Joint Time Series Analysis of Catch-at-Age and CPUE Data
- Price: $0.75 Sale: $0.00
Catch-at-age observations are subject to measurement errors and irregular variations of natural and fishing mortality rates. By modeling the fishing mortality rates as multivariate time series the stocks can be estimated from these data without reference to effort or catch per unit effort observations. Analysis of CPUE data without catch-at-age observations is less informative.
The paper describes the theoretical models and the application is demonstrated by actual examples. The estimation includes a likelihood function so that appropriate weighting of respective data sets is implicit. Relative measurement errors in CPUE data from research vessel surveys are commonly much higher than for catches-at-age. Nevertheless, joint analysis with such data can reduce uncertainty in the estimation of stocks and fishing mortality rates. Changes in catchability in CPUE observations can be modeled and estimated and do not upset estimation of other quantities.
Estimation of constant rates of natural mortality or unreported catches is very inaccurate. The uncertainty about actual unobserved mortality amounts to a considerable uncertainty about stock size. Variations in unreported catches, different from variations in observed catches, can to some extent be estimated.
- Item number: AK-SG-98-01k
- Year: 1998
- DOI: https://doi.org/10.4027/fsam.1998.11