Accuracy of Diver Counts of Fluvial Rainbow Trout Relative to Horizontal Underwater Visibility

Abstract

AbstractWe investigated the effects of variation in underwater visibility on the accuracy of diver counts of rainbow trout Oncorhynchus mykiss in the Salmo River, British Columbia, over a 3‐year time period. A four‐man team of divers, drifting in a downstream direction, made periodic counts of trout along a study reach in which radio‐tagged fish that had also received a visual mark were present. Observer efficiency of divers (number of tags seen relative to the number known to be present) was significantly related to horizontal underwater Secchi disk visibility during 2002 and 2003 but only poorly so for the first year of the study in 2001. Overall, horizontal visibility in the 3 years' combined data set was significantly related to observer efficiency, explaining 62% of the variation. Diver counts of untagged trout confirmed these patterns. Diver counts of trout greater than 30 cm and greater than 40 cm showed precise, significant relationships with horizontal visibility for both 2002 and 2003. Horizontal visibility changes explained 94% and 93% of the variation in counts of trout greater than 30 cm and greater than 40 cm, respectively, in 2002 and 94% and 87% of the variation for the same size categories in 2003. The poor‐quality, nonsignificant relationships between counts of trout and visibility in 2001 were consistent with the poor observer efficiency relationship for that year as estimated from the observations of the radio‐tagged fish. Taken together, these poor‐quality relationships for the first year of the study point to either crew inexperience or some other factor(s) as an important source of error in addition to water clarity.