The composition and internal structure of drumlins: Complexity, commonality, and implications for a unifying theory of their formation

Investigation of drumlins is significant to both glaciology and palaeoglaciology but the sheer diversity of their composition and internal structure is often cited as a major obstacle towards a satisfactory (unifying?) explanation of their formation. This paper presents the first systematic survey of the vast literature on this subject, with the aim of concisely summarising observations and identifying any emergent patterns or commonality that theories of drumlin formation should be able to explain. Results confirm that investigations are often limited by availability of suitable sediment exposures (40% of studies report data from <5 drumlins and 44% do not specify sample size), although borehole data and geophysical techniques can alleviate this problem. However, it is clear that the constituents of drumlins are incredibly diverse in terms of their composition (e.g. a range of lithologies, clast shapes, sizes and fabrics); structure (e.g. sediments that are sorted, homogeneous, surface conformable, unconformable); and evidence of deformation (e.g. ranging from pervasive, to non-pervasive/limited, to absent). Despite this diversity, our review leads us to suggest that drumlin composition can be simplified to five basic types: (i), mainly bedrock, (ii), part bedrock/part till; (iii), mainly till; (iv), part till/part sorted sediments; and (v), mainly sorted sediments. This is a potentially significant step, in that it reduces the oft-cited complexity of drumlin composition and provides a more realistic goal for theories or numerical models of drumlin formation to target. The various types can occur within the same drumlin field, which leaves us with two possible implications for drumlin formation. (1) Different types of drumlin are formed/seeded by different processes, despite being morphologically similar (equifinality?) — investigation of drumlin composition may, therefore, reveal diagnostic processes/explanations for these different types of drumlin and we argue that bedrock ‘drumlins’ are an example. (2) A single process occurs across large parts of the ice-bed interface to create drumlinised terrain in a variety of sediments — investigation of drumlin composition may, in this case, simply reflect pre-existing sediments but, importantly, the way in which the drumlin-forming mechanism modifies/is modified by them. We argue that the latter, simpler, explanation applies to all other types of drumlin (excluding purely bedrock forms) and conclude that the diversity in drumlin composition is no obstacle to a single unifying theory.