Spatial-temporal fish stock assessment.
|Title:||Spatial-temporal fish stock assessment.|
|Abstract:||This thesis explores an alternative method of assessing the status of commercially exploited fish stocks. Aggregate numerical least-squares regression methods are most often used to estimate fin fish population numbers by age of fish within the stock. These methods, commonly called virtual population analysis (VPA) or cohort analysis, provide single point static estimates of stock size. In this thesis, the state-of-age aggregated stock components are tracked over the course of each season. Data about fish stock spatial-temporal migration dynamics and partial observations from catches by the commercial fishery are used to update stock status estimates over the course of the season. Stock status is described in a probability distribution over discrete stock classes. A Bayesian updating procedure is used to take account of the partially observable catch information. The model requires a probabilistic description of the underlying stock dynamics in the form of a Markovian probability transition matrix, and a reliability measure that explicitly accounts for errors in observation. The spatial-temporal assessment approach is applied to the Scotia-Fundy herring stock in NAFO divisions 4WX. The dynamics of the herring are described and the model is developed for a given season.|
|Collection||Thèses, 1910 - 2010 // Theses, 1910 - 2010|