The Pond That Always Held

▶️ Prefer to watch? Every story’s a short video → https://www.youtube.com/@GetSafetyIntel · 📣 Got a near-miss? Share it anonymously (QR + link at the bottom).

Forty-one years ago today, two mud ponds that had sat above an Alpine village for two decades let go at lunchtime and buried 268 people in under a minute. Nobody blew them up. They failed because “it’s held this long” got mistaken for “it’s safe.” Here’s the day it happened — and the discipline the people downstream of every dike, berm, and settling pond live by.

At 12:22 on a bright July afternoon, above the village of Stava in the Italian Alps, the upper of two tailings dams serving the Prestavèl fluorite mine collapsed onto the lower one, and the lower one went with it. Roughly 180,000 cubic meters of sand, silt, and water — mine waste that had been pumped up there for years to settle out — came down the valley at about 90 km/h (56 mph). It ran 4.2 km to the Aviso River and swept away three hotels, 53 houses, six warehouses, and eight bridges. It killed 268 people — families, hotel workers, kids on summer holiday. It took less than half a minute.

The dams didn’t fail from a storm or an earthquake. They failed from being poorly built, poorly drained, and never truly watched. The upper embankment was too steep and sat on the soggy, un-drained tailings of the pond below it. The waste didn’t dewater the way the operators assumed it would, so water built up inside the embankment. Drainage pipes were laid wrong, sagged under the weight of the sludge, and leaked. A 1974 stability check flagged the upper dam as marginal — and nothing was done. In the months before the collapse, the ground itself sent warnings: a sinkhole opened at the toe in January 1985, water and sand seeped from the embankment, another pipe ruptured and opened a sinkhole in June. There was no instrumentation to monitor stability. After they were first built, regulators never inspected the embankments at all. A court later found the plant “designed, built and managed in such a way as not to offer those safety margins that civil society expects.”

Read the American version and it’s the same paragraph: Buffalo Creek, West Virginia, February 26, 1972 — a coal-slurry impoundment built with almost no engineering gave way and killed 125 people in the hollows below. That disaster is why the U.S. wrote real rules for the ponds that hold mine waste: 30 CFR 77.216 — MSHA’s impoundment standard. In plain terms, if a structure impounds water, sediment, or slurry and its failure could hurt people, it has to be designed and certified by a professional engineer, its plans reviewed and approved before you build it or raise it, and it has to be inspected on a set schedule by a qualified person looking for exactly what Stava showed: seepage, soft ground, settling, a rising phreatic line. Stava had none of that regime. Buffalo Creek is why we do. Both were written in mud and blood.

And here’s the part for everyone who doesn’t work near a mine: you almost certainly rely on some engineered thing that holds something back — a secondary-containment dike around a tank farm, a retention or settling pond, an ash impoundment, a berm at the low end of the yard. The Stava lesson isn’t about fluorite. It’s that “that pond has been fine for twenty years” is not a safety record — it’s an unpaid bill. The dam that never failed is not proven safe. It’s just not failed yet, and the weak signals — the wet spot at the toe, the slow settle, the seep everyone’s stopped noticing — are the invoice arriving early.

(A poignant footnote: this February, the 2026 Winter Olympics brought cross-country skiing to Tesero, the municipality that lost so many of its own at Stava. The valley that was buried now hosts the world. The Stava 1985 Foundation keeps the memory — and the prevention lesson — alive on site.)

A worker was killed in a school crawl space — and the willful line is chilling (July 13). OSHA cited a Texas building contractor and a staffing company after a January 7 fatality at Converse Elementary near San Antonio. A worker removing dirt from the school’s crawl space with a mini-excavator was crushed between the machine and a concrete beam. To make the excavator fit in the tight space, the company had removed the rollover protective structure and bolted on fabricated parts — that’s the one willful violation. On top of it: 15 serious confined-space citations — no permit-required confined space evaluation, no atmospheric testing, no ventilation, no rescue plan, no training — plus two more against the staffing firm for putting temp workers in that space untrained. Proposed penalties top $299,000. So what: the second you modify equipment to squeeze it into a space it was never meant to enter, you’ve left the manufacturer’s safety envelope and you’re improvising with a life. A crawl space full of accumulated dirt is a permit-required confined space — test it, ventilate it, plan the rescue, and train everyone who goes in, including the temps.

CSB: two young workers died of H2S because the systems that would have warned them didn’t exist (July 14). The Chemical Safety Board’s new update on the January 27 hydrogen-sulfide release at the Woodland Pulp mill in Baileyville, Maine is hard to read. A 20-year-old co-op engineering student and a 26-year-old engineer new to the company were working an unrelated drawing project when H2S — generated in an acid sewer during a cost-driven shutdown, then freed when the scrubber fan was switched off — flowed into their building. They collapsed. They weren’t found for hours, because no one was tracking who was in the building. There were no fixed H2S detectors in the area and no personal monitors on the workers. CSB says the company knew the acid-sewer H2S hazard existed. So what: H2S blinds your own nose at high concentrations — it gives no warning, so the detector has to. Shutdowns and startups are the deadliest windows in any process unit; that’s when you most need personal monitors, fixed alarms, and a hard count of who is inside and where. Knowing the hazard exists and not building the monitoring around it is the gap that kills.

A crew running a plant’s stormwater and process settling pond had a spot at the toe of the containment berm that was always a little wet. Always had been. New guys got told the same thing on day one: “Yeah, it seeps a little there, don’t worry about it.” One morning after a heavy week of rain, a maintenance tech noticed the wet spot had crept wider and the ground near it felt spongy under the loader — soft where it used to be firm. He almost said nothing, because “it’s always wet there” was practically written on the fence. He flagged it anyway. Turned out a section of the underdrain had partly failed and water was tracking through the embankment instead of under it. They caught it with a work order and a repair, not an evacuation.

Nobody was careless here. The crew had been trained by repetition to read that seep as normal — that’s how normalization of deviance works, quietly, one uneventful day at a time. The system’s hole wasn’t a bad worker; it was that “a little seepage” had no easy, blame-free way to become a report, so for years it never did. The fix was two things: a dead-simple way to flag a weak signal (a photo to a text line, logged, no lecture), and telling every crew out loud that seepage, settling, and soft ground at a containment are exactly the early warnings Stava’s operators had — and buried. The person closest to the work saw it first. He always does. The only question is whether reporting it costs him anything.

Got a fail or a near-miss? Hit reply, or scan the QR below — we’ll feature it anonymously — no names, no company, no blame. Just the lesson, so the next crew doesn’t learn it the hard way.

Walk the low point of your site today — the containment dike, the settling pond, the retention berm, the wall that holds something back — and look at the toe. Any new seep, any damp patch that’s wider than it was, any ground that’s gone soft or is settling: treat it as a report, not a shrug. Write it up the same day. Stava’s dam sent that exact signal for six months and everyone downstream paid for the silence. A weak signal you can report in thirty seconds is the cheapest safety system you own.

🎬 Rather watch than read? Every Safety Intel story is a short video on YouTube → https://www.youtube.com/@GetSafetyIntel — subscribe so you never miss one. 📸 Share YOUR near-miss (100% anonymous): https://scantoconfess.com — or just scan the QR above. 📤 Forward this to a crew member who needs it today.

Please stay Safe & Hydrated!!!