Abstract

Distributive submarine fans contain channel-lobe elements that compensationally stack to build a radially dispersive map pattern. The middle parts of some submarine fans contain juxtapositions of channel elements and lobe elements due to longitudinal and lateral shifts in their channel-lobe transition zones. This article uses an exceptionally well-exposed three-dimensional outcrop of the Ross Sandstone at Bridges of Ross (Ireland) to document the stratigraphic and plan-view manifestation of lateral juxtapositions of channel elements and lobe elements in submarine fans. Observations made herein compare favorably to those in seafloor studies of Navy Submarine Fan (offshore southern California, USA) by William Normark and others, indicating that these systems can be used as paired outcrop-seafloor analogs for distributive fans in which the channel-lobe transition zones are located in longitudinally variable positions. In addition, data from Bridges of Ross and Navy Submarine Fan are integrated to constrain a geometric model that predicts the fractional length of a fan that contains lateral juxtapositions of channel elements and lobe elements. Lateral juxtapositions of channel elements and lobe elements are important because they enhance vertical and lateral connectivity within subsurface reservoirs.

INTRODUCTION

Normark (1970) used data from deep-tow high-resolution instruments collected over La Jolla and San Lucas fans in the California Borderland (offshore southern California, USA) to document distributive channel networks on submarine fans. From these observations, he developed the highly cited submarine fan model, which divides fans into three longitudinal regions: (1) the upper fan, a region characterized by a leveed fan valley, (2) the middle fan, a region containing distributary channels, and (3) the lower fan, a region lacking channels and containing predominantly lobes.

Subsequent work by Normark and Piper (1972), Normark et al. (1979), and Piper and Normark (1983) on Navy Submarine Fan (herein Navy Fan), offshore San Diego, California, USA, documented the detailed plan-view shape and shallow stratigraphy of a distributive channel-lobe system and its mesotopographic features (Fig. 1A). These studies documented a distributive network built by compensationally stacked channel-lobe elements. Each channel-lobe element has a channel-lobe transition zone (CLTZ), which is the interface between the channel element and its genetically related lobe element. In Navy Fan, CLTZs are located in variable positions on the fan, resulting in juxtapositions of channel elements and lobe elements in the resulting stratigraphy (Fig. 1A). Ensuing shallow seismic and outcrop studies of submarine fans have documented similar map patterns: for example, the Brazos Trinity Minibasin 4 (Beaubouef et al., 2003), the Zaire Fan (Droz et al., 2003), offshore Indonesia (Posamentier and Kolla, 2003), offshore Corsica (Deptuck et al., 2008), Gulf of Cadiz (Hanquiez et al., 2010), and the Guaso 1 turbidite system in Ainsa, Spain (Gordon et al., 2015). An alternate map pattern exists in some other submarine fans whereby CLTZs of all channel elements are located at nearly the same longitudinal position on the fan. This pattern is interpreted to occur at an abrupt decrease in seafloor gradient, such as the base of slope position, and/or in small, structurally confined fans in which the size of the receiving basin scales to the lengths of lobes (Gordon et al., 2015). A notable example is the Tanqua Karoo Fan (South Africa; Bouma et al., 1995; Prélat et al., 2009). These alternate map patterns represent plan-form (or geomorphic) end members to submarine fan systems (Fig. 1B). Many other detailed seafloor studies have been conducted on distributive submarine fans (e.g., Hueneme and Dume fans, Piper et al., 1999; offshore Indonesia, Saller et al., 2008; Amazon fan, Jegou et al., 2008); however, these studies were conducted at a larger scale and individual channel-lobe elements were not mapped.

Distributive submarine fans are important because they are significant hydrocarbon reservoirs in many parts of the world (e.g., Weimer and Pettingill, 2007). Although seismic data sets are exceptional for characterizing large-scale map patterns in these fans, small-scale characteristics such as lithofacies and stratigraphic architecture described in the end-member models are below the seismic resolution at reservoir depths (e.g., Saller et al., 2008). This is important because the end-member fans have architectural distinctions that have implications for subsurface reservoirs. For example, a key difference between the end members is that in one end member (Fig. 1B, right side), lobe elements compensationally stack both in the longitudinal and lateral directions, leading to lateral and vertical juxtapositions of channel elements and lobe elements in the resulting stratigraphy, which could serve to increase connectivity in subsurface reservoirs. In contrast, the other end member (Fig. 1B, left side) contains lobe elements that compensationally stack mainly in the lateral direction, leading to a lack of juxtapositions of channel elements and lobe elements in the resulting stratigraphy and more compartmentalized subsurface reservoirs.

This article uses a well-exposed three-dimensional (3D) outcrop of the Ross Sandstone at Bridges of Ross (County Clare, Ireland) to document the stratigraphy of a segment of a distributive channel-lobe system, to address four goals. First, this article documents the detailed stratigraphic architecture and paleogeographic evolution of a distributive channel-lobe system in which the CLTZs of successive elements are located in spatially varying lateral and longitudinal positions on the fan (Fig. 1B). Second, this article compares and contrasts observations from the Ross Sandstone outcrop to those of Navy Fan in order to emphasize the importance of paired outcrop-seafloor analogs. Paired outcrop and seafloor analogs are important because they provide complementary perspectives from which to study lithofacies, stratigraphic architecture, and sequential evolution. Third, this article develops a geometric model for distributive fans that contains CLTZs located in longitudinally varying positions. The model can be applied to subsurface data sets where the internal stratigraphic architecture is not resolvable. Fourth, this article discusses implications of juxtapositions of channel elements and lobe elements to connectivity in subsurface reservoirs. The results of this article can be used to reduce uncertainty in the interpretation of subsurface data sets and can help guide production strategies in subsurface reservoirs.

GEOLOGICAL SETTING

The Namurian (Pennsylvanian) Ross Sandstone crops out along sea cliffs of the Loop Head Peninsula and the Ballybunnion area of western Ireland (Fig. 2A). The Ross Sandstone was deposited in the transtensional Carboniferous Shannon Basin (Martinsen et al., 2000, 2003; Pyles, 2008), which is located above the Iapetus suture (Fig. 2A), a major southwest-northeast–trending structural lineament that passes through Ireland, northern England, and southern Scotland (Phillips et al., 1976; Strogen, 1988; Strogen et al., 1996). Rider (1974) defined the Namurian lithostratigraphic units in the basin and interpreted them to record a shallowing-upward succession. The Ross Sandstone contains interbedded turbidite sandstones and mudstones that are interpreted to have been deposited in an ancient submarine fan (Rider, 1974; Collinson et al., 1991). Paleocurrent indicators from all accessible outcrops of the formation form a large-scale, fan-like, radially dispersive pattern with a modal sediment transport direction to the north (Pyles, 2008). Outcrops at Loop Head Peninsula expose the axial or central part of this broad distributive fan (Pyles, 2008).

The Ross Sandstone overlies and laterally interfingers with the Clare Shale, which contains black shale deposited in an anoxic basin-floor to basin-margin setting (Rider, 1974; Collinson et al., 1991; Martinsen et al., 2000; Wignall and Best, 2000; Pyles, 2008). The Ross Sandstone is overlain by the Gull Island Formation, which contains multiple slumps and slides that record the progradation of an unstable slope system into the largely infilled turbidite basin (Rider, 1974; Collinson et al., 1991; Martinsen et al., 2000; Wignall and Best, 2000; Strachan and Pyles, 2014).

The Namurian strata contain several regionally continuous, black, organic-rich, goniatite-bearing shale beds interpreted as condensed sections (Hodson and Lewarne, 1961, and references therein). Each bed contains a unique goniatite assemblage, allowing them to be used as regional correlation markers (e.g., Hodson and Lewarne, 1961; Collinson et al., 1991; Chapin et al., 1994; Wignall and Best, 2000; Martinsen et al., 2003; Pyles, 2008; Pyles et al., 2011). The Ross Sandstone contains seven condensed-section–bounded stratigraphic units (Pyles, 2008).

Several studies document characteristics of the Ross Sandstone. These studies include regional analyses of large-scale stacking patterns (e.g., Martinsen et al., 2000, 2003; Wignall and Best, 2000, 2004; Martinsen and Collinson, 2002; Pyles, 2004, 2008, 2009; Pyles et al., 2011), detailed documentations of stratigraphic architecture (e.g., Chapin et al., 1994; Elliott, 2000a; Sullivan et al., 2000; Strachan, 2002a, 2002b; Pyles and Strachan, 2014), and stratigraphic and process-response analyses focusing on megaflutes (Elliott, 2000b; Macdonald et al., 2011), Channel Element evolution (Lien et al., 2003), architectural elements (Pyles, 2007), cogenetic debrite-turbidite beds (Pyles and Jennette, 2009), and compensational stacking (Straub and Pyles, 2012).

This study focuses on sea-cliff exposures of the Ross Sandstone at Bridges of Ross, which is located on the north coast of Loop Head Peninsula (Fig. 2), near the hamlet of Ross. At this locality, the sea cliff is notably rugose, although roughly aligned with the local paleocurrent direction, providing a local 3D exposure of the elements in the cliff face (Fig. 2). The exposure is famous for a natural arch (or bridge) that is exposed at the western part of the outcrop. The locality is named “Bridges” of Ross because there used to be two bridges; one collapsed in a storm about 100 years ago. The arch is bisected by the axial trace of an east-west–trending anticline (Fig. 2B; Strachan, 2002a; Pyles, 2008). Previous work on this exposure focused on strata that are from the youngest condensed-section–bounded stratigraphic cycle in the Ross Sandstone (R1a2–R1a5; Gill, 1979) that crop out on the north limb of the anticline (Fig. 2B). Earlier studies interpreted deformed strata in the lower part of this succession as a slide (e.g., Gill, 1979), but Wignall and Best (2000) and Strachan (2002a, 2002b) documented internal deformational features to be more consistent with an origin as a slump and renamed these deformed beds the Ross Slump, the terminology used herein. Gill (1979) interpreted the erosional surfaces above the Ross Slump as slip scars related to slumping. In Lien et al. (2003), Pringle et al. (2003), Pyles (2008), and Pyles and Strachan (2015), the erosional surfaces were interpreted as the lower bounding surfaces of submarine channel elements.

DATA AND METHODOLOGY

Data collected to address the goals of the study area are: (1) an xyz lidar (light detection and ranging) data set collected at ∼2 cm resolution that documents the surface topography of the studied interval (Fig. 3); (2) 170 m of stratigraphic columns that document grain size and physical sedimentary structures at centimeter-scale resolution (Fig. 4); (3) high-resolution photopanels; (4) 393 paleocurrent measurements from flutes, grooves, ripples, dunes, and channel-margin orientations; and (5) 182 paleotransport directions for slumps collected from basal striations, fold axes, and fold hinges. These data were used to document the locations of key stratal bounding surfaces, lithofacies patterns, and to constrain correlations and map patterns of the architectural elements in the study area.

This study utilizes the architectural element approach as described in Pyles (2007), in which four architectural elements are interpreted on the basis of their cross-sectional shape when viewed normal to paleocurrent directions: (1) channel elements, (2) lobe elements, (3) slump elements, and (4) mudstone sheet elements. A common, four-tiered hierarchical approach to characterizing channels and lobes was proposed, i.e., bed, story, element, and complex (Pyles, 2007; Straub and Pyles, 2012). These hierarchical terms are modified by the type of element, which is expressed as an adjective (e.g., lobe bed, lobe story, lobe element, lobe complex). A translation between this hierarchical designation and those from other studies was included in Straub and Pyles (2012). In this hierarchical approach, the boundaries between stratigraphically adjacent elements record abrupt and relatively large-scale changes in (1) the location of the axes of the elements, (2) lithofacies patterns, (3) bedding style, and (4) paleocurrent directions. This study is focused on the element-scale architecture of one part of a channel-lobe complex in the Ross Sandstone at Bridges of Ross.

A west-east–oriented cross section (Fig. 4) was constructed using stratigraphic columns and element boundaries as documented in lidar data and photopanels. The cross section is referred to herein to help depict stratal characteristics and architectural associations.

RESULTS

Introduction

The Bridges of Ross outcrop extends ∼2.0 km in the east-west direction and ∼200 m in the north-south direction (Fig. 2B). Beds in the studied interval dip predominantly northward, and are located on the north limb of an east-west–trending anticline that is exposed across the study area. A small syncline locally deforms strata on the northeastern part of the exposure (Fig. 2B). The modal paleocurrent directions collected from megaflutes, channel margins, cross-beds, ripples, and dunes is to the northeast (060°) with a range from north to southeast (Fig. 2B). This trend deviates from the regional trend in the formation, which is predominantly north (Fig. 2A).

For convenience, the study area is subdivided into four sites corresponding to the four promontories or peninsulas (Peninsulas 1–4), which are ordered in an eastward and approximately down-paleocurrent direction (Fig. 2B). Peninsulas 1 and 3 are locally referred to as Bridges of Ross and Fisherman’s Point, respectively. Bridges of Ross is unique to other exposures in the Ross Sandstone in that the time-equivalent stratigraphy can be correlated across multiple cliff faces for >2 km in the down-current direction. Using crosscutting relationships and superposition, and the recognition criterion described in the Data and Methodology discussion herein, the studied interval is divided into seven discrete architectural elements that record the sequential development of this part of the channel-lobe complex that is a small component of the overall submarine fan (Fig. 4): (1) Ross Slump element, (2) Lobe Element 1, (3) Lobe Element 2, (4) Lobe Element 3, (5) Channel Element 0, (6) Channel Element 1, (7) and Channel Element 2. Each is described in the following text.

Ross Slump

The lowest element in the succession is a mudstone-rich, internally contorted unit that ranges from 3.5 m thick at Peninsula 1 to 6.4 m thick at Peninsula 4 (Fig. 4). The unit is referred to as the Ross Slump (Strachan, 2002a, 2002b) and is exposed and accessible at all four peninsulas (Figs. 4–7). The basal bounding surface of the Ross Slump is a décollement horizon coinciding with the R1a2 condensed section, a mudstone sheet element (Figs. 4–8; Gill, 1979) interpreted to have acted as the dominant shear plane along which most of the strain took place (Fig. 8). Although mainly planar, the basal décollement locally ramps down and up section, where it incorporates underlying beds (Bakken, 1987; Martinsen and Bakken, 1990) or deflects upward over dome and ridge deformational sandstone bodies (Figs. 4–8; Strachan, 2002a, 2002b). The upper bounding surface is sharp, planar, and locally erosional, and is overlain by undeformed tabular sandstone units (Fig. 4) (discussed below). The upper bounding surface of the Ross Slump is well exposed at all four peninsulas but is especially well exposed at the western parts of Peninsulas 2 and 3 (Figs. 5B and 6), where the surface is locally ornamented by sandstone volcanoes (Fig. 8B). Sandstone volcanoes on the upper surface of the Ross Slump are 0.05–3 m in diameter and 0.01–0.4 m in height. The flanks of the sandstone volcanoes are straight to slightly concave upward and commonly contain longitudinal ridges interpreted to result from sand flows (Gill and Kuenen, 1958).

The Ross Slump contains deformation structures including contorted bedding, folds, faults, shear zones, deformed sandstone bodies, sandstone dikes, faults, and highly deformed and sheared mudstone (Figs. 8C–8F). The deformed sandstones, sandstone dikes, and sandstone volcanoes are interpreted to result from fluidization and liquefaction associated with shearing and loading while the slump was emplaced (Strachan, 2002a, 2002b). The best exposed deformed textures in the slump are on the eastern side of Peninsula 2, where sandstone dikes and deformed sandstones are abundant (Fig. 8E). At Peninsula 4, the Ross Sandstone is locally deformed into thrust duplexes wherein the R1a2 condensed section is repeated several times (Fig. 8F).

Paleotransport directions of the slump are based on measurements from lineations on the lower bounding surface (décollement) and orientations of fold axes and hinges. Lineations, which are interpreted to reflect local incision of the slump into underlying strata, are exposed on the northeast end of Peninsula 2 and are oriented east-northeast to west-southwest (Fig. 4). This trend is consistent with other deformation structures such as syndepositional fold axes and fold hinges that collectively record movement toward the east-northeast (Fig. 4). There are some deviations from this general transport trend, perhaps caused by internal shearing and differential flow as the unit interacted with local topography and expanded laterally.

Lobe Elements 1, 2, and 3

The Ross Slump is overlain by tabular strata that are divided into three lobe elements (Figs. 4, 5A, and 9A). The three units are interpreted as lobe elements because: (1) the amounts of erosion on their lower bounding surfaces are less than the thicknesses of the elements; (2) erosion on the base of the elements is limited to a narrow part of the axial regions of the elements, whereas the bases of the elements are conformable elsewhere; (3) their cross-sectional shapes are nearly tabular with flat to weakly erosional lower bounding surfaces with a planar to convex-upward upper bounding surface; (4) their aspect ratios are ∼500:1–1500:1, consistent with other published examples of lobe elements in the Ross Sandstone (average aspect ratio, width, and thicknesses for lobe elements in the Ross Sandstone were reported as ∼750, 1500 m, and 2 m, respectively; Pyles, 2007) and elsewhere (Deptuck et al., 2008; Prélat et al., 2009); and (5) these elements are similar in terms of lithofacies distributions, nature of bounding surfaces, dimensions, and architecture to lobe elements documented in more laterally continuous exposures at Kilcloher Cliff, Horse Island, and Kilbaha Bay (Pyles, 2007, 2008; Straub and Pyles, 2012). The boundaries between the elements weather in outcrop as bedding-plane exposures (Figs. 5A and 9), across which there are shifts in the locations of the axes (or thickest part) of the elements and a shift in the paleocurrent directions (Fig. 4). Following the work in Straub and Pyles (2012), the spatial shifts in the locations of the axes are interpreted as the stratigraphic manifestation of compensational stacking.

Lobe Element 1 is exposed and is accessible at all four peninsulas. Paleocurrents measured from megaflutes, flutes, and ripples in Lobe Element 1 are collectively to the northeast, although there is a systematic lateral change from north-northeast at Peninsula 1 to east-northeast at Peninsula 4. The axis or thickest part of Lobe Element 1 is located near the westernmost part of Peninsula 1 (Figs. 4, 5A, and 9A), where it is 3.2 m thick. In its axis, the lobe element contains two units: (1) a lower unit composed of beds that thicken and coarsen from one to the next in an upward transect; and (2) an upper unit that contains amalgamated, structureless sandstone beds (Figs. 4 and 9A). At the axis, the lower bounding surface of Lobe Element 1 is locally ornamented by megaflutes (sensu Chapin et al., 1994; Elliott, 2000b) that erode <0.5 m in to the subjacent element (Figs. 4 and 9). Erosion is limited to an ∼50-m-wide zone (∼5% of the width of the exposed element) corresponding to the axis. The lower bounding surface is conformable and planar elsewhere. The lobe element and its internal sandstone beds systematically become thinner and finer grained eastward (along strike); at the east side of Peninsula 1, Lobe Element 1 is 2.5 m thick (Fig. 9B); at Peninsula 2, it is 0.85 m thick (Fig. 9C); at Peninsula 3, it is 0.5 m thick (Fig. 9D); and at Peninsula 4, it is only 0.3 m thick (Fig. 9E). At each of these localities the lobe element contains interbedded sandstone and mudstone beds (predominantly Bouma b-c-d divisions) that thicken and coarsen in an upward transect, although the upper, amalgamated part, which is characteristic of the axis, is absent at Peninsulas 2, 3, and 4 due to lateral facies changes into thin-bedded strata. These lateral changes in bed thickness, grain size, and lithofacies reflect axis-to-margin trends. Although the entire element is not exposed in the outcrop, this article uses thinning rates and a presumption of cross-sectional symmetry to estimate Lobe 1’s width and aspect ratio as 2800 m and ∼900, respectively.

Lobe Element 2 is exposed and is accessible at all four peninsulas. Paleocurrents measured from megaflutes, flutes, ripples, and dunes in Lobe Element 2 are collectively to the northeast (Fig. 4). The axis of Lobe Element 2 is located on the east side of Peninsula 1 (∼500 m east of the axis of Lobe Element 1), where it is 2.75 m thick (Figs. 4, 5A, 9A, and 9B). The lobe element systematically thins westward from its axis to 1 m (along strike) where it overlies the axis of Lobe Element 1 (Figs. 4A and 9A), and thins progressively eastward from its axis (along strike) to Peninsulas 2, 3, and 4, where it is 0.95 m, 0.7 m, and 0.6 m, respectively (Figs. 9B–9D). In its axis, the lower bounding surface of Lobe Element 2 is mantled by megaflutes (Figs. 4 and 5A). The megaflutes are up to 1 m (∼30% of the thickness of the element) and as wide as 10 m. Erosion is limited to a 200-m-wide zone (∼15% of the width of the exposed part of the element) corresponding to the axis of the element. The lower bounding surface is planar and conformable elsewhere. Locally at Peninsulas 2, 3, and 4 Lobe Element 2 is eroded by Channel Elements 1 and 2 (Figs. 4–7 and 9). Lobe Element 2 has the same axis-to-margin trends in thickness, bedding, and lithofacies as Lobe Element 1; however, megaflutes are more abundant in Lobe Element 2, possibly indicating that the outcrop exposes a more proximal and more erosive part of the Lobe Element. Although the entire element is not exposed in the outcrop, this article uses thinning rates and a presumption of cross-sectional symmetry to estimate Lobe 2’s width and aspect ratio as 1800 m and ∼600, respectively.

The distribution of outcrops of Lobe Element 3 is strongly controlled by erosion from the overlying Channel Elements. Lobe Element 3 is only exposed on the west side of Peninsula 1, where it is 3.5 m thick and sediment transport directions are to the north (Figs. 4 and 5A). At all other localities, Lobe Element 3 is erosionally removed by Channel Elements 0, 1, and 2.

Channel Elements 0, 1, and 2

The lobe elements are overlain by units interpreted as channel elements (Figs. 4–7) because (1) the amounts of erosion equal the fill thicknesses of the elements; (2) in cross section, the elements contain concave-upward lower bounding surfaces and planar upper bounding surfaces except were they are eroded by younger elements; (3) the aspect ratios are ∼15:1, similar to other deep-water channel elements in the Ross Sandstone (average aspect ratio, width, and thicknesses for lobe elements in the Ross Sandstone are ∼40, 170 m, and 4 m, respectively; Pyles, 2007) and elsewhere such as in the Brushy Canyon (e.g., Gardner et al., 2003; Pyles et al., 2010).

Channel Element 0 is locally exposed at Peninsulas 3 and 4 (Figs. 4, 6, and 7), but is inaccessible at these locations. However, the orientation of the channel element is locally to the east; it can be discerned because its southern margin is exposed on four cliff faces at Peninsulas 3 and 4.

Channel Element 1 is well exposed at all four peninsulas, although it is only accessible at Peninsulas 1 and 2 and the west side of Peninsula 3 (Figs. 4–7). The average paleocurrent is to the east and is approximately parallel to the overall trend of the coastline. The best exposure of Channel Element 1 is at Peninsula 1, where the entire cross section of the channel element is exposed (Figs. 5 and 10). The aspect ratio, width, and thickness of the channel element are 20, 85 m, and 4.25 m, respectively. At Peninsula 1, flutes on the basal bounding surfaces of sandstone beds within the channel element and foreset orientations from cross-beds, together with channel-margin orientations, indicate that the channel element is oriented toward the southeast (Figs. 4 and 10). The western margin of Channel Element 1 is relatively steep (>30°), whereas the eastern margin is relatively shallow and has a step-like shape (Fig. 10A). Strata in the eastern side of the channel element are sigmoid-shaped, laterally dipping structureless sandstone beds that are encased in shale-clast conglomerate and are interpreted as laterally accreting packages (LAPs, sensu Abreu et al., 2003) that record the lateral migration of the channel element (Fig. 10). These intraelement units are restricted to one margin of the channel element. Inclined, thin-bedded strata that drape the margins of the channels are distinctive from LAPs because thin-bedded drapes laterally thicken toward the axis of the channel except in instances where they are eroded, and they are not sigmoid shaped in cross-sectional view. Beds in the western part of the channel element is amalgamated and contains large-scale, cross-stratified and structureless sandstone beds that onlap the LAPs (Fig. 4). Channel Element 1 is highly asymmetrical both in terms of cross-sectional shape and lithofacies distributions (sensu Pyles et al., 2010).

Channel Element 2 is well exposed and accessible at all four peninsulas (Figs. 5–7). The average paleocurrent is to the east-northeast and is approximately parallel to the overall trend of the coastline. At Peninsula 1, the southern margin and axis of Channel Element 2 is well exposed (Fig. 5A). At this location, Channel Element 2 partly erodes into Channel Element 1. The southern margin of the channel element is overlain by sigmoid-shaped, north-dipping structureless and large-scale cross-stratified sandstone beds that are encased in shale-conglomerate interpreted as LAPs (Fig. 5A). Bedding-plane exposures of the sandstone units within the LAPs are ornamented by dunes, from which an eastward paleocurrent direction is evident. This orientation is aligned with the channel-margin orientations. At Peninsula 2 (slightly down-current direction), the northern margin and axis of Channel Element 2 are well exposed. At this location, Channel Element 2 erodes through the entire southern half of Channel Element 1, and Channel Elements 1 and 2 are laterally offset from each other. The northern margin of Channel Element 2 is exposed at a spring low tide as a bedding-plane exposure from which the orientation of Channel Element 2 can be documented to trend to the northeast (Fig. 5B). This margin is abruptly overlain by south-dipping structureless to large-scale cross-stratified sandstone units encased in shale-clast conglomerate interpreted as LAPs (Fig. 5B). These LAPs are located on the opposite margin of the channel element and dip in the opposite direction than those in the same element at Peninsula 1. At Peninsula 3 (slightly down-current direction), the northern margin and axis of Channel Element 2 are well exposed (Fig. 6). The same stratigraphy is exposed on the eastern and western side of Peninsula 3, although the western side of the peninsula is more accessible. At Peninsula 3, Channel Element 2 erodes through the southern half of Channel Element 1, and the axes of Channel Elements 0, 1, and 2 are laterally offset from each other. The northern margin of Channel Element 2 is overlain by a shale drape that was either not present or not exposed at Peninsulas 1 and 2. Strata overlying the shale drape are laterally inclined structureless to large-scale, cross-stratified sandstone beds encased in shale-clast conglomerate interpreted as LAPs (Fig. 6). The paleocurrent directions measured from flutes, cross-beds, and channel-margin orientations are eastward. At Peninsula 4 (slightly down-current direction), the northern margin of the channel element is exposed on the cliff on the northern side of the peninsula (Figs. 7A and 11A) and is overlain by inclined sandstone beds encased in shale-clast conglomerate that are interpreted as LAPs (Figs. 7, 11A, and 11B). The axis of the channel element is accessible where some of the best examples of cross-stratified sandstone in the Ross Sandstone are exposed (Fig. 11D). The upper surface of the channel element is mantled with megaflutes. The northern margin of Channel Element 2 is covered by glacial till at this location, although the margin is exposed on the east side of Peninsula 4 (Fig. 7B), allowing the aspect ratio, width, and thickness of the channel element to be measured as 12.8, 160 m, and 12.5 m, respectively. At Peninsula 4, Channel Element 2 erodes through the southern half of Channel Element 1, and the axes of Channel Elements 0, 1, and 2 are laterally offset from each other.

Synthesis

In summary, the upward succession at Bridges of Ross contains a slump (Ross Slump) overlain by compensationally stacked Lobe Elements, which are in turn overlain by Channel Elements. The axes of Lobe Elements 1, 2, and 3 are laterally offset from each other, reflecting compensational stacking. However, at the scale of the study area, Lobe Elements 1, 2, and 3 collectively become thinner to the east due to depositional thinning (Figs. 4–7 and 9). The three lobe elements are thickest where the underlying Ross Slump is thinnest, and they are thinnest at Peninsula 4, where the underlying Ross Slump is thickest (Fig. 4), possibly indicating that slump-induced topography partly controlled the locus of deposition of overlying lobe elements. This surface topography appears to have been healed by the time Lobe Element 3 was deposited (Fig. 4), as this surface is relatively parallel to the R1a2 surface, a reasonable datum for the cross section. Channel Elements 1, 2, and 3 collectively converge with the Ross Slump in the eastward direction. This convergence is due to the combined effect of depositional thinning of the underlying lobe elements and erosion at the bases of the channel elements. Erosion at the bases of the channel elements serves to laterally juxtapose Lobe Elements 2 and 3 with Channel Elements 0, 1, and 2.

PALEOGEOGRAPHY

Paleogeographic interpretations of the Ross Slump, Lobe Elements 1 and 2, and Channel Elements 1 and 2 are presented in Figure 12. The maps represent our interpretations that unify the currently available data, including mapped contacts on lidar data and photopanels (e.g., Figs. 5–7), which are documented as bold blue lines on the paleogeographic maps; paleocurrent and paleotransport data (e.g., Figs. 2 and 4); measured thicknesses (Figs. 4–7); and stratigraphic columns (Fig. 4).

The paleogeographic map of the Ross Slump (Fig. 12A) depicts the overall observed thickening from west to east, although local variations are evident. The measured transport indicators document northeastward movement of the slump. Local variations in thickness are due to variability on the lower bounding surface of the slump (Fig. 4) where it locally erodes and entrains underlying beds or is deflected upward over dome and ridge deformational sandstone bodies (Fig. 8C).

The paleogeographic map of Lobe Element 1 (Fig. 12B) depicts the axis of the lobe element to be located near the western end of study area where it is 3.15 m thick and is sand rich (Figs. 4, 5, and 9). The lobe element thins to the east-northeast to 0.20 m where it is sand poor (Figs. 4 and 5). The axis of the lobe element is underlain by megaflutes (depicted as white ovals on the map) that are restricted to an ∼50-m-wide part of the axial area of the element. The extrapolations of isopach contours away from the outcrop belt are based on a combination of thicknesses, paleocurrents, and the plan-view shapes of lobe elements from Navy Fan (Normark et al., 1979) and Brazos Trinity minbasin 5 (Beaubouef et al., 2003), which are exceptionally well-documented lobe elements in near seafloor studies (Fig. 1A). Facies patterns and morphometric characteristics including the thickness, estimated width, and aspect ratio are summarized in Figure 12B.

The paleogeographic map of Lobe Element 2 (Fig. 12C) depicts the axial region of the lobe element to be located near the eastern end of Peninsula 1, where it is 2.95 m thick and sand rich (Figs. 4, 5, and 9) and paleocurrent indicators are to the northeast. The axis of Lobe Element 2 is offset by ∼500 m (∼20% of the width of the elements) from the axis of Lobe Element 1, indicating compensational stacking. The lobe element symmetrically thins to the east and west. The axis of the lobe element contains an erosional base composed of closely spaced megaflutes, depicted as white ovals (outlined with red), that are restricted to a 200-m-wide part of the axis of the element (Fig. 4). The abundance of megaflutes and amount of erosion are interpreted to reflect that the outcrop is possibly located near the location of the CLTZ. Facies patterns and morphometric characteristics, including the thickness, estimated width, and aspect ratio, are summarized in Figure 12C.

The paleogeographic map of Channel Element 1 (Fig. 12D) depicts the plan-view shape of the channel element, which is constrained from the locations of the channel element margin and thalweg documented from outcrops (Figs. 5–7 and 10), channel-margin orientations (rose diagrams circled in red; Fig. 12), and paleocurrents (rose diagrams circled in blue; Fig. 12) from all four peninsulas. The width of the channel element (85 m) is constrained from Peninsula 1, where both margins are well exposed (Fig. 10). LAPS (Figs. 5 and 10) only occur on the inside of channel-element bends, similar to those observed at the Beacon Channel outcrop (Pyles et al., 2010, 2012). Facies patterns and morphometric characteristics including the thickness, width, aspect ratio, and sinuosity are summarized in Figure 12D.

The plan-view shape of Channel Element 2 (Fig. 12E) is constrained from the locations of the channel element’s margin and thalweg (Figs. 5–7 and 11), channel-margin orientations (rose diagrams circled in red; Fig. 12) and paleocurrents (rose diagrams circled in blue; Fig. 12). The width of the channel element (160 m) is constrained from Peninsula 4, where both margins are well exposed. The locations of LAPS are based on those in outcrop (Figs. 5–7 and 11) and similar to Channel Element 1 are associated with the inside of channel-element bends. Facies patterns and morphometric characteristics, including the thickness, width, aspect ratio, and sinuosity, are summarized in Figure 12D. Channel Element 2 is larger and less sinuous, at least in the study area, than Channel Element 1.

The sequential evolution of the documented part of the distributive fan is depicted in Figure 12F, where the five elements are superimposed on a single map. The map shows compensational stacking of Lobe Elements 1 and 2. Lobe Element 3 is not sufficiently exposed to document its paleogeography; however, the three lobe elements are overlain by channel elements. The lateral juxtapositions of channel elements and lobe elements documented in Figures 4–12 are expressed on the map as a long, narrow zone whereby Channel Elements 1 and 2 locally incise Lobe Elements 2 and 3. Only the CLTZ of Lobe Element 2 is interpreted from the outcrop observations and is located at the western side of Peninsula 1. However, the CLTZ of the other units can be inferred: the CLTZ for Lobe Element 1 is located southwest of Peninsula 1 and the CLTZs for Channel Elements 1 and 2 are located to northeast of the study area. Collectively, these four elements document that CLTZs are longitudinally offset by at least 2 km in this part of the fan (but <15 km as this value represents the most distal extent of the fan; Pyles, 2008), resulting in lateral juxtapositions of channel elements and lobe elements.

This upward and temporal succession is interpreted to record progradation of a distributive channel-lobe complex. The processes forcing this apparent progradation could be autogenic, perhaps due to morphodynamics (the interplay between flow processes and sedimentation), or allogenic, perhaps due to increased sediment supply induced from climate change, increased tectonic activity, or lowered relative sea level. Resolving which of these, or other, processes are responsible for the pattern is not the focus of this paper.

COMPARISON OF BRIDGES OF ROSS TO OTHER EXPOSURES IN THE FORMATION

The architectural associations at Bridges of Ross compare favorably to some other outcrops of the formation. For example, outcrops of the middle Ross Sandstone at Kilbaha Bay and Horse Island, and the upper Ross Sandstone at Cloonconeen, all contain similar juxtapositions of channel elements and lobe elements (Pyles, 2008; Figs. 2 and 13). Furthermore, the sizes (width, thickness, aspect ratio) of lobe elements and channel elements statistically similar at the 90% confidence interval (Pyles, 2004, 2008; Pyles and Strachan, 2014) and their internal facies distributions are essentially the same. Therefore, the architectural style exposed at Bridges of Ross is interpreted as representative of other parts of the formation. This is important because Pyles (2008) and Pyles et al. (2011) depicted little vertical change in the stratigraphic architecture through the >500-m-thick formation with the exception of an increase in the amount of slumps.

COMPARISON OF BRIDGES OF ROSS TO NAVY FAN

Similarities between Bridges of Ross and Navy Fan may justify the use of these systems as paired outcrop-seafloor analogs for distributive submarine fans. First is the spatial variation in the positions of CLTZs (right side, Fig. 1B). At Bridges of Ross, channel elements and lobe elements are vertically and laterally juxtaposed (Figs. 4–12) due to kilometer-scale longitudinal shifts in the location of the CLTZ. This association of elements is similar to the Normark et al. (1979) documentation of Navy Fan (Fig. 1A). Second, the sizes of channel elements and lobe elements at Bridges of Ross are comparable to those documented by Normark et al. (1979). Channel elements are between 170 and 190 m wide in Navy Fan and between 85 and 160 m wide at Bridges of Ross, and 5–15 m deep in Navy Fan and 4.25–15.5 m thick (deep) at Bridges of Ross. Aspect ratios for both systems vary from 13 to 20. Furthermore, lobe elements are 1100–3050 m wide in Navy Fan and 1800–2800 m wide (estimated width based on the assumption of symmetry) at Bridges of Ross. Lobe-element thicknesses were not reported for Navy Fan, but they average 3 m thick at Bridges of Ross with aspect ratios of ∼600–1000. Third, the ratio of lobe-element width to channel-element width is ∼15:1 (±7) for Bridges of Ross and Navy Fan. Normark et al. (1979) documented crescentic depressions, later termed megaflutes (sensu Chapin et al., 1994), commonly occurring in groups near the CLTZ at Navy Fan. These structures are commonly 5–30 m wide and 1–2 m deep. The Ross Sandstone is known for having megaflutes (e.g., Chapin et al., 1994; Elliott, 2000b). Megaflutes are 3–15 m wide and 0.5–1 m deep are also present at Bridges of Ross, where they occur as a cluster beneath the axis of deposition of Lobe Element 2 (Figs. 4 and 5A). Observations in Navy Fan provides context that aids in the interpretation of outcrops of submarine fans, whereby clustering of megaflutes underlying the axes of lobe elements could be a clue that the exposure is just basinward of the CLTZ.

Whereas numerous similarities exist between Navy Fan and Bridges of Ross, it is important to note that the seafloor and outcrop perspectives, respectively, offer different but complementary observations on distributive fans. First, the Normark et al. (1979) study of Navy Fan documented the entire channel-lobe complex because it is fully exposed on the seafloor, whereas Bridges of Ross exposes only a fraction of a distributive channel-lobe complex due to the manner in which this ancient stratigraphy was exposed. Figure 1A illustrates the size and representative stratigraphy of Bridges of Ross in comparison to Navy Fan. Navy Fan therefore provides a larger scale map view than Bridges of Ross. At Bridges of Ross, the lateral juxtapositions of channel elements with lobe elements and channel elements with channel elements is documented (Figs. 4–7), whereas these geometries and facies juxtapositions could not be imaged in Navy Fan due to data limitations. Bridges of Ross therefore provides a more complete view of detailed stratigraphic relationships. Whereas piston cores and seafloor samples locally document the lithology of Navy Fan (Normark and Piper, 1972; Normark et al., 1979), Bridges of Ross contains superb exposures from which lithofacies associations and bedding orientations can be documented (Figs. 4–11). For example, cross sections of the channel elements and lobe elements are well exposed, allowing one to document the lithofacies relationships and patterns that are unique to each. Bridges of Ross provides an ∼20-m-deep (thick) view into the sedimentation of a distributive fan (Figs. 4–11), whereas due to limitations in imaging and sampling, Navy Fan is only documented for the upper ∼10 m of strata.

GEOMETRIC MODEL FOR SUBMARINE FANS WITH CLTZs LOCATED IN LONGITUDINALLY VARIABLE POSITIONS

Bathymetric data and shallow seismic data sets can, in many near seafloor and shallow seismic examples, image the channel elements and lobe elements in distributive submarine fans (e.g., Normark et al., 1979; Beaubouef et al., 2003; Pirmez et al., 2000; Posamentier and Kolla, 2003; Droz et al., 2003; Hanquiez et al., 2010). However, at reservoir depths, seismic data are often inadequate for imaging individual lobe elements, although the size of the distributive fans and their feeder channels can often be resolved (e.g., Saller et al., 2008). This article presents a simple geometric model that predicts the fractional extent of a distributive fan that contains lateral juxtapositions of channel elements and lobe elements in its stratigraphy due to longitudinal variations in the positions of CLTZs. The model is intended to be used in subsurface data sets where individual channel-lobe elements are below the resolution of the data sets. The model is based on simple geometry and empirically defined patterns from Bridges of Ross and Navy Fan and input variables that some seismic data sets can resolve at reservoir depths.

The model compares the length of the fan that contains juxtapositions of channel elements and lobe elements (lcltz) with the total length of the fan (lf), and is expressed as 
graphic
where ll is the length of individual lobe elements (see Fig. 1B for a graphical depiction of these and other variables in the model);forumla is dimensionless, and ranges from 0 to 1. Whereas lf can often be measured from subsurface data sets, lcltz and ll are inaccessible measurements at reservoir depths due to issues with seismic resolution (Saller et al., 2008). As such, we algebraically derive alternate methods for estimating forumla. Input variables used for this model are channel-element width (wc) and fan length (lf), which can usually be determined from seismic images (e.g., Abreu et al., 2003; Saller et al., 2008). In order to estimate the length of the fan that contains lateral juxtapositions of channel elements and lobe elements (lcltz), the width (wl) and plan-view aspect ratios of lobe elements (Al) need to be defined. We do so using empirically derived relationships from Bridges of Ross and Navy Fan, where 
graphic
[± is reported as standard error; σ/(n)0.5], and wc is 127 (±8) m. Therefore, wl = 1650 (±360) m. The lengths of lobe elements (ll) are empirically defined using the plan view aspect ratio of the lobe element (Al), which is defined as 
graphic
where Al is 1.2 (±0.15) on Navy Fan. Therefore, if the length of a fan can be estimated from seismic data (such as an amplitude image), and the width of one or more of its distributive channel elements (wc) can be measured, the fractional length of the fan that contains juxtapositions of channel elements and lobe elements is 
graphic
where wl is defined in Equation 2. Taking the limits of Equation 4 allows one to predict juxtapositions of channel elements and lobe elements to be absent in small fans where the lengths of lobe elements scale to the length of the fan, which is expressed as 
graphic
and that juxtapositions of channel elements and lobe elements are pervasive in fans that are exceptionally larger than their constituent lobe elements, which is expressed as 
graphic

These geometric limits are geologically reasonable. Because forumla is dimensionless, Equations 1 and 4 can be applied to relatively large and small fans alike, regardless of the sizes of lobe elements.

The efficacy of the Equation 4 is tested by using seafloor and outcrop interpretations from published seismic studies whereby element-scale characteristics of the entire fans are mapped. The studies we identified to fit this criterion are listed in Table 1, in which predicted values of forumla using Equation 4 are compared to forumla values measured directly from the data sets. Measured values and predicted values compare favorably for all of the examples (the difference between measured and predicted values is 0.01–0.15) except for those of the Gulf of Cadiz, which is underpredicted by 0.26. In this example, the aspect ratio of the lobes (Al, Equation 3) is significantly higher than 1.2, which was the empirically derived value from Navy Fan. However, Hanquiez et al. (2010) reported that the plan-view shapes of the lobe elements in the Gulf of Cadiz were modified by contour currents. Regardless, it is critical to consider that Al might need to be refined for some fans. For example, a small proportion of lobe elements measured by Prélat et al. (2010) have reported length to width ratios of as much as 10. These high Al lobes are located in unconfined settings. For clarity, forumla is not estimated for Bridges of Ross as only a fractional part of the distributive channel-lobe complex is exposed, therefore lf cannot be resolved.

APPLICATIONS

This article focuses on the cross-sectional and plan-view expression of lateral juxtapositions of channel elements and lobe elements in distributive channel-lobe complexes due to longitudinal shifts in the CLTZ as the systems evolved. The plan-view manifestations of lateral juxtapositions of channel elements and lobe elements in Navy Fan (from Normark et al., 1979) and Bridges of Ross are illustrated in Figures 1A and 12F, respectively. These maps document compensational stacking to operate both laterally and longitudinally in submarine fans. This style of stacking is also documented, although not emphasized, in recent near-seafloor studies of the offshore Nigeria (Pirmez et al., 2000), Brazos-Trinity minibasin 4 (Beaubouef et al., 2003), offshore Indonesia (Posamentier and Kolla, 2003), and in outcrop studies of the Ross Sandstone (Pyles, 2008; Straub and Pyles, 2012; e.g., Fig. 13) and the Guaso 1 turbidite system (Gordon et al., 2015). Therefore, although this stacking pattern represents an end member for submarine fans (Fig. 1B), it is notably common.

Submarine fans in which CLTZs are located in variable longitudinal and lateral positions are important to consider in the development of subsurface reservoirs in similar types of systems for the following reasons.

First, lateral juxtapositions of channel elements and lobe elements in stratigraphy due to spatial variations in the location of the CLTZs provide a process for connecting sandstones in channel elements to those in lobe elements, resulting in a high proportion of sand-on-sand contacts between adjacent elements and high static connectivity, and therefore an opportunity for fluid and pressure communication between elements in subsurface reservoirs (Funk et al., 2012). This is important because nowhere in Bridges of Ross are sandstones between Lobe Elements 1, 2, and 3 connected (Figs. 4–7 and 9); rather, they are stratigraphically separated by shale beds. The same is true for other well-exposed outcrops that contain predominantly lobe elements, such as Dunmore Head (Pyles, 2004), Kilbaha Bay (Straub and Pyles, 2012; Fig. 13), Kilcloher (Chapin et al., 1994), Horse Island (Fig. 13), and Cloonconeen (Fig. 13). As such, individual lobe elements could operate as independent reservoirs or flow units in subsurface reservoirs. However, through erosion, Channel Element 1 erodes into Lobe Element 3 (Figs. 4, 5, and 10) and Channel Element 2 erodes into Lobe Elements 3 and 2 (Figs. 4–7), resulting in sand-sand contacts between the channel elements and their laterally adjacent lobe elements, and therefore a comparatively well-connected reservoir composed of multiple elements. The same is true in other outcrops where channel elements erode into adjacent lobe elements (e.g., Kilbaha Bay, Horse Island, and Cloonconeen; Fig. 12). Connectivity is therefore greatly enhanced in fans with CLTZs located in longitudinally variable positions. Second, Bridges of Ross and Navy Fan are complementary analogs for distributive submarine fans, and can therefore serve as qualitative analogs for constraining conceptual models of reservoirs of distributive systems and as quantitative analogs for conditions of the sizes, shapes, and associations of architectural elements in geologic models. Third, the Ross Sandstone is one of the most commonly visited outcrops of submarine fans in the world. The work included herein can be used to supplement future field guidebooks to this outstanding locality.

CONCLUSIONS

This article proposes end-member stacking patterns for distributive submarine fans (Fig. 1B). One end member contains CLTZs that are located in varying longitudinal positions on the fan, resulting in distributive fans with relatively large areas with lateral juxtapositions of channel elements and lobe elements in their stratigraphy. This pattern is due to compensational stacking that operates laterally and longitudinally. The other end member contains CLTZs that are located at a common longitudinal position, perhaps at the base of a slope, resulting in fans that do not contain lateral juxtapositions of channel elements and lobe elements in their stratigraphy. The Bridges of Ross outcrop of the Ross Sandstone and Navy Fan are complementary outcrop and seafloor analogs for distributive submarine fans in which the CLTZs are located in variable longitudinal positions on the fan. Outcrops of Bridges of Ross document stacking patterns, lithofacies distributions, architectural juxtapositions, and the paleogeographic evolution of a small part of the distributive fan. Many of the architectural associations documented at Bridges of Ross are common to other outcrops of the Ross Sandstone and to those in Navy Fan, supporting their use as paired outcrop-seafloor analogs for distributive channel-lobe complexes. Observations from Bridges of Ross and Navy Fan were used to develop a geometric model that predicts the fractional length of a submarine fan that contains lateral juxtapositions of channel elements and lobe elements. Lateral juxtapositions of channel elements and lobe elements are important because they can serve to drastically increase connectivity in subsurface reservoirs.

We thank David Piper and two anonymous reviewers for their thoughtful suggestions. We also thank our friends and colleagues Mark Kirschbaum, Jane Stammer, and Linda Martin for reviewing early versions of this manuscript. Financial support for this research was provided by Chevron through the Chevron Center of Research Excellence at the Colorado School of Mines and sponsors of Laser Assisted Analogs for Siliciclastic Reservoirs (LASR) research consortium at the Bureau of Economic Geology (Chevron, ConocoPhillips, EnCana, ExxonMobil, Marathon Oil Corporation, Norsk Hydro, Petrobras, Shell, and Statoil). We thank Tim Wawrzyniec and Amy Ellwein for assistance with and use of their Optec lidar scanner and Renaud Bouroullec for assistance with field work.