- 1French Geological Survey, Orléans, France (y.thiery@brgm.fr)
- 2French Geological Survey, Montpellier, France (b.colas@brgm.fr ; j.guitet@brgm.fr)
Landslides are ubiquitous geomorphological phenomena occurring in various parts of the world, not only in mountainous regions with irregular terrain but also in areas with more moderate relief (e.g., cuesta fronts, plateau slopes, rocky coastal zones). Each year, they cause significant damage to populations and infrastructure. A large majority of landslides are triggered by precipitation.
Currently, there is a growing implementation of early warning systems for these rapid and sometimes destructive events. These tools represent a powerful alternative to mitigate human losses and reduce infrastructure damage. However, such tools rely on precise landslide data catalogs, including accurate location and timing. This information is essential to produce susceptibility maps and establish triggering thresholds. These thresholds enable the construction of destabilization scenarios to assist authorities during crises or emergencies while facilitating prediction and prevention efforts for local populations.
Unfortunately, in many cases, even when landslides are well-located, there remain significant uncertainties regarding their occurrence dates (ranging from weeks to months or years). For instance, the French national database reports that only 21% of landslides are dated to the nearest day, while 69% are dated beyond a month. These temporal limitations complicate the establishment of usable triggering thresholds and reduce the effectiveness of warning tools.
Since 2019, the French Pyrenees have experienced an increase in rainfall events associated with significant geomorphological manifestations on slopes, such as superficial landslides. These phenomena have impacted infrastructure, notably roads and tracks, causing traffic interruptions, as recently observed in the Aspe Valley. Some areas not previously identified as susceptible to landslides highlight the need to improve knowledge and prediction of these events.
This contribution presents a methodology applied to two sectors in the French Pyrenees (Pyrénées-Atlantiques and Hautes-Pyrénées) to establish triggering thresholds probabilities using a recent landslide catalog. The limited and recent temporal data availability raises questions about their relevance. To address this constraint, a strategy was developed to define probabilities associated with specific rainfall episodes and establish vigilance thresholds. These thresholds were spatially applied and coupled with landslide susceptibility maps to obtain triggering probabilities under given meteorological conditions. This methodology represents a first step toward the development of a warning tool for rainfall-induced landslides in the Pyrenees.
How to cite: Thiery, Y., Colas, B., and Jeremie, G.: Precipitation-induced landslides in data-scarce sites: challenges and applications in the French Pyrenees, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18988, https://doi.org/10.5194/egusphere-egu25-18988, 2025.