 Sarah works for the U.S. Fish and Wildlife Service as a refugee biologist at Morse Wetland Management District. Morse is in Minnesota, for those who don't know Morse is at. She coordinates the biology program at the district, including inventory and monitoring, research and planning. Sarah received her BA degree in biology and environmental studies from Lawrence University in Wisconsin and a master's of science degree in wildlife and fishery sciences from South Dakota State University. When she isn't at work in the prairie and in the wetlands, Sarah enjoys spending time with her family, exploring public lands and gardening and cooking. So Sarah's topic today is timing, prescribed fire in tall grass prairie based on plant phenology. So let's welcome Sarah. Well, I feel like I need to manage some expectations after Sean was hyping up this presentation so much yesterday. We, Kami and Jill and Jen and I have been working on this topic of timing management in our tall grass prairies for the lifetime of the NPM program, which is about 12 years now. We don't have any answers. So if you were hoping for the big aha, that's not really gonna happen. But what we do have, I think is a really good story to tell and a journey that we've been on for the last several years. So that's what I just really wanna share with you today. So Kami gave a good overview yesterday of what AMPAM or Native Prairie Adaptive Management Program is all about. So I'm not gonna get too much into the details but just as a reminder, we've got, excuse me, the purple dots on here are all of our management units that are enrolled. It's US Fish and Wildlife Service land for the most part, all remnant prairies, native prairies. We have sites throughout the prairie pothole region, which is that green area on the map. But again, some of them are in the mixed grass prairie, some of them are in the tall grass prairie. Kami was talking about mixed grass stuff yesterday and we'll be focusing today on tall grass. So everything east of that yellow dashed line. And AMPAM is again, a decision support tool that's intended to help guide our managers. So it's not a research project. That's one thing that I think is important for people to keep in mind. The idea is not to necessarily answer all the questions but to provide guidance to our managers based on the current state of their prairie, what the optimal management tool is to apply in the next year. For the, we have separate models that are built into the tall grass and the mixed grass prairies for this project. And so within the tall grass, we have a bit more of a focus on timing of management related to smooth brome that is not included in the mixed grass prairie. So why did we do that? When we first started AMPAM, we really didn't intend to have two separate models or model sets. We were gonna have the whole project be just all together. But as we got to talking and thinking about it, there was this Wilson and Stubedick model out there, this provisional model in the literature that a lot of our managers were using and we're just like fully on board and kind of it was just incorporated into our, sort of what we thought of as our common sense knowledge of how we were managing our prairies. So the idea with this Wilson and Stubedick model is that there's this sort of magical window in the spring. We call it the spring window often. So you'll probably hear me refer to that. That was really looking at the period between tiller elongation and fluorescence for smooth brome as being the best time to burn a prairie. And specifically they were focused on tall grass prairies that had at least a 20% warm season native grass component. And so the reason that this isn't included in the mixed grass prairies, of course, because that doesn't have that warm season, heavy warm season component like our tall grass prairies do. So since this model was out there, it was so widely used and accepted among our land managers, we thought it made sense to actually build this into our models that we used for the tall grass prairie. One thing I wanna point out is, and this was touched on a little bit yesterday, was that in their paper, they for some reason thought that recognizing or identifying when there were above ground nodes on brome tillers was gonna be real hard for land managers to do. I'm not super sure why that is. It doesn't feel difficult to me. I don't know if they just didn't think people wanted to crawl around in the grass or what, but so their recommendation in that paper was to use the five leaf stage. So five leaves on the plant as a surrogate for nodes as identifying when elongation was starting. So just keep that in mind as we keep going here. Ignoring that one. So what we did then was, remember, Kami was talking about for the mixed grass, we had four management alternatives that are in the suite of recommendations for our managers. For mixed grass, it's burn graze, rest and a combo burn graze. So what was different for tall grass was because we were so just all on board with this idea of that spring window being the ideal time and really wanting to push our managers to do the best right thing. We thought that's what that was. So we actually incorporated that spring window right into our management recommendations. So we have rest again, the same as the mixed grass folks do, but then instead of just grazing and burning, what we have is graze in the spring window, burn in the spring window, or we have a fourth option, which is called defoliate, which means burning or grazing anytime outside that window from elongation to inflorescence or hanging at any time. We had a lot of folks that were interested in using hanging as a management alternative. So if we were gonna use those that window to guide our management, we also needed to come up with some rules, which is what's shown on the right here as far as how we would actually define and identify that window when it was occurring out there. And so you'll see that, I did not advance. So you'll see that we did focus on this five leaf, stage as being that surrogate for the above ground nodes, because that again was what was recommended in that paper and that was what most of our managers were doing. They thought they were going out and checking for five leaves before they would burn. And then the window would end when most of the brom on that, on any given unit had inflorescences. So we found some issues right away. I'm gonna back up so you start looking at that. So we found some issues right away. One was that we're a bunch of wildlife biologists who are not range managers. Most folks who are managing land for the Fish and Wildlife Service are wildlife biologists by training or ecologists, conservation biologists, not range scientists, not botanists. That's changing a little bit. But so counting five leaves seemed pretty straightforward to a lot of us. But when it came right down to it, what do you account as a leaves? That there are rules, I'm sure that's pretty familiar and seems like a silly question for a lot of folks in this room who do have that range background. But for us, as we got to talking with our different participants, we realized that there was a lot of confusion out there about what actually counted as a leaf. How are we gonna know when we get to five leaves? So that was an issue, just being consistent. When we got our first year or two of data back from people about when they were identifying that spring window happening, there were some really odd variability. And I mean, we would expect there to be a lot of variation, right? As you go across, do you remember how big that map is? Like from Southern Minnesota all the way up to Northern North Dakota in Minnesota, that's a big range. You'd expect there to be a lot of variability and when that window was happening in, like by dates in spring. But what was happening was it was very unexpected variability and hard to explain variability. So it was like we'd have stuff way up north that the window would pop up and then down south. And it was just not, there wasn't a pattern that made any sense. And so that was when we really started to talk and think through maybe what we were doing wrong as observers again, right? Because recognizing that we're wildlife biologists nest more so than botanists, maybe we needed to do a better job of training our folks. So first thing we did was brought on Justin Dupay who got his masters at SDSU with Andy Smart and he put together a nice little training publication for us that really helps lay out very explicitly, very consistently how to count leaves, what counts as a leaf, when does it count as a leaf? And put together a photographic user guide for us that our folks could use to go out in the spring. He had a couple of technicians that ran around to a bunch of different sites and really felt like that did a good job when we went through that training and use that tool. We really were being a lot more consistent with how we were identifying that window. So that was a good first step. However, we were still having trouble with that window not being very consistent and not being very predictable. And the other issue was that we were starting to lose some of the enthusiasm from our participants because to be able to burn in the window, you needed to know exactly when that was happening. And it required a lot of trips out to their management units every couple of days to check on the progression of that roam that's got pretty time consuming and frustrating during a pretty busy time of the year for us. So next we brought on Sean and Lisa Preister and had her PhD here in 2018 or so. So around that same kind of 2015, she did a project where she was able to really get down obviously hands and knees out in the prairie and do some pretty intensive girl staging on the broom plants out there on our MPAM units and some other sites and had a really nice correlation between accumulated growing degree days and smooth broom. This was really helpful for us both for that efficiency thing that I was talking about. Now because we had these accumulated growing degree day values, we could kind of wait until we got close to those numbers that she identified before we started going out to visit our sites and check for elongation. So that was real helpful. The other thing that really was very valuable to us that came out of her work was that, yeah, we were right, that five leaf stage was not the way to go. She was seeing what some of us had been seeing that quite often those above ground nodes were there and detectable well before the broom plants were hitting five leaves, often three to four leaves. There were, it was pretty common for a broom to not even get to five leaves and also pretty common for a broom to not ever get inflorescence or have enough inflorescence to hit that rule that we'd come up with for what the end of the window was. So having Lisa actually go out in the field for a couple of growing seasons and just really get to know some of those populations of broom was real helpful for helping us think through some of that spring stuff. Another kind of just cool side note, Jen Zorna, our data manager, put together this spreadsheet for us that used accumulated growing degree days at the closest using data from the closest weather station to each of our management units and put this little tracking spreadsheet, kind of live tracking spreadsheet together for us. So we had a nice consistent way to keep track of when we were reaching that sort of early part of when elongation might be happening and we could just kind of keep an eye on this and then run out to our site. So that was pretty neat. However, so we thought we were, I feel like this whole story is like, we did this, but then however this happened. So that's gonna be my transition between every chunk here. So we felt like we had a lot of those sort of logistical things and just consistency things figured out with how we were observing all of this stuff, feeling pretty good about things we touched base with all of our participants in the tall grass for MPAM. And they were like, yeah, let's try to keep going with this, we'll move forward and see what we can do. Keep trying to target that window for fire and grazing. We tweaked our rules just a little bit based on what Lisa had found. And so you'll see here, it's still, we're still looking at like a start and end of the window, but instead of those five leaves, we really were gonna look for nodes so that we got together with everyone and help them understand what that all meant. And then the end of the window because we did have so much trouble with a lot of our brome never reaching inflorescence or there'd be just very few plants that would have inflorescence on them. We ended up coming up with this kind of best professional judgment. I don't wanna say it was totally arbitrary. We did a lot of thinking and looked in the literature and looked at some of our other data from earlier in the project and decided that we were gonna just say that window is 15 days after it starts. So again, all of this is like not based on calendar now, we're basing this off of accumulated growing degree days and presence of nodes. Okay, so all that was going on, we were feeling pretty good about where we were. Jill, our analytical person, then right around that same time we had kind of hit this moment where we had 10 years of data from this project and it seemed like a good time to go back. Remember, Kami was talking yesterday about that three-legged stool of MPAM where we have our adaptive management program that's helping our managers make decisions. We have targeted research, which is like what Justin and Lisa did with us. And then we have these retrospective analyses which is a real benefit from having the standardized monitoring protocol being applied across all these sites that we can then also, in addition to using the data to make management recommendations, we can use that data to answer some of these questions and look back at, maybe with a more kind of traditional statistical approach to think about some of it. So Jill did that kind of right around the same time that we were thinking, okay, we're ready, we're gonna just go ahead and forge ahead. And then we started seeing some pretty surprising things in our data. So just a couple, a little bit of context again, we're talking about as I work through some of these results from Jill. Again, this is that one year time step that Kami had mentioned yesterday. So we go out and monitor, we do our management and then we go out and monitor again. And so that cycle is a time step as a data point for this. Remember I talked about, we changed that rule of what the window was. Jill went back and reapplied our current rules for what the window was. So that was consistent across all of these because that is an important aspect that we're looking at. And then just wanna make sure that you kind of just keep in mind there are some real limitations with the data that we have because this isn't a research project, all of these sites are submitted by participants, whichever ones they wanna put into the project. We do have some sample size issues here. So there's a handful of those that are pulled out there. You're gonna see some of these, I'm gonna step back here so folks can see all of these. You're gonna see some of these abbreviations that are over in the key on the right. G for graze, B for burn, obviously. O-W means outside the window. So that's anything in the year that doesn't fall between elongation and inflorescence. And then WW is within the window. And then we do look at all of this based on what the dominant invader is on the site. And so smooth brome, Kentucky bluegrass, the CEO is co-dominant smooth brome and bluegrass because we are so lucky that we have both on a lot of units. And then RM is remainder. That's any other invasive plants that might be present out there that didn't rise to the level of really building models around but we did wanna keep track of it. For us, that's usually things like quackgrass, red top, some woody invaders. So this is just some context for some of the many box plots that I'm about to show you. Just look, again, keep in mind that there are some low sample sizes in some of these and that's just something that we need to kind of keep in mind. So real similar to the box plots that Kami was showing yesterday, I wanna look at my notes to make sure that I don't forget to say anything here. So these figures have the treatment along the X axis. So again, the OW is outside the window. WW is inside the window. The label on top of any of these is gonna show you which invader or plant community is dominant. That dashed line that's in the middle of it there is our starting value of cover averaged across all of the units for that individual plant. So like in our case, we had about 37% cover smooth brome when we were starting. So then if the box is above that dashed line it means there was an increase in cover. If the box falls below that dashed line there's a decrease in cover, okay? And then we've got figures here for brome, bluegrass that remainder and then native query. And one thing that Jill wants me to make sure that I point out very clearly is that all of those have to add up to 100 just the way that our data is. And so when you see an NP for native query what that really is is 100% minus whatever cover for brome minus whatever cover for bluegrass and remainder, okay? So just another kind of context thing, I guess. Okay, so all of that to say we thought we had things figured out we thought our biggest problems were just these logistical things of not knowing how to count leaves and having a lot of confusion about what nodes were and all that kind of stuff. But there was also just some problems with the data that we found when Jill did some of these analyses. So we realized pretty quickly after Kami and I had a little bit of crying together about it all after spending all this time thinking we knew what we were talking about we made some incorrect assumptions just like baseline baked into all of our models for the tallgrass prairie and Pamunus. So first assumption was that Wilson and Steubenick were right or that their provisional model worked for the Northern tallgrass prairies doesn't seem to be the case. So there's lots of stuff. I'm just gonna ask that we think about the red here, the burning. So we'll, according to Wilson and Steubenick burning inside that magical window should decrease our brome increase our native cover, right? And you'll see that we're actually seeing the opposite here. So we, this is both, both of these are for sites that are smooth brome dominant the left one is the responsive brome the white one is right one is the responsive native prairie. And you can see that when we burn outside of the window, we're getting a decrease in smooth brome cover. And when we burn inside the window we're getting an increase in smooth brome cover. So not only was it like, I mean, so that was a little stressful. Then when we started looking a little further there were a few other assumptions that we got to thinking about that we made on the other was a next one was that we thought we should be able to extend that brome window concept to Kentucky bluegrass. Now we recognize this was a little bit of a stretch but there wasn't a lot of data out there in the literature about Kentucky bluegrass and timing for management. But, you know, I mean, kind of made sense. They're both cool season, rhizomatous invasive grasses. We've got them a lot on the same site. So we'll just go with it. Now, again, just thinking about burning that assumption was not correct. The timing doesn't seem to make a big difference with Kentucky bluegrass. So these are sites that are dominated by Kentucky bluegrass and their left figure shows the response of bluegrass, the right figure shows the response of native prairie. Again, Jill wants me to point out as many times as I can that that's not just native prairie but that's also kind of accounting for all those other, you know, just because it's a bluegrass dominated site the native prairie result is also accounting for brome and other invasives. But anyway, there's no difference in burning outside the window or inside the window. Although it is a little bit heartening to see that burning does seem to decrease our Kentucky bluegrass. It's just that that timing doesn't seem to be an important factor, okay? Assumption number three, that was sort of a little bit of myth busting that we did that we could extend that window that Wilson and Schumann talked about to grazing. They were talking about fire really specifically in that project. Again, we knew this was a little bit of a stretch and we had to come up with all these crazy rules to count, you know, when a graze would count as being in the window or out of the window because obviously a fire is sort of a one day event usually. And so that's easy to say that one day happens in the window or not grazing extends for many days or weeks. And so that was a little bit complicated. But again, didn't work. The green big bars here are the grazes. So this is again that same exact figure that I showed for the first assumption when I was focused on fire. So smooth brome dominated sites and the response of brome on the left and the response of native prairie on the right. And you can see that grazing did seem to maybe decrease our brome a little bit, but the timing didn't seem to be important. Finally, the last assumption that we had sort of built into all of our progress models was that anything outside the window would be equivalent. This just shows our confidence in that window concept, right? So we were so sure that burning inside the window was the ticket that we were happy to say that maybe an April 1st burn was gonna be the same as a September 1st burn. Which now looking back, I was kind of a dumb thing to think, but we were just really all in on that Wilson and Stoomadick model. And so again here, just so now I'm highlighting the outside the window treatments. So the green is the graze, the red is the burn. And you can see again, that is certainly not the case. So grazing outside the window and burning outside the window, for example, had very different effects on the brome cover out there, right? Okay, so that was disheartening. We still were pretty sure that the window was right, that we were wrong. Like maybe we were just thinking about that window wrong. And it did one thing we did notice is that it ended up being a very narrow timeframe for us up here in this part of the world, maybe compared to where Wilson and Stoomadick were doing most of their work down in Nebraska, right at Meade. And so we thought maybe we just had sort of been defining that window incorrectly. So we tried to kind of slice and dice the data in a few different ways. Spoiler alert, it didn't work. But so just kind of again some context and sample size issues that I just wanna kind of have us all keep in mind, this is the distribution of our burns and our grazes in the tall grass, about 40 different sites. You can, this is the don't worry too much about the X axis, it's just, this is broken out by a warm season growing degree days, just cause that was how Jill was doing some analysis. But what you can see is that, oops, I did it. You can see that most of our burns were happening early in the growing season. That's because our managers were trying to do what we were telling them to do and hit that burn window. So they were at least trying that. But anyway, just to sort of give you some context, just that's, again, it's not a research project where we had assigned timing. It was a part of adaptive management is that that flexibility is built right in for the managers to still be making their decisions. Okay, so first thought was, it was very silly to think that outside the window is totally equivalent, that that in the window is just always gonna be better. So our first thought was, okay, maybe we need to split that early season up, split that outside the window up so that we would be looking at the vegetative stage separately from the elongation stage, separately from the reproductive stage. So we, instead of in and out the window, what we decided to look at were, what about our sites that were burned early in the spring, later in the spring or summer and onward? Again, some sample size issues. So you can see we had, there's 382 total data points, but very few of those later burns, because again, we were encouraging people to burn early. And so that's what they were really trying to accomplish for us. So just sample size context. So this does not take the dominance aspect into account anymore. So this is just all of the sites combined together. So those dash lines do represent a co-dominant state. This didn't solve our problem. So burn one again is, if you can't, oh, I'm sorry, I know it's hard from being in the audience, it's hard to see the X axis when it's weighed down there like that. So the first bar is gonna be rest. The green bar is grazing anytime. And then we were just looking at burning. So the red bar, that's the first one is burn one, which is early spring. The second one is burn two, which is late spring. And the third one is burn three, which is summer and on. Okay. So you can see the upper left graph is showing us the response of smooth brome. What we were kind of hoping was that maybe that early spring being combined with summer as outside the window originally was muddling things up and that if we could split that out, maybe we would see that there actually was an effective timing. And that didn't really seem to be the case. Although again, I will just like as a ray of sunshine show you that burning doesn't really seem to matter when you do it for Kentucky bluegrass, it does seem to help. Okay, that didn't work. So then we decided to phone a friend. We were freaking out. Does that seem fair, Kami? At that point, it's like, this isn't working. This isn't working. We have this whole model set all of these years and all of these participants really committed to MPAM. We didn't wanna just be like, hey, didn't work and walk away. So we're really trying to figure out what the right thing is to do. So we brought together a bunch of experts which many of them are in this room and I'm sure our familiar names for lots of you. And what we asked them to do was, okay, we thought Wilson and Stu Medica was right, but that didn't seem to be the case for us. So like set that aside in an imaginary prairie where the starting state maybe is like 25% cover of smooth brome. If we burned during the vegetative stage of that brome, what do you think the outcome would be? It would be increase or decrease, okay? So that's what this figure is showing for sake of anonymity, we don't know who these different colors represent but these are our different experts and their responses. And you can see that there's a little bit of variability in our expert's opinion about when we might see the best effect from burning brome, which was kind of funny but also kind of reassuring for us because we were having all of these struggles thinking we knew what was going on and didn't. So I'm just gonna zip through these. What Jill did was used what we knew about the growth stages and accumulated growing degree days. This is cool season and kind of the average for when we thought our experts thought that smooth brome would be most impacted. And this is, we also asked them about big bluestem. I don't make these funny, fun animation things. And when the best time would be to burn to not hurt our warm season grasses and then overlaid those because of course, cool season and warm season growing degree days aren't the same. And in the end, there is this window that we thought our experts were helping us kind of identify as a good time to burn when we're gonna hurt brome and we're gonna not hurt our warm seasons. We did the same thing for grazing and I'm running out of time so I'm just gonna have to zip through this really quickly. I apologize for that but happy to talk about it again later. So we basically kind of came up with these new ideas of like a good time to burn, a bad time to burn, a good time to graze, a bad time to graze. Took our data, reassigned them into those categories. Again, some sample size issues. We didn't have as many grazes and burns in what our experts had identified as the best time. And again, timing didn't seem to be the thing. It wasn't the most important thing. And so you can see smooth brome up on the top left there. I'll just hit on that one for the sake of time. And so the green or gray is a rest. The first green is a well-timed graze. The second is a poorly timed graze. Well-timed burn, poorly timed burn doesn't seem to matter too much for smooth brome. Graze are complicated. That's just the, it's a realistic thing that we have to think about. Kami, I love this analogy always says we need to think about it as plain chest not checkers. There's a lot of strategy, a lot of thinking. We do feel like confident in saying that active management is important even though we can't pinpoint the timing. So for example, we did see that fire was beneficial in the case of Kentucky bluegrass. We do see overall a common upward trend for our native prairie cover. So managing, managing, managing does seem to be helping but maybe timing isn't the most important consideration. And now we're in another whole big process of trying to figure out, well, what is, how should we guide our managers? So there's ecological lessons and then there's also life lessons for all of us. There is, I think this, the biggest thing is that there is real benefit in having a standardized protocol that all of us were using. If we didn't have that, we wouldn't have had the opportunity to have Jill go back and reassess some of these things and help us answer some of those questions. So I can't say enough, even though it can be disheartening and we're not getting the answers that we think are correct or helpful that there's a lot of value in being able to look at it in a lot of different ways and answer those. Challenge your assumptions just cause somebody says there's big foot, don't go with it. We thought that that five leaf stage was the ticket and everybody just did it and they thought it was right and it turned out maybe not so much. And then I would also say, especially for some of the younger folks in the room, use the experts that are out there. It was really comforting. Again, I will say when we got that figure from all of our experts having very different ideas about and I will say they all felt real uncomfortable answering that question. Like none of them wanted to answer that question for us of when the best time was to burn or graze because it is sort of an unknown. So in a lot of ways, I wish we would have had that conversation earlier in the whole process, but don't be afraid to reach out to the experts that are out there and start forming those relationships and leaning on them when you have crisis moments in your paradigm. The end.