Whatever the scale, no matter the fineness of time divisions that we examine, the geological record throws up evidence for sudden, apparently random, events and for cyclical change. This is as true for huge and lengthy events, such as periods of mountain building and the ups and downs of sea level, as for seemingly trivial events like fluctuations in the width of trees’ growth rings. For some the explanation is well known. Tree-ring cycles and annual rhythms in some minutely layered Precambrian sediments match closely the tiny fluctuation in the Sun’s output caused by an 11-year cycle of sunspot activity. For most there is a great deal more controversy. Perhaps the best documented case of cyclicity extending into the rock record results from the combined signals, with 23, 41 and 100 ka periods, that characterise climate variation within the last 2.5 Ma or so (Chapter 4; Part VII). That astronomical pacing also pops up haphazardly in the sedimentary record as far back as 300 Ma ago. Seeking patterns for curiosity’s sake seems to be as much part of our make-up as paranoia. As regards the second, there are conspiracies and something out there is trying to get you. Most patterns and sudden events in the evolution of Earth and life popped out unannounced from increasingly detailed, sophisticated and comprehensive scrutiny of the record. They are there, despite the incompleteness.
For two centuries antiquarians and palaeontologists have recorded the rise and extinction of animal and plant taxa. Darwinian evolution was not at first the central focus of their work, but the ever finer division of Phanerozoic time since useful fossils first appeared. Elaboration of evolution emerges from that practical objective. ‘Golden spikes’ from radiometric dating calibrate this fossil time-keeping. So, shouldn’t we now signify the passage of geological time properly by using equal divisions, of which years to seconds are at the short end? So we might have, had radiometric dating preceded the fossil hunters’ inquisitiveness, but it would not have lasted long. Fossils are easily found and their sequences of change pose great philosophical conundrums to curious people. Geological time division in the Phanerozoic is essentially a close reflection of biological evolution (Chapter 5).
Chopping up the rock record starts at the finest division of Zones, which are recognised by the rise and fall of individual species that were distributed over large areas. Each Stage, which includes a number of zones, is defined by the appearance and disappearance of globally distinctive associations of several species and higher taxa. Geological Epochs and Periods (Cambrian, Jurassic etc) begin and end with gross, global changes in flora and fauna, as do the Eras (Palaeozoic, Mesozoic, Cenozoic). The fundamental boundary between the Precambrian and Phanerozoic Eons was set where fossils with hard parts first become commonly found in the rock strata. There are other dividing characteristics, such as evidence for mountain building and erosion, and rises and falls in sea level, but the primary global division is biological. Because it came long before our ability to measure time itself, as the only globally useful means of dividing time in sedimentary rocks, which every geologist knows (or should know), the biological time scale is with us for keeps.
Figure 19.1(a) Diversity of marine animals during the Phanerozoic expressed as the number of recorded families. The curve is built from records of families in groups that dominated the Cambrian, the rest of the Palaeozoic and Mesozoic to modern times, with a smaller addition of those found only in rocks showing exceptionally good preservation. (b) Extinction rate per million years of all plants and animals at the genus level.
Splitting time on a life or death basis is sensible, as a graph of marine-animal diversity through the Phanerozoic (Figure 19.1a) shows. Of the nine System and Era boundaries only two are not signalled by significant ups or downs in diversity at the family level. Looking at diversity at the level of the genus, and at the average rate at which animals in this taxonomic division disappeared, not only signals each major boundary but reveals lots of other apparent events (Figure 19.1b). ‘Apparent’ is an appropriate adjective because of the incompleteness of the fossil record. How incomplete it is becomes clear by looking at a couple of details on Figure 19.1a. The right end of the graph suggests about 900 families of marine fossils in the youngest of all sedimentary sequences, but there are about 1900 families alive today. The topmost division of the curve – those families known only from those few cases where delicate animals show as fossils – adds significantly to the whole. The fossil record changes all the time as advancing research adds new discoveries, and even when palaeontologists change their minds and reclassify taxonomic groupings. Comparing the diversity of marine families living today in the Great Barrier Reef with that in the North Sea shows the potential for uncertainty in Figure 19.1; diversity depends on environmental conditions. One reason for the patchiness of the geological record is changing sea level. When it stands high it floods flat continental margins where sediment preservation is most likely. Low stands shrink the depositional area of easily preserved sediment and older sediments are eroded away.
Such cautions form one reason for avoiding the use of variations in the numbers of individual species to divide geological time. Incompleteness results in far more species failing to show up than there are those that become fossils. The broader taxonomic divisions of genus or family, which include several species that are more widespread, are more likely to leave us fossil representatives. For instance, a single specimen from just one of potentially many included species signifies the family’s presence in the record. The incomplete fossil record is all that we have, yet it is odd enough to encourage explanations, speculative as they might be. Every geologist now accepts that there have been five global mass extinctions, and lots of lesser ones. The magnitudes of the ‘Big Five’ are fearsome. At the end of the Ordovician 28% of earlier families disappeared. Scaling up to the level of species, using statistics for living organisms, suggests the extinction at the Ordovician-Silurian boundary of as many as 85% of all life forms. The 57% extinction of families at the close of the Permian, and therefore the Palaeozoic Era, suggests a staggering 96% snuffing out of species. Those ending the Devonian, Triassic and Cretaceous are not so extreme. The mass extinction at the end of the Cretaceous and the Mesozoic Era has had a big press mainly because that is when the dinosaurs bit the dust. But it is also an event where we know with most certainty that life on land as well as marine animals suffered more or less equally.
An extinction timetable
Fossil records are more complete in rocks deposited since the end of the Carboniferous Period than from earlier times. They also contain organisms that are taxonomically more like modern ones and therefore easier to classify. So the succeeding 290 Ma are a focus for more detailed studies. Life after the Carboniferous encompassed 10000 genera of marine animals, of which 6000 are now extinct. The data include all groups from unicellular, shelled foraminifera to vertebrates, from a great variety of marine habitats. Figure 19.2 shows the overall rate of extinction since 260 Ma ago. Extinctions at the genus level came in pulses very approximately 30 Ma apart. Since the time series encompasses all marine animals and habitats, a fairly regular emergence of harmful conditions for marine life seems to be inescapable. Marine arthropods (crabs etc.), echinoderms (sea-urchins etc.) and vertebrates, considered alone, do show some gaps. Perhaps those groups were in some way better equipped to survive whatever factors periodically killed more sensitive creatures in large numbers.
Figure 19.2 Time series since the Permian for extinctions at the level of genus of all marine animals. Vertical arrows mark significant mass extinction; red bars show periods of major flood basalt volcanism; red stars are impact craters >30 km diameter.
Despite life’s immense complexity, there appears to be an element of simplicity and generality in the record of Mesozoic and Cenozoic mass extinctions. What can it be due to? Any underlying ‘pace-maker’ has to encompass all the factors that may have sent trials of fitness among living things. The pace itself appears not to be exclusive to life, but may carry through other features of the evolving Mesozoic and Cenozoic Earth System. The timing of major mountain-building events on all continents throughout the Phanerozoic shows a periodicity of between 30-36 Ma. Igneous events that involve the deep mantle to generate flood basalt provinces come in bursts roughly every 30 Ma (Figure 19.2). The curve of sea-level change (Figure 5.5) comprises many superimposed periodicities, the strongest of which comes out at 33 Ma. Oxygen-isotope records of climate for the last 130 Ma show 30 Ma repetitions. Even the record of changes in the Earth’s magnetic field, produced in the fluid outer core, includes a 30 Ma signal. For good measure we can also throw in occurrences of evaporite deposits, pulses in sea-floor spreading, and times when unoxidised organic matter built up on the ocean floor. Even the most optimistic poker player would gybe at such a run of coincidences, especially since all these seemingly unrelated processes lumped together share a 26.6 Ma periodicity (Figure 19.3) For such a peak to emerge from random data has odds of 10000 to 1, but a roughly common pace does not necessarily imply a common causality.
Figure 19.3 Spectral power of time series for extinctions and seven global geological processes plotted against the logarithm of frequency in events per Ma. The highest peak indicates a periodicity of about 27 Ma.
Two broad conclusions could be drawn from all this. One is that you can make anything out of numbers, especially if the data are imprecise; as popularised by Mark Twain, there are ‘lies, damned lies and statistics’. But applying statistics to scientific information is a ‘respectable’ and often valuable approach, provided we have a clear sense of perspective; why collect the data in the first place, if they are not to be analysed? The other general conclusion is that there may indeed be a causal link between geological events and mass extinctions. But setting out to tie down the connection from the available, diverse records is a bit like a historian trying to prove the Shakespearian implication that Richard III was a nasty piece of work because he had a spinal deformity, ate a lot of fish and his reign was beset by a run of wet summers – all true.
One view takes empiricism a stage further – and probably much too far – by linking the coincidences to the roughly 30 Ma period in which the Sun wobbles through the plane of the Milky Way galaxy in its orbit around the galactic centre. If that was matched by minute changes in the gravity affecting the Solar System it might perturb the Kuiper belt and Oort Cloud or the main belt of asteroids from which powerful projectiles come. As you saw in Chapter 10, the Earth and Moon do have a common history of bombardment and near-instantaneous releases of energy that dwarf the biblical concept of Armageddon. Big impacts are the worst imaginable trials for life. A hill-sized ball of ice, rock or solid metal arriving tomorrow would be pretty good support for that view, if anyone was still around to document its aftermath. The problem is that the timing of cratering on the Moon is almost entirely relative. Radiometric dating of impact-glass samples returned by the Apollo missions from six points on the lunar surface just confirms that bombardment continued almost to the present day; there is no time series.
The analysis of Earth-wide geological events can serve to support a long-held, suspicion that biological extinction is somehow related to inorganic terrestrial events, including igneous activity, mountain building, fluctuations in sea level and land area, climate change and changes in ocean chemistry. In other words, that periodicity is a property of a dynamic and complex Earth system, for which there is no reason to appeal to outside factors. Many scientists remain deeply suspicious of their ‘whizz-bang’ colleagues, saying that extraterrestrial forces lack only sex, major crime and the British Royal family as ingredients for tasty global headlines.
The central question stems not from all the suspected factors fluctuating together to reach dangerous peaks on a regular basis – like the days when the biorhythm app on your smart phone warns you not to get out of bed – but the constant pace at which these peaks have occurred, the equivalent of a metronome in the Earth System. There are two broad possibilities: one due to internal causes; the other due to astronomical forcings from the outside. To explore these fundamentally contrasted hypotheses, we start with the last of the ‘Big Five’ mass extinctions that brought the Mesozoic Era to an end and which caught popular imagination as the Nemesis of dinosaurs.
The K-Pg boundary event
Reconstruction of the lunar and terrestrial bombardment records by analysing craters during the Apollo programme revitalised interest in catastrophes (Chapter 10). Huge energy deliveries during post-Archaean times, as well as in the pre-4 Ga period, affected both Earth and Moon. A pointer to their possible effects on life systems came from geochemical studies of rocks that straddle extinctions. In 1980 Walter Alvarez and his co-workers reported that material at the Cretaceous-Palaeogene (K-Pg; K-T for Cretaceous-Tertiary is still widely used) boundary in ocean-floor sediments and on continents contain unusually high concentrations of noble metals, particularly that of iridium (Ir), which is the easiest to analyse. Noble metals have an affinity for metallic iron, and on Earth are almost certainly concentrated in its core. Meteorites are rich in those metals. The K-Pg Ir anomaly seems to be the chemical equivalent of a ‘smoking gun’ that implicates impact by an asteroid. But that is not all. The boundary layer includes droplets of once-molten rock, high-pressure and shocked versions of common minerals and even tiny diamonds. It also contains sooty residues rich in complex hydrocarbons. The event may have involved global fires (Chapter 8), though some of the hydrocarbon residues have characters that link best to what cosmologists consider to be molecules formed inorganically in interplanetary space.
The K-Pg boundary event has such a weight of evidence in support of the involvement of an impact that it came as no surprise when a candidate was found. Exploration for oil in the Gulf of Mexico off the Yucatan Peninsula displayed a near-perfect circular structure on a map of gravity anomaly. Seismic surveys also revealed that the deeper sediments seemed disturbed in a circular pattern. Reappraisal showed conclusively that a crater lay in the area, now called the Chicxulub (pronounced chick-shulub) structure, about 180 km in diameter and due to impact by a body between 10 and 20 km across. At the time of writing (autumn 2016) drilling into the centre of the circular feature was beginning to yield pertinent information.
That the Earth has been hit by large projectiles is undeniable; we can see their scars. Not many, but on the ground and on aerial and satellite pictures they are unmistakable (Figures 10.1, 10.2). Although the Earth must have been hit by comparable objects to those that so obviously scarred the lunar surface, and they must have arrived with a similar frequency (Chapter 10), they are greatly under-represented on the Earth. There are four main reasons why this is so. Weathering and erosion efficiently mute and remove evidence for all but the youngest and the largest impact structures. About 70 percent of the Earth’s surface is ocean floor, hidden from direct view and subject to more or less continuous sedimentation. Within about 200 Ma all of it returns to the mantle. Mountain building deforms much of the continental crust to induce rapid erosion as well as obliterating the distinctive shapes of impact scars. Volcanic activity and sedimentation repave large tracts of the continents at different times, thereby hiding any preserved scars from earlier periods. Far more than the fossil record, that of Earth’s encounters with extraterrestrial objects is scanty; just 188 confirmed craters. Figure 19.2 shows the timing of the 14 post-Permian craters with diameters >30 km.
Evaluating the likelihood of many animal species having met a swift and terrible end through massive energy delivery from space in the past relies as much on the approach of an insurance company’s actuary as it does on geological evidence (Chapter 10). Assessing the probabilities of past terrestrial impact events relies on evidence from the lunar surface. The Moon’s record of bombardment is little changed by the wear and tear of time. Moreover, it has a very simple geological history, and one event, the formation of the maria between 3 800 and 3 200 Ma, formed a pristine surface on which all later impacts left their mark. This allows the size-frequency distribution of lunar impact craters to be charted for the period represented by geological time on the Earth (Figure 10.5a). For reasons that I shall not burden you with, we can divide the lunar record into two long, post-maria periods, from 3200 to 1100 Ma and from 1100 Ma to the present day. In the first there are 87 lunar craters with diameters greater than 30 km and 45 in the second. On average, one formed every 24 Ma during both time periods. Because it is bigger and has a larger gravitational field, the Earth always collects more debris than the Moon. About 2500 impact events on Earth may have produced craters larger than 30 km during geological time; one every 1.3 Ma. Craters larger than 300 km have an expected frequency of one every 135 Ma, and the largest single crater likely to have affected life could have had a diameter of about 1300 km.
Looking at the power of such events (Figure 10.6b), for it is the rate at which energy is delivered that disturbs day-to-day processes. Impacts yielding the equivalent of more than a year’s sunshine in a second are predicted to have occurred about every 85 million years over the last 3 800 Ma. Powers amounting to instantaneous delivery of the entire nuclear arsenal should have been much more common. A grand total of about ten thousand suggests an average ‘waiting time’ of about 400 thousand years – a possible ten such impacts while humanity evolved.
How do powers associated with impacts compare with those involved in the normal course of geological processes? The energy available annually from the Earth’s internal heat production is equivalent to around 105 megatonnes of TNT, of which only one hundred thousandth is released by earthquakes. Most emerges quietly as heat conducted through the crust or by volcanism on the sea floor. Solar energy flux is very much bigger, but it is distributed over the entire surface and among a large number of ways of doing work. Any single Earth-generated catastrophe has tight limits on the energy involved. Assuming for the moment that it was impacts that triggered mass extinctions, how to judge the size involved in making a dent in the biosphere that we can pick up in the fossil record? Building up a ‘kill curve’ for impacts is quite simple.
Life has been waiting 3 800 Ma for Armageddon and the geological record of life together with genetic relatedness between all living organisms is good evidence that it hasn’t happened. Chances are that some terrestrial life survived one impact that made a hole 1300 km across. That represents the minimum power that might extinguish all life. Geophysical exploration for oil in the North Atlantic off Nova Scotia revealed anomalies with a circular structure, about 45 km wide. Drilling showed it to contain evidence for a major impact around 50 Ma ago. It also provided a complete fossil record bracketing the time of the impact, in which palaeontologists found no sign of significant extinction among marine organisms. The culprit seems to have been a cometary nucleus about 3 km across. This is the largest dated impact for which no tangible effect on life as a whole can be assumed. Its power is the largest from which life can recover unscathed in terms of species diversity. The Chicxulub crater, 180 km across is implicated in the demise of 40% of all animal genera at the K-Pg boundary. That gives three points from which you can easily plot a ‘kill curve’ for impacts, but one with a great deal of leeway in its shape. Chapter 10 described in some detail the various phenomena that large projectiles probably set in motion as they pass through the atmosphere and strike rock or sea.
For the K-Pg boundary event masses of evidence indicate that it coincided with an asteroid strike, yet, more than 3 decades after Luis Alvarez discovered the first evidence with that connotation, many experts still doubt such a Nemesis or would like to see it refuted. They do not deny an impact, but suggest that the connection was a coincidence. They focus their doubts on several features of the fossil record: two groups of molluscs disappeared more than a million years before the boundary, microfossil extinctions began about 100 ka earlier, and the famous dinosaurs appear not to have bitten the dust suddenly. Fossil evidence appeals to processes that lasted for much longer than the few seconds of an asteroid impact.
The Cretaceous closed with a protracted catastrophe of another kind. During about a million years that spans the K-Pg boundary hot-spot magmatism beneath what is now north-west India poured out 1.5 million cubic kilometres of basaltic lavas to build the Deccan Traps. Scaling up the effects of observed basalt outpourings, volcanologists estimate that over this period 10 trillion tons of SO2 and about the same amount of CO2 belched into the atmosphere. While the greenhouse effect of the CO2 was in operation, the acid rain from sulfur gases would have posed a tremendous stress on most life forms, particularly in the upper oceans. Such volcanism also provides another plausible terrestrial explanation for the iridium anomaly itself. Formed from a superplume rising from the core-mantle boundary, the magmas may have carried noble metals from the core, where they are thought to be much more abundant than in the mantle itself. Recent finds of economic platinum and gold deposits associated with flood basalts lends some support to this notion.
An internal model for the K-Pg event and other mass extinctions involves pulsation in the heat-loss mechanisms within the Earth that probably underlies the broad periodicity of many large-scale phenomena. We can envisage many things linking to jumps in the rate and position at which internal heat is released – continental break-up, flood volcanism, subduction and mountain building, changing sea level, climatic change due to release and absorption of CO2, and changes in the bulk composition of the oceans. Several possibilities for a mechanism present themselves. Maybe it has something to do with changes in the dynamics linking the behaviour of the core to that of the mantle. Changes in the Earth’s geomagnetic field show a period around 30 Ma, and this field is generated by electrical and circulation processes in the fluid outer core. Rising plumes in the mantle, responsible for the hot spots that trigger flood volcanism and sometimes continental break up, seem to rise from the core-mantle boundary. The Earth’s interior transfers heat by convection, whose overturn may have several modes that ‘flip’ from one to the other. Mathematicians studying convection through computer modelling have hinted at chaotic behaviour that could show periodicity. The processes whereby sea floor spreads and oceanic lithosphere subducts may have boundary conditions reached in more or less fixed times. For instance, subducting lithosphere may ‘bottom out’ at the base of the asthenosphere, so ‘clogging-up’ the system, or lumps may fall off the descending slab due to density changes. Plates themselves may have a limiting size, beyond which they become unstable.
There are many factors that conspire to blur the debates surrounding the K-Pg boundary event. For a start, different times present different numbers of localities to the scientists who collect fossils. Late-Cretaceous times are short on continuous sequences for some of the bigger animals. They were far better for tiny, rapidly evolving creatures crammed into deep-ocean cores, but their miniscule size has its negative side, as too the thinness of the representative layers. Any later animal grubbing in the mud might transport them to higher, younger layers. Re-enter the statistician. The blur on all the fossil records permits both catastrophic and gradualistic models; oh dear! There is a way to resolve this impasse.
Ken Hsu, formerly of the Swiss Federal Institute of Technology, together with others, picked out different types of planktonic shell from ocean cores that straddle the K-Pg boundary; those of deep water forms and those which lived in the upper ocean. They measured the carbon isotopes in the shell carbonate that the organisms extracted from the water in which they lived. Today, shallow water is rich in 13C because the lighter isotope (12C) is taken up during photosynthesis. The position is reversed in deep water, where dead organic matter enters solution due to decay. There is a normal gradient in carbon-isotope composition from surface to ocean floor, and this shows up in the vast majority of deep-ocean sediment layers. At the K-Pg boundary the gradient switches dramatically, its reverse persisting in some sections for around 500 ka. The switch took less than a thousand years. Hsu and his colleagues saw exactly the same switch in data from modern Alpine lakes, where it forms with the annual winter die-off of minute photosynthetic plants. From the global switch in the oceans at the K-Pg boundary they concluded a catastrophic collapse of this primary biological activity on which all other life depends. To Hsu, the seas then were almost lifeless. He calls such sterile conditions Strangelove oceans after the eponymous, mad advocate of mutually-assured destruction by nuclear holocaust. The method is almost independent of the fossil record, relying only on separating material according to the depth at which it lived and an assumption of continuous deposition on the ocean floor. It is powerful support for a catastrophe, however it was induced, and that it took less than a thousandth of the time during which the Deccan Traps poured out. The Chicxulub impact coincided with ongoing Deccan volcanism; a troublesome time.
Some participants in K-Pg debates, and a great many onlookers, preferred to grind neither axe, favouring inhospitable conditions from the combined effects of both volcanism and giant impact. Chicxulub was perhaps the last straw after several hundred thousand years of Deccan basalt floods. But more detailed and accurate work continues to be done on what remains about the most popular single topic in the Earth sciences. However, the most recent study on the K-Pg by Paul Renne of the University of California-Berkeley deepens, and perhaps muddies, the waters further. It centres on the most accurate dating yet of Deccan basalts, which produced a previously unachievable precision of ±0.1 Ma for times between 65.5 to 66.5 Ma, which nicely bracket the K/Pg boundary age of 66.04±0.04 Ma. Moreover, the study by Renne and colleagues involves a large number of basalt samples. Combined with estimates of the volumes of basalt erupted from flow to flow, they discovered that the rate of magma effusion doubled immediately after the time of the K-Pg event. Does that imply that the Chicxulub impact somehow affected the magma production from the mantle plume beneath the Deccan? It had been suggested early in the debate – and widely rejected – that the roughly antipodean position of the lava field relative to that of Chicxulub may indicate that the huge seismicity from the impact triggered Deccan magma production. Renne and colleagues think that idea deserves another look, at least at the possibility of some linked effect of impacts on magmatism. Perhaps the magma chamber was somehow enlarged by increased global seismicity; other chambers could have been added; magma might have been ‘pumped’ out more efficiently, or a combination of such effects. Chicxulub’s collision was equivalent to an earthquake of Magnitude 12.4, a thousand times more powerful than the largest recorded earthquake with tectonic causes. Even a small proportion of that power transmitted through the Earth and its core could have influenced processes at the antipode.
That may seem a weird concept, and indeed it is hard to grasp because of its magnitude. However, the imagination of geoscientists sometimes beggars belief. In 2004 some at the Geomar Institute of the University of Kiel in Germany, led by Jason Phipps Morgan (now at Royal Holloway University of London), stoked up the controversy by taking a very different view of events. Magma from a mantle plume has to penetrate the overlying lithosphere. Where this is old and thick, as beneath ancient continental cratons such as the Indian subcontinent, it may act like the lid on a pressure cooker. During development of flood volcanism gas pressure might build up to levels sufficient to breach explosively to the surface, when monstrous crustal collapse above the plume head might propagate upwards at hypersonic speeds firing huge masses of lithosphere across the Earth. On re-entry the largest blocks would create craters; the power of the process would create melts, shock the ejecta and create fullerenes and microdiamonds from carbonaceous rock; having begun ascent at the core-mantle boundary the plume itself could release the noble metals that first attracted the Alvarez’s attention. The authors dubbed such hypothetical events ‘Verneshots’ after the founder of science fiction, Jules Verne. In his 1865 book De la terre a’ la lune (From the Earth to the Moon) Impey Barbicane, president of the Baltimore Gun Club, creates a huge cannon that fires a capsule containing three adventurers to the Moon. Interestingly, seismic surveys of the Deccan Traps and other flood basalt provinces in Siberia and Yemen show circular fracture systems in the crust beneath, similar to those formed when large volcanoes collapse after large eruptions, but bigger.
The largest documented mass extinction that brought an end to the Permian Period and Palaeozoic Era, although awesome, presents a simpler picture; despite a global search for a contemporary crater, it has no association with evidence of asteroid or comet impact..
Almost Armageddon – the end-Permian event
The K-Pg mass extinction was but a sideshow by comparison with the devastation that ended the Palaeozoic Era 251 Ma ago. At the time of the Permo-Triassic (P-Tr) boundary life in the sea was all but snuffed out. As many as 95% of all species of marine organisms met their end. So too did many continental vertebrate groups, both amphibian and reptilian; insects were decimated (for the only time since they first appeared); most of the plants that formed the Carboniferous and Permian coal swamps vanished from the record – there are coals from the late Permian but not the early Triassic.
Fruitful research at the K-Pg boundary fuelled analysis of the 190 Ma older event. Despite the search, few chemical pointers to either asteroid or comet impact have shown themselves. Although a large comet would deliver no concentrated noble metals, it would certainly excavate masses of shocked rock and rain droplets of melt across the planet. Shocked grains have been found at the boundary only in Antarctica and SE Australia. But there is a flood-basalt province that brackets the P-Tr boundary. The Siberian Traps represent the largest extrusion of magma in the Phanerozoic, about 2 million cubic kilometres.
The key information for climate change during the Mesozoic Era lies in the muddy sediments of the ocean floors. But, unfortunately, there is no sign of Triassic, let along Permian sediments on modern ocean floors: all have been subducted. However, ocean sediments are sometimes incorporated onto continents – they accompany sheets of oceanic lithosphere thrust upwards instead of down at destructive plate margins. Japan has such an ophiolite with a 50 Ma long sequence of deep-water sediments that spans the critical time across the Permian-Triassic boundary.
The Japanese Permian to Triassic rocks are dominantly mudrocks, now flint-like through lithification. They accumulated from a rain of tiny plankton that secrete shells of silica. They are also bright red, and that is interesting. The colour is due to ferric iron, and indicates that for most of the time deep ocean water was thoroughly oxygenated. At the Permian-Triassic boundary the sediments are very different. Spanning a 20 Ma period they are not red at all, but grey to black, indicating too little oxygen for the conversion of ferrous to ferric iron. This colour change starts at 260 Ma and reverts back to brick red at 240 Ma. Straddling the P-Tr boundary are clays with low fossil contents, and exactly at the boundary is a jet black, hydrocarbon-rich mud that represents perhaps 4 Ma of deposition. Deep ocean waters were devoid of oxygen. Such anoxic hydrocarbon-rich muds occur in all marine sequences of this age, and formed important sources for petroleum reserves when their organic content was ‘cracked’ by later burial and heating. Carbon isotope studies from one of these oil-prone units in Canada shows the sudden gross shift in the proportions of light and heavy carbon so characteristic of Hsu’s Strangelove oceans. At the Permian-Triassic boundary, almost everything living in the open ocean died from top to bottom. The intricate Japanese section shows that its decline started long before, in fact 10 Ma earlier, and this coincides with the dwindling of shallow-water biodiversity elsewhere. Yet the final extinction was sudden, its carbon-isotope signature occupying less than 10 cm in the Canadian section. Assuming a sedimentation rate there of between 40 and 80 m per Ma gives a time of the order of 1000 years; a catastrophe. But rather than coming unannounced, life-threatening conditions had been brewing for up to 10 Ma.
Modelling the effects of ocean anoxia on the oxygen content of the late-Permian atmosphere suggests a fall from around 30 to 16 %. From conditions when it would have as easy to breathe at 6000 metres as it is at sea level today, if fell at Permian sea level to the modern equivalent of an altitude of 2700 m, at which the first symptoms of altitude sickness begin. All land animals that had evolved in the Carboniferous to Permian world had done so when oxygen was at a high concentration. Surviving the late Permian global anoxia would have been possible only in a much reduced, low-altitude land area. Several lines of evidence point to growing CO2 levels in the Permian atmosphere, modelling of which suggests a late-Permian world with mean global surface temperature 8° C higher than now.
Many threats related to hypoxia and anoxia, and to extreme global warming were around during the late Permian. The anoxic ocean-bottom waters indicate that the ocean surrounding Pangaea ceased to circulate and transfer oxygen downwards for 20 Ma; high temperatures had increased overall salinity. Asphyxia is a candidate for long-term killing of marine life but another outcome of anoxic oceans is a switch in the biological mechanism of decay that has a directly toxic product. On the beds of modern anoxic seas, such as the Black Sea, much decay is carried out by bacteria that exploit the energy potential of sulfate to sulfide reduction, the waste product of which is hydrogen sulfide (H2S) gas. Disastrous results are warded off in modern anoxic seas by sulfide-oxidising bacteria in surface waters. But production of H2S on oceanic scales when surface waters were warm and less oxygen-absorbing, at a time of falling atmospheric oxygen is a recipe for disaster from the build-up of acid surface waters and toxic gas escapes to the atmosphere.
While chemical stresses were building up sea level fell to an all-time Phanerozoic low. Since continental ice sheets had long disappeared from Pangaea, the only likely agency is a decrease in sea-floor spreading so that cooling lithosphere sagged to increase ocean-basin volume by the late Permian. That implies a long-term slowing of heat loss from the mantle. But radioactivity continually produces heat. Conceivably, it built up deep down, until released by the superplume responsible for the 251 Ma old Siberian Traps. The volcanism coincided exactly with the final, sudden peak of extinctions after all the other longer-term negative changes for life during the Permian. Not only were the Siberian Traps by far the biggest effusion of basalt lava known from the geological record, with commensurate potential for CO2 and SO2 release, they were emplaced through a Carboniferous sedimentary basin rich in coal measures and oil shales, adding to their potential to drive up climatic warming. The older sedimentary basin is pocked by chimney-like structures signifying violent gas release through partly consolidated sediments. Astonishingly, electron microscopy has shown abundant fly ash in the silica-rich deep-marine cherts in Japan that record the carbon isotope signature of the P-Tr mass extinction; a sign of ignition and burning of coal.
In the absence of any plausible evidence for a major impact, and the wealth of data that indicate slowly deteriorating conditions for life and the coincidence of the Siberian Traps with the final throes of the Palaeozoic Fauna shows that the Earth is quite capable of almost sterilizing itself, with no outside help. Examination of Figure 19.2 shows that of the eight significant Mesozoic and Cenozoic extinctions only that at the end of the Cretaceous is associated with an impact crater wider than 30 km. On the other hand, as well as the P-Tr extinction, six coincide with flood basalt events. Whatever the other possible ‘drivers’ for downturns in diversity, major processes in the deep mantle ‘have form’ as police officers are prone to say. The end-Triassic member of the Big Five is associated with the start of rifting of South America from Africa and formation of the South Atlantic Ocean, and huge flood basalt piles of that age occur on both sides of the ocean. The mass extinction towards the end of the Devonian Period (Figure 19.1) is devoid of plausible evidence for an impact connection, but neither do truly massive flood basalts figure apart from a 320 thousand km2 area of basalts flows, again in Siberia (the end-Permian Siberian Traps outcrop over 7 million km2). The end-Devonian extinction has been attributed to the climatic influence of the colonisation of the continents by large, coal-forming plants. However, it is becoming apparent that the downturn in organisms involved as many as 10 small extinctions spread over ~20 Ma. If vegetation of the land surface was the driving process it is ironic that most of the genera that ceased to be were marine.
The first mass extinction recorded directly by the disappearance of fossil organisms took place around the Ordovician-Silurian boundary, and it has some differences from the other big ones. The first event (443 to 445 Ma) of a ‘double-whammy’ mainly affected free-swimming and planktonic organisms and those of shallow seas; near-surface dwellers such as graptolites and trilobites. The second, about a million years later in the earliest Silurian rocks, hit animals living at all depths in the oceans. Between them, the two events removed about 85% of marine species, close to the extinction level at the Permian-Triassic boundary. Neither impact nor flood basalts are implicated. The double extinctions followed major glaciation on Gondwana when the supercontinent lay over the South Pole, resulting in an 80 to 100 m fall in global sea level (Figure 5.5). Glacially-induced sea-level fall would have reduced the extent of shallow seas. Along with global cooling, that could explain the early demise of shallow water, free swimming animals. Yet thermohaline circulation by sinking of dense surface water to abyssal depths, would have provided oxygen to maintain and even increase life deep in the water column. In the absence of any other obvious mechanism, many palaeontologists look to a return to stagnancy and anoxia as glaciation waned, increasing the production of H2S gas by sulfate-sulfide reducing bacteria to result in the second phase of more universal extinction. Neither mechanism has the awesomeness of impacts and basalt floods, and perhaps in those times conditions were on a fine balance between favouring the metazoan animals of the seas and making life insupportable; an easily upset world. More in frustration than strict adherence to William of Ockham’s law of parsimony (Ockham’s razor: ‘Among competing hypotheses, the one with the fewest assumptions should be selected’) a small minority of scientists have appealed to a ‘hypernova’ close by in the Milky Way galaxy, which would have exposed the Earth to a gamma-ray burst. If that was the case, high-frequency electromagnetic radiation would have left no trace; the ‘perfect extinction’?
Mass extinctions transformed the course of evolution, by leaving few types of organism alive and able to reproduce on land and in the sea. Survivors formed the ancestral stock from which animals and plants of later seas and continents evolved. The Japanese and Canadian sedimentary sequences at the P-Tr boundary show that the oceans at the start of the Mesozoic Era remained almost sterile for maybe a half million years. Nature abhors a vacuum of any sort, and all the vacant niches awaited occupants and competition among them. Evolutionary radiation following shocks to the biosphere forms the theme of the next chapter. Bear in mind that evolution following hard times involves adaptation of those organisms that survived. The traits that fortuitously helped their survival may play a role in later developments.
You may well have lain awake plagued by a nasty cough (or by someone else’s coughing). We cannot help it; coughing is a reflex. Oddly, the cough impulse (and also swearing, biting and the kind of rage directed at the cause of sleeplessness) has a direct link with a part of the innermost brain crudely referred to as the reptilian complex. It has that name because its structure resembles that of the entire brain of reptiles. Neuroscientists consider the enclosing structures of mammalian brains to be later additions. It now seems clear that almost all living land vertebrates descend from very few primitive reptile species that survived the P-Tr mass extinction. During long sleepless nights comfort yourself with the thought that this blessed coughing might be the only reason that your far distant ancestor came through the Permian wasteland made acrid by basalt floods in Siberia.
If you enjoyed reading this chapter online and want to learn more, Stepping Stones is now available to download as an eBook