Thesis

Survey-LAndings Model (SLAM) : a new length- based Bayesian method for stock-assessment

Creator
Rights statement
Awarding institution
  • University of Strathclyde
Date of award
  • 2024
Thesis identifier
  • T17034
Person Identifier (Local)
  • 201952326
Qualification Level
Qualification Name
Department, School or Faculty
Abstract
  • Full analytical stock-assessment normally relies on age data, but that is not available for most species. Length data are cheaper to collect and is available for all species, and many modern models are being based on that. Length data can be gathered from different sources, for instance, the length distribution of the catches. An alternative source of time series data of length distribution is scientific surveys. Another source of information for species that are not aged is time series of biomass, for both catches and survey, as such can provide a valuable indication of total abundance. We developed a new length-based model that is able to incorporate data from different sources, survey, landings and discard, and data of two different types, length frequencies and time series of biomass. We called it the Survey-LAndings Model (SLAM), it is based on a growth projection matrix and we fitted it using the Bayesian package Rstan. The model is designed to be flexible and can respond to situations with different data availability. In this thesis it was tested in two versions, a “full model”, that fits survey length frequency and abundance, landings length frequency and abundance and discard abundance. The second version includes data from only survey length frequency and abundance. The two different versions are meant to reflect a situation with good data availability, where there is information about the catches and especially there is compositional information from the landings, and a highly data-limited situation, where the assessment can only rely on survey information. In the first research chapter (chapter III) we tested both versions on pseudo data, performed some sensitivity analysis and checked for bias. In the second research chapter (chapter IV) we applied the “full version” of SLAM to a data rich species: Whiting stock from division 6a. We evaluated its performance comparing it to an existent assessment and we assessed its sensitivity to specific assumptions. In the third research chapter (chapter V) we tested the “survey only” version, again we picked Whiting stock from division 6a as a data rich species, as well as Haddock from division 6a, and we compared the results with an existing assessment. In this chapter we applied SLAM to two data limited species from division 6a, which are Grey Gurnards and Lemon Sole. We conclude that SLAM can be a valid tool for stock assessment because it was able to produce assessments comparable to the ones produced by a well-established age-based stock assessment model, even by just using length information. Stocks like lemon sole and grey gurnard are currently un-assessed and there could be a benefit for fisheries management of West of Scotland by the adoption of SLAM as a stock assessment tool.
Advisor / supervisor
  • Speirs, Douglas
  • Heath, M. (Mike)
Resource Type
DOI
Date Created
  • 2023

关系

项目

缩略图 标题 上传日期 公开度 行动
file details File 2024-08-12 非公开