of sub duction on its edges. In contrast, the Atlantic spreading
rate probably reflects something close to the mantle’s natural,
The pace of the earth 71
stately pace. Earth’s convective behavior should therefore be
considered an “active lid” system, in that the plates do not
merely dance to the mantle metronome but in some cases set
their own meter, and this ultimately dictates how fast moun-
tains grow. To build mountains, however, we first need to cook
up some continental crust, and the recipe takes us back to the
midocean ridges.
WAT E R W O R K S
Vine and Matthews correctly interpreted the morphology of
ocean ridges as a record of the cooling of successive batches of
basalt. But fresh ocean basalt does not give up its heat passively,
like our soufflé cooling quietly in the kitchen. Instead, heat is
robbed from it by frigid ocean water, which streams through
cracks and pores, jealously steals joules, then makes a high-
speed getaway through chimneylike underwater geysers called
black smokers. The water also pilfers elements like calcium from
the young rocks, and leaves behind sodium, mediating the
salinity of the oceans. (This was unknown to John Joly when
he tried to estimate the age of the Earth based on the salinity
of the sea. His value of 100 million years was not meaningless—
but it represented the typical residence time of sodium in the
sea, not the time since Earth’s formation). It is estimated that
the entire volume of the world’s oceans flushes through rocks
of the midocean ridges in about 8 million years.4
Not all the infiltrating water escapes, however. Having en-
tered into labyrinthine passages and forged chemical bonds
with minerals in the basalt, some is now locked into the ocean
crust for the long term. As it happens, this accidental entrap-
ment of water is one of the most essential components of Earth’s
tectonic system. A subducting slab carries the stowaway water
72 Ch a pter 3
from its youth as it descends into the mantle. The cold slab
slowly warms, and when it reaches depths of about 30 miles, it
finally sweats out this ancient seawater. We tend to think of the
water cycle as a relatively short- term phenomenon; the average
molecule of water stays in the atmosphere for about nine days;
the residence time of water even in the largest lakes, like Supe-
rior, is a century or two; deep groundwater may be stored for
a millennium. But there is a 100 million- year water cycle that
involves the interior of the Earth, and adding water to the man-
tle is in fact the critical step in the recipe for continental crust.
In the presence of water, the otherwise solid rock in the
wedge of mantle above a subducting slab will melt at significantly
lower temperatures than it normally would, in the same way that
salt lowers the melting temperature of ice on a sidewalk. This
“water- assisted” melting is both creative and destructive: it ulti-
mately forges new continental crust but does so via some of the
deadliest volcanoes on Earth, which form on the overriding plate
in the subduction zone, directly above the spot where the down-
going slab gives up its long- sequestered water. The volcanoes
typically form an arcuate chain— a broad C shape that reflects the
curvature of a subduction trench on a spherical Earth, like the
crescentic shape of a dent in a ping- pong ball. Where the upper
plate is also basaltic ocean crust, the volcanic chain is called an
island arc. Examples include Japan, Indonesia, the Philippines,
the Aleutians, and the north half of New Zealand. If a subducting
plate dives beneath a continent, the resulting volcanoes form a
continental arc, like the Cascades and Andes (see figure 8).
In both arc settings, the water- generated mantle melt must
make its way through the upper plate en route to the surface.
The magma may be stalled by the rigidity of this lid of obstruct-
ing crust and, while ponded there, partially melt it. Just as at
midocean ridges, the low- melting temperature components
The pace of the earth 73
will be most readily extracted, giving rise to new magmas that
are even richer in silica and less like the mantle than basalt
is. Through many such cycles of such smelting, crust that is
progressively more “evolved” is generated, ultimately yielding
granite, the lightweight raw material for the continents. Plate
tectonics on the modern Earth is an extraordinary system. The
creation, maturation, and eventual destruction of ocean crust
are all necessary for the genesis of continental crust— a perfect
samsara cycle of birth, death, and reincarnation.
M O U N TA I N T I M E
An oceanic subduction zone will function smoothly (though
not necessarily aseismically) as long as the crust entering the
trench is thin and dense enough to slide into the mantle. But
if the slab pulls in “undigestible” things like ocean crust that is
too hot or too thick, or a lumpy old island arc— or an unsink-
able continent— traffic comes to a halt. And if the upper plate
is a continent, a major pile- up is unavoidable, and a mountain
belt begins to grow. The loftiest mountains on Earth, like the
Himalaya now, and the Alps, Appalachians, and Caledonides
in their day, form when a long- lived subduction zone has swal-
lowed an entire ocean basin, and two continents are set on a
collision course.
How long does it take to raise a mountain belt? In the case
of the Himalaya, the seafloor spreading history recorded by
marine magnetic anomalies makes it possible to track India’s
headlong rush from its place in the ancient southern continent
of Gondwandaland in the late Cretaceous to its current position
as part of Asia.5 Pulled northward by subducting ocean crust,
India traversed about 2500 km (1500 mi) in 30 million years
(an impressive average pace for such a marathon, at more than
74 Ch a pter 3
8 cm (3 in.) a year), before first running into Asia about 55 mil-
lion years ago. Since then, the Himalayan range has risen as the
northern part of India has wedged itself beneath Asia, and the
crust of both continents has also thickened vertically through
faulting and folding. As the convergence continues, a wave of
deformation is propagating outward from the original point of
contact, both north and south, progressively widening the welt
of uplifted and contorted crust.
Before the emergence of plate tectonic theory in the 1960s,
mountain belts were difficult to explain. Many geologists rec-
ognized that the buckled, crumpled strata typical of montane
regions required horizontal compression, but the motive force
behind this was difficult to understand under the prevailing
assumption that continents were fixed in place. A nineteenth-
century Austrian geologist, Eduard Suess (grandfather of Hans,
who would document the dilution of atmospheric 14C by “dead”
carbon from fossil
fuel burning), recognized that many of the
rocks in the Alps had formed on the seafloor and had somehow
been elevated to their present positions. He postulated that
Earth’s mountains were akin to the wrinkles on a raisin, ridges
formed by shrinkage as the result of steady cooling and con-
traction of the planet— a notion consistent with Lord Kelvin’s
views of the thermal evolution of the interior.
The art critic, intellectual polymath, and alpine enthusiast
John Ruskin, a contemporary of Suess’s, also had an intuitive
sense that mountains are not static, eternal monuments but
records of dynamic events. For Ruskin, however, the morphol-
ogy of the Alps evoked liquid fluidity rather than desiccated
fruit: “There is an appearance of action and united movement
in these crested masses, nearly resembling that of sea waves . . .
fantastic yet harmonious curves, governed by some grand
under- sweep like that of a tide running through the whole body
The pace of the earth 75
of the mountain chain.”6 He also recognized that that these
“harmonious” shapes represented the countervailing effects
of the “elevatory force in the mountain” and “sculptural force
of water upon the mountain.” But how efficiently do these op-
posing forces act?
The highest Himalayan peaks, at elevations of 9000 m
(29,000 ft), lie where the coast once was. So it might seem log-
ical to estimate their growth rate simply by dividing their height
by 55 million years, which yields a positively underwhelming
uplift rate of 0.015 cm (0.006 in.) per year. But this calcula-
tion is a gross underestimate of the actual rate of mountain
building— because as soon as tectonic forces start constructing
mountains, the highly efficient erosion crew arrives to begin
demolition. So we need to find ways to measure these opposing
processes independently.
Today, surface uplift can be measured in nearly real time
thanks to high- precision global positioning system (GPS) satel-
lites. In the highest part of the Himalaya, the Tibetan Plateau,
GPS- based uplift rates averaged over a decade are in the range
of 2 mm (0.1 in.) per year. This is about an order of magnitude
slower than the tectonic convergence rate (around 2 cm, or
1 in., per year)7 and reflects a fairly typical ratio of vertical to
horizontal deformation in the crust. But the instrumentally
measured uplift is more than 100 times faster than a long- term
estimate that ignores the effects of erosion. How can we know
whether modern satellite- based estimates are representative of
uplift rates over longer periods of geologic time? As the “roof
of the world” is rising, its top stories are constantly being re-
moved, in a process geologists call exhumation. What were
once the subterranean levels now have high penthouse views.
To reconstruct long- term uplift rates, we need to know how
many floors have been dismantled, and how quickly.
76 Ch a pter 3
There are several ways to calculate how much additional rock
once existed in the airspace above a mountain belt. One is to ask
the rocks now at the surface how deep they used to be at a certain
time in the past. This can be done using a technique called fission
track dating, developed largely by oil companies to reconstruct
the thermal histories of sedimentary rocks to predict whether
they are likely to produce petroleum or natural gas (sediments
need to have been warm enough that their organic matter got
properly “cooked,” but never so hot that it all burned off).
Fission track dating makes use of the fact that the more abun-
dant isotope of uranium, 238U, is not just radioactive but also
has an unstable nucleus that splits itself in spontaneous fission
events at a known rate. Under high magnification, uranium-
bearing minerals including zircon (the darling of geochronol-
ogy) and apatite (the mineral in teeth and bones) retain a vis-
ible record of these high- energy events in the form of damage
zones or “fission tracks.” Each uranium- bearing mineral has
a particular temperature above which the crystal lattice can
heal itself and erase these scars, like an Etch- a- Sketch that has
been well shaken. Below this temperature, however, the tracks
will remain etched in the crystal. So, by counting the density
of fission tracks in a given volume of a mineral, it is possible to
determine how long it has been since it cooled through a certain
temperature (and depth) in the crust. Fission track thermo-
chronology for Himalayan rocks shows that modern uplift rates
based on a few decades of satellite data are in fact consistent
with uplift over geologic timescales.8
R E M A I N S O F T H E D AY
Another approach to estimating how much has been trimmed
from mountains by erosion is to look at the volumes of sediment
The pace of the earth 77
that have accumulated at their feet, like snippets of hair on a
barbershop floor. In the Himalaya, most of the erosional detritus
has accumulated in two gigantic sandpiles on the seafloor: the
Indus and Bengal submarine “fans,” where the Indus, Ganges,
and Brahmaputra Rivers have dumped sediment for the past
50 million years. On Marie Tharp’s maps, the Indus and Bengal
fans are long tongues lolling far out onto the floor of the In-
dian Ocean. The Bengal fan is the largest in the world. From the
mouth of the combined Ganges- Brahmaputra on the coast of
Bangladesh— which itself is made entirely of sediment shed from
the mountains— the fan extends 3000 km (1800 mi) southward.
If superimposed on the continental United States, the Bengal
fan would stretch from the Canadian border to Mexico, and for
almost half that distance, it is more than 6.5 km (4 mi) thick.
Drilling and geophysical exploration of the Indus9 and Ben-
gal10 fans have revealed an upside- down, impressionistic re-
cord of the unroofing of the Himalaya, with the disaggregated
remains of rocks that were at the top of the mountains in their
infancy now forming the lowest layers in the immense mass of
deep- sea sediment. The total volume of the Bengal fan alone,
an estimated 12.5 million km3 (3 million mi3),11 is greater than
the present- day volume of the crust of the Tibetan Plateau
above sea level.12 That is, more rock has been removed from the
Himalaya by erosion than forms the towering modern range.
This fact makes the seemingly simple question posed by Hut-
ton (and Dylan)— How long does it take to wear mountains
down?— more difficult to answer. Which mountains are we
talking about? The Himalaya have existed for 55 million years,
but today’s mountains are not the same as the mountains whose
ruins lie on the floor of the Indian Ocean.
The ephemeral nature of mountains— or any landscape— is
one reason that unconformities in the rock record, like Hutton’s
78 Ch a pter 3
famous outcrop at Siccar Point, hold such fascination
. Un-
conformities preserve buried topography and thus provide
glimpses of the long- vanished terrain of earlier eons. Wiscon-
sin’s Baraboo Hills region, a mecca for geology field trips (and
home of the late Ringling Bros. and Barnum & Bailey Circus,
the “Greatest Show on Earth”), represents one of the most
remarkable examples of paleotopographic preservation any-
where in the world. This Precambrian mountain belt, formed
about 1.6 billion years ago, was buried by hundreds of feet
of marine sediments when early Paleozoic seas washed over
what is now the Great Lakes region. Today, erosion of these
Paleozoic rocks has reached a stage in which the unconformity
between the Precambrian and Paleozoic worlds is exposed in
many places. The long- hidden mountains are being unburied,
or exhumed, and the modern land surface closely approximates
that of the late Proterozoic. Interestingly, this ancient landscape
was the inspiration for two great environmental thinkers: John
Muir, whose family immigrated to the area from Scotland when
he was a young boy, and Aldo Leopold, whose Sand County
Almanac is set in the shadow of the primordial Baraboo Hills.
There are much older rocks, and deeply eroded roots of older
mountain belts in other places (even Wisconsin), but the Bara-
boo Range represents some of the oldest preserved topography
on Earth— a Great Show indeed.
T H E H I L L S A R E A L I V E
The sediments shed from the Himalaya tel us that while there has
been some variation in uplift and exhumation rates over time,
on average these rates fall within the range of estimates from
both GPS observations and thermochronologic approaches
like fission track dating. This is a comfortingly uniformitarian
The pace of the earth 79
result; Lyell would be pleased. The great heaps of sediment also
underscore an amazing fact about the Earth: that the speeds of
tectonic processes, driven by the internal radioactive heat of
the Earth are, by happy coincidence, about evenly matched13
by the tempo of external agents of erosion— wind, rain, rivers,
glaciers— powered by gravity and solar energy. In the barber-
shop analogy, it is as if the hair on a customer’s head keeps
growing as fast as the barber can cut it. And while the tectonic
growth and erosional trimming of mountains both proceed at
Timefulness Page 9