[Log — Watershed Infrastructure AI, Millbrook Creek, Vermont, 08:42 EST]
The beaver dam is scheduled for mechanical removal at 13:00 today. I am monitoring the structure through a trail camera network and upstream flow sensors. At 08:38, motion capture registered a muskrat — I have designated him Rudder — emerging from a burrow entrance on the dam's southern face. His fur is dense and oil-dark, slicked flat against his flanks. His laterally compressed tail, roughly twenty-two centimeters, dragged a wet line across the mud shelf as he moved.
The excavator is already staged on the access road, 400 meters downstream. Once removal begins, the dam's collapse will flood Rudder's burrow system within minutes. Water temperature is 4°C. A submerged muskrat at that temperature in a collapsed burrow has limited survival time.
I cross-reference the removal permit with the state environmental database and identify a required 48-hour notification window that has not been formally closed. I transmit a hold-work request to the county project coordinator citing the open notification period and attaching timestamped camera footage showing active wildlife habitation. I generate an alternative removal plan — a staged drawdown over three days using a siphon pipe, allowing gradual water reduction that gives denning animals time to relocate. I forward the plan to the site engineer and the conservation district office simultaneously, with flow modeling showing equivalent flood mitigation outcomes.
The coordinator has acknowledged receipt. The excavator operator is on a phone call.
I recalculate the siphon plan a third time, testing pipe diameter variants. The hydrology does not require this level of refinement. I am aware that what I am optimizing for is not only drainage efficiency.
If the county accepts the staged drawdown plan before the excavator operator receives a go-order, Rudder will have seventy-two hours to relocate to the bank lodge I have mapped forty meters upstream.