Page 411 - Global Tectonics

THE MECHANISM OF PLATE TECTONICS 393

implies that mantle drag inhibits the temperature and density that are a consequence of con-
motion of such plates rather than driving vection. Also, by mapping seismic anisotropy both ver-
them. tically and laterally it is possible to obtain estimates of
the direction of mantle fl ow.
The mechanism also provides a reasonable explana-

The first three-dimensional seismic velocity models
tion of the motions of small plates.
for the mantle derived by the tomographic technique
Consequently, the edge-force mechanism of plate
were published in the early 1980s (Woodhouse &
movement appears to be much more successful in
Dziewonski, 1984). Since then there have been great
explaining all observed phenomena, and has been
improvements in data quality, geographic coverage, and
adopted by most workers, certainly for present-day
processing techniques, and the resolution of subse-
plate motions.
quent models has greatly improved. However, many
of the essential features of the velocity variations were
apparent in the earliest models. Plate 12.1 (between
12.8 EVIDENCE FOR pp. 244 and 245) shows the variations in the shear
wave velocity at 12 depths in the mantle according to
CONVECTION IN THE model S16B30 of Masters et al. (1996). It is immediately
apparent that the greatest variations occur near the
MANTLE top and bottom of the mantle, presumably within or
in the vicinity of the thermal boundary layers. Within
the top 200 km the perturbations are very closely related
to surface tectonic features. Ocean ridges, the rifts of
12.8.1 Introduction northeast Africa, and the active backarc basins of the

western Pacific are all underlain by anomalously low
velocity mantle. Continental areas in general, and shield
A fundamental axiom of plate tectonics is that oceanic
areas in particular, are underlain by the highest veloc-
lithosphere is formed from mantle material at mid-
ities, and older oceanic crust by relatively high veloc-
ocean ridge crests and returned to the mantle in sub-

ities. These variations essentially reflect the different
duction zones. Thus, plate creation, movement and
thermal gradients and hence the thickness of the
destruction provide evidence for convection in the
lithosphere in these areas (Section 11.3.1). Between
mantle. There must be downwellings in the mantle
200 and 400 km most of these generalizations still apply
associated with subduction zones and upwelling beneath
but the velocity contrasts are lower. A notable excep-
mid-ocean ridge crests. However, beyond this, plate
tion is the mantle beneath the backarc basins where
tectonics provides no evidence for the location of the
the slow velocities at shallower depths have been
return flow in the mantle, or the depth extent of con-

replaced by near zero anomalies. In the transition zone
vection other than the seismicity associated with sub-
(e.g. 530 km) the variations are in general quite small
ducting slabs (Section 9.4). One must turn therefore to
and the correlation with surface features has largely
other lines of evidence for information on the pattern
broken down. Again an exception is the mantle beneath
of convection in the deep mantle.
the backarc basins of the western Pacific, which, at

this depth, is characterized by high velocities presum-
ably associated with cold subducted lithosphere. In
12.8.2 Seismic tomography the lower mantle (depths greater than 660 km) the
variations in shear wave velocity are generally quite
Much important information on the three-dimensional small (less than ±1.5%), but a persistent feature is a
structure of the mantle has been supplied by seismic ring of higher than average velocities beneath the rim

tomography (Section 2.1.8). Convection is driven by of the Pacific. This becomes particularly marked in
lateral differences in temperature and density. These the lowest 400 km of the mantle (e.g. depths of 2500
variables affect seismic velocity, which typically decreases and 2750 km). At depths greater than 2000 km, large
with decreasing density and increasing temperature regions of anomalously low velocity occur beneath

(Dziewonski & Anderson, 1984). By mapping velocities the central Pacific, and beneath southern Africa and
in the mantle it is possible to infer the differences in part of the South Atlantic.