Szilas et al. (2013) present a critical Comment regarding our interpretation (see also Polat, 2013), that the petrogenesis of samples from Innersuartuut (southwest Greenland) was analogous to modern within-plate magmas (e.g., oceanic island basalts [OIB] and oceanic plateau basalts [OPB]). Instead, they conclude that the Innersuartuut suite is more comparable to modern and ancient (the 3.7 and 3.85 Ga Isua suites) convergent margin magmas.

We (Jenner et al., 2013) used a combination of whole-rock projections, systematic correlations between SiO2, CaO, FeO, TiO2, P2O5, Al2O3, and Na2O with MgO, and between CaO/Al2O3 with Sc/Yb, Gd/Yb, Zr/Hf, and Nb/Yb, in addition to primitive mantle-normalized plots to place constraints on the pre-metamorphic petrogenesis of the Innersuartuut suite. The range in compositions is consistent with the expected (experimental) liquid line of descent of anhydrous (non-subduction) basaltic melts at pressures of 0.7–0.9 GPa, prior to subaerial eruption of melts and entrained phenocrysts. Trends in compositions of the Isua suite in whole-rock projections (Jenner et al., 2009, their figure 4), the decrease in FeO with decreasing MgO (Jenner et al., 2009, their figure 2), and differences in trace element systematics indicate that both the initial parental melt compositions and differentiation histories of the Isua and Innersuartuut suites were distinct.

The main criticism presented by Szilas et al. pertains to our apparent lack of consideration of the effects of element mobility during small-scale partial melting associated with granulite facies conditions. They claim that this would produce “selective depletion of Zr-Hf relative to Sm-Gd and enrichment of Nb-Ta relative to La” and refer to modeling presented in Foley et al. (2002). However, we provide a long discussion (in our Data Repository item, DR2013087) regarding evidence for and against post-magmatic processes, and show that the variations in the moderately compatible elements Zr/Hf, Gd/Yb, Nb/Yb, Sc/Yb, and CaO/Al2O3 can be reproduced by simple fractional crystallization, indicating minimal post magmatic mobility.

We would like to clarify the main conclusions presented in Foley et al. (2002) referenced by Szilas et al. Foley et al. (2002) used partial melt modeling to assess whether the generation of the early continental crust can be linked to either partial melting of low-Mg amphibolite in a subducting oceanic slab or partial melting of eclogites/magnesium-rich amphibolites in the lower portions of an over-thickened (originally clinopyroxene+olivine-rich) oceanic crust. Foley et al. (2002) demonstrated partial melting of low-magnesium amphibolite can explain the low Nb/La in melts (high Nb/La in the residue), as required for the early continental crust, whereas the melting of eclogite cannot. Hence, the modeling presented in Foley et al. (2002) would not produce the OIB-OPB–like Nb/La signatures of the Innersuartuut suite during granulite facies metamorphism of Isua-type protoliths in the lower crust. Further, estimates of the trace element pattern of the lower continental crust (granulites and lower crustal xenoliths) show comparable negative Nb anomalies to those seen in the middle, upper, and bulk continental crust, as well as the majority of convergent margin magmas (Rudnick and Gao, 2003), indicating lower continental crustal granulites also typically preserve their arc-like signatures during high pressure-temperature metamorphism.

In the complete absence of arc-like signatures, we do not consider that a subduction-derived petrogenesis with a convenient degree of post-magmatic alteration resulting in selective depletion of Th, U, Zr, and Hf and enrichment of Nb relative to the REE, required to result in the non-convergent margin-type major and trace element characteristics, constitutes a “more reasonable” interpretation of the petrogenesis of the Innersuartuut suite, as suggested by Szilas et al. Although Szilas et al. provide an alternative scenario worth careful consideration, we are drawn to our original conclusion not because it is the only conceivable explanation of our data, but because it is the simplest.

Szilas et al. highlight differences in the compositions of the Innersuartuut suite compared to modern HIMU-, EM1-, and EM2-type OIB. However, we suggest this might be expected considering that distinctive enriched HIMU, EM1, and EM2 reservoirs likely did not exist as early as 3.85 Ga (Nebel et al., 2013, and references therein). The lack of an enriched component in the deep mantle in the Eoarchean, coupled with potential differences in degrees of partial melting during modern and ancient plume-related magmatism, present plausible mechanisms for producing more-pronounced depletions in the light rare earth elements, Th, and U and lower Gd/Yb compared to modern OIB. Indeed, in our figure DR5, we identify two 2.8–2.9 Ga suites from the Baltic Shield, previously interpreted to have plume-source characteristics, with comparable Th-Nb-Yb-U systematics to the 3.85 Ga Innersuartuut suite. Further, while OPB commonly lack the distinctive high Gd/Yb of OIB, they are demonstrably related to plume-related magmatism (Fitton and Godard, 2004), and reflect diversity in the compositions of plume-derived magmas and ultimately, their petrogenesis.