Subaerial sediment-water flows on hillslopes have received growing attention from the scientific community in recent years. Interest in this type of geomorphic process is twofold: they constitute a major risk factor and they have generally been poorly defined in the past. Major classification schemes are considered here, with a particular interest in criteria for discriminating between various subaerial flow type processes. Identification techniques from various fields (morphometry, rheology, geomorphology, sedimentology) are reviewed. Recent accounts of flow processes as part of a continuum of processes as opposed to belonging to distinct categories are discussed. It is proposed that process identification include information drawn from various methods in order to minimize identification error, given that correct identification is essential to appropriate hazard assessment and mitigation measures.

Statistical relationships between weather conditions and the release of snow avalanches in the low-elevation coastal valleys of the northern Gaspé Peninsula are still poorly validated. As such, we explored climate–avalanche relationships through classification tree algorithms applied to tree-ring reconstructions of avalanche events. In order to assess the contribution of local factors on avalanche activity, avalanche regimes on east- and west-facing slopes were analyzed and compared. The results showed that avalanches on east-facing slopes appear to be primarily related to large cumulative snowfall in January, February, and March. On west-facing slopes, avalanches are mainly due to episodic snowfall and warming temperatures. However, both sides of the valleys showed the potential for the release of large avalanches in November and December, which is earlier than expected by the literature. Indeed, the weather variability at that time of the year (temperature oscillation around 0ºC) appears to favor the formation of an early, unstable snowpack and subsequent triggering of avalanches, such as the wet slab avalanche recorded by a time-lapse camera in November 2014. This camera provided a useful insight on the capacities of classification-tree models to link the yearly resolution of tree-ring data with weather triggers at different timescales.