Innovation in Disaster Prevention: Advancing Technology for Prediction and Response transcript

How are you? want toach outd test. Okay, that's fine. >> I'm goingity. hearing room. [music] >> Test test. >> Yeah. I really don't think it's time recently. So What are the CEO right now? You're I'm in the midst of a runaround set up through here and so I was just kind of I'm trying to fix this one. >> It's still like loose. Test. Test. You're not talking. Um yeah. Good morning everyone. Sorry for the tardiness. We had the vice president across the street.

Took me a couple of minutes extra to get over here. Uh the committee on science, space, and technology uh will come to order. Without objection, the chair is authorized to declare recesses of the committee at any time. I would like to now recognize myself for five minutes for an opening statement. Welcome to today's hearing, innovation in disaster prevention, advancing technology for prediction and response. I want to thank our witnesses, every one of you, for being here this morning to discuss how emerging technologies can help protect our communities from natural disasters.

This committee has a long history of examining how science and technology can protect American [clears throat] lives, and few missions are more urgent uh than helping our communities better predict, prepare for, and respond to these events. Just last year, in my home state of Texas, we experienced catastrophic flooding in the Texas Hill Country that claimed more than 130 innocent lives, including dozens of children. In the aftermath, Texas invested millions of dollars in new flood warning infrastructure, including river gauges, outdoor sirens, and a realtime warning system.

When flooding struck the region again this summer, uh county officials reported that this new system performed as intended and no fatalities occurred. This is a powerful reminder of what just what is at at stake here. When communities have better information and more time to act, technology can and does help save lives. Of course, floods are far from the only threat that we face. My district has been hit by hurricanes, tornadoes, floods, ice storms, and an occasional wildfire.

Across Texas, we also face drought, earthquakes, and even landslides. Pretty much everything except volcanoes. And thankfully, we don't have to worry about those quite as much. Fortunately, our ability to understand and respond to these hazards is advancing rapidly. Artificial intelligence can improve forecasting and help emergency managers make sense of enormous amounts of data. Advanced sensors can monitor changing conditions in real time.

Satellites can provide detailed information from areas that may be difficult or dangerous to uh to reach. Autonomous systems can help assess damage and support search and rescue operations. We are well beyond the days when spotting a wildfire meant relying on somebody who was sitting up in a fire tower looking for smoke. Today, cameras equipped with machine learning technology can scan continuously for signs of fires, while satellites can detect and track fires from space.

In 2025, along groundbased systems detected more than 900 fires before they were reported by the public. But developing these technologies is only part of the challenge that we face. A forecast is only useful if emergency managers receive it in time. A warning only works if it reaches the people who are in harm's way and gives them the information that they can understand and act on. The last mile between collecting and analyzing data and getting actionable information to the public remains a very weak link in our disaster response system.

Gaps remain in message clarity, geographic precision, connectivity, and the ability to turn warnings into timely protective actions. Today, we will hear directly from the experts working to close these gaps. We will examine developments in artificial intelligence, autonomous systems, advanced sensing, commercial satellite capabilities, data analytics, and next generation communications. And importantly, what it takes to move promising technology out of the laboratory and into the hands of emergency managers and our first responders.

This means identifying the barriers standing in the way. Lengthy contracting processes, fragmented funding streams, and difficulty integrating new technologies into existing systems can all slow down this deployment. And even when the technology itself is ready to go, this committee has already taken steps to address some of these challenges. Last year, we advanced the weather act reauthorization act to strengthen weather research and forecasting, expand the use of commercial data, and improve how critical weather and flood flood alerts reach the public.

And just this summer, we advanced legislation to modernize the National Earthquake Hazards Reduction Program, including efforts to improve earthquake monitoring, early warning, hazard mapping, and resilience. We must continue that work by breaking down these barriers, supporting critical research, and helping ensure American innovation translates into practical tools that protect protect lives and property. Today's hearing will help us to build on these efforts and identify where emerging emerging technologies can make an even greater difference.

Again, I want to thank our witnesses, each one of you, for being here today, and I look forward to your testimony and our discussion. And with that, I yield back the balance of my time and would like to now recognize the ranking member, uh, Miss Lofrren from California for five minutes. >> Well, thank you, Chairman Babin, for holding this hearing and to the witnesses, uh, for coming today to share your perspective as well as your expertise.

Uh, the topic, uh, of this hearing is timely. Noah announced, I think just uh yesterday or the day before, that we have just experienced the hottest summer on record and a historic uh super El Nino is expected to bring even more extreme uh weather events this fall and uh winter. Climate change is intensifying natural hazards around the nation and we have to prepare uh by investing in our capabilities for forecasting, detection, monitoring, warning and response.

Now, I come from California and I know how quickly natural events can escalate into hazards that threaten public safety. We have wildfires, earthquakes, landslides, drought. uh they all pose ongoing risks uh to our state and to my constituents. Just last February, uh an atmospheric river occurred at the end of a severe fire season in my district. Torrential rainfall uh that was into the recently burned uh areas. It caused landslides, debris flows, flash flooding across the region. forecasting models failed to indicate the storm's severity and where it would make landfall.

Uh that limited the evacuation time. Emergency officials should not have to resort to giving door-to-door evacuation notices. We have to improve our ability to detect and communicate risks before they become active hazards. Science and technology are fundamental to addressing the new reality of extreme hazards. Our current hazard prediction and alert systems uh were built uh through inter agency partnerships and they leverage institutional expertise and federal tools and data to inform risk management.

Now industry is using AI and machine learning to improve our current detection systems. That's important. Uh but dedicated research helps us to understand how natural hazards occur, who they impact and how we can reduce risk and improve response. Translating the science into operational tools into forecasts and warning systems is going to require collaboration across government, uh academia, the public, and of course the private sector.

We've seen the power of these partnerships in California. CalFire is a leader in fire innovation adopting next generation wildfire tech uh wildfire technology like AI powered camera networks uh to enhance firefighter safety and efficiency programs like fire sense at NASA as which has really been spectacular uh can also serve as a model by leveraging technical capabilities at our research agencies to directly support the operational decision decision making of our federal and state wildfire management agencies.

It's good to hear from Dr. Falcowski who previously led the fire sense project at NASA as about how we can move these technological innovations to support operational agencies as well as decision makers. I recall so well a meeting I went to just about a couple years ago where everybody who had a stake in in fire prevention was there and it was clear that somebody needed to take the lead for coordination not taking over the response but to be a hub for making sure that information uh got uh out.

There are many successful federal programs and I am concerned about the trajectory of research and recovery uh efforts because of the extreme cuts uh and I was a partisan decision-making that the Trump administration has engaged in. The administration weakened the uh federal emergency management apparatus across several key agencies including NASA, NSF, Noah, NIST, and FEMA by defunding valuable programs [clears throat] and research are pushing out agency expertise on these uh issues.

In the past 19 months, the president slashed fundamental research that informs our hazard forecasts, haphazardly [clears throat] uh consolidated federal firefighting resources, withheld disaster aid on the basis of partisanship. It became clear to me that public safety is apparently not his priority. I'm most concerned about instances where the Trump administration has withheld disaster response funding from states that have elected Democratic leaders.

Last year, the administration withheld important wildfire funding from Los Angeles after the most destructive wildfire in the city's history. This past February, the administration overruled their own regional FEMA and Noah offices to deny hundreds of millions of dollars in major disaster aid to four Democratic le states, New York, New Jersey, Massachusetts, and Rhode Island. We can't allow federal disaster response to become partisan.

Doing so will certainly lead to a race to the bottom that neither Democratic nor Republican le states can afford. So, at this point, Mr. Chairman, I yield back the balance of my time and look forward to hearing from the witnesses. >> Thank you. Our first witness today is Mr. John Gale, uh the head of the United States government affairs at PanoAI. Thank you for being here. Our second witness is Miss Melissa Roberts, founder and executive director of the American Flood Coalition.

Thank you. Our third witness is Dr. Ali Mustafabi. Is that the way you pronounce it? Okay. Uh, founder and CEO of Resilitics AI. Thank you. Our fourth and final witness is Dr. Falcowski, a lead scientist at Earthfire Alliance. Thank you. I'd now like to recognize Mr. Gail uh for five minutes to present your testimony. You're recognized. >> Thank you, Chair Babin, Ranking Member Loften, members of the committee. Appreciate the opportunity to testify today.

My name is John Gail and I serve as head of US government affairs for Pano AI, an American technology company focused on rapid wildfire detection and real-time situational awareness. Prior to joining Pano, I spent more than 20 years working in public policy and government relations, including most recently at the US Department of the Interior Bureau of Land Management, Congressman Herd's district. That experience taught me that addressing our wildfire crisis requires us to think about the entire life cycle of fire.

We need healthier, more resilient landscapes. We need to reduce hazardous fuels and increase the pace and scale of active forest management. We need better science, better modeling, and continued investment in emerging technologies. The good news is that we already have proven enterprisegrade solutions being commercially deployed across the country today. To deploy them at scale across federal lands, however, we need a true public private partnership pursuing two objectives simultaneously.

First, keep investing in research. Pano applause committee's leadership in disaster prevention. We must continue advancing visual monitoring, AI, satellites, autonomous systems, predictive modeling, communications and sensing. These capabilities have enormous potential not just for wildfire, but as part of the multi-hazard approach to disaster prediction and response, the very thing this committee is examining today.

And second, create a much faster pathway for proven commercial technologies to reach the people who can use them today. Too often the barrier is no longer technological. It's institutional. Fragmented procurement, separate funding streams, duplicative testing, and acquisition processes that can take years, even when a technology is already operating successfully everywhere. We spend billions every year fighting fires after they grow.

But spending just a few million dollars to deploy technologies that spot small fires before they they grow has been a challenge. We need to better align those investments. And just as Congress has given the Department of Defense flexible contracting tools, like other transaction authority to accelerate innovation, we should consider giving our federal land and wildfire agencies similar tools to rapidly evaluate and deploy proven technologies.

Let's move research priorities and technology adoption in tandem. The Department of Defense has long operated from a simple premise. When we send American warf fighters into harm's way, we should equip them with the best technology and intelligence available to accomplish their mission. America's wild and firefighters deserve the same. Wildfire is unique among natural disasters because we routinely put human beings directly in the path of danger.

Yet, too often, firefighters and incident commanders still begin their response with incomplete information. Where exactly is the fire? How large is it? How quickly is it growing? Where is it predicted to grow? What resources should we send? Those are questions technology can answer today. At Pano, we combine artificial intelligence, advanced cameras, near infrared sensors, satellite data, and trained human analysts into a real-time wildfire intelligence platform.

When our technology identifies a potential ignition, our analysts verify it before an alert goes out, providing responders with visual confirmation and precise location information before crews even reach the fire. Today, Pano protects more than 70 million acres across 18 states with more than 1500 fire and emergency response agencies onboarded to our platform. Notably, much of that adoption is being driven by utilities, states, and local governments.

That creates an extraordinary opportunity for the federal government to leverage infrastructure and technology investments its partners are already making and create a shared operating picture across jurisdictional boundaries. Because in wildfire, minutes matter and small fires are less destructive to our communities and cheaper to contain. Our analysis of federal wildfire data estimates that every dollar invested in pano technology can produce approximately$ four dollars in avoided federal suppression costs alone.

Research and deployment should reinforce one another. Deploy what works now, learn from it, integrate the data, continue improving it, and build the next generation of technology on top of that operational foundation. If we can bring the same commitment to technological superiority that we expect for America's war fighters to the men and women protecting our communities and natural resources from wildfire, we can keep more fires small, reduce suppression costs, protect communities, and send firefighters into dangerous situations with better information so they get back home to their families.

They deserve nothing less. Thank you, and I look forward to your questions. >> Thank you. And our second witness today is Miss Melissa Roberts, [clears throat] founder and executive director of the American Flood Coalition. Uh you're recognized for five minutes. >> Thank you, Chairman Babin, Ranking Member Lofrren, and members of the committee. Thank you for the opportunity to testify today. The coalition has over 500 members from more than 20 states who are on the front lines of flooding.

So, cities and towns, local and state- elected officials, and civic, business, and military groups who've come together to protect their communities from flooding. We're a nonprofit that's fully philanthropically funded, so we don't charge dues, and we don't pursue or receive government contracts. And that's so that we can be unbiased advisers in situations like this one. While we're the flood coalition, our work extends far beyond flooding because the communities we work with are facing compounding threats from drought, heat, wildfires, floods, and landslides.

No matter the hazard, our current disaster system is a disaster. We've counted over 130 programs that fund disaster recovery and resilience across 20 different federal entities. This fractured landscape makes it particularly difficult to harness the technological advances we're here to talk about today. So, I have four recommendations for your consideration. My first recommendation is to create an innovation center where we can bring the best data and modeling to bear to identify exactly where we can best reduce risk.

We're seeing more devastating disasters across the country, including in many of your districts, but we still can't answer the basic question of where to build projects to save the most lives and avoid the most damage. Historically, data collection and modeling has been costly, slow, and siloed by hazard. But this is a problem technology can now help us solve. We have the ability to overlay different data set to water quality, floods and droughts to find the best areas to intervene.

But to do this, we need a federal home for disaster innovation. This could fit at NIST be a new ARPA program or a project within the national labs. And while this would require investment, in the last decade alone, we've spent more than half a trillion dollars in taxpayer money on just the federal portion of disaster recovery. And it's much less expensive to pay to reduce the risk of disasters than to recover from the destruction.

My second recommendation is about making federal data accurate and AI ready. There's a fundamental responsibility to ensure that federal data sets are updated, forward-looking, and integrated because this data dictates how we design and build infrastructure all across the country. When data is outdated, every pipes a little bit too small, every bridge is a little bit too low. Updating rainfall data through Noah Atlas 15 is an important step, but we need to ensure that other hazard data sets are also updated.

We should also build on the committee's great AI ready federal data guidelines act to ensure that we can turn data into insights. My third recommendation is to leverage AI, drones and satellite imagery for rapid disaster damage verification. The US leads the world in innovation, but when it comes to how government agencies operate on the ground, especially in the aftermath of a disaster, we're stuck in the past. After a disaster, one of the key functions federal and state governments must do is verify disaster damages to start the flow of aid to communities that are desperately in need.

Right now, we do this through what's actually called a windshield assessment because people are sitting in their cars driving through impacted areas that are often inaccessible and unsafe trying to figure out visually what the preliminary estimate of the damage is. And this slows down the recovery process. It's maddening that this is our approach when we have satellites that can literally see a manhole cover from space.

This is exactly the type of challenge that a data innovation center like the one I mentioned could directly address. And my fourth recommendation is to reform federal procurement processes so we can actually unlock the technology I've suggested. Our current processes are slow, rigid, and riskaverse. They're based on multi-year RFP processes that lock in technology and vendors for 5 to 10 years. Given the rate that technology is developing, Congress should allow for new dynamic procurement processes that create competition and allow for rapid piloting and identification of promising technology.

This could include creating low-risk controlled environments to test emerging technology and bakeoff style procurement processes that are stagegated with clear benchmarks. This would allow us to engage the full innovative potential of the private sector to find the solutions that would bring the most value. So with that, thank you for the opportunity to testify and I look forward to answering your questions. >> Thank you so very much.

Uh great suggestions. Uh now our third witness is uh Dr. Ali Mostafi, founder and CEO of Resilitics. You are recognized for five minutes, sir. >> Thank you. Uh chairman, ranking member and members of the committee. Thank you for the opportunity to testify today. I am Ali Mustafavi, founder and CEO of Resolitics AI at Texas Aesto. Artificial intelligence offer real opportunities to augment our capabilities for hazard prediction and for better preparedness and proactive decision- making towards protective action.

Machine learning and computer vision already support hazard prediction and damage assessment. Generative AI, large language models, and agentic systems now add ways to retrieve information, reason, and coordinate analytical tools. The area we work on is the gap between a warning and understanding of what will be impacted. Earlier warning gives communities time to prepare but understanding and of who and what is in the path consumes that time.

We call this assessment latency. The time between resilience question and defensible answer. I founded Resolitics AI to make integrated disaster assessments accessible to teams with limited analytical capacities. With support from the National Science Foundation's SBIR program, we developed emergency AI that translates fundamental research into operational capability. Early components of this technology were operationalally tested during hurricanes burial, Milton, and Helen.

Emergency AR technology serves as a first pass resilience analyst for emergency management and situational awareness supporting emergency management, healthcare, and infrastructure teams. Its National Geospatial Foundation connects hazard, population, vulnerability and infrastructure data, including federal open data. Users ask questions in plain English. Analytical tools translate that question and calculate a potential exposure and produce numbers, map and briefing and audit trail in minutes.

We call this glassbox AI. Teams can examine evidence, verify findings in making every oper operational decision using our technology before a threat develops. Blue sky assessments inform plans and exercises. A as a threat develops, gray sky assessments use the same foundation to examine potential impacts. For example, a county can assess fl uh potential flood exposure to identify communities needing outreach, facilities needing readiness checks, and where to investigate resource prepositioning.

This helps turn warning into uh protective action. Our ongoing engagement with uh local, state and private sector emergency management have shown significant interest and opportunity for the technology we have developed. Yet adoption barriers are widening the gap between AI advances and everyday practice. Better frontier models do not automatically become disaster capabilities. Three changes would help. First, creating a clearer contracting routes from SBI prototypes to operational customers through p public private partnerships.

Second, coordinating funding and and give emergency management agencies flexibility to use grants for pilots, training and sustained implementation including in small, rural and tribal communities. Third, sustaining invest investment in fundamental research that advances geospatial reasoning, reconciliation of conflicting evidence and coordination of imagery, measurement and tools. American leadership depends on connecting federal open data frontier disaster AI research and models and a streamlined commercialization pathways for disaster technologies.

Disaster technology incubators can connect the startups with practitioners, pilots, and early adopters. In closing, a significant opportunity exists in augmenting disaster preparedness and emergency management processes with frontier AI forward deployed into domain solutions and technologies that strengthen preparedness and situational awareness. Success means early protective action, understanding impacts sooner, reaching people who needs assistance and acting before a hazard becomes a disaster for community.

Thank you and I welcome your questions. Thank you very much. Our fourth and final witness is Dr. Falcowski. Thank you for being here, and I would recognize you for five minutes. >> Chairman Babin, Ranking Member Lorren, and members of the committee. Thank you for the opportunity to appear before you today. My name is Dr. Dr. Michael Falcowsky and I'm the chief scientist at Earthfire Alliance or EFA, a US global nonprofit building the tech technological infrastructure needed to detect, monitor, and understand Earth's wildfires.

Before gen joining EFA, I led NASA's wildland fire program, which focused on developing cutting edge technology, including satellite and airborne systems to drive discovery in global fire science and improve wildfire management. I have dedicated my career to understanding fire and land management needs and ensuring science and innovation serves those in need of the solutions on the ground. Mr. Chairman, there is no doubt that our country is in a wildfire crisis.

The frequency of devastating fires has dramatically increased over the past few decades and this year is no exception. As of September, wildfires have already burned nearly double the acreage of what burned in 2025 and we are just entering the peak fire season in places like California. We are experiencing major impacts on our communities and our economy with homes destroyed in places like Spokane, Washington and Reno, Nevada.

Our cattle industry disrupted in places like Texas and Nebraska and toxic air blanketing communities from Washington State to Wisconsin and even here in Washington DC. However, fire itself is not the problem. Many of the forest and grasslands in our country require regular beneficial fire to maintain healthy ecosystems and reduce fuel loads that contribute to uncontrollable fire. The most devastating impacts occur when fire meets extreme weather conditions near underprepared communities.

This is exactly what happened this August when a series of fires quickly move through Spokane, destroying nearly 850 homes and structures and forcing 65,000 people to evacuate. So, what can we do? The US desperately needs practical solutions to increase the resilience of our communities by applying cutting edge technology that supports beneficial fires and predicts, identifies, and mitigates potentially catastrophic fires before they occur.

In much of the country today, fire is still being managed and fought with the technology developed 20 or 30 years ago. This is largely the result of a significant disconnect between operational fire management agencies and operational or sorry and technology innovators. Although we are seeing major leap forward leaps forward in technology, individual agencies and companies often work in silos which hinders the adoption of innovative tools that are fit for purpose.

Our nation's welfare crisis will not be solved by any one organization or technology alone. And there are glaring gaps in current capabilities that urgently need to be filled. For example, when I was at NASA, local leaders and firefighters repeatedly asked for higher resolution, higher frequency data on fire location, intensity, and spread. I joined EFA to help fill this gap. In collaboration with over 200 firefighters and fire scientists, EFA's flagship program, Firesat, was specifically designed and purpose-built to meet these expressed needs.

Firesat can detect small fires early and provide persistent observation as they progress. When fully operational, the constellation will deliver fresh data for any point on Earth every 20 minutes or less. This level of detail for early detection and frequency of data updates is unprecedented. Current multi-purpose satellites struggle to detect wildfires until they are orders of magnitude larger, and they do not provide updated data at the frequency required by firefighters.

EFA currently has four satellites in orbit, three of which launched in July. And we are now working closely with firefighters and scientists to build an accessible data set that will be invaluable for implementing innovative fire management and resilience strategies to help keep communities safe. [clears throat] Bringing fires from vision to reality required more than innovative technology. It demanded a new way of collaboratively funding, designing, building, and operating satellites.

To achieve this, EFA was founded on what we call a P4 model, which partners philanthropic and public investment with private and governmental innovation. Firesat would not be possible without the techn technological foundation agencies like NASA and Noah developed and derised over the last 30 years or the hundreds of firefighters and fire scientists who informed its design. EFA's novel approach offers a roadmap of what successful innovation and impact can look like.

The nation cannot continue to afford developing technology for disaster resilience in silos. EFA's model shows the importance of driving innovation with the needs of those on the ground, the role that the nonprofit, public sector, and philanthropic communities can play in convening these essential collaborations and the exciting possibilities commercial innovations present when integrated appropriately. I am confident that we have the tools and know how how to solve this crisis crisis.

Thank you for your time and I submit the remainder of my testimony for the record and I look forward to your questions. >> Thank you. Thank you [clears throat] very much. And uh I tell you what, I just got back from Mr. Herd's district in Colorado and we flew over hundreds of thousands of acres that have been burned. So, you're right on you're right on target. Thank you very much for that. I want to thank each one of you witnesses uh for your testimonies and uh we're looking forward to having the questions asked.

Uh and I'd like to recognize myself uh for five minutes uh for questions. Uh Dr. Mustafabi uh for the AI based digital twin and situational awareness platforms discussed today. What is a realistic timeline and a cost estimate for a state like Texas uh to deploy these technologies station uh statewide? >> Thank you for the question. Um I think uh technologically uh we have the technologies to implement uh the situational awareness uh uh using uh AI based solutions.

I think the main barrier remains data readiness of agencies. Uh I think with uh if uh the organizations like Texas Department of Emergency Management is data ready uh within uh within a timeline of two years uh that can be done uh cost estimate uh that's that's hard to estimate for a state like Texas you know but um I think you know from a technology perspective I think we have the technology frontiers to implement that.

Are these platforms interoperable with the systems that Texas is their emergency managers uh in our state are already using uh or would they require significant upgrades or replacement of existing infrastructure? >> I I think the interoperability integration would not be a barrier for implementation [clears throat] of these technologies. >> Good. Thank you very much. And also, I know we hear our counterparts and friends on the other side of the aisle are are very very uh free with talking about cuts uh in spending and in funding.

So, I I I would like to just address some of this uh of the of the administration budget cuts on scientific grants. The integrity of American science depends not just on funding, but on trust, transparency, and accountability. Taxpayers invest tens of billions of dollars each year in fally funded research and they deserve confidence that these funds are being used honestly, transparently, and in accordance with the law.

Recent enforcement actions, I will add, have raised serious concerns about research institutions that have received federal funding while submitting inaccurate data, failing to disclose foreign support, are falsely certifying compliance with federal requirements. The overwhelming majority of researchers act very responsibly, but public trust depends on ensuring that federal funding is awarded and managed honestly. A research enterprise funded by the American people must ultimately be accountable to the American people who are paying the way for this.

Now, Mrs. Roberts, I'd like to ask you a question. Uh, how many separate federal grant or funding programs currently touch hazard detection and warning technology? And would consolidating them speed deployment or slow it down? That's a great question. I mean, we have dozens of programs across federal agencies that are looking at different parts of this data and they can't give us a full picture of what it looks like and we have real challenges with interoperability and as you mentioned going from data to insights continues to really be an ongoing challenge.

So I would urge everyone to think about this massive system of 130 federal programs and really streamlining it to something more manageable that can serve people better. >> Okay. Thank you. And I'd like to follow up with Dr. Mostafia with another one. Should Congress prioritize funding for multi-hazard platforms over singlehazard systems going forward? >> Absolutely. I think disaster many of the areas uh in the US have multi-hazard uh threats.

So I think multi-hazard programs would definitely yeah be the answer. >> Thank you. and Miss Roberts. One one other from the funding side. What would that shift require >> to from the funding side? Um it would to streamline the programs would require obviously quite a lot of thinking on the part of Congress and to move to multihazard programs. I think we need a lot of data sharing and agreements in place to be able to start to move towards what we're discussing on a multi-hazard platform, which I'm also in favor of.

Thank you. And uh in the time I've got left, Miss Roberts, also uh you testified that Noah Atlas 14 rainfall data used for building codes and infrastructure design was decades out of date in parts of Texas before Hurricane Harvey hit in 2017, which devastated uh my district. As a Texan, I take that very seriously. In fact, the national rainfall record is shared by me and my colleague not here this morning, Mr. Weber. uh in our districts of nearly 60 I think it was over 60 inches uh especially after the catastrophic flooding in the Hill Country last year.

So um uh and now that preliminary Atlas 15 estimates are coming out this month, what needs to to happen to ensure that Texas communities incorporate this updated uh data into their local codes and planning before the next flood without incurring significant cost barriers. Yeah, I'd commend Noah for their recent work on Atlas 15 and say that what we need to do is to quickly integrate it into guidance for building design because every time we put in a pipe, we build a bridge, we design a road, we want it to build for the conditions we're facing now and into the future to protect communities in Texas and across the country. >> Yes, ma'am.

And I think that uh I'm already over time, so I'm going to yield to uh our ranking member, Miss Lofra. >> Uh thank you, Mr. Chairman, and thanks to all of the witnesses for your testimony and your insight. Um Dr. Falcowski, you've worked with directly with the firefighters and natural resource managers for a long time. Um now based on your experience including with the fire sense project at NASA, can you talk about what the best practices uh should be for engaging stakeholders and actually getting a product that people will actually use? >> Yeah, thank you for that question, Congresswoman.

So what I learned over my career is that engaging the stakeholders like the firefighters in this case early in the process is paramount to success. All too often we have the the technologists working in isolation really developing cutting edge uh amazing technology that sometimes might might not be quite uh quite fit for purpose because they have not engaged in the dialogue early in the process and they don't understand the problem that they are trying to solve.

On the other side we have the the brave men and women women of our firefighting service that are really busy focused on saving lives and they don't always have time to keep up on what the cutting edge of technology is. They don't know the best of what is out there. And so there really needs to be this forum for an open dialogue between those of both of those communities in order to to understand problems and develop solutions that are fit for purpose. >> You know, we talked about the need to streamline and I think that that is correct, but we have a multiplicity of actors who are playing important roles.

I mean, um, it's not just federal agencies that have specific, um, data that they're collecting or expertise that they have, but also the state. I mean, in in California, we've got CalFire and others, including academic institutions, NASA, Ames is another one. Um, and but also, uh, local government and also the private sector. How do you saying we want to streamline is a worthy goal. Um but h how do you integrate all the information and expertise that is sort of scattered among the various responsible entities who do in fact have a role to play.

I know that doctor that you were overseeing that effort uh at the NASA as and can you share any insights on how that can be successful? >> Yes. So I agree with you that there are benefits of consolidation but what we really need as I said earlier is a forum for which um this collaboration can happen. So really a hub of innovation that has all the different players feeding in to a central uh place where these ideas can be discussed, problems can be uh properly understood and then technology can be developed.

It's a really challenging problem. Uh oftentimes these groups do not speak the same language, right? scientists and engineers speak very differently than disaster response firefighters and that sort of thing. And so you really need people in this in this these hubs of innovation that understand both sides and are able to sort of be a liaison between those two groups. >> And your thoughts on what struck me in the meetings I've been in was everybody working in good faith trying to get this done but nobody assigned to play that coordinating role.

Is there a role that Congress should play in helping to designate that coordinator role? And would it be different in different loces? And also, it's multi-hazard and sometimes they're overlapping. I mean, the uh the rain onto a previously burned area is uh you know, it's connected with with multiple hazards. How would you do that? Yeah, I agree with you that there needs to be some sort of formal mechanism identified um that identifies an an institution or organization where they can play that consolidation role and that role for communication. um speaking from the perspective of wildfire which is you know it's it's very unique from other disturbances um and hazards you know there's this idea of developing a fire intelligence center and I think certainly something like that could very much help develop and transition technology and operations to help our firefighters but again it it needs to be designed very carefully and that is certainly something that Congress and the experts could help um um chime in to make sure it's done correctly.

One of the things that uh you know we've talked about predictions and and warnings and the like but warnings are only effective if people believe them. And one of the things that I'm interested in is whether uh we you Dr. Mustavi or Falcowski believe that the research on human behavior risk communication and decisionmaking in the development of new forecasting and warning technologies would be useful and we've had substantial reductions by the NSF into um social behavioral and economic the SBE directorate has that had an impact on this whole area of understanding how warnings are received so that they can beffected effectively calibrated to be acted upon. >> For sure.

Absolutely. I think uh you know understanding human behavior both individual and collective behavior of a community in terms of decision for evacuation you know and uh you know other protective actions is fundamental and I think deeper deeper understanding especially with integration of technology how uh you know uh communities and humans interact with the technology in response is is an area that needs understanding and deeper research so definitely more research in this area is uh fundamentally needed. >> I see my time has expired Mr. chairman.

So, I'll yield back. [clears throat] >> Yes, ma'am. Thank you. And I'd like to recognize uh Representative Webster from Florida. Five minutes. >> Thank you very much, Mr. Chairman. Mr. Roberts, uh in Central Florida, which where I live, um we're home to one of the largest modeling and simulation and training ecosystems in the world. This same technology was helped has helped drive the development of digital twins, which today are increasingly being used to model weather events and forecast their potential impacts on recovery needs.

From your perspective, um, are we doing enough to take advantage of these dual use capabilities to help communities model and prepare for hurricanes and floodings and fires and other natural disasters? >> First, I'll say I agree with you that Florida and Central Florida are really on the forefront of cuttingedge technology for understanding these risks. And I would say we're certainly not doing enough. We have some incredible technological capabilities, but at some point we have to move from the simulations and the digital twins to actual projects that reduce risk.

And that's what's often difficult. We can know where there's risk, but at some point we need to be able to pay for, you know, the feasibility studies, the engineering, and be able to put steel in the ground to protect people. And that's a place where Florida's actually done quite a lot in funding those projects, but we could be doing more to translate what we know on the technology side to what we've actually put in place to protect people and save lives. >> Yes.

So could uh Central Florida serve as a real world test bed for bringing government, academia and industry leaders uh to validate the next generation of AI disaster prepare uh prediction. >> Absolutely. I think Central Florida has all the key ingredients, including quite a lot of risk, amazing technological capacity, and being in a state that has a lot of incredible data sets and modeling along with the Florida Flood Hub, which I think is one of the leading examples in the country for helping communities to understand risk and pull together data and modeling across the entire state.

So, I would look forward to seeing that in Central Florida and think it's a great place for that. What would it take to to move these capabilities from the modeling and simulation environment into the hands of the communities that actually need to use them most? >> One thing I see working with communities is that, you know, we're working with a lot of small municipalities, you know, over 400 in Florida that might not even have anyone on staff who has the technical capabilities or is fluent in GIS or knows how to translate it.

So I think thinking about as others have mentioned how we take the technology in the insights and translate it into you know a community with a small staff that's doing their capital budgeting is a really important piece of actually reducing risk. >> I got I got a little bit of time left. I I noticed in your resume that uh in times past you served as a a specialist in infrastructure financing and one of those things was uh how to mobilize private sector financing for uh sub sub [snorts] sustainable infrastructure.

So thank you for doing that. I have a bill that actually I'm talking to the chairman about that um would fund some of the needs in space science and technology with private financing which there are people willing to do that they just haven't anyway thanks for that yield back >> thank you gentleman yields back I'd like to [clears throat] recognize the representative Bonamichi from Oregon professor >> thank you very much Mr. chairman and thank you to the witnesses.

I I just want to note at the outset that in this hearing today, we're talking about using AI to save lives while the headlines in the public discourse is that it's going to kill us. So, I I just want to note that um I I appreciate the conversation today, but we need some guidelines and some guard rails so that we're protecting against bias and so that we're protecting uh privacy and security. That being said, um I represent a district in Oregon.

And I know natural disasters are increasingly severe and frequent. Communities that were once spared uh are now left struggling to reco to uh recover. Uh according to the National Weather Service, national hazards caused 645 deaths and close to 5 billion dollars in property damage nationwide last year. Uh this year alone, wildfires have burned more than two and a half million acres across Oregon. Of course, we followed um everything in Spokane, our neighbor to the north. um or b breaking the previous record of of in and residents across the state really rely on a patchwork of camera networks run by state agencies, universities and utilities to catch fires um hopefully within minutes but instead of hours but that kind of speed shows why coordination really matters and the same is true offshore.

I represent a good part part of the beautiful Oregon coast. Uh and the Cascadius abduction zone earthquake um when it happens uh will cause a tsunami uh in as little as 10 minutes uh faster than most residents could reach high ground. Oregon has a lot of continued efforts to install vertical evacuation structures. Um so I want to ask you Dr. uh Falcowski. First [clears throat] of all, um uh for Oregon to detect wildfires, as I mentioned, we leverage uh multiple separate networks operated by universities, state agencies, utilities, and federal partners.

So, how could shared federal data standards improve coordination? And are there instances in which separate systems would provide some useful redundancy? >> Yes, thank [clears throat] you for that question, Congresswoman. um fire and most uh disasters don't respect borders or just jurisdictional boundaries. And so when we think about data for disaster resilience, the data sets really need to be free and open, accessible and standardized and easy to for the agencies and organizations that are protecting our communities to use.

Um and it really to achieve that we really need collaboration across across organizational boundaries to to uh to be successful. >> And what federal policy could help with that coordination? Um, [clears throat] I am not familiar with a specific federal policy, but I would like to take that question back to my team and and bring a answer back for you for the record. >> Terrific. I think that would be helpful. I want to ask you, Dr.

Most Mustafabi. Um, so [clears throat] Oregon has some new state guidance for the coastal residents. I mentioned the risk of tsunami. U, something I've been working on since I've been on this committee. Um, less than 20 minutes in some places to reach safety. oftentimes in in places geographically, it's impossible to to outrun a tsunami. [clears throat] And so the state only has one of the needed sort of requisite vertical evacuation structures.

I just visited it over at uh the Hatfield Hatfield Marine Science Center in Newport. Um, and I wonder how could digital twin technology support modeling for vertical evacuation infrastructure and inundation mapping because it's a real serious concern over there. Our state agency now is is um using technology to put information in hotels for example because oftentimes visitors don't understand the risk of tsunami and it will tell them giving them a map.

Here's where you are, here's the evacuation route, here's how long you need to get out. But that doesn't work for everybody. So how how would it help to [clears throat] model for vertical evacuation? >> Thank you, Congressman. This is a very important question. I think the the uh power of digital twins and air powered solutions is the fast synthesis and integration of various data sets faster than any human being can do that.

So uh uh by analyzing the forecast for the surge evacuation routes uh and in matter of seconds uh areas that needs to evacuate can be identified and warnings can be sent so that people can evacuate and then you know use those vertical evacuations. >> Yeah, I appreciate that. And ranking member Lavkin mentioned we've had a lot of of discussions about the social science of warnings and the efficacy of warnings. I think that's an important part too.

People need to understand the seriousness. Um I I I also want to note that um FEMA has an important role to play. Uh I've always thought that FEMA could have more of a role uh in in preventing rather than just cleaning up after disasters. And that's why we had the brick program, the building resilient infrastructure and communities program, excuse [clears throat] me, in the hospital in Historia, Oregon in the district I represent got a brick grant to build vertical evacuation in the hospital uh that they're rebuilding and constructing.

Unfortunately, uh the former Secretary of Department of Homeland Security canceled all the brick grants, but we're working hard to get that because it is going to save lives. Truly, you save lives, uh if you get the information out to people that they can evacuate vertally vertically. So, um I will also note that the bipartisan weather act, which advanced out of this committee, um last year, would accomplish um a lot of our goals by strengthening disaster preparedness, coordination, and response system.

And it's past time to make that law. And I'm over time, I yield back. Thank you. I'd like to recognize the uh representative [clears throat] from Texas, Mr. Weber. >> Thank you. Thank you, Mr. Chairman. So, this questions for all witnesses. Uh the technologies mentioned in your testimonies offer significant improvements in natural disaster detection, response capabilities, as well as recovery methods. All of which, needless to say, help save countless lives and save billions of dollars in economic losses.

I'm from the Gulf Coast of Texas. We get a lot of should it should be don't y'all think it should be called the Gulf of Texas instead of the Gulf of America? I I think it should. But during disasters, local governments and first responders are the first line of defense. However, many localities in my district and Dr. Babin's district as well struggle with tight budgets and personnel gaps which creates challenges in procuring and implementing these very systems.

So my question is what are some ways localities and technology technology providers can work together to navigate these very difficult realities which can happen in just a week or two. Mr. Gail, we'll start with you. >> Thank you, Representative. Uh my suggestion would be to help facilitate and encourage public private partnerships. So let's bring together private sector interests. Let's bring together state and local interests and let's bring together federal interests, pull resources and develop an approach to multi- disaster scenarios where we are marshalling the resources together to achieve the greatest uh fit for what we need in terms of response and technology adoption.

I I fully think that the more you bring people together, the the more resources you have available and the smarter solutions you can implement the most quickly. >> Mr. Robert, >> I would say that it's really important to think about procuring these technology systems at the regional and state level rather than through hundreds of localities, many of which have never done a similar transaction and have difficulty negotiating them.

I think this is a place where Texas has done incredible work thinking in regions and also with the Texas Water Development Board that's incredibly sophisticated. So I think the more we can think about procuring this data and giving access to it at the statewide or regional level rather than expecting say a small community of 600 people to be negotiating a contract the better we'll do and the more efficient it will be. >> Mh. >> Dr.

Mustafi, you're next. for sure. Um I think uh pilot programs that can help with the last mile adoption uh would significantly help that not only you know help the communities and uh agencies to adopt the technology but build capacity around that technology and uh we have observed this last mile adoption barrier in the uh both the procurement and capacity for technologies that creates a gap between the technologies that exist today and technologies that are used on the ground.

Thank you, Mr. Falcowski. Okay, you're you're next. >> Great. Thank you, Congressman. I'll add uh in addition to the the public private partnership that Phil philanth philanthropies are very important for derisking early technology and they've been really critical to the success of what we're doing at Earthfire Alliance. And then I'll also add engaging end users early. We have an early adopters program where we have 18 agencies uh working with us, fire management agencies working with us across the globe and in fact the Texas State Forest Service is one of our early adopters and we're working very closely with them to help understand what their fire management needs are and then how our data can be uh formatted and fit and tailored to meet their purposes. >> All right, quick question.

I'll go back to Mr. Gail. Have you ever been in a hurricane? >> I've never been in one, but I've exper >> Excuse me. >> How about a Gail storm? That's a joke. [laughter] So, so, uh, Miss, go ahead. You've never been a hurricane? >> I've never been in one, but I've certainly experienced the aftermath and, uh, and helped build communities that have been impacted by hurricanes. >> How about you, Miss Roberts? Ever been in a hurricane? >> Similar. >> I'm trying to pick up your broke.

Are you from Oklahoma? >> No, I'm from Massachusetts. >> That was my next guest. So, doctor, how about you? >> Yes. Hurricane Harvey. 2017 Harvey which hit the Gulf Coast and hit our district pretty hard. Dr. Beaven's district as well. You >> already mentioned it. >> Yeah. Well, you got to give Harvey a little credit. Dr. Vulcowski. >> I have not been in a hurricane, but I have used to be a firefighter, so I have firsthand experience with wildfire. >> In In what jurisdiction? >> I was out west fighting around Yakama, Washington. >> I got you. >> Well, we appreciate y'all's expertise.

You've made a lot of good suggestions and thank you for recognizing that Texas gets it right. You know, I served four years in the Texas House before I got demoted to Congress. So, anyway, thank you for your expertise and we appreciate y'all's um testimony. Mr. Bab, I yield back. >> Gentlemen yields. I'd like to recognize Representative Stevens from Michigan. >> Well, thank you, Mr. Chair, and thank you to our witnesses.

And as a point of personal privilege, I would just like to acknowledge the phenomenal row on the upper disas of women of the uh science committee on uh you know in this important moment. Uh we've just got some amazing voices. I know Miss Fushi just stepped out. She's an AI expert. Uh Miss Selenus who uh is a guru in her respective uh areas of expertise. Nothing gets past Deb Ross and Bonamichi is our dear mother hen.

And of course, we cannot forget our ranking member. So, I'm using my time, but this is an important one. And I just got to say that this is life or death. This is Look, Weber's from Texas. He walked out or he's still hanging around in the back of the DAS. I don't want Michigan looking like Texas. All right. We are not the central plains. And what's happening is these darn tootin tornadoes are blowing through on no notice and knocking over trees and killing people.

They're killing people in their homes. I didn't grow up like this in Michigan. You know, I I I live in neighborhoods with hundred foot trees that year after year after year get slammed by these high wind events. And we had someone in Huntington Woods, our director of public service, he lost his life. I had a a mother in, you know, Leavonia and her family was impacted by this and and there was loss of life. I mean, unspeakable stuff.

And so when I've got a piece of legislation that I know Chairman Babin loves, it's called the Twister Act, the Tornado and Windstorm Innovation for Safety and Tracking Enhancement Research Act to expand Noah's existing tornado research program to improve the prediction and detection of emerging tornado threats. We got to move. You know, I I want to keep in touch with every one of you because what you just shared and the fact that you're coming here to this hearing, it is really important.

The east side of Detroit, lower east side, I just called the mayor of Detroit blown through, blown out millions of dollars minimum of damage. I I mean, it's just costing us. So, we're interested. And I loved um what Mr. uh Webster shared from Florida about the digital twin because I worked on that and I I like all these IoT applications and you know different types of technologies. This is why it's important to have people up on this dis like the women I am serving with on science though to make sure that it is equitable that it is fair that it is accessible to all municipalities.

And so Dr. Mustavi, as we look at improving hazard predictions, can you talk about the relationship between fundamental research sustained data collection and our ability to better develop forecasting tools with AI? What happens to those capabilities when the gaps in the when there are gaps in the underlying data? >> Absolutely. I I think that data foundations are crucial for building every technology that we create in disaster space and in any other applications.

I think uh you know if we uh have outdated data is uh there is a term say you are all you for AI systems and computational models they are all about the data that they have. So uh the gaps will undermine the capabilities for anticipatory assessment forecasting and decision-m for sure. >> Yeah. Yeah. Great. And you know Mr. Gail and Dr. Mustavi. Uh, wildfire prediction also depends heavily on our understanding and observation of weather.

So, how would the loss of consistent weather observations or fundamental research um into changing weather patterns affect your ability to develop and improve AI tools for wildfire detection and prediction? Loss of resources for continuing that kind of research would be devastating. But we are partnering with a number of people in the technology sector to ensure that we have the latest and best weather sensing uh integrations into our platform.

Now, uh, Western Weather Group is our partner on the weather sensor side of things and they do phenomenal work and we want to continue working with the members of this committee and the rest of the congressional body to ensure that we have the resources we need to have the best tech possible on the ground. >> Well, count us in for the effective utilization of the taxpayer dollar to innovate, to save lives. Jeff Masters, who founded weather.com, is a Michigander, a constituent of mine.

Um, I love your coalition. We're all eager to keep in touch and bring solutions to the betterment and for the betterment of community. Thank you, Mr. Chair, and I yield back. >> Representative yields back. I'd like to representative [clears throat] Herd from Colorado. >> Thank you, Mr. Chairman and Chairman Vman. I want to thank you personally for joining me and members of the Western Caucus last week in Colorado's third district.

Part of that tour was uh doing an aerial overflight of the Aspenacres fire. 102,000 acres in PBlo and Kuster counties burned. And so today's hearing on disaster prevention is particularly well timed. Um Mr. Gail, it's great to see you here this morning. Your Pano AI system detected the Bear Creek fire in Douglas County before it became a major incident. Can you walk through what happens uh from the moment your camera detects smoke to the moment a firefighter aircraft is actually dispatched?

Where where in that chain do we lose the most time? >> Time is essential. So learning about the ignition and locating it is the most critical part of keeping small fire small. We average about three to three and a half minutes from the time our technology detects the fire to the time that it gets confirmed by human analyst to when the alert goes out to the response agencies. Bear Creek Fire was no exception. We alerted all the response agencies in Douglas County in that geography which was a remote section of Douglas County deep in the wilderness.

That alert went out. We determined that hot shot crews were about a two and a half hour hike from getting to that fire. Douglas County, which is unique in that it has its own heel attack crew, made the decision to deploy their helell attack team into that fire and get suppression on it right away. The outcome of that was containing that fire to less than three acres. That was a huge example of how the technology works and why finding fires while they're still small is absolutely essential.

So rapid detection, learning about it quickly, deploying resources on it, and also understanding the extent of what resources you do need to deploy is a is a part of the whole picture in how we approach wildfire response and providing our wildland firefighters the visual intelligence and situational awareness they need to do it safely. >> Right. If Congress uh Mr. Mr. G, if I could just stick with you, if Congress could fix one part of that handoff, procurement, communications, um, data sharing, agency coordination, what what should we focus on most?

What would save the most time? >> I mean, all those things you mentioned, I think, are critical to address together. Um, you sit on the House Natural Resources Committee, too. Uh, Chairman Westerman and your colleagues in the Senate have a piece of legislation called Fix Our Forest Act, which has a wildfire intelligence center provision in it. I think a combination of funding, creating uh new entities like the intelligence center to improve the way we coordinate and uh procure technology and deal with data so that we have sharing, integration and collaboration across all jurisdictions is going to be key to doing this quickly and effectively across state and federal entities that that need this kind of coordination.

Like fire does not know boundaries and we need to make sure that we have the resources going to the ground in addition to making sure we can do it quickly and improve the way we procure technology so that we don't have to keep waiting when we know we have proven commercially successful solutions ready to go right now. >> You've you've uh said that the technology is increasingly proven but procurement is too slow. What does the federal government make companies prove repeatedly that's already been demonstrated elsewhere?

Well, one, the technology is effective and it can work well. And and two, there's um I think some catch-up to be done with appropriations like we don't know where AI technology and wildfire technology solutions sit in the budget process. And so I think there's a a bit of a breakdown in administrative hurdles that we need to clarify to help facilitate efficiency in the way we are taking congressionally appropriated resources and putting them on the ground quickly and making sure that we don't have to re-evaluate things that have already been evaluated and proven successful.

In Colorado, uh our most deployed state, by the way, we're detecting and locating fires all the time. We have case studies beyond Bear Creek. We have case studies in your district, Congressman Herd. We are already successful. Many of our partners in wildfire technology are already successful. Let's deploy the proven solutions now while we research on how to make them better and and then work together in Congress to appropriate the resources that we need to scale this appropriately and have coverage over the entire country. >> Great.

Thank you, Mr. Gail. In my remaining time, Mr. Falcowski, you talk about in your testimony beneficial fire. Does the technology that you're talking about does that help with determining where best to also do prescribed fires and prescribed burns or is it only um when wildfires have ignited? >> It's really both situations. So the tech that we're developing can certainly enable more prescribed fire and give give uh prescribed fire managers more confidence in in putting fire on the landscape and then also in the case of a wildfire that's not threatening a community. we can help our use our data and the intensity measurements to understand where we might want to let a fire burn to have uh land management benefits. >> All very very important across the country in particular in Colorado's third district.

Thank you chairman Babin. My time is expired. >> Thank you. Uh I'd like to recognize Representative Ross from state of North Carolina. Well, thank you so much, Chairman Babin and Ranking Member Lofrren for holding this timely and important hearing. And thank you so much to our witnesses for sharing your testimony and insights with us today. Nearly 2 years ago, Hurricane Helen tore through western North Carolina, causing catastrophic flooding, wiping out homes, businesses, and entire communities.

It was a devastating and unprecedented tragedy in North Carolina's history, and communities are still rebuilding. The recovery effort has been a testament to the sheer grit and perseverance of Western North Carolinians, as well as the partnership of local communities, Governor Stein, our state government, and the federal government. As of today, the federal government has slowly awarded about $9.8 million for Helen recovery in North Carolina.

That's about 16% of the total $60 billion in estimated damage from the storm. in addition to delays from the prior head of um of uh Homeland Security and um and HA and um FEMA. Congress has not appropriated any supplemental funding for rec storm recovery, not to mention Helen recovery since 2024. So that was signed into law by President Biden. Congress needs to appropriate more money for disaster relief, including for Helen, and our governor has requested more, but there have been so many disasters that have not gotten any money from the federal government.

The upcoming anniversary of Helen underscores just how important this hearing is today and how important bipartisan work is in this Congress. For example, last Congress I led a bipartisan push to fund next generation hurricane hunter aircraft which are vital to tracking and monitoring hurricane. So, I want to talk about what we can do for weather forecasting, preparation, and response, not just what we do afterwards.

Emerging technologies like we've heard of can make all the difference in the world. And when we combine the technologies with data from traditional sources, we can predict things, save lives, and save property. But we have to invest in it. So, Ms. Roberts, Hurricane Helen demonstrated that catastrophic disasters do not always fit into traditional assumptions about where certain risks are greatest and what communities will need to recover.

Should federal preparedness grants provide states with greater discretion to allocate funding based on state and state and local risk assessments rather than standardized approaches? and if so, how would you put the appropriate guard rails in? >> Yes. First of all, Congresswoman, I want to thank you for your work to support communities impacted by Hurricane Helen. As we've been working with those communities, we've really appreciated the support you've given in your office.

Um, I would say one of the things that, you know, came out after Hurricane Helen was first, as you mentioned, the flood risk really wasn't understood. I mean less than 2% of people in impacted areas had flood insurance because they were told they weren't at risk. So I think we certainly need more flexibility in understanding those risks. And the other thing that became apparent was that there wasn't adequate understanding of landslide risk that followed which is one of the themes we've talked about in the multi-hazard nature of these challenges.

So, I think we need more investment, as you mentioned, upfront and more flexibility in thinking about how the state can address these risks. And North Carolina has made a number of incredible strides to think about this at the regional level and invest in their flood resilience blueprint. And I think those are the types of efforts that the state needs discretion over and to continue investing in and have support for. >> And if if there would be any guard rails, what would they be from the federal government?

I think from the federal government, we want to make sure that money is being spent to figure out how to reduce risk the most and that it's holistic and is being used in an appropriate way. But I think it would help a lot to focus on results and outcomes for risk reduction rather than narrow categories of uses, which is what we see now when there's 130 programs that you have to navigate. places like North Carolina spend all their time figuring out how to mix and match the programs and not get caught in double jeopardy rather than actually getting to focus on how to save lives. >> Thank you for that response and thank you for your excellent work.

I yield back. >> I'd like to recognize the representative [clears throat] from Indiana, Representative Baird. >> Thank you, Mr. Chairman, and welcome to all you witnesses. Uh you know, I'm going to start off uh with Miss Roberts. uh that you noted in your testimony that Noah and the USDA drought risk Atlas hasn't been comprehensively updated since 2017, which is what almost a decade. And even though that data feeds directly into wildfire risk and agricultural planning, my state felt the cost of that.

The historic flooding that hit Indiana this past August killed at least seven people and left farmers still assessing crop damage going into harvest. So my question is, how much is outdated federal data actually costing farmers in bad planting decisions, uninsured losses, and infrastructure built to the wrong standard? And should USDA and Noah jointly prioritize updating the atlas alongside the rainfall data modernization that you described? >> Thank you for the question.

I would say, as you mentioned, this has incredible costs for our farmers. They make decisions every season based on, you know, this sort of weather data that they need. And so giving them inaccurate data is setting them up to fail and is incredibly costly for our farmers and for the entire country. It also means when we have outdated data that every piece of infrastructure we're investing in isn't going to last for its useful life.

So if we put in a pipe that's a little bit too small because it's based on outdated data, instead of lasting say 20 years, it might last only 10. And that's a cost that we all end up paying for. So I think it's really important that we follow in the great footsteps of the updates that Noah is doing to Atlas 15 and make sure that all of our data is modernized. And I would build on um the comments Dr. Mustafari made and say that you know when we're thinking about technology it's garbage in garbage out.

So if we don't have updated data going in we're not going to get good insights even if we apply cutting edge AI systems. >> You know we're just starting the harvest season in the Midwest and it's uh very important that you have the right kind of information in order to make decisions because you can make adjustments in the at planting time and at harvest time and but we're always subject to the weather. So for me this is kind of an important issue and that we use the right kind of data and we have the capability to analyze that through our through our uh uh computing systems that we have today.

So so thank you for that. So my next question goes to uh Dr. Musk agriculture and here again I'm talking about agriculture is uniquely exposed to the hazards of build over an entire growing season rather than striking all at once. drought stress, flooding and planting of harvest, severe storms damaging grain bins and livestock. So, does your platform currently integrate agriculture data such as soil moisture, crop condition, livestock facility locations, or upstream flooding into its situational analysis?

And if not, what would it take to extend this kind of uh monthly hazard real-time platform to serve farmers and rural emergency managers, not just the urban and the coastal communities? >> Thank you for your questions. Currently, our platform does not serve that purpose. However, we need uh quality data. With quality data, integration of uh those capabilities will not be a a big undertaking for us. So my next question my next question uh really goes to um you as well doctor.

So >> sure >> uh my alma mater Purdue University is researchers working on exactly this problem. They contributed to a study published earlier this year on integrating different flood modeling methods across domains including agriculture. and Purdue College of Agriculture has published work using AI and drone imagery to predict crop yield impacts from flooding and moisture stress. So um is restitics drawing on this kind of land grant research and how much of your platform's underlying science comes from publicly funded university research versus proprietary models that you built independently and we got about 35 seconds. >> Absolutely.

I want to recognize that Purdue is my alma matter as well. So um yes I you know actually resolitics >> now we're getting somewhere you know that >> resolitics is coming out of fundamental research from university. It's a spawn off from Texas&M University where my faculty as well. So um fundamental research is foundational especially in disaster tech space where venture capital ecosystem is not uh you know favoring disaster tech outside insurance technology. >> Thank you and thank all of you.

I wish I had time to do the other two, but anyway, I yield back. >> Gentlemen yields. Uh, now I'd like to represent uh to recognize Representative Selenus from Oregon. Yes, ma'am. >> Thank you, Chair Babin, and thank you to the ranking member um for holding today's hearing. As Oregon faces another record-breaking wildfire season, I truly appreciate this committee's consistent attention to natural hazards risk mitigation. from severe wildfire, damaging wind and ice storms to a potentially historic El Nino uh season and the everpresent risk of a major earthquake and thus tsunamis.

Oregonians are acutely aware of the costs of natural disasters and the immense upside if we can better predict, prepare for and respond to these catastrophic events. Recent incidents in frontier AI development have created a lot of concern, especially over the last week around AI risk and are driving renewed urgency in congressional efforts to regulate the industry as a whole. Mr. Gail and Dr. for Mustafabi, in these conversations, how should we think about companies like yours and how can we ensure that we both protect against that catastrophic risk posed by AI without actually hampering efforts to deploy new AI enabled hazard mitigation technologies. >> I'll speak to our technology first.

I believe that there's some education to be done in what artificial intelligence is. The way we apply AI, for example, is computer vision and machine learning. So, we are training an AI model to look specifically for certain features. In the case of our technology, we're training it to look for smoke and then at nighttime, our infrared sensors are hunting for heat signature. And so we're not using generative AI in our model to deploy this technology solution.

It's to mimic human eyes and do so in a highly accurate way. So I think the general public needs to understand the different types of artificial intelligence and the way AI models are applied and recognize that in a technology uh circumstance for disaster response and and predictive behavior that computer vision and machine learning AI is really critical to doing things quickly, efficiently, effectively and helping uh be the AI for good that everybody wants it to be.

Thank you. >> Uh I echo the comment about um literacy. I think uh understanding what AI technologies are at their core is foundational and I think it's the it's the responsibility of the tech developers to communicate that with the users. uh also similar to um the technology uh of Pano AI, we are also using AI in a sense that we have created a foundational uh geospatial uh data layer that AI systems can retrieve from and perform different types of analytics and assessments that a per human data analyst would do in terms of assessing risk and vulnerabilities.

So I think you know setting guard guard rails is is essential also building trust between the users and the technology providers is another big steps. I think from a technology pro provider perspective it's our responsibility to communiate comm communicate the limitations of our system educate our users how to use our system to get the best results and then uh at the end of the day the uh responsibility of making decisions is with humans and human judgment so understanding those limitations will help better use the technologies. >> Thank you both.

Uh Dr. Dr. Falcowski in your testimony you reflected on the challenges around inter agency collaboration which can hinder technology adoption as ranking member of the subcommittee on forestry in house a getting this coordination right has been a key priority for me from your experience working on wildfire at NASA as a wildfire fire yourself what are the barriers to strong collaboration between science agencies like Noah or NASA and their public lands management counterparts such as the forest service and BLM Yeah, thank you for that question, Congresswoman.

And it's something that I've I've observed throughout my career. And just want to remind everybody that technology alone can't solve this. And it really gets down to that uh co-development and discussion among the the different user groups and the technologists and the the agencies like NASA and Noah who are often at the forefront of developing and d-risking cutting edge technology. So from my observation at my time at NASA um a lot of times NASA, Noah, NSF, other agencies that are doing science and tech development were not included in a lot of the conversations at the inter agency level with the wildland fire management uh fighting organization.

So I think that is step number one getting getting the agencies that are doing the engineering and the tech development in inside the conversations that are happening so they can really understand the problems and like I said co-develop solutions that are going to have impact on the ground. >> Thank you. My time's nearly expired. I yield back. [clears throat] >> Gentleoman yields back. I'd like to recognize a gentleman from Alaska, Mr.

Begage. Thank you, Mr. Chair. Uh, my first question, Dr. Mustafabi. Uh, your firm started with an NSF small business research award and turned into a company with paying customers, which is exactly what that program is built to do. walk the committee through where the handoff from research award to operational customer often breaks down and what Congress uh you believe could do to help alleviate that. Absolutely. This this is a very important question I think for for startups coming from SBI program.

I think I want to emphasize the importance of the National Science Foundation SBI program and other SPI programs for supporting innovations especially in the disaster tech space which uh essentially help us go through that you know debt value from foundational you know proof of concept that we have from research that shows fundamental promising models methods and data sets or data products and then turning them into operational technologies and the SP programs going through different iterations of the prototypes minimum on viable products is is the pathway that help us you know iterate show our our solutions to different customers and different users and then use the feedback to iterate on that technology.

I think what would enhance these um this process is going from the the SPIO program to an uh operational uh implementation and adoptions. I think we are kind of missing some sort of um you know incubator type you know mechanisms that help the organizations the state local and private sector to adopt these technologies using uh you know the programs that derisk the adoptions you know and help them test these without you know committing significant investments up front I think that would significantly help with you know translating the pathways to operation >> that's great thank you very much for that uh insight.

Um, Dr. Falcowski, Alaska burns more acreage in a bad season than any other state and much of it is hold over and smoldering ground fire that existing satellites often don't catch. Can you help us understand what fires can see above 60 degrees north and what it would take to get the Alaska Fire Service and the State Division of Forestry uh more integrated into the earlier early adopter program? >> Yes, great. Thank you for the question, Congressman.

And so our technology that we've got deployed on fireside is really ideally optimized for the problem that you uh are citing there. So number one, we have four operational satellites right now in orbit. Um and they're polar orbiters. So we're getting something like 20 18 to 20 measurements a day uh at high latitudes like uh like you have in Alaska there. Uh the technology we have on our satellites can detect fires that are really small.

So it's high spatial resolution, but we also also have high thermal sensitivity and so we are able to detect really cool smoldering fires uh and and like the the holdover fires that that you're talking about there. Um right now we are in the process of uh working with the uh folks at University of Alaska Fairbanks um and the GINA group to understand how our data could feed into their system and and be deployed to the uh fire service in Alaska uh pretty read readily. >> That's great.

I was just uh up in Nana outside of Fairbanks and we had the X-Prise competition out there for for drone uh based fire detection and uh and management. And so it' be really interesting to see how uh how those assets on the ground can can be complemented by your assets in the sky. So thank you for that. Looking forward to continue to work with you on that. Last question. Um, Miss Roberts, uh, western Alaska villages flood from storm surge and river ice rather than rainfall.

How do you recommend Alaska better anticipate these events and what investments do you believe need to be made in Alaska in your view? >> Yeah, thank you for the question. I would say, you know, we work with a whole host of communities that are flooding for all different sorts of reasons. As you mentioned, whether it's hurricanes, riverine flooding, ice melt, or other reasons. A lot of the solutions are the same. The first thing is that we need to make sure communities understand their risk, and the state has a big role to play, especially for a lot of our smaller and more rural communities.

Next, we need to make sure that they actually have resources, whether it's the technology or just the planning resources to start investing in the projects that can reduce risk. I think something really important in this discussion is it's one thing to identify the risk and it's quite another to actually start putting projects in place that can save people's lives. And that's really important. And I think in places like Alaska and many others, it's really important that we're thinking about this as a regional level and not asking every single municipality to figure this out on their own. >> Yeah, it's good feedback.

Uh, one of the challenges that we have, of course, in Alaska is that we lack basic infrastructure that we take for granted in the lower 48. Many of our communities are not connected to the road system and so emergency preparedness is absolutely critical. With that, I see my time has expired and I yield back. Gentleman yields. And now I'd like to recognize Representative Fushi from North Carolina. >> Thank you, Mr. Chairman and ranking member for holding this committee hearing today.

And thank you all for your testimonies and for appearing before us this morning. Dr. Mafavi and Mr. Gail, effective AI systems run on clean, properly structured, and contextualized data. The federal government is a treasure trove of data assets, but many of these assets are not made public in AI ready formats. And when it comes to disaster response, data can reside with states and localities, companies or nonprofits, each of which may maintain their data in different ways.

Can each of you talk about what challenges if any you have faced with data access and interoperability related to hazard data? >> Absolutely. I I think uh for for us the foundational challenge and opportunity uh for AI in disaster domain is a connected data layer across different organizations and across different jurisdictions. I think the challenge we have faced that we had to manually gather different data sets across different sources and manually go through their uh them and do the quality control and do the curation needed to make them AI ready and that was a significant level of effort that we had to put and thinking about you know expanding that with every additional data set that comes becomes available and each time each data set becomes uh uh updated is not going to be efficient.

So I think you know having procedures and you know mandates to have these data sets open connected and AI ready would tremendously accelerate the development of technologies and implementation. >> Thank you for the question. We see some issues with data in terms of how uh data is shared across different jurisdictions. And um and what we would love to see at least in the wildfire space that I could be applied to uh all sort of hazard response mechanisms in the way we host and share data is what is outlined in the fix our forest act for the wildfire intelligence center.

So that recognizes the need for creating uh accessible data that remains secure and available to first responders. But as a company right now, Pano AI is working with all data providers and utilizing our data in secure transmission scenarios to ensure that we're doing our part to be interoperable and making sure that our data is integrated with first response agencies so that they can move quickly and not have to worry about whether or not the data has got there yet.

We would also like to see resources invested in the National Inter Agency Fire Center, NIPY. uh oftent times during high uh fire times like we we've had an intense fire season now you have a lot of entities hitting the NIPY data set all at the same time that slows things down and so we would like to propose ideas and solutions that help improve those efficiencies to help NIPY to help ENMAC which does the national multi- agency coordination let's help all these entities work more efficiently and effectively that way when we have intense fire seasons and we have intense data usage that we're not getting bogged down and the information can flow as quickly as possible. >> Thank you both.

I am proud to lead HR8926, the state and local public sector innovation act to help provide federal resources to local communities to modernize their data and technology infrastructure. How might these challenges be unique to state and local governments and municipalities and how could the federal government help? I think I would just reiterate what I what I just mentioned about uh lifting the wildfire intelligence center off the ground and developing complimentary systems to deal with other disaster response too, not just wildfire.

There's a lot of lessons that can be learned with the way that the National Inter Agency Fire Center and ENMAC work on the fire side of things. Like we can deploy those types of coordinating abilities and the way they utilize data to create something that's more uh solidified, unified, and accessible by all the important people that need to respond to fires, see what's going on, and integrate those public private partnerships that I talked about earlier.

And finally, Miss Roberts, in your testimony, you talk about the need to modernize federal hazard data assets created through the Weather Act and uh NIHRP. Can you elaborate on these recommendations with the time we have left? >> Absolutely. I'll just briefly say, you know, all of the work that's done at the state and local level is dependent on the accuracy and how complete these federal data sets are. We have many federal data sets that don't accurately cover the entire country or are not upto-date and forwardlooking and that's critical for all of the decisions that are being made at the state and local level of what to build and where to build. >> Thank you and I yield back.

[clears throat] >> Representative yields back. Uh Representative Whitesides from California, you're recognized for five minutes. >> Thank you, Mr. Chairman. Let me start by saying uh to you and the ranking member that I really congratulate you for holding this hearing. I think it is one of the most important hearing subjects that we could hold as a science committee and I'm great I'm very glad to see the bipartisan spirit that animates uh today's hearing.

Um let me also say that uh I'm uh very uh interested in the testimony here. We have amazing witnesses and the the substance of your testimony I think should be studied and integrated into bills by this committee and other committees uh to help reduce the risk of these catastrophic disasters on the American people. [snorts] Let me start by talking about wildfires which pose a major threat to Southern California and California's 27th district which I represent.

The cost of the US per year of wildfires is estimated by the National Institute of Standards and Technology to be between 71 billion and 348 billion per year. For the January 2025 Palisades and Eaton fires alone, estimated costs are over 65 billion in economic losses with some estimates as high as $250 billion. Yet we actually don't spend that much on preventing and pre preparing for wildfire. For fiscal year 2026, the US wildland fire service budget request to totaled around 6.5 billion with 2.85 billion reserve for suppression.

Here's the important point. As as you have said, wildfire preparedness and prevention is a good investment for the federal government. We get a huge return for every dollar we spend on mitigation. Research shows that for every $1 we invest, we save $6 in future disaster costs. That's an amazing ROI. Uh, Mr. Chairman, if we can invest at the federal level more in prevention and preparation, we can shift the cost curve and save lives.

That's why I'm proud to have been involved in several efforts to improve our wildfire preparedness before coming to Congress, including supporting organizations like Earthfire Alliance and encouraging the development of technological capabilities like those deployed by Pano. And it's why I'm proud to support the innovative work from across the industry here today. Um, so I want to focus on what stands between you and getting these technologies into the hands of those that need the most.

Uh, Dr. Falcowski. Um, what are the barriers to technological adoption you have witnessed in the firefighting and fire sciences communities throughout your career and how is EarthFire Alliance working to address those barriers? >> Yeah, thank you Congressman for that question. Yeah, so a lot of the barriers that I've observed over my career again go to this communication piece and so not involving organizations that are making decisions on the ground, the firefighting organizations in this case, early enough in the process.

And so what we're doing at EFA is really trying to to meet that um need by including our early adopters in in the decisions that have been made for not only designing and building our satellites, but then also once we get our initial data flowing now um helping them uh integrate the data sets into their systems and understanding the barriers that they might face so we can better create data sets that are fit for purpose. >> Thank you.

You're doing such important work. It's such an exciting capability and I'm looking forward to working with your team to uh get it into a full-scale deployment. Um let's see uh next question um is for Dr. Gale. In your view, how can the federal government play a useful role in helping take nent technology uh ideas and and concepts and scale and operationalize them so that they can make a real difference in the wildfire crisis?

And different way to ask that is what kinds of packages and programs are most helpful to accomplish those goals. >> Thank you for the question. I think the federal government has the most critical role in accomplishing this task. We can learn a lot from the existing proven technologies on the ground. We just need to package it and deploy it on federal lands. We don't have a technology problem. We have an adoption problem.

And we don't need just one single technology solution. We need a layered technology approach that includes satellites, terrestrialbased camera systems like panos that includes fire prediction and modeling behavior so that we know how fires are going to spread. We need to take a layered approach that brings all these technologies together and I think the federal government is extremely well positioned to do so in partnership with Congress.

So, a combination of policy and lawmaking out of Congress to direct the federal agencies on how to do this and appropriators providing the resources and funding to do it will help ensure that federal agencies can do it quickly. We have the new US wild and fire service structured under the Department of the Interior. They have a strong role to play. Uh Chief Feny who's the first director of that agency is a technologist.

He sees the value in deploying all these technologies together. We have the will. We just need to create the mechanism so that we can put the technology on the on the ground quickly and move past some of these procurement barriers that are getting in the way. >> Thank you, Mr. Gil. My time is expired and I yield back. >> Gentlemen yields. I'd [clears throat] like to repres uh recognize Representative Biggs from South Carolina for five minutes. >> Thank you, Mr.

Chairman. Over the past few years, my home state of South Carolina has set the standard for disaster response, recovery, and resilience. In October of 2024, Hurricane Helen devastated communities across the upstate, which caused widespread power outages, home damage, flooding, and road closures. In response, people came together and rapidly proved resilient in the face of an unprecedented natural disaster. First responders, families, and small businesses alike provided resources to restore the damage caused by the storm.

As a lieutenant colonel in the Air National Guard of South Carolina, I am proud of the hundreds of guardsmen and women who stepped in and supported emergency operations, equip chainsaw teams to clear roads and assess damaged on damage on the ground. So my question is to all the witnesses, the South Carolina Office of Resilience has recognized the strength of disaster resilience as the state at the state level. What can you tell us about best practices for disaster resilience at the federal level? >> I'll go first and say that I think we can learn a lot from what states and local communities are doing in terms of disaster preparedness.

I think the first best practice to look at is get ready for what we know is coming. whatever disaster that is, whether it's fire, flood, hurricane, intense weather patterns that affect states like South Carolina, let's learn from what local governments and state governments are already doing to help their citizens prepare for these things and develop the resources that we need to get out in front of them. >> Yeah, I would say the South Carolina Office of Resilience is an incredible example of state leadership.

One thing to learn from is that they were very clear in putting someone in charge of resilience and saying this is who owns it and can coordinate across jurisdictions and across agencies to get to risk reduction. I think another thing they've done really well is actually looking at the data and being able to say these are the places where our landscapes are holding water and helping prevent flooding downstream. And they did that in an incredibly lowcost way.

And they've also been working with Clemson University to deploy lowcost sensors so that they have denser and better real-time innovation. So I think they're doing incredible work and there's a lot other states and the federal government can learn from. >> Thank you. Uh I think from a technology perspective uh adoption and successful implementation of technology at the state level can create uh uh the success story that's needed for federal programs to uh to incentivize other states and localities to adopt technologies.

You know uh the practices in South Carolina in the state of Texas and disaster information systems can uh become uh the models for for other uh you know regions. And I think that could uh uh become a pilot for uh technology implementation. >> Yeah. So I think when it comes to the federal government, predictable multi-year funding into how these collaborative approaches can be um convened is one of the best things that Congress could do.

We really need to agencies to coordinate and convene to derisk big bets uh and move uh to implementation faster by removing bureaucratic bar barriers that slow down investments and new tech and new solutions. Thank you very much. My next question is for Miss Roberts. Can you provide more information about the bake off that you've referred to um and how federal policy changes can help leverage new technologies? >> Yeah.

So right now a lot of our procurements as many of you know are based on you know multi-year RFPs that then lock in a vendor and a level of technology for often five to 10 years. As we've seen how rapidly technologies progressing in the last couple years I think it's clear we don't want to be locked into where we were even a couple years ago. So one thing we've thought about is expanding dynamic procurement. So instead of saying we already know the technology and we're going to just find the lowest cost, we can say as many of my other witnesses have also mentioned, how do we create low-risk environments where we can actually start to test technology and create stage gates where maybe in the first six months you have to meet a certain level.

If you meet that level then you progress to the next stage of the bake off and you get to do the next six months. And it gives us a lot more ability to understand the marketplace and the landscape and to see how the technology can prove itself rather than saying we need you to wait 10 years to prove this before we'll consider locking it in. I think it would just give us a lot more flexibility in a really dynamic landscape. >> Thank you so much.

And my time is expired. So I appreciate your participation here. And Mr. Chairman, I yield back. >> Yields back. Representative [clears throat] McBride. you're recognized for 5 minutes. >> Thank you so much, Mr. Chairman and Ranking Member Lofrren. And thank you so much to all of our witnesses for taking time today and bringing your expertise to share with us. Today's hearing is titled innovation in disaster prevention.

And so I want to focus on that word prevention and how the administration's attacks on federal programming are hurting Americans and specifically farmers in Sussex County within Delaware. Farmers are already suffering due to President Trump's reckless policies. From his war of choice with Iran to his nonsensical trade wars. On top of all of that, this administration has moved to cut funding for essential programs that help communities prepare for and detect disasters.

The president proposed eliminating the National Center for Atmospheric Research. And just a few weeks ago, his administration tried to dismantle the ocean observatories initiative, a network of buoys and instruments that provides coastal communities, including in Delaware, with a real-time data on water temperature, salinity, currents, and more. These continued attempts to dismantle, dissolve, and destroy programs that experts rely on to keep Americans informed and safe is simply disgraceful.

At a time when farmers are already under immense pressure, Congress must support the research, data, and tools that help communities not just respond to, but prepare before disaster strikes. Farmers in Sussex County face a unique threat. During hurricane season, not only do they see flooding, but they also see saltwater intrusion, which poisons crops, harms soil health, and can turn productive land into marsh. Putting technology in farmers hands is important.

A recent report found that 23% of North American farmers use Genai but only 10% of them pay for farmsp specific Genai tools. So uh Dr. Amastafavi farmers face differing climates and growing condition growing conditions. AI models are only as good as the data they're trained on. Consistent data streams are key for improve improving model performance. Yet data collection for key natural hazards has continuously been under attack by the administration.

This includes funding that supports USGS sensors for flood monitoring and seismic centers for earthquake and tsunami data. How will the quality and lead time of AI models and predictions be affected if missionritical data sources are substantially cut? >> Absolutely. Uh I would like to mention that um you know the advancements in frontier AI would not necessarily translate to disaster capabilities for different users.

The forward deployment requires domain specific data and in uh connected quality controlled and curated data sets are foundational for bringing frontier AI capabilities to disaster domain. >> Dr. Felowski, would you like to speak a little bit about how cuts to these programs that are collecting this data with complexity and integrity could impact our ability to use technologies to uh farmers benefits? Yeah, I'll just point out that the Earth observation data that's collected by Noah with the Ver instrument and the GO instrument and NASA before that have been really critical to supporting farmers and fisheries and other uh a other industries that have huge economic benefit to the US economy.

Um and certainly they've also played a big role in understanding uh the impacts that that fire play uh within our country but then also globally as well. >> And Dr. Folks, on top of that, I'm curious if you could speak to what Congress can do to make data collected by the federal government more accessible and useful for researchers and for farmers and for those with non technical expertise who want to leverage it to use AI for detection and prediction in a safe and responsible manner. >> Yeah, that's great.

That's a great question, Congresswoman. I appreciate that. And I think the federal government is already doing a lot to make data openly accessible. agencies like NASA, Noah, USGS for example, especially when it comes to satellite data sets, make them free and open. I think often what happens and what the barrier is is um communication of what these data sets can actually do for the everyday end user and and your local constituents.

And so I think communications is a big thing in helping people understand what's actually available at their fingertips. >> Thank you very much for that and thank you for this illuminating conversation today. My neighbors and I certainly have legitimate questions about what AI means for our f future. From privacy to energy demands, but as Congress, it is our job to establish appropriate safeguards while making sure useful tools and trusted federal data can reach the people who need them, including Delaware farmers and families who deserve the best data and research available to adequately prepare before disaster strikes.

So, thank you so much for lending us your expertise, and I look forward to supporting programs and legislation that do just that. Thank you. I yield back. Thank you. Uh, Representative Van Eps from Tennessee. You're recognized for 10 minutes. >> Thank you, Mr. Chairman, and thank you to our witnesses for being here today. I'm grateful for the work the chairman and this committee have done to prioritize natural disaster preparedness and response.

In 2023, Middle Tennessee was hit by an EF3 tornado, resulting in four fatalities and significant damage. In 2021, deadly flash floods struck Humphre County, tragically killing 20. When a catastrophic weather event strikes, the lives of Americans can change dramatically. I've seen it firsthand work in working in Tennessee state government. That's why we must use the best technology at our disposal to equip first responders to appropriately prepare and respond to these natural disasters.

I'm glad this committee is focused on productive ways to use AI that can help protect American lives and communities. However, the public sector continues to face challenges acquiring these advanced technologies at both the federal and state level. During your remarks, you all mentioned the need for agile acquisition vehicles for our federal disaster response agencies and just talked about the multi-year RFP uh challenges.

Mr. Gale, you specifically noted the DOW's use of other transaction authorities to accelerate the deployment of proven technologies in the field and the prospect of using a similar model for our disaster response agencies. I'm very supportive of these types of acquisition vehicles that allow the federal government to move at the speed of our technology advancements. Such vehicles have greatly improved our war fighter capabilities and the opportunity is there for our disaster response agencies as well.

Can you discuss the potential institutional barriers for granting these types of acquisition authorities to to disaster response agencies such as in-house procurement expertise and risk tolerance? I don't know that there's institutional barriers. The authority simply doesn't exist yet for all of the federal agencies. So I think it would be great to see Congress extend that same authority to agencies like the Department of the Interior Department of Agriculture so that they have the same tools in the toolbox as the Department of War. >> Great.

Thank anyone else any of the other witnesses want to respond to that? >> I would agree with what Mr. Gail [clears throat] said. I think it um a lot of these tools have been demonstrated within the Department of War and there's some great examples in the fire realm. uh the DOW certific program often does prize challenges and and and moves people that are successful uh in those challenges to award really quickly with OTAAS.

And so I think that's a model that could be replicated along with the the bakeoff idea that Miss Roberts keeps talking about is is this these prize challenges approach that can really move to rapid acquisition and contracting of of of proven technologies. I would also just add I think the challenge in the procurement process is especially clear when we think about the fact that we're not just talking about frontier AI capabilities.

We're also talking about some pretty rudimentary technologies that have been deployed for decades in the private sector. We're also talking about use of drones, use of satellites. And so I think you know it's not just the most kind of challenging and new technologies we're struggling with. it's quite old ones that we're not able to bring into the federal government yet. >> And I just add two points that um uh we need also capacity building for the users of the technology beyond procurement uh that uh kind of incentivize the adoption of these technologies to overcome the uh pre-existing risk aversions towards technologies. >> Great.

Thank you. [clears throat] Operational pivoting here. Um, operational deployment cannot happen without a transparent R&D to product pipeline. The small business innovation research program is essential for agencies to build this pipeline. Do you do you have any recommendations on how to make the program more transparent for researchers and developers in the emergency management space? uh as a company that uh has uh an active National Science Foundation SPR award.

I think uh the uh the this is a foundational uh mechanisms for bringing innovation to to this space. I think the recommendation would be going beyond the SPI so can create the proven technology but the last small adoption and uh the uh translation into operational practice requires additional programs such as incubators and innovation hubs that uh will help the the agencies adopt these technologies. >> Great. Thank you.

Anyone else in 20 seconds? Awesome. Thank thank you all so much. Thanks for thanks for being here, Mr. Mr. Chairman, I yield back. [clears throat] >> Gentleman yields. I'd like to recognize uh Representative Meny from great state of Texas. 5 minutes. >> Thank you very much, Mr. Chairman, and thank you to ranking member everybody for putting this together. When a disaster hits, government has to show up. You can't outsource the responsibility to keep people alive fully to the market.

But [snorts] the Trump administration is dismantling the very infrastructure that makes disaster response possible. They tried to cut FEMA's workforce by 50%. Literally half of the people who are responsible for coordinating disaster relief across the entire country. A federal judge ruled just last week that was unlawful. The judge wrote that the 50% staffing cut number appeared and is a quote as if pulled from thin air.

They've also proposed slashing the National Oceanic and Atmospheric Administration budget by 26%, eliminating satellites, weather models, and the forecasters who tell us when a storm is coming and how bad it'll be. That's cutting the warning system. Earlier this year, the Secretary of Energy came before this very committee and testified that he doesn't believe that severe storms and disasters in this country are getting worse.

I'd like to ask our witnesses today to think about that because everything that I've read in your testimony would suggest the exact opposite. For those of us on the Gulf Coast, this is personal. Houston and Southeast Texas have hit by disaster after disaster. Harvey, Emldda, Barrel, and every major storm that moves through the Gulf is coming towards us. FEMA is how our communities rebuild, and Noah is how we see it coming.

When you gut those agencies, you're gutting our ability to survive. And here's what I think about constantly. The Houston Ship Channel is home to the largest petrochemical complex in the entire country. Dozens of facilities, refineries, chemical plants packed along that waterway. A major hurricane going directly down the Houston ship channel would devastate more than just Houston. It would disrupt the energy supply for the entire United States of America.

That's a national security threat sitting just adjacent to my district. And the Trump administration is increasing the risk of that threat becoming a reality. Take the chemical disaster rule. It required chemical companies to conduct independent safety audits, share emergency planning information with local firefighters and police, and explain to the public what chemicals are stored near their neighborhoods. That rule came out of the 2013 West Texas fertilizer explosion that killed 15 people.

The Trump administration reversed that rule, said these protections were no longer reasonable or practicable. We [snorts] took them to court over that. And because rolling back chemical safety regulations along the most concentrated prochemical corridor in America right before hurricane season is the kind of decision that gets people killed. This administration is moving in the wrong direction on disaster prevention.

And I'm genuinely excited about the technology that we're discussing today. I know uh Dr. Mosavi uh your company is based in Houston and I'm very excited about that. preventing and mitigating disaster through AI, satellites, and real-time intelligence. Uh, this is the future, and I want communities across the country to benefit from it. We've got this Houston company at the table uh with AI deployed during Hurricane Barrel that, as I understand, helped emergency teams understand in real time which neighborhoods were hardest hit, whichies and grocery stores were still opening, where to focus resources first.

That's the kind of innovation that I'm excited to hear about, and that saves lives. But what I want the public to understand is that the private sector cannot replace the federal government on this. These AI systems run on data and the bulk of that data comes from Noah satellites, federal research programs, and FEMA systems. You can't defund the foundation while celebrating the building that's on top of it. These tools make government better.

They don't make government optional. They make it better. Disaster preparedness should be above politics. And the people who suffer when we get this wrong are the families in my district when we get back home. I want to start with uh Dr. Falcowski. Uh Houston experiences increasingly intense rainfall events. How can forwardlooking rainfall data such as Noah's Atlas 15 help design storm water infrastructure that is adequate for future flood risk? >> Um I'm not a rainfall expert.

I'm I'm a wildfire expert. So I would um take that question back for the record or let one of my co-panelists answer. I would think if you got major fires, you got water coming down on it. Um I I'll open this up uh for anybody on the panel. Um we've talked a lot in in Houston um about addressing updating our rainfall data uh because it's so integral to our being able to be prepared for the next flood and dealing with the major hurricanes.

So Miss Roberts or anybody on the panel uh who's comfortable answering? >> Yeah, that's a great point. And I would say, you know, ahead of Hurricane Harvey, we know that Texas was using outdated rainfall data. And I think the advances in Atlas 15, which are soon to come out, are incredibly important because we know local governments like Houston are making difficult budgeting decisions. And every time they're building infrastructure, we need to make sure that it's going to last for its full service lifetime.

And if we don't have accurate data, I mean, that's how all local and state governments decide what to build, where to build, and how to create the specs. And so, if we're giving them act outdated data, we're making sure that that infrastructure just can't last as long as it needs to. So, this is critical, as well as the other hazard data sets that others have mentioned. >> Thank you. And as a former local government official, I really appreciate your testimony on making sure that we're all working together to move forward and and be prepared for the next storms.

And with that, I yield back. Thank you, Mr. Chairman.