Can we infer avalanche–climate relations using tree-ring data? Case studies in the French Alps

Abstract : Dendrogeomorphology is a powerful tool to determine past avalanche activity, but whether or not the obtained annually resolved chronologies are sufficiently detailed to infer avalanche-climate relationships (in terms of temporal resolution) remains an open question. In this work, avalanche activity is reconstructed in five paths of the French Alps and crossed with a set of snow and weather variables covering the period 1959-2009 on a monthly and annual (winter) basis. The variables which best explain avalanche activity are highlighted with an original variable selection procedure implemented within a logistic regression framework. The same approach is used for historical chronologies available for the same paths, as well as for the composite tree-ring/historical chronologies. Results suggest that dendrogeomorphic time series allow capturing the relations between snow or climate and avalanche occurrences to a certain extent. Weak links exist with annually resolved snow and weather variables and the different avalanche chronologies. On the contrary, clear statistical relations exist between these and monthly resolved snow and weather variables. In detail, tree rings seem to preferentially record avalanches triggered during cold winter storms with heavy precipitation. Conversely, historical avalanche data seem to contain a majority of events that were released later in the season and during episodes of strong positive temperature anomalies.