Fisheries management models typically assume that fish populations fluctuate around a stable equilibrium — disturbed by fishing, weather, predation, but tending to return to a baseline. The models use linear dynamics: proportional responses, predictable oscillations, recoverable deviations. Quotas are set based on the assumption that reducing fishing pressure by a known amount produces a proportional recovery in population.
Published in Nature Ecology and Evolution, Hechler and Krkosek analyzed 243 recruitment time series and 266 spawner time series across 143 marine fish species worldwide. Eighty-one percent of populations showed nonlinear dynamics. Temperature variation amplified the nonlinearity: warmer, more variable conditions produced larger, more unpredictable fluctuations. Fast-lived species were more nonlinear than slow-lived species.
The structural insight is about the mismatch between the model and the system. Linear models produce proportional predictions: reduce catch by 20%, expect 20% recovery. Nonlinear systems do not respond proportionally. A small change in fishing pressure can produce a large population response, or no response, or a response in the opposite direction, depending on where the system is in its dynamics. The same management intervention applied to the same species at different times can have qualitatively different outcomes.
The 81% figure is the important number. This is not a phenomenon that appears occasionally in unusual species. It is the dominant mode of marine fish population dynamics worldwide. The linear models are not wrong in some cases and right in others — they are the exception applied as if they were the rule. The equilibrium assumption is not a simplification. It is a mischaracterization of the system being managed.