Origin of bottom-simulating reflectors: Geophysical evidence from the Cascadia accretionary prism

Vertical seismic profile (VSP) data from two drill sites on the Cascadia margin show low-velocity zones, indicative of free gas, beneath a bottom-simulating reflector (BSR). Offshore Oregon, at Ocean Drilling Program (ODP) Site 892, velocities drop from an average of 1750 m/s above the BSR to less than 1250 m/s below it. Sonic logs confirm that seismic velocity in the sediments adjacent to the borehole is less than that of water for at least 50 m beneath the depth of the BSR at this site. Similarly, at ODP Site 889 offshore Vancouver, velocities range from 1700 to 1900 m/s in the 100 m above the BSR and drop abruptly to 1520 m/s in the 15 m just beneath it. The low velocities observed beneath the BSR are strong evidence for the presence of l%-5% free gas (by volume). The BSR at these two sites results from the contact between gas-free sediments containing a small quantity of hydrate above the BSR and a low-velocity free-gas zone beneath it. Although the BSR is associated with the base of the hydrate stability field, hydrate appears to account for relatively little of the velocity contrast that produces the BSR. Velocity above the BSR at Site 889 is only about 100 m/s greater than that expected for sediments of similar porosity. Sediments above the BSR at Site 892 appear to have normal velocity for their porosity and may contain little hydrate. INTRODUCTION Bottom-simulating reflectors (BSRs) broadly mimic the relief of the sea floor and, in many settings, are believed to mark the pressureand temperature-dependent base of the methane-hydrate stability field (e.g., Shipley et al., 1979; Hyndman et al., 1992). Below this boundary, methane may be present as dissolved or free gas, but not as hydrate. Hydrate is an icelike substance, consisting of gas molecules within a lattice of water molecules, that is formed at high pressure and low temperature. Appropriate P-Tconditions for methane-hydrate stability are widespread in the shallow sediments of the sea floor of continental margins, and BSRs are commonly observed in seismic reflection data from accretionary prisms (Shipley et al., 1979). Because hydrate may store large quantities of methane, an understanding of the origin of BSRs and the amounts of hydrate or free gas associated with them is important in assessing both their economic potential and their effect on global climate change (Kvenvolden, 1988). BSRs have also been widely used to estimate geothermal heat flow from the P-T conditions for hydrate stability, thereby providing constraints on determinations of fluid movements within accretionary prisms (Yamano et al., 1982; Minshull and White, 1989; Davis et al., 1990; and others). BSRs are typically strong, negative-polarity reflectors. Reflection amplitude can be as great as that of the sea floor, and reflection polarity of BSRs is consistently opposite that of the sea floor, indicating a large decrease in acoustic impedance (velocity × density) below the reflector. In water-saturated sediments, the presence of either hydrate or small amounts of free gas will lower the density slightly. The presence of hydrate increases sediment velocity, although the relation of concentration to velocity is not well known GEOLOGY, v. 22, p. 459-462, May 1994 (Stoll, 1974; Pearson et al., 1986); even very small amounts of free gas in a sediment will greatly reduce its velocity (Domenico, 1976). Consequently, the impedance contrast between sediment containing hydrate and sediment containing free gas (or non-hydrate-bearing sediments) is predominantly caused by the change in velocity. The origin of this velocity contrast, and the relative importance of hydrate vs. free gas in the formation of the BSR, has remained controversial. Models for BSRs have, to date, been based almost exclusively on analysis of surface seismic data. In various locales, the BSR has been interpreted as either the base of a high-velocity layer due to hydrate above the BSR (Hyndman and Davis, 1992) or the top of a low-velocity zone due to free gas beneath the BSR (Bangs et al., 1993; Singh et al., 1993) or a combination of both (Dillon and Paull, 1983; Minshull and White, 1989; Miller et al., 1991). Although Deep Sea Drilling Project (DSDP) and ODP drilling has penetrated BSRs at several sites (Kastner et al., 1991), the rapid dissociation of hydrate and loss of gas during core recovery have thwarted attempts to determine the in situ concentrations of hydrate and free gas. Bangs et al. (1993) presented a partial suite of downhole logs that show good evidence for the presence of free gas despite poor borehole conditions. In this paper, we present velocity measurements from vertical seismic profiles (VSPs) and downhole logs that unambiguously document the presence of a surprisingly thick zone of free gas beneath the BSR. GEOLOGIC SETTING During ODP Leg 146 (Westbrook et al., 1993), we obtained in situ velocity measurements from VSPs and downhole logs at two sites that penetrated BSRs in the Cascadia accretionary prism, offshore Oregon and Vancouver Island. Oregon margin Site 892 is situated near the crest of a large ridge —16 km behind the thrust front, at a water depth of 670 m (MacKay et al., 1992). Multichannel seismic (MCS) data show a strong BSR with few other coherent reflectors (Fig. 1). The site is located in the hanging wall of a minor thrust over which the BSR rises locally, presumably because of the movement of warm fluids along the fault (Moore et al., 1991). The structural and hydrologic setting of the Oregon margin in the vicinity of Site 892 has been examined by means of MCS data, Seabeam bathymetry, GLORIA and SeaMARC 1A side-scan data, and Alvin dives (Carson et al., 1991; Moore et al., 1991; MacKay et al., 1992). Nonetheless, the subsurface velocity structure at the site remains poorly determined because of the lack of coherent reflections for use in velocity analysis of the MCS data. Vancouver margin Site 889 lies on a broad mid-slope terrace at a water depth of 1322 m. A veneer of slope sediment 130 m thick overlies the more deformed accreted sediments at the site (Fig. 1). The BSR is within the accreted sediments at Site 889, but in nearby regions it extends across both prism and slope-basin deposits (Hyndman and Spence, 1992). Based on detailed velocity analyses, estimates of heat flow, and modeling of the BSR, the hydrologic setting at Site 889 was inferred to be dominated by diffuse upward flow of Reprinted by permission from Geology, 22:459-462, 1994, The Geological Society of America.