The root cause of unexpected data patterns

In the early years of soil moisture measurement, anomalies in soil moisture data may have been easier to attribute to faulty sensors. Now that sensor rate of failure is ~0.5%, how do you explain deviations from expected data patterns? In this 30-minute webinar, followed by a ~15-minute Q&A, research scientist Dr. Colin Campbell will discuss unexpected data patterns and how to determine potential root causes.

Understand what’s happening at your site

Join Dr. Campbell as he walks you through data set examples, including:

• How temperature impacts readings
• Diurnal patterns of soil moisture distribution
• Infiltration and drying events, when to expect them, and how to understand what changes might not be attributed to those events
• Soil moisture behavior under saturation conditions or permanent wilting point
• How soil moisture varies with depth

Next steps

• Link to slides
• Watch Part 1: How to Interpret Soil Moisture Data
• Article: The researcher’s complete guide to soil moisture
• Article: Why soil moisture sensors can’t tell you everything you need to know
• Related webinar: 8 Research-ruining data mistakes
• Request a ZENTRA Cloud live demo

Questions?

Our scientists have decades of experience helping researchers and growers measure the soil-plant-atmosphere continuum.

• Talk to an expert
• Request a quote

Transcript

BRAD NEWBOLD
Hello everyone and welcome to part two of how to interpret soil moisture data. Today’s presentation will be about 40 minutes, followed by about 10 minutes of Q&A with our presenter, Dr. Colin Campbell, whom I will introduce in just a moment. But before we start, we’ve got a couple of housekeeping items. First, we want this webinar to be interactive. So we encourage you to submit any and all questions in the questions pane, and we’ll be keeping track of these for the Q&A session toward the end. Second, if you want us to go back or repeat something you missed, don’t worry. We will be sending around a recording of the webinar via email within the next three to five business days.

BRAD NEWBOLD
All right. With all of that out of the way, let’s get started. Today we’ll hear from Dr. Colin Campbell, who will discuss how to interpret the data you collect from soil moisture sensors to achieve the most thorough and accurate conclusions possible. Dr. Campbell has been a research scientist at METER for nearly 25 years following his PhD at Texas A&M University in soil physics, and is currently serving as president of METER. He is also adjunct faculty with the Department of Crop and Soil Sciences at Washington State University, where he co teaches environmental biophysics, a class he took over from his father, Gaylon, about 25 years ago. Dr. Campbell’s early research focused on field scale measurements of CO2 and water vapor flux, but has shifted toward moisture and heat flow instrumentation for the soil plant atmosphere continuum. So without further ado, I’ll hand it over to Colin to get us started.

COLIN CAMPBELL
Yeah, thanks, Brad. It’s great to be with all of you today, and exploring this topic a little more closely — interpreting soil moisture data. This is part two — hopefully, you had a chance to watch the virtual seminar we gave a couple years back on part one — I don’t think there’s going to be a ton of crossover here, but maybe a few. But in general, what I’m wanting to do today is act like you and I are friends, we’re sitting down, maybe we’re at a conference together, we’re sitting down, and you’re bringing some of your data to me and saying, “Hey, what exactly is going on with this?” And so I’ve got a bunch of soil moisture data here, and we’re just gonna go through it all. Now most of the stuff I’m going to show you are part of different experiments that that I’m a part of — a lot of different places here in the United States, particularly in the western half the United States — and I’m not going to be able to go deeply into what’s going on in each experiment. So what I would tell you is, if you’re interested in any area that you see here, and you haven’t seen a virtual seminar on that from us in the past, go ahead and ask, and we can queue you up a virtual seminar in the future kind of going more in depth into exactly what we’re trying to accomplish in any one of these experiments. And that’s one of the reasons it’s so great to work as a part of this team here at METER, is we get to do real science every day, and we’re making sensors and systems that are going to help you do the same thing. So we kind of do what you do.

COLIN CAMPBELL
Without further ado, let’s jump into the topic. My dad loves to tell a little joke to us. So here, here’s how the joke goes. So there’s a geotechnical engineer and a geologist out standing by some geological formations, and the geotechnical engineer says, “If I hadn’t seen it, I wouldn’t have believed it,” and the geologist says, “If I hadn’t believed it, I wouldn’t have seen it!” And sometimes when we look at soil moisture data, we have to understand a little bit about what’s going on, and things do actually exist to really believe that they’re happening in our dataset. So maybe we have to wear a little bit of the hat of the geologist, as we sit and talk today. Now, again, we’re gonna go through a smorgasbord of data, and the thing we’re going to do first is look at some customer data. And I’m just going to show this to your raw. So here’s a dataset that all of these just came into our support group, we have an amazing support group here at METER, they’re always helping you to try to get your best data out of the systems that we produce. This is one of the things that that we got, actually, fairly recently into our support area, just said, “Hey, we got a sensor here, and it seems to be working quite nicely, And then suddenly, it’s bouncing around all over the place. What’s going on?” You know, what I would say about this data set is that we have something failed in the sensor. Now this happens to be a very old sensor. It’s one I don’t think we even build very much anymore. But this has been in the field a long time. I don’t know what exactly happened, but I do know that it’s not behaving as it should. Maybe there’s a failed component on there. Things do happen.

COLIN CAMPBELL
So another dataset, we got sent in and this is a little while back, but sometimes in the field unexpected things happen, and one of the traditional ways of talking about unexpected thing is like, well, there’s a lightning strike. Well, this was an actual lightning strike that happened at this research site. And from sensors that were working quite well our TEROS, 12 sensors, these are our brand new offering our kind of top tier offering that I use all the time, these data are not good data, and presumably on the y-axis, we’ve got water content, and we have some time variable, just customer graph data, and what we see is that the sensor is dropping out all the time. And we had a lightning strike and something must have gone wrong with the circuitry because of that strike. What I would say here is, by the way, if you’re going to set up your system on a Campbell Scientific data logger — this was done on a Campbell Scientific CR 1000, those are awesome data loggers — we love Campbell, Scientific — please go ahead and read METER Group’s lightning search and grounding practices. It’s a white paper we’ve got on our website.

COLIN CAMPBELL
And so here’s one, this is an ongoing experiment. It’s pretty old now though, I think we started this experiment probably can help save you this particular problem that’s really eight or nine years ago, with some colleagues where we’re unique to people using third party data loggers with our looking at invasive species, rodents, fire, several of these system. So it doesn’t happen very much. But that’s something things out in a desert location where the ground cover with plants is fairly sparse. So this is a great place to start that happened to this particular customer. Here’s another thing talking about temperature impacts, because not only does that that we got from a customer, they said, “Hey, it get pretty hot in the day, and cool down at night there, but also we buried sensors very near the surface, and there’s what’s going on here, we can see the temperature going up and not a lot of plant coverage. So perfect opportunity to look at down, that’s in orange, and then we got a water content that temperature effects on sensors.

COLIN CAMPBELL
Okay, so here’s the typical temperature behavior, the water content sensor is exposed to looks a lot like temperature in blue.” And if you go and take a extreme temperature will show small fluctuations. Now when little peak to peak calculation there, if you divide the water you’re looking at this graph, you might be thinking initially, hey, wait a second, that does not look like small fluctuations content changes by the temperature changes, it comes in water content, those look big. Well, I want to turn your out to almost 1% water content per degree C and I can tell you, attention to first the left hand, y-axis, soil temperature, we’re going to look at some of these data and how it actually that’s ranging, just a little less than 20 degrees, maybe 18 degrees, peak to peak, 16 degrees peak to peak out there, affects sensors in the field — that’s not common, it should be that is soil temperature at five centimeters depth. On the right about at least an order of magnitude less than that. But hand side we have water content, and that’s ranging only from about 4.8% up to 5.5% or so.

COLIN CAMPBELL
So essentially, I’ve blown up the when you see this kind of thing, there’s probably something going water content axis to show you some of the things that happened wrong with a sensor and something that needs to be with temperature. And clearly there’s a correlation between looked at a little bit more clearly. Because that’s not the two just visually — I didn’t make a scatterplot. But this is, you know, looking at it visually there’s a expected, we shouldn’t see such clear patterns of temperature on correlation. But in the end, this is more expected the water content sensor, there could be something else going temperature change. It’s about five hundreds of a percent per degree C five hundreds of a percent water content and Under on. Sometimes that’s possible in the field, but I’d be a little debt down there at the bottom, that’s more of what we’d expect suspect of these data. So the answer to a lot of these to see, whenever you bury sensors, so close to surface, I mean five centimeters, that’s pretty close to the questions is, hey, when problems happen, call METER support. At surface. And we don’t have very much plant coverage there. If the end of the presentation, we’ll have a contact email for any at all, we probably are talking about bare soil surface. And in the end, what’s going on there, as we just get hot them, please get ahold of those guys, and they can help you and cold there, that maybe water movement, you know, we get maybe through any of your challenges. Really what we’re going to be get some deposition at night, who knows, but it may be just kind of something that’s going on with directly related to talking about today, though, is how to make sense of data during temperature.

COLIN CAMPBELL
Okay, so I want to blow this same chart up, now the other 99.5% of the time when you’re out there and you don’t we’re looking at the exact same data, except we’re looking over a six month period, again, soil temperature on the left hand y have problems, but weird things might be happening, or just axis water content on the right hand y-axis now going from a standard things that you really don’t understand how to range of 0 to 20%. And this is what we see in terms of water interpret. So that’s our goal here, and we’re going to jump in content at five centimeters depth and a desert. Again, same exact soil, we’re just looking at more data. And now we see and just go through a bunch of cases. Remember, this is kind of trends that make a little more sense, we’re going to come back like drinking from a firehose, if this gets overwhelming at any to this graph later and look a little bit more closely at what’s generating those bumps, you’re probably assuming that point, just understand that what we’d like to do is have you look they’re some kind of precipitation, that’s what I’m at some of the other virtual seminars where we actually talk assuming, but we’re gonna make sure that’s really true.

COLIN CAMPBELL
But we can see here, how much diurnal changes get imposed on that about the experiments that these sensors are used in. Most of water content signal at five centimeters over a six month these are actually stuff that I’m working on with colleagues. period of time. If we drop deeper in the soil, we still see a little temperature related change. But now down here at 10 centimeters. So in the green, we’ve got temperature, ranging between, let’s say, 21, and 25. So about a four degree change in temperature, at 10 centimeters depth, and then on the right hand y-axis, we got water content, but instead of the comparatively larger one at five centimeters, this range, it’s going up and down is about two tenths of a percent. So very small changes, again, the temperature sensitivity is even slightly smaller, approximately the same, it’s about for 4.5 hundredths or we’ll round up to five hundredths of a percent water content per degree C. So we can see that there is a certain amount of sensitivity of temperature, but it’s actually extremely low. And when we get back to these data, you’ll see that that at 10, and certainly 20 and 30 centimeters deep, you don’t seem that much at all. But that’s kind of what to expect from temperature behavior of water content measurements in the soil.

COLIN CAMPBELL
Okay, let’s go on to another example. So this is in a silt loam soil, it’s actually buried here in our area, we have a silt loam soil here, we do freeze in the winter, we typically get a nice layering of snow, and it kind of comes and goes and a lot of what we’re seeing here — so that green line that’s kind of going up and down, that’s water content. But let me draw your eyes to the time here, we’re going from December, down to really close to right now in March. And what we see is, is a lot of, you know, we’re — sorry, that green line said it was water content, it’s temperature, the blue too, lighter and darker blue, those are those are water content — we see the temperature just going up and down in the range, on the left hand y-axis is only zero to six degrees Celsius. So not a big range. The reason I put this graph up here was not to show you how much temperature changes at five centimeters in the soil, but to focus you in on what happens when that temperature drops down to zero to the water content.

COLIN CAMPBELL
So I’ve shaded a blue area there. And what happens is this interesting thing, so this darker blue line that’s at 20 centimeters, the lighter blue line is water content at five centimeters. And when we get that freezing event, the water content just drops down to around 15% for where it was at about 35%. Now, to be clear, that is not a true reading of water content when the soil freezes. That water ostensibly disappears to a dielectric measurement. The reason is the dielectric of liquid water is way, way higher than the dielectric of ice. And so when that drops, we’re just seeing a lot of ice in the measurement area, or the volume of the sensor. You cannot depend on that measurement and you should summarily throw those data out. And that sounds really firm, but I actually reviewed a paper within the last year where they actually were talking about the frozen soil water content in Canada, and they didn’t even realize it was frozen. The temperature was below zero, the water content dropped from 30% down to like 7% it was clearly frozen, and their idea was to use techniques to kind of model that the frozen soil, or frozen water in the soil — don’t do that, it doesn’t make any sense, you’ve got to throw these data out. So that’s just — so what would we do here — buy a sensor with a temperature sensor on there and a water content measurement, and when the temperature drops into freezing range, and you get a drop out in the water content, you got to get rid of those data.

COLIN CAMPBELL
Okay, so some other strange things. So this is a research site I have. This is a 3000 meters on in a Montaigne environment. This is right at treeline, we just have these little patches of trees in these tree islands were studying the climate change impacts on these tree islands in the meadow areas. And these are the water content data we get. So in an orange and purple, that’s temperature, that’s soil temperature and it flatlines for a long time because we got snow cover there. And then it starts to pop up and you can see kind of normal temperature behavior. And so what we’re wanting to look at here is two different situations, the one in gray, and the one in orange, that’s water content in gray. Orange is the temperature in the meadow. Dark blue is the water content. And purple is the temperature inside one of these tree island areas. And this giant spike that happens around the middle of May, that might just really surprise you if you don’t understand what’s going on. We jumped from about 50% water content up to 40% water content. You might say oh my gosh, no rain event could really do that to the soil, we get runoff, we get all kinds of stuff. Well, it’s not a rain event. It’s snowmelt. And one of the reasons we know it’s snow melt is because we have a game camera up there, that’s taking pictures three times a day and I know what’s going on, I have a little snow stick, I see the melting snow. And now once we understand that metadata about the area, now we can understand what’s happening that the meadow it’s getting a lot of solar radiation, the gray line is water content is dropping fast toward the end of May. And at the same time the blue that’s in the island that holds for a little while because we got the snow drifting up against the trees there. It’s blocking some of the solar radiation with shade. And we just hold water there much longer. So big deal to know exactly what’s going on at that site, just by happening to have a game camera up there.

COLIN CAMPBELL
Okay, another question: air gaps. This is one of the biggest challenges we get to good soil moisture measurement. And what happens what do we see when we get air gaps? Well, we had this great experiment that just happened to highlight this — wasn’t necessarily on purpose. But we know there are air gaps around a sensor. And I’m going to show you what that looks like. So this is in a silt loam again, we have volumetric water content on the y-axis with time, over a couple month period, it looks like here on the x-axis. And we have water content data at five centimeters here, the orange line has no air gaps, the blue line has air gaps, it’s pretty easy to see that a lot that happens with that blue line looks similar to the orange line, except that a early spring mid spring water content that sitting down at less than 5% in a Palouse silt loam just doesn’t make any sense. We know that soil is wet here during the spring. So we know that the orange line has to be reading better. Not exactly sure why the water content on the air gap sensor, per se might be reading low until it gives one of the telltale signs of an air gap sensor. There in around the 15th of April. So there hopefully you can see my cursor we get this just jump from 5% to over almost 60% And then a fast drain back down to 5%. That is a great signal that you’ve got an air gap in there because the water just fills that air gap and then it drains away quickly.

COLIN CAMPBELL
If we look at 10 centimeters deep with the exact same x and y-axis here. Same timeframe we’re a little bit less error gappy — is that a word — is a little less air gaps there but we see the same thing we see down around the 15th we get that big rain event the 10 centimeters sensor that has no air gaps shows a little bit of that event, but man the water just fills around that air gap sensor jumps up not quite as high up above 40% And then jumps way back down, get a little bit better contact where the water content is around 10% But it should be more like 30% And so that’s a really great sign when it’s reading quite a bit lower than we might expect. And also we get these nice large spikes pretty easy to see. You might have problem, there probably is an air gap around your sensor. Now, if you did this well, which we did it 20 centimeters, things work out quite nicely. Look how well those agree. Now I said, quote an air gap sensor in the blue. Technically, it’s not because that got installed well and didn’t have any air gaps around the sensor. So there was a rain event, we did see a little bit of a jump here a little bit higher than the orange sensor, not much. I think we have great contact there and they’re reading together well, and they’re more in the range that we’d expect they’re a little bit deeper. So no surprise that we’re sitting about 35% water content. So better insulation better data, that’s how to recognize that when it’s not telling you things that are reasonable, start asking questions, especially if you get those really high jumps, somehow waters getting to that sensor, maybe preferential flow, maybe air gaps. Okay. So those are all kind of things happening in the soil that are, I don’t know, that may be unexpected.

COLIN CAMPBELL
Dealing with maybe your site, let’s talk about some of the impacts of plants and what they have on the measurement of water content. This is a picture years and years ago, so it’s 15 years ago that we did this, we installed 42 sites around an agricultural rain fed wheat field, not so far from where we are at METER. And we were pretty surprised at the root water uptake, I’d never actually measured it back at that time. It was less common to make water content measurements. And I’d never seen data like this. So this is rain fed winter wheat. And again, we’re measuring water content on the on the y-axis. And now this time, we’re going across a summer — so starting in about April and ending at the end of July. Now by the way, if you’re watching this outside of the US sorry about the way dates are shown here, I know that the day and the month are switched, but bear with us. That’s our format here. And that’s the way I always graph it. So now let’s look at — so we have labeled here 30 centimeter sensor, that’s a 30 centimeters deep and 60 centimeters deep, etc.

COLIN CAMPBELL
And I’m going to just call these out with highlights. So the 30 centimeter sensor, when we actually buried the system, we already started to have water uptake by this wheat crop at the 30 centimeter level, because that’s relatively late in the year. You know, we started in early May, late April, early May. And you know, the wheat crop has already taken up some of the soil moisture started 35%. And it’s going down. As we start looking though at 60 centimeter that starts higher up kind of what we’d expect to be fairly wet down there around 40% water content. And as the 30% tray starts to flatten out lower water availability in terms of the content 60 centimeters starts and is that drops down low, the 90 centimeters starts and is it drops off the 120 centimeter stops. And as it kind of bottoms out. Now the 150 centimeter doesn’t do anything. It’s just stays relatively fat. So does that mean kind of in the middle of July that the wheat we just kind of gives up and hits maturity? Well, it could mean that we didn’t have a camera that I could show you exactly what what’s going on out there. But we dug down in that soil to see Well, why didn’t actually take any water from 150 centimeters, those roots seem really robust, reaching so deep. Well it turns out between 120 and 150 there’s a thin but very rigid hardpan layer that the roots can’t get to. And in fact, it actually helped us understand the reason why — I was really surprised it started up at 50% water content just to start the season. The reason was, is we actually have a perched water table up at 120 centimeters over this hardpan layer. So there’s this great reservoir of water for the wheat to go down and get and you can see it it just sucked it dry. So a little surprising that the wheat roots went down that far. But when we started looking at these data over a season now made a lot of sense, especially with that hardpan layer.

COLIN CAMPBELL
Here’s another — so this is actually relatively recent. This was just a year a couple of years back, this is irrigated hemp. This is here in our region, drip irrigated, similar soil to what I just showed you. But this was way under irrigated — it was irrigated, but they didn’t realize how much you need to replace. ET was around six they only replaced it with around two centimeters of water per day. So the crop was pretty thirsty for water and had to get its roots deep. Here is it 15 centimeters so you can see it goes down and then the irrigator knowing that there really behind in the water because I told them that we’re having problems they tried to catch up in a Palouse silt loam in the middle of the summer, that’s mid August, that is really, really hard. And you can’t — high evaporative demand, low hydraulic conductivity of soil that went down to 15 centimeters a little bit to 30, but it never reached 60. So 15 centimeters water content, we got 30 centimeters water content, taking up water after the 15th. And finally, that kind of ran out of that. So the roots go down to 60. And take up a bunch of water down there as well. So we can see a kind of a similar behavior with this poorly irrigated hemp crop. Okay, hydraulic redistribution. And by the way, that’s a picture of that site that I just showed you some data from. Hydraulic redistribution, diurnal wavy traces that are not related to temperature. So I already told you if you start seeing waves on your temperature signal that you ought to look at temperature. But this is — maybe I don’t know, often, sometimes, at least regularly enough to pay attention to it — a case of hydraulic redistribution, and maybe you’re sitting there saying, You know what, I don’t believe what you’re saying. All diurnal signals on water content are just temperature signals, I’d like to convince you otherwise.

COLIN CAMPBELL
And so here’s some rain fed data from that same wheat. But at another location in that field, again, water content on the y-axis time, over the summer. Actually, I did this, I just told you about the dates. I did reverse this. This is a paper we published in a Japanese soil science journal. So this is actually May 2nd, all the way to the end of July into August. And I’ve highlighted in blue, just little snippets of data where we see these diurnal wavy patterns that I think we’d assume were related to temperature. But I’d ask you, you know, so on the 30 centimeter, I would say, Okay, you may be right, you know, even though there’s not much here while the crop isn’t taking up water, and then we see it a lot when, when we’re taking up a lot of water, okay, it’s 30 centimeters. And if we go to the graph on the right, which is temperature, over that same period, for each one of those depths, we do see that that there are some diurnal changes during that time. So could I argue really convincingly that it’s not temperature related? No. But if we look at 60, the relatively flat through the early part of the season, when we really start taking up a lot of water and changing the water content, we see really pronounced waviness, also at 90, and also at 150 centimeters, there was no plow layer, there or this hard pan layer. And before that absolutely flat. And if we look over these data, none of the data past 30 centimeters really show this this diurnal variation, especially 90 and 150. Don’t have any in there at all.

COLIN CAMPBELL
And so how do you account for these data right there 150 centimeters that are showing diurnal patterns? Well, I account for it in that the roots have grown down that deep, they’re taking up water there, they’ve got a strong pull on this transpiration stream during the day. And at night, that transpiration poll relaxes, and we’ve got some water at least accumulating in the roots, if not moving back a little bit into the soil. It’s not a purposeful effort of the plant to try to put water here or put water there. But it is just a fact that water, when it relaxed in that stream, in my opinion, may be showing up back in those roots and getting back in the soil. So why doesn’t it show up at 120? Even though we have a strong uptake of water? Well, this may suggest that that water still is in the roots, because it’s certainly being taken up at 120. But maybe we simply don’t have any roots around there. I don’t know. I’m not not really sure why it doesn’t show up there. It certainly doesn’t show up on all my data when I’m measuring plants in the field. Just some.

COLIN CAMPBELL
Okay, so maybe you don’t believe me, maybe that 15-year-old research isn’t that convincing for you, so let’s talk about something that we got out of another wheat crop. This is in a completely different area is still a silt loam soil but down in southern Idaho. And here I’ve just highlighted in gray, this is at 65 centimeters, so so very deep temperature changes are not very active down there at 25% water content all the way across the season here into the first part of July. And then that water content starts dropping and then we see stair stepping. Okay, that suggested to me that we’ve got water uptake at that level by the wheat, and we’re just seeing behavior related to that transformational stream stretching and then relaxing each day. See that also, again, we see maybe in all these patterns. But if you say, Oh, the 15 centimeters 30 centimeters, we’re not going to look that that’s too close to surface, that’s fine. But look over the 60 centimeters, we see that diurnal pattern here where there’s certainly no temperature change that at that level on a daily basis, so again, I think just a little bit of that trace transpirational string, you can argue with me, maybe we’ll do a full virtual seminar, if everybody rises up and says, we don’t believe you. And I’ll try to gather more data to convince you, but don’t do that. I want to do other virtual seminars. Okay.

COLIN CAMPBELL
So Okay. What about unique sensor behavior? So right there in the blue sweatshirt is one of my good friends, Ryan. And one day, when I met Ryan, the very first time he’s a really good grower in southern Idaho — grows potatoes and wheat. And I walked up to him and he says, Colin, I’ve been using your sensors for a couple years, and I don’t think they work. And I’m like, Oh, great. Now I was like, Okay, tell me a little bit more about why you don’t think they work. And he says, My sensors always read the same when I bury them under my potatoes. I said, Well, Ryan, let’s, let’s consider that, because it’s always possible that a sensor doesn’t work. But if all your sensors aren’t working, you know, maybe there’s some other explanation. And so we buried sensors in his field, and I watched the traces all through the year, and in this irrigated — variable rate irrigation potato field. He was right. The water content — he wasn’t right about the sensors not working. He was wrong about that. But he was right that the water content didn’t change much. So here water content on the y-axis time, from June to the end of August on the x-axis. And look at the water content. These are six locations all in the same center pivot irrigated field, and the maximum change over the whole season was 3%. And he’s like, Colin, how am I supposed to schedule my irrigation based on that? And I said, Well, Ryan, those sensors are working. But let’s take a look at other measurements he had in the soil. And luckily, on that experiment, it was just at the start. So this was 2018, about five years ago, just when I was starting to do work, co-locating water potential sensors with water content sensors.

COLIN CAMPBELL
And now here was the real data that was interesting. This is the same field. Now we’re looking at the same time range, but we’re looking at the available water, the energy state of water, which tells us whether or not the water is available for plant growth, or it’s difficult for plants to get out. And so I’ve drawn a picture — kind of colors across this graph. If you’ve watched other virtual seminars, you’ve seen this picture before. The green part means you got optimal water, we have three sites reading in the optimal range. The yellow and red are the stressed and the permanent wilting areas. And we had three in that range. And he came to me one day, when I said, Hey, I think we got some stressed areas in your field. They said, No, I went and kind of grabbed a handful of soil. And it seems good to me. And I said, well, the sensors that you’ve got your water content sensors, you’re right there, they’re pretty flat, but their water potential sensors are getting in a stress range. We agreed to disagree at this this point, because he felt like things were going just fine. But he had a really accurate yield monitor on his potato harvester, the end of season, we looked at the data. And in fact, the places that we said were stressed yielded 25% less potatoes than the non stressed areas. Immediately he was like, Okay, now you’ve converted me and he puts it in all his fields now, and it’s super happy with this analysis. But the takeoff point here is that, hey, sometimes water content sensors have a unique behavior.

COLIN CAMPBELL
And one of the things you can do is add water potential sensors to make sure to understand what’s going on with this very critical thing we call the moisture release curve. It’s the relationship between the amount of water or the water content and the energy state of water, the matric or water potential. This is a very important relationship and tells you something about your soil that’s unique to your soil. Every soil has a unique soil moisture release curve. Okay, we’re just going to quickly throw a couple of these up here, we actually show how we can create this in the field and in the lab and how they match up pretty well. So here’s one for a sandy soil — water content and on the y-axis, we have matric potential, which is a negative value in my little negative kPa it kind of dropped there. That’s that little black line. And this is this is not a log scale on the matric potential it often is the next graph it will be this is just a straight linear scale. And there’s lots of water available up at the top of that water content percent at the very high or near zero matric potential values. So it’s available to the plants but it easily drains below the root zone. And essentially, that often takes nutrients, other things out of from where the roots can get at it and down to our aquifers, which is not a great thing. But on the other side as that matric potential decreases more and more, let’s say below negative 100, kilopascals and below, that’s low water availability, and the soil height holds on to it more and more tightly as that drops. The art of irrigation, the art of kind of growing plants is to keep this in the range, right about here from about negative 10 to negative 100 kilopascals. Where the plants can take it up. It’s not draining, but it’s not too hard for the plant to grab. It’s a great thing to actually know.

COLIN CAMPBELL
Okay, here’s some data from a silt loam soil, we did the same thing. The solid line is it lab generated curve. The blue dots are a field generated curve, again, water content on the y-axis and matric potential water potential on the x-axis, this time, this is a log scale. But now we’ve defined that for the silt loam, these are very, very different if I kind of clicked forward and back, you see how different they are, they are unique to the soil and help us understand how that soil is — how much water the soil will retain at certain availability levels, which is vital if you’re wanting plants to grow in it or to understand the availability.

COLIN CAMPBELL
Okay, the last thing we’re going to cover in this this virtual seminar is, just talk about what is typical soil moisture sensor behavior. I’ve taken you through all these different things to talk about this problem and that problem, this behavior and that behavior. But what can we expect, we just have a normal day out there? And I’ve got three examples. From various experiments, I get to look at tons of these, what am I, my friends, Dean, and I sat down at the American Geophysical Union show and just looked through a ton of these data that he had taken. I think he’s got 1,000 soil moisture sensors out there. So we got lots of things to choose from, I’m only going to take a few here. Maybe we can do it again, if you want to see more. But here I go. This is back to that experiment in the desert. It’s a native desert soil. Sandy loam, low plant density. As I mentioned, this, again, is a drone picture, those weird little bar things, those are our rain-out shelters, we’re trying to understand what the impact will be to vegetation, invasive species, etc. If we change our rain patterns, with climate change. Okay, in this desert soil, this is during the summer, going from May all the way to the end of October, so technically into the fall. And also spring, six months. You can see the diurnal temperature patterns there that we already discussed. But we do see these spikes coming along. Let’s say right at the beginning of August right here, we got another one right here. And then we see a few odd bumps that make us wonder what’s going on out there.

COLIN CAMPBELL
So if we start looking deeper in the soil, so that’s five centimeters, really impacted by straight up evaporation from the surface. Okay, going on, this is at 10 centimeters and maybe you’re a little surprised that it’s not doesn’t seem to be as sensitive to these potential rain events that are coming in. We don’t know if they’re rain events. But that’s maybe our guess, we see that that 10 centimeter does go up a little bit. But we also see during the summertime, it’s that evaporative demand is just driving water out of that topsoil, five centimeter soil, so a lot of it’s not making it down, although some is, and that’s indicated by our 20 centimeter sensor. This is really interesting, that 20 centimeter sensor is sitting at about 20% water content. And even a little above, through those, what we presumed to be rain events. There in August, we actually see the 20 centimeters sensor increasing. What that tells me is that because we are maintaining that water so high, we don’t have a root zone that’s going down there. And there could be plants on the surface, but chances are they’re things like cheatgrass that are very shallow rooting, very quickly take up spring moisture and die and put on seeds so they can reproduce for the next year. And so we actually do have some significant moisture at that level that plants could use. If they were they were actually there to grab that. So here’s one of the great things that I’ve got an experiment. I’ve got an Atmos 41 all in one weather station and so I do have precipitation. Once I pop that on the graph, I think your eyes should be drawn to all the times we’ve got rain and the soil moisture sensors respond.

COLIN CAMPBELL
One thing that you ought to do in every experiment, in my opinion is throw a weather station out there. So that you know what’s going on. It’s so helpful in data interpretation. And sometimes years ago, when these weren’t as readily available when I didn’t have an Atmos 41 to throw out anywhere that was so simple, I would just try to guess at it grab a weather station from a few kilometers away. But it doesn’t really help especially in this situation, where I’m guessing most of these were little small passing thunderstorms that just hit our site. And here’s the soil temperature at 30 centimeters, just to give you a feel for what’s going on, one thing that you’ll notice is important take home is that that temperature is changing over a decent range eight to up to 24 degrees, that overall the water content does not respond to kind of that yearly change in temperature, which suggests those small blips in temperature going up and down are related something else than this, this long term change. So there’s no kind of arcing water content trend to see in those data.

COLIN CAMPBELL
Okay, a couple more sports turf. We’ve got turf grass, sand, heavily irrigated. The data here, this was irrigated by calendar and by hand watering. And one thing that is for sure, is that our optimum moisture at 10% meant that these that we were way over irrigating this field, the consequence is just that we’re wasting water, that’s not great. But we’re also wasting fertilizer and fertilizer costs a lot of money. And so what we’re trying to do in this situation is help these guys who have a really challenging job to try to dial that in better. And so what we’re seeing there is just all these irrigation events across the time from first of July, up to the 18th of July, I think we are irrigating about every other day out here, you could see it at five centimeters, your hand watering occasionally when things didn’t look good. You can see right before the Fourth of July, they irrigated a lot because they were going to have their big Fourth of July celebration on that turf. And they needed to kind of wet it up. So it was it was robust enough for a couple of days of not watering. So in general, there’s some work to do on this. But this is what you’d see in an irrigated situation in sand, lots of quick movement because the water drains out relatively quickly. Okay, so here’s another sports turf. But in this case, we’re in a native silt loam. There’s actually us installing the sensors out in that field.

COLIN CAMPBELL
And this is just a picture of that field. And what it looks like so nice and green. The water applied here is estimated from the ET lost from the previous day. And that shouldn’t be lost on you, you should pick up on some of this information. If you’re working in irrigated areas, ask people questions, because when you look at the data, you’re gonna say, okay, that makes a lot more sense now, because look at how consistent this data is. I didn’t know if you could irrigate very well, just on previous days’ ET. But here, it’s showing one thing: that it’s consistent, right, that we’re sitting around just a little above 30% water content, which we know is a little on the full side for a silt loam. But every day, they irrigate to exactly the same level. That’s the good news. This by the way, these are two, they’re in inches there at 5, 15, and 30 centimeters depth there. So similar to some of the other stuff we’ve done. Now, the other question is, and we’re not going to get into this, it’s in another virtual seminar that is that the right irrigation? Well, the truth of matter is that was way too much. Our water potential works well above the optimum level. We’re trying to dial that in as well. But of course, you’re wanting a green field. So how do you do that? We’re working with these irrigators on that our water potential sensors are just showing that we’ve, we’ve got a little work to do.

COLIN CAMPBELL
Okay, so last couple of slides. So a native montane area, this is the one I was talking about earlier on. Here’s some pictures of that site. On the left side, we’ve got some tree islands at the METER ZL6, that little white thing in the middle of those trees. We’re measuring some photo indices in there. We’re doing some water content right at the edge of the tree island and then in the middle, and then we’re doing it on the left hand side. That’s the meadow that’s in the fall looks really pretty in the spring and early summer. That’s the fall and that’s dried up and you can see why in just a second. So we got a weather station out there helping us know what’s going on. And here’s the data from there. And I already talked about this as water content now on the on the y-axis and this another six month period going from the first of April to the end of September. And now I’ve got three water content traces. The one in yellow, that’s on the outside of the island, like I was just mentioning, the one in green is inside the island and the one in gray, that’s actually out in the meadow.

COLIN CAMPBELL
And once we understand the metadata around where these sensors are located, these data make a ton of sense. So in the gray area, we’ve got this initial snow melt right here, probably got a rain event somewhere in here. And maybe the snow wasn’t all melted, I think it was, I’ve got a game camera up there to tell me whether that’s true. I think we probably got a rain event there. And then we’ve got drying out. And eventually that thing turns brown, right? That you saw that the island holds its water a lot better here on the outside of the island, presume we got dressed, we’ve got a few different melting events here. Although this may be precipitation, I don’t know, I’d have to check the camera. But then we got a fall off. So we’ve got water there quite a bit longer than out in the meadow. And then we’ve got this amazing curve that’s inside the tree island where we get very little radiation so the snowmelt doesn’t melt very fast. Essentially, we just got this gradual snowmelt, gradual increase in water content to a peak, and then it starts falling off. So we’ve got a much longer time, have water available to growth, it doesn’t really kind of bought him out till late July, which is kind of fun. Okay, here’s a little deeper now we only buried one on the outside of the island. So this is just an inner Island sensor here. And in green now instead of 10 centimeters, it’s a 20 centimeters. This is in the meadow and we can see it holds its moisture better mostly because our plants are not super deep rooting. So it does take up that moisture but it takes quite a bit longer. And it’s not so imposed upon by evaporative demand. I’m not saying transpirational demand, just specifically evaporative demand.

COLIN CAMPBELL
Okay, one last thing, just this week, METER Group released the TEROS 54 profile probe. This is something I’m extremely excited about. METER Group held off releasing a profile probe for many years because we didn’t want to install a probe in slurry and just a straight hole, we wanted to figure out a way that we could slice our water content measurements into undisturbed soil. And we’ve done it with this instrument where you drill a pilot hole, and then that sensor will just push through into native soil and measure water content at four levels, 15, 30, 45, and 60. And I just grabbed a few graphs, I accidentally didn’t grab the 30 centimeter depth I’m sorry about that. It’s not because it didn’t look good. But I’m just going to take you — this is in the winter, again, we got some snow melt going on here got a lot going on. And the y-axis here is water content, very focused in we’re just 33% to 42%. So this is kind of zeroed in, just so that you can see the behavior of the sensor at 15 centimeters, from early December to late January, that’s 15 centimeters. Here’s 45 centimeters and here is 60 centimeters. So you can see a comparison here between the TEROS 54 at the same depth and one of our standard TEROS 11 sensors, and you can see the great agreement we get we’re pretty excited about the performance on the sensor and looking forward to using it.

COLIN CAMPBELL
Okay, points to ponder things to think about now we’re done. And I’m just going to shoot these at you. When looking at data use all the information you have at your disposal. Think about that going in — site pictures are awesome. I always go and look at those. You can never — I mean, there probably is a limit to taking site pictures, you know the number. I haven’t reached that yet. I’m always looking back and saying what pictures do I have at that site? So go nuts on that. Look at soil temperature, you can buy temperature along with your water content, we make them in the same sensor, so why not? Use a game camera, they’re great. We’re using them up about that 3000 meter site. Two of them got stolen. That wasn’t very fun. We still got one. They lasted for like three years up there so it’s not terrible. These things happen but I love the data we created from there. And also make site visits you know, with our new data loggers and things we get the data on the cloud, so you don’t have to go out there nearly as much which is wonderful for me. It’s excellent. But when you do make a site visit, you know take pictures, learn about the situation and make notes. Second install a weather station where possible we’ve got Atmos 41 and our brand new Atmos 41W that’s completely wireless. You just throw that up, pull a tab and you’re good to go. This makes it so a lot of your questions about soil moisture should be answered easily by making a weather station measurement.

COLIN CAMPBELL
Okay, the next one is consider installing both water potential and water content sensors together, they give you a moisture release curve, I use that all the time. Maybe not every single location, that’s not as big a deal. But consider doing that, you know, maybe at least one depth in your moisture profile. Next, never use moisture data when soils freeze, it’s just you can’t use it, you got to throw that away. Next, be careful assuming moisture fluctuations are all due to temperature. I showed you that in some cases that was true. But diurnal cycles are sometimes indicative of hydraulic redistribution if the roots are present. So you need to be at the site, you need to take pictures. You need to know enough about that to make that judgment. Finally, don’t assume you know everything about your soil; roots can reach deep into the soil and take up water or — everything even about your plants. Roots will also stop at soil barriers like hardpans, sometimes you got to dig down. And maybe it just takes a probe just to dig just a little hole down to that depth. And you can pull a sample out and see oh, yeah, there’s a hard pan there. So go ahead and look there. And with that, that’s all I got. I don’t know if we have time to take some questions, but I’ll turn it back over to Brad.

BRAD NEWBOLD
All right, okay, thank you, Colin. Yeah, we are at the end of our time, we do have a couple — we’ll take a few questions. We got some time. And also, thank you, for everybody who submitted questions already, we’ve got a ton of them. And there is still time to submit your questions. If we are not able to answer your questions here live with Colin, there will be somebody from our METER Environment team who will be able to reach out to you via the email you registered with. And to be able to answer your question directly. There are quite a few questions in here that are very project specific that people from our team would love to be able to discuss with you. So these first couple questions. There’s a couple of early questions here, Colin, talking about frozen soil. And mainly they’re trying to ask, How do we know when our soil is frozen? So say, for instance, the soil temperature is still above zero? Is the soil frozen? And we’ll talk a little bit more about that. But how would you assess that,

COLIN CAMPBELL
This’d be a fun one just to do a specific virtual seminar on talking about that specifically, because, as you noticed in the data, I showed the temperature was around zero, but it wasn’t below zero. And sometimes in your measurements, you can even say that you expect it to be a little below zero, just because of the latent heat of freezing, you know, it doesn’t necessarily nucleate. And then — maybe it’s not the latent heat of freezing we’re talking about but just that it super cools, down below zero and it takes a certain energy and a nucleation to go up to freezing. So depending on the situation, what I’m looking at is both the behavior, the water content sensor, and the behavior of the temperature sensor, and you’ve got to kind of put those together to get a sense for whether that’s freezing. I also, you know, I use a weather station there, look at the air temperature, if you have a camera, looking at the snow on the surface, a lot of those things are going to lead you to better conclusions about what’s freezing and not freezing. Especially if you got super cold temperatures. If you’ve got snow coverage, very unlikely you’re getting a lot of water movement in the soil, maybe some you always get some but not a lot. And you see a drop in a water content measurement that’s got to be freezing. So that’s what I’d suggest.

BRAD NEWBOLD
All right. There was another question on the flip side for very hot, dry summer seasons, where they’re seeing zero water content. Is that Is that still a possibility?

COLIN CAMPBELL
Yeah so, you know, I’ve seen it before where you got really hot really dry, and you’re near the surface, and you could get zero as a percent. And, typically, that’s because — so we calibrate the sensors in a bulk density of let’s say, around 1.3 grams per centimeter cubed. Now, near the surface, it’s really hard to compact the soil to that level. And let’s say you’ve actually got kind of I don’t know, this is not a technical term — fluffy soil on top low density soil, right? Um, that I can see a situation where, if the soil is not compacted very well, or maybe you’ve got got a sand there. This is a typical situation where you’ve got, you know, at least in our desert soil that’s happening, we maybe we’ve got a little bit different situation than maybe we haven’t have calibrated in what I would suggest is spending a little time doing a specific calibration around that dry point that you’re interested in. And I imagine that when you do that you’ll get a few percent above zero.

BRAD NEWBOLD
All right, another one here is asking about how much can organic soil amendments influence soil moisture?

COLIN CAMPBELL
Yeah, good question. Yeah, that may also be worthy of a nice little virtual seminar talking about some unique soils but, but organic soils, you know, I know you’re expecting this — they’re gonna hold water better than maybe at times than mineral soil, but also that they have a different density, as I mentioned, changes up a little bit the calibration, you know, also the moisture release curve is quite a bit different. And so in those situations, I tend to to want to add those two sensors together. Although sometimes one of the challenges with high organic soils is getting a good contact so water moves into and out of the matrix that something like that TEROS 21 matric potential sensor has on it. So that’s probably something that we ought to answer just a little bit more in depth, but but we do see a change in moisture release curve, and a potential change in calibration.

BRAD NEWBOLD
All right, and maybe we’ll end on this one, just because it’s talking about, you know, the future of soil moisture. And one of one of those new, you know, quote, unquote, sensor types is comparing sensors to satellite data. And what kind of insight do you have, you know, along with that?

COLIN CAMPBELL
Oh, man, that sounds like another virtual seminar. Ah, that would be fun. You know, that’s something that we’re doing quite a bit right now. We have a data science team that works along with us. And we’re doing some evaluation of infield sensors with satellite data and looking at the right trying to find the right indices to, to do well, one of the challenges is that, although our five centimeter sensor that we keep talking about, that is in the depth that satellites are sensitive to, they don’t go very deep. So 5, 6, 7 centimeters is as deep as you’re gonna get with a satellite soil moisture measurement, any one of those indices. And a lot of our measurements are deeper than that. And so, more and more, we’ve got some of these available, we’re trying to do some comparisons, you know, at conferences like HEU, that I mentioned earlier, there’s a lot of papers coming out on remote sensing, especially in soil moisture. And that’s certainly something that we’re thinking deeply about. If — the specific answer to that question of how well do they connect. So far, what we’re seeing is, we’re seeing some correlation when we, you know, some good correlation when we measure it at five centimeters. We’re also trying to understand how can we maybe create a, mid level mid footprint level measurement of soil moisture, using some potential other techniques to try to match the point measurements of water content with a satellite? So what I’d say yeah, totally Brad, it’s a you know, it’s a future of soil moisture measurement. And, and we’re working in that direction, and there’s a lot to learn.

BRAD NEWBOLD
All right. Okay, well, that’s gonna wrap it up for us, we’ve hit the top of the hour. Thank you for joining us today. Thank you for those who have stuck around for the Q&A here at the end as well. We’ve hoped that you have enjoyed this discussion. And thank you again for all your great questions. Again, if we did not get your question here live, we do have them recorded and somebody will be getting back to you to respond directly to your question via email. Also, please consider answering the short survey that will appear after the webinar is finished, just to let us know what types of webinars you’d like to see in the future. And for more information on what you’ve seen today, please visit us at metergroup.com. Finally, look for the recording of today’s presentation in your email. And stay tuned for future METER webinars. Thanks again, stay safe and have a great day.