Introduction: Spherule layers, particularly those hosted by deeper-water Precambrian
successions, are emerging as important sources of information on large
terrestrial impacts [1-3]. Searches in the late Archean to early
Paleoproterozoic Hamersley basin of Western Australia have been especially
fruitful as spherule layers from a minimum of 3 large impacts have already been
found [4]. Strata of comparable age and deposited in similar environments are
preserved over large areas in the Transvaal Supergroup of South Africa. An
initial search resulted in the discovery of a late Archean layer in the
Griqualand West basin hosted by the Monteville Formation of the Campbellrand
Subgroup [5]. Here we report the discovery of a second late Archean spherule
layer 250 to 300 m above the Monteville layer stratigraphically. Preliminary
data will be presented with an eye to comparing the new layer to known spherule
layers of roughly comparable age.
Stratigraphic and Geographic Setting: The new spherule layer was first recognized by one
of us (DYS) in cores GH6-3 drilled by Doe Run Exploration and GKP01 sponsored
by the Agouron Institute from 10 and 40 km south of Griquatown respectively.
The layer is in the Reivilo Formation [6] and is 83 and 54 meters below the
base of the Kamden Member in GH6-3 and GKP01, respectively. However, it is
closer to the Kamden than that stratigraphically because layers in the cores
are inclined 10-60û from horizontal. The strata surrounding the spherule layer
consist of carbonaceous shale and carbonate with abundant microbial structures
indicative of deposition below wave base (like those described by [6,7]). In
GH6-3, the spherule layer is 2 cm thick and consists mainly of well-sorted
spherules. It is abruptly overlain by a millimeter-scale lamina of silty
detritus consisting of angular crystals with other clast types admixed (Fig.
1), including small pieces of broken spherules. In GKP01, spherules are mixed
with 20 cm-thick carbonate breccia, which is the only indication of high energy
deposition in this stratigraphic interval. Age constraints are sparse [8], but
we estimate the Reivilo layer formed at approximately 2.56 Ga.
Description of Spherules in GH6-3: Originally the spherules appear to have been
well-sorted spherical grains in the coarse sand size range. The original sizes
and shapes have been obscured by compaction, especially pressure solution along
grain-to-grain contacts (Fig. 2). The spherules consist almost exclusively of
K-feldspar with the low sodium content typical of authigenic phases plus minor
amounts of finely crystalline mica. Late Archean to Paleoproterozoic impact
spherules have very distinctive textures [3]. The Reivilo spherules display
some of these textures, most notably an abundance of confocal sprays of highly
elongated feldspar crystals, many radiating inwards from spherule margins (Fig.
2). Another similarity is that the feldspars in the Reivilo spherules are
shaped like skeletal plagioclase crystals grown rapidly from a melt under
conditions of strong supercooling [9-11]. On the other hand, the feldspar
crystals in the Reivilo spherules are thicker and coarser and make up a higher
percentage of the spherules than those in other Late Archean to Paleoproterozoic
layers. In addition, infilled vesicles, relict glass cores replaced by clear
phases, and botryoidal fans of acicular feldspar crystals are a common feature
of spherules from other layers [3], but quite rare in the Reivilo spherules.
Fig. 1. Photomicrograph of spherule in basal part of
silty lamina on top of Reivilo layer in core GH6-3. Long axis of spherule is
0.85 mm.
Interpretation of Reivilo Spherule Layer: All of the known Archean to Paleoproterozoic
spherule layers occur as discrete layers rich in coarse sand-size grains with
predominantly spherical shapes and a heterogeneous suite of internal textures
indicating they were formerly molten; geochemical evidence of extraterrestrial
material has been detected in most of them [1-5]. We interpret the Reivilo
spherule layer as impact ejecta because it fits the pattern texturally, even
though we have no geochemical data as yet. If the Reivilo spherules are like
those in the other layers, they were probably generated by an object very
roughly the size of the K/T impactor and represent silicate melt droplets with
a low silica composition along the lines of basalt [12]. The well-sorted nature
of the layer and the abrupt transition to silt at the top further suggest that
the layer was deposited under the influence of high-energy waves and/or
currents, perhaps generated by the impact itself, as shown by many of the other
layers [2,4]. After the Reivilo spherules came to rest on the seafloor, they
were compacted as overburden slowly increased and original crystals were
replaced with authigenic K-feldspar. However, shapes and internal textures were
sufficiently well preserved for them to be recognized as impact spherules.
Fig. 1. Photomicrograph of more typical spherule from
main body of Reivilo layer in core GH6-3. Note its pressured-solved contacts
with adjacent spherules. Long axis of spherule is 0.9 mm.
Broader Implications: The discovery of the Reivilo spherule layer has
interesting ramifications both for correlations between the Hamersley and
Transvaal successions and for the genesis of impact spherules. A persuasive
case has been made that 3.47 Ga spherule layers in South Africa and Western
Australia are products of a single impact [13]. The Jeerinah and Monteveille
spherule layers are both close to 2.63 Ga in age and could likewise be ejecta
from a single large impact [14]. The next layer above the Jeerinah in the
Hamersley succession is in the Wittenoom Formation, which appears to be around
2.54 Ga in age [8,15]. Given the significant uncertainties in their ages, the
Wittenoom and Reivilo layers could be contemporaneous. However, the spherules
in the Wittenoom and Reivilo layers differ texturally as. vesicles and
botryoidal to acicular crystal formations are much more abundant in the former
than in the latter (provided our sample is representative). This leaves only
two alternatives. The first logical alternative is that the Wittenoom and
Reivilo layers represent ejecta from a single large impact within which there
were downrange changes in the textures of spherules. Major textural changes
occur downrange in the K/T boundary spherules, but this does not offer a good
analog for the Wittenoom-Reivilo correlation. In the K/T boundary layer, the
more distal spherules (known as microkrystites) are much more highly
crystallized [16], whereas the more crystalline Reivilo spherules appear to be
larger than the Wittenoom spherules, suggesting they are the more proximal.
Perhaps the Reivilo spherules spent a longer time in a higher-temperature part
of the ejecta cloud, allowing them to crystallize more rapidly and thoroughly.
Prolific early crystal growth could also preempt the appearance of vesicles and
decrease the abundance of glassy cores. The second logical alternative is that
the Wittenoom and Reivilo are not correlative, in which case the Reivilo layer
represents a major impact not recognized before. Either result is interesting,
and whatever their mode of origin, the recognition of the Reivilo layer
reaffirms the fact that impact spherule layers, though thin, are highly
distinctive and can be recognized by very Òlow-techÓ methods.
References: [1] Simonson B. M. and Glass B. P.
(2004) Ann. Rev. Earth Planet. Sci.,
32, 329-361. [2] Lowe D. R. et al.
(2003) Astrobiol., 3, 7-48. [3] Simonson B. M. (2003) Astrobiol.,
3, 49-65. [4] Hassler S.W. and
Simonson B. M. (2001) J. Geol., 109, 1-19. [5] Simonson B. M. et al. (1999) Geol. Soc.
Amer. Spec. Paper 339, 249-261. [6]
Beukes N. J. (1987) Sed. Geol., 54, 1-46. [7] Sumner D. Y. (1997) Palaios, 12,
300-318. [8] Nelson D. R. et al. (1999) Precamb. Res., 97,
165-189. [9] Lofgren G. (1974) Amer.
J. Sci., 274, 243-273. [10] Lofgren G. E. (1983) J. Petrol., 24, 229-255. [11]
Simonson B. M. and Sweeney D. (2002) Geol. Soc. Amer. Abs. Prog., 34/6,
81. [12] Simonson B. M. and Harnik P. (2000) Geology, 28,
975-978. [13] Byerly G. R. et al., Science, 297, 1325-1327. [14]
Simonson B. M. et al. (2000) Austral. Jour. Earth Sci., 47,
315-325. [15] Woodhead J. D. et al. (1998) Geology, 26,
47-50. [16] Smit J. (1999) Ann. Rev. Earth Planet. Sci., 27,
75-113.