I think what's broken is the very concept that salmon runs can be forecast with sufficient accuracy as to be the basis for structuring fisheries. The results are in, and indeed, the current model is performing about as should be expected (which means poorly). One year, we come in well over (which I like, because that usually means lots of escapement goals are being met or exceeded). The next year, if the factors in the model suggest a similar surplus, we all too often see the fish fail to materialize, leading us to overfishing already depressed runs.

It made a lot of logical sense that this year's coho run should be expected to be poor to nearly absent. According to everything we thought we knew about the ocean and what makes conditions poor or ideal for salmon, combined with what we actually witnessed last year, that outcome seemed inevitable.

We still don't know what the final numbers will look like, but it seems safe to say the coho are reminding us just how little we understand about the ocean, as they have apparently weathered the Blob much better than expected.

Rather than try to fix the model (which I agree is clearly broken), I'd rather they change the whole paradigm to one that prescribes terminal fisheries over mixed stock fisheries, to the greatest extent possible, with no fisheries authorized over a given stock before required escapements are met or imminent. Reported data over pie in the sky. People in the know will tell you that's a non-starter, since nobody wants to give up their ocean fisheries, so there you have the reason why we will continue to use forecasts that have little hope of being accurate to model our fisheries.

In other words, what Carcassman said (while I was busy typing my novel).