Overwintering and migration behaviour of post‐spawned <scp>Atlantic</scp> salmon <scp> Salmo salar </scp> in a large hydropower‐regulated river and reservoir

Tracking 47 post-spawned adult Atlantic salmon Salmo salar L. in a hydropower-regulated river through autumn, winter and spring revealed that winter survival was 56% and 75% in two study years, respectively, with higher mortality of males than females (50% vs. 33% and 100% vs. 13%, respectively). Some kelts (n = 7) displayed nondirected movements that were interpreted as a reconditioning period for an average of 9-17 days prior to directed downstream movements indicating the initiation of migration. Survival after the initiation of migration in spring was 83% and 94% to the hydropower dam in the first and second study years, and decreased to 60 and 63%, respectively, after dam passage. There were no further losses in the downriver reach in the second year, with the first year having a cumulative survival estimate of 53% to the river mouth. Kelts approached the dam when the spillway gates were available as a passage option most of the time (64%-75%), but some kelts arrived at the dam or had not yet passed when spillways were closed (n = 6) and the only remaining passage option was restricted to the turbines. However, all but one kelt that must have passed via turbine were successful in reaching the river mouth. Migratory delay presumably due to searching behaviour caused by low water flow was estimated at approximately 6 days as migration rates were significantly slower in the reservoir (median &#177; s.e. 8.5 &#177; 2.5 km day-1 ) than up- (29.7 &#177; 5.0 km day-1 ) or downriver (22.1 &#177; 3.1 km day-1 ). The proportion of time (median 30%) that kelts spent swimming upstream (searching behaviour) in the reservoir was a significant variable for migration success.