1
Growth, destruction, and
preservation of Earth’s continental
crust
C.J. Spencer
1
, N.M.W. Roberts
2
, M. Santosh
3,4,5
1
Earth Dynamics Research Group, Department of Applied Geology, The Institute for
Geoscience Research (TIGeR), Curtin University, Perth, Australia
2
NERC Isotope Geosciences Laboratory, British Geological Survey, Keyworth, Nottingham,
UK
3
School of Earth Sciences and Resources, China University of Geosciences, Beijing, China
4
Department of Earth Sciences, University of Adelaide, Adelaide, Australia
5
Faculty of Science, Kochi University, Kochi, Japan
Abstract
From the scant Hadean records of the Jack Hills to Cenozoic supervolcanoes, the continental
crust provides a synoptic view deep into Earth history. However, the information is
fragmented, as large volumes of continental crust have been recycled back into the mantle
by a variety of processes. The preserved crustal record is the balance between the volume of
crust generated by magmatic processes and the volume destroyed through return to the
mantle by tectonic erosion and lower crustal delamination. At present-day, the Earth has
reached near-equilibrium between the amount of crust being generated and that being
returned to the mantle. However, multiple lines of evidence support secular change in
crustal processes through time. Though a variety of isotopic proxies are used to estimate
crustal growth through time, none of those currently utilized are able to quantify the
volumes of crust recycled back into the mantle. This implies the estimates of preserved
continental crust and growth curves derived therefrom represent only a minimum of total
crustal growth. We posit that from the Neoarchean, the probable onset of modern-day style
plate tectonics, there has been no net crustal growth (and perhaps even a net loss) of the
2
continental crust. Deciphering changes from this equilibrium state through geologic time
remains a continual pursuit of crustal evolution studies.
Keywords: Continental crust; Tectonic erosion; Crustal recycling; Secular change; Earth
history
1. Introduction
The processes and rates governing the formation of Earth’s continental crust have been key
questions in Earth science and remain debated today (e.g. Armstrong, 1981; Taylor and
Mclennan, 1985; Rudnick, 1995; Cawood et al., 2013). Continual development in measuring
new isotopic systems and the application of new minerals with greater spatial and analytical
resolution have fuelled the debate surrounding these questions (e.g. DePaolo, 1980; Condie,
1998; Kemp et al., 2007; Condie and Aster, 2013; Dhuime et al., 2015). The debate over
general crust-forming processes have continued despite the availability of tools to look at
the intricacies of crust formation in varying settings, as against the fewer techniques with
poor resolution available in the past (e.g. Harrison, 2009; Reimink et al., 2014; Johnson et al.,
2016; Santosh et al., 2016). Lessons from Phanerozoic plate tectonics have also informed us
about the destruction and recycling of crust (e.g. Kay and Kay, 1991; Huene and Scholl, 1991;
Stern, 2011; Vannucchi et al., 2016), but this aspect of Earth’s evolution has seen less
studied when referring to deep time, and remains a keystone in understanding the volume
of Earth’s continental crust through time. Recently, contributions that bias in the geological
record has made our archive of continental crust to sharp focus (Hawkesworth et al., 2009;
Spencer et al., 2015), although quantification is still lacking.
In this overview, we attempt to cover processes governing the formation, destruction and
preservation of Earth’s continental crust, focussing on the evolution of these processes
through Earth history.
2. Nature and loci of modern crustal growth
3
The Earth is currently comprised of two major silicate reservoirs that are defined broadly by
igneous differentiation. The mantle is generally defined as having undergone no (or minimal)
igneous differentiation and is ultramafic in composition, whereas the crust has undergone
varying degrees of differentiation and is mafic to felsic in composition. For the purposes of
this paper we consider herein the transfer of melt products from the mantle to the crust as
‘growth’. Growth of the continental crust currently takes place primarily by arc, hotspot, rift-
related, and spreading-ridge magmatism (Scholl and von Huene, 2007, 2009; Clift et al.,
2009; Stern, 2011; Cawood et al., 2013). On the modern globe, the composition of the crust
exposed at the surface is broadly bimodal (and corresponds with topography) in that the
oceanic crust is predominately mafic and the continental crust is predominantly
intermediate and felsic in composition (Cawood et al., 2013). The largest volume of magma
produced in the crust occurs at mid-ocean ridges (Jicha and Jagoutz, 2015). As oceanic plates
diverge, decompression melting of the upper mantle produces new mafic oceanic crust. This
results in between ~230 and 325 km
3
km
-1
Myr
-1
or ~155,000,000 to 220,000,000 km
3
of
ocean crust per Myr (after Jicha and Jagoutz, 2015). This newly formed oceanic crust is
subsequently hydrothermally altered by continued mid-ocean ridge magmatism and is
hydrated. Within at most ~200 Myr, the oceanic crust is consumed at subduction zones
where recycling of the oceanic crust into the mantle is driven by negative buoyancy. During
subduction of oceanic crust, the flux of water and other volatiles from the subducting slab
causes melting in the mantle wedge above the subducting slab along with minor
contributions from the subducting slab and sediments (Coats, 1962; Plank and Langmuir,
1993; Currie et al., 2007; Spencer et al., in review). The resulting magma rises through the
mantle and is emplaced into the overlying crust where it forms a magmatic arc marking the
birth of continental crust. Ocean-margin arc magmatism produces the largest volume of
newly formed continental crust (Scholl and von Huene, 2009 and references therein).
Estimates of global additions to the continental crust vary among different workers. It is
estimated that arc magmatism accounts for ~60% to 80% of continental additions (Clift et
al., 2009; Stern and Scholl, 2010; Cawood et al., 2013) with oceanic and continental large
igneous provinces and hotspot/rift-related magmatism accounting for the remainder. Stern
and Scholl (2010) further distinguish the balance of continental crust produced to be ~30%
higher in oceanic arcs than in continental arcs. From a suite of representative oceanic arcs,
4
Jicha and Jagoutz (2015), estimate arc productivity between 157 and 290 km
3
km
-1
Myr
-1
or
~5,000,000 to 10,500,000 km
3
per Myr (assuming oceanic arc length of 17,449; Bird, 2003).
In contrast, arc productivity in the Andes since ~20 Ma is estimated at ~35 km
3
km
-1
Myr
-1
or
~280,000 km
3
per Myr (Haschke and Gunther, 2003 and assuming 8000 km arc length). If the
magma production during this time in the Andes is broadly representative of continental
arcs globally, this equates to ~1,200,000 km
3
per Myr or less than a quarter of the magma
volumes produced in oceanic arcs (Fig. 1).
3. Nature and loci of modern crustal
destruction
In an ironic balancing act, the subduction zone not only is the loci of dominant continental
growth, but also acts as the primary driver for continent destruction. As stated previously,
the magmatic transfer from the mantle to the crust is herein considered ‘growth’ and the
converse is ‘destruction’. In this context, destruction can be misleading as several lines of
evidence support the preservation of crustal reservoirs in the mantle and likely along the
mantle transition zone (Kawai et al., 2013) or the core/mantle boundary (Lay and Garnero,
2011; Zhao et al., 2015; Ma et al., 2016; Garnero et al., 2016). For the purposes of this paper,
we simply use the term destruction to represent vertical transportation of crust into the
mantle. The vast majority of oceanic crust returns to the mantle via subduction processes
with ophiolite and oceanic asperity obduction transferring minor volumes of oceanic crust
onto the continent (<<1%; see Dilek and Furnes, 2011; Furnes and Dilek, 2017). A number of
mechanisms are responsible for the removal of continental crust in subduction zones. We
use the catch-all term ‘tectonic erosion’ to refer to these processes. Tectonic erosion
generally removes the continental crust through bottom-up processes (e.g. basal erosion
and delamination) however, sediment deposited in the accretionary prism or grabens in the
ocean crust can be incorporated into the subduction channel. Due to changes in the
subducting slab angle, basal erosion can remove large volumes of the subcontinental mantle
lithosphere as well as the continental lithosphere (Kay and Mpodozis, 2002; Yamamoto et
al., 2009; Chapman et al., 2016). Basal erosion can also occur due to extension in the
5
subducting oceanic crust wherein horst structures form asperities at the subduction
interface which ‘bulldozes’ continental material from the upper plate into grabens that are
subsequently subducted (Ballance et al., 1989; von Huene et al., 1999; Wells et al., 2003;
Azuma et al., 2016). As the continental crust along a subduction margin thickens, it is
theorized that the root of the arc can gravitationally founder and delaminate into the mantle
(Kay and Kay, 1993; Kay et al., 1994; DeCelles et al., 2009). This is likely due to the transition
of the lower crust to eclogite, which is denser than peridotite (Kay and Kay, 1991; Ducea,
2002; Lee et al., 2006).
The Andean orogeny is one place on the planet where the balance of magmatism and
tectonic erosion is clear. Modern volcanism in the Andes is located between 200-400
kilometers from the trench (Götze et al., 2006). Several studies have shown the clear
eastward migration of the magmatic arc through time (Ramos, 1988; Stern, 1989; Stern,
1991; Scheuber and Reutter, 1992; Atherton and Petford, 1996; Yáñez et al., 2001; Kay et al.,
2005; Ramos and Folguera, 2005). Currently there is late-Paleozoic and early-Mesozoic
mélange exposed within ~50 km of the present coastline of the western continental margin
(Kato, 1985; Bell, 1987; Rebolledo and Charrier, 1994; Willner et al. 2004; Godoy and Lara,
2005; Kato and Godoy, 2015) implying a significant amount of continental lithosphere has
been removed from the continental margin (Scholl and von Huene, 2007).
The buoyancy differential between the continental crust and mantle is traditionally thought
to prevent the deep subduction of continental material, However, the discovery of ultra-high
pressure mineral phases (e.g. coesite and diamond) in metapelitic lithologies confirms the
subduction of continental material to depths greater than 200 km (Chopin and Sobolev,
1995; Ye et al., 2000a; Ye et al., 2000b). Furthermore, the chemical composition of many
intraplate hotspot magmas carries isotopic signatures akin to continental material (Dupre
and Allegre, 1983; Loubet et al., 1988; Eiler et al., 1995). Despite these empirical constraints
of deep subduction of continental material, the theory controlling the subduction of large
continental masses was established by Molnar and Gray (1979). It was postulated that the
gravitational force of the subducting oceanic lithosphere (and its subsequent eclogitization)
might also exert a force on the leading edge of the attached continental lithosphere.
Evidence supporting this hypothesis is seen along the northern margin of Australia where
