The Editor, Journal of Glaciology

SIR,

Subglacial deformation of sediment and the importance of such deformation to the dynamics of large ice sheets is a relatively new and exciting development (Reference Boulton and JonesBoulton and Jones, 1979; Reference Alley, Blankenship, Bentley and RooneyAlley and others, 1986). In a recent paper, Reference BegetBeget (1986) discussed these developments and, based on the strength properties estimated for a diamicton formed by a late Wisconsinan sediment flow in central Illinois, concluded (1) that sediment beneath the outermost part of the Lake Michigan lobe at the time of the sediment flow had similar low-strength properties, (2) sediment was deforming beneath the lobe, and (3) the southern Lake Michigan lobe of the Laurentide ice sheet had a low profile at that time. Although we suspect that some of Begefs conclusions may be valid, we question the premise on which he approached the problem and based much of his discussion, and therefore suggest his conclusions are not justifed.

Beget (1986) used data and observations from a paper by Reference Hester, DuMontelle and GoldthwaitHester and DuMontelle (1971) that described a large sediment-flow deposit in front of the Shelbyville moraine at the terminus of the Lake Michigan lobe ( Fig.1). We agree with the general origin of the sediment-flow deposit as originally interpreted by Hester and DuMontelle and described by Reference BegetBeget (1986). However, it should be noted that Hester and DuMontelle’s description of the diamicton is not adequate to determine whether it resulted from one thick, extensive flow or several thinner, possibly overlapping flows; such a difference has significant implications with respect to the interpretation of the flow’s properties and strength.

Fig.1. Inferred geologic relationships at the terminus of the Lake Michigan lobe at the time of the deposition of the large sediment flow. Modified from Reference BegetBeget (1986).

One aspect of Beget’s discussion that we question is the origin of the debris on the glacier surface. He suggested that it was “derived from subglacial till which had been sheared into the glacier and up to the glacier surface and further on (p. 237) stated “The large volume of the flow till [sediment-flow deposit] suggests it can be considered a representative sample of the basal till, brought to the glacier surface along multiple intraglacial thrusts …”. We believe that the supraglacial debris in the terminus zone was more likely derived from ablation of debris-rich basal ice exposed at the active margin by compressive flow and/or thrusting over stagnant ice and marginal deposits, processes that have been widely documented at modern glacier margins (e.g. Reference BoultonBoulton, 1968; Reference LawsonLawson, 1979). The existence and importance of thrusting as a sediment entrainment and depositional mechanism has never been documented (see discussion in Weertman (1961)). This difference in origin is significant with respect to Beget’s model. If his conclusions regarding the strength of the subglacial till are correct, his proposed “thrusting of till” to the surface cannot be accurate because the material would deform by flow rather than by failure.

Beget calculated an approximate yield strength of 8 kPa for sediment-flow material using the geometry of the sediment-flow deposit, an estimation of its wet density, and an equation developed by Reference JohnsonJohnson (1970) (although we note that Reference Johnson, Rodine, Brunsden and PriorJohnson and Rodine (1984) deleted the analysis developed by Reference JohnsonJohnson (1970) and used by Beget). Because the resulting diamicton was similar in texture and composition to Shelbyville till (the assumed source), he concluded (p. 237) that “The rheological properties of the flow till [sediment-flow deposit] also characterized identical Shelbyville till; shear strength of water-saturated till beneath the southern Lake Michigan lobe therefore was also approximately 8 ± 2 kPa”. Beget then went on to assume that the outer 400 km of the Lake Michigan lobe rested on material of this approximate strength. Observing that glacier profiles would be more or less adjusted to the strength of the subglacial sediment (because it was much weaker than the overlying ice), he calculated a profile for the Lake Michigan lobe that indicated it was much lower (thinner) than modern ice sheets.

The critical question is: does the strength of the material in a sediment flow off the terminus of a glacier (position “a” in Figure 1) tell us much, if anything, about the properties of subglacial sediment and whether deformation is occurring beneath that glacier (position “b” in Figure 1) Beget would have us believe that it does. Certainly, whether subglacial sediment deformation took place beneath Pleistocene glacial lobes in Illinois or elsewhere, and what effect sediment deformation had on ice dynamics, are critical research questions. However, we do not think that Beget’s analysis provides much enlightenment to them.

We agree that the diamicton of the sediment-flow deposit is generally similar to the basal (Shelbyville) till that was being deposited beneath the lobe. However, three points should be considered further in relating the strength of the two materials: (1) the number and size of particles >2 mm; (2) the structure or fabric of the subglacial sediments; and (3) the water content of the sediment flow. With respect to the first, Reference LawsonLawson (1982) has shown that significant down-current textural modification of sediment flows is common in glacigenic environments. For example, larger clasts may remain as a lag in the source region or move as traction load in the lower part of a flow. In their original work on the sediment-flow deposit, Reference Hester, DuMontelle and GoldthwaitHester and DuMontelle (1971) demonstrated that matrix texture (<2 mm) of the deposit and till were similar. However, they reported that the maximum particle size was generally <2.5 cm, and not “erratic boulders” as curiously reported by Beget (1986, p. 236). The absence of larger clasts suggests that the flow may not have had the competence to tranport them. Large clasts (cobbles and a few boulders) have been observed in Shelbyville till by one of us (W.H.J.); they would have had an important effect on both ice flow and subglacial sediment deformation (Reference Boulton, Wright and MoseleyBoulton, 1975, 1987; Reference Brown, Hallet and BoothBrown and others, 1987). Their absence in the sediment flow thus makes it less appropriate as an “analogue” material for Shelbyville till.

Secondly, debris in the supraglacial environment undergoes mixing through re-sedimentation processes (Reference Lawson, Goldthwait and MatschLawson, 1979, in press), and thus primary structure (fabric, joints, soil structure, stratification, etc.) that subglacial sediment (including till) may have had is lost or modified. These structual and stratigraphie characteristics would be significant during initiation and early stages of deformation, and are not taken into consideration by using a sediment flow as an analogous material.

Thirdly, many workers have recognized that water content plays an important role in debris-flow rheology (e.g. Reference HamptonHampton, 1975; Reference LawsonLawson, 1982; Reference Johnson, Rodine, Brunsden and PriorJohnson and Rodine, 1984). Sediment flows in the glacigenic environment exhibit significant variations in flow strength and behavior primarily because their water contents and resulting wet densities vary widely (Reference LawsonLawson, 1982). Reference BegetBeget (1986) assumed a “water-saturated” debris flow with a wet density of 2000 kg/m³, a value that probably is reasonable given the available data. However, along the ice margin there must have been other contemporaneous flows of similar texture with both lesser and greater water contents. Which one, among a continuum of flows of varying strength, is representative of the rheology of the subglacial sediment? Is it just the one flow that Hester and DuMontelle described?

Our most critical questions concern Beget’s extrapolation of the flow’s strength to the subglacial sediment and the resulting implications with respect to subglacial hydrology. For Beget’s analysis to be valid, subglacial sediment beneath the Lake Michigan lobe must have supported essentially no load, i.e. the normal load (the weight of the glacier) was supported by pore-water pressure in the subglacial sediment. Such a situation is required because Beget’s analysis assumes no contribution to strength from internal friction (zero effective stress) and any strength in the subglacial sediment comes only from cohesion.

Effective stress in this case is determined by two major unknowns: the pore-water pressure in the sediment and the thickness of the ice at the location. The latter clearly is in part dependent on the former, which in turn is controlled by several other inter-related factors, i.e. basal thermal regime, local basal melt rates, nature of subglacial drainage (whether in thin water layers, Röthlisberger channels, Nye channels, or subglacial sediment), and the properties of the subglacial sedimentary sequence, particularly those that affect drainage. Pore-water pressure may have approached or equaled glaciostatic pressure as Beget assumed, but the strength of the sediment flow indicates only that it is possible for material similar in texture and composition to Shelbyville till to have low strength, and nothing with respect to subglacial conditions, particularly hydrology, and certainly not to the thickness of the glacial lobe.

Beget offered two independent “field tests” of his proposed low profile for the Lake Michigan lobe, but neither is valid. First, he referred to moraine profiles from Hester and DuMontelle (1971) which “indicate the terminal parts of the Lake Michigan lobe rose to the north at approximately 7.6 m/km” (Beget, 1986, p. 238). Slopes re-ported by Hester and DuMontelle are for the front of the Shelbyville moraine, and have no direct bearing on the gradient of the paleo-ice surface. These slopes cannot be used to support a low profile as Beget did.

Secondly, Beget used an approach originally suggested by Reference FlintFlint (1971, p. 484) to estimate a minimum ice thickness at the center of the Lake Michigan lobe 400 km north of the Shelbyville moraine. This test involves using the elevation (450 m) of marginal deposits along the eastern edge of the Driftless Area in central Wisconsin (not “Michigan”) and the depth of Lake Michigan (−180 m) to obtain a minimum ice thickness of 630 m (Beget, 1986, p. 238). Beget, as did Flint, neglected to mention that these are marginal deposits of the Green Bay lobe, which was located between the Driftless Area and the Lake Michigan lobe. Nevertheless, Beget acknowledged that the medial thickness of the Lake Michigan lobe would have been greater than this minimum value, but he observed that an ice thickness of 835 m predicted for the location from his reconstructed profile “is a reasonably good fit to the field data” (p. 238). In fact, it [630 m] is a minimum value and only that; it is possible that the ice thickness was significantly greater. In addition, the elevation figures do not consider the possibility that differential isostatic adjustments during and following glaciation may have been significant.

We do not deny that subglacial sediment deformation took place during the last glaciation in Illinois (e.g. deformed subglacial sediment and stratigraphie contacts at a locality 80 km from the glacier margin; see Figure 5d and e in Reference Hansel, Johnson and SochaHansel and others, 1987). It is noteworthy that at this same locality, where evidence for subglacial deformation has been observed, there is also evidence that at some times subglacial deformation was not occurring (e.g. relatively undeformed subglacial channel deposits and stratigraphie contacts; see figures 5f, 6a and d, and 7a in Reference Hansel, Johnson and SochaHansel and others, 1987). In addition, one of us (W.H.J.) has observed several (>six) sections within and near the Shelbyville moraine where the A horizon of the Farmdale geosol and/or a moss layer on Morton loess immediately subjacent to or within 15 cm of Shelbyville till have not been deformed (Reference Frye, Glass and WillmanFrye and others, 1962; Reference Johnson, Glass, Gross and MoranJohnson and others, 1971; Reference Follmer, McKay, Lineback and GrossFollmer and others, 1979). Subglacial channel deposits within till in the Shelbyville moraine generally are not deformed. However, the A horizon of a Farmdale geosol at a locality 70 km north of the Shelbyville moraine shows much evidence of internal deformation. We suspect there was considerable temporal and spatial variability in subglacial deformation beneath the Lake Michigan lobe during the last glaciation.

Beget used an apparently novel approach for estimating the Theological properties of till beneath the Lake Michigan lobe during the late Wisconsinan glaciation. We do not question the possibility that the Lake Michigan lobe was thin, nor the probability that deforming sediment played an important role in affecting the dynamics of the lobe. We do not accept Beget’s analysis as being valid, however, and urge caution in using minimal data and such a questionable approach and gross generalization to modeling a large glacial lobe. In particular, it will require careful and detailed study of Lake Michigan lobe deposits and geomorphology before the contribution of subglacial deformation to flow of the lobe will be fully known.

This discussion was written while W.H.J, was a visiting scientist at the U.S. Army Cold Regions Research and Engineering Laboratory, Hanover, New Hampshire. We thank T. Arguden, D. Lawson, K. Prestegaard, and K. Rodolfo for valuable discussions, and D. Lawson and N. Smith for reviews that improved the manuscript. The opinions expressed, however, are our responsibility.