Latest / The Alerting Authority / Wildfire Evacuations: Why Your Emergency Plan Might FAIL

Transcript

Wildfire Evacuations: Why Your Emergency Plan Might FAIL · The Alerting Authority

• 0:06
  SPEAKER_01: Hi, I'm Jeanette Sutton.
• 0:08
  SPEAKER_02: And I'm Eddie Bertola.
• 0:09
  SPEAKER_01: And we welcome you to another episode of the Alerting Authority.
• 0:13
  SPEAKER_01: And as always, we encourage you to subscribe, follow, listen, and most importantly, participate in these podcasts.
• 0:20
  SPEAKER_01: We want your questions, your concerns, your ideas, your problems, and your pain points, and your success stories so we as alerting authorities can better do our jobs and make our communities safer.
• 0:31
  SPEAKER_02: That's right.
• 0:31
  SPEAKER_02: And Jeanette doesn't just say that every episode just to say it.
• 0:35
  SPEAKER_02: We really do hope to continue to get your feedback.
• 0:38
  SPEAKER_02: In fact, we have gotten feedback from different listeners in the last little bit, and it's fun to see.
• 0:44
  SPEAKER_02: And it really helps with topics like the one you're going to see today.
• 0:49
  SPEAKER_02: And I want to thank sponsors.
• 0:51
  SPEAKER_02: This one is a self-sponsored episode from the Alerting Authority, which is, again, the podcast.
• 0:59
  SPEAKER_02: And we want to sponsor it because the quote unquote, we'll figure it out when something happens is not an emergency communication strategy.
• 1:06
  SPEAKER_02: We help organizations strengthen their alerting programs through practical training, governance, and policy reviews, and real-world guidance that prepares people to make good decisions under pressure.
• 1:18
  SPEAKER_02: When every second counts and it does, confidence should come from competence, not crossed fingers.
• 1:26
  SPEAKER_02: So if you have questions about your policy, procedure, actions, or just want to run through different scenarios, please reach out to us because not only can we get you information from us, but also just a wealth of knowledge from people like the one you're going to hear from today.
• 1:43
  SPEAKER_02: So I'll pass it back to Jeanette.
• 1:46
  SPEAKER_01: Okay.
• 1:46
  SPEAKER_01: Um, so I'm super excited about this because um it hasn't been until just probably the last couple of years that I've gotten to work directly with our guest, which has been a lot of fun to exchange ideas.
• 1:59
  SPEAKER_01: Our guest today is Tom Kova.
• 2:01
  SPEAKER_01: He is a professor in the School of the Environment, Society, and Sustainability at the University of Utah.
• 2:09
  SPEAKER_01: He has a Bachelor's of Science in Computer Science from the University of Oregon and a PhD in geography from the University of California, Santa Barbara.
• 2:18
  SPEAKER_01: His primary research interests are wildfire evacuation analysis, emergency management, and geographic information science.
• 2:28
  SPEAKER_01: And it's that wildfire evacuation analysis that we are going to talk about today.
• 2:33
  SPEAKER_01: I'm this is going to be a fun conversation.
• 2:36
  SPEAKER_01: Yeah, clapping in the background.
• 2:38
  SPEAKER_01: So I also have to share a couple of personal reflections about Tom, and you didn't know I was going to do this, but here it goes.
• 2:50
  SPEAKER_01: We met many years ago.
• 2:51
  SPEAKER_01: I'm pretty sure that I was a grad student at the Hazard Center and you were an enabling fellow with the National Science Foundation.
• 2:58
  SPEAKER_01: So this was like early 2000s.
• 3:01
  SPEAKER_01: I started grad school in 2000.
• 3:03
  SPEAKER_01: So I looked you up and it said you were an enabling fellow with NSF from 2003 to 2005.
• 3:10
  SPEAKER_01: And for those who are listening, enabling fellows are trained by some of the top scholars in the field of disaster research.
• 3:18
  SPEAKER_01: And so Tom was getting this introduction to the whole world and then of disaster scholarship, and then went on to become a very successful professor, a very successful principal investigator, doing lots of studies on evacuations and wildfire.
• 3:35
  SPEAKER_01: And then I got to work directly with Tom around 2021, when I was doing the work with FEMA iPods, and we started to look at wildfire messages that were being issued as WIAs.
• 3:49
  SPEAKER_01: So he and I got to be co-authors on this awesome paper that Erica Kuligowski led, and it's called Ember Alerts, where we analyzed all of the contents of the wireless emergency alerts over two years and found that they were lacking and a lot of content.
• 4:07
  SPEAKER_01: And they had that really interesting wildfire evacuation trigger language that we got to talk about a lot with the U.S.
• 4:16
  SPEAKER_01: fire administrator on the evacuation roundtable that we were then participating in over the last few years.
• 4:22
  SPEAKER_01: So you are probably one of the most frequently interviewed social scientists about evacuation modeling, especially in the last few years.
• 4:33
  SPEAKER_01: But what I love is that you, Tom, you make it the complex really easy to understand.
• 4:40
  SPEAKER_01: And so we welcome you to the Alerting Assurance Authority show today.
• 4:45
  SPEAKER_01: Thank you for being here.
• 4:47
  SPEAKER_00: Well, uh, thank you for that introduction.
• 4:48
  SPEAKER_00: And um, I'm happy to be here too.
• 4:51
  SPEAKER_02: And as always, any stories that you have of Jeanette, we are more than happy to hear.
• 4:57
  SPEAKER_00: Well, I'll tell you one, a short one is I knew I met Jeanette when I met Lori in 2003, I think, in my first Boulder Hazards workshop.
• 5:07
  SPEAKER_00: And I just knew those two were going places.
• 5:10
  SPEAKER_00: So sometimes I'm right.
• 5:14
  SPEAKER_01: Oh, thanks.
• 5:15
  SPEAKER_01: And that's Lori Peak, and we will have her on sometime in the fall when that's coming up.
• 5:21
  SPEAKER_01: Yeah, I should talk about children and alerts and warnings.
• 5:25
  SPEAKER_01: And um yeah, it's fun to see where we've all ended up after 20 years of research and scholarship.
• 5:32
  SPEAKER_01: Um, so yeah, Eddie, back to you.
• 5:36
  SPEAKER_02: Yeah, no, I'm I'm excited.
• 5:38
  SPEAKER_02: So again, for my history, um, I have not ever met Tom, although I am very interested in his work and and his um areas of study.
• 5:49
  SPEAKER_02: Because for those that know me, I was with the California Highway Patrol and and now I'm with uh P.
• 5:56
  SPEAKER_02: Ridge, Arkansas for law enforcement.
• 5:58
  SPEAKER_02: I cannot tell you how many evacuations um we have had to be involved in and how many times for wildfire, since that's what we're gonna focus on now, that we have dealt with.
• 6:12
  SPEAKER_02: And I am ready to learn.
• 6:14
  SPEAKER_02: And I think the the question for me is um help us out for again, practitioner level.
• 6:20
  SPEAKER_02: Um, we know there are a lot of people that are academics that listen that can probably get everything from the start.
• 6:26
  SPEAKER_02: But for those of us that are new to evacuation modeling, what is it?
• 6:31
  SPEAKER_02: Um, and how does it handle traffic flow, exit timelines?
• 6:35
  SPEAKER_02: How is it um how does it impact us in a practical way?
• 6:41
  SPEAKER_00: Sure.
• 6:42
  SPEAKER_00: Um, well, a quick history lesson, if I can keep it quick.
• 6:46
  SPEAKER_00: Um there's really before and after Three Mile Island, which is um 1979 in Pennsylvania.
• 6:53
  SPEAKER_00: If any of your uh listeners don't know, it was a uh a partial meltdown of a nuclear reactor.
• 7:01
  SPEAKER_00: And the reason why I say that is that before that, uh, there wasn't really much in the way of evacuation modeling.
• 7:08
  SPEAKER_00: It was, you know, maybe some we call it manual capacity analysis, where you just say, okay, there's this many people and the road can handle, or this many vehicles, and the road can handle this many vehicles per hour.
• 7:19
  SPEAKER_00: So maybe it'll take an hour, right?
• 7:21
  SPEAKER_00: Just sort of like an Excel spreadsheet these days kind of calculation, which that's where that's always where I start.
• 7:28
  SPEAKER_00: If I'm if somebody's asking me about a community, I'll just look at how many households there are, how many exits, how many make assumptions about the exit capacity.
• 7:38
  SPEAKER_00: But that's what was going on in the 1970s.
• 7:40
  SPEAKER_00: Uh and because of Three Mile Island, uh not unlike uh the whole social science thing with Quarantelli and looking at uh earthquakes as or looking, sorry, yeah, looking at earthquakes as sort of like bomb simulators in the 50s, right?
• 7:55
  SPEAKER_00: It came from the government.
• 7:57
  SPEAKER_00: So when the U.S.
• 7:59
  SPEAKER_00: uh nuclear regulatory commission uh realized, wow, we had some real evacuation issues with Three Mile Island in terms of not having a standardized size zone.
• 8:12
  SPEAKER_00: You know, I think they started with five miles and just it was not, it didn't go very well.
• 8:18
  SPEAKER_00: And it was also when they discovered the evacuation shadow, that's where that concept comes from, is all the people that evacuated who were never told to, and um they didn't they had issues with whether or not they could issue an evacuation.
• 8:31
  SPEAKER_00: So after that, they said every nuclear power plant in the United States needs to have a study done for evacuation time estimation.
• 8:39
  SPEAKER_00: And they just instantly launched a cottage industry because they didn't give anybody the software and they didn't give anybody the method, and so it was just consulting galore.
• 8:51
  SPEAKER_00: Um some people that were there, familiar names.
• 8:54
  SPEAKER_00: Mike Lindell was there at that time.
• 8:57
  SPEAKER_00: Uh he was my mentor for the enabling program, which uh brought up.
• 9:01
  SPEAKER_00: Thomas Urbanac is a big name from that era, and also Ronald Perry, uh Mike's collaborator from ASU.
• 9:11
  SPEAKER_00: Um there were others too.
• 9:12
  SPEAKER_00: Oh yeah, actually, Susan Cutter wrote a paper about evacuation behavior in Three Mile Island.
• 9:18
  SPEAKER_00: Um so out of that came the first sort of broad transportation-based simulator called NetVac.
• 9:27
  SPEAKER_00: And uh that was I think Joseph Sheffi, I don't know, um, that developed that.
• 9:34
  SPEAKER_00: Uh so uh that system was really set the standard for this idea of simulating traffic.
• 9:42
  SPEAKER_00: You set up a 10-mile radius, was what came out of that, 10-mile radius around a power plant.
• 9:49
  SPEAKER_00: You run these traffic simulations, and you kind of try to come up with what they called an ETE, an estimated time of evacuation.
• 9:57
  SPEAKER_00: Um, so that that kind of launched the whole field.
• 10:01
  SPEAKER_00: In the 1980s, um, people like Anton Hobaika created a program called MassVAC.
• 10:09
  SPEAKER_00: And MassVac, well, he really branched out into hurricanes, tsunamis, floods, um, hazardous materials.
• 10:18
  SPEAKER_00: And so it became much more than just this nuclear power plant thing.
• 10:23
  SPEAKER_00: Um so by the 19, yeah, so that that kind of covers the 80s.
• 10:29
  SPEAKER_00: Uh in the 90s, uh, Oak Ridge National Lab became a big player.
• 10:33
  SPEAKER_00: They created the Oak Ridge Evacuation Modeling System, ORAMS.
• 10:38
  SPEAKER_00: And two two uh big names.
• 10:41
  SPEAKER_00: Uh Frank Southworth uh was a real hero of mine when I was a grad student, and just learning my way around, you know, because he wrote a very nice guide on how-to, kind of a how-to evacuation modeling guide.
• 10:54
  SPEAKER_00: And then Oscar Franzisch.
• 10:57
  SPEAKER_00: I'm not sure I'm pronouncing his name right, but um he also was publishing and he was doing kind of a roadshow with OREMS.
• 11:05
  SPEAKER_00: A couple things about that era is that they were really into mass evacuation.
• 11:10
  SPEAKER_00: That term came up a lot.
• 11:11
  SPEAKER_00: It was just how big of an evacuation can we simulate?
• 11:16
  SPEAKER_00: And the systems were macro, macro simulation.
• 11:20
  SPEAKER_00: So that in transportation, that means you're modeling flows, you're not modeling and modeling individual vehicles or individual agents, you're really modeling flows.
• 11:29
  SPEAKER_00: Um, and so yeah, the mass the mass evacuation theme was a big deal.
• 11:35
  SPEAKER_00: Also, there wasn't really any publicly available software.
• 11:39
  SPEAKER_00: Like if someone like me wanted to do it, you know, you just there was no way to call someone up and say, send me your system.
• 11:46
  SPEAKER_00: You know, it was each per each group was rolling their own, so to speak, in terms of software, and nobody knew what anybody's interface looked like, and it was really reports and publications were your only window into what others were doing.
• 12:02
  SPEAKER_00: Um, so by around 2000, I would say, oh, yeah, yeah, that's right.
• 12:08
  SPEAKER_00: Around 2000, I would say that you know, this whole micro simulation, that's what what it was called in transportation, but later it kind of changed to agent-based modeling.
• 12:16
  SPEAKER_00: So now you can model individual vehicles, and then you get into things like car following models and lane changing and intersection behavior and decision making and so on.
• 12:27
  SPEAKER_00: Um, and that's about when I came along.
• 12:30
  SPEAKER_00: I uh wanted to use something like O-RIMS, but I just didn't have the resources or the connections.
• 12:38
  SPEAKER_00: It was a networked kind of thing.
• 12:40
  SPEAKER_00: And so I switched to using just a regular transportation simulation program like engineers would use to simulate daily traffic or any other journey to work kind of situation.
• 12:52
  SPEAKER_00: And there were many of those at the time.
• 12:54
  SPEAKER_00: I was I chose one called Paramex, and I put an evacuation front end on it.
• 12:59
  SPEAKER_00: So uh it was the behavioral models would have been just everyday traffic.
• 13:04
  SPEAKER_00: Uh so that was one drawback.
• 13:06
  SPEAKER_00: It wasn't like there was any sort of uh specialized evacuation car following models or lane changing and so on.
• 13:14
  SPEAKER_00: Um, but it got the job done.
• 13:16
  SPEAKER_00: Uh and that that kind of caught on after that.
• 13:19
  SPEAKER_00: So I did that like in around 2001, 2002.
• 13:24
  SPEAKER_00: And those those early models, I guess, there were a lot of assumptions.
• 13:28
  SPEAKER_00: It was all assumptions, actually.
• 13:30
  SPEAKER_00: You know, um, you can't really predict behavior with these systems, but you can say if people behave this way, what will this the evacuation look like?
• 13:42
  SPEAKER_00: And so they had uh in in the 80s, they had what I call the Le Mans start, which is if you wanted to visualize it, it would be like imagine all the evacuees are on their lawn, they have everything packed, they're just standing there waiting to go, and they would say, go.
• 13:59
  SPEAKER_00: And everyone would jump in their car and get out on the roads.
• 14:04
  SPEAKER_00: And that was pretty normal.
• 14:06
  SPEAKER_00: So no behavioral component really at all, right?
• 14:09
  SPEAKER_00: The evacuees are robots that jump in their car and the cars instantly spill out onto the roads everywhere.
• 14:15
  SPEAKER_02: Um, and so everybody follows all the traffic laws, and nobody is irased or upset or anything.
• 14:22
  SPEAKER_02: Of course, we all know that you know it doesn't happen.
• 14:25
  SPEAKER_02: Okay, no, this is good.
• 14:26
  SPEAKER_02: Keep going.
• 14:27
  SPEAKER_00: No, that that's exactly right.
• 14:29
  SPEAKER_00: Yeah, assumptions about what where each household, what exit they will take, everyone will go to their closest exit.
• 14:35
  SPEAKER_00: Everyone is gonna drive calmly and everything and not aggressively.
• 14:40
  SPEAKER_00: Uh so that was that was standard.
• 14:42
  SPEAKER_00: And I know sometimes people look back at it and still say that this is so unrealistic, but we were just like back in the days of uh Three Mile Island stuff, we were interested in benchmarking the infrastructure.
• 14:56
  SPEAKER_00: That's really what it was about.
• 14:57
  SPEAKER_00: It was about what can this road handle?
• 15:00
  SPEAKER_00: Not really, can we get a realistic time of what it will actually take?
• 15:06
  SPEAKER_00: Um, I should mention uh the first really significant behavioral model I saw was uh from a group called or a team called Stern and Sinewani Stern out of Israel in in 1989, and they had a paper called Behaviorally Based Evacuation Model.
• 15:22
  SPEAKER_00: And their general finding was the more you integrate realistic behavior, the longer it takes.
• 15:28
  SPEAKER_00: That's pretty much the rule.
• 15:29
  SPEAKER_00: So if you don't send everyone out on the roads immediately, and you now you start thinking about giving them time to decide whether they want to go, and some of them don't comply, and you know, you just keep on rolling in more and more realistic behavioral insights from the social science side, um, you will just get longer and longer evacuation times.
• 15:50
  SPEAKER_00: It's very rare that you would add realistic behavior and it would be quicker.
• 15:56
  SPEAKER_00: It didn't leave much room for quicker.
• 15:58
  SPEAKER_02: Well, and it seems like like you talk about assumptions, and assumptions are are sometimes just what you have to do when we we don't know or we're trying to, as you said, benchmark.
• 16:08
  SPEAKER_02: I think it's a it's often what emergency managers and law enforcement are doing.
• 16:13
  SPEAKER_02: They are using similar assumptions.
• 16:15
  SPEAKER_02: Um because I remember as and when we're learning about evacuation modeling here, there are evacuation notices, warnings, alerts that go out or orders all the time.
• 16:28
  SPEAKER_02: I mean, you guys can look on PBS Warren and others and see them.
• 16:32
  SPEAKER_02: And I'm thinking to myself, now all the ones that that I've helped craft or send, I don't know if I've ever thought through what you're talking about.
• 16:42
  SPEAKER_02: And in my mind, I've made assumptions that, well, you know, I'm I'm sure the roads can handle it, I'm sure all these things, when in reality, it's probably not like that.
• 16:53
  SPEAKER_02: Um, and so that's why I'm I'm uh really excited to get into this um because from what you mentioned from Three Mile Island to where we are today, and you held a book up um before we started the show um talking about you know the paradise fire.
• 17:08
  SPEAKER_02: And it is like this is one of those things that is very applicable.
• 17:14
  SPEAKER_02: And and I think the sad part is that a lot of people, um, like myself included, haven't been exposed to this type of science and and this type of research so that we haven't uh you know applied it or really thought through it and and what we're doing.
• 17:32
  SPEAKER_02: So anyway, I'm super excited for this one.
• 17:34
  SPEAKER_01: I had a question about what you were talking about, Tom.
• 17:37
  SPEAKER_01: Um and then I'm um when you said t exit timelines, are you talking about the initiation of exiting, or are you talking about how long it takes for everyone to leave the community?
• 17:50
  SPEAKER_00: The latter.
• 17:52
  SPEAKER_00: And really they had a specific term for it, which kind of gets rid of the whole behavioral side, but they called it network clearing time.
• 18:00
  SPEAKER_00: And so it was really they weren't thinking about behavior at all.
• 18:04
  SPEAKER_00: In fact, Susan was way ahead of her time with that study in '79.
• 18:08
  SPEAKER_00: But um, in terms of traffic, they were thinking like, what if, you know, what are what if we control the intersections?
• 18:14
  SPEAKER_00: What if we assign people the different routes?
• 18:17
  SPEAKER_00: Um, what if we reverse lanes?
• 18:19
  SPEAKER_00: Contraflow became a big area.
• 18:22
  SPEAKER_00: Brian Wilshon at Louisiana State is kind of Dr.
• 18:25
  SPEAKER_00: Contraflow.
• 18:26
  SPEAKER_00: Um, and so, you know, they were thinking it was civil engineering, transportation engineers, right?
• 18:32
  SPEAKER_00: So, and you know, you guys, I know Jeanette knows these communities well, but they were very separate.
• 18:38
  SPEAKER_00: There was the social science group, you know, Kathleen Tierney and Dennis Milletti and so on, Mike Clendell and that group, and then there was the civil engineering side, you know, the traffic modelers.
• 18:51
  SPEAKER_00: And definitely evacuation modeling comes out of the ladder.
• 18:53
  SPEAKER_00: You know, it's more of a traffic modeling kind of thing.
• 18:56
  SPEAKER_00: And that's been one of the most fun things, I think, for me over my last, you know, a couple decades is watching those two communities get together, some of it inspired by the NSF and their whole interdisciplinary push that they did.
• 19:10
  SPEAKER_00: But uh now we have a huge overlap uh between those two groups in terms of what they would call nowadays convergence, I guess, or synergy.
• 19:19
  SPEAKER_00: Where uh I think I think Mike, back to him again, he was one of the first really to bring all the behavioral stuff that he had done.
• 19:27
  SPEAKER_00: He literally wrote the book, right?
• 19:28
  SPEAKER_00: Behavioral foundations uh in emergency management in the early 90s.
• 19:33
  SPEAKER_00: But he brought all that to the table to the modeling side in the early 2000s and started creating these uh behaviorally based, mostly hurricane, I think, uh inspired simulation software where you know now you can say, okay, we've that well, that was kind of the dawn of the behavioral era, I think, but lately it's just exploded.
• 19:54
  SPEAKER_00: Honestly, it's uh I really so uh makes me feel so good.
• 19:58
  SPEAKER_00: How many civil engineers do social science now?
• 20:02
  SPEAKER_00: They are just, you know, Stephen Wong, Xile Zhao, and then there's also the social sciences that do the simulation stuff, like Erica Kuligowski, as we know, working with the whole Woonity crowd.
• 20:15
  SPEAKER_02: I I'm probably using too many acronyms, but uh well, and I'll say for anybody who's gonna Google this, Mike Lindell, who he's talking about, is not the My Pillow guy.
• 20:24
  SPEAKER_02: He is not referencing that that is not who it is.
• 20:27
  SPEAKER_02: No, he's the mother for like this top.
• 20:31
  SPEAKER_02: You know, just to help.
• 20:33
  SPEAKER_01: Yeah, and we've we've invited Mike to be on the podcast, but he was so busy in uh last fall working on a chapter or a book or something.
• 20:41
  SPEAKER_01: So I'm gonna come back to him and ask him again.
• 20:43
  SPEAKER_01: He is literally a walking encyclopedia of knowledge.
• 20:47
  SPEAKER_00: He Yeah, he he's so much better with the history than me because he was living it.
• 20:51
  SPEAKER_00: He was there.
• 20:52
  SPEAKER_00: I mean, I I read about these guys, but he lit he literally was there, I think four or five years out of his uh PhD at Boulder, right, where you went, Jeanette, um, was when he he rode both the three mile island thing and also the Mount St.
• 21:10
  SPEAKER_00: Helens.
• 21:11
  SPEAKER_00: So he had the natural and the technological uh case studies that really changed everything.
• 21:17
  SPEAKER_00: And he was he was just right there, yeah.
• 21:19
  SPEAKER_01: Yeah, we will get him.
• 21:21
  SPEAKER_01: All right, so let's let's talk about adding behavior into these models.
• 21:26
  SPEAKER_01: And I want to talk specifically about alerts and warnings.
• 21:30
  SPEAKER_01: So how do how are alerts incorporated into these evacuation models now?
• 21:36
  SPEAKER_01: What do we know?
• 21:38
  SPEAKER_00: Well, okay, so the just going back in time again, the main thing they had for quite a while, including me, was the departure curve, we called it, which was like an S shaped curve that would go up to whatever the compliance level was.
• 21:52
  SPEAKER_00: So 90%, 80%, 100% compliance.
• 21:56
  SPEAKER_00: That would be the where the S leveled off.
• 21:58
  SPEAKER_00: Uh and you could You could speed that up or slow that down.
• 22:02
  SPEAKER_00: In other words, p what happens if people leave fast, or what happens if they leave gradually?
• 22:09
  SPEAKER_00: Well, that uh was a kind of uh behavior in that context is kind of exogenous to the model, right?
• 22:16
  SPEAKER_00: It's the modeler decides how everybody's gonna behave.
• 22:20
  SPEAKER_00: It's shifted, it's shifted in the agent-based world to uh to getting just increasingly enriched in that whole area you're talking about, uh where now you can think about modeling uh the milling process is one that people have looked at, social networks and how important that is, right?
• 22:40
  SPEAKER_00: Warning is a social process, is the the mantra.
• 22:43
  SPEAKER_00: So uh, you know, instead of just having someone warned in a vacuum in their home, you know, they're gonna interact with other people.
• 22:50
  SPEAKER_00: Um I think uh there's more and more behavioral-based data finding its way into informing the modeling.
• 22:58
  SPEAKER_00: So uh this is you know, uh survey data, you mentioned earlier, focus group, all that social science data is is increasingly informing more realistic uh behavioral-based modeling on that side.
• 23:16
  SPEAKER_00: Uh so um trying to think.
• 23:20
  SPEAKER_00: Yeah, I mean, nowadays it's also even getting, you know, starting to morph into machine learning and AI and and that kind of thing.
• 23:28
  SPEAKER_00: So uh it's getting harder to say that, yeah, they're not they're not incorporating all the behavioral sides to it.
• 23:35
  SPEAKER_00: Um in terms of alerts, I don't know if I've seen a fully endogenous alert sort of uh, you know, with all of the important warning uh response model, you know, the uh Meleti and Sorensen world, as you know very well.
• 23:55
  SPEAKER_00: I don't know if I've seen that actually incorporated into a model yet where you could say, like, what happens if we leave out the source of the hazard?
• 24:03
  SPEAKER_00: Or what happens if we don't make it very clear, you know, clarity, completeness, all that stuff.
• 24:10
  SPEAKER_00: If you started taking that out or you know, improving it, how would that affect the behavior?
• 24:16
  SPEAKER_00: I don't know if I've seen that explicitly represented in an evacuation model.
• 24:21
  SPEAKER_00: Oh yeah.
• 24:23
  SPEAKER_01: Would it be possible?
• 24:24
  SPEAKER_01: Is it so how how would uh oh Eddie, we're gonna geek out now.
• 24:27
  SPEAKER_01: Okay.
• 24:28
  SPEAKER_02: I know that's I'm I'm processing.
• 24:30
  SPEAKER_02: Keep going.
• 24:30
  SPEAKER_02: And just remember us, remember us non-academic.
• 24:35
  SPEAKER_01: Because this makes me excited because one of the papers that we just got published is called content mapping, where we did exactly what you just talked about of taking out pieces of the message and looking at how an incomplete message that's missing the source, how does that affect the warning response model outcomes versus an incomplete message that's missing the location?
• 24:57
  SPEAKER_01: How does that affect things?
• 24:58
  SPEAKER_01: And so we now know.
• 25:00
  SPEAKER_01: We know what is the most important thing.
• 25:02
  SPEAKER_01: And we know when you're missing pieces of information, what causes more milling behavior and less milling behavior?
• 25:09
  SPEAKER_01: So, okay, so if we've got that answer from these experiments, how would somebody take that and incorporate it into an evacuation model?
• 25:19
  SPEAKER_01: And is it something we could do?
• 25:21
  SPEAKER_00: Yeah, the simplest, simplest approach, I'm always for starting simple, is you know, you would have like, what if we sent a weak warning message, maybe one that didn't meet all of the uh hallmark warning response model features, and then a strong message.
• 25:38
  SPEAKER_00: And hopefully, well, if you had the agents receiving the messages, you know, you would want, you would expect uh higher compliance with the stronger message.
• 25:48
  SPEAKER_00: I'm not sure about earlier departure.
• 25:52
  SPEAKER_00: Is there anything in the social science literature that stronger warning messages lead to earlier departure for households?
• 26:01
  SPEAKER_01: You know, I don't I don't know.
• 26:03
  SPEAKER_01: I don't know if um if Mike Lindell did something in that space because he's done a lot of the evacuation as one of the outcomes that he measures.
• 26:11
  SPEAKER_01: What we do know is one of our studies on the use of evacuation warning versus evacuation order that came out of the what we learned from the Ember Alerts work, we found that uh people have a very difficult time differentiating between those two categories.
• 26:30
  SPEAKER_01: They essentially believe that they understand what those things are telling them to do, um, but they don't recognize that a warning is more of a preparedness, prepare to take action, and an order is take action now.
• 26:47
  SPEAKER_01: They thought our population that did this experiment thought that they both meant leave immediately.
• 26:53
  SPEAKER_01: So there's actually a misunderstanding of the language, which could prompt behaviors that I don't I don't know if that's being taken into consideration when this jargon approach to evacuation is being used and how that could affect people's decisions to leave, if it leads people to look more information up before they make a decision to leave, or if they misunderstand it and they think, oh, I gotta go now.
• 27:23
  SPEAKER_02: And that's all in English, by the way.
• 27:25
  SPEAKER_02: That's that's that's without adding other languages.
• 27:29
  SPEAKER_01: Right.
• 27:30
  SPEAKER_01: Right.
• 27:31
  SPEAKER_00: Yeah, from a from a modeler's point of view, like when you want them to do things, they see the software engineering time.
• 27:37
  SPEAKER_00: So I'm imagining like um saying, well, there's a hundred hours.
• 27:42
  SPEAKER_00: Maybe it's easier under the AI era, you know, to generate the the software for some of this.
• 27:47
  SPEAKER_00: Um you did make me think, like, if I said, are you if I asked a modeler, are you modeling evacuation rescues?
• 27:54
  SPEAKER_00: Because those really take a long time, right?
• 27:56
  SPEAKER_00: To go to people's houses and help the low mobility population.
• 28:00
  SPEAKER_00: Are you thinking about special facilities like hospitals and medical facilities?
• 28:04
  SPEAKER_00: And no.
• 28:07
  SPEAKER_00: The evacuation modelers are not dealing with a lot of these things uh because it's like, well, I guess I would have to have vehicles going into the community, and then they would, you know, there there are great exceptions though, like uh family gathering was one that was a theme for quite a while, picking kids up from schools.
• 28:28
  SPEAKER_00: And and I, you know, Pamela Murray Tweet's name comes to mind.
• 28:32
  SPEAKER_00: She had that theme going on for a while, um, in terms of just how much longer it takes for families to gather, and they're hardly ever together, right?
• 28:41
  SPEAKER_00: Unless it's like a Sunday morning evacuation.
• 28:44
  SPEAKER_00: Um, you are gonna have the parents uh possibly elsewhere and so on.
• 28:50
  SPEAKER_00: So in addition to the social networking, people have addressed the family gathering behavioral side of that.
• 28:57
  SPEAKER_00: But I don't know about rescue evacuations.
• 29:01
  SPEAKER_00: Um there's still lots of things that don't get explicitly modeled endogenously.
• 29:07
  SPEAKER_02: So what you're saying is there's opportunity for future research.
• 29:10
  unknown: Yes.
• 29:29
  SPEAKER_00: Yeah.
• 29:30
  SPEAKER_00: Yeah.
• 29:30
  SPEAKER_00: No, I I think there's lot lots more to do, yeah.
• 29:34
  SPEAKER_02: So I'm gonna put you in, I know we're talking on a podcast, but now I want you to imagine you are, you can say you're at IAEM, you can say you're at another conference, another thing where you have a bunch of emergency managers for um municipalities or pick your area, you have law enforcement, you have fire, other first responders there, and they are coming to you and asking just a simple question of okay, how can I use evacuation modeling to help me get better results in the way that um, well, to get people to to comply, right?
• 30:14
  SPEAKER_02: But how can it, how should it impact the way that I planned these types of alerts, uh, whether it's timing, the geography, or wording, how can this help me?
• 30:27
  SPEAKER_00: Well, I'm I'm glad you brought up timing there.
• 30:29
  SPEAKER_00: Um that's another big trend these days is linking a hazard model with an evacuation model uh to get like the question that they were answering back in the 70s, or sorry, 80s and 90s was how much how long will an evacuation take?
• 30:45
  SPEAKER_00: And leaving out a big other question, which is how much time do we have?
• 30:49
  SPEAKER_00: And so uh along came uh Hurevac was a FEMA model that they came up with that I think might have been the first one to not just say it's gonna take this long to get a particular city uh metro area cleared, but um, let's think about how much time we have based on the hurricane's predicted track so we know when we should initiate the evacuation.
• 31:17
  SPEAKER_00: So anything to do with, and this gets back to the warning message because that's ultimately what it is, is is the when question.
• 31:24
  SPEAKER_00: When do we send out the order to evacuate?
• 31:28
  SPEAKER_00: Uh that is a timing question that can't be answered just by how long will it take us to do it.
• 31:33
  SPEAKER_00: You got to know how much time is available.
• 31:36
  SPEAKER_00: Um, so there has been a connection there.
• 31:39
  SPEAKER_00: But to get back to your question, I was I was in a really exciting meeting this summer.
• 31:43
  SPEAKER_00: It was firefighters, police.
• 31:45
  SPEAKER_00: It was a Zoom meeting, firefighters and police, law enforcement, emergency management from Ashland, Oregon.
• 31:52
  SPEAKER_00: This is another project uh led by uh Boise State and Clemson.
• 31:59
  SPEAKER_00: And uh this the idea here was let's put an evacuation model and a fire model in front of these people and see what they think.
• 32:08
  SPEAKER_00: Let's run some scenarios and find out what they would like to do.
• 32:13
  SPEAKER_00: And it was kind of slow at the beginning, you know, here's the fires coming towards the community, here's our little cars.
• 32:19
  SPEAKER_00: Visualization has come a long way too.
• 32:21
  SPEAKER_00: So when you're dealing with emergency managers nowadays, instead of just giving them a number and a table like two hours, which is what they would do in the 80s, you can actually show them the car is moving across the landscape and the fire spreading down, you know, Google Earth type thing.
• 32:36
  SPEAKER_00: And it was funny, after they sort of got it, if you will, they were like, well, what if you there you you have roads there that we would use that aren't represented?
• 32:45
  SPEAKER_00: Oh, okay, that's a major addition to the model, right?
• 32:48
  SPEAKER_00: There's other exits that weren't shown in the model.
• 32:52
  SPEAKER_00: Uh what if we reverse this lane?
• 32:54
  SPEAKER_00: What if we control that intersection?
• 32:56
  SPEAKER_00: What if we phase things differently?
• 32:58
  SPEAKER_00: Maybe our zone one is too big.
• 33:00
  SPEAKER_00: You know, maybe we should start with a smaller zone.
• 33:02
  SPEAKER_00: So they got into this what if it explodes, basically.
• 33:06
  SPEAKER_00: It becomes, you can see that this is where the software people go, ah, because they're just like 10 hours, 100 hours, two months, six months.
• 33:16
  SPEAKER_00: So you have to kind of say that, you know, these what if scenarios have a cost, you know, time, effort, uh, programming, and so on.
• 33:26
  SPEAKER_00: But it was very nice to see these two communities.
• 33:30
  SPEAKER_00: I've been in the business so long, it's always about the paper, you know, the journal article.
• 33:35
  SPEAKER_00: Once that thing is in print, you go on to the next project.
• 33:38
  SPEAKER_00: In this project, there was an explicit community outreach and community collaboration partnership thing where we're trying to do much more than just publish something.
• 33:51
  SPEAKER_00: It's an NSF project.
• 33:53
  SPEAKER_00: So um, so that's that's what happens, is they want to know what if we try this route, you know, send more people this way?
• 34:00
  SPEAKER_00: What if I think we had a scenario where everybody was supposed to go to the airport uh in Ashland related to a drill, and instead it's like, what if we sent people different places?
• 34:12
  SPEAKER_00: You know, can we speed things up under a uh really high pressure situation?
• 34:17
  SPEAKER_00: And they had a fire in 2020, the Almeida fire.
• 34:23
  SPEAKER_00: It started in Ashland, but it quickly went to Talent and Phoenix, which are small towns just north of Ashland.
• 34:30
  SPEAKER_00: So it may have started in Ashland, but it really wasn't an Ashland evacuation.
• 34:35
  SPEAKER_00: So that's another reason why they made a good partner.
• 34:38
  SPEAKER_00: They have a lot of experience with fire.
• 34:42
  SPEAKER_02: Well, I can see the value of it.
• 34:43
  SPEAKER_02: Like as I picture myself in that room, I would be definitely one of the people that probably are driving people crazy saying, okay, what if we did this and what if we did this?
• 34:53
  SPEAKER_02: Because there are so many variables.
• 34:55
  SPEAKER_02: Um, I think one of the big benefits, though, is the fact that you guys did this, because just like how we always tell people, don't check the box, just check the box that if you're doing a tabletop or another exercise and you say, okay, and we would write a message and then you move on, go through the process, write the message.
• 35:13
  SPEAKER_02: For this, I think if you're doing any type of evacuation as a part of your exercise, it's one of those, okay, we're going to ask people or tell people to evacuate to actually take a look and take a look at the research, take a look at the nuances, um, and and go through and try to see what impact it's actually going to have.
• 35:38
  SPEAKER_02: And if what you're proposing through your planning, with your timing, the way that you're sketching it out with geography, if it actually makes sense.
• 35:47
  SPEAKER_02: So good job with the firefighters and police.
• 35:52
  SPEAKER_00: Thanks.
• 35:52
  SPEAKER_00: I'm just an advisor on this project mostly, but yeah, I'm I'm really happy to be involved in it.
• 35:58
  SPEAKER_01: So when they when they were asking about changing the exit routes and opening the streets and things like that, were they also taking into consideration different human factors, like things that affect delays?
• 36:13
  SPEAKER_01: I think you mentioned earlier everybody being together or having a go bag pact or the networks, um, access and functional needs.
• 36:22
  SPEAKER_01: Like, how did they ask about those things?
• 36:24
  SPEAKER_01: And what do we know about how those actually affect the models that have been developed, other than it just takes longer?
• 36:32
  SPEAKER_01: Like, do we have any specifics from those things?
• 36:36
  SPEAKER_00: Um, it's no, actually, I feel like you're just jumping way ahead with uh access and functional needs, as I mentioned like before, with the special facilities, you know, medical facilities, retirement homes, daycare centers, just schools in general, right?
• 36:52
  SPEAKER_00: That uh don't uh don't have the the vehicles to move to move all the students they have.
• 37:00
  SPEAKER_00: Uh they usually end up phoning in a rescue.
• 37:05
  SPEAKER_00: Um so it's uh it's just uh sort of a level of detail that still uh you know could be researched, I guess.
• 37:14
  SPEAKER_00: Uh so like uh Eddie said, it's a lot it's a big opportunity for a lot of junior scholars to think about all the things you just listed.
• 37:24
  SPEAKER_01: Right.
• 37:25
  SPEAKER_01: But were the practitioners asking questions about how those things might affect their planning?
• 37:31
  SPEAKER_00: Not not so much.
• 37:32
  SPEAKER_00: And I wonder, you know, you make me wonder if uh if the thing about an evacuation model, if it does really focus you on the transportation management side of things, just because it's about vehicle simulation and car following and lane changing and intersections and stuff, that they do tend to drift over to focusing on that, you know, key bottlenecks, all the way back to the um Three Mile Island stuff, they were looking at where might bottlenecks emerge that we can't see.
• 38:03
  SPEAKER_00: I mean, obviously if you have a Bay Bridge or a Brooklyn Bridge, that's a bottleneck.
• 38:08
  SPEAKER_00: But sometimes if you're in a Phoenix or a or a Houston, you don't really know exactly where these bottlenecks may emerge, and they can be pretty bad when you have millions of people.
• 38:18
  SPEAKER_00: So um no, I don't don't remember them mentioning that much.
• 38:24
  SPEAKER_00: Um but just uh I know Eddie brought up that Paradise book I have, but I was just uh there's a page in there with a tribute to the victims, and it's like 75 years old, 76, 82, 69.
• 38:40
  SPEAKER_00: I think I saw one person under 30.
• 38:43
  SPEAKER_00: So elderly people who uh again fall under uh uh under other categories too, right?
• 38:50
  SPEAKER_00: Mobility issues.
• 38:52
  SPEAKER_00: Um and just they just take longer.
• 38:56
  SPEAKER_00: When you have a fast moving hazard like a fire, uh most of the time that would be the most effective thing to do, I think, to reduce fatalities would be thinking about access and functional needs, uh populations and how to take care of them, you know, how to uh how to pick them up if you're there your neighbor.
• 39:16
  SPEAKER_00: Uh-huh.
• 39:17
  SPEAKER_00: Or dealing with the facilities and and making sure they all have the transportation needs.
• 39:23
  SPEAKER_00: It's amazing what they did in paradise.
• 39:25
  SPEAKER_00: You know, I think a lot of people thought what a disaster, and it was a disaster, but the mayor said we moved twenty six thousand people out of the town.
• 39:35
  SPEAKER_00: So wow.
• 39:39
  SPEAKER_00: You know, it's just a different perspective on the same thing.
• 39:43
  SPEAKER_00: But yeah, it could have been so much worse than uh than that.
• 39:47
  SPEAKER_01: Yeah, and it sounds like in some cases you're talking about timing for people who need more time to get out, but in other cases it's about actually going and physically helping them to get out.
• 39:59
  SPEAKER_01: So there's two kinds of elements there related to their ability to evacuate.
• 40:04
  SPEAKER_00: Yes.
• 40:06
  SPEAKER_00: Yeah, that's right.
• 40:07
  SPEAKER_00: And it's you know, they always say like with uh earthqu earthquakes, you know, the first responders, there's not enough, right?
• 40:13
  SPEAKER_00: If you look at like what just happened in Colombia.
• 40:16
  SPEAKER_00: Uh so the first responder is the person who was in the middle of the street and not buried.
• 40:21
  SPEAKER_00: They just gotta go start digging people out.
• 40:24
  SPEAKER_00: And it seems like that's that's the case for some of this too.
• 40:27
  SPEAKER_00: It's uh who on your block needs help?
• 40:30
  SPEAKER_00: You know, who if everybody checked on one person, that would that would help a lot.
• 40:38
  SPEAKER_02: Are there and and again, because you mentioned the Paradise book, and and again, again, from my experience with the California Home Patrol, there were so many times that that we assisted with evacuations, and and we were mentioning before the Orville Dam, um, that large evacuation that was fairly short notice.
• 40:59
  SPEAKER_02: I feel like there were a lot of things we learned from it.
• 41:02
  SPEAKER_02: Are is that a part of this type of research too?
• 41:06
  SPEAKER_02: Is looking at that like that after action.
• 41:08
  SPEAKER_02: Um, because sometimes when we're introduced to a horrible situation and we just have to respond, we respond, we do our very best.
• 41:19
  SPEAKER_02: I I don't know, again, for my part as that practitioner, if I've seen a lot of um after action reviews of evacuation modeling.
• 41:29
  SPEAKER_02: Um, are those do those happen all the time?
• 41:32
  SPEAKER_02: Are they available for people to look at?
• 41:35
  SPEAKER_00: Um, I think uh, well, modeling is, you know, there are a few companies now that are doing evacuation modeling.
• 41:43
  SPEAKER_00: Um, but it's it really isn't done uh very much in terms of like in Paradise.
• 41:48
  SPEAKER_00: I don't know if they ever ran any models before before this fire.
• 41:53
  SPEAKER_00: So I yeah, it should be clear that uh obviously access and functional needs and all these other areas that you're talking about, they find their way into planning.
• 42:01
  SPEAKER_00: They're definitely involved.
• 42:03
  SPEAKER_00: That's that's a that was a very experienced community.
• 42:05
  SPEAKER_00: They had had like 13 near misses since 1999, so they understood fire and they had done a lot of preparation.
• 42:13
  SPEAKER_00: And without a doubt, they were an A plus for preparedness.
• 42:17
  SPEAKER_00: Uh but then you can have a scenario that's you know a tsunami of fire like that one hour away, and you know, it's just all the planning in the world couldn't have prepared them for that.
• 42:28
  SPEAKER_00: So it uh yeah, so I want to separate that that all this stuff is going on in the actual world of planning, but on the modeling side, um, and there's there's so many communities that haven't done any modeling and don't, you know, just don't have the resource or whatever, but I do think that it's becoming more uh more common for uh counties to try things like evacuation modeling.
• 42:54
  SPEAKER_00: It's becoming more available.
• 42:56
  SPEAKER_02: Well, and even if they don't they don't have the software like you, they can listen to you, they can read um the comments, they can listen to this and start thinking through the things that you're bringing up.
• 43:07
  SPEAKER_02: Um, and at whatever level they can.
• 43:12
  SPEAKER_02: So um I was gonna ask, and again, my mind is going through so many experiences and and different events that that I've been a part of, and I'm thinking, okay, like I wish I could go back um and adjust.
• 43:26
  SPEAKER_02: Um, and one of the things is about phased messaging.
• 43:30
  SPEAKER_02: This is a topic that um is very much um present right now amongst practitioners as hey, is there evidence?
• 43:40
  SPEAKER_02: Is there not?
• 43:40
  SPEAKER_02: Is it something we should do, or do we just tell everybody at once, one thing?
• 43:45
  SPEAKER_02: And so my question to you is what evidence do we have that phased messaging or evacuation staging is important?
• 43:54
  SPEAKER_02: And what can be done to plan for rapid onset events like this tsunami or others?
• 44:02
  SPEAKER_00: Okay, well the first the first question, yeah, there's evidence in practice and also in research that phasing evacuations uh reduces risk.
• 44:13
  SPEAKER_00: Not not all the time in terms of saving time or m you know making it quicker, but if you just think about it from the point of view of a fire, evacuating a stage or a phase zone closest to the fire, people that are at the greatest risk, while having some of the others wait gets them away from the fire, you know, that first group.
• 44:36
  SPEAKER_00: So I know that and I think Delaware, I think Rachel Davidson did some work on this on uh instead of minimizing evacuation time, minimize risk, minimize overall risk.
• 44:47
  SPEAKER_00: And that you can really reduce the risk to loss of life if you get the people closest to the fire out first without having them hit traffic cues because everyone left, right?
• 44:59
  SPEAKER_00: So get them out, then phase two.
• 45:02
  SPEAKER_00: It's it's easier to think about it if the fire is coming from one direction, which it usually is, and then you have phase one is the closest, phase two is the next closest.
• 45:12
  SPEAKER_00: And I always think about uh I think it was Meletti who used to say, you know, tell the people you want to leave, tell the ones you don't want to leave not to leave.
• 45:21
  SPEAKER_00: And you won't get 100% compliance in either way, but if you can tell the people that you know, if you're not in zone one, don't leave yet, then that can the roads can stay clear so those people can get out.
• 45:35
  SPEAKER_00: So and then there's also been a lot of simulation studies uh where they have shown that uh yeah, you know, you want to prevent gridlock.
• 45:45
  SPEAKER_00: Once if you tell everyone to go in some of these low egress communities that you know in the wildland urban interface, you can get gridlock at the intersections, and then you've kind of lost your ability to move anyone, really.
• 45:59
  SPEAKER_00: So phasing and staging is definitely better.
• 46:04
  SPEAKER_00: Now there are a lot of questions about the predefined zones.
• 46:07
  SPEAKER_00: Jeanette and I have talked about these.
• 46:10
  SPEAKER_00: Um I did see a talk a few years ago by uh Chief Honia from Butte County, and he managed uh both Paradise and the Orville Dam.
• 46:21
  SPEAKER_00: How about that?
• 46:22
  SPEAKER_00: Not to mention a ton of other events.
• 46:24
  SPEAKER_00: I don't know what's going on with Butte County, but they have a lot of evacuations.
• 46:28
  SPEAKER_00: And he said, uh when I asked him about that, he said, I don't have time to design evacuation zones.
• 46:35
  SPEAKER_00: I'm in a helicopter looking down on the community.
• 46:38
  SPEAKER_00: I'm talking to my person in the cube farm at the EOC.
• 46:43
  SPEAKER_00: If we have to split a zone, we'll split a zone.
• 46:45
  SPEAKER_00: If we have to change something from our predetermined zones, we'll do that.
• 46:49
  SPEAKER_00: But I I really got the message, you know, it helps to talk to the people that actually do it versus the theoreticians in the ivory tower.
• 46:59
  SPEAKER_00: That's for him, pre-evacuations or pre-designed zones are a godsend.
• 47:05
  SPEAKER_00: And he'll adapt them if he needs to to deal with the actual fire.
• 47:10
  SPEAKER_00: But um, so yeah, stages uh stages are a big deal.
• 47:16
  SPEAKER_01: I think we've talked about zones in different ways.
• 47:21
  SPEAKER_01: And so we'll just you know, just to open it up a little bit more for those who are listening.
• 47:27
  SPEAKER_01: I think when I find that zones are problematic is not the use of them operationally, like a pre-designed area, that that area is going to get a message going only to that area.
• 47:42
  SPEAKER_01: Um, but I mean that's wonderful because it is pre-identifying, not having to figure out the do they all can they all get out, have can do they have the the streets that they can use to leave the area that's already defined, which reduces that decision making at the time operationally.
• 48:02
  SPEAKER_01: Where it becomes a problem for me is when the message that goes out, it only says what zone needs to evacuate.
• 48:12
  SPEAKER_01: It's it's that, you know, if they're evacuating zone 14B, it says in the message evacuate one's one four B.
• 48:22
  SPEAKER_01: And so unless the people who have been predefined in that zone know that they are in zone one, four B and that it applies to them, they could be wondering what in the world is going on here.
• 48:34
  SPEAKER_01: And it's the use of the jargon that's not clear that is the concern that I have.
• 48:41
  SPEAKER_01: I love the idea of using zones operationally for making those decisions.
• 48:46
  SPEAKER_01: It's the communication of that information to the public that gives me a little pause about how effective is that going to be for people.
• 48:55
  SPEAKER_00: Yeah, they do have uh the whole know your zone thing, right?
• 48:59
  SPEAKER_00: So in the in the peacetime or whatever it's called, uh they will have people, you know, remember you're in easy for Q7IJ.
• 49:09
  SPEAKER_00: And uh now I've been told they can learn them by people that do the evacuations, but I also think that Butte County has the luxury of not being a Santa Barbara or a uh Tuolummy, you know, kind of uh tourist destination.
• 49:28
  SPEAKER_00: You know, you don't have a lot of uh maybe you don't have as many international visitors there, or maybe you don't have as many, I don't know, just tourists in general, right?
• 49:38
  SPEAKER_00: And so a lot of these zones, what happens if you're in Tahoe and it's the peak tourist season, what do you do with your one E7Q5J then, you know?
• 49:50
  SPEAKER_00: So uh I'm I'm with you on that.
• 49:54
  SPEAKER_00: But it it does seem like with Genesis and you know, that's the company, right, that creates a lot of these uh zones for counties that it's just growing.
• 50:04
  SPEAKER_00: You know, it's not it we haven't seen that moment yet that you and I are fearing where everyone is like, what zone am I in?
• 50:13
  SPEAKER_00: You know?
• 50:14
  SPEAKER_00: Um so it it it maybe it will happen and there'll be a moment where it's like this doesn't work, but it hasn't hasn't happened yet, I don't think.
• 50:22
  SPEAKER_00: Not with wildfires.
• 50:25
  SPEAKER_01: Well, it would be interesting to go back into those places that have received those messages to ask how did how did they know that that zone was where they were?
• 50:35
  SPEAKER_01: Is it because they had memorized their zone or because they had signed up for an alert to come to them and if they got an alert, they knew that they were in the zone.
• 50:43
  SPEAKER_01: Um, or if they had to go and look it up to figure it out, or if somebody called and said, hey, it's time for us to go, like there's different pathways to knowing your zone to leave when it's time to leave.
• 50:56
  SPEAKER_01: Um and uh you know, it would be a fascinating post post-event survey to find out more about that.
• 51:04
  SPEAKER_01: And I think it's wonderful that more organizations are adopting new strategies to get the word out to people who need it.
• 51:12
  SPEAKER_01: Um and just, you know, if we follow the MLEDI model of the warning response model, like we know that messages need to be really clear to be actionable.
• 51:23
  SPEAKER_00: So yeah, I think I think uh, well, sometimes a couple things you made me think of.
• 51:29
  SPEAKER_00: Sometimes the zones are followed by a brief description in text.
• 51:33
  unknown: Yes.
• 51:34
  SPEAKER_00: So you'll see, you know, this you know, just you and I have talked about one, two, three, ready, set, go, right?
• 51:40
  SPEAKER_00: And they'll never say just level three.
• 51:43
  SPEAKER_00: They'll say level three, go now, you know, or you know, so they can never just say the number.
• 51:50
  SPEAKER_00: They'll always say.
• 51:54
  SPEAKER_00: I guess they could, but I have I usually say every time they say level one, this summer I watch it, or level two or whatever, they always accompany it with the text.
• 52:03
  SPEAKER_00: And I think it's similar with the zones, you know, they might say like give the license plate version of it, and then say, you know, south of highway twenty, east of the river, that kind of thing.
• 52:16
  SPEAKER_00: They should name the river, but you know, I'm not.
• 52:19
  SPEAKER_00: So uh what was the other thing?
• 52:22
  SPEAKER_00: I guess that was it.
• 52:23
  SPEAKER_00: That was it mainly.
• 52:24
  SPEAKER_00: That uh yeah, they do have the ability to look up their zone and also sometimes the you know, with 360 in the in the Wii A now you can actually include some text along with the code for tourists especially.
• 52:40
  SPEAKER_00: Even they would have trouble though, with some of the sort of spatial uh reasoning.
• 52:45
  SPEAKER_00: I saw a zone this summer from Pueblo, Colorado that was if you live, if you are within two miles of such and such intersection, and you know, that geographers know that as a GIS query that's very hard to do in your head.
• 53:01
  SPEAKER_00: Line of sight, am I two miles from where is that intersection, first of all?
• 53:05
  SPEAKER_00: And then second, am I two miles from it?
• 53:08
  SPEAKER_00: I don't know.
• 53:09
  SPEAKER_00: As the crow flies, we don't fly like crows, so we really don't, you know, doing Euclidean distance buffers in our head is not something we're known for.
• 53:19
  SPEAKER_02: So I find it problematic when people use zones and then they tie in a specific evacuation route to that zone.
• 53:27
  SPEAKER_02: And even if there are, let's say I'm evacuating, I won't go super creative, I'll say zone one and zone two without all the other fun things you can attach to it.
• 53:37
  SPEAKER_02: And they're saying zone one, zone one, head southeast, zone two, head southwest.
• 53:43
  SPEAKER_02: It because it really just seems like there are so many assumptions uh being made.
• 53:47
  SPEAKER_02: And that's where it gets to, I think, Jeanette's point, which I we're on the same page in the same book that we want to sell, is operationally all of that makes sense.
• 53:58
  SPEAKER_02: But for people receiving it, whether they're permanent residents or even the tourists, and Santa Barbara is a prime example because for those that don't know, it is absolutely gorgeous.
• 54:09
  SPEAKER_02: Santa Barbara is a wonderful place.
• 54:11
  SPEAKER_02: But it's prone to fires, prone to floods, prone to multiple other disasters.
• 54:16
  SPEAKER_02: And so their Office of Emergency Management is very good at all these things.
• 54:22
  SPEAKER_02: Um, but for people who aren't familiar, um, you really have to give that plain language.
• 54:29
  SPEAKER_02: Let me explain it to you as much as possible so you can take that action.
• 54:35
  SPEAKER_01: Yeah, I think one of the things we were talking about is um maybe related to this last question that we have about these new smart systems that are starting to come out that might give more information to a person who needs to determine Southeast versus Southwest.
• 54:53
  SPEAKER_01: I mean, the instead of like a static escape plan that maybe there are some new technologies that are coming online, dashboard alerts, digital signage, like what's next that can help us to overcome this particular problem of directing people the right way.
• 55:12
  SPEAKER_00: One thing that uh, well, yeah, I got a bunch in that area, but one that I think would be helpful, although it does require cell phones to work, would just be if everyone had a way to click a check mark when they were out of the area, or click a check mark if they were out of the area before the evacuation even started.
• 55:33
  SPEAKER_00: You know, there there have been uh many evacuations.
• 55:37
  SPEAKER_00: I can think of one in Lake Tahoe and Gorafire in 2007, where we did a survey, we were studying family gathering with one of my master's students of mine, and we got all these surveys back.
• 55:49
  SPEAKER_00: We weren't in the area, we weren't there.
• 55:50
  SPEAKER_00: We were hiking, we were shopping, we weren't even home.
• 55:54
  SPEAKER_00: And so it's like, yeah, in the middle of the day in South Lake Tahoe, what percentage of the people were actually home?
• 56:01
  SPEAKER_00: It turned out we couldn't study family gathering if they were already together and hiking.
• 56:06
  SPEAKER_00: So uh, and they weren't even in the zones.
• 56:08
  SPEAKER_02: Well, no, I mean you think though, as so Jeanette, like we've talked about Apple's recent developments, but there's other technologies that are happening where right now you can get essentially an advanced type alert from Apple.
• 56:22
  SPEAKER_02: And it's this is new and it's still in testing.
• 56:25
  SPEAKER_02: Um, but it's essentially a secondary alert.
• 56:28
  SPEAKER_02: They take information from the cap uh messaging, and they can actually show a map where you are in relation to the polygon that is being established.
• 56:38
  SPEAKER_02: But what it doesn't show yet is a safe direction of travel, um, potential evacuation route, which Tom, you probably know way better than me.
• 56:47
  SPEAKER_02: Dealing with wildfires, a safe evacua uh safe direction of travel can change almost immediately because they're so dynamic.
• 56:55
  SPEAKER_02: So I don't know if you know other things headed our way.
• 56:59
  SPEAKER_00: If if there was an app where somebody could say, I'm in the zone and I need assistance, I'm out of the zone, don't worry about me.
• 57:07
  SPEAKER_00: Uh my son is in the zone at this address, and I can't reach him because you've blocked the entrances, you know.
• 57:14
  SPEAKER_00: So just if there was some accounting to figure out who needs help, who's who's already out.
• 57:20
  SPEAKER_00: Uh, or you could say, we're just leaving now, we have our go bag and we're going, you know, some sort of uh quick app to tell the emergency managers, you know, they could have a map showing them here's the here's the rescues we need to do, here's the people we don't have to worry about, you know.
• 57:39
  SPEAKER_00: Because they do go door to door and put stickers sometimes on houses they know are clear, uh, but that takes a lot more time and a lot more uh manpower, people power, uh than uh than having an app.
• 57:54
  SPEAKER_00: So uh some of the other technologies.
• 57:57
  SPEAKER_00: Well, I love Wii, everybody loves Wii.
• 57:59
  SPEAKER_00: Um you know, you've got uh mo many lifetimes of a career, Jeanette, helping people learn how to use it correctly.
• 58:08
  SPEAKER_00: It's like taking on world poverty, you know, generations will be able to work on it.
• 58:14
  SPEAKER_00: Uh but but for the instant notification of everybody right where they stand, including tourists, that if it's used effectively, uh so that's and then now with the 3.0 polygons, um, you know, to s to get it so it's targeted like that is amazing.
• 58:32
  SPEAKER_00: Um I think I know you had a session on AI on this channel.
• 58:37
  SPEAKER_00: Uh, and so I do I have a lot of questions about when can AI do better than a person a human, and when can a person do better than AI?
• 58:47
  SPEAKER_00: You know, so when we see a lot of the sort of you know, think of them maybe as warning mistakes or failures even, uh it seems like something where AI could have done better.
• 58:59
  SPEAKER_00: In other cases, anybody that's worked with AI knows how many weird mistakes it can make when it hallucinates.
• 59:08
  SPEAKER_00: And you don't want a system that hallucinates telling neighborhoods where to go when and things like that, you know.
• 59:15
  SPEAKER_00: Uh so I but I think there's a it's more of a research area.
• 59:20
  SPEAKER_00: I'm not ready to deploy AI yet, but it's more of a new research area for uh what what or where can it assist?
• 59:28
  SPEAKER_00: Decision support.
• 59:29
  SPEAKER_00: That's what we used to call it, you know.
• 59:32
  SPEAKER_00: Spatial decision support was a term in GIS, but the GIS field.
• 59:36
  SPEAKER_00: But where how can it assist?
• 59:38
  SPEAKER_00: You know, it would be in some cases it would be nice where if it said, you should send the warning now, based on the progress of the fire and based on the estimated time of this community that I've come up with, this would be a good time.
• 59:52
  SPEAKER_00: And the person could either override it and say, No, I want to wait longer, uh, or they could say, you know, this the system is telling us that we should do it now, and and uh what do you guys think?
• 1:00:04
  SPEAKER_00: And everyone says, we agree, let's warn them.
• 1:00:08
  SPEAKER_00: You know, so I I don't know, this is kind of the future, maybe.
• 1:00:13
  SPEAKER_01: I think we need to have somebody who actually does the in the dashboard alerts come and talk to us about how that the the data that people input like through ways or um put a notifying that there's a traffic accident up ahead kind of information as you're driving, um, how that can all be pulled together to think about routes and giving people live instructions.
• 1:00:42
  SPEAKER_01: I would bet that the that there's somebody really smart who's thinking about that possibility in the future because it really does feel a little far out there, and yet we get our directions are changed while we're driving all the time because of traffic accidents.
• 1:00:56
  SPEAKER_00: I I did write a paper about how we need to make our scenarios more challenging.
• 1:01:01
  SPEAKER_00: You know, inspired by Paradise and other scenarios that are simulations are too optimistic.
• 1:01:08
  SPEAKER_00: It's kind of a, you know, all the roads are available, nothing ever changes, nothing bad ever happens, and everybody just progresses to the edge of the evacuation zone, and you clock the time and say, look at how great my plan is.
• 1:01:21
  SPEAKER_00: That was a great era, but we need more challenging scenarios to be uh able to keep up with the real world, which is playing hardball, obviously.
• 1:01:31
  SPEAKER_00: The real world's playing fastball, and we're sort of playing T-ball in our journal articles.
• 1:01:38
  SPEAKER_01: I think that's a great place to stop and reflect on how important it is, as Eddie was mentioning earlier, about the inclusion of actually doing this kind of training activity in your alert and warning training practices, of putting in the right kind of scenarios that really challenge your assumptions and not just checking a box of write saying, I wrote a message and I issued it, and everybody evacuated immediately.
• 1:02:11
  SPEAKER_01: They all complied.
• 1:02:12
  SPEAKER_01: It was perfect.
• 1:02:13
  SPEAKER_01: But actually thinking through those challenging things.
• 1:02:16
  SPEAKER_01: And that is one of the things that the alerting authority does in trainings, coming to communities and helping to think about the anatomy of an alert from beginning to end in challenging scenarios, not just evacuation, which is a big thing, but also other really challenging scenarios that have high uncertainty and a lot of fear.
• 1:02:40
  SPEAKER_01: We also conduct message audits and reviews of message templates to help to make sure that all of those key contents are in there to really help to motivate people to take act, uh, take action, um, doing policy assistance as well.
• 1:02:55
  SPEAKER_01: And we have a new boot camp coming up.
• 1:02:57
  SPEAKER_01: Um, the we just completed one in July, and our next boot camp will be the week following Thanksgiving, November 30th through December 4th.
• 1:03:08
  SPEAKER_01: So please, if you're interested in that, um, reach out.
• 1:03:11
  SPEAKER_01: But um, Tom, this has been a really interesting conversation.
• 1:03:15
  SPEAKER_01: I think you've raised a lot of concerns and questions, a lot of opportunity for more research.
• 1:03:21
  SPEAKER_01: Grad students, I hope you were listening to this one.
• 1:03:24
  SPEAKER_01: There's a lot of opportunity there.
• 1:03:26
  SPEAKER_01: But um, thank you.
• 1:03:27
  SPEAKER_01: Thank you for joining us today.
• 1:03:29
  SPEAKER_02: And my pleasure.
• 1:03:31
  SPEAKER_02: I might want to do a part two, but how I would want to do a part two, Tom, because again, the insights I'm I'm working through, I would literally put up a picture for those that were would be watching this, we'd describe it.
• 1:03:44
  SPEAKER_02: And I would love to essentially bring up a scenario and have you walk me through all the nuances and things that I should think about in in a in a scenario.
• 1:03:57
  SPEAKER_02: And because I can I can listen to this and I can be like, okay, I'll do a scenario.
• 1:04:02
  SPEAKER_02: But I but I know I need help thinking through this because this is not my specialty, but it's something that I have to get good at and get better at.
• 1:04:11
  SPEAKER_02: Um, so if you're up for that in the future, um, I'd love to to bring a couple and again, we can do wildfire and some others, and maybe some of the viewers have some suggestions of things they would love to see you walk through.
• 1:04:25
  SPEAKER_02: Um and just to help.
• 1:04:27
  SPEAKER_02: So I it won't turn it into that firefighter police and we'll change 15 things, but we can at least talk through it so that we can model what what we're hoping practitioners will start doing as well through their their tabletops.
• 1:04:42
  SPEAKER_00: Sure.
• 1:04:43
  SPEAKER_00: Yeah, that'd be great.
• 1:04:46
  SPEAKER_02: Okay, well, just thank you again.
• 1:04:49
  SPEAKER_02: It was a pleasure, and I just want to echo everything Jeanette said.
• 1:04:52
  SPEAKER_02: Thank you guys for letting us do this.
• 1:04:54
  SPEAKER_02: Thank you for listening.
• 1:04:55
  SPEAKER_02: Um, every second has a story, and it's our pleasure to be here to help share yours.
• 1:05:00
  SPEAKER_02: So thank you.