Combining geostructural and retrospective analyses for rockfall susceptibility zoning

Abstract Rockfall is one of the most common hazards in mountainous environments, threatening infrastructure, and lives. However, a factor limiting preliminary rockfall hazard mapping at the cliff scale is often the lack of knowledge of potential source areas. In the best case, rockfall sources are known from short-term observations (a few years) based on LiDAR, photographic, or monitoring surveys. The presence and characteristics of discontinuities within the rock mass can then be used to identify source areas with differing release susceptibilities. In this investigative study, we used two approaches to assess spatial patterns of rockfall activity along the Täschgufer cliff (Swiss Alps). In a first step, we combined tree-ring and trajectory analyses to estimate the probability that rockfalls occurring from a given pixel of the cliff reach the sampled trees. This retrospective analysis of past activity also quantifies a rockfall release frequency, homogeneously distributed among the pixels susceptible to affect the trees. In a second step, we applied a high-resolution geostructural analysis to classify the cliff pixels according to their level of fracturing and to highlight source areas potentially more prone to rockfall detachment. Both approaches agree upon the absence of clear compartmentalization and a uniform weathering over the cliff. These convergent results, in line with existing observations, confirm the robustness of our results. By contrast, we hypothesize that the retrospective analysis yields an overestimation of release frequencies resulting from site peculiarities, such as the absence of preferential propagation corridors or the considerable distance between the cliff and the upper edge of the forest. Despite these site-related limitations, the methodology developed here clearly evidences the complementarity of both approaches for rockfall source area susceptibility zonation: while the retrospective analysis allows the delineation of areas potentially threatening the sampled trees, the geostructural investigation identifies, amongst these source areas, the most susceptible to trigger rockfalls. In that sense, we believe that the methodology deployed here provides a robust overall assessment of rockfall activity which could be used to prioritize mitigation measures and applicable for hazard management strategies in the future.