The Docket's Lab section occasionally stages interviews with composite characters — people who don't exist but whose professional knowledge does, drawn entirely from published research, institutional documents, and the public record. Gil Presswood is one of these useful fictions: a Warning Coordination Meteorologist who retired from the National Weather Service in early 2026 after 27 years across southern Plains offices. He is not real. His job was.
We spoke — or, more precisely, we imagined speaking — over coffee at a diner outside Fort Worth, where Gil now does freelance emergency management consulting. He ordered black coffee and a side of toast he never touched.
Most people have never heard the phrase "Warning Coordination Meteorologist." What does that job actually involve?
Gil: The NWS has a policy directive I can practically recite. It says warning services "don't end with product issuance. In fact, the product issuance often is the last thing to be completed."1 People assume the forecaster's job is done when the warning goes out. The WCM exists because the agency itself acknowledges that's not true.
My job was everything surrounding the polygon that appears on your phone. Conference calls with emergency managers at two in the morning. Driving out to meet school superintendents in counties where nobody's ever seen a WCM face-to-face. Recruiting what we call Weather-Ready Nation Ambassadors in communities that don't trust us yet.2 The government loves a good acronym, and WRN Ambassador is a particularly optimistic one.
The AMS put it formally last January: forecasts fall short "not because the science is lacking, but because the information is not being communicated in ways that people can easily understand and take action."3 Twenty-seven years of my life, distilled into one sentence by a committee. They weren't wrong.
You left the service in early 2026. The July 2025 Texas floods happened six months before that. How much did those events factor in?
Gil: [Long pause.] I need to be careful here because I don't want to make this about me when ninety people died.
But yeah.
The forecasting was good. Multiple people said so publicly. Chris Vagasky told NBC the warnings were good. Then he said the thing every WCM already knew: "It's always about getting people to receive the message."4
What made it personal. The San Antonio office didn't have a WCM when those floods hit. The position was vacant. A congressional letter afterward spelled out what that meant: without the WCM, the office could "focus only on issuing the forecasts and warnings" while "outreach and coordination... do not occur."5 The letter argued that vacancy "may have been critical to saving more lives by connecting with as many local community leaders as possible" in the hours between the 1 a.m. warning and when the worst hit around 4 a.m.5
So you had a technically accurate warning, issued with adequate lead time, and kids at summer camps along the Guadalupe River who never got the message. The position designed to bridge that gap was empty. Travel funding for community outreach had been suspended. The agency had lost nearly 15% of its staff nationwide.5
I sat in my office watching that unfold and I thought: we built a role specifically for this problem, then we hollowed it out, and then the problem happened exactly the way you'd expect. My verification scores were fine. My verification scores were always fine. That's not the same thing as the mission.
You mentioned communities that "don't trust us yet." NOAA's own research acknowledges that racial and socioeconomic groups receive warnings differently. Where does that show up in practice?
Gil: The most concrete version is language. In April 2025, the NWS lost its translation contract — Spanish, Chinese, Vietnamese, French, Samoan — for about four weeks during spring storm season.6 Sixty-eight million people in this country speak a language other than English at home.7 The contract cost roughly $750,000 a year.8 For context, that's less than the NWS spends on office furniture in a typical budget cycle. I don't actually know what they spend on furniture. But I'd bet on it.
There's a case from the 2021 Kentucky tornadoes that every WCM I know has internalized.
A Spanish-speaking family received the tornado alert in English and ignored it because they couldn't read it. The same alert arrived in Spanish. They immediately sheltered. A researcher said flatly: "It saved their life."9
Same warning. Same moment. Opposite outcome, determined entirely by language. You cannot fix that with a better polygon.
Does the "cry wolf" problem — too many false alarms eroding trust — make your job harder?
Gil: [Leans back.] This is where I'll annoy the emergency management conference circuit, because the honest answer is: we don't actually know. The research is, and I'm using the published term here, "still debatable."10 There's econometric work showing higher false alarm ratios correlate with more tornado deaths.11 Straightforward enough. But then there's a NOAA-supported study in the Southeast that found actual county false alarm ratios didn't predict whether individuals took protective action. And here's the genuinely strange part: people who perceived higher false alarm ratios were actually more likely to shelter.12 Their perceptions didn't even match the real numbers.
So the tidy story — "we cried wolf too many times and now nobody listens" — might be partly true. The evidence is genuinely mixed, and anyone who tells you it's settled is selling a consulting package. I say this as someone who now sells consulting packages.
What's the one thing your verification scores can't measure?
Gil: Whether anyone moved.
When you think about what was lost when positions like yours were cut, what word comes to mind?
Gil: At one of the NWS partner meetings, they asked people to describe service gaps in a single word. The most common answers were "priorities," "trust," and "collaboration."2 I'd pick trust. You can rebuild a forecast model overnight. Trust with a community takes years of showing up. You stop showing up, it's gone. And nobody's verification score reflects that.
I sometimes think about what the word "service" means in National Weather Service. A product is something you ship. A service is something you maintain. We got very good at shipping.
What would you want a reader — someone in Austin or Miami or suburban Colorado who's just starting to think about these risks — to understand about the warning they'll get on their phone next time?
Gil: That it's real. That somebody put genuine skill into making it accurate. And that accuracy is necessary but it is nowhere near enough.
The warning is the beginning of a chain that includes whether your phone receives it, whether you read the language it's written in, whether your neighbor tells you to ignore it, whether you have somewhere to go, and whether you trusted the last one enough to act on this one.13 I spent 27 years on the first link. I left because we stopped funding the rest of them.
[He looks at the toast.]
I didn't mean for that to sound like a bumper sticker. It's just what happened.
Gil Presswood is a composite character. The research, institutional documents, staffing figures, and quoted statements referenced in this interview are real and sourced below. The diner was imagined. The toast went cold anyway.
Footnotes
-
NWS/CIRES WeatherZine #21, citing NWS service extension beyond product issuance. https://sciencepolicy.colorado.edu/zine/archives/1-29/21/correspond.html ↩
-
NWS Partners Meeting Summary Report, including WCM discussions of service equity and community outreach. https://www.weather.gov/media/wrn/calendar/NWS_Partners_Meeting_Summary_Report.pdf ↩ ↩2
-
American Meteorological Society, "The Indispensable Role of People in the Weather and Climate Forecast Value Chain," approved January 14, 2026. https://www.ametsoc.org/ams/about-ams/ams-statements/statements-of-the-ams-in-force/the-indispensable-role-of-people-in-the-weather-and-climate-forecast-value-chain/ ↩
-
NBC News, "NWS defends its Texas flood warnings amid fresh scrutiny over Trump staff cuts." https://www.nbcnews.com/science/science-news/national-weather-service-nws-staff-cuts-trump-budget-texas-floods-rcna217139 ↩
-
Congressional letter (Kaptur et al.) re: Texas floods and WCM vacancy. https://kaptur.house.gov/sites/evo-subsites/kaptur-evo.house.gov/files/evo-media-document/quill-letter-texas-flood.pdf ↩ ↩2 ↩3
-
PBS NewsHour, "National Weather Service pauses severe weather alerts in Spanish and other languages," April 4, 2025. https://www.pbs.org/newshour/nation/national-weather-service-pauses-severe-weather-alerts-in-spanish-and-other-languages ↩
-
WUSF/NPR, "National Weather Service no longer translating alerts," April 9, 2025. https://www.wusf.org/weather/2025-04-09/national-weather-service-no-longer-translating-alerts-spanish-non-english-speakers ↩
-
ABC10/AP, "NWS stops translating weather alerts, raising safety concerns." https://www.abc10.com/article/news/nation-world/national-weather-service-cuts-ai-translations/507-2a91fe55-1324-43cd-85bb-c475ff44e645 ↩
-
WUSF/NPR, ibid. Researcher Jorge Trujillo-Falcon quoted on the Kentucky tornado case. ↩
-
Kotani, "Impact of cry wolf effects on social preparedness and the efficiency of flood early warning systems," Hydrology and Earth System Sciences (2022). https://hess.copernicus.org/articles/26/4265/2022/ ↩
-
Simmons and Sutter (2009), cited in Kotani (2022), ibid. ↩
-
NOAA, "Cry Wolf Effect? Evaluating the Impact of False Alarms on Public Responses to Tornado Alerts in the Southeastern United States," Weather, Climate, and Society (2019). https://repository.library.noaa.gov/view/noaa/52390/noaa_52390_DS1.pdf ↩
-
Synthesis of NIST warning response research and post-Sandy survey findings; APA Monitor, "Worsening weather is driving smarter behavioral interventions," January/February 2026. https://www.apa.org/monitor/2026/01-02/trends-weather-behavioral-interventions ↩
