 Wherever you're joining, thank you so much for being here today for this NCAR Explorer Series special event, leaving it all behind evacuation lessons from wildfires in Colorado with Kat Ashley, Will Cannon, Timothy Giuliano, and Scott Pierce. I am Dr. Evie McCumber, and I am an educator here at the National Center for Atmospheric Research, or NCAR. NCAR is a world-leading organization that is dedicated to understanding earth system science, including our atmosphere, weather, climate, the sun, and the importance of all of these systems to our society. I am very excited and so glad that you all are here today. Just to learn more about how scientists from different disciplines can all come together to determine when and how to issue evacuation orders. For this event, some logistics, you will be able to ask for panelists questions following this panel, and Olia and Alexandra will help moderate that so we can ensure that we hear from both our in-person and virtual audience. If you are in-person, you can raise your hand, and we will give you a microphone, and then you can ask your question. If you're joining us virtually, hello. You can ask your questions using the Slido platform. If you scroll down the webpage, you will see the Slido window just below where you are seeing the live stream video of the event. If you haven't already, go ahead and click on the green join event button, and then you can ask questions on the Q and A tab and answer questions on the poll tab. Kat, Will, Tim, and Scott also have a few poll questions for us, so for both our in-person and virtual audience, you can definitely respond on Slido. For those in-person, you can go ahead and feel free to use your phone, and if you have your laptop, you can also do that, so you can respond on Slido. Go to Slido.com and enter the code hashtag NCARMSR. But please, please, please be sure to join Slido to add your thoughts to our word cloud question. What do you think of when you hear the word evacuation? Because we are definitely going to get to that really soon. What do you think of when you hear the word evacuation? This event is also being recorded and will be available on the NCAR Explorer Series website. We have a series of very accomplished scientists, so the bios are going to be long. Just make sure that you know that. With us today, we have visiting scientists, Captain Etchley and William Cannon, in addition to NCAR scientists, Timothy Giuliano and Scott Pierce. Dr. Captain Etchley is an assistant professor of natural research sociology in the school of forestry at Northern Arizona University and an early career faculty innovator with NCAR in the program second cohort. As a wildfire social scientist, she explores how communities are adapting to and recovering from wildfire. Her research date has span nine U.S. states and several notable wildfires, including the devastating 2019 campfire in California. She earned her Ph.D. natural resources from the University of Idaho and holds a master's in science, a master's in risk and environmental hazards, and a bachelor's in geography from Durham University in the U.K. William Cannon is currently a Ph.D. student in the school of forestry at Northern Arizona University, working with Dr. Captain Etchley as part of the early career faculty innovator program. His current work in wildfire social science explores how socially diverse communities interpret and respond to fire in evacuation and recovery context. Using a qualitative interview approach, he interviews, he investigates the diverse perspectives, knowledge, and experiences of communities affected by wildfires aiming to identify effective strategies and approaches that can facilitate a more inclusive and informed decision-making process. He earned a master's in geography at Northern Arizona University and a bachelor's in geography from the University of Utah. Timothy Giuliano is a project scientist at the research applications laboratory of NCAR. He joined NCAR in 2019 as St. Louis doctoral fellow after earning his bachelor's in meteorology from Millersville University and his master's and Ph.D. degrees in atmospheric science from the University of Wyoming. His research focuses on numerical weather prediction with interest spanning a variety of lower atmospheric problems, including boundary layer dynamics and turbulence, while wildland fire prediction meteorological impacts on renewable energy, and aerosol cloud interactions. Scott Pierce is a software engineer at NCAR where he helped develop the vapor 3D visualization package for the geophysical sciences before joining NCAR. Scott designed and deployed remote sensing systems for atmospheric physicists and produced quantitative precipitation estimation analyses for various government agencies. He holds the bachelor's in electrical engineering and a master's in computer science from the University of Colorado Boulder. Scott and Tim's work in visualization of the east troublesome wildfire just won best overall visualization at this year's practice and experience in advance which are computing conferences. So congratulations on that. Hey, so I am almost on talking elsewhere. I'm going to pass it on to them. You can hear more interesting things in a second. But before I turn it over to our speakers, let's shake out your thoughts on our word cloud. So could Paul and Jesse please share a slide up with us and the word cloud answers. Whoa. So I'm going to pass the mic to Cat so they can start discussing your answers. This is very representative I think the community we're in right now. This recent history of fire in the area. You've had a lot of evacuations recently in this area. I'm seeing a lot of great themes here, emergency safety, and then there's a lot of leaving really fast which I think is very unique to the experience that this area has had. Usually we might see some kind of planning stuff and then some very visual language. So we can see a little bit of that. A lot of concern and conversation that's really shining through in this. Can we go back to, if I pressed this, will we go back to the PowerPoint? Hey, thank you. Okay, so you heard a little bit in our introductions about our very diverse backgrounds. We're social scientists, modelers, we're visualization folks, and we all came together because of a program here at NCAR for the innovators program. It brings social science professors from universities across the U.S. to NCAR so they can work with modelers in the different labs here. And we have this idea of convergence research. So what can we achieve together that we couldn't have achieved if we kept in our disciplinary silos? So when I wrote the proposal for this project, my background is wildfire social science and really interested in how communities are responding to and adapting to fire, how different experiences might influence different responses. So we all came together around this interest in fire and how we can make people safer in the long run. So we have some other collaborators who couldn't join us tonight, Franco Placovic, Rajesh Kumbar, Gaby Frister, and there's many more who have dipped in and out, so the work we'll share tonight is a combination of everyone's efforts in that. So this question was how many wildfires are there on average in the state of Colorado? Most folks are leaning towards at 3,600, and the correct answer is actually the highest one. Oh, not quite the highest one, 5,600. So a lot of those are probably less than 100 acres. We often think that large fires really kick off around that 100-acre mark, so that number is probably a lot of lightning emissions at a small acreage. Maybe 95% of wildfires stay under a few acres, so it's that 5% that we really want to be concerned about. Could we now head to the why might some people not evacuate question? Great, so we've got a broad range of things here. Pets, leaving things behind that can create a lot of anxiety, understandably limited mobility, attachment to home, our connections to place are really important in the ways that we navigate and think about these fires. Uncertainty, no call came. That is a big conversation right now. I'm sure it's the same in this community with recent fires, whether we should expect something formal or not, or if there's time for that. Great, so I think then we'll head back to the cal points, if that's okay. So I want to start with a little overview of what we know about evacuation. In those slides we really talked about, in those questions you just answered, you can kind of see the story of increasing fires. We're having more than 5,000 fires start just in this state. We're talking about tens of thousands across the US, and then if you scaled it up, think about what's happening in Canada. We see a lot in Australia, increasingly in Europe and South America. This is a big conversation right now about how do we keep communities safe and how do we help them make decisions that keep them safer. A lot of that story also goes with more frequent fires and larger fires happening in locations that didn't happen before. So I'm based in Arizona. We're seeing a lot of fires in the Sonoran desert right now that there hasn't really been a track record of before, and it's telling us a lot of new information about how different kinds of vegetation burn, what a fire might look like in a community that's predominantly cacti and grasses, very different environments to what we think of when we think about fire. And if you think back maybe even just, well let's say 10, 20 years ago, a 50,000 acre fire was considered really big. Now we're in the 6th digits, we're seeing multiple two to 300,000 acre fires a year. That's a really concerning pattern that we've seen, almost an order of magnitude in the size of fires. We had our first gig of fire which I think is a million acres a couple years ago, so we're on this roller coaster for a long time to come I think. When these communities are increasingly at risk because of these fire activities and people are developing out into these areas, there are a lot of different factors at play in how we become at risk and how we change our risk for fire. Most of the time you'll hear officials say evacuation is the number one safest thing to do during a fire and that's for a number of reasons but predominantly we can take people out of the equation, priority is always life, life of the public and of firefighters. It's a lot safer in fire risk area when we don't have to worry about people coming in and out and making decisions we might not expect them to. But as we saw on that slide a little while ago, you can have some very different reasons for why you might not evacuate and stories behind that. We touched on some great ones in that slide. I want to add a few extra ones and I'd say the reasons behind this are pretty infinite. There's a lot of reasons that are valid or out of circumstance. The first is you find out about fire too late to evacuate safely. You really want to think about where the safest place is and if driving through a lot of fire is your only option it may be safer if you've done a lot of work around your property to stay there. Some people have a lot of experience fighting fire. The most common folks I see staying off the fire in my work are often people who've had a career in firefighting. They've been a volunteer firefighter. They think they have a sense of what to do and then they are good enough at reading fire behavior that hopefully they're humble enough to know when to step out of that situation or leave. Then we've got lower risk perceptions. Sometimes you might hear about a fire and it's a way away and you make that decision for yourself of what risk you're willing to take and that's a big part of the conversation we'll talk about today. Another one is lack of trust in professionals especially in rural areas. There can be a lot of distrust for the Forest Service, for the government. Maybe you want to stay on your property because you don't trust them to protect it and that may or may not be true especially in fast moving fires. We can't always be everywhere all at once but that can be a big driver of why people might stay behind without a skill set around fire suppression. So all of these together you can see tell a little bit of a story about experience with wildfire. Do you have it or not? Does it drive your risk perception? Have you been in a fire before and now you have a clearer idea of what you do in the next fire or you want to change your approach completely because the outcome of your recent fire wasn't quite where you wanted to be. It turns out there's a huge amount of research out there about people and the decisions they make during evacuation and a lot of studies that happened before fire so we're asking members of the public about hypothetically what they would do during fire. A lot of people say oh fire event happened yep I'd make a decision and I'd act on it immediately. I'd evacuate as soon as they heard of something in my area or I'm ready to stay and fight fire on my property or maybe you've done a lot of work on your property and you want to shelter in that. In reality we know that there's this kind of wait and see period instead fire happens and then you're in this information gathering cycle. What do we know about the fire? How far away is it? Is it moving fast? What are my other neighbors doing? You're collecting information to decide what you're going to do in that specific scenario. Fires behave very differently so it can really vary depending on that information that's coming in and while you're doing that we find that a lot of people look for two sources so they might hear it from somewhere else then they want to triangulate it with another person or another source. So you might see it on TV, you call your neighbor what do you think about this? What are you doing? Or you see all your neighbors packing their stuff then you go huh maybe I should think about this and you call your cousin who's in Washington state and is a firefighter and they're going to say I think you should leave also but we very rarely act on one source of information unless the time period is really condensed with very very obvious that the threat is imminent. I think now I want to turn to there's a slide from the Slido how many of Colorado's largest fires happened in the past 20 years or how many of the 20 largest fires happened in the last 20 years? Absolutely is this closer? How many of the top 20 biggest fires in Colorado's history have taken place in the last 20 years and we've got a good group of people who've got it all 20 have happened in the last 20 years so this situation is really condensed in time at least in recorded history to the last 20 years. Can we go back to the slide sorry? The picture I'm trying to paint here is how complicated fire decision-making can be. We have all these different sources all these different ideas about what might necessitate actions and what won't and that's why I think our team here kind of brings together things in a different way so this project that we're talking about today falls under this category of convergence research which I mentioned earlier it's about bringing together different disciplines integrating our methods so that we can find out new information that maybe we wouldn't have known before. In our project we really wanted to leverage all the wonderful atmospheric and fire modeling that's happening at ENTAR. So in our team we looked at some fires that have happened in Colorado recently and we drew upon all the modeling information from them to make visualization so you can see that on the left hand side there Scott will talk more about that in a minute that is wonderful work and we turned them into videos and then we took them out to community members and we used them to talk about their experiences so we met with folks in this area and also over in the Grand Lake area and we asked them tell us about your evacuation experience from start to finish and then when they were done we'd bring this out and say does this help you remember anything else does this tell you something else and our goal was we wanted to learn from them about how they make decisions but we also wanted to give them information about the event they experienced they can process it they can think about it so that they have a clearer picture of the story that they can take forward as well. We today have done almost a hundred interviews with different folks and we are still a work in progress on the Marshall fires so if you experience that and you like to talk to them my email is at the end of this would love to hear from you. Then I think we'll go back to the next slide on Slido which is the how fast are the east troublesome wind speed. For the fastest wind speed we actually had the fastest wind speed being the east troublesome at 60 miles per hour. It's kind of surprising it looks like for this crowd. Oh oh that's the one I had but it said 60. Oh might have a discrepancy to discuss in the Q&A. Yes we can discuss in the Q&A. Can we go back to the slide space great so for a little background on the two fires we studied the first one the east troublesome happened in a much more rural area than here so people were a lot more spread out you can see from this evacuation map covered quite a large area you can see some areas were left on pre-evac there was also a lot of second home ownership which created a lot of unique evacuation conditions as we'll get to a bit. So overall the fire on October 21st 2020 made a large run of approximately 90 acres in one day so covered significant ground made a lot of challenging conditions for evac. Most residents experienced about 15 minutes from getting any pre-evac notice to getting a mandatory evac notice because of that significant run and that was during the evening near dinner time for these people which created other challenges overall 192,000 acres with 366 homes lost onto a fire I'm sure most people in this crowd are familiar with the Marshall fire smaller footprint 6026 acres but very destructive over a thousand homes lost over two billion dollars in damage and most people that got evacuation notices in this area had less than 20 minutes between that notice and needing to leave immediately so some comparisons between the two fires both were very fast moving at the time of evacuation little difference is the east troublesome kind of sat for a while for a week and was burning off in a forested area so people were aware of it and maybe kind of made up their mind that it was going to stay over there that might not necessarily be the case but and then both occurred outside of a traditional fire season so October and then December 30th obviously for the Marshall fire kind of might change how resources are allocated or what resources are available for suppression changes how evacuation planning is and then there's also the just challenges of the weather they came with that and getting to that the wind driven aspect of these fires feed was a big factor in the rate of spread for both fires little bit of a difference the Marshall fire obviously grass fire so it spread really fast versus the forest fire where you have these larger fuels that can burn for longer time sitting in area it was also the urban versus rural comparison so for evacuation you have this issue of condensed people in a single area or you have people really spread out that you all need to get out through a single egress point so different challenges for both but both had very distinct challenges and then fire behavior also changed once it hit structures for the east troublesome this resulted in combination with the wind in the flume collapsing which kind of acted as a kind of cascading force for embers to spot fire further from the flame front and then for the Marshall fire once it hit structures you have these wooden fences that are spreading the fire you have different structures burning at different rates and then there's also very localized changes due to those structures in their placement all right so before we get to the beautiful animations that that scott's created I wanted to first discuss the numerical modeling that we do here at NCAR and that we've done for this project that really kind of you know makes the visualizations possible so just for some kind of general background you know wildfires of course are very complex and there's this intimate relationship between wildfires and the atmosphere so you know many folks have had a campfire you know fires are very hot now imagine scaling that up to a wildfire right there's a lot of energy heat and moisture that's released when when fuels are being combusted and so we need to really capture that behavior in the model accurately if we want to you know depict how the fire is behaving right so as it says here the wind and the fuel moisture are some of the most important contributors to fire spread the wind of course will affect how fast the fire moves but also in which direction it moves winds can be highly variable and so yeah capturing that well is important and yeah this feedback can also lead to development of what are called pyro cumulus clouds it's an example that's shown in the top right so it's this link between the emissions from the fire they release a bunch of aerosol heat and moisture and that can actually create pyro cumulus clouds which which can be really detrimental to for example firefighters on the ground so really it's this kind of tight interaction that we need to capture in the numerical model so in order to predict fire spread we need to use what's called numerical weather prediction now many folks are probably familiar with weather models right you tune in to the news or go to the national weather service and you're getting information about the weather from weather models that are run and so what we need to do is we need a couple of that type of model with a fire behavior model so that we can inform the fire model how the fire will spread based on you know the atmospheric winds and and different variables so on the left hand side here is just an example of kind of a numerical weather prediction model output that you may be familiar with this example is just precipitation so we have this information covering a large area but we really need to get down to the fine scales right fires are occurring at relatively local and regional scales and so we need to somehow go from this kind of broader picture down to really fine scale fire behavior that depends on these different meteorological parameters and it really aims to predict the fire spread the amount of heat that's released the amount of smoke that's released and all of this is dependent on yeah wind speed terrain slope and the fuels on the ground and in the canopy potentially so this picture here is a low intensity kind of ground fire as we call it you can see that it's just kind of the ground fuels that are being burned whereas you can also have you know events like this where the canopy ends up being consumed and you can have kind of this really large fire and in this case of course you have a lot more energy being released to the atmosphere and more intense fire so again we go from this kind of kind of bigger picture weather model down to much finer scales in order to really capture the wildfire so the idea here is we take outputs from yeah national weather service or national center for environmental prediction and we kind of zoom in hone in on an area that we're interested in for you know where a particular fire may be occurring in this case you know we just are showing Colorado as an example because of course we're interested in a lot of the fires that are happening here and so we refine down in our numerical model to this smaller region and then what happens is we can go down kind of in increments and we can refine further and further and ultimately we can get down to the scales that a wildfire may may be covering and so in our model we need really fine grid cells to be able to resolve the fine structure of the fire but in any case we can do this over different regions you know we we really can adapt to to where the fires may be so just to show you an example of some of the model outputs from a simulation that we ran for the cold spring's fire which occurred in 2016 up near netherland some of you may be familiar with that what we're showing here is an animation from the model outputs and the the color volume rendering there is the smoke that's generated from the fire and there are wind arrows down near the surface which shows the wind speed and direction and you can see really how much variability there is in where the smoke's going how much smoke is being produced kind of at what levels it's being transported at and so really in order to get this this accurate picture we need to be coupling the atmosphere with the fire and also of course the fire is feeding back on the atmosphere right so it's this two-way coupling that that we need to really be able to capture and from the simulation on the right hand side there is just an example of the observed perimeter in the black and the predicted perimeter in the red from our numerical simulation overall it's doing pretty well we consider that a pretty good simulation the model does tend to overestimate the perimeter this was because in the model we don't have suppression efforts and that was actually a big reason why the fire didn't spread further in this particular case so just some some general thoughts about the modeling side of things of course the models and the simulations are just one piece of the very large puzzle right we really rely on measurements and observations from the ground from satellite and that help constrain our models but also making our model outputs useful for the consumer whoever that may be whether it's the public whether it's decision makers whether it's for research right so we need to think carefully about how you know this modeling piece fits into the the larger picture again this is just one element of the decision support and then a quote here all models are wrong but some are useful right so really understanding you know certain qualities about your model that may be good and it performs well in certain situations but understanding its limitations and when it may not be performing well that's really important and also just want to mention that as cats kind of alluded to you know this is really convergent research and it involves and requires collaboration between many groups of people from you know fire ecologists fire behavior analysts incident meteorologists emergency responders and also the public right so this is a an ecosystem that requires input and feedback from all parties and just another quote in theory there's no difference between theory and practice and practice there is right so oftentimes what we think is going to happen isn't what actually happens and and we need again that's why we need input from everyone right because we as scientists tend to be kind of siloed in our offices doing our research but really broadening out and getting input from other folks especially those who are on the ground and can observe things anecdotally that's that's really critical so I think at this point I'll pass it on to Scott still oh all right I think this is working I'm actually going to use the podium because I'll be using the cursor to point some things out I have some visualizations to show but I also want to talk about why we produce these visualizations at NCAR and also get into the technical details of how it's done so NCAR is a is the National Center for Atmospheric Research and it's comprised of eight labs or divisions and the one that I'm a part of is called CISL it's the Computational and Information Systems Laboratory and CISL's main role at NCAR is to support the scientists with the the science that they're they're conducting mainly through making sure that our super compute clusters remain operational to run the massive simulations that are run continuously 24 hours a day and CISL we also develop software to visualize the simulations that are run on the super compute clusters that we have and that's what I do I'm a part of the team of four engineers and scientists and we develop an application called vapor and let's test test is that better okay so vapor is a desktop application that can be run on the supercomputers but you can also download it and install it on your home PCs we support macOS windows and Linux so most operating systems out there are supported and it's also open source which means that you can look at the software the code itself modified if you want and it's just free to use anyone can download it from our website www.vapor.ucar.edu and you can ask support questions if you like at vapor at ucar.edu we also have sample data that you can download so tim was kind enough to provide us with Marshall wildfire sample data it's pretty big we only contain two time steps in the entire simulation and it's almost a gigabyte I believe the original simulation was more along the lines of two terabytes so that's one of the reasons that we need vapor to run on our super compute clusters it's just impractical to bring two terabytes of data onto a laptop and it'll just take forever to render it's just so much data we need the processing power of our super compute clusters though you might ask why is in car developing a visualization application and the fact of the matter is that 3d visualization is extremely difficult and most scientists just don't have the time to practice their science as well as do 3d visualizations simultaneously so with vapor our main mission statement is to make it as easy as possible to do visualization for geophysical scientists like the ones that work here at income um let's see here's one of our first visualizations of the east troublesome wildfire that's looking I believe to the northeast and it's using four renderers vapors comprised of 11 renderers and each renderer takes the data and depicts it according to color and opacity in different ways renderers that we use for these fires are the volume renderer for smoke the image renderer for the base map 2d data renderer for the fire perimeter and then the wind barbs are the arrows on the bottom that show the wind direction the east troublesome wildfire was a pretty notable use case well let's see did I miss one slide well it looks like one of the slides didn't get updated but I will uh note that the east troublesome simulation was a notable use case because during the time of the uh the fire the forecast models were not as accurate as they really could have been and the reason for it is if you live out here in Colorado you know about the pine beetle infestation up in the woods that caused all these dead down trees that were not taken account for in the original model simulation so uh Tim's associate Amy developed an AI algorithm that used the Sentinel 1 and Sentinel 2 satellite data from the european space agency to identify where this beetle kill existed there's just not enough resources by the national forest service to survey all this wilderness on a continuous basis I think the previous fuels were updated five years ago so they were that far out of date which led to the under prediction of the fire spread of the east troublesome fire so like I mentioned earlier vapor is comprised of several different renderers each one uh applies color and opacity to the data in different ways one of the flagship renderers is the volume renderer which uses a technique called ray casting and how ray casting works is it basically what vapor will do is it will load the data into the application and for each pixel that exists on the screen it will fire a ray or like a laser into the data domain and that laser will sample the data incrementally until it saturates and says we've collected enough data it's completely opaque or it leaves the domain if you play video games like minecraft ray casting works in the same way one laser per pixel and then that thing just samples the domain whether it's simulation data or video game so what the grid looks like the actual data is like a cube of grid of gridded data with a data point I guess at each grid point and so in vapor the ray will enter the domain from wherever your perspective is and it will start sampling that gridded data until it reaches saturation or it exits the domain and in this example it's kind of simple you can see that this ray coming down from the top maybe can sample the data at every 25 meters and then march down to the next 25 meters and continue accumulating whatever variable we're looking at in this case it would be smoke and eventually the ray will exit and then we will render a pixel to this to the screen Tim's model is called work it's the weather and research forecast model and it's a little bit different if I go back to one slide you can see this very nice regular grid equally spaced at all points and there's even gridded points underneath this hill the way that work works is that it uses what's called a terrain following grid where the grid will follow the shape and the contour of the earth and it will also a little curve around the earth but it will also be more condensed at the surface and the reason for that is because I think correct me if I'm wrong on this there's more convective mixing towards the surface so you need a finer resolution to resolve that and it's also where we care about what's happening with the weather so with these wharf models you'll see this is a very simplified example towards the surface the wharf model is very discrete very small grid cells at the bottom and then up in the stratosphere you get very big grid cells but the problem is with our ray we can't just sample at every 25 meters and so one of the things that vapor does to make it as easy as possible for scientists is accommodate for these special grids that scientists use so we have an algorithm that samples this grid correctly which isn't always available with other visualization packages basically how you use vapor in the volume renderer is through this one mechanism we call it the transfer function it's kind of the bleeding or the beating heart of vapor you can see over here this histogram is just a distribution function of one of the variables that we're looking at the minimum value is about minus 20 and the maximum value is about 45 and we have this this distribution of the variable there's a lot of values I would say this is maybe around zero a few values by negative 20 and a few values at 44 but basically vapor is applying this color palette to all the values that lie above it in the histogram so did I skip ahead so yeah so basically that's how the color is applied to a variable but we also have to apply opacity so right here on the right you can see the flames and the smoke you can kind of see this transfer function over here if we take all these dots and maximize them all the way to the top everything is completely opaque so all data values in our big brick are completely opaque the ray caster is going in there and it's just bouncing off the surface of this cube it's not penetrating into the actual data so we have to use this a series of control points to reduce the opacity of certain values of the data to make it more transparent you can see here by dragging them down a little bit the domain gets a little bit more transparent and then if we drag them down even more that's how we get to our final see-through version of the volume renderer and so yeah I think that's it for the basics on how vapor works there's youtube videos online to that go into more detail on all this and again it's free to download sample data is there so that's it thank you great yeah so moving on to kind of results that we found talking with people affected by both fires for the east troublesome we found that there was a large influence of not having a recent fire evacuation at least of the fire of the size there were previous fires but none that really warranted a large evacuation like the east troublesome that post challenges for residents planning as well as managers in some aspects it was burning for about a week prior to evacuation so again some people weren't necessarily thinking that it was going to reach them they thought they had it kind of under control at that point that led to texts kind of coming at short notice just because of how fast the fire was moving there was a lot of uncertainty about the fire location and that was due a lot of to wind speed and then the smoke especially when the plume collapsed the smoke got denser darker it was harder to see the fire another aspect was the topography there were a lot of mountain ridges that blocked view of the fire and then on the professional side the wind speeds were so high that it limited aerial reconnaissance so they were also uncertain some aspects about the fire at least from an overhead view an interesting aspect we found was that highway 125 in that area was a mutual trigger point both for professionals and residents this was due to the topography how it channeled wind the presence of fuels expecting it as a fuel break and then also its geographic proximity to the town professionals were trying to use it as a fuel break but then when that plume collapsed it just went right over and they weren't able to since it was a rural community a lot of people had family member or friends that were either working on the fire line or in emergency management in some aspect so they were able to get information ahead of time at least a little bit from those people they were saying hey it's cross this highway residents knew just based on their local knowledge that that meant that they needed to start thinking more seriously about evacuation the late seasonality of the fire for the east trouble some posed some challenges and some benefits again due to that second home ownership in the current outside of peak tourism season there wasn't as many people as there necessarily would have been in the summer however you also had people that were still up there visiting who maybe weren't keyed in to the local evacuation notices they didn't have that local knowledge of highway 125 so that posed some different challenges in that way similarly with the marshal fire people weren't necessarily thinking about a fire occurring people weren't necessarily at home maybe they were out shopping it was near new years people are getting ready for new years it's sudden onset posed household challenges so kids being home from school maybe they're at home parents are out doing something they need to go back and get them for evacuation people that can't drive being at home we spoke to some people that needed to coordinate with neighbors to get people out there was limited receipt of official warning this was due in part to how fast the fire moved and also how the companies that manage evacuation notices operate it's on an opt-in service so people weren't opted in they didn't get the notice there was uncertainty about the communication channels people were expecting media or official notices but again it moved so fast media was behind maybe on the coverage people weren't necessarily getting that information in that wait and see stage that they needed to make that decision to make the switch maybe until they actually saw flames in their backyard in some cases as we talked to people and then traffic for both was a challenge um east troublesome fire they had one point of egress essentially for the major town that created a lot of challenge martial fire in a densely populated area a lot of challenges there getting people out so when we took the visualizations to people we did find some uses for them for different groups different uses um for fire pressure professionals and emergency managers we found that it's really useful for rebuilding trust justifying decisions explaining challenges again getting that broader scale view of the fire why why it behaved the way it did how they responded how they were able to respond it's just giving additional context and it could be used as a tool in that way for fire professionals specifically we found that they could use it as a tabletop training exercise um so pulling up the fire um with maybe onboarding new folks um or people that aren't used to fire in that area showing the fire asking them how they would respond and then showing how they did respond or how they should respond and then proceeding through the simulation in that way another use along that line would be onboarding instant command teams when they switch over managing the fire they need to onboard the other professionals on how to respond um so you could use it in that way pulling up the fire as it's behaved so far again that's a time sensitive resource intensive simulation so that would be a future use down the line uh on the mental health side um I did want to know in our interviews when we brought up the um visualizations we would after exhausting our protocol we would try to be very um tactful on how we brought them up we would um ask if they were okay seeing the visualizations explaining what was in them they didn't want to see them we didn't bring it up we didn't want to um do that to people they didn't want to if they were we would bring them up and we found that for some people it was really helpful for processing the event um discussing the traumatic experience um it kind of was a visual something else to look at something else to kind of talk about their experience within a broader context um so we think that this could be useful for mental health professionals used in a really intentional way to help process the event uh for residents kind of similar to the mental health um just helps them understand the fire behavior as a whole rather than just their experience contextualizing things okay so you get in the sense now um with these visualizations we were getting a little bit deeper into these uh conversations than perhaps we would have done with um just the conversation alone when we bought out those visualizations a lot of times people were really starting to go ah that's what was happening over there or my uh my friend on the other side of town told me um this was happening now I see how that happened so the story of the fire's dynamic really started to unfold in these conversations and you're hearing a little bit from all of us about this uncertainty that shows up in models and in data sometimes it's out of date sometimes we know it's not as granular as um maybe it might be in the future but there's also a lot of uncertainty on the human behavior side people are really unpredictable and um can act in ways that maybe they didn't expect during a fire it was a really good article uh on the BBC a few years ago when the Portland Murray fire happened in uh Alberta where they went and interviewed people on what they took in the hurry when they left and people were just grabbing like dinner trays one person put their spaghetti in a tupperware and that's all they took like you don't have much time to think so you just go so you don't always know what you're going to do in those events but what we were able to find out is by talking about that uncertainty and seeing what the story is there some of the factors that might influence where uncertainty takes control and we act more emotionally than maybe logically like um models of how people behave might assume you do so there are a few factors that we found um maybe drove some more emotional responses across these two fires the first was how long has it been since the fire began have we known about it for a little while and we're getting a bit comfortable with that risk being around did it just come out of nowhere we didn't know that fire could happen in that area and it's really evolving rapidly without us having time to process it for ourselves um then do we know where that fire is when it's moving quickly um you heard from Will there was a lot of difficulty identifying where the fire was I think for both of these fires that we studied because of that fast movement the conditions the volatile behavior of the fire um it makes collecting data really challenging and then communicating it really challenging and if we're communicating how much longer is it going to be before that information is out of date which leads us to what kind of information we have and how accurate do we think it is does it come from someone we trust does it come from Joe down the road who heard it on facebook so it must be true where does that information come from and can we rely on it to make safe decisions um and if we only have information from Joe should we act on it just in case uh then what do we think is that risk it can be a very different story um for folks in ranching communities who maybe structures aren't a risk but crops are their main income the fire comes through they want to protect those crops because that's the only way they're going to be able to afford to build their house there are these very varied values and um some of the values in these places are more about the place and the relationship to them these are special places to people and you have generations of family history in these areas connections that go back um the places that are burning up precious to you um there's also that challenge of kind of infrastructural values in uh the east troublesome fire people thought they might be able to go through trail ridge out through the national park that ended up not being the case it got um stuff started burning over there really quickly it wasn't really time to communicate it um so there's kind of that back and forth of how you negotiate all these values you have all these different things you're thinking about in this network of getting information in the system of fire behavior it's a really challenging time to make decisions and feel confident about them so when we bought these models to folks and talked with them about it we actually learned a lot to bring back to our modelers and we worked with scott when we did the east troublesome fire first we got a lot of feedback from folks there on how to improve it so that they can navigate it we went back a few different times on different versions different viewpoints what we thought would be most useful when we came to do the marshal fire study um so a few things highlight that we took away from this that um hopefully modelers can adopt moving forward the first is making it local it's really cool to see high level stuff people want to see the fall fire but they also want to know where their house is on that or for the marshal fire people really want to know where cosco was because that was so central to a lot of people's experience you got stuck in traffic there they came out and saw it for the first time there it told a big piece of the community's story so understanding where that was on the map helped people orient themselves to it really quickly similar story in grand lake people wanted the lakes uh labeled they wanted the key county roads labeled just so they could see where they were at in the maps and they could zoom in and and get a sense of how their story fit in with the bigger fire we also found less is more in some earlier versions of our visualizations we had a lot of arrows or wind barbs a lot of detail because that was really exciting to us as scientists I think that's a lot for someone who's just been slapped down and I've had to look at for the first time so we kind of drew back a little bit we lessened the amount of information we made separate visualizations one that focused on smoke one that focused on the fire one that focused on the wind so that we could really pass out that story in ways that didn't overwhelm people with really cool visualizations then we had this back and forth about how realistic should we make the fires look because that can be really traumatizing in itself to live that experience even if it's just on a tv screen folks in the east troublesome really wanted it to be realistic they they had processed we went a couple years after that fire they wanted to see their story accurately represented whereas they think with the martial fire we were coming in just over a year after this fire and folks were a little hesitant about seeing it too close they could put maps and things like that so I think the timing of when we bring visualizations to people is important too then the last one I want to touch on viewpoints that matter to people where was everyone standing when they saw that fire a lot of people were kind of around the south end of Grand Lake looking over folks from Costco want to see viewpoints from Costco or you had a visualization that looked from Lewisville from Superior from the top of town towards the flat iron so that people could orient themselves we want to give people as many opportunities to connect their story to this visualization so that we can both learn I think we might visit our last two slides if that's okay and the first one we had was how do you plan to access information during a fire it's fantastic we've got a whole range of things here emergency alert system notifications we're seeing maybe a little bit of a story about how reliable they can be so I like seeing that there's a variety you want to probably look in different sources although emergency alert systems are where you're going to find the most useful information you might remember a few years ago in Tennessee there was a fire near Gatlinburg near Dollywood and all of that stuff in that area the fire took down telephone communications really early on so less than 17 percent of people in that community got texts and we can see really similar stories we can plan for these with good intentions the fire holds a narrative when these events are happening so it's great to see that we have a diversity of different places that we're going the second piece is thinking about is it trustworthy if it's from a neighbor or social media is it someone posting who has a background or works for the fire department where do they get that information from don't be shy to ask those questions to figure out the story behind the information you're getting next slide I think the last one is you know where you'd go during your evacuation interesting so we're very split and maybe a show of hands if you said yes was it somewhere you've been before is it a family member is it a location you know for the people who said no is that because you haven't evacuated before I've seen a lot of nods yeah when you go through an experience like this you have an idea now of what's happening if you haven't it's kind of a lot of effort to think about all these hypotheticals and what that would be um for you where's the fire coming from what does that look like so if we head back to the slide here you go back to the few things that we've learned across all of these studies about how to prepare for evacuation and I think this preaching to the choir for some folks in the room good food for thought for others um and hopefully there's something new you can take away from this the big one which it's exciting to see this is where most people are going to go for information sign up for those local emergency alerts so many counties have to rely on those opt-in ones uh the technology where you can kind of draw on the map you want everyone with cell service in that area to get it that can be um temperamental not everyone can afford to use it it's only accessible within certain levels of the government sometimes so you might not always see that so sign up for those but it's good to rely on other sources or your own um decision making as well a really common piece of advice is to pack a go bag if you google online you'll see there's seven p's of evacuation and that's stuff like people pets prescriptions paperwork um most common things you want to grab if you have a lot of that in a bag that you keep near your door that's one less thing to worry about and you can plan on packing your spaghetti and other stuff you can focus on that stuff instead um so that is a lot of effort and um I always shudder when I think about my go bag because it has clothes I never wear in it so I'd be horrified if I was stuck in all these really old clothes but um one thing that I've heard a lot of people do is they take a laundry basket because that is the clothes you wear all the time you can go to laundromat and have something to do while you're waiting for information especially if you have a big family and that fills up quickly that's probably going to be your best um call of action if everyone in your household bothers to put it in there instead of on the floor um then related to that one of those p's is paperwork think about what documents you'd really need to kick start um your recovery or your return after a fire um you might think about grabbing passports and things birth certificates a lot of times at recovery centers if it is a serious fire you'll be able to access a lot of those so as long as you have one form of ID I'd say um that's a great starting point then think about what you're going to need to maybe start insurance claims and a lot of times that evidence that you own your house or that you have an insurance policy so there are a few documents that you can really prioritize email to yourself so that you can kind of get to the front of the line in that um in that kind of condensed timeline of how do you set things back up again if you've lost things then uh we're kind of split in this room on where we go have an idea of where you want to go and then all the different ways you can get there if you're trying to head to Fort Collins and you're going up the main highway what happens if that's closed do you have a route around it or um if the end of your neighborhood is closed you have another way around um we spoke with people who drove their vehicles over Davidson they said I get out during the Marshall fire there's a lot of um urgency to that so think about if you go this way is that going to be you locked in traffic are they going to be other ways kind of thinking through those scenarios in your head I know it's not fun but it can be really helpful and with that I'd say think about your own trigger points so we heard that in the East troublesome fire um there was highway 125 a lot of people felt whoops sorry that if the fire crossed that personally they didn't feel comfortable staying at home anymore so what is it for you in your household when you think about where you live but when it gets too close that's when you start to get uneasy because you should listen to that instinct and act on it don't wait for someone to tell you um and a lot of people that did seem like the Davidson Mason area um and the highway were big trigger points for the Marshall fire the last one is uh talk about this with members of your family I really encourage if you're having dinner with your family or everyone in your household just talk about it for five minutes one day and then it's done you've had that conversation I've done this kind of research across a lot of different communities you'd be surprised by how many couples have completely different ideas of what they do uh the husband often wants to stay it's very gendered men often want to stay women often want to leave in fire don't wait until you get that text to make the discovery that if the other members of your household are thinking about doing something different that's a great thing to talk about um ahead of time and what kind of your own personal comfort level is and when you leave across all of this and in both of our fires we really heard people say at the end of the day it was their intuition to leave they started to get uncomfortable that was their decision to leave early or to leave as the fire was coming um so don't listen uh don't ignore that if you feel uncomfortable that's a great sign that it might be time to leave and you don't need other people to tell you better to be safe than sorry always I think so we just shared a very tiny glimpse of this project there are so many different pieces of it we have components that are very into the evacuation pieces and recommendations we have pieces about recovery of the marshal fire we'd love to talk more with you about those things or if you have an experience with the marshal fire that you want to share we're in town for a couple weeks we'd love to meet with you for coffee and make sure that your story is heard in this one of the strengths we have as scientists is being able to tell your stories on broader platforms so they can create change in other places um please reach out if you're interested in chatting um I think maybe we can go to questions if there's nothing else to add yep okay uh if anyone has a question please raise your hand so the folks online can hear as well you talked about building trust with the community after this was that left to police or the fire people or like our county commissioners that's a good question who's responsible for building trust I think it's the same as who's responsible for risk of fire in the beginning it's a little bit of everyone um but it's going to vary across places so um some places I've been their county has a history of conflict with the community it they need to work on that in other places at the forest service so it depends on the fire and who people feel uh have lost trust in that area so that in a lot of places that ends up being who they'd expect to receive an emergency notification from and when it doesn't come um what's going on there a few things I think that professionals can do to regain that trust is talk really frankly about why that didn't happen I think it's really easy to sweep under the rug oh uh we can't talk about that or whatever but if you have that open conversation about the system we pay for didn't work or um we thought we were going to have more time or maybe we just need more planning and we weren't in a place where we plan for this fire um I think that transparency can go a long way but it definitely decreases trust a little before we can go into that so when you hear authorities say they make mistakes hopefully that's the beginning of a conversation where you can press them to um to change their activities does that answer your question that's a good question and I would I know that there's a group of residents who are coming together to talk about evacuation um and I think that kind of community pressure is really valuable um because change doesn't happen without people asking advocating and keeping it fresh um it sounds like it varies from my read on different authorities like there's different level of trust in Lewisville versus superior governments so I think I I don't know the full back stories but a lot of times these histories um have been around a history of conflict for a long time and it takes a fire to really bring that conflict to the surface so I'd say for some of the locations affected by the Marshall fire their fire was the most recent catalyst they need to talk about the full story and have that out in the open we went to um there was a round table at the Lewisville recreation center or community center where they had different governors there to answer questions I think that's a good start but it can't just be a drop in for an hour on TV and that's not a criticism appreciate that they did that um but it still leaves a lot on the locals to fix and then there's that challenge of what's feasible in policy versus um in what do we think it's feasible versus what's realistic policy and some of these things that we no need to change take a long time having these conversations igniting these discussions about what needs to change can repair some of that trust and then you're unified moving forward to try and affect change at those larger levels I think so really varies on their place but I think the pieces are starting to come together from what I see in this area to really have those open conversations and I think it'll get easier over time too thank you my question is regarding fire mitigation I live right down the road here so I was evacuated during the NCAR fire last year I'm also on the board of directors for our homeowners association and we are dealing with fire mitigation right now two questions that came up in my mind right away were you had mentioned something about wooden fences around the Marshall fire we're dealing with that right now our community is I think it was built in the early 1970s and we have wood that is falling down wooden fences it's very expensive to replace it and my my question my first question is what do you think about that my second question is one that came up in our meeting last month and it regards charcoal grills if that is dangerous or we've been asked and right now we can't find anything that says that you shouldn't allow charcoal grills but we did a walkthrough with two fire mitigation people right after we were evacuated and they recommended that you tell the community not to have charcoal grills I wrote that down in my notes last month the question came up because we have a resident who said that he can't find anything online that talks about it and so I can't find anything either so my second question is what do you think about charcoal grills for fire mitigation two great questions the first question was about the pence replacement the cost and the challenges around that I think there's a real tension with a lot especially in HOAs about how far you want to push your residents versus involve that and like have them come to the table and a lot of times that depends on what funding is available so I'd say if you can look into like firewise does have grants if you're going the firewise community route there can be funding therefore things like that you could go the code route and require it I know that would not make you popular as the HOA person but there's kind of this window opportunity after fires where people are willing to negotiate and I think you're you're towards the end of that now often it's only six months after a fire it really is short and then we're moving on thinking about the next fire season so I think thinking about using those windows of opportunity to engage those questions you're absolutely doing the right thing by bringing people in and having that conversation that's fantastic to hear so I think looking for sources to cost share that if that's private events having those conversations or talking about would it affect insurance is that going to make a lot of times financial conversations they won't at all then it may just come down to a level of comfort and could you prioritize households that maybe are the wicks with fences and start with them because if they're the starting points you get those votes on board ah and the fires that we have with NCAR have supported my house I'm an advocate for I know it becomes a lot but yeah but trying to get that you know the whole community yeah if it was my own private home I could do whatever I want yeah but you know oh yeah I it's a difficult kind of either or I wonder if you could say okay it's either you replace the fence or it's a conversation about removing everything away from the fence and clearing that space in which one would you rather do I wonder if there's that and the question about charcoal grills that's interesting I didn't know there wasn't any documentation about that and I wonder if you could push for okay if there's no documentation around that then the area you have a charcoal grill on needs to be fire-wise it needs to be brick you can't be having it on grass and you kind of set those parameters instead that's a workaround um that is interesting and I wonder if because it's an enclosed flame technically that's how it hasn't been the focus of more study yeah they had studies that said that uh propane gas propane is just as equal um as charcoal grills and they're not forbidding propane tanks so if you can't if you can do propane why can't you do charcoal grills yes but to me it just it's not a smart decision but you know yeah I don't I don't make those decisions forgive the lighting it is on a timer and they will get it turned back on in just a second or not moving enough here um oh I am going to go ahead and ask a question from our online audience um just to make sure that we are getting their thoughts in the room as well I'm going to bring up if you could bring Slido please and we can look at Catherine's question yeah so how often are the models that are used by response groups like fire EMS and city planners um during well how I think she meant how often are they refreshed I want to say that um if Catherine is still there um there's a verb missing again I'm just asking um are they being used yeah how often are they being used that's better yeah that's a great great question um you know obviously in research world we like to think that they're being used frequently but of course that doesn't happen without close collaboration with these folks um there were recent efforts by the state of Colorado to utilize some of our fuel models fire models for operational purposes and we're hoping that in the future there of course will be more funding to extend those those types of efforts I think there's a lot going on in the research side of things to try to um improve the the speed up of these models right because oftentimes they can be fairly expensive if you want to run them for a larger area so I would say in general at the moment they're certainly not being as used as much as they should be oftentimes what happens is a fire starts and there will be an incident meteorologist who is assigned to that fire there they will have a group of people with them and kind of on the scene they'll just see what the conditions are like what the winds are like um and use really fast simple models um which don't take into account these more complex interactions oftentimes they do a decent job of giving uh you know a broader picture of what's happening but at fine local scales uh it's really missing key things and so yeah incorporating the models that we're using that are uh that are coupled together are critical and um I yeah I think moving forward yeah we need to certainly use them more I will say that we had uh run these simulations previously in the cloud so now you know you have Amazon cloud um we can actually get them to run in the cloud at this point which will help um with cost and and speed so yeah there's a lot of work to be done uh so yeah a few thoughts certainly more than happy to engage there let's move into an end the room question um so on the Marshall fire there was um a hotel that was on I guess the uh west side of us 36 and um I remember talking with people and thinking that went up in flames and it supplied a lot of ash and debris that got blown over onto the rest of Lewisville and what model would that be your model or maybe these smaller models that you were talking about that could model an impact of that kind of structure located on the western side of a town that gets 100 mile an hour winds going in that direction like um could that be used to help with city planning or kind of making sure that doesn't happen again yeah so you're you're thinking more so you know after the fact you know looking at hypothetical situations to see how you know certain wind patterns and certain structures how they're burning would affect you know parts of the town downwind right so yeah that's certainly possible um the model that we use doesn't get down to kind of like structure level combustion uh chemistry and physics there are models that do that they're extremely expensive and they're typically used to look at you know individual buildings or clusters of buildings so trying to bridge that gap you know between those types of models and the models that we're using I think is an important piece moving forward another thing is our model currently only accounts for wildland fire fuels so you know grasses shrubs trees the like and it doesn't include actually urban areas and buildings and so that's one area of active research that we're currently pursuing because you know as you know that's critical and and once that's in the in the place you know the atmospheric model is already kind of ready to to take that information in so if a structure were to burn then we can see you know how potentially debris or smoke would be transported so yeah it's not quite there yet but there's there's definitely efforts going on right now yeah and I think if you look and I'm not going to remember the acronym but it's something like IBHS there's a institute that does research for insurance companies about how structures burn and things like that they're based in Maryland and folks from that organization came out to the Marshall fire afterwards and wrote a report about their thoughts that's why we know some of the fences were a huge risk or a concern a driver perhaps but I think if you look at their kind of approach that's where we're going and it's it's a really big shame that we don't have a accurate read on how multiple buildings burn and interact with each other yet I think that's a really important area of growth that we want to see more modeling in because we only see half of the story when we can't model those urban areas so you're highlighting a really key need right there as well going to move on to a slider question it's going to be a question there's an add-on to it um so if we could please bring on Kaylee um they asked how often did you interview those with homes affected that were second homes um versus primary homes how did their attitudes or answers differ and we have a follow-up also from Karen who said regarding that I highlighted it regarding Kaylee's point second home owner thoughts may vary in circumstance was a renter a friend or family that the owner planned to make it a main home in the future so those are a conglomerate of questions for y'all those are great questions so I would say um in the east troublesome fire maybe a third of the people we spoke with were second homeowners um and we really focus on getting their perspective for the very reason that they are often different they have somewhere else they can go when they evacuate they have a home that they can live in if that one is lost um those opinions do vary but in a lot of studies we've done also in Arizona um a lot of people feel like their second home is the place they really want to be living in their main areas like some uh I'll say for Arizona it's Phoenix and no one talks positive about Phoenix I don't know if that's the case for Denver but um they they're much more excited about their second home and they care about that location and a lot of times they are thinking about a long-term vision where they're putting money into that property because it's their retirement home and they see a future there so I would say while their thoughts and their views of risk um are often different they're not realizing that there is fire risk there or they're not yet willing to invest money in those mitigations they are concerned and aware of what they're losing and when it came to evacuation um not as many were there and we actually met with a few people who have ringed doorbell cameras and they were so concerned about their house they log into that and watched the impacts of their property on their phone from hundreds of miles away or were able to figure out if it was standing or not when electricity came back on so we see those crowds relying a lot more on um that kind of technology to figure out if the house is okay when they're away for the Marshall fire I don't think we met with this many second home just a couple um and they were in both of these places I think a lot of people who have second homes but they're Airbnb's so it's a source of income um that maybe they're saving for something or they can't afford to have both houses without doing that so um one thing we did hear a lot about in the Marshall fire was um if you had multiple properties affected in the fire it's really hard to claim insurance on multiple properties so one person had at a public meeting we went to said he had six rental units um he wasn't getting at he was actually going into the red and he wasn't going to make any of that back so investments are really difficult to hold on to so it's not all sunshine and rainbows when you have multiple homes there's a much greater risk that you're going to lose something just statistically um but it's safe for these studies um it was they they cared the same about the place even if their story and their risk around that place wasn't the same questions in the middle of the room um yeah no I was signaling you Aldi is going to come with a mic but my hook is on for a reason first of all thank you for the presentation this has been really great I think fire is near and dear to the heart of many people in this area at this point I had a rather specific question and that is I have heard that ember casting is a major threat to structures well ahead of the fire front is that model data field available and if so have you tried to visualize it so we do now have ember spotting in the model recently Maria frediani who's a colleague with us implemented that into the model um I don't know did we I think we added the outputs to the to the vapor visualization or okay yeah yeah the particle data yeah so I guess we haven't yet visualized it um coming soon yet there's certainly you know a lot of interest in that area because as you say that's that's the reason these these fires tend to spread rapidly um it happened of course in the east troublesome fire and then in the marshal fire that's how it jumped the highway and was able to get into to louisville um so that's another area where there's a lot of research going on right now because there's so much randomness and uncertainty associated with you know how these embers are initially produced and how they're transported and where they land and what fuels they land on and all of this so um yeah that's that's something that you know we're actively looking at and uh yeah it's a very important consideration and correctly if I'm wrong but um some of the ember behavior we see models really can't wrap their computer minds around so when we did the east troublesome one I believe when it jumped the constant continental divide you guys have to go and manually and start it on the other side because it the way the model thinks it there's no way that it could jump that right in the in the marshal fire that was oh oh and in the marshal fire as well yeah right so yeah in the east troublesome fire you're right you jump the continental divide um to the east side and the model so at the moment what we have in the model it's it's not starting new fires because you know we hope that we get the ember spotting right if we don't and it starts new fires and it can just kind of get out of hand quickly so so that's a whole separate thing um but in the east troublesome fire it was able to uh actually spot over the divide a little bit in the marshal fire it didn't part of that is just because of the fuels right the urban uh component of the fuels is an important piece that that is kind of missing at this point so yeah that was that was a kind of complicated uh simulation uh that the model really isn't intended to do but we're we're kind of expanding it now to to account for these types of things yeah thank you we have a slider question also asking about models so you might as well keep the mic where can we find the model for the marshal fire and does the model include both initiation points oh yeah so uh actually we did just recently publish uh our our modeling results on the marshal fire um and i'm yeah happy to share that uh after the fact and and in the paper we have the data sets um that that were used to generate the outputs and so um as uh scott said we have a lot of outputs in the order of terabytes so um you know we're happy to share those with you we just can't upload all of those to our repository it's just not possible um but yes we have the information about kind of the model inputs and and where the fire was uh ignited and what time and all of that so yeah if you'd like to reach out more than happy to uh to discuss that further thanks we'll take one more question from the room i guess i do hi um my name is emma i actually work over at the car center for science education so i'm just in a neighboring program um thank you for the presentation i tim one of the examples that you brought up in your slides um one of the models you had talked about how it didn't account for suppression efforts in that one is that anything that's possible with modeling or visualization or is that something that you're looking into i imagine it's probably complicated because people who are on the ground are i don't know adapting every second to the situations but i'm curious yeah suppression is one thing that we haven't started to tackle at this point um i think as you say it's it's complicated for a number of reasons um i think the biggest challenge is that the folks who are on the ground are of course most concerned with you know saving life and property which is what they should be concerned with so in order to get you know that into the model we need kind of real time feedback on where the fire is being suppressed how it's being suppressed so until i think there's more efforts uh like on the ground to you know have a job that specifically doing those types of things it's it can be i think pretty hard to get in the model but i think yeah going back and like historically looking at simulations to see how suppression would affect the the model simulation is something that we can do we haven't really gone that route too much some people are looking at that at this point but yeah that's that's another important aspect so you guys are pointing out all of the the parts of the model that we're working on so thanks i'm gonna take the last question um it's my question um fine it's fine it's not about fires um but is that if there are any students watching this today and they want to be as great as all of you are what advice do you have for them the hard-hitting question here i know i know i'll start that um i think don't be shy to send an email like my start in research was just sending an email to someone i didn't know don't be afraid to reach out and then don't be afraid to look stupid the one time you get away with that for free is when you're a student and you can use that to your advantage to learn things that maybe just asking the simplest questions can reveal new information so i think when you're a student you're in a really unique place to leverage um assumptions that people have about you to get new information that people further wrong in their career could not um so don't be shy to try something yeah i think piggy backing off of that um try a lot of different things when you're starting out again when you're a student that's your best chance to like make mistakes and try things out obviously down the road you want to specialize in a few things and be really good at it but um i think i've found that i have a lot of interest and when i pursue those it can open a lot of interesting doors and i think that's how i ended up in this position working with convergence research i guess i'll add to this um i would say one thing that's really nice about today's world is there's a lot of you know virtual seminars virtual talks virtual things so granted it's not the same as you know being in person and interacting with people certainly take advantage of those things you know oftentimes in listservs they'll they'll put out these these events and even if it's not in your uh i think area of expertise are like what you're most interested in it's good to branch out and learn some other topics and and understand kind of what else is going on it's you know there's there's a lot of crossover now between disciplines and so not being too narrow i think uh is important and really trying to trying to broaden your your skill set as well i guess i would say depending on what level you're at in your education if you're before college i would say really pay attention to what you're learning in math uh mathematics i think we all have some mathematical background statistically no well maybe tim and i do um but i thought i'm sure i'm sure you guys have a background in that but i don't know i would say cling on to the mathematics and and science um because it's applicable i think in most if not all fields well thank you um and with that i'm just going to thank you all again um and let's thank her speakers first for that yeah they managed to make this an amazing talk and i'm so happy i got to be here for this um thank you all again for attending this great panel on evacuations as part of our explorer series um we have a great talk coming in august 30th about total solar eclipses so if you want to learn about how total darkness is the answer to seeing the sun this is it it's going to be here on august 30th um if you're interested in more n car explorer series events definitely check out our website for lectures and conversations and if you want to see this again that recording is going to live there um now for the survey part if you are 18 years or older please take a moment to fill out our three to five anonymous survey that is going to help us understand the impact of the program and how we can improve for next event the survey will close on monday august 7th um you can find the survey by scanning that qr code you can also ask the staff member either aliyah or alexandra um if you would like some help to take the survey using one of our tablet computers i really hope to see all of y'all next time i have memorized all your faces um and have a great rest of your evening thank you so much for being here