Interview with Sean Griffin
Sean Griffin is an expert in disaster intelligence, data, mapping, and analytics and serves as President and CEO of Disaster Intelligence Inc.
CBD: Camilla Barker-DeStefano
SG: Sean Griffin
CBD: Sean, thank you very much for joining us.
SG: Thanks for having me.
CBD: Okay, so tell us about your background. How is it that you got to where you are today?
SG: Thanks for the question. So, I started out in emergency management as an offshoot of my time on active duty military when I was in the Navy. At the time the United States military, after 911, was going on the journey of upgrading their emergency management programs at the installation level. So from a base to base the Navy, and the Marine Corps, all the services we’re updating and integrating a new emergency management program. At the time that’s where I became interested. I then left active duty and had an opportunity to join federal service with the National Institute of Health. And because of that experience that I gained in the Navy with their new emergency management program I decided to join the Office of Emergency Management at the National Institutes of Health as their emergency management specialist. That was 11 years ago now.
And then I spent two years at NIH responsible for the campus there. Everything from the medical facility, to animal facilities, to bio safety level three laboratories, et cetera, and then moved to a defense logistics agency, was the emergency management director there for the headquarters at Defense Logistics Agency. Then went to the state department also in the office of emergency management. From there, I worked on a number of White House committees. The gentleman who chaired one of those committees went to the department of energy. I then, he asked me to join him there so I went to the department of energy as an emergency manager. And then I was detailed during my time at DOE to the White House national security council, as you mentioned a director for incident management policy. Actually the same position that Tom Roston, the gentleman who brought me over to DOE had, so I was relatively familiar with the portfolio.
After my time at the White House under President Obama and President Trump I left March of 2017, went back to the department of energy, and they didn’t really know what to do with me. So, I worked on a number of projects to include an exercise that we ran down in Texas, Houston, Texas, modeled after hurricane Ike, similar hazard characterization to hurricane Ike. And then as you know we had hurricane Harvey that year. So, the exercise was in May, we had the the event in August, and it was a new administration, a different administration. It’s one where you at the latter end of eight years of a presidency with respect to president Obama the wheels are well underway, the skids are greased so to speak, and folks really understand what they’re doing.
Well with the new administration you have folks that were just put into place Brock Long, who is a professional, a great emergency manager. He was on the job for two months when hurricane Harvey hit. Jeff Byard, who is now the administrator who has been nominated, it was his second day after hurricane Irma.
CBD: [inaudible 00:03:50].
SG: Right. So, these individuals, great men and women who decided to serve in these positions but they’re new and it doesn’t matter how many years of experience it’s still a new position, you have to get acclimated. You also have to wrap yourself into the environment. Jeff Byard came from Alabama which is where he knew Brock Long. He was I believe the operations director there. And he’s just trying to understand the federal system which is very complicated and we talked about this previously on our last call is that the challenge is all of these laws and all of these regulations it’s very difficult to navigate. And one of the things that was very illuminating during my time on the national security council staff is that I was asked to put together this transition exercise.
And so, the Presidential Transition Act of 2015 Congress said it’s a good idea to train the folks who are incoming whether it was the Clinton team or the Trump team, it doesn’t really matter, this is about America and our national security and our homeland security. And so, what we did is based on that legal requirement we put together the incoming team with the outgoing team and we train them on four different scenarios. But really we didn’t train them on the scenario we trained them on the legal aspects of their job. If they were the secretary of energy, or the secretary of state, or the secretary of homeland security here are the key authorities regulations that you need to know Mr. Or Mrs. Secretary, incoming secretary, on how to perform your duties.
Because a lot of folks forget or maybe overlook the fact that on January 20th at 12:00 PM eastern standard time all of those officials become a pumpkin like the Cinderella story. They no longer have any legal authority. And all of those incoming professionals whether it’s Rick Perry with respect to secretary of energy, he’s not confirmed yet. And I believe he was confirmed March of 2017 so a few months later. So, you have this vacuum or really this institutional knowledge that left. And so, that’s what made me more appreciate the understanding of things such as interdependencies not just with respect to physical infrastructure because a lot of people like to talk about interdependencies of power systems and communication systems and how a failure of one can impact the other and that’s pretty well understood.
But I think what’s less understood is the interdependent nature of the legal systems not just here in the United States but internationally. What we had talked about is how does the United States onboard say urban search and rescue teams that have dogs, and medicine, and licensed professionals that aren’t necessarily accepted here in the United States. And given our federalists system that the states are the ones who credential medical doctors, and paramedics, and EMTs, and firefighters, and so forth. And customs and border patrol deal with the imports of medicines and so forth.
And so, we have this overly complex system and what became very evident to Brock Long, and to Jeff Byard, and to Dan Kaniewski, and to all these individuals who are seasoned professionals, these aren’t rookies. No, they’re not. It’s overly complex. And I give them a lot of credit because if you look at FEMA’s three, the Federal Emergency Management Agency’s three strategic objectives it’s build a culture of preparedness, ready the nation for catastrophic disasters, and reduce the complexity.
The reduction of complexity is focused on FEMA but really it’s the entire system in disasters. This whole world is overly complex on things like climate policy, cyber policy, but disasters in its own domain is complicated. And whether we want to reflect on 2011 Japan and how the Japanese onboarded assistance there which was overly complicated and diplomatically challenging or whether you want to talk about Haiti and how the first world inundated that country with all of this assistance without any bounds of the law and completely threw caution to the wind on how we were going to help or fix the problems that many of these people have never even been to Haiti before. How do you get folks? In the United States we have somebody like me who lives in the the Washington beltway area. I’m not from Louisiana, I’m not from Texas, I don’t live there on a day to day basis, I don’t understand the local cultural context, and I am expected to solve these problems from a very far distance completely removed from those local communities.
So, where I am today is with disaster intelligence is that those three goals I think are salient ones. Building a culture of preparedness, we need a culture who or that embraces disaster risk and disaster risk literacy, getting better educated and understanding the local laws as well as international law on how to deal with the complexity of a disaster that impacts the community. Secondly, readying the nation for catastrophic disaster, we’ll talk about my company and what we do to help resolve that. But remember the Carrington event happened in 1812 right, space weather. We didn’t have the grid. We had the Telegraph system maybe but we didn’t have this large machine called the U.S. National or the North American grid. What happens when we have a space weather event that potentially fries electronics within our grid system.? How do we deal with that type of catastrophe when everything from the internet of things, to refrigeration, to the cold chain in medical supply? How do we deal with the failure of a large grid outage across a massive area where millions of people in urban centers rely on air conditioning and everything else that’s power dependent?
So and then, reducing the complexity. In order for us to have a prepared culture we have to understand how to communicate risk better, how to be better stewards of risk, and make it simple enough where my children or my neighbor can understand how to take action, how to have a call to action, and actually do something about the risk in their community. And then lastly, I’ll say that because of all these laws something that I have passion for, we talked about this, is peeling back the many layers of the onion and understanding what actually exists from a legal perspective and how do we refine that and create a better environment in which we can enable ourselves and self-actualize during disaster so that we can solve these critical issues. Because otherwise we’re leaving the survivors behind even though they’re the ones that we’re serving so.
CBD: Yeah, I mean that’s the background that you just explained to us is absolutely fascinating to me and I have just so many questions that have come up just in relation to your views on certain things. Because I feel like you’ve been in so many different areas of this mechanism so whether it’s a private sector, or it’s the federal government, or whatever it is and that’s given you these unique insights into how things overlap and then where there are gaps. And I think that is absolutely fascinating.
Before we start to talk about more your kind of day to day with what you’re doing with disaster intelligence. I just want to just track back to something that you said just now about the complexity of onboarding assistance. Say if the United States needed assistance, if there’s another Katrina, if there’s another Cascadia style event, the reality is I think most of us are well aware, although some don’t want to accept, but I think most of us are well aware that the U.S. Will require international assistance of some kind. And you mentioned search and rescue teams, having search and rescue teams, and the difficulties in even just getting a dog into the country. What do you think politically is stopping the conversations on that? Because we’re aware that there is a need to make this process easier, to make this on boarding process easier, and I think we’ve got the research, we’ve got the data to support that need. So, what is it politically that’s stopping us from going forward and coming up with solutions to make it easier to get assistance into the United States when we need it?
SG: I think the primary reason why this is hard or why it’s hasn’t been resolved is that we haven’t faced it yet where the public is engaged to the point where they’re asking us to solve it. I think that’s just true in public policy making all the time. It’s the topic du jour, the topic of the day, and then folks focus on that one shiny object and then we go to the other. So, I think the lack of political will is there because folks just don’t know that this is something that’s a challenge and a problem.
And so, when we conduct these major exercises and we build these scenarios the thing I like to say is that we love problem admiration versus problem solving. We’re so enamored with the challenge of a Cascadia or a New Madrid that we’re constantly like, “Well this bridge is going to be out and that bridge is going to be out. Okay, well that’s yeah that’s almost self evident.” But what we need to do better as planners and folks within the disaster management community is understand what are the discreet issues and what are we actually trying to achieve. And if we’re trying to achieve a large area, say a search and rescue type of scenario, well where is that assistance going to come from?
One of the interesting statistics from the Nepal earthquake in 2015 is that of all the highly technical urban search and rescue team members that came into the country from around the world those USAR teams only saved 17 people. Whereas the local populace, the neighbors helping neighbors, sort of speak, they saved tens of thousands of people because they’re closest to the problem and it takes time to get these teams in. And so, by the time you onboard them are they even necessary or effective? To call up Virginia Taskforce One well these people have families, they have day to day. As much as they want to go to Nepal and help respond well it takes them time, a few weeks to mobilize. And then they finally get on the plane, and they show up, and they have to integrate, into in this case the UN cluster system. And then the UN cluster system has to figure out where to place these people and where there’s the greatest need.
And so, this complexity of organization I think gets in the way but to come full circle on your question why is it a challenge is because we’ve never had to face it here. We’ve never had to onboard tens of thousands or thousands of international workers into the United States to help with a large scale disaster. Because these hurricanes as horrible as they impact these communities they’re regional events. They are much more focused on the immediate impact area.
Whereas when we looked at the Cascadia subduction zone we did the analysis using the department of energy national laboratory, modeling and simulation capabilities. Because of the impact to the bulk power system we would have rolling blackouts or at least some voltage collapse east of Colorado and south to Mexico for a subduction zone incident that happened off the coast of Oregon. That’s a significant amount of area. The United States has never ever had to deal with the large scale loss of power across that type of wide area. Now we had the 2003 blackout but that was a couple days and it was resolved. And keep in mind that was a tree that impacted a wire that created a cascading effect to cause the 2003 black out, a tree.
CBD: This just shows you how vulnerable the system is. When it comes down to the day to day this is such a vulnerable system and I don’t think the U.S. Is alone in that respect. It’s the same in the U.K. Where I’m from. It’s like you take down one power line and it’s the domino effect, what else is going to be then impacted straight after that.
SG: Exactly. And so, we lack the ability to provide the context for how complex and how complicated these situations are and because we traditionally haven’t been able to unpack all of that to elevate the discrete problem sets that will need to be solved such as onboarding international search and rescue teams. That’s just one discreet problem area. Or in Puerto Rico, the fact that we lack the visibility of understanding that there was one medical oxygen facility that supplied all of the medical oxygen to every single hospital on the Island. And, we had the hair on fire moment at the national response coordination center where they’re asking me, “Hey, how do I get a generator to this medical oxygen facility?” I say, “Okay, well what’s the load requirement? Where is it located? Has anybody done an assessment?”
These things take time. It’s not like we could just purchase a generator, and throw it at the medical oxygen facility, and it’s magically going to work. So, where do we need to go moving forward and this is why we’re so focused on these particular issues at disaster intelligence is how do we leverage technology and data to be able to lift up through the noise of problem admiration to parse out the individual discreet issues that we can assign a task force, or assign a working group, or assign a series of leaders or a group of leaders to just focus on fixing one discrete problem in blue skies.
We can’t be all things for all people but we can at least try and solve some things now in blue skies. So when the thing does happen the Cascaded, the New Madrid, the space weather event. Hopefully not but an EMP from an adversary or a coordinated physical cyber attack on our infrastructure from an adversary. If one of those things happens we don’t just want to reduce the complexity when we’re dealing with it we want to mitigate the risk now so that when it does happen it’s not such an unbounded problem that it’s ultimate chaos and we’ll never be able to recover from it.
CBD: Absolutely. Absolutely. I mean, one of the themes that I’m guessing from what you’re saying is that we are so enveloped in these multiple levels of, layers of the onion to use your words. There are so many layers packed on top of each other that one of the things that we really need to do is just sit down and say, “Okay, let’s figure this out, let’s strip this back, let’s pair it back, and figure out what the core issues are and what we actually need data ways, analytics ways, legal, all these different inputs. What do we actually need to feed into this problem so that we can actually make better decisions? I mean, do you think that’s a fair kind of conclusion. Are we taking all of this data and we’re doing all of this data analysis to be able to make better decisions?
SG: Yeah, absolutely. That’s the most simplified way to describe what the challenge is and what we need to do is we have to be able to parse these things out so that we can problem solve today in blue skies and have … It’s again, to use the Eisenhower perspective it’s not the plan it’s the planning. And so, one of the things that I’m doing or I’m trying to do is get better engaged in my local community and understand how I can contribute not just as a business owner but just as a normal person on the earth who owns a home, and has neighbors, and has family, and day to day responsibilities.
So, next week I’m meeting with my local emergency manager. I didn’t even know him until two weeks ago. I met him at the Florida governor’s hurricane conference. I said, “Oh hey, you’re my local emergency manager. What are we doing about the local hazards that we face?” And so again, I think what it requires it’s a whole of society, whole of community I guess is the term of art that we use in the United States. And when I was working with Swedish government that’s the same term that they used. The whole whole of society approach is it starts in the primary schools, it starts in the home front. We have to have a better sense of risk and risk literacy at an individual level so that we can better apply ourselves to these large problem sets.
Because with respect to whether it’s a change in climate or whether it’s a changing geopolitical world the world is not necessarily getting safer from the threats and hazards but we can do something about reducing the risk and being prepared to ensure that we have continuity of community and continuity of main street. Because without continuity of community, without continuity of main street, we’re going to lose a sense of ourselves and it’s going to be difficult to move on and unify ourselves in the sense of disaster.
And so, I believe that not to get to overly theoretical or esoteric but we are so enamored and focused on the day to day minutiae of what’s in front of us that we’re losing sight of what’s ahead. And I think that we need to as a friend of mine said take a strategic pause and really understand our environment so that we can come together as a community to better engage with it and make decisions that benefit the collective and the whole so that we can all benefit from these advancements in technologies, in the modern conveniences and everything else. Because we become so dependent and so lost in the illusion that it will be there tomorrow. It’s just a false sense of reality.
And I struggle with trying to communicate that so if there’s a call to action I suppose for myself, for whomever is going to engage with this video or what this discussion is that help me become a better communicator of risk, help make a better everybody understand how we can better communicate risk. Because I want to understand how do I break it down for my two sons where it’s difficult to just communicate to them, “Hey, you probably should eat this versus that.” That’s hard enough.
But how do I break down climate hazards, and how do I break down disaster risks, and communicate that at a simplified level that my seven year old can better understand it. And then, he can take that knowledge and translate it to his friends at basketball or on the hockey rink. And so, I think that’s where I struggle and we probably could be doing better at is understanding how do we communicate. And to me I think some of it is simple as saying storytelling. We can be become better storytellers. One of the aspects of disaster management that I think we fail to capture more often is just the story of the survivor and telling that over and over again.
CBD: Yeah. It’s powerful, it’s very powerful yeah.
SG: It’s very powerful and to use that as a way to become more engaged and more empathetic that disasters at the end of the day affect people and disaster management is all about people. It’s not about the grid, it’s not about the refrigerator, it’s not about the medical oxygen facility, it’s about people. And how do we reduce risks to those systems to ensure that people can continue to live another day and benefit from all this advancement that has taken us thousands of years to get to.
CBD: Yeah, absolutely. Absolutely. I mean it tracks back all the time to that core instinct of all of us which is survival and the survival in the instant in which the threat occurs. But then there’s also this pre survival phase where we can actually do things about our chances of being able to survive an incident. So yeah, I mean I absolutely agree with everything that you’re saying in terms of why we need our priorities to be I guess. So, let’s just narrow this in a little bit to what you are doing with disaster intelligence. Because we’ve talked about a lot of different issues in this video already whether it’s communications, or it’s political troubles, or whatever it is. So, now let’s narrow into what it is that you are doing on a day to day basis or a short term basis with your company to make a difference to these problems.
SG: Sure. So we, disaster intelligence, were just awarded Department of Homeland Security’s Small Business Innovation Research award where we will be building a power and communication system, network risk assessment model or tool. It’s not just about the model but it’s a platform in which we can again going back to what we said peel back the layers of the onion, understand how grid systems interact with communication systems. And in fact we’re less focused on the modeling because with our partners in the Argon National Laboratory or we’re seeking a partnership with Idaho national lab, MIT national lab, Lincoln laboratory who we’ve been collaborating with is that there are many models out there. And as a friend of mine said, “All models are wrong, some models are useful.” And so, the thing that we provide to that ecosystem is we’ve built a backend, data science platform in which we can ingest and analyze billions of records what we like to say at the speed of thought.
And so, because of our partnerships with Nvidia and with OmniSci in particular we’re leveraging the graphics processing unit and these really fast databases to parse through all of this data that we’ve never been able to do. In fact, a lot of the memory frameworks that we’re dealing with didn’t even exist over a year ago. These are leading edge new technologies that certainly in the emergency management community has never engaged with. Then the other great thing is it’s distributed. So, we’re doing this using web based tools in distributed computing environments so you don’t have to be at the supercomputer at the national lab. The same GPU’s that Oak Ridge national lab is using for the summit supercomputer are the same GPU’s that we’re using in a distributed environment.
And so, what we want to do is take these models and take the academic research and all the peer reviewed work that has been done in these laboratories, our university environment, and then translate that and transform that into operational code. Into a distributed compute environment in which you and I or the emergency manager locally or the emergency manager or disaster manager in Japan, or in Naipaul, or in Italy, it doesn’t matter that we have this environment where we can leverage these models and these systems to be able to understand risk in a way that we’ve never been able to do ever in the past.
So, the idea with the DHS project is we want to build a use case in which we understand one service area. So for example, just for argument’s sake let’s say we take the state of the Commonwealth of Virginia where I live, Dominion Energy is the power provider to my home. We are served by Verizon, by AT&T, by Sprint. And so, if Dominion loses power irrespective of the hazard precipitating event, it could be space weather, it could be a physical cyber attack, it doesn’t really matter, but they lose system integrity and they can’t deliver power to communication networks. How do those communication networks survive? How do they operate in an austere environment where they’re dependent on power? Because keep in mind many of the cell towers, they might have 24 hours of power, battery backup power. Some of them were run on generators, not all of them. Some of them are less than 24, maybe four hours, maybe 30 minutes. So the survivability of these individual nodes, how does that all work in concept?
So, what we want to do is build this integrated model to understand that we can run these scenarios iteratively and then figure out okay if we lose power here how does it affect communications there? And then how does that loss of communications and power impact the rest of the community or community essential functions? In the continuity sense they call it mission essential functions. So what essential functions of the community are lost, whether it’s hospitals, whether it’s nursing homes, whether it’s pharmacies, grocery stores, distribution centers, whatever. And then, we can start to understand those cascading second, third, fourth, fifth order effects. Because it’s, I believe, we believe that starting with powering communications makes a lot of sense because almost everything like what we’re doing now, talking on this computer via the web, it’s all dependent on communication systems and power. Without those two systems in place none of this is possible, none of the modern conveniences that we enjoy are possible.
So, we’re going to start there as the tip of the iceberg and then understand below the water level how power and communications impact everything else and then being able to run those scenarios anywhere in the world to be able to understand that risk. So, there are some assumptions being made. One assumption is that we have access to all the key data which we don’t today. Another assumption is that the data is not biased.
CBD: Right. I was going to ask about that. Yeah.
SG: So, we need to make sure that the data that’s being put into the model is unbiased and that we can strip the bias out. That’s not trivial. These are hard problems. I recognize that but you’ve got to start somewhere.
SG: And secondly we will employ, or not secondly, thirdly …
CBD: Wherever we’re at yeah.
SG: Wherever we’re at. Third, we will employ machine learning, deep learning applications to be able to parse out anomalies which that goes beyond my pay grade as far as intellectual capacity goes but that’s why we have folks like Jack on our team who’s PhD computational mathematics, former senior research scientist at Google, former professor who will help unpack that from a technical perspective. So, we got a good team and I’m really looking forward to it but that’s just one use case. The other use case is the university of Delaware disaster research center and a number of other universities such as Cornell University Oklahoma and UNC Chapel Hill, they’ve built advanced [inaudible 00:33:34] evacuation models which they can better identify based on the characterization of the hazard, the tropical storm, this storm surge and so forth which counties to evacuate first, versus second, versus third.
And how do you build basically a tree of decisions in which the emergency manager can understand the risks and the trade offs of whether I evacuate these counties first versus those counties or these evacuation zones versus those evacuation zones. Because right now it’s completely arbitrary based on the political will and whim of the local electorate. And that’s important and should not be a discredited. We need strong political leadership so that when the message is communicated you need to evacuate that it’s done confidently. But how can we make it more confident is we take this great research and we put it into operational code because right now local emergency managers aren’t using this. So, this goes to the overall thesis of the company is how do we unlock all of this research and scientific progress that’s basically been behind closed doors in these laboratories and these academic environments and pull all of this into a distributed computing environment so that folks like you and me can leverage these decision making tools and these models and this analytics to be able to make decisions where it’s simplified, and easy, and intuitive to understand.
CBD: Yeah. Awesome. I mean congratulations on that. It’s a fascinating project and I think the value of it is really only going to increase, and increase, and increase the more we get deeper into this data. So, and into these kind of decisions and these issues about what we’re actually, where we’re actually going in the next, not just 10 years, but 20 years, 30 years. I mean it’s really, really fascinating stuff. I feel again like we could probably talk about this all day long if we were left to our own devices but I know you’re busy and I’ve got a dog that sounds like he wants to go out. So, I think that’s probably a great place to stop so thank you so, so much, really thank you so much for the time that you have spent with us today. I am 100% sure that people are going to find this to be fascinating eye opening stuff. So, thank you very much.
SG: Well, wonderful. I’m appreciative of your time. Certainly. It’s mutual and thank you so much for the opportunity.
SG: If folks want to reach out to us it’s info, I-N-F-O, at disaster-ai.com.
SG: And also see us at disaster-ai.com on the internet. Please reach out to us. We would love to provide you access to our tools and to give us feedback because ultimately this is to serve the communities. We’re a public benefit corporation. We’re looking to provide tools in which are user friendly and meet the practitioner’s needs so the more people that we can get eyes on it, and engaging with it, and providing us feedback and the better chance that we have to deliver tools that meet your needs and ultimately the community so.
CBD: Absolutely. [inaudible 00:37:00] make sure that all of your information, all of the information about the company, or an email addresses, I will make sure that all of this is on there ready and accessible for anybody that wants to get in touch with you. Definitely.
CBD: All right. Awesome. Well, thank you very much.
SG: All right. Take care.
Want to see the full video interview for free?
All interviews featured on DISPATCH are available in video form at the Crisis Academy Insights Library. To access the video, simply visit the Crisis Academy website and sign up. It is 100% free and only takes a few seconds.
Interested in learning more about disaster intelligence?
We regularly post news and analysis that relate to disaster intelligence. Check out our Technology category for more interviews and articles like this.
Please note all interviews are transcribed by humans, and we humans make mistakes. If you have noticed one, please accept our apologies.