An incipient slipface avalanche was triggered and the resulting tongue of grainflow deposit was examined in cross-section.
Fig 1. Avalanching sand
These observations involved a 7m transverse dune at Coral Pink Sand Dunes State Park in southern Utah, USA. The most recent wind capable of moving sand had left a cornice deposit on the upper slope of the slipface, which at some places along the brink had already avalanched. An array of bamboo measuring sticks was inserted in the lower slope beneath an unavalanched section, then an avalanche was triggered upslope by disturbing the lower part of the cornice deposit (by tossing a small stone). The resulting flow was filmed (figure 1, to the right) (sometimes it doesn't load the first time -- press 'refresh' on your browser).
The main sand flow was continuous, fed at the top by a scarp eating into the cornice deposit, until reaching the bottom and backing up (more or less coincident, in this case, with exhausting the upper deposits).
(The dark string-like object in the top-left of the video frame is lint on the camera's CCD. The video is contrast-enhanced.)
The tongue of sand that had flowed into the array of stakes was then examined. Figure 2 below shows the array and the height of the grainflow deposit at each stake. Dark-coloured organic material pushed by the flow marks the boundaries of the grainflow tongue.
Fig 2. Array of stakes, and the height of the grainflow deposit at each stake
Laminae at each stake were made visible using a refinement of Bagnold's water seep method. The three figures below show the results at the three rows of stakes. The estimated position of the pre-avalanche surface (based on marks on the stakes) is indicated by a yellow line in the photos. The photo have been resized so that they are all the same scale, and aligned to the former surface (the yellow line). The stakes were inserted normal to the surface.
Fig 3. Laminae at stakes in row 1 (upper-most row)
Fig 4. Laminae at stakes in row 2 (middle row)
Fig 5. Laminae at stakes in row 3 (lowest row)
The prominent shelves or flanges in the lower portion of many of the columns is clear evidence that pinstripe laminae (layers of finer grains) are present in this sand.
However, it is hard to associate any of the pinstripe laminae with the observed flow. If finer grains were deposited by kinetic seiving, we'd expect a lamina at the base of the flow. Likely positions for the base of the flow are: 1) at the former surface, or 2) beneath the former surface. The data here doesn't clearly support either, but lean toward non-erosive flow (ie., the avalanche flowed over the existing surface without disturbing it). Column 1-3 could be interpreted as non-erosive flow (but why is the whole thickness of new grainflow apparently a better conductor of water than the layers beneath it?). Other columns in row 1 seem to support non-erosive flow.
In row 2, there is a strong pinstripe at a location about 4cm beneath the former flow. Is that the shear layer for the triggered flow, or (more likely) was that a previous deposit? There is no hint of a pinstripe at the former surface, but columns 2-2 and 2-3 hint at a shear layer at about 1.5cm beneath the former surface -- this could be the base of erosive flow. (Black marks on the stakes are 1cm apart.)
Flowing sand will deposit a layer of finer grains by kinetic seiving only if that sand has finer grains (ie., only if it has a wide range of grain sizes). It's possible that by the time the sand in this artificially-triggered avalanche travelled to the lower slopes of the dune, all the finer grains had deposited above and the remaining grainflow's range of grain distribution was narrow (and hence forms no basal laminae). Maybe laminae at the toe of the dune (such as observed in the columns here) are typically created by finer grains from grainfall and/or saltation (which would likely be contemporaneous with avalanching in natural conditions), or supplied by grainfall and then modified by subsequent grainflow via kinetic sieving.