Skip to main content

Box 1 Example of accelerometer transmitter data from a sea-run brown trout (Salmo trutta) in a lake in western Norway covered by acoustic receivers

From: Acoustic accelerometer transmitters and their growing relevance to aquatic science

Accelerometer transmitter data for one brown trout (Salmo trutta) in a lake in western Norway are presented from September to December 2021. The trout were tagged with Thelma LP13-ADT transmitters with 27 s sampling intervals at 10 Hz in summer 2021 and this individual spent several months in the exorheic lake Vasbygdivatnet, where an array of 12 acoustic receivers recorded the transmissions in near-continuous time. The data series illustrated here comprises sixteen weeks of data through the spawning season of trout and into the overwintering period when the fish ostensibly conserve energy before returning to the ocean to recondition the following summer

This individual, 4718 (Fig. A), had an average acceleration of 0.17 m/s2 (Fig. B). The time series of acceleration was regularised to a fixed interval of 120 s using the approx function in R and a Hidden markov model was fit to the data to identify two discrete states using the depmix function in the depmixS4 package in R. The first state identified seemed to be slow, steady-state swimming whereas the second state seemed characterised by more rapid burst-type activity with higher sustained acceleration values. State transition probabilities were 17% from state 1 to state 2 and 4% between state 2 and state 1 (Figure C)

View full size image

View full size image

Positions of the individual in the lake were calculated using YAPS (Baktoft et al. [3]) following standard procedures to synchronise receiver clocks. A generalised additive model was fit to the acceleration data to examine the spatial effect using a gamma family and an interactive longitude, latitude smoother with k = 25. Model predictions illustrated relatively high predicted acceleration in the northwestern part of the lake, closest to the outlet that flows into the river Aurlandselvi. Predicted values also tended to be higher in the middle of the lake indicating the fish possibly used the centre of the lake to transit to the edges where it swam less vigorously

View full size image

  1. View full size image
  2. Simple data summaries and analytics derived from the accelerometer transmissions are presented