Modelled subglacial floods and tunnel valleys control the life cycle of transitory ice streams

Abstract

Abstract. Ice streams are corridors of fast-flowing ice that control mass transfers from continental ice sheets to oceans. Their flow speeds are known to accelerate and decelerate, their activity can switch on and off, and even their locations can shift entirely. Our analogue physical experiments reveal that a life cycle incorporating evolving subglacial meltwater routing and bed erosion can govern this complex transitory behaviour. The modelled ice streams switch on and accelerate when subglacial water pockets drain as marginal outburst floods (basal decoupling). Then they decelerate when the lubricating water drainage system spontaneously organizes itself into channels that create tunnel valleys (partial basal recoupling). The ice streams surge or jump in location when these water drainage systems maintain low discharge but they ultimately switch off when tunnel valleys have expanded to develop efficient drainage systems. Beyond reconciling previously disconnected observations of modern and ancient ice streams into a single life cycle, the modelling suggests that tunnel valley development may be crucial in stabilizing portions of ice sheets during periods of climate change.

Large-scale integrated subglacial drainage around the former Keewatin Ice Divide, Canada reveals interaction between distributed and channelised systems

“…The Keewatin sector of the LIS exhibits a relatively low frequency of ice streams (Margold et al, 745 2015;Stokes et al, 2016). While this may be partially attributed to the resistant bed of the shield (Clayton et al, 1985;Kamb, 1987;Stokes and Clark, 2003a, b), we also suggest that efficient evacuation of meltwater through the dense channelised network that developed in this region during the final stages of deglaciation, as the climate warmed (Storrar et al, 2014b), would have inhibited the development of fast flow and 750 potentially contributed to the shut-down of existing ice streams (Lelandais et al, 2018). This is consistent with modern observations that link decadal-scale ice-flow decelerations with more pervasive and efficient drainage channelisation driven by increased surface meltwater inputs to the bed (Sole et al, 2013;Tedstone et al, 2014;van de Wal et al, 2015;Davison et al, 2019).…”

Glacial geomorphology of the Neutral Hills Uplands, southeast Alberta, Canada: The process-form imprints of dynamic ice streams and surging ice lobes

“…Patterson, 1997;Margold et al, 2015;Livingstone et al, 2016), such transitory and mobile ice stream behaviour is to be expected, especially given the role subglacial hydrology plays in the spatial and temporal variability of contemporary ice stream operation (e.g., Gray et al, 2005;Peters et al, 2007;Vaughan et al, 2008;Carter et al, 2013;Elsworth and Suckale, 2016;Siegfried et al, 2016). Moreover, modelling experiments by Lelandais et al (2018) have indicated that ice streams can switch off when the drainage capacities of tunnel valleys are capable of suppressing subglacial water pressures. More importantly, in considering apparent ice stream surging, these modelling results demonstrate that ice streams switch on and accelerate in response to the build-up, migration and subsequent marginal outbursts of subglacial water reservoirs.…”

Ribbed bedforms in palaeo-ice streams reveal shear margin positions, lobe shutdown and the interaction of meltwater drainage and ice velocity patterns

“…It seems that imbalances are common leading to ice stream lobes typically being characterized by the alternation of surge, stagnation and retreat phases, and which control the types of landforms that develop at the margin (Evans et al, 1999(Evans et al, , 2008. Surge phases are marked by distributed meltwater discharge, which when it switches to channelized drainage reduces the 105 widespread lubrication of the ice-bed interface, thus stabilizing the lobe through stagnation of its front (Patterson, 1997;Lelandais et al, 2018). Ice lobe retreat tends to follow the peak-phase of ice sheet collapse, characterized by the maximum advance of the lobe (e.g.…”

“…Three-dimensional numerical modelling of bedforms remains a challenging enterprise however, and investigation is far from complete, mostly because the involved components show drastically distinct thermo-dependent and strain-rate dependent rheologies, and therefore temporal 200 scales of activity (Paterson, 1994). To circumvent the challenge of numerically modelling these complex interactions, Lelandais et al (2016Lelandais et al ( , 2018 developed a physical laboratory model able to simulate simultaneously ice flow, subglacial hydrology and subglacial erosion/transport/sedimentation/deformation. This model contributed to better constrain the link between subglacial meltwater drainage and the lifecycle of ice streams, but subglacial bedforms did not arise.…”