LAUSR

Abstract Dendrogeomorphic analyses provide long and continuous chronologies of mass movements that are useful for the detection of trends related to climate change. Socio-environmental changes can, however, induce non-stationarities. This study addresses the following questions: (1) How does the evolution of forest cover induce non-stationarities in tree-ring based reconstructions of snow avalanche activity? (2) How are trends inherent to tree-ring approaches distinguishable from real fluctuations in avalanche activity? Using dendrogeomorphic techniques, we reconstructed snow avalanches in six adjacent paths in the French Alps. Results show two distinct trends in process activity between 1750 and 2016. In the southern paths, the frequency of snow avalanches increased sharply in the 1970s. The distribution of tree ages, as well as old topographic maps, allow an attribution of this trend to the destruction of large parts of the forest stand by a large snow avalanche in the 1910−20 s. This extreme event induced a sudden change in the capability of newly colonizing trees to record subsequent snow avalanches. In the northern paths, by contrast, progressive afforestation starting in the mid-19th century, as well as colonization of the release areas after World War II, resulted in a strong reduction in snow avalanche activity since the 1930s. Even if global warming remains a possible additional driver of snow avalanche activity at the study sites, the rural exodus and the abatement of pastoral practices during the 19th and 20th centuries are the main explanations for the observed trends in process activity. Results also illustrate a need to clarify the complex interrelations among forest evolution, global warming, social practices, and the process activity itself when interpreting trends in mass movements.