Deformation of the Batestown till of the Lake Michigan lobe, Laurentide ice sheet Jason F. THOMASON, * Neal R. IVERSON Department of Geological and Atmospheric Sciences, Iowa State University, Ames, Iowa 50011, USA E-mail: thomason@isgs.uiuc.edu ABSTRACT. Deep, pervasive shear deformation of the bed to high strains (>100) may have been primarily responsible for flow and sediment transport of the Lake Michigan lobe of the Laurentide ice sheet. To test this hypothesis, we sampled at 0.2m increments a basal till from one advance of the lobe (Batestown till) along vertical profiles and measured fabrics due to both anisotropy of magnetic susceptibility and sand-grain preferred orientation. Unlike past fabric studies, interpretations were guided by results of laboratory experiments in which this till was deformed in simple shear to high strains. Fabric strengths indicate that more than half of the till sampled has a <5% probability of having been sheared to moderate strains (7–30). Secular changes in fabric azimuth over the thickness of the till, probably due to changing ice-flow direction as the lobe receded, indicate that the bed accreted with time and that the depth of deformation of the bed did not exceed a few decimeters. Orientations of principal magnetic susceptibilities show that the state of strain was commonly complex, deviating from bed-parallel simple shear. Deformation is inferred to have been focused in shallow, temporally variable patches during till deposition from ice. 1. INTRODUCTION Numerous authors have suggested that soft-bedded glaciers can move primarily by widespread shearing of their beds over thicknesses exceeding several decimeters (e.g. Alley and others, 1987; MacAyeal, 1992; Jenson and others 1995; Boulton, 1996b; Clark and Pollard, 1998; Licciardi and others, 1998). This mechanism of basal motion may help instigate and sustain fast glacier flow, with associated effects on ice-sheet stability and climate (e.g. MacAyeal, 1992; Clark, 1994; Clark and others, 1999). High rates of basal sediment transport (Alley, 1991; Hooke and Elverhøi, 1996; Dowdeswell and Siegert, 1999; Anandakrishnan and others, 2007) and a wide variety of glacial landforms, including some drumlins, end moraines, Rogen moraines and boulder pavements, have been attributed to this style of movement (e.g. Boulton, 1987; Clark, 1991; Hindmarsh, 1998a,b; Johnson and Hansel, 1999). The vast basal till sheets (commonly >10 000 km 2 ) of Wisconsin episode mid-latitude ice sheets are thought by some to have been transported by this mechanism (Alley, 1991; Clark, 1997). Others believe that, although these tills were deformed locally, they were transported largely in ice and deposited through lodgment (e.g. Clayton and others, 1989; Piotrowski and others, 2001), the process whereby debris is released from sliding basal ice and accumulates on the bed. Both processes undoubtedly involve shear of basal sediments, but the bed-deformation model, through its requirement that most basal motion occur by simple shear of the bed, must result in ratios of bed-surface displacement to shearing-bed thickness (bed shear strains) in excess of 100, even for slow glaciers (10 m a –1 ), short occupation times (100 years) and thick shear zones within the bed (5 m). The bed-deformation model also requires that strain in the bed extend to depths greater than those associated with deformation caused by particle ploughing during lodgment. Deformation depths due to ploughing are conservatively one to five times the particle diameter, as indicated by wedge and cone-penetration studies (Baligh, 1972, 1985; Koumoto and Kaku, 1982), and are likely proportional to the coarseness of particles that constitute the bed (Tulaczyk, 1999; Thomason and Iverson, 2008). Thus, given the small volume fraction of cobbles and boulders in many tills, deformation depths due to ploughing and lodgment should not generally exceed a few decimeters. Testing the bed-deformation model using basal tills of the geologic record therefore requires evidence that allows inconsequential shear strains (100) to be distinguished from the higher strains of the model. Additional evidence should distinguish shallow deformation of the bed from deep deformation to high strains. Also, evidence that sheds light on the three-dimensional state of strain in the bed should be sought to try to establish that bed-parallel simple shear dominated strain. Developing definitive testing criteria is difficult. Qualita- tive descriptions of till micromorphological features from many studies (see Menzies and others, 2006, for a review) indicate that basal tills have commonly been deformed, but to strains that are poorly known. Macroscopic heterogene- ities in till can provide indicators of strain direction and magnitude but are expected to be homogenized at the high strains required of the bed-deformation model (e.g. Clark, 1997; Piotrowski and Tulaczyk, 1999; Van der Wateren and others, 2000), so the magnitude of deformation of com- monly massive tills is unclear. The degree of particle mixing at till contacts has been used to demonstrate bed deform- ation (Carlson and others, 2004; Piotrowski and others, 2004) but has not generally yielded quantitative estimates of strain magnitude, despite the potential for such estimates (Hooyer and Iverson, 2000b). Efforts to assign unique particle-fabric signatures to ‘deformation tills’ and ‘lodgment Journal of Glaciology, Vol. 55, No. 189, 2009 *Present address: Illinois State Geological Survey, 615 East Peabody Drive, Champaign, Illinois 61820, USA. 3B2 v8.07j/W 20th January 2009 Article ref 08j026 Typeset by Ann Proof no 1 1