Moving Beyond Environmental-Recruitment Relationships For Ecosystem Based Management

Understanding and predicting the effect of climate driven changes in the environment on exploited marine populations is a necessary component of ecosystem based management. Yet, valid relationships between environmental and biological variables are difficult to find using standard statistical tools. We present multivariate state space reconstruction (SSR) as an alternative approach for using time series to explore the effects of environmental variability on populations. In the case of the Pacific sardine (S. sagax), harvest limits are currently adjusted to changes in the temperature (SST) at the Scripps Institution of Oceanography (SIO) pier. However, the temperature-recruitment relationship motivating this policy has failed to hold up to retesting with newer data, and this has thrown the current policy into question. We demonstrate how multivariate SSR can be used to detect the influence of environmental variables on population dynamics and then integrated with dynamic programming to create a management framework which takes environmental variables into account. The methods reveal that the SIO pier SST is in fact a useful environmental indicator for sardine management if the system is viewed nonlinearly.