With each year of added information, stock assessments that use analytical population models result in revisions to the estimates of the fish stock and fishing mortality for previous years. Normally, these revisions would be expected to vary randomly. However, a number of ICES stock assessments produce estimates that are biased in that successive yearly assessments result in revisions of previous estimates of the population that are consistently lower or higher than the previous assessments.
This retrospective bias poses a particular issue in providing catch advice to clients because the forecasts will produce either optimistic or, in some cases, pessimistic estimates of the advised catch. If the advised catch is overly optimistic, this can result in excessive exploitation and ultimately bring the stock outside of safe biological limits. Pessimistic estimates, on the other hand, can result in a loss of yield that could be obtained from the stock.
Participants at this workshop will examine potential approaches for addressing retrospective bias in ICES stock assessments. The workshop will aim to document the extent of the retrospective bias in category 1 and 2 assessments based on data produced by assessment working groups during 2019. Additional objectives are to identify and compile possible causes for retrospective bias and to develop approaches for corrections and guidelines for acceptability of a stock assessment with retrospective bias. The outcome of the workshop will be considered for implementation in developing the 2020 advice on fishing opportunities.
The workshop will be held at the Northeast Fisheries Science Center in Woods Hole, MA, USA, 11–15 November 2019. For more information contact the Co-Chairs Larry Alade and Chris Legault.