 Well, first I really want to thank the organizers for the invitation to speak here today and just say it was very hard to find exactly which topic I was going to talk about today because by and large my work deals with modeling how vegetation responds to extreme events. So with this I've looked at how vegetation responds to fires, floods, hurricanes, drought. I wanted to focus on a particularly vulnerable ecosystem these montane cloud forests and their vulnerability to climate change. So we'll focus on that today. And why it's so important is these ecosystems occur at high elevations that typically are the headwater catchments for a lot of our major river basins. There are also hot spots of biodiversity and the forests here are very highly productive. Oh, yeah. And the processes that I'm going to talk about the gaps in our ability to model these systems and how they're going to respond in climate change are not just local to montane cloud forests but really affect forest environments worldwide and particularly forest that occur in mountainous regions. And so forest, 25% of all forests actually occur in mountainous regions as you can see here globally. Now all of these are going to be cloud forest but a large amount of the processes are going to be very similar. And here in my talk today I'm going to focus on two particular cloud forest environments. The first is a temperate cloud forest that occurs in the southern Appalachian mountains and the second is, and so you can see here, you can't see it yet, now you can. This beautiful image of what it looks like to be inside of the cloud forest when you're enveloped in the low cloud and fog immersion. And the second that I'm going to focus on is work that I've just recently started in Costa Rica in Monteverde and this is what it looks like to be in this particular cloud forest. So it's a really nice place to do some field research as long as well as some modeling. So what are some characteristics of cloud forests that make them unique and also challenging to add to our earth system models? The first is this presence of low level cloud and fog. This creates an environment where you have very, very humid microclimates within your canopy. The plants are very well adapted to these humid environments where there's very low ridity. It means that they can sort of consistently exchange gas because they're in this ideal state. Along with this low level cloud and fog, you can see that the quality of light from sun looks a bit different. So you have more diffuse radiation rather than direct radiation and so the plants are also very adapted to this quality of light rather than just the quantity of light that they receive from the sun. The second challenging thing is because you're constantly involved in low level cloud and fog, your leaves tend to be wet. And why this is troubling is that gas can't diffuse through a liquid. And so if you have liquid on the surface of your leaf then theoretically you should not have any gas diffusion. And so this also adds complexity that we're not typically accounting for in our system models. Finally, within these canopies you also get the presence of epiphytes. And so epiphytes are vascular and nonvascular plants. They can look anything from a moss to a small tree or shrub growing on top of your tree in a rather symbiotic fashion. And so these epiphytes also really enjoy the low level cloud and fog just being in this diffuse light environment, this very humid environment and under projected climate change. What we think is going to happen is that you're going to have increasing temperatures in these environments. You're going to have an increase in the cloud base height and we're not exactly sure how that's going to impact how these force environments are adapted to these particular characteristics. And so today I'm going to sort of illuminate some of these gaps and maybe talk through some of the ways that we can improve how we stimulate these processes in cloud forest environments but also any forest environment where these types of processes occur. And so before I really dive in, I need to first talk about two key ways that we incorporate plants into our Earth system models. The first is through some model conductance. Sorry, on our leaves we have these small pores and through those pores is where we exchange gases. So water vapor and CO2 get exchanged through these leaf pores and some model conductance is how we model essentially how open or closed those leaf pores are, how difficult or how easy it is for gases to move between the atmosphere and the leaf pore. And typically how we model this is sort of as an electric circuit where you have a resistance which is the inverse of conductance that moves water vapor through different potentials, water potentials. So water vapor is going to want to move from higher pressure to lower pressure. And so in this particular case it's envisioning this resistor or conductor between the leaf pore, whatever the pore pressure is there, and the atmosphere. The other way that we incorporate it is sort of a bigger circuit model where now we're considering how water moves through the entire plant perhaps all the way from the soils. You could even go down to the groundwater if you have roots that can tap into groundwater up through the stem, the leaves and then out through the atmosphere as transpiration. And so the little conductance term on the side is also accounting for the fact that you can have water storage that your plant can release if it's underwater stress. And so these are modeling water stresses critical for these environments under climate change because we want to know how we're going to change our gas exchange processes and particularly carbon uptake within these regions. And so this was sort of the first question that I had because while we're not necessarily modeling or not a lot of work has been done modeling gas exchange processes in these environments, quite a bit of work has been done through observation and data collection from tree physiologists. And so this is work that was done in the southern Appalachian Mountains observation data. And this has been demonstrated with both field and greenhouse experiments. And so what you see on the left hand side is a plot of photosynthesis under cloud immersed conditions. And you see at 9am and then at 2pm in the afternoon local time. And then on the right hand side is what this, what net photosynthesis looks like at these same times of days, but on a clear day. So when you have no cloud immersion, no low level cloud and fog. And the biggest, the key difference here is that once you get into the middle of the day, through the peak temperature, under the cloud immersed days, you have higher levels of net photosynthesis. Your plants are able to be more efficient at drawing down CO2 from the atmosphere under these cloud immersed conditions. And while it might look like a small amount over the entire growing season, this could be as much as 20% of your overall carbon uptake for these ecosystems. So having cloud immersion is quite crucial to sustaining these forests as carbon sinks. So my first question was, well, how well are we actually able to quantify gas exchange rates in these ecosystems with our current models? With how our current models are simulating gas exchange processes? And so one of the typical empirical models that we use for somatoconductance is called the Jarvis somatoconductance. And it's an empirical model where you're simply simulating your somatoconductance as a series of multiplicative terms where the first is your maximum somatoconductance for that species or canopy. And then you have a series of index functions that represent how your somata want to be open or closed under sunlight conditions. Really, this is only considering the quantity of sunlight, not the quality. So we're not tackling that diffuse light condition. Air temperature, leaf water potential, which again is the water potential within your leaf. It's an indication of your plant water stress, any given moment of time, and then vapor pressure deficit, which is a measure of atmospheric aridity. And so what we wanted to do was use data from the field for these different conditions and what they look like under immersed low cloud and sun conditions. And so what you're seeing are each of these index functions from our Jarvis somatoconductance model in all of these small panels. And you can see they all go from zero to one. And so zero means that gas exchange is not occurring. Your leaf pores are completely closed because the conditions are such that your plant is under too much stress to exchange gas. And then one means that your pore is completely open and you're freely exchanging gas. And so you can see that obviously your plants want to be in low aridity conditions or high relative humidity conditions. So for your leaf to air vapor pressure deficit, when you're close to zero, you're at one. For air temperature, there's sort of an optimum around, I think, 18 degrees Celsius that your plants prefer for keeping their pores open. For the water potential, you want to have lower water potentials and your plant is going to be happier. And then for sunlight, obviously you want more light. And at some point that becomes saturated and you don't get any additional benefit of having more sunlight. And then what you can see in these horizontal bars are showing the ranges that we've seen from observation data experienced under cloud immersion, under low cloud conditions, and then under sunny conditions. And so what we wanted to do was search this entire parameter space and see under any of these conditions that have been observed, if we run them through this model, are we ever able to recreate the gas exchange rates that have also been observed under these conditions? And the answer is no, we actually cannot. And so on the left-hand side are the results from our model. And so this blue is indicating the entire parameter space under all of the conditions possible for cloud immersion. And then on the x-axis, you have your leaf to air vapor pressure deficit and then on the y-axis is the model conductance that we're actually interested in. And on the right-hand side are the results from observation data collected at the site in the Southern Appalachians. And it's a little misleading because the x-axis is not quite the same. But the key is that when we're looking at our cloud immersed conditions, the high end of the range when you're around 0.5 for your leaf to air vapor pressure deficit is somewhere around 0.45 moles per meter squared per second. And if you look over at our model results, you're never there. So we're never able to capture this. And for me, this really points to the fact that we're using an empirical modeling approach and perhaps for these ecosystems, you really need something that's more process-based and explaining how the leaves are actually responding to these conditions rather than just doing this rough approximation that seems to work really well for other environments. And on top of this, if you're using this type of model for your gas exchange rates, well, if you're going to model transpiration, you're also going to be underestimating. And in this case, you're not just underestimating for your immersed conditions, but for all day conditions, both low cloud and sun conditions for these environments. And so this is really pointing to a gap in our ability to actually use our Earth system models to simulate what climate change is going to look like for these environments. So next, I'm going to talk about a new project that recently got funded to look at how local microclimate, but also host tree water status, is impacted by the presence of epiphytes. And so to simulate what we expect to happen with increasing temperatures and increasing cloud-based heights, essentially, we don't expect epiphytes to move to higher elevations with these raising cloud-based heights, because they actually don't move very quickly, is what the biologists have told me. So they expect them to just disappear. And so with some collaborators, they are doing, and they're actually in Costa Rica right now, doing these huge removals where they're stripping the epiphytes from the branches of the trees. And so we'll have some control and some experimental trees, some with and without epiphytes, and we're fully instrumenting them to sort of see how plant water status, but then also how microclimate is changing with this removal. And as a modeler, I'm not in the field, obviously, stripping these trees, but my role is trying to figure out a way to better represent the role that these epiphytes play in our host tree water status and also in microclimate. And so going back to the different ways that we represent our vegetation or our system models, the second approach using a soil plant atmosphere continuum model, that's what we're going to do. And now add to this capacitance term, this plant storage term, this additional storage of water in the canopy because these epiphytes hold a lot of water consistently. And so this additional water term influences water cycling through interception and evapotranspiration from these epiphytes. It influences how much water actually gets down to the soils. And then they also, what's very interesting is that a lot of these trees have roots in their branches that go into the epiphyte maps. So the trees can also directly uptake water through their branches from these epiphytes. And so we're just not sure, and we're hoping with this model to be able to explain and understand how much these contributions of water from epiphytes influence our host trees. And so our approach is to figure out the storage term by using a water balance modeling approach. So we're constructing this epiphyte water balance model where we're considering the epiphytes as a water tank that is held in the suspended in the canopy. And our inputs can come from interception of rainfall and also from fog. So both vertical and horizontal precipitation also do deposition. And then our main outputs are going to be through evapotranspiration from our water tank and also host tree water uptake through those roots that are in the branches. And we have a little idealized simulation running since we don't have our trees play instrumented yet, not able to compare to our data in the field. But we are able to represent the change over say three days and what that water storage looks like using this approach, considering that you have no rainfall input or no fog input. So you can see you have a dry down and the tank is emptying. So I told my grad student working on this project that makes physical sense. So this is a good step. And so these are sort of smaller scale ways to think about how the specific adaptations in these environments are going to influence gas exchange processes. But I also just wanted to highlight and think about some ways this also what else we're missing in our earth system models but at larger spatial scales. And so the first is looking at say regional scales. And this is some earlier work that I did when we first sort of were looking at these Appalachian forests and trying to figure out why our land surface hydrology models were so bad at estimating carbon fluxes over this inner mountain region. And we proposed, well maybe we're just not in our rainfall forcing data capturing the input that you get on foggy base in these systems. And that fog can account for say 0.5 millimeters over three hours in the middle of the day. That's what our fog gauges in this region told us. And so we thought let's just add this to our rainfall forcing and see how much this increases our carbon uptake response. So these are results at four kilometers using a land surface hydrology model where we're physically simulating water budgets, energy budgets, and then using a photosynthesis model that's based on the bar core formulation. And so what you're going to see here, and sorry for the confusing axis, but basically the darker the color, the more carbon assimilation you're getting in the simulation where we've added fog to our rainfall input. Let's see if I can get this going. And this is going to be over the sort of water season or warm season for southern Appalachians. Maybe we're not going to see it. Okay. No worries. Well, if it ran, what you would see is that progressively is you got to the from May to June to July or warmer season months, you would see that this the amount of carbon uptake that you got during the fog simulations increased quite a bit between two to three grams of carbon per meter squared per day, which could be somewhere between 10 to 15 percent of your overall carbon uptake for this ecosystem over its warm season. So that's quite a bit. And so this is proposing, right, so maybe at these larger scales, you don't want to necessarily get into the negrity of modeling some model conductance better or modeling the role that epiphytes play better, but you could get this increase and match better observations just by improving the rainfall data set that you're using for say, making sure that our rainfall data sets are including the rainfall inputs from that we get from fog in these environments where fog is frequent, because here during the warm season, we have foggy days 60 percent of the time. So that's a big water input that we're missing in our rainfall data. So going up even more in scale. And so now we're getting out of just cloud forests or mountain forests, but just thinking of ecosystems where you have wet leaves. And so I mentioned earlier that these this wet leaf problem, I think of it as a problem in our models, but this wet leaf problem really is a bigger problem than just within these montane cloud forests. And so you can see in this figure that a lot of parts of the world have experienced wetting events for most of the days. And currently our models are our system models either completely ignore what it means to have a wet leaf, or others assume that if you have some percentage of wetting in your canopy, then your model conductance is going to get reduced or go to zero, which would be true if we didn't have some model on the bottom of our leaves, which don't get as wet. And so this is sort of a jumping off point for some recent work that was funded with a little seed grant, where we're going to do some tests, actually gathering data on how much gas exchange we get with wet leaves. And then my part is going to be to hopefully first build some empirical models to explain what we're doing and then hopefully get to process based models for explaining how our gas exchange looks like when we have wet leaves, but only partial or no wetting on the bottom of the leaf. And so just to summarize some of these gaps and challenges in modeling plant adaptations to low cloud fog and wedding events. And we're thinking about this at different scales. So the key gaps are first at the leaf scale, getting some model conductance, getting this gas exchange right. And one of the key processes that is also has been observed to occur in these cloud fours that we also don't think about, and might be a solution for thinking about how to build a process based model for some model conductance for these systems is thinking about fuller water uptake. And so this is an adaptation that these leaves have where they're actually able to uptake water through the pores in their leaves. This can improve the water status, which will, you know, through nonlinear feedbacks influence your leaf water potential, and then hopefully increase your your model conductance. That's one way to perhaps close that gap. Next, we talked about at sort of the branch and leaf scale, the role that epiphytes play, and all of the ways that they influence, not just the water budget, but the energy budget by their presence and interception of water, how much gets down into your soils. So really, not just the water and carbon cycle side, but also the energy budget side. At the canopy and plot scale, how this diffuse light and high humidity environment is going to influence your microclimate is a critical need that we're still trying to figure out. And then at these regional scales, this improving our representation of rainfall inputs and an interception of fog water that's not currently being considered, and then an even larger and global scales, this wet leaf issue that we still need to find a better way to deal with. So with that, I want to thank you all for the opportunity to speak and particularly acknowledge my collaborators on the NSF project. That's us in the field in January riding horses, which was really fun. Civil Goach, who's at Franklin Marshall, Nalini, Nicarney, who's at the University of Utah. Todd Dawson, who's at UC Berkeley and missed out on the fun of riding horses, but hopefully next time when we go back this summer, he'll get to join us. In the middle, that's, you know, me with my students and I want to particularly highlight my students who've been working, who've contributed to the work that I presented today, Paul, who's a grad student, and then two undergrads, Haley and Louise, who have been integral to getting this work done. So thank you very much. So I was wondering if you take away or remove those epiphytes because to simulate climate change, right? Would that also change the evaporation and therefore less fog will come in and therefore less water will be there for the tree, actually. And you see kind of it's a spiral, right? So you get less trees and therefore maybe more erosion at some point in the landscape. Can you explain that if that process will occur or not? Yeah. So those are sort of the hypotheses that we listed in our proposal. Some of the, there are a lot of competing effects that we're trying to disentangle with these experiments, but then also with the modeling work. So yes, remove what we, what we hypothesize is that removal of the epiphytes, you're going to have, yes, less evapotransferation from the epiphytes, but instead you're going to get more water that or less water that gets intercepted in the canopy, because now you've removed those epiphytes, but you'll have more water going to the soils. So perhaps the trees will just sort of adapt, that's what we're curious about, whether they're going to adapt to having higher soil moisture and be less reliant on that water storage within the canopy. But also removing the epiphytes in some initial experiments that they did, it seemed to have an impact on that vapor pressure deficit. So basically making it a slightly more arid environment overall. So it's also having this impact on the micro climate itself. So maybe you would get fewer wedding events, so you would collect less water and then your soil moisture would be lower. So we're hoping to disentangle some of these. Hi, very, very nice talk, but also it highlights the difficulty of doing these global upscaling things. But I wonder, in how many of the current global models, is there some sort of representation of these epiphytes? There's no representation of epiphytes in global models. None at all. And do you think it'll change soon? Maybe I'll have a big enough impact, but it will change people's minds. I think that hopefully this work will sort of tell us. And I guess with the work that I do, looking at how we represent vegetation processes, it's not always about, say, building this new component and adding it to these earth system models, but also trying to understand, well, if you don't incorporate this process, what's the size of the uncertainty in your estimate? And also, what are the sort of spatial and temporal scales that it's associated with? Because perhaps, say, epiphyte removal doesn't really matter as much in dry season. It's just in the wet season, these environments, right? And so if you're only interested in dry season processes and maybe incorporating epiphytes isn't necessary for what you're doing. But yeah, so I think the goal is not necessarily to have every vegetation process represented well, but being able to quantify our uncertainty by including or not including these processes.