The origin of the name “Turtle Mountain” has never been definitely explained.
Between 1810 and 1870, Métis hunters from the Red River area followed trails north and south of the feature, to reach the buffalo herds. When viewed from the south, the upland appeared to the Métis as a turtle on the horizon with the head pointing westward and the tail to the east. Another account says that the feature was named for an Ojibwa Indian, “Makinak,” (turtle) who walked (ran?) its entire length in one day. The Ojibwa often took their names from things in nature, and the turtle was an important figure in their religious tradition. Other names that have referred to Turtle Mountain include Makinak Wudjiw, LaMontagne Torchue (French for ‘Turtle Mountain’), Turtle Hill, Beckoning Hills, and the Blue Jewel of the Plain.
Still another possible origin for the name might be the painted turtles, which are plentiful in the area today. The only “semi-official” information I could find that referred to the origin of the name was included in the early accounts of government cartographers, who noted that, from a distance, the profile of the plateau resembles the back of a turtle. Patrick Gourneau, in his book History of the Turtle Mountain Chippewa explained why “Turtle Mountain” is not “Turtle Mountains.” He states the following: “The naming of Turtle Mountain goes back a long time, versions from white men and Indians. To mention only three Chippeway versions, it indicates that it was the early Chippeway migrants from the woodlands of the east who named it Turtle Mountain. None of the three versions carry the name Turtle Mountains. As far back as my memory goes, I have not ever heard a full blood term the hills as Turtle Mountains, and same applies to the “Mechifs.” The Chippeway name is “Mekinauk Wudjiw” (Turtle Mountain). If it was Turtle Mountains it would be “Mekinauk Wudjiw wum” (plural). The “Mechifs” referred to the hills as “La Montagne Torchue.” “La Montagne Torchue” is French meaning Turtle Mountain. — quoted in Trail of Misgivings by Daniel F. Jerome, 2006, 280 p.
The name “Turtle Mountain” has been misused for so long, that it has become common practice to use the plural or the singular interchangeably to designate the area. I will use the singular form in this article.
For many people, mention of Turtle Mountain brings to mind the International Peace Garden, which straddles an area of three and a half square miles on the U.S.- Canada (North Dakota – Manitoba) border. North Dakota, after all, is known as the Peace Garden State. The Peace Garden was established in 1932 as a symbol of the peaceful relationship between the two nations. The North Dakota portion of the Peace Garden is in Rolette County on the west side of U.S. Highway 281.
Turtle Mountain rises 600 to 800 feet above its surroundings, high enough to receive significantly more precipitation than the surrounding grassland. As a result of the heavier precipitation, Turtle Mountain is forested. The hills cover an area of about a thousand square miles, half in North Dakota, and half in Manitoba. Along with river bottom land and the forested Pembina Hills to the east, Turtle Mountain is one of the few extensive wooded areas in the region. The predominant covering of aspen is interspersed with black poplar, ash, birch, box elder, elm, and bur oak. A large part of the vegetation consists of shrubs like hazel, chokecherry, saskatoon, nanny berry, dogwood, highbush cranberry (pembina), and pincherry. Fire played an important role in the development of present-day vegetation. Prior to settlement, Turtle Mountain was periodically swept by fire caused by lightning and by human activity. Plains Indians recognized that a heavy growth of new plants appeared in burned areas. They knew too that forests did not attract bison, an important food source, so they routinely set fire to the wooded areas. Prairie winds then carried the fires for many miles. This practice may represent one of the earlier attempts by humans to attract animals by manipulating the environment.
Turtle Mountain is basically an erosional feature, a broad area, resulting when younger sediments were left standing when the surrounding older materials were eroded away. Unlike the Killdeers, though, Turtle Mountain was then glaciated and the resulting glacial landforms greatly changed the area. Had the area not been glaciated, Turtle Mountain might be more like the Killdeer Mountains, although much broader and probably not so prominent a feature. The area of Turtle Mountain is underlain by rocks of the Cretaceous Fox Hills and Hell Creek formations and the Paleocene Cannonball Formation, all covered by a thick layer of glacial sediment. In early Pliocene or earliest Miocene time, five or six million years ago, the area that is now Turtle Mountain was part of a broad, northeast-sloping plain. Rivers and streams flowed over the plain from the west and southwest, making their way to Hudson Bay. Then, in Pliocene time, maybe four million years ago, erosion increased markedly and large amounts of material were removed as deep valleys dissected the plain. I am unsure why this cycle of erosion began. Perhaps the area was uplifted by geologic forces so that streams began to cut down and into the sediments over which they had been flowing or (more likely) the climate may have changed.
The erosion removed sediment and shaped new hills and valleys. Gradually, as streams carried the sediments surrounding Turtle Mountain away to Hudson Bay, a large mesa, or perhaps a range of buttes, remained where the hills that comprise Turtle Mountain stand today. The reason the outlier developed where it did is not clear. The uppermost bedrock unit (beneath the covering of glacial sediment) of Turtle Mountain is the Tertiary Cannonball Formation, which is not notably resistant to erosion. It is possible that some kind of resistant layer was present throughout much of the erosion cycle, perhaps a part of the lower Bullion Creek Formation. Additional drilling in the area may eventually penetrate a remnant of some resistant material that has not yet been found. If any resistant layer exists, it is everywhere buried beneath glacial sediments.
About three million years ago, the climate turned colder and, as snow built up to great depths near Hudson Bay, glaciers formed and the ice flowed southward, out of Canada into North Dakota. As the climate fluctuated during the Ice Age, glaciers advanced and receded, flowing over and around Turtle Mountain several times. About 25,000 years ago, the Late Wisconsinan glacier flowed southward over Turtle Mountain for the last time. During the most recent major glaciation, Turtle Mountain was continuously buried beneath the actively moving glacial ice for about 10,000 years.
The movement of the glacial ice over the obstruction formed by the Turtle Mountain upland caused the ice to become compressed, resulting in shearing within the glacier, especially on the west and north sides of the area. The shearing of the ice at the edge of Turtle Mountain caused large amounts of rock and sediment to be incorporated into the ice. As the climate moderated between 15,000 and 13,000 years ago, the glacier became thinner and its margin receded northward. Because Turtle Mountain rises 600 to 800 feet above the surrounding area, and because ice 200 or 300 feet thick can flow under its own weight, the flow of glacial ice on the lowland adjacent to Turtle Mountain continued for a while. At the same time, the glacier on top of Turtle Mountain stagnated, leaving several hundred feet of debris-covered ice covering the surface on the upland.
In areas surrounding Turtle Mountain, where shearing of material into the glacier had not been as intense, the ice was cleaner and it simply melted away, leaving only a small amount of sediment. In contrast, as the debris-covered stagnant ice over Turtle Mountain melted, the debris it contained gradually became concentrated at the surface of the ice, resulting in an increasingly thick insulating layer that greatly retarded the rate of melting. Thus, even though the glacier had stopped flowing, and had stagnated over the Turtle Mountain upland by 13,000 years ago, the layer of insulation that built up on top of the stagnant glacier kept it from completely melting for another 3,000 years. It was not until about 10,000 years ago that the last glacial ice on Turtle Mountain melted.
The glacial sediment on the stagnant glacier covering the Turtle Mountain upland was irregularly distributed and, for this reason, the ice there melted unevenly. This uneven melting caused the upper surface of the stagnant ice to become hilly and pitted with irregular depressions. The glacial sediment on and within the ice was saturated with water from the melting ice and it was highly fluid. It slid down the ice slopes in the form of mud flows and filled the depressions. Thick accumulations of debris in depressions on the stagnant glacier insulated the ice beneath, keeping it from melting quickly. Newly exposed ice, from which the insulating debris cover had recently slid, melted rapidly. The result was a continual reshaping of the surface of the stagnant, sediment-covered glacier.
The environment over Turtle Mountain gradually stabilized and the lakes flooding the sediment-lined depressions on the stagnant glacier became more temperate. Most of the water in the lakes came from local precipitation, rather than from melting ice. Precipitation at the time was greater than it is today; probably more than 50 inches a year, and the mean annual temperature was a few degrees cooler than it is today. Eventually, all the stagnant ice over Turtle Mountain melted, and all of the material that had been on top of and within the glacier was distributed in its current position, forming the hilly “collapse” topography found in the area today. These landforms are referred to by geologists as “hummocky collapsed glacial topography,” or “dead-ice moraine.” The modern landscape on Turtle Mountain is characterized by hundreds of lakes and ponds, by hummocky topography, and also by some broad, flat areas that stand well above the surrounding rougher land, along with some flat, lowland areas. Many of the higher flat areas are old lake plains, underlain by silt and clay that were once surrounded by glacial ice. These areas are referred to as “elevated lake plains.” Some of the lower flat areas are covered by stream deposits of gravel and sand. No streams flow for any great distance throughout the area.
Geologists tend to concoct unusual names for the things they study. “Dead-ice moraine” may sound odd to some of you. It’s a name for a kind of landform found in parts of North Dakota. Dead-ice moraine sounds odd enough, but can you believe it is found along with things called “doughnuts” and “puckered lips”? First of all, the word “moraine” is an 18th century French word. It was coined by Horace de Saussure to refer to “a heap of earth or stony debris” (de Saussure did not initially realize he was referring to glacial deposits). I’ll explain the “dead” part of dead-ice moraine later.
Dead-ice moraine is also referred to as “hummocky collapsed glacial topography” or “stagnation moraine.” It has irregular topography, formed as the last glaciers were melting at the end of the Ice Age, between about 12,000 and 9,000 years ago. The most extensive area of dead-ice moraine in North Dakota is found on the Missouri Coteau, which extends from the northwest corner to the south-central part of the state (coteau is French for “little hill”) Other extensive areas of dead-ice moraine are Turtle Mountain in north-central North Dakota and the Prairie Coteau in the southeast corner of the state near Lidgerwood. All three areas are uplands that stand above the nearby lower land. The landforms on Turtle Mountain are identical to those on the Missouri Coteau and Prairie Coteau, but Turtle Mountain has a woodland cover, the result of several inches more annual precipitation than the other areas.
North Dakota’s areas of dead-ice moraine generally make for poor farmland as they are rough and bouldery. They do, however, include a lot of excellent rangeland and thousands of depressions, which may contain lakes, ponds, and sloughs known as prairie potholes . The dead-ice moraine of the Missouri Coteau is known as the prairie-pothole region (the so-called North Dakota “duck factory”). The dead-ice moraine is essentially undrained, except locally. No rivers or streams flow for any appreciable distance in any of the three dead-ice regions – Turtle Mountain, the Missouri Coteau, or the Prairie Coteau. Dead-ice moraine formed when glaciers advanced against and over steep escarpments as they flowed onto the three upland areas. The land rises as much as 650 feet in little more than a mile along parts of the Missouri Escarpment, which marks the eastern and northeastern edge of the Missouri Coteau. Similar prominent escarpments border the Prairie Coteau and Turtle Mountain in North Dakota and Manitoba, especially the west side of Turtle Mountain near Carbury. When the glaciers advanced over these escarpments, the internal stress resulted in shearing in the ice. The shearing brought large amounts of rock and sediment from beneath the glacier into the ice and to its surface.
Eventually, as the Ice Age climate moderated, the glaciers became thinner and could no longer flow over higher land, although they kept flowing through lower areas. When the ice on the uplands became detached from still-flowing ice, the glaciers on the uplands stopped advancing and stagnated (or “died”). As the stagnant glacial ice melted, large amounts of sediment that had been dispersed through the ice gradually accumulated on top of the ice, which was several hundred feet thick. The thick covering of sediment on the stagnant glacier helped to insulate the underlying ice, helping to preserve it and prolonging the time it took to melt. As a result, it took several thousand years for the ice to melt. Geologists have determined that insulated, stagnant glacial ice continued to exist on the Turtle Mountain and Missouri Coteau uplands until about 9,000 years ago, nearly 3,000 years after actively moving glaciers had disappeared from North Dakota.
In places where the debris on top of the ice was thickest, the glacier was slowest to melt. If little or no insulating debris covered the glacial ice, melting was quicker and the ice had entirely melted away by 12,000 years ago.
As the stagnant ice on the uplands slowly melted, the glacier surface became more and more irregular. The soupy debris on top of the ice continually slumped and slid, flowing into lower areas, eventually shaping the hummocky, collapsed glacial topography – dead-ice moraine – found today over the uplands. As the stagnant glacial ice melted, and debris slid from higher to lower places, a variety of unusual features resulted. Long ridges formed when sediment slid into cracks in the ice. Such ridges may be straight or irregular, depending on the shape of the cracks. Often, cracks that formed in a rectilinear pattern when the glacial ice was disintegrating, became partly filled with debris that slid into them. Today, we see nearly straight, intersecting ridges, where the ice cracks had been. These ridges are called “disintegration ridges.” Mounds of material collected in holes and depressions in the ice. If the mounds were cored by ice, when the ice cores melted, the centers of the mounds collapsed, forming circular-shaped ridges – “doughnuts.” Some of the doughnuts are breached on two sides because the debris cover on a mound of ice slid off two sides of the mound. Some geologists have referred to such features as “puckered lips.” Wherever part of the covering of debris slid off an area of ice to a lower place, the newly exposed ice then melted more quickly transforming what had been a hill into a hole or depression. Such reversals of topography continued until all the ice had eventually melted.
The insulating blanket of debris on top of a stagnant glacier was so thick in places that the cold temperatures of the ice had little or no effect on the surface of the ground. Trees, grasses, and animals lived on the land surface overlying the stagnant glacial ice. As conditions gradually stabilized, water collected in lakes in depressions on the debris-covered glacial ice. Most of the water in the lakes was probably the result of local precipitation rather than from melting ice. Precipitation at the time was greater than it is today, probably 50 or 75 or more inches of rainfall a year. The mean annual temperature was only a few degrees cooler than it is today.
Surrounding the ponds and lakes, the debris on top of a stagnant glacier was forested by spruce, tamarack, birch, and poplar, as well as aquatic mosses and other vegetation, much like parts of northern Minnesota today. This stagnant-ice environment in North Dakota, 10,000 years ago, was in many ways similar to stagnant, sediment-covered glaciers in parts of Alaska today. Fish, clams, and other animals and plants thrived in the numerous lakes. Wooly mammoths, bear, caribou, wapiti, and other large game roamed the broad areas of forested, debris-covered ice.
During the years I was mapping North Dakota geology, I occasionally came across Ice Age fossils in North Dakota’s dead-ice moraine: caribou bones, mammoth teeth, fossil fish (mainly perch), and various kinds of snails, but paleontologists studying the Ice Age fauna and flora in detail have found many more kinds of Ice Age fossils than I noticed.
Prehistoric people probably lived on the insulated glaciers in North Dakota 10,000 years ago without realizing the ice lay only a few feet below. Or, if they did realize it, they likely accepted it as a normal situation (and I suppose it was normal for that time). Eventually, all the buried ice melted, and all the materials on top of the glacier were lowered to their present position, resulting in the hilly areas of dead-ice moraine we see today.
Every summer, even during the coldest part of the Ice Age, some melting took place on a glacier’s surface and along its margin. Melting occurred during each summer season – even more when the climate warmed for periods of several years at a time – 20 or 30 years – periods of time comparable to the kinds of swings we see in North Dakota’s climate today. The farther south the glacier advanced, into more temperate zones, the more the amount of melting challenged the health of the glacier in those areas until a balance was finally struck between 1) the rate at which the glacial was ice advancing, 2) warmer climates to the south, and 3) overall climate warming due to the approaching end of the Ice Age. Gradually, the balance among these three factors shifted farther north (and east) and ice began to disappear in those parts of North Dakota that were glaciated.
The position of the edge of an ice sheet at any given time was determined by the balance between melting and the rate at which the glacier was flowing. While the climate remained cold (at average annual temperatures below freezing), a continental ice mass became thicker and the edge of a glacier advanced. When it warmed a little, perhaps with average temperatures a bit cooler than those we have now, the glacier margin melted back about as fast as new ice could be supplied. Even though the glacier was moving, its edge neither advanced nor receded (but melting was taking place on the glacier surface). Given still warmer conditions, the surface of a glacier melted more rapidly; the ice thinned, and the glacier’s edge melted back faster than new ice was being supplied. Areas the ice had covered gradually became deglaciated.
As the margin of a glacier melted, debris that had been frozen into the ice many miles to the north was freed and deposited on the ground. This “glacial sediment” consisted of a blended sampling of the various kinds of rock and sediment over which the glacier had flowed. Glaciers advancing into North Dakota from the northeast deposited mainly sandy, granite-rich materials they had picked up as they flowed over the Precambrian rocks of northwest Ontario. Ice coming from the northwest brought a mixture of sandy and clayey materials it had accumulated as it flowed over broad expanses of Cretaceous shale and Tertiary sands of southern Alberta and Saskatchewan. Ice advancing straight southward from Manitoba, up the Red River Valley, deposited carbonate-rich sediment it had picked up north of Winnipeg where Paleozoic limestone and dolostone are exposed today. If you travel north of Winnipeg to Stony Mountain or Stonewall, Manitoba, be sure to notice the quarries, now producing some of the same materials that glaciers brought to North Dakota, perhaps 17,000 years ago. Studying the composition of the glacial sediment is one way that geologists can determine the direction from which the ice came, and the kind of land it flowed over.
Eventually, as each glacier melted (North Dakota was probably glaciated between 10 and 20 times during the past three million years), the land gradually became free of ice. No reversal of ice flow is involved when the glacier recedes; I emphasize that “retreat” of a glacier refers to the melting of the ice. Three different kinds of ice wasting occurred, at different times and places in North Dakota. The first occurred when the glacier margin may have been far to the south, in Iowa and South Dakota. The result of melting, perhaps over a period of relative warmth of hundreds or thousands of years, resulted in the loss of much of the ice mass off the glacier’s surface. The “view from above” in North Dakota would not have changed much – everything was still all ice – but the thickness of glacial ice covering the land was diminished in thickness by hundreds or thousands of feet, even before the glacier margin had receded into North Dakota.
The second way a glacier wasted (at least from our North Dakota perspective) was when the ice margin was nearby. When that happened, wasting involved frequent change in the position of the edge of the active glacier. As a glacier melted, and after it had become thinner, its active margin gradually receded because the volume of ice arriving was insufficient to replace the ice lost at the edge due to melting. Shrinkage of this kind caused the ice margin to melt back, sometimes in a step-like fashion, the flow of ice pausing long enough at times for the forward movement of the glacier to deliver piles of sediment (moraines) to the receding ice margin. A year of glacial activity might involve the margin moving forward a short distance during a winter; then, during the following summer, the margin receding a slightly greater distance. During this phase, one of the most important things taking place, at least during summer seasons, was the deposition of large amounts of gravel and sand being deposited in front of the glacier by water flowing from the melting, sediment-laden ice. The net effect of this second phase of glacial melting was deglaciation; land that had been covered by ice saw the light of day again, after about 20,000 years.
The third way a glacier wasted involved large-scale stoppage of ice movement, leaving large parts of the glacier stranded, sometimes over broad, mainly upland areas, detached from the main body of still-actively-flowing ice on surrounding lowlands (the plains surrounding Turtle Mountain, for example). In North Dakota, this was important over upland places like the Missouri Coteau and Turtle Mountain. Areas of “stagnant,” or “dead” ice on the uplands then continued to melt slowly. Landforms resulting from the melting of such stagnant ice are distinctive and much different from those that were constructed during the step-wise retreat of active glacial ice I described earlier.
Much of North Dakota’s modern landscape reflects its latest encounter with glaciers during the Ice Age. While glaciers flowed into and over the state, carrying the pulverized rock and soil debris they had picked up along their routes, they sheared off old bedrock landforms, smeared on new layers of sediment, and built new landforms. They filled old river valleys with sediment at the same time rivers of meltwater were flowing from the glaciers carving new valleys. In some places, the glacial ice forced existing rivers to follow different routes; in other places it completely obliterated and concealed what had been rivers and valleys. Cold winds blowing over sand that had earlier been deposited on floodplains and in lakes built dunes and spread a veneer of silt (loess) over much of the state.
Most of the sediment associated with the action of glaciers of the most-recent glaciation is soft. It is “unconsolidated,” and does not hold together well (you can dig it with a shovel). An exception: earlier glaciers also deposited sediment. Nearly all of this earlier sediment has eroded away, but in those places where we have found it exposed, or drilled into it, it may be cemented. A jackhammer may be more appropriate than a shovel for digging in such cemented deposits. However, the softer, looser materials that form most North Dakota glacial deposits are much more common. Sediments related to glaciation in North Dakota can be grouped into three main types: till, lake sediment, and outwash.
1. Till was deposited directly from the ice, mostly in the form of mud flows, which slumped or flowed into their current position as the ice melted. Till consists of silty, sandy, pebbly clay, as well as cobbles, or even large boulders.
2. Lake sediment is layered material that was deposited in lakes, which formed on and near the glacier. Such sediment consists mainly of layers of fine-grained silt and clay, deposited on lake floors, along with some sand and gravel, which collected as beaches along the shores of lakes , many of which were dammed by glaciers.
3. Outwash consists of material deposited by running water. Some outwash may be cemented into a kind of stony concrete, but most of it is loose sand and gravel that was washed out of the melting glacier (hence the name “outwash”). Outwash was deposited by streams and rivers flowing through meltwater valleys or as broad, often nearly level sheets of sand ahead of a melting glacier.
Where they are present, sediments deposited directly by glaciers, and by wind and water associated with glaciation, form a thick covering on top of much of the preglacial (bedrock) surface. In central North Dakota, in Sheridan County, the glacial sediment is over 700 feet thick in places. Ten miles northwest of Tolley in northwestern Renville County, it is at least 800 feet thick, the thickest I can document in the state. The amount and thickness of glacial sediment can vary considerably over short distances so it is likely that even thicker deposits than I mentioned exist in places. Over much of the glaciated part of the state, the glacial materials average 150 feet thick.
At any given location, the glacial deposits may consist of two or more layers of till, interbedded with lake beds, alluvial sediments or other materials. In some places, soils, which had developed on the surface of an earlier glacial, river, lake, or wind-blown deposit, were buried when a new layer of glacial material was deposited. These old, buried soils (paleosols) were formed during long intervals of weathering and exposure, like the one we are enjoying now. Paleosols are among the best indicators in the geological record for multiple episodes of glaciation during the Ice Age. The characteristics of a paleosol also help us understand the climatic conditions (forest or grassland, wet or dry, cool or warm, etc.) at the time it formed.
The best places to see several multiple layers of glacial deposits in North Dakota are near Riverdale, at the Wolf Creek inlet to Lake Sakakawea in McLean County and in Beulah Bay, about 15 miles north of the town of Beulah in Mercer County. In both locations, two and, in some places, three discrete till units, separated by cemented gravel layers or paleosols, are being eroded by waves along the lake.
Drill-hole data in eastern and southeastern North Dakota provide evidence that at least a half dozen glacial advances have occurred there since the Ice Age began. Parts of southwestern North Dakota were glaciated during some of the earlier glaciations, but (apart from some rare exceptions) glacial landforms are not found there today because they were eroded away long ago. Glacial lake sediments and river gravels containing glacially derived materials can be found as far southwest as Dickinson and near the Killdeer Mountains, and I have tentatively identified patches of hard, cemented till and glacial river gravels near Bowman and Rhame, places usually considered never to have been glaciated.
Modern soils are an important link to our geologic past. Fresh glacial deposits consist of a mixture of materials, and, because their sources are so varied, they provide the combination of nutrients necessary for fertile soil. In the glaciated part of the state, North Dakota’s soils consist of the weathered exterior of materials left by glacial action. In the thousands of years that have elapsed since the ice sheets disappeared, constantly changing climate, physical and chemical weathering, accumulation of prairie and woodland plant litter, development of root systems, and burrowing activity by organisms have all contributed to the transformation of glacial deposits into the rich soils that form the basis for much of our agricultural wealth.
Agreement on the origin of the name of the “Missouri” River is difficult because too many contradictory explanations exist. The name apparently comes from a Siouan Indian word, “ouemessourita” or “emissourita,” translated by early French explorers as “those who have wooden dugout canoes,” or “river of the large canoes,” or “town of the large canoes,” or any of several other possibilities. One source says the term was the name the Illinois Indians used for the native people who lived in the Mississippi River Valley, probably mainly on the eastern (Illinois) side of the river; another source says they lived in what is now the State of Missouri.
The Missouri River originates near Three Forks, Montana, where the Gallatin, Jefferson, and Madison rivers come together. It flows 2,341 miles to St. Louis, where it joins the Mississippi River. This makes the combined Missouri-Mississippi River, at 3,709 miles, the fourth longest river in the world, after the Nile, Amazon, and Yangtze. The entire length was once riverine environment but, due to the dams that have been built along its route, approximately a third of the length is now reservoirs – lake environment rather than river. Listed from upstream to downstream, the dams are: Fort Peck in Montana, Garrison in North Dakota, Oahe, Big Bend, and Fort Randall in South Dakota, and Gavins Point on the South Dakota-Nebraska border.
Along with its valley, the Missouri River is largely a product of glaciation. Before North America was glaciated, all the drainage in North and South Dakota, eastern Montana, and northern Minnesota was north or northeastward into Canada. There was no “Missouri River” carrying drainage from the northern mid-continent region to the Gulf of Mexico. The way I define the Missouri River requires that its water ultimately reach the Gulf of Mexico as it does today, and that it carry water draining from the Rocky Mountains and northern Great Plains. Prior to glaciation, no such river existed. Why is the situation today so different than it was before the Ice Age?
The modern Missouri River Valley in North Dakota consists of several discrete valley segments that differ markedly from one another. Some of the segments are broad: six to twelve miles wide from edge to edge, with gentle slopes from the adjacent upland to the valley floor. Others are narrow: less than two miles wide, with rugged valley sides – even badlands slopes in places. Most of the wide segments trend from west to east whereas the narrow segments are mainly north-south. The Bismarck-Mandan area is one of only a few exceptions, and I’ll explain why shortly.
The west-east segments of the Missouri River Valley are wide because they coincide with much older valleys that existed long before the area was glaciated. Old, mature river valleys, which formed over long periods of time (hundreds of thousands or millions of years), tend to be broad with gentle slopes. Younger valleys formed more quickly (tens of thousands of years), and are usually narrower with steeper sides. An example of a wide segment is the forty-mile-long, west-east segment of the Missouri River Valley upstream from Garrison Dam. This part of the valley, now flooded by Lake Sakakawea, was once the route of a river that flowed east to Riverdale. However, the river didn’t turn south at Riverdale, as it does today. Rather, it continued eastward past Riverdale, and on past Turtle Lake and Mercer, flowing into northeastern North Dakota. For convenience, I’ll refer to this ancient river as the “McLean River.”
East of U. S. Highway 83, the route of the old McLean River valley is a broad, low area, partly buried beneath tens to hundreds of feet of glacial sediment. Lake Audubon, Turtle Lake, Lake Brekken, Lake Holmes, Lake Williams, Lake Peterson, Pelican Lake, Blue Lake, Brush Lake, and other smaller lakes mark the former route of the McLean River through eastern McLean County. However, continuing farther east, the McLean River valley becomes so deeply buried beneath glacial deposits that it would be nearly impossible to know its route from a study of the surface topography. Fortunately, hundreds of test holes were drilled during studies of the ground water resources of the glacial deposits so we have a good idea of the route the river followed into northeastern North Dakota.
Another wide, west-east trending segment of the modern Missouri River Valley, between Stanton and Washburn, is an eastward continuation of the modern Knife River. Prior to glaciation, the Knife River flowed east in its modern valley to Stanton, but it continued eastward from there, past Washburn. A few miles east of Washburn it turned slightly northeastward. The ancient Knife River joined the McLean River near the town of Mercer and the combined Knife-McLean River continued northeastward to the Devils Lake area. It then flowed north along the east side of Turtle Mountain area into Canada.
Still another wide segment of the Missouri River Valley in northwestern North Dakota extends from near the modern Missouri River /Yellowstone River confluence, northeastward to Williston.
This six-to-eight-mile-wide section of the valley coincides with the pre-glacial route of the Yellowstone River through that area. Prior to glaciation, the Yellowstone River continued to the north, past Williston, following a route that is now mainly buried. The pre-glacial route coincides with the modern route of the Little Muddy River as far as Zahl, about 30 miles north of Williston. North of Zahl, the old Yellowstone River valley into Canada is so deeply buried that its route is known only through drill-hole data. The river entered Saskatchewan about six miles north of Crosby.
The Missouri River Valley between Williston and New Town, now flooded by Lake Sakakawea, follows the same route as did an east-flowing, mid-Ice Age — but probably not pre-glacial – river. This part of the Missouri River Valley is somewhat narrower than most other east-west segments of the valley in North Dakota, and it is also younger than most of them. It is a continuation of a mid-Ice Age river that flowed east from Montana. In Montana, the route of this river coincides with the modern route of the Missouri River past Wolf Point, Poplar, and Culbertson. The Montana segment of the mid-Ice Age river joined the north-flowing Yellowstone River near Buford.
At Bismarck-Mandan, the Missouri River Valley is about two miles wide at the Interstate Highway 94 crossing, but on the south side of Bismarck the valley broadens to six miles wide. The widening southward seems contrary to my earlier comment that north-south segments of the valley tend to be narrow. There is a reason for this exception though. The valley widens at Bismarck-Mandan because, prior to glaciation, the Heart and Little Heart rivers, which today flow into the Missouri River, joined a few miles east of Bismarck. The combined (preglacial) Heart/Little Heart River continued flowing eastward, joining the Cannonball River in southern Burleigh County, near Moffit. The old, combined Heart/Little Heart valley still exists as a broad lowland south and southeast of Bismarck. It is now a wide spot in the Missouri River Valley.
The Heart/Little Heart river system was probably dammed several times by glacial ice advancing westward. Each time a glacier advanced, a lake formed ahead of – west of – it in the Heart/Little Heart valley. The lake (or lakes) are referred to as glacial Lake McKenzie. At least once, and possibly several times, glacial Lake McKenzie overflowed, carving what is now the Missouri River valley south of the Bismarck-Mandan area.
When the (preglacial?) Heart River flowed eastward, through the Bismarck area, it deposited a thick gravel deposit which now lies buried about 100 feet beneath the Missouri River. Bismarck’s new (2013) water-intake structure withdraws ground water from this old Heart River gravel deposit.
When the McLean River valley was blocked by a glacier in the Riverdale area midway through the Ice Age, a large proglacial lake formed ahead (to the west) of the ice in the valley. This lake might be considered to be the “original” Lake Sakakawea: an early ice-dammed lake that predated the Corps of Engineers version of Lake Sakakawea by thousands of years. When the lake overflowed, near where Garrison Dam is today, the resulting flood quickly carved a narrow spillway trench south to the Stanton area.
Similarly, the Knife River, which flowed past Stanton and on to the Washburn area, was dammed by glacial ice just east of Washburn and the valley was flooded upstream beyond Washburn. The resulting lake overflowed and spilled southward into the Burnt Creek-Square Butte Creek drainage, carving a narrow trench from a few miles east of Washburn to the Bismarck-Mandan area. The modern Missouri River flows in that trench today.
And, as I noted, when the Heart/Little Heart River was dammed by a glacier, which probably advanced as far west as Sterling, glacial Lake McKenzie formed. The lake overflowed southward, forming a new valley, now flooded by the northernmost part of Lake Oahe.
The youngest and narrowest segment of the Missouri River Valley in North Dakota is at New Town, between Four Bears Bridge and Van Hook Bay. As recently as 13,000 years ago, a glacier blocked the Missouri River from its route around the north and east side of New Town. The old river route (prior to 13,000 years ago) is now a broad valley, known as the “Van Hook Arm,” flooded by Lake Sakakawea. The glacier dammed the valley, causing a lake to form upstream (to the west) of the point of blockage. Thick layers of lake sediment, known as the “Crow-Flies-High silt,” were deposited in the ice-dammed “Crow-Flies-High Lake.” Crow-Flies-High Lake extended westward from the New Town area to near Williston. In many places between these two cities, exposures of the bedded lake silt deposits occur at elevations as high as 70 feet above the modern, maximum reservoir level (1850 feet) of Lake Sakakawea. The lake rose until it overflowed southward, cutting the channel now spanned by the Four Bears Bridge.
Other “Missouri” River Routes
Up to now, I’ve tried to explain the origin of the modern route of the Missouri River. That’s not the end of the story though. The modern route of the Missouri River is only the most recent of many routes that earlier “Missouri” rivers followed through North Dakota at various times during the Ice Age. These rivers also carried runoff water from as far away as the Rocky Mountains, through North Dakota, on its way to the Gulf of Mexico. However, most of these routes, mainly in northern and eastern North Dakota, are now buried beneath thick accumulations of glacial sediment. Whatever routes these rivers followed, they had to have flowed generally eastward and southward because their original, northerly and northeasterly routes into Canada were blocked by ice each time glaciers advanced into the state. Test drilling, done to study ground water resources, has helped us identify least least some parts of the old “Missouri” River routes. There are dozens of them.
One of several early routes of the Missouri River, determined by test-hole drilling, took the river southward past Cooperstown and Valley City to the southeastern corner of the state. Another route took the river southeastward past Jamestown. In the northern part of the state, rivers like the Yellowstone were diverted from their northerly routes to easterly and southeasterly routes, past places like Columbus, Kenmare, and Minot. These buried valleys can be considered to be early “Missouri” River routes. The array of buried river valleys is really amazing – and so complicated – and such a great number of possible routes exist, that it is impossible to work them all out. All of them are now buried beneath hundreds of feet of glacier sediment, and most of them have no surface evidence whatsoever.
However, not all of the early “Missouri River” routes through North Dakota are deeply buried. In the western part of the state, a version of a Missouri River formed when an early glacier advanced at least as far southwest as the Hebron area. The margin of that glacier coincided with what is now a prominent, broad valley, known as the Killdeer-Shields channel. The channel extends southeastward from the Killdeer Mountains, past Hebron and Glen Ullin, to the Fort Yates area, crossing the modern Missouri River Valley, and continuing through southwestern Emmons County into South Dakota. No river flows through the Killdeer-Shields channel today, but an early Missouri River flowed in it, perhaps for a longer period of time than the current Missouri River has flowed in its modern route. Interstate Highway 94 crosses the valley about half way between Dickinson and Mandan. Good views of the Killdeer-Shields channel can be seen just north of Richardton and between Hebron and Glen Ullin. Old U.S. Highway 10 and the Burlington Northern Santa Fe Railroad follow the old channel from Hebron to Glen Ullin.
I realize that my description of the changes in the routes the various “Missouri” Rivers have followed since the Ice Age began is complicated. Even so, it doesn’t begin to account for the evolution of all of the changes in the vast array of routes that rivers followed during the Ice Age in North Dakota.
Most of the narrow, north-south segments of the modern Missouri River Valley correspond to places where glaciers diverted then-existing rivers southward. Glaciers in the central part of the state diverted northeast-flowing rivers, like the Knife, McLean, and Heart-Little Heart, and Cannonball, forcing them to flow southward from the points of diversion, forming the north-south segments of the modern Missouri River. Glaciers advancing into northwestern North Dakota diverted mainly north-flowing rivers, like the Yellowstone and Little Missouri, away from their routes into Canada, forcing them to flow to the east and south.
The modern Missouri River Valley is a “composite” feature, consisting of older, wide pre-glacial segments, formed over long periods of time prior to the Ice Age, along with younger, narrow segments that were cut relatively quickly at various times during the Ice Age. The parts of the Missouri River Valley that extend mainly from west to east are wider and much older than are the narrower segments that extend from north to south. Many of the early “Missouri” River routes followed for varying periods of time during the Ice Age in northern and eastern North Dakota were later buried beneath thick deposits of glacial sediment.
The current route of the modern Missouri River Valley is only the latest in a continuing series. After the next glacier has come and gone, a new version of the Missouri River will likely follow a different route than does the river today.
Eskers and kames are among the best-known of the various features formed by glaciers and by the running water associated with melting glaciers. Eskers come in all sizes: ridges snaking across the countryside ranging from a few hundred feet to several miles long, and up to 50 or 100 feet high. Kames may be cone or pyramidal-shaped hills as high as a hundred feet, or they may be simply small mounds of material. Kames and eskers are found in most parts of North Dakota that were covered by the Late Wisconsinan glacier.
Eskers were deposited by streams and rivers flowing 1) on the surface of a glacier, 2) in cracks in the glacial ice or, sometimes, 3) in tunnels beneath the ice. Imagine a river flowing in a valley or crack in the glacier. The banks of the river were formed of ice and, in some places, the river floor might also have been on ice. These Ice-Age rivers and streams deposited gravel and sand in their ice-walled valleys, just as a modern stream deposits sediment in its earthen valley. However, the ice banks of esker rivers eventually melted away, leaving the gravel deposits that had been deposited in the ice-walled valleys, standing as ridges above the surrounding countryside.
North Dakota has thousands of esker ridges. Most are small and non-descript, apparent only from a height (air view) or on air photos, but some eskers are impressive.
One of the best examples of an esker in the U.S. is the Dahlen Esker, located midway between Fordville and Dahlen in the northeastern part of the state. It can be seen as a prominent ridge off to the west where State Highway 32 crosses the Grand Forks-Walsh County line. If the weather is dry, you can drive about a mile on a section-line trail to the crest of the Dahlen Esker (if the fields on either side of the trail are being farmed, be careful not to drive on the crops).
The Dahlen Esker was deposited by a meltwater stream flowing in an ice-walled channel, or possibly through a tunnel in the ice, near the edge of the glacier. The stream flowed mainly southward, toward the margin of the glacier. The esker is about four miles long, 400 feet wide, and as high as 50 to 80 feet. In some places, native prairie covers the surface. The Dahlen Esker has been described in various geologic reports dating to the late 19th century, but the best description and discussion of the feature was provided by Jack Kume for the North Dakota Geological Survey (“The Dahlen Esker of Grand Forks and Walsh Counties, North Dakota,” Miscellaneous Series 32, 1966).
Other large eskers include ones near Benedict in northeastern McLean County; another about ten miles southwest of Carrington in Foster County; one immediately west of Hannaford in Griggs and Barnes counties; one near Dazey in Barnes and Stutsman counties; and an unusual one in Frankhauser Lake in northern Sheridan County. The Frankhauser esker is particularly interesting because it winds its way through a lake, which floods a depression that was formed by ice-thrusting. The esker formed during the ice-thrusting process.
Kames are similar in many ways to eskers. Like eskers, they consist largely of gravel and sand, but they are conical or irregularly shaped hills, rather than long ridges.
Water flowing on the surface of the glacier, or in esker valleys, plunged into holes in the ice, filling the holes with a mixture of materials. When the ice eventually melted, the water-deposited materials slumped down, resulting in mounds and conical hills. Kames occur in several places, mainly on the Glaciated Plains, usually in association with ground moraine, but they are sometimes found in areas of dead-ice moraine. A few examples include one in southwestern Richland County, one south of Lidgerwood, and one west of Cayuga in Sargent County. A prominent kame, visible from miles away, is located four miles south of State Highway 200 in eastern Foster County along the west side of the James River. About half of the 60-foot-high feature has been mined for sand and gravel.
Most eskers and kames are composed of coarse, poorly sorted materials, a mixture of sedimentary textures ranging from silt and sand, up to large cobbles or boulders. As they were forming, flowing water deposited flat-lying beds of sand and gravel. Later though, when the ice-walls melted and the bedding collapsed, the bedding became contorted. While a stream was flowing in an ice-walled tunnel or valley, or water was flowing into a hole in the glacier (a “moulin”), cobbles and boulders fell from the melting ice into the water-lain deposits, into the stream or down the moulin. Some eskers have a nearly complete covering of boulders on their surfaces. Most kame or esker deposits contain so many cobbles they are not suitable for construction purposes. Some of them consist of materials suitable for rough fill work.
Glaciation was the main geologic influence on much of North Dakota’s landscape. The Ice Age, a time geologists also refer to as the Pleistocene Epoch, includes most of the past three million years of geologic time. Glaciers advanced over the northern plains several times during the Ice Age, reaching northern and eastern North Dakota. When it wasn’t glaciated, the state had a climate much like the one we enjoy today or possibly even milder at times. the Ice Age wasn’t one long “deep-freeze.”
During their studies of the geology of the state, geologists have found evidence for at least seven separate glaciations, but there may have been more. The most recent of these glaciations is known as the Wisconsinan (because deposits typical of that glaciation are widespread in Wisconsin). The Wisconsinan glaciation began about 100,000 years ago and ended about 11,000 years ago. Some geologists debate whether the Ice Age has really ended yet. After all, large areas of the earth’s surface are still covered by extensive glaciers (Greenland, Antarctica, etc.). It’s likely that we are currently enjoying a lull between major glaciations.
Even though North Dakota was glaciated many times during the Ice Age, it is the Wisconsinan glacial deposits, the most recent ones, that are most obvious to us. These are the ones that form the hills and valleys in eastern and northern North Dakota and they are the ones in which our prairie potholes and wetlands are developed. Most of our richest farmland is developed on the Wisconsinan glacial surface.
Early glaciers, which advanced into North Dakota before the Wisconsinan glacier, also had a profound effect on the state. The materials they deposited have been largely eroded away, and about all that remains of them are occasional boulders — “erratics.” I will discuss erratics elsewhere. It was an early glacier that diverted the course of the Little Missouri River eastward more than 640,000 years ago (possibly earlier). Before that time, the Little Missouri River flowed northward into Canada. In fact, all of North Dakota used to be drained by rivers that flowed into Canada. When it was diverted, the Little Missouri River began to carve the badlands we see today.
All of us who have traveled around North Dakota know that the landscape varies considerably from place to place. Southwestern North Dakota, with its badlands, buttes, and broad vistas is largely the result of hundreds of thousands of years of erosion. The landscape there is not glacial. It has been carved from layers of flat-lying sandstone and other materials.
The Missouri River marks an approximate boundary between the eroded landscape of southwestern North Dakota and the entirely different glacial landscape north and east of the river, where we see small hills – small at least compared to large buttes like Sentinel Butte and Bullion Butte found in southwestern North Dakota. Eastern North Dakota is characterized by thousands of potholes, poorly developed drainage in places, and remarkably fertile farmland.
When the glaciers advanced over the state, they picked up some of the materials over which they flowed. The glaciers contained a variety of kinds of soil and rock, which they eventually deposited as thick layers of sediment. The exposed surface of these sediments has been weathered for the past several thousand years (since the glaciers melted) and it forms the rich soils our farmers work today.
Over much of eastern North Dakota, the glacial sediments were laid down as an undulating plain (think of the Carrington, Finley or Kenmare areas, for example). In other places, a more hilly landscape resulted (think of Turtle Mountain or the Missouri Coteau — places like Belcourt, Hurdsfield, Max, Ryder and countless others). In still other places, water from the melting glacier became ponded, forming huge lakes. Today, most of these areas are flat. Examples of the flat topography may be seen in places like Fargo, Hillsboro, Grand Forks or Grafton. The old floor of Glacial Lake Agassiz (the Red River Valley) is the classic example of flat. Hundreds of smaller glacial lake plains are found in North Dakota too.
As the glaciers flowed over North Dakota, they tended to smooth off and wear down the hill tops and fill in the lower areas with sediment. The overall result is a fairly level landscape. The layers of glacial sediment underlying that landscape are extremely complex, containing buried river channels, blocks of sandstone and shale, old landscapes that were covered many times by fresh glacial sediments. Buried layers of gravel and sand, deposited by water flowing from the melting glacial ice, constitute aquifers. They contain some of our best sources of fresh water.
As the ice flowed, in some places it picked up large chunks of material and moved them short distances before setting them down again. A good example of this is at Devils Lake, where a large amount of material was picked up and moved southward a few miles. Today, Devils Lake lies in a broad lowland. South of the lake is a high range of hills, including Sully’s Hill. The hills consist of materials that were once in the lowland where Devils Lake is now.
In some places, huge floods of water from melting glaciers carved deep river channels. Countless small meltwater valleys, along with some large ones too, are found throughout eastern and northern North Dakota. The Sheyenne, Souris, and James River valleys are good examples of large meltwater valleys. Valley City, Minot, and Jamestown are nestled in meltwater valleys. The Missouri River valley is another example of a glacial river channel, but it had such a complicated history that I’ll plan on writing a special article about it.
How thick were the glaciers that covered North Dakota? Certainly, they were more than a thousand feet thick in the east and north, so thick that the Earth’s crust beneath the ice buckled and sagged downward, eventually rebounding when the ice melted.
Who or what lived in North Dakota during the Ice Age? Mastodons and wooly mammoths lived along the edge of the glacier. Elk, caribou, and horses were common. Horses became extinct in North Dakota and in North America at the end of the Ice Age, They survived, worldwide, because they had migrated to Asia via the land bridge between North America and Asia prior to then. During my field work over the years, I’ve found mastodon teeth, caribou bones and, in the Lake Agassiz deposits, fossil fish bones, mainly perch. It’s likely that early humans also lived here while the most recent glacier was still melting.
I’ve mainly been discussing North Dakota’s glacial landscape. Part of the state, the southwest quarter, was not glaciated, but the glaciers also left their mark there. The badlands along the Little Missouri River owe their existence to early glaciers that diverted the river eastward from its northerly route into Canada. This diversion triggered greatly increased erosion by the Little Missouri River, which resulted in the formation of badlands. Some places that were not glaciated are marked by polygons, formed when permafrost froze the land beyond the limit of the glaciers.