Technical Comments

Response to Comments 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.1144231

Abstract

The comments by Nutman and Friend, and Hamilton, question our evidence for the presence of the Isua ophiolite. Their critical remarks are particularly directed at the veracity of our inferred sheeted-dike complex, the cogenicity of pillow lavas and dikes, and the nonexistence of modern equivalents. Here, we expand on our explanations in response to each of their comments to better justify our arguments and interpretation.

Our study (1) emphasized that the components of the Isua ophiolite are tectonically dismembered and that its reconstruction is therefore not a trivial task, as compared with many younger (Phanerozoic) ophiolites. Hence, we welcome the comments by Nutman and Friend (2) and Hamilton (3) as an opportunity to provide further details, to clarify the re-construction of the Isua ophiolite, and to respond to the following issues raised in their comments: (i) the sheeted dikes, (ii) the geochemical composition of the dikes and the pillows, (iii) the facing (stratigraphic) relationships between the dikes and pillows, (iv) geochronological constrains, (v) comparison with modern equivalents, and (vi) implications for Archean plate tectonics.

Nutman and Friend (2) state that “At Isua, copious Paleoarchean Ameralik dike swarms cut all Eoarchean rocks, including all components of the 3.8 to 3.6 Ga orthogneisses that envelop the Isua supracrustal belt” and imply that we may have misidentified the Ameralik dikes as (part of) an ophiolitic sheeted-dike complex. The sheeted dikes from our study are clearly distinguishable from the Ameralik dikes. The latter occur generally as thick tabular bodies with ophitic texture and plagioclase megacrysts. They are largely undeformed and, in the area of question, lack cleavage or schistosity. They cross-cut metacherts, pillow lavas, ultramafic units, and the garbenschiefer rocks (including the boninitic metaba-salts), which flank the sheeted-dike complex. The sheeted-dike complex in contrast forms tabular mappable units which contain thin screens of pillow lava, volcanic breccias [see figure 2C in (1)], and occasional plagiogranite, and comprise highly deformed and foliated amphibolites that are in places crosscut, often orthogonally, by undeformed Ameralik dikes. The Ameralik dikes hence postdate deformation of the ophiolitic sequence and, in the area of question, are easily distinguished from the highly deformed ophiolitic dikes. Although the “low-resolution” sketch maps (1, 2), both modified directly from (4), are not at appropriate scales to illustrate these details, we are confident that we have correctly differentiated the younger Ameralik dikes from the amphibolites interpreted by us to be sheeted dikes.

Our chemical data [table S1 in (1)] show that the dikes of the sheeted complex are less evolved than the pillow lavas. Nutman and Friend (2) and Hamilton (3) state that the pillow lavas cannot be derived from the dikes because they are geochemically too different; hence, they question the inferred cogenicity. We disagree with this assessment, because immobile trace elements (e.g., Ti-V and Ti-Zr) covary along common fractionation trends, and the ratios of incompatible elements (e.g., Zr/Y) are very similar between the dikes and the pillows. Moreover, the dikes must necessarily be closer to the magma source than the pillow lavas that they fed, and we argue that the magma underwent fractional crystallization en route to the surface. Although the major element variations of the Isua rocks are likely complicated by metamorphic overprints, they can largely be explained by fractional crystallization of Mg-rich olivine and Ca-rich pyroxene leading to the chemical differences between the dikes and pillow lavas seen. Such chemical differences are common in most volcanic systems.

Hamilton (3) uncritically uses the major element composition of the dikes to classify them as “pyroxenitic komatiites.” Such classifications should not, however, be made on the basis of elements that may have been mobile during metamorphic conditions. For example, the process of spilitization—which causes local redistribution of principally Ca and Mg, together with again in Si, Na, and H2O and often a loss of Ca, K, Sr, Rb, and Ba—is a well-known process [e.g., (5)] that has misled many geochemists. Spilitization is commonplace in Archean volcanic rocks, and we are currently documenting such chemical changes in the little-deformed, low-grade metamorphic lavas from the Mesoarchean Barberton Greenstone Belt in South Africa [the type location of (pyroxenitic) komatiites], from which we have geochemical data for more than 500 samples. Large chemical changes may have occurred in the Isua complex and, hence, we reject this classification. Hamilton (3) further draws attention to the similarity between the little-deformed basalts of the ∼2.7 Ga Fortesque Group of the Pilbara Craton (Western Australia) and the geochemistry of the Isua dikes and lavas, and to the former having been misclassified on the basis of Ti/Y/Zr discriminant diagrams (6). We do not see the relevance of these observations extrapolated from the Pilbara Craton to the discussion about Isua.

Nutman and Friend (2) point out that the facing (younging) direction of the pillows is opposite to the transition from the dike to pillow facies. This younging relationship, which has been determined in one outcrop by them and by us, has little relevance because the ophiolite units cannot be traced continuously and the pillows and dikes are strongly flattened and stretched, and thus tectonically rotated, a feature which is also common in Phanerozoic ophiolites [e.g., (7)]. Furthermore, the supracrustal belt has experienced several episodes of deformation and folding [e.g., (4)]. In such a setting, the attitude of dikes, bedding, and the younging direction of pillows can attribute little, if any, relevance to their original disposition. It is thus not possible to use a local facing direction as a regional reconstruction tool without a more complete understanding of the structural geology.

Nutman and Friend (2) claim that the existing geochronological constraints do not support a cogenetic relationship between the components of the Isua ophiolite that we have identified. There are, however, no radiometric dates on the individual mafic-ultramafic components (i.e., pillow lavas, sheeted dikes, gabbro, and ultramafics) that constitute the Isua ophiolite to corroborate this. As we pointed out in (1), current whole-rock Sm-Nd age data from metasediments and enclosing garbenschiefer yield an age of 3,779 ± 81 Ma (8), and the pillow lavas and metagabbro an age of 3,777 ± 41 Ma (9). Indeed, detailed chronostratigraphy within any of the Isua tectonic packages is still lacking, and several geochronological aspects of the Isua belt remain unresolved. Lack of geochronological data cannot therefore be the basis for discounting cogenicity.

Hamilton (3) states that we stretch the term “ophiolite” both in terms of construction and petrological development and also questions whether modern equivalents for Archean oceanic crust, and hence the Isua ophiolite, exist. If the term “ophiolite” is to be restricted to the typical Phanerozoic “Penrose-type” ophiolite only, we might agree. There exists, however, a much broader classification scheme for ophiolites that includes a wide variation in terms of lithological construction and petrological evolution, depending on their environment of formation and stage of evolution (10). For example, the Caribbean-type ophiolites that represent the oceanic crustal assemblages common in Large Igneous Province (LIP) [e.g., (10, 11)] distinctly differ from common Penrose-type ophiolites, having thick volcanic piles (exceeding 5 km) that include substantial volumes of ultramafic rocks [e.g., (12, 13)]. It is also pertinent to mention that the Tortuga ophiolite complex, a Chilean-type ophiolite (10), contains high-Mg basalt dikes cutting the deeper diabase and gabbro units below the sheeted-dike complex (e.g., 14).

Hamilton's views on the Archean are well known [e.g., (15)], and we do not wish to repeat the discussion of Hamilton (16) and counterarguments by de Wit (17). Our Isua field observations and new geochemical data put this debate on a new footing, and we are striving to allow the rocks themselves to provide the answers to how the Earth worked at 3.8 Ga. Our physical evidence for a spreading center [figure 2, C to F, in (1)] and geochemical evidence for subduction-related magmatic activity from Isua (3, 18, 19) offer evidence that is consistent with Archean plate tectonics, but much remains to be done to understand its full importance.

In summary, many of the comments by Nutman and Friend (2) and Hamilton (3) are pertinent, focusing on uncertainties in the complex geological history of the Isua supracrustal belt, including its inferred oceanic crustal component (1). Although we maintain our interpretation that these rocks represent part of a dismembered ophiolite, we agree that there are still many uncertainties left, and we reiterate from our paper (1) that there is a need for more in-depth work, based on detailed field observations.

References and Notes

View Abstract

Navigate This Article