Abstract

High-resolution multibeam bathymetry and chirp (compressed high-density radar pulse) seismic data acquired from an autonomous underwater vehicle outline in unprecedented detail the shape and near subbottom character of the axial channels within upper Monterey and Soquel Canyons (offshore California, USA). In Monterey Canyon, the bathymetric data span water depths from 100 m to >2100 m, and include the confluence with Carmel Canyon at ∼1900 m water depth. The bathymetric data for Soquel Canyon begin close to the canyon head at 100 m water depth and extend down to the intersection with Monterey Canyon. The seafloor within the axis of Monterey Canyon is covered with sediment fill out to 910 m water depth. Below this water depth exposures of underlying strata are common, presumably because of decreasing sediment drape and generally increased erosional resistance of the pre-canyon host strata. The seafloor within the axial channel of upper Soquel Canyon is smooth and contains horizontally layered sediment fill. In contrast, the sediment fill within the incised portions of the axial channel of Monterey Canyon is characterized by distinctive crescent-shaped bedforms down to the limit of the surveys. These differences in morphology and texture correspond with the contrasting cohesive strength of the sediments filling these canyons and the increased propensity for weakly cohesive sands and gravels in Monterey Canyon to fail. Episodic movement of coarse-grained sediments down Monterey Canyon maintains a longitudinal gradient of ∼1.6°. The more cohesive fine-grained sediments in Soquel Canyon stabilize the seafloor and maintain a substantially higher longitudinal gradient (3°–6°) than that measured in Monterey Canyon. The textural and lithologic data, plus previously published observations, indicate that upper Monterey Canyon is currently active, whereas upper Soquel Canyon appears to be inactive as a coarse sediment transport conduit. Episodic seabed sediment failures in active submarine canyons are hypothesized to control the gradient of the axial channel. The propensity for sediment failure in weakly cohesive coarse-grained sediments results in shallower horizontal gradients compared to submarine canyons stabilized by more cohesive fine-grained sediments.

INTRODUCTION

Submarine canyons are among the most important conduits for sediment transport on the Earth (Shepard and Dill, 1966; Normark, 1974; Normark and Carlson, 2003; Normark et al., 2009b). The shape of the seafloor within a submarine canyon inevitably reflects the processes that transport sediment through the canyon, deposit sediment within the canyon, and erode into the underlying host rocks of the canyon. The existing paradigm is that turbidity currents and other gravity flow events are the dominant sediment transport processes acting within submarine canyons (e.g., Twichell and Roberts, 1982; Farre et al., 1983). While the shapes of submarine canyons evolve in response to a variety of changes driven by external factors such as climate, sea level, tectonics, and human alterations (e.g., Covault et al., 2007; Lamb and Mohrig, 2009; Mountjoy et al., 2009), their shapes ideally converge to reflect equilibrium conditions during their active phases (e.g., Pirmez et al., 2000; Mitchell, 2005; Gerber et al., 2009). Obtaining accurate and detailed bathymetric maps is essential for documenting the seafloor morphology and understanding the sedimentary processes within submarine canyons. Progress in the studies of submarine canyons has come with our ability to collect detailed bathymetric maps and use these maps to guide high-resolution sampling and accurate placement of instrumentation within submarine canyons (e.g., Paull et al., 2005, 2010).

Over the past decade, our collective understanding of the processes that occur within submarine canyons in general and specifically in Monterey Canyon (offshore California, USA; Fig. 1) has expanded enormously; there have been many surprises as the technologies have improved. Major steps include the acquisition of a regional grid of 30 kHz multibeam data that accurately delineated the sides of Monterey and Soquel Canyons (Greene et al., 2002), sampling using remotely operated vehicles (ROV) to collect Vibracores that showed that the axial channel contains coarse-grained sediment (Paull et al., 2005), and monitoring experiments that revealed the occurrence of numerous sediment transport events (Paull et al., 2003; Xu et al., 2004). The discovery that the recurrence frequency of significant sediment transport events in upper Monterey Canyon is subannual (Paull et al., 2010) was unanticipated.

Beginning in 2004, the California State University Monterey Bay Seafloor Mapping Laboratory collected 200 kHz multibeam surveys of the heads of Monterey and Soquel Canyons down to water depths of ∼250 m with sufficient resolution to image bedforms within the axial channel. These multibeam bathymetric data revealed that large crescent-shaped bedforms (herein referred to as CSBs) occur on the seafloor within the axial channel of Monterey Canyon (Smith et al., 2005, 2007). These CSBs have wavelengths up to 80 m, amplitudes to 2 m, and well-defined asymmetrical crests oriented concave down canyon and roughly perpendicular to the local canyon axis. We know that similar CSBs occur further down Monterey Canyon, and within Santa Monica, Mugu, Redondo, Carmel, and La Jolla Canyons (Paull et al., 2008). These bedforms appear to be common features on the floor of active submarine canyons.

Perhaps the greatest surprise came from the comparison of repeated multibeam mapping surveys showing that the CSBs are active, because their locations change between successive surveys (Smith et al., 2005, 2007; Xu et al., 2008). While it is clear these bedforms change position between repeated surveys conducted only months apart, differing interpretations about the migration direction have been offered. Smith et al. (2007) argued that these bedforms are sediment waves that migrate up canyon, while Xu et al. (2008) questioned whether this apparent migration direction is an artifact of the sampling interval (e.g., Hughes Clarke et al., 2009) and suggested that the CSBs may move down canyon. Both of these models assume that sediment movements are driven by the semidiurnal tidal currents that sweep through Monterey Canyon (Xu et al., 2002, 2008).

The processes that form the CSBs were discussed in Paull et al. (2010); while the CSBs were initially inferred to be migrating sand waves, that paper presented evidence that the CSBs move during brief high-energy events. The areas of the canyon floor where CSBs occur are composed of poorly sorted, coarse-grained sediments that appear to have been deposited en masse (Paull et al., 2005, 2010). Two contrasting mechanisms are proposed to form these bedforms: (1) erosion and deposition during cyclic steps in large turbidity flows as they pass over the canyon floor, or (2) failures within the canyon floor. Failure of the canyon floor may involve some combination of fluidized flow of shallow seafloor sediments and slumping of larger intact segments of the seafloor. Moreover, these mechanisms are not viewed as being exclusively one or the other, as both may occur within individual events (e.g., Anderson et al., 2006). The origin of the CSB features still remains poorly understood, but further documentation of the types of morphologies, and the extent and distribution of CSBs, will add to the understanding of how sediment moves down submarine canyons.

Physiography of Monterey and Soquel Canyons

Monterey Canyon begins in the littoral zone at the mouth of Moss Landing Harbor (Fig. 1). From there a continuous channel can be traced for more than 400 km seaward into >4000 m water depth on the Monterey Fan. Soquel and Carmel Canyons are the most prominent of the tributary canyons that coalesce with Monterey Canyon at 980 m and 1970 m water depths, respectively. This paper focuses on all of Soquel Canyon below 100 m water depth and the axial channel of Monterey Canyon from 100 m to 2100 m water depth, including the intersection with Carmel Canyon (Figs. 1 and 2).

Previously available surface ship data for Monterey Canyon show that in cross section the axial channel is a distinctly flatter area than the adjacent canyon walls (Greene et al., 2002; Paull et al., 2005; Smith et al., 2005; Fig. 3A). The axial channel width in upper Monterey Canyon varies between 100 m and 350 m and gradually increases down canyon until ∼1850 m water depth. The San Gregorio–Palo Colorado fault zone crosses the canyon axis in ∼1900 m water depth. Below the crossing of the San Gregorio–Palo Colorado fault zone the axial channel widens to more than 1000 m (SGPCFZ, Fig. 1; box 11, Fig. 2).

Upper Monterey Canyon has a sinuosity of 1.9, similar to that of a meandering river (Leopold et al., 1964; Paull et al., 2005). In comparison, Soquel Canyon has a generally straight axial channel that runs downslope to the southwest. However, at ∼450 m water depth in Soquel Canyon the general trend of the axial channel is offset by ∼650 m to the northwest, creating two ∼90° bends. These bends account for most of the sinuosity of Soquel Canyon.

Stratigraphic and Tectonic Framework of Monterey Bay

The stratigraphy of Monterey Bay and the host rocks into which Monterey Canyon is cut are known from sediment coring and HOV (human-occupied vehicle) and ROV sampling of the strata exposed on the canyon walls, boreholes drilled within the bay, and regional mapping (Greene, 1977; Greene, 1990; Stakes et al., 1999; Orange et al., 1999). The offshore stratigraphic framework is summarized in a geologic map of Monterey County, California (Wagner et al., 2002), that includes the floor of Monterey Bay. The Wagner et al. (2002) map was constructed in part using seafloor textures seen in multibeam data. The major formations believed to be exposed on the flanks of the canyon or in subcrop include sedimentary rocks of various Tertiary ages and Mesozoic basement rocks of both metamorphic and igneous origins. Of relevance here is that the erosional resistance of the units increases with their age. For example, the upper canyon is hosted in the relatively soft Pliocene–Pleistocene mudstones (e.g., Purisima Formation), while in places in the lower canyon much more solid erosionally resistant granodiorite of the Salinian Formation crop out.

The San Gregorio–Palo Colorado fault zone (SGPCFZ, Fig. 1), the westernmost active dextral fault of the Pacific–North American plate boundary, extends underneath Monterey Bay with a north-northwest–south-southeast orientation (Clark et al., 1984; Greene et al., 2002). Additional delineation of this active fault zone within the bay is provided by the pattern of low-magnitude earthquake epicenters (Greene, 1990). A zone of southwest-dipping reverse faults referred to as the Navy fault zone (NFZ, Fig. 1) crosses the inner bay with a northwest-southeast orientation. Greene (1977) mapped this using structural offsets seen in seismic reflection profiles.

Canyon Floor Cores

Previous studies (Paull et al., 2005, 2006, 2009, 2010) undertaken using an ROV-deployed Vibracoring system document the nature of the sediment facies that occur within the axis and flanks of upper Monterey Canyon (<1500 m water depth). Coarse-grained deposits form a narrow trail of sediment that is tightly restricted to the axial channel floor. Sand found on the floor of upper Monterey Canyon is supplied from the modern beach and nearshore environment (Paull et al., 2005, 2010).

The lower sidewalls and terraces on the sides of the uppermost portion of Monterey Canyon are largely cloaked with a thick drape of recent sediment supplied from the Salinas and Pajaro Rivers (Eittreim et al., 2002a, 2002b; Paull et al., 2006). This drape is composed of laminated fine-grained sediments that have accumulated in the relatively calm environment sandwiched between the wave-washed continental shelf and the sediment gravity flow–scoured axial channel on the canyon floor (Paull et al., 2005). The extent and thickness of this drape decreases with distance from shore. Outcrops of the underlying country rock on the canyon walls become common ∼18 km from shore where the canyon axis is below 910 m water depth.

Survey Goals

The data we present here were collected to explore and document the morphology of the Monterey Canyon system. A particular goal was to image the canyon's axial channel where sediment transport processes are concentrated and to determine how far into the canyon system CSBs extend. An additional impetus was to provide detailed surveys of the deep-water sections of two recently established U.S. National Marine Protected Areas (Soquel Canyon and Portuguese Ledge) near the time that these were established (Fig. 1).

METHODS

AUV Surveys

High-resolution seafloor surveys of the axis of Monterey Canyon were conducted using an autonomous underwater vehicle (AUV). This AUV was developed at the Monterey Bay Aquarium Research Institute (MBARI) specifically for seafloor mapping (Kirkwood, 2007; Caress et al., 2008). The vehicle carries a Reson 7125 200 kHz multibeam sonar and an Edgetech 2–16 kHz chirp (compressed high-density radar pulse) subbottom profiler. The AUV simultaneously collects high-resolution multibeam bathymetry (vertical precision of 0.1 m and a horizontal beam footprint of 0.75 m diameter at 50 m survey altitude) and chirp seismic-reflection profiles (vertical resolution of 0.11 m). The AUV navigation and attitude derive from a Kearfott Seadevil inertial navigation system initialized by the global positioning system at the surface and aided by a Doppler velocity log (DVL) during missions. These surveys were conducted at 3 knots with the AUV programmed to fly 50 m off the seafloor on missions as long as 18 h. To cover the floor of the axial channel required flying 3 or more overlapping swaths that were nominally 150 m apart.

The data presented here were collected during 16 AUV dives in Monterey, Soquel, and Carmel Canyons (Fig. 1). To enable continuous DVL bottom tracking, the AUV was launched from the R/V Zephyr on the continental shelf in <100 m water depth. For these surveys four different launch points were used. A series of surveys were conducted in Monterey Canyon on 22 and 24 September 2008 from the canyon head down to ∼950 m water depth. Additional surveys were conducted on 1 and 2 December 2008 from ∼1400 m water depth up canyon to overlap with the previous surveys. On 11 and 12 March 2009, 2 surveys were conducted to continue mapping down canyon to ∼2000 m water depth. On 2 December 2009 a survey was conducted to map the floor of Monterey Canyon at the intersection with Carmel Canyon. Two additional surveys were conducted on 9 December 2009 to include more of the lower sidewalls of upper Monterey Canyon, and on 25 June 2009 the southeastern sidewalls of the canyon south of Navy slump were surveyed. Six surveys were conducted in Soquel Canyon (28 and 29 March 2008; 22 and 23 July 2008; and 5 and 10 March 2009), covering the canyon floor and sidewalls from the canyon head on the continental shelf down to the confluence with Monterey Canyon.

The AUV data were processed using MB-System, an open-source seafloor mapping software package (Caress and Chayes, 1996; Caress et al., 2008); 1 m grids of the data were generated. The long profile following the thalweg of these canyons was constructed by connecting the up-canyon apex of 1 m contours along their axial channels.

ROV Sampling and Observations

MBARI has operated one or more ROVs (Ventana, Tiburon, and Doc Ricketts) within Monterey Bay for more than two decades. These ROVs have made more than 4000 dives within the bay; however only a modest fraction of the total dive effort has been focused on benthic observations and an even smaller portion of the dives targeted the seafloor of the Monterey Canyon system. Some results of these investigations have been published (Paull et al., 2003, 2005, 2006, and 2009). Because the bottom water of Monterey Canyon is characteristically quite turbid, it is usually difficult to make visual observations and collect sediment or rock samples on or near the seafloor. During clear water periods, the nature of the canyon floor has been visualized and periodically rock samples have been collected from strata outcrops on the canyon walls using the mechanical arm of the ROV. ROV dive observations and samples associated with particular morphologic features seen in the multibeam data are used to provide ground truth.

RESULTS

The new AUV-acquired high-resolution bathymetry focuses on the axial channels of Monterey and Soquel Canyons. The surveys cover the axial channel of Monterey Canyon from 100 m to 2100 m water depths, Soquel Canyon from between 104 m to its intersection with Monterey Canyon in 1210 m water depth, and Carmel Canyon below 1680 m water depth (Figs. 1 and 2). Terminology used to describe canyon morphology is illustrated in Figure 3. The areas below ∼350 m water depth in both canyons have never been mapped in this detail before. Images of the entire data set are in Supplemental Maps 1–5 in the Supplemental File1. Figures 4–15 illustrate the bathymetry of selected areas from Monterey Canyon; they have been arranged and numbered to extend down Monterey Canyon in order of increasing water depth of the canyon thalweg. The character of Soquel Canyon is illustrated in Figures 16 and 17. Although the surveys were not designed to cover the lowermost portions of the canyon sidewalls, small sections of the canyon flanks were inevitably imaged. ROV-collected video images of features characteristic of the axis of Monterey Canyon are shown in Figure 18.

High-Resolution Bathymetry of Monterey Canyon

Canyon Walls

From the canyon head out to an axial channel depth of 910 m (Figs. 4, 5, and 6), the texture of the canyon walls is generally smooth, suggesting that the underlying host strata are draped with recent sediment. Below 910 m depth (Figs. 7A, 7B) the texture of the bathymetry indicates that the canyon walls are rougher. In places, protrusions from the sidewalls extend into the path of the axial channel, and isolated canyon floor topographic highs occur (CFTH; Figs. 7, 9, 10, 11, and 12). In places these rougher-textured canyon wall surfaces, protrusions on the sidewalls, and canyon floor topographic highs have a distinct linear fabric that is suggestive of bedded strata outcrops (BS; Figs. 6, 7, and 9).

In a few areas the AUV-acquired bathymetry covers some sections of the canyon walls above the canyon floor (Figs. 4, 8, 14, and 15). The walls of the canyon contain numerous slope-parallel ridges and troughs. Many of the troughs contain stairsteps of arcuate features (AF) that suggest they were formed by submarine mass-wasting events, consistent with previous interpretations of surface vessel multibeam surveys within Monterey Canyon (Greene et al., 2002; Smith et al., 2005) and elsewhere (e.g., McAdoo et al., 2000; Canals et al., 2004). Here AF is used to describe features on the canyon walls and CSB to describe features on the canyon floors.

The occurrence of AFs and exposures of bedded strata, indicative of erosion, are especially common on the canyon sidewalls along outside bends of the Monterey Canyon axial channel (AF; Figs. 4 and 5). However, many of the scars on the canyon walls adjacent to the channel occur on the upslope side of the bends (Figs. 4 and 5), rather than on a path that is directly downslope of the previous straight canyon segment.

CSBs

The AUV-collected multibeam bathymetry show that distinctive bedforms occur within the axial channel of Monterey Canyon down to at least 2100 m water depth (Figs. 4–14; Supplemental Maps 1–4 [see footnote 1]). Repetitions of CSBs with a concave down-canyon shape are common within the incised portion of the axial channel of Monterey Canyon down to 1280 m water depth (CSB; Figs. 3B, 3C, and 4–10). Below 1280 m water depth some large reverse curvature bedforms (RCBs) (e.g., concave up-canyon crests) also occur (RCB; Figs. 10–12).

In water depths <1280 m the CSBs are consistently asymmetric, with a relatively steep (∼15°) 1–2-m-high scarp on the down-canyon side and a longer flank on the up-canyon side that is approximately level or dipping gently (to 1°) up canyon (CSB; Figs. 4–9 and 19). In some places the CSB scarps are composed of two or more concave-downward segments (CSB; Figs. 6, 7, and 8).

Individual well-formed CSBs are easily identified in map view (Figs. 4–9). We counted 371 well-formed CSB scarps that crossed the Monterey Canyon thalweg between 160 m and 1280 m water depth (Supplemental Maps 2 and 3 [see footnote 1]). These are spaced at intervals of 50 ± 28 m (1σ) with minimum and maximum separations of 12 m and 167 m. However, this count is subjective, as other less distinct features could also be counted as CSBs.

The CSBs appear to occur in groups of 4–20 that are morphologically similar in size and shape (CSB, Fig. 19). These groupings are distinguished by an uppermost scarp that is distinct by being somewhat higher and thus forming small knickpoints (SKP; Figs. 5A, 9A, and 19) or distinctly narrower than the CSB scarps immediately upstream (CSB; Figs. 5A and 6A). The appearance of the CSBs in the multibeam images is characteristically sharper immediately down canyon from the distinctive scarps identified as small knickpoints and tend to become less distinct down canyon within an individual group of CSBs. The length of these groupings of CSBs varies from ∼200 m to >2 km and their down-canyon termination is commonly associated with another distinct small knickpoint scarp, which marks the beginning of the next group of CSBs further down canyon.

When traced laterally, the crests of individual CSBs curve down canyon and merge with the sides of the axial canyon or the slightly taller scarps that bound the benches on the flanks of the incised part of the axial channel (Figs. 3B, 3C, 4, 5, 6, 8, 9, and 10). Multiple CSBs commonly merge with the same sidewall scarp.

The cross-channel widths of the CSBs vary considerably over relatively short distances along the canyon. For example, in 690 m water depth (CSB, Fig. 6A) the width of a distinct CSB is only 30 m. From this location, the widths of the CSBs increase down canyon until they extend over 200 m across the channel 1 km further down canyon. In some places the changes in width of trains of CSBs are associated with constrictions in the channel (C, CSB; Figs. 9A and 10A), but in other areas the changes in CSBs do not appear to be related to bedrock controls within the axial channel axis (CSB; Figs. 5A, 6A, and 7A).

The CSBs are usually in the incised and deepest part of the axial channel associated with the canyon thalweg. However, some groups of CSBs appear to be continuous with AFs on the canyon sidewall. For example, in ∼1920 m water depth, well-developed CSBs within the axial channel bifurcate upslope in at least four separate directions (CSB, AF, Fig. 13). Only one of these trains is connected with the main channel of Monterey Canyon. The features identified as AFs are morphologically similar to the CSBs in that they are concave-downward scarps. While there is a greater range in the size of the AFs, the primary distinction is their position with respect to the canyon floor versus its flanks. Another example of the CSB is illustrated in ∼1310 m where a train of CSBs can be traced up canyon into a headless reentrant on the canyon floor (CSB; Figs. 10 and 19).

The deepest occurrence of CSB-like features is below the confluence with Carmel Canyon (CSB, Fig. 14). However, here their appearance in the multibeam data is less distinct than at shallower depths further up canyon.

RCBs

Although much of upper Monterey Canyon seafloor is covered with trains of CSBs, which have concave down-canyon scarps, there are two areas where somewhat similar bedforms occur with the opposite curvature (RCB; Figs. 10A, 11, and 12). One of the areas with the RCBs is in 1265–1284 m water depths (RCB, Fig. 10A) just upstream from a large constriction within the canyon axis referred to as the Navy slump (Greene et al., 2002). The other area with RCBs is between 1810 and 1875 m water depths where the walls of the upper canyon open out into an ∼1-km-wide basin (RCB; Figs. 11 and 12). Unlike the CSBs, the RCBs are not associated with local incision in the canyon floor fill (CSB, Figs. 3C, 3D). Instead, the crests of the RCBs fade out laterally within the canyon floor rather than terminating along a side scarp. The RCBs are also larger (to 10 m high and to 120 m long), less asymmetric, and lack the distinctive downslope scarps characteristic of CSBs (Fig. 19). ROV transects conducted during Tiburon dive T1136 and Doc Ricketts dives DR17, DR18, DR108, DR110, and DR112 show that the seafloor where these bedforms occur is composed of angular gravel and boulders of granodiorite that crop out through a veneer of sediment (Figs. 18F–18H). Many of the observed boulders are >1 m on an edge.

Benches on the Margins of the Incised Channel

Common features observed along the margins of the incised channel of Monterey Canyon down to ∼1100 m water depth are elevated benches (B; Figs. 4–9). The benches separate the canyon sidewalls from the slightly deeper incised channel on the canyon axial channel floor that contains CSBs (Fig. 3C). These benches have relatively smooth surfaces that dip down canyon with slopes similar to the seafloor of the incised channel (<2°). Linear scarps (LS; Figs. 4 and 6) 2–10 m higher than the incised axial channel floor define the boundary between the benches and the incised channel (Figs. 4, 6, and 8). Smaller scarps that form the crests of CSBs curve into and often merge down canyon with the scarp that defines a bench. The linear scarps that bound the benches can approach 1 km in length and are generally oriented parallel to the incised channel axis. In some cases benches form triangular-shaped wedges on the inside bends of meanders (B; Figs. 4, 5A, and 6A). Elsewhere benches are absent on the inside bends of meanders (Figs. 6A and 7A).

Outcrops on the Canyon Floor and Canyon Constrictions

The shallowest occurrence of pre-canyon strata outcrops within the axial channel of Monterey Canyon is in 910 m water depth, where a distinct linear ridge occurs within the incised channel. Here channel floor topographic highs up to 10 m across rise ∼4 m above the seafloor. They form a discontinuous ridge ∼100 m long (CFTH, Fig. 7A). Additional ridges 6 m high and 20 m long rise above the canyon floor ∼120 m further down canyon. The adjacent canyon walls show lineations and apparent stairsteps that are interpreted as outcrops of subhorizontal bedded strata (BS, Fig. 7A). These mid-channel topographic highs and outcrops of bedded strata form a constriction in the canyon floor.

Several other distinct constrictions of the incised channel occur further down canyon (C; Figs. 9A, 10A, and Fig. 13). A constriction in Monterey Canyon in 1060 m water depth is created by 2 blocks or outcrops more than 60 m wide that form a partial dam across the incised channel (C, Fig. 9A). A 36-m-wide slot cuts through this dam and there is more than 10 m of relief immediately down canyon. Observations and sampling during ROV Ventana dive V3347 show that the down-canyon face is composed of massive carbonate-cemented sandstone. Similar appearing outcrops or large blocks that rise above the canyon floor are common below this feature and define other constrictions in the incised channel. Just 200 m further down canyon a headland protrudes 200 m from the northern wall into the canyon axis, reducing the channel width from >350 m to <150 m. Observations made during ROV Doc Ricketts dive DR40 show that the face of this feature is composed of lithified bedded rock (Fig. 18C).

The most distinctive constriction in upper Monterey Canyon (Fig. 10) is associated with the feature called the Navy slump (Greene et al., 2002), a large promontory that juts out into the canyon from the east, forming an ∼1000-m-wide, 900-m-long, and ∼125-m-high obstruction in the axial channel between 1300 m and 1400 m water depths. The canyon thalweg follows a circuitous path around the northwestern edge of the Navy slump and passes through an ∼1-km-long straight constriction where the width of the channel decreases to 38 m. ROV dives in this narrow constriction show that the seafloor is composed of gravel, and the steep, nearly parallel sidewalls are composed of massive competent rock (ROV dives V2077, V3001, and V3302; Fig. 18E). ROV transects over Navy slump reveal that it is largely sediment bare and lacks loose rock. A few rock samples have been obtained from both sides of the constriction, and show that the Navy slump is composed of Salinian granodiorite on both sides (Wagner, et al., 2002). The bathymetry shows that the surface of the Navy slump is rough, and distinct linear troughs (DLT, Fig. 10) cross much of the feature. The largest of these troughs is >350 m long, 10–20 m across, and as much as 40 m deeper than the surrounding rock mass (DLT, Fig. 10B). These open troughs appear to have relatively flat floors that form a surface that connects downslope to the axial channel floor, suggesting that these troughs are partly filled with sediment.

In 1850 m water depth one area with a diameter of ∼40 m is ∼3 m higher than the surrounding seafloor (CFTH; Figs. 11 and 12A). Observations and sampling on ROV dive DR132 show that this canyon floor topographic high is composed of massive granodiorite and the seafloor inside and outside the moat comprises gravel and angular boulders of granodiorite covered with a discontinuous veneer of fine sediment.

Bottom Roughness and Canyon Floor Debris

The 1 m bathymetric grids show that the seafloor within the incised channel of Monterey Canyon has considerable roughness (Figs. 5B, 6B, 7B, 9B, 10B, and 19). The data show an irregular texture, suggesting that tens of centimeters to meter-scale local relief is common within the incised channel (Fig. 19). Rough textures occur within areas where CSBs are identified and in areas where CSBs are not developed. In places, the distinctive textures in the bathymetric data are easily identified as debris on the canyon floor (e.g., Fig. 5). The detailed longitudinal gradient plots (Fig. 20) also illustrate how local bottom roughness obscures the identity of the CSBs.

Fault Zones

Two regional fault zones cross Monterey Canyon (Fig. 1). The feature called the Navy slump (NS, Fig. 10A) occurs where the western margin of the Navy fault zone extends across the canyon (Greene et al., 2002). The long straight narrow constriction that extends alongside the Navy slump is parallel with the northwest orientation of the Navy fault zone as mapped on the shelf adjacent to both sides of the canyon (Greene, 1977).

A linear section of the incised channel between 1900 and 2000 m water depths is oriented north-northwest–south-southeast. Here there are two series of isolated, elongate canyon floor topographic highs that are ∼250 m apart (CFTH, Fig. 13). These linear arrays of topographic highs are nearly parallel with the general north-northwest trend of the lower portion of Carmel Canyon (Fig. 1) and have been used to map the San Gregorio–Palo Colorado fault zone using preexisting bathymetry (Greene et al., 2002). Higher-resolution multibeam bathymetry shows that there are north-northwestward–oriented gullies cutting the base of the northern wall of Monterey Canyon oriented parallel to the San Gregorio–Palo Colorado fault zone and Navy fault zone (G, Fig. 15).

Chirp Subbottom Profiles from Monterey Canyon

The AUV-collected chirp profiles from the flanks and benches on the side of Monterey Canyon resolve numerous nearly seafloor-parallel reflectors indicating that thinly bedded ∼15–30-m-thick sediments drape the canyon walls (0.02–0.04 s two-way traveltime) (Figs. 21A, 21B, and 21E). In places on the canyon walls, these thin-bedded sediments are offset by small faults (Fig. 21A) that appear to be rotational faults that sole out toward the canyon axis (Smith et al., 2005, 2007). Strike profiles show that the draping layers are frequently disrupted by gullies on the canyon flanks (G; Figs. 15 and 21E). The coincidence of the interruptions in the continuity of the drape and the existence of the gullies make it difficult to establish if the discontinuities seen in the chirp profiles are associated with recent displacements along the San Gregorio–Palo Colorado fault zone. The AUV-collected chirp profiles over both the typical CSBs and those with reverse curvature from the floor of Monterey Canyon show the seafloor as being one strong reflector, and lack indications of internal layering below (Fig. 21C).

High-Resolution Bathymetry of Soquel Canyon

Soquel Canyon is an ∼10-km-long tributary that extends northeast of Monterey Canyon (Fig. 1). The canyon head is located mid-shelf ∼10 km from the nearest shoreline. AUV-acquired multibeam bathymetric data reveal the shape of the axial channel from between 104 m water depth down to its intersection with Monterey Canyon in ∼980 m water depth. Surveys of Soquel Canyon were collected using various AUV survey patterns, with some focused on imaging the floor of the axial channel, and others covering portions of the canyon flanks.

Upper Soquel Canyon comprises two straight segments (2.5 km and 4 km long) oriented north-northeast–south-southwest, offset by a shorter 0.6 km segment oriented west-northwest–east-southeast, forming two right-angle bends (Figs. 2 and 16; Supplemental Map 5 [see footnote 1]). The high-resolution bathymetry shows the floor of upper Soquel Canyon as being a relatively smooth flat-floored channel 200–350 m wide and with a thalweg slope of 3°–5° down to 650 m water depth (Fig. 16).

Only a few scarps were imaged within the upper Soquel Canyon surveys, and they are primarily on the canyon sidewalls. One cluster occurs on the lower flanks of the canyon in ∼420 m water depth, upstream of the right-angle bend above the axial channel (AF, Fig. 16). These scarps are apparently slide scars associated with failures on the canyon walls.

The boundary between upper and lower Soquel Canyon occurs where the axial channel water depth is 650 m (Supplemental Map 5 [see footnote 1]). Below 650 m water depth the channel topography becomes more rugged and the slope increases to ∼7° (Fig. 17). This is the first indication of possible exposures of country-rock outcrops on the canyon floor; the bathymetry suggests that these outcrops are the eroded edge of gently dipping bedded strata.

The transition from smooth seafloor to an incised channel with outcrops of bedded strata occurs below 720 m water depth (Fig. 17; Supplemental Map 5 [see footnote 1]). The bottom texture of the lowermost flanks of the canyon sidewalls become distinctly rougher, and the outcrops appear to lack the layering observed slightly up canyon (Fig. 17). Two very tight constrictions occur within lower Soquel Canyon. At 760 m there is a constriction that narrows to as little as 5 m width, which is ∼30 m deeper than its adjacent sides (C, Fig. 17). Another similarly narrow constriction occurs at 790 m water depth (C, Fig. 17). Observations made on Tiburon dive T235 (conducted by H.G. Greene in ∼800 water depth) revealed that the walls of the incised canyon at these constrictions are composed of massive competent rock, which the ROV was unable to sample. The Wagner et al. (2002) geologic map indicates that a Tertiary volcanic unit occurs in this area.

The western edge of lower Soquel Canyon contains two morphologic features that resemble submarine plunge pools (Lee et al., 2002b). They are nearly circular, closed depressions with a diameter of ∼300 m (PP, Fig. 17). They are aligned along a north-northeast–south-southwest trend parallel to the canyon axis spaced ∼500 m apart in a branch of Soquel Canyon. The deeply incised channel of Soquel Canyon bypasses the plunge pools.

The multibeam bathymetry reveals that the characters of the southeast and northwest flanks of Soquel Canyon are different (Fig. 16). The southeast side is covered with comparatively smooth sediment drape with only an occasional exposure of what appears to be bedded strata. In contrast, most of the northwest side is a rugged topography with numerous deeply dissected gullies filled with recent sediments.

The confluence of Soquel Canyon with Monterey Canyon occurs in a water depth of 960 m (Fig. 8), forming a 200-m-deep, ∼100-m-wide broad embayment on the north side of the canyon. The gradient in Monterey Canyon does not change, nor is there a break in the pattern of the CSBs within the incised channel at the intersection of Soquel Canyon. Seafloor within Monterey Canyon forms a smooth gentle slope that extends beyond the confluence without a noticeable change in the character of the axial channel.

Chirp subbottom profiles show that the canyon floor fill in upper Soquel Canyon is associated with discontinuous horizontal reflectors, which continue through the depth resolved in these profiles (Fig. 21D). These reflectors suggest that the axial channel is filled with subhorizontal layered sediments that are laterally discontinuous.

High-Resolution Bathymetry of Lower Carmel Canyon and Monterey Canyon Confluence

Carmel Canyon joins the southern side of Monterey Canyon at a water depth of ∼1950 m (Fig. 14). Carmel Canyon heads at the shoreline in Carmel Bay in close proximity to the mouth of the Carmel River (Fig. 1). However, the AUV surveys covered only the lowermost 3 km of Carmel Canyon from below 1680 m water depth to the confluence with Monterey Canyon (Fig. 11; Supplemental Map 4 [see footnote 1]).

The AUV multibeam images show two distinct trains of CSBs in Carmel Canyon between 1730–1850 m and 1900–1920 m water depths (CSB; Fig. 11). These trains of CSBs are similar in amplitude and wavelength to those that occur throughout most of the incised channel of upper Monterey Canyon.

The coalescence between Carmel and Monterey Canyon occurs within the San Gregorio–Palo Colorado fault zone (Fig. 1) and the orientations of distinct linear scarps (LS, Fig. 11) along the canyon walls have been used to map this fault system (Greene, 1977). The orientation of the last segment of Carmel Canyon (Figs. 11 and 14; Supplemental Map 4 [see footnote 1]) and its main trend (Fig. 2) are parallel with Monterey Canyon at their confluence, but opposite in slope direction. The confluence area is one of the few sections where the incised channel of Monterey Canyon is smooth and without CSBs. Down canyon from the confluence the seafloor becomes rougher and CSBs reappear (Fig. 14).

Soquel and Monterey Canyons Thalweg Comparison

The high-resolution multibeam bathymetric data show a considerable contrast in the morphology of the axial channels within Monterey and Soquel Canyons. Soquel Canyon has thalweg gradients that range from 3° to 7° (Fig. 20). Steeper slopes occur within lower Soquel Canyon below 650 m where tight constrictions in the incised channel occur and outcrops of indurated host rock are exposed on the canyon walls. The thalweg in upper Monterey Canyon is distinctly less steep, averaging 1.6° between 160 m and 1216 m water depths (Fig. 20). Over this section of the Monterey Canyon our trace of the thalweg is essentially linear (R2 = 0.999). The slope within the narrow constriction that bypasses the Navy slump is 3.7° between 1290 m and 1432 m water depths. Below this constriction, the slope of the thalweg averages 2.3° between water depths of 1470 and 1756 m.

DISCUSSION

Changes in Seafloor Texture along the Canyon Walls

The textures observed in the high-resolution multibeam bathymetry from the canyon walls reveal the distribution of exposed outcrops. The change from the characteristically smooth seafloor texture (indicative of sediment drapes) near the heads of Monterey and Soquel Canyons to rougher textures (indicative of strata outcrops) further offshore is attributed to diminished sediment drapes with distance from the primary nearshore sediment sources, including the outflow from the Salinas and Pajaro Rivers (Paull et al., 2006).

Asymmetries in the roughness of the canyon sidewalls are also noticeable in the segments where the meanders in their axes trend approximately north-south and parallel to the shoreline (Figs. 11 and 16). These asymmetries are attributed to variations in the sediment supply from nearshore sediment sources, which preferentially bury the eastern walls of these canyons and leave more sediment-bare rock outcrops exposed along their western flanks.

Exposure of Host Strata

The textures observed in the multibeam bathymetry from the lower walls of Monterey and Soquel Canyons also show areas of bedrock outcrops and reveal trends consistent with the exposure of increasingly competent host-rock lithologies with depth. There is no evidence of bedrock exposures occurring on the seafloor of the axial channel of Monterey Canyon shallower than 900 m water depths, and the canyon walls are largely draped with recent sediments (Supplemental Map 2 [see footnote 1]). In a few places older strata of the Purisima Formation or Santa Cruz Mudstone (Wagner et al., 2002; Powell et al., 2007) may be exposed or occur in subcrop, but they are poorly lithified and easily eroded formations. The canyon wall forming the interior bend of the meander at 900 m water depth (BS, Fig. 7) is mapped as being composed of lithified diatomites of the Monterey Formation (Wagner et al., 2002), moderately competent rocks that are erosion resistant (Garrison and Douglas, 1981). A few blocks of material protrude from the axial channel fill that could be either outcrops or debris from the canyon walls (CFTH, Fig. 7). The massive outcrops on the walls of lower Soquel Canyon (Fig. 17; Supplemental Map 5 [see footnote 1]) are composed of graywackes and peridotites, presumably part of the Franciscan Formation (Wagner, et al., 2002; Wentworth et al., 1984).

Below 1060 m water depth (BS, Fig. 9A) occasional large blocks of strata of unknown age rise above the axial channel fill. Here the host rock is believed to be sedimentary in origin and the AUV multibeam data show what appears to be outcrops of gently tilted bedding surfaces (BS, Fig. 9A). These exposures of stratified rock could be either in place or huge blocks of material that have fallen onto the canyon floor. However, the broadly consistent orientation of apparent bedding between 1060 m and 1280 m suggests that these outcrops are part of a coherent stratigraphic sequence that extends intact into the subsurface under the canyon.

Further down Monterey Canyon, the feature referred to as the Navy slump forms the most distinct and longest constriction in the axial channel (C, Fig. 10A). Multibeam data show that the surface of the Navy slump has numerous open, headless gullies that appear to be developed along jointing planes or faults in well-indurated rock (Fig. 10B). The name Navy slump implies that it is a large slump mass that has fallen off the canyon wall and partially dammed the canyon; however, an alternate interpretation is that this feature was formed by differential erosion of rocks juxtaposed by movement along the Navy fault zone (Fig. 2; Supplemental Map 4 [see footnote 1]). Upstream of the Navy slump canyon floor topographic highs are layered, suggesting a sedimentary origin for host rock in this portion of the canyon; the Navy slump is the first known occurrence of Salinian granodiorite in the canyon. Salinian granodiorites continue to crop out down canyon and around the major meander bend to where the San Gregorio–Palo Colorado fault zone crosses Monterey Canyon in ∼1900 m water depth (Figs. 2, 11, 13, and 14; Supplemental Map 4 [see footnote 1]).

Insight into Canyon Carving

The high-resolution multibeam images of Monterey and Soquel Canyons provide a snapshot of the morphology of the axial channels and lower walls of these canyons. This snapshot provides some insight into the ongoing erosional processes shaping the seafloor and their potential long-term impact on canyon morphology.

Erosion within Monterey Canyon

The documentation that multiple sediment transport events have occurred recently in Monterey Canyon shows that it is an active sediment transport conduit (Smith et al., 2007; Xu et al., 2002, 2004, 2008; Paull et al., 2003, 2005, 2009, 2010). The significance of the recently observed events for canyon incision is less clear.

Outcrops of the host rocks into which Monterey Canyon is cut occur intermittently along the walls of the canyon and occasionally protrude through the sediment fill within the axial channel below 900 m water depths (Figs. 7 and 9–14). ROV observations show that these exposed outcrop surfaces are scoured (Fig. 18D), and the lack of colonization by encrusting organisms also indicates recent abrasion (Paull et al., 2009). While turbidity currents that scour and undercut the canyon walls will act to increase the width of the canyon, locally steepen the canyon walls, and even stimulate slumping on the canyon walls, this process will not directly contribute to increased canyon incision where the host rocks are covered with fill. In order to increase the canyon incision the host rocks underlying the axial channel need to be directly exposed to the erosive forces (Baztan et al., 2005).

The majority of the seafloor of Monterey Canyon is currently covered with coarse-grained sediment fill (Paull et al., 2005). Unfortunately, the thickness of the canyon floor fill and the depth to the contact between the canyon fill and the preexisting country rock into which Monterey Canyon is carved are unknown. Neither surface ship seismic reflection profiles nor AUV chirp subbottom profiles have imaged the base of the canyon axis fill. If the fill is only a relatively thin veneer overlying the pre-canyon host rocks, periodic scours may reach through this cover and erode into the host rocks. In this case, canyon incision may be an ongoing semicontinuous process affecting many small segments of the canyon at different times, not dissimilar to the progressive erosion in terrestrial incised valleys (Blum and Tornqvist, 2000; Strong and Paola, 2008).

The thickness of the unconsolidated sediments of the canyon fill in upper Monterey Canyon can be estimated by projecting the canyon walls downward into the subsurface. Projections of the sidewall slope could accommodate as much as ∼30 m of fill (Paull et al., 2005). If the present sediment fill is both immobile and thicker than the level that the common events scour, the presence of the fill will insulate the underlying host rocks from erosional downcutting. Thus, while Monterey Canyon is currently an active sediment transport conduit, the observed gravity flow events may not be effective for cutting the canyon into its host rocks. If there are times when the volume of canyon fill is reduced substantially, allowing scours within periodic gravity flow events to cut into the host rock, it is unclear under what conditions this occurs. Conversely, there may be movements within the sediment fill that erode the host rocks without requiring their exposure on the canyon floor (discussed in the following).

Erosion within Soquel Canyon

The high-resolution multibeam data from Soquel Canyon reveal an erosional history different from that of Monterey Canyon. The seafloor of upper Soquel Canyon is relatively smooth, and covered with relatively fine grained layered sediments out to 610 m water depth (Fig. 21D), suggesting that upper Soquel Canyon is not currently undergoing erosion. The absence of coarse-grained sediment is presumably because the mid-shelf position of Soquel Canyon head leaves it isolated from direct access to a supply of sand associated with the beach and nearshore environment. Below 610 m water depth, the longitudinal gradient becomes steeper (Fig. 20), sediment cover decreases, and outcrops become more abundant toward the confluence with Monterey Canyon. Deep incision of the channel into host rock occurs between 750 m and 850 m water depths. Arcuate features on the canyon walls cut into the pre-canyon host rocks suggest the importance of sediment failure on the development and enlargement of Soquel Canyon (AF, Fig. 17). The series of closed depressions similar to submarine plunge pools described by Lee et al. (2002b) (PP, Fig. 17) and scours described by Normark et al. (2009a) appear to be erosional features. Although the longitudinal gradient in lower Soquel Canyon (to 7°) far exceeds that of Monterey Canyon (∼1.6°; Fig. 20) at the intersection of the two canyons, Soquel Canyon does not appear to be an active conduit for coarse sediment transport. The lower portion of Soquel Canyon shows evidence of deep erosion into its host rock. Whether erosion within the lower portion of Soquel Canyon is ongoing is unclear. However, the long profile shape of Soquel Canyon (Fig. 20) is probably inherited.

Monterey Canyon Long Profile

The contrast between the long profiles of Monterey and Soquel Canyons is notable (Fig. 20). A 38.7-km-long segment between 160 and 1280 m water depths, where the long profile shape is distinctly linear, has an average slope of ∼1.6°. The long profile shape outlined in Figure 20 smooths out the smaller scale variations associated with the canyon floor bedforms (Fig. 19).

The persistent 1.6° average slope of the canyon axis between 160 and 1280 m water depths is the lowest sustained average slope in either canyon. While this segment of Monterey Canyon appears to be pinned by the bedrock control (Mitchell, 2006) associated with the Navy slump at its downstream end (Fig. 10A), this slope is developed within a segment of the canyon that is mostly free of bedrock control (Figs. 4–8). The 1.6° sloping segment of the canyon is where the recurring trains of CSBs are best developed and where the incised portion of the axial channel is covered with coarse-grained event deposits (Paull et al., 2005, 2010). It is curious that this segment of the canyon has had documented sediment transport events at a subannual recurrence frequency. The persistence of the 1.6° slope throughout the most active segment of the canyon floor may reflect equilibrium with respect to modern conditions. Apparently the ongoing sediment transport events within this section of the canyon do not generate sustained average slopes that exceed 1.6°. We speculate that sections of the canyon floor filled with the relatively cohesionless sediments that significantly exceed slopes of 1.6° are more susceptible to periodic failure.

The segment within the linear long profile in upper Monterey Canyon differs from the concave profiles characteristic of terrestrial river channels and many submarine turbidite channels (e.g., Pirmez et al., 2000; Kneller, 2003; Pirmez and Imran, 2003; Mitchell, 2005; Gerber et al., 2009). Models suggest that submarine canyons with linear long profiles may reflect a balance between erosion and deposition within the canyon floor (e.g., Gerber et al., 2009).

Canyon Floor Bedforms

Repetitive trains of large wave-like bedforms have been found in a number of marine settings (e.g., Piper and Savoye, 1993; Lee et al., 2002a; Wynn and Stow, 2002; Spinewine et al., 2009). Understanding the mechanisms responsible for their formation has been a long-standing goal of the geological community.

When the CSBs were discovered within the axial channel of uppermost Monterey Canyon, data with adequate resolution to see these features were only available to 250 m (Smith et al., 2005). The existing AUV data show that CSBs occur throughout most of Monterey and Carmel Canyons down to 2100 m, the limit of the existing mapping AUV surveys. However, the less crisp appearance of the CSBs in the multibeam data below the confluence of Monterey and Carmel Canyons suggests that these CSBs are perhaps somewhat older features that have been modified by sediment draping. This observation is consistent with the recurrence interval of transport events within the axis of Monterey Canyon decreasing below 2 km water depths (Paull et al., 2009).

Sand Waves?

The first hypothesis as to the origin of CSBs on the floor of Monterey Canyon proposed that they were sand waves and/or dunes. The CSBs were inferred to migrate along the canyon floor, driven by traction associated with the tidal currents (Smith et al., 2005, 2007; Xu et al., 2008). Large seafloor sand waves are known to form and migrate in areas where the bottom undergoes strong (e.g., >100 cm/s) currents (e.g., Wynn and Stow, 2002). However, in Paull et al. (2010), it was shown that boulder-size objects buried within these features periodically move, the CSBs are formed as en masse deposits, and they lack a predictable internal architecture characteristic of migrating bedforms (e.g., Wynn and Stow, 2002). While some relatively high current velocities have been measured in discrete events within the axis of Monterey Canyon, the maximum daily tidal current is usually considerably <80 cm/s (Xu et al., 2008). Such current velocities are marginal for moving coarse sand, and inadequate for moving gravel, let alone boulders (Hjulstrom, 1935). The inclusion of angular boulders on and within the axis throughout the canyon (Figs. 18F, 18G) is further confirmation that these bedforms are deposited in discrete high-energy gravity flow events. Moreover, the materials within the bed move more than one wavelength (i.e., as much as 900 m) down canyon during individual brief sediment transport events. Thus, the hypothesis that the CSBs progressively migrate due to the tractive forces associated with regular tidal currents was rejected (Paull et al., 2010).

Cyclic Steps, Soquel Canyon Plunge Pools, and CSBs

Recently the CSBs within the axis of upper Monterey Canyon have been used as examples of bedforms developed by cyclic steps in turbidity currents (Cartigny et al., 2011; Kostic, 2011). The concept that repetitive hydraulic jumps, referred to as cyclic steps, occur within flows where alternations between shallow, swift supercritical flow (Froude number > 1) and thick, tranquil subcritical flow (Froude number < 1) occur was initiated by Parker (1996). Originally, cyclic steps were invoked to explain repetitive bedforms within river channels. Theory, laboratory experiments, and numerical models indicate that cyclic steps in energetic flows can create repetitive bedforms in a variety of environments (Koyama and Ikeda, 1998; Parker and Izumi, 2000; Sun and Parker, 2005; Taki and Parker, 2005; Kostic and Parker, 2006). An understanding of the impact of cyclic steps within turbidity currents on the generation of large bedforms is still developing.

In the marine realm trains of scour depressions, erosional scarps, and sediment waves have been attributed to both erosional and depositional processes generated by cyclic steps in turbidity currents (e.g., Lee et al., 2002a; Fildani et al., 2006; Lamb et al., 2008; Normark et al., 2009a; Heinio and Davis, 2009). Numerical models have simulated the formation of repetitive bedforms associated with cyclic steps in overlying flows under marine conditions (e.g., Fildani et al., 2006; Kostic and Parker, 2006; Cartigny et al., 2011; Kostic, 2011). Unfortunately, there are no criteria to definitively identify bedforms generated from cyclic steps in turbidity currents and to confidently discriminate them from bedforms developed by other mechanisms. Moreover, the terminology is confusing, as the term “cyclic steps,” one with a mechanistic connotation, has been applied to both the bedform morphologies and a turbidity flow mechanism proposed to be responsible for their formation. Here we have tried to separate the description of the observed morphologies (i.e., CSBs) from inferences about the processes that may have created them.

While the cyclic step theory does not restrict the size of the bedforms that could be generated, most of the features that have been attributed to cyclic steps in marine turbidity currents have wavelengths of ∼500 m and typically occur in sets of 3–6. Many of the published examples are similar in size and morphology to the plunge pool–like features seen in lower Soquel Canyon (PP, Fig. 17). The long-term impact of repetitive cyclic steps occurring in the same location during multiple turbidity flows may generate these plunge pool–like features.

Models of Cartigny et al. (2011) and Kostic (2011) illustrate how incremental erosion and deposition stimulated by cyclic steps in turbidity currents can result in the progressive migration of bedforms. The CSBs within the axis of upper Monterey Canyon have been used as examples of bedforms developed by cyclic steps in turbidity currents. Cartigny et al. (2011) compared predictions from numerical models to the data available at the time about CSBs observed in upper Monterey Canyon. The new empirical data presented in Paull et al. (2010) and outlined here extend the range of observations down to 2100 m depth, providing significantly different and more detailed constraints than those that were matched in the existing models.

The possibility that CSBs in the incised channel of Monterey Canyon are generated by cyclic steps in turbidity flows cannot be discarded. However, models need to explain the following new observations. The CSBs are developed within thick beds of very poorly sorted debris flow–like, en masse facies lacking a predictable internal architecture. The occurrence of angular boulders on the canyon floor (Figs. 18A, 18B, 18E, and 18F) and observed movements (∼170 m) of large (∼1360 kg) instrument frames deployed on the canyon floor during discrete events (Paull et al., 2010) seems inconsistent with the formation of CSBs by balanced oscillations in the turbidity flow regime over short distances (∼50 m). Abrupt lateral facies changes occur on the margins of the CSBs to the adjacent benches, imposing serious constraints on the thickness of the high-energy component of these flows (Paull et al., 2010). If the cyclic flow theory were to apply here, regular alternations between high-energy turbulent flows capable of transporting boulders and conditions adequately tranquil to allow deposition of thick blankets of debris flow–like, en masse deposits would have to reoccur over just tens of meters. However, the boulder-sized monuments and instruments move multiple CSB wavelengths during what appear to be individual sediment transport events (Paull et al., 2003, 2010).

The high-resolution multibeam data presented here also significantly extend the known depth range of CSBs in Monterey Canyon out to 2100 m water depth and show that there are hundreds of CSBs in the incised channel of Monterey Canyon that are associated with a consistent 1.6° slope (Fig. 20). The models of Kostic (2011) depend on the occurrence of a change in slope to set up conditions for generating cyclic steps. The modeling efforts should also consider why the average slope of the thalweg remains consistent at 1.6° over hundreds of CSB cycles.

Failures within the Canyon Floor

The possibility that sediment failures occur within the floor of Monterey Canyon was introduced to explain CSBs (Paull et al., 2010). Inherent in this concept is that while there may be density-driven flows on the canyon floor, movements also occur within the canyon fill. Such movements may either be occurring along defined slippage surfaces or be associated with liquefaction (e.g., breaching) of the canyon fill.

Breaching within weakly cohesive sediments is common in river channels and during suction-dredging operations in unconsolidated water-saturated sands (Van den Berg et al., 2002). Breaching is associated with retrogressive slope failures and produces crescent-shaped scarps. Such scarps propagate upslope. Breaching has been inferred to occur within submarine canyons, notably Scripps Canyon, an arm of the La Jolla Canyon system (Mastbergen and Van den Berg, 2003). CSBs are known to occur in La Jolla Canyon (Paull et al., 2008). The thick, poorly sorted en masse facies observed within the canyon floor are consistent with deposition downslope of retrogressive failures associated with breaching (Paull et al., 2005, 2010). This concept also has the appeal of being capable of accounting for the episodic movements of very large objects without the need to invoke very high current velocities that are not measured in Monterey Canyon.

The upslope propagation of these retrogressive failures is not constrained to follow the canyon thalweg. Thus, scarps associated with breaching in the canyon floor can climb up the canyon walls irrespective of the path of individual or repeated turbidity currents. However, failures are more likely to propagate up the channel, where the unlithified coarse sediments are most susceptible to breaching failure (Paull et al., 2005, 2010).

Movement within the unconsolidated sediment filling the canyon axial channel may have an impact somewhat analogous to that of alpine glaciers. Erosion into the canyon host rocks will occur if the failures within the canyon floor periodically contact the underlying country rock (Fig. 3C). The subsurface depth affected by breaching failures with the canyon floor is unknown, and may vary between events. However, the contact between the canyon fill and the underlying country rock may be a favored surface. Movement along this interface would inevitably cause erosion of the underlying strata. Thus, canyon incision may be occurring without having to remove the canyon fill and expose the underlying strata.

Two areas within the channel of Monterey Canyon contain large bedforms with crests showing reversed curvature (i.e., concave-up canyon crests). Both of these areas are upstream of features that may be restricting down-canyon sediment movement. This observation suggests that RCBs may form where sediments build up (RCB; Figs. 10A, 11, and 12).

Some similarity between the RCBs and the surfaces of glaciers and rock glaciers is noted. Both are covered with large angular blocks of material and commonly have large furrows and ridges with a parabolic shape similar to the RCBs (e.g., Swift et al., 2006; Johnson et al., 2007; Harrison et al., 2007; Serrano et al., 2010; Evans, 2010). In the case of glaciers and rock glaciers, the surface material has ridden downslope accommodated by flow within the underlying valley fill, and the ridges develop by compression. If analogous processes occur within the Monterey Canyon floor, it may help explain how large blocks of material can be moved without requiring that turbidity flows with adequate velocity to transport boulders occur regularly within the canyon axis.

AF and CSB Similarity

The appearance of the CSBs that occur throughout most of the sediment-filled sections of Monterey and Carmel Canyons is similar to many of the AFs that occur on the canyon walls, in that they form crescent-shaped concave down-canyon scarps. The AFs occur at a number of scales. Some are large enough to be easily recognized in the previously collected surface ship bathymetry (Supplemental Map 1–4 [see footnote 1]), whereas smaller AFs are resolved only in the AUV surveys (AF; Figs. 4, 5, 13, 14, 15, 16, and 17). Given their occurrence on steep canyon walls, most workers would interpret the AFs as being slide scars produced by submarine mass failures.

The morphology of AFs is difficult to distinguish from the CSBs on the floor of Monterey Canyon (CSB; Figs. 2, 4, 7, and 12) on the basis of bathymetry alone. The similarity suggests that the CSBs may also be slide scars within the incised channel of the canyon. The slightly more distinctive CSB scarps identified here as small knickpoints (Figs. 5A, 6A, 9A, and 19) may be headwall scarps of an individual train of canyon floor failures.

In several places within Monterey Canyon, trains of CSBs extend from the seafloor up the canyon walls (CSB; Figs. 10, 13, and 14). This suggests that trains of CSBs may be continuous with AFs on the canyon walls. Where the CSBs extend from the canyon floor up the canyon walls into the presumably more cohesive fine-grained sediments, the slopes typically increase from an average of 1.6° on the canyon floor to 5°–25° (Greene et al., 2002; Paull et al., 2005) on the canyon walls.

The locations of many of the AFs on the canyon walls are not directly associated with the outside bends of the channel, which is where gravity driven flow passing down through the canyon might have the greatest impact. Instead, many of these AFs are on the upstream side of the bends, where the impact of failures within the canyon floor propagating upslope would be greatest.

Lithologic Control on CSB Occurrence

The distinctly steeper slope of the canyon floor where competent bedrock is exposed (e.g., within the narrow constriction that bypasses the Navy slump or in lower Soquel Canyon) is easily attributed to bedrock control. However, over most of the floors of these canyons the slopes of the axial channel may reflect equilibrium conditions between the processes of erosion and deposition.

The lithologic composition of channel fill may have significant control over the occurrence of the CSBs. Previously published ROV-collected Vibracore data have shown that where CSBs occur in the incised channels of upper Monterey Canyon and Carmel Canyon the primary lithology is unconsolidated and relatively cohesionless coarse-grained sediment. These sediments are interpreted to be gravity flow–event deposits (Paull et al., 2005, 2010). ROV observations show that debris flow deposits are exposed on the canyon floor within the CSB and RCB fields (Figs. 18A, 18B, 18E–18H). Debris flow deposits characteristically lack internal layering, which is shown by chirp subbottom profiles in Monterey Canyon (Figs. 21A–21C). In contrast, the axial channel in Soquel Canyon is covered with internally layered fine-grained sediment (Fig. 21D) and lacks evidence for internal disruptions associated with gravity flow events (Paull et al., 2005). Fine-grained material such as that found in Soquel Canyon is cohesive at shallow depths.

The differences in the average slopes of the sediment filled portions of Monterey and Soquel Canyons may be attributed to contrasting cohesive strength of the sediments. The cohesionless sand and gravel within the incised channel of Monterey Canyon have an increased propensity for failure, resulting in the average slope of ∼1.6°, which is self-maintained. The more cohesive fine-grained sediments in Soquel Canyon stabilize the seafloor and allow substantially higher longitudinal gradients (3°–5°) to occur.

Significance of Benches Bounding the Incised Channel of Monterey Canyon

Soquel Canyon (Fig. 16) lacks the obvious benches that are common along the sides of the incised channel within Monterey Canyon (B; Figs. 2–8). The acoustic character of the chirp subbottom profiles shows nearly horizontal, thin, and discontinuous reflectors from benches within Monterey Canyon and the entire floor of Soquel Canyon (Fig. 21D). This acoustic characteristic suggests that channel fill deposits underlie the benches on Monterey Canyon and the seafloor of Soquel Canyon. While Vibracore data show that the sediment fill in Soquel Canyon is finer grained than the sediments that cover the benches in Monterey Canyon, these sediments appear to be abandoned channel fill deposits and/or thin turbidites. The benches in Monterey Canyon are areas that have not undergone recent sediment failure events and have not been profoundly affected by the high-energy deposits associated with the formation of CSBs within the incised sections of Monterey Canyon (Fig. 3). Although the benches in Monterey Canyon are in immediate proximity to CSBs, they are notably distinct from the en masse movement and deposition of thick poorly sorted sediments that form the CSBs. These benches may be remnants of older canyon fill or accumulations of thin fine-grained turbidite sequences deposited by high-energy turbidity flow events (Fig. 3). The benches show some similarity to models describing the development of inner levee deposits in the Congo (Babonneau et al., 2010).

CONCLUSIONS

The MBARI mapping AUV has provided bathymetry of the axial channels of Soquel Canyon and Monterey Canyon in unprecedented detail down to 2100 m water depth. Textures interpreted from these multibeam data indicate that the drape of sediment on the walls of the canyon thins with increasing distance from the head of the canyon and nearshore sediment sources. The Monterey Canyon floor is covered with sediment fill out to 910 m water depth. Below 910 m water depth intermittent outcrops of resistant strata protrude through the channel fill, and in places form constrictions within the incised channel.

The detailed bathymetric data show trains of CSBs within the incised channel fill in Monterey Canyon down to at least 2100 m water depth and in places on the lower sidewalls of the canyon. Similar CSBs occur on the seafloor of lower Carmel Canyon, but not in Soquel Canyon, which has a smooth seafloor. Both Monterey and Carmel Canyons are currently active conduits for coarse-grained sediment transport, whereas the seafloor of Soquel Canyon comprises fine-grained sediments and does not appear to be an active conduit for sediment transport. The CSBs occur on areas of low slope (∼1.6°) and where the canyon fill consists of cohesionless sand and gravel. Where CSBs are absent, the canyon floors are smooth and filled with horizontally layered sediments. We hypothesize that the nature of the fill (weakly cohesive sand and gravel versus cohesive fine sediments) within these canyons controls where and whether CSBs form.

The walls of Monterey and Soquel Canyons have numerous arcuate features, which by morphological appearance are identified as slide scars. The distinction between the arcuate features on the canyon side and the CSBs on the canyon floor is unclear; they may form a continuum.

While the existing paradigm about the processes that occur within submarine canyons emphasizes the importance of turbidity currents and other flows that pass through the canyon, the new multibeam data evidence can be interpreted to indicate that sediment failures are also occurring within the canyon floor fill. Sediment failure within the sediment fill in the incised channel of Monterey Canyon may be an important process in generating the morphology within this channel and for moving sediments, including boulders, down the canyon. Benches on the margins of the incised channel have been isolated from the sediment failures in the incised channel long enough to accumulate a significant thickness of sediment.

The detailed bathymetric data set presented here is a critical step in the exploration of these submarine canyons and provides fundamental constraints on the physical processes that are important in their evolution. The existence of many of the morphologic features illustrated and discussed in this paper was unknown less than a decade ago, and this paper greatly expands their known extent. However, the multiple hypotheses about the dynamic processes responsible for generating the observed morphologies can be inferred from static images of the seafloor. Distinguishing between the various hypotheses to explain the origins of the observed features, even with the highest resolution bathymetry and subbottom profiler data, is difficult (e.g., Duarte et al., 2010). Facies data help constrain the possibilities, but only present a static view that needs to be interpreted to infer the actual processes of deposition. Modeling of these features is under way, but ultimately can only differentiate what could or could not happen given the assumptions. Unfortunately, few data on the physical conditions during these events are available to constrain models. Ultimately direct measurements of the physical processes that erode and deposit material within submarine canyons within the episodic sediment events need to be made in situ.

The David and Lucile Packard Foundation provided support. Special thanks to the crew of the R/V Zephyr and the Monterey Bay Aquarium Research Institute Autonomous Underwater Vehicles (MBARI AUV) Operations Group. The paper benefited from the thoughtful review comments of Roberto Gwiazda, Roger Urgeles, and an anonymous reviewer. We especially acknowledge the enormous impact the late Bill Normark made on the field of marine geology. Bill was involved in the initiation of the AUV mapping program within Monterey and other submarine canyons.

1Supplemental File. PDF file of five full maps of the autonomous underwater vehicle (AUV)-collected multibeam data from the floor of the Monterey Canyon system. Supplemental Map 1 shows the location of detailed bathymetric maps in Supplemental Maps 2–4 with respect to both location in Monterey Bay and relation to each other. Supplemental Maps 2–4 are progressively colored bathymetric maps showing 2 m grids of AUV–collected bathymetry overlying a 15 m bathymetric grid derived from the regional EM-300 data (Greene et al., 2002) along with 100 m contours. If you are viewing the PDF of this paper or reading it offline, please visit http://dx.doi.org/10.1130/GES00636.S1 or the full-text article on www.gsapubs.org to view the Supplemental File.