Detect temporal variations of population growth rate: A time varying population growth model for Atlantic weakfish and an exploration of the possible driving factors on the population dynamic variation

Background/Question/Methods

Atlantic weakfish (Cynoscion regalis), is one of the important fisheries along Atlantic coast.  Its dynamics has been a problem to model and predict.  As for many other fisheries, stationary biomass dynamic models and catch-at-age models have been used in its assessment.  The weakfish population has been found to be very dynamic and a stationary density dependent population growth model fails to capture its dynamics.  Nonstationary population dynamics are rarely applied in population dynamics but have been discussed among research scientists.  No generally applicable models/methods have been developed to test the possible hypotheses of nonstationarity.  In this study, using the Atlantic weakfish (Cynoscion regalis) fishery as an example, we compared a nonstationary population dynamics model with the commonly used stationary biomass dynamic model to assess the population dynamics.  The available information on weakfish and these used in this study include catch data from 1920s till now, and survey relative abundance indices from 1970s and 1980s till now.  Both fishery-independent and dependent relative abundance indices were used in the assessment.  Because of the high dynamic characteristics of weakfish, process error is a vital characteristic to consider in this situation.  A Bayesian state-space surplus production model was developed to model the dynamics of the weakfish.  A random walk process was used to model the time-varying population growth which simulated the trend of the productivity over time.  The parameters and population abundance were estimated using a MCMC algorithm of Metropolis-Hasting within Gibbs, and performance of the models was compared by goodness-of-fit (deviance information criterion, DIC, here) of models. 

Results/Conclusions

The lower DIC value when nonstationary population dynamics model was used suggests that the weakfish population dynamics is probably nonstationary.  Our study found that weakfish population growth rate varies over time and the trend of the population growth rate strongly correlated with the Atlantic Multi-decadal Oscillation (AMO) other than North Atlantic Oscillation which implied that AMO is probably the drive factors that cause weakfish population to vary other than weakfish itself.