eBird Power User Files Formal Appeal After App Flags 'Impossible' Sequence of Shorebirds

The dried mud had ridged itself into a fingerprint across the confirm button of a phone held over the flats at 06:14, and Nora Beaulieu pressed it anyway, the way you press anything when the birds are still out there and the tide is going to turn.

Nora is what the local count calls a power user: 41 checklists from the same mudflat since spring, submitted before the coffee is finished, each one annotated with wind and tide because she read somewhere that annotating wind and tide makes a list look considered. On the morning in question she logged fourteen shorebird species in a single morning session, dunlin, semipalmated plover, stilt sandpiper, white-rumped, western, pectoral, least, and the rest of the usual autumn parade in rough plumage, southbound birds still in heavy molt and picking along the wrack line where the previous night's onshore wind had pushed them.

The app's reply arrived before she reached the parking lot: This sequence of species may be improbable for this location.

Nora read it four times, then read the fine print beneath it, which explained that the platform compares a submitted list against lists already accepted from the same location and flags submissions that fall far outside the expected pattern. The word improbable, she later wrote in her formal appeal, "implies a judgement about my eyes that the software has not made about anybody else's, and I would like it either withdrawn or given a reason."

She submitted the appeal at 09:48 with eleven attached items: a tide table, a wind log, six photographs of birds in the kind of distant, backlit, low-contrast conditions that make every shorebird a silhouetted argument, and a short paragraph explaining that a strong overnight onshore wind had compressed the outer flat's birds onto the narrow strip of exposed mud she happened to be standing on. The photographs are, by her own admission, "not great." One of them appears to show a dunlin. The others appear to show weather.

Petra Evers, who speaks for the standards board whenever birding requires a ruling delivered with the confidence of a building inspector, weighed in at the Tuesday meeting. "An appeal is not a rebuttal of arithmetic," she said, and let the sentence stand in the room until somebody else found a way to break it. "The algorithm does not owe the observer a conversation. The observer owes the algorithm a reason, in writing, and preferably with a photograph of the bird in question rather than the general area in which the bird might plausibly have been."

Nora's photograph of the general area is now in the file.

Elsewhere in the county, Vera Iverson, who has, by her own account, tried every piece of birding advice published online simultaneously since March, has meanwhile discovered that reporting the same flock twice, once counting left to right and once right to left, produces a checklist the platform flags for an entirely different reason. She has opened a second appeal. She would like it noted that she is following all recommendations.

The record itself has already moved on. Vera Ortega, assembling the group's monthly update, entered Nora's fourteen species in the summary table without annotation, because the table does not have a column for onshore wind, and because Vera's instruction to herself each month is to report what the group now believes happened. The number is in the document. The reason the number exists is in Nora's appeal folder, awaiting review.

As of this filing, the flag stands, the appeal is open, and the fourteen species remain visible on Nora's own checklist, where she can see them by scrolling. She scrolls most mornings before work. "They were there," she said, out on the flats at 06:14 with the tide already turning, phone held up in the salt air like an object making a claim. "I don't need the platform's permission for the tide. I would, however, like an explanation of the word improbable."