Reconstructing subglacial meltwater dynamics from the spatial and temporal variation in the form and pattern of eskers

Abstract

Meltwater drainage beneath glaciers and ice sheets is intimately linked to their dynamics. Meltwater may increase ice velocity if it acts to lubricate the bed; conversely, an efficient subglacial meltwater drainage system may preclude meltwater induced ice acceleration by limiting the amount of water available to facilitate sliding. Thus, understanding the nature of meltwater flow beneath ice masses is crucial for predicting how ice sheets and glaciers will react to increased meltwater input. However, direct observation of subglacial meltwater drainage systems is extremely difficult, meaning that indirect methods such as remote sensing, numerical modelling, dye tracing and geophysical survey are the only way to observe this environment. These methods often suffer from excessive uncertainty and poor spatial and, particularly, temporal resolution.
This thesis presents the results of an alternative approach, using the geomorphological record of eskers to understand the former behaviour of meltwater beneath the Laurentide Ice Sheet (LIS) in Canada, and at Breiðamerkurjӧkull in Iceland. Eskers are elongate, sinuous ridges of glaciofluvial sand and gravel deposited in glacial drainage channels. Despite a large body of research on eskers, no systematic analysis of the large-scale properties of eskers, or the implications this may have for understanding subglacial meltwater, has yet been undertaken. Eskers are mapped at the ice sheet (continental) scale in Canada from 678 Landsat ETM+ images and at high resolution (~30 cm) from 407 aerial photographs of the Breiðamerkurjӧkull foreland, in order to address three outstanding questions: (i) What controls the pattern and morphology of eskers? (ii) How did subglacial drainage systems evolve during ice sheet deglaciation? (iii) How can eskers be used to further our understanding of subglacial hydrology?
Over 20,000 eskers are mapped in Canada, revealing that esker systems are up to 760 km long, and are surprisingly straight. The spacing between eskers is relatively uniform and they exhibit little change in elevation from one end to another. As the LIS deglaciated between 13 cal ka and 7 cal ka, eskers increased in frequency, which is interpreted to represent an increase in meltwater drainage in channelized, rather than distributed, systems. Eskers are abundant over the resistant rocks of the Canadian Shield and also show a strong preference for formation in areas covered with till. Esker length, sinuosity and spacing appear to be unrelated to the underlying geology. Finally, two types of complex esker systems are proposed: esker fan complexes and topographically constrained esker complexes. The formation of esker complexes is dependent on sediment and meltwater supply and the pre-existing topography controls the overall shape of the esker systems.