Diel cycles in marine predator diving behaviour centre around the light-mediated diel vertical migration (DVM) of prey, and are considered critical for optimizing foraging and limiting competition across global seascapes. Yet our understanding of predator diel behaviour is based primarily on examining relative depth usage between constant day/night cycles with no formal investigation of how varying light regimes interact with abiotic factors to shape diel activity. The extreme seasonal light regimes (midnight sun, polar night, day/night cycle) in the Arctic provide a unique natural experimental setting to empirically investigate the occurrence and intensity of diel behaviour in marine predators relative to changing light levels while concomitantly assessing interacting abiotic factors. Depth time series data from satellite-linked tags deployed on six belugas for up to 12 months were used to quantify diel behaviour by calculating dissimilarity in time-at-depth between periods of low and high solar altitude on each day. Generalized additive mixed effects models were used to examine the influence of hours of daylight across extreme light cycles, coupled with bathymetry and sea ice concentration; focal diel patterns were further examined relative to the thermal structure of the water column. As predicted, belugas exhibited cathemerality during the midnight sun, and initiated diel behaviour with the onset of the fall day/night cycle, with a marked increase in its intensity with the progression to equal day/night length. Occurrence of diel patterns, however, was complex; ceasing in regions with seafloor depths < 700 m, and occurring with greatest intensity when the water column was thermally homogeneous within the upper 150 m. Through empirical investigation, this study demonstrates that the onset of day/night light cycles and presumably associated prey DVM can modulate predator diel dive behaviour under certain circumstances, but highlights how the complex interaction of abiotic factors with light regime shape dynamic spatiotemporal patterns. These findings, building on a body of recent work, emphasize that the traditional view of the ubiquitous occurrence of diel behaviour tied to DVM at the base of the food web oversimplifies vertical predator-prey interactions, identifying the need for more structured investigation. 

Six male beluga whales were tagged with SPLASH10-F-238 satellite-linked transmitters (Wildlife Computers Ltd., Redmond, WA, USA) from Hendrickson Island, Northwest Territories, Canada, in July 2018. Tags collected Fastloc GPS locations and estimated Argos locations each time a transmission was made. Tags sampled depth at 1 s intervals and transmitted the data as time series in hour-long messages subsampled at 75 s intervals. 

Locations were filtered (see Storrie et al., 2022), and continuous-time correlated random walk (CTCRW) models were used to re-estimate locations at regular 15 min intervals (Johnson et al., 2008). The datasets provided here are a subset of CTCRW-modeled locations and depth time series data for one beluga (beluga ID LC2018#3) between August 4th and October 30th 2018, and can be used to replicate the methods for identifying diel patterns in dive behaviour. R code can be downloaded from a GitHub repository (https://github.com/storriel/Diel_Dive_Behaviour). 

These are csv files, which can be opened with Microsoft Excel, a text editor, or a programming language such as with the R code associated with the manuscript (https://github.com/storriel/Diel_Dive_Behaviour).