A report (Gersonde et al., 1997) of recent shallow piston coring around the estimated site of a late Pliocene asteroid impact identified originally from early Eltanin piston cores (Kyte et al., 1981, 1988; Kyte and Brownlee, 1985; Margolis et al., 1991) has renewed interest in an event that might be detectable at Leg 178 sites on the continental rise drifts. The event has been dated at 2.15 Ma, close to the Reunion geomagnetic polarity event (C2r.1n) and in oxygen isotope Stage 82 (Gersonde et al., 1997). It is proposed that an asteroid 1 to 4 km in diameter struck an area close to 57.8ºS, 90.8ºW, ejecting large volumes of seafloor sediment and also providing an impact-generated airborne cloud of melted asteroidal debris. The debris has a distinctive geochemical anomaly, unmixed with ocean-floor material, and the ejected sediment is reported to be early Eocene to Pliocene calcareous and nannofossil ooze with minor diatoms and radiolarians. In addition, an impact of this magnitude is considered likely to have generated a "megatsunami" of 20-40 m amplitude in deep water, increasing manyfold in the shoaler waters of a continental margin (Gersonde et al., 1997). There have been many recorded occurrences of reworked continental shelf material transported inland by tsunamis and of a "homogenite" on the floor of the enclosed basin of the eastern Mediterranean with a wider grain-size distribution and an origin at least partly in sediments from the surrounding continental shelf and slope (Cita et al., 1996). Suspension of fine-grained sediment in deep water is not ruled out.

Irrespective of the amplitude and depth penetration of such a phenomenon, we would anticipate a significant effect at the Antarctic Peninsula margin, about 1500 km away, and probably also toward the crest of the drifts, some 1300 km away, where the water depth is less than three-quarters that of the direct path of the megatsunami from the impact site. The effect at the Antarctic margin would depend crucially on the stage of glaciation: well into an interglacial, with the ice-grounding line far inshore and the continental slope in a relatively stable "starved" situation, the effect might have been minimal. During a glacial period, however, with an ice sheet grounded to the continental shelf edge and an upper slope loaded with potentially unstable glacially transported sediment, the effect could have been far greater, generating unusually large debris flows and turbidity currents everywhere along the margin. Subsequent events on the shelf would most probably have erased the shelf record of such an event, but the resulting effect on the continental rise drifts should have been preserved. We might therefore be looking for a geochemical anomaly, involving air-fall asteroid debris (spherules and a derived noble metals anomaly) together with early Eocene to Pliocene biogenic material, also airborne, above a possible but not inevitable megaturbidite or other anomalous sedimentary expression of a large-scale tsunami.

Sediments within the drifts on the continental rise exhibit a strong bottom-simulating reflector (BSR) at 550-650 ms depth. It is clear beneath the scarp faces (e.g., Fig. F14: shotpoints 1400-1700 and 5100-5400) but less so beneath the gentler flanks where bedding is parallel to the seafloor. It caused some concern before Pollution Prevention and Safety Panel review as a potential gas hydrate BSR, as did the possibility of stratigraphic traps on the continental shelf. We had aimed to drill through or to the BSR at the prime rise site.

We did not expect thermogenic hydrocarbons at the sites, either on the rise or on the shelf. In general terms, the offshore Antarctic Peninsula has seen considerable activity from expeditions, national scientific base relief, and marine scientific surveys over the past 80 yr--and from whalers and sealers for 80 yr before that. No hydrocarbon shows have ever been reported, although there have been indications of methane in Bransfield Strait to the northeast. This last is a region of high biogenic productivity as well as active extension leading to high geothermal gradients and is not an analogue for the regions of the proposed sites.

The key factors governing hydrocarbon potential of the shelf are probably

1. The absence of terrigenous vegetation on the Antarctic Penin-sula margin during glacial times;
2. The accumulation of immature clastics in an accretionary wedge before ridge-crest subduction; and
3. The onset of fully glacial conditions with erosion at the ice-sheet base, and deposition of (unsorted) diamictons on shelf and slope, at some time between 10 and 6 Ma.

Also, the subducted ridge crest was a heat source but caused uplift (focused on the mid-shelf high) and consequent subaerial erosion of the existing sedimentary succession (S4 in Fig. F12). This happened shortly (1-4 m.y.) after collision. Collision ages are 11 Ma for the shelf transect (Sites 1100, 1102, and 1103) and 16.5 Ma for the shelf interlobe site (1097). Thus, although there are probably stratigraphic traps now at the top of S4 close to the mid-shelf high, these were outcropping at the time of high heat flow following ridge-crest collision.

Only two BSRs recognized around Antarctica have been attributed to gas hydrates: one is on the landward slope of the South Shetland Trench (Lodolo et al., 1993). On the South Orkney microcontinent (Lonsdale, 1990), a BSR considered beforehand to be from the base of a gas hydrate was shown by ODP Leg 113 drilling to have resulted from silica diagenesis. South of the Polar Front, biogenic silica is abundant and widespread within pelagic and hemipelagic sediments, and diagenetic alteration of silica at depth is likely to be common.

It should be straightforward to distinguish between the two: the base of a gas hydrate is related to pressure and temperature-almost all known marine hydrate BSRs involve the simplest mixture, methane and seawater-and heat flow over the ocean floor is a known function of ocean-floor age and sedimentation rate. Thus, a preliminary estimate of the depth of a methane hydrate BSR can be made. Also, with a methane hydrate, the bedding above the BSR commonly shows a lower range of acoustic impedance contrast than the bedding below. The acoustic impedance reduces downward, leading to a negative polarity reflector. In contrast, with a silica diagenetic reflector, the range of impedance contrast is reduced below the reflector, impedance increases downward, and the reflector has positive polarity. Silica diagenesis is in general terms the product of a time-temperature integration, so depths cannot be predicted precisely (see Lonsdale, 1990). Although there are occasionally strange chemical effects, possibly biologically mediated, depths of silica diagenesis BRSs are generally greater than would be predicted for a methane hydrate BSR in an area of known age.

We believe the BSR that is widely seen over the drifts is caused by silica diagenesis. We have considered both origins for the reflector and strongly favor diagenesis, for the following reasons:

1. Reflector polarity is difficult to establish because the reflector is diffuse where it crosses bedding and is hard to distinguish where it is parallel to bedding. We would say it is positive, which favors a diagenetic origin. The reflector is also in places diffuse, which might favor a diagenetic over a pressure/temperature-sensitive phase boundary origin.
2. Impedance contrast is sometimes difficult to see on the continental rise, but for most occurrences, impedance contrasts above are clearly greater (supporting a diagenetic origin).
3. Depth of a gas hydrate BSR provides a measure of heat flow. Over ocean floor of known age, heat flow is predictable. There is minor uncertainty in heat flow attaching to the history of sedimentation in a thickly sedimented section, but that too is calculable. The ocean floor beneath these rise sites is well dated and ranges in age between 42 and 35 Ma, 20 and 14 Ma beneath the two drifts targeted (Sites 1095, 1096, and 1101). If these BSRs were gas hydrates, with the sediment thicknesses present, they should lie in the range of 300-500 ms depth at these sites. In fact, the BSR depths lie in the range of 550-700 ms.
4. Piston cores from the drifts show abundant biogenic silica (mostly diatoms) within interglacials, and there is every reason to believe it continues to depth. If so, a silica diagenetic reflector should occur within the drifts, and the measured depth would be reasonable.

Nevertheless, it was intended to measure heat flow in the upper part of Site 1096, where the BSR is well developed, to provide an early test of its likely origin there.