Monthly Archives: April 2018

Atlas of slope, shelfbreak gullies, and submarine canyons



The Atlas, as are all blogs, is a publication. If you use the images, please acknowledge their source as indicated below (it is the polite, and professional thing to do).  I retain copyright of all images presented herein

Brian Ricketts –


Marine slopes are bona fide geological settings in themselves, but from a geotectonic perspective they are the region where continental crust is transitional to oceanic crust, and where sediment bypasses the shelf as it heads towards the deep ocean floor – typically as submarine fans.  Slopes, as their name suggests, have significantly greater dip than an adjacent shelf; the break between the shelf and slope is  defined by this break in sea floor gradient.  Slopes frequently are cut by gullies and submarine canyons; the gullies tend to be localized across the shelf-slope break, whereas canyons extend across the shelf (sometimes coming within a few 100m of the shore), to the full depth of the slope.  Gullies and canyons focus sediment transfer to the ocean deep. The Black’s Beach and Point Lobos canyons were visited on an AAPG trip with Tor Nilsen; the Bowser Basin examples I worked on in the late 1980s – early 90s.

This link will take you to an explanation of the Atlas series, the ownership, use and acknowledgment of images.  There, you will also find links to the other categories.

Click on the image for an expanded view, then ‘back page’ arrow to return to the Atlas.

The images:

The iconic, Eocene Pt. Lobos submarine canyon, California, where canyon-fill conglomerate (brown hues) is in abrupt contact with Salinian granodiorite (white weathering) –  an example of a steep canyon wall.

Looking south, along the Pt. Lobos canyon axis. Conglomerate at the base, overlain by turbidites.




Layered Pt Lobos canyon-fill conglomerate against the blocky weathering granodiorite bedrock. California. The canyon wall is indicated by arrows.



Discordant packages of conglomerate canyon-fill, Eocene Pt. Lobos submarine canyon, California.




Interbedded canyon-fill conglomerate and turbidites, Eocene Pt. Lobos submarine canyon, California. Some of the conglomerate beds have debris flow characteristics, others may be down-canyon traction current deposits.



Local slope facies between channelised, canyon-fill conglomerate, presenting delicately laminated siltstone-mudstone, starved ripples with mud-drapes, thin graded beds (looking more like distal turbidites),soft-sediment load structures, and a few sand-filled burrows. Eocene Pt. Lobos submarine canyon, California.


Slump discordant packages of interchannel, thin-graded fine-grained sandstone, Eocene Pt. Lobos submarine canyon, California




A muddy debris flow consisting almost entirely of slope facies mudstone rip-ups, plus a few pebbles, overlain by clast-supported, canyon-fill conglomerate. Eocene Pt. Lobos submarine canyon, California



Black’s Beach, iconic coastal cliffs that reveal sediment gravity flow deposits (mainly turbidites and debris flows), and the remnants of an Eocene submarine canyon. This view is north of Scripps Pier, California.



Pebble-lined canyon floor at Black’s Beach, cutting into estuarine and other paralic facies (root structures and burrows are common). Eocene, California

Basal conglomerate filling the canyon floor, Black’s Beach, California




Typical channel conglomerates eroding into thick (proximal) turbidites and thinner channel overbank facies, Black’s Beach submarine canyon. Signs at the beach entrance warn of rock falls,  house collapses, and other exposures.

Discordant canyon-filling conglomerate and thick proximal turbidites, Black’s Beach, California.




A shelfbreak gully, incised into slope deposits, overlain by cyclothemic, and progressively shallowing shelf facies. Gully fill is mostly conglomerate. It is thickest at the waterfall (about 40m). Initiation of gullies was by fluvial erosion during sea level lowstands, aided by slumping in inherently unstable slope deposits. The gullies delivered gravel and sand to the basin beyond the slope.  Upper Jurassic, Tsatia Mt, Bowser Basin, British Columbia.

A paper on this topic: Ricketts, B.D. and Evenchick, C.A. 1999.  Shelfbreak gullies; Products of sea-level lowstand and sediment failure: Examples from Bowser Basin, Northern British Columbia.  Journal of Sedimentary Research, v. 69, p. 1232-1240.

The base of ‘waterfall’ shelfbreak gully, overlying slope mudrock and thin turbidites.  Bowser Basin, British Columbia.




A closer view of the ‘waterfall’ gully margin (Tsatia Mt), showing numerous discordant contact within the slope mudrock facies, and minor slumping of the gully fill. At least two major episodes of fill are recorded here. Bowser Basin, British Columbia.



Shelfbreak gullies extend down slope. Here, two packages of channelized conglomerate (along the ridge line) occur entirely within slope facies.  The small lenses of conglomerate below are thought to represent channel spillover lobes. Joan Lake, Bowser Basin, British Columbia.


A large slump block of gully-fill conglomerate, embedded in slope mudrocks, shows the inherent instability of the gullies and associate slope deposits. Bedding within the block are also disrupted. The block is located just below the right margin of the ‘waterfall’ gully, shown in the above images.  Tsatia Mt, Bowser Basin, British Columbia.


Slope facies, here consisting of relatively undisturbed thin, graded, very-fine grained sandstone-mudstone (thin turbidites), and a few small starved ripples in the laminated mudstone-shale. Bowser Basin, British Columbia.

Thin graded sandstone beds, starved ripples, laminated sandy mudstone, small slump folds, syn-sedimentary pull-aparts or boudinage, and microfaults, all features that are  typical of slope facies mudrocks. Bowser Basin, British Columbia.

Left: laminated mudstone-siltstone and a few thin graded sandstone beds.  Slope facies, Bowser Basin, British Columbia. Right: stratigraphic discordances occur at all scales in the Bowser Basin slope deposits. Many are caused by slumping, but discordant mudrock packages also arose from flows spilling over the channel-gully margins.

Slump-induced, listric-style fault in slope mudrocks, Tsatia Mt.  The fault flattens out along a thin turbidite bed; displacement decreases towards the fault tip at top right, where overlying beds are continuous. Bowser Basin, British Columbia.



 Slope mudrock and thin sandstone beds are truncated by a synsedimentary fault (just above the lens cap). Bowser Basin, British Columbia.



Upper Jurassic Submarine canyon complex of stacked channels, within a slope assemblage, Todagin Mt, Bowser Basin, British Columbia. Like the shelfbreak gullies, although on a grander and more prolonged scale, the canyon delivered mud, sand and gravel to the deeper Bowser Basin This view taken from Tsatia Mt.



View from the center of Todagin canyon-fill, showing the step-like stacking of successive channels. Maximum thickness of the conglomerate-fill exceeds 300m.  On this ridge, an equivalent thickness of slope mudrock overlies the canyon. Bowser Basin, British Columbia.



Two views of the Todagin canyon base, and bedded conglomerate-fill, most of which was deposited by debris flows, sometimes separated by thin turbidites. Each view shows about 25m of section. Bowser Basin, British Columbia.

Stacked channel conglomerate; the channel margin is almost vertical through about 20m thickness. The steep margin may be synsedimentary fault controlled – the overlying beds are not displaced. Opposite the margin are typical slope mudrocks and thin turbidites. Bowser Basin, British Columbia.



The upper section of Todagin canyon, showing back-stepping channel stacking. Bowser Basin, British Columbia.




Near the top of the canyon succession, two cycles of thinning- upward turbidites, that may have formed as the active channel moved across the canyon floor, away from this site of deposition. Note the slump discordance in the lower cycle. Bowser Basin, British Columbia.


Turbidites overlie the main Todagin canyon-fill conglomerate, about 10m thick, capped by a smaller channel. Note the slump discordance in the lower cycle.  Bowser Basin, British Columbia.




The top of the main canyon-fill conglomerate here is overlain by slope mudrock, cf. the image above.  Turbidites, at the location in the image above, have thinned significantly or pinched out completely in this exposure. The overlying conglomerate forms a smaller, more isolated channel, pinching out to the right. The overall influence of the submarine canyon is waning at this stage.  Bowser Basin, British Columbia.

Mudstone rafts, captured by debris flows, near the base of the Todagin canyon succession. Bowser Basin, British Columbia.




Contrasting debris flow textures. Left: mud-supported clasts in a more plastic debris flow. Right: clast-supported frameworks that probably formed in a more fluid, sheared debris flow. Both types are common in the Todagin canyon succession, and in the gullies. Bowser Basin, British Columbia.

Well-developed layering in this debris flow, probably formed during prolonged, quasi-continuous surging flow of grit to cobble sized clasts. The whole unit is about 8m thick. Hammer bottom left. Todagin canyon.





Atlas of shelf deposits



The Atlas, as are all blogs, is a publication. If you use the images, please acknowledge their source as indicated below (it is the polite, and professional thing to do).  I retain copyright of all images presented herein

Brian Ricketts –


The term ‘shelf’ is used here loosely – it covers a range of submarine settings, mostly shallower than about 300m, from the upper slope to shoreline, the shoreface, fairweather and storm wave-base.  There is some overlap with the ‘Paralic’ category, but the context of the shallowest examples (like beach, shallow subtidal) is in their relationship to their deeper counterparts.  The separation of the ‘Shelf’ and ‘Paralic’ categories is a bit artificial, and one of convenience.

This link will take you to an explanation of the Atlas series, the ownership, use and acknowledgment of images.

Click on the image for an expanded view, then ‘back page’ arrow to return to the Atlas.

The images:

Coarsening- and bed-thickening upwards shelf (about mid shelf) to shoreface cycle, Jurassic Bowser Basin, northern British Columbia.  The coarser facies contains hummocky crossbeds (HCS) at storm wave-base, and subaqueous dune-ripples above fairweather wave-base.  There are numerous trace fossils indicative of high energy, such as Ophiomorpha, Rosellia, and Thalassinoides.


Coarsening=upward cycle at about outer- to mid-shelf – some HCS at the top of the sandstone. This is a more seaward cycle to that shown above.   Jurassic Bowser Basin, northern British Columbia.



This shale to thinly bedded sandstone cycle occurs close to the shelf edge, at the transition to slope deposits.  There are a few bottom current ripples, but no HCS or larger dune structures. Jurassic Bowser Basin, northern British Columbia.



The chert-pebble conglomerate accumulated in a shelfbreak gully.  The uninterrupted transition from shale-dominated slope to shelf is located immediately to the right of the gully margin.  Jurassic Bowser Basin, northern British Columbia. Details of the gullies have been published here: Shelfbreak gullies; Products of sea-level lowstand and sediment failure: Examples from Bowser Basin, northern British Columbia. 1999,  Journal of Sedimentary Research 69(6):1232-1240

Hummock cross stratification (HCS) in a typical lower shoreface shelf cycle (storm wave-base),  Jurassic Bowser Basin, northern British Columbia. Hammer rests on a thin pebbly debris flow that immediately underlies the HCS unit.  It is generally thought that HCS forms during storms, from the combination of a unidirectional flowing bottom current, possibly as a sediment gravity flow, that is simultaneously moulded by the oscillatory motion of large storm waves.

Possible swaley bedding, formed in much the same way as HCS, but where the hummocks have been eroded leaving the concave-upward swales. Jurassic Bowser Basin, northern British Columbia.



Storm rip-ups of shelf muds in a mid-shelf cycle.  Jurassic Bowser Basin, northern British Columbia.




Many shelf cycles in the Bowser Basin succession, terminate abruptly and are overlain by a bed of fossiliferous (ammonites, trigoniids and other molluscs), pebbly, mudstone.  This marks the transition form a highstand (HST) to succeeding transgression; the mudstone is the TRansgressive Systems Tract (TST).


Transition from a sandy HST, to fossiliferous mudstone (small ammonite near the lens cap) of the TST. The top of the TST corresponds to a maximum flooding surface (MFS) – the stratigraphic record of maximum transgression.  Jurassic Bowser Basin, northern British Columbia.



The upper portion of this coarsening upward shelf cycle, the highstand systems tract, contains low-angle planar lamination and some hummocky cross-stratification (HCS). The base of the transgressive unit (TST) is an erosional surface. Jurassic Bowser Basin, northern British Columbia.


Two views of a lenticular, trough crossbedded pebbly sandstone that has cut into the top of a shelf cycle. This has been interpreted as a lowstand fluvial channel, that traversed and eroded the shelf as it was exposed during falling sea level.  This was one mechanism for transporting gravel and sand to the slope and deeper basin, via shelfbreak gullies (like the one pictured above).  Jurassic Bowser Basin, northern British Columbia.

The same fluvial, lowstand channel shown in the images above. The channel is about 2m thick.  Jurassic Bowser Basin, northern British Columbia.



Panorama of a slope-shelfbreak gully-shelf-to fluvial transition, beautifully exposed at Mt Tsatia, Jurassic Bowser Basin, northern British Columbia. Conglomerate on the immediate right are equivalent to the rusty beds near the opposite summit. The shelfbreak is located at the top of the wedge-shaped gully (corresponds to the top of the waterfall) – below the gully are slope deposits. The thickness of strata in this view is more than a kilometre.

A really nice (folded) succession of coarsening upward shelf cycles, Eocene Eureka Sound Group, South Bay, Ellesmere Island. The Eocene shelf was laterally equivalent to river-dominated deltas (Iceberg Bay Fm.) to the north and east.



Coarsening upward mid-shelf – shoreface cycles at South Bay, Ellesmere Island (same location as image above). Small subaqueous dunes, ripples and HCS are common.


Coarsening upward muddy shelf cycles, mostly below storm wave-base, but the occasional cycle extending into lower shoreface (some HCS).  Eocene, Eureka Sound Group, Ellesmere Island

Downlap of muddy outer shelf siltstone and mudstone, Eocene Strand Bay Fm, Ellesmere Island




Sandy, Paleocene shelf dunes forming part of large sandwave complexes. Most of the crossbeds are the planar, or 2D type. The right image shows detail of crossbed foresets, with some reactivation surfaces (probably tidally induced); crossbed is about 40cm thick.  There is some indication here of tidal (flood-ebb) couplets.  Expedition Fm, Eureka Sound Group, Ellesmere Island.

Sandwave complex on a Paleocene sandy shelf, made up of multiple dunes. Eureka Sound Group, Ellesmere Island.




The abrupt, corrugated surface here is a Late Pleistocene wave-cut platform, eroded across Pliocene mudstones (Tangahoe Fm). The wave-cut platform and overlying estuarine-dune sands are part of the Rapanui Formation, near Hawera, New Zealand.  The eroded corrugations and channels contain wood, shells and pebbles.

Late Miocene – Early Pliocene coarsening upward shelf cycles, from outer-mid shelf siltstone-sandstone, to shoreface, tidally induced sandy coquina sandwaves (left image).  The 3 images show part of the highstand systems tract. The carbonate facies are part of the classic, cool-temperate water limestones of Wanganui Basin, New Zealand.  Matemateaonga Fm, Blackhill.

Thick HST calcareous sandstone – limestone, Late Miocene – Early Pliocene Matemateaonga Fm, Blackhill.




Large planar crossbeds in shelf sandwaves (HST), overlain by a pebbly shellbed deposited during the next transgressions (TST).  Late Miocene – Early Pliocene Matemateaonga Fm, Blackhill.



Typical transgressive systems tract (TST) shellbed, Late Miocene – Early Pliocene Matemateaonga Fm, Blackhill.




Detail of shelf dune foresets with backflow ripples climbing up foreset dip. Late Miocene – Early Pliocene Matemateaonga Fm, Blackhill.




Subtidal sandstone with lenticular and wavy bedding deposited during ebb-flood tides. Late Miocene – Early Pliocene Matemateaonga Fm, Blackhill.




Large planar crossbedded calcareous sandstone, formed either as shelf sandwaves or platform of a tidal inlet flood delta. Late Miocene – Early Pliocene Matemateaonga Fm, Blackhill.


Atlas of delta deposits


A beautiful Landsat image of Lena River Delta, Siberian Russia. The entire delta complex
is about 200 km wide. At present, the most active part is the right-center lobe.
Image credit:

The Atlas, as are all blogs, is a publication. If you use the images, please acknowledge their source as indicated below (it is the polite, and professional thing to do).  I retain copyright of all images presented herein

Brian Ricketts –


Deltas come (and go) in all shapes and sizes.  They form where a trunk river discharges into a largish body of water – mostly shallow seas, but modern and ancient deltas also form in large lakes. Early delta facies models (1960s-70s in particular) were based primarily on the Mississippi birds-foot delta.  As time, and alternative models were developed, it became apparent that the classic river-dominated birds-foot geometry was part of a much larger spectrum of deltas, including those that are tide-dominated and wave-dominated.  The resulting facies tend to be quite different in each of the categories, particularly at the seaward margin of delta accumulations. For example, wave-dominated deltas tend to be higher energy environments at the point where river-derived sediment is dispersed at the seaward margin.

The term ‘delta’ is also a kind of catchall – there are many different kinds of sedimentary facies in deltas, ranging from strictly fluvial to strictly marine. So, for example where fluvial deposits are clearly associated with a delta, they are included in the latter category.

The examples here include the classic Carboniferous, river-dominated deltas from Kentucky; and Late Cretaceous – Paleogene wave and river-dominated types from the Canadian Arctic . I have a few examples from the lacustrine deltas in Ridge Basin, although the field trip to that wrench basin focused on sediment gravity flow deposits (a great AAPG 10-day trip to several ‘turbidite’ basins in California, led by Tor Nilsen, 1988).

This link will take you to an explanation of the Atlas series, the ownership, use and acknowledgment of images.  There are a couple of NASA images that are in the public domain. If you copy these please credit NASA accordingly.

Click on the image for an expanded view, then the ‘back  page’ arrow to return to the Atlas

The images:

Sediment plumes discharged from Mackenzie Delta, Yukon, into Beaufort Sea, are a maximum during spring-early summer thaw .  The locus of mouth-bar deposition is probably in the whitish region immediately outboard of the lower delta plain. Taken July 19, 2017 by the Operational Land Imager on Landsat 8.

Image credit: NASA Earth Observatory, image by Jesse Allen. Landsat data from the U.S. Geological Survey.

Aerial view of Mackenzie delta plain, in less clement weather.  Not far from the northern Yukon town of Inuuvik.




A spectroradiometer image of Mississippi Delta, January 17, 2016 shows sediment plume distribution at the iconic birds-foot, and at other drainage points along the coast.

Image credit: NASA, MODIS sensor on Terra satellite,

A classic Kentucky Carboniferous highway exposure of channel sandstone cutting into floodplain siltstone-mudstone and thin overbank sandstone. This is overlain by point-bar accretionary foresets (building away from viewer).  Mostly upper delta plain.



These two images create a kind of panorama of Carboniferous of point bar – channel sandstones overlain by interdistributary bay mudstone-siltstone. The point bar overlies and is partly equivalent to thin floodplain coal (left image, base of outcrop).

Lepidodendron log in fine-grained floodplain deposits, Carboniferous, Kentucky. A much younger John Horne adding know how and levity to the field trip stop.



A compaction fault has juxtaposed Carboniferous delta plain channel sandstone against floodplain silts and muds.  Near Hazard, Kentucky.




Delta plain channel migration and down-cutting has left this coal ‘island’ (top right) (Number 7 seam).  The channel overlies interdistributary bay muds.   Daniel Boon Parkway, near Hazard, Kentucky.



Channel down-cutting of the Carboniferous Number 8 coal, subsequently overridden by accretionary point bar foresets.  Hazard, Kentucky.




Proximity to a basement fault has focused sedimentation, resulting in the stacking of successive delta plain fluvial channel-point bars. Point bar accretion was to the right. Carboniferous, near Louisa South, Kentucky


Channel margin and low-relief levee, overlying floodplain silts and muds. Carboniferous, Hazard, Kentucky.




This is the only example I have seen of a channel and twin levees. The levees are overlain by interdistributary bay muds, and at the top of the outcrop, distributary channels. Ivel, Kentucky



Channel margin slump block, delta plain, near Rush, Kentucky





Crevasse splay (whitish unit above bus in left image), that has broken through a distributary channel levee and distributed fine sand, silt and mud across the interdistributary bay.  Right image shows a more general view. Below the splay is bay fill muds. Above are more bay fill muds and thin coal seams.  Betsy Layne, Kentucky.

Broader view of thin crevasse splay (people standing on it), overlain here by accretionary point-bar foresets.  Betsy Layne, Kentucky




Crevasse splays tend to be thicker near the breached levee. Here, two splays are overlain by interdistributary bay muds.  Ivel, Kentucky.




The distal section of a crevasse splay (i.e. farther into the interdistributary bay) – the pencil (right of center) spans the entire splay thickness. It is sandwiched between two coals. Betsy Layne, Kentucky.



Detail of crevasse splay deposits, shows laminations of fine-grained sandstone-siltstone,  and abundant bioturbation that commonly obliterates primary layering.  Betsy Layne, Kentucky.



Prodelta siltstone – mudstone overlain by distributary mouthbar sandstone, near Pikeville Kentucky.




Prodelta mudrocks, with a few thin fine-grained sandstone lenses. Pikeville, Kentucky.




The frequency and thickness of sandstone beds increases towards the distributary mouth bar. Pikeville, Kentucky.




Laminated and thin-bedded fine sandstone, with a few crossbeds and ripples that may indicate mouth-bar deposition above wave base.  Pikeville, Kentucky.



Thin, graded sandstone beds are more common in the transition from distributary mouth-bar to prodelta. Pikeville, Kentucky.




A late Cretaceous-Paleocene wave-dominated delta in the Canadian Arctic. Here coarsening- and bed thickening-upward units cycle through prodelta to slower shoreface, with abundant evidence of traction currents, including hummocky cross-bedding. Expedition Fm. Eureka Sound Group, Axel Heiberg Island.


Closer view of a coarsening-upward prodelta-shoreface cycle, Strand Fiord, Axel Heiberg Island. Hummocky cross-bedding occurs with associated with ripples, small sand waves, and current scours. The abrupt upper surface (top right) marks the beginning of transgression,with a maximum flood surface about 1m above.


Coarsening-upward cycles, from prodelta to lower delta plain, in an Eocene river-dominated delta, Axel Heiberg Island. Channel sandstone and thin overbank coals occur in the upper part of each cycle. Iceberg Bay Fm, Eureka Sound Group. The river-dominated system is separated from the earlier wave-dominated system by a 3rd-order transgression that influenced deposition through the entire basin.

Small coarsening upward, shoreface cycle capped by coal, representing marine incursion of the min river-dominated delta. Eocene Iceberg Bay Fm, Axel Heiberg Island. Coal at the base of outcrop is the top of the preceding cycle.



Trough crossbedded fluvial channel, delta plain, in the Eocene Iceberg Bay Fm. Ellesmere Island




Panorama of multiple fluvial channel-swamp-bog coal cycles composing the main Eocene delta plain, Iceberg Bay Fm, Axel Heiberg Island.



Late Miocene, lacustrine prodelta (base) to distributary channel-bar facies, Ridge Basin. Ridge Basin is a strike-slip basin wrenched by the San Gabriel fault strand of the San Andreas transform. The delta, on one side of the basin, is equivalent to the Violin Breccia that formed continuously along the active faulted margin.


Prodelta to channel and mouth-bar sandstone, Late Miocene Ridge Basin, California.




Trough crossbedded channel and pebble lags, lacustrine delta plain, Ridge Basin California.





Late Miocene delta top, channel – mouth-bar sandstone, lacustrine delta, Ridge Basin, California.  Interfingering pebble and cobble bands in the left image were derived from the Violin Breccia

Crossbedded delta top sandstone, Ridge Basin California.


Atlas of glaciofluvial – periglacial deposits


Atlas of alluvial fans


The two main groups of alluvial fans illustrated here are from humid and arid environments. Arid climate fans I have visited or worked on are from Death Valley in eastern California (part of Mojave Desert in the Basin and Range geological province, and the mountains of Atacama, northern Chile.  The Atacama examples are about 4000m above sea level. The Death Valley photos were taken in 1996 during an SEPM Research Symposium.

The Atlas, as are all blogs, is a publication. If you use the images, please acknowledge their source as indicated below (it is the polite, and professional thing to do).  I retain copyright of all images presented herein

Brian Ricketts –

This link will take you to an explanation of the Atlas series, the ownership, use and acknowledgment of images.

Click on the image for an expanded view, then ‘back one page‘ arrow to return to the list


The images:

Headwaters of active alluvial fans in Tertiary Hills, Northwest Territories.  Bedrock here is Paleocene fluvial conglomerate, sand and sub-bituminous coal, that is being recycled by modern alluvial-fluvial drainage. Clast-size range in the fans is determined by the clast sizes in the eroding conglomerates.


This small (humid) alluvial fan drains into Peel River, east Yukon. The fan’s outer edge dips its toes in Peel River. Inactive segments of the fan are incised by the river, but active segments provide new gravel, sand and mud to active river side bars (river flow is to the top). Recently active fan channels and flooded swaths are mainly in the central part of the fan, having migrated from the fan edge farthest from the viewer.


Incision of a gravel-sand flow unit on the alluvial fan that merges with Peel River (image above). Deposition as bedload was probably generated by stream flood.




Part of a large, coastal (humid-cold) alluvial fan complex along the north coast of Yukon (west of MacKenzie River delta).  The active channel at this time was itself, a largish braided river. Field of view across the coastline is about 3 km.




Several, small alluvial fans merge with the braided stream that drains into the south end of Canon Fiord, Ellesmere Island.  Potential paleoflow directions in the fans would be oriented about 90 degrees to indicators in the braided river. This is an arid setting, with most flow during spring and early summer thaw.



Incision of a Late Pleistocene (very humid) alluvial fan at Franz Joseph, New Zealand.  The gravels are very coarse; boulders up to 3m across. The sediment source is in the immediate background – the western edge of the Southern Alps (here, mostly greenschist).



Thick, poorly bedded debris flows in Middle Eocene alluvial fans that accumulated outboard of rising thrust belt during the Eurekan Orogeny, Axel Heiberg Island (Arctic Canada).  Source rocks consist of various Triassic and Jurassic sandstones and diabase.



Death Valley from Dante’s View, looking east towards the Panamint Range (a block-faulted and uplifted metamorphic core complex). Salt flats in mid-view (mostly halite, some gypsum and borax), and a nice succession of (arid) alluvial fans that interfinger with the saline facies.  This is one of the classic Basin and Range couplings between fault blocks and intervening basin.


Death Valley, looking north from Dante’s View – the fault block here lies immediately east of Panamint Range.  Alluvial fans merge with the salt flats. The dimly visible whitish area in the distance is Mesquite Flat sand dunes, near Stovepipe Wells.



The view east of Dante’s View, to successive Basins and Ranges.





Excellent exposure of Hanauphan fan, Death Valley. There are dozens of debris flow and sheet flood events recorded in this outcrop.




Stacked debris flow and sheet flood conglomerates in Natural Bridge fan, Death Valley.  The red colour of most sediments here is another testament to the arid environment.



Crudely layered debris flow conglomerates in Natural Bridge fan, Death Valley. Most flows developed during flash floods.  Person’s elbow for scale, bottom right.



Finer grained flow units, Natural Bridge fan, Death Valley. Some of these may have been deposited by hyperconcentrated flows – sand-gravel-mud-water mixtures that have a rheology somewhere between water-bedload, and debris flows. A more recent example is shown in the image below.


Section through a recent flash flood, hyperconcentrated flow, Death Valley. Texturally, the flow resembles a muddy debris flow; poorly sorted, mud-support of clasts, but the range of clast-sizes is much smaller.




Arid alluvial fans merging with gypsum-halite salars, Atacama, northern Chile. Most fans fringe Eocene and younger volcanic cones.



Looking down-slope along inactive parts of fans.  Left image shows levees of cobbles and boulders deposited by a debris flow. Atacama, northern Chile. Eocene and younger volcanic edifices in the distance.



Alluvial fan lobes encroaching a gypsum-halite salar, Atacama, northern Chile. Right image shows an elevated fan (left centre) that represents deposition during a phase of higher lake levels; the older fan is now partly degraded. The cuspate and indented distal fan margin record a succession of fan lobes. Grey-brown colours reflect basalt-andesite clast compositions, enhanced by desert varnish.

Cross section through inactive fan segment, showing multiple debris flow and sheet flood layers. The outcrop is about 2m high.  This is a more distal part of the alluvial fan, and clast sizes are usually less than 15cm. Atacama, northern Chile.



Two distinct debris flow units, separated by thin bedload-deposited sands, or possibly a thin hyperconcentrated flow.  Arid alluvial fan, Atacama, northern Chile.  The outcrop sheen is caused by films of halite and gypsum from the nearby salar.



Well-bedded ephemeral stream deposits, with a few bedload ripples, some clast imbrication, and scour-and-fill structures around large clasts.  Distal, arid alluvial fan, Atacama, northern Chile.



Well bedded-laminated sheet flood sands are probably traction deposits (a few bedload ripples). The intervening poorly-sorted, pebbly sands are hyperconcentrated flow deposits. From a more distal part of an arid alluvial fan, Atacama, northern Chile.



Desert varnish affects most clasts in Atacama, northern Chile alluvial fans.  Darker varnish hues generally indicate older deposits, and longer surface exposure.  The relative ages of successive or overlapping fan lobes can often be determined from varnish colour mapping.


Large gypsum books on an alluvial fan surface, close to its contact with a salar. The crystals average 5-6cm across.  Atacama, northern Chile.



Atlas of fluvial deposits


This collection of images from various modern and ancient fluvial settings, includes meandering (high sinuosity) and braided (low sinuosity) rivers.  Where possible I have paired modern analogues with ancient examples.

The Atlas, as are all blogs, is a publication. If you use the images, please acknowledge their source as indicated below (it is the polite, and professional thing to do).  I retain copyright of all images presented herein

Brian Ricketts –


This link will take you to an explanation of the Atlas series, the ownership, use and acknowledgment of images.

Click on the image for an expanded view, then ‘back one page‘ arrow to return to the list

The images:

Where transverse sea-waves (generated by wind shear) meet riverine standing waves. Manawapou River, west coast NZ.



An Arctic braided river, Strand Fiord, Axel Heiberg I. Flow is seasonal, and at a maximum during spring-early summer thaw. The closer view (right) shows a bit more detail of within-channel bar surface features. The sediment load is mixed sand-gravel.



Bonnet-Plume River (Yukon) braided reaches, that, unlike high Arctic examples, are framed within vegetated overbank and some inactive gravel bars. The sediment load is mixed sand-gravel. The left image also shows an ephemeral meandering reach that cuts through inactive braid-bars.

Vegetated braid bar gravels exposed in a cut bank. Bonnet-Plume River, Yukon.




Gravelly, braided reaches of Ahuriri River, Otago, NZ. The headwaters are in the distant Southern Alps.




Gravel bar, Ahuriri River, Otago, New Zealand, with falling-stage sand deposits along the bar trailing edge (downstream edge).




Gravelly, braided glacial outwash near the terminus of Strand glacier, Axel Heiberg Island. Erosional chutes, that form during falling stage, are filled with crossbedded sand.





2-dimensional pebbly dunes, each with simple planar tabular crossbeds, formed along the margins of larger gravel braid bars during river falling stage. Monster River, Yukon.


Braid bar-top ripples (flow to the right), imprinted by rain drops. Monster River, Yukon.




Glacial outwash flow during spring thaw, between gravel bars, Strand Glacier (Axel Heiberg Island). The water is mud-laden. The standing waves are in-phase with antidune bedforms.




Strong imbrication  of flat clasts along an exposed river bed.  Flow was to the right.




Bank-full conditions in an ephemeral, gravelly, Franklin Pierce Bay (Ellesmere Island) stream, during a rare mid summer Arctic rainstorm.  About 25mm fell in a matter of hours.  The stream rose very quickly and forced us to move our camp.  There were many rockfalls from nearby cliffs during the storm.



Eocene gravel bar deposits, Otto Fiord, Ellesmere Island. The sandy interval mid-picture was interpreted as small, crossbedded, falling stage sand bars, like those shown in the modern analogues.





Left: Thick, Eocene, syntectonic gravelly braided river deposits (Buchanan Lake Formation; Franklin Pierce Bay, Ellesmere Island). Sediment was shed from evolving thrusts during the Eurekan Orogeny.  Right: Large gravel crossbeds (up to 3m thick) in the same formation.



Two examples of trough crossbedded sandstone in low sinuosity channels (braided) in the Middle Eocene Buchanan Lake Fm, Axel Heiberg Island (Arctic Canada).  These channels are the more distal equivalents to syntectonic gravelly deposits (Eurekan Orogeny).



Festooned trough crossbeds in Proterozoic low sinuosity sandy channels, Loaf Fm. Belcher Islands (about 2 billion years old). The right image shows small ripples that probably developed during waning stream flow.


As river deposits settle and consolidate, the water between grains is forced out by the weight of the sediment (this is called dewatering). The process commonly disrupts and contorts the sedimentary laminae, forming structures that superficially resemble pillows; these structures are given the general name ball and pillow.  Loaf Fm. Belcher Islands (about 2 billion years old).


Reddened (iron oxides), desiccated mudrocks interbedded with channelized sandstone in the Proterozoic Loaf Fm, Belcher Islands; proof that the level of oxygen in the ancient atmosphere had increased significantly.




Paleocene Sandy braid channel and associated bar crossbeds (Summit Creek Fm.) exposed in right bank of MacKenzie River near Fort Norman, Northern Canada.



Active sand bars attached to a semi-permanent, vegetated bar, mid-stream, Clearwater River, Alberta.  This is a possible candidate for an anastomosing river.


Large chunks of rock can be carried significant distances across open water while embedded in the tangle of tree roots. A possible answer to the mystery of some drop-stones.



Small meandering stream, point bars, and oxbow lake, north of Calgary, Alberta.




Point bar and overbank deposits in a high sinuosity (meandering channel) in the Upper Cretaceous Dunvegan Formation, Peace River, northeast British Columbia. Flow was to the right.





Crossbeds in Upper Cretaceous Dunvegan Formation, Peace River, northeast British Columbia. Left: stacked planar tabular crossbeds (2D subaqueous dunes); Right: Trough crossbeds (3D subaqueous dunes).  Both types are associated with the Dunvegan point bar (image above), and the sand-filled channel below.


Sandstone-filled channel in the Dunvegan Formation, Peace River, northern British Columbia. This view shows overall asymmetry of the channel.




Detail of sandstone channel cutting into slightly older, muddy overbank deposits. Upper Cretaceous Dunvegan Formation, Peace River, northeast British Columbia.



Sandy point bar deposits in high sinuosity channel (meandering), overlain by thin floodplain lignites in the Middle Eocene Buchanan Lake Fm, Geodetic Hills, Axel Heiberg Island.  The lignites contain abundant, well preserved conifers (Spruce), Hickory, and Metasequoia fronds, cones and seeds. The point bar is about 4m thick.


Small crossbeds and laminated sandstone in point bar deposits, Middle Eocene Buchanan Lake Fm, Geodetic Hills, Axel Heiberg Island. There is abundant plant material throughout. Pen (mid image) is 15cm long.



Multiple sets of climbing ripples, Middle Eocene Buchanan Lake Fm, Geodetic Hills, Axel Heiberg Island. These form when a significant suspended load of fine sand settles and becomes part of the bedload  on the channel floor.



Multiple thin lignite – subbituminous coal beds with exquisitely preserved tree trunks in growth position, Middle Eocene Buchanan Lake Fm, Geodetic Hills, Axel Heiberg Island.  The succession here represents a stacking of flood plain and forested areas adjacent to meandering rivers.  These deposits accumulated in a more distal position to the emerging mountain front during the Eurekan Orogeny.  See my post on the Fossil Forests



Some of the exquisitely preserved plant material from the Geodetic Hills Fossil forest. Left: Metasequoia cones looking like they were buried yesterday; Right: Metasequoia fronds and hardwood leaves. Middle Eocene Buchanan Lake Fm, Geodetic Hills, Axel Heiberg Island.



Large meandering river point bars and overlying floodplain-swamp muds, associated with the Princess Coals, Carboniferous of Kentucky, near Rush (Highway I-64).  Point bar ‘foresets’ consist of very laminated, rippled, and slumped, fine grained sandstone-mudstone.  I visited these outcrops during an excellent AAPG field trip run by John Horne, 1984.


Detail of the inclined point bar layers show numerous discontinuities in sandstone lenses and wedges, and truncation surfaces that indicate shifting sediment distribution across the bar, and possibly some erosion.  Carboniferous of Kentucky, near Rush (Highway I-64).



Slumping and rotation of laminated sandstone-mudstone in point bar foresets (Carboniferous of Kentucky, near Rush, Highway I-64). Small synsedimentary faults cut the middle layers.





Odd-ball (sic) structure sometimes found in fluvial deposits, are armoured mudballs. These form when a chunk of sticky mud slumps from a channel margin, is rolled by currents along the channel floor, and in the process picks up small pebbles and bits of wood. The recent example on the left is from Mackenzie River, near Fort Norman. The Paleocene example, to the right of the lens cap,  (right image) is from the Summit Creek Fm, in an outcrop fortuitously nearby the modern analogue – small pebbles impregnate the mud ball surface.



Atlas of beach, lagoon, bar, estuary, tidal flats


Shallow marine – estuarine – tidal flat – lagoon and coastal dunes

This collection of images spans the shallowest marine environments including beach, lagoon-bay with all the associated environments such as sand-spits and barrier-bars, tidal flatsestuaries and coastal dunes.  Where possible I have paired modern analogues with ancient examples.

The Atlas, as are all blogs, is a publication. If you use the images, please acknowledge their source as indicated below (it is the polite, and professional thing to do).  I retain copyright of all images presented herein

Brian Ricketts –


This link will take you to an explanation of the Atlas series, the ownership, use and acknowledgment of images.

Click on the image for an expanded view, then ‘back one page‘ arrow to return to the list

The images:

Tairua estuary, east coast Coromandel Peninsula, New Zealand.  Two images taken during falling tide, exposing attached and semi-detached sand-shell bars. The main channel is at bottom of each image.




Raglan ebb tidal delta, bedforms on the platform attached to the permanent shoreface


The estuary on the south side of Galway Bay, County Clare, Ireland, near New Quay. The boulder-cobble beach consists almost entirely of Burrens limestone (Carboniferous). Left view from Abbot Hill.

Mudflats and algae, Kinvara, at the head of the estuary, south side Galway Bay.


Karst in Burren Limeston at Flaggy Shore, New Quay, County Clare, has been accentuated by salt corrosion and mechanical erosion. It is overlain by boulders of locally derived Carboniferous limestone


Boulder storm ridge at Black Head, County Clare – the heart of the Burrens. All boulders are locally derived Carboniferous limestone


Potholes in Burren Limestone, Flaggy Shore, County clare.


Seagrass, ripples, and burrow excavations in a tidal pool, Flaggy Shore, County Clare


Large, 2D dunes, intertidal Minas Basin, Fundy Bay




Shallow subtidal to intertidal, 2D subaqueous dunes, Rowat Fm, Belcher Islands (Aphebian, about 2 billion years old). Hammer for scale.



Cross-sectional view of 2D subaqueous, intertidal dunes, showing complex migrating dune-formed crossbedding, and dune reactivation, Rowat Fm, Belcher Islands (Aphebian, about 2 billion years old). The sands are mixed siliciclastic-carbonate (dolomite).



Reactivated 2D dunes with superposed ebb tide ripples, Minas Basin, Fundy Bay




Proterozoic tidal inlet, showing cross-section of reactivated subaqueous dunes (mid image), possible herringbone crossbeds, and smaller ripples. Rowatt Fm, Belcher Islands (about 2 billion years old). Lens cap bottom right.



Multiple dune sets, intertidal, Minas Basin, Fundy Bay




Sandy tidal flat ripples, Minas Basin, Fundy Bay




Paleocene, straight crested and bifurcating intertidal ripples, Expedition Fm, Axel Heiberg Is;and, Canadian Arctic




Straight-crested ripple train in Paleocene intertidal deposits, Expedition Fm, Axel Heiberg Island, Canadian Arctic. Hammer left-mid image.




Flood tide ripples over-ridden by smaller ebb tide ripples sets. Minas Basin, Fundy Bay




Interference ripples in Proterozoic tidal flat facies, Belcher Islands




Large 2D and 3D dunes, and superposed run-off ripple sets, Minas Basin Fundy Bay




Tidal flat, interference ripples, Minas Basin Fundy Bay




Ebb tide run-off & reactivation of 3D dunes, Minas Basin Fundy Bay




Large 3D dunes, Minas Basin Fundy Bay




2D flood tide dunes and small ebb tide ripples,  Minas Basin Fundy Bay




2D and 3D intertidal dunes, Minas Basin Fundy Bay





Eroded salt marsh cycles, Minas Basin Fundy Bay


Small meandering tidal channel in very muddy estuarine tidal flats, Whitford Estuary, south Auckland. Bank failure is common.




Salt marsh, sedges and small mangroves being transgressed and eroded by tidal flat. This is a modern example of a ravinement surface. Whitford, south Auckland



Eroded salt marsh depsosits, transgressed by sandy tidal flat – beach. The erosion surface is a modern, active ravinement surface. Galveston, Texas.




Two examples of Paleocene tidal bedding (mostly lenticular and wavy bedding) interfingering with  lagoon and marsh. Eureka Sound Group, Ellesmere Island


Paleocene tidal bedding interfingering with marsh-lagoon-bay sediment. On the right, the thicker sandstones may represent storm washovers into the bay. Eureka Sound Group, Ellesmere Island


Coursening- and sandier-upward bay or lagoon subtidal to beach, cut by small tidal channels (lenticular sandstones). Eocene, Eureka Sound Group, Ellesmere Island.




Ebb tidal delta at the mouth of Waikato River, south Auckland.




Paleocene subaqueous dunes up to 2m amplitude, in tidal inlet-delta, overlain by thin tidal flat-salt marsh deposits. Expedition Fm, Axel Heiberg Island, Canadian Arctic




Large within-channel dunes in a tidal inlet associated with a sand spit facies; the spits and bars were attached to (paleotopographic) headlands across an unconformity eroded into Ordovician carbonates. Paleocene, Eureka Sound Group, Ellesmere Island.



Two views of the unconformity between Ordovician carbonates and Paleocene estuarine-tidal channel-spit facies. Eureka Sound Group, Ellesmere Island.


Beach cusps, Long beach, Vancouver Island, formed during a prolonged period of fine weather. The next storm will erode the cusps and redistribute the sand



Typical beach stratification in an eroded berm; primarily laminated sets with low-angle truncations, parallel, or slightly inclined to the beach face.




Tidal inlet standing waves (antidunes) in an outgoing tide, Mangawhai Heads, north Auckland. The antidunes migrate upcurrent (against the current) and gradually build until they break, subsequently reforming.


Proterozoic tidal channel – inlet trough crossbeds; this outcrop gives a 3-dimensional view of individual sets. Paleoflow was into the image. Rowatt Fm. Belcher Islands.



Flaser, lenticular and wavy bedding in late Pleistocene deposits near Ihumatao, Auckland.  White muddy sediment overlies and envelopes grey sandy ripples, and fills troughs between ripples.



Coastal dunes, Galveston coast, Texas




Washover fan breaching coastal dunes, Galveston coast, Texas




Paleocene washover-fan sandstone associated with barrier island and tidal inlets (see images above from the same formation), Expedition Fm, Axel Heiberg Island, Canadian Arctic.



Stacked storm deposits associated with an upper tidal flat, each layer consisting of ripped up muds. Rowatt Fm, Belcher Islands. Proterozoic.



Mudcracks in salt marsh, Kaiua, NZ




Proteroaoic supratidal desiccation cracks and voids in multiple layers of delicately laminated dololutite. Some curled slabs may be coated with crpytalgal laminae. A layer of storm-derived lutite rip-ups at the bottom of the image. Rowatt Fm. Belcher Islands. See below for a modern analogue.


Desiccated, curled, algal mats from a salt marsh near Galveston Texas. The mats are easily disturbed during high or storm tides.




Mangroves: Left image: stabilizing shell banks (storm ridges) adjacent a tidal Inlet, Auckland Harbour; Right image: in a salt marsh, Kaiua, bordering Hauraki Gulf.


Everglades Mangroves, Florida.  A tangle of roots and pneumatophores  that are living quarters for so many species.  On the right, an epifauna of barnacles, small snails and bryozoa.

Everglades alligators, including the little guy on its parent’s head.




Sea grass on a tidal flat, Savory Island, coastal British Columbia.




Gravel bar formed at the intersection between a high energy, West coast New Zealand beach, and the Tangahoe River mouth






Atlas of aeolian deposits


Here are some examples of modern and ancient sand dunes from continental and coastal settings. Click on the images for an expanded view.

The Atlas, as are all blogs, is a publication. If you use the images, please acknowledge their source as indicated below (it is the polite, and professional thing to do).  I retain copyright of all images presented herein

Brian Ricketts –


This link will take you to an explanation of the Atlas series, the ownership, use and acknowledgment of images.

Click on the image for an expanded view, then ‘back one page‘ arrow to return to the list


The images:

Close up of Jurassic Navajo Sandstone dune crossbeds with tangential toe-sets, Zion National Park, Utah. Image height is about 2m. Large sand dune complexes in a continental desert, about 180 million years ago.



Really big dune foresets (lee-face deposits) in the Jurassic Navajo Sandstone, Zion National Park, of southern Utah




Closer view of the exposure in above image, showing large foresets and individual dune truncation. The person (bottom left) is about 2m tall. Navajo Sandstone, Zion National Park, Utah




Coastal dunes bordering Great Exhibition Bay, northernmost New Zealand. The white sand contains 90-95% quartz.  Much of the sand for the modern dunes comes from reactivation of late Pleistocene dune complexes, and sand stored on the sea floor.


Dune sets in Late Pleistocene deposits, Great Exhibition Bay, northernmost New Zealand.




Large, active barchan dunes traverse Kokota sandspit that forms a protextive barrier to Parengarenga Harbour, northernmost New Zealand. This view is of the main lee face. The dune is about 8m high.



Small lakes and ponds that are sometimes perched between sand dunes, accumulate mud, sand and peaty material from plants. Here are at least 6 interdune pond mud layers, some with fossil roots (also replaced by clays). See also the image below. Late Pleistocene, Great Exhibition Bay.


Interdune lake near the inland margin of stabilized, Pleistocene sand dunes, Kariotahi, south Auckland, NZ.




Multiple dune sets in a Pleistocene sand barrier, Kariotahi, south Auckland. Formation of the old sand barrier and dune field eventually created the straightened coastline that typifies the west coast of North Island.  The barrier formed during repeated rises and falls in sea level over about 2 million years.


Multiple dune sets in Pleistocene sands (Same cliff face as image above), indicating a fairly consistent orientation of the preserved lee faces, and the direction of dune migration (to the left – east). Kariotahi, Southwest Auckland, New Zealand.



Barrier and dune sands, Kariotahi, south Auckland. Here an older set of dunes (orange weathering) are capped by an thin soil and bands of iron oxide (along the surface that slopes steeply to the left). The soil would have stabilized the older dune sands. A younger set of dunes formed over the top of the old soil.



A more panoramic view of the old soil and dunes in the image above.




The landward margin of Pleistocene sand dunes, now stabilized. The steep slopes are the old lee slopes (the steep surface of the dune that faces down-wind). When active, the dunes would have migrated towards the viewer. Kariotahi, south Auckland.



Reddened sandstone with large dune crossbedding, in the New Red Sandstone on Arran, west Scotland. The red colours are formed by iron oxides, after the sand was deposited during the Permian through Triassic – 280 to 200 million years ago.



Multiple dune crossbed sets in New Red Sandstone, Arran, west Scotland, 280 to 200 million years old.




Salt-tolerant grasses and shrubs extend their roots deep into the dunes to capture as much moisture as possible, and in so doing help to stabilize the dune.



Mesquite Flat dunes, Death Valley, near Stovepipe Wells.  An inland ‘sea’ of sand, usually organized as dunes, is called an erg.




Dune sand at Mesquite Flat, migrating over desiccated pond silts. Death Valley




At least two generations of wind ripples in Mesquite Flat dunes. All those trails attest to numerous insects and the larger critters that hunt them.


Submarine mud flows and landslides modified Kaikoura canyon during the 2016 M7.8 earthquake


A slashing blow by some mythical behemoth, knifing effortlessly through earth’s rocky foundations; a (seemingly) bottomless chasm, a canyon, with nothing but the wind between you and whatever lies below. A bit over the top perhaps, but canyons often spark the imagination – standing on the lip can feel like being perched on the edge of the world, vertiginous for some.  Most canyons have been carved by the relentless churning of stream and river, incising the layers of rock and removing the sediment to distant shores.

Terrestrial canyons have their submarine counterparts, that transect the submerged, outer margins of continents and volcanic islands. Submarine canyons commonly mark the transition from shallow continental shelves and platforms to ocean basins, acting as conduits for sediment delivery from rivers, deltas and shallow seas, to the deep oceans.  And like their land-based cousins, they are deep (1 to 2 km, or in the case of Grand Bahama Canyon, 5km), steep-sided incisions in the ocean floor.  More than 600 submarine canyons have been identified world-wide from bathymetry maps.

The 3-dimensional bathymetric reconstruction of Monterey Canyon (top image), about 100km south of San Francisco, illustrates common attributes of these structures. The canyon cuts deeply into the break between the shelf and steeper marine slope – in this image the break is a definite line separating light blue from darker shades.  The Monterey Canyon head encroaches onto the shallow shelf to within a few 100m of the shoreline; this is actually atypical of most other canyons where incision of the sea floor usually begins closer to the shelf edge. The canyon channel snakes down slope, eventually flattening out on the deep ocean floor; the main channel is joined by several smaller tributaries. Several smaller gullies are also incised into the shelf edge and slope.

The general opinion among earth scientists is that submarine canyons are formed by two main processes: Erosion by sea-floor hugging flows of mud and sand (given the general name sediment gravity flows), and by collapse of the steep margins, producing submarine landslides (and potentially, tsunamis).  Common triggers are thought to include storm surges and earthquakes. The primary basis for this interpretation is abundant geological evidence of past events, combined with some experimental work, but it remains a largely theoretical interpretation because there have been very few direct observations of either process in action.  The reasons for this disparity are that submarine flows of mud and sand are relatively rare events (at least on a human time scale), and because of the difficulties inherent in witnessing such processes in deep water. For this reason, recent events in Kaikoura Canyon (southeast New Zealand) have sparked significant international interest.

Kaikoura Canyon (New Zealand), 60km long and up to 1200m deep, is located along the tectonically active Hikurangi margin, close to the Alpine Fault system that transects northern South Island and the adjacent submarine shelf.  At its deepest extent (about 2100m) the main canyon channel merges with Hikurangi Channel, which at more than 1500km, is one of the longest deep-sea channels in the world; Hikurangi channel wends its way across the more subdued ocean floor towards the abyssal Pacific Ocean.  The submarine canyon head is an uncomfortable 1000m from the coast, a spitting distance that elevates the risk of destructive tsunamis that can evolve from submarine landslides along the canyon walls. November 2016, and the magnitude 7.8 Kaikoura earthquake, provided a rude reminder of the potential for disaster. The seismic jolt activated slope collapse and sediment movement down the canyon slopes and main channel; fortunately, the ensuing tsunami was small, but the bonus for science was huge. Mapping  of the canyon head and main canyon channel, fortuitously three years before the earthquake and three months after the event, has enabled scientists to track the changes to channel morphology and sediment distribution that can be attributed solely to the earthquake (The project was coordinated by NIWA – New Zealand’s National Institute of Water and Atmospheric Research).

The first two images show before (2013) and after conditions at a location near the canyon head (closest to shore). Large swaths of muddy sediment were dislodged from the ridges and slopes, cascading into the main channel; most of the canyon head is now devoid of its sediment mantle.  Parts of the canyon floor are 50m deeper than before the earthquake, because of erosion by the moving sediment.

The second set of images show before-and-after scenes of the canyon floor at 1800 to 2100m water depth. The striped pattern is formed by large, ripple-like gravel waves, or dunes, that under normal conditions would migrate slowly downslope. However, most of the gravel dunes were moved at least 500m downslope by the rapidly transiting muddy flow.

Much of the dislodged sediment continued as a turbulent muddy flow down the main canyon channel and thence to the deeper Hikurangi Channel; the flow had sufficient momentum to carry it more than 680km from its source. Evidence for this comes from deep-sea cores taken 4 days, 10 weeks, and 8 months after the earthquake.  Cores were taken from the floors of both canyon channel and the more distant Hikurangi Channel, plus the flatter area, or overbank, beyond the channel banks (analogous to a river floodplain).  The reasoning here is that, if sediment gravity flow deposits can be identified in the overbank region, it means that the flow itself was deeper than the channel and, given that we know how deep the channel is, an estimate can be made of the minimum depth of the actual flow.  Overbank deposits were detected in cores, indicating that the moving flow was at least 180m thick, 680km from Kaikoura Canyon. As Joshu Mountjoy (one of the project leaders for NIWA) has pointed out, this has proved to be one of the few occasions in which actual flow dimensions in a deep-sea channel could be measured.

From the Kaikoura event we have confirmed that seismicity can trigger physical modifications to submarine canyons and submarine slopes, and that sediment is flushed from canyons to the deep ocean by far-travelled, turbulent muddy flows (i.e. sediment gravity flows). We have learned something of the stability of the canyon itself and the sediment that gradually mantles the sea floor. The legacy of the Kaikoura earthquake (or any major earthquake for that matter) is often voiced in terms of broken lives, disrupted highways, and the costs of rebuilding. There should be no attempt to minimise these outcomes, but we should also remind ourselves of the advances in scientific understanding of earthquakes, and the geological consequences that accrue from an event like this. We should applaud these gains in knowledge because ultimately such knowledge will help save lives and property.

Most of the information for this post is gleaned from NIWA news articles and publications, linked in the text above.