Geology Reference
In-Depth Information
There are other plausible sources of low-amplitude topography. Plumes are an
obvious source, but they can usually be identified fairly clearly. There are also deep
thermochemical 'superpiles' in the lower mantle, one under Africa and one under
the central Pacific [102-104]. Simmons
et al.
[104] infer that the relatively high
elevation of Africa is due to the pile under Africa.
There is a broad positive anomaly stretching from the central Pacific to the
Japan Trench. This anomaly has contributed to the impression of flattening, but its
regional character can be seen in Figure 8.4. It was called the Pacific Superswell by
Davies and Pribac [91] and interpreted as a residue of the Mesozoic Darwin Rise
inferred by Menard [105]. The mantle structure envisaged by Davies and Pribac is
illustrated by Figure 8.1.
8.3 Layered convection?
Perhaps the most persistent and sometimes heated debate about mantle convection
has been whether the mantle convects in two separate layers or as a single, whole-
mantle layer. The initial conception was that the upper mantle, above 660 km depth,
convects separately from the lower mantle [106, 107]. As evidence against this
picture emerged, a deeper layer, within about 1000 km of the core, was proposed
[108], but there are also strong arguments against such a layer. The only layer for
which there is clear evidence is the so-called D
zone within the lowest 200-300 km
of the mantle [109, 110].
The original notion of layered mantle convection grew out of the idea, mentioned
earlier, that the lower mantle is so viscous as to be immobile, so in the late 1960s
convection was presumed to be confined to the upper mantle [111]. When the lower
mantle turned out to have a moderate viscosity [46], as presented in Chapter 4, it
was presumed by the mid-1970s to convect separately [106]. There were no strong
reasons for this presumption [112], but one suggestive reason was that earthquakes
in the Wadati-Benioff deep earthquake zones under ocean trenches are confined to
the upper mantle. This was interpreted to mean that the plate-related flow was also
confined to the upper mantle. In fact, most of the deep zones give no indication
of deflecting horizontally and are consistent with flow continuing into the lower
mantle [1]. The cessation of earthquakes can be plausibly attributed to the phase
transformations that occur at this depth. The geophysical case for whole-mantle
convection was being argued by 1977 [112, 113].
When, in 1976, some of the first measurements of neodymium isotopes were
interpreted to indicate a primitive zone in the mantle [114], the layered picture
was invoked, the lower mantle being presumed to be geochemically primitive,
and chemically isolated from the upper mantle [107]. Soon it became evident that
