Crust delamination

In (i)-(iii), delamination would be unnecessary if primitive cumulates or residues of lower crustal melting were ultramafic, and remained present below the seismic Moho. However, the absence of pyroxenite layers —10 km in thickness in the Talkeetna and Kohistan arc sections suggests that large proportions of ultramafic, igneous rocks are not present at the base of arc crust. 

Thus, all three explanations for the genesis of continental crust require delamination of garnet granulite, eclogite and/or pyroxenite. Moreover, recent, dynamical calculations support the hypothesis that delamination is possible—even likely—where Moho temperatures exceed —750 °C, where crustal thicknesses reach 30 km or more, and where ultramafic cumulates are present (Jull and Kelemen, 2001). The base of arc crust fulfills all of these criteria. Finally, our data on the Talkeetna arc section, and more limited data on the Kohistan arc section, support the hypothesis that substantial proportions of pyroxenite, and perhaps also garnet granulite, were removed by viscous delamination from the base of the arc crust. 

If delamination of dense lower crustal rocks has been essential to continental genesis, and the delaminated rocks constitute —20-40% of the mass of the continents, one might expect to see evidence for this component in magmas derived from the convecting mantle. However, continental crust comprises only —0.5% of the sihcate Earth, so that—even if it represents 40% of the original crustal mass— recycled lower cmst might comprise a very small fraction of the convecting mantle. 

Tatsumi Y. (2000) Continental crust formation by delamination in subduction zones and complementary accumulation of the enriched mantle I component in the mantle. Geochem. Geophys. Geosys. (G-cubed) 1, 1-17. 

The growing consensus that ancient crust is always underlain by ancient lithospheric mantle has been challenged recently by Wu et al. (2003). These authors report an unusual absence of xenoliths with Archaean ages beneath the Archaean North China Craton. They propose that in some cases, therefore, ancient subcontinental mantle can be removed from beneath ancient continental crust by delamination - a process which has previously been postulated but never demonstrated. The subject of mantle delamination is discussed more fully in Chapter 5, Section 5.5.2. 

An additional, unquantified contribution is also made by the delamination of the lower crust and the subcontinental lithosphere (Chapter 4, Section 4.5.2.1). 

Delaminated continental crust and SCLM 1.3% Assume delaminated blocks comprise 50% crust, 50% SCLM. Located in the upper mantle on buoyancy grounds... 

Similar observations have been made in the younger crust of the Sierra Nevada batholith in the USA, where Ducea and Saleeby (1998) proposed that the very thin granite crust (30-40 km), underlain by peridotitic upper mantle, can be explained by the delamination of a thick eclogitic root. Boyd et al. (2004) support this view with evidence from a seismic tomographic study which shows the descent of a two layer slab comprising an eclogitic upper part and a peridotitic lower part, into the mantle. This may be the only place on Earth where dense material is currently being removed from the continental crust into the mantle. 

It is clear, from the examples cited above, that lower crustal delamination takes place. What is less certain is the relative importance of this process, for the number of convincing examples of lower crustal delamination is small. Kramers and Tolstikhin (1997) argued on the basis of their Pb-isotope forward transport model that lower crustal delamination is a minor process and not important in modifying crustal compositions. In contrast, Plank (2005) argued that between 40 and 50% lower crustal loss is required over geological time, to achieve the present-day Th/La ratio of the bulk crust. Whilst this volume equates to only 0.3% of the mantle mass, and less than 10% of the mass of subducted slabs over geological... 

Niu James, 2002), all of which tend to be TTG in composition. As already outlined, Plank (2004) has shown that the low Th/La ratio of Archaean felsic crust (0.18) does not require an origin by fractionation and delamination but is instead a product of (two-stage) mantle differentiation. 

Ducea, M. and Saleeby, J., 1998. A case for delamination of the deep batholithic crust beneath the Sierra Nevada, California. Int. Geol. Rev., 40, 78-93. 

---