Évaluation du stock de bocaccios (Sebastes paucispinis) de la Colombie-Britannique en 2019 et lignes directrices relatives à l’élaboration de plans de rétablissement

Abstract

Bocaccio rockfish (Sebastes paucispinis, BOR) is ubiquitous along the British Columbia (BC) coast (at ~60-300 m depth), occurring in low densities along the west coast of Vancouver Island (WCVI), across Goose Island and Mitchell gullies in Queen Charlotte Sound and into the lower parts of Hecate Strait. In 2002, the Committee on the Status of Endangered Wildlife in Canada (COSEWIC) assessed the status of the BC population of Bocaccio rockfish (BOR) as “Threatened”. In November 2013, BOR was reassessed by COSEWIC as “Endangered”. Under COSEWIC and SARA (the Species at Risk Act), an endangered species is defined as one that is facing imminent extirpation or extinction. BOR has been assessed twice by DFO Science (in 2009 and then in 2012) using a Bayesian surplus production model which evaluated this species to be in the DFO Critical Zone. This led to setting mortality caps at the minimum acceptable levels of harm while still allowing the multispecies trawl fishery, which takes BOR in conjunction with other trawl species, to continue operating while still rebuilding. Beginning in 2017, there has been mounting evidence of a strong BOR recruitment event in BC waters, leading to difficulty in 2018 and 2019 for commercial operators to stay within the existing low catch limits that were set to encourage rebuilding in the BC BOR population. The purpose of this stock assessment is to evaluate the current BOR stock status and to assess the potential impact of this new recruitment on the DFO Precautionary Approach (PA) status of BOR in the future. This stock assessment evaluates a single BC coastwide population harvested by multiple fisheries. Analyses of biology and distribution did not support separate regional stocks for BOR. A single coastwide stock was also assumed for the two previous BOR stock assessments. We use an annual catch-at-age model tuned to six fishery-independent trawl survey series, a truncated bottom trawl CPUE series, annual estimates of commercial catch since 1935, and age composition data from survey series (31 years of data from four surveys) and the commercial fishery (12 years of data). The model starts from an assumed equilibrium state in 1935, and the survey data cover the period 1967 to 2019 (although not all years are represented). Two fisheries are modelled: one a combined bottom and midwater trawl fishery and an ‘other’ fishery, which combines halibut longline, sablefish trap, salmon troll, rockfish hook and line, etc. The second fishery is a compromise that acknowledges other methods capturing this species while keeping the complexity to a minimum, given the lack of good information from these additional fisheries. Three base runs using a two-sex model were implemented in a Bayesian framework (using the Markov Chain Monte Carlo procedure) under a scenario that fixed natural mortality to three levels (0.07, 0.08, 0.09) while estimating steepness of the stock-recruit function, catchability for the surveys and CPUE, and selectivity for four of the six surveys and the commercial trawl fleet. These three runs were combined into a composite base case which explored the major axis of parameter uncertainty in this stock assessment. Nine sensitivity analyses were performed to test the effect of alternative model assumptions. The composite base case suggests that the BOR spawning population lies in the Critical Zone (with a probability >0.99), as do the three component runs. This is in spite of the stock being moderately productive and exploitation rates being uniformly low. For instance, the median exploitation by the trawl fishery, which accounts for 95% of the catch, in the final year is estimated to be 0.025 (0.012-0.044) even at the very low biomass levels. A strong cohort, estimated at 44 times the long term average recruitment (range: 30-58), was born in 2016 and is projected to bring this stock out of the Critical Zone by the beginning of 2023 and will have a better than 50% probability of being in the Healthy Zone in that same year. These predictions are entirely dependent on the assessed size of the 2016 year class, which is highly uncertain. However, there is evidence, beginning in 2017, that this cohort is large and dominates the available data. Three of the synoptic surveys, particularly the Queen Charlotte Sound survey in 2019, show strong quantitative increases in abundance and in distribution. This cohort dominates the age and length frequencies in the commercial trawl, beginning in 2018. Similar strong recruitment (in 2010 and 2013) in the US BOR population, located south of Monterey, has lifted that stock out of an ‘overfished’ designation and was assessed in 2017 to be approaching 0.5B0. We suggest that the demonstrable capacity of the four active synoptic surveys plus the high quality monitoring of the trawl fishery catches and discards will verify the progress of this strong cohort as it recruits to the fishery.