Science Advisory Report  2016/015

A revised operating model for Sablefish (Anoplopoma fimbria) in British Columbia, Canada

Summary

• The British Columbia (BC) Sablefish (Anoplopoma fimbria) harvest strategy is designed around a management procedure that is simulation-tested against operating model scenarios that capture quantifiable uncertainties in Sablefish stock dynamics and fisheries. A revised operating model that accounts for structural mis-specifications and lack of fit to key observations recognized in the 2010 operating model (Cox et al. 2011; DFO 2011) was reviewed at a regional peer review meeting held January 20, 2016.
• The revised operating model is a multi-gear, age-structured, statistical catch-at-age model with modifications that include:
1. Implementation of a two-sex growth model to account for differences in growth, mortality, and maturation of male and female Sablefish,
2. Application of an ageing error matrix to model age proportions,
3. Estimating prior distributions for fishery selectivity from tagging data to allow time-varying selectivity for each fishery, and
4. Revisions to the way in which age composition data are modeled  to reduce sensitivity to small age proportions, particularly in the age-35+ group.
• The revised operating model was fitted to Sablefish landed catch data (1965-2015), three indices of total abundance and age-composition (1990-2015), and at-sea releases (1996-2015) in each of the commercial fisheries to evaluate model quality and fit to historical data, both of which are important for closed-loop harvest strategy simulations. The sensitivity of operating model biomass estimates to input data and prior assumptions was evaluated by running the model with different combinations of data weighting on abundance indices, male and female age composition data, and the inclusion of an ageing error matrix, resulting in seven data scenarios (D1-D7).
• Estimates of Sablefish stock status, productivity, and biomass trends over the past several years produced by the revised operating model are consistent with previous harvest strategy simulations. The revised operating model estimated harvest rates for sub-legal Sablefish (<55 cm fork length) that are much higher (7-8%) than the sub-legal harvest rates estimated by the 2010 operating model (1-2%).
• Although analysis of tagging data supports the hypothesis that size-selectivity in all commercial Sablefish fisheries has changed over time, models with constant selectivity assumptions were more stable than time-varying models, and tended to produce better fits to the historical data, especially at-sea releases.
• The new features in the revised operating model (two-sex structure, ageing error matrix) are appropriate and follow standard practice in fisheries science. Recruitment estimates with reduced auto-correlation, improved fits to the age-35+ group and the temporal pattern of at-sea releases in the trawl fishery from 2000 to the present are notable achievements. However, the lack of fit to some at-sea releases in the trawl fishery (1996 to 1999) and in the trap fishery (2006 and 2007) requires further investigation. Based on these findings, the revised operating model is an improvement relative to the 2010 operating model and is recommended for BC Sablefish harvest strategy simulations.
• Six of the seven data scenarios (D2-D7) represent plausible operating model scenarios for stock size and productivity and are recommended for future evaluations of Sablefish management procedures. Model D2, which is a two-sex model with an ageing error matrix applied to model ages, was adopted as the base operating model. It is recommended that the D3 scenario, which is identical to D2 but with a shorter estimated recruitment time series (1990-2015), be explored as an alternative base operating model because it may reduce poor model behaviour that occurs due to low information content in the early part of the historical period.
• The revised operating model does not account for spatial dynamics among Sablefish stocks on the west coast of North America. Interactions with stocks in United States (US) waters (Alaska, south of BC) through trans-boundary movements may have important effects on Sablefish stock dynamics in BC that are not currently captured by the revised operating models.
• It is recommended that future model development explore ways to more fully integrate tagging data into the sex- and age-structured operating models to address issues related to correlations among estimates of unfished biomass (B₀), the steepness of the stock-recruitment relationship (h), and length-based selectivity.
• Inconsistent sampling of age composition data in commercial catches (only Trap fishery age composition data are currently available) has contributed to model issues that have been consistently identified during BC Sablefish harvest strategy evaluations (Cox et al. 2011; DFO 2011, 2014). It is recommended that a properly designed commercial catch sampling program be developed and applied across all fishery sectors.
• Harvest strategy evaluation is an iterative process of change and improvement. It is recommended that BC Sablefish operating models be reevaluated for suitability at 5-year intervals.

This Science Advisory Report is from the January 20, 2016 regional peer review on A revised operating model for Sablefish (Anoplopoma fimbria) in British Columbia, Canada. Additional publications from this meeting will be posted on the Fisheries and Oceans Canada (DFO) Science Advisory Schedule as they become available.

Complete document (PDF, 276 KB)

Accessibility Notice

This document is available in PDF format. If the document is not accessible to you, please contact the Secretariat to obtain another appropriate format, such as regular print, large print, Braille or audio version.