Propagation of Process Error Through the Linkage of Forecasts to Assessments

Assessment of marine fisheries in the U.S. strives to quantify the probability that a specified future catch limit will prevent overfishing.  This probabilistic approach has been most recently elucidated in the work by Prager and Shertzer (NAJFM, 2010).  While the population and fishery forecast model to apply such methods can be done subsequent to completion of the assessment of the historical data, Maunder et al (2006) advocates linkage of the forecast with the assessment to better propagate uncertainty in the assessment into the forecast.  The Stock Synthesis (SS) assessment program (www.nft.nefsc.noaa.gov) has adopted this approach of incorporating the forecast into the assessment.  Essentially, there is a continuous population process and the demarcation between the assessment and the forecast eras is simply a transition from estimation of fishing mortality rates that match observed catch, to estimation of catch that will match estimated target fishing mortality rates.  This presentation will report on recent advances in this forecast implementation in SS.  The forecast now proceeds in a two stage process.  In the first stage, the forecast calculates the expected catch level according to a target harvest policy and expected population dynamics.  In the second stage, the forecast recalculates the fishing mortality rate needed to attain that forecast level of target catch after first applying implementation error to that intended level of catch and while taking into account the degree of expected fluctuation in important population processes, particularly annual recruitment.  Model outputs include the probability, for each year of the forecast, that the fishing mortality rate will exceed some limit or that the population abundance will fall below a specified level.  The feasibility of including other fluctuating processes (e.g. growth and selectivity) in the forecast will be discussed.