Acoustic telemetry predation sensors reveal the tribulations of Atlantic salmon (Salmo salar) smolts migrating through lakes

Abstract

Smolt migration through lakes is hazardous, as the predation pressure can be extreme and the hydrology a great contrast to that of a riverine area. However, the mechanisms yielding these challenges have been scarcely investigated. We conducted an acoustic telemetry field study in Lake Evangervatnet, Voss, Norway, utilising Vemco V5 predation tags. Atlantic salmon (Salmo salar) smolts (N = 20) were tagged with the novel predation sensor tag to investigate mortality, the lacustrine migration behaviour of smolts, and the applicability of these tags for smolt studies. A total of 60% of tagged Atlantic salmon (Salmo salar) smolts perished in the lake. Half of the mortalities (30% of tagged fish) were directly attributed to predation by brown trout (Salmo trutta) based on predation sensors. The surviving smolts were slow to traverse the 6.5 km lake, with progression rate between lake inlet and outlet on average 0.016 m/s over a mean of 7.9 ± 6.2 (SD) days. Acoustic detections revealed a consistent pattern of nocturnal migration and multidirectional movements within the lake. By running a series of correlated random walks under varying parameters and comparing the simulated travel times to the observed travel time used by the tagged smolts, we emulated the observed behaviour of the smolts. These simulations suggested that smolts lacked the ability to efficiently navigate through the lake, instead swimming in random directions until they reached the lake outlet. Predation sensors can offer improved resolution when tracking the behaviour and fate of smolts and can facilitate better mitigation efforts by identifying survival bottlenecks and separating predation from non‐predatory mortality.

Investigating the phenology of juvenile potamodromous brown trout (Salmo trutta L.) in two large lake catchments

“…Kennedy et al (2018) showed high predation losses on S. salar smolts entering Lough Erne in the spring, due principally to the close proximity of post-spawned E. lucius to marginal shorelines and river mouths at that time of year. Autumn migrants may experience lower predation pressure than spring migrants, and more telemetry research into this issue using calibrated predation tags (Hanssen et al, 2021), to compare seasonal predation rates, would be valuable. Lough Neagh showed higher overall lake survival rates than Lough Erne.…”

Application of machine learning and acoustic predation tags to classify migration fate of Atlantic salmon smolts

“…The lag-time between tag consumption and the activation of the predation signal is ~ 5.8 hrs at 20°C (S. Smedbol, Innovasea Systems Inc., pers. comm., January 2020) or 35.4 ± 17.7 hrs at a mean temperature of 11.8°C (Hanssen, 2020).…”

“…Consultation with the tag manufacturer con rmed the possibility of undetected predation events (Type II error). Additionally, a previous validation study has shown V5D predation tags to have only 50% accurate detection of predation (Hanssen, 2020). Therefore, machine learning methods were also applied to the detection data to classify smolt fate.…”

Predation research with electronic tagging

“…Understanding predation dynamics of salmon has benefited greatly from electronic tagging. Analytical tools (Gibson et al 2015), field protocols (Skov et al 2014) and tag technologies (Halfyard et al 2017, Lennox et al 2021a) have been developed to assist calculations of predation to identify predation hotspots and bottlenecks for salmon smolts (Hanssen et al 2021). Combining tagging with biopsies and trait measurements has provided evidence for selective predation (Miller et al 2014).…”

“…Atlantic salmon kelts tagged with temperature‐sensitive archival tags revealed predation based on rapid temperature warming attributed to cetaceans and endothermic fishes. Another animal physiological indicator is the use of pH‐based sensors that can be used to indicate whether the animal has been digested in the stomach and intestines of the predator (Halfyard et al 2017, Hanssen et al 2021).…”