Contrasting glacial landscapes in Wisconsin

D.M. Mickelson and P.M. Colgan


In the next 2 days you will see landform assemblages left by the Superior, Chippewa, and Green Bay lobes of the southern Laurentide Ice Sheet. The wealth of glacial features in Wisconsin has led to the development of Wisconsin's Ice Age Scientific Reserve, a series of nine parks jointly administered by the Wisconsin Department of Natural Resources and the National Park Service. The purpose of these parks is to preserve and illustrate features produced by glaciation, and it is the first of its kind in the United States. The Ice Age Trail, which follows the outermost moraine of the late Wisconsin glacier throughout most of its length, connects the parks. We will visit 3 or 4 units of the Ice Age Reserve on the field trip. Although not the focus of the trip, we will cross pre-late Wisconsin deposits and border the Driftless Area, so we begin with a brief introduction to those areas.

There is no record of when the earliest advance of ice took place in Wisconsin, but it was probably more than 700,000 years ago. How much earlier is not known, but there is evidence of glaciation in the western plains, as far south as Nebraska and Kansas, by 1.8 million years ago. The conclusion that early glaciations entered Wisconsin more than 700,000 years ago is based on lake and glacial sediment west of Eau Claire that has reversed magnetic polarity of lake sediment and till (Hersey and parts of the Kinnikinnic Members of the Pierce Formation (Baker, et al., 1994).

It was recognized before the turn of the century by many glacial geologists (including T.C.Chamberlin and R.T. Salisbury) that there had been multiple glaciations in the mid-continent region. They suggested (probably based mostly on what was at that time becoming accepted in the European Alps) that there were four glaciations, each separated by a interglacial period. The four glaciations were named the Nebraskan, Kansan, Illinoian, and Wisconsin, after states where deposits of each glaciation were well exposed. We now know that continental ice sheets may have entered this area between 12 and 15 times during the last two million years, and much work remains to be done to interpret the very incomplete stratigraphy.

The Wisconsin Glaciation, was the final--and in some places in southern Wisconsin, the most extensive ice cover of all. Although there is evidence of early Wisconsin ice advances farther east in Ontario, there is little evidence of the extent of isotope stage 3 or 4 ice in the western Great Lakes area. That lack of evidence probably does not mean that glaciers failed to enter the region; rather, almost everywhere earlier deposits probably were mostly eroded and the remnants covered by younger materials laid down by the late Wisconsin glaciers. Ice almost certainly entered the Mississippi River basin because wind blown silt, or loess, associated with glaciation blew off the Mississippi valley and into southwestern Wisconsin during this time (Leigh and Knox, 1994). It was evidently a very cold period, with permafrost and barren, wind-swept tundra plains in northern and central Wisconsin.

Another ice sheet began to advance as early as 27,000 or 28,000 years ago, and probably entered the Wisconsin soon after. Because ice flow was controlled by the shape of the underlying topography, the southern margin of the ice sheet was lobate (Figure 1, inset). The Lake Michigan lobe extended far down the Lake Michigan basin into central Illinois, but westward only a few tens of kilometers into Wisconsin, where it joined the Green Bay lobe. The Green Bay lobe was funneled down the Green Bay and Lake Winnebago lowland to just south and west of Madison, and ice of this lobe actually flowed toward the northwest in Portage, Marathon, and Langlade County, where it met the southwest-flowing Langlade lobe. The interlobate zone between the Lake Michigan and Green Bay lobes (over what is now the Kettle Moraine) was lower than the ice in the axis of the lobes on either side and formed a trough that sloped southwestward toward Walworth County. Our last stop of the trip will be close to the southern Kettle Moraine.

Because of the east-west trend of the > 300 m deep Lake Superior trough and the upland south of Lake Superior, ice flowing from the north was diverted toward the southwest to beyond Minneapolis and toward the southeast into the Langlade and Green Bay lobes (Figure 1, insert). Model results suggest that the ice was very slow to cross Lake Superior into Wisconsin because of calving losses (Cutler, et al., 2002) In fact, this inability of the ice to rapidly cross Lake Superior is probably why the Driftless Area of southwest Wisconsin was never glaciated. The east-west trending Lake Superior basin has always diverted glaciers to the east and west, and allowed only limited ice to cross the highland of the Upper Peninsula and northwest Wisconsin to and flow into the Wisconsin Valley and Chippewa lobes (Figure 1, inset).


Mickelson (et al., 1983) considered the area we will cross in the next two days to be two landform assemblages, but we have combined them here because changes in relief of moraines appears gradational from north to south and streamlined hills dominate the landscape behind the marginal zone.

Ice Marginal Zone

This is a landscape that has usually been mapped as end moraine. The zone is 2 to 20 km wide and consists mostly of moderate to high-relief hummocks in end moraines. In the end moraine this topography stands 5 to 40 m above the surrounding landscape. Both distal and proximal sides of the end moraine have slopes greater than 2%. The internal relief of moraine hummocks is much higher than the broad, smooth moraines seen in Illinois and Iowa earlier in the trip. In most places the moraines are narrower (less than 2 km), but about the same height as moraines in Illinois and Iowa. Erratic boulders are scattered on the moraine surface. In many moraine exposures beds of gravel and till are steeply dipping in the up-ice direction, indicating bulldozing, ice push, and localized thrusting at the margin. Located within the wide hummocky end moraine areas are large flat plateaus of sorted sediments and glaciolacustrine sediment that have been interpreted as ice-walled lake plains (Ham and Attig, 1996; Johnson,1986; Johnson, et al. 1995). These ice-walled lake plains will be seen at stops 1 and 2. Farther south in the Green Bay lobe of Wisconsin, the end moraine is narrower and in many places layers of sand and gravel, supraglacial sediment, and basal till are stacked up in this zone (Johnson et al.,1995; Lundqvist et al., 1987). Some of the gravel in the moraine appears to have been deposited in subglacial tunnels close to the ice margin (Lundqvist et al., 1987; Cutler, et al., 2001).

Large volumes of outwash were deposited as fans from the mouths of tunnel channels. These tunnel channels interrupt the outermost moraine every 5 to 10 km in Wisconsin and Minnesota and it appears that much of the sediment in the outwash plains in front of the ice was delivered by these tunnel channels (Attig et al. 1989; Clayton et al. 1999; Culter et al., 2001). The tunnel channels are about 0.5 to 1 km across where they intersect the moraine. Sometimes a steep sided channel is present, but often it is filled with hummocky sand and gravel, which collapsed as ice melted. Mapping of the channels in the Green Bay lobe indicates that water was driven up 80 m in elevation as it approached the ice margin (Clayton et al., 1999). Some of the sediment within the fans is extremely coarse. Intermediate axes of boulders of up to 2 meters have been described in the proximal part of the fans, grading into coarse to fine sand the distal part of the fan (Cutler et al., 2001).

Subglacial Zone

There is a narrow zone (a few kilometers) between the moraine zone discussed above and the distal side of the drumlin field. In this area there appears to be thin diamicton, probably much of it basal till, that commonly covers palimpsest features. In places where bedrock is close to the surface, bedrock hills with only a thin cover of till dominate (Clayton and Attig,1997; Colgan, 1999). A few eskers are present, as well as hummocky sand and gravel.

The landforms in most of the subglacial zone of Illinois and Iowa (Johnson and Hansel, 1999) are strikingly different than in Wisconsin and adjacent Minnesota. Here most of the surface has been streamlined by iceflow. Drumlins and megaflutes up to 50 m high and as long as 9 km occur throughout this zone. Thousands of drumlins occur in Minnesota and Wisconsin. The composition of most of the drumlins is not known, although some detailed studies have been done in Wisconsin some of the drumlins are composed of sand and gravel (Upham,1894; Alden, 1905; Whittecar and Mickelson, 1980; Stanford and Mickelson,1985). Other drumlins are composed of diamicton, some have bedrock cores, and some contain lake sediment (Alden, 1905; Colgan and Mickelson, 1997). The drumlins appear to be composed to a great extent of material that pre-existed the drumlin-forming phase or was deposited in front of or at the advancing ice margin. This consists of bedrock, sand and gravel, older diamicton, and lake sediment. The drumlins appear to be partly erosional features, although certainly part of their height and length is attributable to the growth beneath the ice by inward movement of sediment and the deposition of diamicton and down ice accretion of eroded material (Boulton, 1987).

Other Observations

Throughout most of Wisconsin, diamicton is fairly thin and in many places the diamicton of the last glacial maximum lies directly on bedrock. Abundant local bedrock seems to have been incorporated into the till. This is distinctly different than the nature of till in Illinois and Iowa and indicates much more local erosion in the marginal zone and just behind it. Striated rock surfaces are present even in the marginal areas, but striated boulder pavements are rare. Many bedrock exposures show indications of plucking and removal of large clasts as well as abrasion. Although there are signs of erosion of the bedrock in the marginal zone, the amount of erosion was probably very limited based on cosmogenic dates of striated surface (Colgan et al.,2002).

In summary, there are major differences between the glacial landforms in northern Wisconsin and to the south in Illinois. The low-relief till plain so common in Illinois, Indiana and Ohio is completely absent, and in its place is a landscape with striking streamlining in the direction of ice flow.


We believe that the following observations are all consistent with an hypothesis that the glacial assemblages in Wisconsin developed because the glacier had a frozen bed in a zone near the margin when ice was at its maximum position (Fig. 1). Much of the volume of moraines in this assemblage was developed by the stacking of basal till slabs on top of layers of sand and gravel and supraglacial sediment (Lundqvist et al. 1993). This indicates that for some time when ice was at its maximum extent, outwash was being generated along the ice margin as well as from tunnel channels. We interpret the tunnel channels as indicators of large discharges of subglacial water, perhaps stored as an extensive subglacial layer and then released during a catastrophic event (Cutler et al., 2001). In the Green Bay lobe they clearly were active when the ice was as its outermost margin because fans head at the distal side of moraines where tunnel channels intersect (Attig et al., 1989). Where they have been studied further west in Minnesota they are interpreted to have formed during ice retreat as well.

The abundance of local material and the prevalence of streamlined forms suggest that erosion was the major process occurring behind the ice margin. Another indication of the prevalence of subglacial erosion here compared to Illinois is the rarity of paleosols in the north. Wood is not present in the LGM age tills in Wisconsin. We believe the absence of wood is an indication that continuous permafrost and tundra conditions existed when ice advanced to its maximum position. There simply was not sufficient wood on the landscape to be preserved. Both outside and inside the outermost margin of the ice in Wisconsin there are indications of former permafrost (Clayton et al., 2001). Patterned ground can be observed on aerial photographs, especially on outwash surfaces created during the LGM (Black, 1976; Johnson et al., 1995; Colgan 1996; Clayton et al., 1997). It also appears that in Wisconsin and Minnesota this permafrost lasted until about 13,000 B.P. in the south and then slowly disappeared northward about 3,000 years later (Clayton, et al., 2001). Not only was permafrost overridden, but it must have reformed as ice retreated between 17,000and 13,000 B.P. At least discontinuous permafrost may have also been present in some areas of Illinois (Johnson, 1990), but we believe this permafrost was not extensive enough to have the same impact as it did in Wisconsin, and that it likely developed along the ice margin after the ice reached its maximum extent.


Stop 1. Turtle Lake ice walled lake plain This stop is in the outer part of the St. Croix moraine, where high hummocks and ice walled lake plains are common (Johnson, 2000). The low section contains foreset beds and offshore sediment common in the ice walled lake plains of the St. Croix moraine where a stream off the ice margin presumable flowed into the standing water. Surrounding hummocks are a mix of hummocks composed of supraglacial sediment and meltout till (Johnson et al., 1995) and ice walled lake plains. How did the hummocks originate? Are they mostly composed of meltout till? Is thick supraglacial sediment necessary for the formation of ice walled lake plains? Can flowing water be involved in eroding the landscape as proposed in other areas?

Stop 2. Ice Age Visitor Center in Chippewa moraine This area is classic, high relief hummocky topography similar to that in the St. Croix moraine. The Chippewa Moraine Visitor's Center was dedicated in fall 1992, and exhibits designed by the National Park Service were installed during summer 1997. The views from here include ice walled lake plains, hummocks composed of diamicton, kettle lakes, and an outwash plain. The well for the Chippewa Moraine Visitor's Center penetrated 86 m of sediment, much of it fine-grained and not good for producing water, and did not encounter bedrock (Syverson, 1998). The Visitor's Center ice-walled-lake plain has two crests. This suggests that two ice-walled lakes coalesced as the ice walls confining the lakes slowly melted back, and then more lacustrine sediment was deposited at lower levels within the new ice-walled lake (Syverson, pers, comm., 2002).

Stop 3. Bob Lake ice walled lake plain. The pit displays well developed deltaic foreset beds and silt drapes, as well as a vertical exposure of reddish-brown, sandy loam Copper Falls Formation diamicton (supraglacial sediment) in the north face of the pit. The diamicton is quite uniform, and Johnson (et al., 1995) proposed that this is supraglacial meltout till formed as the sediment melted out slowly at the glacier surface beneath a thick insulating layer of supraglacial sediment (Syverson,pers.comm, 2002). Is this meltout tilll? Was thick meltout till required for the development of ice walled lake plains?

Stop 4. Johnstown moraine at Devils Lake. We will walk the moraine along its crest. Note the lower internal relief (about 10 m) here than at stops 1 and 2. The glacial history of this area has been recently interpreted by Attig and Clayton (Attig at al., 1990; Clayton and Attig, 1990). If exposure is good we will see melt out till that is steeply dipping up ice, interbedded with sand and gravel that was presumably deposited just in front of the ice margin and overridden. Note the steep dip of the gravel beds and the large amount of variability in the exposed section. The moraine was built by stacking of debris-rich ice, supraglacial sediment and stream sediment during several very minor readvances (Lundqvist et al.,1993). This variability is typical in this part of the Green Bay lobe and contrasts with the uniformity in clay till moraines close to Lake Michigan and with the moraines in Illinois that are made up mostly of basal till. Supraglacial sediment is characterized by an abundance of gabbro and basalt that was derived more than 300 km up ice. In places very coarse tunnel deposits are present surrounded by the meltout till. These are expecially rich in gabbro, probably indicating a supraglacial source for the water (Lundqvist et al, 1993). Was the supraglacial sediment thinner here than at stops 1, 2, and 3? Is this meltout till?

Stop 5. Observatory and Bascom Hills. These drumlins lie within a few kilometers of the outermost edge of the Green Bay lobe drumlin field and about 15 km behind the Johnstown moraine. They were described by Upham (1894) when he visited Madison in 1892 and saw exposures of thin till over sand and gravel. He named sand- and gravel-cored drumlins "Madison type drumlins" based on his observations here. Length-to-width ratios of these hills and the drumlin on which the State Capitol stands are about 1:1, contrasting greatly with the drumlins to be seen at stop 7. The low, flat surface west of Observatory Hill is the former lake bed of Glacial Lake Yahara.

Stop 6. View stop on Highway M drumlin. This drumlin is in the central part of the Green Bay lobe drumlin field. Length -to-width ratios of the drumlins here are well over 10:1. These may be gravel or till cored. We interpret the length of the drumlins to be related to ice flow velocity, and that the ice was slowing down from here toward the ice margin. Is velocity the reason for the differences in the extent of streamlining? If not, it is length of time the area was ice covered?

Stop 7. Waterloo Drumlin.

The exposed part of this well-developed drumlin is composed entirely of Holy Hill Formation diamicton. Note the uniformity of the till and its local source (mostly dolomite) suggesting that it is till deposited at the base of the ice and not let down from a supraglacial position. Was this diamicton deposited in the drumlin form by subglacial deformation or lodgement, or was it deposited during ice advance and then overridden and shaped into the drumlin form? Was it deposited directly beneath the ice as a streamlined feature? Compare with the drumlins at stops 5 and 8 that have a gravel core. The drumlin is oriented nearly north-south and is in the middle of the Green Bay Lobe.

Stop 8. Sullivan drumlins. These low drumlins are among the best to illustrate advance and retreat tills if well exposed (Whittecar and Mickelson, 1979) Note that gravel is folded (overturned) and truncated by an erosion surface and till in the northernmost Sullivan pit. Above the basal till (mostly in the soil) is and abundance of Precambrian sandstone that was probably transported englacially. The Stephan pit is in a poorly streamlined drumlin as is typical of those close to the outer edge of the drumlin field.

Stop 9. Woodward pit. This deposit is ice-contact stratified sand and gravel in a hill that is not streamlined. Note the bedding indicates deposition of sand and gravel as foreset beds into standing water. This formed as ice withdrew from the Kettle Moraine and traces of diamicton near the surface may be mudflow deposits off the ice surface. Alternately, may be a protodrumlin consisting of sediment that pre-existed the drumlin forming advance that was overridden but not streamlined significantly. In this case, the thin diamicton near the top of the section may be basal till analogous to that seen at the top of the section at stop 8. There has not been sufficient exposure to make that determination.


Alden, W.C., 1905. Drumlins of Southeastern Wisconsin. U.S. Geological Survey

Bulletin, 273, 111 pp.

Attig, J.W., Clayton, Lee, Mickelson, D.M.., 1985, Correlation of late Wisconsin glacial phases in the western Great Lakes area: Geological Society of America Bulletin, v. 96, p.1585-1593.

Attig, J.W., Clayton, Lee, Lange, K.I. and Maher, L.J., 1990, Ice age geology of Devils Lake State Park: Wisconsin Geological and Natural History Survey, Educational Series 35, 28 p.

Attig, J.W., Mickelson, D.M., and Clayton, Lee, 1989, Late Wisconsin landform distribution and

glacier-bed conditions in Wisconsin: Sedimentary Geology, v. 62, p. 399-405.

Baker, R.W., 1994, Pre-Wisconsin glacial stratigraphy, chronology, and paleomagnetics of west-central Wisconsin: Geological Society of America Bulletin, v. 94, p. 1442-1449.

Black, R.F., 1976, Quaternary geology of Wisconsin and contiguous Upper Michigan, in Mahaney, W.H., ed., Quaternary Stratigraphy of North America: Stroudesburg,

Pa., Dowden, Hutchinson, and Ross, p. 93-117.

Boulton, G. S., 1987, A theory of drumlin formation by subglacial deformation: In

Menzies, J, and J. Rose (eds.), Drumlin Symposium, Balkema, Rotterdam, p. 25-80.

Clayton, Lee and Attig, J.W., 1990, Geology of Sauk County, Wisconsin: Wisconsin Geological and Natural History Survey, Information Circular 67, 68 pp.

Clayton, Lee, Attig, J.W., Mickelson, D.M., 1997, Conditions around the margin of the Green Bay lobe during the height of the Wisconsin glaciation: In: Mudrey, M.G., Jr., Guide to Field Trips in Wisconsin and adjacent areas of Minnesota, 31st Annual meeting of the North-central section, Geological Society of America: Wisconsin Geological and Natural History Survey, p. 23-30.

Clayton, L, Attig, J.W., and Mickelson, D.M., 1999, Tunnel channels formed in Wisconsin during the last glaciation: In Mickelson, D.M. and Attig, J.A., (Eds.), Glacial Processes Past and Present: Geological Society of America Special Paper 337, p. 69-82.

Clayton, Lee, Attig, J.W., and Mickelson, D.M., 2001, Effects of late Pleistocene permafrost on the landscape of Wisconsin, USA: Boreas, v. 30, p. 173-188.

Colgan, P.M., 1999, Reconstruction of the Green Bay Lobe, Wisconsin, United States, from 26,000 to 13,000 radiocarbon years B.P.:In Mickelson, D.M. and Attig, J.A., (Eds.), Glacial Processes Past and Present: Geological Society of America Special Paper 337, p. 137-150.

Colgan, P.M. and Mickelson, D.M., 1997, Genesis of streamlined landforms and flow history of the Green Bay lobe, Wisconsin, USA: Sedimentary Geology, v. 111, p. 7-25.

Colgan, P.M., Bierman, P.R., Mickelson, D.M., and Caffee, Marc, 2002, Variation in glacial erosion near the southern margin of the Laurentide Ice Sheet, south-central Wisconsin, USA: Implications for cosmogenic dating of glacial terrains: Geological Society of America Bulletin, v. 114, p. 1581-1591.

Cutler, P.M., Mickelson, D.M., Colgan, P.M., MacAyeal, D.R., Parizek, B.R., 2001, Influence of the Great Lakes on the Dynamics of the southern Laurentide Ice Sheet: Numerical Experiments: Geologv, v.29, no. 11, p. 1039-1042.

Cutler, P.M., Colgan, P.M., and Mickelson, D.M., 2002, Sedimentologic evidence for outburst floods from the Laurentide Ice Sheet margin in Wisconsin, USA: implications for tunnel-channel formation: Quaternary International, v. 90, p. 23-40.

Ham, N.R., and Attig, J.W., 1996, Ice wastage and landscape evolution along the southern margin of the Laurentide Ice Sheet, north-central Wisconsin: Boreas, v. 25, p. 171-186.

Johnson, M.D., 1986, Pleistocene geology of Barron County, Wisconsin: Wisconsin Geological and Natural History Survey Information Circular 55.

Johnson, M.D., 2000, Pleistocene geology of Polk County, Wisconsin: Wisconsin Geological and Natural History Survey Bulletin 95.

Johnson, M.D., Mickelson, D.M., Clayton, Lee, and Attig, J.W., 1995, Composition and genesis of glacial hummocks, western Wisconsin: Boreas, v. 24, p. 97-116.

Johnson, W.H., 1990, Ice-wedge casts and relict patterned ground in central Illinois and

their environmental significance: Quaternary Research, v. 33, p. 51-72.

Johnson, W.H. and Hansel, A.K., 1999, Wisconsin episode glacial landscape of central Illinois: A product of subglacial deformation processes?: In Mickelson, D.M. and Attig, J.A., (Eds.), Glacial Processes Past and Present: Geological Society of America Special Paper 337, p. 121-135.

Leigh, D.S. and Knox, J.C., 1994, Loess of the upper Mississippi Valley Driftless Area:

Quaternary Research, v. 42, p.30-40.

Lundqvist, Jan, Clayton, Lee, and Mickelson, D. M, 1993, Deposition of the late Wisconsin Johnstown moraine, southcentral Wisconsin: Quaternary International, v. 18, p. 53-59.

Mickelson, D.M., Clayton, Lee, Fullerton, D.S., and Borns, H.W., Jr., 1983, Late glacial record of the Laurentide Ice Sheet in the United States, In: Late Quaternary Environments of the United States, Wright, H.E., Jr. (Ed.), Volume 1: The Late Pleistocene (S.C. Porter, Ed.), University of Minnesota Press, Minneapolis, p. 3-37.

Stanford, S.D. and Mickelson, D.M., 1985, Till fabric and deformational structures in drumlins near Waukesha, Wisconsin, U.S.A.: Journal of Glaciology, v. 31, no. 109, p. 220-228.

Syverson, K.M., 1998, Glacial geology of the Chippewa Moraine Ice Age National Scientific

Reserve Unit, Chippewa County, Wisconsin, in Syverson, K.M., and Havholm, K.G., eds.,

Geology of Western Wisconsin: Guidebook for 61st Annual Tri-State Geological Field

Conference and University of Wisconsin System Geological Field Conference, p. 57-65.

Syverson, K.M., in preparation, Pleistocene geology of Chippewa County, Wisconsin: Wisconsin

Geological and Natural History Survey Bulletin xx and 1:100,000-scale colored map.

Upham, W., 1894, The Madison type of drumlins: American Geologist, v. 14, p. 69-83.

Whittecar, G.R. and Mickelson, D.M., 1977, Sequence of till deposition and erosion in drumlins:

Boreas, v. 6, p. 213-218.

Whittecar, G.R. and Mickelson,D.M., 1979, Composition, internal structures and an hypothesis of

formation for drumlins, Waukesha County, Wisconsin, U.S.A.: Journal of Glaciology, v. 22, p. 357-372.