The scale pattern of Chum Salmon is considered to contain ecological information representing the rate of growth, physiological state of fish, the river stocks and so on.
In the present paper, the authors propose an analytical method which uses Linear Discriminant Function associated with Spectrum Analysis of the scale pattern for river stock discrimination of chum salmon.
The scale samples used were those collected in 1988 from the two rivers, about 30kms apart from each other, in Iwate prefecture as is shown below in the Fig. 1. The data obtained from two groups of fish with these known stocks were firstly converted to those of power spectrum elements, then they were applied to the linear discriminant function as variables.
Satisfactory results of separation of the stock origin of the two with over 74 percents' accuracy were obtained by using the Linear Discrimination Function along with Power Spectralized Data.
It can be concluded from the results that the proposed analytical approach is of considerable value in distinguishing the river stocks of chum salmon.
View full abstract