Terry Gerton Denice, let’s start with you. We are entering peak hurricane season here, and on Saturday actually will be the anniversary of Hurricane Katrina. So it’s a good time to think about what’s going on with weather data. And most people think about forecasts and weather models, but from a federal operations perspective, when we’re talking about weather disasters, what kind of data becomes more important once a storm is approaching?

Denice Ross When a storm is approaching, and I know this from personal experience having lived in New Orleans for many years, and including that Katrina time period, data accuracy is the most important thing. And there are two things that federal data really focus on. One is what’s coming and who will be impacted. And I just wanna give a couple of examples here, some data that help us get more accurate forecasts, I just learned that the saltiness of the ocean gives insights into how quickly a storm will intensify. Who would have known? And the Argo fleet of floating buoys out of NOAA provides real-time information on water salinity. And why that matters is because that cone of uncertainty about hurricane landfall, timing and strength. That knowing the saltiness of the water allows that model to be less uncertain. And that has an impact on real people because then officials and individuals can make better calls about whether to evacuate or shelter in place. And that’s especially important for people with disabilities and the elderly, for example.

Terry Gerton Weather buoys, who would have thought?

Denice Ross I know, who would have thought? Another example that’s really important about knowing who’s in the path of the hurricane. Right now, there’s a lot of talk about changes to the census that might impact response rates, especially from people in mixed immigration households, for example, and also the demographic details of people living in a specific area. And emergency planners need to know how many people are in harm’s way and what their characteristics are so they can get them to safety. For example, when I worked in local government at the city of New Orleans, we use data from the American Community Survey on households without access to a vehicle to determine how many evacuation busses we needed to get and what the ideal locations for pickup were. This makes sure that there’s a seat on the bus for everyone who needs it. But if there’s as a significant under count for any reason, there might not be enough seats.

Terry Gerton Chris, Denice has just talked us through a couple of data sets that you might not think about when you’re thinking about a hurricane inbound. What about some of the sets that will be essential as planners deal with the aftermath of a hurricane?

Christopher Marcum Oh, that’s a great question. So one of the data sets that we think about that’s not quite one of those hurricane data that we typically think about that is really essential for both the planning and also the response further down the road are Medicare data. So we typically think about Medicare data as being about medical expenses, but they have an off label use during hurricanes. One of the most at-risk populations during emergency response that we really need to plan for in advance are individuals with health conditions like emphysema, ALS, or kidney disease. And they often, those individuals rely on at-home medical equipment that plugs into the wall. Ahead of a hurricane, planners can use the Health and Human Services emPOWER data to map projected power outages to outer zones and to help preposition mobile generators or staff or medical shelters that can send direct outreach to those vulnerable residents so that no one who happens to be relying on self-sustaining medical tech in their home gets left in the dark.

Terry Gerton Denice, there’s a lot of things to think about here and recalling that you were on the ground iin Katrina, you’ve got that firsthand perspective. As agencies move from response to recovery, how do their data needs change? What information needs change over time?

Denice Ross Well, federal data are obviously relied on ahead of a storm to take actions to protect people and property. And then during the storm, the focus shifts more toward the local emergency operations center that’s managing the response. And they’re taking in lots and lots of local data at a rapid pace. And of course, they’re still relying heavily on federal weather data. And they probably are also using data from different federal agencies like Department of energy on the locations of critical infrastructure. And then as things shift to recovery, the feds start assessing the impact of the disaster so that they can deliver resources where it’s needed. And we usually think of recovery in sort of short-term and longer-term. And during that short-term period, damage assessments create a lot of data. In some cases, that comes from FEMA contractors walking the streets and documenting damage to buildings. There’s a lot in Intel about community needs that come from forms and applications such as FEMA’s individual and housing assistance programs that include demographic details like age, whether you own or rent your home, what your income level is, and whether you had flood insurance. And then for the longer term recovery, that’s where it gets really interesting because there you’re trying to target federal funds in a way that will make people whole and build future resilience. And so an interesting collection of data sets comes into play at that point. Three examples from three very different agencies include HUD, NOAA, and the Fish and Wildlife Service. So, HUD, one of the most creative data sets I’ve seen come out of that shop is that they have something they call the U.S. Postal Service Address Vacancies Data. It’s an agreement that HUD has with the Postal service to aggregate data on how different neighborhoods are recovering from a storm over time. They do this by measuring vacancies, but most importantly, housing units that are uninhabitable. The postal service gives them the title of no stat. And so a no stat house is one that cannot receive mail because it’s too blighted to live in. And so this data is useful in highlighting parts of town that need extra assistance in the recovery, and it helps local government and nonprofits target redevelopment grants and housing assistance programs where they’re most needed. Another data set is from NOAA, and that’s their sea level rise data. And that shows how specific properties might be impacted by rising waters. And so, for example, this would be used by the owner of a coastal business that flooded so that as they’re rebuilding, they know how high they should elevate their new HVAC. So when the waters come, they don’t have to replace the HVAC again. And then another example is the National Wetlands Inventory out of the U.S. Fish and Wildlife Service. And that’s really essential because wetlands are the first line of defense against incoming water and storm surge. And these data are important for municipal planners and also Army Corps of Engineer projects to identify natural buffer zones that can absorb that storm surge, and it allows states to prioritize coastal restoration projects that can shield low-lying residential areas.

Terry Gerton Denice Ross is the director of the Data Policy Institute at the Federation of American Scientists, and Dr. Chris Marcum is a senior advisor for federal data policy, also at the Federation of American Scientists. Chris, let me turn to you. We’ve been talking up to this point about data that planners use to prepare for and respond to weather disasters, but Congress and the administration have both raised concerns about how long disaster assistance takes to reach survivors. How much of that is a data problem?

Christopher Marcum Yeah, one of the big challenges in disaster prep and response is knowing in advance who’s at greatest risk of being left behind. Agencies really do need to do a better job at collecting and responsibly sharing disaggregated data for that purpose. And there was a 2021 Government Accountability Office report that emphasized how FEMA, Department of Housing and Urban Development, and the Small Business Administration need to work together more effectively to improve data quality and coordination for the disaster prep and response. And that challenge still remains today, even though it was a 2020 report. Most often though, the data really do help shine a light on some of these problems. And an example on how the rest of us might sort of look at sort of accountability of disaster response planning and prep is through the usaspending.gov. USA Spending makes it really easy to see the gap between money the government promises and the money it actually delivers before or during and after a disaster. Data allow you to track what’s known as obligations, that’s the money that they promise and the outlays, or the money they’re actually spending by agency, by the location and program of delivery. And during Hurricane Maria, which affected Puerto Rico, FEMA, HUD, and DOE (Department of Energy) had obligated about $14 billion for recovery and modernization, which we know through USA spending, but they’d only spent about $3.5 billion during that response. So in other words, like three quarters or so of the promised funding hadn’t actually reached people in need. And that information can be used to identify types of situations where there are bottlenecks in funding programs that Congress might want to respond to or the administration can fill, and it can really help modify the design of programs to help getting that money flowing more efficiently.

Terry Gerton Denice, we’ve talked a lot of here about the data that agencies use during disasters, but how resilient is the data ecosystem itself? I mean, what happens when a critical source of information isn’t there when responders expect it to be? Maybe it’s been taken out by the storm as well.

Denice Ross Well, normally, in everyday situations, disruptions do happen to data that may lead to gaps in research or less trustworthy statistic or inconveniences. But in a disaster, data disruptions like that can be life or death. And we have been monitoring the health of individual data collections at dataindex.us. And there is certainly a need for shoring up the resilience of the data ecosystem, from collections to sharing to privacy protections. And we do know that even before the disruptions of the last couple of years, the federal data ecosystem was not well equipped for the data needs for the rapid changes that happen during a disaster. For example, when a storm hits, officials need immediate infrastructure data to coordinate evacuations or locate critical assets. And the Department of Homeland Security has had a data set called HIFLD Open. It’s the Homeland Infrastructure Foundation level data, not the best acronym in federal government. But it was a tool that brought together critical geospatial data from many sources and I like to think of it as like a digital go-bag for emergency managers at the local level. And it had information about locations of resources like hospitals and fire stations, hazards like nuclear facilities and concentrated animal feeding operations. If a storm hits them and disrupts them, then toxic stuff might get into the environment. And it also has information on vulnerable populations, like people in nursing homes and where the childcare centers are. During emergency responses, local emergency operations centers relied on HIFLD Open data, and especially those who have lower capacity, like, they don’t have maybe a full GIS team, for example. So they really depend on this digital go-bag that DHS had collected. Fortunately, civic tech groups restored access to the tool and have created sort of a next version of it. So there is still a place that local governments can go. But my concern is, is that they may not be checking to see if it’s there. They’re just, you know, like most of us, we take these federal data for granted. And then when you need them and they’re not there, the stakes can be really high in a disaster.

Terry Gerton And Chris, then, looking ahead to the next major storm, even though FEMA has predicted a less volatile hurricane season, where is the biggest opportunity to improve disaster response through data?

Christopher Marcum Yeah, that’s true. Although if we look over to the Pacific and at Hawaii, they’ve certainly had their fair share of runs with big storms this season. We live in a world that relies more heavily on the internet, on satellite, on cellular connectivity, and all of that connectivity is about sharing and access to data and information. So I do think that we need better technological preparedness for disaster response in a way that allows emergency managers to leverage all those resources, including a large tranche of private sector data in responsible ways that can help with prep and response. For instance, data collected through the GasBuddy app can be used and has been used to help reveal which gas stations are open and supplied during a disaster. And that is in contrast with how slower federal process might unfold with the Energy Administration, calling gas stations to ask them if they’re open. That GasBuddy app, the private data, will help fill that gap in a more real time technical capacity.