Do Not Believe Your Model Results
Early in my career as a fisheries scientist, I recognized that mathematical modeling was a powerful way to study how fish populations behave in response to fishing because fish populations are too remote to directly observe. I took as many population dynamics, stock assessment, and statistical modeling classes as I could, and I became a zealot of stock assessment modeling. It was an exciting time to be learning about population modeling. The field was transitioning from a descriptive exercise to a more analytical challenge. The 1980s and 1990s have been termed the “Golden Age” of fishery stock assessment because statistical modeling was rapidly advancing in the wake of the technological revolution (Quinn 2003).
I was working in New England, contributing to the Northeast Regional Stock Assessment Workshops, and results from our stock assessment models told us that many fish stocks were depleted from overfishing, but we were frustrated that fishery managers were ignoring our advice to substantially cut back on fishing. I developed an extremely technocratic view of how fisheries should be managed. I believed that fisheries would be much better off if the fishery managers would simply follow the recommendations from scientists. I was wrong to think it was so simple. Of course, scientific advice should be considered in resource management decisions, but I believed that our stock assessment models accurately represented the complexity of the ecosystem, and I did not consider the people, their livelihoods, or the policies that attempt to represent a broad set of fishery objectives.