Mantle Mixing: Processes and Modeling - Physics and Chemistry of the Deep Earth - page 353

50

British Columbia
Mexico

70

Cas cadia

90

110

N Peru gap

130

C Peru gap

Izu

New Zealand Nankai

Solomon

150

Colombia-Ecuador

C Chile gap

C Sumatra

N-C Chile S Chile

N Sumatra

170

Nicaragua

Alaska

Scotia

S-C Chile

Bonin

N Chile

New Britain

190

C Chile

N Kurile
S Kurile

Costa Rica

E Banda Sea

Kamchatka
Bali-Lombok
N Philippines
Hokkaido
W Banda Sea
Sunda Strait

Peru

AK Peninsula

Java

S Marianas
S Lesser Antilles
Aegean
N Lesser Antilles
N Marianas

210

N Vanuatu
S Vanuatu
Calabria
W Aleutians

Ryukyu
Kyushu

C Honshu

N Honshu

Guatemala-El Salvador

Tonga

S Sumatra
S Philippines

Kermadec

E Aleutians

230

C Aleutians

250 0

5 10 15

slab H 2 O loss (Tg/Myr/m)

20 25 30

Fig. 12.3 (Modified from van Keken et al ., 2011). The amount of water loss in Tg/Myr per m along strike in 56
segments of the world's subduction zones is plotted as a function of depth. These estimates are based on high
resolution dynamical models of these subduction zones and a prediction for water loss using phase diagrams for the
lithologies in the subducting crust and uppermost mantle. Warm subduction zones (e.g., Cascadia) lose significant
amounts of water in the forearc. All subduction zones see significant water loss when the slab gets into contact
with the hot mantle wedge (assumed to be at 80 km depth here). Water loss at larger depths is diverse. Slabs do not
lose all of the water they carry down and the present day subduction system brings significant amount of water into
the deeper mantle. Reproduced with permission of the American Geophysical Union.

and the characteristics of EM-II (e.g., Taylor &
McLennan, 1985; Rudnick, 1995; Kelemen et al .,
2003; Plank, 2005). Korenaga (2004) suggested
that the continental breakup mechanism itself
can explain the hotspot diversity. While the
Earth's accretion, large scale differentiation and
moon formation are generally considered to
have happened before the ''initial condition'' of
mantle mixing, there may be prolonged effects
of a magma ocean, particularly if a deep magma
ocean as suggested by Labrosse et al . (2007) was
present in the early Earth.
A fundamental shift in lithospheric recycling
likely happened in the mid- to late Archean.

While it is not clear when plate tectonics started
on Earth, there is a growing body of evidence
from Archean geology, petrology and geochem-
istry that the tectonics of the Early Earth may
have been dominated by plume-driven or de-
lamination tectonics (e.g., Zegers & van Keken,
2001; B edard, 2006; Shirey & Richardson, 2011).
The transition to modern plate tectonics is esti-
mated to have taken place near the late-Archean
(Smithies et al ., 2005; Shirey & Richardson, 2011).
Early plate tectonics likely was dominated by rel-
atively shallow and weak subduction (e.g., van
Hunen & van den Berg, 2008) and without recy-
cling significant amounts of water to the deep