Snorkelling as a method for assessing spawning stock of Atlantic salmon, Salmo salar

Reliability of underwater snorkel counts of adult Atlantic salmon, Salmo salar L., was analysed in the tributaries of the River Teno, close to the spawning period. In small (width 5-20 m) rivers, the replicated total counts of salmon were reasonably precise (CV = 5.4-8.5%), while in the medium-sized river (width 20-40 m) the precision of the counting method was considerably lower (CV = 15.3%). Low precision in a medium sized river was also observed in an experiment using marked live fish where the observation efficiency varied between 36.4% and 70.0%. In a small river, the detection efficiency of artificial fish silhouettes (test salmon) was almost perfect in pools (98%), but decreased in rapids (84%). Separate counts of males, females, grilse and large salmon were usually more variable than total counts, indicating that divers were more capable of locating a fish than properly identifying its sex and sea-age. The behaviour of adult salmon was favourable to conduct snorkel counts, as fish normally stayed still, or after hesitating, moved upstream (> 95% of the cases) when encountering a diver. The high observation efficiency (> 90%) and precision, favourable behaviour of salmon and congruence between snorkel counts and catch statistics in small rivers suggest that reliable data on Atlantic salmon spawning stock can be collected by snorkeling provided that the environmental conditions are suitable and the divers are experienced.