Analysis of travel paths taken by many S waves that travelled beneath the African continent, largely by geophysicists at the California Institute of Technology, shows that beneath it is a large zone of anomalously low wave speeds. Part of the zone dips down obliquely from the rough location at shallow depths of the Afar plume beneath Ethiopia/Yemen to the core-mantle boundary between the surface locations of Africa and South America. The structure is well placed for seismic tomography, by virtue of its good match with useful earthquakes and the world-wide network of seismometers. More advanced analysis (Ni, S et al. 2002. Sharp sides to the African superplume. Science, v. 296, p. 1850-1862) shows up a strangely sharp-sided part of the plume that rises from the core-mantle boundary for about 1500 km below southern Africa. There its boundary with more normal mantle is little more than 50 km wide. Modelling suggests that the upward flow has caught up a dense layer with possibly different chemistry, which would result in a tilt towards the direction of movement so that instead of rising vertically, the plume would have an oblique trajectory. The tilt also fits with Africa’s north-eastwards drift (in an absolute frame of reference, relative to other hotspots) since 100 Ma ago.
Whatever its origin, a rising, hot mantle zone beneath Africa is consistent with the continent’s high overall topography, which has encouraged the lithosphere to rift. This extension has resulted in the East African Rift, which further encouraged partial melting in the underlying mantle and the resulting volcanism. By far the most important aspect of Africa’s recent volcanic activity has been the Eocene to Oligocene flood-basalt event of the Ethiopian Plateau and the current activity in the Afar part of the Rift.
Subduction metamorphism and earthquakes
The recently commissioned Hi-net array of 600 digital seismometers in Japan paid dividends in an unexpected way during 2001, by picking up long-lived vibrations rather than discrete seismic events Obara, K. 2002. Nonvolcanic seep tremor associated with subduction in southwest Japan. Science, v. 296, p. 1679-1681). The tremors occurred in a part of Japan where there are no active volcanoes, with which protracted vibrations are usually associated. Their epicentres define a clear zone, at about the depth of the Moho and on the Wadati-Benioff zone where the Philippine Plate is being subducted. This region is where dehydration reactions that convert cold, wet oceanic crust to dense eclogite, the driving force for plate tectonics through slab pull, are predicted to occur by thermodynamics. Kazushige Obara, of Japan’s National Research Institute for Earth Science and Disaster Prevention, suggests that this correlation might fit with the release and rise of hydrothermal fluids released by dehydration of the slab. Part of his evidence is that such tremors seem not to occur where the much older (and therefore cooler) Pacific Plate is being subducted beneath the northwest of Japan. It probably does not undergo such reactions until it has reached about 100 km depth, where temperature would be sufficient to enter the field of eclogite stability. Detecting fluid motion at 3 times the depth of that beneath southwest Japan might emerge with more specialized procesing.
See also: Julian, B. 2002. Seismological detection of slab metamorphism. Science, v. 296, p. 1625-1626.
Crustal shortening and thickening in collisional orogeny produces mountain belts with a root of crust beneath them. This truism is central to isostasy, where the mass of uplifted mountains is balanced by a compensating mass of low-density root material beneath that penetrates the mantle lithosphere. The classic story of the reduction of mountain belts to a peneplain involves continuous isostatic uplift as the topography is eroded away. Finally, no root remains and the exposed rocks reflect in their high-grade metamorphism a steady upward passage from the root. Later cover rests with profound unconformity upon this peneplain. Yet this essentially simple theory does not hold in many cases, especially for older collisional orogens. As Karen Fischer of Brown University, USA has shown (Fischer, K.M. 2002. Waning buoyancy in the crustal roots of old mountains. Nature, v. 417, p. 933-936), there is a crude correlation between the age of orogens and their ratio of elevation to root thickness. The ratio decreases from around 0.15 (root about 7 times thicker than surface elevation) in active orogens to zero before 1 Ga ago, when peneplained orogens still have a substantial root.
In order for this to happen, either the roots’ buoyancy must somehow decline with age or the mantle lithosphere which it penetrates becomes too rigid to allow isostatic uplift to occur. Resolving which has most effect depends on analysing the gravity anomalies above orogens. It is no easy task to model the two processes, and this is what Fischer has achieved. She finds that mantle viscosity is not responsible, and that the cause is variation in root density. This is probably a result of slow decline in heat flow, and the resulting mineralogical equilibria in the root. For mafic granulite roots, a change from heat flow values of 70mWm-2 to around 40 mWm-2 could increase their density by 100-150 kg m-3, by an increase in the proportion of garnet, perhaps to the extent of producing eclogites at the deepest levels. Eclogites would be seismically very similar to mantle lithosphere, so that even thicker, hidden roots may be present. Reduction in buoyancy by this means could take as little as 20 Ma, before which the elevation to root thickness ratio has declined below that in active orogens.
One implication of this process is that orogenic collapse by lateral extension of highly elevated crust, which might lead to rapid root thinning, is not the general process that many structural geologists believe. If it was, orogenic roots would be removed relatively quickly. Decrease in root buoyancy is also a plausible explanation for the creation of cratons, where quite low-grade metamorphic rocks, formed at shallow crustal levels occupy vast areas of low-lying shields.