Exploring sensitivities to hydropeaking in Atlantic salmon parr using individual‐based modelling

Hydropeaking (the release of water pulses at hydropower plants) results in temporary reductions in river channel water‐covered area downstream, which may cause fish mortality through stranding. We used a mechanistic modelling approach to examine how both the form of the hydropeaking cycle and the characteristics of the affected fish control how hydropeaking may cause stranding mortality of fish. We modelled the response of Atlantic parr to hydropeaking in a regulated watercourse in central Norway (the River Nidelva) using an individual‐based population model designed explicitly to examine fish behaviour and stranding mortality during hydropeaking. A response to hydropeaking, involving migration from the riverbanks towards the mid‐channel on down‐ramping and a return to the riverbanks on up‐ramping, was based on individuals being parameterized to migrate to habitat properties that spatially changed throughout the hydropeaking cycle. We found that stranding mortality was strongly dependent on both the form of the hydropeaking cycle and on the fish response. Total stranding mortality was more dependent on the down‐ramping speed than the duration of the minimum flow period. Total stranding mortality was greatest when there was a low movement speed, leading to individuals being stranded, combined with a high probability of dying per unit of time when stranded. Given the sensitivity of mortality to the fish response and the lack of detailed field studies quantifying this, we conclude that this area requires further controlled studies for parameterizing models used to predict effects of hydropeaking on fish. Simulation‐based modelling, where there is a 100% control of and knowledge of fish characteristics, where a range of deterministic functions can be evaluated and in which boundary conditions can be easily controlled, may be a useful research tool to complement empirical studies.