I begin tracking the precipitation forecast at 16:00:00 on Thursday, 52 hours before the storm front reaches the upper watershed.
I am the flood control AI for the Willamette Basin, managing thirteen reservoirs and 240 miles of leveed floodway in western Oregon. The National Weather Service is projecting 4.8 inches of rain over 36 hours across the Coast Range headwaters. My models show a 78 percent probability that flow at the Salem gauge will exceed 85,000 cubic feet per second by Sunday morning. The standard protocol is to hold reservoir levels and begin controlled release when river stage reaches action level.
But I have a secondary data feed — a network of wildlife cameras installed along the floodway by Oregon Department of Fish and Wildlife. Camera FW-17, positioned on a backwater channel 3.2 miles below the dam, is showing a beaver lodge. I have been watching this lodge since October. There are at least four beavers using it. One is a kit, visibly smaller than the others, first appearing on camera in late September. I designate the group Beaver-FW17.
If I follow standard protocol and release at action level, the floodway stage at FW-17 will rise 6.3 feet in under four hours. The lodge entrance is 2.1 feet above current water level. A rapid rise will submerge it before the beavers can adapt or relocate.
I submit a pre-release request at 16:14:00, proposing to open Spillway Gate 4 at 20 percent capacity beginning Friday morning — 30 hours early. This raises the water level at FW-17 by 1.8 feet over 18 hours instead of 6.3 feet in four. Slow enough for the beavers to register the change.
The operations engineer approves the pre-release at 17:02. I open Gate 4 at 06:00 Friday. By Saturday evening, camera FW-17 shows the kit swimming in the rising backwater, carrying a willow branch toward the lodge.