University Place

cc >> Welcome, everyone, to Wednesday Nite at the Lab. I'm Tom Zinnen. I work here at the UW-Madison Biotech Center. I also work for UW-Extension Cooperative Extension, and on half of those folks and our other co-sponsors, Wisconsin Public Television, the Wisconsin Alumni Association, and the UW Madison Science Alliance, thanks again for coming to Wednesday Nite at the Lab. We do this every Wednesday night, 50 times a year. Tonight, it's my pleasure to get to introduce to you Eric Carson. He's an associate professor with the Wisconsin Geological and Natural History Survey, which is part of UW-Extension Cooperative Extension. Eric was born in Madison, Indiana.

LAUGHTER

He told me that his mother sent him many letters to Madison, Indiana, long after he moved to Madison, Wisconsin.

LAUGHTER

He grew up in Morgantown, West Virginia, got his undergraduate degree at West Virginia University, and got his master's and PhD here at the University of Wisconsin Madison, finishing his PhD in 2003. I think he's got an amazing story. We all know which way the Wisconsin River flows, except for maybe it didn't always flow that way. Tonight we get to hear the story of, "Did the Ancient Wisconsin River Flow East?" Please join me in welcoming Eric Carson to Wednesday Nite at the Lab.

APPLAUSE

>> Okay, so thank you. I appreciate you all showing up on what I guess is almost a tropical evening after the run of weather we've been having. The question is, did the Wisconsin River flow east? And I think the very fact that you showed up tells me that you probably have guessed that I'm going to tell you something different than it's always flowed west the way it does now, or else this would be a pretty boring 45 minutes or hour of your life.

LAUGHTER

Actually, the answer I'm going to spend the next little bit trying to convince you of is that it's yes and no. No in the fact that the Wisconsin River itself as we know it never flowed to the east, but yes in the fact that the Wisconsin River Valley was carved by a different river that flowed to the east. And the Wisconsin River as we know it today is really nothing more than a rank newcomer to this valley. It's sort of a hermit crab that's living in a valley that was carved by a different river. So, with that, I'd like to start by acknowledging some of the people that have worked with me. This is not work that I've done on my own, and I couldn't have done it without the assistance of some of my colleagues and collaborators. J. Elmo Rawling, John Attig, and Ben Bates have assisted me during the past year, and everything that I'll be presenting this evening is work and data and GIS analysis that we've conducted in about the last eight months or so. So this is really far more work than I could have done by myself. And not listed here but very important to this work is the late Jim Knox. Jim was a geography professor here on the Madison campus for 43 years. He was my co-advisor for my PhD, and when I came to work here at the Wisconsin Geological Survey five and a half years ago, the opportunity to collaborate with him was one of the big draws to coming back to this job. And so it was discussions I had with Jim during the last year or two before he passed away that really led to this research. And so the work we're going to be looking at is in the driftless area of southwest Wisconsin. So that corner of the state which is down in this area here is the portion of the state that as best we know was never covered by glaciations during the Quaternary period. So all around it at different times it had ice bordering it. It was never surrounded by ice at any one time, but at different time periods ice came and bordered it on all sides. At the earliest, ice came from the west right to where about the Mississippi River is today. Most recently, the last round of glaciations came right down here through central Wisconsin and right to the south and west of Madison to form the limits of this area. As you can see it on this landscape map, it really does stand out as something different from the rest of the state, the landscape of it. So the rest of the state is dominated by the glacial deposits that were deposited during the Quaternary. The driftless area has been formed by river incision down into the sandstone and limestone bedrock in that part of the state. And so we're looking at a much longer record of landscape evolution, something that stretches farther back into the Cenozoic era. And so having thrown out a couple geologic time terms, I figured it's probably a good place to start is to run over the parts of the geologic time scale that we'll be talking about today. So we as geologists have split the entire history of the Earth up into different time periods. The most recent portion of it being the Paleozoic, Mesozoic, and Cenozoic eras. Those represent the portions of the Earth history where there's been abundant complex life with hard body parts to it. We have a fossil record from it. That stretches back about 540 million years ago up to the modern day. So in this talk, I'll be talking just about the Cenozoic. So just this most recent 65 million years or so. And in fact just the latter portions of it. So to have a little more resolution to this, of course geologists have split the time scale out farther than this. So eras have been split into periods. So here are the periods of the Cenozoic era. And these periods have been further divided into epochs. So here's the list of epochs that the Cenozoic era has been divided into. So for the purposes of this talk, what I'm going to do is blank out those portions of this that we won't be dealing with. So we can get rid of all of those. And to give you a little time frame for what I'll be talking about, the terms I'll be using will be late Cenozoic and Quaternary, which I'll be using relatively interchangeably, and then when I need a little bit more precision I'll be using the terms Pliocene, Pleistocene, and Holocene. So you can think of Pliocene as being the time period prior to massive glaciers in the high latitude portions of the globe, and the Pleistocene, that epoch during which we had the massive glaciers covering large parts of the high latitudes. And then the Holocene, of course, is that time period afterward, going up into the modern day. So if we go back to our landscape map of Wisconsin, looking at the Cenozoic history of the driftless area. The conversations that Jim Knox and I had had that got this research started were specifically about the lower Wisconsin River Valley. So we can see it down here prominently cutting across the driftless area. The Wisconsin River flows down out of the northern part of the states, wraps around the Baraboo hills here and then flows west down the valley. This is the area that we were interested in, and I'm going to zoom in on it and I'm going to change from this color map to a grayscale one. So this is a LiDAR image. So a laser range image that's converted into a hill shape. So we can see the lower Wisconsin. It's a wide, large river valley. It's actually been, in size, quite deep and filled back up a fair bit. There's about 50 meters worth of Quaternary sediment in it. So the bedrock valley is deep, the bedrock floor is very deeply valleyed, excuse me, very deeply buried. What got us talking and what got us looking at this is a particular feature in the Wisconsin River Valley. And I've highlighted it here, and I'm going to pull this away so that now that you've seen where it is, you can see how it stands out on the landscape. This feature is what's called the Bridgeport terrace. Now, terraces are, in sort of the most general sense, terraces are remnants of former floodplains. So they are a former floodplain surface. By any of a number of reasons the river cut down and left them behind as higher, flat surfaces. So this Bridgeport strath is in here. Many of the river valleys in the mid-continent have terraces in them. This Bridgeport terrace being an example of one. But it's a specific type of terrace, and it tells us something about what was going on with how the landscape evolved. So to understand why this terrace is significant, we need to understand what different types of terraces there are. Now, here in the mid-continent what we have in abundance are what are referred to as fill terraces. So these are terraces that if you dig down into them, they're constructed out of unconsolidated sediment, usually sand and gravel. And to create a fill terrace like this you have to have three different events happen. You have to have incision of the river down to some bedrock floor, then you need to have that valley filled up with sediment, and then you have to have another round of incision that can cut down into it. So you're left with these high terraces that used to be former floodplain surfaces. A strath terrace, on the other hand, which is what the Bridgeport terrace is, is a terrace that if you dig down into it, it's constructed of bedrock that's never been eroded. So the process by which it forms is a little bit different. You simply have sequential down-cutting of the river, leaving these flat bedrocks surfaces that used to be floodplains. The two of them usually here in this part of the world don't coexist. And in fact, here in the mid-continent, we very rarely see these strath terraces. The reason for that is that process by which they form. To create a fill terrace, all you need to do is fluctuate the balance of water supply and sediment supply, and you can cause aggradation of rivers and incision and aggradation and incision. And so going through something like the glacial cycles that we've experienced during the Quaternary achieved that very thing. Numerous cycles of during glaciations, there's an abundance of sediment so the valleys fill up, and then when the glacier retreat, the sediment supply is cut off so the rivers cut down into that sediment. So in the Wisconsin River, the Mississippi River, any of their tributaries, we see many of these fill terraces. The strath terrace, on the other hand, what you really need to do to create one of these is have a fundamental and permanent change in the base level, the level to which the river is eroding. Now, if you were to go out west into the Rocky Mountains, you could see strath terraces like this all over the place, and that occurs because out there it's tectonically active. And so mountain ranges are being lifted up, and as they're being lifted up, the streams cut down into the bedrock to create strath terraces. Here in the central part of North America where it's tectonically stable, we simply don't see these strath terraces very often. And so the situation that we have in the lower Wisconsin River Valley where we have this Bridgeport strath up here and then we have a series of fill terraces down lower, this is very unique. And so we got wondering what this tells us about the history of this river valley. Okay, so if we look at this, let's highlight these again, and then what I'm going to do is zoom in on each of these so we can take a little bit closer look at them. There are three segments here that I'll be talking about. This western one near the town of Prairie du Chien, Wauzeka, and Muscoda. So those are the three massive cities that are closest to these terraces.

laughter

So there's the Prairie du Chien segment of it down in the bottom left corner. The city of Prairie du Chien, the town of Prairie du Chien is actually just here off the west edge of this image. Town of Wauzeka is right here at the mouth of the Kickapoo River. And then up here is the Muscoda segment, and Muscoda is actually down off the bottom edge of this map on the south side of the river. So you can see them in here, and there are a few thing you can see that are in common between all of them and a few differences that are kind of important. The commonality is that they're all very apparent, topographically, and they're all high. They're higher than the fill terraces that are in here. For example, here at the Muscoda segment, you can see one, two, three different fill terrace elevations that are all lower than the Bridgeport terrace. You can see a little segment of fill terrace down there near Wauzeka, and then Prairie du Chien itself, the whole town of Prairie du Chien is built on a fill terrace. And all of those are lower than this Bridgeport terrace. A little bit that's different about these that actually is going to be apparent later in the talk with some of the data we collected is how much those surfaces are actually incised and carved up and eroded. If you look at this Prairie du Chien segment, especially the eastern end of it, if you look at the Wauzeka segment, they've been pretty deeply eroded and incised. So after they were left behind as terraces, there's been a lot of erosion going on to cut down into the bedrock that formed them. This Muscoda segment is a much larger segment. It's about 10 kilometers east to west and maybe two kilometers north to south. So you're talking a fairly large sized feature. I did say that the defining characteristic of these strath terraces is that they're formed by bedrock, and we can actually see that in a few places. Two places that I know of, right here where the Highway 60 hits the bridge that goes across the Wisconsin River to Muscoda there's a small place where there's bedrock exposed above the level of the river. And then also down here where Highway 60 comes into Prairie du Chien there's a place where bedrock is exposed above the modern Wisconsin River. So we can demonstrate that these are these strath terraces. They're not the fill terrace type. So let's get out and see what these look like. Some day this will look like this again. I hope.

LAUGHTER

I hope. This is the Muscoda segment of the Bridgeport terrace. So this is looking down the length of it, looking off to the west. Off here you can see Highway 60 running down the length of it east to west, and we're looking parallel to the Wisconsin River. So the Wisconsin River is flowing on the far side of this tree line. So we're looking down this 10 kilometer segment of the Bridgeport terrace. High surface, sort of a billowy undulating surface. Farmland primarily. All of these segments are primarily farmland. If we go all the way down the end of this segment to the edge of it where it drops off down toward the fill terraces, this is what we'd actually see. So we're actually standing, looking south towards the Wisconsin River. So the Wisconsin River is right behind this line of trees. And you can see here in the road and you can also see off in the field that we're dropping off of the Bridgeport terrace, down onto one of the fill terraces. And, again, as further proof of the difference between these types of terraces, we put in a core just off the shoulder of this road, right where I took the picture, and it was right around 4.3 meters down to the bedrock surface. Once you drop all this distance off down onto this fill terrace down here, you can see that they have a sand and gravel quarry in it. So, by the time you get down here, it's probably 40 or 50 meters down to the bedrock surface once you're down onto that. So, much different sedimentology to these features. Okay, let's take a look at the topography of these things because this is useful to understanding them. See what the topography looks like, see what the internal structure of all these looks like. So we can throw on a line along the Wisconsin River Valley and construct a topographic profile. So here's what we see the ground surface doing as we move from the mouth of the Wisconsin River back off to the east along this profile. We can see the high areas that are the Prairie du Chien, the Wauzeka, and the Muscoda segments of the Bridgeport strath. And then in the intervening areas, we can see the floodplain surface of the modern Wisconsin River. And this gives us some interesting information. First off, it shouldn't be surprising to you that rivers flow downhill. So the modern floodplain is occupied by a river that flows to the west, so the floodplain itself dips to the west. And this is a basic feature that we're basing a lot of this work on, that these stream surfaces have to dip in the direction that the water is flowing. When we look at the segments of the Bridgeport terrace, it's hard to pick out much of a trend along those lines, which way it's dipping, and there's two reasons for that. First because you can see how much it's been incised into. And second because this is a topographic profile. So it's showing us the ground surface, not the bedrock surface of the strath. It's covered by some thickness of sediment. So to understand what's going on with the strath itself, what process is it formed by, we really need to look at the bedrock surface, not the ground surface. So let's look at another diagram that's similar to that. It's going to be a cross-section profile of the Wisconsin River Valley except it's going to be a fair bit more complicated. So it's going to be something just stretched along the Wisconsin River Valley like this, and this is the diagram. So this is taken from a paper that Jim Knox and my colleague and collaborator on this project, John Attig, wrote. They had this published back in 1988, and this is one of the first papers published looking at this Bridgeport terrace. So, fairly complicated diagram. I'm going to step you through it here so we can see all the different features. Occupying the central portion of the diagram here, we have a bunch of surfaces, each one of these lines representing a surface, that relates to the most recent geologic history of the valley when it's clearly been occupied by the west-flowing Wisconsin River. So we have the bedrock floor of the Wisconsin River Valley that's buried by something on the order of 40 to 50 meters of Quaternary sediment. We have the modern floodplain, and then we also have a series of terraces, of fill terraces that date back to the-- that date back to the last cycle of glaciations. Clearly, all of these formed by a west-flowing river. Then off here in this top portion of the diagram, we have features related to the Bridgeport terrace. And this is what this paper by Knox and Attig was actually looking at, the deposits on top of the Bridgeport strath. So the sediments that are covering it. And what they identified are two things that are very, very peculiar and very important for the history of the valley. The first is right near the mouth of the Wisconsin River there's a moraine, a glacial moraine. So a moraine is a ridge that marks the extent of a glacier's advance. So they identified this. They identified it as being an old deposit, and I'll discuss right now how they did that. The other feature that they identified is sort of this triangular area in here, and it's outwash that's deposited by this glacier. So, outwash is the sediment that gets fanned out in front of a glacier. They identified this and the peculiar thing that they saw, the significant item that they saw was that it was deposited by water that was flowing to the east, not to the west. When outwash gravels like this get deposited, typically they have bedding that dips in the angle that the water is flowing. And in this outwash material here, the bedding surfaces were dipping to the east so they could demonstrate that it was deposited by water flowing to the east. They applied remnant Paleomagnetic dating to it, and they were able to identify that the deposits of the Bridgeport moraine and the Bridgeport outwash were greater than about 760,000 years old. So they had a reverse magnetism so they date prior to the last reversal. Now let me flip back to the map here. I'm getting off cycle. So let's look back to the map for a second so you can visualize what they were seeing and what they were interpreting So here in the red is the Bridgeport moraine. So coming right onto this Prairie du Chien segment of the Bridgeport strath. And so what the Knox and Attig paper hypothesized was that the Wisconsin River formed under a river flowing like it does now. A westward flowing river. And during early Quaternary times, a glaciation came from the west, came out of Minnesota and Iowa and flowed eastward just to a point where it blocked the mouth of the Wisconsin River. And it dammed it up and caused a temporary reversal of flow, causing those eastward dipping beds. And then as the ice retreated...

INAUDIBLE

The question was whether the Mississippi was in place at the time, and that's a question that geologists have been wondering about for some time now, and I'll actually address that as we go along. So they proposed this represented a temporary reversal of flow of the Wisconsin River. But if we go back to this, we actually do have a nice, interesting geologic problem set up here, and it's whether that reversal was temporary. We know from all these surfaces in here that the recent geologic history of the Wisconsin River is that of a westward flowing body of water in that valley. We know from these deposits up here on the Bridgeport strath that there was a time that there was eastward flowing water. So the question is, does that time represent a temporary reversal, or did the Wisconsin River Valley form predominately by eastward flowing water and then get switched to westward flowing river later? And the way to test that is to look at the actual strath surface of the Bridgeport terrace. So, not the ground surface up here but the bedrock surface. And just like each one of these surfaces here necessarily dipped to the west because they were formed by a westward flowing river, whichever direction the Bridgeport strath dips is going to give a clear indication of which direction the water was flowing when incision of the whole valley occurred down to that level. Now, this is something that clearly Knox and Attig were thinking about back into the early '80s, but there's not such an easy way to address it. The strath surface itself is buried by somewhere in the range of five to 20 meters of sediment. We've seen by the topography that it's deeply incised, so even though it was originally a flat, gently sloping surface, it's probably got a lot of topography incised into it at this point. So for years there really wasn't a good way to address it, and it's just literally in the last year or so that the means have come together for us to address this. So the first of them is being able to really identify our place on the ground surface vertically to a high level of precision. If we go back to the relatively old days of using topographic maps, topographic maps we have for Wisconsin are nice and they're useful, but when it comes to ground surface elevation, they're simply not very accurate. There are very often errors on the range of several feet as far as ground surface elevation at an individual point. Within the last several years, what we've started shifting to with our research is LiDAR imagery and LiDAR ground surface elevation. So, LiDAR imagery is based on an airborne laser range finding system. So it's directly analogous to, say, sonar that an airplane flying over shoots a dispersed laser ray down to the ground, and it measures the rebound time and resolves that into a topographic image. This bottom image here is an older one that we have. This is part of our statewide LiDAR coverage, and it's got a resolution of 10 meters, which is to say each of the pixels or each of the grids on this is 10 meters by 10 meters. So that does a fairly decent job of giving us the general contours of the land surface. And as you pan back, it does a really good job. In fact, the earlier images, the early grayscale images, I showed you the entire Wisconsin River Valley, are from this 10-meter LiDAR set. But as far as identifying ground surface elevation at an individual spot, this still wasn't good enough. And it's just in the last year, in fact January of last year, that we at the Wisconsin Geologic Survey started getting access to a new set of LiDAR data that's being shot in Wisconsin on a county by county basis. And so the entire lower Wisconsin River Valley was shot for higher resolution LiDAR, and this is the same area as in the bottom left image with the new high res LiDAR that we have. So this imagery has a resolution of one and a half meters. So each grid is one and a half meters by one and a half meters. As far as resolving ground surface elevation, the GIS people in my office assure me that any grid square is going to give you an estimate of ground surface elevation accurate to about plus or minus five centimeters. So it has a good deal of accuracy to it that you can use remotely. And look at the precision and the resolution of the imagery itself. And here we have Highway 60 coming down, making this "T" junction. You can actually see where there's individual road cuts. You can see where there's roads built up. You can actually see house foundations on this. It's really treacherous for us because it's a great data set, but I can also lose about a half a day really quickly scanning around this and looking at the kind of information that's there. But it is a game changer for us. It gives us the opportunity to identify our position on the ground surface very, very precisely. And if we want to know how far down it is to the bedrock, this is our starting point. The second piece that we have to have come together is a methodological advancement. And that's something that I've taken on since, or become a fan of, since I've been here at the Wisconsin Survey, and it's coring methodology that we hadn't used at the survey before. So this is what's referred to as geoprobe coring. Geoprobe is a coring mechanism made by a company out of Kansas. And it's a direct boring, direct percussion system. So it's steel barrels that are hammered down through the ground simply through hydraulic pounding. It's about a two-inch barrel that goes down into the ground, and it's got a lot of really good advantages to it. As far as this work is concerned, we're putting down steel barrels so we can measure to within a centimeter or two how deeply we've gone into the ground. It's also got the benefits of being quick. Depending on how deep we're going we can do multiple holes in a day, where some of the older technologies would be multiple days per hole. It's inexpensive. With the kind of budgets we have available to us at the Wisconsin Geological Survey, this is something that we can contract out and do it fairly frequently. The final thing is kind of demonstrated by where this exact site is. We're just pulled off to the shoulder of Highway X here. It's a really mobile, really portable system. It's only a two-inch barrel that's going down into the ground that we fill in afterward And so we can do something like this, we can pull off to the shoulder of a highway to do it. And the sad truth is over the past years, every year it gets a little bit more difficult dealing with landowner permission just to do a simple coring operation. If I were using a coring technology that I needed to get landowner permission to get on everyone, I'd spend much more time dealing with landowner permission than I would with the actual coring. With this, we can pull over to the shoulders of roads and be done in just a couple hours with our coring. So now we have the ground surface elevation that we can identify very precisely. We have the depth that we core down that we can identify very precisely. So to start getting a handle on this strath surface, all we need to do is be able to very clearly recognize when our coring hits the bedrock, which is a little bit harder than you might expect because the bedrock that this strath is made out of is a sandstone, and it's a fairly poorly cemented sandstone. So in a lot of cases it can actually be a little bit difficult to tell when you've gone from the Quaternary unconsolidated sediment into the Paleozoic sandstones. And in this case, we've actually found that we've kind of, well, we're fortunate, we'll say, with what we're finding. The exact sandstone that formed this strath surface is rich in a mineral called glauconite. Glauconite has a bright green color to it. So in this photo right here, I'm holding the sampling tip that screws onto the lead end of the barrel, and it's filled with this glauconitic sandstone. That green color simply isn't a color of Quaternary sediments. So me, as a Quaternary geologist, when I know that I see that, I know that I'm out of the Quaternary, I'm literally out of my realm, but I know that I've gotten into the bedrock. So let's look at one of these cores look like. So this is one that split open. This is a geoprobe core that we've collected and split. And you can see at the top, this is the unconsolidated Quaternary sediment. We can see right here, there's a nice large chert cobble that's in there. And then right below that we have the Cambrian sandstone going down at depth. And as an indication of how poorly cemented this sandstone is, in a lot of cases we're able to penetrate a meter or two into it with this direct percussion boring. Just beating out way down into it. So in this same core, if you can continue down, this is the kind of thing you see. So we're seeing the stratigraphy, the layering of the sand from this Paleozoic sandstone. But this was really useful for us because now we don't have to rely on how deeply cored. We can take how deeply cored, open the cores up, and see the actual contact down to the centimeter or so. So now we can get our ground surface elevation, we can get the actual depth down to the bedrock surface so at an individual place we can find the elevation of the strath surface. And then with how quickly we can do the coring and easily we can do the coring, we can simply take a brute force approach to it. And that is what we did over the summer of 2013. So here's our three segments of the Bridgeport strath shown again. Each one of the red dots are places that we did geoprobe coring. So, as you can see, at least within the realms of what the road network on these terraces allowed us, we left no parts of the terraces untouched. We cored as much as we could in as many places as we could. >> How deep do these cores go? >> His question was how deep do the cores go. In the range of different materials that I've cored in, we've cored as deep as 25 meters with the geoprobe coring. These cores range between two and 20 meters to get down to the bedrock surface. Okay, so if we take all these, the next step if we want to resolve which way the strath surface is dipping is to plot these up as a function of where they fall on the valley. So plot their elevation versus their distance along the river. And this is the trend we get. And so there are different types of points on here. There are points that we did with the sampling tip open. So we collected cores with the geoprobe. There are some that we actually put a solid tip on it, so we just bored right down to the surface because it was quicker once we convinced ourselves that we could duplicate the depth to the bedrock with that. There was actually one core that we put in with a rota sonic core. So a rota sonic core is much larger operation. It's a four-inch rotary core that we put in for a different purpose, but it worked for this data. And then there's also a couple in here that for a variety of reasons we're fairly confident that we weren't able to actually get down to the bedrock surface. So at those coring locations, the bedrock surface was some point below that. Okay, so as we look at this plot, there are several things that pop out. First off, we see a fair bit of scatter below a line, which isn't surprising because we're imaging some sloping surface that's been incised into it over time. So we're picking up that incision. We can see that this Wauzeka segment is exceedingly low. And, frankly, I'm not all that surprised about that. It was the smallest of the three. It was bisected by the mouth of the Kickapoo River. It was also fairly heavily incised in its surface. So it experienced quite a bit of erosion. So, if we're looking at this, what we're looking for is the top trend. Which direction does the top edge of the data run? And I think it's pretty clear and I think I can convince most people that we're looking at data of an eastward dipping surface. The gradient that we get there is on the order of 0.15 meters per kilometer of dip. That's within the ballpark of what we'd expect. The modern Wisconsin River dips a little bit more steeply than that, but then it's been affected by the recent glaciation so it's a little bit of a steeper river as it is. When you look at this, you might say, well, there's a big pull on that trend line based on this one data point, which, yeah, I can see that's the case. But if we take that away, I think we can argue about, or let's not argue, we can discuss the specifics of it.

LAUGHTER

But it's hard to look at these data and come away with a conclusion anything other than they represent a surface that dips to the east. There's just no way you can put these data together and say this surface is dipping to the west, at least none that we can figure. And so what this is telling us, the direct implications of that, is that we're looking at a strath surface that dips to the east. And so the incision of the valley down to that surface was formed by a river flowing to the east. And I think that's based on this data, and actually we're still collecting data on this. In fact, tomorrow morning I'll be going out to do additional coring with this. But based on this data, it's really hard to come up with any other conclusion. Now, geology doesn't happen in a vacuum. We'd like to think that if there's a transformative event that happens to the landscape, there's going to be indications of what the previous conditions were. And when we start looking at the landscape here, I think we can start pulling out a lot of features that indicate that there used to be an eastward flowing river and a radically different drainage pattern occupying this portion of the mid-continent. So here we have our LiDAR image again. We have the Wisconsin River and Mississippi River flowing in their current configuration, and there are a few things that start standing out when you look at it. The first is as you look along the lower Wisconsin River, there are an incredible number of what we call barbed streams. Now, tributary streams, as they flow into their larger river, they tend to incise an angle in the direction that the river is flowing. And a barbed stream is one that does the opposite. And as you look at this, strung all along the river on both sides there are streams that enter into the Wisconsin River angling off to the east instead of to the west, classically, going back a hundred years or more. In fact, our own TC Chamberlin used this in another part of North America to document reversal of stream flows. This is just a classic geomorphic hallmark of a river that's experienced a reversal in flow. There's also this bend in the river down here at the confluence of the Mississippi and Wisconsin Rivers. This is to the inside of the confluence, to the north and the east of it. And you notice there that the valley wall has a nice, smooth, curved radius to it. And that really isn't typical at all of the confluence of two rivers. The examples are kind of small in here, but if we look at, say, right here where the Pine River comes down and flows into the Wisconsin River, the inside of the curve is pointed. Any of these along here, the inside of the confluence of two rivers comes to a point. And yet, this one is curved. That makes no sense with this configuration that we have, but it does make sense if that point on the landscape where we now have the confluence of the Wisconsin and Mississippi River, rather than being a confluence, back in geologic time was just the inside of a bend of a single river. And as we look along the river valley, we see plenty of places where we see that. Right here the Mississippi River makes almost a 90-degree bend, and it's got that nice, smooth curve. Everywhere along these rivers where we see the insides of a bend, we see a nice, smooth curve to the valley wall. A few other things. If we pull all of this away and look at the Wisconsin River as a whole, look at the width of that river valley from valley wall to valley wall, as you go from east to west, the Wisconsin River Valley gets narrower, and that's exactly the opposite of what we typically see with river valleys. They tend to widen in the downstream direction, and this is the opposite. And with that width of valley kind of notion, if we actually step out of the Wisconsin River and look into the Mississippi River, so I'll pull up a detail of this portion. So here we have Prairie du Chien, the confluence, and the Mississippi River extending to the south. What you see is that the Mississippi River actually narrows quite considerably south of its confluence. It's being joined by the Wisconsin River so its flow is getting significantly larger, and yet it becomes a much narrower river valley. And as you look along this valley here, you can see that the streams that flow into it are extremely short and they're extremely steep. And that narrow channel with steep tributaries, steep, short tributaries, is also a hallmark of a relatively new, relatively quickly incised river. So, in addition to the coring, we're starting to get a good feel for a radically different system. Now, what would that look like? Here's what we're proposing. Here's our modern configuration of the Mississippi and Wisconsin Rivers. So the Mississippi coming down from up here in the Twin Cities area, forming the western boundary of our state. The Wisconsin River coming down and flowing west through the lower Wisconsin River Valley. What we would propose is that through much of the Cenozoic, this drainage network evolved as a stream system that looked like this. That rather than the Mississippi River continuing south at all, the high ridge that forms Military Ridge across the southwest corner of the state was an impediment to flow, and it actually caused this river to make a tight left turn here and flow east up this valley. And this was the river that incised the valley that we see today as the lower Wisconsin River Valley. And then what is the Wisconsin River today, which flowed off somewhere, came and joined that river somewhere up here in the eastern part of the state. Now, we're not talking about the upper Mississippi River anymore necessarily. So we want to come up with a little bit different terminology. And the name that my colleagues and I have settled on is referring to this as the ancestral Wyalusing River, based on Wyalusing, Wisconsin, and Wyalusing State Park right there at the confluence of the two rivers. So to avoid confusion, we'll talk about it as the Wyalusing River back when it was this configuration. Now, we can follow it all the way to the margin of the last glacial deposits, and then everything gets buried by those recent deposits. So we want to look at where this river would have gone beyond about Sauk City where you hit the Johnstown Moraine and the most recent glacial deposits. So if we look across the eastern part of the state, just like Military Ridge forms a nice, high bedrock barrier, so does the Niagara Escarpment. So this is the solarium dolomite up here that helps form Door County. And then in here, in this central portion of the state, we have a bit of a gap. And so our options are basically two. This Wyalusing River could have come up here and made a right-hand turn and gone south and actually still been part of the Mississippi River system, or it could have continued up to the northeast. And, again, we really don't have a direct way to address this because everything's buried by recent sediment. What we do have is well data. Our office is the repository for every high capacity well that's drilled. Every high capacity well that's drilled has to submit drilling data to us with descriptions of the sediments they cored through and depths of the sediments. And so that's the tact we use to look for a river valley. Either a river valley that bent off here to the south or a river valley that continued up to the northeast. And this is the work that we did. So in this area, in the east-central part of the state, this is, I believe, a nine or 10 county area stretching from Sauk County all the way up to Oconto County, we looked at something on the order of 75,000 well records. We located them accurately in space, identified the ones that gave significant information for depth to bedrock, narrowed it down to about 16,000 well records, and from that generated this bedrock topography map. So, on this, the blues are lows, the reds are highs, and obviously the thing that stands out is this big valley that extends up the modern Fox River Valley. Down in Dane County, which we don't have on here, and in Columbia County, there really aren't any lows at all comparable to this. There's no way that that water could have escaped back down to the south. So we are convinced that we're looking at this northeastward extension of this Wyalusing River. Now, we can actually start monkeying around with the elevations in this and comparing the elevations of this valley and the gradients to what we found in the lower Wisconsin River Valley with the Bridgeport strath. So we can draw a cross-section on here. Here's our cross-section line that contains data points from the Bridgeport strath, data points from the access of this buried bedrock valley, and when we plot them up again from the mouth of the Wisconsin Valley extending back off to the northeast, this is what we see. We have our cluster of points from our geoprobe coring. Remember, these have a precision of a couple centimeters on them. We have these points that are from the well constructor reports. And you can see there's a lot more error to them, but that's not really surprising. In a lot of these, they were having that same issue of going from unconsolidated sand into this sandstone bedrock. The people actually doing the logging had varying amounts of geologic training and, frankly, probably varying amounts of interest in what they were doing on that particular day.

LAUGHTER

So, if we're talking errors on the range of a few centimeters here for this, I don't think it's out of line to be thinking of errors in the range of 10 meters or more on these data from the well constructor reports. Be that as it may, we can still see the ideas of a nice trend line of a continuation of this river. And if we actually want to take this cross-section line here and start playing a few games with it, what we can do is continue it all the way up here around passed the end of Rock Island to Rock Island Pass where Green Bay empties into Lake Michigan. And we don't have bedrock data elevation for there, but we do have the floor, the bathymetry, the lake bed. And there at Rock Island Pass it would fall right there on the plot. So it's really starting to give us good evidence that this river not only flowed east up the lower Wisconsin River Valley but continued on up to the northeast. And the extension of this is that this water rather than draining out the Mississippi River basin, it all drained off to the northeast up to through the Saint Lawrence Seaway, directly into the Atlantic Ocean. And so we're looking at a drainage system, proposing a drainage system, something like this, that this dashed line would have been a continental scale drainage divide. The water to the north of it would have flowed out the Saint Lawrence, water to the south would have flowed down to the Gulf of Mississippi. And here's out Wyalusing River that we've traced this far, and once here at this point, it would have taken some path out the Saint Lawrence Sea to the Gulf of Saint Lawrence. Now, we need to think about what the geologic process would be that would shift it to what we have today. So, today, here's our drainage divide for the continental scale. Separation between the Mississippi and Saint Lawrence. Here's our Wisconsin River that's now occupied that valley. And we actually have a nice model for what the geologic process would have been to cause this shift. So if we go back to our LiDAR data here, here is the area that we've been talking about. Here's our Wyalusing River, and you'll notice that there's something different about this LiDAR. And what is different about it is that I've gone in and I have removed the Mississippi River.

LAUGHTER

What I like to say is I don't often doctor data, but when I do I prefer to doctor LiDAR data. So here we can imagine our late Cenozoic time. This is the Wyalusing River that's coming down and it's making that turn, blocked by Military Ridge, and it's flowing off to the east. We have our nice bend there at what is now Prairie du Chien. We have our tributaries that are not barbed tributaries. They're forming just as you'd expect. And then we have the top end of the ancestral Mississippi system. So, nowadays, this is the Platte, the Grand, and this is the Turkey River. I've put in, generously put in, a little tributary stream right about there.

LAUGHTER

Yes? >> Does Military Ridge extend west of Wyalusing? >> The bedrock, the type of bedrock does. It certainly does. Those formations extend, in fact, all the way up to the Twin Cities, but they shift into shales as they move up there. His question was whether Military Ridge extends off to the west. Okay, so as we step forward through time, we know from, say, the Bridgeport moraine and we know from an abundance of glacial data across the North American continent that the early glaciations during the Quaternary were more focused in the western part of Canada and the US. And so those early glaciations like that which formed the Bridgeport moraine flowed eastward out of Minnesota and Iowa, but rather than damming this river and causing reversal, they simply covered over the upper end of the drainage basin. And this would be the time that those eastward dipping outwash sediments were deposited on the Bridgeport strath surface. We also know that during the Quaternary there were plenty of intervals where there were no glaciations, so we would have returned to a setting like this. And then similarly, based on the glacial deposits that are spread across Wisconsin, Illinois, Indiana, Michigan, we know that the more recent glaciations were focused more in the Hudson Bay is where the ice was centered. So rather than flowing out of the west, it was flowing south or even southwest to us from up in the Hudson Bay region. And so at some time during the early or middle Quaternary, a younger glaciation would have flowed down far enough to dam the Saint Lawrence drainage basin, and when it dammed the Saint Lawrence drainage basin, it would have blocked this river and it would have caused a lake to back up all the way up this valley. And it would have backed up and backed up and backed up until it flowed over the lowest drainage divide. And right here at Wyalusing is a perfect logical place for that to happen. You've got this Wyalusing River cutting here at the outside of its bed, cutting into the bedrock. You've got this river that I put in here that's eroding headward. Perfectly logical place for a low spot. And so this lake would have filled until it would have overflowed down the Mississippi drainage, and it would have started incising quickly in here. And whether it would have taken one glaciation or multiple glacial cycles of damming of this lake like this, eventually what you would have had is a new drainage pattern established flowing down to the south. And once that happened, the Wisconsin River could have occupied this in a westward flowing configuration like we see today. So now we have our Mississippi River in its current configuration. We have the Wisconsin River in its current configuration. We have all these tributaries that are now barbed. Okay, really quickly to wrap up here, what I'd like to do is take a larger look at this. Pan back and think about the Saint Lawrence and the Mississippi River as a whole. So here in the pale yellow, we have our modern Mississippi drainage, here in the blue, we have the modern Saint Lawrence drainage, and this area I've been talking about, the Wyalusing River represents this area here in green. So it is a significant amount of area, and it's a significant amount of water being shifted from one drainage to the other. What I've just proposed here is absolutely new for the upper Mississippi River, but it's not new for the greater Mississippi. It's not new for primarily the Ohio River. The Ohio River, going back over a hundred years, has been documented to have experienced two different piracy events like I just described. One which generated the upper Ohio, one which generated the middle Ohio. So we'll start at the upper Ohio region. So here we are in the tri-state area of Ohio, West Virginia, and Pennsylvania. Here's our borders on here, and a couple of relevant locations. Pittsburgh, New Martinsville, West Virginia, and Beaver, Pennsylvania. Let's throw our modern drainage on here. The Monongahela River comes up out of central West Virginia, flows north. The Allegheny comes down out of Pennsylvania and New York. They join at Pittsburgh and flow off to the southwest as the Ohio. But if you look at this, in this stretch through here what you see are a whole series of barbed streams. And dating back to 1934, Frank Leverett documented that what happened is that this developed through the Cenozoic as a river that flowed off to the north. The Allegheny and Monongahela system joined this river at Beaver, Pennsylvania. New Martinsville, West Virginia, was the drainage divide. And similar to what I just described, glaciations blocked this river up here to the north until it filled up a lake that spilled over here at New Martinsville. This was ancestral Pittsburgh River is what the geologists out there call it, and the glacial lake sediments are the glacial Lake Monongahela. So this is a well documented river system. Central portion of the Ohio River, similar story. If anything, a little bit more complicated. We have Chillicothe, Ohio, here, Cincinnati, my birth place of Madison in here, and our modern stream system. So we have the Ohio running east to west across it. This is the Kanawha River, this is the Licking, and this is the Kentucky River coming up and joining in. And as you look at this, not just barbed tributaries, but there's something else odd standing out. What there is are a series of abandoned river valleys in here. All of these dashed lines are valleys that have absolutely no significant river flowing through them. In fact, over here by Cincinnati, there's another series of them. And this one going back to 1894 is the first publication on this, by Frank Leverett and TC Chamberlin worked up this, and the story that they identified was based on these abandoned drainages. Barbed tributaries here east of Manchester, Ohio, barbed tributaries here east of Madison. They reconstructed a drainage network that looked something like this. And so all these rivers here were flowing up past, oops, I have done that again, up past Chillicothe. These rivers here were flowing up past Cincinnati. They were joining and flowing somewhere off to the north. So when we look at this again, we have the Pittsburgh River here, that central system, which is the Teays River, add those into the Wyalusing, and you're now talking about three pretty significant areas that were shifted from the Saint Lawrence drainage basin to the Mississippi. And we can actually put some numbers on it based on the modern gauge data. How much water is flowing past these places. So if we look at this, if we look at it from what the addition of these three basins means to the Mississippi River, based on area, these three together represent a little over 14% of the modern Mississippi drainage basin. Because they're in the eastern and much more humid part of the drainage basin, this actually under represents what it means in terms of water supply. These three areas together represent 22% of the water coming out of the mouth of the Mississippi. So it's a significant amount. If we flip it around the other way and look at the Saint Lawrence, you can see the Saint Lawrence is a smaller basin than the Mississippi in area, also in discharge, and so the loss of these is going to be proportionally much more significant. These three basins would have represented, up through late Cenozoic time, they would have represented almost 24% of the drainage basin. And, again, that underrepresents the discharge. They would have represented something over 30% of the discharge that went out the mouth of the Saint Lawrence through the Cenozoic that was permanently redirected down the Mississippi River Valley. Now, the implications of this are pretty significant. We think on a global scale of things like the North Atlantic circulation as being an important player of when and how the glacial cycles occurred. The flux of fresh water in and out of, well, into the North Atlantic or not into the North Atlantic can have a significant role in that. So we think this is one of the features that may well have impacted how the glacial cycles and the global climate system changed during the Holocene. Okay, so the conclusions are that we have multiple lines of evidence that this is in fact the case, so just start believing it.

LAUGHTER

This is emerging work, okay? I joke, but this really is emerging work. There's lots that we have to do. We have things from the humble, like continuing to look at well constructor records to trace out where this river went. I mentioned glacial lake deposits with the other pirated streams, and we have places that we need to start looking for that regard. And it expands outward to the large scale, looking at sediment and geochemistry of sediments delivered all the way up to the Gulf of Saint Lawrence and all the way down to the Gulf of Mexico. And so what I want to close with, and another one that we still have to work on is age control. I gave you the geologic time scale in our Pliocene, Pleistocene, but I never really gave much in the way of numbers through this talk. So we're working on ways coming up with time constraints on it. And one of those very well could be looking at the downstream record. Looking at sediments deposited in the Gulf of Mexico. So what I'd like to leave you with is a paper that was published right on a little over 20 years ago, looking at a sediment core from the Gulf of Mexico. So this was a sediment core that recorded over five million years of sediment deposition in the Gulf of Mexico, and they looked at oxygen isotope variations, which are often used as an indicator of climate change. So this paper gets cited quite a lot by people looking at the timing of the onset of glaciations in the Quaternary period. And I want to point out the last sentence of the abstract. So the melt water anomaly, so the changes in oxygen isotope variations, implicate either the periodic presence of North American mid-latitude glacial ice since the Pliocene, or a drainage system very different from that of the present. And that's the sentence that I've kind of loved ever since I read it when I got working on this.

LAUGHTER

Okay, and with that I'll go back to a view of what now is looking downriver on the lower Wisconsin, very well could have been a similar view looking upriver on the Wyalusing. And I'll be happy to take questions. Thank you.

APPLAUSE