Unbiased and precise estimates of commercial catch data such as discard volumes, length or age composition are essential inputs to many stock assessment methods. This training course is an applied statistical methods course, concerned almost exclusively with the estimation of commercial fishery data used in ICES assessments but is also relevant to member states needs for data reporting.
The course aims to provide national staff with the level of expertise required to improve data collection and provision at national and international levels in a way that meets ICES demands for both quantity and quality of catch information, while exponentiating progress towards statistically sound sampling of ICES stocks.
The course will examine common problems experienced by national scientists when designing and estimating commercial catch data for assessment. After a brief review of common sampling strategies and estimators used to characterize commercial catches, the course will focus extensively on more complex sampling designs (e.g., stratified multi-stage cluster designs with equal and unequal probability) and methods for comparing those designs in order to optimize the sampling effort.
Methodologies used to correctly calculate inclusion probabilities, handle missing observations (missingness), for post sampling data usages (e.g., domain estimation), for extracting information from older datasets that have been collected in a statistically rigorous manner or for data-limited stocks, and non-parametric methods for obtaining confidence intervals for estimators will be covered. The latter includes modern approaches such as bootstrapping or Monte Carlo simulation and can also be used to assess bias in the estimators such as that due to incorrect specification of inclusion probabilities or for the older datasets obtained without probabilistic sampling.
Alternative estimation approaches, such as model-assisted and model-based inference, will be reviewed and compared.
Go to online registration.
Instructors:Mary C. Christman, MCC Statistical ConsultingJon Helge Vølstad, Institute of Marine Science, Norway
Course fee:750 Euros for ICES member country affiliated participants1250 Euros for non-member country affiliated participants