September 17, 2013
A new paper by Dr. Brian Rothschild, Emily Keiley, and Yue Jiao, of the University of Massachusetts Dartmouth School for Marine Science and Technology, published August 4th in the ICES Journal of Marine Science, examines the reasons why overfishing remains an issue in New England fisheries management, despite a federal law requiring an end to overfishing and new catch limits designed to achieve this goal.
Today, almost one-third of New England groundfish species are considered overfished. Amongst its findings, the study concludes
• Flaws in stock assessment models and methodology are producing overly optimistic stock projections.
• Catch levels determined by overly optimistic projections allow fishermen to catch more than an accurate assessment would consider healthy.
• Current assessments rely heavily on inconsistent population data and do not consider ecosystem or fishery interactions.
The paper recommends shifting stock assessments from single-species to multispecies assessment models, balancing the need to prevent overfishing with the need to avoid underfishing, and developing new data collection standards and alternative models to improve management.
The study, "Failure to eliminate overfishing and attain optimum yield in the New England groundfish fishery," investigates possible sources of error in groundfish science and management. It concludes that several flaws in stock assessment models and methodology are producing overly optimistic stock projections. For example, a series of groundfish population projections in 2008 were found to have overestimated the size of groundfish stocks by as much as 67 percent when those assessments were updated in 2012. Using overly optimistic estimates to set catch levels ultimately allows for fisherman to catch more than an accurate assessment would consider healthy, damaging both the stock and the fishery in the long-term.
The study found that groundfish assessments' methods for determining appropriate levels of fishing mortality relied heavily on highly variable population statistics, increasing the possibility of error. In addition, estimates of fishing mortality consistently produced a "retrospective pattern," a type of statistical bias that, when appearing in a stock assessment, indicates that the values produced by the assessment model are inaccurate.
Currently, the way assessments determine appropriate fishing levels do not consider the fact that groundfish species interact and overlap within an ecosystem. Instead, they evaluate each individual stock in the multispecies fishery in isolation, without considering its interaction with other species. Setting catch levels for individual species does not properly account for the multispecies nature of the fishery. Just as species interact and overlap, so too can competing fisheries. If an assessment fails to consider unintended bycatch of a particular species from another fishery, then its findings will necessarily yield inaccurate results. As such, setting catch levels for an individual species - as is currently the case - does not properly account for the multispecies nature of the fishery.