AN EXTENSION OF STOCK COMPOSITION ANALYSIS TO INCLUDE MARKING DATA

Population characteristics and man-made marks have been applied separately to gather information on the stock composition of mixed-stock fisheries. This paper develops an extension of the finite mixture problem with learning samples model of Fournier et al. (1984. Can. J. Fish. Aquat. Sci. 41: 400-408) to include marking data as well as morphometric, meristic, or biochemical genetic data for stock composition analyses. Simulation studies are used to test the model on a previously described ''problem cluster'' that consists of stocks that exhibit high levels of genetic affinity. The test results show that the inclusion of marking data can improve the accuracy of stock composition estimates.