AN EVALUATION OF THE STEREOCINEMATOGRAPHIC METHOD TO ESTIMATE FISH SWIMMING SPEED

I evaluated the precision and accuracy of the stereocinematographic (SCG) method for estimating fish swimming speed. The SCG method implements the differences in images recorded by two cameras to determine the position of a target in an x, y, z, coordinate system. Movements and speeds were determined using variations in the position of the targets over time. Movements of rulers (95% Cl = XBAR +/- 4.6%) estimated in the laboratory did not differ significantly from measured values. The accuracy of the SCG method in the field was assessed by comparing simultaneous estimates of the speed of the head and of the tail of individual fish observed in in situ enclosures. Differences between these descriptors of fish swimming were always < 2 body lengths (bl).s-1 and, on average, did not differ significantly from 0. Swimming speeds (95% Cl = XBAR +/- 4.5%) ranged from 0.6 to 20.7 cm.s-1 (0.1-3.8 bl.s-1). Speed variations between two consecutive 1-s intervals ranged from -23.9 cm.s-1 (deceleration) to 23.6 cm.s-1 (acceleration). Positioning fish at 1- to 6-s intervals tended to decrease the variance of swimming speed estimates. A sample size of 100-150 speeds per hour was sufficient to accurately describe fish swimming in an in situ enclosure.