|Diagnostics for multiyear tagging models with application to Atlantic striped bass (Morone saxatilis)|
Latour, R.J.; Hoenig, J.M.; Olney, J.E.; Pollock, K.H. (2001). Diagnostics for multiyear tagging models with application to Atlantic striped bass (Morone saxatilis). Can. J. Fish. Aquat. Sci. 58(9): 1716-1726
In: Canadian Journal of Fisheries and Aquatic Sciences = Journal canadien des sciences halieutiques et aquatiques. National Research Council Canada: Ottawa. ISSN 0706-652X, more
Rates; Rates; Rates; Recovery; Selection; Selection; Survival; Animalia [WoRMS]; Marine
|Authors|| || Top |
- Latour, R.J.
- Hoenig, J.M.
- Olney, J.E.
- Pollock, K.H.
Information on age- and year-specific survival can be obtained from multiyear tagging data using one of three classes of tag recovery models. Two of the model types yield information on total survival, while the third allows separation of total mortality into its fishing and "natural" components if information on the tag reporting rate is available. The performance of each class is usually assessed using goodness-of-fit tests, Akaike's information criterion, and similar measures. However, we propose that examination of model residuals is also important for the evaluation of model performance and contend that at least four types of problems are potentially detectable via patterns in residuals. Those presented in this paper include nonmixing of newly tagged animals, emigration of older animals, cohort effects associated with initial tag-induced mortality or tag shedding, and a change in the natural mortality rate. We present the diagnostic procedures by analyzing a hypothetical tagging data set and discuss the various constraints inherent to the residuals of each class of models. The diagnostic procedures are also used to evaluate striped bass tagging (Morone saxatilis) data from the Hudson River and Chesapeake Bay.