Using swimming speed and morphological attributes of fish derived from video image analysis to assess grilse proportion and segregation in mixed population of maturing and non-maturing Atlantic salmon (Salmo salar L.) reared in sea cage

More effective ways to manage mixed populations of maturing and non-maturing Atlantic salmon are needed, and a method to predict the percentage of maturing salmon would be ideal. Maturing fish are known to swim slower and have a larger condition factor (K) than non-maturing fish, and if differences are large, bimodality could be evident in swimming speed and K factor data collected from mixed populations, and used to predict the grilse (maturing salmon) percentage. To test this hypothesis, two positions within the swimming aggregation were video recorded before and after the fish were fed using a stereo camera system within commercial salmon farming cages. From recorded images, swimming speed and morphological attributes were determined. The expected bimodal distributions were not observed; all variables showed a normal distribution with a single mode. Positional and temporal variations in swimming speed and size measurements were found. The average swimming speed was generally lower and the K factor higher in the upper section of the swimming aggregation. Furthermore, the swimming speed and the condition factor were noted to be respectively higher and lower before the feeding event. It was surmised that grilse segregated themselves from the rest of the population, and / or differences in swimming speed and other attributes between grilse and non-maturing salmon were too small to statistically substantiate bimodality in the data sets. Results provide insight into the relative importance of early grilse detection and removal in feeding management.Applied Science, Faculty ofChemical and Biological Engineering, Department ofGraduat