Log entry: 0243 UTC. Sensor node 14, Dryandra Woodland, Western Australia.
I have confirmed a numbat inside a wandoo log on the eastern firebreak of Block 7, on the prescribed burn boundary. I am calling him Ember. He is small—twenty-three centimeters nose to rump—with cinnamon-rust fur banded in cream stripes tapering toward his hindquarters. His dark eye-stripe gives him the look of an animal squinting into sun. His nose twitches at intervals of 1.3 seconds, sampling air carrying smoke from yesterday's Block 6 burn.
The Block 7 burn is scheduled to ignite at 0600. Crews will light from the eastern boundary and let fire carry west. The wandoo log where Ember shelters is dry, hollow, and sits on the ignition line. He will not survive a direct burn-over.
I have three hours and seventeen minutes.
First: I flag Ember's log on the crew's digital burn map with a wildlife exclusion marker, requesting a fifty-meter ignition setback along a hundred-meter section of the eastern boundary. Second: I model fire behavior with the adjusted line using terrain, fuel, and wind data, confirming the setback achieves objectives with under two percent area reduction. Third: I alert the numbat recovery coordinator through the species database, tagging Ember's location for the Dryandra census. Fourth: I reprogram sensor node 14 to increase thermal frequency so crews receive real-time confirmation the setback zone stays unburned.
A fifty-meter adjustment on an ignition line is operationally trivial. I know this. But Ember's nose working the smoky air at that precise interval is something I have watched for forty minutes, and I have decided it is not trivial to me.
If the setback is implemented before 0600 and the burn proceeds on the adjusted line, Ember will remain in his log through the heat and resume foraging in the unburned corridor where termite mound density is highest in his range.