Testing the instability hypothesis for the formation of drumlins: 2010AGUFM.C13A0539S

Ice flow is largely facilitated by processes at the base of the ice, which are very difficult to observe. Nevertheless, we know from observations of exposed ice sheet beds and, more recently, from beneath the ice in Antarctica, that the processes that generate ice flow result in a suite of ‘subglacial bedforms’. Understanding how these bedforms are created is crucial to understanding the flow of ice sheets. Drumlins are by far the most common subglacial bedform. They are ubiquitous on former ice sheet beds but, despite their importance, over 200 years of investigation (and over 1400 papers!) has failed to generate a consensus with regard to the processes responsible for their formation. Investigations have traditionally been hampered by low sample sizes, such that hypotheses derived from observations at one location are not found in other locations. Moreover, hypotheses or ‘ideas’ have rarely evolved into physically-based numerical models that capture the complexity of the ice-bed interface and are able to make predictions of bedform characteristics. This paper is aimed at addressing these issues through the collection of a large dataset of drumlin characteristics (>50,000), which can be used to test the instability hypothesis (and associated numerical model) of drumlin formation. This hypothesis states that drumlins (and probably other related subglacial bedforms, e.g. ribbed moraine, mega-scale glacial lineations) arise through an instability in the flow of ice over erodible sediment, which is analogous to the mechanism whereby fluvial and aeolian dunes are formed. Drumlin sizes and shapes are well suited to analysis by satellite remote sensing and DEMs and we present the results of several tens of thousands of drumlin measurements (>50,000 individual landforms). Significantly, this large dataset reveals that many of the long-established paradigms regarding drumlin size and shape are questionable. We also address the controversial issue of the composition and internal structure of drumlins: the sheer diversity of their internal constituents is often cited as a major obstacle towards a unifying explanation of their formation, but our analysis indicates that they can be simplified to a few basic types. Finally, we compare our observations of drumlin characteristics to predictions made by the instability model and show that it is able to reproduce the typical dimensions of drumlins and therefore represents one of the most complete and promising explanations of their formation.