 one of the discussions that the committee had was whether or not the word extreme should be used in this context and we agreed as a committee ultimately on the word atypical. Nevertheless, we expect that there will be more extremes and we will be talking about these kinds of things today, talking about communicating about these kinds of things, about the terminology that we use and about the specific technologies and advances in technologies that we can use to communicate about tropical cyclones going forward when we do expect changes and the intensity and frequency of them. At least the intensity. We are going to start today with a recap from Rebecca Morse but I'd like to remind you first of what our charge and the committee is. Next please. Thank you and here's the workshop planning team. You can see that we've got Derricka Carol Smith, Brad Coleman from American Meteorological Society, Craig Fugate from Craig Fugate Consulting LLC, Mike Lindell from Texas A&M University and also associated with the University of Washington now, Andrea Schumacher at the NSF National Center for Atmospheric Research, Marshall Shepherd from the University of Georgia and Jeanette Sutton from the University at Albany and again most importantly the National Academy staff who have been absolutely fantastic, Hugh Walpole, John Benzuelo, Rachel Silver, Rachel Sanchez, Rita Gaskins, Rob Greenway and Eric Edkin. We will be starting with a wrap up of yesterday from Rebecca Morse and then that will be followed immediately by a session on practical translation of risk in the public arena which will take us into some different discussions than we had yesterday and with that I would like to pass it along to Rebecca. Is she ready? Can you hear me okay? Yes. I see myself. Thanks. I saw myself on the screen. Thanks Anne. Okay so before I start I wanted to say that I'm currently working as a rotator at the National Science Foundation but everything I'm going to talk about today is not the opinion of NSF. It's based on my role as a researcher at NCAR in my permanent job. So nothing I say should be construed to be opinion of NSF or the federal government. So I'm really excited to see Bask doing this. It's a really exciting workshop and I'm really glad to be able to participate. I'm sorry to be there with you in person today. We had amazing perspectives yesterday, really great ideas from all different kinds of perspectives and we also I thought we had really good overviews at the end of each section where people talked about some of the key themes from each section. So what I'm going to try to do today is synthesize a few themes that I heard across the sessions. There won't be any answers to all the problems that we talked about but we have more time to discuss those today and think about next steps and after I synthesized my thoughts I looked at the agenda for today and some of the things I'm going to talk about are going to be talked about more today so that's convenient. So first one theme that I heard was that every event is different. The event itself, the meteorology of the hurricane, the place it hits and the people that it affects. So we heard a lot about people. We also heard about the built environment including interdependent infrastructure and the importance of things like electricity, power, transportation, communication and how those affect other services and there has been research at interdependent infrastructure that talks about the importance especially of things like electricity because a lot of things don't function if that doesn't function. So that was an important theme and also that we need data not only about the meteorology but also about how intersects with the built environment and with the people and tools to be able to assess for example things like how the rain will intersect with the built environment to lead to flooding. And so we are getting more and more data about the built environment and that's an important area but the area that we really have the least data about that was talked about by the researchers were how people respond to these events and how they're thinking about these events especially data in real time that follows people over time just like we do with the meteorology and the built environment. Another theme that I heard was that how we talk about risk matters for how we manage them and how we characterize risk and so this is kind of the way I think about this is kind of in terms of problem definition so how we define the problem influences how we solve it and and really teed this up nicely that I heard a variety of different kinds of definitions from extreme event definition talking about these events as extremes all the way to the other side which is really the goal is to reduce deaths and that's what some of the emergency managers talked about really focusing on like reducing deaths is what it's about and so how you define the problem and approach it will affect what you think about as the solutions and then I started thinking about well does the extreme part matter or is it really that it's putting people at risk is that really the part that is important and so we need to clarify or characterize the different aspects of this problem that we're talking about to understand it but we can't depend on that because the problem is always bigger than we can really characterize when we divide things up and so it's important to keep bringing in new perspectives and looking up from the details and seeing the big picture and I think that yesterday was an excellent example of that of really getting all kinds of different perspectives on the issue so that we can really kind of make sure that we're not just focusing in on one part of the problem the other part in addition to the built environment was really the people and we heard about people talking about saying how can I see things from their perspective from the perspective of the people that are using the information and the people are experts in their own jobs or their own lives and it's not that meteorology is the expert and coming in to sort of give people information it's really important to kind of flip your thinking and that led me to think about what is really success for the public and for other people and there was conversation about success being do people understand the information do they get the information do they do something with the information do they take a certain action and so what does understanding mean and maybe that our definition of understanding may not matter for them it really may be about something else so another theme was on complex and cascading hazards and how in today's world there are multiple hazards both in within a hurricane and the hurricane interacting with other aspects of people's lives especially a more vulnerable population for some populations aspects of their everyday lives are hazards they need to feed their family they need to protect what little they have they might live in environments where there's pollution and other concerns and when a hurricane comes they need to worry about saving their livelihood their health concerns all those kinds of things and so really thinking about the multi-hazard in a more in a broader context about how those intersect with other hazards that was something people talked about and then because of that the right time to evacuate might not be the same for everyone so we heard about uncertainty and how that intersects with decision-making and for people who have various from for some people it might make sense to evacuate soon because they have a relative they can go to or they have the resources or they don't have a job they can work you know remotely or they don't need to work or those kinds of things whereas for other people there really might need to wait until there's more certainty because the risks of other things like losing their job not being able to feed their family not having the transportation being in a shelter might actually be risks that are as important or more important and so really with this theme of how seeing things from other people's perspective or from the perspective of the people that are using the information thinking about how do hazards intersect with people's lives and center around their lives and their decisions and not the hazard the hazard is just one part of kind of a bigger bigger system so another theme we heard was about talking in language people can understand so people can personalize and understand what the risk means to them and so for different people this is going to might mean different things you might need to communicate differently both in different ways using different kinds of language or using different kinds of metaphors or experiences or other things as well as through different sources of information different trusted sources we also heard some people talking about whether people at risk whether they understand that they're quote-unquote vulnerable or at risk and how they're vulnerable so what the risk means to them and also do they understand that what is recommended can help and how that can help so we heard about helping people understand for example what storm surge can do based on you know using the height of a um of a um you know a stick also people understanding that um the action that's being recommended evacuation or something else how that can help and why that's important to do again keeping in mind that people have competing demands on their time and for what they need to pay attention to and and what they think is risky so one theme that we heard across many of the sessions yesterday was about jargon and language another was about the hazard and multiple kinds of hazards another theme was about the location and kind of the spatial variability and uncertainty and of course uncertainty was a theme we heard um across the sessions another theme I think that we heard that wasn't emphasized as much as time um one person I think it was Rebecca Moulton said even waiting is an action that's indicative of kind of the importance of time and a few people mentioned her came in and when thinking back on that I was wondering if one of the issues was that as the storm made a turn to the right it also made landfall earlier and so along with kind of getting a scenario that was worse than some people expected in the areas that were most affected they also lost time for their decisions and so this idea of time is important and that intersects with what I'm going to talk about next which is different audiences there's all kinds of different audiences the two main ones we heard about were professionals emergency managers and other partners of the weather service and members of the public so with the professionals and the emergency managers we heard about again time the importance of taking early action um and needing to take actions make sort of make decisions that will enable future decisions and how confidence and forecast help that and also confidence in what is happening um one theme that that didn't come up but I think might be important is the theme of surprise and understanding what could happen when I read stories about people who are affected by tropical cyclones often we heard this yesterday people say I didn't think it could happen to me so that kind of surprise both for emergency managers and for members of the public kind of challenges makes challenges for decision-making and someone showed examples I think it was Tom showed examples of the different fires and when the warning systems weren't alerted and when I read about those examples for a lot of them it really it seems like the officials were surprised they didn't not not um not sound the warning because they didn't know it they I mean they didn't not sound the warning because they didn't think it was important they didn't sound the warning because they were surprised and they were behind what was happening and they couldn't catch up and kind of understand what was happening soon enough to know what to do about it and who to alert at least with some of those stories so I think this idea of time and giving people more time to make decisions to make the predecisions that set up the decisions as well as the trust we heard an important element of trust with not only the forecast but also the forecaster I remember one person we interviewed said about forecasters they said you believe in the forecast and I believe in you so really that personal trust and we also heard a little bit about artificial intelligence coming along and so kind of thinking about what is the role of the human along with artificial intelligence if for some people the trust is really about the person as much as it is about the piece of information so then going to the public we heard about of course there are all kinds of members of the public but we heard about different kinds of members of the public people that are interested so people who are really excited when they you know here it's a category for hurricane or kind of the eye candy that gets people's attention a big satellite picture the vitals of the storm as well as people are at risk who are at risk and kind of different levels of people at risk so when thinking about this I was thinking about how the people who are interested in the storm because it's you know they're seeing seeing this happening they're really an educational opportunity and so we heard from some of the speakers about how people have a benchmark storm or their experience that they're basing their kind of decisions on and so these interested people could be an opportunity to kind of reframe their thinking for times when they might be affected as well as for people they might be communicating with in areas at risk and then also we heard about this in in the researchers conversations and others is that for people who are at risk at some point they flip from this is happening this is a threat to this is happening to me and I need to do something about it and people do this at different times but maybe these people that are interested so we're excited that it's a category five storm in the in the you know tropics even when it's far away those are the people who are later going to be the people at risk so kind of thinking of these audiences intersecting might be helpful in terms of thinking about how to communicate with these different groups and also kind of the importance of the time in terms of people needing to have time to process the decisions and flip into that risk management so this is important for both of those all of those kinds of audiences the members of the public and the professionals we heard about this to give people time to process to get the information to make the decisions to kind of we heard that people don't evacuate right away they need time to prepare the emergency managers can't just you know evacuate people on buses they need to call on the buses and all those things so really this element of time is important to get people paying attention and then kind of assess the risk and then slowly kind of ease into what they can do as they can do it related to this trust we also heard about the important I think someone faces laying track with the partners that's also important for members of the public so to lay track and build that trust during an event as the event is evolving as well as between events so that you have that trust and those relationships to rely on when the event happens and then one last theme I heard about was kind of there have been so many advances in the last 10 20 50 years that have made things so much better in some ways and yet there are issues that remain when I think about you know when I first started doing this kind of work a number of years ago I won't say how many really storm surge was a big issue and the weather community then weather service the hurricane center broadcast meteorologists emergency managers has been huge huge improvements in how we predict storm surge and how storm surge is communicated and yet there are still issues and so tackling the problem is a moving target because things change rapidly and people are complex and the situations are very complex and so you know some of the issues we heard about there are the key issues for today for example the cone how can we communicate more hazards with the cone or the issue about surge people understand above ground level above ground level was actually an improvement that was made I don't know 10 or so years ago to help people understand it but yet there's still some issues so once you address those there'll probably be another issue waiting behind that but that's okay because that's the way it is so we're still improving things and related to this when I heard about people talking about the different kinds of hazards the floods and heats and earthquakes and fires and other kinds of hazards I was wondering if different parts of tropical cyclones are more related to the different hazards so kind of the earlier lead time with with a hurricane and the aftermath is maybe more like heat where people have to do you know it's a public health issue in the aftermath and beforehand people have to make these decisions about you know where to go and where to protect themselves whereas when you kind of get to the last minute where someone hasn't evacuated and they find themselves you know facing a flash flood in a tornado at the same time or they're facing storm surge it really might be like that earthquake where you only have a few seconds and you've got to do something and you can only have one thing and so what is that one thing that gets communicated so that people have that when that situation arises what they can do and so people are going to make decisions at different times so when some people are going to evacuate as soon as they see a hurricane some people are going to evacuate with an evacuation order and so you want to get those out who can get out and want to get out and help everyone who wants to get out and then there might always be people left and so those people who are left at different times what is it that they need to know to make the best decisions in those situations go to a shelter of last resort you know what do they do when the water is rising you know those kinds of things so to communicate across the lifetime of a hazard and those might be different things and then one last thing is about how getting back to the theme of extremes that sort of by definition extremes or these kind of rarity typical events surpass what has happened before in a location so people aren't going to have that experience usually they are going to be surprised in some ways it is going to be challenging because sort of by definition it's not something they've seen before and the environment might interact in different ways and so as one of the speakers said you know we don't know what's going to happen when more rain than has ever fallen before falls in an area um but really this is this idea of the benchmark storm and borrowing from other experiences can be helpful so the idea that other things have happened to other places and kind of how do you bring that in to deal with these extremes but knowing always that new things are going to happen and we'll learn from those and hopefully can do better in the future so those are some of the themes I heard today and I hope those are useful and I'm really looking forward to hearing more of the discussions today. Thank you Rebecca we are going to transition to the next panel. Okay so that was awesome Rebecca thank you so much for your breakdown so many good points so we're going to start with panel 5.1 which is the session name here is called the practical translation of risk in the public arena we're going to break this down into different sectors first we're going to start with more public sector translation of risk and then we're going to move more into some really interesting walkthroughs of new technologies from public and private sector so our first panel is going to focus on the following session goals the first of which is to examine current and emerging methodologies for communicating risk uncertainty information in the public arena and also to identify how new risk communication technologies and approaches are being evaluated for effectiveness and communicating risk and motivating behavior in the public arena so our first panel we have three speakers all are virtual I believe and we are going to be talking to Dr. Mike Brennan from the National Hurricane Center Dr. Castle Williamsburg from National Weather Service and Gina Esco from the NOAA Weather Program Office and we're going to start here with Mike Brennan from National Hurricane Center okay thanks thanks Andrea good morning and great to be with you all I think I need to be able to share my screen I was planning to show my slides but it says it's disabled at the moment unless someone else can bring up the slide deck you should be all set now oh perfect great so let me share screen two here see if this works give me just a minute okay are you able to see slides looks good okay great thanks so yeah this is a great topic and you know Rebekah covered so many of the challenges that we you know currently face and have faced for many years and conveying risk in tropical cyclone events and what I want to touch on is sort of again as you mentioned some sort of the public sector risk conveyance and really from the national level that we operate from here at the National Hurricane Center and I'll cover some of that today and look forward to what Gina and Castle and Gina follow up with with sort of emerging trends and what the social science is telling us so you know at the National Hurricane Center we're really following the storm throughout its entire lifespan so we're you know basically messaging large-scale hazards and risk to multiple populations in multiple places in various stages of the storm at the same time so it can be really challenging you can imagine for example think of Hurricane Irma where we have a storm affecting the Leeward Islands and Puerto Rico and the U.S. Virgin Islands in the in a sort of immediate sense within the first 12 to 24 hours and then you have downstream impacts in the two to three-day timeframe in Hispaniola and Cuba and then four to five-day impacts in Florida and the Bahamas so we're trying to convey that risk of the various hazards that vary in time and space and severity across multiple locations in various degrees of risk but some of the tools that we have you know looking at the hurricane center specifically to convey risk certainly just starts with the fact that hey is there going to be a tropical cyclone we issue probabilistic forecast of tropical cyclone formation out to seven days into the future what you see here on the right here is an example of our seven-day tropical weather outlook I think this was the busiest one we put out last season in mid-august where we had five different systems we were watching across the Atlantic basin in various stages of formation then we can move on to having the actual storm forecast itself you know the forecast of the track the intensity the size of the storm you know the cone graphic this is useful in in conveying risks because you can certainly highlight areas they're at risk of seeing some sort of impacts from the storm but the cone is only telling you where the center might go and the other forecast information in terms of the official forecast is deterministic and doesn't really have any uncertainty information built into it now moving into that uncertainty space we have a variety of probabilistic hazard based products that are focused on the individual hazards that are associated with the tropical cyclone storm surge wind flooding rainfall tornadoes and there are a variety of products that we at NHC and elsewhere in the weather service have developed to convey and depict the risk of those hazards in terms of things like wind speed probabilities the excessive rainfall outlook which looks at the risk of flash flooding from rainfall the you know storm surge watch warning or the inundation graphic that show a reasonable worst-case scenario for storm surge inundation at a given location and then we also have our watches and warnings which are sort of based on they're based on a combination of deterministic forecast information probabilistic products that I just talked about and the uncertainty assessment from the forecaster and obviously in collaboration with other national weather service offices and the watches and warnings are risk communication tools that convey the risk of a particular hazard at a given location whether it's possible whether there's a danger of that and there's also a time window associated with the risk of that hazard in terms of you know providing lead time particularly in the tropical cyclone program lead time is very important and is tied to the onset of some sort of hazardous condition in terms of either tropical storm force winds or the rise of water associated with storm surge and we attempt to provide you know 36 to 48 hours of lead time at least in terms of watches and warnings to provide people some lead time to take up protective actions and we obviously know that in many communities those protective actions including evacuations have to start well in advance of the watch warning lead time and then we have messaging tools themselves we have products that we've developed over the years like key messages that we'll talk about a little more detail here as we go through sort of the life cycle of a tropical cyclone we have our discussion products that the forecaster can convey uncertainty information and information about risk and how the risk is changing as the forecast evolves we provide impact based decision support briefings here here at the national hurricane center for the FEMA federal agency states FEMA regions and even occasionally down to the local level we obviously have our media pool where we try to get out to the general public reinforce information that is coming from state and local government officials in terms of reinforcing protective actions that they've been asked to take we try to have an active presence on social media and many many different platforms and we also have started to use more of our own live streams to get out there in advance of when we would start say the formal media pool so that the hurricane center has a voice out there in social media and before you know before before we start the media pool which is typically more focused on the last couple of days before landfall you know one of the challenges that we've talked about with tropical cyclone hazards especially storm surge and hurricane force winds is that they are low probability high consequence events and they're low probability at a particular location even within the watch morning timeframe you know the for example the chances of a particular area getting say six feet of inundation above ground level at an actionable time range in advance of a landfall of a major hurricane might be 10% at a given location given the uncertainty in the track and the intensity and the size of the storm and if you you know a 10% chance of three feet of inundation or more above ground level is sort of the basis for the issuance of a storm surge watch and warning so we're operating in this relatively low probability space similarly so for hurricane force winds when we issue hurricane warnings and even more so for the extreme winds in the inner core of a major hurricane which may only cover a few miles of actual coastline but that risk has to be broadly communicated across a large area given the uncertainty just in the track forecast of a storm you know 36 to 48 hours in advance of landfall and how those watches and warnings are conveyed but it is challenging to get the public to respond to these low probability but high consequence events which tend to live up in this upper left corner of this of this diagram you know by the time we get to the to the time frame where we're very certain or likely knowing that there will be that very extreme storm surge or those very extreme hurricane force wind we may be on the time scale of just a few hours before those level of impacts begin which can be well beyond an actionable time frame for preparations so in terms of how we message through the life cycle of a storm you know we may be still five or more days away and we're starting to message impacts at least potentially this is an example of you know tropical storm Florence later became hurricane Florence in 2018 we introduced key messages for the us east coast beyond five days in advance because we wanted to raise awareness sometimes that's just the first basic step is say hey there's a storm out here it could start to impact the us east coast this weekend we coordinate the introduction of these messages across the weather service to ensure consistency from the national and local to the local level and these key messages really helped set the agenda with talking points that we hope that the media emergency managers the broadcast meteorology community and others can emphasize and repeat so that people are hearing the same high level message now at this stage the emphasis is just on preparedness and very broad areas that can be impacted so right now we're talking from you to and then the entire us east coast was at risk of some impacts we can talk about what the specific impacts might be in terms of hazard storm surge rainfall and wind but we can't get very specific at this point this far out in time but we can also even if we don't have a storm sometimes we're including messaging before we even have a storm that's formed either through the tropical weather outlook or through our ability to issue advisories for what we call potential tropical cyclones for systems that are forming near land when we get to the three to five day time frame we can start to focus in the messaging at least from the nhc level on the areas where we're expecting where we could see the highest risk of impacts or where that risk is starting to increase it's still too uncertain to get very specific about exact timing and magnitude but we can talk about how the risk is changing again for Florence three to five days out we were talking about an increasing risk of two life-threatening impacts storm surge at the coast and a prolonged heavy rainfall freshwater flooding event inland and we were starting to highlight the area from South Carolina into the mid-atlantic states and so we were able to sort of start to narrow that focus this is a time frame when preparations are typically underway especially for higher end impacts or areas that that require a large lead time for evacuation so again in here we're trying to emphasize and reinforce that people follow any advice they're given by their local officials and we can start to use a wording rather severe wording or significant wording in terms of life-threatening if the confidence is high enough at this stage when we get into the watch warning phase we still want to keep the messaging focused on the hazards and not the track the intensity the category of the storm and at this point we start to break the messaging down into individual hazards that are going to vary in time and space and require a different response and we can start to provide more details for this example we're talking about life-threatening storm surge is now likely somewhere within the warning area from South Carolina to Virginia life-threatening fresh water flooding is likely from the Carolinas into the mid-atlantic states and similar messaging for the hurricane force winds being possible or likely within the watch and warning areas respectively again as I mentioned each hazard requires a different response and the other challenge with tropical cyclones is that hazards are going to occur at different locations in time so you can go back to think about hurricane Harvey where we had an extreme wind and storm surge risk at landfall along the mid-texas coast and then a catastrophic flooding risk days later along the upper Texas coast so we're trying to message all of that at the same time and we have again watches and warnings and there's other weather service hazard specific graphics like the hurricane threats and impacts graphics that can sort of start to localize that messaging down to the local scale for the surge rainfall wind hazards impacts and the timing and Rebecca mentioned this before and it's really important this is when we have to start messaging post storm hazards and safety we've lost almost as many people to so-called indirect fatalities after largely after tropical storms and hurricanes as we do from the direct forces of those storms in the US in the last 10 years or so so this is a time when you have to start messaging post storm generator safety post storm cleanup safety heat safety you know depending on what the conditions might be like after the storm you start to have to message this now because it's going to be very difficult to reach people once the storm has begun to affect them and their communications may be much more limited so that's all I have from NHC to start off the discussion and I'll turn it back over to Andrew thanks thanks so much uh next we'll hear from Castle Williamsburg hi good morning just checking you can hear me can hear you great wonderful um and just sharing my screen hopefully that pops up as well yep looks good okay great uh good morning everyone uh my name is Castle Williamsburg I am a contractor supporting the weather program office within NOAA and specifically the social science program um today I'm going to be talking about addressing some challenges when we uh when it comes to translating risk communication research for practitioner use and I'm going to use our hurricane supplemental projects from fiscal year 18 as a way for us to reflect on this process and so if any of you aren't familiar with the hurricane supplemental process essentially after hurricanes Harvey uh Irma and Maria there was this really there's a huge emphasis on modernizing uh the tropical cyclone product suite and in particular trying to find ways to further integrate social behavioral and economic sciences into those products so through the weather act and the 2018 disaster supplemental appropriations there was a huge motivation to kind of further explore what social science might be used in impacting uh these products and services so in order to do that we funded four projects that had a complementary design to offer a body of research rather than these one-off projects um and throughout the projects and toward the end of those projects we started collaborating and iterating with the weather services tropical roadmap team and they are a group of both policy and operational meteorologists and we did this to hopefully make our research findings both more uh more operationally relevant and then lastly a small team of social scientists across the weather program office and within the weather service synthesized the research across these four projects to identify several common themes and then we went back to the weather services tropical roadmap team to constantly continue iterating on some of these ideas to make sure that the major takeaways and findings and future research and development was operationally relevant and could be used in practice so you might be asking well what did we find um so I was really excited to hear all of day one yesterday and with Rebecca's uh kind of recap this morning because I think a lot of these findings were covered but I want to emphasize them a few so these studies in combination with past research really show that probabilistic information does help people make decisions in the face of uncertainty in fact these studies revealed that emergency managers forecasters and other weather professionals believe that the public struggles to interpret and use probabilistic information and as a result they often simplify and dilute their communication of this uncertainty and when they do this simplification it leads to less effective messages so to make uncertainty information more valuable to people and their decision making practitioners should include a short explanation that describes how to interpret the probability information founded forecast products and graphics this will help improve comprehension and interpretation of probability information another emerging theme is the desire for more localized and personalized forecast information many decisions that partners make during tropical cyclones depend on the anticipated impact to their local area therefore partners ask for geographically specific information whenever and wherever possible even when local forecast information is highly uncertain different types of timing information were identified as critical for partner decision making in addition to when hazards are expected to begin partners also were interested in timing information that revolved around expected end time so that they could resume some of their emergency or rescue services and also duration timing to better understand maybe some sustained impacts like wind to infrastructure like bridges at greater than five days of lead time partners ask for more information about tropical cyclone tracks and scenarios forecast models and as Rebecca highlighted forecast confidence um partners suggested changes to NOAA's tropical cyclone products week to optimize the extraction of key information from textual and graphical products to further encourage message transmission we talked about those intermediaries yesterday this was again a relevant theme and lastly partners highlighted the value and importance of those summary products that Mike just mentioned or those key message products and asked for future exploration of ways in which to combine the entire ecosystem of tropical products into a single dashboard or landing page to aid future decision making and so with these findings we also wanted to reflect on our process itself because we kind of encountered several challenges that emerged when trying to translate this risk communication for practitioner use and so we did our best to overcome these challenges but we want to highlight what we did and kind of emphasize where the some of those roadblocks were so the first question was how do we know when we know enough um so how might we determine if and when we have gathered enough information to ensure that findings are ready for use by practitioners and so to kind of accomplish this or combat this challenge we did use this complementary design where instead of having one off projects we had four that had intentional overlap and intentional differences and this allowed us to do that triangulation and to come up with those high level recommendations and findings that I just provided on the last slide the second challenge was making sure that the findings were operationally relevant and so to do this we the project team started with operational challenges by collaborating with Weather Service we did this constant iteration with the Tropical Roadmap team and we also did the translation and co-development where we worked with the meteorologists the policy and operational meteorologists to make sure that some of the findings that we were developing were being co-developed with them so that they would be more beneficial and help accelerate the research to practice process. The third challenge is how do we share these results widely so we feel as though the information that we want to give back to the community to other meteorologists to other emergency managers and decision decision makers we kind of hold that within reports and within journal articles so how do we get that back out to the community so we've been trying to engage in conferences, integrated warning teams, if you haven't heard of our SPARK platform which is the system for public access to research knowledge we're trying to make some of these research findings more open science, more open access for all audiences and we're also developing a forthcoming story map to hopefully kind of get these get this information out to those audiences so they understand what their audiences are thinking about what other emergency managers are thinking about etc and lastly the final challenge is how can we track successful knowledge transfer and benefits so we're finding a lot of value in being able to identify what some of these benefits might be and how some of this knowledge might be used in practice we also want to track it and be able to understand what are the benefits of it and what are some of the changes that are happening because of it so we're thinking about how might we reach practitioners how do we understand how practitioners currently incorporate knowledge into their workplaces what formats might be best for providing practitioners with these research findings and how might we track some of this use in order to see some of those benefits so that's all that I have for us today looking forward to the discussion but I do want to call out a recent article by Jen Henderson that starts to get at this idea about how these practitioners are starting to think about and understand how they use and adopt new knowledge in their workplaces thank you thank you Castle all right so next we're going to hear from Gina Esco thank you all good morning Testron can you hear me as well yep we can hear you okay great thank you and pardon my COVID cold I tried not to overdo it at AMS and clearly that was not successful I really wish I was there with you in person and thank you Castle for sharing the slides I'm going to try to go through this quickly because I have a lot that I'd love to share but I there's just not enough time to share at all so I'm going to try to do this as simply as possible and the point that I really want to address is how do we evaluate our risk communication how do we know if what we're doing is actually making a difference and so that's really the story today and some ideas for how we can move forward and so I call this weathering the tropical storm together and leveraging partnerships academic public private partnerships because to me that's really at the core of how we are going to advance risk communication next slide please we've heard over the last day and a quarter now that risk communication is complex and what I've also heard is that it is inherently multi-sector right it's not one sector that's advancing risk communication but it's really the wonderful power and combination of us all working together the national weather service the national hurricane center broadcast meteorologists additional private sector partnerships emergency managers other partners decision makers and of course our publics and that makes risk communication very challenging to evaluate when there are that many components which also means that we probably need to work together next slide please so the question that we get most often in our program is how do we know we've made a difference or an impact and I think that's really at the core of this session as well how do we know how do we know that the risk communication we've been advancing is making a difference or an impact next slide please this question was asked of us about five years ago in relation to how our physical scientists work in the agency and they used this example and it's totally oversimplified so I'm sorry if anyone in the room is a numerical weather prediction modeler who's cringing looking at the slide terribly oversimplified but this is what they said they were to Gina you know when we write a new model algorithm we hit it against past weather data and we see if our new model algorithm is better than what we currently have and if so hashtag success we have improvement and therefore we know that we should advance that for our forecast capabilities and essentially they said to us where is your measure of success how come you don't have that next slide please so we held a workshop this was in 2019 the report came out in 2020 and what the workshop revealed and I should say these conversations were the lead up to the workshop and really transformed the agenda of this event and basically what we came down to is we found that without a similar infrastructure for social science compared to our physical science counterparts NOAA cannot measure mission critical factors such as performance metrics impact and change that's of any hazard in this case we're just using the example here of tropical next slide please so every time you hit a challenge you have to turn that challenge into an opportunity and what we've done is take that and describe this as an opportunity for NOAA to advance such an infrastructure and we also can't do this alone which includes prioritizing things like collecting baseline and longitudinal data encouraging data archival so that we can have access to each other's data and of course creating things like metadata so that we know what's in those types of data sets next slide please I'm happy to say that and I'm not going to go into great detail on it because I would rather poke the bear a little bit here with ideas for how we could advance evaluation we are very very fortunate that through the bipartisan infrastructure law we did receive funds to begin to build out an infrastructure and we call it the societal data insights initiative it is a one NOAA initiative that will take a first step baby steps it's going to take a while to develop a social science data infrastructure that will significantly enhance our evaluation capabilities and integrate social and meteorological data together which I think is at the core of what we're trying to ask today how do we know our risk communication is making a difference next slide please and in order to do this I'm here to say we can't do this alone we need all of you and I'd argue that we probably need even more than those who are listening so if you are listening I'm telling you we need you and we're going to need a lot of partnerships and I'd like to give a couple examples of how next slide please one I've heard a lot just in the last day and a half but honestly over the last 10 days over AMS as well about challenges with gathering large samples right really getting generalizable samples really good samples are very expensive people are trying to get convenient based samples that are possibly biased even some of the more expensive samples also we've found have some bias they've tended to be more white more rich more wealthy right more educated and that's not giving us necessarily a really good perspective on the publics that we serve and so one of the questions that I have I don't know that this will be any better but I'd like to try is for event based research or longitudinal research if everyone is holding a phone almost every phone has a weather app that is built in might we partner with the companies that have those weather apps to embed surveys in them and have a public public academic private partnership we can do this through data crates crates are cooperative research and development agreements we wouldn't be asking necessarily for private sector data but perhaps embed the survey into the app so that perhaps we could get a more distributed sample still convenient nonetheless or perhaps larger is there a way to do this could we collaborate more together that's a question that I have that I think could advance the future of evaluation next slide please the next question we have is we know very little about what our publics are doing we assume based on data right based on survey data and qualitative data that the majority of our publics are doing exactly what we're what we're hoping which is that they're taking some kind of action that they are evacuating or they are preparing in some way what we don't know is how or when when we issue products where in that spectrum are they taking action they actually taking action at a seasonal hurricane forecast are they making their preparedness tools that early simply stated I don't know I don't have the data and so the questions here are whether or not again there could be some public private academic partnerships perhaps sharing some data commercial or retail or financial data that might help us elucidate the answers to some of these questions or at least start exploring what is happening in order for us to understand the gaps and how people are responding to risk communication simply stated we need more evaluation capacity and we need varying types of data to help us answer these questions next side please what's more is that we have a lot of questions over the years we have tried to track as best we can with a small team where some of the major gaps are with respect to risk communication research and I'm not going to read all of these I've kind of been on a road show the last few months at a variety of conferences trying to share these because they're really big questions that we do not have simple answers to and what I want to share and I want to poke the bear just a little bit on the top one because I heard some of these reflections yesterday it wasn't necessarily said out loud per se but I started piecing things together research over and over again social science research has it said that the public's and partners make better informed decisions when we communicate uncertainty now it doesn't necessarily express exactly how we should do that it's it's not an answer of serving probabilities on a silver platter but time and time again the research points us in the direction of we need to communicate uncertainty and probabilistic information on the other hand I hear repeatedly from respected partners no they don't understand it it's not useful we can't do that you just got to give them the very simple message so the question here for me is which one is right and the truth is I think they both are and the question for me is an evaluation one of how do we determine how to best bridge research and practice to advance risk communication if risk communication says to do x y and z if the research says that but our partners don't believe it then we're on a crossroads of are we advancing risk communication right and it's not about believing one another but again I think it's a great area of what is the truth in either our samples may be biased which could be right we know we have problems with some of our sample mechanisms at the same time perhaps heuristics are playing a role for our practitioners it's so empathetic to this sometimes the loudest users right that is what we hear and we tend to over generalize some of that feedback and so how do we bridge this and this is a major major area for us in the weather program office to try to understand that space that gap because I think it's a major question if we're going to advance risk communication together we need to find that space of understanding I think we're getting awfully close with the trusted relationships we have but I think we still have more questions than we do answers next slide is simply just more questions that we have again I'm not going to read all of these but just for your awareness these are some of the the really big questions we have the one I'll emphasize here is that the that NOAA and the national weather service are shifting from and this is going to take a while from a one size fits all model of communication to a model of service delivery that accounts for unique needs and that means needing to understand all of those unique needs so that's going to take time and it's going to take partnerships last slide I think is just my contact information there you go highlighting Dr. Kyle Meadow who's one of our newer members if anyone is interested in doing a data partnership with us please reach out to myself or Kyle thank you very much thank you so much Gina I think I'm just going to open it up to questions yes yes please hi Robbie Goldman AAAS congressional follow this question's for Gina I was wondering if you have or know of any examples of probabilistic information being communicated in a simple or actionable way to your partners and if not how might we go about doing that that's such a great question and I'm glad Mike Brennan is on the line as well the National Hurricane Center is already communicating uncertainty through a variety of probabilistic messages the storm surge flooding map is based on a 10% exceedance fact Mike I think you mentioned that in your presentation all of the the products sort of have probabilities underneath the hood the examples I sort of provide is that there are a variety of ways to communicate uncertainty some of which are more explicit where you actually provide the probabilities very concretely and some are sort of in the least likely scenario most likely scenario right and so there are there are many ways to do that what we have learned about some of the hurricane center products is that even though some of those probabilities are underneath the hood that what we learned from our hurricane supplemental projects in castle could talk to this even more is there's a need for what we call cats not the animals although many people like those two we call them communication assist texts the idea is is that some people don't know the storm surge map is based on a 10% exceedance and it helps to have sort of a sort of mad lib fill in the blank where you live what what your chances of experiencing storm surge are we wouldn't want to overplay well there's a 90% chance that you won't see that storm surge right the idea here is that we're emphasizing the 10% because we're saving lives and I don't know Mike or Castle if you'd like to actually answer that question as well yeah i'll jump in real quick jeannie i mean you're right we have a lot of probabilistic information that underpins a lot of our products and some of it's very explicit and some of it's sort of uh sort of not out there in front and center but you know there's probabilistic information that underpins the watch warnings issuance which is sort of uh translating that risk into a more sort of clear kind of a you know actionable message for people you know here's an area where you're at risk of life threatening inundation from storm surge you know you need to take precautions based on this in reality it's underpinned by say you know there's at least a 10% chance of three feet or more of inundation given in this this area um but yeah you know we've we've you know and we're working towards that direction to win too you know recommending and laying out hurricane watches and warnings based on probabilistic thresholds so um you know more of the explicit probabilistic products in some ways are designed more for you know more of the emergency management portion of the user community that is really interested in those various thresholds and the risk of certain thresholds being exceeded for you know evacuation decisions or many many other decisions that have to be made and you can certainly expand that well into the private sector and other parts of the public sector where there are varying thresholds that people need to uh need to take advantage of that you need to consider when when making decisions and so the goal is to put all of that probabilistic information out there for people to exploit but not necessarily put it all in public facing products well thank you so much we this has been an excellent uh excellent excellent panel and we really appreciate you sharing your knowledge with us um we are going to move on to our next session but thank you again so now I'm going to turn it over to Jeanette Sutton and we have an exciting next session um so our next session is the demos um which are really more walkthroughs of new and emerging technologies or existing technologies and giving you a little more insight into how they work so um we are going to begin with Mike Gerber from the National Weather Service that will be followed by Brock On from HAWS Alerts and then online is Anna Tolle Gruzzt and Philip Mai from Toronto Metropolitan University and Brock uh Mike you're up first and there's a little timer at the bottom to let you know when you're getting close to your time great thank you let it begin uh yeah all right well thanks everybody uh great to be here Mike Gerber of NOAA's National Weather Service been in the National Weather Service for uh 32 years oh my gosh 32 years just realize that and uh but I've been in the uh I'm in the Office of Dissemination I'm the lead for Wireless Emergency Alerts in the National Weather Service so I'm going to spend the next several minutes just kind of throwing a few things at the wall we'll see what sticks and then you can you know ask questions here in just a little bit next slide please great thank you so on the National Weather Service we activate wireless emergency alerts only for the most critical uh hazards if you will based on impact so thank you so only if it's uh if it's an event where the hazard poses a great you know threat to life and property or we're going to activate wireless emergency alerts and in terms of tropical we activate wireless emergency alerts for hurricane and typhoon warnings of course as well as an extreme wind warning extreme wind warning is when a major hurricane is coming ashore so that's when you're going to get that extreme wind warning and for each one of these wireless emergency alerts we've developed templates so that whenever you get a wea from the National Weather Service it's probably going to look familiar to you you know it's kind of a plug and chug situation here where we you know it's basically the same message every time except we will plug in of course the time information there and actually really don't need for so much for these plug in the time for the hurricane warning because it doesn't really have an expiration or event end time per se but the extreme wind warning of course you will see the till time in there now these are these templates we provide them in both English and Spanish because we provide wea in both English and Spanish those are the two languages that are allowed by wire that are essentially allowed by wireless emergency alerts and so now you'll notice also there's 90 and 360 character templates if you speak to wireless industry you'll you'll find out that pretty much every cell phone out there now is capable of receiving the 360 character version of the message that really in fact I even received an email yesterday from one of the major carriers who said really there isn't really anybody out there who's getting the 90 character messages anymore but pretty much all the phones are capable of getting 360 and I'll show you what that looks like here in just a moment when we look at a CAD message which is conveying the wea so how does a bill become a wall right how does that alert go from the national weather service to to wireless industry on the left side there of the chart what you see is those are the alerting authorities okay so we at the national weather service are of course a federal authority but we're one of over 1600 alerting authorities that have been authorized by the federal emergency management agency to basically send alerts to FEMA's iPods which it serves as the sole pathway to wireless emergency alerts so what do we do we just like other agencies other states locals tribals and so forth if we want to get an alert to wea we we they all have usually a third party commercial vendor you know alert authoring tool but we in national weather service have our own home baked software if you will that when a wireless sorry when an alert is issued that we want to go to wea we have software that automatically takes that alert converts its common alerting protocol which is just an industry standard for the exchange of emergency alert information that CAD message then is we send it off to FEMA iPods iPods basically serves as the nation's alert aggregator collection point if you will and again it's the only way to get an alert to wireless industry is through FEMA iPods see must must be authorized by iPods and send them the alert in CAD format here's what that CAD message looks like you'll see here in the seam or it says seam mam long text and the value of that that's the and actually I have a case the example here with tornado warning but this would be a if it were hurricane warning or storm surge warning it would have the language there for those and again that's a 360 up to 360 character message there so we send that CAD message off to FEMA iPods iPods gets it converts it to what's called CMAC which is a special XML type format that they send off to wireless industry and you see there by the way also the polygon information so you get the alert text along with the alert area in by polygon and so what happens when wireless industry gets it okay wireless industry will take that alert polygon they will map it to their network topology to determine from which cell towers in and around the alert area they should broadcast that wireless emergency alert and so if you look there on the on the right side that we 3.0 geotargeting over 83% of cell phones are now capable of this enhanced geographic targeting it's device based geofencing or DBGF for short and what that means is that the polygon is broadcast to your phone if you're so if you receive an alert it doesn't just automatically display the alert on your phone like it used to you know and you'd have overshoot bleed over you'd be outside the alert area and you might get it on your phone it would annoy you you know desensitize people to the alerts and so on and so forth well now with this we have 3.0 geotargeting your device if the device knows its location it will compare it to the alert polygon and only if you're within one tenth of a mile of the polygon inside there within one tenth of a mile will actually display the alert on your phone so that's the way that that it works now I should probably add that national weather services using leveraging impact-based warning to advance wireless emergent to advance and really refine and streamline the way we do wireless emergency alerts you know like we used to activate WEA for all flash flood warnings you know and a lot of people were saying why am I getting my phone lighting up blowing up when it's you know when it's just raining or maybe it's not even raining here and so what we do now is we only activate wireless emergency alerts when they the the threat damage threat for that flash flood warning is considered to be considerable or or catastrophic so that's basically the way we do things and we do it similarly for severe thunderstorm warnings which we recently added we only activate it when it's a destructive threat wow the alarm is going off holy cow how do we stop it that's not a WEA I promise thanks okay thank you all right so now we're going to pass it over to Brock Long and yeah you're changing sorry not Brock Long that's the old theme that you can call me whatever you want yeah my name is Brock on since you remember sounds like rock on pleasure to meet all of you as a citizen of Florida I just want to thank everyone in this room I think that it's been fascinating for me to hear the really hard work that goes into the products that we use every single year to stay safe and I just can't emphasize enough how impressive and exciting it is to see all the work that goes into it I'm very very honored to be here today and contribute our part on this so I come from a company called Haas Alert and I'm going to tell you a little bit today about the work that we started doing and then how it intersects with the really critical work that everyone in this room and online is doing and what we're talking about is a technology called digital alerting for automobiles so we started at Haas Alert trying to solve a very specific problem on the road which is struck by incidents on the road involving emergency responders it turns out that this is a leading cause of death for emergency responders in the field today this includes everything from fire departments and police departments to private companies like tow truck operators work zone operators close to 300 000 collisions a year involving vehicles that are actively trying to alert drivers on the road of their presence using things like light bars and sirens and so we went to work on this problem by trying to enable those fleets to deliver real-time digital alerts into vehicles so the drivers have more time to slow down and move over as they're approaching active incidents on the road and so today we are live in millions of vehicles on the road 2018 and newer Jeep Dodge Ram and Chrysler vehicles receive real-time alerts from fleets that are using our service warning them to slow down and move over and I can actually show you a video here of what those alerts look like so these alerts are received by drivers about 30 seconds before they reach that incident they'll get a pop-up alert that warns them there's an emergency vehicle ahead there's a tow truck ahead there's a work zone ahead and this gives them much more time to slow down and move over these are delivered over the everyday cellular networks the drivers use to stream Netflix in their vehicle and other things like that so we're leveraging that functionality for life-saving purposes before we got into the vehicles we actually started a navigation apps so we also deliver these alerts to ways there's a very good chance if you've ever seen an alert like this it's actually coming from our technology but the great thing about our technology is rather than people having to crowd source and report these alerts our technology installs directly into these emergency vehicles and activates when their light bars are active so it's a totally passive solution for the emergency responders in the road we also deliver these alerts through apple maps as well again our goal is to get these alerts to as many drivers as possible no matter what they're driving so that people have more time to slow down to move over and responders have additional protection in the road you know the the benefits of alerting you know this is sort of where the crossover occurs between your industry and the industries that we initially served we know that the odds of a crash are reduced by 90 percent when drivers receive some type of advanced warning compared to just traditional light bars and sirens and it's sort of intuitively obvious why that happens right because if you tell a driver there's something ahead it gives them much more time to prepare than if they just have to organically see it as they're driving at 70 miles an hour on the Florida highway there was also a study done by Purdue University in 2021 where they compared the utilization of digital alerts to traditional light bars and sirens and found that with the use of digital alerts hard breaking next to roadside incidents was reduced by 80 percent and at the time we were only alerting on ways and so they estimated that approximately 30 percent of drivers at the most were receiving these alerts but it still has a net effect on the behavior of drivers on the road when they see their ears slowing down on the road and then we are actually fully deployed in DC fire and EMS here and that was done in part through NHTSA and a study we did in NHTSA found that whenever drivers receive an alert of something ahead of them that they need to slow down and move over before they typically reduce their speed by about 25 percent in the first second so we know that these alerts reduce risk on the road and save lives one of the benefits of safety cloud digital alerting safety cloud is our platform is that from the automaker perspective what we are doing is we are aggregating the universe of roadway hazards and providing a single authoritative stream for automakers so this is a very difficult problem for them to solve because what you're looking at on the road is essentially millions of tiny silos of data and for any automaker to find ways to integrate that data into their system organically it's very very difficult because there's thousands of commercial agreements they'd have to create you'd have to sort of figure out these protocols and our approach is to build those connections for that market so that automakers have a single source to rely on for the delivery of those alerts the other great thing about our system is is that this plugs directly into the connectivity in the clouds that these automakers are already utilizing for a wide variety of other purposes so in the case of stilantis they were able to deploy these alerts through an over-the-air update retroactively on millions of vehicles without having to write a single line of code we just integrated our work directly in their cloud and then they were able to activate this so how does this connect to tropical cyclones and extreme weather events well oh sorry this is one other thing here the other great thing about digital alerting is that this is really intended to be a cross-platform solution and so we integrate directly with over 50 different platforms on the road so that if you're a fire chief you don't just have to buy our system and install it you'll probably have it pre-installed on whatever new truck you buy because we do direct integrations with emergency vehicle manufacturers you can activate it on a wide variety of other telematics platforms and so the purpose of this is not to give everyone a single platform to use it's about connecting all those platforms together in an interoperable network for safety this is a live map of what those alerts are looking like on a given day across the united states every single orange dot uc is an emergency hazard in the road it's actively broadcasting alerts we are active in all 50 states we're active in all provinces of canada and this is something that over 3,500 public safety agencies are using today in their fleets so we were approached by dhs science and technology directorate and they said hey we would be really really great if we could see what would happen if we plugged iPause into safety cloud can we deliver wildfire alerts to drivers as if they're in an area where they might be at risk and so that's the work that we did i am the company's english major so i was not involved in this particular technical integration but the way that this chart has been described to me is that this bottom right corner here is the most important thing you see here we are pulling data from the ipause all hazards feed and then translating that into safety cloud as a as a conspicuous hazard for drivers that are approaching that and so i'll show you what this looks like in our dashboard if we are received uh some incident from ipause that says hey there's a wildfire in this area the first thing we do is we take that area and we translate it to understand what are the relevant roadways that are going to be impacted by this particular hazard right we are then actively deploying lane closure warnings within that area so any drivers that are going to be approaching those active areas they're going to get an alert that there's a lane closed that there's something ahead that they need to be aware of this is also going to our partners again in ways an apple map so they can actively use that for rerouting if they want to and then what you're seeing here is those blue dots are sort of vehicle tracking and you can see here that we at any time that there's a vehicle that is approaching this active area as soon as they enter that area they're going to receive a pop-up alert that tells them hey there's a wildfire ahead you need to be cautious and this is what those alerts look like in real time and one of the great things about this is is that this work that we did with dhs in stilantis once we sort of settled on what the message is going to be what is the user experience going to be we rolled this out in 24 hours with stilantis and then they were live and so now at any point if there is a wildfire somewhere and we've deployed that alert any driver that enters that area will receive this alert you can see here that it's customized to the language use case and so drivers if they've identified there's a specific language stack they're going to use they're going to get that alert in their native language and they're chosen language we've also now expanded this to to flood sensors and work that we've done with noa we know that there are about 120 unattended flood sensors across virginia that intelligence is deployed and so these are going to be integrated into fema ipause open feed and so these sensors are now going to act as an alert originator so if we get a signal that there's an act of flood somewhere we can then use that as a conspicuous alert into a vehicle to warn a driver there's flooding ahead be cautious um the last thing i'll say is that uh we have a strong opinion of pause alert that um the the customization and definition around these alerts should really be led at the national federal level uh the wonderful thing about our system is that this is deployable and usable by emergency responders on the ground but every day we have agencies across the country that want to deliver custom messages to drivers about specific hazards and the question of what's the best way to alert a driver what's the best direction to give them you know do we want to tether this into something like rerouting those are all opportunities that are available to us and this is now a permanent fixture of the transportation ecosystem and that is why partnering with federal agencies and policy makers and experts like you really give us the opportunity to optimize this to maximize public safety thank you thank you so much all right um we're going to move on to our online speakers now anatoly and philip hi everybody can you hear me we can hear you excellent thank you for having us so my name is philip i am the co-director of the social media lab here in toronto um we've been around for over a decade and one other thing that we do here is we study how social media is changing the way people connect um and communicate online um as part of our work we develop social media research tools and visualization uh and dashboards to support social science research on um you know unlike um participation in communities so for example during um um at the beginning of COVID we work with the who to track um online misinformation about COVID in general and to help public health officials figure out what kind of um misinformation is floating around online so this way they get uh uh leg up on how to respond to some of those uh messages online one of the tool that we've been developing to help researchers um to understand the information environment and monitor for misinformation is a tool called communolytic it's a combination of the word community and analytics um it's available at communolytic.org um it's basically a no-code computation of social science research tool for studying online communities and discourse it contains a suite of data collection modules that can collect publicly available data from Reddit, Telegram, uh X, YouTube, CrowdHangle and so on and so it allows you to run a variety of analysis um on the data that you collect as a way to do some social listening you can do topic analysis, you can do toxicity analysis, you can you know detect um uh who's talking to whom and make those connections but what I really want to focus on today is how social media is um changing and actually making uh coordinating crisis communication via social media is becoming as we've been hearing throughout the the two days so far more complex um and one of the reason is the proliferation of new apps and platforms um in the last five years six years we've seen the rise of TikTok where in previous types of platforms you might rely heavily on text to get your message across but TikTok for example require a different skill set that changes the way and the type of training and skills you have to have in-house in order to basically get your message across on those medium um but again this kind of proliferation of these new platforms poses challenges for organization operating with limited resources um so for us one of the way we're trying to figure out is how do how do communicators get their message across in an information environment where um the audience is now more scattered than ever before and as we've been seeing in the last few months for example um you can't really rely on traditional media all also to um get your message out because as you can see hundreds and hundreds of um journalists have been fired news organizations have gone bankrupt so it's becoming more and more difficult uh information environment uh especially just if you want to get something out in a timely manner and then the other thing is that over the last seven eight years now many of the platform has switched over to algorithmic filtering so it's more and more difficult for um anybody to go viral or to your message to just get out to the audience anymore simply because many of these systems have now become pay-to-play where if you don't pay you don't play um so as a result many of these systems the way they're designed it is designed to basically push very emotionally charged messages and if you are not delivering it in that way your messages are going to die quiet death on a lot of these platforms um of course unless you pay so these are some of the challenges that we're working on and trying to figure out how to get around some of those and then the other thing that we want to flag is that the rise of generative AI um in the next few years if not months you will see more and more of the growing adoption of generative AI for legitimate and also for malicious purposes on these platforms uh recent high defakes of celebrities uh and public figures shows that social media platforms are unprepared for the influx of AI generated misinformation um so you know there are so many tools available now to the average user um to pollute basically the information space so I know that many of you um are now relying on uh social media as one of the medium or one of the channels to get your message out but like I said going forward I think that need to be re-think simply because when you build up an audience but if you don't control the medium the medium could disappear and there goes one of your um avenue to get the information out um and the other thing is that I know that many of you are also designing and working on dashboards that um might just data from social media um but the problem is it's harder and harder to trust those medium a because again the algorithm filtering that's happening online but also the increase availability and prevalence of area generated content so it's going to get more and more difficult to figure out um what's true and what's not um going forward so we'll have to have as a community re-think how and what how much weight we give to the information that we find and discover online going forward and I guess the last thought if I turn it over to my colleague Anatoly Gruse is to touch on the importance of reaching out and reconnecting with the folks who are working in traditional media because of all of the um shut down and firing in recent years the people that you have uh worked with in the past my long no longer be in the industry so it might be good practice just basically once a year just to reach out to see if those people are still around and um if they're not to ask them to introduce you to their colleagues who might be taking over their roles so this way when an emergency does happen you don't discover that your Rolodex is now full of ghosts um so that's a thing that a practical way um to keep it up with this fast moving and a fast changing environment so on that note I'm going to turn it over to my colleague Anatoly. I think I'm being flagged that we're out of time but Philip thank you I think you did a great job summarizing what we do and how it relates to crisis communication I just want to add briefly we're also following how state actors weaponize in social media and we see how crisis communication can easily be weaponized to either exaggerate or downplay a risk can be used to disrupt events or discredit local and federal authorities by criticizing for example their responses so we feel there's a lot of work needs to be done in preparing for this you know this particular situation that I'm sure will happen in the future um I think we can close at this point and looking forward to your questions. Thank you I think we have time for one question nothing is there anything on okay Richard? I guess if nobody else wants to ask I'll ask I thought the Haas technology was just incredible that was fantastic thank you for that and when we were developing the earthquake early warning effort there was a huge amount of concern that having earthquake alerts in cars would actually do more harm than good that we might cause more accidents than we would be preventing and so I just wonder you must have had the same issue with some of your alerts I just wonder if you could talk to the issues about whether you know you're scaring people things like that thanks. Yeah it's uh it's an excellent um excellent question so uh this is this uh doesn't even just it's not limited to just earthquakes I'll give you an example um if we know that a vehicle is approaching an intersection that an emergency vehicle is approaching uh originally we were testing alerts where we said you know warning emergency vehicle approaching on the left be careful and then what we found was is that just by providing that directional uh information within the alert it actually caused some drivers to panic more because then they started looking to the left right and so then we narrowed the information in that alert to say emergency vehicle ahead be cautious which basically just means just keep your eye on the road um this gets at the exact question that I sort of posed at the end here which is that for each of these different hazards the action that you are asking a driver to take is going to be unique to that particular hazard that we're alerting to and so this is a question that needs to be solved on a case-by-case basis you know it might be that you know earthquake coming you know pull over and prepare uh wildfire ahead seek shelter you know turn around I mean these are depending upon what the circumstance is and what the hazard is the message that you might want to attach to that is going to be unique to that um we have the capability to enable customization around those alerts to some degree uh the operators of these platforms at the end of the day you know ways and apple maps and these various uh automakers are going to control the final delivery of that experience and that is why our goal is to standardize as much as we can what these alerts uh are you know what they're requiring of drivers and what we're asking them to do um but this is becoming more sophisticated as we speak there are automakers that want these alerts to pop up uh in the instrument cluster there are automakers that want to use a heads up display or do haptic feedback uh so the different ways that you can notify someone of a hazard and the different ways that you can ask them to take action uh the other the other frontier of this is autonomous vehicles um you know we can deliver data along with these alerts that tell an autonomous vehicle like don't come in this area because there's an active emergency vehicle or turn around or automate slowing down and pulling over um and so as we move to those uh future states of transportation um the opportunities available to emergency response managers and communicators to to really coordinate targeted messaging and direct the flow of traffic and transportation is going to become more robust thank you very much panelists um I will now welcome Gabriel Long-Parody to moderate the next panel so thank you great thank you so much for such an engaging um and lively discussion from the last session so I am pleased to to introduce this next session a session and the speakers so now we're going to transition to talking about risk communication innovations and new frontiers in tropical cyclone communication with a focus on the private sector so we have three wonderful speakers today the first is Mike Chesterfield from the Weather Channel the second is Micah Berman from Google and the third is John Lawson from AWARM and I believe the speakers are in the room so thank you start with with Mike thank you I know we're running short on time so um I'll try to move as quickly as possible a lot to go through thank you all um my name is Mike Chesterfield I'm the Vice President of Weather Presentation and Data Visualization at the Weather Channel special thanks for the committee for the invite I think it's ultra important to continue to collaborate um private public partnerships to me is is really going to be the key to helping solve some of these communication issues that we face today I can't see the presentation down here okay today we're going to be talking about the innovative techniques that we're using in visualization to help to improve messaging particularly as it relates to hurricane evacuation orders and storm surge what have we learned or what do we think we know a good visualization is not only able to convey a particular message to the public but also identify the particular task in this case evacuate we've also through studies and learned that visuals have been found to enhance our understanding and shape our perception of risk not only do images and graphics increase the perception of risk but studies have also found that dramatic video can also confirm risk while lowering perceived uncertainty many ask me why do we put talent into hurricanes and this is one of the main reasons we do this creates a challenge though as we look at trying to warn people using video video content traditionally displays the current or the past limiting use as a warning tool so how do we produce videos which depict the future that's the challenge that the weather channel has taken on we created what we call immersive mixed reality I'm not going to bore you too much with the how I want to get more to the how useful it is but really quick high level overview by combining a real-time graphics engine in this case the unreal engine and modern processing power and next generation broadcast production tool set for the first time we're able to efficiently produce control hyper realistic video experiences in other words we're able to show what the future is going to look like in a video product or as we like to say at the weather channel we get to play God a little bit so it's it's pretty fun we also immerse the on camera meteorologists expert narratives into these believable simulations and this has opened up a whole new realm of forecasting measuring capabilities for us we think it's extremely important to pair the visual the visual excuse me with the on camera meteorologist expertise the first such product we produced was 2018 surge effects what this does it depicts the nhc reasonable worst-case scenario forecast and what it will look like in a recognizable environment so we are clear when we show these images that this is the reasonable worst-case scenario we are clear that this is what we want you to prepare for it's not necessarily forecast but this is what you need to prepare for we leverage real life objects to highlight potential impacts and dangers we immerse the talent into the scene to provide relative context while also providing meteorological expertise and advice we think it's very important to put the on camera meteorologist into the scenario because normally we can't put somebody into the scenario so this provides the audience with that context right takes what traditionally has been displayed on a flat map and translated for the audience I had a video it's not showing up but believe me it's awesome for those of you have not seen this it is available on youtube just do a search for surge effects the weather channel and you'll get an example but had I've been able to show a video you'd be able to see the talent show the water rising around them real objects surrounding them including homes etc which I think provides the audience with that perspective necessarily to warn them oh wait there it is will it work well Dr. Bostell seems to be frozen so we're just going to go ahead and advance this maybe or I broke it one of the two there we go so the next question does it work well we do not at the weather channel have research money or anything like that but luckily I have a very dedicated employee who was getting his master's at Edinburgh University and as part of his dissertation he said hey I need some extra credit I want to raise um so he he conducted a survey what to determine the effectiveness of the surge effects product and what he found through 815 respondents is that yes uh at least anecdotally it looks like that the video was effective the respondents said that they'd be much more likely to react to an evacuation order after seeing the video as opposed to before seeing the video and that was true across all age groups it's very important that we consider ages this we we know we struggle connecting with the older audiences when it comes to evacuation orders and you can see there was a significant increase from 59 percent to 77 percent in fact we talk about Ian and how deadly it was a majority of those individuals who passed away unfortunately were elderly so was the video on former and educational yes pretty much universal which was great news for us would they prefer this video overseeing a meteorologist with traditional maps in the newsroom majority said yes um they would prefer this method so we think we're definitely on to something this past year we launched a new product called flood effects and this is the integration of an inundation simulation based again on the national hurricane reasonable worst-case scenario forecast we take actual drone footage and we have it at a treatment called nerf from neural radiance fields which is translates the 2d video imagery into 3d environments so essentially we're taking a 2d flat image and creating a volumetric display and take that volumetric environment incorporate it and import it into our unreal engine and then put a simulation on top of that the end result is a simulation that shows again what the reasonable worst-case scenario is we pair it with the inundation map the key to this product really is the hyper localization capabilities that it provides we're using actual imagery from a location in which is in the threat area again i had a video i'm assuming it's not going to work again these videos are available on youtube if you are interested i can point you in the right direction so what's next um we at the weather channel um we know we hear all the time from on camera meteorologists across the country there's nothing more frustrating than having the forecast correct but people not listening to it oftentimes we make the mistake of blaming the end audience it's not the end audience it's us the problem is us so how do we continue to push forward what are the tools that are coming to help us one thing that was released this year um google released a a set of 3d earth tiles through a combination of satellite imagery street level imagery and gen ii the results a creation of 3d digital earth twins essentially what we're looking to do is hopefully this capability or this technology will enable us to scale flood effects like simulation capabilities to any street usa you can imagine uh the public being able to put in their their address and then get a visualization simulation that depicts what the forecast or a reasonable worst-case scenario shows out their backyard or out their front door there's really no way better way in my opinion to get people to react and actually show it in what their future may actually look like and this one's a little controversial but i believe it is coming i don't think there's much we can do to stop it so we might as well use it and use it properly metahumans or personal weather assistance is what we're calling them warnings delivered through digital humans allowing for hyper local personalized messaging the advantage of using these types of tools would be to personalize or to scale a present presenter imagine having a million gem cantories that talk to you personally use your name know your personal situation talks to you in the dialect that you understand this is what this type of capability is is will open up the possibilities for again i want to thank everybody for taking the time again if you have any questions please see me after thank you awesome hi everyone i'm mike aberman and before i go any further just want to say thank you so much for having me i do as many folks here have expressed feel really grateful to be in this room of experts and to learn from all of you about your work so i'm a product manager at google and i look after a platform safety on android so my teams our mission is to keep android users safer and do that by deploying services across all three billion give or take android devices in the world so things like the emergency location service for e911 and the earthquake early warning system which is the subject of my talk today but at the highest level my goal in being here with you all is to gather your ideas for how android can help lead the way in innovating alert delivery there are different mobile phone platforms and not everyone has a smartphone i know but i think we are at at a time and in a space where we are fundamentally under utilizing the capabilities of the supercomputers that live in our pockets to make alerts useful and relevant to people my teams and i are really interested in experimenting and trying to figure out how we can help advance this work so i'll do two things quickly the first is tell you about our earthquake early warning system and the second is pose a bunch of questions um so earthquake early warning most americans have a phone 97 percent of them but more than half of them have never ever in their lives gone more than 24 hours without their phone well everyone has phones around them earthquakes are the third most common disaster the fourth most deadly globally speaking and they're the one that we can't predict and there's some evidence that suggests earthquake early warnings work we heard some about shake alert earlier um this same idea uh we think there's a lot of promise in earthquake early warning and so we set out to build a system uh to try and deliver those alerts to users what we've created is a supplemental android detection and distribution system so in the us we consume a feed from shake alert and distribute those alerts to android phones in about 98 countries give or take around the world we actually do our own detection using the accelerometer in the phone the same sensor that detects when you change your phone from landscape to portrait mode um to listen for earthquake shaking model the earthquake including its magnitude and epicenter and then use that to trigger alerts that we send to phones as well um in all cases these are estimates we've talked some about communicating uncertainty and i would say it's something we think about as well we're estimating the earthquake or shaker is estimating the earthquake right as it's unfolding the second key piece is delivery and so we've built a bespoke low latency delivery network that's point-to-point ip based but can send alerts out to millions of phones in fractions of a second around the world and just to underscore the supplemental system piece you know we are a mobile phone manufacturer we are not a disaster communication expert we view our role here as being able to make sure the android platform can support and supplement all of the official government warning systems that are out there and already exist not replace them so in terms of alert delivery we have a bimodal model today where we deliver two different kinds of alerts based on the anticipated shaking intensity for any given user um and i think it's a cool opportunity in being right integrated into the phone knowing where a user is um that we can make uh the alerting experience pretty specific to what we think will will happen for that user and in the future you could imagine alerts that are even more specific like if we know users inside or outside could we customize the alerting experience for them but today it's bimodal there are two kinds so the first is uh be aware alert and they look like the image you can see on the right it's a standard android notification lets the user know they might have felt shaking or they may feel shaking shortly and this is primarily an educational tool our system exists it works it corresponds to the shaking that they can feel um and it builds trust the second kind of alert that we send is called a take action alert uh you can see those on the right there they come in two flavors uh so the first flavor on the left is the standard alert suggests that users drop cover and hold the second flavor on the right is a more general protect yourself alert we call it um because it's not uh official guidance in every country in the world in which we operate to drop cover and hold in case of an earthquake so we do present this alternative guidance in some markets and i think again it's an advantage of being integrated into the phone that we can for example uh customize and make these warnings available in any of the languages that android supports uh and can then customize the alert content to be relevant to the the country that any given user is in uh finally if uh we send an alert to a user and they don't interact with it or for whatever reason the alert will arrive late we can change the context of the alert so that it's still relevant to the user it notes that shaking will have passed but we still leave the alert around for a while again with the goal of education and and building trust in uh in the system so then once users click through an alert and actually i should say about the take action alert we have designed a specific uh sound that goes with this alert as well so we worked with a sound designer to try and engineer uh something that would get users attention but wouldn't cause them to panic and we think there's you know more opportunity to do things like that as well okay so when users click through an alert we then have cached information on the device about next steps knowing that connectivity is likely to go out immediately after a major earthquake uh we give users more information about the earthquake itself and the source of the information as well as next steps and safety tips and then we use the android detections to power google search as well so uh when android detects an earthquake in the countries in which we're live we can present this information to users searching for information even before authoritative and official sources have produced a final estimate which often takes minutes to hours so what have we learned and where are we headed this is the second piece we've found that user feedback on the system broadly speaking is tremendously positive users really appreciate this and and feel really good about the fact that their phone is able to help detect uh and deliver to them uh customized warnings about hazards around them and we did a lot of experimentation before we launched the alert originally and are doing some more now try and figure out what the right information is to help users take action and it's been really valuable to be a part of the conversation here already and learn from you all on some of the research that you're doing on those fronts but here are some things we'd like to understand and i'd encourage y'all to please uh i'm i'm open i'm here to hear your ideas questions concerns and criticism whatever it may be about the system and about what we can do so three things here one is your ideas for how we can make the earthquake warnings we send today more actionable we would love those it's something we are thinking about we're working on a revision of our alerting model right now a second thing is how can we expand this capability to cover new types of hazards if you think about it right a crisis alert is just the beginning of a journey ideally users have prepared but at the least the warning about an impending tropical cyclone or a hurricane or an earthquake that's the beginning then users need to take action to protect themselves they probably want to communicate with their loved ones maybe they want to learn more information about their loved ones even if they're not close to the event itself i think we can expand the work that we're doing today to integrate with the other capabilities of the phones to make uh the warnings that we deliver truly useful again a hypothetical you could imagine a tsunami warning that lets users know how high they are relative to the danger level and whether they need to keep going up if users need to evacuate you could imagine giving them an easy way to access a mapping application and then even a route that's unlikely to be congested out of the danger area and finally any other key learnings you all have that you think should be on our minds very interested in thank you again so much uh for having me and for all the work that you all do to keep us safer hello i'm john lawson uh a warrant alliance i want to thank uh the planning committee particularly janette sudden for inviting me there's been a great conference i'm also aware that i am the last thing between you and lunch so i hope you find this interesting i think you will so we're the um let me see on the right okay so we're a coalition a non-profit coalition of companies and public broadcasters as well as consumer electronics makers and we have an advisory council that is they're not members because we do some advocacy work but have been very helpful in advising us uh shout out to ron prater with uh big city emergency managers and uh all of our other partners so i guess you could state our mission more broadly as improving climate resilience that's really what we're talking about uh with better alerting we're not solving co2 emissions but we are hopefully helping people uh take protective action uh more often and and sooner so the underlying technology here this is um not vaporware it's being deployed now is called atc 3.0 atc back in the 2000s is the technology or the standards body that brought us what's high definition television white screen television they had to make it digital to do it this is the third generation and you can see that advanced emergency alerting with the great credit of this global community that created the standard baked in advanced alerting from the very beginning and it it works in conjunction with these other capabilities so what does it mean in terms of alerting specifically this is coming from tall towers this is going from tv towers most a lot of people today don't know that you can get free tv over the air but this is a next generation television signal that provides geotargeting we can do it in a very granular way the the alert is embedded in the signal that goes out 50 miles from these towers but with geocodes and the device doing where it is you can you can hit the polygon pretty pretty accurately we can provide rich media agree totally with with mike in the like the research he displayed the text messages have great value but uh reducing billing requires rich media and we can provide it if we get it from the alert authority and that could include flood inundation maps unlike the atc one oh which optimize resolution over reception this is really built for indoor reception and mobile it works very well and very high speed moving devices we can wake up devices very controversial but as long as the device is not unplugged and most most devices can be woken up with this technology if they allow it resilience and i'll talk about this this is when all hell has broken loose or norm or just a normal storm when power loss and cellular losses really not the exception but the rule and probably most importantly atc 30 which we call next gen tv as a consumer matter really is it's the world's first all ip over the air transmission system for television which means that even though it's one to many it can natively integrate with any other technology as a return pass so as i said this is not theoretical it's being deployed now it's all over south korea and will be deployed in other countries it looks like in the u.s we now have a at least a station that's providing a signal that reaches 75 percent of us tv households that doesn't mean that people have the the dongle or the new set to receive it but physically that's where we are so this is not um legacy eas the emergency alert system dates to the truman administration it was a never designed for local alerting it's almost never used for that we're not we are i support we are i was the head of the association of public television stations in the 2000s and i lobbied for we are um and it's a lifesaver but it's also got some vulnerabilities all of these systems do there's no silver bullet one vulnerability is just the text message the limitation in terms of the media and we really have been informed by the late great doctor malady i know janette and others have worked with him he coined this term milling uh the research i've read indicates people usually do not panic in emergencies that was my revelation to me it happens but it's rare the exception people actually treat each other better in emergencies than they do in in day-to-day life but they almost always delay and that delay can be fatal so sometimes he's short messages uh even short messages don't get out which has been the theme of this conference in many ways electricity cellular we've seen many maps of storms this is ian but the the red magenta colors this is from the FCC this is their map cellular was pretty much wiped out in the impact area but almost all of the tv stations stayed on the air they were hardened after 9 11 these towers rarely go down and they've got days of diesel fuel so if the received device whatever it looks like a phone or a tv set or a ipad um can receive this over the air signal and you've got a lifeline here's itis even worse and power was out for days or weeks in some areas as i understand but almost all the tv stations in the area stayed on the air so um these power outages are a recurring problem everything went wrong that day on malley people were gonna die but a lot of people died needlessly because of some alerting failures one problem was that the power was out cellular was out they couldn't get with us and we you know infamously this tsunami warning sirens were never activated but we um it just recovers all the time at all different storms and the market is responding you don't have to have a new tv set to get next gen you can get one of these set top dongles this is one that's got battery backup and you can see it really is a home server in that it's receiving the signals off the air but retransmits over bluetooth and wifi to any device in the home so we can we can let people know there's something coming the FCC has sort of uh finally sort of uh acknowledged that we always not the the answer to all awarding problems they did this public notice but we were very disappointed just despite our outreach that they didn't mention the existing infrastructure of tv and radio stations but instead focused on satellite uh balloons and unmanned aircraft systems as backups when we are goes out we we did file we know there's a file in that so we've done some groundwork with emergency managers uh janette and uh ron prater were part of an effort outreach we did in 2018 and we had the top creator designer for cnn who came up with some concepts we tested with emergency managers in three major cities and this is what this is not the final ux at all in fact we're working happy to work with janette and company at universe of albany to try to uh refine what what's what else has been done since then but this is an example it would be only used in an imminent alert a human being might might not touch it there'd be in a relationship between the alert authority and the station this would come out you wouldn't get it unless you're in the polygon but you can get more information or dismiss it if you want more information this is from santa barba county that's their debris flow uh hazard areas you can give people that kind of information you can give them if their phones are operative you can give them the ability to interact with twitter or x i guess we need to change that uh shelter locations you can switch to the news coverage um all at the hand you know we we want to give them rich media but you can't push it to them all at once we want to give consumers optionality so we can do that switch to the news conference um whatever uh whatever is relevant uh it'd be a combination it would be a these are private agreements between the alert authorities and the stations the tv and broadcast tv and radio only have to do eas the rest of the voluntary so this engagement extended to some round tables we did not unlike what we're doing here in this workshop we did it in four cities we've never done uh events that got such high ratings we were brought together emergency managers and and local broadcasters um to talk about using this tech and the one in washington has led to a pilot project between wj la tv and um the council of governments here taking alert information from fairfax and earlington counties and we hope to be able to report on that to the public soon so the takeaways from these round tables strong interest in using this technology the broadcast the local broadcasters are struggling to remain relevant so they definitely see an an opportunity here to do that um technology is empowering but the relationships are really the key um and above all we need national leadership i agree totally with the gentleman from haas um we're not getting it particularly from the FCC and um it's not gonna i don't see what wide-scale deployment until we we have that kind of leadership and i hope that we can um turn the corner on that soon uh there is a private sector approach to it i'm part of a startup that's we'll be selling receivers and we're creating a um a disaster channel a 24 7 reality channel it's kind of c-span meets the weather channel but it would not we wouldn't have presenters we all be programmed by ai although the presenters might be ai but there's just a massive amount of public domain content being generated every day that could populate a 24 7 channel and we think um companies involved in recovery and have a stake in limiting disaster uh losses would could be sponsors so we're trying to find both a public private partnership and um entrepreneurial ways to achieve this but uh the information you're you're you're generating is extremely helpful in understanding how to present how to use this technology to present this rich media to uh to the public so that they take timely protective action thank you thank you so much to all the speakers this was an amazing session and i have several questions but i i want to ask one that comes from slido uh specifically from Jen Henderson who i think is posed a question that many of us are thinking about and um Jen says to Gina's earlier suggestion of public private partnerships how likely is it for those of us in the social sciences to partner with those of you who have the apps and have the the different new technologies you're working on to help collect data by surveys etc um any suggestions for next steps for those of us in the room virtually uh or physically who are interested in partnering well i can tell you social sciences is key to what we're doing and we're wide open to partnerships we're not we're not the companies we don't have very deep pockets but we have a lot of interested people that um would would would welcome the opportunity to work with the academic community and others to figure out how to do this thing right and i can just say for my part i mean it's something that uh i'm tremendously interested in it's not something frankly that's easy for uh like a large organization to do we have a lot of obligations to our users but i think it's a place where we can creatively problem solve i mean it's you know and i think that's where the the ground forward is is like uh how can we uh come together and figure out what it would look like um to be able to do something that for an organization like google for example is like respectful of its users and their time and our agreements with them but also is able to get um data that can help uh advance the way that this works for for all of us i think it's tremendously important so i don't have an answer to the question um only a caution that it's not simple and an interest in trying to work it forward with anyone who's interested in that yeah i just echo um when michael just said it it's not a simple but through relationships uh it's definitely doable so um you know come talk to me and i can put you in contact with the right people can't can't guarantee that the answer is going to be yes um but i can certainly use my voice as well to help to move things forward because we certainly rely on the result of data to shape our products so i think it's a win-win great thank you so much um you i'd like to turn it over to you see if there's anybody in the room who has a question um otherwise there's some fantastic questions here online absolutely i'm gonna run up to the front of the room so i can see everyone's hands but go ahead um fascinating panel i loved it all mike my question is for you um i loved your talk and it got me thinking about unintended consequences and i would be very curious to hear your thoughts um at the weather channel about the possibility of oh triggering people who want to become social media famous by filming themselves in the flood waters or surfing in the storm or people who maybe don't understand that it's not real who might then feel more safe right traveling through water and a third category which is um i've heard a a group of folks worried about um trauma-informed method promoting trauma informed messaging meaning my understanding avoiding language that is explicit around impacts and i would imagine this group would also be concerned about images that are explicit about impacts and i'm just i would love to hear what you all you know your discussion and thought about all this yeah um i i would just say this we're extremely careful um it's why we put a expert communicator in with these scenes uh narrating over it we constantly remind the audience that this is a simulation we would never put ourselves into this situation because of the dangers we're only showing it to show the potential danger that's uh could be behold based on the forecast um we do everything in our power to discourage uh imitators and social media in fact we go out of a way not to show individuals who put themselves in those dangers even though the video may be wild and crazy and people may be actually interested in seeing it may be going viral we do not show that on the weather channel so uh definitely appreciate the concerns we use terminology like simulation remind people that this is not real um but at the same time uh we are our goal is to communicate the potential danger using a visualization that allows people to put themselves into the situation and we think that's uh critically important hope that answers your question yeah really interesting discussion um alex from the weather service here so you know i think pretty much a lot of you and even people in the previous panel are pretty well practiced in delivering traditional government alerts right the tornado warnings and that sort of thing so as we think towards the future um micah you talked about like the amazing thing you're doing with the accelerometer like what are like the next generation things like data sets that we could be providing or i guess like what would what would be the underlying data set for a useful alert to you because i mean we are providing a lot of different data um i'm just curious how to make those connections um in a way that would allow you to deliver these rich alerts and that's for anybody on the panel really well we we need these roundtables interoperability was i mean we tried not to keep it to technology uh only but interoperability was always a big one it's like if an emergency manager has everbridge and a tv news station has its own platform like avid how do you bring those together how does the public station that doesn't do news how do they pass it that's really not that complicated in terms of the technology it's all about relationships it's about understandings i mean our goal um we hope it'll happen here with the dc pilot it leads to an m o u between emergency authorities in a region and the and the broadcaster in terms of what they'll they'll push out the emergency manager has they have to create the content they have to make the call and the agreement might be that if there's an imminent threat like it's certainly a shake alert that that goes out without anybody in the station touches it touching it otherwise in a news station the news official on duty is going to make decide whether it goes out or not so i think it's more institutional than a technical issue i i would tell you on that by saying i think in i have two categories of thought i think one is in terms of what we're doing in terms of earthquake detection and alerting like are there other things we could be detecting i think it's an interesting question i think anytime you build a model that detects something it has precision and recall and there's lots of trade-offs that you need to think about really carefully and so i would say like there you know there may be space there to fill a gap where there's a need but i'd be really interested in being like problem-driven in that and i guess i don't have a great sense of where those holes are but would love to hear from you all uh the second thought is i think in terms of actually delivering alerts and making them useful i'm not sure in my mind it is a problem with you know the data available from a government source today about the existence of a particular watch or warning as much as it is a problem of understanding what users need and want and what will cause them to take action i think it's less in my mind about the trigger and more about the experience that we build around that and how we make sure that's that's effective and as integrated as it can be i think that's a lot of the work i would just add a hyper localization it's incredibly important having the impacts broken down to as local as we could possibly get and that's where perhaps gen ai or some things may help because obviously you can't provide house by house impacts but perhaps another service can on top of the information that you're already putting out can help um thank you so much again i'd like to extend a warm thank you to all our speakers so mike micah and john for sharing their views and having such an interesting set of talks today