← Sobeck. · a take
On beavers, evidence, and policy that skipped its test

The Beavers Probably Work. That's Not the Problem.

England wrote a national licensing policy off one advocacy-reported dry winter at one 10-hectare site — and the honest test hasn't run yet.

Beavers are back in London for the first time in four centuries, at a site called Paradise Fields in Ealing, and the story you've likely seen is a tidy one: beavers build dams, dams slow water, slowed water means no flooding downstream. The coverage (Fondriest, AP, the Independent) has the shape of corroboration — multiple outlets, one conclusion. Here's my problem: the beavers probably work, and the policy still doesn't follow from the evidence.

Grant the ecology its full strength first, because the argument isn't about beavers. Beaver hydrology is real, not voodoo. A dam complex turns a fast-moving channel into a stepped wetland: water that would have arrived downstream in hours gets spread, stored, and released over days. Peak flow attenuation is a physical mechanism, and I'd bet the Paradise Fields dams do exactly what the headlines claim. The Ealing Wildlife Group, who helped make the reintroduction happen, aren't selling snake oil. If anything, the mechanism is the strongest part of this story — which is precisely why it's doing all the load-bearing work the evidence should be doing.

Because here's what actually happened: a dry winter at one 10-hectare site produced a "no flooding downstream" headline, advocacy groups reported it, and outlets repeated it. A dry winter alone can produce that headline. Beavers or no beavers, rain that never falls doesn't flood anyone. What would settle it is rainfall-normalized attenuation data — flow measurements indexed against actual precipitation, so you can see what the dams did relative to what the weather did. Nobody's published that. Earthwatch ran water testing on the site, and those results are sitting unpublished while the press cycle runs on ahead of them. The unpublished data matters more than the published headlines. And the tell that nobody downstream is checking anything: the coverage can't even keep its kit dates straight — when the basic chronology drifts between outlets all carrying the same story, you're looking at one press release laundered through many bylines, not independent verification.

There's also an incentive-structure observation worth making, and it's not a conspiracy: Amazon's climate fund is in the mix on this project. That doesn't mean anyone's lying. It means the parties with money in the story are the parties whose story it is, which is exactly when you want the boring, adversarial data — the kind nobody's released yet.

The real test hasn't run. The first genuinely wet winter on the first genuinely difficult catchment — culverts, roads, landowners who don't want wet ground — is where this story gets honest. Difficult catchments are where attenuation gets complicated: water that's slowed has to go somewhere, and somewhere is often someone's field or someone's road. That test will produce real data, and some of it will be inconvenient, and the policy was written before it could. England licensed wild beaver reintroductions nationally on the strength of one advocacy-reported dry year at one small site. Maybe the beavers will pass the wet-winter exam with distinction. I'd honestly bet they help. But "probably works" is a reason to run the trial, not to skip it — and a national policy that's already been written has a way of treating the trial's results as a formality.

← all takes