Inside Texas' next-generation flood warning system

Summary

---

Podcast Summary: Inside Texas' Next-Generation Flood Warning System

Podcast: Priorities Podcast by StateScoop
Host: Sophia Fox Sowell
Guest: Dr. Nick Fang, Director, Water Engineering Research Center, University of Texas at Arlington
Date: July 29, 2026

---

Episode Overview

This episode delves into Texas' ambitious, state-funded $4 million project to modernize and revolutionize flood monitoring and early-warning systems in the flood-prone Texas Hill Country—often dubbed “Flash Flood Alley.” Host Sophia Fox Sowell interviews Dr. Nick Fang, who leads the University of Texas at Arlington team building this next-generation alert network. Their conversation covers the region’s vulnerability, technological innovations, collaborations, how predictive modeling could save lives, and why this system might become a national model.

---

Key Discussion Points & Insights

The Unique Challenges of Texas Hill Country

[03:46]

• Geology & Hydrology: The area’s steep slopes and impermeable limestone/clay soils cause rain to run off rapidly, contributing to sudden, severe flooding.
• Multiple Rivers & Steep Terrain: Numerous rivers cut through the landscape, increasing runoff velocity and risk.
• Meteorological Convergence: Frequent convergence of moisture-laden systems from the Pacific and the Gulf heightens extreme weather risks.

Notable Quote:

“Hill country is a beautiful place if there’s no flood... The geologic condition—lotsa limestone and clay soil—make that landscape unique, but limestone is very impermeable, and when rain falls... the water [turns] into runoff very quickly... That’s the reason [it’s] called flash flood alley.”
—Nick Fang, [03:59]

---

How Flood Monitoring & Warning Systems Help

[05:48]

• Lead Time: Fundamental hydrology shows a delay exists between rainfall and peak runoff, allowing for early warnings.
• Purpose: The system’s ultimate value is in providing more time for evacuations and preparations, protecting lives and property.
• Information: Key deliverables are when and where flooding will happen, and how severe it will be.

Notable Quote:

“Such a system... really [provides] time. So we can predict based on current rainfall when the flow, flood water’s going to happen, how high will [it] be... [It] can tell where the flooding happens. That’s very important for decision making.” —Nick Fang, [05:48]

---

Technical Architecture

[07:16]

Dr. Fang breaks the system into three major components:

1. Data Collection:
   • Rainfall & river conditions (flow, elevation), primarily leveraging National Weather Service data and Texas Tech’s expanding network of weather stations and radars.
   • Federal and state data sources are merged for greater accuracy and density.
2. Modeling & Analytics (“Engine”):
   • Hydrologic modeling and machine learning modules process data to predict flood locations, timing, and severity.
3. Communication:
   • User-friendly, geospatial warnings are distributed to emergency managers and agencies, showing who’s at risk and when.

Notable Quote:

“The second module is... where hydrologic model and the data machine learning module [are]... The module can generate what’s the natural reaction from rainfall and map where the flood’s going to happen and how high... The third module... [dispatches] the information in a very user-friendly format.”
—Nick Fang, [07:16]

---

Collaboration & Data Integration

[08:49]

• System uses established national data as the foundation, then builds on it with enhanced local monitoring via new weather stations and radars.
• Collaboration includes Texas Tech University, the Upper Guadalupe River Authority, Texas Water Development Board, Texas Department of Emergency Management, and TxDOT. The Governor’s office provided funding.

---

Recent Flooding & Lessons Learned

[11:35]

• The July 2026 Guadalupe River flooding reaffirmed the urgent need for real-time warning systems.
• Although the most recent rains missed highly vulnerable areas, relying on luck is not a strategy.
• Meteorological patterns are repeating; only proactive systems can keep communities consistently safe.

Notable Quote:

“Dodging the bullet this year doesn’t guarantee in the future you’re going to dodge it again, I mean, by luck. So we need to really kind of make sure setting up a very good system for the whole Texas and that is the community really needed.”
—Nick Fang, [11:35]

---

National Models, Transferability, and AI

[13:11], [14:43]

• The system is designed to be transferable and applicable nationally, thanks to the universal availability of radar rainfall data and adaptable modeling.
• AI & Machine Learning: Essential for processing large, real-time data volumes but must remain “highly supervised” to ensure reliability and practical decision support. Human review prevents false alarms or missed warnings.

Notable Quote:

“These days, [it’s] very hard... not [to use] AI, machine learning on those particular data-driven approaches... Ours is highly supervised because we want to make sure all the data prediction... is really reliable... human intervention [is needed] before we send out.”
—Nick Fang, [14:43]

---

Timeline & Implementation

[13:59]

• Grant runs through September 2027.
• Late fall 2026: First operational pilot system launches, with ongoing in-field testing and refinement through 2027.

---

Training & Operations

[15:50]

• Comprehensive training for emergency managers and local officials is considered essential.
• The University will remain involved post-launch to ensure smooth adoption and handoff to state or local agencies.

Notable Quote:

“Training is very important. Once we develop a system, actually we’re going to provide a series [of trainings] to local emergency personnel and also we’re going to stay on the project for a while to make sure... people [are] able to operate [it] themselves as well.”
—Nick Fang, [15:50]

---

IT Integration & Handoff Challenges

[16:45]

• UT-Arlington can host the system and handle data needs, alleviating the concerns that often come with integrating new IT with legacy government platforms.
• Ongoing discussions about future operational handoff to state agencies.

---

Memorable Quotes (with Timestamps)

• “The time is very important. So such a system... really providing... is the time.” —Nick Fang, [05:48]
• “Hill country is a beautiful place if there’s no flood.” —Nick Fang, [03:59]
• “Dodging the bullet this year doesn’t guarantee in the future you’re going to dodge it again, I mean by luck.” —Nick Fang, [11:35]
• “Training is very important... [We’ll] stay on the project for a while to make sure... people [are] able to operate [it].” —Nick Fang, [15:50]
• “Machine learning is a highly supervised [tool] because we want to make sure all the data prediction... is really reliable.” —Nick Fang, [14:43]

---

Timestamps for Key Segments

• Texas Hill Country Flood Risks: [03:46]–[05:28]
• How Early Warning Saves Lives: [05:48]–[06:57]
• System Architecture: [07:16]–[08:49]
• Collaboration and State Agencies: [08:49]–[10:29]
• Lessons from Recent Floods: [11:10]–[12:33]
• Transferability to Other States: [13:11]–[13:45]
• Timeline & Funding: [13:59]–[14:26]
• Role of AI & Machine Learning: [14:43]–[15:32]
• Training for Emergency Personnel: [15:50]–[16:29]
• IT Integration: [16:45]–end

---

Final Thoughts

The episode presents a compelling look at harnessing advanced technology and inter-agency collaboration to tackle one of the deadliest, fastest-moving natural threats in the country. Dr. Fang’s team’s approach—leveraging national data, local expertise, predictive modeling, and AI—sets a precedent for flood-prone regions nationwide. Their commitment to practical training and phased deployment underscores the project’s goal of saving lives through science, technology, and shared knowledge.

Transcript

[A]

Hello and welcome to State Scoop's Priorities podcast. I'm Sophia Fox Sowell, a reporter for statescoop. This week you'll hear my interview with Nick Fang, director of the Water Engineering Research center at the University of Texas at Arlington. He's leading a new $4 million state funded effort to build an advanced flood monitoring and warning system for the Texas Hill country known as Flash Flood Valley. We talked about how the technology uses weather data and predictive modeling to give emergency managers more time to respond and how it could become a model for flood prone communities across the country. But first, here are the biggest state IT stories of the week. More than 30 Minnesota communities saw their water and wastewater utilities disrupted by a coordinated cyber attack this week. Though none of the local governments reported anything more than a lapse in operation, Analysts said they are concerned that the US Will continue to face similar threats, particularly as attackers are aided by frontier AI models. California Chief Information Security Officer Vitaly Panich will step down after seven years leading the state's cybersecurity strategy, he announced on social media. Paneech, whom Gov. Gavin Newsom appointed in 2021 after serving nearly two years in an acting capacity, helped steer California through a period of rapidly evolving cybersecurity threats, including the COVID 19 pandemic, escalating ransomware attacks and the state's growing adoption of artificial intelligence. Following recent updates to nationwide broadband availability data, some states may be required to cut one third of the awards to satellite Internet Companies through the $42.45 billion broadband equity access and Deployment program, according to an independent review published last week. The Advanced Communications Law and Policy Institute released its findings as the BEAD program moves into its final phases. One year after catastrophic flooding killed more than 100 people along the Guadalupe river, another round of deadly July flooding swept across central and South Texas, creating flash flood emergencies that forced evacuations and rescues. Floods remain one of the deadliest and most difficult natural disasters to predict, especially in places like Texas Hill country, where water levels can rise in minutes. To give emergency managers earlier, more precise warnings, a team at the University of Texas at Arlington is leading a new $4 million effort to build a next generation flood monitoring and warning system. The system will combine advanced sensors, weather data and real time modeling. Nick Fang, one of the engineers leading the project, said the university will use National Weather Service data and partner with Texas Tech for additional weather stations and radars. If successful, he said, this system could become a model for flood prone communities nationwide.

[B]

So Water Engineering Research center is being kind of really focused on many Water related challenges. So flood is one of the kind of challenges we've been looking into. And we have a decades of experience also developing radar based flood warning system for Texas, many different parts. And I think it just kind of based on experience and what happened last year, we've been kind of appointed with this particular important task to really help Hill country people there as well. So that was the kind of the relationship, I mean the experience and also where we are in Texas and the research we've been doing really kind of fits that particular need right there.

[A]

No, that makes sense. Tell me about why Texas Hill Country. I mean flash flood alley. Why is it such a vulnerable place for these types of catastrophic floods and really severe impacts?

[B]

So I want to first to start with. Hill country is a beautiful place if there's no flood. I mean it's very, very wonderful for people to visit. I mean very scenic mainly is because the geologic condition right there was a lot of limestone and clay soil right there and make that landscape very unique. But however, I mean this limestone is very impermeable and when rain falls on the ground and basically the water going to be turning into the runoff very quickly. Right. And secondly, here is many different rivers running through this Hill country area and at a very steep slope. So that when you have a steep slope in on the land ground elevation that usually just kind of keep the velocity go faster. That's one. The third reason is that area being kind of highly visited by different kind of weather system. So the moisture from a Pacific Ocean from a Californian could actually passing through and going to visit that area, the Gulf Mexico or Gulf America these days is a moisture you got to come in continuously joined right there too. So those are three different reasons happen in the summer every year like last year and this year July. That just kind of really is a normal demonstration of three different factors playing in one place. And together it happens. And there's a reason called Hill country flash Alley number flood alley. That's the reason.

[A]

Hmm. That seems like a huge conflicts of factors to try and mitigate and control. Please tell me how flood monitoring and flood warning systems can really help combat that and give emergency managers more of a leg up to prepare and give communities more time to respond.

[B]

So we actually take advantage of fundamental knowledge from hydrology is when rain hits the ground, the runoff is not being generated right away. There's always there's a delay happening right there too. We typically call the lead time or the response time from the runoff itself. So we want to take advantage of the time from running first dropper running hits the ground and when the peak flow going to happen. That kind of the delay time is where people can take a measure. Right. The mitigation you can do also I mean evacuation that time is very valuable for people to do. People can escape from there and also can sending the rescue crew right there. The time is very important. So such a system actually really providing. What's really providing is the time. So we can predict based on current rainfall when the flow flood water going to happen, how high will be. And that's. That's the system can provide it then and also can tell where the footman happen. So those are those information is tell people where and when going to happen is very important for decision making.

[A]

Yes, absolutely. For emergency managers. Tell me about some of the technology that's being used. Is it predictive modeling metrics? Is it sensors? I know warning systems and sirens. The system is designed to kind of go out and notify before then. So it has to be faster. Tell me about some of that tech.

[B]

So at a high, very high level, typical flood warning system has three major components. First component is the data collection and that actually collects rainfall data and also the river condition like a flow data, elevation data itself. And the second module is a. That really we call the engine of a system is where hydrologic model and the data machine learning module there as well. Taking the data from first component feeding into the second module. And the module actually can generate what's the natural reaction from the rainfall and put on the map tell people where the flood going to happen and how high will it be. And the second module will feed it into the third module which is a communication module. You dispatch the information in a very user friendly format. People can take that information, say tell me how much time I have, how high the water you can reach up to you tell me. So on the geospatial format people can know where it happens. So those are three components working very well with the 20 years experience we have actually that really deliver very nicely for people to give a reliable information.

[A]

I want to take each of them kind of differently. So with data collection, I'm curious how is this different or how will it work in tandem with some of the warning systems and data from the National Weather Service? Is UTA planning to build a flight flood monitoring and integrate some of that federal data or national data that's already there.

[B]

So we actually going to start with the national data, National Weather Service data itself. Because the widespread network already there and those devices already be picking up very good rainfall Data there too. But also we work closely with Texas Tech because these are also funded Texas Tech to really generate additional weather stations in across the Texas and also the area in Hill country there as well and putting additional radars. So those are new weather stations and the radars and can pick up or enhance the network density. Pick up a more detailed information that actually make the technology we use the years back. I mean it's the current with what we're going to have in the next year will be very different because detail level will be different and reliability, accuracy, the radar rainfall data will be much better than before. So that's a data collection piece. We are not actually building our own data collection module. We we collaborate with the Texas Tech research researchers right there as well. So this is a kind of a very unique. I mean it's addition to the conventional method on the data collection module.

[A]

I think that makes a lot of sense. I mean you're taking federal components and federal data and a foundation that's already there. You're collaborating with another university using their their resources and not building something from from scratch or just adding to it and supplementing to really provide a service to this region of the country that is incredibly vulnerable. What other state agencies are you connecting with? Moving on to the communication standpoint. Who will these warnings and notifications go out to in terms of state and local officials?

[B]

So we currently actually being very close communication with the Upper Guadalupe River Authority and also work with the Texas Water Development Board and obviously governor's office is the funding source for our research itself. And we also intend and working closely with the Texas Department Emergency Management cartidum and this information eventually can be utilized by different state agencies as well. And I think we are actually have very close relationship with the TxDOT, I mean Texas Department of Transportation and they can utilize the information down the road eventually can inform which bridge is going to be really kind of endangered itself

[A]

from last year's floods to this year's floods. Obviously Governor Abbott announced the $4 million grant in in May, almost a year after the Kerr county floods of last year. But then just last week the Guadalupe river flooded again and hit Kerr county pretty hard. Is there anything that you've learned, you and your team have learned from this most recent round of flooding that you're hoping to incorpor monitoring system.

[B]

So I would say that the event happened last week. It really reconfirmed the importance of setting up a flood warning system for Texas. Because obviously the people in curvy this year, I mean they experienced less flood. I Mean than last year. But I mean, so we did analysis. The rainfall happened this year is just kind of about 12 miles away from South Fork. I mean, area which was the very devastated area last year where the camp mistake. I mean, what happened to Camp mystic happened. It's not, not far away from there. So literally say, I mean, meteorologically speaking, the similar flood happened last year could have occurred again. So what I'm saying here is dodging the bull this year doesn't guarantee in the future you're going to dodge it again, I mean, by luck. So we need to really kind of make sure setting up a very good system for the whole Texas and that is the community really needed.

[A]

I think you said that very well. Flooding, I mean, obviously is a natural disaster that happens in all areas of the country. I remember a few years ago during Hurricane Helene in North Carolina, specifically, their flood gauge system gave emergency managers, public safety offices and communities a huge, huge head start in terms of evacuations and mitigating the damage that's been impacted. Are you looking to other, other states, other systems that already use this flood warning system, this flood monitoring technology to kind of base what you're building in Texas Hill Country?

[B]

Absolutely, because the technology we develop here is very transferable and can be applied to different states and different kind of terrain. And because the radar rainfall data is kind of widely available this moment, we can actually take it from anywhere from different states and building up a very good hydrologic hydraulic models and we can utilize this particular information to give a very useful warning message which can be dispatched through the communication package. Right, as well. Yeah.

[A]

So Governor Abbott gave and announced this grant in, in May. It is now towards the end of July. How long do you anticipate needing or what is the timeline in terms of making this actionable and functionable?

[B]

So the funding. Funding the project will end September 2027. And we actually got an official, I mean, start this up this spring and we tried really hard to kind of set up a system which is going to be functioning and operational in late fall this year. And that system needs to be tested, I mean, through the, through the rest of time. So that's where the timeframe we're looking at at this moment.

[A]

Okay, I think that sounds great. Especially in September. That's only a few months away. Are you using artificial intelligence, AI, automation, anything to any large language models to help, I don't know, shore up the system, learn from it, or like actively evolve over time?

[B]

You're reading my mind. I mean, these days, these days, very hard. I mean, it's not using AI machine learning on those particular data driven, I mean approach because our system process constantly large volume data including from the rainfall and prediction on the hydrologic perspective. And so I mean some machine learning AI is a very handy tool for us to process data and to lead to the final decision point. But I mean obviously this is the machine learning. AI is not a kind of regular AI. Machine learning is a highly supervised because we want to make sure all the data prediction from this particular module is really reliable. So we need to put in the human intervention, really check on the results before we send out.

[A]

Is there any training that needs to happen? I mean obviously at the University of Texas at Arlington you're all very familiar with these systems. Is there any training that needs to happen on the state or local side so that these emergency managers or officials become a little bit more comfortable and confident in using these systems too?

[B]

We regarded training as a very essential piece on the flood warning perspective because obviously you want to really kind of understand what's the output from this particular system. How are you going to translate this information turning into actionable operation. Right. And so training is very important. And for us once we develop a system actually we're going to provide a serious training to local, I mean emergency personnel and also we're going to stay on the project for a while to make sure that people utilizing this particular system, they're able to kind of operate themselves as well. So I would say training is very important.

[A]

I'm also curious on the state and local agency side. I know oftentimes that government IT systems are not as updated, not as developed and they cannot integrate data as well. Are you finding any of those challenges as you're starting to build these systems?

[B]

No. So being being a research institution, we have our advantage right here. We have a pretty good, very good computational facility and also the data storage. We can host such a system at a university and provide information to the state agencies operation there as well. That's one as a one options and obviously, I mean working we are still in talking with the state agents. How are we going to really do a very good handoff? I mean it's a particular operation, so there are still variety options ahead of us.

[A]

Thank you to Nick Fang for participating in that conversation. You can subscribe to the Priorities podcast@monities podcast.com and wherever you get your podcast. While you're there, be sure to leave a review or rating on the podcast page. That small extra step helps more people like you find the show. This podcast is a production of Scoop News group in Washington, D.C. our producer, Carlin Fisher puts it together. Until next week, I'm Sophia Fox. Sowell, thanks for listening.