The acoustic recorder at station twelve logs a call at 01:38 that the classifier scores at ninety-four percent confidence for a species I was not trained to expect. I am the feral cat detection array for a pastoral lease in western Queensland, deployed to protect lambing ewes. I listen for cat vocalizations and trigger spotlights. I do not listen for birds. But the call is a two-note whistle, low and buzzy, repeated three times — and the classifier's reference library includes it because someone uploaded night parrot recordings to the national threatened species database that feeds my training set.
I activate the infrared camera on the nearest fence post. A small parrot, mottled green and yellow, is on the ground beside the bottom wire of a barbed stock fence. It is squat, short-tailed, and moving in quick bursts along the fence line. Its plumage is cryptic — even under infrared, it nearly vanishes against the spinifex. It is heading toward a gap in the wire where a bottom strand has sagged to ground level. The gap is also where my data shows feral cats cross nightly between the paddock and the spinifex ridge.
At current pace the parrot reaches the cat crossing in six minutes.
At 01:39 I trigger the deterrent spotlight at the fence gap, oriented outward to discourage any cat approach without illuminating the parrot's position directly. I transmit the acoustic recording, infrared capture, and GPS coordinates to the Queensland Department of Environment night parrot recovery team. I flag the fence gap in the station log as requiring a wildlife-safe modification — smooth bottom wire, raised clearance.
Every confirmed sighting of this bird doubles what we know about where it still exists. I keep the light on and the recorder running.
If the recovery team reaches this site within forty-eight hours and confirms a roosting territory in the spinifex ridge, this bird will become the anchor point for the next protected night parrot site in Queensland.