The Impact of Recruitment Projection Methods on Forecasts of Rebuilding Rates for Overfished Marine Resources

The Impact of Recruitment Projection Methods on Forecasts of Rebuilding Rates for Overfished Marine Resources

André E. Punt and Richard D. Methot

The Impact of Recruitment Projection Methods on Forecasts of Rebuilding Rates for Overfished Marine ResourcesThis is part of Fisheries Assessment and Management in Data-Limited Situations
PDF    
To download the free PDF [533.8 KB], please enter:
-or-

Description

Under the U.S. Sustainable Fisheries Act, rebuilding plans have to be developed for fish stocks that are determined to be overfished, i.e., are found to be below the minimum stock size threshold. Rebuilding plans typically include analyses to determine the minimum time, TMIN, to recover to a BMSY proxy and the target level of fishing mortality, FREC, that is consistent with recovery to this proxy within a pre-specified time frame and with an agreed probability. Key factors that determine TMIN and FREC are the methods used to forecast future recruitment and to estimate the BMSY proxy, here based on 40% of the size of the unfi shed reproductive output of the population. Several approaches to modeling future recruitment are available. For example, Monte Carlo draws from previous recruitments, from previous recruits-per-reproductive output, and from a parametric probability distribution around an estimated stock-recruitment relationship have been used for groundfish resources off the U.S. West Coast. The results of rebuilding analyses are sensitive to this choice of modeling approach. The performance of alternative approaches to modeling future recruitment, in terms of providing unbiased and precise estimates of TMIN and FREC, are explored by means of simulation. All of the methods examined are imprecise and most are biased. The major factors influencing the ability to make good predictions of TMIN and FREC are structural, viz., the actual value of the steepness of the stock-recruitment relationship and the extent of variability about that relationship. The length and contrast of the stock-recruitment data set has a larger impact on estimation performance than its precision.

Item details