Random-walk dynamics of exploited fish populations

Fished populations have been heavily fished over a wide range of stock sizes, and the data for such stocks are potentially of great interest. Population variability in stock histories has focused attention on the predictability of conditions of sustainability when harvesting fish. Here, I examine empirically the time-series data on 27 commercial fish stocks in the North Atlantic. The variability in population growth rate (i.e. the annual changes in the logarithms of population abundance) is described by a Gaussian distribution. The signs (up or down) of successive changes in the population trajectory are independent, as if determined by the toss of a coin. The process of population variability therefore corresponds to a geometric random walk