A 15-day thermal window now defines survival for California's most endangered salmon

Each nest starts its own clock. Winter-run Chinook spawn across the summer season ^[5], and the window the study identifies is counted from the fertilization of a given redd ^[1], so what faces a reservoir operator is not one cold fortnight but a rolling series of them, one per spawning date ^[15]. That is what makes the timing framing attractive: if only 15 days per nest decide the outcome, cold water released outside those days looks, on paper, like waste. In warm years with less water, the researchers report, the cold water barely covered the window at all ^[2].

Rachel Johnson of NOAA's Southwest Fisheries Science Center and UC Davis, who helped lead the work, went out of her way to close that door. She cautioned against aiming limited cold water at the 15-day window for only part of the spawning season, which would deliberately favor a reduced number of redds, and said that selecting the winners at this stage narrows the chance that salmon encounter favorable conditions later in life ^[10]. The objection is not hydrological. It is that a bottleneck this tight already reduces population size and erodes the diversity that lets the run absorb bad years ^[14], and hand-picking which nests get protection does the same thing on purpose.

The durable part is the measurement. Oxygen isotopes in otoliths record the temperatures a fish was exposed to, read here by an ion microprobe at a UCLA laboratory ^[12]. Combined with hatch dates from otolith growth increments, temperature models and spawning surveys, that let the team name the redds that actually produced surviving juveniles ^[13]. Fisheries biologists have long judged the year by how many adults came back to spawn, assuming many succeeded ^[8]. Kohma Arai, who authored the paper as a UC Davis postdoctoral scholar, says only a small fraction of spawning events produced the juveniles that survived ^[5]. Johnson puts it more bluntly: in some years the number of successful spawners can be shockingly small, and that is the number needed to assess extinction risk ^[9]. For a run averaging a few thousand returning adults with one surviving population ^[3], the gap between the two counts is the whole question ^[16].

What the write-up does not supply is a threshold, a fraction, or the years involved, so nobody can yet turn this into a release rule for Shasta, where cooling flows compete with deliveries to farms and cities and the balance tightens in dry years ^[6]. The authors' own recommendation points away from finer reservoir control and toward reintroduction to the historical habitat above the dam ^[11], blocked since Shasta Dam cut the run off from the cold streams draining Mount Shasta ^[4]. That is a permitting argument, not an operations one.