Geophysical Constraints on the Surface Distribution of Authigenic Carbonates Across the Hydrate Ridge Region, Cascadia Margin

Available online at www.sciencedirect.com

Marine Geology 202 (2003) 79^120 www.elsevier.com/locate/margeo

Geophysical constraints on the surface distribution of authigenic carbonates across the Hydrate Ridge region, Cascadia margin

Joel E. Johnson a;Ã, Chris Gold¢nger a, Erwin Suess b

a Oregon State University, College of Oceanic and Atmospheric Sciences, 104 Ocean Admin. Bldg., Corvallis, OR 97331, USA b GEOMAR Research Center for Marine Geosciences, 24148 Kiel, Germany

Received 28 June 2002; received in revised form 30 June 2003; accepted 18 August 2003

Abstract

On active tectonic margins methane-rich pore fluids are expelled during the sediment compaction and dewatering that accompany accretionary wedge development. Once these fluids reach the shallow subsurface they become oxidized and precipitate cold seep authigenic carbonates. Faults or high-porosity stratigraphic horizons can serve as conduits for fluid flow, which can be derived from deep within the wedge and/or, if at seafloor depths greater than V300 m, from the shallow source of methane and water contained in subsurface and surface gas hydrates. The distribution of fluid expulsion sites can be mapped regionally using sidescan sonar systems, which record the locations of surface and slightly buried authigenic carbonates due to their impedence contrast with the surrounding hemipelagic sediment. Hydrate Ridge lies within the gas hydrate stability field offshore central Oregon and during the last 15 years several studies have documented gas hydrate and cold seep carbonate occurrence in the region. In 1999, we collected deep-towed SeaMARC 30 sidescan sonar imagery across the Hydrate Ridge region to determine the spatial distribution of cold seep carbonates and their relationship to subsurface structure and the underlying gas hydrate system. High backscatter on the imagery is divided into three categories, (I) circular to blotchy with apparent surface roughness, (II) circular to blotchy with no apparent surface roughness, and (III) streaky to continuous with variable surface roughness. We interpret the distribution of high backscatter, as well as the locations of mud volcanoes and pockmarks, to indicate variations in the intensity and activity of fluid flow across the Hydrate Ridge region. Seafloor observations and sampling verify the acoustic signals across the survey area and aid in this interpretation. Subsurface structural mapping and swath bathymetry suggest the fluid venting is focused at the crests of anticlinal structures like Hydrate Ridge and the uplifts along the Daisy Bank fault zone. Geochemical parameters link authigenic carbonates on Hydrate Ridge to the underlying gas hydrate system and suggest that some of the carbonates have formed in equilibrium with fluids derived directly from the destabilization of gas hydrate. This suggests carbonates are formed not only from the methane in ascending fluids from depth, but also from the shallow source of methane released during the dissociation of gas hydrate. The decreased occurrence of high-backscatter patches and the dramatic reduction in pockmark fields, imaged on the eastern part of the survey, suggest gas hydrate near its upper stability limit may be easily destabilized and thus, responsible for these seafloor features. High backscatter along the left-lateral Daisy Bank fault suggests a long history of deep-seated fluid venting, probably unrelated to destabilized gas hydrate in the subsurface.

* Corresponding author. Fax: +1-541-737-2064. E-mail address: [email protected] (J.E. Johnson).

MARGO 3397 29-10-03 80 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120

Keywords: SeaMARC 30; sidescan sonar; gas hydrate; £uid £ow; accretionary wedge

1. Introduction ROVs, tv-camera tows, and coring and dredging devices. Pore £uid expulsion is a common process in The ¢rst documentation of authigenic carbon- accretionary wedges on active continental margins ates and chemosynthetic biological communities and is coincident with the dehydration and com- associated with pore £uid expulsion on an active paction of the sediment column during accretion- continental margin occurred at the ¢rst accretionary wedge development. Fluid expulsion can (1) ary ridge west of Hydrate Ridge, o¡shore central be episodic along high-porosity stratigraphic ho- Oregon (Fig. 1), in the mid-1980s (Suess et al., rizons and/or faults exposed at the sea£oor (e.g. 1985; Kulm et al., 1986; Ritger et al., 1987). Since Kulm et al., 1986; Lewis and Cochrane, 1990; then, other sites along the Paci¢c rim have also Moore and Vrolijk, 1992; Sample, 1996; Sample been discovered (e.g. Henry et al., 1989; Le Pi- and Reid, 1998), (2) occur during mud volcanism chon et al., 1992; Ohta and Laubier, 1987). The (e.g. Brown, 1990; Kopf, 2002), and/or (3) occur occurrence of authigenic carbonates on Hydrate during di¡use, intergranular £uid £ow (e.g. Ridge was documented during ODP (Ocean Drill- Moore and Vrolijk, 1992 and references therein). ing Program) Leg 146 drilling in 1992(see West- An abundant component in the £uids escaping brook et al., 1993), inferred from reprocessed from organic-rich accretionary wedges is dissolved GLORIA sidescan sonar imagery by Carson et thermogenic and biogenic methane. When sub- al. (1994), and further constrained by submersible jected to the lower temperature, lower pressure, observations and sea£oor sampling (e.g. Torres et and oxidizing bacteria-rich environment near the al., 1999; Bohrmann et al., 2000; Greinert et al., sea£oor surface, these methane-rich £uids can 2001). Gas hydrates are also well developed at precipitate cold seep carbonates of aragonite, cal- Hydrate Ridge and were ¢rst inferred on seismic cite, and/or dolomite compositions (Ritger et al., re£ection pro¢les collected as the site survey for 1987; Greinert et al., 2001). In addition, some of ODP sites 891 and 892( MacKay et al., 1992). The this methane is temporarily incorporated as free widespread distribution of the BSR (bottom sim- gas into the gas hydrate fabric near the sea£oor ulating re£ector), marking the base of gas hydrate prior to expulsion (Suess et al., 2001), if at high stability (Hyndman and Spence, 1992), or the top enough gas concentrations and appropriate pres- of a free gas zone trapped beneath the hydrate sure (water depths at least 300 m) and temper- (MacKay et al., 1994), suggests the ridge is ature (bottom water temperatures approaching capped by gas hydrate (Tre¤hu et al., 1999). The 0‡C) conditions (Kvenvolden, 1993). ¢rst recovery of gas hydrate on Hydrate Ridge Both surface and deep-towed sidescan sonar occurred during ODP Leg 146 at site 892( Hov- surveys o¡er a unique method for regional map- land et al., 1995) when the crest of the northern ping of sea£oor £uid venting sites because the summit was drilled through a hydrologically ac- acoustic impedence contrast (densityUsound tive fault (Fig. 1). Since then, several researchers velocity) between the authigenic carbonates pre- have returned to collect samples of the gas hy- cipitated at these sites and the surrounding drate and authigenic carbonates and to conduct hemipelagic sediments is detectable. Sidescan so- other geophysical surveys to characterize the be- nar surveys are valuable because they provide a havior of the hydrate system in this active tectonic regional-scale survey of sea£oor £uid venting oc- setting (e.g. Torres et al., 1999; Bohrmann et al., currences and distributions, and they also help 2000; Suess et al., 1999, 2001; Tre¤hu et al., 2002; guide later sea£oor observation and sampling this study). e¡orts conducted with manned submersibles, On Hydrate Ridge and in other accretionary

MARGO 3397 29-10-03 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120 81

Fig. 1. Shaded relief bathymetry of the Hydrate Ridge region. Contour interval is 100 m and bathymetric grid is 100-m pixel resolution. Inset shows Paci¢c Northwest bathymetry and topography (Haugerud, 1999) and the location of Hydrate Ridge region on the lower continental slope of the Cascadia accretionary prism. Hydrate Ridge is a NE^SW-trending thrust ridge with northern and southern summits; (NHR) Northern Hydrate Ridge; (SHR) Southern Hydrate Ridge. The ridge is located V10 km from the deformation front (DF) and bordered on the west and east by slope basins (HRB-W) Hydrate Ridge Basin-West and (HRB-E) Hydrate Ridge Basin-East. ODP (Ocean Drilling Program) site 891 on the crest of the ¢rst accretionary ridge (FAR) and site 892on NHR are shown. Daisy Bank (DB) is also shown.

ridges within the hydrate stability zone, a close preted on sidescan sonar records and observed on association between gas hydrates and carbonates the sea£oor. may exist because £uids dewatered from the prism Our recent research e¡orts have been focused not only supply methane to the gas hydrate stabil- on identifying the structural controls on the dis- ity zone, but also transfer heat to shallower tribution of authigenic carbonates in the Hydrate depths, which can induce the destabilization of Ridge region and their relationship to the under- the gas hydrate (Suess et al., 2001). On both the lying gas hydrate system. In this paper, we present northern and southern summits of Hydrate Ridge, the results of our 1999 SeaMARC 30 (SM30) the authigenic carbonate and pore water carbon deep-towed sidescan sonar survey coupled with and oxygen isotopes support this association by sea£oor observations and samples and subsurface suggesting the carbonates are precipitated in part geologic mapping, based on seismic re£ection from methane derived from destabilized gas hy- data, to determine the distribution of authigenic drate (Bohrmann et al., 1998; Greinert et al., carbonates and their relationship to large-scale 2001). This observation suggests destabilized gas subsurface structures and the underlying gas hy- hydrate can contribute to the total accumulation drate system across the Hydrate Ridge region. of authigenic carbonate precipitated during accretionary wedge dewatering and compaction. With this in mind, an understanding of the regional 2. Tectonic setting distribution of subsurface structures (faults and folds) and the extent of the gas hydrate stability 2.1. Cascadia accretionary prism zone is necessary to make interpretations about the possible origins of £uid venting patterns inter- The Juan de Fuca Plate is currently being sub-

MARGO 3397 29-10-03 82 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120 ducted obliquely beneath the North American sal plain (Gold¢nger et al., 1992, 1997). The ac- Plate along the Washington, Oregon, and North- cretionary wedge widens from 60 km o¡ southern ern California continental margins (Fig. 1). The Oregon to 150 km o¡ the northern Olympic Pen- Cascadia accretionary prism evolved in response insula of Washington, where the thick Pleistocene to this oblique subduction and is composed of Astoria and Nitinat fans are presently being ac- folded and faulted abyssal plain turbidites and creted to the margin. The active accretionary hemipelagic sediments (Kulm and Fowler, 1974). thrust faults of the lower slope are characterized This oblique convergence also creates a right-lat- by mostly landward-vergent thrusts on the Wash- eral shear couple within the upper to lower con- ington and northern Oregon margins and sea- tinental slope and o¡ the Washington and Oregon ward-vergent thrusts on the central and southern margins nine WNW-trending left-lateral strike- Oregon margin (Gold¢nger et al., 1992; MacKay slip faults, antithetic to the shear couple, have et al., 1992; MacKay, 1995). The landward-ver- been identi¢ed on the continental slope and abys- gent province may be related to the subduction of

Fig. 2. Structure map of the Hydrate Ridge region (interpreted from multichannel seismic re£ection pro¢les collected as the site survey for ODP Leg 146, inset), overlain on 100 m shaded relief bathymetry. Deep-seated left-lateral strike-slip faults (Daisy Bank and Alvin Canyon faults) dominate the deformation north and south of Hydrate Ridge and thrusts and folds are common in the region between the strike-slip faults. The distribution and style of structures between these strike-slip faults and the apparent clockwise rotation of Hydrate Ridge suggest deformation may be in£uenced by oblique subduction-driven right-lateral shear. The eastern end of the Daisy Bank fault is mapped based on the previous work of Gold¢nger et al. (1996) and the SM30 sidescan sonar data presented in this paper. The major geologic and geographical features are labeled as follows: (DF) deformation front; (FAR) ¢rst accretionary ridge; (HRB-W) Hydrate Ridge Basin-West; (NHR) Northern Hydrate Ridge; (SHR) Southern Hydrate Ridge; (HRB-E) Hydrate Ridge Basin-East; (DB) Daisy Bank.

MARGO 3397 29-10-03 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120 83 rapidly deposited and overpressured sediment sin-East (HRB-E) and Hydrate Ridge Basin-West from the Astoria and Nitinat submarine fans (HRB-W). (Seely, 1977; MacKay, 1995). O¡ Washington and northern Oregon the broad accretionary 2.2. Structure of the Hydrate Ridge region prism is characterized by low wedge taper and widely spaced accretionary thrusts and folds, The Hydrate Ridge region is a highly deformed which o¡scrape virtually all of the incoming sedi- portion of the accretionary wedge that results mentary section. Sparse age data suggest the from oblique subduction-driven compression. prism is Quaternary in age and is building west- The faults and folds in the region were initially ward at a rate close to the orthogonal component mapped by Gold¢nger et al. (1992, 1997) and of plate convergence (Gold¢nger et al., 1996). MacKay et al. (1992) and MacKay (1995) and This young accretionary wedge abuts a steep document the landward to seaward structural ver- slope break that separates it from the oceanic gence change across the region as well as the pres- and volcano-clastic Siletz terrane that underlies ence of two deep-seated left-lateral strike-slip the continental shelf (Snavely, 1987). Above this faults (the Daisy Bank and Alvin Canyon faults basement is a modestly deformed Eocene through of Gold¢nger et al., 1997). Using the same multi- Holocene forearc basin sequence (Snavely, 1987; channel seismic data set presented in these papers McNeill et al., 2000). Hydrate Ridge lies within and in Tre¤hu et al. (1999) across Hydrate Ridge, the northern end of the seaward-vergent province, we have created a regional structure map that o¡shore central Oregon, and is the second sea- includes the previously mapped structures and ward-vergent accretionary thrust ridge from the the smaller-scale folds and faults previously iden- deformation front (Fig. 1). It is bordered on ti¢ed, but not correlated across the region (Figs. 2 east and west by slope basins, Hydrate Ridge Ba- and 3). Based on the mapped structures, Hydrate

Fig. 3. (A) Example section of a seismic re£ection pro¢le (line 9) from the ODP Leg 146 site survey (location shown in Fig. 2, inset). The data are time-migrated and depict relative true amplitudes. The amplitude scale grades from white to black. (B) No- tice the small-scale folds (a = anticline; s = syncline) and strike-slip faults (sense of slip not detectable). A small thrust fault is also present. Variations in the degree of deformation across the Hydrate Ridge region can be seen by mapping these small-scale features along with the previously mapped larger structures (Gold¢nger et al., 1992, 1997; MacKay et al. (1992); MacKay (1995). This line in particular shows the concentration of deformation near the northwestern corner of Hydrate Ridge.

MARGO 3397 29-10-03 84 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120

Ridge appears to be bounded at its north and 3. Methods south end by the left-lateral strike-slip Daisy Bank and Alvin Canyon faults. The geometry 3.1. Imaging the carbonates and slip direction of these faults implies clockwise rotation of the block contained between them The observed authigenic carbonates present in within an overall right-lateral shear couple (Gold- the Hydrate Ridge region are exposed at the sea- ¢nger et al., 1997). Oblique subduction-driven £oor or slightly buried by a few centimeters of right-lateral shear of the Hydrate Ridge block hemipelagic mud (Bohrmann et al., 1998; Kulm could be responsible for the apparent clockwise and Suess, 1990; Suess et al., 2001). Because of rotation of Hydrate Ridge (Johnson et al., 2000; the acoustic impedence (densityUsound velocity) Johnson et al., in preparation). The eastern edge contrast between the carbonates and the sur- of the block occurs at the contact with the Siletz rounding sediments, sidescan sonar can be used terrane, which forms a strong backstop for the to image the distribution of authigenic carbonates accretionary prism (Fleming and Tre¤hu, 1999) across the region at the surface and in the shallow and is likely the eastern edge of the right-lateral subsurface (Johnson and Helferty, 1990). In re- shear couple responsible for the strike-slip faults. gions bearing sea£oor gas hydrate, like Hydrate The concentration of tightly spaced structures at Ridge, there is also an acoustic impedence con- the north end of HRB-E, the apparent bulge in trast between the gas hydrate and the surrounding the abyssal plain protothrusts southwest of Hy- hemipelagic sediments. Although smaller than drate Ridge, the wider spacing of structures at that due to carbonate, the impedence contrast be- the south end of HRB-E, and the decrease in tween gas hydrate at the sea£oor and hemipelagic abyssal plain protothrusts to the northwest may sediments may be su⁄cient to produce an inter- serve as evidence for clockwise block rotation, mediate backscatter signal. Because of the limited however, in this paper we focus on the distribu- bandwidth of current analog sonars and variation of structures rather than their origins. tions in the tow¢sh depth during the survey, how- Although the main fault or faults responsible for ever, the sonar gain must be adjusted frequently, the formation of Hydrate Ridge are not imaged which makes post-cruise quantitative analysis of on the existing seismic re£ection pro¢les, the variations in backscatter strength problematic. asymmetric morphology of the ridge, with its Because of this, we do not attempt to di¡erentiate more steeply dipping and eroded western £ank, between authigenic carbonates and sea£oor gas suggests it is cored by one or more seaward-ver- hydrate of intermediate backscatter strength gent thrust faults, similar to the thrust anticlines across the survey, except where con¢rmed by sea- seen throughout the wedge and commonly ob- £oor observations. served in fold-thrust belts (Suppe, 1983). The greater relief of the northern summit of the ridge, 3.2. SeaMARC 30 survey compared to the southern summit, also suggests net slip along the underlying structures may be To maximize our sea£oor resolution and in or- greater toward the north. der to image carbonates buried by a thin veneer

Fig. 4. The SM30 sidescan sonar mosaic overlain on 100 m shaded relief bathymetry. The SM30 survey is gridded at 1-m pixel resolution. Notice the three categories of backscatter present across the region (inset map). Category I backscatter is characterized as blotchy to circular with apparent surface roughness, category II backscatter is blotchy to circular with no apparent surface roughness, and category III streaky to continuous with variable surface roughness. The three regions of each backscatter category are shown in the inset and depict locations in the survey where each category is dominant. Note the coincidence of most of the category III backscatter with the regions of greatest slope. The major geologic and geographical features are labeled as follows: (HRB-W) Hydrate Ridge Basin-West; (NHR) Northern Hydrate Ridge; (SHR) Southern Hydrate Ridge; (HRB-E) Hy- drate Ridge Basin-East; (DB) Daisy Bank (a high-resolution PDF ¢le of this ¢gure is available from the author).

MARGO 3397 29-10-03 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120 pp. 85^104

MARGO 3397 29-10-03 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120 105

Fig. 5. Slope map of the Hydrate Ridge region created from swath bathymetric data gridded at 50 m (see Section 4.2 for data sources). Slope variations were calculated from the 50-m grid using Erdas Imagine software. Slopes are shaded in 5‡ increments. Category I and II backscatter patches are outlined in black. Notice the coincidence of the backscatter with the regions of lowest slope. Category III backscatter occurs in the low-slope regions only on the tops of uplifts associated with the Daisy Bank fault zone. The coincidence of high backscatter with low slope suggests the backscatter signal is most likely due to changes in surface roughness and sediment composition rather than bathymetry.

of hemipelagic mud we chose the low-frequency bonate and gas hydrate in the Hydrate Ridge re- (30 kHz), deep-towed, SM30 sidescan sonar sys- gion, spanning a corridor from the deformation tem, operated by Williamson and Associates in front on the west to beyond the predicted upper Seattle, WA, USA. The sonar was towed at a hydrate stability limit (450^500 m; Tre¤hu et al., depth of V200 m above the sea£oor and col- 2002) on the east. Included on the eastern edge of lected data in V3.0-km swaths across the entire the survey was the southeast extent of the Daisy region and V1.5-km swaths across the crest of Bank fault zone, a deep-seated left-lateral strike- Hydrate Ridge. The frequency on the port side slip fault, and likely £uid £ow conduit, spanning is 27 kHz and on the starboard side 30 kHz. here, just above the upper depth limit of hydrate The gain of the sonar was adjusted manually in stability (Gold¢nger et al., 1996). Carson et al. 10, 3-dB steps to gain approximately equal record (1994) attempted to remove the bathymetric sig- intensity across the survey. Navigation was by nal from shallow-towed GLORIA regional side- Sonardyne USBL (Ultra-Short BaseLine acoustic scan data in an e¡ort to constrain the extent of positioning). The R/V New Horizon was used to authigenic carbonate on Hydrate Ridge. These tow the SM30 at 2^3 knots. The sidescan images low-frequency (6.5 kHz) GLORIA data can re- were acquired and processed using Triton Elics cord deeply buried features, however, making sur- International Isis sonar processing software, and ¢cial interpretations problematic. Nevertheless, ultimately georeferenced and gridded at 1-m pixel our deep-towed SM30 data support the general resolution for the entire survey using Erdas Imag- interpretations of Carson et al. (1994), but at ine software. The survey was designed to image much higher resolution and across not only Hy- the surface and shallow subsurface authigenic car- drate Ridge, but the slope and shelf to the east.

MARGO 3397 29-10-03 106 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120

MARGO 3397 29-10-03 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120 107

3.3. High backscatter and carbonates lines were collected in order from south (track 1) to north (track 11), beginning at the southeastern The intensity of backscatter on sidescan sonar corner of the survey. All odd numbered tracklines records is a function of (1) the angle of incidence were collected towing the sonar from east to west of each beam (the bathymetric variations on the and even numbered tracks from west to east. sea£oor^the slope), (2) the physical characteristics Thus, at the overlap between tracklines the same of the surface (microscale roughness), (3) the in- geologic feature is imaged twice, but with insoni- trinsic nature of the surface (composition^density) ¢cation from opposing directions. At the overlap, and (4) the frequency and pulse characteristics of the more detailed swath is generally shown. The the sonar (Blondel and Murton, 1997). One of the 1.5-km swaths over the crest of Hydrate Ridge important e¡ects on the backscatter signal re- were towed NE^SW along the axis of the ridge. ceived by the sidescan sonar is the e¡ect due to Dark gray to black lines along the centers of each bathymetric slope (1, above). Steep sea£oor ba- trackline are the nadirs (no data recovery beneath thymetry sloping toward the passing sonar has the tow¢sh). White thin continuous lines across enhanced backscatter strength compared to those the slope basin east of Hydrate Ridge are surface slopes dipping away from the sonar. Because of returns recorded by the sonar. Continuous high this signal enhancement, di¡erentiation between backscatter along the nadirs on E^W lines 6^8 is sediment types based on backscatter strength is an artifact of the sonar. best determined in regions of low slope. On Hy- drate Ridge submersible dives, tv-camera tows, 4.2. Backscatter patterns and bathymetric slope and sea£oor samples (see Section 4.3) have documented that authigenic carbonates are present Examination of the survey mosaic reveals the where bathymetric variation is minimal and high high-backscatter patterns (light tones) generally backscatter is dominant on the imagery. Similarly, can be divided into three categories: (I) circular across much of the survey, the sea£oor slopes are to blotchy with apparent surface roughness, (II) less then 5‡ (see Section 4.2), suggesting high circular to blotchy with no apparent surface backscatter in these regions is more likely related roughness, and (III) streaky to continuous with to changes in rock and sediment composition on variable surface roughness (Fig. 4). Category I the sea£oor rather than a bathymetric e¡ect. high backscatter is concentrated mainly on the northern summit and northeastern end of Hy- drate Ridge and to a lesser extent on the southern 4. Results summit of Hydrate Ridge. Category II high backscatter is concentrated on the eastern edge of 4.1. Sidescan sonar survey HRB-E, and eastward up to the 700-m bathymetric contour. It also extends spatially from north to The complete mosaiced survey gridded at 1-m south across the entire survey in this region. Cat- pixel resolution for both the V3-km and V1.5- egory III high backscatter is present on the west- km swaths is presented in Fig. 4. The E^W track- ern edge of the survey in regions of steep bathym-

Fig. 6. (A) SM30 coverage and ground truth across Hydrate Ridge. TVG (tv-grab) and OFOS (tv-camera tow) tracks from Sonne Leg 143 are shown. Mud volcanoes MV1, MV2, and MV3 are shown (see Section 4.4) as well as the Southern Hydrate Ridge (SHR) pinnacle (note the acoustic shadow on the imagery). The intermediate backscatter on SHR represents sea£oor gas hydrate as observed in Alvin dives (Torres et al., 1999). (B) OFOS track 216 across NHR (from Bohrmann et al., 2000). The diagram was constructed from deep-towed video observations of the sea£oor. Note the coincidence of the chemoherm carbonates and carbonate crusts with the regions of highest backscatter along the track. (C) A bottom type map constructed from Alvin observation on NHR (from Torres et al., 1999). Again note the coincidence of high backscatter on the survey and the carbonates observed on the sea£oor. (D) Alvin photograph of the carbonate on the SHR pinnacle (photo courtesy of Marta Torres, Oregon State University). Notice the large fracture in the middle of the image.

MARGO 3397 29-10-03 108 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120

Fig. 7. (A) SM30 coverage at the western edge of HRB-E. Notice the circular category II high-backscatter patches. Tv-camera tow tracks and the gravity and multicorer sites described in the text are shown. (B) Methane distribution in surface sediment multicorer samples taken at some of the backscatter patches shown in A. Samples taken at a bright-backscatter patch (SO148/75- 1B-MUC) yielded high methane, at a dark-backscatter patch (SO143/31-1A-MUC) lower methane, and at a low-backscatter reference site, at the center of HRB-E (SO143/63-1A-MUC), an even lower methane concentration. (C) Methane distribution in two surface sediment gravity core samples (locations shown in A). High methane was found in both gravity core samples, one taken from a dark-backscatter patch (SO143/32-2-SL) and a second from a bright-backscatter patch (SO143/35-1-SL). Gravity core (SO148-76-SL) taken at a very bright-backscatter patch recovered carbonates and gas hydrates. Tv-camera tow track SO148/9 also documented authigenic carbonate at the surface (on the same backscatter patch sampled by gravity core SO148/76-SL). The remaining tv-camera tows documented only sediments and bacterial mats at the surface, suggesting that some of the high-backscatter patches may be caused by authigenic carbonates and/or gas hydrates slightly buried by hemipelagic sediment.

MARGO 3397 29-10-03 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120 109

Fig. 8. (A) SM30 coverage across Daisy Bank. 1992 Delta submersible dive tracks are shown. Video observations from the dives document cobbles, blocks, and slabs of well-lithi¢ed sediments and carbonates (B, inset) on the nearly £at top of Daisy Bank, coincident with the apparent surface roughness and moderate backscatter observed on the SM30 survey. etry associated with the large submarine canyon grid. Additional bathymetric data sets (EM300 and steep slopes on the western £ank of Hydrate and EM120 swaths from Jack Barth at Oregon Ridge, and in the southeastern corner of the State University and lower-resolution data sets, survey, associated with subtle breaks in slope. NOAA EEZ SeaBeam 16 and National Ocean Category III high backscatter is also present Service hydrographic soundings) were used to ¢ll along the Daisy Bank fault zone to the north- in gaps in coverage. The inclusion of lower-reso- east. lution data in the grid limits the resolution to To determine the e¡ect of slope on the back- 50 m, however, yields coverage coincident with the scatter intensity, a slope map was created from SM30 survey. Examination of bathymetric slope 50-m gridded bathymetry, which is the highest- reveals much of the high backscatter (category I resolution grid available across the entire survey and II) on Hydrate Ridge, HRB-E, and the sea- area (Fig. 5). Newly acquired high-resolution £oor west of Daisy Bank occurs in regions with EM300 data across Hydrate Ridge and HRB-E the lowest slope, 6 10‡, mostly 6 5‡. The high (from Clague et al., 2001), EM120 data across backscatter observed in regions with the highest HRB-W, and EM300 from HRB-E to Daisy slope, on the western edge of the survey where Bank (collected by the authors in 2002) are the slopes exceed 15‡ surrounding HRB-W and along three primary data sets used in this bathymetric some of the Daisy Bank fault uplifts, is likely

MARGO 3397 29-10-03 110 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120

Fig. 9. SM30 coverage and structures across Hydrate Ridge. Three mud volcanoes (MV1, MV2, and MV3) are present on North- ern Hydrate Ridge (NHR). MV2and MV3 are associated with high backscatter suggesting carbonate precipitation during expulsion. MV2does not appear to be associated with high backscatter, which is consistent with an overall decrease in patchy backscatter toward the saddle region. Notice MV3 breaches the crest of an anticline on the eastern £ank of Hydrate Ridge and MV1 and MV2exist near the crest of NHR. The acoustic shadow of the Southern Hydrate Ridge (SHR) pinnacle is visible as a black speck in the middle of the largest backscatter patch on SHR. The saddle region between NHR and SHR appears highly frac- tured, perhaps a result of extension at the crest of Hydrate Ridge. enhanced by the steep bathymetry. The only ma- slope e¡ect to the backscatter intensity minimized jor exception to this is observed on some of the in regions of low slope, the backscatter can be low-slope, high-to-moderate backscatter seen on interpreted to result from contrasts in surface the £at-topped ridges of the Daisy Bank fault roughness and/or harder or denser sediment com- zone (Figs. 4 and 5; see Section 4.3). With the position across the survey.

MARGO 3397 29-10-03 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120 111

Fig. 10. (A) SM30 complete survey coverage. The pockmark ¢elds and the extent of the BSR are shown. The easternmost pockmark ¢eld extends to the V550-m contour and is coincident with the termination of the BSR at nearly the same depth. This suggests the pockmark ¢elds may represent the gas hydrate stability limit near the eastern part of the survey. Also notice the abrupt termination of most of the pockmark ¢elds and category II backscatter patches at the 700-m contour (see Section 5.3 for discussion). (B) Close up image of one of the pockmark ¢elds (location shown in A).

4.3. Groundtruthing the sonar northern and southern summits of Hydrate Ridge have con¢rmed the presence of extensive authi- 4.3.1. Category I backscatter ^ Hydrate Ridge genic carbonates and gas hydrates. During R/V Samples from the sea£oor and deep-towed vid- Sonne Legs 143-1, 2, and 3, deep-towed tv-cam- eo camera data (Bohrmann et al., 1998; Greinert eras, termed OFOS (Ocean Floor Observation et al., 2001; Suess et al., 2001) as well as Alvin System) and TVGs (tv-guided grab samplers) observations (Torres et al., 1999) on both the were deployed across the Hydrate Ridge region

MARGO 3397 29-10-03 112 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120

(Fig. 6A). A graphic log representing the video acoustic signal, as the SM30 has previously im- observations along OFOS track 216 is shown in aged objects of high-impedence contrast buried by Fig. 6B. We observe a close correlation between several meters of soft sediment (M. Williamson, high backscatter on the sidescan sonar and the personal communication, 1999). slabs, cobbles, and boulders of authigenic carbonates observed in the OFOS videos along all three 4.3.2. Category II backscatter ^ Eastern Slope tracks, 216, 213 and 223 (Fig. 6A). OFOS track Basin 223 documents carbonate only near the crest of Tv-camera tows conducted in 1999 across some Hydrate Ridge, where the backscatter is of inter- of the high-re£ectivity category II backscatter mediate intensity but the pattern resembles a patches indicated only sediments and bacterial pavement of carbonate that extends well into the mats at the surface, suggesting carbonates and/ saddle between the south and north summits (Fig. or hydrates if present here are buried beneath a 6A). In all of the TVG samples shown (Fig. 6A), thin veneer of sediment (Fig. 7A). A tv-camera authigenic carbonates were also recovered from tow conducted during the R/V Sonne (SO148) the sea£oor, again coincident with locations of cruise in 2000, however (Linke and Suess, 2001), high backscatter. documented bacterial mats, clam ¢elds, and some Observations made during ¢ve Alvin dives on authigenic carbonates present at the sea£oor in the northern summit of Hydrate Ridge resulted this same region (Fig. 7A). In addition to the in a map of the bottom types, characterizing the tv-camera tows, three category II circular back- sea£oor as consisting of massive carbonates, scatter patches were cored (by gravity or multi- mostly carbonates, mixed sediments and carbon- corer) during R/V Sonne cruises SO143 and ates, sediments, or clams and/or microbial mats SO148 (Fig. 7A). All of the cored sites showed (Torres et al., 1999; Fig. 6C). On the southern unusually high methane contents in the sediment summit of Hydrate Ridge, six Alvin dives resulted compared to a reference site (SO143-63-1A- in the discovery of a very large carbonate chemo- MUC) in the center of the basin (Fig. 7A^C). herm pinnacle, which stands nearly 50 m above There appears to be a qualitative relationship be- the sea£oor (Torres et al., 1999). The dive sites tween the brightness of the circular patches in the were concentrated at the largest high-backscatter sidescan sonar image and the methane distribu- patch on the southern summit of Hydrate Ridge. tion with depth in cores. The highest methane In the middle of this patch, the acoustic shadow content was recorded at the bright patch at site cast by the large pinnacle carbonate can be seen SO143/35-1-SL with almost 1000 nmol/g within (Figs. 4 and 6A,D). Investigation by Alvin of the the ¢rst 100 cmbsf, whereas the reference site in intermediate-intensity backscatter present north- the center of HRB-E barely exceeds 1 nmol/g east of the pinnacle resulted in the identi¢cation (Fig. 7B). At the dark patch (sites SO143/32-2- of gas hydrate at the sea£oor, which was previ- SL and SO143/31-1A-MUC) the methane con- ously sampled with a large TVG (Suess et al., tent gradually increased towards 10 nmol/g in 1999). This suggests the sidescan sonar may the shallow part, but exceeded 200 nmol/g at have imaged sea£oor gas hydrate here and re- s 500 cmbsf (Fig. 7B and C). The brightest corded it with intermediate backscatter intensity patch (site SO148/75-1B-MUC) showed by far (Fig. 6A), alternatively, the intermediate backscat- the highest near-surface methane concentration ter could be caused by slightly buried authigenic (V8 nmol/g at 10 cmbsf) and at greater depth, carbonate not recovered by the TVG. During the reached by gravity coring (SO148/76-SL), con- same Alvin dives, the small high-backscatter circle tained disseminated gas hydrates and authigenic northwest of the pinnacle was also investigated, carbonates. Although limited sampling and sea- however no carbonate or gas hydrate was ob- £oor observations have occurred across the entire served at the sea£oor surface (Torres et al., category II backscatter region (Fig. 4), the similar 1999). We suggest sediments cover the authigenic shape and backscatter intensity of the circular carbonate or hydrate likely responsible for this patches suggests they may represent authigenic

MARGO 3397 29-10-03 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120 113 carbonate, as they appear to have the same back- thymetry (Clague et al., 2001). Mud volcanoes are scatter strength as the known carbonates on the one of several types of surface expressions of mud crest of Hydrate Ridge. The lack of visible surface intrusions, which occur as overpressured, multi- roughness and blurry nature of the circular phased pore £uids (methane and water) and sedi- patches, however, may indicate many of them ments are expelled at the sea£oor surface (Brown, are buried by a thin drape of hemipelagic sedi- 1990). The two mud volcanoes present near the ment. The close association of the authigenic car- crest of Hydrate Ridge (MV1 and MV2) are co- bonates and gas hydrates recovered in gravity incident with the abundant carbonate chemo- core SO148-76-SL also suggests gas hydrates herms documented on the northern summit of may be present at many of the category II circular the ridge, while the mud volcano on the eastern patches. £ank of the ridge (MV3) has breached the crest of a smaller secondary anticline (Fig. 9). The two 4.3.3. Category III backscatter ^ Daisy Bank northern mud volcanoes (MV2and MV3) show Category III high-backscatter sites are coinci- high backscatter on the sidescan imagery, perhaps dent with regions of high slope (Figs. 4 and 5), suggesting methane-rich £uids were expelled dur- however, a major exception to this occurs on the ing their eruption and consequently oxidized to £at-topped ridges associated with the Daisy Bank precipitate carbonate (Fig. 9). The more southerly fault zone (Fig. 4). Sea£oor observations by the mud volcano imaged with more subdued back- Delta submersible document abundant carbonate scatter is coincident with the general decrease in chimneys, doughnuts, and slabs, within 100^150 m backscatter toward the saddle region of Hydrate of the fault traces and tabular carbonate blocks Ridge. up to 6 m in length on the top of Daisy Bank Active £uid venting in the form of hydrate- (Gold¢nger et al., 1996). Recent examination of coated bubbles has also been observed on both the 1992 Delta submersible dive videos with the the northern and southern summits of Hydrate new SM30 sidescan data indicates the extensive Ridge during Alvin dives (Torres et al., 1999) and high-to-moderate backscatter with apparent sur- tv-camera tows (Suess et al., 1999) and acoustical- face roughness observed on the top of Daisy ly imaged in 12- and 18-kHz subbottom pro¢les Bank is due to cobbles, blocks, and slabs of (Torres et al., 1999; Bohrmann et al., 2000; Hee- well-lithi¢ed sediments and carbonates (Fig. 8). schen et al., 2003). Methane has also been mea- Although the uplifts associated with the Daisy sured in high-concentration plumes in the water Bank fault northwest of Daisy Bank have higher column above Hydrate Ridge (Suess et al., 2001). backscatter than the top of Daisy Bank, the slope Additional evidence for £uid or gas escape, im- e¡ect due to their steep-sided, linear nature likely aged on the SM30 survey, is present in the central enhances some of their backscatter signal, even to eastern region of the survey as a band of pock- though lithologically they may be similar to Daisy mark ¢elds (Fig. 10). The pockmark ¢elds are Bank. limited on their eastern side by the V550-m contour and on the western end by the V1050-m 4.4. Other £uid venting manifestations contour. Some of the pockmarks are coincident with the locations of category II high backscatter Also present on the crest of Hydrate Ridge and and others appear to be isolated in lower-back- along its northeastern £ank are three mud volca- scatter sediments. noes, MV1, MV2, and MV3 (Fig. 9). The circular craters at the centers of mud volcanoes MV1 and MV3 are clearly visible on the sidescan imagery, 5. Discussion however, MV2is less obvious because of the high backscatter caused by the surrounding carbon- 5.1. Backscatter distribution ates, but a circular crater associated with this high backscatter is visible on high-resolution ba- Con¢rmation of the backscatter patterns on

MARGO 3397 29-10-03 114 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120

Hydrate Ridge by abundant sea£oor observations ter sites investigated at the sea£oor, further makes them the most well-constrained features in groundtruthing in the central and eastern part the survey. In addition, the highest level of detail of the survey is needed to truly verify their com- can be seen by looking at the 1.5-km swaths, positions. which only cover the top of Hydrate Ridge (Fig. 4). Examination of this imagery in particular re- 5.2. Relationship to geologic structures veals the variations in backscatter strength and sea£oor features from the northern to the south- 5.2.1. Hydrate Ridge region ern summit of the ridge. The northern summit is The pattern of extensive authigenic carbonates, characterized by extensive authigenic carbonate the presence of three mud volcanoes, and ob- chemoherms and gas hydrates of high to inter- served sea£oor gas discharge suggest £uid venting mediate backscatter intensity and the surface of is highly concentrated on northern Hydrate the ridge appears to break apart toward the sad- Ridge. Structurally, the northern summit is shal- dle region, between the two summits, coincident lower, at a depth of V600 m, than the southern with a dramatic decrease in large carbonate che- summit, at V800 m. This di¡erence in bathyme- moherms. The carbonates appear to become rees- try suggests the ridge is e¡ectively plunging to the tablished, to a much lesser degree, toward the southwest creating a structural trap at the north- southern summit and result in the large pinnacle ern summit for any £uids and gases to migrate structure (Fig. 4). toward and accumulate. A large accumulation of Category I and II high-backscatter patterns free gas is inferred beneath the northern summit are both blotchy to circular and only di¡er in of Hydrate Ridge based on seismic velocity and their apparent surface roughness. They are both attenuation (Tre¤hu and Flueh, 2001) and it is this also only present in the regions of the survey pool of gas, which likely supplies the gas hydrate^ where gas hydrate is either inferred by the pres- authigenic carbonate system in the near-surface ence of the BSR (Tre¤hu et al., 1999; Tre¤hu and sediments. The extensive occurrence of authigenic Flueh, 2001) or con¢rmed by recovered samples carbonates on the northern summit suggests it has (Hovland et al., 1995; Suess et al., 2001). On recorded either a longer history or a higher rate of Hydrate Ridge the apparent surface roughness methane expulsion than that recorded by the of the category I high backscatter has been southern summit. U/Th ages from northern sum- groundtruthed by Alvin, deep-towed video obser- mit authigenic carbonates may indicate a long his- vations, and sea£oor samples and is likely due to tory of methane expulsion is more likely, as sam- the fresh exposure of the authigenic carbonates ples of the carbonate carapace there yield ages and gas hydrates on the sea£oor. In contrast to between 68,700 and 71,700 years, while samples the erosional environment of Hydrate Ridge, the from the southern summit pinnacle range in age eastern slope basin and region to the east likely from 7300 to 11,400 years (Teichert et al., 2003). have a higher sedimentation rate, which may tend The conduits for methane expulsion at the north- to bury authigenic carbonates and hydrates pre- ern end of the ridge could be faults and fractures, cipitated near the sea£oor surface. This could similar to those imaged on the 1.5-km swath side- explain the smooth texture and the variable back- scan data in the saddle between the northern and scatter intensity of the category II high-backscat- southern summits of Hydrate Ridge (Fig. 9). Per- ter sites in these regions. Sea£oor observations haps, some of these conduits are bending-moment and core also suggest at least some of the category normal faults created during £exure at the crest of II high-backscatter sites contain authigenic car- Hydrate Ridge. Bedding-parallel faults created bonate and gas hydrate buried by hemipelagic during £exural slip folding, out of sequence thrust sediments. The similar nature of the category II faults, or high-porosity stratigraphic horizons backscatter patches across the survey suggests could also serve as conduits. Unfortunately, de- they may all have this same composition, how- tails of the shallow subsurface structure at the ever, with only a few of the category II backscat- crest of Hydrate Ridge cannot be resolved on

MARGO 3397 29-10-03 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120 115

Fig. 11. Schematic diagrams depicting the environments likely responsible for each of the backscatter categories. Suggested pathways responsible for the delivery of £uids to the shallow subsurface and gas hydrate stability zone (BSR labeled) are shown in the lower block diagram. Fluids migrate and accumulate at structural highs like Hydrate Ridge and the ¢rst accretionary ridge (FAR) and to a lesser extent on the eastern slope east of HRB-E. The structural highs are local sites of tension and it is likely faults and fractures permeate their crests. Category I backscatter occurs in this environment and, in the presence of abundant gas hydrate, £uid £ow responsible for the carbonate precipitation at the surface is characterized as structurally in£uenced and gas hydrate related. Category II backscatter occurs on the eastern slope, east of HRB-E. It is characterized by variable intensity and a patchy distribution of backscatter patterns, suggestive of sporadic sea£oor venting. It also lies in a region of the wedge that is mildly deformed, thus structural control on £uid expulsion sites is unlikely. Fluid £ow responsible for this backscatter is likely di¡use and related to disruptions in the underlying gas hydrate. Category III backscatter across the Daisy Bank fault zone suggests long-term deep-seated £uid £ow. Its location above the hydrate stability zone and the linear backscatter patterns suggest this region is dominated by structurally controlled non-gas hydrate related £uid £ow. the existing seismic re£ection data, so the true and the location of the mud volcano (MV3) on nature of the conduits can only be speculated at the eastern £ank of the northern summit of Hy- this time. The important observation, however, is drate Ridge also coincides with the crest of a that the carbonates appear to be well developed at smaller anticline (Fig. 9). These associations sug- the northern summit of Hydrate Ridge, where gest anticlinal folding may help to focus £uids concentrated structural deformation is likely. until faults or fractures, potentially caused by The ¢rst accretionary anticline just west of Hy- the folding, or exposed high-porosity stratigraphic drate Ridge shows abundant evidence for £uid horizons create the conduits needed for subse- expulsion at its crest (Kulm and Suess, 1990) quent £uid expulsion. Future correlation of the

MARGO 3397 29-10-03 116 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120 high-backscatter sites on Hydrate Ridge with egory II high backscatter, circular to blotchy with higher-resolution seismic re£ection data, however, no apparent surface roughness, dominates the re- will ultimately yield better insight into these pro- gion along the eastern edge of HRB-E and east- cesses. ward toward the Daisy Bank fault zone. The association between the pattern of £uid venting and 5.2.2. Daisy Bank fault zone bathymetry suggests some of the high-backscatter Most of the category III high-backscatter sites patches may be coincident with abrupt changes in are coincident with high slope, however a major slope, perhaps caused by small-amplitude folds in exception to this occurs on the £at-topped ridges the subsurface (similar to those shown in Fig. 3). associated with the Daisy Bank fault zone (Fig. Much of the category II backscatter, however, 4). High backscatter along the fault scarp and appears as small, scattered patches in areas of uplifts along the fault zone suggest either authi- low slope across the region, suggestive of low-vol- genic carbonate presence or an older more lithi- ume di¡use £uid venting. The coincidence of the ¢ed and uplifted section of the stratigraphy capa- category II backscatter with the hydrate stability ble of high backscatter, or a combination of the zone may indicate £uid venting is associated with two. Sea£oor observations in the Delta submersi- shallow disruptions in the underlying gas hydrate ble document abundant carbonate chimneys, system. The lack of signi¢cant surface roughness doughnuts, and slabs, within 100^150 m of the on the category II high backscatter, however, and fault traces and tabular carbonate blocks up to subtle variations in backscatter strength suggest 6 m in length on the top of Daisy Bank (Gold- these features may be buried by variable amounts ¢nger et al., 1996). Authigenic carbonates have of hemipelagic sediment. This may indicate that also been observed at the Wecoma left-lateral either (1) £uid venting is currently not active at strike-slip fault to the north and isotopic measure- some of the sites across the region (where the ments on them suggest the £uid source is deep backscatter is circular but of intermediate re£ec- within the sedimentary section (Sample et al., tivity), (2) £uid venting and carbonate precipita- 1993). The high backscatter at Daisy Bank is shal- tion are continuous, but occur at a slower rate lower (200 m) than the predicted hydrate stability than hemipelagic sedimentation, or (3) there is zone (V450^500 m; Tre¤hu et al., 2002), and like an episodic nature to the £uid venting, permitting the Wecoma fault, it likely taps a deeper methane active £uid escape at some sites (e.g. the highest- source to precipitate the carbonates present near backscatter sites) while abandoning others (e.g. the sea£oor surface. A long-term history of struc- intermediate-backscatter patches, possibly cov- turally controlled deep £uid venting along the ered by hemipelagic sediment since their last ac- Daisy Bank fault is also suspected due to the per- tivity). We suggest the apparent variability in £uid vasive high backscatter present along its length. venting and authigenic carbonate precipitation Such a large active fault conduit is likely to deliv- across the eastern slope could be attributed to er a signi¢cant volume of £uid to the sea£oor disruptions in the subsurface gas hydrate system. surface, perhaps accommodating much of the re- In this region, perhaps larger faults at depth are gional dewatering of the accretionary wedge. responsible for the delivery of pore £uids to the These focused £uids, once expelled, result in the gas hydrate stability zone within the upper sea- signi¢cant carbonate precipitation observed on £oor sediments. Once there, the £uids only escape the sea£oor and inferred by the high backscatter by (1) direct fault conduits to the sea£oor, which along the fault zone. seem unlikely here since there are few linear manifestations of £uid venting, (2) rapid £uid expul- 5.2.3. HRB-E and eastern slope sion, resulting in pockmarks on the sea£oor, or Highly focused £uid £ow is evident on Hydrate (3) di¡use £uid £ow, through high-porosity con- Ridge and along the Daisy Bank fault, however, duits or microfractures. We suggest the latter two in the region between these structures the back- processes occur across this region, as circular scatter pattern is much more di¡use (Fig. 4). Cat- high-backscatter patches can be seen associated

MARGO 3397 29-10-03 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120 117 with pockmarks (Fig. 10) and high-porosity con- intergranular £uid £ow. Authigenic carbonates duits are likely in a wedge characterized by turbi- can precipitate from the methane in these £uids dite sedimentation (Kulm and Fowler, 1974). as they interact directly with seawater in the shallow subsurface and/or as methane is released dur- 5.3. Landward limit of gas hydrate stability ing the destabilization of gas hydrate. The distribution of high backscatter across the Hydrate Pockmarks terminate near the eastern end of Ridge region suggests £uids pond in anticlinal the survey area at V550 m, coincident with the structures like Hydrate Ridge and the ¢rst accre- eastern extent of the BSR, as interpreted on the tionary ridge, while migrating away from or com- only multichannel seismic re£ection pro¢le (line 8) pletely abandoning more undeformed portions of that extends this far east from the ODP Leg 146 the wedge (e.g. HRB-E, where there is only one site survey (Fig. 10). This suggests the pockmarks high-backscatter patch, Fig. 4). The crests of anti- may represent the surface manifestation of gas clinal structures like Hydrate Ridge are under lo- hydrate dissociation near the landward limit of cal tension and it is likely that fractures and faults gas hydrate stability in this region. Because pock- serve as the major vertical £uid £ow conduits to marks typically result from rapid £uid or gas es- the sea£oor. However, because Hydrate Ridge lies cape through a relatively thin sedimentary section within the hydrate stability zone and carbonates (Brown, 1990) the shallow source of methane in derived from destabilized gas hydrate have been gas hydrate near its stability boundary may be a identi¢ed there (Bohrmann et al., 1998; Greinert likely source for this £uid expulsion. The V550-m et al., 2001) it is likely that both structurally in- contour corresponds with only the easternmost £uenced £uid migration and the destabilization of pockmark ¢eld on the survey. Most of the pock- gas hydrate contribute to the authigenic carbon- mark ¢elds appear to terminate at the V700-m ates imaged on the sidescan imagery. Therefore contour to the west, coincident with the abrupt we classify the category I backscatter environment termination of category II backscatter (Fig. 10). as structurally in£uenced and gas hydrate related The coincidence of the termination of the BSR on (Fig. 11). Category II backscatter occurs in a seismic line 8 with the easternmost pockmark ¢eld mildly deformed portion of the wedge within the and the other pockmark ¢elds suggests the hy- region of gas hydrate stability. The hydrate sys- drate stability limit extends to the V550-m depth tem here is likely fed by up-dip £uid £ow along near seismic line 8, but lies deeper to the west, at porous stratigraphic horizons or deeper faults out 700 m, across the rest of the survey (Fig. 10). We of HRB-E (evidenced by the abundant backscat- suggest the distribution of pockmarks and cate- ter on the eastern edge of HRB-E and on Hydrate gory II backscatter patches and their eastern ter- Ridge, both up-dip of HRB-E). In the middle mination serves to delineate the landward limit of portion of the survey (HRB-E to the Daisy gas hydrate stability, and thus £uid venting re- Bank fault) the bathymetric expression of major lated to gas hydrate destabilization in this region. subsurface structures is minimal, suggesting the subsurface is only mildly deformed (Fig. 1). The high-backscatter patterns in the category II region 6. Backscatter patterns and £uid £ow suggest £uid pathways to the surface are di¡use or £uids advect at slower rates. Variable backscat- Schematic diagrams describing the possible en- ter intensity, due to buried authigenic carbonates, vironments responsible for the three categories of may also suggest a sporadic history of £uid vent- backscatter and the potential pathways for the ing, although high methane content in the near- £uids are shown in Fig. 11. Fluid £ow, driven surface sediments suggests active upward £uid by compaction and dewatering of the wedge, in £ow. The lack of signi¢cant subsurface structure the Hydrate Ridge region supplies £uids to the in the category II backscatter region and its loca- hydrate stability zone via faults and fractures, tion within the hydrate stability zone suggest au- dipping stratigraphic horizons, or during di¡use thigenic carbonate precipitation is most likely re-

MARGO 3397 29-10-03 118 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120 lated to the destabilization of gas hydrate in this II backscatter patches on the eastern edge of the region. The category II backscatter environment survey area may serve to delineate the landward is thus characterized as di¡use and gas hydrate limit of gas hydrate stability across this region. related (Fig. 11). Category III backscatter, not related to major slope changes, occurs along the Daisy Bank fault zone. Here, high backscatter Acknowledgements patterns suggest £uid £ow has been continuous and likely deep-seated. In addition most of the We thank the 1999 R/V New Horizon crew, the Daisy Bank fault zone lies above the eastern hy- shipboard scienti¢c party, and Williamson and drate stability limit (V550 m at the shallowest). Associates of Seattle, WA for their hard work The environment responsible for the Daisy Bank during the SeaMARC 30 data acquisition and category III backscatter is thus classi¢ed as struc- processing phase. We also thank Gerhard Bohr- turally controlled and non-gas hydrate related mann (GEOMAR), for helpful comments and (Fig. 11). suggestions during the sidescan sonar cruise and use of the R/V Sonne Leg 143 data, Peter Linke (GEOMAR) for use of the R/V Sonne SO148 7. Conclusions OFOS data, and Marta Torres (OSU) for use of her Alvin observations on Hydrate Ridge. Addi- Based on sea£oor mapping using high-resolu- tional thanks to Jason Chaytor (OSU) for compil- tion SM30 sidescan sonar data coupled with sea- ing the 50-m bathymetric grid used to make the £oor observations, samples, and subsurface geo- slope map and Anne Tre¤hu and Matt Arsenault logic mapping the following conclusion can be (OSU) for assistance with the 1989 Digicon seis- made: (1) the blotchy to circular category I high mic data. This article also bene¢ted from very backscatter present on Hydrate Ridge is indeed thorough and thoughtful reviews by Nathan authigenic carbonate, (2) the category II high Bangs, David Piper, and an anonymous reviewer. backscatter present along the eastern side of the GEOMAR collaborators on cruises SO143 and survey is likely authigenic carbonate, with some SO148 were funded through the TECFLUX proj- gas hydrate, slightly buried by hemipelagic sedi- ects by the Federal Ministry of Science and Tech- ment, (3) both category I and II high-backscatter nology (BMBF; Grants 03GO/143/148). Funding sites represent carbonates that may have precipi- for the TECFLUX project at OSU for the data tated from destabilized gas hydrate and show an collection and processing of the SM30 data was intimately linked gas hydrate carbonate system, provided by NSF Award #9731023. Acknowl- however on Hydrate Ridge, direct £uid migration edgement is also made to the Donors of The Pe- from depth also likely contributes to the carbon- troleum Research Fund, administered by the ate precipitation, (4) breached anticlines not only American Chemical Society, for partial support serve to trap and aid in the migration of £uids of this research. and gases through the sediment column, but also serve as escape pathways, providing a local extensional environment at their crests for £uid References expulsion, (5) di¡use gas hydrate-related £uid Blondel, P., Murton, B.J., 1997. Handbook of Sea£oor Sonar £ow is likely responsible for the category II high Imagery. Wiley-Praxis Series in Remote Sensing. John Wiley backscatter on the eastern slope, east of HRB-E, and Sons Ltd. in association with Praxis Publishing Ltd., (6) category III backscatter associated with the Chichester, 314 pp. Daisy Bank fault zone is likely derived from Bohrmann, G., Greinert, J., Suess, E., Torres, M., 1998. Au- deep-seated £uids that have a long history of es- thigenic carbonates from the Cascadia subduction zone and their relation to gas hydrate stability. Geology 26, 647^650. cape along the fault and are likely unrelated to Bohrmann, G., Linke, P., Suess, E., Pfannkuche, O., 2000. RV the destabilization of gas hydrate, and (7) the SONNE Cruise Report SO143 TECFLUX-I-1999. GEO- abrupt decrease in pockmark ¢elds and category MAR Report 93.

MARGO 3397 29-10-03 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120 119

Brown, K.M., 1990. The nature and hydrogeologic signi¢cance region, Cascadia accretionary prism. Geol. Soc. Am. Ab- of mud diapirs and diatremes for accretionary systems. stracts with Programs 32, A36. J. Geophys. Res. 95 (B6), 8969^8982. Johnson, H.P., Helferty, M., 1990. The geological interpreta- Carson, B., Seke, E., Paskevich, V., Holmes, M.L., 1994. Fluid tion of side-scan sonar. Rev. Geophys. 28, 357^380. expulsion sites on the Cascadia accretionary prism: Map- Kopf, A.J., 2002. Signi¢cance of mud volcanism. Rev. Geo- ping diagenetic deposits with processed GLORIA imagery. phys. 40, B1^B49. J. Geophys. Res. 99 (B6), 11959^11969. Kulm, L.D., Fowler, G.A., 1974. Cenozoic sedimentary frame- Clague, D.A., Maher, N., Paull, C.K., 2001. High-resolution work of the Gorda-Juan de Fuca Plate and adjacent con- multibeam survey of Hydrate Ridge, o¡shore Oregon. In: tinental margin- A Review. In: Dott, R.H. Jr., Shaver, R.H. Paull, C.K., Dillon, W.P. (Eds.), Natural Gas Hydrates: (Eds.), Modern and Ancient Geosynclinal Sedimentation. Occurrence, Distribution, and Detection. AGU Geophys. Soc. Econ. Paleontol. Mineral. Spec. Publ. 19, 212^229. Monogr. 124, 297^303. Kulm, L.D., Suess, E., 1990. Relationship between carbonate Fleming, S.W., Tre¤hu, A., 1999. Crustal structure beneath the deposits and £uid venting: Oregon accretionary prism. central Oregon convergent margin from potential-¢eld mod- J. Geophys. Res. 95 (B6), 8899^8915. eling: Evidence for a buried basement ridge in local contact Kulm, L.D., Suess, E., Moore, J.C., Carson, B., Lewis, B.T., with a seaward dipping backstop. J. Geophys. Res. 104 (B9), Ritger, S.D., Kadko, D.C., Thornburg, T.M., Embley, 20431^20447. R.W., Rugh, W.D., Massoth, G.J., Langseth, M.G., Co- Gold¢nger, C., Kulm, L.D., Yeats, R.S., Appelgate, B., chrane, G.R., Scamman, R.L., 1986. Oregon subduction MacKay, M.E., Moore, G.F., 1992. Transverse structural zone: Venting, fauna, and carbonates. Science 231, 561^ trends along the Oregon convergent margin: Implications 566. for Cascadia earthquake potential and crustal rotations. Kvenvolden, K.A., 1993. Gas hydrates ^ Geological perspec- Geology 20, 141^144. tive and global change. Rev. Geophys. 31, 173^187. Gold¢nger, C., Kulm, L.D., Yeats, R.S., Hummon, C., Huf- Le Pichon, X., Kobayashi, K., Crew, K.-N.S., 1992. Fluid tile, G.J., Niem, A.R., McNeill, L.C., 1996. Oblique strike- venting activity within the eastern Nankai Trough accretion- slip faulting of the Cascadia submarine forearc: The Daisy ary wedge: A summary of the 1989 Kaiko-Nankai results. Bank fault zone o¡ central Oregon. In: Bebout, G.E., Earth Planet. Sci. Lett. 109, 303^318. Scholl, D.W., Kirby, S.H., Platt, J.P (Eds.), Subduction: Lewis, B.T.R., Cochrane, G.C., 1990. Relationship between Top to Bottom. AGU Geophys. Monogr. 96, 65^74. the location of chemosynthetic benthic communities and Gold¢nger, C., Kulm, L.D., Yeats, R.S., McNeill, L., Hum- geologic structure on the Cascadia subduction zone. J. Geo- mon, C., 1997. Oblique strike-slip faulting of the central phys. Res. 95 (B6), 8783^8793. Cascadia submarine forearc. J. Geophys. Res. 102(B4), Linke, P., Suess, E., 2001. RV SONNE Cruise Report SO148 8217^8243. TECFLUX-II-2000. GEOMAR Report 98. Greinert, J., Bohrmann, G., Suess, E., 2001. Gas hydrate-as- MacKay, M.E., 1995. Structural variation and landward ver- sociated carbonates and methane-venting at Hydrate Ridge: gence at the toe of the Oregon accretionary prism. Tectonics Classi¢cation, distribution, and origin of authigenic litholo- 14, 1309^1320. gies. In: Paull, C.K., Dillon, W.P. (Eds.), Natural Gas Hy- MacKay, M.E., Jarrard, R.D., Westbrook, G.K., Hyndman, drates: Occurrence, Distribution, and Detection. AGU Geo- R.D., Shipboard Scienti¢c Party of Ocean Drilling Program phys. Monogr. 124, 99^113. Leg 146, 1994. Origin of bottom-simulating re£ectors: Geo- Haugerud, R.A., 1999. Digital elevation model (DEM) of Cas- physical evidence from the Cascadia accretionary prism. cadia, latitude 39N^53N, longitude 116W^133W. U.S. Geo- Geology 22, 459^462. logical Survey Open ¢le Report 99-369. MacKay, M.E., Moore, G.F., Cochrane, G.R., Moore, G.F., Heeschen, K.U., Tre¤hu, A.M., Collier, R.W., Suess, E., Kulm, L.D., 1992. Landward vergence and oblique structur- Rehder, G., 2003. Distribution and height of methane bub- al trends in the Oregon margin accretionary prism: Implica- ble plumes on the Cascadia margin characterized by acoustic tions and e¡ect on £uid £ow. Earth Planet. Sci. Lett. 109, imaging. Geophys. Res. Lett. 30, 1643^1646. 477^491. Henry, P., Lallemant, S.J., Le Pichon, X., Lallemand, S.E., McNeill, L.C., Gold¢nger, C., Kulm, L.D., Yeats, R.S., 2000. 1989. Fluid venting along Japanese trenches: Tectonic con- Tectonics of the Neogene Cascadia forearc basin: Investiga- text and thermal modeling. Tectonophysics 160, 277^291. tions of a deformed late Miocene unconformity. Geol. Soc. Hovland, M., Lysne, D., Whiticar, M., 1995. Gas hydrate and Am. Bull. 112, 1209^1224. sediment gas composition, Hole 892A. In: Carson, B., West- Moore, J.C., Vrolijk, P., 1992. Fluids in accretionary prisms. brook, G.K., Musgrave, R.J., Suess, E. (Eds.), Proc. ODP Rev. Geophys. 30, 113^135. Sci. Results, 151^161. Ohta, S., Laubier, L., 1987. Deep biological communities in Hyndman, R.D., Spence, G.D., 1992. A seismic study of meth- the subduction zone of Japan from bottom photographs ane hydrate marine bottom simulating re£ectors. J. Geo- taken during ‘Nautile’ dives in the Kaiko project. Earth phys. Res. 97 (B5), 6683^6698. Planet. Sci. Lett. 83, 329^342. Johnson, J.E., Gold¢nger, C., Tre¤hu, A.M., 2000. Structural Ritger, S.D., Carson, B., Suess, E., 1987. Methane-derived development and deformation styles of the Hydrate Ridge authigenic carbonates formed by subduction-induced-pore-

MARGO 3397 29-10-03 120 J.E. Johnson et al. / Marine Geology 202 (2003) 79^120

water expulsion along the Oregon/Washington margin. methane plumes at the Cascadia convergent margin. Earth Geol. Soc. Am. Bull. 98, 147^156. Planet. Sci. Lett. 170, 1^15. Sample, J.C., 1996. Isotopic evidence from authigenic carbon- Suess, E., Torres, M.E., Bohrmann, G., Collier, R.W., Rickert, ates for rapid upward £uid £ow in accretionary wedges. D., Gold¢nger, C., Linke, P., Heuser, A., Sahling, H., Hee- Geology 24, 897^900. schen, K., Jung, C., Nakamura, K., Greinert, J., Pfann- Sample, J.C., Reid, M.R., 1998. Contrasting hydrogeologic kuche, P., Tre¤hu, A., Klinkhammer, G., Whiticar, M.J., regimes along strike-slip and thrust faults in the Oregon Eisenhauer, A., Teichert, B., Elvert, M., 2001. Sea £oor convergent margin: Evidence from the chemistry of syntec- methane hydrates at Hydrate Ridge, Cascadia Margin. In: tonic carbonate cements and veins. Geol. Soc. Am. Bull. Paull, C.K. and Dillon, W.P. (Eds.), Natural Gas Hydrates 110, 48^59. Occurrence, Distribution, and Detection. AGU Geophys. Sample, J.C., Reid, M.R., Tobin, H.J., Moore, J.C., 1993. Monogr. 124, 87^98. Carbonate cements indicate channeled £uid £ow along a Teichert, B.M.A., Eisenhauer, A., Haase-Schramm, A., Bock, zone of vertical faults at the deformation front of the Cas- B. and Linke, P., 2003. U/Th systematics and ages of authi- cadia accretionary wedge (northwest U.S. coast). Geology genic carbonates from Hydrate Ridge, Cascadia convergent 21, 507^510. margin: Recorders of £uid composition and sealevel Seely, D.R., 1977. The signi¢cance of landward vergence and changes. Geochim. Cosmochim. Acta (in press). oblique structural trends on trench inner slopes. In: Talwa- Torres, M.E., Bohrmann, G., Brown, K., deAngelis, M., Ham- ni, M., Pitman, W.C. (Eds.), Islands Arcs, Deep Sea mond, D., Klinkhammer, G., McManus, J., Suess, E., Tre¤- Trenches, and Back-Arc Basins, Maurice Ewing Series. hu, A., 1999. Geochemical observations on Hydrate Ridge, American Geophysical Union, Washington, DC, pp. 187^ Cascadia margin, July 99. Reference 99-3, Oregon State 198. University Data Report 174. Snavely Jr., P.D., 1987. Tertiary Geologic Framework, Neo- Tre¤hu, A.M., Bangs, N.L., Arsenault, M.A., Bohrmann, G., tectonics, and Petroleum Potential of the Oregon-Washing- Gold¢nger, C., Johnson, J.E., Nakamura, Y., Torres, M.E., ton Continental Margin. In: Scholl, D.W., Grantz, A., Ved- 2002. Complex subsurface plumbing beneath southern Hy- der, J.G. (Eds.), Geology and resource potential of the drate Ridge, Oregon continental margin, from high-resolu- continental margin of western North America and adjacent tion 3D seismic re£ection and OBS data. Proceedings of the Ocean^Beaufort Sea to Baja California. Circum-Paci¢c Fourth International Conference on Gas Hydrates, Yoko- Council for Energy and Mineral Resources Earth Science hama, Japan, pp. 90^96. Series 6, 305^335. Tre¤hu, A.M., Flueh, E.R., 2001. Estimating the thickness of Suess, E., Carson, B., Ritger, S., Moore, J.C., Jones, M.L., the free gas zone beneath Hydrate Ridge, Oregon continen- Kulm, L.D., Cochrane, G.R., 1985. Biological communities tal margin, from seismic velocities and attenuation. J. Geo- at vent sites along the subduction zone o¡ Oregon. The phys. Res. 106 (B2), 2035^2045. hydrothermal vents of the eastern Paci¢c: An overview. Tre¤hu, A.M., Torres, M.E., Moore, G.F., Suess, E., Bohr- Biol. Soc. Washington Bull. 6, 474^484. mann, G., 1999. Temporal and spatial evolution of a gas Suess, E., Torres, M.E., Bohrmann, G., Collier, R.W., Grei- hydrate-bearing accretionary ridge on the Oregon continen- nert, J., Linke, P., Rehder, G., Tre¤hu, A., Wallmann, K., tal margin. Geology 27, 939^942. Winckler, G., Zuleger, 1999. Gas hydrate destabilization: Westbrook, G., Carson, B., Musgrave, R., Shipboard Scienti¢c Enhanced dewatering, benthic material turnover and large Party, 1993. Init. Rep. ODP 146, 630 pp.

MARGO 3397 29-10-03