A sand model landslide compared to the 2018 Llusco event (with coordinates of the Llusco slide!)

by Philip S. Prince, Virginia Division of Geology and Mineral Resources

If you Google the word “landslide”…the first search result you get is the Fleetwood Mac song. I suppose this says something about the place of Earth Science in the 2019 world, but whatever (more on this at the end of the post!). Clicking the “Images” tab improves things, assuming you are indeed seeking information about the geologic feature. The very first thumbnail that appears is the Llusco landslide of 2018, which occurred outside of Cusco, Peru, destroying the village of Lutto Kututo (apparently without casualties due to its slow, week-long emplacement). Wikipedia has chosen this feature as the headline image for their “landslide” page, and its striking appearance makes it a good choice.

Wikipedia’s headliner “landslide” image, with annotations of notable features. Note how the top of the slide  tilts back towards the bare earth headscarp. The headscarp and cracks are obvious, but the lower portion of the slide is relatively undamaged and is not obviously displaced in the image. The outermost point on the slide toe would be directly above the “y” in “comparatively.” See the Google Earth images near the bottom for an overall view of the slide extent. Image sourced here.

In addition to a very crisp and obvious headscarp, the upslope portion of the slide mass is conspicuously broken into numerous blocks separated by impressive cracks and fissures. I first heard about this slide through Dave Petley’s outstanding Landslide Blog post, which connected the fracturing of the slide mass to the shape of the failure surface.  I thought this was a really interesting example of using subsurface geometry and general landslide know-how to explain the very dramatic surface characteristics, and I tried to produce an analog model slide with similar features using moderately cohesive granular media (which means sand with some flour mixed in to make it stickier). The main purpose of the model is to put the rounding and cracking of the upslope portion of the slide into context, as the cracks themselves have made the Llusco slide a very popular feature around the internet (also more on this at the end).

The model result

The video linked here shows the model and generally narrates the concepts it illustrates, which are further described below. Note that the model slide is emplaced in a single, rapid event, unlike the real thing (unless it could be considered some version of scaling time down as well as physical size). Even so, the final product shows the appropriate features, which result from controlling the shape of the slide’s failure surface.

The Landslide Blog post references the non-circular failure surface on which the slide detached and moved. “Non-circular” means the curvature of the failure surface changes along its length; the appearance of the slide suggests the failure surface was strongly curved at its upslope end and much less curved downslope. Accordingly, the bottom of the upslope portion of the slide mass is strongly curved, and when it moves onto the less curved part of the failure surface, the entire upper part of the slide mass must change shape. Fracturing of the top of the slide thus begins.  I find reading that confusing and I wrote it, so hopefully the following cross section-style images do a better job.

The red dashed line shows the failure surface. It is steep and curving in its upslope portions, and becomes nearly planar and gently tilted on the downslope end.

As the slide moves along its failure surface, the base of slide mass will stay in contact with the failure suface due to gravity (meaning the gap shown here certainly won’t form!!). For this to happen in the non-circular scenario, the upslope portion of the slide with the curved base will have to deform. This occurs progressively during movement, with the collective result of extending and cracking the top of the slide (yellow arrows) so the once-curved part of the base can fit the flatter failure surface.

The end result is a slide mass that is heavily damaged on its upslope end, and relatively intact near the toe (depending on where on the slope the failure occurred, etc. etc.) Note that the upslope portion of the top of the slide is curved, and the very head of the slide tilts back towards the steep scarp…this is an easy-to-see characteristic of the Llusco slide. This is a reasonable approximation of what the cross section of the model would look like.

Note that the upslope portion of the top of the slide is rounded and tilts back towards the scarp, forming a trough or depression. Compare this shape to the Llusco slide…it’s the “rollover” or back-rotation marked in the first image of the post.

The model slide can be seen to start flexing and cracking just after it begins moving…

Cracks are just starting to form as the rounded slide base moves onto the less curved part of the failure surface. Note how the top of the slide is starting to round at its upslope end; its surface was initially planar like the intact slope. Rounding means extension, and extension means cracking.

Extension and fracturing continue to the point that the slide begins to take on complex characteristics, with its downhill portion starting to move independently of the broken uphill blocks.

The block-and-fissure portion of the model turned out well, and resembles the general pattern seen in Llusco. Note this requires strong and cohesive model media; as you might imagine, regular dry sand cannot support vertical slopes and blocks of this shape.

The cohesive sand-flour mixture allows individual blocks with vertical sides to remain intact. I think a component of lateral extension created the isolated blocks instead of parallel ribs.

The real thing, from the best YouTube video.

The cohesive nature of the model material is also reflected in the steepness of the model headscarp. The head of the slide appears to have back-rotated 30 degrees or more.

The model slide is steeper than the Llusco slide (typical due to the inadequate weakness of the microbeads), but the shape of its failure surface produces similar geometric features.

The pictures and video links in the Landslide Blog post are by far the best resources for looking at the actual slide. It’s hard to find on Google Earth, and it is necessary to look at the imagery history to get a good, cloud-free view once you do find it. The March 2018 imagery is pretty good. The slide can be found at 14.390336S  72.109662W. Like the model slide, the fracturing and extension are very intense in the upslope portions of the slide mass. The lower part of the slide, which was presumably situated above a much less curved portion of the failure surface, is comparatively undamaged and its displacement is not immediately obvious until you look at building damage.

The yellow circle marks a cluster of trees in the intact lower portion of the slide. In the bottom image, it’s easy to see where the toe of the slide overran the lower road. Follow the road from the bottom of the image up, and you can see a conspicuous row of buildings on the right side of the road that barely escaped destruction; the road was clearly blocked by the slide.  Presumably, the large red-roofed building visible in the top image was not so lucky. Note also the pond that formed (or was enhanced?) by damming from slide emplacement. The most recent imagery on Google Earth shows a new road cut across the slide toe.

Another interesting aspect of the Llusco slide is how many conspiracy theory/fake news/supernatural phenomenon websites picked it up, apparently due to the appearance of the cracks and fissures. I think there used to be even more, as a few links go to deactivated Twitter accounts. The cracks are definitely alarming and eye-catching, but it is important to remember that even the most dramatic expressions of Earth process have rational, and sometimes simple, explanations. I think a major part of creating or accepting the explanation is learning to visualize the overall function of large, slow, or even out-of-sight systems. This is what I try to accomplish with physical models. If you were totally unfamiliar with landslide dynamics or rock mechanics, or if you just saw a single picture of the cracks, it could be very difficult to attribute them to a basic geometric explanation as Dave Petley was able to do in the Landslide Blog post. For me, seeing it all at once makes it easier to work out the cause-and-effect relationships. That said, I somehow doubt that failure surface geometry will ever be as popular as apocalyptic polar reversal frequency change predictions!