Establishing regimes of landslide activity – Analysis of landslide triggers over the previous seven decades (Western Carpathians, Poland)

Long-term analyses of landslide patterns and triggering factors, covering several decades of continuous data, including periods of both acceleration and stability, are a key to understanding landslide activity, background, and variability. This study focused on the long-term relationships between landslide activity and the triggering factors, precipitation in particular, for three landslides in the Western Carpathians, Poland (temperate climate conditions of Central Europe). We determined the activity and triggers for the studied slopes based on dendrochronological reconstruction covering 68 years (1951–2018), including tree-ring eccentricity and compression wood dating for 107 Norway spruce trees. Through statistical comparison between the dendrochronological record of landslide activity and meteorological data (based on daily precipitation totals), we established types and seasons of precipitation that are critical for triggering landslide activity. We found regional similarities but also significant local differences between the three studied slopes. We were able to explain the origin of the differences through the individual features of landslide topography and structure, such as the depth of the shear zones, disintegration of landslide blocks resulting in a plastic, flow-like movement of the material, and location of the landslide blocks in relation to high groundwater levels in the valley floors. Finally, we determined the optimal sequences of triggers leading to heavy landsliding for each slope, therefore establishing the regimes of their activity. Furthermore, we used the same methodological approach on a landslide located in a highly seismic and monsoon area in the Hengduan Mountains, Sichuan, China, where we also revealed the pattern of landslide-precipitation dependence. Based on these results, we argue that the long-term regularities in landslide response to triggers can be generalised into regimes, as is commonly done with river discharge, groundwater levels, and their hydro-meteorological background. We propose establishing “regimes of landslide activity” that are based on decades of observations and reconstructions. Our study demonstrates that such a long-term approach can be an efficient tool for describing and explaining the variability of landslide activity and hazards over space and time.