Searching for giant, ancient impact structures on Earth: The Mesoarchaean Maniitsoq structure, West Greenland

Abstract

A 100 km-scale, circular region in the Archaean North Atlantic Craton centred at 65°15′N, 51°50′W near Maniitsoq town in West Greenland comprises a set of highly unusual geological features that were created during a single event involving intense crushing and heating and are incompatible with crustal orogenic processes. The presently exposed features of the Maniitsoq structure were buried 20–25 km below the surface when this event occurred at c. 3 Ga, during waning convergent orogeny. These features include: a large aeromagnetic anomaly; a central 35×50 km² large area of comminuted quartzo-feldspathic material; regional-scale circular deformation; widespread random fractures with featherlike textures; intense fracture cleavage; amphibolite–granite-matrix breccias unrelated to faulting or intrusions; formation and common fluidisation of microbreccias; abundant evidence of direct K-feldspar and plagioclase melting superimposed on already migmatised rocks; deformation of quartz by <c> slip; formation of planar elements in quartz and plagioclase; and, emplacement of crustally contaminated ultramafic intrusions and regional scale hydrothermal alteration under amphibolite-facies conditions. The diagnostic tools employed to identify impacting in the upper crust are inadequate for structures preserved deep within the continental crust. Nevertheless, the inferred scale, strain rates and temperatures necessary to create the Maniitsoq structure rule out a terrestrial origin of the structure.

Introduction

Extraterrestrial impacts shaped and modified the early Solar System, and all of its rocky and icy bodies preserve records of past impacting. In order to improve our knowledge about these processes it would be desirable to obtain information on the deep effects of giant impacts on rocky planets. The Moon displays extensive evidence of giant Archaean impacting with a peak at 3.9–3.8 Ga, followed by exponential decline (Cohen et al., 2000, Ryder, 2002). The Earth has a larger cross section and gravitational mass than the Moon and must have received even more early impactors, with potential profound effects on its crust and upper mantle. The Earth is also the only planet in the Solar System that has sufficient uplift and erosion to expose and allow direct observation of the deep parts of giant impact structures, prior to their complete destruction by continued erosion or plate-tectonic events. However, despite decades of searching on several continents, no Archaean craters have yet been established. Vredefort in South Africa aged 2.02 Ga is also the largest (d∼300 km). Yarrabubba (∼50 km) in Western Australia may be ≤2.4 Ga (Macdonald et al., 2003), whereas Setlagole in South Africa (2.79 Ga, ∼25 km; Anhaeusser et al., 2010) remains to be confirmed. All known impact structures are relatively small and surficial, with diameters <300 km and only shallow or moderate erosion. Older and possibly larger impacts are evidenced by rare ejecta layers (Lowe et al., 1989, Simonson and Glass, 2004, Simonson et al., 2009), but the locations and sizes of the source craters remain poorly constrained.

There are few places on Earth suitable to search for the deep remains of impact structures larger than Vredefort, Sudbury and Chicxulub. Well-exposed, old continental cratons provide the longest exposure and possible overlap with the early active impacting. But what impact-related features might be expected, deep below the floor of a giant impact structure? Hydrocode modelling indicates that simple upscaling of the 3D structures known from mapping and drilling at Vredefort, Sudbury and Chicxulub is not feasible, due to Earth's strong gravitational field (Ivanov and Melosh, 2003). With an impactor twice the size of that assumed for Vredefort, most of the impact melt would be trapped in a vertical neck of partially to completely melted material penetrating into the deep crust and possibly upper mantle, instead of spreading out on the crater floor (Ivanov et al., 2010, Garde et al., 2011). Shock-metamorphic, melting and re-equilibration processes known from shallow targets would also be modified by the ambient lithostatic pressure and preheating of deep-crustal targets.

Currently accepted diagnostic criteria for the recognition of impact structures comprise (1) geochemical/isotopic tracers of extraterrestrial material, (2) high-pressure polymorphs of quartz, (3) shatter cones, and (4) shock lamellae (Planar Deformation Features, PDFs), particularly in quartz (French, 1998, French and Koeberl, 2010). However, these criteria are probably deficient with respect to deeply eroded, giant impact structures. This problem has received little attention in the literature.

First, it is unlikely that material from a giant impactor would penetrate far into the crust. Traces of impactors have been found in the melt sheets or suevites of complex impact structures, but structures with diameters of tens of kilometres or larger are commonly eroded and may not retain identifiable traces of the impactor (Palme, 1982, Koeberl, 1998, McDonald et al., 2001). Unless a giant impactor arrived at a highly oblique angle and low velocity it would vaporise or be expelled as melt droplets (Pierazzo and Melosh, 2000), only leaving material traces in distant ejecta such as observed in spherule layers (e.g. Lowe et al., 2003, Glikson, 2005). Furthermore, shock-metamorphic coesite and stishovite formed in deep-seated, preheated host rocks would readily revert to quartz. Shatter cones are common in the surficial parts of impact structures and also known from the Vredefort structure eroded to about 10 km below the impacted surface (Lana et al., 2003, Pope et al., 2004), but they are best developed in fine-grained sedimentary rocks and probably restricted to the upper crust. Finally, deep-crustal PDF-formation would be restricted by the thermal state of the target. Experimental shock deformation of preheated quartz shows that the PDF-forming window and hence the geographical area of PDF-formation narrows with increasing temperature. With preheating to just 600 °C, PDF-formation is replaced by melting already at ∼26 GPa instead of ∼35 GPa without preheating (Langenhorst, 1994, Langenhorst and Deutsch, 1994). In addition, PDFs formed in the deep crust are likely to be poorly preserved. Impact-related and subsequent hydrothermal activity would convert them to trails of fluid inclusions (‘decorations’) with a very high likelihood of additional blurring and/or annealing over time, and their straightness may be affected by crystal-plastic deformation during impacting, crater reorganisation and/or subsequent tectonic events. Such PDFs would hardly constitute the same straightforward and clearcut evidence for impact identification as in near-surface structures.

The presently accepted diagnostic criteria are thus inappropriate for identification of deep-crustal remains of giant impacts. Still, shock waves penetrating into the deep crust would create short-lived brittle deformation that does not otherwise occur in the lower crust, and the accompanying shock heating might result in widespread diaplectic glass formation and/or direct melting of rock-forming minerals such as feldspar, as observed in near-surface craters (e.g. Machado et al., 2009). In the deep crust extensive subsequent recrystallisation would be expected, with possible impact-related generation of anatectic granitic melts. Finally, a shock wave extending all the way into the mantle might lead to discharge of mantle melts into the crust, although the strong gravity field of the Earth would prevent formation of large igneous provinces, even with impacts much larger than Vredefort (Ivanov and Melosh, 2003, Ivanov et al., 2010).