Page 9 - i1052-5173-26-9
P. 9

between Rb/Sr and crustal thickness is observed in subduction         remain difficult to test, and the upwellings may have been more        GSA TODAY | www.geosociety.org/gsatoday
settings (Dhuime et al., 2015), it implies that crustal thickness     localized and likely originated at shallower levels, perhaps as in a
may be a reasonable proxy for crustal volume at least over the        heat-pipe model (Moore and Webb, 2013). In today’s world, such
past 2.0 b.y.                                                         magmatism would be regarded as intraplate; it has no association
                                                                      with subduction, but in a time before plates, such terminology is
  Figure 4B (dotted blue curve) illustrates such a model for the      arguably misleading. Destruction and recycling of early crust
changes in crustal volumes through time. It is constrained by the     occurred through a combination of delamination (Johnson et al.,
present volume of the continental crust, the presence of 70% of       2013) and meteorite impact, the latter continuing until the end of
that volume at 3.0 Ga, and the present-day distribution of crust      the late heavy bombardment ca. 3.9 Ga (Gomes et al., 2005;
with different model ages (Condie and Aster, 2010). It starts at the  Marchi et al., 2014), and so perhaps we are fortunate to have the
end of the heavy bombardment ca. 3.9 Ga; it assumes that the rate     few zircon grains that have survived from the Hadean.
of crust generation was greater before 3.0 Ga than subsequently,
and that the compositions of the crust generated before and after       The onset of stage 3 is taken to be the stabilization of Archean
3.0 Ga were different. Pre–3.0 Ga crust was preferentially recycled   cratons and the change in crustal growth rates ca. 3.0 Ga. It is
from 3.0 to 1.5 Ga following the onset of subduction at 3.0 Ga.       envisaged that Earth had cooled sufficiently for “hot” subduction
Such models are not unique, but they indicate that changing           to take place (Sizova et al., 2010) from ca. 3.0 Ga to ca. 1.7 Ga
volumes of continental crust can be modeled with two types of         (Fig. 4). The continental crust became thicker and more evolved,
crust that are destroyed at different rates. The model outlined here  plate tectonics resulted in collisional orogenies, supercontinent
predicts that nearly twice the present volume of the continental      cycles developed, and there was increased erosion to the oceans
crust has been recycled since the end of the heavy bombardment.       due to thickening and subaerial exposure of continental crust
                                                                      resulting in an increase in Sr isotope ratios in seawater (Fig. 1B).
  The increase in thickness at the sites of crust generation from     Models in which ~70% of the present volume of the continental
3.0 to 1.7 Ga indicates that more of the continental crust was        crust was present at 3.0 Ga also require that large volumes
elevated above seawater (Flament et al., 2013), and hence suscep-     of continental crust have been destroyed, presumably in the
tible to sub-aerial weathering and erosion. This increase in crustal  late Archean and the Proterozoic (Hawkesworth et al., 2013).
thickness is accompanied by a steady increase in the Sr isotope       However, the net growth of crust requires the rate of crust genera-
ratios of seawater (Fig. 1; Shields and Veizer, 2002), implying       tion to exceed that of recycling (Fig. 4).
increased continental runoff, which would increase the amounts
of CO2 draw down due to continental weathering (e.g., Kramers,          The fourth stage is from 1.7 to 0.75 Ga, referred to as the
2002). Preliminary models indicate that in the mid-Proterozoic        “boring billion” (Holland, 2006), and more recently as Earth’s
the volume of continental crust may have been up to 20% greater       middle age (Cawood and Hawkesworth, 2014). It is marked by a
than at the present day, and since 1.0 Ga the crust has become        paucity of preserved passive margins, an absence of significant
thinner, and we infer that the crust decreased in volume (Fig. 4).    anomalies in the paleoseawater Sr isotope record and in Hf
The rates of crustal growth appear to have decreased as Earth         isotopes in detrital zircon (Fig. 1), a lack of orogenic gold and
cooled, and “cool” subduction began to dominate, which, on the        volcanic-hosted massive sulfide deposits, and an absence of glacial
basis of preserved metamorphic and other rock units, is taken to      deposits and iron formations (Cawood and Hawkesworth, 2015).
have commenced around 0.75 Ga (Brown, 2006; Cawood and                It appears to have been a period of environmental, evolutionary,
Hawkesworth, 2014), such that the rates of crustal recycling were     and lithospheric stability, which has been attributed to a relatively
greater than the rates at which new crust was generated.              stable continental assemblage that was initiated during assembly
                                                                      of the Nuna supercontinent by ca. 1.7 Ga and continued until
DISCUSSION                                                            breakup of its closely related successor, Rodinia, ca. 0.75 Ga. It is
                                                                      also marked by abundant anorthosites and related rocks perhaps
  There appears to have been five stages in Earth’s evolution, with   linked with the secular cooling of the mantle. The overlying
the last four being recorded in the geology of the continental crust  continental lithosphere was strong enough to be thickened and to
(Fig. 4). Stage 1 included the initial accretion of the Earth, core/  support the emplacement of large plutons into the crust, yet the
mantle differentiation, the development of a magma ocean, and of      underlying mantle was still warm enough to result in widespread
an undifferentiated mafic crust. Most models suggest that a           melting of the lower thickened crust and the generation of anor-
magma ocean may have persisted on Earth for 5–10 m.y., and            thositic magmas (Ashwal, 2010).
continuing volcanism, along with deformation, would have
progressively thickened the initial mafic crust (Kamber, 2015;          The termination of Earth’s middle age, and onset of stage 5,
Kamber et al., 2005). Once the crust was at least 15–20 km thick,     corresponds with Rodinia breakup at 0.75 Ga and the develop-
remelting could take place (Fig. 2; Kamber et al., 2005), and the     ment of “cold” subduction. The latter is recognized by the onset of
resultant felsic magmas represented the high silica component in      high- to ultrahigh-pressure metamorphic rocks (Brown, 2006).
the distinctive bimodal silica distribution that characterizes the    Falling mantle temperatures enabled deeper levels of slab breakoff
Archean crust (Fig. 3). This second stage was marked by elevated      in collision zones and the resultant greater depths to which conti-
mantle temperatures compared to the present day (Fig. 4C) that        nental crust was subducted prior to exhumation, allowing the
resulted in lithosphere weakened by the emplacement of melts          development of the ultrahigh-pressure metamorphic assemblages
(Gerya, 2014; Sizova et al., 2010). This inhibited subduction, and    (Brown, 2006; Sizova et al., 2014). Stage 5 is marked by a strongly
hence plate tectonics, and magmatism were driven by mantle            episodic distribution of ages linked to the supercontinent cycles of
upwellings that percolated the lithosphere. These might have been     Gondwana and Pangea (Fig. 1). Oxygen levels in both the atmo-
associated with deep-seated mantle plumes, but such models            sphere and deep oceans increased, phosphate and evaporate deposits

                                                                                                                                             9
   4   5   6   7   8   9   10   11   12   13   14