Technical Comments

Comment on "A Vestige of Earth's Oldest Ophiolite"

See allHide authors and affiliations

Science  02 Nov 2007:
Vol. 318, Issue 5851, pp. 746
DOI: 10.1126/science.1144931


The claim by Furnes et al. (Reports, 23 March 2007, p. 1704) that Greenland metavolcanic rocks require Paleoarchean sea-floor spreading is incompatible with their own data. The purported sheeted dikes have the composition of pyroxenitic komatiite and could not have fed the adjacent ferroandesitic pillow lavas. Neither type has ophiolitic analogs, and both are likely ensialic.

When did plate tectonics (the rifting and aggregation of internally rigid plates of lithosphere, as oceanic lithosphere formed by sea-floor spreading between some of them and was subducted beneath others) begin? One bit of the voluminous evidence for a late start is the lack of sections of ancient oceanic lithosphere, ophiolites, abundant in Phanerozoic terrains and consisting of submarine mafic and ultramafic igneous rocks erupted directly upon residual mantle. Furnes et al. (1) espouse the contrary view that plate tectonics began very early and reported the discovery of a partial Paleoarchean ophiolite in the form of “cogenetic” metamorphosed and deformed sheeted dikes and pillow lavas of “intraoceanic island arc and mid-ocean ridge basalt (MORB) affinities.” However, their data neither support these inferences nor provide the claimed “compelling structural evidence of horizontal extension by dike injection at a spreading ridge.”

Neither the Isua pillow lavas nor the purported dikes fit into the broad array of modern islandarc and seafloor basalts for which Furnes et al. use the term “ophiolitic.” The nine similar analyses of “dikes” [table S1 in (1)] are ultramafic, with CaO > Al2O3, 48 to 51% SiO2, ∼15 to 20% MgO, ∼12 to 13% Fe2O3* (total iron as Fe2O3), and ∼1% Na2O. This is the bulk composition of pyroxenitic komatiite, an Archean lava type (2), probably ensialic elsewhere, common at Isua (3, 4) (primary textures have been obliterated by metamorphism). The adjacent pillow lavas cannot have been erupted from the dikes: The five subuniform analyses of lavas [table S1 in (1)] have Al2O3 >> CaO, 54 to 57% SiO2, ∼5% MgO, ∼12% Fe2O3*, and ∼ 4% Na2O and have the composition of mafic ferroandesite, an Archean rock type that is ensialic elsewhere. The near uniformity of each type makes gross metasomatic change of relatively immobile elements implausible. Neither rock type has common modern equivalents, in or out of ophiolites. Sheeted dikes in Phanerozoic ophiolites are, by contrast, compositionally similar to, or moderately less feldspathic than, associated pillow lavas [e.g., (5, 6)], and are not ultramafic.

These Isua pillow lavas do reasonably match, in major elements and minor-element ratios, widespread rocks in the Australian kilometers-thick Neoarchean Fortescue sheet of mostly mafic flood lavas, which overlies 150,000 km2 of older Archean continental crust (7). Fortescue rocks are unlike any modern continental flood basalts, and Ti/Y/Zr discriminants similar to those used by Furnes et al. misclassify them as mostly oceanic-arc basalts (7), thus invalidating conjecture that numerology can define tectonic settings of Archean rocks in the absence of geologic information. Isua stratigraphy is undefined, but the lower parts of Archean supracrustal sections commonly have, like little-deformed Fortescue, subregional sheet stratigraphy where constrained by good mapping and dating.

Archean igneous rocks contrast strikingly with modern ones in both bulk compositions and lithologic and statigraphic assemblages, precluding the direct analogies with modern tectonic settings possible in Phanerozoic terrains. Investigators [e.g., (1)] who nevertheless seek uniformitarian plate-tectonic explanations commonly appeal to selected minor-element ratios as discriminants, with the assumption that those ratios somehow define tectonic settings. That the discriminants misclassify not only Fortescue, but also many modern assemblages of known settings, and that none of the structural and associational predictions implicit in derivative plate-tectonic assignments are fulfilled in Archean terrains, has not deterred the widespread practice by geochemists. The supposed suprasubduction Isua “boninites” (1, 8) are defined only with trace-element rationales, and their basaltic to high-Mg-basaltic, non-boninitic major-element compositions (e.g., low SiO2) are ascribed to gross additive and subtractive metasomatism (8).

The Isua sheeted dikes inferred by Furnes et al. from a small area are severely flattened decimeters-thick layers of relatively massive rock, separated irregularly by more schistose rock assumed to represent chilled margins and volcanic rocks. If the layers indeed are derived from dikes, they indicate crustal extension but not its setting, because sheeted dikes occur locally in non-oceanic settings; or the layering may be inherited from planar decimeterto meters-thick internal facies of close-following pyroxenitic-komatiite lavas (9) at a locality of simple triaxial strain and favorable preservation.

Archean greenstone belts commonly have, like Isua, thick sections dominated by mafic and ultramafic lavas. Only felsic basement, mostly tonalitic, trondhjemitic, and granodioritic gneisses, has ever been found depositionally beneath them, and clastic sediments derived from that basement commonly intervene between it and the volcanic sections where not cut out by younger magmatism [e.g., (10)]. No Archean supracrustal rocks are proved ensimatic. Furnes et al. did not consider non–plate-tectonic alternatives incorporating delamination (11).

References and Notes

View Abstract

Navigate This Article