It's actually a pretty reliable estimate compared to some other open water fish populations. We run the same hydroacoustic transects each year, give or take one or two depending on reservoir levels. We run a total of 45 transects starting at the Confluence and end up at the Dam two nights later. All sampling is conducted under the night(s) of the new moon in August to reduce sampling variability due to patchy distribution. We also only estimate kokanee under 14-inches, being the adults are difficult to catch in trawls, which are used to "truth" the counts generated by hydroacoustics.
The major drop in 2007-08 was pretty scary and it really broke the cycle you observe. We were hopeful that the kokes would rebound and were rewarded in 2009. There's a lot of things we know are impacting kokanee in the reservoir like predation (lake trout & some burbot) and the spawning runs have been really reduced over the last several years. We have also recently found whirling disease in reservoir kokanee. I think poor catch and release tactics could be having a toll, being so many smaller fish are released in hopes of harvesting bigger fish, but that mortality hasn't been quantified to date.
It's a pretty dynamic system, much like any other kokanee water. We'll continue to promote the harvest of small lake trout and burbot to reduce mortality, stress good C&R, and acquire eggs and stock as many kokes as possible to keep their numbers up for the future.
Hope that helps.