A critical source of noise in tree-ring based chronologies of landslide movements

Abstract

Tree-ring based analysis is currently a widely used tool for obtaining data on the spatio-temporal behaviour of past landslides, which is key information for determining hazard or estimating potential triggers. However, even this approach has its limitations, which may induce a lot of noise in the obtained results. In doing so, the sources of this noise, which in some cases can even simulate a spurious landslide event, are not well known. Thus, this study highlights one of the possible significant sources of noise in tree-ring based landslide chronologies, which is the effect of wind combined with the soft landslide surface, which causes instability of trees that can easily lean to simulate a landslide movement. For this purpose, 197 individuals of common spruce (Picea abies (L.) Karst.) on the flat parts of two flow-like landslides were analyzed. 280 growth disturbances were identified and 17 events were reconstructed. In order to assess the landslide origin of the collected data, this study included analyses of the nature and relative contribution of different types of growth disturbances, their spatial distribution, analyses of the direction and magnitude of tree trunk leaning, as well as analyses of extreme precipitation and wind events. By combining the results obtained, it can be concluded that most of the detected growth disturbances and reconstructed events were induced by wind in combination with the soft landslide surface into which the trees sink. Since these growth disturbances cannot be filtered out by standard reference trees, they are an extremely serious source of noise that can completely invalidate the dendrogeomorphic results of the study. The study thus identifies which parameters should be newly considered in dendrogeomorphic analyses of landslide movements.