91av

The Pacific Ocean, averaging 16 000 kilometres wide and 4600 metres
deep, is the greatest expanse of ocean on Earth. It may also be the oldest.
Geologists who have matched up the geology of the lands that edge the Pacific
think that this ocean has persisted for more than 600 million years. Starting
when the rocks that now form the cores of North America, Antarctica and
Australia began to separate, ocean floor began to form; it has continued
ever since. In that time the Tethys Sea – lying roughly where the Mediterranean
is now – stretched and shrank, the Atlantic opened, shut and opened again,
and continents rifted and collided on the complicated paths to their current
positions.

Since the acceptance of plate tectonic theory in the early 1960s, geologists
have known that the size and shape of the world’s oceans change constantly.
Many of our familiar oceans did not even exist in the past and those that
did were very different in form.

Plate tectonics was not the first theory to suggest that an ocean may
have had a beginning; in 1881, Osmond Fisher, an English clergyman with
an interest in natural sciences, suggested that the basin of the Pacific
was the depression left over when the Moon broke away from the Earth.

Now geologists know that the Pacific and other oceans were born when
continents rifted apart to make new plate boundaries. Molten rock welled
up into the rift, forming new seafloor that spread to build the ocean. Such
rifting is a response to changing circulation patterns in the mantle beneath
the Earth’s crust. As the rifting proceeds, molten rock accumulates at the
rift, basalt spreading like a carpet on each side of the new plate boundary.
The two continents gradually move apart, separated by this new ocean crust.
Older oceanic lithosphere – the outer rigid layer of the Earth – returns
to the mantle along ‘subduction zones’, boundaries where plates converge.
They are marked by the vast ocean trenches that form the deepest parts of
the oceans, and by continuous chains of active volcanoes known as volcanic
arcs.

The five major lithospheric plates that make up the modern Earth have
both oceanic and continental parts. The continents move across the surface
of the Earth in response to rifting and subduction, and the oceans wax and
wane between them. When an ocean closes, two continental masses collide
and create ranges of mountains; this happened 45 million years ago in the
eastern Tethys Sea, when India collided with Asia to begin the growth of
the Himalayas. Understanding the journeys of these ancient continents in
the geological past is an exercise akin to imagining the past 10 moves made
with a Rubik cube: modelling the old geometries becomes more and more difficult
the further back in time one peers.

Geologists agree on plate reconstructions as far back as early Jurassic
times, about 180 million years ago. This is because most of the evidence
they use for reconstructions comes from the rocks of the ocean floor. The
world’s oldest sea floor – in the Pigafetta basin east of the Mariana Islands
in the western Pacific – is between 160 and 170 million years old. At its
simplest, plate reconstruction amounts to no more than putting continents
back together. The similarity of the shapes of the coastlines around the
bulge of South America and the West African bight was one of the things
that, in 1915, compelled the German meteorologist Alfred Wegener to propose
his hypothesis of continental drift – at the time an unthinkable idea –
to a gathering of perturbed German scientists. Now, the shape of the sea
floor at a depth of 2000 metres is considered to mark the average point
of breakage when two continents split. Joining up these puzzle pieces gives
the best match between continents.

Information on the direction in which the plates have moved also helps
reconstruction. Basaltic ocean floor commonly contains fractures, easy to
see with detailed bathymetry, which are oriented parallel to the direction
of plate motion. To understand how a particular ocean opened, simply return
opposing continents to their positions before rifting using the fractures
as a guide. A more precise reconstruction technique employs the magnetism
preserved in the rocks of the ocean floor. Systematic reversals in the polarity
of the Earth’s magnetic field have left their mark in the basalts that built
the ocean crust. As the ocean grows, stripes of normal and reversed magnetism,
varying in width, spread out from the ridge to form a symmetrical pattern.
Fossils preserved in the sediments immediately above the igneous rocks give
each stripe a date, and the pattern of thick and thin bands forms a timescale,
much like the variable width of growth rings in trees. In plate reconstructions,
particular stripes are pulled back to the ridge to show where the continents
were at that time in the history of the ridge. Computer simulations offer
very precise fits and detailed sequences of plate motions. This is a far
cry from the first tentative reconstructions. Warren Carey of the University
of Tasmania, for instance, made some of the first reconstructions by moving
meticulously cut out and scaled paper continents around on a globe.

Reconstructions of the world before Jurassic times are a more difficult
proposition. There is no intact ocean floor that dates from this time; it
has all been subducted. As a result no information is available from ocean
floor fractures and patterns of magnetic stripes. Matching coastlines is
not always a reliable guide because they tend to be changed by erosion,
sedimentation, faulting and folding; the older they are, the greater is
the disparity. So geologists have had to concentrate on the continents themselves,
using palaeomagnetism, correlation of the geology and ancient climates from
before rifting, radiometric dating, the distribution of distinctive fossils
and contemporaneous deformation – and liberal use of Sherlock Holmes’ principle
of least astonishment.

Palaeomagnetism became a valuable tool once geologists realised that
the orientation of the magnetic field preserved in rocks could be related
to the latitude where the rocks lithified. Minerals in some igneous rocks
record the orientation of the Earth’s magnetic field at the time and place
where they formed: crystallisation near the equator produces magnetic field
vectors that are roughly horizontal, while close to the North or South Pole
the field lines are much more inclined. The tilt tells geologists the latitude
of the rock when it formed; but it conveys nothing about longitude, because
the Earth’s field is roughly symmetrical about its axis. (This technique
assumes that the magnetic field has always been aligned with the Earth’s
rotational axis.) Measurements on rocks formed at different times give a
sequence of latitudes for a continent through time, reflecting the meanderings
of the continent across the Earth. The method becomes less reliable where
a rock has been subjected to heat and pressure after its formation, because
these processes can reset the orientation of the magnetic field in its minerals.

Palaeomagnetism is not the only tool; it is perhaps most valuable when
combined with other techniques. The most common method for reconstructing
the positions of the continents is to correlate geological units and structures
formed before rifting, which now appear on more than one continent. Linear
features of substantial width, such as ancient fold-belts oriented at moderate
to high angles to the rift, are most useful. They have a chance of surviving
the deformation which commonly follows rifting and subduction and are big
enough to be clearly visible on satellite photographs and easily detected
by geophysical surveys.

Ian Dalziel of the University of Texas has used this type of correlation
to gain a new understanding of the history of the Pacific Ocean. But another
approach sparked his initial interest in the beginnings of the Pacific.
Dalziel had been working on the rocks along the Pacific margin of Antarctica
in terms of the classic sequence of rocks expected to form as a continent
breaks up. He then indulged in the hypothetical exercise of standing on
the shore of the southernmost Pacific wondering where the matching rocks,
from the other side of the rift, had gone. His answer came one day in the
summer of 1990, when Eldridge Moores, his colleague at the University of
California at Davis rang and asked how he would react to a correlation between
the Precambrian rocks of western North America and eastern Antarctica, on
the basis of his extensive knowledge of the geology of both areas? Dalziel’s
reaction was swift and positive. ‘I liked it,’ he reflected. ‘It had an
element of surprise – previously people thought that no continent had ever
crossed the Pacific Ocean.’

Dalziel and Moores had begun to discuss problems in Antarctic geology
on an International Geological Congress field trip to the Scotia Arc in
Antarctica in 1989. They realised that any correlation with Antarctica also
involved Australia, because the two were one landmass until about 95 million
years ago. They therefore examined the geology of both Australia and Antarctica
to test the idea of a link with North America in late Precambrian times,
some 750 to 600 million years ago. In fact, such a link was proposed by
Richard T. Bell and Charles W. Jefferson of the Geological Survey of Canada
in 1987, providing a useful base for them to build on. Dalziel maintains
that ‘the histories of the eastern Antarctic/Australian margin and the west
coast of North America immediately prior to and during the early Cambrian
were extremely similar. The difference is that western North America is
still a well-preserved passive margin, whereas the trailing edge of the
Australian/Antarctic continent immediately went into compression because
of the initiation of subduction’.

There was subduction in Australia at least 520 million years ago, welding
the deformed crust which now underlies most of Victoria, New South Wales
and Queensland, to the craton or stable part of the crust that makes up
most of Australia. This process camouflaged its earlier history. If you
imagine removing the younger rocks and reversing the displacements of the
biggest of the younger east-west faults, you find a fairly straight, rifted
margin stretching for 4500 kilometres and similar in dimensions to the rocks
of the same age on the west coast.

In searching for a more exact idea of the relationship between the two
continents, Dalziel and Moores found that the Grenville Front, a continent-wide
linear geological feature in North America extends into Dronning Maud Land
in eastern Antarctica. This line, which stands out dramatically on geological
and aeromagnetic maps of North America, probably represents the line of
collision between two older continents about 1100 million years ago. It
separates crust formed between 1600 and 1750 million years ago from crust
with structures that are much younger, between 1000 and 1300 million years
old, aligned in a different direction.

The Grenville Front can be traced down from Canada as far as Texas,
where it forms the huge, rounded, rocky hills of the Llano Uplift, the source
of the stone used for many prominent Texan buildings, including the State
Capitol in Austin. The critical evidence for the extension of the zone in
Antarctica lies in three lonely peaks or nunataks projecting through the
icecap near the mountains of the Shackleton Range. Dalziel notes that this
area has been ‘long regarded as anomalous by Antarctic geologists because
its physiographic and structural trend is at right angles to other mountain
ranges in the region’.

Extending markers such as the Grenville Front into Antarctica is obviously
fraught with difficulty because the rock is almost completely covered with
ice. In this instance, the unusual age of the nunataks and geophys-ical
trends beneath the ice together provide strong evidence for the link across
the Pacific. Radiometric dating of different portions of the igneous rocks
of the nunataks has yielded two ages, 1023 and 840 million years old, which
differ from those found for other formations in Antarctica. According to
Dalziel, this is evidence that the rocks are part of the Grenville Front.
But he and Moores originally disagreed on the path of the Grenville Front
beyond the Shackleton Range. Moores believes that it follows the Antarctic
coast, where it acted as the basic structural weakness along which Antarctica
and Australia rifted apart, before entering Australia along a zone called
the Albany-Frazer Belt. Dalziel now accepts this in the light of recent
comparative work on the geology of Antarctica and South Africa.

But Dalziel and Moores always agreed on the major point, that the Grenville
Front can be matched across the Pacific. Other similarities between the
two coastlines soon became evident. For instance, Bell and Jefferson pointed
out that rocks in northwest Canada have been deformed so that they have
a distinct ‘structural grain’. The same pattern and orientation can be traced
into rocks of similar ages and histories in southern Australia and Adelie
Land in Antarctica. A belt of rocks that probably originated as exhumed
ocean floor – an ophiolite – can similarly be traced in a discontinuous
band running from west Antarctica into southern North America. And if the
continent rifted apart in late Precambrian times, the geometry of some sedimentary
basins of similar age in both continents makes sense. Several of these basins,
filled with limestone, lie parallel to the postulated rift zone that predated
the Pacific Ocean and most likely formed as the crust was thinning before
the continents separated. There are similar Cambrian limestone belts along
the eastern margin of North America and Dalziel recognises these as evidence
of a second great rift that opened close to the present position of the
Atlantic. In doing so he confirmed conclusions of previous researchers in
the same area, but violated his own espoused axiom, that ‘oceans don’t need
to close where they open’.

At present the available palaeomagnetic data supports the proposed geological
relationship of North America to Gondwana in Cambrian times. Dalziel says
that there is no ‘smoking gun’ in the palaeomagnetic data to prove that
this reconstruction is impossible. On the other hand, the palaeomagnetic
data are hardly conclusive. Dalziel and Moores readily admit that better
palaeomagnetic data would provide a good test of the hypothesis and rate
this area a high priority for future research.

FOSSIL EVIDENCE FORCES RETHINK

The fossil record from these times presents a more significant problem.
There are distinct differences between Cambrian trilobites on either side
of the Pacific Ocean, a type of disparity usually taken as a sign that two
areas were separated by an expanse of open sea. After hearing Dalziel’s
presentation of his theory in June 1991 at the eighth Gondwana conference
in Hobart, Australia, John Long, of the Western Australian Museum in Perth,
noted that these differences were already well established at the beginning
of the Cambrian Period, 540 million years ago. Long implied that the continents
must have separated at a significantly earlier date for the species to have
had a chance to evolve separately. Australian redlichiid trilobites were
free-swimming organisms inhabiting shallow coastal seas, and only a sizable
ocean would have prevented them mixing with distinct ollenelid trilobites
living in similar western North American seas at the same time.

In the months since the publication of the new rifting model, the palaeontological
evidence has forced Dalziel and others to consider an older age for the
rift to drift transformation, perhaps as old as 750 million years. This
is the age of the oldest sediments which can be reasonably attributed to
the break-up event. However, this may not be the only way to account for
the trilobite differences: narrower oceans with currents running parallel
to the continental margins could also have prevented faunal migration, and
would permit a younger break-up.

The implications of the new reconstruction are not confined to palaeontology.
The model indicates that Precambrian mineral belts such as the Broken Hill
and Mount Isa regions of Australia, containing valuable deposits of lead
and zinc, might continue in north-western Canada.

Paul Hoffman of the Geological Survey of Canada shares the wide view
of the Moores-Dalziel model. Hoffman believes that the opening of the Pacific
Ocean was only part of the break-up of a vast supercontinent which literally
‘turned inside out’ over a period of 200 million years. The process began
when the giant continent split into three wedge-shaped pieces which rotated
away from one another, like an opening fan, around a point close to what
was then the equator. The Pacific Ocean and a precursor to the Atlantic
formed in this way. Laurentia, the name given to the continent made of North
America and Greenland, formed the middle segment. Ultimately the wedge comprising
Australia and eastern Antarctica rotated to the point where it collided
with the wedge containing East Africa and South America, which was itself
spinning in the opposite sense. In the process, coasts were pushed together
to create mountain ranges in the interior of a new continent known as proto-Gondwanaland.
An ocean the size of the Pacific – termed the Mozambique Ocean by Dalziel
– was consumed, and its sediment was folded and incorporated into the huge
landmass. Radiometric dating indicates that the whole process occurred very
quickly, geologically speaking, perhaps in as little as 200 million years.
This speedy rearrangement may have been responsible for the sudden changes
in oceanic life recorded by fossils from this time. Hoffman suggested that
the subsequent radiation of Cambrian metazoan life owed something to the
many new shallow seas that surrounded the fragmenting continents.

Hoffman’s overview may seem grandiose, but it is based on many interlocking
pieces of evidence. It takes a special ability to step back from the jumble
of small-scale evidence and assemble a working hypothesis of continental
amalgamation. Perhaps the greatest challenge is to invent reliable and discerning
tests of such hypotheses, and to be prepared to abandon them if they fail
to hold water.

Dalziel and Moores have provided good evidence for the hypothesis that
the Pacific Ocean began as a trickle 750 million years ago and has stayed
open to the present day. In the intervening period the Atlantic Ocean, or
its equivalent, has opened twice and closed once; at present it is growing
at the expense of the Pacific. Ruud Weijermars, of Uppsala University in
Sweden, has predicted that at current spreading rates the Pacific will disappear
completely in another 50 million years. By that time, Japan will be compressed
against China as Australia continues its relatively fast northward movement.
But in view of the charmed life the Pacific has led in the past – not to
mention the vagaries of plate motions – who can really be sure?

Garry Davidson is a post-doctoral fellow at the University of Tasmania
Centre for Ore Deposit and Exploration Studies in Hobart, Australia.

Further reading: ‘Southwest US – East Antarctica (SWEAT) connection:
a hypothesis’, by Eldridge Moores in Geology, vol 19 p 425, May 1991, was
followed by a paper by Ian Dalziel in the same journal in June 1991 (vol
19 p 598). Paul Hoffman wrote about early plate reconstructions in Science,
vol 252, p 1409.

* * *

The 500 million year waltz of the continents

The Moores-Dalziel reconstruction of the opening of the Pacific Ocean
is important because it fits so well into the accepted models of how the
continents have moved over the past 500 million years. The collapse of proto-Gondwanaland
was an early stage in a cycle of continental amalgamation and separation
that runs through Earth history.

Names for fossil continents and fossil oceans have proliferated since
studies of continental drift began, and there is now a bewildering variety.
Unhappily, the boundaries of each dead landmass are often disputed, and
the same areas sometimes have different names.

John Veevers of Macquarie University in New South Wales, believes that
Earth history can be viewed broadly as repeated cycles of supercontinent
amalgamation and fragmentation, at intervals of about 400 million years.
The new understanding of the origins of the Pacific bring into focus an
early episode of fragmentation.

Each episode may correspond to a Wilson Cycle, a term which describes
the sequence of ocean opening, closure by subduction and subsequent continental
collision and deformation. The cycle was named in honour of Tuzo Wilson,
a Canadian geophysicist who was a pioneer in the theory of plate tectonics.

In late Precambrian times, 750 to 700 million years ago, the break-up
of an un-named supercontinent led to a new continental configuration. Today’s
southern continents and India were loosely clustered together as Gondwanaland.
Most of Europe, known as Baltica, and Asia were separate.

Dalziel and Moores posit that during this period, cracks appeared in
the anonymous supercontinent, which they have named proto-Gondwanaland.
Laurentia, made up of North America and Greenland, broke out of the assembly,
leaving a widening Pacific Ocean in its wake. A narrow seaway called the
Iapetus Ocean probably developed at the same time or soon after, separating
Baltica from the side of Laurentia that is now Greenland.

A new cycle of amalgamation began in Lower Silurian times, about 420
million years ago. The Iapetus Ocean and the forerunner of the Atlantic,
which lay between Europe and North America, closed up, linking Scotland
to England and forming a landmass that is known as Greater Laurentia. Fusion
of Asia and Greater Laurentia along the line of the Ural Mountains formed
the bulk of Euramerica soon afterwards.

The supercontinent Pangaea (from the Greek, meaning ‘all Earth’) was
ultimately formed when Euramerica and Gondwana-land joined. This amalgamation
remained stable for nearly 150 million years, as the landmass drifted steadily
north.

Pangaea began to break up in late Triassic to early Jurassic times;
North America split from West Africa and Europe between 180 and 200 million
years ago and the final major separation was complete about 90 million years
ago. Some continents broke away long after others; North America separated
very early on, whereas eastern Gondwanaland, made up of Australia, Madagascar,
Antarctica and India, stayed together until about 100 million years ago.
Then India broke free and headed north across the Indian Ocean.

Australia and Antarctica were thought to have begun to separate about
45 million years ago, but seismic and sea floor drilling have shown that
this happened at about the same time as the departure of India. Africa and
South America were also estranged about 90 million years ago.

Today, the plates continue their steady movement. Two places in particular
could be the sites of oceans of the future. There is currently an active
rift between West Africa and East Africa beneath the Red Sea, and the major
oceanic ridge, the East Pacific Rise, is propagating into the west coast
of North America.