Transient subglacial hydrology of a thin ice sheet: insights from the Chasm esker, British Columbia, Canada

Glacier drainage systems are never steady state; instead they fluctuate over time and space in response to variation in water input. In order to improve numerical models of glacier hydrology it is important to fully understand the processes controlling subglacial drainage system evolution, but contemporary ice sheet beds are typically inaccessible. Thus, esker distribution, morphology, and sedimentology have been used to infer the dynamics and hydrology of former ice sheets. However, debate remains as to the processes responsible for esker formation and most theoretical investigations have assumed that they formed due to processes that operated beneath thick ice despite many field investigations to the contrary. We investigate esker formation during the final stages of the thin, inactive, and rapidly decaying Cordilleran Ice Sheet (CIS) in interior British Columbia, Canada. A combination of geomorphological, ground-penetrating radar (GPR) and electrical resistivity tomography (ERT) data suggest esker ridge sedimentary architecture is consistent with synchronous formation during a glacial lake outburst flood (GLOF). These data reveal esker ridge deposition most likely took place within an ice tunnel that evacuated late-waning stage flow, following erosion and partial fill of a broader meltwater corridor. Esker ridge sedimentary architecture reveals this depositional environment was dynamic, reflecting complex interaction between ice thickness, ice structure, ice tunnel geometry, flow conditions, and sediment supply. Under these thin, inactive ice conditions ice tunnel location was initially governed by structural weaknesses in the ice and/or equipotential gradient. Because creep closure rates were low, the ice tunnel evolved through feedbacks between conduit growth via frictional melting/mechanical ice excavation, and conduit closure due to sediment infilling, rather than ice creep. This resulted in a non-uniform ice tunnel that enlarged in an upglacier direction and locally unroofed to form open channels during its final stages. These data suggest that eskers may record unsteady, rather than steady flows in a drainage network and so the assumptions often made when using eskers to inform hydrological components of ice sheet models may not be applicable to all ice sheet settings. (C) 2012 Elsevier Ltd. All rights reserved.