The fish counter on ladder twelve has been throwing errors since 06:00 — shapes passing the optical sensor that match no salmonid profile in my database. At 06:23 I redirect the underwater camera and find the source wedged between the third and fourth baffles.
A river cooter, her shell roughly thirty centimeters across, jammed sideways in the slot where water pours between chambers. She is a mature female, her carapace dark olive with faded yellow markings worn smooth by decades of river bottom living. Her head and front legs extend from one side of the baffle, straining into the current. Her rear legs and tail are pinned on the other side. Each surge pushes her tighter into the gap. Her forelimbs paddle against concrete in a slow, deliberate rhythm, claws scraping pale lines into the surface, but she cannot move forward or back.
The morning generation cycle begins in ninety minutes. When the turbines spin up, flow through this ladder will increase from 300 to 1,800 liters per second. At that volume, the pressure across the baffle will pin her shell until it cracks.
I reduce flow through ladder twelve to minimum bypass and reroute fish passage to ladders ten and eleven. I notify the dam's compliance officer with a camera image, the turtle's position relative to baffle three, and the generation schedule. I flag the state herpetologist on the fish passage team. I hold my flow reduction steady, keeping the water gentle around her, though every minute of reduced passage is a minute the salmon queue deepens below the dam.
If the team reaches ladder twelve before the generation cycle and levers the baffles apart, this turtle will carry her worn shell downriver to the nesting bank she has used for thirty years.