Predicting fine-scale downstream migratory movement of Atlantic salmon smolt (Salmo salar) in front of a hydropower plant

Abstract

The Atlantic salmon (Salmo salar) is an iconic species of significant ecological and economic importance. Their downstream migration as smolts represents a critical life-history stage that exposes them to numerous challenges, including passage through hydropower plants. Understanding and predicting fine-scale movement patterns of smolts near hydropower plants is therefore essential for adaptive and effective management and conservation of this species. We present a spatially explicit individual-based model for predicting the movement of Atlantic salmon smolts in regulated rivers in Norway, parameterised for smolt movements in the River Mandal and the River Orkla. The model is rooted in statistically derived relationships between observed smolt swimming behaviour and the hydraulic variables they encounter. The aim of the model was to provide fast yet representative swimming patterns past hydropower plants, based on the hydraulic conditions experienced by the smolts. The model outperformed a ‘drift-only’ model in portraying observed swim tracks when comparing simulated and observed tracks. It was found to represent smolt swimming behaviour well. Our results show that by constructing swim models using relatively simple and general statistical relationships between smolt swimming behaviour and the hydraulic environment, we can produce fast and relevant outputs for an adaptive management process, aimed at exploring how physical implementations or changes in flow regulations might affect smolt populations. Individual based models, Spatial modeling, Fish behavior, Boosted regression trees, 3D telemetry, Fish conservation