 Do you want to hear me in the back? All right, good. All right, so this is my first CSM-DMS meeting. So I've actually been studying in the last couple of days, and I learned a lot, and I think everything just amazing. So I'm grateful to have this opportunity to share my work on the post-racial landscape evolution. So my advisor, Alison Anders, has talked about the bigger picture problem of this study yesterday. So I want to just remind some key points here. So first of all, Illinois and other parts of the Midwest, northern Midwest are flat, so very flat. So here in Colorado, you have a lot of mountains. But I think one thing we can share in common in Illinois is just we can skin in Illinois. So I think this is probably the steepest, slow thing, champagne abandoned area. As you can see, it's still very flat. And this is Cecilia Collins back. So she's the person behind all the groundwater models I just showed you yesterday. All right, so a consequence of this flat surface is that a large area of the upland is dominated by closed depressions that are not connected to the fluvial drainage networks. So the reason we have this is because the repeated declassation of the Laurentide ice sheet basically disrupts all the pre-existing drainage networks and deposit a thick pile of glacier tails. Another thing that this glaciation did is during the declassation, the meltwater from the ice sheet can carve some deep valleys. So here, so basically the two important components for post-glacier landscapes are, first, the meltwater valleys, and second, the load relief uplands. So here, this is the DM of the Upper Sangamon River in Illinois. You can see that the river valley here is about 30 meters deep. Then figure on the right, this is the soil drainage data of the same area. Dark blue means the soil is polygrain. As you can see, a lot of places on the uplands is very polygrain. And the consequence of the polygrain soil is that here, this is the area photo of the cornfields. A lot of the crops are dead here because there are too much water out there. So Alison talked about how the water in the depressions can connect to the drainage network through groundwater. But here, I want to explore another possibility so how the water can route to the drainage network through the overland flow, so surface flow. So let's think about two and number scenarios. So first, disconnect case. So in this one, we assume all the closed depressions are disconnected from the drainage networks. So as you can see, these three depressions are disconnected. Then on the other end, we can imagine that all the depressions are filled with water up to their spillover point. So the depressions are connected to the drainage networks all the time. So my question in this study is so I want you to know whether the disconnect case versus the disconnect case, they can give us a very different path of the landscape evolution. So to do that, I build a numerical model using the LandLab platform. So this is my model domain. The evolution of the topography is dominated by a three-power urban law and linear diffuser for the geosolope process. So my model domain on the left boundary, this is an open boundary that represents Mount Water Valley and the elevation, the depth of the valley is 40 meters deep. And the major part of the model domain is flat surface with some random noise. So representing the low relief upland. Then on the other three boundaries, the top bottom and red are closed. So all the water can only flow out to the domain from the left boundary, so the Mount Water Valley. So that's how we calculate the topography. But how do we differentiate the disconnect case versus connect case? So for the disconnect case, I use the standard B8 algorithm in LandLab. I use the flow router module. So the flow direction is determined by the direction of the steepest down slope. So in this way, all the closed depressions will remain disconnected from the drainage network. So then for the connect case, I implemented two additional modules in LandLab. The first is paid filler, that can fill all the depressions up to the spillover points. So you can imagine that this algorithm will flood all the hold on main inward from the Mount Water Valley from the open boundaries. Then this algorithm use a priority queue to guarantee that all the depressions is filled up to their spillover point. Then again, I implemented another algorithm that described by Barnes et al. This is a flow router over flat surface. So this algorithm will calculate gradient away from the higher edges of the depressions and also a gradient towards lower edges. So by adding these two gradients, you can have a convergent flow field. So this paper Barnes et al actually have a detailed proof why this algorithm works. So I just skip that part in this talk. So I want to mention one thing here is that these two modules actually don't change the topography. They actually create additional mask that only use to determine the flow direction. So the other topography, the depressions is still depressions in the model. Then another thing I want to mention that these two algorithm actually equivalent to the depression funder and the router module in LandLab. The reason I have to implement these two algorithm is that when I start to work on this project, this module is still in development. So I can use that. All right, so now let's look some results. So first, this is the model topography after 50,000 years evolution. So I presenting four cases here. So the C base is soft and connected and the C hard are connected. So C represent connected. And the D base here is disconnected case. Hard, soft and base just representing the different irritability coefficient in the stream power urinal. So as we can see in this figure, so first the disconnected case, you basically cannot see any channels. The channels are very short, just getting 100 meters long here. So most of the upland still remain unchanged after 50,000 years evolution. But for the connected case, as you can see, there are several long channels that have cut into the uplands. Even though the C hard case, the irritability is one other magnitude lower than the disconnected case. You can still have some longer channels than the disconnected case. So a very basic finding from this data is that the hydraulic connection, so the disconnected versus connected can have a huge impact on the landscapes. But we want to quantify this stuff by calculate this person's integrated. So this is a person of the domain that is connected through a down slope path to the pre-existing meltwater valley. So even though in the connected case, I force all the water in the depressions connect to the drainage network valley. But when I calculate the person integrated, I still use the D8 algorithm, the closer depressions is not considered integrated. All right, so that's a person integrated. So the increase in person integrated is basically equivalent to the decrease of NCA. I'm sorry, I forgot to mention the NCA, NCA representing the non-contributing area. So that means the water in the depressions are usually considered not contributing to the river networks. Although in this model, I'm exploring whether the non-contributing area is non-contributing. All right, so let's look at how the person integrated evolves over time. So in this figure, I plot the person integrated as a function of time. The green line here is the disconnected case. Then all the three black lines are the connected case with different aerobility. So as you can see, all the four cases predicted the increase of integration. So the non-contributing area becomes integrated over time. Oh, that's not so surprising. But what's the difference here is that the rate of the evolution. So in the disconnected case, the integration happens in much slower rate, especially if we want to compare this disconnect case with the C hard case. So the aerobility in this case is one auto of magnitude lower than the disconnected case. We still have higher integration rate compared to the disconnected case. So for the integration, the hydrological connection is primary control on the integration. And the other thing in this lower panel, I'm plotting how the net erosion, how that erosion involves at the function of the person integrated. So in this figure, the x-axis here is now person integrated not the time. So a very interesting finding in this figure is that if we want to have the same degree of integration, let's say 10%, the disconnected case, this green line actually requires more erosion. All the connected case need less amount of erosion compared to the disconnected case to achieve the same amount of integration. So I think that's a very interesting finding. So they have different amount of erosion. How does that reflect in the landscape? So in this slide, I figure I'm plotting the model of the landscape when 15% of the model domain is integrated. So let's go back to a couple of slides. So as you can see, for the connected case, even after 200,000 years integration, the person integrated is still less than 10%. So it actually takes over 500,000 years for the disconnected case. To get this 15% integrated. So as you can see, there's a big difference between them. So for the connected case, we have several long channels that cut into the domain, but for the disconnected case, there are many, there are more channels compared to the connected case, but most of them are very short. So then I also plot the histogram of the channel length. So again, as you can see in the disconnected case, in the disconnected case, most of the channels are shorter than one kilometer. But for in the connected case, we have some long channels that over that even, almost four kilometers or five kilometers. All right, so that's all my model renounce. Then next we want to compare these renounce with some real landscapes. This is a upper sangman river basin in Illinois. We identified some channels in this basin and plot that histogram of the channel length. So this is the histogram in this figure. The blue color is the channels from this bigger box here. Then the orange colors are the channel length from this smaller. Oh, I can't even see that. This one, yes, the smaller box. So the size of this smaller box is roughly same as my model domain, so five kilometer by five kilometer. So as you can see in the channel in this histogram, we have some short channels, but we also have several long channels. That's very similar to what the disconnected or the connected case predicted. So now let's just have a direct comparison between them. So as you can see in the disconnected case, we don't have long channels, but in the connected case, we have some channels. So the figure on the left, the real landscapes is more similar to the connected case than the disconnected case. Then the other thing I want to point out here is that I said earlier, the disconnected case requires over 500,000 years to get 15% of the model domain integrated. But for the connected case, the time usually, I think the time request on order is on order of tens of thousands of years. This time scale is actually similar to the time since the deglazation in the NOE area. So the last time, the Laurentiad attitude covered the northern part of Illinois was about 20 to 25,000 years ago. The time scale also suggests that the connected case might be a possible scenario here. Okay, so then that's all my model results. So in conclusion, I have kind of three bigger, or three big findings in my research. So first, my model results suggest that the connected case can have faster and the faster rates of Euroran integration and also the morphology different, the channels are longer and more seniors. Then secondly, that the connected case requires less amount of Euroran to accomplish the same degree of integration compared to the disconnected case. And the third point is that a simple comparison with the upper San Maria basin in the NOE suggests that a hydraulic connection between the depressions and the mouthwater valleys could have had a very important impact on the landscape evolution. All right, so that's all my talk. So I have my contact information over here. If you want to use my module instead of the depression fund and router in land lab, you can contact me and it's on my GitHub page. All right, so that's it. Thank you. Thanks, Ting-Tang. Yay for open source. Questions for Jing-Tang? Yeah, so the question was, the model results suggest that in the real landscapes, the channels looks more like the connected case, but if you brought the flow over the real landscapes, whether the channel would be connected. So the answer I think is no. I think yesterday she had a poster shows how use the basic the same DEM I used in yesterday that you show to try to route flow over the landscape. Yeah, basically I think her results so that show that they're not connected. Yeah, so the question was, do I have a potential? I'm sorry, did you say explanation or experiment? Oh, yeah, yeah, right. So yes, actually the talk I just gave yesterday is talking about how the groundwater may be a connection between the closed depressions and multiple valleys. And that work is still going on. I think, but I'm not currently involved in this project anymore. I'm moving to glacial landscape evolution. So I don't know whether they have a plan to move this part forward. Yeah, so I guess Alison can answer that question better than sorry.