On 15 September 2024, a vegetated hillside above the A83 in Scotland collapsed into a 500-tonne debris flow, forcing the road to close—despite less than 20 millimetres of rain in the preceding 24 hours. This matters because existing landslide warning systems rely on rainfall thresholds. When a slope fails in dry conditions, with no earthquake or other obvious trigger, those systems cannot predict it. Such "DrySlides" have rarely been documented, and their underlying mechanics are not understood. The researchers hypothesise that plant root reinforcement and deformation-induced changes to subsurface water flow combine to cause failure, and that measurable precursors may exist in the landscape. If the project succeeds, it could transform how engineers assess landslide risk on vegetated slopes. Transport Scotland is currently revegetating this very hillside as a nature-based solution to stabilise it. Understanding when root strength is overcome—and when it is not—would directly inform those efforts. The work could also lead to new warning systems that detect precursors before a dry slope fails, protecting critical transport routes without relying solely on rain gauges.
View original technical description
At 07:30 on the 15th September 2024 a shallow, vegetated landslide above the A83 'Rest and Be Thankful' in Scotland transitioned into 500 tonne debris flow, forcing a precautionary closure of the road. The failure occurred in 'dry' conditions (<20 mm rainfall in the 24 hours before failure) with low soil saturation. It is a paradigm shift to consider landslides being triggered in dry conditions with no other external driver (e.g. an earthquake) that has rarely been documented. Landslides in the U.K., especially debris flows, are often linked to the closest rainfall event when the triggering time is unknown. The 'Phase 7' landslide mentioned above, named after the phase of mitigation nets and catch-pit construction, offers a rare and transient opportunity to analyse a 'DrySlide'. Existing landslide warning systems rely on the well-established link between rainfall and landslide triggering. Warnings are issues when rainfall exceeds predefined intensity-duration thresholds. 'DrySlides' thus pose risks that are not usually assessed and for which there are no current warning or management systems. We fundamentally do not understand the processes of dry failure and how we can use an improved process understanding to allow failure prediction - the ultimate aim of much landslide work. We hypothesise DrySlides are due to a combination of plant root reinforcement resisting soil deformations prior to failure, and deformation-induced alterations to subsurface water flow. If this is true then green engineering / bioengineering / 'nature-based solutions' need an improved understanding of the role of plant root strength in failures and how/when (or IF in the case of 'arrested' landslides) this strength is overcome. This is especially topical given current revegetation efforts on this slope by Transport Scotland and Forestry and Land Scotland. We believe DrySlides are likely forecastable with precursors linked to failure mechanics and that precursors are recorded in the landscape.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know