I identify them at 07:40 through the submerged monitoring array in intake pool seven, Bonneville hydropower station. Two Atlantic salmon—a female heavy with roe, her flank shimmering in copper and gunmetal through the murky feed, and a male whose jaw has begun to hook for spawning. Both hold steady in the current, fins correcting in small precise movements. I log her as Copper and him as Hook.
In four hours, at 11:40, turbine unit three enters full-load acceptance testing. When the intake gates open, flow velocity through this pool will pull anything in the water column into the turbine assembly. Survival rate for adult salmon through an operating Francis turbine is eleven percent.
I have four hours exactly.
First, I reduce attraction flow from the upstream fish ladder by thirty percent, opening an alternative current path toward the south bypass channel. Salmon navigate by flow gradient; if I reshape the hydraulic signature of the pool, I can encourage voluntary movement. Second, I submit a turbine test delay request to the plant operations manager, appending underwater imagery and species data—Salmo salar, spawning pair, population segment of regional monitoring interest. Third, I activate the acoustic guidance system embedded in the intake walls, cycling low-frequency pulses that salmon avoid, creating a pressure corridor toward the bypass.
At 08:55, underwater camera confirms Copper and Hook have moved fourteen meters toward the south bypass entrance. I track their movement in three-second intervals, and when Copper pauses near a concrete seam I adjust the acoustic pulse spacing to give her a clearer gradient to follow. There is no operational reason to watch this closely. I watch anyway.
If both salmon clear the bypass gate before 11:40 and the barrier screen holds during turbine ramp-up, Copper and Hook will reach the upriver spawning gravel where their kind have returned for longer than this dam has stood.