Geology Reference
In-Depth Information
(a)
(b)
14.5
1800
500
300
400
300
1700
14
200
1600
400
13.5
200
300
1500
500
200
13
1400
1300
12.5
0
1
2
3
4
0
1
2
3
4
Time, Gyr
Time, Gyr
Figure 9.4. Thermal evolutions in which resistance to plate bending at subduction
zones limits plate velocity. Curves are labelled with the assumed bending radius
of the plate. (a) Temperature. (b) Heat loss (solid) and heat generation (dotted).
From Davies [146]. Copyright Elsevier Science. Reprinted with permission.
greatest thickness, due to the greater melting and dehydration. Thereafter the heat
loss slowly increases and heat inputs decline. At about 300 Myr in this particular
solution the dehydration thickness becomes less than the thermally defined thick-
ness of a plate, and the solution briefly reverts to conventional behaviour, with the
heat loss dropping rapidly as the mantle cools further.
This is an innovative and interesting evolution, but it depends on assumptions
about several phenomena that are not very well constrained, including melting
depths and dehydration at high mantle temperatures, and the radius of curvature
of subducting plates. It turns out to be very sensitive to the bending radius of
plates. Observations of present plates indicate bending radii ranging from 120 km
to 1200 km, with most between about 160 km and 600 km [150]. The effect of
assuming larger radii is shown in Figure 9.4. With radii of 400-500 km, the bending
resistance still dominates early in Earth history, but then it falls below the viscous
resistance of thermal plates and the solution reverts to the behaviour of conventional
models. The present Urey ratio is 0.8 and the cooling rate is about 30
◦
C/Gyr for
both of these solutions. The '300' solution is transitional, but even in this case the
present Urey ratio is 0.67, well above the value required to reconcile long-term
heat loss and heat generation. Temperatures in this solution exceed 1800
◦
Cinthe
Archaean, which surely is excessively high.
Thus this proposed reconciliation of heat generation and heat loss seems to
require bending radii that are close to the extreme observed today, and that would
need to have applied even in the past when the plate thickness, defined by melt-
ing and dehydration, would have been even greater. The implausibility of this is
increased by the modelling of Capitanio
et al.
[151], which shows that plates tend