 So, the last talk of the morning session is going to be given by Jordan Adams, who's from the University, on integrating a 2D hydrodynamic model into the Landslap modeling framework. Systems for inviting me to talk to you today about my research, which is broadly titled Integrating a Hydrodynamic Model into the Land Lab Modeling Framework. Okay, so, in geomorphology, we often use numerical methods of landscape evolution models to understand the dynamics of climate, topography, and tectonics through time. And so, my role in the Land Lab project is to think about how traditional landscape evolution models have captured hydrology and how the choice of hydrology method in these landscape evolution models can impact your model output. So, most landscape evolution models use what we call the steady-state hydrology assumption. So, what you get in steady-state hydrology is something that you see here. So you have a single precipitation rate that is constant throughout the storm duration, and it's uniform across the watershed, so it's just a very idealized watershed. And so, when we dot the single rainfall rate through time in the steady-state assumption, we get a single discharge rate at every point in the watershed. So, at every point in this watershed, we calculate our runoff rate as a product of your precipitation rate in your drainage rate. So, again, we get one steady runoff rate through time. And when you're calculating incision or trying to involve your topography, you use that discharge rate and you get, again, a steady incision rate. So, on the opposite end of the spectrum, we have non-steady hydrology, and many rainfall and off-models use these types of non-steady methods. Where there's profiting here, I'm just using a steady precipitation rate. But instead of calculating ones that we discharge rate through time, what you actually get is a hydrograph at every point. So you have your discharge versus time, and you calculate this discharge as a function of things like water depth, surface water slope, surface thickness, and you get your incision rate as a function of that discharge value. So, again, at every point moving from low-grain ageries to high-grain ageries, you're calculating these hydrographs, which often persist much longer than your precipitation rate. But there can be problems with this steady-state assumption, particularly when you have really large watersheds or you have short-duration precipitation events. And the problem is when you have a really far away point upstream, the travel time of water as it moves towards the outlet could be much longer than the precipitation event. And so, Stolium and Tucker tried to calculate how these steady-state assumptions were moving from steady to non-steady state. And so what they did was they approximated a hydrograph using the maximum flow length of your watershed in strong duration. And what you can see here is as they increased the non-steady hydrology, there are pretty good start differences when you compare these topographies. And what they noted was that in their non-steady cases with increasing non-steady hydrology, the watershed for marked by greater relief, increased channel convexities, and low-valve densities. So I wanted to try to address this using the LandLab modeling framework. So LandLab is an open-source, high-lab modeling library that's been geared towards our surface dynamics. If you're interested in learning more, I invite you to visit our website to talk to any member of the LandLab team. I hope many of you are attending our clinic tomorrow. And so what I've been doing for my PhD is actually developing what we call the Overlay Flow Component in LandLab. And so what I've taken is an urban coordination model that was developed out of the University of Bristol. So they developed a model that would route a wave across low friction urban areas. And so they came up with this algorithm here, which is an explicit algorithm, and you see that discharge is calculated as a function of things like water depth, surface water flow, surface thickness, and when you bring discharge values. And so what this algorithm does is it routes a hydrograph of all of your group locations with flow moving in four directions. And so they've also developed, or they also used, excuse me, adaptive time step, which keeps your model stability and efficiency in place throughout the model done. Okay, and of course we wanted to test this against non-analytical solutions. So here you have water depth versus distance, and I tested this for a variety of wave velocities and surface thicknesses, and you can see that the model approximates the predicted inalytical solution quite well. So to understand the quantitative differences between being steady and non-steady cases, I had to route a flow across a model domain. So what I did was I developed this geomorphic steady state water depth, so that means up close is equal to a water rate throughout the entire watershed. It's 36 square kilometers in drainage area, and I evolved it using the simple stream power. So in simple stream power, we have the decisionary calculated as a function of your bedrock or loadability, your discharge to some exponent and your surface water slope to some exponent. And for simplicity theory, I didn't use any sort of limiting threshold, just kept it very simple and straightforward. And so obviously I needed about some sort of precipitation event across this watershed to get one off. So I had a steady and non-steady precipitation event. So in the top here you can see your steady state precipitation. So in these models, we calculate precipitation rate at a low constant intensity of time and calculate your discharge and incision rate as a function of that. So what I modeled was half a meter a year, this is our character, so we send the other climates, and I modeled this rainfall rate for 10 years. And so to compare this to the non-steady case, I used the same amount of water, so the same total water over those 10 years. But because the hydrographs persist they're much longer than the precipitation event, I had to break out the precipitation into discrete storm events. So to capture the same amount of rainfall, I did 50 storm events per year for 10 years. So I had 500 hydrograph events that evolved with the cover, using these, what I call the incisor graph. So the incision was a function of these hydrographs through time. And just to make sure that intensity and duration weren't big factors in my model results, I modeled three storms that had the exact same rainfall rate through time, or excuse me, that approximated the same amount of rainfall here, it was half a meter a year. And so what you get when you route these storms across the main stream is something like this. So here you can see the model domain and claim here, and you can see the wave propagating from the upstream to the downstream portion. And so here at the bottom, I actually have a hydrograph that's taken at the outlet, so about right here. And so what you can see is that when you compare it to the steady state case, the solid blue box, the peak discharge actually exceeds what we would predict by steady state, but it persists for a much longer time than the steady state duration. So we can compare this again, so you are steady state predicted discharge values, and then this is, and discharge is calculated as a function of precipitation rate in drainage area. You can see particularly in these upstream portions, you're actually going to exceed the predicted steady state discharge. So then if we look at how discharge scales across the entire watershed, what I've done here is I'll plot average peak discharge at all drainage areas. So here I've plotted two storm cases, so the most intense cases here in blue, with the steady state predicted discharge value as the solid line, and so here's the least intense case in green, with the Xs representing the non-steady peak discharges across the watershed. So what you can see is in the upstream most portion, the patterns are pretty much the same. You're going to over, your non-steady case is going to overestimate your peak discharge when compared to steady state, but as you move downstream, these differences between these storms are less clear. Particularly if you're focusing right near the outlet, in the most intense case, your peak discharge is actually less than the predicted steady state value, a pattern that is not seen in your least intense cases. And still again, I say I'm tying this all back into landscape evolution, so we want to look at this in terms of institution. So here is the total in size depth versus drainage area. So when I say total in size depth, I say over the 10 years of model run, what would be the predicted erosion that all of these drainage areas? So up here, this bad sign is the predicted geomorphic steady state. So if I'm using that steady state precipitation and discharge rate to calculate incision, we would predict, because uplift was much better as we rate, that you would have the same emerging rate across the watershed. And this pattern isn't seen as the non-steady method. So you can see that these slopes here represent the three non-steady storms that I routed. And you can see that in our cases throughout the watershed, even though the hydrograph persists for much longer than the steady state case, we actually experience incision much less than predicted by that steady state value. But there aren't still some important trends to take away from this, which is as you move from the upstream to downstream portion of your watershed, you go from less incision in the upstream portion to more incision in the downstream portion. So to kind of just back into what Soryam and Tucker thought, we would predict using a non-steady system, we would have greater overall relief when we use non-steady hydrology in the model landscape revolution or compared to steady state. However, we would predict due to this upstream stifling due to less incision in the upstream portion of the watershed, we would actually predict increased channel concavities. So just to kind of wrap it all up, what we saw using the non-steady model was that peak discharge in these non-steady storm cases can exceed the predicted steady state. However, because those peaks last much longer than predicted, the impact on incision is not as great as predicted steady state incision. And the imperfections for landscape revolution models suggest that when we use these non-steady cases, we would predict greater overall relief and increased channel concavities. And so moving forward with my PhD, I'm interested in capturing these short-term landscape revolution events such as post-fire flooding. And what I take away from this is if I use the steady state case, I might not be capturing some of the complexities of landscape revolution model if I use the steady state case. And so I'd like to thank my grandmother for all the news and the funding sources. And again, I'd like you to visit the grandmother website. Thank you. Thank you. Thank you. Thank you. Questions for Jordan. We have time for a couple of questions. Some were somewhat counterintuitive, but I thought the smaller catchments where you saw the real changes in action happening would be more of, well, better approximated by a steady state because there's less distance to flow or to get differences. Why is that? Why am I missing in that counterintuitive? So I think as we get smaller catchments, we would be better approximate with a steady state. Larger catchments for a steady state can be problematic because of that problem so bad. So, and I've tested this model on different shapes of basins, and we can kind of get closer, but it's still not all over there. Reminds me of your question. I'm really curious. There are various published relationships, peak discharge or painful discharge versus basin area, where variety of different sorts of areas for rivers, which obviously have different geometries. Have you looked at comparing your model results to those published empirical data sets? So my next step for the next chapter of the situation was actually using this model against field data and trying to see if it works with the published empirical results, yes. So I haven't done that yet, but it is going to be done. All right. Pretty good. So I know there is a no-builder, or is there a component for under the groundwater flow? There's no groundwater flow yet. So we are going to try to tie this in with an infiltration model, but as of right now, the only plans are to use sandless hydrology, no-bound hydrology. Okay. Let's thank Jordan again.