Recruitment in randomly varying environments

Abstract

The dominant role of environment variability in the recruitment dynamics of exploited marine populations has long been recognized. However, studies of the structural features of recruitment variability and its implications for management are comparatively recent. The objective of the present study is to develop stochastic analogues of several stock-recruitment models in an attempt to characterize the expected form of recruitment fluctuations. Approximate expressions for the mean and variance of recruitment are first developed for several models including generalized Ricker and Beverton-Holt models. The mean recruitment is higher under stochastic variation in mortality rates than for the corresponding deterministic case. The approximate variance of recruitment increases with increasing population fecundity (initial cohort size), duration of the pre-recruit phase, and variability in mortality rates, but decreases as mortality rates increase. Extension of these results to multistage recruitment processes is also considered.

Conditional probability density functions of recruitment are next developed for models under the assumption of normally distributed density-independent mortality rates. The general conclusions of the approximate analysis are shown to hold under more specific assumptions regarding the probability distribution of mortality rates. The implication of autocorrelated random variability in mortality rates is also explored. These models provide a general context for interpretation of empirical recruitment distributions.