What is glacier sliding

Glacier and ice-sheet motion is fundamental to glaciology. However, we still lack a consensus for the optimal way to relate basal velocity to basal traction for large-scale glacier and ice-sheet models (the 'sliding relationship'). Typically, a single tunable coefficient loosely connected to one or a limited number of physical processes is varied spatially to reconcile model output with observations. Yet, process-agnostic studies indicate that the suitability of a given sliding relationship depends on the setting. Here, we suggest that this arises from myriad overlapping setting- and scale-dependent sliding sub-processes, including complicated near-basal stress states not captured by large-scale models, reviewed here as comprising a basal 'sliding layer'. A corresponding 'bulk layer' then accounts for ice deformation only minimally influenced by bed properties. We provide a framework for incorporating arbitrarily many sub-processes within a given region -- separated into normal ('form drag') and tangential ('slip') resistance at the ice-bed interface, stressing that the maximum scale of cavitation is an important contributor to the division between the two. Under reasonable assumptions, our framework implies that sliding relationships should fall within a sum of regularised-Coulomb and power-law components, with a rough-smooth distinction proving more consequential in dictating sliding behaviour than a traditional hard-soft transition.