For the third time, a judge tossed out the federal government’s plan to protect salmon on the Columbia River system. Apparently, the ruling was largely because the plan ignored the impacts of dams. The Seattle Times summarized the scale of the problem this way:

The Bush administration then issued a plan to spend $6 billion over 10 years to tinker with the operations of the dams, but would still allow up to 51 percent of the young spring chinook to die while migrating downstream from the Snake River, and up to 92 percent of the young fall chinook migrating down the river. And – in a dramatic shift from his predecessor – Bush said removing the old Snake River dams was out of the question, even if all else failed.

Remember, both spring and fall Snake River chinook are listed as "threatened" under the federal Endangered Species Act. Can allowing 92 percent mortality of fall Chinook really be a serious attempt at restoring the run?

The judicial answer being "no," the ruling was a victory for salmon advocates. But the legal battle will undoubtedly drag on for years longer.

One thing that’s fishy, however, is how often parties on both sides of the debate refer to recent salmon counts as "proof" that salmon populations are either surging or about the collapse. Salmon numbers naturally fluctuate wildly from year to year. It’s the long-term trends that are important; and, incidentally, it’s the long-term trends for salmon that are really alarming.

Here’s what I mean. On a year to year basis, the average change in Snake River Chinook is a whopping 82 percent. That is, in a typical year (and there aren’t many "typical years"), the number of returning salmon is 82 percent higher or lower than the previous year. The chart below shows returning spring and summer Snake River Chinook numbers from 1960 to 2003 (the blue line shows the real annual numbers, the yellow line shows my calculation of 5 year rolling averages). Data courtesy of