 My name is Courtney Meyer. I'm a senior staff scientist here at Neon, and I've been responsible since 2010 for the design of the plant productivity biomass and leaf area index sampling that we do on the on the ground. So from the very beginning, I've been wanting to make decisions about that sampling that allows the integration of those datasets with airborne remote sensing datasets. So hopefully that's been successful, and I'm going to show you a little bit today about how that's played out on the ground and what we're doing in terms of the data that we're collecting that will, I hope, be of some use to your community and as well as the more classical sort of terrestrial ecologists. So to start off, Leah mentioned that I am a co-team lead of the FSU team, the fundamental sentinel unit. That's the team that builds the terrestrial observation systems. That's what TOS is right there. And essentially what that is, it's all the sampling, as you can see, it's fairly diverse. It's not just plants and integration with remote sensing. It's a lot of different organisms and what we're called sort of modules. We have a suite of measurements that we make from plants to soils to microbes to birds, infectious agents and so on. And so a small part of that is the stuff that I've been responsible for. So to give you a sense of how this looks on the ground, we have here just an idealized site and we have a hierarchy of plots. So we have, at the highest level of a hierarchy, we've got distributed plots, gradient plots, and tower plots. And so the distributed plots are these sort of white these are the base plots, I should say, these white little symbols here. But there are also mammal grids, which are these little red symbols, and then bird grids. And so we try to co-locate those sampling units on the ground whenever possible, so that we have these spatially co-located measurement streams. And there are some cases where that breaks down. So mosquito points tend to be by themselves, but we're not going to get into those details too much. What's, I think, perhaps most interesting for this community are the base plots of the distributed nature and the gradient plots, because what we're going to do is we're going to employ gradient plots when we get a remote sensing dataset from a site. We'll be able to look at that site spatially and understand where the extremes of a particular measurement are. So the ones that we're making paying the most attention to are the, I want to say, leaf area index, canopy nitrogen, and biomass. Those are going to be the three variables that we don't have that many base plots that we can install. I think we have a maximum of five, but we do have 20 of these distributed plots that we'll be making these measurements in. So we have some flexibility to try and create a good calibration or validation, depending on what measurement you're talking about, of these airborne remote sensing products, that the airborne AOP team, airborne observation platform, is going to be producing. There are also tower plots which support some similar measurements, but those are all clustered here within the air shed of the tower. So they're at the landscape scale, at the site scale, they're a little bit less interesting with respect to the stuff that you guys think about. But they're incredibly important for tying those ground measurements that the terrestrial observation system makes to the tower measurements if you're trying to think about linking NPP and NEE and things like that. So those are all just different components of what NEON does to keep in mind, depending on what your interests may be. And this slide right here is really meant to illustrate the fact that for plants and the measurements that we're talking about that want to link to remote sensing measurements, we have to consider a wide range of environments and biomes and ecosystems when we're thinking about the standardized methods that we need to bring to bear as an observatory. So for plants it's been particularly challenging because you really can't measure biomass of this system the same way you measure biomass of this system. There's no one protocol that you can employ that'll get you there. So instead we're looking at a broader, and this ties into the uncertainty questions that Lee was talking about. I took a broader approach and our aim is to sample these systems to more or less the same level of uncertainty. So using uncertainty to standardize is kind of a strange concept, but that is the only thing that really leapt out that made any sense. So we could say we're going to estimate the biomass of the system to the mean plus or minus 10% if we can everywhere. And if we can't, we'll at least quantify where we are once we have some data from the site. So that's our strategy for uncertainty with respect to these biomass productivity and biogeochemistry measurements. There's some other QAQC type things that we'll get into, I'll mention briefly, but that's mainly how it's going to work. So this is the site on the ground, no longer a schematic. We have a tower right up here. This is in Virginia at the Smithsonian Conservation Biological Institute. And so this site has a long history of research, and so we have to place these plots around the existing research, because we are the same as any other researcher, essentially, at these sites. And you can see then how the distributed base plots of these bright green dots, and they show up all across that landscape. So those are the plots that are going to be generating the data that will be those, you know, LI, canopy, chemistry, biomass, all those things that kind of feed into the marrying of the ground and the remote sensing datasets. So that's just a snapshot of how it ends up working out in real life. It's a fraught process with lots of iteration with the site host and the permitting team and all that stuff. So we can't just place plots wherever we want. It ends up being kind of an interesting dance. So I'm going to spend the rest of the time ignoring what the rest of my illustrious team does and focusing on the plants, because those are mostly what end up integrating with the remote sensing datasets, although not exclusively, to be sure. So at the plot level, we'll be measuring LI percent cover and richness. And LI is measured on the ground with a hemispherical photo approach. So we've got upward and downward facing hemispherical photos so that we'll be able to understand canopy contributions, as well as understory contributions to leaf area index. And that'll give you a nice big plot level number. And in fact, the data that you're going to get on the ground from those cameras is going to be of a higher spatial, well, coarser spatial resolution than you would get from the instruments on the airplane. The airplane is going to be giving you on the order of one to one and a half meters pixels, whereas when you're taking these pictures, you're at the plot scale and you integrate 12 pictures into one number, and that actually gives you more of like a 40 by 40 meter or a 20, 30 by 30 meter number. So let's keep that in mind when you're thinking about the spatial scales of the data. We'll also be measuring canopy traits by species. So we're going to get leaf mass per area for the dominant species in the plots. We're going to get canopy chemistry from those same species. And then for a different subset that overlaps, we'll have phenology data. The rest of those protocols that enable us to understand the total biomass of a plot involve looking at herbaceous biomass, which is in a forested system, not very important, but in grassland systems where we also will be flying and working, it's the major driver of the biomass of that system. So that's why that's there. We'll look at litter production, not so important for this group potentially, but it's a big part of productivity and what we need to measure to get that. In terms of structure, here is probably the bread and butter of what you'd be thinking about. So we'll be measuring the diameters of different kinds of woody vegetation as well as the canopy diameters shown here and then the heights of those individuals. We'll also record taxon ID and we'll be mapping the location also so I'll show you how that works in a second. And then finally, we'll also be measuring coarse down wood in our attempts to understand total mass, whether it be biomass or necromass within these plots. Oh, I forgot coarse and fine roots. So roots are often, well, they can be estimated from remote sensing data sets because if you use allometries that get you from above ground biomass to below ground biomass and if you're comfortable with lots of assumptions, you can get that too and we'll have coarse and fine root core numbers that we can add into the picture in terms of what the below ground stocks of mass are. So when we're thinking about plant diversity, so here's how our distributed base plot is laid out. This is just a diagram of the base plot. It's got this 40 meter by 40 meter perimeter and this sort of annular ring around the outside that's used for biogeochemistry and microbe sampling in a subset of plots. And so those data come from as part of the plot where we've said it's okay to take out these destructive soil samples. The diversity numbers come from this inner core here. So we have up to 30 of these per site and 30 is a maximum. Some sites don't have the vegetation diversity to warrant 30 plots. So for example, we use NLCD to determine where these plots go in the landscape using a spatially balanced recursive raster algorithm. So if you are working in a cornfield, you do not need 30 plots. And so the more diverse the plot or the site is, the more of these plots you have. For diversity, they're sampled one to two times per year depending on the number of greenest peaks. So for forest ecosystems, that usually means peak greenness according to MODIS EVI is usually what we've used to target those dates. And for grassland systems, you sometimes get two times per year. For example, if the Central Plains experimental range out here in Colorado, we've got kind of an early season C3 peak and a late season C4 peak. So we'll see two sampling bouts for diversity in those systems. And then the tower plots also support plant diversity, but only in a subset of them. So phenology takes place in the tower footprint. There's this red box here, which is a phenology plot that is 200 meters on a side. And so that is walked multiple times a week during transition phases and I think once per week during the green low during the middle of the year. We monitor and during, we have what's called phase one and phase two monitoring. So phase one starts off where we don't know very much about the species at the site yet because we haven't gathered very much data. And so we choose three dominant species. And then later on in phase two, we pick it up and use rank abundance curves to select additional species to put into that data product. The other component of that is we have a phenocam and we can therefore because the phenocam looks out from the tower here over the same place where we're making the observations along the phenology loop, we can then do things like link the actual phenophase transitions of species up to a community level and up to a landscape level and then therefore make links with remote sensing. So those are all very interesting analyses that we've tried to design the ground based sampling to support. So here we have some of the canopy foliage chemistry. So I'm going to spend the next few slides just talking about the biomass productivity and biogeochemistry components of the design. And we have not too many plots because these are expensive measurements to make. We have lots of different species we want to try and capture. We have spatial representation you want to try and get. So we're working at four tower plots and up to 10 distributed plots and these measurements are made every three to five years on the ground. So at the moment the plan is five years, but we're trying to get approval from NSF to push to three years. So we don't know where that's going to end, but hopefully it's three. The thought around these measurements is that these are these frequent but not every year calibrations of the AOP with the understanding that AOP will be flying every year. So we'll have the reflectance data every year if the logistic nightmare of organizing that is successful, which so far it looks pretty good. Dave's in the back there give me the thumbs up. Yeah, I know it's been an impressive undertaking. So the ground measurements then don't need to be made annually because we'll have this proxy variable that we can link to. But we just want to verify that the relationships between the wet chemistry and the reflectance measurements aren't changing through time, which they might depending on plant physiological changes or other things. Or canopy structural changes or any there are any number of things that could lead to changes in those relationships. So within the least samples that we get, we're going to be measuring these constituents from sunlit foliage only. So you can imagine in systems like this where you've got old growth, very tall trees. That's hard. Whereas if you're in a grassland, it's not so hard. So but those are the things we're going to be measuring. And then we'll also be measuring some of these biogeochemistry measurements here that the stars in litter and roots as well. So if you're interested in biogeochemistry, there are ways to kind of track these nutrients through the system potentially with those isotope measurements. Okay, so in addition to the chemistry, we will be measuring these things from individuals that are dominant species within the system. And if a plot is codominant, then we'll choose up to I think it's three different species within the plot and we'll get samples from all of those canopies. And we're getting samples from multiple subsamples from different spots in the canopy and compositing those into one. But they're all sunlit leaves. So we're not mixing in the shade leaves for this sampling effort. These trees, if possible, will also be tagged for vegetation structure measurements. So we would target those first. There are some cases where that may fail and then it's possible to have chemistry measurements from a tree that was not tagged for vegetation structure measurements. But that's what we're aiming for. We'll record the where in, you know, which plot the data come from, what tax on it is. And then we will also record which point it was mapped from and then the distance and azimuth from true north from that point. So the spatial data are all available on the website. So you'll be able to take the azimuth and offset information combined with the spatial information from the points and calculate the position of individual canopies within plots if that's something that floats your boat. So I think that's going to be one of the more interesting components of the data set to be able to do the spatially explicit analysis that way. For herbaceous clip, this part for biogeochemistry, the reason this is in here is because in some systems like Southern California where you've got a mix at our site there in the San Joaquin ecosystem, it's a mix of Savannah oak and open grassland shrubby systems. So if you were to focus exclusively on the trees, you'd be missing probably 50% or more of the area of the site, which isn't ideal if you're interested in nutrient fluxes or anything like that. So we do have a modified protocol for biogeochemistry where we will clip all of the plants within a little strip. And typically we do some sorting into functional groups, but for this work we won't. We'll just bulk all of that material together, grind it up, subsample it, and then get the same chemical analysis that you saw in the previous slide. So, VED structure here, we choose 20 of those distributed plots because VED structure is a time consuming measurement to make. And we measure only in that 20 meter by 20 meter core, so not the entire area. And we sample here one time every three years in the distributed plots. So again, we'll have the lidar measurements every year, but we didn't, we have the resources to make some decisions about how frequently plots are going to be sampled. And we are going to go with annual sampling of the VED structure measurements within the tower plots because we have real-time annual NEE coming off those towers. And the distributed plots are more focused on biomass, and so we'll be able to make those integrated measurements episodically so every three years. So essentially for a domain that's got three sites, it's just one site per year to keep kind of rotating around the sites as you go. The other main feature to note here is that we use nested subplots to standardize the sampling effort. So when you're looking at these vegetation structure data, once we get them served to the portal, I don't think they're not there yet. There's a big glut of them though, don't worry. We use these nested subplots, which you can see as these little different colored squares here within the core area of the plot, to standardize the sampling. So what that means is some plots are just thick with tiny shrubby things that all have a diameter greater than one centimeter, which is our cutoff for dealing with whether we're going to measure or not. And it would take years to just do one site with a system like that. So we've allowed the use of these things to say we want to get a minimum of 20 stems within that growth form, 20 individuals that we're going to measure. And that's going to be good enough because you start looking at your degrees of freedom. You can at least get a reasonable estimate of that growth form within that plot. And sometimes those smaller shrubs that are used in these nested subplots are mapped. It depends on whether they're visible to remote sensing instruments. So we've tried to set up the sampling so that if the individual has a line of sight and in general, across the plot, that growth form is visible to remote sensing, then we will map it. If not, if we're just looking at an understory underneath these large trees, then we won't map those. So just be aware that there's a lot of this comes down to if we had infinite time and resources, we would surely measure and map everything, but we don't. So we tried to make decisions that would enable the best use of the data by the user community. So these are the methods that we use to measure those variables that I listed before. So there's standard DBH tapes and laser range finders to measure the heights and canopy diameters of these individuals. So those again, that's just summarizing those mostly what we're getting off of the woody vegetation at the site. So below ground biomass is also measured. So I'm not going to go into that too much, but those are that's only done in tower plots. So it's probably less relevant, but I just wanted to make sure to give you a complete picture of what we're doing. And then leaf area index. So here, this one, I think is going to be an important ground data set because AOP while they fly, we hope every year they get a snapshot of what's going on. And depending on the logistics of how that season is playing out, you may or may not hit peak greenness at a given site because you have to rush on to the next site and you can't wait around and it was a late frost and blah, blah, blah. So what we're going to do is we're going to have three tower plots because the tower is visited very frequently by the technicians who maintain instruments and so on. We can ask them to go out and take photos every two weeks. So we'll have three plots from the tower airshed every two weeks from leaf out to senescence of the understory or the overstory depending on the system. Distributed plots then will be done similar to biomass every three years in the same plots that biomass is done on. So there's spatial co-location of those measurement streams. And then gradient plots as needed at a given site. So gradient plots may not get installed if the 20 that we choose from the distributed plot pool happen to span the range of the variables of interest that we talked about. And then again to fill in the picture here course down would will be measured every three years in the same distributed plots. And then we'll get bulk density so we will also have mass of those things. So we'll get volume and mass of course down would in the same distributed plots that we're getting the above ground biomass leaf area index and the canopy sampling. So tower plots. I'm going to skip over this now because I think we're going over but just know that we can talk about tower plots if you're interested in using the data from those. A little bit of nuance to how they're set up but as you can see a little nuance so we'll skip through that. But the summary of these of these sampling designs is that we produce independent estimates of biomass and productivity both above and below ground. And we parse that into defined vegetation components based on growth form in terms of herbaceous versus woody versus you know the down versus live and all that sort of stuff. We get species diversity we get percent cover and richness and there the data that we collect can be used to derive other metrics of diversity that you can select to suit your heart's content. And then phenology we will get these different phenophages we get leaf out maturity of leaves we get senescent states we can take measurements from we take measurements from individuals but we sample multiple individuals within species so we can get population level estimates and then those can be combined into these community estimates via the links with the other measurement streams from the tower. And then LAI will get LAI estimates and I think you guys probably all know why those are interesting. So in summary these measurements we're making on the ground will be able to be combined with this is an image from Crystal Schaaf from Boston University of the Redwoods that she did with one of your kidneys which is pretty cool. And so we'll be able to validate that kind of thing and potentially in the future depending on how the world works I think it'd be really interesting to transition to collecting these sorts of data on the ground and moving away from the diameter at breast height type measurements. But there's a lot of I think finagling that one might need to do between now and then. So and then the other measurements that we're making would hopefully be able to support by those canopy chemistry data maps such as this is from Greg Asner of the canopy chemistry map of the Amazon. So Greg's figured everything out and we're playing catch up a little bit in that respect. So that's all I have so I just wanted to open up for questions.