Gulf of Alaska | Fishes and Fish Habitats MASTER’S POSTER PRESENTATION Estimating abundance trends by integrating data from multiple fishery-independent surveys
Presenter: Tristan Sebens , tsebens2@alaska.edu Curry Cunningham , cjcunningham@alaska.edu
Fishery-independent Catch-Per-Unit-Effort (CPUE) data are a critical and cost-effective tool for stock assessment and sustainable fisheries management. However, these CPUE data have inherent limitations: Fishery-independent surveys are typically subject to government funding, the availability of which can fluctuate. This can result in spatial or temporal gaps in survey coverage as sample designs are adjusted to meet budgetary requirements. Additionally, choices in areas and depths sampled, or more importantly not sampled, can lead to non-representative subsamples of a stock. Size-specific gear-selectivity can produce both CPUE and age-composition data which is not fully representative of the stock. We may be able to compensate for these challenges through the intercalibration of observations from multiple surveys, wherein the relative catch-efficiency and selectivity of multiple surveys are directly estimated from catch rate observations that overlap in time and space. By intercalibrating data from multiple-fishery independent surveys employing different (i.e. trawl vs. fixed) gears, that operate at different depths and with differing spatial footprints, and with different interannual frequencies we hope to develop a more robust understanding of trends in abundance for Alaskan groundfish species. Using data from three Alaskan surveys (The NMFS Bottom Trawl Survey, the AFSC Sablefish Longline Survey, and the IPHC Setline Survey), our analysis is conducted across four species-region case-studies: Pacific cod in the Gulf of Alaska, Arrowtooth Flounder in the Gulf of Alaska, Rougheye Rockfish in the Aleutian Islands, and Greenland Turbot on the Eastern Bering Sea Slope. In this analysis, we compare the performance of three classes of statistical models for intercalibrating fishery-independent survey data: Timeseries random-walk models, generalized additive models (GAMs), and vector-autoregressive spatiotemporal (VAST) models. These models were fit to CPUE data collected by three fishery- independent surveys conducted in the marine waters surrounding Alaska. Model performance is compared based on their goodness-of-fit to the CPUE data, and comparison to official published abundance estimates from the NOAA- NMFS Bottom Trawl Surveys. Through this analysis, we hope to gain a better understanding of the potential benefits of combining data from multiple fishery-independent surveys when developing indices of abundance to inform stock assessment, as compared to standa
Alaska Marine Science Symposium 2023 201
Made with FlippingBook flipbook maker