 Yeah, good morning everyone. My name is Silke. I'm currently a postdoc at Boston University, but today I wanted to talk about some modeling work that I did during my PhD at Delft University of Technology in the Netherlands, which I finished last year. So I'll be talking about Henry Dynamics. So maybe the most obvious question is, what is a tenure or a chenille depending on how you want to pronounce it. Basically, it's a sandy ridge found on muddy coastlines. So there's this body of sand that's lying on top of and surrounded by mud. Some of you might be familiar with chenille's already part of the coastal plain, but I specifically studied the chenille's in their early phases of their life when they're still in the intertidal zone. And what makes them different from sandy ridges on sandy coastlines or barrier islands is the fact that they're on muddy coastlines. So what you see the chenille part, that's the only sand in the area. It can be sand, it can also be coarser particles, shells, but in this example it was sand. Now, why are we so interested in these chenille's? Well, my research was part of a larger research effort looking at restoring eroding mangrove mud coasts. And we can see that where these chenille's are present, as you can see in this photo, they actually dampen waves, and they can slow down eroding the erosion of mangrove mud coasts. But not only can they slow down erosion, they can even enhance recolonization. And I know if you're familiar with the theory of window of opportunity, but basically for mangroves to be able to recolonize, they need a certain time window in which the conditions are favorable for growth. And that starts basically with an inundation-free period. And this is where chenille's can play a big role because von Weißelfeldt all showed that the presence of chenille's enhances mudflat growth. So they can elevate the bed level until it's high enough for mangroves to grow. And of course they also dampen the waves, so they make sure that those really young seedlings cannot be uprooted. But the thing is, we actually know very little about these chenille's. Why are they there, and how do they move? So for this, we went to Indonesia. This is our coastline that we studied. It's part of, it's bordering the Java Sea, which is very shallow and sheltered sea. So don't expect extreme high waves. It's really a low energy coastline. To give you a bit of context, this is what the coastline looked in 1994. This is what it looks like nowadays. So the dotted line is where the coastline used to be. And as you can see, there's been a lot of erosion. Part of that is also subsidence, but at some points the coastline retreated with almost five kilometers. You might also notice that close to the old coastline, you see some dark green patches. Those are patches of mangroves that either survived or have recolonized. And often that's also where you can find a chenille right in front of that patch of mangroves. So if we zoom into one of those areas, I'll show a time lapse and I want you to focus on the area within the White Oval. Yeah, so over a period of 15 years, we saw a lot happening. Part of that is also of course because the photos were taken at different tidal elevations. But the chenille, it moved. It was sometimes a little bit more landward, sometimes a bit more offshore. There were photos where it was totally gone. But all in all, in those 15 years, most of the time there was a chenille more or less in the same location. That's actually pretty surprising observation because what we see all around the world is that chenille once they form, they will migrate onshore, and they will end up quite high on the shoreline where they can no longer be reached by waves and tides. So the fact that the chenille that we observed in Indonesia did not do this suggests that maybe for some situations there's an alternative stable state where the chenille can actually be relatively stable in the intertidal zone. So the first thing we did is we went to Indonesia and we wanted to measure this. Even though we planned our fieldwork during the northwest monsoon storm season, these were the waves that we measured. The highest waves we measured were around 50 centimeters at deep water, so that was very disappointing. But surprisingly, we actually still measured quite a lot of change at the chenille. So my hand is indicating the bed level only one day earlier. So we had one meter erosion in one place. And then the other set sensor was almost completely buried. So we had a lot of accretion. Actually the chenille pressed the highest part migrated up to eight meters per day, despite these really low conditions. So that was a bit surprising. And what made it even more surprising is that we know over the last 15 years that chenille was more or less in the same location. So what's going on? What are the processes driving? So for this we developed an idealized chenille model. You might recognize the setup. It was inspired by the barrier island model, but the big difference is that we are now dealing with a muddy coastline. So the substrate is mud and we assumed it to be static and we only looked at the dynamics of the sandy part because in the end that's what matters for the mangroves. If the sun can be there and create that shelter that it needs. And the objectives of our model was to investigate what are the processes that are driving these cross-short dynamics and also can we somehow explain a possible alternative stable state. Our reasoning behind developing an idealized model instead of using the existing process based models, for example, the L3D is that those models are computationally very expensive and we wanted to model a range of time scales going from seconds to decades. Also sometimes because there's so many processes involved, it's hard to differentiate which processes are actually driving the changes that we see. And then finally also some processes are just not well resolved right now. An example that is bar dynamics. They're known to generally flatten out bars and we know that the chenille remained relatively steep. So this was our reasoning and because we know so little of chenille dynamics itself, we had to find a way to drive our idealized model. We found kind of a hybrid approach where we still use the strengths of process based models like the L3D to calculate the migration rates. But then our idealized model is really fast to look at the longer time scales. So let me explain you how we did it. So if this is a profile, we simplified it by representing it by one dot, which is a chenille center we call it. It's the middle of the crest. And at every time step that crest center will move horizontally and vertically. And also because it's a muddy coastline and there's just that one body of sand we assume a conservation of sand volume. So if it moves it might also have to change in height and width so that the total volume of sand is still the same. And that delta X and delta Z is an hourly migration rate. And it's a function of the offshore wave height because of course waves are what most often drive sediment transport. It's also a function of the tides and the tide plays both a vertical role. Basically it determines whether or not the tenure is submerged. And if it's submerged, how high the water is on top of the tenure which we represent by H top. It also has a potential horizontal effect through currents. And basically if it's up or flood that determines direction. But if we think back about that coastline of the mark, what used to be a coastal plain is now flooded by each tidal cycle. And then during at all that water needs to flow out to a couple of narrow channels. And so the water encounters a lot of friction by the mangrove trees, but also houses and old ponds and everything that's now in a way and so basically that slows down the flow during that. And as a result, we actually get a water level gradient because the water level on the seaward side is already lower than the outflow from the coastal plain. And what happens in our model is by introducing a phase leg. So the blue line is the seaworth water level which is just a very simple type, and then landlord during app, we actually caused high tide for a bit. And then it followed the same curve so as a result, at the same time, you have a different one level on both sides of the tenure. Those Delta X and Delta Z we calculated with Delta 3D. But first we had to calibrate our model in Delta 3D using the measurements that we had. And then we simplified our profile. So now you can see the dash dotted line is the profile that we measured. The solid line is our idealized tenure profile. And then we have the dashed line which is kind of a generalized tenure profile, which was most suitable for our calculations and Delta 3D. And what those three profiles have in common is that one dots the tenure center so they're as high and they're in the same position. And we model the range of scenarios. All these values are what you can expect at our location in the mock during the Southeast Monsoon season. We did not have data to calibrate our model during the storm season. So that's why we focus on the season that we did have data for. And then the way we modeled it is we calculated for every scenario. We computed six hours first hour spin up and then four hours over which we calculate the average migration rate. And all those different scenarios led to different values of Delta X and Delta Z, which we stored in so called migration matrix. So on the left hand plot, you'll see the horizontal displacement for a set of scenarios, right hand plot is a vertical displacement. Horizontal axis is the water level on top of the tenure and vertical axis is the wave height offshore. And then you'll see the color bar of the rates. As you can see for, if we look at the left plot, for example, higher we waste mean more offshore onshore migration. On the same hand, on the other hand, the low water depth combined with high waves lowers the tenure quest. These were kind of expected results. But that was for a situation without that face like, if we include, for example, a face like of 10 minutes. So that means that now the land boundary is lagging behind on the seaward boundary. We see that now the horizontal displacements, if we have a large water depth combined with relatively slow, low waves, instead of the tenure migrating onshore we have a tenure migrating offshore. So it's interesting that we see this dynamic because that's something that we were interesting. And also for some situations, the tenure quest actually heightens, so there's not only lower. The next step was to calibrate or validate our idealized model. The first ones that we did are idealized model always predicted that the tenure quest would lower. We think that that was a result of del 3d actually flattening the bars in their model. So we mitigated this by introducing a small vertical correction of only 1.5 millimeters per hour. And with that manual correction, we actually are able to reproduce barely well it's not perfect, but we can see the trends for the first couple of days, you can see that there's barely any change and then suddenly there's a landward migration and lowering after the change is slowed down again. So now that we have our idealized tenure model set up, it's time to do some longer runs. So we developed some very idealized boundary conditions first to really understand what was going on. We had a very simple tide with only two constituents. We had a very regular wave climate in this area. You mostly have waves in the afternoon as a result of a sea breeze. So we basically just apply the constant wave condition every afternoon for a period of 10 years, I will start the animation. The line will show every year, and you'll see a dot appearing the dots are every month. So we have reproduced the classic tenure behavior of waves pushing the tenure onshore until they're up on the on the shoreline. Now if we instead of having a set of zero we introduce this phase leg. Let's see what happens then. So now we have a balancing act between waves trying to push a tenure onshore, and then that title outflow that pushes the tenure offshore. It makes this very regular loop and that's a result because in this area you have the sea breeze so the waves are always at the same time of the day. And then, because it's mostly diurnal tide, you have your timing of your high tide that shifts throughout the year. So, depending if the high tide coincides with the waves, or the low tide coincides with waves makes the waves dominant or the tide dominant. But all in all, this is actually what we were looking for because you can see that within the year the tenure is quite dynamic heightening low lowering, moving onshore moving offshore. But on the longer term 10 years, it's still more or less in the same location. But yeah, I admit it might look a bit too idealized. So, we also looked into a bit more realistic conditions. The way we did that is one we included all the title constituents not only this very regular tide. But then also, not every day we have the same wave height so the sea breeze is more or less every 50% of the days we see a sea breeze, and then also the strength of that wind and therefore the height of the waves carries. So we made a stochastic boundary as a stochastic wave climate. And then the last thing that we were having a trouble with is that right now for those 10 years we just assumed. Southeast monsoon conditions the whole year but that's not true that's true for eight months of the year and a four months of the year, you have the northwest monsoon season, but onshore winds and therefore higher waves the whole day, and occasional storms. But we did not have a way to validate our model, however, based on satellite data we know that the net effect of that storm season is that after the storm season the January is more seaworth. So we implemented this by adding a delta x storm so we would run our model for eight months, and then have a seaworth displacement and then run for another eight months. So first I'll show you again a situation where you don't have a storm season. I don't have a face like but you do have full title set of constituents and varying waves. We still see that same onshore migration but instead of being the same rate every year you now have more or less migration depending on the way so that already looks more realistic. The next thing we do we do not include here is for example storm search, because in the end that's what causes a tenure to be so high up on the coastline that it's totally stable because it, it cannot be reached anymore. So that will be like the next step to add to the realistic boundary conditions. But then we also added the situation with a small phase like and a storm season effect. It's a nice animation because it basically was a bit chaos, but what we would see is that the tenure was moving all the way everywhere, but all in all it's still pretty stable. So yeah, in summary, we developed an idealize tenure model. We now understand that the main drivers are the waves and the tides and especially the timing between waves and tides is very crucial. So this model is able to capture the classic tenure behavior of a tenure moving onshore until it's high on the on the on the shoreline. We're also able to explain an alternative stable state, which in this case is caused by substance of the coastal plane, which causes a phase lag. And depending on the timing between the waves and tides, you'll see that either onshore migration or offshore migration is dominant, which leads to on a short term very dynamic tenure. Which is good news, because even though the window opportunity for mangrove recolonization is pretty short in the order of days to weeks to build up a much that that is high enough to actually allow mangroves to recolonize, you need a tenure that's present for months to years and we think that with this, this dynamic that could be the case. And we hope that therefore mangroves are able to recolonize naturally. Yeah. Other questions for silk question. Thanks so good. That was a fantastic talk. I think I heard you say at one point that Delta 3d was flattening the junior and the you are putting a constant rate of accretion. Did that. Did you run simulations with without that. And was that kind of a consistent thing across all wave. Yeah, so if we did not apply this vertical correction, which is still a very small correction, it's only one and a half millimeters per hour. What would happen is that the tenure would lower and because we have a conservation of sand volume. It would make it a very wide tenure and in the end we would just always end up with a layer of sand over the entire shoreline shore face. We don't think that was a realistic behavior, but we would see it across all scenarios. Do you think there is an opportunity for parameterizing that in the future and what and if so, how would, how would you go about doing. I think there's already a lot of efforts trying to improve the bar dynamics in Delft 3d. Um, so I have not really looked into those, but I was hoping once that improves, we would not have to have to apply this very. It's a very stupid, just silly constant correction. Well anyway, thanks so much that was great. If time for another question. I think you put into your model that you always conserve mass is that. And so then your senior remains. But is that always the case that you always come to us. Not really. It's a simplification. What you will see is also similar to barrier islands during very energetic storms you might have overwashed. And then actually in the mud flat behind on the back barrier side of the tenure. It's mostly mud, but you have these very thin sand layers due to overwash. So there you lose a little bit of sand. On the other hand, on the seaward side. You might still have the winnowing process is basically is basically sediment is being sort of by waves. The very fine mud remains in suspension, the sand forms a very thin layer on the bed, which is then by wave of symmetry, push towards the shoreline. So you have always a little bit of adding sediments and I also have a little bit of losing for this model we kind of assume that those balance and that therefore our conservation of volume was realistic. Will you move to modeling those other so like I know where I live, which is the Gulf Coast. We're losing our chinears. And so I'm not sure if something like that could be modeled. Yeah, I actually would love to study those chinears, because it's a whole different system they're much bigger than what we saw here. But yeah, yeah, I don't really know what's going on there. It's a good good place for you to do your next. I encourage you. So I might have one question. So I noticed that so you incorporated a lag time. And I saw in your presentation that it varies between, I think 3.5 minutes and 10 minutes. And I might have missed it. Can you explain why you experimented with that time and is it something you have tried to measure it in the field as well to see if that's realistic time like or not. Yeah. So the total range remodeled was between zero and 20 minutes. We don't really have right now a physical explanation of what would be a realistic time also we're not sure whether that's truly constant over the title cycle it might also depend on title range. So we tried measuring it, but our model, our measurements were not accurate enough to really be able to say this is a title range. I think it's like the 3.5 minutes is more realistic than the 10 or the 20 minutes, but to be able to capture the dynamics we also tried to figure out. Thank you. Thanks for the questions.