 Sai, the last speaker is Sai Nudu Rupati. And Sai is gonna present a modeling, a transit eco response to climate variability since the late Placio Pleistocene using landlabs sort for that. Sai. Thank you, Albert. Thank you very much. It's an honor to be here and presenting at the stage. I'm Sai, and I did this work during my PhD at University of Washington, Seattle, and I would like to thank my advisor, Professor Arkan Isthan Balolu. And I also want to thank the landlab team. I have used landlab extensively, and I'm also a developer of landlab, and I've learned so much from the team. And thank you, CSDMS. I have been coming to CSDMS for a good time. I was actually gonna say, it feels like home if I was gonna be there, but I'm literally at home, so. And this research is funded by NSF, and thanks to them for helping us with that. So, okay. My area of research is a semi-arid grasslands of central New Mexico, and I've been investigating the effects of changes in climate and disturbance trends. By disturbances, I mean wildfires and grazing. And the grasslands here are really sensitive to changes of these order. And what happens to these grasslands is that they are converted into woody land, or woodlands, that can be either shrublands or forests. For example, if the grazing is stopped and suddenly the grasslands were turned towards forest rather than a shrubland. And the current consensus is that the natural fires are good for the grass cover and not good for the woodlands, like shrubs and trees. And that the suppressed fires actually are not good for grass, but better for shrubs that are invasive. And grazing is, by default, removing grass in these grasslands. And wetter climate favors forests, drier climates favor shrublands. So to study these, we created different components in LandLab. If you have heard of LandLab, LandLab has a set of process-defining components. And we created hydrologic and ecohydrologic components and put together a model by coupling them. And we modeled the plant functional type wedge cover as a function of water stress and availability of seeds. So this is a two-light model that I'm showing you here. The Interstorm Vegetation Dynamics model looks at each individual cell in the domain and looks at how the hydrology is affected and how the soil moisture decay happens over time. And this information is used to understand how much vegetation biomass is created depending on what vegetation type sits on there. And there is a second layer of the model that is run at an eerily camp step. And this looks at the similar automated plant establishment and mortality. Depending on some rules, such as, say, if there is too much water stress at a cell, then it is very likely that the plant type is gonna be replaced by bare soil at that location. And for establishment, a bare soil cell is occupied by a well-doing neighbor. And there are a lot of rules that give certain advantages to certain plant types. For example, trees can propel their seeds longer than shrubs, so on a plain field, the trees would dominate if trees and shrubs were in the ecosystem. So we brought this model and we ran it with the modern day climate in the central New Mexico area. So the mean angle of presentation is around 250 millimeters, 10 inches of rain. And the top left, I'm showing you a typical flat surface here in central New Mexico. There are sporadic trees, juniper pines, and some shrubs and grass. In the top right figure, I'm showing you a simulation where we started it with an equal random condition where equal plant covers were there. So each cell is occupied by one plant type and orange cell is a shrub, a green cell is grass, a black cell is a tree, and the white cell is bare soil. The shrubs here are invasive and trees cannot survive in these flat surfaces because there is not enough moisture. So the trees die off with time, but the shrubs dominate because they are drought resistant and they also have an advantage over grass. They could send chemicals underground to kill grass. Keeping the climate constant, we brought in the topography and suddenly we see that the trees could actually survive. And this is typical in the ecosystems. On the top left, I'm showing you a watershed where the north facing slope is occupied by juniper pines and the south facing slopes is mostly dominated by shrubs. And we can actually observe that here in the model. And the reason for this is that the cumulative decrease in the solar radiation in between the north facing and south facing slopes could conserve enough moisture for trees to survive the heat. Now with this models in hand, we wanted to ask these questions. Does the Holocene vegetation change, can this be attributed to climate variability since the lake pleased to see what happened in the last 13,000 years? Could it be the reason why we're seeing what we see right now? And do fire grass feedbacks and climate trends interact to reinforce shrub expansion or invasion? And we also wanted to look at the role of topography on ecosystem change. We upgraded our model with another component which could explicitly process fires and grazing. The fires would start, it would simulate lightnings and it would ignite the plant type in the cell depending on its susceptibility. And the fire would spread to the vegetated neighbors based on individual plant vulnerabilities. And grazing would remove random grass cells and it could be upgraded again to like introduce grazing trends. The current understanding in the region is that over the last 14,000 years there were cold and wet trends and there were warm and dry periods. During the last, during the times 14,000 years before present to 8,000 years before present, there were woodlands here and then grasslands took over because of initiation of a warm period. And then the shrubs made an appearance around 8,000 years before present and they have been dominating some of the water stress areas in this region. We found a wonderful dataset from Holland Benner 2013. What they did was they sampled an exposed chartigraphy and found out the percentages of carbon. So they could distinguish between a cool grass and a warm grass. The C3 is the warm grass and C4 is the warm, the cool grass. And they reconstructed mean annual temperature and mean annual precipitation at 37 different times in the past. We use this information and we corrected our daily metrological data from a Sevier to LTER, new data information data. And we created, we reconstructed a paleo dataset and it's summarized here. And we use that to calculate mean annual PET, potential of upper transpiration for different plant types using Benner-Monty equations. And we use this information to run our models. So we first started exploring the model on the flat surface and we tuned the model initially with the climate that was at 13,000 years before present and ran it until it went to an equilibrium. And then we let the climate record that we created drive the model. In the past, there was a COVID period, like I told you, and the grasslands were doing well. And in gray, I'm showing you the fire frequency. Since the fires are explicit, you could see the blobs of white appearing in the vegetation maps. And when there was enough grass, we could see a frequent natural fire system and a warm and dry period gave an advantage to the drought resistant shrubs and the shrubs finally took over and the grass cover went down. Trees could not survive anymore and they completely were wiped off from the flat surfaces. And there was a brief cold and wet period after that, which kind of helped grasslands recover a little but it was not enough. And the modern day warm period again is helping the shrubs dominate in this region. And this is actually very consistent with what was observed in the region. To look into this in a little more detail, I'm showing you a vegetation maps at different years. And I'm also showing you on the top left the vegetation connectivity for grass neighbors. So what this is, this for every grass cell, I'm looking at how many grass cells are in its first neighboring ring. And for the years that have a lot of grass, we can observe that there was a better connectivity compared to a shrub dominated landscape. So we brought in topography again and we wanted to see how topography changes the scenario and we suddenly see that the trees could survive in the warm and dry periods. And there is a consistent tree presence, especially in the north and south facing slopes. And what happens here is if we look closely, there was a warm period during the 9,000 year before present, but the trees survived from being wiped out. And there were again some droughts and there was an extended warm period, but the north facing slopes gave a refuge to trees. The grass cover came down with the shrubs invaded but a slight increase in precipitation again, gave an advantage to grass and trees. So on a level playing field, shrubs would dominate grasslands and trees would dominate shrubs. So I think the trees dominating shrubs also helped grass to come back a little faster. To look at this simulation in little detail, I'm showing you the evolution of the vegetation in the domain from 7,800 years before present. And the warm period kicks in, the shrubs take over the flat lower surfaces. And then when the minimal precipitation goes up, the grass comes back in the flat regions and then the trees survive the warm period. I'm stressing the fact in the top figure, I'm grouping all the snapshots where there were more shrubs and we could find that the connectivity is low. And in this bottom figure, I'm showing you the figures, the maps where there were more grass and the connectivity is good. We want to stress the point that the better connectivity is good for fire spreads, natural fire trends. And that is also good for grasslands in this region. And this is what we infer from our modeling endeavors. So a grassland, when a wet to dry climate change happens, then the trees and grass die and shrubs expand that reduces the connectivity and also reduces the fire size, which in turn improves the favorability of shrubs and then a woodland or a shrub land is formed. But if there is a dry to wet transition, then the trees are favored which dominate, which kind of have a competitive advantage over shrubs. And this would also help grass and improve the connectivity and then increases the natural fire frequency and keeps the grassland healthy. To summarize, natural fires are good for healthy grasslands. Suppressed fires are good for woody plant encroachment. Grazing is again a favoring woody plant encroachment. Wet to climate helps keeping savannas healthy and the drier climates again help the shrub lands grow. Thank you and open for questions. Thank you very much, Sai. That was great. So we have a minute for questions. And again, you can raise your hand or you can place it in the chat while we're waiting. Let me ask you one question, Sai. So you reconstructed climate and you find colder periods and drier periods versus wet and warm. And you correlated the vegetation with fires. And I'm wondering actually if those fires are probably triggered by lightning, right? And did lightning, the frequency of lightning, did that change over time? And were you able to capture that with a proxy or something like that? That's a great question. I did try to find a lightning dataset, but we used a constant lightning frequency. It can be altered in the model and we did experiment with few, but we used a lightning frequency to correlate it with the current natural fire trends. But it can be altered in the model. It can be given as an input. And I see Mike's declares and question. It's moving up and there are a few questions there. So from Mike Stegler, when the climate changes, what is the time scale for the landscape to adjust to a new equilibrium? Does it differ from different types of changes? That's a very good question. So what we did is that we took the climate, there were 37 pines in the past that we got information from and we took the mid pines of those locations and we changed the climate and we let the climate change naturally occur in the sense we just abruptly changed the climate and then we let the vegetation orientation change by itself. Finding equilibrium is it might depend on the climate change itself and I don't think the coarseness of this model wouldn't allow us to exactly predict how much time it would take. But in this area, we have read that the shrub invasion has happened or is observed a lot over the last 150 years. And when we did these studies individually, we could actually recreate a lot of that. But we didn't actually study exactly how much time it took for each of the climate change trends to bring the ecosystem to an equilibrium. What we can actually see, so here the warm and dry period came but when the shrubs were there, there were some trash-hole propelled, like there was a trash-hole that was exceeded and then the shrubs suddenly shut up and then there was an equilibrium for a certain time. And then again, the climate changed but it took a little bit of time for the equilibrium to set in. So it may be because of the time it takes for the connectivity to improve, the grass to populate and then get connected and then the natural fire trends to come back. So there are many factors that would play over there.