How Much Difference Can a Little Change in Productivity Make? Identifying Productivity Parameters That Matter to Assessments

For decades, we have attempted to understand fluctuations in marine fish population productivity, trying to find links to measurable environmental conditions or population characteristics that could improve predictions for future catch. These efforts have been largely unsuccessful, in part because of the myriad of factors influencing reproduction and productivity that will be discussed in this symposium. Mechanisms that affect larval production, survival and dispersal are complex and vary in space and time. But recruitment can be handled as a simple random variable, and with enough assumptions and age-structure information, the mechanisms that cause recruitment variation are unimportant to assessment results. More biological realism does not necessarily improve model fit and predictive power, particularly when measurement error is high and density-dependent survival negates most of the variance in recruitment, regardless. The real question, then, is under what conditions do variations in reproductive output of individuals translate into differences in assessment results? One of these conditions is cyclical variance in stock-recruit relationships, which may be common in marine environments that experience decadal shifts in productivity. I will illustrate this with examples from salmon and rockfish, species with different life histories that show a lack of stationarity in key life history parameters.