Before becoming administrator for the Federal Emergency Management Agency in May 2009, Craig Fugate was a customer. As director of the Florida Division of Emergency Management, he oversaw the state's response to many hurricanes. Since coming to FEMA, he has responded to disasters such as the tsunami in American Samoa and the massive floods in Tennessee last year. Fugate spoke recently about the need for emergency data feeds, how social media can play a role in disaster response, and his vision of a future that includes a proactive, location-based warning system that contacts cell phone users in harm's way and provides detailed instructions on what to do.
In what ways does FEMA use IT to accomplish its mission? We're trying to change the way we're using technology. A lot of the things we were doing are no longer relevant, particularly when we look at things like GIS. For a long time, the attitude here was, "Well, we use GIS to print maps." You have to change the mindset of the managers that the tools are a lot more powerful than that. That's where we are now, just getting people to understand it.
We have a Web page, but what good does it do in a disaster if people don't have Web access? But they may have a smartphone. So we did a mobile Web page last year. It's really a simple page. If you're in a disaster area, you don't want to see our org chart and you don't care about our mission statement. What you want to do is find the information pertinent to the event and how to prepare against that threat. So we did a very stripped-down, low-bandwidth page that has pertinent headline information about the big event that's occurring.
FEMA is now pushing out information through social media channels such as Twitter, Facebook and a blog. Are you looking to have a conversation with people during a crisis, or just feed them information? We don't have answers, but we can't wait for answers. Nobody has come up with a blueprint that says this is how social media must and will be used in all disasters, because it changes fast. We're trying to figure out how to get into conversations with the public without getting into one-on-one transactions, which would be next to impossible. We do get tweets back and we do try to respond to them, but that's very difficult in a disaster. It's still a learning process for us.
It may be more practical for a mayor or a local responder. But for FEMA, it's more about -- is this disaster and the information coming out fitting up with what we know and do we have resources to address these issues? Are there things coming up we need to be anticipating? And where are the gaps? In our blog we started up recently, we were looking for a way to get information out there without being constrained by the traditional, "here's a press release, look at what FEMA is doing" approach. I'm still concerned that our blog looks like, "here's what we're doing," but at least we're trying to get information and links and talk about things beyond what you would normally get in the press release. With Facebook, we are allowing people to post there. We don't censor that piece of it.
Prior to your appointment as FEMA administrator, you were director of the Florida Division of Emergency Management. You were there when Hurricane Katrina hit. Has the role of technology -- or how FEMA used it during that disaster -- changed since those days? FEMA came in with its own systems, and the rule was you had to adapt to its systems. FEMA had always been Brand FEMA and didn't really look at the partnership -- that they were working with a team.
How has your traditional Web site changed? We had the Web site, but we didn't have a strategy. We had stuff all over the place. Last fall one of the state emergency management directors' criticisms was that on our Web site we had places where policy guides and policy documents were inconsistent. Everybody had their little bit of the Web site, but nobody really owned it. So I'm like, "OK guys, this is unacceptable." So we went back and scrubbed our Web pages, which had grown over time. We made people take ownership for those pages, and if we had orphan pages we took them off.
We finally got all of our policies updated in one place. You might say, "You're just now getting around to it?" Unfortunately, the answer is yes.
Does everything still have to be done FEMA's way? We're trying to make sure when we go into a state that we're supporting the governor and the team. If we can bring tools to the table that they're not using, we'll do that. But if they already have something, we're not going to come in and say you have to use ours.
We had a lot of data that we would routinely put out as PDF files. Now it's a data layer, so you can just pull it into your GIS system. I had four different hurricanes in Florida in 2004, and if I wanted to get a map of the declared counties, I could go to FEMA.gov and get PDFs. But what I really wanted to get was a GIS layer that I could bring into my map so I could have it in my situation, my common operating picture. It would have been a lot cleaner if I had those as a GIS layers because some counties were hit as many as three times by different hurricanes.
Are those PDFs and GIS layers available to the public or just state and local agencies? Whatever data feeds we're putting out there people can look at. If you have these maps are you also interested in getting public feedback onto those maps based on what people are seeing in a disaster area? Yes, but we have to be very careful in the federal government because of the perceptions of abuse of collecting private information. What we're looking at here is not private information but people posting in a public forum.
We're looking at how to take that information and locate it in a way that it gives you a reference to what you know and what you're looking for. ESRI built a couple of maps of disaster response where, for example, they've taken wildfire information in data layers on a map but also social media tools you can click on that show where people are posting in Facebook or Twitter or other social media. They're using the hashtag "wildfire" and it shows where those [geocoded] Tweets are coming from on the map.
How is that useful to FEMA? Once we're there and working in a disaster area this is not as critical. But let's say an earthquake occurred. We get initial reports from the USGS, but it takes hours to days to get data points about how severe the damage is. But people are already posting that this bridge is out, that road is closed.
In the first minutes or hours, social media gives you a better idea of where the worst areas are and the magnitude, so you're not waiting to make that response. If we're not hearing from areas that also may mean significant damage. When you start with government, they're going to start not with the calls they're getting but what they can immediately see and feel. But the public covers a bigger area than that during these events.
What are you doing now to facilitate this vision of public reporting through social media? Right now it's a lot of exploring. I don't think there's a definitive roadmap of how to do this. This isn't something FEMA is doing so much as something that we are participating in. You're seeing a lot of local and state emergency managers looking at how to use social media with volunteers, the tech community, social scientists and others. One hashtag that's pretty active about this is #smem.
How do people know what hashtags to use in an emergency? If you can identify a hashtag that makes sense for the event and people start using it, you take all of that energy and produce a firehose of data. We were doing this with the tropical storms this year, and we talked with the hurricane center about how to standardize hashtags. It turned out that the public standardized it for us. They just took the name of the storm with a hashtag in front of it.
But those are common names. How's the signal-to-noise ratio when you search on those hashtags? With Hurricane Earl, you could get some interesting cross-posts, but everyone who was posting about Hurricane Earl was doing #earl. People were posting about evacuations, updates from the Hurricane Center, evacuation routes in North Carolina, about traffic.
How can people make those hashtags more specific -- and more useful? You can do this in such a way that, for example, everyone in Florida tweets the storm name and the hashtag #FL. If you want your state to know what you're sending up, do that [and] then they can follow you and see your reports.
Is everyone at FEMA on board with social media? We still run into the naysayers who say you can't trust the public, you can't do this, you'll have bad outcomes. I've always looked at it the other way: It's going to happen, and you need to deal with it.
Are you planning to use cell phone signals to locate victims of future disasters in the U.S.? Yes. One of the lessons learned in Haiti was that a lot of survivors were merely trapped in the rubble. They weren't injured, but they were trapped for relatively long periods of time -- far beyond what we had seen in many other types of earthquakes. But if you have cell signals, you've got a better chance.
How can FEMA use technology to be more proactive? Two things are happening. One is the evolution of the emergency alert system under the Integrated Public Alert and Warning System. It's in the implementation phase, and commercial cell phone providers will have to provide emergency alert warnings to customers based on the geographical location of those warnings. FEMA and the FCC have been working for some time to provide geocoded warnings that cellular carriers would release based on the locations of their cell towers. This has been in process well before social media got going and is just coming to fruition.
The next piece is, if you're in a hurricane warning area and an evacuation order occurs, wouldn't it make sense that your phone could tell you you're in a warning area, and this is where you need to go, here's turn-by-turn directions to the nearest shelter, all without having to go to different Web pages?
It's always based on what that threat means to you, and the directions are specific to what your needs are -- not something so generic that it may not even apply to you. With broadcast information, people don't always hear which one they need to do.
Who's going to develop that? The way to do that is to produce data feeds that would show that information and look at how to build apps or get third parties to build the apps.
How tightly to you want to partner with social networking sites like Facebook? I've talked with Facebook and Twitter. They are looking for a data feed. Recently Facebook signed up to provide Amber Alerts. Those are the kinds of things we're talking about. If we just produce the data feeds, we won't preselect how people use it with social media.
Young people reply less and less on voice communications on smart phones. Are FEMA and its partners ready for a world in which people call for help through things like texts and Facebook updates? The FCC is starting that discussion with text messaging. When the 911 system was created no one ever envisioned social media. So the question is how can you take that information and put that in front of a dispatcher? Text messaging is a little more straightforward than, say, Twitter or Facebook. But if this is the way people are communicating -- and particularly in low-bandwidth situations, it may be the only way they can get through -- it makes sense to look at how that would get routed, how we'd get there, and on the other end, do they have the equipment to receive it and act upon it?
You've had some interactions with volunteer groups such as Random Hacks of Kindness. How has that worked out? They said, "Give us a challenge," and I said, "[In a disaster], wouldn't it be really great if you had one app where I could just hit a button that sends out 'I'm OK,' and it would send the information to all of my lists, whether updating my Facebook status or Twitter or sending a text? Just one button, one app, boom." So they did that.
How does that work? It's like a Web page aggregator they built that someone can use to build an app for an iPhone or Android phone. There are some folks who have built those "I'm OK" apps.
So the second thing is, what if I build an app that says "I'm not OK"? We really want to give people a "how to prepare" app. So if something is happening, what steps are needed to make sure I am protecting myself or my family? So these are things we're looking to build or get someone to build.
We need to ask what kinds of things would be useful to people and then build it in such a way that it's the least amount of effort for the person using it, because in a crisis you don't want to have to get out the instruction manuals.
Today, if you were to use an "I'm not OK" app, your friends would know. But would FEMA or your first responders? Not necessarily. And quite honestly, letting FEMA know you're not OK is not useful. FEMA is not a response agency. We're in a support role to governors and local responders. You want those calls for help to go to local responders. We want to build a system that will give that information to the closest unit. FEMA also wants to look at that so we can anticipate if we have enough resources in the area.
In the aftermath of Hurricane Katrina, FEMA had problems with fraudulent claims for assistance. How are you using technology to reduce fraud? Previously, when people applied for assistance, we took their information and determined whether they were eligible. If the information was bogus, we didn't have an immediate way to validate that. After Katrina, we tapped into credit history resources to verify employment and addresses to make sure there was an actual home there.
We can be dealing with tens of thousands of applications a day, so if we're not using technology, you end up slowing down the process. We use these tools to quickly identify the most obvious ones that have questions without slowing up the people who need help.
In the future, how do you envision FEMA operating, and what role will technology play? I want to tear down the walls between what we're doing with our data and our partners and the public. There's privacy issues if I'm collecting information about you for assistance. I need to keep that protected. But I'd like all of the things we're doing in trying to help people be much more transparent.
The other part is to make sure we have data feeds in a standard format so we're not seeing data entered at multiple points repetitively. We get information from the closest place where it's shared, so we don't have to call people to ask, "How many shelters do you have open?" As I update my shelter information, I'm publishing it, and you should see it in your GIS maps. Those are the kinds of things we need. FEMA can't do all of it, but at least we can be the catalyst, the grit that makes the pearl.