I don't know how they decide where to spawn.
I wonder when there is a stream upstream of the lake they spawn in the stream and when there is not, they spawn in the lake.
Each "run" of sockeye has a unique genetic code specific to its watershed that allows it to flourish.
For example it is generally true that sockeye spawn in a river and the newly hatched fry migrate downstream to rear in a nursery lake for a year before migrating to marine waters. However, take, for example, Weaver Creek sockeye. The newly hatched fry migrate downstream through Weaver Creek and Morris Creek to the Harrison River, hang a left, and migrate UPSTREAM to rear in Harrison Lake. Or the sockeye that spawn in the Harrison River Rapids. The fry of that spawn migrate the Fraser River estuary shortly after hatching and actually rear in the estuary, not a lake. I'm sure there are many other examples that don't fit the norm. For example, lake spawning sockeye in Cultus that Ralph referrred to. (Interestingly, I also read that Harrison Rapids sockeye first migrate up to Harrison Lake and spend some in the depths of the lower lake before dropping back to the river to spawn)
In my dog eared copy of "The Run of the River" author Mark Hume writes about how this unique genetic code makes it difficult for biologists to re-introduce sockeye into a system after it has been extirpated. His example is the sockeye of the Upper Adams River. The Upper Adams had a massive run of sockeye that was extinct by the early 1950's. From the mid 1950's to the 1980's, biologists transplanted millions upon millions of eggs from similar systems, with negligible adult returns. He wrote "In the Upper Adams, biologists are working to rebuild a system that was completely wiped out. All the sockeye stocks of the Upper Adams are extinct. The Fraser River slides and the logging flash-dam had destroyed not just a run of salmon, but a unique genetic code. The code to unlock the timing of Adams Lake had been erased from nature's memory bank"
He also writes about the sockeye of Karluk Lake in Alaska where biologists were trying to rebuild a run that was destroyed by overfishing, from 5M+ in the 1880's to 0.5M in the 1960s. After decades of rebuilding failure, they realized that the key wasn't putting more fish on the spawning grounds, but rather the key lay in Karluk Lake itself and the timing of fry emergence. They learned that Karluk stock was comprised of 20 separate groups of sockeye "The groups spawned at different times and the timing of the fry emergence was perfectly synchronized to the forage production in the lake" Overfishing had reduced these 20 groups to just 3 or 4 and "enhancement efforts had concentrated propagating those groups. The result was that the fry arrived in the lake in a few great lumps causing over competition for food at some times and leaving the lake lightly used at others. The fish were out of synch with the lake and the result was poor survival for fry"
In the 1980's the Upper Adams biologists switched to transplanting stocks from the nearby Momich-Cayenne system "hoping to find fish that were similar enough to re-establish an Upper Adams genetic code" Very modest returns resulted and a DFO biologist noted "If success is measured by the creation of even a small population, adapted to Adams Lake watershed environment and capable of providing seed for further expansion, then these transplants were successful. The seeming unsuccessful transplants attempts in early years may have been important in establishing a genetic base from which survival can be enhanced..."
On a side note, I quickly reviewed a 2003 paper on the Upper Adams, and to put it simply, the jury was still out. I couldn't find anything more recent.
So, I went WAY overboard in my reply (like Spawn Sack overboard
) but, to summarize, spawning is determined by a unique genetic code for a particular watershed.
At the very least, I enjoyed reading The Run of the River again