Tag Archives: appalachian-mountains

Lidar change detection shows where huge landslides have woken up and started moving

Posted on by
Reply

While fast-moving debris flow landslides pose the most immediate risk to human life in the North Carolina Mountains, huge, slow-moving landslides also cause plenty of problems for property and infrastructure. The gradual effects of these landslides (some of which may be hundreds of feet wide) on roads and homes can be devastating. Comparison of 2017 lidar imagery to 2025 lidar imagery, which was collected to capture Helene’s impacts, vividly illustrates the effects of these “big slow movers” on mountainsides over several years. The GIF below shows one such landslide in Haywood County. The first image is 2017 lidar, which is followed by 2025 lidar, in which the movement is visible. The third panel is color-coded lidar change detection, with red indicating areas where the land surface has dropped and blue indicating areas where the land surface has risen due to the slope being pushed outward and upward. Check out how far up the slope the red areas extend…this shows just how big the slide actually is.

Watching the GIF a few times is necessary to fully appreciate the extent of this gradual slide movement. As the lower parts of the slide push outward (blue areas), support for upslope areas is reduced, allowing slide movement to spread upslope. The upper limit of movement is visible as subtle scarps, or downward steps, on the land surface. These scarps can be hard to see without the color-coded lidar change detection imagery, particularly from fieldwork in the forested mountains. The detail image below shows the upper scarps of this huge slide. Look for the visible surface change associated with the upper left red areas. The movement is subtle, but it is clearly visible once you know to look for it. The broken logging road grade just up and left of center is another good marker.

This slide is a good example of slow movement within areas of previously displaced material, generally referred to as “landslide deposit” material by geologists. These deposits cover vast areas of mountainsides in the region. While the exact geologic history of these huge landslide deposits is unknown, they represent accumulated rock and soil that started out much higher on the mountain and progressively slid or fell and piled up on the lower slopes over long periods of time. Most appear to be the product of continual accumulation, meaning that the pile of material is constantly growing and being loaded with more weight. The irregular deposit material is not as strong or stable as soil developed from in-place rock weathering, making it prone to natural landslides. This tendency towards landsliding is enhanced by roadcuts or other disturbances. The GIF below summarizes an overall geologic evolution of one of these deposits in which an old landslide within the greater deposit (brown material) is “woken up,” or reactivated, by development.

In this case, an old, dormant landslide within the accumulated landslide deposit was reactivated by the roadcut at its toe. While the house was built just above the old landslide, the reactivation allowed the slide to expand within the steeper portions of the accumulated deposit material and impact the house. The detail below shows this relationship.

The size of the roadcut alone provided plenty of instability, but this situation may have been made worse if large areas of forest were cleared for the house, allowing more rainfall to infiltrate and further weaken the deposit soil. The toe of the landslide will push out onto the road until it is able to provide enough support for the upslope areas, stopping movement. If the toe is cleared without major support, like a large retaining structure, movement will likely resume. The image below shows what the final sketch would look like with lidar change detection applied, with red showing elevation loss and blue showing elevation gain.

Geologists in the region are very interested in finding old landslides with lidar imagery, as well as mapping the larger zones of accumulated landslide deposit, because of the potential of these areas to begin experiencing slow, but ongoing, movement after disturbance. Haywood County, like all North Carolina mountain counties, hosts plenty of these slow-moving slides. Post-Helene lidar imagery has allowed their movement to be visualized by lidar change detection for the first time. The GIF below shows another example a few miles from the first. Look for the slight movement at the center of the GIF, highlighted by the lidar change detection colors.

In this case, a large roadcut removed an existing landslide toe within a larger landslide deposit area. This toe loss set the stage for reactivated movement. Whether the movement occurred as a result of rainfall during tropical system Fred in 2021, Helene in 2024, or is a steady, ongoing issue, is not known. The movement is subtle, but the colored lidar change detection highlight its clearly. The upper portions of this slide are also worth a look in detail. Look for the thin, narrow area of movement just up and right of center. This narrow lobe is spreading upslope towards a focused pile of fill material which was presumably produced by the roadcut running across the image. The roadcut appears to be cut by erosion or slide movement just above the fill pile.

A third example shows the striking similarities between the geologic aspects of these features. Compare the setup to the sketches above. The rugged, upslope terrain with bedrock outcrops is visible, along with the large, lumpy landslide deposit areas below. The moving slides in this example are both likely reactivations of older, dormant slide, though road building and development have obscured some of the geologic details.

An oblique view is also useful to see where the reactivated landslides occur on the mountainside.

The uppermost scarp of the slide at right is a particularly good example of landslide movement spreading upslope. The GIF below shows the uppermost scarp in detail–watch for the scarp and associated band of red to develop just above the center of the GIF. This feature would be very difficult to identify without lidar change detection. The photos following the GIF shows this feature in the field. It could easily be missed during clearing of the site for development, which could cause plenty of ongoing problems down the road. The scarp is indicated by a yellow arrow in each photo. It looks insignificant, but a shift in the ground surface of its size beneath a house is (or any other constructed feature) is a major issue, to say the least.

Lidar alone is a game-changer in terms of understanding the landscape, allowing widespread and accurate identification of accumulated landslide deposit areas and currently dormant landslides that could be reactivated by development. The ability to locate and track areas of active movement with the new lidar change detection tool further expands our ability to interact with the challenges of building and living in the mountains. Locating areas of active movement can also provide information about what might be triggering slide reactivation (road cuts, tree removal, concentrated runoff from developed areas, etc.), helping to guide engineering response when necessary.

Landslide deposits areas are also notable because they are frequently the final destination of those fast-moving debris flow landslides, adding to the difficulty of safely developing them. Locations experiencing a big, slow-moving slide reactivation often also have significant debris flow hazard. In this case, the house with slow-moving landslide issues just happens to be downslope from a convex “spur” on the mountainside, protecting it from a full debris flow impact but making for an uncomfortable close call.

Many deposits may actually represent the long-term accumulation of debris flow deposits, continuing to receive debris flow material during big storm events. Geologists spend a lot of time mapping deposit areas because they are so closely tied to so many kinds of slope movement. Deposit areas certainly shouldn’t be off limits to development, but any kind of construction on them should look out for old landslides and debris flow pathways and account for the reduced strength of the soil in the deposits. Lidar, along with the expertise of geologists and geotechnical engineers, makes it possible to address all of these potential issues to avoid years of future problems.

Field mapping of landslides is still important, even in the era of LiDAR and AI

Posted on by
Reply

Estimating where landslide debris is going to come to rest is an important part of hazard delineation. LiDAR change detection imagery has made evaluating landslide runouts much easier, with color-coded loss and gain zones providing at-a-glance information about where slides started and stopped. The Haywood County, North Carolina, debris flow shown below is a great example. Red colors indicate loss in surface elevation, showing the initiation zone and scoured track of the debris flow. Blue colors indicate elevation gain, showing where soil and rock (along with woody debris) were deposited, offering a detailed picture of material transport.

Unfortunately, not all slides are this easy to interpret from LiDAR imagery, making field work necessary to better understand a slide’s movement and potential impact area. Just over the ridge from the slide above, a small but significant Helene landslide left a scar on the slope but little additional information about where the fluidized slide material went. In a situation like this, field work is still necessary to really understand what happened. Even though this is a relatively small feature, slides like it caused massive damage to structures during Helene, so understanding its material movement has real-world value. Though it’s hard to visualize, this slide released a few dump truck loads of fluidized soil and rock at once, so we definitely still care about these smaller failures and where they go (it would have looked very much like the slide in the video link)!

This slide left no visible deposit or significant scour on the slope. Some of the fluidized soil would have followed the gully downslope, but a slide of this size could easily overwhelm such a small gully. The slope below the slide is generally intact, with understory plants growing and no clear zone of soil loss due to passage of the fluidized slide over the land. The images below show the slide scar and the view downslope. The Art Loeb trail passes just below the base of the scar (really fine line below the slide in the image above), offering easy access! The details of the slide scar are also worth a look. Colluvial soil with a distinctly dark matrix failed at the bedrock-colluvium contact. The contact zone was damp and seeping on the day of the visit, despite a period of dry weather.

So, where did everything go? Useful evidence still remains at the downhill edge of the green area in the image above. Deposits of cobbles, small boulders, and rafts of vegetation and woody debris are still present well away from the gully (first image below), indicating a much wider, overland “wave” of slide material. Rock debris pinning mats of vegetation against trees is typical in deposits from this sort of slide. The surprisingly delicate stacking of rocks behind saplings (second image) is always interesting.

Looking back up at the initiation zone of the slide, the overland travel of the slide debris is impressive. The deposits shown above are just off the left edge of the image below. The wavy yellow lines indicate the approximate limits of the area over which slide debris moved. Notably, the volume and momentum of the fluidized debris kept it from following the gully, allowing it to wash over a broader area of the slope.

On the lidar change imagery, the affected area would something like what is outlined below. While the affected area looks quite large for the size of the initiation zone/scar, it seems typical of observed behavior of fluidized slides on more open slopes in the area.

The takeaway message is that fluidized, “blowout”-style failures like this one are extremely mobile and can spread significantly across the landscape without leaving any lidar-visible trace. Even though they appear small, they have significant destructive potential as observed with the Sugar Grove blowout in the link near the beginning of the post, and anticipating their potential movement over the landscape is important in hazard areas. Restricting potential impact areas to the edges or mouths of small channels or gullies may not capture the full extent of impact potential. The GIFs below compare the observed behavior (first one; green checkmark!) to the more restricted, “down the gully” model, which falls short in this case (red X).

This type of combined remote and field-based analysis of Helene’s effects is important to get the most out of post-storm datasets. We also have to keep the field work rolling, as some of the finer details of slide behavior are quickly overwhelmed by the regrowth of vegetation and continued stream erosion!