THIS is, in part, the story of a glorious failure. Vincent Courtillot and his
colleagues set out to study the volcanic rock formations of the Deccan Traps,
the volcanic deposits that cover rather a lot of western and central India. They
originally aimed to clarify what actually happened in the collision between
India and Asia, which took place over the 50 million years during which the
Traps were thought to have erupted. Courtillot found, however, that the bulk of
the volcanism took place within less than half a million years.

Something much more interesting came out of the project. The Deccan eruptions
shed light on the best-known catastrophe in the story of life on Earth: the mass
extincition at the end of the Cretaceous. Courtillot tells the tale in
Evolutionary Catastrophes, subtitled “The Science of Mass Extinctions”.
Over the past 300 million years our planet has been battered by at least seven
major ecological catastrophes.

That most fundamental of geological concepts, the stratigraphic column, has
recorded the pulse of these extinctions. When the geological pioneers of the
19th century divided up rock sequences, they chose units that could be
recognised easily because they contained distinct types of fossils. It is no
surprise, then, that the boundaries between the different periods, subsequently
turned out to mark the mass extinctions.

The most famous of these occurred at the boundary between the Cretaceous and
Tertiary, some 65 million years ago—known as the K/T boundary (“K” from
the German for “Cretaceous”). A large part of its fame stems from human
self-interest: the K/T extinction probably facilitated the evolution of mammals,
including us, by ending the domiantion of the dinosaurs.

So what dunnit? Were the extinctions due to the sudden impact of an
extraterrestrial body, or the result of normal, ongoing developments in the
evolution of Earth—global changes in sea level or outbursts of volcanic
activity? This debate between catastrophists and uniformitarians is as old as
geology, and has been re-invigorated by the discovery of a massive impact crater
beneath the Caribbean, off Chicxulub, Mexico. There are telltale remnants of
that impact around the globe at the time of the K/T extinction. But the Deccan
lava flows were busy covering well over 1 million square kilometres at just the
same time, geologically speaking.

Courtillot gives a reasonably balanced discussion of the strengths and
weaknesses of the impact and the volcanic theories. As he notes, his own
prejudices as a vulcanologist come through.

Much of the debate centres on the speed at which things happen. New dating
methods give ever shorter durations of the major volcanic episodes that
correlate with mass extinctions. The injections of huge masses of dust and
aerosols into the stratosphere would lead to climatic disruption and the
destruction of food chains on a global scale. These effects were extensively
modelled in the 1980s, not least by researchers investigating the possibility of
a “nuclear winter” triggered by war. They would occur with either volcanic
activity or an asteroidal impact.

Courtillot’s work on the Deccan Traps found hundreds of huge but relatively
brief lava flows—which would have caused enormous climate disruptions
every few hundred years or so. Imagine a dinosaur species and the plants on
which it depended surviving one such event. The next catastrophe, or the one
after, would likely wipe it out. In this view, the multiple hammering from
repeated massive volcanic eruptions might be more effective agents of global
extinctions than a single event.

Perhaps it is bad luck that the best studied mass extinction suffered the
double-whammy of the Chicxulub impact right in the middle of the Deccan
volcanism. No clear evidence of impacts has been found at the time of any other
extinction. So Courtillot scores it: 1 for impacts, 7 for volcanism.

The game isn’t over: watch this space and read this book. Read it, too, to
any bureaucrats and research council members who love “Foresight” exercises and
directed research with tightly defined “deliverables”. The unexpected result is
often the most exciting and scientifically rewarding.

THE summer of 1783 was miserable in northern Europe: cold with interminable
foggy days. Benjamin Franklin, then US ambassador in Paris, sent a three-page
paper to the Manchester Literary and Philosophical Society that set an agenda
still raging today. He wrote that “the cause of this universal fog is not yet
ascertained. Whether it was [due to…] consumption by fire of one of those
great burning balls or globes which we happen to meet with in our rapid course
round the sun or whether it was the vast quantity of smoke, long continuing to
issue during the summer from Hecla in Iceland . . . is yet unknown”.

In other words, he speculated that either a meteorite impact or volcanic
eruptions had caused the weather to deteriorate. In this case, it was the
latter, and the consequences were far worse than miserable summer weather: 25
per cent of Iceland’s population died of starvation as the noxious gases from
the eruption disrupted the food chain.

The Great Dinosaur Extinction Controversy addresses an identical
problem: was the extinction of the dinosaurs and other species about 65 million
years ago caused by massive volcanic eruptions or by meteorite impacts? There is
evidence for both.

But do not expect a balanced argument from this book: it is firmly on the
side of the volcanic hypothesis. It sets out to discredit the impact
hypothesis—describing it as “pathological science arising from
self-delusion”. Strong stuff indeed. The book has more in the same vein. For
instance, the authors found a quotation calling the impact hypothesis
“codswallop”. They manage to mention this three times. Repetition alone hardly
strengthens their argument.

They argue that the dinosaur extinctions resulted from dust and gases emitted
from huge volcanic eruptions which built the Deccan Traps of India: these flows
extend over a million square kilometres, and were erupted over a half a million
years just at the time the dinosaurs disappeared.

But there is also clear evidence at exactly the same time of a massive
impact. Interestingly, the earliest evidence came when Luis Alvarez, a
Nobel-prizewinning physicist and his geologist son, Walter, detected enormously
high concentrations of iridium in a thin clay layer exactly at the
Cretaceous-Tertiary boundary. Iridium is an obscure element related to platinum,
and meteorites were the only major source of it known at the time of this
discovery. Thus was born the impact hypothesis. Subsequently, abnormally high
iridium concentrations have been reported from many other rocks of identical age
around the globe.

Further investigations revealed that the same rock layers that exhibit high
iridium contents also contain minute quartz grains that, under the microscope,
reveal characteristic internal features that can only have been caused by
deformation under extremely high pressures. This is seen as more evidence of
massive impact.

If the volcanic scenario is correct, the iridium and shocked quartz evidence
has to be explained away. This book makes a spirited attempt to do so, claiming
that giant volcanic eruptions cause both features. Perhaps the iridium derives
from deep in the Earth’s interior, being released by large volcanoes such as
those in Hawaii. But even if this is true, there remains the tricky problem that
the really high concentrations of iridium—as much as ten thousand times
normal background levels—occur over extremely thin layers of sediment just
a few centimetres thick: just what you would expect from an impact. Although
there is some evidence of lower concentrations of enhanced iridium,
corresponding to intervals of perhaps 100 000 years or more, the short-lived but
massive spike still has to be explained.

The best explanation of these patterns, it seems to me, is that there was an
extended period of volcanism, such as the Deccan eruptions, with an impact event
superimposed.

The shocked quartz evidence is harder still to explain away as the effect of
volcanism: some similar features can be created by volcanoes, but the pressures
are insufficient to reproduce some types of deformation seen. Charles Officer
and Jake Page’s best attempt is to say that some of the shocked quartz is due to
impacts and some due to volcanism. But even they conclude that it is “a
relatively murky picture”.

What about a smoking gun: has anyone found the impact crater? For a long time
the answer was no, and this book recites many of the claims that were eventually
discredited by accurate dating. But recently a good candidate has been found. It
is the 160-kilometre-wide subsurface Chicxulub feature off Mexico, which is
exactly the right age.

So much for the scientific debate. Officer and Page spend the rest of the
book discussing the way that the media and other scientists responded to the
rise in popularity of the impact hypothesis. It is undoubtedly a fascinating
topic, but is inevitably reported here in a rather one-sided way. I am sure we
will hear more on this subject from historians of science, and Officer and
Page’s traditionalist views will undoubtedly form part of the evidence.

One other element in their argument caught my attention, that of the risk
associated with possible impacts. Officer and Page cite an estimation of the
chance of a large meteorite hitting a US city in the next ten years as one in
ten million. Not something to lose much sleep over, they report approvingly. Yet
the chance of winning the British lottery is about one in fourteen million, and
millions spend their money on such probabilities every week.

Officer and Page correctly remind us that there have been no known human
fatalities from impacts in recent times. One tale they do not mention is worth
reciting. In 1911 a 40-kilogram meteorite killed a dog at Nakhla in Egypt. That
was bad luck, certainly. But perhaps even more unlikely was the source of the
meteorite: it was an extremely rare type thought to be a lump of Mars knocked
off by an earlier impact, then sent spinning on its way to Earth. That seems to
be double bad luck. But it happened.

The book has a good set of references to original literature. It is generally
accurate in its scientific statements, although I was disappointed by the
authors’ incorrect understanding of mantle plumes, the very cause of the massive
volcanism they espouse so strongly. Mantle plumes do not consist of molten rock
rising from the boundary with the liquid core of the Earth, as they assert
several times. The mantle only melts as it reaches shallow depths just beneath
the Earth’s outer skin. But that aside, the facts are stated clearly and it is
easy to go back to their sources and make up your mind about their
interpretations.

I started off with warm feelings towards the volcanic hypothesis. But there
is increasing evidence for one or more impacts at the Cretaceous-Tertiary
boundary. I rather think that Officer and Page have overstated their case.

Should I end by speculating that this book is destined to go the way of its
subjects, the dinosaurs? Officer and Page are certainly putting their heads on
the block by publishing such an anti-impact book at the very time when the
Chicxulub studies seem likely to strengthen the hand of the impact thesis. But I
do not want to be unfair on the authors. They have produced a very readable book
which stimulated my interest, not only about dinosaur extinction, but also about
risk analysis and the way science progresses.

So I will end instead by reiterating Franklin’s uncertainty over the cause of
the dreadful summer in 1783: we still don’t know whether the dinosaurs were done
in by volcanic eruptions or by enormous impacts. I suspect both played their
part. It looks as if the Earth was hit 65 million years ago by a double whammy
of massive volcanism in India and a huge impact in Mexico.

This book left me wondering. The dinosaurs lasted 150 million years. Modern
humans have been around for less than 150 thousand years, yet are showing
vigorous signs of being able to wipe themselves out—without waiting for
either volcanic eruptions or extraterrestrial impacts.

When continents collide, they make mountains; when they divide, a new ocean forms. Geologists can study the history of collisions, because erosion gradually reveals the structure of the rocks inside mountain ranges. But when a continent tears into two, the broken edges vanish beneath the waves. Now geophysicists have found a way to examine the remnants of these continental edges or margins. Seismic surveys of the rocks beneath the sea floor show that the continental break-up which generated the North Atlantic Ocean was accompanied by extensive volcanism – but only in certain parts. From these surveys, geophysicists have begun to piece together the history of the splitting of one continent which took place tens of millions of years ago. Last week, an international team of geologists, working on board the research ship Joides Resolution, started to collect samples of rocks from a continental margin off the coast of Portugal. These samples, geophysicists hope, will help them to confirm their theories, and to understand why the continents split apart in the first place.

When a continent breaks, the process is known as rifting. It is happening today in the Red Sea, where the Arabian Peninsula is moving slowly away from Africa. The continent stretched and thinned until the continental crust finally broke into a series of faults and then subsided. New ocean crust is now forming under the Red Sea as molten rock wells up from below to fill the gap. Millions of years ago, Greenland and North America were near neighbours of Britain. But here, too, the thick continental crust stretched like toffee until it finally broke, allowing the North Atlantic to form. A few years ago, geophysicists from the Universities of Cambridge and Durham put to sea on the Natural Environment Research Council’s research ships, the Charles Darwin and the Discovery, to carry out seismic surveys of Hatton Bank, a continental margin formed when Greenland broke away from northwestern Europe 55 million years ago.

Seismic profiling has become a powerful tool for earth scientists . Seismic waves generated underwater by explosive charges or bursts of compressed air travel into the sea floor where they reflect and refract as they pass through rocks with differing physical properties. Researchers can deduce the structure of the rocks below from the signals that return to the surface. Seismic waves have low frequencies (between 5 and 50 hertz) and propagate well through rock, giving researchers a chance to ‘see’ a long way into the Earth. For the Hatton Bank project, researchers used seismometers placed on the sea floor, a refinement of the technique enabling them to pick up signals that had travelled up to 100 kilometres through the Earth’s crust.

Computer simulations of the seismic profiles at Hatton Bank led the team to a model of the structure of the continental edge that included far more volcanic rock than expected. Continental crust was stretched and thinned towards the new ocean, as expected, but the seismic reflections indicated that there were many thin sheets of rock near the surface, gently sloping out to sea. These have since been drilled and identified as lava flows, but at the time no one suspected that there would be so much igneous rock, solidified from lava, at the edge of the continent.

Startling lava flows

The association of continental break-up and volcanism is not a new observation. Signs of extensive volcanic activity have been found across western Britain and Ireland dating from the time when the North Atlantic opened – examples include the volcanoes and basaltic lavas of Skye and the Hebrides, and rock formations such as the Giant’s Causeway in Antrim. It was the scale of the lava flows that was startling: in places, the layers of rock were 5 kilometres thick, making up much of the thickness of the crust at the margin. Mapping the sea floor by seismic reflection methods shows that there are similar lava flows along the continental edge on both sides of the North Atlantic. They stretch from Edoras Bank to the northern tip of Norway on the east side of the Atlantic and along the entire Greenland coast to the west. When the Atlantic first opened, perhaps as much as 2 million cubic kilometres of lava poured out in only one or two million years – almost overnight in geological terms.

The seismic sections across Hatton Bank held another surprise: there is between three and four times as much igneous rock trapped near the base of the continental crust. This means that some 10 million cubic kilometres of igneous rock were generated when the North Atlantic first broke open. This is an enormous amount of rock: if spread across the continental landmass of the US, it would make a layer 700 metres thick.

In 1987, a joint French and British team working aboard the French research ship Le Suroit, recorded data across another part of the Atlantic margin, near the Goban Spur at the mouth of the English Channel. Ocean started to form here as North America broke away from Europe, about 120 million years ago and some 65 million years before the Hatton Bank margin. The Le Suroit results showed that there was very little igneous rock, either as lava flows near the surface or as intrusions lower down in the crust. The structure in the upper levels of the crust was different, too. Faults, and the tilting of blocks of rock associated with fault movements, played a much bigger part in the break-up of the continent than they did in the Hatton Bank section.

Why did these two parts of the continental margin behave so differently? The answer lies in the interior of the Earth. The crust is part of the lithosphere, the rigid outer layer of the Earth which is about 100 kilometres thick. Beneath the lithosphere is the mantle, which continues 3000 kilometres down to the core. Although mantle rocks are hot enough to flow – typically reaching 1350 °C at 100 kilometres down – the pressure is so high at these depths that they remain solid. But if the pressure drops for some reason, the mantle will begin to melt. This is what happens as the lithosphere stretches when continents break apart: the mantle beneath wells up to fill the gap, decompresses and partially melts. The melt is less dense than the mantle around it, so it trickles upward to the surface. Eventually it solidifies to make new ocean crust.

At the Goban Spur, the crust of the continent was stretched and grew thinner and thinner until the upper part finally fractured and broke into series of faulted blocks. The same stretching took place lower in the crust, but there the rocks were hotter and under higher pressure, so they deformed without breaking up into faults. The crust continued to stretch until it was only 5 kilometres thick, about one-sixth of its original thickness, before molten rock welled up to form new ocean crust.

The amount of molten rock generated when the mantle decompresses is highly dependent on the temperature of the mantle. If the mantle heats up by 140 °C, a rise of just 10 per cent, it produces three times the volume of molten rock. Normally the ductile part of the mantle is at a uniform temperature and only small volumes of molten rock form under rifts. But from time to time, enormous plumes of unusually hot mantle rise towards the surface. This is what happened beneath the region of the present North Atlantic that lies between Ireland and Greenland. Shortly before the continent split, a mantle plume had begun to form, giving rise to volcanism over an area 2000 kilometres across. This plume is still going strong; it is responsible for the active volcanoes of Iceland, having already built the whole island from volcanic rocks.

Why, though, did the bulk of the molten rock stay within the crust at Hatton Bank? Why did it not all erupt as lava? The answer is a matter of density. The molten rock is less dense than the mantle that melted to form it, so it rises. But magma is more dense than continental crust. Most of it is trapped beneath the edges of the rift, where it solidifies to form intrusions of igneous rock injected into the continental crust.

Hatton Bank represents one extreme type of split, called a hot rift; Goban Spur the other, known as a cold rift. The mantle plume beneath Hatton Bank generated large-scale volcanism and the igneous rocks thickened the crust. The rocks were warmer, so faulting was a less important mechanism in the overall thinning of the crust. At Goban Spur, there was no plume and little volcanism, so faulting was a significant mechanism in the stretching of the colder and more brittle rocks.

The temperature differences in the underlying mantle also affect how quickly the rifted continental margins sink to form the new sea. The continental crust is thicker than oceanic crust, so it ‘floats’ higher in the mantle. This means that the formation of a new ocean always involves subsidence; the stretched and thinned crust on the continental margin usually sinks 2-3 kilometres, while the even thinner ocean crust eventually sinks more than 5 kilometres. At the Goban Spur margin, for example, the thinned continental crust margin subsided by several kilometres as it stretched and continued to sink over the following 50 million years or so. But at Hatton Bank, the underlying plumes added so much new igneous rock to the crust that, initially, the continental margins were elevated above sea level. Hot rifts therefore tend to rise as the continent splits. They do subside later as the ocean widens, but never by as much as a cold rift.

The differences between hot and cold rifts have major implications for researchers investigating how continental margins form and develop. In the North Atlantic, the contrast between two areas, just 1000 kilometres apart on a continental margin 10 000 kilometres long, represents a challenge to the theorists. How did the Atlantic margin change as it grew northwards? Do these two types of margin represent very different types of rift, or are they just variations on a larger overall theme?

Geophysicists are still conducting experiments to try to shed light on these and other problems. On the Charles Darwin last summer, a team of scientists from the Universities of Cambridge and Edinburgh took seismic soundings at Edoras Bank to probe the layers of lava and the underlying crust. Edoras Bank, once joined to the southern tip of Greenland, was at the edge of the Iceland plume when the North Atlantic was opening, so the results should help them find out how the presence of the plume affected the development of the continental margin.

Last week, an international team of scientists, led by Bob Whitmarsh of the Institute of Oceanographic Sciences in Surrey and Dale Sawyer of Rice University, Houston, started to gather physical evidence of what happened at a cold rift. They are aboard Joides Resolution, a ship especially adapted for drilling boreholes and collecting samples of rock from beneath the sea floor as part of the international Ocean Drilling Program. Equipped with 9150 metres of drill pipes that can be suspended through a hole in the hull, the ship will spend eight weeks drilling into the thinned continental crust at the Iberian margin, an area off the west coast of Portugal that was once joined to Newfoundland in North America.

During October and November, the Joides Resolution will attempt to drill into a hot rift off the east coast of Greenland, an area that was joined to Hatton Bank before the North Atlantic was formed. Geophysicists hope that the samples from the two sets of drilling will provide the physical evidence on the subsidence history of the continental margins and composition of lava flows to confirm their theories based on the seismic soundings. In addition, analysis of lava samples should also tell them more about the mantle plume that was responsible for the formation of 10 million cubic kilometres of igneous rock that poured out when Greenland broke away from northwest Europe.

Combining data from these different investigations into the geology of rifted margins should soon help geologists to reconstruct the history of the break-up which generated the North Atlantic. It may even illuminate a different, and important aspect of the geology of Britain’s western seaboard: possible oil and gas deposits far out on the continental edge . Understanding how the continents rifted in the first place, and whether they were hot or cold rifts, will give oil geologists a head start when their search takes them to these remote and dangerous waters.

Bob White is professor of geophysics at the Bullard Laboratories, University of Cambridge.

* * *

1: Seismic waves strike deep

If you feel an earthquake, you are feeling seismic waves that have travelled through the Earth from the movement of a fault. Seismic waves are low frequency, penetrating waves that can travel enormous distances through rock. They reflect and refract at the boundaries between different types of rock, by amounts that depend on the contrast in the seismic velocity and density between the two rocks.

Geophysicists generate seismic waves at a particular spot using explosives or compressed air, and then record when and where they pick up the reverberations from within the Earth (‘Sound waves reflect Britain’s deep geology’, New ĐÓ°ÉÔ­´´, 4 February 1978).

There are two main seismic techniques for determining the structure deep in the Earth’s crust. The first is seismic reflection, used extensively in the search for oil and gas. In this technique, reflections are recorded from directly below the detectors – called geophones if used on land and hydrophones if the experiment takes place at sea. The simplest signal comes when a boundary between two rock layers reflects seismic waves particularly well. This results in a strong signal reappearing at the surface after a time equal to that taken for the wave to travel down to that reflector and back up again.

Experiments in which there are many sources of seismic energy and many geophones or hydrophones often show distinctive reflected signals over a wide area that can all be ascribed to the same reflector. Things become more complicated when there are many boundaries that act as reflectors. In this case, seismologists have to take into account waves that travel along ever more complicated paths on their return to the surface.

The second technique is seismic refraction, in which the source of the seismic waves and the receiver are moved increasingly farther apart as the experiment progresses. In consequence, the recorded seismic energy has travelled deeper and deeper into the Earth. The speed at which the seismic waves travel – called the seismic velocity – tends to increase deeper in the Earth, so the waves are eventually bent back to the surface.

To understand the paths followed by the seismic waves as they travel back to the surface, researchers devise models of the layers in the Earth and assign to each layer properties such as density and seismic velocity. Then they work out all the paths that the seismic waves would take through this model and compare the signals that reach the surface in the model – the synthetic seismogram – with those measured in the field. The model must be continually adjusted to produce synthetic seismograms that give a closer and closer match to the data.

Once geophysicists have found the changes in seismic velocity deep in the Earth from synthetic seismograms, they will know the distribution of different types of rocks. This method can be used to discover the thickness of the crust and the shape of its base, for example.

The technique is relatively new and would be impossible to use without the aid of powerful computer programs. Modern software can simulate the passage of seismic waves through complex structures, including those that change laterally – an important property for areas such as continental margins where the rock layers and faults vary enormously along a traverse from continent to ocean.

* * *

2: Oil, gas and the birth of the oceans

In common with other rifted margins such as the west coast of Africa, Britain’s western seaboard is an area where hydrocarbon deposits are likely to be found. The pattern of faulted and disrupted layers of rock created when a continent splits can produce good reservoirs for oil and gas. And the heating and subsidence of the continental margin that accompanies the birth of an ocean can also help to generate hydrocarbons. Whether oil and gas form and whether they migrate into these possible reservoirs depends on the geological history of the area, particularly how the continent rifted.

Petroleum geologists search for oil by looking for areas which have the right history of temperature and subsidence. Continental margins such as Britain’s western seas and the continental shelf off western Africa are among their targets. These areas combine likely sources of hydrocarbons with relatively rapid subsidence. The shallow seas that formed as rifting began were teeming with life, eventually producing thick sediments rich in organic matter. Speedy subsidence soon brought these sediments into conditions where the reactions that produce oil and gas could work.

The different temperature and subsidence histories of hot and cold rifts markedly affect these processes. Although the heat that results from the injection of igneous rocks beneath a hot rift may speed the reactions in the sediments, it may lead to the reactions continuing for too long. All the original organic matter may be driven off, resulting in an area with very little oil, or with gas only. A cold rift might preserve more hydrocarbons, but the sediment might never reach a temperature which promotes oil formation.

The history of subsidence and temperature now emerging from seismic exploration of these continental margins, and now from the drilling of the Joides Resolution, will help in the discovery of these increasingly valuable deposits. As exploration for oil moves farther from Britain’s shores, understanding exactly how continental margins formed will help to target costly exploration efforts in the most likely places.