents were clustered together in the supercontinent Pangea (Figure 5.9), which means also that there was a giant ocean, Panthalassa, which covered 70% of the Earth's surface. We understand modern ocean circulation rather well, but that does not mean that we can predict how Panthalassa would have worked.
Yukio Isozaki studied sediments from Japan and Canada which were laid down on the deep-sea floor of Panthalassa. Normal oceanic conditions (as we understand them today) deteriorated around 260 Ma, as the deep ocean water became anoxic. Yet the surface waters remained normal: they supported abundant radiolarians, whose bodies were deposited as chert on the sea floor. However, after about 255 Ma the radiolarians become rarer toward the P-T boundary, and across the boundary there are none at all, suggesting that Panthalassa became anoxic right to the surface. Long after the extinction, around 245 Ma, the surface waters again became oxygenated enough to support radiolarians, but the deep waters were anoxic until about 240 Ma.
If Panthalassa became anoxic right to the surface, this in itself would cause a catastrophic extinction of marine organisms. The side effects of a marine extinction on atmosphere and climate, not yet spelled out, would in turn have affected land plants and animals.
In Isozaki's model, the extinction is linked with a chemical crisis in the waters of Panthalassa that is symmetrical in time. There is no major trigger associated with the Siberian Traps eruptions. Instead, the crisis occurs in stages, as he sees it. The first, in which the deep ocean becomes anoxic, can be associated with a major extinction event at the end of the Guadelupian stage of the Late Permian. The second, with full oceanic anoxia, occurs at the PT boundary. The oceanic crisis resolves itself in stages too, with full recovery taking about 10 m.y.: certainly it took this long before reef-dwellers appear again in the Triassic fossil record.
Isozaki admits that the ocean features he infers do not fit our understanding of the way our modern oceans work. But, he argues, he has the data to make his statements, and the fact that we cannot explain how they occurred simply reflects the fact that we do not understand how Panthalassa worked as a superocean. Isozaki also notes that the rapid changes in carbon isotopes and in sea level across the PT boundary itself are not explained by his data: something else is going on at the boundary. (And that something else looks much more sudden.)
In 1996 Henk Visscher and his colleagues reported extreme abundances of fossil fungal cells in land sediments at the PT boundary. There are hints that the fungi-enriched "layer" is the record of a single, world-wide crisis, with the fungi breaking down massive amounts of vegetation that had been catastrophically killed (there were no termites yet). Such a fungal layer is unique in the geological record of the past 500 m.y. The best evidence we have suggests that there were major extinctions among gymnosperms, especially in Europe, and among the coal-generating floras of the Southern Hemisphere. The vegetation of the early Triassic in Europe looks "weedy," that is, invasive of open habitats. Andrew Knoll and his colleagues have suggested that the extinction was caused by a catastrophic overturn of an ocean supersaturated in carbon dioxide. This would result in tremendous, close to instantaneous, degassing that would roll a cloud of (dense) carbon dioxide over the ocean surface and low-lying coastal areas. An analog might be the recent catastrophic degassing of Lake Nyos, in the Cameroon, where hundreds of people were killed as carbon dioxide degassed from a volcanic lake and cascaded down valleys nearby. The difference is that the proposed PT disaster was global.
In this scenario, the carbon dioxide build-up results from the global geography that included the gigantic ocean Panthalassa. Knoll and colleagues speculated that the abnormal ocean circulation in Panthalassa did not include enough downward transport of oxygenated surface water to keep the deep water oxygenated. With normal respiration and decay of dead organisms, the deep water evolved into an anoxic mass loaded with dissolved carbon dioxide, methane, and hydrogen sulfide. Carbon continued to fall to the sea floor from normal surface productivity, but it was deposited and buried because there was no dissolved oxygen to oxidize it. As carbon dioxide levels fell in the atmosphere, the earth and the ocean surface cooled. Finally, the surface waters became dense enough to sink, triggering a catastrophe as the CO2-saturated deep waters were brought up to the surface, degassing violently. The event would trigger a greenhouse heating and a major climatic warming.
In 1998 Samuel Bowring and colleagues reported that the carbon isotope change at the P-T boundary in South China was probably very short-lived: a "spike" only perhaps 165,000 years long. This suggests a major (catastrophic?) addition of non-organic carbon to the ocean, rather than just a failure in the supply of organic carbon. They suggested three possible scenarios. Two of them are variants of the Siberian Traps scenario above, except that in addition the climatic changes could have set off an overturn of Panthalassa and a carbon dioxide crisis. Their third suggestion is an asteroid impact, but there is not much evidence for that.
Most recently, Greg Retallack and colleagues have found evidence in Australia that suggests a prolonged greenhouse warming set in right at the P-T boundary. Several paleoclimatic indicators suggest the same story, which implies that the role of carbon dioxide was the vital link between any environmental disasters and the extinctions. Carbon dioxide in the atmosphere could have been increased by volcanic eruptions, by oceanic turnover, and it would have been accentuated and prolonged if plants were killed off globally. (World floras and oceanic plankton would have to recover before the carbon dioxide could be drawn down out of the atmosphere.) We may be getting close to the answer here!
So far there is only enough evidence to fix a major impact at one of the mass extinctions in the fossil record, the KT extinction, and perhaps to suggest another at the FF boundary. Likewise, there is only enough evidence to connect giant flood basalt eruptions with two mass extinctions, at the PTr and the KT boundary extinctions. It is clear, however, that the largest known impact and the largest known eruption coincide with undoubted mass extinctions. It would be amazing if that was a coincidence. These questions are still open!
Easy Access References Becker, L., et al. 2001. Impact event at the Permian-Triassic boundary: evidence from extraterrestrial noble gases in fullerenes. Science 291: 1530-1533, and comment, pp. 1469-1470. [Evidence of impact, based on gas samples inside fullerenes (buckyballs)].
Bowring, S. A., et al. 1998. U/Pb zircon geochronology and tempo of the end-Permian mass extinction. Science 280: 10391045. [But see Munila et al. 2001.]
Erwin, D. H. 1993. The Great Paleozoic Crisis: Life and Death in the Permian. New York: Columbia University Press. [Excellent: Erwin clearly labels thoughts and inferences so that we participate along with him as co-thinkers. More geology and paleontology than the title suggests.]
Isozaki, Y. 1997. Permo-Triassic superanoxia and stratified superocean: records from lost deep sea. Science 276: 235238.
Jin, Y. G., et al. 2000. Pattern of marine mass extinction near the Permian-Triassic Boundary in South China. Science 289: 432-436. [A very rapid event around 251.4 Ma.]
Raup, D. M. 1991. Extinction. Bad Genes or Bad Luck? New York: W. W. Norton.
Renne, P. R., et al. 1995. Synchrony and causal relations between Permian-Triassic boundary crises and Siberian flood volcanism. Science 269:14131416. Volcanic scenario.
Retallack, G. 1995. Permian-Triassic life crisis on land. Science 267: 7780.
Ward, P. D., et al. 2000. Altered river morphology in South Africa related to the Permian-Triassic extinction. Science 289: 1740-1743, and comment, p. 1666-1667.
More Technical Reading
Erwin, D. H. 1988. The end and the beginning: recoveries from mass extinctions. Trends in Ecology & Evolution 13: 344349.
Looy, C. V., et al. 1999. The delayed resurgence of equatorial forests after the Permian-Triassic ecologic crisis. Proceedings of the National Academy of Sciences 96: 13,857-13,862.
Looy, C. V., et al. 2001. Life in the end-Permian dead zone. Proceedings of the National Academy of Sciences 98: 7879-7883.
Munila, R. et al. 2001. Timing of the Permian-Triassic biotic crisis: implications from new zircon U/Pb age data (and their limitations). Earth and Planetary Science Letters 187: 131-145. [They don't, and probably can't, tell us whether the extinction was catastrophic or not, contra Jin et al.]
Pfefferkorn, H. W. 1999. Recuperation from mass extinctions. Proceedings of the National Academy of Sciences 96: 13,597-13,599. [Recovery was slow.]
Rampino, M. R., et al. 1988. Volcanic winters. Annual Reviews of Earth and Planetary Science 16: 7399.
Retallack, G. J., et al. 1998. Search for evidence of impact at the Permian-Triassic boundary in Antarctica and Australia. Geology 26: 979982. [They did not find any.]
Retallack, G. J., et al. 1999. Postapocalyptic greenhouse paleoclimate revealed by earliest Triassic paleosols in the Sydney Basin, Australia. Bulletin of the Geological Society of America 111: 5270..
Twitchett, R. J. 1999. Palaeoenvironments and faunal recovery after the end-Permian mass extinction. PPP 154: 27-37. [Recovery was slow.]
Visscher, H., et al. 1996. The terminal Paleozoic fungal event: evidence of terrestrial ecosystem destabilization and collapse. Proceedings of the National Academy of Sciences 93: 21552158.
Copyright RC 2001
Posted here 3/2002
Return to 107 class page