| Summary: | Passages through hydropower plants can be fatal for downstream migrating salmonids (Salmo spp.). This is particularly true for large adults that are returning to the ocean after spawning. Physicalstructures such as racks can be used to guide fish towards fishways that sidestep high-mortality passages through turbines, but these structures are often too logistically challenging and economically burdensome to deploy at large scales. If functional, non-physical guidance structures, such as bubble barriers, could prove important for salmonid populations in regulated rivers, as they represent a low cost alternative that could also be deployed in larger rivers. This thesis aims to: i) quantify the diverting effect of bubble barriers on downstream migrating salmonids; and ii) disentangle the different sensory cues and behavioral traits that give rise to this potential diversion effect. My experiments showed a strong repelling effect of bubble barriers on downstream migrating Atlantic salmon (Salmo salar) in multiple contexts, ranging from laboratory flumes to 50-100 m long barriers deployed in a regulated river. My field experiments showed that bubble barriers can successfully guide both juveniles (smolts) and adults (kelts) of Atlantic salmon and sea run brown trout (Salmo trutta) at discharges exceeding 500 m3 s-1. The relative fish guidance efficiency ranged from 28% to 86%,and was negative correlated with water velocity. Based on laboratory flume experiments, adding stroboscopic lights reduced the guiding efficiency of the barrier, and interestingly, the repelling effect disappeared entirely when evaluated in darkness. These findings strongly suggest that visual cues are crucial for the repelling effect of bubble barriers. I subsequently hypothesized that the visual appearance of bubbles barriers might be perceived as an area associated with risk for fish, suggesting that more bold and active individuals could be more likely to pass through. However, I found no correlation between the probability of being ...
|
|---|