Early Archean Asteroid Impacts on Earth

Home , Imbrian

dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 1 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

Earth’s Oldest Rocks Edited by Martin J. van Kranendonk, R. Hugh Smithies and Vickie C. Bennett Developments in Precambrian Geology, Vol. 15 (K.C. Condie, Series Editor) 1 © 2007 Elsevier B.V. All rights reserved. DOI: 10.1016/S0166-2635(07)15085-4

1 Chapter 8.5 1 2 2 3 3 4 EARLY ARCHEAN ASTEROID IMPACTS ON EARTH: 4 5 STRATIGRAPHIC AND ISOTOPIC AGE CORRELATIONS 5 6 6 7 AND POSSIBLE GEODYNAMIC CONSEQUENCES 7 8 8 9 ANDREW GLIKSON 9 10 Department of Earth and Marine Science and Planetary Science Institute, Australian 10 11 National University, Canberra, ACT 0200, Australia 11 12 12 13 13 14 14 15 8.5-1. INTRODUCTION 15 16 16 17 The heavily cratered surfaces of the terrestrial planets and moons testify to their long- 17 18 term reshaping by asteroid and comet impacts and related structural and melting processes. 18 19 Following accretion of Earth from the solar disc at ca. 4.56 Ga (see Taylor, this volume), 19 20 the Earth–Moon system is believed to have originated by a collision between a Mars- 20 21 size planet and Earth, followed by episodic bombardment by asteroids and comets, with a 21 22 peak documented at ca. 3.95–3.85 Ga – the Late Heavy Bombardment (LHB) – recorded 22 23 on the Moon, but not on Earth due to the paucity and high grade metamorphic state of 23 24 terrestrial rocks of this age (Wilhelm, 1987; Ryder, 1990, 1991, 1997; see Iizuka et al., 24 25 this volume). Major impact episodes on Earth are recorded from rocks dated at 3.46, 25 26 3.26–3.24, 2.63, 2.56, 2.48, 2.02 Ga (Vredefort), 1.85 Ga (Sudbury), 0.58 Ga (Acraman), 26 27 0.368–0.359 Ga (late Devonian), 0.214 Ga (late Triassic), 0.145–0.142 Ga (late Jurassic), 27 28 0.065 Ga (K–T boundary), and 0.0357 Ga (late-Eocene) (World Impact List GSC/UNB; 28 29 Grieve and Shoemaker, 1994; Grieve and Pesonen, 1996; Grieve and Pilkington, 1996; 29 30 French, 1998; Simonson and Glass, 2004). Some of these impact events coincided with, 30 31 and are referred to as the direct cause of, mass extinctions and/or the radiation of species 31 32 (Alvarez, 1986; Keller, 2005; Glikson, 2005a). These impacts represent a small part of 32 33 the extraterrestrial impact record, as the retention of any impact-generated structure is a 33 34 direct function of the preservation and/or destruction of crust. By analogy, the continu- 34 35 ous re-accretion of sulphur on Io and ice on Europa obliterates their impact records, as 35 36 does volcanic re-surfacing on Venus and plate tectonics on Earth. Estimates of the impact 36 37 incidence on Earth is inferred from the lunar cratering record, and the current asteroid 37 38 and comet ﬂux, which suggests the formation of some 100–200 craters of diameter (D) 38 39 D 100 km, and 5–10 craters of D 300 km since ca. 3.8 Ga (Neukum et al., 2001; 39 40 Ivanov and Melosh, 2003; Glikson et al., 2004). Most of these craters would have impacted 40 41 on hitherto subducted oceanic crust on Earth. The larger continental impacts are less likely 41 42 to be destroyed by subduction, and resulted in deeper excavation of the continental crust 42 43 and thicker ejecta layers, thereby being generally better preserved than their counterparts 43 dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 2 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

2 Chapter 8.5: Early Archean Asteroid Impacts on Earth

1 that impacted oceanic crust. Glikson (2001) estimated that only between 8–17% of the 1 2 larger impacts (D 100 km) is recorded as impact structures and/or ejecta. However, as- 2 3 suming a relatively smaller proportion of large projectiles, over 50% of the larger impacts 3 4 have been recorded at this stage (Ivanov and Melosh, in Glikson et al., 2004). 4 5 Models suggest impact triggering of igneous, tectonic, or plate tectonic events (Green, 5 6 1972, 1981; Grieve, 1980; Alt et al., 1988; Jones et al., 2002; Elkins-Tanton et al., 2005; 6 7 Glikson and Vickers, 2005), but this correlation has not been directly demonstrated on 7 8 Earth. Ivanov and Melosh (2003) questioned impact-induced volcanic activity, whereas 8 9 Glikson et al. (2004) state: ‘Indeed, a few giant impact events certainly occurred on 9 10 Earth in post-LHB history, and a few might have encountered the fortuitous circumstances 10 11 necessary for them to enhance magma production. This is an important topic for future 11 12 study.’ Lowe et al. (1986, 1989) remarked on the potential tectonic significance of the co- 12 13 incidence between 3.26–3.24 Ga impact horizons recorded in the Barberton Greenstone 13 14 Belt (Kaapvaal Craton) and the abrupt the break between the underlying mafic–ultramafic 14 15 volcanic crust (3.55–3.26 Ga Onverwacht Group) and the overlying turbidite, felsic vol- 15 16 canic, banded iron formation and conglomerate-dominated succession (3.26–3.225 Ga Fig 16 17 Tree and Moodies Groups), which includes the earliest observed granite-derived detritus 17 18 (Fig. 8.5-1). This age period correlates broadly with a similar break in the Pilbara Craton, 18 19 Western Australia, where the mafic–ultramafic volcanic rocks of the ca. 3.255–3.235 Ga 19 20 Sulphur Springs Group are unconformably overlain by Soanesville Group clastic sedimen- 20 21 tary rocks, including a local unit of olistostrome breccia (Glikson and Vickers, 2005; Van 21 22 Kranendonk et al., 2006a). Widespread 3.27–3.24 Ga plutonic igneous rocks (dominantly 22 23 granitic) occur in both the Kaapvaal Craton and Pilbara Craton (see Poujol, this volume, 23 24 and Van Kranendonk et al., 2007a, this volume). 24 25 25 26 26 27 8.5-2. 3.47 GA IMPACT EVENTS 27 28 28 29 Microkrystite spherule-bearing lenses in chert and in the matrix of diamictite have been 29 30 reported from the 3.47 Ga Antarctic Creek Member (ACM) of the Mount Ada Basalt, in 30 31 the North Pole Dome area of the East Pilbara Terrane, Pilbara Craton, Western Australia 31 32 (Lowe et al., 1989; Byerly et al., 2002; Van Kranendonk et al., 2006a). The diamictite, 32 33 defined as ACM-S2, consists of a 0.6–0.8 m thick unit of spherule-bearing, chert-intraclast 33 34 conglomerate, which is separated from the main ACM-S3 unit by dolerite and felsic vol- 34 35 canics/hypabyssals. Zircon dating from the spherule-bearing sandstone unit yielded an 35 36 age of 3470 ± 2 Ma (Byerly et al., 2002). The microkrystite spherules are discrimi- 36 37 nated from angular to subangular detrital volcanic fragments by their high sphericities, 37 38 inward-radiating fans of sericite, relic quench textures, and Ni–Cr–Co relations. SEM- 38 39 EDS (scanning electron microscope coupled with electron-dispersive spectrometer) and 39 40 LA-ICP-MS (laser induced coupled plasma mass spectrometer) analysis indicate high Ni 40 41 and Cr in sericite-dominated spherules, suggesting a mafic composition of source crust. 41 42 Ni/Cr and Ni/Co ratios of the spherules are higher than in associated Archaean tholeiitic 42 43 basalts and high-Mg basalts, rendering possible contamination by high Ni/Cr and Ni/Co 43 dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 3 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

8.5-2. 3.47 Ga Impact Events 3

1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 10 11 11 12 12 13 13 14 14 15 15 16 16 17 17 18 18 19 19 20 20 21 21 22 22 23 23 24 24 25 25 26 26 27 27 28 28 29 29 30 30 31 31 32 32 33 33 34 34 35 35 36 36 37 37 38 38 39 39 40 40 41 Fig. 8.5-1. Composite columnar section of the Barberton greenstone belt stratigraphy, showing the 41 42 principal groups and formations and U-Pb zircon isotopic age determinations (after Byerly and Lowe, 42 43 1994; Byerly et al., 1999; Lowe et al., 2003). 43 dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 4 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

4 Chapter 8.5: Early Archean Asteroid Impacts on Earth

1 chondritic components. The presence of multiple bands and lenses of spherules within 1 2 chert, and scattered spherules in arenite bands within ACM-S3, may signify redeposition 2 3 of a single impact fallout unit or, alternatively, multiple impacts. Controlling parameters 3 4 include: (1) spherule atmospheric residence time; (2) precipitation rates of colloidal silica; 4 5 (3) solidification rates of colloidal silica; (4) arenite and spherule redeposition rates; and (5) 5 6 arrival of an impact-generated tsunami. The presence of spherule-bearing chert fragments 6 7 in ACM-S3 may hint at an older spherule-bearing chert (?S1). Only a minor proportion of 7 8 spherules is broken and the near-perfect sphericities of chert-hosted spherules and arenite- 8 9 hosted spherules constrain the extent of shallow water winnowing of the originally delicate 9 10 glass spherules. It is suggested that the spherules were either protected by rapid burial or, 10 11 alternatively, disturbance was limited to a short term high energy perturbation such as may 11 12 have been affected by a deep-amplitude, impact-triggered tsunami wave. 12 13 An impact spherule unit precisely correlated with spherules of the ACM occurs in a 13 14 30 cm to 3 m thick chert unit, H4c in the upper part of the Hooggenoeg Formation of the 14 15 Onverwacht Group, Kaapvaal Craton (Lowe et al., 2003). Zircons in this unit have been 15 16 dated at 3470 ± 6.3 Ma, identical in age to the Pilbara unit (Byerly et al., 2002). The 16 17 spherules constitute a bed of medium- to coarse-grained, current-deposited sandstone, 10– 17 18 35 cm thick and interbedded with fine-grained tuffs and black, or black-and-white, banded 18 19 cherts. In sections where H4c is ∼100 cm thick, the unit generally rests directly on, or just 19 20 a few centimeters above, altered basaltic to komatiitic volcanic rocks. In the type section, 20 21 H4c is 2.8 m thick, but varies from ∼50 cm to 6 m thick within 500 m along strike. In this 21 22 section, the unit is underlain by up to 160 cm, and overlain by 90 cm, of pale greenish to 22 23 light gray sericitic chert and black carbonaceous chert. 23 24 24 25 25 26 8.5-3. THE 3.26–3.24 GA BARBERTON IMPACT CLUSTER 26 27 27 28 The uppermost, predominantly mafic–ultramafic volcanic sequence of the Onverwacht 28 29 Group consists of an assemblage of komatiitic volcanics and their hypabyssal and altered 29 30 equivalents, capped by ferruginous chert. This unit is known as the Mendon Formation 30 31 (Fig. 8.5-1) and has been dated by U-Pb zircon from a middle chert unit as 3298 ± 3Ma 31 32 (Byerly, 1999). Unconformably overlying the Mendon Formation is the Mapepe For- 32 33 mation, the basal unit of the Fig Tree Group, consisting of a turbidite–felsic volcanic– 33 34 ferruginous sedimentary rock association dated in the range of 3258±3 Ma to 3225±3Ma 34 35 (Lowe et al., 2003) (Fig. 8.5-1). The S2 impact fallout unit coincides with the base of the 35 36 clastic Mapepe Formation of the Fig Tree Group, whereas the S3 and S4 impact fallout units 36 37 occur within clastic sedimentary rocks and felsic pyroclastic rocks, 110–120 m above the 37 38 basal contact. In places, the S3 unit occurs unconformably above deeply incised Mendon 38 39 Formation komatiites. 39 40 Detailed documentation of the S2,S3 and S4 impact fallout units (Fig. 8.5-1)revealsa 40 41 wide range of diagnostic field, petrographic, mineralogical, geochemical and isotopic cri- 41 42 teria identifying extraterrestrial components mixed with volcanic and other detritus (Lowe 42 43 and Byerly, 1986b; Lowe et al., 1989, 2003; Kyte et al., 1992, 2003; Byerly and Lowe, 43 dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 5 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

8.5-3. The 3.26–3.24 Ga Barberton Impact Cluster 5

1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 10 11 11 12 12 13 13 14 14 15 15 16 16 17 17 18 18 19 19 20 20 21 21 22 22 23 23 24 24 25 25 26 26 27 27 28 28 29 29 30 30 31 31 32 32 33 33 34 34 35 35 36 36 37 37 38 38 39 39 40 40 41 41 42 42 43 43 Fig. 8.5-2. dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 6 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

6 Chapter 8.5: Early Archean Asteroid Impacts on Earth

1 1994; Shukloyukov et al., 2000; Kyte, 2002; Glikson, 2000b, 2000c; Reimold et al., 2000). 1 2 Principal criteria include: (1) impact fallout units that can be correlated in below-wave 2 3 base environments between basins and sub-basins, independently of facies variations of 3 4 the host sedimentary rocks (Lowe et al., 1989, 2003); (2) microkrystite spherules that have 4 5 inward-radiating quench textures and centrally offset vesicles (Glass and Burns, 1988) 5 6 (Fig. 8.5-2), distinct from outward radiating textures of volcanic varioles, which may also 6 7 contain microphenocrysts of feldspar and quartz (Glikson, 2005b, 2005c); (3) microkrys- 7 8 tites that contain quench-textured and octahedral Ni-chromites (NiO < 24%), including a 8 9 high Co, Zn and V variety unknown in terrestrial chromites (Byerly and Lowe, 1994); and 9 10 (4) microkrystites that contain nickel and PGE nanonuggets (Kyte et al., 1992; Glikson 10 11 and Allen, 2004; Glikson, 2005b). The Ni-rich chromites are distinct from terrestrial type 11 12 Ni-chromites associated with sulphide (NiO < 0.3%) (Czamanske et al., 1976; Grove et 12 13 13 al., 1977), and have low Fe2O3/FeO ratios, possibly consequent on condensation under 14 low-oxygen Archaean atmospheric conditions. 14 15 Attempts at estimating the magnitude of the impact events and the composition of the 15 16 target crust are based on inferences derived from spherule radii (O’Keefe and O’Hara, 16 17 1982; Melosh and Vickery, 1991) and geochemical mass balance calculations of as- 17 18 teroid size, based on PGE (Kyte et al., 1992; Byerly and Lowe, 1994) and 53Cr/52Cr 18 19 isotope ratios (Shukloyukov et al., 2000; Kyte et al., 2003; Glikson and Allen, 2004). 19 20 Early estimates based on spherule size distribution (mean 0.85 mm), and the iridium flux 20 21 (5.3 mg/cm2), deduced a 20–30 km diameter size of the parental asteroid (Czamanske 21 22 et al., 1976; Grove et al., 1977). Lowe et al. (1989) referred to the maximum size of 22 23 the microkrystite spherules (<4 mm) to estimate projectile size at 20–50 km in diame- 23 24 ter. From the high mean Ir levels (116 ppb) in impact layer S4, Kyte et al. (1992) suggested 24 25 a FeNi projectile with Ir levels higher than chondrites (∼500 ppb), whereas high mean 25 26 Cr level (1350 ppm) suggest a komatiitic target rock. These authors estimate the Ir-flux 26 27 for S4 at 2.5 gr/cm2 – some 30 times greater than that for the K–T boundary Ir flux 27 28 2 2 28 (0.08 gr/cm ). Byerly and Lowe (1994) estimate an Ir flux of 5.3 gr/cm for the S2 29 29 and S3 impact layers, consistent with a chondritic impactor (Ir 500 ppb) with a diam- 30 30 31 31 32 32 33 Fig. 8.5-2. (Continued.) Microphotograph of microkrystite spherules in sample SA306-1 (A–D) and 33 ∼ 34 SA315-5 (E–H) from the 3.24 Ga S3 impact fallout unit, Mapepe Formation, Barberton green- 34 35 stone belt. The spherules are dominated by chlorite, displaying inward radiating fans, central to 35 offset vesicles and rims of microcrystalline quartz and are set in a micro-fragmental matrix con- 36 36 taining angular clasts of volcanic rock and chert. Plane polarized light. A – spherule with chlorite 37 37 rim, microcrystalline mantle and siliceous core; B – chlorite spherule with central siliceous vesicle, 38 showing LA-ICPMS burn traces; C – spherule consisting of inward-radiating chlorite fans, inter- 38 39 preted as replacing quench textures of original pyroxene and olivine; D – detail of C; E – Spherule 39 40 occupied by radiating palimpsest needles after ?olivine and/or pyroxene representing quench texture; 40 41 F – spherule with inward-radiating chlorite sheafs and a central quartz-filled vesicle; G – spherule 41 42 with inward-radiating ferruginous chlorite fans, light-coloured chlorite mantle and siliceous core; 42 43 H – fragment of spinifex-textured komatiite. 43 dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 7 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

8.5-3. The 3.26–3.24 Ga Barberton Impact Cluster 7

1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 10 11 11 12 12 13 13 14 14 15 15 16 16 17 17 18 18 19 19 20 20 21 21 22 22 23 23 24 24 25 25 26 26 27 27 28 28 29 29 30 30 31 31 32 32 33 33 34 34 35 35

36 ) 36 37 37 38 38 39 39 Continued. 40 40 41 41 42 42 43 43 Fig. 8.5-2. ( dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 8 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

8 Chapter 8.5: Early Archean Asteroid Impacts on Earth

1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 10 11 11 12 12 13 13 14 14 15 15 16 16 17 17 −4 −2 18 Fig. 8.5-3. Correlation between the flux of iridium (in units of 10 mg cm ) and the diameter 18 19 of chondritic projectile (Dp), based on mass balance calculations assuming mean unit thickness, 19 mean Ir concentration and global distribution of fallout ejecta (after Glikson, 2005). Unit symbols 20 20 3 – Rp (projectile radius) = Vp/[4/3]π; Vp = Mp/dp; Mp = FG/Cp; FG = (AE/DS ∗ Ts). 21 21 (AEE – measured element (E) abundance in fallout unit (ppb); DS – mean specific density of fall- 22 3 3 3 22 out sediments (mg/cm ) (assumed as 2.65 gr/cm ); CE – weight of element E in mg per cm 23 23 (CE = AE ∗ DS); TS – mean stratigraphic thickness of spherule unit (in mm); FE – local mean 24 2 2 24 flux of element E in mg per 100 mm (cm ) surface (FE = CE ∗ TS); FG – inferred global flux 2 25 of element E in mg per cm surface; FG = AE/DS ∗ TS ∗ Earth surface area (SE); Cp –as- 25 26 sumed concentration of element E in projectile (ppb) (C1 – Ir-450 ppb; Mp – weight of projectile 26 3 27 (Mp = FG/Cp); dp – assumed specific density of the projectile (C1–3.0gr/cm ); Vp – volume 27 28 of projectile (Vp = Mp/dp). 28 29 29 30 30 31 eter ∼30 km. A FeNi impactor would be 4–10 times smaller, but inconsistent with the 31 32 maximum spherule sizes. From a bimodal size distribution of spherules in the SA306-1 32 33 sample, with a diameter range of 0.4–2.0 mm, Byerly and Lowe (1994) suggested early 33 34 flux of small, Ir-rich, Ni spinel-bearing spherules (mean diameter ∼0.65 mm) and a late 34 35 flux of larger, mostly Ir-poor, spinel-free spherules (mean diameter ∼1.25 mm), suggest- 35 36 ing early ejection of projectile-rich components and a projectile 24 km in diameter. These 36 +3 +2 37 authors showed the SA306-1 impact spinels have higher Ni/Fe and lower Fe /Fe than 37 38 Phanerozoic impact spinels, hinting at a less oxidizing Archaean atmosphere. From Cr 38 39 isotope-based estimates of the global thickness and mean Ir concentration of impact units 39 40 (S2; ∼20 cm, 3 ppb: S3; ∼20 cm, 300 ppb: S4; ∼10 cm, 100 ppb), Kyte et al. (2003) 40 41 suggested a parental asteroid 50–300 times the mass and 3–7 times the diameter of the 41 42 ∼10 km-large K–T Chicxulub asteroid, consistent with estimate made by Glikson (2005b: 42 43 Fig. 8.5-3). 43 dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 9 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

8.5-4. Stratigraphic and Isotopic Age Correlations Between the 3.26–3.24 Ga Impacts 9

1 Given projectile/crater diameter ratios in the range of 0.05–0.1, these impacts would 1 2 result in impact basins at least 300–700 km large. The dominantly mafic to ultramafic com- 2 3 position of the impact ejecta requires that these basins were formed in mafic–ultramafic- 3 4 dominated regions of the Archaean Earth, probably oceanic crust. Impacts of this mag- 4 5 nitude can be expected to have had major consequences on the impacted crustal plates, 5 6 greenstone–granite terrains, and underlying mantle regions. 6 7 7 8 8 9 9 10 8.5-4. STRATIGRAPHIC AND ISOTOPIC AGE CORRELATIONS BETWEEN THE 10 11 3.26–3.24 GA IMPACTS, UNCONFORMITIES, FAULTING AND IGNEOUS 11 12 EVENTS 12 13 13 14 Lowe et al. (1989) observed: ‘The transition from the 300-million-year-long On- 14 15 verwacht stage of predominantly basaltic and komatiitic volcanism to the late orogenic 15 16 stage of greenstone belt evolution suggests that regional and possibly global tectonic re- 16 17 organization resulted from these large impacts.’ Evidence consistent with impact-induced 17 18 effects in the Kaapvaal Craton is provided by U-Pb zircon ages of the Nelshoogte trond- 18 19 hjemite (3236 ± 1 Ma) and the Kaap Valley tonalite (3227 ± 1 Ma), suggesting plutonic 19 20 20 emplacement within about ∼5–10 My after the S3 and S4 impacts (3243 ± 4 Ma). The 21 tonalite-trondhjemite geochemistry of the granitoids suggests partial melting of mafic 21 22 crustal sources. 22 23 23 Close temporal correlations are observed between the Onverwacht Group – Fig Tree 24 24 Group transition and the transition between the Kelly Group (end of felsic magmatism 25 25 at ca. 3290 Ma) and the 3255–3235 Ma Sulphur Springs Group in the East Pilbara Ter- 26 26 rane of the Pilbara Craton (Fig. 8.5-4). In the Sulphur Springs Group, there is a change 27 27 from ∼3255–3235 Ma ultramafic–felsic volcanic rocks and silicified epiclastic rocks of 28 28 29 the Sulphur Springs Group and deposition of overlying sedimentary rocks (banded iron 29 30 formation and turbidites) of the Soanesville Group (Fig. 8.5-5). Eruption of the Sulphur 30 31 Springs Group has been variously interpreted in terms of back arc rifting (Brauhart, 1999), 31 32 or mantle plume events, the latter including local caldera subsidence above a syn-volcanic 32 33 laccolith (Van Kranendonk, 2000; Van Kranendonk et al., 2002; Pirajno and Van Kranen- 33 34 donk, 2005; Smithies et al., 2005b). Whereas in places the boundary above the marker chert 34 35 (silicified epiclastic rocks) at the top of the Sulphur Springs Group appears conformable, it 35 36 was clearly lithified prior to deposition of the overlying, local unit of olistostrome breccia 36 37 (Hill, 1997) and more widespread turbiditic sandstones of the overlying Soanesville Group 37 38 (Van Kranendonk et al., 2006a). 38 39 The following stratigraphic and age relations are outlined between the Barberton Green- 39 40 stone Belt (BGB) and greenstone belts of the East Pilbara Terrane (PGB) (Fig. 8.5-4): 40 41 (A) BGB – S2 impact spherule unit directly underlying 3258 ± 3 Ma felsic tuff. PGB – 41 42 sedimentary rocks of the Leilira Formation (ca. 3255 ± 4 Ma: Buick et al., 2002) at 42 43 the base of the Sulphur Springs Group, but no impact spherules identified; 43 dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 10 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

10 Chapter 8.5: Early Archean Asteroid Impacts on Earth

1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 10 11 11 12 12 13 13 14 14 15 15 16 16 17 17 18 18 19 19 20 20 21 21 22 22 Fig. 8.5-4. Isotopic age correlations between 3.28–3.22 Ga units in the Kaapvaal Craton and Pil- 23 bara Craton. Solid squares and error bar lines – U-Pb ages of volcanic and plutonic units; stars – 23 24 impact ejecta layers; circled crosses – ferruginous sediments; MF – Mapepe Formation; UF – Ulundi 24 25 Formation; NT – Nelshoogte tonalite; KPG – Kaap Valley granite; LF – Leilira Formation; KCF – 25 26 Kangaroo Cave Formation; EPG – Eastern Pilbara granites; NRF – Nickol River Formation; KR – 26 27 Karratha Granodiorite. 27 28 28 29 29 30 (B) BGB – the S3 and S4 impact spherule units of the Mapepe Formation, Fig Tree Group, 30 31 directly overlie 3243 ± 4 Ma felsic tuff. PGB – 3235 ± 3 Ma felsic volcanic rocks 31 32 in the Kangaroo Caves Formation, Sulphur Springs Group, but no impact spherules 32 33 identiﬁed. 33 34 In the West Pilbara Superterrane of the Pilbara Craton, a similar lithological change is 34 35 seen in the Roebourne Group, where komatiite and basalt of the >3270 Ma Ruth Well For- 35 36 mation are overlain by 3.27–3.25 Ga sandstone, shale, banded iron-formation, and felsic 36 37 volcanic rocks of the Nickol River Formation, and accompanied by emplacement of the 37 38 3.27–3.26 Ga Karratha Granodiorite. Indeed, major plutonic activity at ∼3.275–3.225 Ga 38 39 (Fig. 8.5-4), deﬁned as the Cleland Supersuite, is documented throughout much of the 39 40 Pilbara Craton (Van Kranendonk et al., 2006). 40 41 The S2 impact ejecta unit in the south-eastern part of the Barberton greenstone belt is 41 42 overlain directly by 20–30 m thick banded ferruginous chert, the Manzimnyama Jaspilite 42 43 Member, of the basal Mapepe Formation (Lowe and Nocita, 1999). These widespread sed- 43 dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 11 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

8.5-4. Stratigraphic and Isotopic Age Correlations Between the 3.26–3.24 Ga Impacts 11

1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 10 11 11 12 12 13 13 14 14 15 15 16 16 17 17 18 18 19 19 20 20 21 21 22 22 23 23 24 24 25 25 26 26 27 27 28 28 29 29 30 30 31 31 32 32 33 33 34 34 35 35 36 36 37 37 38 38 39 39 40 40 41 Fig. 8.5-5. Schematic section through the top of the Sulphur Springs Group (Kangaroo Caves For- 41 42 mation – 3235 ± 3 Ma), overlying olistostrome, siltstone, ferruginous siltstone and felsic volcanics 42 43 of the Pincunah Hill Formation (after Hill, 1997). 43 dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 12 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

12 Chapter 8.5: Early Archean Asteroid Impacts on Earth

1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 10 11 11 30 km penetration depth of a 20 km

12 ∼ 12 13 13 14 14 15 15 16 16 17 17 18 18 19 19 20 20 21 21 22 22 23 23 24 24 25 25 26 26 27 27 28 28 29 29 30 30 31 31 32 32 33 33 34 34 35 35 36 36 37 37 38 38 39 39 40 40 km).

41 / 41 C 42 ◦ 42 43 43 Fig. 8.5-6. Large blocksPilbara of (after felsic Hill, volcanics, chert 1997).and and oceanic Pressure-temperature siltstone geotherms diagram in and portraying the the olistostrome, the position basal position of Pincunah dry of Hill and the Formation, wet Sulphur lithosphere-asthenosphere peridotite Springs, boundary solidi. Central for Downward arrow continental marks the asteroid, equivalent to decompression and rebound.0.33 Upward arrows mark alternative adiabatic melting paths of mantle diapirs (adiabatic cooling dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 13 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

8.5-5. Possible Geodynamic Consequences 13

1 imentary rocks in the southern part of the Barberton greenstone belt include oxide facies 1 2 BIF, jaspilite, and hematite-rich shale. The S3 impact ejecta unit in the northern part of 2 3 the Barberton greenstone belt underlies ferruginous sediments of the Ulundi Formation, 3 4 including jaspilite, ferruginous chert and black chert deposited under quiet, deep-water 4 5 conditions (Lowe et al., 2003). By contrast, deposition of S2 and S3 impact units along the 5 6 Onverwacht anticline, which formed an antecedent rise located between deeper basins to 6 7 the northwest and southeast, occurred under high-energy, shallow water conditions (Lowe 7 8 et al., 2003). 8 9 In the Pilbara Craton, uppermost stratigraphic levels of the felsic volcanic Kangaroo 9 10 Caves Formation contain BIF, as rafts in olistostrome and as a more widespread unit, up to 10 11 1000 meters thick (Van Kranendonk, 2000). The sequence is disconformably overlain by 11 12 arenites and turbidites and ferruginous shale and siliciﬁed equivalents up to 100 m thick 12 13 in some areas. The potential signiﬁcance of a relationship between impacts and BIF is 13 14 underpinned by similar observations in 3.46, 2.63 and 2.56 Ga impact units (Table 8.5-1; 14 15 Glikson, 2006). 15 16 16 17 17 18 8.5-5. POSSIBLE GEODYNAMIC CONSEQUENCES 18 19 19 ◦ − 20 For a present-day mid-ocean geotherm (25 Ckm 1), assuming yet higher Archean 20 21 geothermal gradients and possibly smaller-scale convection cells and plate dimensions 21 22 (Lambert, 1983), thin (<5 km) crustal regions underlain by shallow asthenosphere 22 23 (<50 km) can be expected to have occupied perhaps a quarter of the Archean oceanic 23 24 crust. Low angle impacts in deep-water ocean will be partly to largely absorbed by the 24 25 water column. A high-angle impact by a 20 km diameter projectile on thin oceanic crust un- 25 26 derlying <2 km water depth would result in a 300 ± 100 km diameter submarine crater. As 26 27 originally modeled by Green (1972, 1981), downward compression of the transient crater 27 28 followed by rebound of asthenosphere originally located at ∼40–50 km would result in in- 28 29 tersection of the peridotite solidus by rising mantle diapers and in partial melting. Insofar 29 30 as morphometric estimates after Grieve and Pilkington (1996) may apply to asthenospheric 30 31 uplift, the amount of rebound may be similar to that of a central uplift SU = 0.086Ds1.03, 31 32 namely ∼30 km. However, Ivanov and Melosh (in Glikson et al., 2004) point out that 32 33 such large degrees of rebound, deduced from craters formed in high-viscosity crust, may 33 34 not apply to the mantle. These authors state: ‘Indeed, a few giant impact events certainly 34 35 occurred on Earth in post-LHB history, and a few might have encountered the fortuitous 35 36 circumstances necessary for them to enhance magma production.’ 36 37 Elkins-Tanton et al. (2005) modeled the magmatic consequences of very large impacts, 37 38 suggesting that a 300 km radius crater excavating 75 km thick lithosphere will produce 38 39 106 km3 magma through instantaneous in-situ decompression of the mantle and updom- 39 40 ing of the asthenosphere, triggering longer-term mantle convection and adiabatic melting, 40 41 in particular under high geothermal gradients. These volumes are similar to those esti- 41 42 mated by Ivanov and Melosh (2003). The short eruption spans of large volumes of some 42 43 plateau basalts (Deccan: 2.106 km3 in 0.3 Ma; Emeishan: 9.105 km3 in 0.3 Ma; Siberian: 43 dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 14 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

14 Chapter 8.5: Early Archean Asteroid Impacts on Earth

1 1 2 2 3 3 D A A B B 4 4 5 5 6 6 7 7 8 8

9 + 9 2 10 10 60 cm thick shale 20 m) and shale ∼

11 < 11 jaspilite

12 + 12 13 13

14 is overlain by ferruginous 14 3 15 15

16 10 m thick jaspilite overlying 16 Marra Mamba iron-formation, immediately above unit overlying JIL ferruginous shale ( BGB–S Overlain by felsic hypabyssal/volcanics ∼ spherule unit ACM-1 –B sediments of the Ulundi Formationthe in northern part of the BGB 17 Member): BIF common above BGB–S 17 18 18 19 19

20 14 m 20 ∼ 21 21 1 m thick

22 ∼ 22 23 23 24 24 25 25

26 – 15 cm thick arenite with 26 4 27 27 28 28

29 – 10–15 cm-thick to locally 2–3 m 29 310 cm thick silica–sericite spherules MJM (Manzimnyama Jaspilite 3 5 cm thick spherule layer Carbonate hosted E Hesta – 80 cm thickspherules carbonate-chlorite and spherule-bearing breccia; 60 cm thick overlying debris ﬂow S < chlorite-rich spherules thick silica-Cr sericite-chlorite spherules S 30 Silica-sericite spherules in chert breccia/conglomerate Silica-sericite spherules within thick chert, arenite and jaspilite Two units of silica-chert spherules within 30–300 cm thick unitand of arenite chert 30 31 31

32 3 32 33 33 34 34 35 35 3 Ma: BGB–S

36 ,baseof 36 2 37 ± 37 38 38 39 39 9 Ma: ACM-2, Antarctic 3 Ma: BGB–S1A and 9 Ma: ACM-1, Antarctic . . 40 . 40 1 2 1 5 Ma: JIL, top Jeerinah 3 Ma: BGB–S 4 Ma to 3225 ± ± 41 ± 41 1 4 1 ± ± ± . . 42 . 42 43 43 ?2.63 Ga: Monteville Formation, West Griqualand Basin, west Kaapvaal Craton Formation, Fortescue Group, Hamersley Basin. 3470 3470 Table 8.5-1. Archaean andtralia, early and Proterozoic Kaapvaal asteroid Craton impact (Barberton fallout greenstone units belt and – associated BGB),Age/stratigraphy ferruginous South sediments, Africa Pilbara Craton, Western Aus- 3470 Impact unit/sand BGB-4, lower Mapepe Formation, OverlyingFig ferruginous Tree sediments Group, Kaapvaal, Craton Reference 2629 the Mapepe Formation, Fig Tree Group, Kaapvaal Craton 3258 Creek Member, Mount Ada Basalt, Warrawoona Group, Pilbara Craton BGB-S1B, upper Hooggenoeg Formation, Onverwacht Group, Kaapvaal Craton Creek Member, Mount Ada Basalt, Warrawoona Group, Pilbara Craton 3243 dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 15 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

8.5-5. Possible Geodynamic Consequences 15

1 1 2 2 3 3 E D E D D 4 4 5 5 6 6 7 7 8 8 9 9 10 10 11 11 12 12 13 13 14 14 15 15 16 16 Located 37 m above baseironstones of banded Located 38 m above theBrockman base Iron-formation, of Hamersley the Basin Carbonate megabreccia-hosted microkrystite spherules Ferruginous siltstone (Sylvia Formation), banded iron formations (Bruno Member) 17 Ferruginous siltstone (Sylvia Formation), banded iron formations (Bruno Member) 17 18 18 19 19 20 20 21 21 22 22 23 23 24 24 25 25 26 26 27 27 28 28 29 29 20 cm thick K feldspar 5cmthickK 20 cm thick spherule unit Carbonate-hosted spherules E 1 cm thick spherule unitcm overlain breccia by 80 10–20 cm K-feldspar-stilpnomelane spherules at top of 2–3ferruginous m volcanic of tuffs K-feldspar-carbonate-chlorite spherules in tsunami-generated carbonate-chert megabreccia < carbonate-chlorite spherules within turbidite < feldspar-carbonate-chlorite spherules in carbonate turbidite 30 1.8 cm thick spherule unit Carbonate-hosted E 30 31 31 32 32 33 33 34 34 35 35

36 ) 36 37 37 38 38

39 Continued 39 40 40 4 Ma: S4, Shale Macroband, 6Ma:SMB-2,topofBee 41 6Ma:SMB-1,topofBee 41 ± ± ± 42 42 2.5–2.4 Ga: lower Kuruman 43 43 1.85–2.13 Ga: Graensco, Vallen, Ketilidean, southwest Greenland ∼ Formation, West Griqualand Basin, west Kaapvaal Craton Dales Gorge Member, Brockman Iron-formation, Hamersley Group, Hamersley Basin 2481 ?2.63 ?2.56–2.54 Ga: Carawine Dolomite, Hamersley Group, Hamersley Basin Gorge Member, upper Wittenoom Formation, Hamersley Group, Hamersley Basin 2562 Gorge Member, upper Wittenoom Formation, Hamersley Group, Hamersley Basin Table 8.5-1. ( Age/stratigraphy?2.56 Ga: Reivilo Formation, West Griqualand Basin, western Kaapvaal Craton 2562 Impact unit/s Overlying ferruginous sediments Reference References: A – Byerly et al. (2002), B – Lowe et al. (2003), D – Trendall et al. (2004), E – Simonson and Glass (2004). dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 16 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

16 Chapter 8.5: Early Archean Asteroid Impacts on Earth

1 3–4.106 km3 in 0.4 Ma) are consistent with volumes deduced by these authors for exca- 1 2 vation of large-radii impact craters. According to Elkins-Tanton et al. (2005), impacts of 2 3 this magnitude will have world-wide effects for which ‘ample evidence could remain in 3 4 the rock record.’ 4 5 Ingle and Coffin (2004) suggested a possible impact origin for the short-lived ∼120 Ma 5 6 Ontong Java Plateau (OJP). Jones et al. (2002) conducted hydrocode simulations (Wünne- 6 7 mann and Ivanov, 2003) to test whether impact-triggered volcanism is consistent with the 7 8 OJP, using dry lherzolite melting parameters, a steep oceanic geotherm and a vertical inci- 8 9 dence by a dunite projectile (diameter 30 km: velocity 20 km/s). The thermal and physical 9 10 state of the target lithosphere is critical for estimates of melt production. Model calcula- 10 11 tions suggest impact-triggered melting in a sub-horizontal disc of ∼600 km diameter down 11 12 to >150 km depth in the upper mantle forms within ∼10 minutes of the impact, whereas 12 13 most of the initial melt forms at depths shallower than ∼100 km. The volume of ultra- 13 14 mafic melt would reach ∼2.5 × 106 km3, ranging from superheated melts within 100 km 14 15 of ground zero, to partial melting with depth and distance. The total melt volume would 15 16 reach ∼7.5 × 106 km3 of basalt if heat were distributed to produce 20–30% partial mantle 16 17 melting. 17 18 Lowe et al. (1989) observed the possible tectonic significance of the S2–S4 Barberton 18 19 impact layers in view of their juxtaposition with, and immediately above, the boundary 19 20 between the ∼3.55–3.26 Ga, predominantly mafic–ultramafic volcanic rocks of the On- 20 21 verwacht Group and the overlying Fig Tree Group association of turbidite-felsic volcanic- 21 22 banded iron formation. The evidence presented here and in earlier papers (Glikson and 22 23 Vickers, 2005) shows that the 3.26–3.24 Ga impact cluster coincides with termination of 23 24 the protracted evolution of greenstone–TTG (tonalite-trondhjemite-granodiorite) systems 24 25 25 in the BGB and was accompanied by volcanic activity of the Sulphur Springs Group (Pil- 26 26 bara Craton) and by plutonic events in both the Kaapvaal Craton (Nelshoogte and Kaap 27 27 Valley plutons) and the Pilbara Craton (Cleland Supersuite; Fig. 8.5-4). The full scale of the 28 28 ∼3.26–3.24 Ga bombardment may not have been identified to date, and the three recorded 29 29 impact events may represent a minimum number. Iridium and chrome isotope-based mass 30 30 balance calculations, assuming a global distribution of the impact layers – an assumption 31 31 justified by the consistently tens of cm thick Achaean spherule layers as compared to the 32 32 global ∼3–4 mm thick K–T boundary layer (Alvarez, 1986). Spherule size analysis and the 33 33 ferromagnesian composition of the impact ejecta suggest asteroids diameters of 20–50 km 34 34 diameter, excavating impact basins 200–1000 km large and located in mafic–ultramafic 35 35 36 crustal regions. 36 37 37 38 38 39 8.5-6. LUNAR CORRELATIONS 39 40 40 41 The Late Heavy Bombardment (LHB) of the Earth and Moon has been interpreted in 41 42 terms of the tail-end of planetary accretion or, alternatively, a temporally distinct bom- 42 43 bardment episode during 3.95–3.85 Ga (Tera and Wasserburg, 1974; Ryder, 1991). Some 43 dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 17 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

8.5-6. Lunar correlations 17

1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 10 11 11 12 12 13 13 14 14 15 15 16 16 17 17 18 18 19 19 20 Fig. 8.5-7. (A) Frequency distribution of Rb-Sr isotopic ages of mare basalts; (B) frequency distri- 20 21 bution of Ar-Ar isotopic ages of mare basalts (A and B after Basaltic Volcanism of the Terrestrial 21 22 Planets, 1981); (C) Frequency distribution of Ar-Ar ages of lunar impact spherules based on data by 22 23 Culler et al. (2000). 23 24 24 25 25 26 of the largest lunar mare basins contain low-Ti basalt, which possibly represents impact- 26 ± 27 triggered volcanic activity, including Mare Imbrium (3.86 0.02 Ga) and associated 27 ± 28 3.85 0.03 Ga K, REE, and P-rich-basalts (KREEP; Ryder, 1991). Isotopic Ar-Ar dating 28 29 of lunar spherules (Culler et al., 2000), when combined with earlier Ar-Ar and Rb-Sr iso- 29 30 topic studies of lunar basalts (BVTP 1981: Fig. 8.5-7), suggest the occurrence of post-LHB 30 31 impact and volcanic episodes. Possible cause-effect relationships between impact and vol- 31 32 canic activity pertain in Oceanus Procellarum (3.29–3.08 Ga) and the Hadley Apennines 32 40 39 33 (3.37–3.21 Ga). Such relationships gain support from laser Ar/ Ar analyses of lunar 33 34 impact spherules (sample 11199; Apollo 14, Fra Mauro Formation: Fig. 8.5-7), showing a 34 35 significant age spike at 3.18 Ga, near the boundary of the Late Imbrian (3.9–3.2 Ga) and the 35 36 Eratosthenian (3.2–1.2 Ga) as defined by the cratering record (Wilhelms, 1987). 34 lunar 36 37 impact spherules yield a mean age of 3188 ± 198 Ma, whereas seven spherule ages with 37 38 errors <100 My yield a mean age of 3178 ± 80 Ma. The small size of the sample and the 38 39 large errors limit the confidence in these data. However, the combined evidence suggests 39 40 that the period 3.24 ± 0.1 Ga experienced a major impact cataclysm in the Earth–Moon 40 41 system, resulting in renewed volcanic activity in some of the lunar mare basins approxi- 41 42 mately at 3.2 Ga, as well as a Rb-Sr age peak about 3.3 Ga. As in the case of lunar impact 42 43 spherules, large errors on the Ar-Ar and Rb-Sr isotopic ages preclude precise correlations 43 dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 18 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

18 Chapter 8.5: Early Archean Asteroid Impacts on Earth

1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 10 11 11 12 12 13 13 14 14 15 15 16 16 17 17 18 18 19 19 20 20 21 21 22 22 23 23 24 24 25 25 26 Fig. 8.5-7. (Continued.) 26 27 27 28 with terrestrial events. Further isotopic age studies of terrestrial and lunar materials are 28 29 required in order to establish a ∼3.2 Ga bombardment in the Earth–Moon system. 29 30 30 31 31 32 8.5-7. SUMMARY 32 33 33 34 (1) The early Archean rock record from the Kaapvaal and Pilbara Cratons preserves ev- 34 35 idence for two major impact clusters on Earth, one at 3.47 Ga, the other at between 35 36 3.26–3.24 Ga. 36 37 (2) A substantial body of petrological, geochemical, mineralogical and isotopic evi- 37 38 dence for an extraterrestrial impact origin of the Barberton 3.26–3.24 Ga spherule 38 39 units includes inward-radiating quench-textured and offset vesicles of microkrystite 39 40 spherules, PGE abundance levels, PGE ratios, Ni/Cr, Ni/Co, V/Cr and Sc/V ratios, 40 41 occurrence of quench-textured and octahedral nickel-rich chromites distinct from ter- 41 42 restrial chromites, and meteoritic 53Cr/52Cr anomalies (Shukloyukov et al., 2000; Kyte 42 43 et al., 2003). 43 dpg15 v.2007/06/04 Prn:6/06/2007; 11:09 F:dpg15041.tex; VTEX/JOL p. 19 aid: 15085 pii: S0166-2635(07)15085-4 docsubty: REV

8.5-7. Summary 19

1 (3) Mass balance calculations, based on the Ir flux and the 53Cr/52Cr anomalies (Byerly 1 2 and Lowe, 1994) and on maximum spherule radii (Melosh and Vickery, 1991), suggest 2 3 asteroid diameters in the order of 20–50 km, implying impact basins 200–1000 km 3 4 large. 4 5 (4) Lowe and Byerly’s (1989) perception of potential significance of the 3.26–3.24 Ga im- 5 6 pacts in view of their juxtaposition with, and position immediately above, the top of the 6 7 >12 km thick, predominantly mafic–ultramafic volcanic sequence of the Onverwacht 7 8 Group is corroborated by precise U-Pb ages. 8 9 (5) The occurrence of banded iron formation, jaspilite and ferruginous siltstone above the 9 10 3.26 Ga and 3.24 Ga impact units in the Barberton Greenstone Belt, and at equivalent 10 11 stratigraphic levels in the Pilbara Craton, may indirectly suggest soluble Fe-enrichment 11 12 of sea water closely following major impact events, possibly representing erosion of 12 13 impact-triggered mafic volcanics under the low-oxygen fugacity conditions of the Ar- 13 14 chaean atmosphere, similar to other impact units in the Pilbara Craton (Glikson, 2006). 14 15 15 16 16 17 17 18 18 19 19 20 20 21 21 22 22 23 23 24 24 25 25 26 26 27 27 28 28 29 29 30 30 31 31 32 32 33 33 34 34 35 35 36 36 37 37 38 38 39 39 40 40 41 41 42 42 43 43