Page 141 - Global Tectonics

OCEAN RIDGES 127

(iii) a temperature-dependent phase change. The high
temperatures beneath ocean ridge crests might cause a
transition to a mineralogy of lower density.
Suppose the average temperature to a depth of
100 km below the Moho is 500°C greater at the ridge

crest than beneath the flanking regions, the average
−3
density to this depth is 3.3 Mg m and the volume
−5

coefficient of thermal expansion is 3 × 10 per degree.
In this case the average mantle density to a depth of
−3
100 km would be 0.05 Mg m less than that of the

flanking ocean basins. If isostatic equilibrium were
attained, this low-density region would support a

ridge elevated 2.2 km above the flanking areas. If the
degree of partial melting were 1%, the consequent
−3
decrease in density would be about 0.006 Mg m .
Extended over a depth range of 100 km this density
contrast would support a relative ridge elevation of
Fig. 6.7 Alternative model of the structure beneath the 0.25 km. The aluminous minerals within the upper

Mid-Atlantic Ridge from gravity modeling. Profile at mantle that might transform to a lower density phase
−3
46°N. Densities in Mg m (redrawn from Keen & are also the minerals that enter the melt that forms
Tramontini, 1970, with permission from Blackwell beneath the ridge crest. They are absent therefore in
Publishing). the bulk of the mantle volume under consideration,
which consists of depleted mantle; mantle from which
the lowest melting point fraction has been removed.
It is unlikely then that a phase change contributes
members of a suite of possible interpretations. They significantly to the uplift.

demonstrate without ambiguity, however, that ridges Partial melting of the upper mantle clearly is a
are underlain by large, low-density bodies in the upper reality because of the magmatic activity at ridge crests,
mantle whose upper surfaces slope away from the ridge but its extent was a matter of conjecture. However, in
crests. the mid-1990s a very large-scale experiment, the Mantle
Electromagnetic and Tomography (MELT) experiment,
was carried out on the crest of the East Pacifi c Rise

specifically to define the vertical and lateral extent of

6.3 ORIGIN OF the region of partial melting beneath it (MELT seismic
team, 1998). Fifty-one ocean bottom seismometers and
ANOMALOUS UPPER 47 instruments that measure changes in the Earth’s

magnetic and electric fields were deployed across the
MANTLE BENEATH ridge, between 15° and 18°S, in two linear arrays each
approximately 800 km long. This location was chosen
RIDGES because it is in the middle of a long, straight section of

the ridge between the Nazca and Pacific plates, and has
−1
one of the fastest spreading rates: 146 mm a at 17°S.
The extent of any partial melt in the mantle should
There are three possible sources of the low-density therefore be well developed in terms of low seismic
regions which underlie ocean ridges and support them velocities and high electrical conductivity. Seismic waves
isostatically (Bott, 1982): (i) thermal expansion of upper from regional and teleseismic earthquakes, and varia-
mantle material beneath the ridge crests, followed by tions in the Earth’s electric and magnetic fi elds, were

contraction as sea floor spreading carries it laterally recorded for a period of approximately 6 months. Anal-
away from the source of heat, (ii) the presence of ysis of the data revealed an asymmetric region of low
molten material within the anomalous mantle, seismic velocities extending to a depth of 100 km, with