No recorders plotted on the marsh yet
Eleven stations are registered in the season roster, but none has a locked GPS fix — deployment coordinates are what turn a recorder into a point on this map. The console is ready before the hardware is.
AU-04 stopped uploading two nights ago
The SD card filled at 64 GB. The recorder keeps listening, but nothing leaves the card until a sync succeeds — everything since Sep 20 is buffered on-device, not lost. This failure is scoped to one station; ten others synced on schedule.
| Time | Station | Species | Conf. | File |
|---|---|---|---|---|
Zero otter detections in 30 days — and that's a resultAbsence is data. Occupancy models consume the full detection matrix, zero-cells included — silence across 11 stations for a month is a finding about the marsh, not a broken query. The matrix row for otter just wrote itself. WHY IT MATTERS · A zero here narrows the occupancy estimate exactly as a detection widens it.
| ||||
This week is three stations short of final
Hatched rows are provisional. Otter and water-rail recordings from AU-02, AU-07 and CFG-02 still carry classifier v4.1 labels; those two rows move when the v4.2 re-run lands tonight. The other six species are final.
Season wrapped — 214 nights, zero gaps
Every station filled every night of the Mar 1 – Sep 21 window. The occupancy model can run on an unbroken matrix — no imputation, no gap-flagging. This is the dataset the season was for.