Wow, on cue, it's just like class. Everyone gets quiet. So good afternoon, everybody. Welcome. My name's Jamie Phillips, professor and chair of electrical and computer engineering. And it's my honor today to introduce Rachel Davidson as the Donald C. Phillips professor. I think they picked me to do this just because of the name. We actually have no relation. So the professorship honors Donald C. Phillips as an alumnus of the University of Delaware's class of 1948, through his estate provided a generous gift to the Department of Civil and Environmental Engineering, and Mr. Phillips earned his bachelor's degree in civil engineering, went on to lead an accomplished career as a construction engineer, and I'm sure he would have been very proud today to see the recent renaming of the department to the Department of Civil Construction and Environmental Engineering. So Rachel Davidson is Associate Dean for Academic Affairs for the College of Engineering and a professor in the Department of Civil Construction and Environmental Engineering and a core faculty member in the Disaster Research Center at the University of Delaware. She's also the PI and director of the NSF -funded Coastal Hazards, Equity, Economic Prosperity, and Resilience, or CHEER Hub. So anybody that works on disasters and turns it around to be named CHEER obviously is an eternal optimist and a very positive thinker. So I can clearly remember what I interviewed at the University of Delaware in 2020. And Rachel was in the college dean's office as a representative to speak with me. And it was obvious to me at the time that she cared deeply about the college community, our students, faculty, and staff. And honestly, that was one of the moments that very much confirmed this is the place I wanted to be. So thank you, Rachel. I've had the privilege of working with Rachel as the department chair and as interim dean. and I've been amazed at her commitment to the college and honestly much of what we do with faculty affairs, mentoring, college recognitions and awards, tenure and promotion, and as a result of her leadership, it's a lot of what is really done behind the scenes, out of the limelight, without a lot of the credit that is richly deserved. So I want to take a moment to say that at this point. So in particular, she's been an outstanding leader and champion in the college for efforts in an area that has been under scrutiny in recent days, but I'm not afraid to say it here, diversity, equity, and inclusion. And as I've heard Rachel referred to it recently, excellence together. So one of our strengths in this college is definitely our community, and Rachel has been at the forefront in fostering an outstanding community here in our college. So I think many of you have worked with Rachel in her administrative role, and today we have the fortune of hearing about Professor Davidson's research scholarship, and in addition to directing the Cheer Hub, I wanted to point out a few other awards and honors. She's a fellow and past president of the Society for Risk Analysis, winner of the ASCE Charles Martin Duke Lifeline Earthquake Engineering Award, and a fellow of the Executive Leadership in Academic Technology and Engineering, or LATE Program. So with that, I would like to welcome up Professor Davidson to the podium to present her with this medal for her professorship that she can proudly wear at our academic events. The floor is yours, and I would like to invite you to stay for some questions and answers after the talk and also reception over in DuPont Hall afterwards. Thank you, Jamie. Thank you. It's so fun to see all of you here. So I thought I would start today with just a little story. i know some of you have heard this but it was it has to do with a student i had when i was just starting out his name was Hiro he was from Japan he was a master student very bright he worked really hard and was quite earnest about his work and it was getting close to the time he was going to be graduating he came to my office and he said thank you so much and i'd like to get you a present to thank you and I said well Hiro that's really nice of you but it's it's not at all necessary don't worry you know best of luck to everything and he said no no no I really want to I'm gonna make you some business cards and I said well Hiro that's that's very nice of you I actually already have business cards but you know thank you and he said no no yours aren't good I'm going to make you some better ones and at this point I realized I was not going to win this debate so I said fine that would be great you know I gave him the you know my office number and so forth a couple days later he came back and said okay we should take a picture for your business cards and I was like Hiro I don't know if I want my picture on the business card he was like no no no it's much better if you have the picture because then people will remember you and like this is what we do in Japan and this I was like okay final you know we could take a picture I get up and I'm like going to pose for this picture and he says at that time in my office I had a globe that showed where disasters of different types happen and he said how about you hold the globe because then like you know everybody will see you'd study disasters and I was like this is gonna look a little silly but he was very persuasive so I said okay I'll hold the globe so I took the picture I thought we were done at that point he said okay last thing we need is your motto and I was like hero I don't have a motto he's like oh you need a motto like that way people remember you and they'll know it so it was back and forth and finally I said look how about you propose a motto. You suggest something for me and I'm sure it'll be great. And so he did. And this has been my motto for the last 20 years. Freedom from disaster for all people. And it served me really well and I never got to thank Hiro for that. So I'm going to use today as an opportunity. So we're going to talk about disasters. You all know what they are. You read about it in the newspaper and so forth. This is just a good summary to give you an idea of the costs of disasters this is just for climate and weather related events the number of billion dollar events each year in the u.s and you can see there are a lot of them last year's 27 they cost a lot and they're very serious problem and it's growing but this doesn't really tell you why they're so important you need to try to understand what happens to a community when there's an earthquake or a hurricane or a wildfire. So these are all photos I took on various post-earthquake reconnaissance trips over the years and I think start to give you an idea of what it's like when one of these events occurs. And it's one thing to see them on the screen, but when you see it in person and you see a building like this size just crumble, it sticks with you, and it's been very motivational for me. The Haiti earthquake, like that, that's one of my students, and that's basically like their White House in the background. In New Zealand, there was liquefaction that was widespread. That was like a new house in the upper right that just sunk a few feet into the ground when the soil liquefied and lost its bearing strength. There were rock falls and so forth. That's tsunami damage from in Japan but at the end of the day it's really not even about the building damage it's about the people and if you just imagine what it's like for that man who is looking at the remains of his house after Hurricane Helene just last year try to imagine what that would be like that's that's what motivates me and I think most of our colleagues and what we're trying to work on so what I wanted to do today is try to give you a real brief overview of three main areas of work that I've done in this general topic so hurricane evacuations infrastructure services and regional hurricane risk management so just like a real quick kind of flavor of the kinds of questions we ask the kind of methods we use to investigate these things and everything I'm going to talk about is work that was done by a huge group of people, which you'll see. So hurricane evacuation. The main question we've really looked at is where should official evacuation orders be issued, when, and what type, voluntary or mandatory? So as a hurricane's approaching, like this is Irma approaching southern Florida, a lot is happening. There's a lot of uncertainty about what's going to happen. It's a dynamic situation. people are moving the forecasts are changing and there's basically this trade-off should they issue evacuation orders sooner that will give people more time to react but there's also a lot more uncertainty and so you might end up with some unnecessary evacuations which can be expensive and and dangerous too or if you wait later you have less uncertainty but then you have less time to evacuate so the traditional approach and what's actually done in practice now is on the left here there are hurricane evacuation studies where they basically take information about the hazard like where we think flooding will occur the behavior how we think like people are likely to leave or not and there's a transportation analysis that says how long do we think it'll take to get the people from these potentially flooded areas out to safety so So that's the clearance time. So they basically look at the clearance time, they look at the forecast time until hazardous conditions and say, okay, we gotta make a decision before one is less than the other. So what we try to do is reframe that and instead like really explicitly account for the uncertainty and the dynamic and the interactions among all these decisions in what we call the integrated scenario -based evacuation approach. And that's just one of many acronyms you'll see. So we're trying to mimic the way emergency managers actually make these decisions now. So as a hurricane's approaching, they're kind of repeatedly deciding, should I issue evacuation order now or wait a little longer? Should I issue it now or wait a little longer? And they're kind of stepping through time. And as they do that, they're thinking about, okay, what has the hurricane done until this point? What's the forecast? How much time do we have left? How much time is it gonna take to evacuate those people? what is everybody else doing um and so we're trying to develop we tried to develop a tool that kind of helps integrate all that information and it turns out it's very well suited to what's called a multi-stage stochastic program so this is um a computational framework so each of those boxes represents a mathematical model they're all integrated and together they provide at the end the output is this tree of evacuation recommendations, so like a set of contingency plans. So I'll just step through briefly what each of these are for those of you interested. We start with the modeling the hurricane itself in the top five steps, and then we run the evacuation model, which includes an understanding of the behavior of the population and the traffic. So the first step is linking models of the meteorological model that shows how a hurricane is likely to evolve over time. That also produces precipitation forecasts and wind forecasts. The precipitation goes into a hydraulic model. We used crests. And then they both go into a coastal flooding model. So they're all linked together. And then by perturbing the initial conditions, we generate some uncertainty in how the hurricanes are going to evolve and what the hazard will be and so we get like an ensemble of scenarios like that and then we put it all together we get a total water level and a total wind speed for each scenario at each point in time because it's evolving the next step then is to take that ensemble of scenarios and create what we call this scenario tree so we said we've got this set of scenarios that shows like how that hurricane might evolve and we say if it evolves the way those red ones do over the next 12 hours say starts to go more north then we think eventually the hurricane will be like one of the red scenarios so at the end of the day it'll it'll evolve like one of these scenarios we don't know which one yet though if it starts to go more south it'll be like one of the blue scenarios and then in the next time step if it starts to like say it stays south like the green ones it'll end up being like scenario 22 10 8 or 21 so we're sort of gradually getting more information as the uncertainty is resolved so the last piece the evacuation model includes a lot of study that help understand the behavior of the population like who we think is going to leave and when. Using statistics and machine learning, we've done some work on that. And then a traffic model says if you know the origins and destinations, what's the route that different people are going to take? So if we put all that together, we can get kind of where people are going to be at each time and what's the hazard going to be at that time and overlay them. And so any one location is not inherently safe or dangerous. It's just dangerous if you're there when the flooding happens or when the wind speeds are strong so it all goes into this multi-stage stochastic program with the objective of trying to minimize a balance of the risk and travel time and it makes the assumption that we know what's happened until that point and we know what the possibilities are in the future but it but we don't have a crystal ball about what's going to happen so for an example this is Hurricane Florence so if you imagine that it's September 9th, 11 p.m., the hurricane's at that location. You would run the model and you would get a whole tree of recommendations. So instead of just one plan, it's the set of contingency plans. And what it says is basically like at that point, say September 10th, 12 p.m., you shouldn't issue any evacuation orders yet. Basically, you have enough time to wait. But if it starts to go down the red path, then you want to issue evacuation orders in a few places because you're starting to get to a point where either we're sure those places are going to be dangerous or we're going to run out of time if it went down the blue path though you would not issue any orders if it goes down the blue and then the green then you wouldn't issue any orders again so you're sort of trying to adapt as the hurricane evolves as you can see. So, for example, by the end of the day, if it evolves like scenario nine, this blue one, you would have issued evacuation orders in all those places. But if it had ended up like this red one, you would have not issued any. So we're trying to only do them when we have to. And then we can look at the performance of these. So the risk reduction and the travel time kind of go together. Each of these points represents one of the ways the hurricane might evolve so if if it evolves like scenario nine you're going to end up issuing a lot of orders you're going to reduce the risk a lot but you have a lot of travel time which is sort of the cost but if it ended up like 10 there's no risk to reduce but you also haven't made people travel for no reason if by comparison we just used one plan based on the best forecast we would have had a lot of people evacuate and so depending on how it evolved a lot of that might have been unnecessary So there's a lot of travel time, no matter what. So at the end of the day, this is a very different approach. And basically, some of the benefits of it is we can explicitly trade off risk and time, and we can get this well -hedged, robust solution that's good no matter how the hurricane ends up evolving. So we're leveraging the value of that increasing information you get as the hurricane moves forward. okay so that's hurricane evacuation gives you an idea of like one type of work we've done second one has to do with infrastructure services so this is like utilities basically now ideally we would not have any infrastructure system service interruptions ever right but that's obviously not feasible so we always have this trade-off again this familiar theme you're going see some themes um we're trading off like how long how much would it cost to improve the performance versus what are the consequences if we don't and so because consequences are so important we've my colleagues and I who are here actually have argued that we need to start thinking about the societal functioning not just the system functioning so not just who gets water or power or gas, but what are the daily activities of life that those support? This is going to make it more complicated so I think we should be clear that there's a reason to do this. And the reason really is that minimizing the societal objective, that's our true objective. That's what we actually care about. And it's very much related to the system impact, but they're not always the same. So for example, if you have the water is out for two hours in the middle of the night in a residential neighborhood, it could be that nobody even notices. But if you have the water out for two hours after an earthquake on a windy day, it could be the difference between a small fire and a major conflagration. So they're not always the same. It also gives us more options to reduce the impact. So if we think of this chain of events, kind of, you know, you want to cut it off as early as you can, we can improve the design of the components of the networks or upgrade them. We could develop ways to reroute around damage once we have it. But if we add the societal impact, we can also think about even if there are outages, how can we manage them to minimize the societal impact? So we end up with these questions. How many people do we think will be affected each year? How likely is it? And what are the interventions we can use. So the standard approach is a Monte Carlo simulation. We simulate, for example, for earthquakes, a ground motion. Then we simulate the damage to the components of the network we're interested in. And then we evaluate the system performance. So who's getting services and who's not. And then the thing is, you have to repeat that a whole bunch of times because there's a lot of variability in it. And if we're worried about earthquakes, we need to repeat it a lot because in order to get the big ones that we really care about you have to simulate it for like a hundred thousand years because by definition they don't happen very often so that's where one of the biggest challenges comes in so what we've tried to do is is change make a little bit of a change to the the simulation framework so instead of looping around over and over we're developing a set of hazard scenarios very carefully and we're making it a small number and artificially changing the occurrence probability so that that small set matches the hazard of the full set you would get but it's just a much smaller number so everything else is much more computationally efficient and we basically do the same thing then for the damage scenarios and then model the functioning of the system and now we've incorporated also the way that that utilities restore services and then translate it also into the societal impact so step by step again the hazard scenario so this is um what I mean by by finding a small set um so we start by generating a candidate set of earthquake scenarios so each earthquake scenario is like a fault rupture a certain magnitude we generate a whole lot of them and then we want to select a small number and modify their occurrence probabilities and so we do that in the first step is just by based on their contribution so the ones that aren't very likely or that aren't causing much damage we can get rid of and then we have this optimization which is really the key contribution by by minimizing the difference between the hazard curves we would get with the large number and with a small number, and then using that also to determine those probabilities, we can get a really small number of events, and there's almost no difference in the results. We do the same thing for the ground motion maps. So this is just an example for Los Angeles. On the left, we have the ground motion map you would get if you just directly simulated like 300,000 events, and in the middle there is 351. So we get almost the same maps. And then this was for liquefaction for the same thing, which is 350. So now everything is much, much easier. Because those later models in the series are very computationally intensive. So we did the same thing in New Zealand. And we went from like 700 ,000 events to 124. We're only worried about ground motion here. And this is showing the distribution of losses. It was almost the same for both. So it's an enormous computational savings. It just makes everything simpler without much of a trade-off. So the damage scenarios is kind of the same idea. We want to come up with a set of damage scenarios. So each of these would be like one damage scenario, a map showing where there's damages, and each has an occurrence probability that we're going to determine. So we want to do it in such a way that we'd get the same answer as if we had simulated a lot of them. So it's a similar idea. we just we actually milked this idea a lot for a lot of different applications but just we just minimized the difference between the curve you would get if you simulated lots of damage scenarios and then if you simulate just a few or take a small subset of that so here we ended up with like 568 um for the system functioning part then so for each of those 568 damage scenarios we want be able to determine who's going to get water service in this case so one of my recent students developed this model to simulate the system functioning of the water supply system when it's heavily damaged and then how it's restored over time so it's um what's challenging here is that when you've got a lot of pipes damaged and a lot of water is leaking the hydraulics are not the same as normal so we have to be very careful about that and use what's called a pressure demand analysis instead of the typical one. But then we also tried to tightly couple the hydraulic model and the restoration process so we're capturing the idea that as the operators are restoring services, they're kind of making their decisions based on what's happening at that moment. And finally, the societal impact scenarios. So this part is like all very new. The other stuff was sort of helping around the edges to try to understand how people respond we spent a lot of time looking at what we called user adaptation so when the water goes out or the power goes out what do you do you don't just sit there and cry right what do you do you you adapt in different ways so we some of our colleagues at the University of Washington we looked at Twitter data to see if we could find evidence of people making these adaptations so here are a couple examples after Hurricane Maria people charging their phones either in the car or at a wendy's for example we also looked at a lot of tiktok videos to see if we could find evidence of people making these kinds of adaptations and this was one example after the texas winter storms of 2021 and so they they had power in this case but no water they had snow because it was a winter storm so she's melting the snow so that you can you can meet the needs that are the really important needs you can't maybe do everything but you can do the ones that you you really need to survive like make the coffee so by looking at these sort of systematically we found like basically these aren't isolated clever workarounds that people do it's actually a very common phenomenon that we saw over and over again and people do the same kind of things in different types of events in different locations and implicitly we're pretty much relying on this you know I mean after the Christchurch earthquake there were they were out of power I'm sorry out of water for like a week and people can only survive without water for a few days so there was something happening there right we're somehow relying on these activities so we created a typology to start to get a handle on you know what are all these adaptations so we divide them into the ones that work by modifying the supply available some work by modifying the demand available so the infrastructure system focus so some like for a generator will will supply electricity for anything you need to do some are more use focused you could get a lantern or a candle that'll maybe serve your lighting needs but you can't charge your computer with that so we kind of tried to understand what the range of them are we also classified them based on the service they provide so some are just a partial substitute some are a complete substitute of what what kind of cost is involved in implementing them so we could kind of start to get a handle on you know they're not just millions of different adaptations there's actually some rhyme or reason to what people doing um and then we we collected survey data stated preference and revealed preference survey data and fitted some statistical models um mixed logic models to to try to estimate the probability that that a person would do a certain adaptation as a function of how long the outage was and characteristics of those people and so we're starting to get an idea quantitatively of like how likely it is people will do different kinds of adaptations and and then we also looked at their self -reported unhappiness during these so we could then put all those pieces together again the earthquake occurrence the damage the system functioning and the societal impact and start to see like if an earthquake happened in LA how many people do we think would be going to hotels how many people do we think would be trying to do different things and to capture the uncertainty so these are still pretty early preliminary they're not perfect but we can kind of start to see how we could think about describing the societal impact okay are you still with me last example and you're probably starting to see some similarities in these things. So the way we manage hurricane risk or disaster risk in general in the U.S. is it's a whole system right just like the educational system or the criminal justice system. The current system has limitations I think from all stakeholders point of view. If we just think of Hurricane Ian which was a couple years ago this woman Maribel Gutierrez lost her home the house was destroyed the family had nowhere to go but from the insurer point of view the united property and casualty went out of business because they had higher than expected losses and governments also suffered there's the fort myers had 29 billion dollars they had to spend on recovery that you know they had to take out of their other normal expenses they were planning on and this is just one example it happens over and over again in general we see households tend to not strengthen their homes they tend to not buy enough insurance and so they have they don't have adequate resources to recover insurers don't like these events at all they find them very difficult to manage hard to make a profit and they worry about going bankrupt we're seeing this to a very large extent now especially in Florida, California where they're pulling out of the market um and governments also they end up with these large unplanned expenditures that just wreak havoc so the thing is we actually know a lot about disasters we've we've been studying this like the drc's celebrated 60th anniversary um but somehow we're not seeing a lot of that knowledge in the real world and I would argue is partly because in order for that knowledge to be reflected in the real world just one reason people have to take some action to do that and the truth is they just often don't so the CHEER hub that Jamie mentioned we're trying to focus on how can we support policy making or programs that specifically are more likely to actually be implemented? Because A, they align with the way people naturally make decisions. So rather than saying everybody should do this and then thinking like, oh, why are people so silly that they're not strengthening their homes? Let's try to understand how they're making decisions and kind of make a strategy that aligns with that. We want win -win solutions. So we have, ideally, solutions where all the different stakeholders can end up better off, so there's more likely to be accepted and implemented. And we address some of these competing challenges. So, for example, economic prosperity, we don't want, you know, we could say, I know how to reduce hurricane losses. Don't build anywhere near the coast. So I've solved the problem. But, I mean, obviously, that's not really a tenable solution, because people want to build near the coast. There's economic reasons and so forth. So let's find solutions that kind of bake that in from the beginning and realize that that's a competing objective. So this is my last computational framework where each box is a model again, and this has evolved over time, but the green ones are kind of basically estimate losses due to hurricanes to households. this is we've all focused on housing so far but what we've done and that's that's fairly standard in the field now to have that kind of model from the green boxes what we've done though is added these boxes at the top that represent the way different stakeholders make decisions and we've integrated those together so we have a model that that tries to describe how we think the government makes decisions and we have a model that tries to describe how we think insurers make decisions and one that shows how how households of different types make decisions and they interact with each other in this game theoretic model so rather than saying households you should do this we're just trying to find the solution assume they'll behave the way they want to behave and let's find a solution that works under those search circumstances so what it provides us results is a recommended government policies and a description of how we think the insurers and households will respond. And then importantly, it also shows what the outcomes are for each stakeholder. So we can kind of see, are there going to be winners and losers? So just very quickly to give you an idea of kind of what's in the black box. The loss models are simulations. We come up with a set of 30 years of hurricane scenarios. We have a component -based model of damage to houses and with and without different kinds of retrofits the economy is modeled as a computable general equilibrium program to try to show what the effects of capital stock disruption and labor supply disruptions are and that's broken down by sector of the economy and type of household for the government now again we're gonna model the government the way they make decisions which is different from how other groups do so the government we've modeled with a stochastic program because again the key issue here is there's a lot of uncertainty about whether hurricanes are going to occur so they have they have certain objectives like maybe they want to minimize the total loss to the whole community and they have different alternatives available to them they could issue grants to encourage homeowners to strengthen their homes they could do property acquisition offers or regulate the insurers by contrast the insurers have a different kind of decision they they're also do using a stochastic program because there's still uncertainty but their objective is to minimize or to maximize their expected profit so they don't care about uninsured losses they just they want to maximize their profit and the things that are in their control or they can decide how to price their insurance and how much they're gonna transfer through reinsurance and they're trying to avoid insolvency oh and then there's insurers interact in a Corneau Nash game which is a game theory to to capture the fact that it's not a monopoly and then the households make very different decisions so they they can decide if they're offered a property a buyout are they going to accept it or not if they are they going to buy insurance or not are they going to strengthen their home and maybe if they're offered a grant would they strengthen it so their decisions are based like the probability they'll take those decisions based on the terms of the deal so when the insurer for example changes the premium it changes the chance that the household will it will buy the insurance so that's sort of where the interactions come in um so as i said we end up with kind of the decisions for each type of stakeholder and the outcomes so just real quick a case study for the eastern half of North Carolina if we look at the government decisions this could show us for example as the government invests more money a billion dollars two billion three billion how much should they spend on the acquisition program versus how much should they spend on retrofit grants and then what will the expected losses be or the reduction and expected losses be under those different policies to look at the households we can say so if we start out imagine that there's no acquisition there's no retrofits there's no government investment some houses have for some houses insurance is affordable for some it's not now if we introduce say a three billion dollar government investment some of the households will be acquired like the ones in red some will be retrofitted the The one's in green, and then some won't have either, and the affordability changes because we've changed some of the losses. Now this is assuming that we're forcing the insurer to keep the price the same, but actually they're gonna wanna raise the price if the government makes this investment. So if they do, it changes what the households do. So there's all this interaction, and this is basically what the value of this whole modeling thing is. there are so many interactions among it within the system among the different players and we end up sometimes with weird unintended consequences and we really want to understand how the whole system works in order to come up with the best policies so so the last just example of output the on the left side is the distribution of what the homeowners might pay the total of what they pay for insurance and any residual loss and so forth so when there's no investment the blue line the they're going to be paying a little more when the government makes an investment it goes down and if the government makes the investment and they also forced the insurers to not change their price if it was fixed then it would go down even more so we can kind of see okay how is this playing out for the government how's it playing out for the insurers and then on the right side is an example for for the insurers with those three scenarios the probability of insolvency changes so we're trying to find solutions where like everybody is okay so maybe there won't be as much opposition and this is work basically that the CHEER hub is is extending and improving this work in a lot of different ways so that's what I hope gives you a flavor of the type of work that I've been doing as I said this is work over many years, and it's all very interdisciplinary, very collaborative. I'm really afraid I forgot some names, but this was as many as I could remember for the different types of work, and I know some of you guys are here, and so that's actually one of the things I like best about doing this kind of work is that it's so collaborative, and there's so many people involved. I also want to make sure before I forget that almost all of this has been funded by the National Science Foundation. So I thank them. So it's sort of become a tradition, I guess, in these talks. People give a little bit of an idea of kind of how they got to this point and say some thank you. So I thought I could tell you literally how I got to this point. I kind of moved around to different places over the years. But instead, I thought I'd just tell you kind of... The truth is, I did not plan this at all when I was a little girl. I didn't dream of studying disasters or any of this. It was more of a, like, make one decision at a time type thing, and here I ended up. But looking back, it makes a lot of sense to me. When I was a kid, I had a subscription to Games Magazine. I don't know if any of my friends here did. It was basically a puzzle magazine. It was very nerdy, and I loved solving puzzles and stuff like that. I also really wanted to do something where I thought I could, you know, make the world a better place. I could help people. And I loved being a student. And so this was kind of the closest I could find to keep being a student for the rest of my life. And I feel like this career has helped me kind of satisfy all of these parts of me. And so I'm very grateful for it. I also am grateful for all the mentors I've had. people have hosted me on sabbaticals especially my my undergraduate and graduate mentors but i've i could call a lot of people mentors but i thought i should at least think of these people who have played critical roles at critical points in my life and i thank them probably the most important is to thank students and postdocs i see some of them here because I kind of strongly encourage them to come. But they're the best, like, Hiro. I love working with them. They're all so different, and they, as most of you know, actually do the bulk of the work that I presented. I also have some really great communities, professional communities that I've worked with. And at the moment, like, I work a lot with the SimCenter, which is a group of universities that does simulation computational work and disasters um the cheer hub you mentioned um I mean we mentioned earlier is like a fantastic group of people um and I thought I would also call out my Wednesday morning crew so these guys we've been meeting at nine o'clock on wednesday morning for years like a long time including this morning And I think that's how we've, like, really been able to integrate the different disciplines. It's two economists, a sociologist, and two engineers. So really importantly, the Disaster Research Center here at UD is amazing. I am so lucky to be part of it. They're world leaders in disaster studies, and we celebrated our 60th anniversary. I see the two co-directors back there. so that the DRC is is amazing and in the College of Engineering as Jamie said I've had the opportunity to work in the dean's office for a long time and I that's I've learned so much from that I need to I'm sorry it's not a great picture Wendy but I got a call out Wendy who's here who like actually organized all this because it's been such a joy to work with her over the last five years my department who kind of helped raise the kids I don't know if that's an old throwback picture but and and Jamie mentioned the DEI work I know I'm missing people here also but I know it's it's it's under attack now but I feel very proud of what we've done and I'm really thankful to all the people who've worked on that what you're hearing in the news bears very little resemblance to what we've actually done in the college which has mostly been just about trying to remove barriers so more people could participate in engineering and and trying to create a climate where everyone can contribute at their fullest and that's that's really all it is um so i'm thankful to all those people who i i really love working with um and then outside of work i'm very blessed to have some really wonderful friends who don't care at all about engineering or academics and so that helps me kind of turn that off and i they sustain me and i i love them and last but not least of course my family my parents and brother are here so thank you for supporting me my whole life including today um for teaching me to be true to myself try to make the world a better place to have fun while i'm doing it and Rob and Tess and Eli i love you guys so much you make me so happy and you make me laugh every day and Thank you and Tess just surprised me and showed up from Virginia this morning so thank you for coming And in case you were Wondering if I was joking about that story at the beginning I was not joking So thank you very much And I guess we have a reception We can go to right Do I have to answer questions first I'm willing to take any easy questions You have or we can chat over food across the yes michael so my question was about the impact of social media recently on on disaster planning and the response and how it might get in the way of things because information's coming out and you have plans you're so yeah um so i mean it's definitely changed the landscape in a lot of ways that we're trying to account for there's been a lot of work using trying to see how social media can be used to better understand the sense making of how like in the aftermath of an event how do people learn what they learn how do they gather information you know we were trying to use it in this way which was a little bit harder um in evacuations obviously it changes things gps changes things entirely the way people decide whether to leave now you've got a really precise traffic forecast maybe you're more likely to go off the beaten path like it used to be that that we would assume people would take the main roads that they normally take when they evacuate now you don't have to you just follow the purple line on ways right so it certainly had a big effect and i think um i think we have you know we're trying to take that into account and it's always changing it's also you know there's always hope when there's new data that like oh this is going to answer all our questions and you know it typically doesn't there's a lot of challenges with it like the smartphone location data data we were using to try to understand hurricane evacuation behavior you know we thought that it would be like a panacea and it turned out there was a lot of challenges with it so um you know but still helpful so uh yeah so that's i mean that's in general though it's you know it's a changing environment so every time you think you've got an answer to something the context changes yeah daniel disasters can be a little oh sorry disasters can be a little pessimistic so i wanted to ask over the course of your career are there any moments that you've had like whether it's a technological advancement or really just like a like commutative thing where you've kind of felt a real burst of optimism or some kind of push where you're like I can manage this and we are moving in the right place that's an interesting question I hadn't thought of what Jamie said that we named this cheer I think I mean that the it's a balance like I think we've certainly made huge advances in a lot of ways like think about like what we know now that compared to what we knew a couple decades ago or 50 years ago. But at the same time, the challenges are growing quickly. So I don't know if we're keeping up. I mean, if you looked at that very first plot I showed, the losses are going up. So that can be disheartening sometimes. But there are, you know, if we look at specific examples, there are a lot of really great success stories. And just a couple that come to my head, hurricane forecasting has been a game changer. I mean, just as recently as 1900 when the galveston hurricane like people had no idea a hurricane was coming until it was literally on them so that's a very different situation than having three days of warning to be able to do something another example is building codes which have made an enormous difference in the the damage and losses that we see they only apply typically to new construction So we still have a big challenge with existing buildings, but there's lots of data to show that building codes have made an enormous impact. And I remember in Christchurch also, they had done a lot of retrofits to the electric power system, and it was very clear that that translated into a very much improved performance of the power system. So there's lots of wins like that. so we gotta you know focus on those jenny how willing do you think state and county emergency managers are to do things differently um that's a good question i think i think a lot of them are willing and they're interested in in getting help um and or working together like we also have a lot to learn from them you know sometimes we sort of conceive of what we think the challenge is and we find out there's other challenges um with the hurricane evacuation stuff we had hit a roadblock and actually getting it in implemented um i haven't given up entirely yet but it's been a lot harder than i had hoped the county and local emergency managers were very interested in it but we kind of hit some roadblocks in some of the politics higher up so yeah I don't know if that answers but you know I mean they don't want like every academic is coming up with an oh I've got the answer for you you know it takes a while to like earn some trust and develop a relationship but everybody wants the same things I think. Hey Rachel so I'm not sure if you know but I was we were in Christchurch two weeks before the earthquake hit and so you talk about two extremes the hurricanes where you can actually kind of maybe predict where it's going to hit and the earthquake which arrives unannounced. What about intermediate type disasters like wildfires in which you know I was thinking pacific palisades is a great example of unpredictability there yeah um so i've been really interested in kind of cutting across this different disaster types so obviously like if you're a meteorologist you just do climate related things it's very different than an earthquake which would be a geological phenomenon but for someone interested in the risk like me it's we can kind of go back and forth and it's always very interesting i think to take lessons from one hazard and try to adapt them or translate them to it to other ones so wildfire obviously has become much more of a front and center concern um and i think there are some things that are similar like you know we evacuate from wildfires so it's similar to hurricanes but it's also different um the it's also you know the protecting buildings there are certain things we can do But there's different challenges there also. So, yeah, I think, I don't think we can do the same kind of scenario-based approach for wildfires yet that we talked about here. They're very difficult to predict where they're going exactly. One of the main ways they spread, so I actually did a lot of work on post-earthquake fires at one point. and one of the main ways they spread is these these embers can fly like a kilometer and so you just get this spotty spreading that's very hard to predict and it makes a huge difference whether it's windy or not so you know some of it's also just getting bad luck but I suspect there will be a lot more research in that area in the coming years and so we'll probably hopefully get better at managing the wildfires yeah yeah thank you okay
Rachel Davidson Inaugural Lecture April 9, 2025
From Andrew Brett April 28, 2025
7 plays
7
0 comments
0
You unliked the media.
Donald C. Phillips Professor of Civil and Environmental Engineering Rachel Davidson will explore an interdisciplinary, systems-based approach to understanding and managing the risks posed by hurricanes, earthquakes, and fires. Through real-world examples, she will highlight innovative strategies for improving hurricane evacuations, minimizing utility disruptions, and informing regional disaster policies—offering both challenges and reasons for optimism.
- Tags
- Department Name
- College of Engineering
- Department Division
- Date Established
- April 09, 2025
- Appears In
Link to Media Page
Loading