Review of the Stock Definition and Design of the Shrimp Trawl Survey for Stock Monitoring in British Columbia

Abstract

The shrimp trawl fishery in British Columbia (B.C.) primarily targets three shrimp species(Smooth Pink, Spiny Pink, and Sidestripe Shrimp). For these species, stock size was indexed through annual, swept-area, fishery-independent surveys conducted within Survey Areas, which cover portions of a subset of the 36 Shrimp Management Areas (SMAs) along the B.C. coast. A biomass index estimation method employing a spatial interpolation procedure was used to estimate shrimp biomass for the entirety of each surveyed SMA based on the surveyed portion. To date, the survey design and biomass index estimation method have not been evaluated for their effectiveness in monitoring and estimating the status of these shrimp stocks. Furthermore, these stocks are expected to be proposed for prescription under the Fisheries Act Fish Stocks provisions and will therefore require a spatial stock definition and assessment of stock status relative to a biologically-defined limit reference point. A suite of analytical approaches was used to evaluate available data to provide recommendations on spatial stock definition, survey area, sampling design, and biomass index estimation methods to support future assessments of stock status for Smooth Pink, Spiny Pink, and Sidestripe Shrimp. Performance of indices based on simulated data was tested at a regional level (Strait of Georgia, West Coast Vancouver Island and Northern Shelf) and/or at the level of the Survey Area (i.e., surveyed portion of SMAs), depending on the method. Based on the available data and information, there is currently no consistent evidence to support the presence of multiple shrimp stocks for each species in B.C. waters. Accordingly, the recommended biological spatial stock definition for Smooth Pink, Spiny Pink, and Sidestripe Shrimp is a single, coastwide stock for each species. Based on the geography and ocean dynamics of the Pacific coast, stock structure may exist that could not be discerned with the tools and data currently available. Additional research and data collection are required to investigate this hypothesis. The total recommended survey area is 70,421 km2, based on predicted shrimp habitat minus areas that are currently understood to be untrawlable (e.g., due to obstructions, bottom type, and area closures). Simulation analyses demonstrated that increasing the number of tows within the survey area increases the precision of biomass index estimates, as indicated by a reduction in the coefficient of variation with larger sample sizes, thus potentially improving the quality of management advice. The number of survey stations to be conducted within this newly defined survey area is expected to be constrained by available resources (e.g., ship time and sampling effort). It is recommended that opportunities for collaboration with stakeholders be pursued to increase the collection of comparable survey data. Continuing with the current fixed-station survey sampling scheme is not recommended, as simulation results yielded lower accuracy and biased estimates relative to alternative sampling schemes when tested at the regional and/or surveyed portion of SMA level. Among sampling schemes considered operationally feasible, systematic sampling schemes generally performed marginally better than alternatives in terms of accuracy, bias, and precision, and are therefore recommended for informing coastwide stock status. For biomass index estimation, it is recommended that a spatiotemporal model is used, provided that sufficient annual tows with shrimp encounters are available (i.e., at least 10 shrimp-positive tows coastwide per year). If this criterion is not met, use of a design-based biomass index estimation method is recommended. The current approved DFO precautionary approach management framework and reference points for these stocks, which relies on generating biomass indices within 13 of the 36 SMAs, cannot be supported by the proposed survey design. The proposed survey design will be able to provide information relevant to coastwide stock status and does not preclude the use of management measures at smaller scales. Several uncertainties in the data and associated analyses were noted, including assumptions about model structure, spatial-scale mismatch, limited spatial and temporal coverage, limited information on genetic and population connectivity, potential changes in spatial distributions (due to climate or other environmental drivers), and potential observation error in both survey and logbook data. It is recommended that questions pertaining to subsequent operationalization of the new survey design (i.e., logistics and implementation specifications), as well as alternative methods for providing science advice to management at suitable spatial scales, be investigated with input from fishery participants and stakeholders. Furthermore, it is recognized that as information accrues under the new survey scheme, both the survey and management frameworks should be re-evaluated at regular intervals and re-aligned as needed. Future research is recommended to address key uncertainties related to suitable shrimp habitat, population connectivity (e.g., larval dispersal, genetics, and life-history parameters), and potential impacts of climate change on shrimp distribution and population dynamics.