University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to The Journal of Geology. http://www.jstor.org The Sedimentary Record of Extraterrestrial Impacts in Deep‐Shelf Environments: Evidence from the Early Precambrian Author(s): Scott W. Hassler and Bruce M. Simonson Source: The Journal of Geology, Vol. 109, No. 1 (January 2001), pp. 1-19 Published by: University of Chicago Press Stable URL: http://www.jstor.org/stable/10.1086/317958 Accessed: 04-01-2016 01:46 UTC Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/ info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. This content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions [The Journal of Geology, 2001, volume 109, p. 1–19]
2001 by The University of Chicago. All rights reserved. 0022-1376/2001/10901-0001$01.00 1 ARTICLES The Sedimentary Record of Extraterrestrial Impacts in Deep-Shelf Environments: Evidence from the Early Precambrian Scott W. Hassler and Bruce M. Simonson1 Department of Geological Sciences, California State University, Hayward, California 94542, U.S.A. (e-mail: shassler@juno.com) ABSTRACT Impact ejecta layers in four formations in the Hamersley Basin (Western Australia) and in one formation from the Transvaal Supergroup (South Africa) show striking evidence for impact-related reworking. Each layer contains sandsized spherules of a former silicate melt that resemble those found in well-documented impact layers. Given available isotopic age dates of associated strata and uncertainties in stratigraphic correlation, these layers represent a minimum of three and a maximum of five impacts between ca. 2.49 and 2.63 Ga. All of these layers were deposited below wave base in deep-shelf environments, yet they show a common suite of sedimentary features indicating deposition and reworking under high-energy conditions. These features occur in a consistent order: (1) extensive erosion, including the transport of meter-scale rip-up clasts, (2) reworking by waves, (3) synwave to postwave offshore-directed bottom return flow, and (4) later reworking by sediment gravity flows. We interpret the consistent association of erosion, wave reworking, and bottom return flow as a result of tsunami triggered by the impact. The sediment gravity flows may have been triggered by impact or may occur much later. The wave features in these layers indicate they are the result of oceanic impacts, and their sedimentological similarities suggest a consistent set of depositional processes that can be used to recognize the distal ejecta layers of marine impacts, particularly those deposited in deep-shelf settings. Given the relatively rapid tectonic recycling of oceanic crust, such layers probably constitute our best source of information on the frequency and effects of large impacts in open-ocean basins. Introduction Extraterrestrial impacts on the earth produce a range of sedimentary deposits that may be preserved in the earth’s geologic record (Grieve 1998). These deposits range from the proximal layer of debris mantling the area surrounding the crater, often called the ejecta blanket, to more widespread distal ejecta fallout layers. These distal deposits range from local to global in extent. In addition, the energy released by impacts often directly or indirectly triggers a variety of high-energy surficial processes capable of reworking sediment, including wave systems, subaerial or subaqueous sediment gravity flows, and possibly hyperstorms. The effects of these processes are often catastrophic and may yield deposits that are likely to be preserved Manuscript received August 3, 1999; accepted August 29, 2000. 1 Department of Geology, Oberlin College, Oberlin, Ohio 44074-1044, U.S.A.; e-mail: Bruce.Simonson@oberlin.edu. in the sedimentary record, particularly in lowenergy environments. The preservation potential of thin distal impacts is heightened in sedimentary strata deposited on the Precambrian seafloor because the organisms responsible for bioturbation did not evolve until ca. 610 Ma, based on the trace fossil record (Crimes and Droser 1992). Two of the most extensive and best-preserved early Precambrian sedimentary successions on Earth are the Hamersley Basin of Western Australia and the Transvaal Supergroup of South Africa. We have identified layers that we interpret as the products of impact events in five different stratigraphic units: four in the Hamersley Basin (fig. 1) and one in the Griqualand West Basin of the Transvaal Supergroup (fig. 2). Given the stratigraphic relationships of the units involved, these layers represent no less than three separate deposits formed between ca. 2.49 and 2.63 Ga, based on interpolaThis content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions 2 S. W. HASSLER AND B. M. SIMONSON Figure 1. Maps for Hamersley Basin spherule layers. A, General map of Western Australia: outline shows Pilbara Craton. Black area indicates the Hamersley Basin outcrop area shown in figure 3B and 3C. B, C, Section location maps for Wittenoom and Carawine spherule layers and Dales Gorge and Jeerinah spherule layers, respectively. Solid line on both maps shows outcrop area of Hamersley Group strata, after Myers and Hocking (1988). Capital letters indicate sections shown in figure 5. tions between isotopically dated zircon-bearing tuffs (Arndt et al. 1991; Altermann and Nelson 1998; Trendall et al. 1998; Nelson et al. 1999). All of the layers interpreted as impact deposits were deposited below wave base in deep-shelf environments, yet none of the layers show the graded bedding one would expect by deposition via direct suspension sedimentation. All five layers show evidence of extensive reworking, and we have detected a suite of sedimentary features common to most or all of them. We believe this consistent depositional pattern is best explained as a result of impact-related processes. The goals of this article are to describe the sedimentological features of these layers and to suggest that they provide distinctive criteria for the recognition of distal impact This content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions Journal of Geology I M P A C T EJECTA IN DEEP-SHELF ENVIRONMENTS 3 Figure 2. Transvaal Basin spherule layer maps. A, General map of southern Africa showing Transvaal Basin. Box outlines the Griqualand West Basin. B, Section location maps for the Monteville spherule layer in the Griqualand West Basin. Solid line outlines outcrop area of Campbellrand Subgroup strata. The Ganyesa Dome represents a probable source of the continental material in the Monteville spherule layer. Capital letters indicate sections shown in figure 6. PM, Pering Mine. MF, Monteville Farm. layers within successions deposited in deep-shelf environments. This article has four sections. First, we summarize the age, depositional setting, and evidence for impact of the Hamersley Basin and Transvaal Supergroup spherule layers. Second, we document four distinctive sedimentary features of these reworked impact deposits: (1) extreme substrate erosion, (2) reworking by impact-generated waves, (3) evidence of postwave offshore bottom flow, and (4) reworking by sediment gravity flows. Third, we present a depositional history for each layer. Finally, we discuss mechanisms for generating these processes and speculate on the nature of the target areas for the Hamersley and Transvaal Supergroup impacts. Spherule Layers Hamersley Basin. The four Hamersley Basin spherule layers for the stratigraphic units in which they occur are the Wittenoom spherule layer, the Carawine spherule layer, the Dales Gorge spherule layer (for the Dales Gorge Member of the Brockman Iron Formation), and the Jeerinah spherule layer (fig. 3). The Wittenoom spherule layer (previously named the Spherule Marker Bed, or SMB, by Simonson 1992) has a preservational area of 121,000 km2 . Sedimentary carbonates in and adjacent to the layer have been dated at 2541
18/15 Ma by direct Pb-Pb dating (Woodhead et al. 1998). We have recognized the Dales Gorge spherule layer at five sites covering an area of 8000 km2 . Interpolating between U-Pb zircon dates from nearby tuff layers (Arndt et al. 1991; Trendall et al. 1998), we estimate its depositional age at 2490 Ma. To date, we have only identified the Jeerinah spherule layer in a single core from the eastern Fortescue River valley (fig. 1C). The layer here comprises a single bed !6 mm thick that shows several complex internal subunits (Simonson et al. 2000a). Again, using U-Pb zircon dates from nearby tuff layers, Simonson et al. (2000a) estimate the depositional age of the Jeerinah spherule layer at 2630 Ma. Given limited knowledge about the Jeerinah spherule layer, we do not include it in our subsequent discussions in this article. The Carawine spherule layer occurs near the base of the Carawine Dolomite, a Hamersley Group unit restricted to the northeastern part of the Hamersley Basin (fig. 1B). It is exposed in multiple outcrops along an east-to-west line roughly 12 km long and in a core located roughly 4 km north of this strike line. We have argued that the Carawine and Wittenoom spherule layers are two parts of a single deposit, based on their sedimentological characteristics (Simonson 1992; Simonson and Hassler 1997) and a Pb-Pb isotopic age at 2548
26/ 29 Ma from Carawine Dolomite carbonates (Woodhead et al. 1998). However, we have yet to find occurrences of any other spherule layer in the Carawine area, leaving open the possibility that the Carawine spherule layer may be equivalent to the Jeerinah spherule layer; it could even be a fourth spherule horizon distinct from the other three. In this article, we treat the Wittenoom and Carawine spherule layers as separate deposits to highlight their sedimentological features better. Griqualand West Basin. The Monteville Formation, at the base of the Campbellrand Group (fig. This content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions 4 S. W. HASSLER AND B. M. SIMONSON Figure 3. Generalized stratigraphic columns for the Hamersley and Transvaal Basins showing positions of spherule layers, lithologies, and geochronology. SL, Spherule layer. The northeastern and main Hamersley Basin outcrop areas are indicated on figure 1B. Stratigraphic thicknesses and lithologies are from Trendall (1983), Beukes (1983), and Klein and Beukes (1989). Hamersley Basin geochronology is from Arndt et al. (1991), Trendall et al. (1998), Woodhead et al. (1998), and Nelson et al. (1999). Transvaal Basin geochronology is from Altermann and Nelson (1998) and Nelson et al. (1999). 3), contains one horizon rich in spherules (Simonson et al. 1999). Our work to date has recognized this layer over an area of ∼17,000 km2 . Simonson et al. (2000a) argue that the age of the Monteville spherule layer is between ∼2640 Ma and ∼2520 Ma based on published dates from tuff layers elsewhere in the Transvaal Supergroup. The stratigraphic successions in Hamersley and Griqualand West Basins display striking similarities (Button 1976; Cheney 1996). In addition, recent radiometric age dates (fig. 3) also indicate the two are in large part contemporaneous. This raises the possibility that the Monteville spherule layer may correlate with one of the Hamersley Basin layers. At this time, given the age constraints outlined above, the Monteville spherule layer may be correlative with the Jeerinah, Wittenoom, and/or Carawine spherule layers in the Hamersley Basin (Simonson et al. 2000a). Depositional Setting. Three lines of evidence indicate that all of the spherule layers in the Hamersley and Griqualand West Basins were deposited below wave base in basinal or deep-shelf environments. First, the layers are interbedded with This content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions Journal of Geology I M P A C T EJECTA IN DEEP-SHELF ENVIRONMENTS 5 sedimentary rocks that were originally clastic or chemical mud, including carbonate (especially dolomite), shale, chert, and/or banded iron formation (Beukes 1987; Simonson et al. 1993a). Second, thin laminae are ubiquitous in these muddy strata, which in some cases exhibit extreme lateral continuity (Trendall and Blockley 1970; Ewers and Morris 1981), indicating that they were deposited in quiet water, well below the influence of shallow water processes. Finally, rare coarse-grained strata, stratigraphically close to some of the layers, also indicate deep-water depositional settings. These include isolated carbonate grainstone beds near the Wittenoom and Carawine spherule layers and beds of volcaniclastic ash and lapilli both above and below the Wittenoom, Carawine, and Monteville spherule layers. These carbonate and volcaniclastic arenites display sedimentary structures typical of turbidites, including normally graded bedding, partial to complete Bouma sequences, and cross lamination formed by climbing current ripples (Hassler 1993; Simonson et al. 1993b). A few of the volcaniclastic layers are normally graded settle-out deposits deposited directly from eruptions; one of these can be traced across the entire basin, again indicating deposition in quiet water (Hassler 1993; Simonson et al. 1993a; Trendall et al. 1998). Whatever their origin, none of these sandy deposits show evidence of reworking by shallow marine processes such as waves or tides. Since there are no quantitative indicators of water depth in the Precambrian, such as foraminifera tests, depth estimates for the strata containing the spherule layers are generally based on their depositional features. Through comparison with Phanerozoic carbonates, Simonson et al. (1993b) argued that water depth during deposition of the Wittenoom Formation was at least 200 m. In contrast, Klein et al. (1987) suggested a value on the order of 40 m for water depth during the deposition of the Monteville Formation. Although parts of the Monteville Formation were clearly deposited above wave base, such as the wave-rippled Motiton Quartzite that caps the formation, the water column must have been deeper when the spherule layer formed (Simonson et al. 1999). The lack of wave-formed sedimentary structures in close proximity to the spherule layer (see below) indicates the sea floor was below storm wave base at the time of deposition. We thus suggest that all of the spherule layers were deposited in water that was hundreds of meters deep. Evidence for Impact The key criterion for our recognition of the Hamersley Basin and Griqualand West Basin spherule layers is the presence of unusual spherules. The spherules are restricted to one single layer in each of these formations, and the spherules in all of the layers are very similar in size, shape, composition, and texture (fig. 4A, 4B). They are fine to very coarse sand-sized spherical particles that consist largely of acicular- to lath-shaped crystals of K-feldspar. Most commonly, these K-feldspar crystals are organized into radial-fibrous sprays that diverge inward from grain edges (fig. 4C). We attribute these textures to postdepositional devitrification, based on their close similarity to devitrification textures in both lunar impact spherules (Lofgren 1971) and devitrified spherules in the K-T boundary layer (Bohor and Glass 1995). Many of these spherules with inward-radiating textures have central clear spots consisting of coarse, nonfibrous crystals of K-feldspar, quartz, and/or carbonate, and/or finely crystalline muscovite. Some of these central spots represent filled-in vesicles (based on their smooth, circular outlines). Others probably represent relict glass cores that devitrified or were replaced later in diagenesis than were the rims. More rarely, the spherules contain acicular- to lath-shaped K-feldspar crystals that are randomly oriented rather than radial (fig. 4D). Based on their close similarity to plagioclase crystallites grown in experimental charges (Lofgren 1977), we interpret these random textures as a product of primary crystallization during the cooling of melt droplets that were originally basaltic in composition. Replacement by K-feldspar also suggests a basaltic composition because the only other units in the Hamersley and Transvaal successions where K-feldspar replacement is commonplace are basaltic tuffs (Hassler 1993). Given these interpretations, the spherules with radial textures probably consisted entirely of glass prior to replacement, analogous to microtektites, while those with random textures probably consisted of a mixture of glass and crystals, analogous to microkrystites. Carbonate has replaced portions of some spherules in all four layers; in addition, spherules in the Dales Gorge spherule layer are commonly replaced by stilpnomelane (Simonson 1992). The latter presumably reflects the influence of the iron-rich sediments that enclosed the Dales Gorge macroband during diagenesis. The quartz and muscovite in the spherules are clearly secondary diagenetic phases. These minerals may represent one or more later generations of replacement, as has been docuThis content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions This content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions Journal of Geology I M P A C T EJECTA IN DEEP-SHELF ENVIRONMENTS 7 Figure 4. Close-ups of spherules and spherule layers. A, Bedding-plane exposure of spherules in the Wittenoom spherule layer at section BB (fig. 1B). Spherules weather out in full relief because they consist of almost pure Kfeldspar and are encased in a matrix of carbonate and shale. Diameter of coin is 17 mm. B, Cross section of a surface sample from the Dales Gorge spherule layer at the DG section (fig. 1C) showing contact between spherule-rich ripple layer and upper spherule-poor graded bed. Note that ripples are not visible on scale of photograph. White material is K-feldspar, which has replaced the outer parts of the spherules and displays radial-fibrous textures. Darker material is largely stilpnomelane whose iron has weathered to brownish oxyhydroxides. Pencil point in upper left is for scale. C, Photomicrograph in plane-polarized light of spherule from the Wittenoom layer at section MG (fig. 1B) locality showing inward-radiating textures typical of the majority of the spherules. Fans consist of K-feldspar with submicroscopic inclusions; clear central spot consists of pure K-feldspar and was probably a vesicle originally. Material surrounding spherule is primarily K-feldspar and carbonate but appears dark due to iron oxyhydroxides formed by surface weathering. Diameter of spherule is 0.71 mm. D, Photomicrograph in plane-polarized light of spherule from the Dales Gorge spherule layer at the Yampire Gorge mine locality showing a texture present in a minority of spherules—randomly oriented laths to needles of K-feldspar. Darker material between laths and filling pores surrounding spherule is stilpnomelane. Adjacent particles have been replaced by a combination of stilpnomelane and K-feldspar. Longest dimension of spherule is 1.08 mm. mented in Archean mafic volcaniclastic rocks (Dimroth and Lichtblau 1979). The spherules are readily distinguishable from carbonate ooids or volcanic accretionary lapilli, which are the other types of near-spherical, millimeter-scale sedimentary particles in the stratigraphic units that host the spherule layers in both the Hamersley and Griqualand West Basins. The ooids consist entirely of carbonate, display both concentric laminae and radial structures internally, and show crystal sprays that diverge outward from grain centers rather than inward from grain edges (Beukes 1983; Simonson and Jarvis 1993). The volcanic accretionary lapilli in both basins consist of fine-grained ash that is massive or displays crude concentric zoning, unlike the radiating to randomly oriented crystallites in the spherules (Boulter 1987; Hassler 1993; Altermann 1996). Neither ooids nor accretionary lapilli are abundant in any of the units that host the spherule layers. Three other lines of evidence support an impact origin for one or more of the layers. First, positive iridium and other platinum-group element anomalies have been reported for the Wittenoom, Carawine, and Monteville spherule layers (Simonson et al. 1998, 2000c). Second, in addition to spherules, the spherule layers also contain particles with other shapes typical of ejecta. These include two or more spherules welded together and dumbbell- to teardrop-shaped grains. Finally, the Carawine spherule layer contains irregularly shaped melt ejecta particles up to 2 cm in length, which resemble Muong Nong–type tektites (Simonson et al. 2000b). Externally, these particles are much more angular and irregular in shape than spherules with the typical fibrous textures. Internally, many of these larger, more irregular particles exhibit replaced flow banding and/or schlieren. Some also have fine crystals or inclusions of what appear to be typical spherules as inclusions. These particles again consist largely of K-feldspar, many with fibrous to acicular habits. The presence of mineral grains showing planar deformation features (PDFs) is commonly cited in support of an impact origin. The Wittenoom spherule layer contains small quantities of sand-sized quartz and feldspar crystals, but to date we have not recognized unequivocal PDFs in any of these grains (Simonson et al. 1998). Individual sand-sized quartz crystals are also present in the Monteville and Carawine spherule layers (Simonson et al. 1999, 2000b); they also lack evidence for PDFs. We interpret these tectosilicate grains as detrital material eroded from sources at the margins of the Hamersley and Griqualand West Basins by impacttriggered sedimentary processes; this is discussed more below. We have argued (Simonson et al. 1998; Simonson and Harnik 2000) that the apparent absence of shocked continentally derived material may indicate a mafic (i.e., ocean floor) target area. This line of reasoning applies to the Monteville spherule layer as well. The only quartz grains containing candidates for PDFs are in the Carawine spherule layer. One of the irregular melt particles that resembles a replaced Muong Nong–type tektite contains a few lenses with quartz crystals showing internal parallel lines that could be PDFs (Simonson et al. 2000b). This suggests an impact into continental rocks may have been responsible for producing the ejecta in the Carawine spherule layer. This content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions 8 S. W. HASSLER AND B. M. SIMONSON Criteria for Recognizing Impact Deposits All three of the diagnostic impact features—melt ejecta particles, a geochemical anomaly, and mineral grains with PDFs—are difficult or impossible to use as field criteria for recognizing reworked impact deposits. The identification of millimeterscale melt ejecta may be difficult. They may compose only a minor fraction of a reworked deposit and may be badly altered by even moderate weathering. The geochemical work needed to establish an anomaly is, of course, laboratory based and cannot be used to recognize impact deposits in the field. Finally, confident identification of mineral grains with PDFs requires careful petrographic examination. In contrast, larger-scale sedimentary features may be readily used for field identification of potential impact deposits. We recognize four field criteria from our work in the Hamersley and Griqualand West Basins: evidence of (1) intense substrate erosion, (2) wave activity below storm wave base, (3) evidence of offshore-directed bottom flow, and (4) redeposition by sediment gravity flows. These features can of course all be generated by nonimpact processes. We believe their presence, singly or in combination, especially in deep-shelf successions, is a good identifier of potential impact deposits that merit closer scrutiny. Figures 5 and 6 show representative stratigraphic columns for each spherule layer that illustrate these criteria. Evidence of Intense Bottom Erosion. All four of the spherule layers considered here show two types of evidence for intense bottom scour: rip-up clasts of local basinal muds and/or sharp basal contacts that locally cut across substrate layers. The Dales Gorge spherule layer provides a dramatic example. At sections in the central Hamersley Basin (fig. 1), the lower part of the layer contains boulder- and cobble-sized pieces of laminated and ferruginous chert and sand- to pebble-sized argillite and chert clasts (fig. 7A). These clasts are identical in lithology to local basinal muds; we interpret them as rip-up clasts eroded by the high-energy waves and/or currents associated with the deposition of the spherules (see below). Clasts are typically slab-shaped, with their long axes oriented roughly parallel to bedding. Clast margins are generally rounded but are locally frayed and brecciated, indicating that they were further broken up during transport. The basal contact of the Dales Gorge spherule layer is sharp and undulating. It cuts vertically up to 5 cm into the underlying strata in a lateral distance of 45 cm. Simonson et al. (1999) document similar features in the Monteville spherule layer. The lower portion of the Monteville spherule layer contains many ripup clasts, here composed of carbonate, shale, and syngenetic pyrite concretions (fig. 7). The largest of these are slabs of carbonate lutite as much as 180 cm long. These slabs are identical in lithology and thickness to the carbonate layers underlying the spherule layer; they are clearly locally derived ripups. In addition, many clasts have well-rounded ends or show the injection of spherules and sandy matrix along some bedding planes, producing a frayed, plastically deformed texture. Both of these textures indicate that the clasts were only semilithified when they were eroded. In addition, the basal contact of the Monteville spherule layer scours down across the underlying strata with local relief of up to 21 cm. Bedding at the margins of these pockets is truncated along subhorizontal to vertical surfaces by the spherule layer and, in places, shows in situ brecciation. Material from the spherule layer has been injected along bedding planes and has broken parts of the beds away from the substrate. These incipient clasts are of the same scale as the intraclasts in the spherule layer; we suggest many of them may have formed via this injection and brecciation mechanism. The Carawine spherule layer also contains abundant intraclasts, largely laminated to massive carbonate and chert lutites that lithologically match the substrate. However, a few of the carbonate clasts show shallow-water features, including stromatolites, seafloor-encrusting precipitates, and oolitic grainstones (Simonson et al. 1993b). These clasts must have originated in a carbonate platform, a different setting than the deeper-water strata that host the Carawine spherule layer. Therefore, this material was either eroded from the shelf and transported into the deeper-water part of the basin during the formation of the spherule layer or was reworked within the basin from previous deposits. The base of the Carawine spherule layer is also erosive; it truncates the underlying laminated carbonate, with relief of up to 25 cm locally. Finally, the Wittenoom spherule layer contains carbonate, argillite, and laminated chert intraclasts up to 62 cm long (Simonson 1992; Simonson and Hassler 1997). These match the lithology of the underlying substrate, again indicating that the deposition of this layer involved extensive substrate erosion. Evidence of Wave Activity below Storm Wave Base. Three of the four spherule layers locally display wave-generated sedimentary structures. First, in exposures in the southwestern part of its known area of occurrence, the spherules in the Wittenoom spherule layer are concentrated into a series of symThis content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions Journal of Geology I M P A C T EJECTA IN DEEP-SHELF ENVIRONMENTS 9 Figure 5. Representative measured sections of Hamersley Basin spherule layers. Shaded areas indicate the spherule layers. The Carawine spherule layer comprises three coarse-grained, poorly sorted, meter-scale debris flow deposits; impact ejecta is concentrated at the tops of all three flows and at the base of the lowest deposit along its erosional contact with the underlying carbonate. The Wittenoom and Dales Gorge spherule layers both vary laterally from mass-flow deposits (sections MG and WG) to wave ripples (sections BB and DG). Stratigraphic evidence, discussed in the text, indicates that wave deposition preceded the mass flows. Dashed lines indicate correlative horizons. Bar scale at the base of each section shows grain size, ranging from silt (s) through sand-sized grades (vf–vc) to pebbles (p) and cobbles (cb). Rare boulder-scale clasts are shown at their approximate stratigraphic locations. Note changes in scale between sections. metrical ripples (Hassler et al. 2000). The majority (190%) of the sand grains in the ripples are spherules; the remainder are millimeter-scale, plateshaped ferruginous intraclasts transported by the ripple-forming event. The ripples range from 23 to 40 cm in wavelength and from 1.0 to 3.5 cm in height and are symmetrical to slightly asymmetrical in cross section (fig. 7C). No internal cross laminae were observed in the ripples; this may reflect the small grain size variation in the spherules. Given their dominantly symmetrical cross sections, Hassler et al. (2000) interpreted these ripples as wave-generated bedforms. Not only are they the only wave-generated bedforms in the entire Wittenoom Formation but they also occur in the only layer in the entire formation that contains spherules. They record an exceedingly rare high-energy sedimentation event that also caused extensive seafloor erosion. Because the wave ripple layer consists almost exclusively of spherules, Hassler et al. (2000) suggest that waves formed the layer shortly after the impact-produced spherules reached the floor of the Hamersley Basin, presumably in response to an impact. In the western and southern Hamersley Basin, the Dales Gorge spherule layer also contains symmetrical dunelike bedforms that we interpret as wave-generated structures. These structures are unique to the Dales Gorge spherule layer; no other strata in the Dales Gorge Member show evidence of wave activity. Where we observed these structures, the layer is composed entirely of coarse to This content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions 10 S . W . H A S S L E R A N D B . M . S I M O N S O N Figure 6. Representative measured sections of Griqualand West Basin spherule layers. Shaded areas indicate the spherule layers. Section MF contains abundant locally derived intraclasts and shows hummocky cross stratification, reflecting seafloor erosion and deposition by impact-generated tsunami waves. Sections RF1 and P11 (both from locality PM; shown on fig. 3) consist of spherules admixed with continentally derived material, suggesting they represent the deposits of syntsunami to posttsunami offshore-directed flows. Dashed lines indicate correlative horizons. Bar scale is the same as in figure 5. very coarse sand-sized spherules; no rip-up clasts are present. The ripples show wavelengths ranging from 39 to 50 cm and heights from 2 to 9 cm. The layer has a sharp planar basal contact, showing no evidence of erosion into the underlying chert bed. Wave deposits are also present in the Monteville spherule layer. At Monteville farm, the upper 2–45 cm of the layer show hummocky cross stratification and bedforms with symmetric crests, which Simonson et al. (1999) interpret as wave generated (fig. 7D). Evidence of Offshore-Directed Bottom Flow. We inThis content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions Journal of Geology I M P A C T E J E C T A I N D E E P - S H E L F E N V I R O N M E N T S 11 terpret the presence of sand-sized quartz and/or feldspar grains in the Monteville, Wittenoom, and Carawine spherule layers as evidence of offshoredirected bottom flow. This flow was initiated as water piled up in coastal areas returned seaward. Three lines of evidence support this interpretation. First, the tectosilicate grains have a continental provenance. Simonson et al. (1998) indicate that quartz and feldspar in the Wittenoom and Carawine spherule layers were eroded from granitic, regionally metamorphosed rocks like those exposed on the adjacent Pilbara Craton. The quartz in the Monteville spherule layer was derived from a continental basement source area, presumably part of the Kaapvaal Craton (Simonson et al. 2000a). Second, tectosilicate sand grains are rare in these formations; their presence in the spherule layers indicates an unusual depositional event. In the Carawine Dolomite, sand-sized quartz grains are present only in the spherule layer. In the Wittenoom Formation, sand-sized quartz and feldspar grains are also present in a number of fallout tuffs, but these grains are petrographically distinct from the ones in the Wittenoom spherule layer (Simonson et al. 1998). The Monteville Formation is capped by a layer of quartzose sandstone (the Motiton Member of Beukes 1987), but this layer is separated from the spherule layer by a minimum of 38 m rich in shale (Simonson et al. 1999). Finally, Simonson et al. (1999) interpret a portion of the Monteville spherule layer as an offshore bottom flow deposit. In cores from the Pering Mine area (PM, fig. 2), the layer consists of an ∼5-cm-thick basal zone of nearly pure spherules overlain by a sandy unit rich in quartz with scattered spherules. Across a northsouth distance of 65 km, this quartz-rich layer decreases from ∼2000 to 5 mm in thickness; the size of the largest quartz grains also decreases from north to south, from 0.6 to 0.05 mm in diameter. We interpret the lower spherule-rich layer as an earlier deposit, likely related to the impact. The presence of quartz and rare spherules in the upper deposit, plus its thickness and grain size changes, indicate that it was formed by a postwave offshoredirected bottom current. The Dales Gorge spherule layer shows evidence of wave action but lacks signs of offshore-directed bottom currents. This may be because this layer formed during a period of sea-level highstand (Simonson and Hassler 1996). At this time, possible source areas for continental material may have been submerged or too distant to influence deposition of the Dales Gorge spherule layer. Evidence of Redeposition by Sediment Gravity Flows. Portions of all four spherule layers show characteristics typical of sediment gravity flows. Each layer is different in its inferred flow and depositional processes, thickness, and abundance of spherules. Besides describing the deposits, this section attempts to highlight these differences. We summarize our previous work on the Wittenoom, Carawine, and Monteville spherule layers and present new data and interpretations for the Dales Gorge spherule layer. In the central and eastern Hamersley Basin, the Wittenoom spherule layer consists of a package of turbidite deposits (Simonson 1992; Simonson and Hassler 1997). These beds show a mixture of normally graded bedding, abundant planar lamination, and climbing ripple cross stratification that represent high- and low-density turbidite deposits (fig. 7E, 7F). The turbidite package is 130 cm at its thickest in the eastern Hamersley Basin, where it consists of three beds. The lower two beds consist largely of sand-sized carbonate and silicate peloids and spherules (Simonson 1992). The spherules are concentrated in a basal centimeter-scale zone at the base of the lowest bed. More rarely, isolated spherules are mixed in with the overlying material. The third, uppermost bed in the sequence is composed entirely of carbonate. It contains no spherules. In the central Hamersley Basin, the turbidite package thins to a single 9-cm-thick carbonate bed that does not contain spherules. This bed overlies the spherule ripple layer described above and is separated from it by up to 7 cm of argillite (fig. 7C). Paleocurrent data from the Wittenoom spherule layer indicate that the spherule-bearing turbidity currents traveled consistently south and west across the Hamersley Basin. The Carawine spherule layer ranges from 12.4 to 24.6 m thick and consists of three debris flow deposits (Simonson and Hassler 1997) (fig. 5). Each deposit shows two internal units. The lower unit varies from 9.8 to 19.4 m thick. It is composed of poorly sorted pebble to boulder-sized clasts set in a recrystallized carbonate matrix. Bedding is massive, and clasts show no preferred orientation. The clast-to-matrix ratio of the lower units varies from 1 : 9 to 1 : 1, suggesting these layers are matrix supported. The upper unit is from 0.6 to 1.7 m thick. It is fine-grained and is composed of the same material as the matrix in the lower unit but lacks clasts. The upper unit shows weak normal grading and discontinuous planar laminations. Spherules are present throughout all three debris flow deposits but are only abundant in the upper fine-grained units (see fig. 5). The large, irregular melt ejecta particles described earlier occur within 2.5 m of the base of the lowest debris flow. This content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions Figure 7. Depositional structures in surface exposures of spherule-rich layers. A, Dales Gorge spherule layer at section WG (fig. 1C). Photograph shows the upper coarse-grained bed and overlying thin graded beds; the basal coarse bed is covered. The visible coarse bed consists of softer material with blocky fracture pattern rich in spherules (e.g., lower right) and resistant, sand- to boulder-sized intraclasts of cherty BIF. Two of the largest boulders are outlined in black; irregular reentrant of boulder on right is due to change in direction of outcrop. Dashed line indicates approximate upper limit of spherule-rich material. The thin graded beds are above this level. Basal contact is hidden beneath rubble at base of outcrop. Layers in the cherty BIF deposited on top of the spherule-rich layer are differentially compacted over the BIF boulders. Rock hammer, 26 cm long, in lower left for scale. B, Basal portion of Monteville spherule layer and underlying thin beds of limestone and shale at section MF (fig. 2B). Dashed line follows basal contact of spherule-rich material; upper contact is more or less below continuous layer of limestone at top of photo. Solid line outlines slab of limestone ∼10 cm thick and 1.8 m long that is both under- and overlain by spherule-rich sediment. The slab was being lifted from the seafloor when deposition ceased. Same layer is present in situ, just to This content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions Journal of Geology I M P A C T E J E C T A I N D E E P - S H E L F E N V I R O N M E N T S 13 the left, where it is truncated by a step in the basal contact. C, Wittenoom spherule layer at section TP (fig. 1B). Symmetric lens below coin consists almost entirely of spherules and is a cross section through one of a chain of evenly spaced bedforms. Bed on which coin rests is distal, fine-grained phase of carbonate turbidite, which lacks spherules at this site. The two layers are separated by a few millimeters to centimeters of weathered shale. Coin is 2.1 cm in diameter. D, Cross section of sample from upper part of Monteville spherule layer at section MF (fig. 2B). Sample consists of coarsely crystalline dolomite and finer spherule-type debris and shows thin, wavy lamination to cross lamination of probable wave origin. Cross section is orthogonal to a symmetrical, hummocky bedform whose crest appears at top. Pencil point is for scale. E, Turbidite phase of the Wittenoom spherule layer at section WG (fig. 1B). Layer is ca. 1 m thick, and the top of the layer is the thin white layer, even with the head of the geologist. The layer is both over- and underlain by shales with interbedded limestones and dolomites, some of which are also turbidites. Geologist is 1.98 m tall. F, Closer view of upper part of turbidite phase showing a partial Bouma sequence in same outcrop as figure 7D. At and below level of coin, sand shows planar lamination (division B) and is rich in spherules, whereas higher up it has relatively few spherules and shows climbing current ripple cross lamination (division C) and then consists of finer, thinly laminated carbonate (divisions E and possibly D). Diameter of coin is 2.1 cm. We have also recognized a spherule layer in a core located halfway between the edges of the Carawine and Wittenoom outcrop areas (Simonson and Hassler 1997). This layer is 2.8 m thick and constitutes two to three high-density turbidites. Grain size ranges from fine sand through large pebbles. The commonest sedimentary structures are reverse grading and planar lamination; ripple cross lamination is present but rare. If the Wittenoom and Carawine spherule layers are correlative, this unit may represent a sedimentological and stratigraphic transition between them (Simonson and Hassler 1997). If the two layers were formed by different impacts, this unit could be the upflow equivalent of the Wittenoom spherule layer or the downflow equivalent of the Carawine spherule layer. The Monteville spherule layer shows features typical of sediment gravity-flow deposits at only one locality, a deep borehole drilled on the western edge of the Griqualand West Basin (Simonson et al. 1999). Here, spherules are present in two layers that are separated by approximately 1.8 m of shale and carbonate, which represent typical basinal facies. The stratigraphic arrangement of these layers and the thickness of the lower layer are suspect, due to abnormalities in core preservation (Simonson et al. 1999). The lower layer consists of intraclasts of laminated carbonate and pyrite set in a dark matrix of carbonaceous mud rich in quartzose silt. Spherules are quite rare. Most of the larger clasts in this layer appear to be matrix supported; no distinct stratification is present. The upper spherule layer is 60 cm thick. It is a composite of seven closely spaced sandy layers that both thin and fine upward. Spherules are abundant only in the basal 18 cm of this layer and even there are subordinate to sandsized intraclasts of carbonate, shale, and pyrite. We interpret the matrix-rich lower layer as the product of a mass movement such as a debris flow and the well-sorted, laminated, sandy upper layer as the product of either a turbidity current and/or wave action in deep water. If undisturbed, the presence of almost 2 m of fine-grained strata between the layers indicates a fair amount of time elapsed between their deposition. In the central Hamersley Basin, the Dales Gorge spherule layer consists of 2-dm-scale coarse-grained beds overlain by a series of partial Bouma sequences (fig. 5). The lowest coarse bed rests on an erosive substrate contact and consists of argillite intraclasts and 20% spherules (visual estimate). The intraclasts are matrix supported and show weak coarse tail grading, changing from flat pebbles at the base to very fine sand at the top. The top of this bed is erosionally truncated and overlain by the second coarse bed, which is massive and composed of pebble-sized argillite and banded iron formation (BIF) intraclasts and very rare spherules. In both core and outcrop, this second coarse bed contains cobble and boulder-sized intraclasts, which are concentrated at the top of the bed. They are supported by the smaller clasts around them. The overlying partial Bouma sequences either truncate against or drape over the outsized clasts. Twodecimeter-scale, normally graded beds cap the uppermost partial Bouma sequence. Both are composed of fine sand to silt-sized clasts of argillite with very rare chert, BIF, and carbonate grains. Spherules are not present. These two beds form the uppermost part of the Dales Gorge spherule layer throughout its exposure area; they also top the Dales Gorge wave ripples described earlier. The matrix-supported basal layer, boulders and cobbles “floating” above the base of the bed conThis content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions 14 S . W . H A S S L E R A N D B . M . S I M O N S O N Figure 8. Sketches showing synthesized depositional history for the Hamersley and Transvaal Basin spherule layers. The thickness of the deposits is exaggerated. A, Ejecta (spherules plus coarser tektite-like particles) are delivered to the basin after the impact. The particles settle through the water column and form a fallout layer on the seafloor. Due to reworking, none of the Hamersley or Transvaal Basin layers record this event; we infer its occurrence as the simplest way to transport impact-generated material from the target area to the basin. B, Within hours to tens of hours, depending on the distance to the target area, impact-generated tsunami waves arrive in the basin, in some areas causing substrate erosion and in others reworking the impact fallout layer into wave ripples. We infer that erosion took place shoreward of ripple formation where the effects of wave energy on the seafloor would be greater. C, Synwave to postwave offshore-directed flow transports loose intrabasinal sediment and continental debris offshore and reworks preexisting deposits. Hummocky cross stratification may have formed during this phase or earlier during B. HCS, hummocky cross stratification. D, Sediment gravity flows, either debris flows and/or turbidity currents, erode and redeposit the preexisting impact deposits of A through C. These flows may represent postimpact mass movements of marine sediment weakened by the impact or slumping of shelf sediment piled up by impact-induced waves. taining them, and the capping partial Bouma sequences suggest that this part of the Dales Gorge spherule layer was deposited by bipartite sediment gravity flows. These flows were likely composed of, respectively, a lower phase dominated by grainto-grain interactions, corresponding to the basal massive layer, and an upper turbulent phase, indicated by the outsized rip-up clast-rich zone and the overlying partial Bouma sequences. The rip-up clasts were eroded upflow and transported within the subsequently developing bipartite flows. The dual fine-grained normally graded beds that cap the Dales Gorge spherule layer represent dilute settling of fine-grained material after deposition of the main flows. The presence of these beds on top of the symmetrical spherule-rich bedforms suggests that this dilute material traveled farther than the underlying coarser-grained flows. Depositional Sequence and Depositional Histories The features presented in “Criteria for Recognizing Impact Deposits” demonstrate that a consistent suite of processes is involved in the formation of the Hamersley and Griqualand West Basin spherule layers. These processes are substrate erosion, wave activity, offshore-directed bottom flows, and sediment gravity flows. Collectively, these processes record anomalously coarse-grained, high-energy sedimentation in deep-shelf, quiet-water environments. The order of these processes is remarkably consistent from layer to layer. They demonstrate a recurring sequence of (1) wave activity and substrate erosion, (2) synwave to postwave offshoredirected flow, and (3) sediment gravity flows, often with additional substrate erosion. We propose that this sequence itself is diagnostic of distal impact deposits when it is preserved in deep-shelf successions. We suggest that this sequence may be analogous to the Bouma sequence, which describes a consistent series of depositional structures formed by low-density turbidity currents. The sequence we describe may be similarly indicative of marine impact deposits. Figure 8 shows a synthetic depositional history for the spherule layers we have studied that demonstrates how we believe this sequence formed. The components of our proposed depositional sequence differ among the spherule layers in two ways. First, the relative importance of each process varies from one layer to the next and even from one place to another when a given layer can be traced for distance on the order of 100 km. Second, not every layer shows evidence for all the processes in the sequence. Each of the layers we have discussed thus records the history of a specific impact and its subsequent sedimentological events. To document this, the remainder of this section presents a depositional history for each spherule layer. This will demonstrate the consistency of the sequence as well as showing its variability. A necessary first step in the history of all four layers was the delivery of the spherules and other melt ejecta from the impact target area to the depositional basin. Such an event should form a fallout layer if deposited below wave base on the seafloor, but none of the four spherule layers show the features one would expect of such a layer, such as simple distribution-graded bedding. This leaves two possibilities: either the ejecta were deposited initially in a fallout layer that was completely reworked by the processes we have documented or reworking began while the ejecta were still arriving and settling through the water column. The Wittenoom spherule layer records a four-part This content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions This content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions 16 S . W . H A S S L E R A N D B . M . S I M O N S O N depositional history. The first event was delivery of the spherules into the Hamersley Basin. Second, was formation of the symmetrical ripples preserved in the central Hamersley Basin by what could have been impact-generated waves (Hassler et al. 2000). Third, was postwave offshore bottom flow, which transported material into the deep-water part of Hamersley Basin. Finally, in the eastern and central parts of the basin, the layer was eroded and redeposited by a series of sediment gravity flows. Both wave activity and the sediment gravity flows caused substrate erosion. The presence of several centimeters of argillite between the spherule ripples and the overlying sediment gravity flows suggests a period of time, perhaps several thousand years, passed between the wave-related activity and sediment gravity flows. Alternatively, some of the shale could represent suspension sedimentation of fine-grained material after the passage of tsunami, in a period of days rather than years. The Monteville spherule layer shows a four-part history similar to the Wittenoom spherule layer. The first event was delivery of the spherules into the Griqualand West Basin. Second, was redeposition of the spherules by wave-related processes and sediment gravity flows (Simonson et al. 1999). The timing of these two processes is more ambiguous than in the Wittenoom spherule layer. Third, in the Monteville spherule layer, wave and sediment gravity-flow deposits occur in separate areas with little or no overlap. Finally, substrate erosion resulted from both wave and sediment gravity-flow activity. The stratigraphic features of the Dales Gorge spherule layer indicate a three-part depositional history. The first event, as in the Wittenoom spherule layer, was spherule delivery into the Hamersley Basin. The second event was wave reworking of the spherules into symmetrical bedforms. The deposits left by this event are preserved in most of the outcrop area of the Dales Gorge spherule layer. The third and final event was redeposition of the waveformed beds into a series of decimeter-scale, compositionally heterogeneous turbidites, along with extensive substrate erosion. This event took place only in the central Hamersley Basin. The two finegrained, normally graded beds that cap the Dales Gorge spherule layer throughout its outcrop area represent dilute settling of fine-grained material after deposition by the main turbidity currents. The presence of these beds on top of the symmetrical spherule-rich bedforms, as well, suggests that this dilute material traveled farther than the underlying coarser-grained flows. The direct contact between the ripples and these thin graded beds suggests that the second and third events took place in rapid succession. Insufficient time passed for any significant accumulation of shale. The Carawine spherule layer shows only a twostage history. The first event was delivery of spherules and large irregular melt ejecta to the Hamersley Basin. This was followed by substrate erosion and the transport of the spherules and other ejecta by a series of closely spaced debris flows. If any wave structures were formed, they were subsequently completely reworked throughout the known area of occurrence. The presence of continentally derived tectosilicate grains in the layer suggests that offshore-directed flow, possibly due to impact-induced wave activity, may have occurred before the debris flows were active. Discussion: Causes of Events Although we can reliably identify wave and sediment gravity-flow deposits in the spherule layers, the generating causes for these processes are harder to decipher. Three mechanisms have been proposed for the generation of large waves by impacts. Most commonly cited is the creation of wave systems by the collapse of the transient water crater produced during an oceanic impact (Strelitz 1979). Other suggestions include impact-triggered seismic activity leading to submarine landsliding and water displacement by ejecta (Smit et al. 1996; Wallace et al. 1996). Alternatively, recent models suggest that oceanic impacts may trigger hypercanes that could produce anomalously large waves (Emanuel et al. 1995). Distinguishing which of these mechanisms may have been responsible for the waves associated with the Hamersley and Griqualand West Basin layers may be an impossible task due to preservational constraints and the lack of modern analogues for impact-generated tsunami. Whichever mechanism(s) was responsible for tsunami generation, the offshore-directed flow deposits in the spherule layers resulted from the back surge of water piled up in coastal areas by tsunami, such as described by Cita et al. (1996). Impact-related sediment gravity flows also have at least three causal mechanisms. First, like wave systems, they may be seismically induced. Second, postimpact mass movement of marine sediment weakened by the impact may generate flows. Finally, slumping of shelf sediment piled up by impact-induced waves may also evolve into flows. For all four spherule layers, the first mechanism is unlikely; impact-related seismicity should trigger sediment gravity flows at the same time that wave systems are active. This indicates that their deposits should be interbedded, or depending on disThis content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions Journal of Geology I M P A C T E J E C T A I N D E E P - S H E L F E N V I R O N M E N T S 17 tance from the impact, sediment gravity-flow deposits may form first and be reworked by wave activity. In contrast, all of the Hamersley and Griqualand West Basin spherule layers show sediment gravity-flow activity after wave action and its related offshore-directed flow. Distinguishing between wave-related flows and postimpact slumping is more difficult; both mechanisms could produce the sequence of events we have observed. However, for the Dales Gorge spherule layer, sediment gravity flows occurred shortly after wave action, suggesting that they may be wave-related. In contrast, a time gap of up to several thousand years probably occurred between these processes in the Wittenoom spherule layer, indicating postimpact slumping as a likely mechanism. The lack of waveformed deposits in the Carawine spherule layer makes it difficult to distinguish which, if any, of these mechanisms caused the debris flows of this layer. Based on the presence of possible planar deformation structures and replaced Muong Nong-type tektites, Simonson et al. (2000b) argue that the Carawine spherule layer resulted from an impact into continental crust. In contrast, we suggest that the Wittenoom, Dales Gorge, and Monteville spherule layers resulted from large oceanic impacts. The abundance of spherules indicates large impacts. Only impacts that create craters tens to hundreds of kilometers in diameter appear to be capable of producing large volumes of melt ejecta and distributing it over a large area (Glass and Pizutto 1994; Smit et al. 1996; Hassler et al. 2000). As discussed earlier, oceanic target areas are suggested by the absence of continentally derived shocked material in these three layers and the presence of basaltic textures in some spherules (Simonson and Harnik 2000). In addition, the smaller volume of continental crust in the Neoarchean to Paleoproterozoic (Eriksson 1995) makes an oceanic impact more statistically likely than at present. We suggest that these target areas were in the deep ocean, because impacts in shelf areas would be unlikely to produce the extensive wave systems such as those documented in the Hamersley and Griqualand West Basins. Conclusions 1) Based on melt ejecta spherule occurrences, we recognize a minimum of three and a maximum of four reworked spherule layers in the Hamersley Basin, Western Australia, and one in the Griqualand West Basin, South Africa. The layers range from Neoarchean to Paleoproterozoic in age. 2) All these spherule layers are anomalously coarse-grained beds hosted by muddy sediments deposited below storm wave base in deep-shelf environments. We therefore interpret all of these layers as high-energy events in a dominantly lowenergy setting. 3) All of the layers show the effects of two or more different types of processes, including (a) substrate erosion and intraclasts formation, (b) reworking of ejecta by impact-related waves, (c) basinward transport of continental material by synwave to postwave bottom flow, and (d) reworking of earlier deposits by sediment gravity flows. 4) If our sampling is representative, we predict most distal ejecta layers from oceanic impacts deposited in marine shelf settings should (a) record stratigraphically anomalous coarse-grained, highenergy sedimentation events, (b) document substrate erosion through the presence of rip-up clasts and/or basal scour surfaces, and (c) show evidence of reworking by anomalously large wave systems. In addition, they commonly (d) show evidence of redeposition by sediment gravity flows. These flows may or may not closely follow wave activity in time. 5) We propose this as a model for workers to test on other spherule layers to see if the model has general applicability. This may be more difficult in Phanerozoic deposits; the prevalence of bioturbation since the Neoproterozoic may obscure the mechanical structures of impact deposits. In addition, the abundance and importance of wave structures may be less, if the continents are growing in size relative to the ocean basins, as some believe. ACKNOWLEDGMENTS The National Geographic Society and Oberlin College supported fieldwork. We thank CRA Exploration (especially Adrienne Meakins and Andrew Connor) for permission to examine and sample cores in Australia. We thank Eugene Siepker and J. Brouwer (Gold Fields of South Africa), Frans Dooge (Pering Mine), and H. Peter Siegfried (Geological Survey of South Africa) for access to cores. We also are grateful to Derrick and Anthea Shaw for access to outcrops on Monteville farm in South Africa and Dan and Rose Daniel for hospitality in Dampier, Western Australia. Brooke Wilkerson and Jean Hetherington provided valuable field assistance. Miriam Hornstein provided laboratory help; Darian Davies, Paul Harnik, and Paul Jambor provided both field and laboratory help. This content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions 18 S . W . H A S S L E R A N D B . M . S I M O N S O N REFERENCES CITED Altermann, W. 1996. Sedimentology, geochemistry and palaeogeographic implications of volcanic rocks in the Upper Archaean Campbell Group, western Kaapvaal craton, South Africa. Precambrian Res. 79:73–100. Altermann, W., and Nelson, D. R. 1998. Sedimentation rates, basin analysis and regional correlations of three Neoarchaean and Paleoproterozoic sub-basins of the Kaapvaal Craton as inferred from precise U-Pb zircon ages from volcaniclastic sediments. Sediment. Geol. 120:225–256. Arndt, N. T.; Nelson, D. R.; Compston, W.; Trendall, A. F.; and Thorne, A. M. 1991. The age of the Fortescue Group, Hamersley Basin, Western Australia, from ion microprobe zircon U-Pb results. Aust. J. Earth Sci. 38: 261–281. Beukes, N. J. 1983. Ooids and oolites of the Proterophytic Boomplaas Formation, Transvaal Supergroup, Griqualand West, South Africa. In Peryt, T. M., ed. Coated grains. Berlin, Springer, p. 199–214. ———. 1987. Facies relations, depositional environments and diagenesis in a major early Proterozoic stromatolitic carbonate platform to basinal sequence, Campbellrand Subgroup, Transvaal Supergroup, southern Africa. Sediment. Geol. 54:1–46. Bohor, B. F., and Glass, B. P. 1995. Origin and diagenesis of K/T spherules—from Haiti to Wyoming and beyond. Meteoritics 30:182–198. Boulter, C. A. 1987. Subaqueous deposition of accretionary lapilli: significance for palaeoenvironmental interpretations in Archaean greenstone belts. Precambrian Res. 34:231–246. Button, A. 1976. Transvaal and Hamersley Basins— review of basin development and mineral deposits. Miner. Sci. Eng. 8:262–293. Cheney, E. S. 1996. Sequence stratigraphy and plate tectonic significance of the Transvaal succession of southern Africa and its equivalent in Western Australia. Precambrian Res. 79:3–24. Cita, M. B.; Camerlenghi, A.; and Rimoldi, B. 1996. Deepsea tsunami deposits in the eastern Mediterranean: new evidence and depositional models. Sediment. Geol. 104:155–173. Crimes, T. P., and Droser, M. L. 1992. Trace fossils and bioturbation: the other fossil record. Annu. Rev. Ecol. Syst. 23:339–360. Dimroth, E., and Lichtblau, A. P. 1979. Metamorphic evolution of Archean hyaloclastites, Noranda area, Quebec, Canada. I. Comparison of Archean and Cenozoic sea-floor metamorphism. Can. J. Earth Sci. 16: 1315–1340. Emanuel, K. A.; Speer, K.; Rotunno, R.; Srivastava, R.; and Molina, M. 1995. Hypercanes: a possible link in global extinction scenarios. J. Geophys. Res. 100: 13,755–13,765. Eriksson, K. A. 1995. Crustal growth, surface process, and atmospheric evolution on the early Earth. In Coward, M. P., and Ries, A. C., eds. Early Precambrian processes. Geol. Soc. Lond. Spec. Publ. 95:11–25. Ewers, W. E., and Morris, R. C. 1981. Studies of the Dales Gorge Member of the Brockman Iron Formation, Western Australia. Econ. Geol. 76:1929–1953. Glass, B. P., and Pizzuto, J. E. 1994. Geographic variation in Australasian microtektite concentrations: implications concerning the location and size of the source crater. J. Geophys. Res. 99:19,075–19,081. Grieve, R. A. F. 1998. Extraterrestrial impacts on earth: the evidence and the consequences. In Grady, M. M.; Hutchison, R.; McCall, G. J. H.; and Rothery, D. A., eds. Meteorites: flux with time and impact effects. Geol. Soc. Lond. Spec. Publ. 140:105–131. Hassler, S. W. 1993. Depositional history of the Main Tuff Interval of the Wittenoom Formation, Late Archean–Early Proterozoic Hamersley Group, Western Australia. Precambrian Res. 60:337–359. Hassler, S. W.; Robey, H. F.; and Simonson, B. M. 2000. Bedforms produced by impact-generated tsunami, ∼2.6 Ga Hamersley basin, Western Australia. Sediment. Geol. 135:283–294. Klein, C., and Beukes, N. J. 1989. Geochemistry and sedimentology of a facies transition from limestone to iron-formation deposition in the early Proterozoic Transvaal Supergroup, South Africa. Econ. Geol. 84: 1733–1774. Klein, C.; Beukes, N. J.; and Schopf, J. W. 1987. Filamentous microfossils in the early Proterozoic Transvaal Supergroup: their morphology, significance, and paleoenvironmental setting. Precambrian Res. 36:81–94. Lofgren, G. E. 1971. Devitrified glass fragments from Apollo 11 and Apollo 12 lunar samples. In Proc. 2d Lunar Sci. Conf., vol. 1, p. 949–955. ———. 1977. Dynamic crystallization experiments bearing on the origin of textures in impact-generated liquids. In Proc. 8th Lunar Sci. Conf, p. 2079–2095. Myers, J. S., and Hocking, R. M. (compilers). 1988. Geological map of Western Australia. Geol. Surv. Western Australia, Perth. Scale 1 : 2,500,000. Nelson, D. R.; Trendall, A. F.; and Altermann, W. 1999. Chronological correlations between the Pilbara and Kaapvaal cratons. Precambrian Res. 97:165–189. Simonson, B. M. 1992. Geological evidence for a strewn field of impact spherules in the early Precambrian Hamersley Basin of Western Australia. Geol. Soc. Am. Bull. 104:829–839. Simonson, B. M.; Davies, D.; and Hassler, S. W. 2000a. Discovery of a layer of probable impact melt spherules in the late Archean Jeerinah Formation (Fortescue Group, Western Australia) Aust. J. Earth Sci. 47: 315–325. Simonson, B. M., and Harnik, P. 2000. Have distal impact This content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions Journal of Geology I M P A C T E J E C T A I N D E E P - S H E L F E N V I R O N M E N T S 19 ejecta changed through geologic time? Geology 28: 975–978. Simonson, B. M., and Hassler, S. W. 1996. Was the deposition of large Precambrian iron formations linked to major marine transgressions? J. Geol. 104:665–676. ———. 1997. Revised correlations in the Early Precambrian Hamersley Basin based on a horizon of resedimented impact spherules. Aust. J. Earth Sci. 44:37–48. Simonson, B. M.; Hassler, S. W.; and Beukes, N. J. 1999. A Late Archean impact spherule layer in South Africa that may correlate to Australia. In Dressler, B. O., and Sharpton, V. L., eds. Impact cratering and planetary evolution, II. Geol. Soc. Am. Spec. Pap. 339, p. 249–261. Simonson, B. M.; Hassler, S. W.; Davies, D.; Wallace, M.; and Reeves, S. 1998. Iridium anomaly but no shocked quartz from Late Archean microkrystite layer: oceanic impact ejecta? Geology 26:195–198. Simonson, B. M.; Hassler, S. W.; and Schubel, K. A. 1993a. Lithology and proposed revisions in stratigraphic nomenclature of the Wittenoom Formation (Dolomite) and overlying formations, Hamersley Group, Western Australia. Geol. Surv. W. Aust. Rep. 34, Prof. Pap., p. 65–79. Simonson, B. M.; Hornstein, M.; and Hassler, S.W. 2000b. Particles in late Archean Carawine Dolomite (Western Australia) resemble Muong Nong-type tektites.In Gilmour, I., and Koeberl, C., eds. Impacts and the Early Earth. Lecture notes in earth sciences. Heidelberg, Springer, p. 181–214. Simonson, B. M., and Jarvis, D. G. 1993. Microfabrics of oolites and pisolites in the early Precambrian Carawine Dolomite of Western Australia. In Rezak, R., and Lavoie, D., eds. Carbonate microfabrics. Heidelberg, Springer, p. 227–237. Simonson, B. M.; Koeberl, C.; McDonald, I.; and Reimold, W. U. 2000c. Geochemical evidence for an impact origin for a Late Archean spherule layer, Transvaal Supergroup, South Africa. Geology 28:1103–1106. Simonson, B. M.; Schubel, K. A.; and Hassler, S. W. 1993b. Carbonate sedimentology of the early Precambrian Hamersley Group of Western Australia. Precambrian Res. 60:287–335. Smit, J.; Roep, T. B.; Alvarez, W.; Montanari, A.; Claeys, P.; Grajales-Nishimura, J. M.; and Bermudez, J. 1996. Coarse-grained, clastic sandstone complex at the K/T boundary around the Gulf of Mexico: deposition by tsunami waves induced by the Chicxulub impact? Geol. Soc. Am. Spec. Pap. 307, p. 151–182. Strelitz, R. 1979. Meteorite impact into the ocean. In Proc. 10th Lunar Planet. Sci. Conf., p. 2799–2813. Trendall, A. F. 1983. The Hamersley Basin. In Ironformations: facts and problems. Amsterdam, Elseiver, p. 69–129. Trendall, A. F., and Blockley, J. 1970. The iron formations of the Precambrian Hamersley Group of Western Australia. Geol. Surv. W. Aust. Bull. 119, 366 p. Trendall, A. F.; Nelson, D. R.; de Laeter, J. R.; and Hassler, S. W. 1998. Precise zircon U-Pb ages from the Marra Mamba Iron Formation and Wittenoom Formation, Hamersley Group, Western Australia. Aust. J. Earth Sci. 45:137–142. Wallace, M. W.; Gostin, V. A.; and Keays, R. R. 1996. Sedimentology of the Neoproterozoic Acraman impactejecta horizon, South Australia. Aust. Geol. Surv. Org. J. Aust. Geol. Geophys. 16:443–451. Woodhead, J. D.; Hergt, J. M.; and Simonson, B. M. 1998. Isotopic dating of an Archean bolide impact horizon, Hamersley Basin, Western Australia. Geology 26: 47–50. This content downloaded from 129.100.58.76 on Mon, 04 Jan 2016 01:46:08 UTC All use subject to JSTOR Terms and Conditions