MSU Extension MI Ag Ideas to Grow With

Now we're going to talk a little bit about some efforts to help quantify where we are as far as irrigation impact to natural resources. So I have with me today to Todd Feenstra from Tritium Consultants. And he's going to explain his relationship to Midwest Water Stewards and what they do. All right. Thanks everyone for allowing me and giving me the chance to join and share some of this today. Midwest Water Stewards is group that spun off of what used to be known as the Southwest Michigan Farmers for Responsible Water Use. And then also combine them with a group called link out of Indiana. Both groups our grower funded and grower members. And they were really dedicated to beginning the process of collecting data to fill some of the data gaps that were needed in the state. So in both states actually. So I'm just going to give a little bit of background and we're just going to jump into it. So with Midwest water stewards. Here we go. In Michigan specifically, this is a map from one of Jim Mills presentations to the water use council. As of the end of 2021, there was almost 6,000 high-capacity withdraws registered in Michigan. As Lyndon had said earlier, you can clearly see the bulk of them are in Southwest Michigan. Of those withdraws, 85% of them are related to ag. So there's a very large discrepancy between ag and then the next even closest one is industrial use. As we go down. So the MI-WWAT process itself, the Michigan water withdrawal assessment tool, or the process really does emphasize. The emphasis is really on agricultural withdraws. It's really focused on those as the as the largest number of withdrawals within the state. So typically in Michigan, what we find is we work all over the state. For the last, I've been doing this for 25 years now. This is a very typical profile that we would see cross-section of geology. It's usually a three-layer system at least. And that layer would be the upper here in yellow would be an upper aquifer that may or may not be connected to the surface water features. Then there's a second layer down, an aquitard or layer that just does not conduct water very well. And then the third layer underneath would be a confined or often in Michigan or leaky confined aquifer. So yellows are potential aquifers. The browns are a restrictive unit. And so the question then becomes, in this type of geology, How well do the layers communicate with each other hydraulically and what is in the MI-WWAT. Does it, does the MI-WWAT account for this? So the MI-WWAT itself is built like this. This is the fundamental model and the assumptions that are behind the online screening tool for the MI-WWAT. It's built-in to that online screening tool. It is a single layer system. And there's a couple of assumptions that go into this model that need to be understood. One is that the water levels in the aquifer system are directly connected to the streams. So that's assumed that they are connected. It is also assumed that all the water that is pumped from the well, the only source of water that eventually makes it to that well is the stream. We may see some depletion from storage around that well, but the only recharge even to that storage in the aquifer is the stream itself. So it's very clear in the documentation and the author who wrote up the model, that's one of his foundational assumptions, is that the only source of water to that well under pumping conditions is the stream itself and then the layer, the layers, obviously it's just a single layer. So there's a tabulated values for the rate of flow in the aquifer that are that are assigned to the different lithology types. And so an aquifer. And then you have an aquitard and then the lower aquifer is just simply an average of those three. And then the storage value, the average dictates how fast water moves through the aquifer. Here in this model, the storage value is just assigned the same across the whole state and that's at 0.01. So that doesn't vary. So those are the foundational assumptions to the model. As we look at that. Why is this becoming so important now, why do I stress that now understanding how that model is built with the assumptions are this is as of October of 22, this was presented to the water use council. Green in this would represent. Clear pumping. You're welcome to pump. Installer will get a registration, shouldn't be an issue. The ones who are more concerned about are the Zone Cs and the zone Ds. The zone Ds are a negative status, meaning something has got to be done to restrict the existing or current water use because it's currently, according to the model, harming the environment.  Zone Cs are a concern for me. As I look at this on behalf of growers because it is zone C, We may be one or two wells away from tipping into a Zone D, but There's no way to get a well registered in a zone, C that would cause it to go to a zone D. Zone C is really the last stop for it for new withdraws, typically. The other one of concern that grows as well, especially as we look at Northern Michigan or areas where there's cold transitional streams. In those streams, there is no zone C it simply goes from a zone B to a zone D. And so in effect, there's a lot of watersheds getting really close to 6800 watersheds in the state of Michigan, that according to the model for all intents and purposes, either one well away or we're not able to add any more withdraws. And so what we're starting to see now, Lyndon mentioned it earlier in the last talk that he gave a few minutes ago, was that we're now starting to see real-world restrictions or denials. We're starting to see real-world issues where we have the zone D is we've got to figure out who's gonna give up how much water. We're, we're right on the precipice of water users groups forming in. So now we've got conflicts. So we've got model predicted conflict. Because the model is now predicting that we either can't put more water in or that we're actively damaging the environment, actively changing those fish populations. And that's not allowed by law. So we're, we're in a situation now. Where is this kinda been creeping up on us for the last 15 years? And in here we are and it's really starting to get over into that tipping point now where we're gonna have to take some steps to address some water conflicts in Michigan. So when we do get to a water conflict, and when that happens, we're going to have to enter into some sort of remediation. Historically, what it's been is the reasonable use doctrine. Combined with the riparian rights? In absolute, there's a bunch of different opinions on how those flow together. But historically what's happened is it's just done with a neighborly arrangement. If you ran your irrigation well, you dried out your neighbors well, then you were responsible and you worked it out with your neighbor that you would lower the pump in his well or you'd replace his well with a new one. Or if it was two irrigators that were impacting each other, they would work together and schedule their irrigation so that the water levels weren't being impacted. So everybody's wells would continue to work. MI-WWATs is a little bit different now because now it's not that we're seeing an observed problem or there's a conflict between wells one drying out or the other. Now we're seeing is that the model is predicting that these wells are impacting streams. And they've impacted them enough that they're going to cause environmental damage. Or we're right on the edge of that. So therefore, we can't put another well in. We're essentially out of water according to the MI-WWAT model. And that is going to trigger something new that has not happened in the state yet. And to my, according to my research, it hasn't happened in other areas of the country really, either. Where we have a model that's predicting that there's gonna be impacts. We don't have a lot of data that supports the predictions of that model is we'll talk about in a little bit. But here now we have a conflict issue. We're going to bring a model into a mediation, some sort of remediation event to try to rectify how do we share this water if we assume that the model is correct and there is no more water to be had. Water users groups may be that venue. Water users groups. And you can probably talk about that some more extent than I can. Water user group essentially is going to be moderated by EGLE Staff in the water withdraw owners for whatever particular watersheds in trouble at that point are going to have to meet together and figure out how to share the water that the model says is available. So you're looking at distance. I mean, a lot of things, distances from the stream, how deep is your well, how hard you pump it? What is your schedule? What is your annual use? Do you have a direct withdraw from the stream? Do you have a groundwater withdraw? What is your impact and what is the model saying about this or that? And it's gonna get a little bit trickier. get hairy, I think and it's for sure going to get contentious. And I can say that with 100% certainty just from looking around the world that where there's been water conflicts. And it's gone into a strong type of mediation like this, and it's going to get it's gonna get pretty hotly contested rather quickly. Perhaps what will happen is it may escalate out of water users group or never even get there. There may be, someone tries to do a legal lawsuit. Maybe the judge says, Hey, we don't really need to do a lawsuit over this. How about we assign you to a professional mediator? And now we're going to have a different type of mediation where someone's going to try to get the two sides to get to some sort of agreement. Or if the nuclear button is hit in, there's a lawsuit filed. And now we're in court. And now both sides are trying to present their case for who's done what, who's impacting how much, um, are we using the model? Is there data to back it up? Are the predictions accurate? Are they not accurate? There's gonna be a lot of questions about the science in the modeling and the data that's behind it, they're going to have to be answered. So I want to back off really quick and give a real-life example here of Clover River in Wisconsin. They measured, this is a trout stream in Wisconsin in the  Central Sands area. And they measured how to declining stream. And so they built this groundwater flow model, a combination of the Wisconsin DNR, the USGS, there was industry was involved, agriculture was involved. Watershed users groups were involved, residential neighborhoods, and residents were involved. And at the end of the day, they built the model. They did some predictions, they did some measurements, and then they tried to get this resolution and determined from different scenarios. What were the things that would impact this the most? How can we restrict the water in, and really who should be doing that? Interesting part about this to me, was it, the authors wrote this up in a scientific journal after it was all over. And one of the things that they stressed and said that they wish they had known how to time in that they stressed for future water conflicts was that there was a sore need for trust. That everybody brought their own opinions in, everybody brought their own theories in. People had different ways of presenting and talking. But the fundamental thing that they said was absolutely critical in getting through this was learning to trust each other and be able to communicate with each other. That really fed into this whole trust issue, really fed back into the models. Are you going to trust the model or who put it together? Are you going to trust the model itself and how do you go about doing that? And really, what we need to do is we need to take the heat out of the room by having some commonality here. The easiest way in the scientific method way is when we build a model, we test it against a real-world measurements. So is the model reasonably representing the real-world? Is it reasonably representing what we've seen, especially as we do predictions? If the MI-WWAT says that we've gone from the Zone A with 24% allowable depletion, and now we're down into whatever zone D limit may be for that watershed. Have we actually seen a decline from a zone, a situation all the way down to a zone D. Do we see depletion in the stream, long term? Are we seeing it even short-term for the month, the drier months or the irrigation season. Are we seeing impacts? You know, that we can trace back to the irrigation? Or have we really comprehended all the other parameters and other things that influence that stream. Have we accounted for? The wetlands, the lakes, the feeders, the field drains, have we accounted for precipitation? Have we accounted for, as was mentioned in earlier speech, what the spring water levels were are they 3 ft higher than last year or 2 ft lower? Are they near the average? Have we accounted for all the other variables that we would see and expect to include in the model. Then at the end of the day, does the model actually match what we're seeing in the field? That's a problem right now because there's, there's a very dire lack of data across the state and we can throw numbers around about what's been collected, what's not been collected. I know the water use council worked with EGLE. EGLE received about $10 million to spend on trying to collect more data and update the models. Um, I believe about a quarter of $1 million has been earmarked for new monitoring wells. Last meeting that I was in when they were reviewing the final budget on that, they were talking USGS was talking and presenting. And EGLE was presenting that there would be about ten wells a year that we're going to be reactivated at a cost of about $250,000. There was a need for about 143 and more wells. We have a long ways to go. We also recognized you know, Dr. Sears had talked about 150 stream gauges. And where do we have discrete measurements? What's in the stream beds? Where do we have aquifer testing that tells us about the characteristics of the aquifer, not just derived from a well log with an assumed or a tabulated value assigned to lithology. Where do we have real aquifer tests that give us the parameters of how fast water flows in, how much water is stored at these, these various sites. And when you dig into that, you'd find out that those datasets were really, really limited. Midwest water was formed specifically to account and to start addressing some of those issues. So we believe very strongly that the base of the pyramid is the data. The data is critical in understanding the groundwater system and the surface water systems. That data then can be used in modeling. We can use it to verify an existing model and check and see, Hey, here's the predictions. How does it match with the real-world measurements we've been collecting? Or we can look at it from a different perspective and say, Hey, we might need additional modeling or improve modeling in some area. Here we've got a nice dataset in that area. Can we use that and that understanding of that area now to build a model that does reasonably represent that area. Then the last part I think, is a very, very critical piece that often gets overlooked. And that's the education, the outreach. There are journal articles and I understand that technical journal articles. There's no one that I know outside of my industry that picks those up to read those. We as experts need to be communicating. What we learn, what we find, how we're doing things, what are the assumptions in our model? How does our model work? Why does our model fit the real-world? Does it fit the real-world? As experts, we need to be able to explain that in plain layman's term in language. That makes sense. That that is also honest and truthful. Because again, our goal is to build trust. We don't want to have that torn down by doing a poor job. And so that whole platform right here in the whole structure of what Midwest waters about is first collecting that data, then checking or building the models, and then making sure that we're sharing that information in public venues. Were sharing it with regulatory, regulators where it's needed with legislators, with watershed groups, lake associations, grower to grower sharing that. We're making sure that people understand. We are looking at it. We are taking a look at it. We are actually measuring it. And we do have something to share and to show you about what we've found. So to that end, we'll talk a little bit later also. We've also created a Facebook page a Twitter page and a a webpage where folks can go to take a look at some of that information and learn more about what we're doing at what needs to be done. With all that. As my long preamble, Let's talk about the monitoring wells and what we are finding. We put monitoring wells in irrigated fields within a couple of hundred feet of the irrigation wells on purpose. Because we want to know we wanna be able to directly measure the impact, the interference effects when that well, that mug or that irrigation while kicks on. Each one of those will give us the ability to run an aquifer test. So we'll actually characterize the aquifer, the transmissivity, the conductivity, the weakens value, storage value, the parameters that we need to do predictions. And then we have the water level data. So we'll see the trends during the irrigation season. The annual trends. What are the highs, What are the lows? How much does it vary? How is it connected to precipitation? And we'll see the long-term trends from year to year to year. If the MI-WWAT is predicting five model, five years. Down the road or what depletion is going to look like. We should be able to see that in this long-term data. And is it varying? How does it vary? 2012 was dry We would expect to see lower groundwater levels. 2018, 19, 20 were very wet. 21, 22, a little bit more towards normal. This year is a little bit of a toss up. You heard about that earlier. So what do we see and what where are we doing this app? We have right now 186 monitoring wells installed. We put one in in January. We are already slated to get to 240 in 2023, and we're not done. We continue to get calls from growers that are interested in installing additional monitoring wells. And as you can see, we're starting to cover more and more of the state. So we've got, we've got a ton of resources to look at, which covers a large variety of geologic environments at various depths. In various types of formation, both bedrock and  glacial as well. So we're really covering the gambit and getting a good handle on how as water responding to irrigation in natural impacts from year to year to year across a variety of geologic terrain and across a variety of geographical terrain as well. We use pressure transducers. They are accurate to about an eighth of an inch to collect the water level measurements every 15 min in these wells, we verify those with hand measurements at least quarterly in each well so we can keep track that the transducers are collecting good information and we check the flow rates. I know there's pivot specs assigned to pivots that the flow rates are measured in the field. So we know what actually is being pumped. Just because a well is equipped with a 500 gallons a minute pump doesn't mean it's pumping 500 gallons a minute. Just because your pivot spec says that you should be putting 450 gallons a minute on through your system. Doesn't mean that that's actually what's happening in real life. So we double-check those numbers with our ultrasonic flow meter. So here's what we're finding in these are some long-term records for us to take a look at it. This is in South Eastern Cass County, Michigan. We plot precipitation on the bottom. So you can see when the rain came and how dense or lacking it is. Then up here we've got about five foot with 1 ft tick marks or the gray lines are 1 ft. So in this case, he's only 100 ft roughly away from irrigation well. And when that well pumps at about 800 gallons a minute, the water levels are drawing down. About 8 ft is off. But interestingly, we can see different patterns. We can see groundwater declines beginning before the irrigation season starts. Sometimes we see the groundwater levels are rising during the irrigation season. Again here last year, there was a pretty significant drop before the irrigation season even started. But long term, do we see depletion of the aquifer system? We just don't see it back in January of 14 we're at 795, in January of 2023 we're at 796. So we're actually a foot higher than we were in 2014. Here's another example, and again, these have been all updated. We just did downloads back here in January of 15 we're at 771. And here we are today. We're slightly below in January, within about 3" of that same mark here in January. But you can clearly see when the wet years occurred, when the more average years occurred. But we can also see a very long-term decline right here that happens to be associated with a significant change in precipitation. They are directly correlated to each other. So these graphs gives us a tremendous amount of information about both the seasonal and the long term water-level trends. And it gives us the ability to monitor for any depletion. One more here. This is also quite an old one that we have. This goes back to 2014. So about nine-year period of record. This one is a little more dramatic. We're actually a little over 2 ft higher today than we were back in 2014. But again, dispels a lot of this, does a lot of misconceptions. He's running this system is running at almost 1,000 gallons a minute and he's only producing three-and-a-half feet to draw down 130 ft away. So very little minimal drawdown on this particular one here. And it's very interesting to, to go back and look at, see how does the irrigation relate to water level trends. And so is it going up? Is it going down? Is it causing depletion? Is it, does it take a long time to recover when the  irrigation stops? And we're just not seeing the impacts that the model is predicting at this point in the groundwater system. So another, to go back to this example here. If we have a three-layer system, in some of the fields we've installed what's called a nested set. We'll put a monitoring well in the upper aquifer. and we'll put one in the lower aquifer. So here's an example out of Cass County as well. This is the deeper aquifer here you can see the irrigation was clearly screened in the deeper aquifer. And you can see the pumping cycle is occurring. What about long-term trends? Long-term trends 746. And then we're up to almost 747 as of this month. And the same thing in the shallow aquifer or about half a foot higher than the shallow aquifer that we were in 2017. So we've got pretty pretty stable, really stable groundwater conditions. You can clearly see it's reacting to precipitation events both near and far as they go through the system. But of also particular importance is do we see pumping from the lower aquifer? Impact is shallow and I prefer if I zoom in on that curve a little bit. Here's a pretty strong pumping cycle. It we're just not seeing the impact in the upper aquifer makes it really difficult. When you see a graph like this to say, Well, pumping the deep one must impact the upper one, which then must be attached to you and pump the local stream. This is pretty clear example and pretty concrete proof that that lower aquifer is not impacting the upper aquifer they're hydraulically disconnected. One of the things that we check in one of those individual pumping curves then the backup, this would be an individual pumping curve. Is we see the drawdown occur in it. It is pumped for about 2,600 min or so. And then we turn the well-off and then we watched to see they only watch to see how long does it take to recover. And we've observed full recovery about 2,000 min after we turn the well-off. And so we can also see internist does this recovery portion of the graph, turn it upside down and plot it right over the top of the drawdown and you can see how they react to each other or relate to each other. You can see the variation. They track really, really well on top of each other in almost every single situation that we have a monitoring well in, we see this same pattern. If you pump for 10 hours it recovers within 10 hours. You pumped for 48, it recovers within that 48 hour period. And we're seeing drawdown curves being mimicked by the recovery curves. So they're behaving the same way both during the pumping and during the recovery periods. It's another critical piece of information. We can also take that then and do a curve analysis on the drawdown or the recovery. And that gives us all the parameters I was talking about as we go to look at the model, input parameters into the model. So in this situation at those well fields where we have monitoring wells, we have that data. We don't have to take it from a table. We don't have to estimate it from a well log. We're actually measuring it in that field. We know what those parameters are for those geologic layers. Then we can go ahead and do our own predictions based on those. So this is a very simplified cone of depression. At 500 ft, 1,000 ft and 2,000 ft. We'll turn the well on and run it as though it's in drought conditions. So we'll run it for three times the normal period of time that it would run. And then we can plot out on the map or on an area of photo, what is the prediction? What is the drawdown predictions that we should see at distance from that well. So at the end of the day, we really know what the impact is. And so we've gone a long ways to building trust. When we say, Look, we've been measuring this. Here it is. Here's how we did the work. Here's the equipment we use. It's valid, it's done to industry standards or above. And here's the results of what we've, we've measured. You can observe patterns. You can directly measure what the impact is. We can build a very simple 2D model now and say, hey, we can predict inside this cone of depression, we can very reasonably predict what the drawdown and the impacts on the aquifer system and the other layers or even stream may be. So we really do at the end of the day know what the impact is at those well fields. So here's a shortlist. Take a quick screenshot if you're online. But here's a shortlist of what the monitoring wells can do for you. And if you look at these items here and you start thinking, how are these items going to help in a water conflict or in a resolution or in some sort of a mediation. This information is really, really good and critical to working towards building trust, especially when things get to the contentious level. We also go into the streams. I'll touch on those and we'll look at those and what are we finding there? We measure two things in streams. We're measuring stream discharge, how much water is flowing through that stream. And we will do that at three to four times at each site. Throughout the year. Last year we were in 400 streams sites and we had almost 550 measurements. Expect that number of measurements this year to about double to 1,000 and will probably be up around six or 700 stream sites. So obviously were we're have some monitoring wells coming in the central portion of the state. We're going to add that in. And we've done a lot of work in the south east portion of the state as well. So we're really emphasizing a lot of the areas where we did see a lot of the primary areas that had Crop Irrigation, a lot of the high capacity water withdrawals. So what are we finding in the streams? So again, this goes back to 2014. And if we look at the overall trend, we see the trend is up, knocked down. So in this case we have a stream where we are actually showing an increase in discharge. If the assumption is that we're pulling groundwater away, assuming precipitation is relatively stable across this time period, we should see water levels declining. If we were to only look at the second half of this graph, we would say, wow, we have a system that's definitely in decline until we factor in 18, 19, 20 were very wet years. We had much dryer, more average years, the last two. And so we see elevated levels here. We see more normal water levels here also corresponds to 14, 15, 16, 17 pretty well. But overall the trend is upward on our water levels in streams. We also collect stream temperature here. So we can actually check against the stream temperatures as well. And you can see just from the, all the dots, there's the trend line. If we look at all the data as a whole, but also all the dots are where we take manual measurements. So you can see where the manual measurements are and you can also see how often we take with the green dots discharge measurements. So we're verifying all of this stuff out in the field on a very regular basis. Another stream, this one is over in northern St. Joe County. And I would maintain we're seeing the same trends a few wetter years in 18, 19 and 20. But here again, we do see rising water level rather than falling trend in the stage here as well. Then the last one here, this is one we're keeping a much closer eye on. This is exactly what I was referencing. If I backup to this graph here, I said, what if we only were to look at the second half of this graph? We would say it was a water-level decline. That's exactly what I'm demonstrating here. The period of record goes 18to 23 are 3-period wettest years, 18, 19, and 20 We got more normal precipitation in 21 and 22. So correspondingly we see a little bit lower water levels, but we are keeping an eye on this one. We're all, so again, watching stream temperature at each one of these gauges as well. One of the other things we do with the discharge, this is Dowiagic Creek in Cass again, Southwest Michigan is on one day. If we pick the purple line here and May 20 of 22, we went and we started at the headwater here and we did six measurements on the same day, from upstream to downstream. So here we have a pattern of a stream gaining water over its entire length. We go back and we do it in September. Near the end of the irrigation season. Are we seeing the same pattern of gaining? And are we seeing the same amount of gain? So obviously the red line and the blue line both here represent September events. And so we see slight discrepancies between them, but they're actually within a measurement error of the instrument. We're seeing in gaining situation throughout the whole season. I would be awfully concerned if what we saw was we saw sudden this trend of gaining, really start to flatten out, not see the same linear trends. Here we have a tailing off, but we also have a tailing off here. I would be more alarmed if I saw that we had a reversal of trends that now this gate, this stream is not gaining during irrigation season. That would be a clear indication that hey, there's something going on here. We need to keep a close eye on this because it's just not behaving the same way. So we do this at all the streams that we monitor as well. So we have these gain-loss surveys all across the state. Also. Most of our streams have at least five gain-loss surveys done. Some of them have closer to ten or 11 at this point. Again, we've got a long history. Year to year to year, season to season of measuring real impacts. So again, this is pretty definitive data that look, we're just not seeing the changes that we should expect to see in our gain-loss surveys if there is truly depletion occurring out there, it's one of the indicators. Last thing we do in the streams I mentioned before, we check the stream bed themselves. If you have a clay bottom stream, there's a disconnect between the aquifer in the stream itself that stream that water is just going to flow across the ground surface. So as it does that, I don't want to run out of time here, Lyndon, you have to let me know if I got to start wrapping it up. As the clay as the water with the clay blind bed  would flow across that surface and really not interact with the underlying groundwater system. We see a whole mixture and I'll show you a couple of pictures in a second here. But we collect these conductivity tests Also at all the places that we go visit. These are current. And we're obviously adding, expectoing double this number by the end of this summer. So we haven't even plotted our most recent ones over in Hillsdale County and in Branch County. In 2022, we ended up with a total of 350 sites that we had tested. Like I said, I expect that number to double this year. We've got a lot of work in front of us. But at the same time, it's becoming, it's just getting bigger and bigger. We have more people joining the group and the group is expanding. And I think people are really buying into and appreciating what value that data brings to the table. Stream beds themselves. These are pretty typical of what we see. This would be more of a sandy mix with some organics thrown into it. Water will go through this. This is not a really great Sand like the underlying here on the bank. But we see a lot of this in the streams. Typically the only section of the stream that is really sandy and gravel is the high flow area. The banks are often very silty, very muddy. So when we test stream bed conductivity, we're always in the active portion of the channel, so we're being the most conservative. But what I mean by that is we're allowing the most water to go across that stream bed. We're just assuming the entire stream bed then conducts it that way. That will be a conservative estimate because the irrigation to obviously have the biggest impact if the stream that is the most conductive. So we're erring on the conservative side. On the other side, what we see over here is you see this very frequently to, in an area where we're flowing across the glacial till or an old in marine, we see a lot of clay in the stream beds or very dense till material. Very often again, there's organics imbedded in it. In some cases, there's so much debris in the stream bed, there's so much lead and organic that it's truly a very layered system. And water is just not getting through that there. So we use a vertical tube seepage meter. We have a very specific construction of that meter. So it's same. Everywhere we go. We have very specific protocols we follow for each time out. And we actually measure water recovery every second so that we get it very quick. And very short-term. We've got tests that recover in five-minutes. And so they're very, very high. But then we've also got, we've run a couple of tests that we've left the metering for two weeks and we've had zero recovery. So it's variable across the state. We can't just apply a simple number to it. In the MI-WWAT, the stream bed conductivity is assumed to be one-tenth of the aquifer conductivity. And that's just not a reasonable number to put into the models. If we want a better picture on an individual well, so we do this in streams. We've also started doing this in some of the lakes around the state. Sometimes will wade in, sometimes we'll do it from off the edge of a pontoon. And we'll do the same methodology, we'll drive it into the lake bed and do the conductivity testing there as well. Then here, the reason for this picture here is we will seal off the top of the tube, but when we extract it from the bed, we get an institute sample. And here clearly we were in, we were in a clay environment here. We had a really nice clay. But this is what we do at those sites. So education, I'm gonna take a quick little break here, which is really cool for me. A great experience this past summer, we had a grower in Kalkaska, Michigan who worked with us and we brought 50 FFA kids out into, FFA kids out into the farm field. We had an irrigation well set up with a monitoring well. Clearly we put our equipment in their hands. They were measuring water levels, they were measuring drawdown recovery. We had them out walking underneath the pivots, right up to the pivots. They were running our flow checks, checking to see how much the well was pumping, how much how much water was coming through the pivot. It was just a really great field day for them, very, very practical. They got to see everything we did live out in the field. We even showed them on a computer live ahead, had it set up here. They were actually able to see the drawdown occur. As it happened, it was directly connected to the transducer and the well. And they were able to observe the drawdown in the recovery. So it was a great day out in the field. The second day they came back and we went into one of the streams right next to the field. And they ran our flow instruments. They ran their own stream bed conductivity testing. They were in their waiters in and out of the streams in groups and just had an absolute blast out in the field and it really opened a bunch of their eyes. I was really excited the teacher called me later he said he had three students that said, Hey, I want to go into hydrogeology now, it was so much fun. I enjoyed what you guys were doing and that was a really interesting field day, but it was a great educational opportunity. We've hosted these field days. We did a field day for EGLE staff. We've done a field day for legislators. We've done field days for conservation districts. We've done them for growers. We even did it for a technical committee for the Cass County pilot project. We love bringing people out into the field so that now we're talking in common terms and we're understanding each other better. So that's, that's about midwest water, that's about what we're learning and what we're seeing. I'm going to end by showing you two sort of case studies. In a sense, the Prairie River Watershed Lyndon mentioned before, the red portion here of the River and its tributaries was up to be reclassified from a warm stream or river stream to a cold transitional. And then there's also a future proposed regulation or change in reclassification of the downstream end of it as well. In this area, if this had happened, there are about 70 registered water withdrawals since 2008. Estimate that the number is probably double of the actual number that are out there. Whether year 2008 or newer or whether you're pre 2008 doesn't matter when it comes to the water conflict because everybody has access to that water. And so we've all got to share it. At some point in time, we probably would have ended up with about 140 water withdrawals that would have to come to the table and decide how to share water and what impacts were. We looked at the DNR reports, and we looked at their classifications. We looked at their temperature logging and their fish studies. And then we graphed it up differently than was in the report to try to get a better visual. This is one of almost identical to a figure that they generated in their, in their reports. And all three of them that showed they had 19 temperature login stations in this watershed. Seven of them were on the Prairie River itself. 12 of them turned out around tributaries. That wasn't, they actually installed it in 2012, which they didn't know, but there was a drought coming in, so 2012 was a clear outlier, 13, 14, they were still out there. In 2015. They reduce the sampling network to three sites on the Prairie River and one into two tributaries. We're down to five. Then in 2016, it was reduced to two sites. In those two sites, there's no tributaries, but those three sites there were different than the previous year. Then when they ran the stream check tool, they did fish studies here, Orland road and fish study at bowers. They did have a temperature logger here at station six. Didn't have a temperature logger at this station but they use one upstream and downstream. We found issues here where air temperatures weren't monitoring stream side. They were monitored in Lansing or three rivers. Instead of being monitored stream side. We also found that at I believe two of these, one of the stations, nor looking at average means, we're looking at mean July temperatures. We found at one of the stations that the loggers weren't installed until July 20, which negated getting a July mean temperature. So we found kind of a mismatch with some things. There was a few of them are located immediately downstream of the tributary. There was a few of them that were located immediately adjacent to a wetland. They had cited EPA protocols, but there was a number of those protocols that they had cited that weren't followed. There really wasn't a clear explanation of the why. And so there was just some issues that we saw with the data. The other thing that we saw when we looked at the stream check was we have temperature loggers here in 2000 I'm sorry. We had fish surveys in 2000. Sorry,  still got there. The temperatures we did have temperatures in 2014.in 16, at sites 6 and 7. But both of those sites in 2015 were the two sites with a loggers were deployed in July 20, so those weren't valid. Then at site 11, there's temperature loggers in 13, 14, 15. So we had a mismatch in dates and temperature loggers at these two sites. So there's just some inherent mismatches that if you look at it, you're kinda like, why was it is it explained? Was it explained? It really wasn't. And so it really did bring some of the brought the work into question. As of last week, I had heard that reclassification has been postponed and the DNR is considering or at least planning at this point of doing three additional years of data collection before they attempt to do another reclassification for or attempt to reclassify this. So it clearly sparked a lot of interest. A group, one group formed and they went out and hired a lawyer and were getting ready to sue. We were contacted by a number of growers in the area asking for some help and what to do both now and into the future. So there's some future work that we're gonna be doing out there to complement some of this. One of the big things that happened was up in this area up here in St. Joe County. We actually had a stream gauge and we had a stream gauge from 2017 to 2017. So we had good water temperatures at that gauge. It was shielded, radiation protected, there were verified temperatures. So everything was done according to good protocol. What we found was the three years that DNR looked at for transition happened to be the three coldest years out of the eight year period of record. And so we had a mismatch where we were able to clearly show that the July mean temperatures is moving around. This is clearly not a stable cold transitional stream. It's clearly not staying within a three-and-a-half degrees temperature range. That data, again, very, very useful when that's put on the table. They also found USGS had well, we'll talk about the monitoring well first. We had a monitoring well right near that gauge. This is a background monitoring well, so it's not near an irrigation system. And from 2010 through 2022, here's the period of record and you can see clearly again, we have rising water levels again that are being demonstrated and measured in this monitoring while USGS has a stream gauge on M66, and the Prairie River  this goes from 1962 and is current. And again, the trend line is up. It's not, we're not seeing the depletion occurring that the model has been predicting. The last one that I'll go through quickly here. We talked about the MI-WWAT having a single layer system. This is a site-specific review that I worked on. And this is the monitoring well record that we put in. You can see clearly we have sand and gravel aquifer at the top, 24 ft of clay than another aquifer than 19 ft of clay, another aquifer than 8 ft of clay. And then the irrigation wells actually screened in this lower aquifer. So we have a seven layer at least system here. This is what it looks like in cross-sectional view. So there's a stream right here. We had a nested set wells. Here's the upper aquifer. Here's what we believe is a discontinuous aquitard. Then we had a monitoring well screen below it. And then we believe is a continuous aquitard here. Then another aquifer. Then another aquitard, and then the irrigation well screen down here. So we had data, we ran a full-scale aqua tests. We had data from both the upper aquifer and the lower aquifer. In the MI-WWAT. This would be a single layer system. In 2018, we helped get a new law approved that added another addition to the political system here. This is the most complicated one that can be used offline in the batch tool, where it's a three-layer system. So we made this work into a three-layer system. We put the data from our nested set where we had an upper and a lower aquifer. We put the data from our nested set into that analytical model and we got a good curve match. We were able to show the drawdown data mimicked both in the deeper aquifer and the shallow aquifer that we got a good curve match can see here this is a tenth of a day. So this is, this is only about 15 min. This is the first 15 min of the test. Then this jumps up to a full day right here. So we turn it off after a full day of pumping. This is the predicted depletion using that model. You can see again here, after about 15 min in, we got to one day and we turned the well-off. The dashed line here is what the model is predicting. The model is predicting that it will run out. And we will be up around ten days before we see for recovery. This right here is about a quarter of an inch. And so it's inside the measurement here. So the model is predicting it'll take about ten days. However, in the field we observed, it fully recovered within 24 hours we had a discrepancy between what the model is predicting would happen and what we are measuring in the field. And so our argument was here. Here's the information, here's how it fits the curves. Here's what the predictions are. Here's how the predicted recovery occurs. And then at the end of the day, this one will still denied. And the reason for that was the theory presented in the meeting by EGLE was that even though it had recovered here it still hadn't recovered 3,000 ft away. Which is a little concerning, confusing to me because they thought, well, do I need to monitor 3,000 ft away? And if I do, which direction do I go? North, south, east, or west of here? Or do I need four of them for additional monitoring wells? And once they get 3,000 ft away, now I'm not going to have to talk to other neighbors and try to locate property because his fields not 3,000 ft in radius here. So there's still some mismatch that's going on here, but I think there's great lessons to be learned from looking at this and having these types of conversations. Um, because at the end of the day, we're coming all the way back in full circle. Can we trust the model? Does the model accurately predict what happens in the field? And if we want to start talking about sharing water or restricting water use, what is our basis of that discussion? If the basis is a model, can we trust the model? Or are we going to need to have more of this type of real-world data that to take the heat out of the room and say, Hey, look, this is what we see out there. So I appreciate the fact you have a model, but if your model is not matching here, now we're having a little bit different discussion.