Geology Reference
In-Depth Information
model, which would have contained about two-thirds of the heat sources. The heat
budget for this model is shown in Figure 8.7(c). By this estimate, the hotspot
topography should be comparable to the mid-ocean ridge topography in extent and
volume. Clearly it is not. This argument was made in 1988 [54], though it did
not seem to attract much notice, perhaps because it requires a little explanation,
whereas tomographic images require little explanation.
Thus the topography of the sea floor not only tells us about the heat transported
by the two main modes of mantle convection, but also implies a strong constraint on
the structure of mantle convection. The relative smallness of hotspot swells, and the
absence of any other topography that can be interpreted as due to buoyant mantle
upwellings, implies that only a modest amount of heat can come from below
the source of mantle plumes. Because plumes would be generated above any
interface within the mantle, not much heat can come from below that interface.
Thus any deep layer containing a substantial fraction of Earth's radioactive heat
sources is precluded. This conclusion creates a major difficulty in the conventional
view of mantle geochemistry. The potential resolution of that difficulty will be
taken up in Chapter 10.
8.4 Other modes and causes
Although a fairly clear picture has been painted by the foregoing arguments, there
are of course still debates about the mantle convection system. These tend to focus
on specifics and details. For example, our knowledge of plate motions for the past
100 Myr or so allows direct modelling of the pattern of mantle flow [119], with
application to rather specific questions. This topic is intended to convey a basic
general understanding, so such topics are not pursued here. However, some issues
have played large roles in past debates, and some point to significant lessons, so a
few will be reviewed here.
8.4.1 Rifting model of flood basalts
White and McKenzie [72] proposed a theory for the formation both of very thick
sequences of volcanic flows found along some continental margins and of flood
basalt eruptions. The theory can usefully be separated into three parts. The first
part is that the marginal volcanic provinces are produced when rifting occurs over a
region of mantle that is hotter than normal because it is derived from a plume. This
seems to give a very viable account of such provinces, and it has been quantified
successfully using a plume head model [77]. The second part is that all flood basalts
can be explained by this mechanism. The third part is that the plume material is
