This conversation was recorded in April 2034 at a kitchen table in East Asheville, North Carolina, eight years after the National Centers for Environmental Information lost its federal funding. NCEI was once the largest archive of weather and climate data on Earth. Constance "Connie" Whetstone, 64, spent twenty-six years there as a quality control specialist. She now coordinates the Southern Appalachian Weather Cooperative, a volunteer network of 412 amateur weather stations, retired meteorologists, and hobbyists maintaining a regional climate record across western North Carolina and eastern Tennessee. On the table between us: a mug of coffee, a Davis Instruments rain gauge with a cracked funnel, and a spiral-bound notebook filled with station visit logs. The air smelled like the loquat tree outside her window, a species that, she noted without emphasis, did not fruit in Asheville when she moved here in 2003.

Seven years of running this network. What does a normal day actually look like?

Connie: I get up at five-thirty. I got up at five-thirty when I worked at the center, so that part's the same. Check the overnight uploads. About 280 stations report digitally now; the rest are phone-in or, God help me, paper mail. There's a woman in Burnsville who mails me index cards. Beautiful handwriting. I enter those myself.

By seven I'm running QC scripts on the overnight batch. Flagging outliers, checking for sensor drift, comparing neighboring stations. Two hours if nothing's weird, four if something is. Then I usually have a station visit in the afternoon. Someone's radiation shield is warped, or their rain gauge is reading low, or they moved it six feet to the left and didn't think to mention it. I drive out, I look at it, I fix what I can. Home by dinner.

It's bookkeeping. That's what people don't understand. The whole thing is bookkeeping.

What's the hardest technical problem?

Connie: Calibration. Not even close. When I was at NCEI, every observation had been through at least two layers of quality control before I touched it. Instruments maintained on a schedule. Siting documented. You could trace any reading back through the whole chain: who measured it, with what, when the sensor was last calibrated, what reference standard was used.

Now I have 412 stations, and maybe sixty of them have instruments I'd call properly calibrated. The rest are consumer-grade thermistors that drift.¹ They drift warm, usually. Half a degree, a degree. Doesn't sound like much. But if you're comparing this July's average to 1990's and your instrument reads consistently high, you've just manufactured a trend that isn't there. Or masked one that is. Either way, you've corrupted the thing you're trying to protect.

Pressure sensors are worse. We used to calibrate against airport altimeter readings; you'd call the ATIS line and get a clean reference.² Most of those automated systems are technically still running, but the maintenance... I don't know who's maintaining them now. I don't trust them the way I used to. So I've got three mercury barometers that I drive around in a padded case like a lunatic. My husband calls it "the baby."

How much of what your volunteers submit is actually usable?

Connie: On any given day? Maybe half.

Half.

Connie: I know how that sounds. The literature says 53 percent data loss after quality control in citizen networks, and we're right about there.³ But that number misleads people. It's not that half my volunteers are doing it wrong. Quality control is subtractive. You throw out the readings you can't verify. Doesn't mean they were wrong. Means you can't prove they were right.

The thing that took me a while to accept is that the network is the instrument, not any individual station.⁴ If I have twelve stations in a twenty-kilometer radius and eleven agree and one doesn't, I know something. That redundancy is the only reason any of this works. At NCEI we had precision. Now I have density. It's a different kind of knowing. Worse in some ways. Better in a few I didn't expect.

What can't your network do that NOAA used to?

Connie: (long pause)

Anything vertical. Nobody in my network can launch a radiosonde. Nobody can measure stratospheric ozone or free-tropospheric CO₂.⁵ Nobody can cross-validate against satellite data because the satellites are still up there, but the ground-truthing infrastructure, the people who reconciled what the satellite saw with what the surface measured... that expertise dispersed. Some went to universities. Some to the private forecasters. Some of it just stopped.

I can tell you what the temperature and precipitation were at 412 points in western North Carolina yesterday. I can tell you that with reasonable confidence. I cannot tell you what the atmosphere was doing above those points. I cannot produce Climate Normals that would meet the old standard.⁶ I cannot do what NCEI did.

I want to be clear about that. Anyone who says citizen science replaced the federal climate record is being kind to themselves.

Does that make the work feel futile?

Connie: No. It makes it feel partial. I can live with partial.

Tell me about your volunteers. Who shows up for this?

Connie: Oh, everyone. I've got retired NWS forecasters who are meticulous. They'll email me about a two-tenths-of-a-degree discrepancy at three in the morning. I've got a goat farmer in Madison County who's been recording rainfall since before CoCoRaHS existed because his father did it and his grandfather did it.⁷ I've got a high school in Waynesville that built their own station from a kit. The data's terrible, but the kids are learning to read a psychrometer, so I keep them in the network and flag everything.

The hard part isn't getting people to start. It's getting them to keep going. Year two is when you lose them. The novelty wears off, the rain gauge clogs, they go on vacation and forget to arrange a substitute. So I send birthday cards. I send holiday cards. I drive out and fix their equipment for free. It's retention marketing, basically. My old career finally useful. I was in university fundraising before NOAA.

That's a strange pipeline into climate science.

Connie: (laughs) You should see the pipelines now. My best station operator in Swannanoa is a retired dental hygienist. Steadiest hands in the network. Perfect calibration logs. I asked her once why she's so meticulous and she said, "Honey, I spent thirty years looking at things people neglected. This isn't different."

There's a line from Louis Uccellini, former NWS director, that I keep returning to. "Forecasts are for today, observations are forever."⁸ Does that still feel true?

Connie: (quiet for a moment)

The observation is the only thing that's real. A forecast is a guess. A model is a simplification. The observation is what actually happened, at that place, at that time. If nobody writes it down, it didn't happen. Not for science. Not for the record.

Whether "forever" means anything... I don't know. I don't know who's going to use this data. I don't know if the quality will be good enough for what they'll need.

So it's not faith in a future recipient.

Connie: No. I know what it costs to stop. I was in the building when the record stopped. I know what a gap looks like, and I know you can't go back and fill it in.

(She picks up the cracked rain gauge from the table, turns it over in her hands.)

This one's from a station in Hot Springs. Volunteer called me last week, said a bear knocked it off the post. I'm going to epoxy it and drive it back up there Saturday. It'll read a little low after the repair. The funnel's not perfectly round anymore. I'll note that in the metadata. That's what I do. I note it in the metadata and I keep going.

Constance Whetstone's station data is archived on servers maintained by the University of North Carolina Asheville's Atmospheric Sciences department and mirrored on two personal hard drives she keeps in a fireproof safe. The Southern Appalachian Weather Cooperative accepts volunteers with any level of experience. "Bring a rain gauge," she says. "I'll teach you the rest."

Footnotes

1. Citizen weather stations commonly use thermistors rather than the platinum resistance thermometers found in professional reference instruments, leading to calibration drift over time. See Aston University intercomparison study: https://pure.aston.ac.uk/ws/files/18872960/Citizen_weather_stations.pdf ↩

2. The Citizen Weather Observer Program historically encouraged members to calibrate pressure sensors against nearby airport altimeter readings. CWOP Guide: https://weather.gladstonefamily.net/CWOP_Guide.pdf ↩

3. Comprehensive quality assessment of crowdsourced air temperature data found that erroneous metadata, data collection failures, and unsuitable sensor exposure reduced usable data by 53%. Earth System Science Data: https://essd.copernicus.org/articles/14/4681/2022/ ↩

4. The Bulletin of the American Meteorological Society documented that treating citizen observations as a network rather than individual stations is central to resolving quality-control challenges: https://journals.ametsoc.org/view/journals/bams/101/1/bams-d-18-0237.1.xml ↩

5. NOAA's Global Monitoring Lab, which tracked greenhouse gases, solar radiation, aerosols, and ozone levels, faced mass furloughs beginning in 2026. Colorado Sun: https://coloradosun.com/2026/04/03/half-of-staff-at-boulders-noaa-global-monitoring-lab-face-furloughs-as-funding-freeze-drags-on/ ↩

6. Climate Normals are developed from daily observations quality-controlled first by NWS meteorologists, then by automated procedures at NCEI. NOAA NWS: https://www.weather.gov/news/212304-citizen-science-climate ↩

7. The NOAA Cooperative Observer Program has included families maintaining stations for generations; the Maddox Family of Rome, Georgia, holds the record at 165 continuous years. NOAA NWS: https://www.weather.gov/news/212304-citizen-science-climate ↩

8. Quote from former NWS Director Louis W. Uccellini. NOAA NWS: https://www.weather.gov/news/212304-citizen-science-climate ↩