You accurately restate the 2 main issues - as I mentioned above, California.Nor do I, but I'm pretty sure removing 15-20% of the biomass of forage fish right before they spawn is not a great thing for the ecosystem, particularly for those organisms that depend on the eggs and larvae to grow and survive. They do not remove them uniformly, they are hit hard in certain areas, and since herring are local populations they can be wiped out in one area but still be abundant in others. Howe sound populations are only now beginning to rebound after being wiped out decades ago.
The 15-20% removal comment gets into the debate about precautionary triggers on TAC as benchmarked by DFO using Limit Reference Points. The suitability of using Limit Reference Points is entirely related to the accuracy of the biomass estimates - which is under debate also for 2 reasons: 1/ methodology (field AND office), and 2/ assumptions (esp. Q and priors) when extrapolating those field data into a biomass estimate. Both have issues, also.
The comments around areas and impacts gets into the discussion around subpopulations (metapopulations) and management actions being detailed and specific and fine enough to conserve local populations verses broad-scale population dynamics. Again - much debate here wrt the ability of DFO to be able to discriminate and manage metapopulations.
I certainly wouldn't disagree w your comments - as there is an unfortunate history of DFO going through a learning curve wrt herring management.