¿ìè¶ÌÊÓƵ

Emile Argand was one of the first geologists to realise that the spectacular
structures of ancient mountain belts had a broader significance. In the
early part of this century, he recognised in the Alps signs of a great horizontal
convergence, decades before the theory of plate tectonics drifted into the
earth sciences. Argand’s significant achievement was to realise that the
4 kilometres of relief revealed in the Alpine mountains was not the limit
to geological thinking. He projected the major folds beneath their neighbours,
to give a three-dimensional picture of the mountain structure reaching as
far as 8 or 10 kilometres below the surface of the Earth. But that was as
deep as Argand could go from the rocks he could see at the surface.

Geophysical techniques such as analysis of seismic waves from earthquakes
tackled the problem of the Earth’s structure from the opposite direction;
they presented a picture of the Earth as a planet, its rocks divided up
into layers tens and thousands of kilometres thick. The two views are difficult
to reconcile; each seems almost irrelevant to the other. But now there is
a project linking the two visions of planet Earth-the European Geotraverse.
The geotraverse will be a collection of data on a cross section 500 kilometres
deep, through a 250-kilometre-wide swathe of Europe from Norway’s North
Cape to central Tunisia-the first section on this scale in the world.

The project began in the early 1980s, and the final experiment finished
last year. So far, the geotraverse has produced its first summary cross
section through Europe to the base of the lithosphere (the crust and part
of the mantle) as well as a burst of theses and research papers. Still to
come are a book, an atlas with the data from geology, density, gravity,
heat flow and other techniques and a host of new hypotheses as the data
reach a wider scientific audience. The results will take many more years
to analyse, but the geotraverse has already opened up European geology to
new ideas.

For the first time, geophysicists have data from a slice right through
the lithosphere of a continent. This picture is strikingly similar to geological
ideas about the structure of Asia south of the Himalayas, where collisions
have swallowed up several oceans, building line after line of ancient orogenic
belts. Europe shows a progression from old, stable crust and thick lithosphere
in the north, below Scandinavia, through increasingly younger deformation
farther south, ending in the most recent earth movements in the Mediterranean.

In 1980, at an international meeting in Paris, a group of European geophysicists
realised that earth sciences was becoming increasingly specialised, and
its researchers more and more isolated. To try and halt this trend, several
geophysicists, inspired by Stephan Mueller from the Federal Geophysical
Institute in Zurich, conceived the idea of combining all the techniques
for looking deep into the Earth to give a multidisciplinary picture of the
structure of Europe. Thus began the European Geotraverse, which got under
way in 1983, with support from the European Science Foundation (ESF), which
was itself funded by a consortium of national research bodies and money
from the Swiss government. The experimental part of the project finished
in 1990, by which time the geotraverse had carried out 13 major experiments
and prompted many more.

Derek Blundell, a geophysicist at Royal Holloway and Bedford New College,
part of the University of London, who has been involved with the geotraverse
since the mid-1980s, was in no doubt about Mueller’s key role: ‘In those
days, a decade ago, geologists looked at the surface and could extrapolate
down to about 3 kilometres, but few would dare to do anything else. At the
same time, geophysicists were setting up models that had no relation to
real rocks. Stephan Mueller realised that we needed to talk to the geologists
and put together a complete picture. It’s very much down to him.’ Blundell
also credits Mueller with piloting the project through the early stages
when representatives of national research bodies, who were sponsoring the
ESF, were pressing for the geotraverse to focus on their own parochial problems.
Mueller reorganised the disparate proposals into a coherent scientific programme.

Mueller himself is pleased with progress so far. ‘What we promised eight
years ago we have realised to a great extent. But we have not done everything.
There will be plenty to do for the next 10 years.’ Other scientists associated
with the project echo his satisfaction. ‘We now have the opportunity to
put together geology and geophysics on a new scale,’ says Blundell. ‘There
are a lot of mental gaps that act as barriers to understanding; we can bridge
the gaps by thinking about geology on the lithosphere scale. Most people
are fairly myopic; with the geotraverse, they have to look below and around
them as well.’

The geotraverse concentrates on integrating results from different disciplines,
often finding out most from where they disagree. This is a rare approach;
most deep crustal experiments concentrate on only one type of data, and
leave ambiguities unresolved. Enthusiasm about the results is palpable;
at a meeting in Strasbourg this spring, it was hard to get a word in edgeways
as people dashed from one geotraverse poster to another to compare interpretations
and dragged new-found colleagues to corners to discuss their own data and
models.

Perhaps the most lasting achievement of the project will be the contacts
established among the hundreds of scientists who contributed to both its
experiments, and the intensive and productive discussions that characterised
its scientific workshops. The method of working was a significant part of
the project’s success and the good news is that it need not cost the earth.
The ESF, the project’s paymaster, spent a mere £700 000 over the
ten years of planning and execution. And the support ESF continues to give
to the collation and publication of results will help to bridge the gap
between different disciplines.

The ESF got its bargain by taking an unusual approach to a project on
this scale. There was no institute, little traditional infrastructure, and
only a couple of paid staff. The scientific coordinating committee invited
scientists working on regions of Europe, and in varied fields and methods
to join different parts of the project. The response was enthusiastic, and
the success of the experiments depended on individuals from universities
and institutions contributing their instruments, equipment and computing
power.

Before each major experiment, such as a seismic profile, the researchers
would meet for two or three days to plan the campaign. The ESF paid travel
expenses, and national research funds contributed board and lodgings for
the workshops and study centres. After the experiments, 60 or so researchers,
new faces as well as those who planned the experiments, spent one or two
weeks away together working out what it all meant. Computers on site meant
that ideas could be tested immediately, often resulting in radical new models.
And the informal atmosphere meant that everyone from postgraduates to professors
had to argue through their ideas on equal terms. Participants questioned
the assumptions and methods used in producing the different types of data,
and as a result understood far more about each others’ disciplines when
the time was up. They also made friends. This was Mueller’s other great
achievement, according to Blundell. ‘The network of friends and contacts
that we now have will serve as a modus operandi for European science in
the future.’

One of the simplest yet most intriguing things to come out of the geotraverse
so far is the variation in the thickness of the Earth’s crust, the cold,
relatively light outer layer of the planet, which effectively floats on
the denser mantle below. The geotraverse has shown that northern and southern
Europe have distinctly different crustal thicknesses; below Scandinavia
the crust is 45 or 50 kilometres thick, whereas beneath Western Europe it
extends down only 30 kilometres. Like blocks of wood floating in water,
thicker crust should float both deeper and higher, as happens in the foundations
or roots of mountain belts. But there is little difference in the height
of the land in north and south, overall, so what accounts for the thicker
Scandinavian crust? Geophysicists are now questioning whether other forces
are acting on either segment of crust.

The character of the Moho itself, the boundary between crust and mantle,
has also been an intriguing feature of the geotraverse results. In the past
decade, many deep seismic experiments have produced images of the Moho (see
Box for more details of these experiments and what they found). A deep seismic
refraction profile called Fennolora provided a springboard for much of the
geotraverse work in Scandinavia. The experiments were carried out in 1979,
and the analysis of the data it produced is consequently the most refined.
Models of the lithosphere based on the Fennolora data reflect the power
of the integrated approach to geophysics that the geotraverse fostered.

Fennolora led to a model for the structure of the crust and mantle beneath
Scandinavia in which the Moho was an irregular shape, rather than flat or
smoothly curved, as previous geophysical profiles had inferred. In fact,
the best fit between model and data came for a Moho on which there were
steps, like irregular zigzags, between 10 and 20 kilometres high and roughly
200 kilometres long. Topography of this sort on the Moho was unexpected
and at first no one was sure how to explain it. The problem was compounded
when geophysicists measuring the Earth’s gravitational acceleration along
the same line found that their model of rock densities worked best without
anything like a Moho step. Their Moho had to be smooth and roughly flat.

Models that matched

The geotraverse approached this inconsistency head-on and its integrated
approach has brought insights into the structure and processes at the base
of the crust. The seismologists produced a model of layers in the crust
that transmitted seismic waves at different velocities. Once you have set
up a model of seismic velocities, you can derive a similar model of different
rock densities, taking into account temperature and other factors, and using
typical rock densities for each velocity. Variations in g, the acceleration
caused by gravity, depend on the distribution of rock that is more or less
dense than average; less dense rock makes g less than usual, for example.
Gravity modellers match their data against a model of layers of rocks with
different densities.

To reconcile the two models derived from seismic and gravity data along
the Fennolora line, researchers sought a rock type that has an unusually
low seismic velocity for its density. If the lower parts of the steps were
filled with such a rock, they could show such low seismic velocities that
they looked like crust on Fennolora, giving the irregular Moho, but high
enough densities so that on the gravity profile, they looked like mantle,
topped by a smooth Moho. This is not such a far-fetched an idea as it may
sound; a rock known as eclogite could fit the bill.

The petrology is accepted as feasible. If rocks of the lower crust,
originally above the Moho, were pushed into the mantle in several places
to form the lower parts of the steps, their mineralogy would change. Crustal
rocks cannot exist unaffected for long at the higher temperatures and pressures
of the mantle. As the millions of years passed, lower crustal rocks could
slowly transform to eclogite. Eclogite is metamorphic and can have a range
of compositions depending on the original lithology. Suitable lower crustal
rocks could form eclogites with the properties-low seismic velocity for
its density-that the Fennolora evidence strongly suggests.

If the interpretation of the Fennolora results is on the right track,
they mean that the shape of the Moho beneath this part of Scandinavia has
persisted since the last bout of deformation there, some thousand million
years ago. The seismologists have evidence that the Moho has kept this shape
even when the density of the rocks within them increased so that the gravity
modellers could not tell them from mantle.

In the past decade, deep seismic reflection sections have shown that
the Moho is distorted and offset by faults, as Simon Klemperer and Richard
Fifield described in ‘Sound waves reflect Britain’s deep geology’ (¿ìè¶ÌÊÓƵ,
4 February 1988). These sections imply that the Moho changes shape in response
to forces in the lithosphere; the new results raise the possibility that
chemical and petrological changes also alter what we see as the Moho through
various geophysical techniques. As Blundell says: ‘The Moho must not be
treated as a passive marker. It is dynamic mechanically, and, I believe,
also dynamic chemically.’

Another first for the geotraverse along the line of Fennolora was the
image of the three distinct layers about 30 kilometres thick within the
Scandinavian lithosphere; each had a seismic velocity distinctly lower than
the lithosphere above and below. Seismic velocity depends on density, which
in turn varies with the type of rock and its temperature. Because the layers
were so thin, they are unlikely to represent temperature variations; they
could be layers of different types of rock, although no one is prepared
to say exactly what. And again, these layers are persistent; the lithosphere
has held its shape and its layering for many millions of years.

More information pertinent to the lithosphere came from a deep seismic
reflection line called Babel, carried out east of the line of Fennolora
in 1989. Babel was again not part of the geotraverse, but one of the many
parallel projects. This experiment produced an intriguing cross section
beneath the Gulf of Bothnia, showing structures at depth matching the surface
traces of an orogenic belt that formed 2300 million years ago. Surprisingly,
it looked like the section produced by the French national seismic team
(ECORS) across the Pyrenees-a mountain belt that formed only 40 million
years ago by collision. Both had a deep crustal and lithospheric root, despite
the difference in age. ‘A looks like B does not mean that A equals B, I
know,’ says Blundell, ‘but the Gulf of Bothnia is so similar to the ECORS
Pyrenees line that there is a strong presumption that they formed from the
same kind of process.’ Here is evidence that the same sort of plate tectonic
processes that formed the Pyrenees could take place more than two billion
years ago, in early Proterozoic times. This means that geologists working
on rocks of this great age can have confidence in interpreting their structure
in terms of today’s plate tectonics.

But the very existence of these roots today is perhaps more significant.
Whatever the origin of the strong signals from these depths, they have kept
their shape for billions of years. Geophysical theory in the past has said
that rocks at these depths should flow easily, like a nicely ripe piece
of Camembert; the differences in density between crust and mantle would
produce buoyancy forces that should remove the roots of the mountain belt
once the compressive forces diminish. But in the Gulf of Bothnia the lithosphere
is strong enough to keep its shape for billions of years-its rheology is
closer to Parmesan than Camembert.

A cautionary note comes into this overhaul of geophysical theory from
results in some younger mountain belts to the south. The geotraverse crossed
both the Alps, formed in the past 40 million years and the Variscides, some
200 million years older. Traces of the Variscides at the surface stretch
across Europe from Germany to Ireland, although erosion has ensured that
they are no longer mountainous terrain. Both orogenic belts formed in a
similar way, under compression from movement of the Earth’s tectonic plates.
But they look very different in the geotraverse. The Alps have their crustal
and lithospheric roots, but the Variscides have none. There could be many
reasons for this. Perhaps the Variscan mountains formed in a different way
and never had roots, or maybe the roots have had time to flow away, and
the crust return to its normal thickness. But if the latter reason, why
have the far older roots to the north stayed? The geotraverse data cannot
yet answer these questions, partly because so much of it is unique. We may
have to wait for data from many more orogenic belts.

The Alps have been a testing ground for many a geological theory, and
home to some of the most intensively studied rocks on Earth. They continue
to test geophysics along the line of the geotraverse. Well over a century
of geological observation and analysis has produced consensus only on the
simplest points of Alpine geology: the Alps are the remnants of two sides
of an ocean which collided and then contracted to form the folded and faulted
mountains we see today.

Gravity surveys have long shown that the Alps have a crustal root. The
debate has focused on what happened when the continents met. Were the strains
that built the mountains vertical movements in the main, or did one side
slide over the other, pushing rock along and upward to build the peaks?
Both ideas, and everything in between, have had their advocates. Some models
even implied that continental crust must have followed the subducted ocean
crust down into the mantle, contrary to the original tenets of plate tectonics.

Plate tectonics says that continental crust is too buoyant to be pushed
down into the mantle. Yet this controversial idea popped up in the Alpine
literature several times in the 1980s. Hans Laubscher, formerly at the University
of Basel, and Dan Bernoulli, then his colleague and now at the Swiss Federal
Institute of Technology in Zurich, looked at the thickness of individual
fragments of crust folded and faulted to form the central Alps. They found
that none of the individual panels was more than 10 kilometres thick, and
they were all upper crustal rocks. If the crust had been the typical 30
kilometres thick, they wondered where the lower crust had gone.

Rob Butler, now at the University of Leeds, took a more quantitative
approach in the western Alps, working out the kilometres of contraction
by unravelling folds and faults visible at the surface as if straightening
out a rumpled and torn tablecloth. Butler calculated that the continents
on each side of the Alps had converged by at least 400 kilometres in the
past 40 million years. ‘But while we can locate all the upper crust of the
old continental margins,’ said Butler, ‘we cannot account for all the lower
and middle crust that we estimate existed. There’s just not enough root.’

He speculated that some of the crust may have been able to subduct because
it transformed to a much denser mineralogy at mantle temperature and pressures-dense
enough to sink and stay down. The rock he invoked was another form of eclogite.
Butler put the missing crust below the Po plain in northern Italy, but points
out that if it kept its normal buoyancy during subduction, the surface above
should have been pushed up, as seems to have happened in the Himalayas.
If this was the case, he says, ‘Milan and Turin should be on a high plateau
like Tibet, not close to sea level.’

Conversion of the lower crust to eclogite is a convenient way to solve
the buoyancy problem and allow continental subduction. But, like all good
theories, it raises further questions, especially about the thermal structure
of the subducting slab. There seems to be only a limited group of circumstances
in which the right metamorphosis could happen. If the transition to eclogite
happens as continental crust subducts, it must take place before the subducted
slab heats up too much. Otherwise the crust would melt and generate huge
volumes of granite, with a density typical of crustal rocks; no way to solve
the buoyancy problem.

The geotraverse through the Alps incorporated data from several seismic
reflection and refraction sections, including ‘wide angle’ seismic surveys
in which receivers were positioned far from the shot points to pick up strong
reflections from Moho depths. Combining these results gave the geotraverse
a profile that showed the pattern of slices of rock overlapping each other,
confirming the important role played by horizontal movement in building
the Alps. The Moho bulged downwards, outlining the crustal root, and in
the central part had been pushed up on itself, bringing a wedge of mantle
up into the crust.

But the deepest part of the root was not where the geophysicists expected
to find it. Instead of lying centrally under the high Alps, the root was
asymmetric, with its deepest part beneath the Po plain. The slab of continental
crust from the northern margin dips underneath the southern side, and continues
down as far as 80 kilometres to the south of the Alps, whereupon it vanishes
from the seismic sections. The data reinforces the question already raised
by geologists on the basis of the rocks exposed at the surface-where has
all the lower crust gone? Blundell’s interpretation was that the lower crust
had indeed been pushed down, but then transformed to an eclogite with a
seismic velocity indistinguishable from other mantle rocks-roughly 8 kilometres
per second. As a consequence, the seismic sections could see only part of
the subducted slab.

The details of the geotraverse section back up this idea. Moving south
on the section, seismic velocity in the continental crust gradually increased
from the normal 6.5 metres per second to 7.4 kilometres per second before
the slab vanished. But a velocity of 7.4 kilometres is not sufficiently
different from the mantle for the slab to show up anyway on the seismic
sections. ‘Perhaps this gradual increase in velocity represents a metamorphic
transition between normal and eclogitic crust,’ suggests Butler, ‘we just
don’t know.’

One problem with interpreting this part of the Alpine section is that
although the velocity contrast suggests that the slab disappears on the
section, there is very little data from precisely this area. The geotraverse
designed their experiment to look for a root centrally beneath the Alps-none
of the planners expected an asymmetric root. In consequence, there were
few recorders in the right places to pick up signals from below the Po plain.
Perhaps resolution of the deep structure of the Alpine root will have to
wait for a deep seismic section with seismometers in the best places to
pick up signals from the root zone, in the Ligurian Sea in Italy, for example.

Ask new questions

The Alpine arguments are nowhere near resolved by the geotraverse data.
If anything, the new information has fuelled the debate in areas such as
how one part of the Alps relates to another, why the compression appears
to be directed in different directions around the arc and how the lower
crust could be metamorphosed to eclogite. There are plenty of ideas to be
explored across the rest of the continent, too: the persistence of lithosphere
roots in the Gulf of Bothnia and their absence in central Europe, the long-lived
shape of the Moho, the mysterious layers in the lithosphere and whether
eclogite metamorphism can work all the wonders researchers have suggested.
Further deep seismic sections are already planned, as a consequence of the
expertise and contacts built up in the geotraverse. And there is no shortage
of ideas for more.

But the European geotraverse will be more than just a mine of local
information; it has already raised fundamental queries about form and process
in the top few hundred kilometres of the Earth. The extra clues it provides
have already proved to be enough to alter questions from ‘what is going
on?’ to ‘why does that happen?’-a much more satisfactory query. As an example,
this first look at the deep lithosphere has already shown that both layering
and the structure of the lithosphere can persist for hundreds of millions
of years. The lower crust is not the weak layer envisaged in rheological
models of the past; it must be strong enough to hold its shape. We may have
to revise more ideas in the light of future data, for the real work of interpretation
has only just begun.

Much of the excitement about the geotraverse has arisen because there
is no other comparable data. But because the section is unique, some of
its features are especially difficult to interpret. Is a structure such
as the relatively thin crust in the Variscides of central Germany intrinsically
unusual, or just the only example captured in this one data set of a process
common elsewhere in the world? Only more data on this scale will resolve
such questions. Across the world, more cross sections are planned in the
Global Geoscience Transects project run by the International Lithosphere
Program. Perhaps in a few decades we will be able to compare the first continental
geotraverse with others. The final geotraverse study centre involved scientists
from eastern Europe and increasing links across the continent can only strengthen
future work. If this reveals as much as the European geotraverse, the world
will be a more interesting, if not a better place.

But perhaps the most lasting and significant achievement of this project
was the communication that has developed between researchers from different
disciplines, fields and countries. Geology is a subject that crosses borders
more obviously than most. It also suffers most from the parochial approach.
If the network of contacts fostered in Europe through the geotraverse experiments
and study centres survives, they will have more lasting significance than
any theory.

* * *

Layer upon layer: the physics of the deep Earth

The Earth’s interior is completely inaccessible. Geophysicists have
had to infer the properties of rocks below the surface through a battery
of diverse techniques, resulting in some ambiguities and uncertainties about
what is really there. When two measurements made using different techniques
appear to come from the same region underground, it is tempting to infer
that they originate from the same feature, perhaps an unusual layer, or
a fault. Often they do, but this is not always the case.

For the past 100 years or so, geologists and later geophysicists have
analysed the pattern of vibrations from earthquakes across the planet. Layers
of different density within the Earth reflect and refract the seismic energy.
Comparing the record from different places gives a picture of the major
layers of the Earth.

The Moho, recognised by the geophysicist Andrija Mohorovicic in 1909,
is one of the most obvious boundaries between layers in the Earth. It coincides
with the depths expected from gravity data for the transition from light
crust to denser mantle. The terms Moho and crust-mantle boundary are often
interchangeable, but they need not refer to the same surface.

As seismologists developed their science, they began to analyse different
types of seismic waves. They noticd a difference in arrival times between
pressure waves, in which particles of rock move only in the direction in
which the wave travelled, and shear waves involving particles moving at
right angles to the wave. They found a sharp dis-continuity rather deeper
than the Moho along which the shear waves only were noticeably attenuated.

The geophysicists inferred that this discontinuity marks the depth at
which rocks begin to melt. Fluids cannot support shear stress, so partially
molten rocks should be less efficient at transmitting shear waves. This
discontinuity divides the upper Earth mechanically: above it lies the lithosphere,
the strong outer layer that deforms elastically, and below is the asthenosphere,
where rocks are weaker and can flow.

The distinction is important in the theory of plate tectonics, in which
plates of lithosphere, not crust, move on the weaker asthenosphere below.
The lithosphere is thicker that the crust, and so includes both crust and
mantle rocks, with their different densities. Lithospheric and asthenospheric
mantle can be the same type of rock, either solid or partially molten.

As well as using natural signals from earthquakes, geophysicists set
up their own seismic experiments. To look deep into the continents, they
use explosions as sources of energy, and concentrate on signals either reflected
from layers vertically below, or refracted and picked up tens or hundreds
of kilometres away. Reflection and refraction produce different types of
results: reflection gives a good picture of the shape of layers, while refraction
is the best way to see the overall layered structure.

Boundaries will affect the passage of seismic waves most strongly where
they involve a marked contrast in physical properties. Density is the most
important variable for interpreting the results. A sharp change in density
will reflect or refract the energy, sending a signal to the surface to betray
the existence of the discontinuity. Boundaries that send back a strong signal
are known as reflectors.

The Moho shows a seismic signal almost everywhere. In some places, it
forms a strong, relatively narrow reflector. Elsewhere, the mantle sends
back few signals, but the lower crust shows a distinctive pattern of short,
nearly horizontal reflectors. Where these reflectors stop marks the Moho.

With refraction profiles, the contrast in seismic velocities between
crust and mantle usually results in a strong signal. When reflection and
refraction data are available, the position of the Moho usually matches.
It is easy to recognise the lower crust, but no one is sure what the unusual
reflectors represent. Nor does anyone know what makes the Moho appear different
in different places, sometimes only 20 kilometres apart.

The various different methods used to explore the deep Earth-and these
are but a few-call to mind the old story about the four blind men describing
the elephant. One felt the trunk, one the skin, and so on, and each gave
a totally different description of the same beast. Geophysicists are one
step removed from this. Rather than seeing even part of their quarry, they
have to work out its nature from its effect on gravity, earthquakes, magnetism,
heat flow and so on.