Bridget: Welcome to the Water Resources Podcast. I am Bridget Scanlon. In this podcast, we discuss water challenges with leading experts, including topics on extreme climate events, over exploitation, and potential solutions towards more sustainable management. I would like to welcome Clint Dawson to the Water Resources Podcast.
Bridget: Clint is the leader of the Computational Hydraulics Group at the Oden Institute at the University of Texas and also serves as chair of the Department of Aerospace Engineering. I've known Clint for many years and we're currently collaborating on a a DOE project on coastal issues. Clint's research focuses on numerical methods and parallel computing and today we're going to discuss applications to coastal flooding using the ADCIRC model, Advanced Circulation Model.
Bridget: Clint has received numerous awards throughout his career and most recently the President's Research Impact Award at the University of Texas for his research on data driven storm surge. So today we're going to talk about extreme hurricanes, storm surge, modeling using ADCIRC, operational forecasting. And then if we have time, we'd maybe talk about ADCIRC applications for infrastructure design.
Bridget: So thanks a lot, Clint, for joining me today. I really appreciate it.
Clint: Thanks for inviting me, Bridget.
Bridget: So you've been involved in, I mean, your research focuses on numerical methods, but you've been applying those to coastal modeling and storm surge modeling for a couple of decades, several decades now, and maybe you can describe some of the recent applications to storm surge, maybe a recent hurricanes like in 2022 or whatever, how you applied ADCIRC in those situations, and what the situation was like, how much storm surge they had and relative contributions of other components to sea level rise and things like that.
Clint: Right. So we've been doing ADCIRC. Well, I've been doing ADCIRC modeling for about 30 years and we got into the storm surge application in the late 90s and early 2000s. So pretty much any of the major hurricanes since then we've had. In the United States, we've had some hand in modeling those either from a historical perspective, looking backward or in a forecast perspective. The most recent hurricane that's probably on everyone's mind was Hurricane Ian that hit the Fort Myers, Florida area. And then, of course, there was Hurricane Ida and Laura that hit in Louisiana. And so even going back to Hurricane Harvey, and then back earlier to Katrina, Gustav, all of those storms,
Hurricane Ike was a big Texas storm, and was a kind of a difficult storm to capture primarily because the hurricane was way off for a long time. It showed the storm making landfall in the Florida panhandle and about, I guess, 2 days before the storm, it actually started curving to the east and it hit the Fort Myers area. Which was an area that had not experienced a major hurricane in a long time, many decades, and so they were really not prepared for it and they had, I had a lot of coastal infrastructure that was like housing and and so forth that was not well protected. The hurricane also was a bit of a surprise in the sense that it started out kind of a category one hurricane, and it went over, I guess, a warm spot in the Gulf of Mexico, and it just sort of blew up to a very strong storm very quickly. So, I think it caught a lot of people off guard.
I would, I would refer people to a website at LSU, the CERA website (Coastal Emergency Risks Assessment), where they can look at the results, the answer results for Ian. Those showed several feet of storm surge, especially on the coast, and then going up into the sort of channel that's in the Fort Myers area. And it was, I mean, it was a devastating storm. I guess we haven't done a really systematic study to at least, I don't know exactly the contribution of sea level rise. I mean, it, there's a lot of factors in these storms. So 1 is whether it happens at high tide or low tide. So, there can be a, that can either contribute to storm surge, or it can help subtract from the storm surge if it's at low tide or high tide. So that's 1 factor. And then, of course, sea level rise is a recurring thing that we're trying to get a handle on, but probably even more importantly. Then just the sea level rise itself is that the in the summer, the ocean expands so sea level goes up and down anyway, because of the expansion of the ocean.So that expansion could be as much as a foot or so. And that can add to the storm surge and that's kind of hard for us to capture as well, because we don't always know. what the exact expansion should be in a given area. But it was definitely a wind driven event, so a lot of storm surge driven by wind, and then always some rainfall on top of it.
Bridget: So, so you mentioned, so it seemed like a pretty dynamic process then. So you can see these hurricanes, or these storms coming across the Atlantic, and, but then they can change as they come towards the coast. So it's difficult then to predict where they might landfall and stuff like that. And then they can amplify or decrease in strength. And so I think I read that maybe Ian was category five and then fatalities, 150 fatalities, which is because maybe they weren't ready for it. And $13 billion in damages, so a lot of impacts. But so, with your ADCIRC simulations then, and you said it ended up landfalling at a quite a different place, was that included in your simulations and when, or you just had to expand a couple of days ahead of time then to capture that?
Clint: Yeah, so the way it works is in a forecast model, we get information from the National Hurricane Center, just like everyone else. So there's the cone of uncertainty. And then there's the what they call sort of their best guess of the track at any given time, and that's based on a number of meteorological models from around the world that the, that are run. And then the Hurricane Center takes all that data and puts out this forecast every 6 hours or so. And so we take that information and we process it and we generate a synthetic model of the wind based on a few parameters. So it's a vortex model, but the vortex can have different shapes and different parts of the storm. So it's a sector based. There's like four sectors of the hurricane and we generate a vortex in each sector because one side of the hurricane may be stronger. Hurricanes are not necessarily completely symmetric, even though they look symmetric a lot of the time, but what was happening in Ian was I we were relying on those forecasts, so our storm surge predictions were tracking the predicted track of the storm and we so we start modeling as soon as there's a forecast. So that may be really early in the storm. So, of course, when the forecast started to shift, then our model results also start to shift. But we have to depend on those model results, those wind predictions. We're not generating those ourselves. That's done by a different community. And so once once we had the, forecast that had kind of locked in on the area, the simulation was was okay, but there were still errors in the wind field that, that precluded us from making really, I think, super accurate predictions in the forecast setting, but generally, we got the forecast correct. So what happens after the storm is the meteorological community does a reanalysis of the event. So they take data. They measure data, they take the actual storm parameters, and they feed it into various models, and this is done by, for example, NOAA does this, but there's also some private companies that do this, and then they release sort of data assimilated or reanalyzed wind fields. Those are typically more accurate, much more accurate than what we get in forecast mode. So then you can go in and try to do a sort of a true analysis of what happened. During that storm and then see how accurate your model is for that given area are those reanalysis products and the hind casting that you do.
Bridget: Is that used for damage assessment or with the insurance.
Clint: I know that's a good question. So that's really up to the state emergency managers. And I don't know if that's the case in Florida, but I do know in Texas, very important that they get as much data as quickly as possible because they have to do a rough order of calculation damage assessment for FEMA for the federal government so they can do a Disaster Declaration and so that can make loans available to people or make money available so that people can start to recover So in Texas, I think our results are used along with visual evidence and what people are seeing on the ground and what other impacts there might have been.
It's also difficult. I mean, it's separating the wind impacts from the flood impacts is important from insurance perspective, because wind damage might be covered under 1 type of policy, but flood damages it's usually not covered in a standard homeowner's policy. You have to have separate flood insurance. And so they really want to know, like, how much was wind and how much was like if a house is destroyed, was it destroyed by wind or was it destroyed by the flood?
Bridget: And I was talking to this morning to a colleague, and he mentioned that oftentimes the flood damage is the biggest component. It could be like 70 or 80 percent and wind damage might only be 20 percent on average.
And so. I guess that's important too. Maybe step back a little bit and there's a lot of terminology and stuff that people might get confused about and the classification of hurricanes, tropical cyclones, storms, they're all kind of based on wind speed. And now I think there's some discussion about adding another category to hurricanes for very intense storms.
So I guess tropical cyclones are winds less than about 40 miles per hour, 38 miles per hour. And then tropical storms then extend from that to 73 miles per hour. And then hurricanes then are above that and with five different categories. So are they talking about another category. I mean, the category five was greater than 157 miles per hour.
Are they talking about possibly adding a category six?
Clint: Yeah, that's what I heard recently is that they were thinking about addin a category six, which for us, the category of the storm is just. I guess it's just, it's what people are used to talking about, right? So it's what they focus on, but it's a bit misleading when it comes to the damages from the storm.
I mean, for example, Hurricane Ike, which was a very damaging storm. In terms of monetary damages and property damage and so forth was only a Category 2, a strong Category 2 hurricane. It wasn't a Category 5. It was big, it had this very big wind field and it had all this momentum behind it. And so all of that water just got pushed into, of course, a very densely populated area.
And, but it's just a Category 2, whereas Harvey. When it made landfall, it was a Category 4, but it was like, it was wound really tight around the eye, because that's kind of what happens when the, as it gets stronger, it tends to wrap more tightly around the eye, and then as it kind of, it may spread out. I mean, it really depends on how big, how, what the radius of the storm is.
That's probably more important than just the, the wind speed at the center of the storm, because that's really local. That's only going to affect people that are really caught in those high gusty winds, but outside of that, the winds may be. No, you could be in a category 5 storm. That's really small, very tightly wound.
You could be 20 miles away from it or 30 miles away from it and not even could just be a breeze going on. So, I think people get caught up in this notion of category, and it's easy for weathermen to talk about because people are used to that, but it's, it really doesn't correlate with, well, with the storm surge.
Bridget: Maybe you should gin up your own, Clint.
Clint: Well, we've had conversations. We had a, we did write a paper on some years ago with the group at Rice University using, it was a kinetic energy factor that, we, We did some correlations between kinetic energy of a storm, which is a quantity that one can measure from various parameters and the potential storm surge and it, I don't remember the paper was written about a decade ago. So, it seemed like it was a better indicator of storm surge than just going with the pure category. Category is just such a simple thing. It doesn't really mean a whole lot. It means it's one data point, right? Right? Massive event. Right, right. What they really should say is that they should give you, not only the central, the wind speed, but also the radius of the storm.
How far out do these winds extend? And which they do, but people don't pay attention to that. On a short weather forecast, people don't pay attention to that. So they don't know exactly what their risk is if they're not right in the eye of the storm.
Bridget: But so, I mean, when you think about risk, then we're talking about the hazard aspect.
What's the hazard from the event? But then you've got the exposure. What's in its path and how densely populated? What's the infrastructure and everything? So, so really for a true risk analysis. You'd need all of that. And the vulnerability then. So hazard exposure and vulnerability would need to be incorporated.
Clint: Right, right, absolutely.
Yeah. Which is a mix of sort of social science and the physics, which. Yeah, which is, which definitely, and also identifying the vulnerable populations that need to be evacuated and then those that can actually shelter in place. That's another, and that's of course where emergency management comes in. Right. Yeah.
Bridget: And you mentioned earlier about wind damage versus the flood damage. And I guess, and you also talked about Harvey in 2017 in Texas. So Harvey was an outlier, a very unusual.
Clint: Well, yeah, it's the only hurricane I can remember that made landfall several times. I mean, it went in, made landfall initially. It went inland for a while, and then it got pushed back out into the Gulf, and then it moved up the coast, and that's when it dumped all that rain in the upper coast, and of course, made landfall again around the Texas Louisiana border. But it was a, it was a compound, what we call a compound event. It was a storm surge event, and then followed by an extreme rainfall event.
And it could have been just one or the other. I mean, if it had been a, sort of an Quote unquote normal hurricane, it would have made landfall, gone inland, dissipated. It might have rained a bunch along its track, but it wouldn't have just hung around for four or five days like it did and just sort of moseyed along the coast and did so much damage.
It was an outlier in the sense of the way that it tracked. It wasn't so much an outlier now. And how much rain we could expect to get from a hurricane. I think that is what we are going to have to get used to is that hurricanes are probably going to be a lot more devastating in terms of just the amount of water, the total amount of water that gets dumped on the land from either the ocean or from rain
Bridget: and because Harvey stalled and then just dumped like 40 to 60 inches of rain and in those regions that the flooding was off the scale.
Clint: Yeah, I mean, they're not that part of the Texas coast is not, It can't possibly handle that much water. It's too flat a place for the water to go and there wasn't much storm surge up there in that on the upper coast, but there might have been just enough. I mean, you still had when pushing water. Inland to some degree, and if that had been a little bit more extreme, let's say the wind had been a little bit stronger so that there was some significant storm surge. That's just going to make the event worse because it's going to keep the water from being able to drain out as quickly, so it's, it's just going to it's not going to be able to follow its normal flow paths. Those will be blocked, and so it has no choice but to just spread outward. And another thing that we saw in Harvey, which I don't think we've ever really observed, is we had flooding across watersheds.
So it actually, water, the water was so deep and so spread out that it actually went from one watershed to another. So that was. I mean, we're not used to that either.
Bridget: That's it. And I guess when we do those simulations, I mean, normally we would do the pluvial flooding from the rain on the ground and then the fluvial flooding in the rivers and then the storm surge.
So all of those contributing to this compound issue, a compound flooding. And then, so normally we look at the watershed and in a river basin, but we don't really take into account the water can come from an adjacent watershed and really amplify it.
Clint: Right. So that, I think they recognize that and they realize that when they do the modeling of the rainfall part of the storm, we have to consider the holistic system and not just each watershed as an individual.
Bridget: Right, and I was reading some places, maybe the storm surge was about 10, 12 feet in some areas or something like that. But then I was also reading that Brazos River at one of the Richmond gauges, it crested at like 55 feet. Yes, that's correct. Yeah, and so in these flat areas, then there's no place for the water to go.
And so then it hangs around for a long time. So it takes a long time to dissipate. We're always maybe focusing on the upfront part of the hurricane and the impact, but really the long term and, how to manage, long term is, is important aspect of it too.
Clint: Absolutely. It's, it's a very complicated issue because of so many different constituencies that sort of have to be satisfied in these kinds of mitigation studies when you try to plan for what maybe should be done in the future. That's when it gets very political.
Bridget: Yes. And, and you also mentioned Hurricane Katrina and, and I can remember that because I think a lot of people from Houston came up to Austin that weekend and, I got caught in an accident coming off of 35 because the people didn't know how to get off 35.
And then, thank goodness there was a vehicle behind me because there was an 18 wheeler behind them. And man, I would have been dead meat in my little vehicle. So evacuating also has its difficulties.
Clint: Well, yeah, I think the evacuation you're thinking of was Rita. Oh, right after Katrina, but it scared everybody to death.
So they all of Houston came to central Texas. Basically, people were stuck on the highways for like 14 hours. They ran out of gas. It was a night. And then, of course, it didn't even hit Houston. It was just a complete non event. But yeah, Katrina, of course. Was, that was a failure of the levee system that wouldn't have that Katrina wouldn't have been perhaps so devastating except in Mississippi.
It would have been anyway, but to New Orleans, if it hadn't been for the failure of the seawalls and the protection system there.
Bridget: Right. And the fatalities were huge from that, like about 1,800 people, 1, 800 people. And from Harvey, I guess about a hundred. So that was it. And you mentioned in our previous discussions that Katrina was sort of a turning point for you guys with the ADCIRC modeling. It happened in like 2005. And so maybe you can describe a little bit about how that affected your modeling and what you did in response and stuff
Clint: Sure. So the, so if we go to say the early2000s, first of all, let me just, since we're getting into the modeling discussion, I just want to explain that I, I don't, I'm part of a large group of researchers and I got into this field through a collaboration with Joannes Westerink at the University of Notre Dame, who's been a long time collaborator And then Rick Luettich at the University of North Carolina, there were the, those two were the early developers of ADCIRC back in the early 90s.
And I got to know them in around 1993 or 1994. And we've been working together ever since then. But we, so my sort of initial contribution to the group was to take an existing code. And make it run fast on high performance computers. That was my, my initial, contribution to this modeling effort.
And what that allowed us to do was to run much bigger cases. I mean, we could, before that, if we ran a model with, I don't know, 20, 000 degrees of freedom, it would take a week on a computer. And then, on the parallel computers that we now have, it would, it would basically run as fast as, I mean, however many processors we could get.
We could, you divide that number by the number of processors, and that's how fast the code would run. So, maybe if it was a hundred processors, it would run a hundred times faster. Well, that makes a big difference in your productivity. And so we were able to go from much smaller models to much larger models, which had a lot higher resolution, a lot more detail and so forth.
Clint: So that happened in the late nineties and early two thousands. And then we were funded to, by the army corps of engineers in New Orleans to develop a storm surge model for Southern Louisiana. And that's because they knew they had problems, right? They knew that they were vulnerable to storm surge. And then it happened and then so we did that.
We actually, we had to validate it for a historical hurricane. And the 1 that we validated it on was Hurricane Betsy, which happened in the 60s. So they hadn't had a major hurricane in a long time in New Orleans. So then Katrina comes and opened it was such a shocking cataclysmic event that it really changed the way people thought about hurricanes.
And people started paying attention, and I think the federal government then woke up and realized that there needed to be significant amount of research from different angles poured into understanding hurricanes. And then, of course. Katrina, it wasn't Katrina, then it was just kept one after the other and multibillion dollars in damages and many lives lost in hitting important areas of of our coasts that are economic drivers.
For our entire economy have a lot to do with the oil and gas industry and chemical industry, especially were impacted. And so after Katrina happened, there was questions of, well, how can we prevent this from happening? What went wrong? And how can we prevent this from happening in the future? So there was a study called the iPad study with the government different agencies, I guess, contributed and an answer was used in part of that study to study the failures of and what could be potentially done to improve storm surge mitigation in Louisiana, and that led to the building of a new protection system for especially for New Orleans. That is a system of gates and levees and pumps and so forth that has proven to be pretty effective.
For example, in Ida, New Orleans didn't get any storm surge. They had rain, but they were able to pump it out. They lost electricity because of the wind, but it wasn't a major surge event for them.
Bridget: Right. And you mentioned parallel computing and then increasing the speed by about a hundred times. So then that allowed you to use ADCIRC in operational mode then to produce these operational forecasts. And then also, I guess, another aspect of it is unstructured grids or variable grid resolution. Maybe you can describe that a little bit.
Clint: Yeah. So that just, that means that if say to, to accurately model a hurricane, you have to include large chunks of the ocean in your computation. So our domain typically, I mean, this. It's the, we're getting now to sort of the global where things are all domains are global, but back when we were doing Katrina, our domain went out into the, about the middle of the Atlantic Ocean, so Western the North Atlantic. And then included all of the eastern seaboard, as well as all the Caribbean islands in the Gulf of Mexico.
So that's important because you want to get the complete track of the hurricane. Hurricanes tend to come off the coast of West Africa, or they develop in the Caribbean, or sometimes even in the Gulf of Mexico. And they track for several days before they make landfall. And so that grid resolution allows us to extend our domains way out into the ocean, where it can be very coarse.
And then as we get closer to the coast, we can have finer and finer resolution so that we can capture all of the continental shelf and then even inland, because of course the flooding happens inland, we capture like very small-scale channels. So it goes from, like, the kilometer scale to the meter scale to the tens of meter scale in the inland parts.
Bridget: Right. And also, when you were doing this in parallelizing the code and everything, then we had, you could also take advantage of the Texas Advanced Computing Center and their high performance computing. So all of these things sort of come together then to help facilitate this rapid turnaround in the simulations that allowed it to be used operationally.
And so the national hurricane center puts out a forecast every six hours, typically, and then you then simulate based on that. And it takes you a couple of hours or whatever. And then you provide it depends.
Clint: So first of all, it's a great, there's a great that. It's sort of in charge of the running this sort of because the operational forecasting model is a system.
It's a system of script of coding. It's a coding framework that does everything automatically. So running answer and is one Component of this system, but it's got to operate basically without a lot of human intervention. It's got to run 24 hours a day. It can't break. It's got to download data process it spawn off computer runs when the computer runs are finished, take the output process, post process it and then post it automatically to a website. So all that happened. So that took many developers many years to get this to where it doesn't break and so, yes, we use tech extensively in that in that effort.
But, yeah, you're right. So, as soon as they release a forecast, we take that data. We process it. We run, so the forecast includes a nowcast, so they correct their previous forecast based on the data that came in the last 6 hours. And so they give you an accurate what they call best track. Of the storm up until the present moment.
So we take the the corrected nowcast for the last 6 hours. We run that too and then we start the forecast and that forecast thing goes out maybe 5 days or whatever. However long the forecast. And so there's multiple steps that have to be done in that process. And yes, we, so the nowcast takes a few minutes because it's only 6 hours.
And then the forecast might take anywhere from 15 minutes to an hour, depending on the resolution of the model that we're using. So, as the storm gets closer to land. We start using higher resolution models to try to capture it. Maybe in the early part of the storm, we might use a coarser model because one, you don't know where the storm is going to go exactly.
And two, it's just runs a lot faster. And the wind is probably wrong anyway. So you're just trying to get a rough order of magnitude at that point.
Bridget: So, so those simulations in that forecast is kind of like your best guess based on the hurricane center output. And then, as you mentioned earlier, then the output from your forecast then are released through the Coastal Emergency Risk Assessment, CERA website at Louisiana State University.
And then anybody can look at those and use those. And I think you mentioned one time there was so many downloads that it kind of, Kind of crashed. Yeah.
But that's great that you can do something that people are interested in and it's so impactful and the emergency managers incorporated into their planning and everything that's really incredible.
And I mean, I know you mentioned a team and everything, but really an academic setting and I guess initially maybe up to five to 10 people and now maybe a hundred that's still relatively small considering all of the things that you actually do. So ADCIRC is advanced circulation is one storm surge modeling effort.
Did there are others, you mentioned national hurricane center and others, maybe you can describe some of these other codes that are used for a storm surge.
Clint: Sure. So, well, SLOSH (Sea, Lake, and Overland Surges from Hurricanes), which is their forecast model. And SLOSH wa developed specifically as a storm surge code. It runs extremely fast. It's uses sort of a simplified physics of the system, and they can, because it runs so fast, they can run like many ensembles of different wind fields, and then they can give you sort of the statistical picture of the storm, but the way that SLOSH is. I guess it's, it's physics. It tends to, in many cases, in some cases, it tends to, well, let me back up. We have seen where it has over predicted the storm surge in some areas. That doesn't mean that it does it all the time. But, for example, in Harvey, if you took the SLOSH results just by themselves, you would have thought that Corpus Christi was going to be flooded.It turned out that Corpus Christi was not flooded. So our model, which is. running on a different mesh, or the ADCIRC model, which is running on a different mesh, and use it, but still using has and also has some more, I guess, full physics isn't just another, is another decision tool in the toolbox, right?
So there may be some storms where our model is, turns out to be more accurate or maybe SLOSH is equally accurate. Or, I mean, it's a trade-off between the experience that people have with our model versus SLOSH or, what, but it's just like with the wind model. They run multiple wind models, and then they take all of that data into account in their decision making.
So ADCIRC is just one tool in the toolbox. There are other models. There are not very many models that are run in operational mode just because of the well, it's It's You know, we're not an official, we're not a government, we don't, we provide information if government agencies, if state emergency managers use it, that's great, but we're not an officially blessed.
I'm not trying and we're not an effort by the Hurricane Center to provide a storm surge guidance. That's SLOSH. It's what they're supposed to use. However, other parts of government use our model. They may use it at different times during the event, but ADCIRC is a component. The Army Corps of Engineers uses it.
Other models that are used are, like, after the fact, the, a lot of models can be used to study what happened during the storm. So, for example, Delft3D is one. SCHISM is one. For different parts of the coast, it might be FVCOM or ROMS or, anyway, there's a number of storms, and usually they're sort of, They were developed for a specific area like FECOM is used more in the northeast because of where it was typically what it was developed for other codes may be used along the eastern seaboard as CERA used a lot in the Gulf of Mexico, just for historical reasons. And so, but in terms of operational forecasting, there's answer has cemented its place as an operational forecast as a pretty well, well, trusted operational forecast model.
And that's just through years of doing it and showing that our results are accurate, right.
Bridget: Right. So I think it's amazing what you guys have accomplished in an academic setting, relying on students and postdocs and cobbling funding together.
Clint: Yeah. Different sources and getting by on a shoestring and doing some things for free, by the way.
Bridget: Right. Right. So, I guess one of the things that's important when we're looking at these storm surges is what are the controlling factors or what are the important controls on storm surge? And I think through your modeling efforts, you kind of have identified some of those things. Maybe you could describe those a little bit, Clint.
Clint: Yeah, so the, the major factors are, of course, the wind speed, the, and then in the coastal area, it depends a lot on the, the coastal topography, the, the bathymetry. So the, the height of land above sea level, if you. In Texas, it's, we have a very low lying coast, so storm surge can propagate way inland. Now, we do have barrier islands, but they're only a few feet above sea level, so they're easily overtopped.They can get eroded during a storm, which I think happened during Harvey. And, but they serve as sort of the first line of protection. And then behind that, though, you have big bays like Galveston Bay, which can generate its own storm surge. And then beyond that, you have the what's called the bottom roughness, which is the once the water starts to flow over shallow in either And once the storm surge gets into sort of shallow water, or even over land, it really matters whether you're whether it's flowing over grass, like low lying grass, or whether it's flowing through like a mangrove forest. A mangrove forest is going to act like a barrier, whereas grass is just going to get, it's just going to, it's just going to lay over and it's just, then a storm surge has nothing to prevent it from going. And so in Texas, unfortunately, we are mostly coastal plains with not much vegetation to protect us from storm surge. So there's these questions of, can we build more natural barriers, but like, if you want to introduce mangrove into the ecosystem here, well, it's not a native plant. So what are the, do you really want to, do you really want to introduce an invasive, potentially invasive species into a environment where it's not in its native environment. So that's always a question like, okay, it seems like a good idea maybe, but who knows what the long term impacts will be. And it's the same thing for structural solutions. I mean, those have long term impacts as well.
Bridget: So one aspect of your work is looking at risk and you work with DesignSafe on natural hazards. So maybe you can describe that a little bit, Clint, it's an interesting concept.
Clint:. DesignSafe was funded by the National Science Foundation through the Natural Hazards Engineering Research Infrastructure, or NHERI, program. It is a multi-pronged, NHERI is a multi pronged effort, so DesignSafe is one part of it. And I've been involved in DesignSAFE TACC. Heavily involved in DesignSafe, Ellen Rathje in the civil engineering department is the PI, but it also involves people at UCLA and Rice and University of Washington and other places. So it's got different components to it. Earthquakes, wind storms, and then storm surge is one component.
So we use DesignSafe. It's both a simulation framework. It provides a simulation framework as well as data storage and archival capabilities. It's it's been very useful for us for doing research on parameter estimation and uncertainty quantification and machine learning. It has a lot of capabilities. Like you can run Jupyter notebooks and you can so you can set up sort of a framework of like running multiple ensembles of answer. Which you would need to do in a, if you were doing like a risk assessment study. And then we use it a lot to store our data where we can store different projects and then we can publish them so that they get a DOI and you can, and then they're not lost.
People can reference them and they're not lost, which a lot of our data, unfortunately just gets lost over the years. We don't know where it is.
Bridget: Well, that's very interesting. And another aspect when we spoke previously, you mentioned that ADCIRC is used also for designing infrastructure design, and there's a lot of discussion these days about the Ike Dyke and those sorts of things.
Maybe you can describe that a little bit.
Clint: Right. So yeah, so the Ike Dyke or the coastal spine was. I think the idea for it was even pre Ike. I think there was a report done by the Army Corps of Engineers about building a coastal spine that would protect Galveston East on. And then after Ike happened, it just seemed like, it had to be done.
Of course, here we are 16 years later and nothing has even started. But yeah, so it's a system of dunes and gates that would extend along the upper Texas coast, but primarily from the lower part, the Western part of Galveston Island, through the peninsula. That's the main focus of some of our research.
So. Depending on how high they build it and exactly how it. What it's made of, it will probably hold back some modest storm surge the problem. The, the sort of downside of that is that it can be overtopped. And so storm surge can still get, it can still flood Galveston Island and Bolivar Peninsula. And once that storm surge gets into the Galveston Bay, it can be propagated up the Houston ship channel and into the city of Houston.
And the bay itself is big enough to generate storm surge. So just having a hurricane go over the bay. So there's a another idea that's come out of Rice. That is to build another, to add on another system of sort of islands and gates inside the, Galveston Bay along the ship channel that would protect the Western part of Galveston Bay, which is where most of the infrastructure, the chemical plants, the oil refineries, the shipping, the port of Houston, that's where they are.
So, I mean, if those got knocked out, imagine if the port of Houston shut down or. The impact on the economy would be Which, I mean, it would be in the billions. that port is essential to our economic vitality. So there, there's, it's still in the research stage, but we've had many years now of discussion where to the point where they're getting a lot of public input on it. And of course it's controversial. This is anything. Anytime you're going to build a structure that looks like it might interfere with recreation or someone's beach house view, people are going to. Of course, react perhaps either positively or negatively.
So now it's sort of in the, it's been rolled out to the public. The public is aware of it. And, to what extent that it gets built, it's still, I'd say, a decade away from still probably the building of it is still probably a decade, probably sometime in the 2030s.
Bridget: Right. Well, I really appreciate your taking the time today to talk about all these issues and it was interesting to hear how your contribution to the ADCIRC modeling and helped result in operational forecasting and speed it up and apply high performance computing and parallelize the codes and all of that sort of thing. And then, so that helps with operational forecasting. And then just now, I guess in design mode, that's also extremely interesting. So, thanks for your time today, Clint, and IS really appreciate it. Good luck with your work.
Clint: Thank you, Bridget
