Results

Summary statistics of diel activity

We continuously recorded ~35,000 dives from 54 U. aalge individuals and ~8,000 dives from 16 A. torda individuals, ~36,000 prey depths and bins of 10° azimuth. Both predators showed higher foraging efficiency during twilight, with a higher peak during dusk compared to dawn (+28% and +17% for A. torda and U. aalge respectively; Fig 1A), but for A. torda the elevated efficiency also lasted much longer during afternoon/dusk (110° azimuth) than in the morning (60° azimuth). Both predators concentrated depth of their dives to the depth of the largest abundances of prey (median of prey depth = 16m, A. torda depth = 12m and U. aalge = 15m; Fig 1B). Still, U. aalge , but not A. torda , also used deeper depths (~20% dives at >50m) but only during daylight hours, while prey depth ranged from 4 to 96m both during the day or under low light levels. Both predators had a higher number of dives in light conditions with high foraging efficiency (Fig 1C), however U. aalge utilized a much wider range of light levels, both with the highest effort performed in lower light periods and with a higher proportion of dives throughout the day as compared to the A. torda . A. torda instead concentrated its effort during peak of twilight, with much more dives during dusk than dawn, and a more rapid decline in effort after the dawn peak. The median ° azimuth for dives were 73.7° for A. torda and 68.8° for U. aalge before noon (i.e. 0-180° azimuth), and 289° versus 290° after noon (180-360° azimuth), respectively. In summary, both predators foraged more in shallower waters, and were more efficient in twilight hours, most marked at dusk compared to dawn, and for Alca T. also more throughout the afternoon. Numbers of prey aggregations were highest around dawn and dusk, lowest during the night and at intermediate levels during daylight periods (Fig 1C). There was a distinct difference in prey abundance in the upper 50m of the water column between low-light and daylight periods (Table A3) as a result of vertical fish migration. This led to 22% increase in abundance in the upper 50 m of the water column, and an overall increase of 20% in detected abundance, during low light periods versus daylight (Carlsen et al., 2024).

Light-related activity and distribution patterns

Effort (number of dives) typically started increasing before the efficiency (BDCr) increased in the morning (0-50 versus 30-70 azimuth), and decreased only after the efficiency started decreasing (Fig 2A). Dive depth and bout lengths had opposite relations with solar position for U. aalge , where deeper dives coincided with shorter bouts in high light conditions and vice versa in the dark (Fig 2B). For A. torda bout lengths were best described as linearly increasing with azimuth. While the raw data previously showed an equal median dive depth across A. torda , the models depict a depth trend reflecting the vertical distribution of prey by azimuth similar to the trend inU. aalge . The U. aalge dives were markedly deeper during the day than night, and with higher peaks during vertical migrations, slightly more so during dawn than dusk (Fig 2B). When investigating the trends in diel PDPs, the depth distribution of prey was symmetrical around noon/midnight (Fig 2C), with inflection points at 73° and 286°, and an estimated average depth change of but was not symmetrical around noon/midnight (Fig 2C). The number of aggregations increased before the migration started in the early mornings (0-100° Azimuth), but already by 160° aggregations started dissolving and were reduced by ~25% by early afternoon (Fig 2C) and kept decreasing towards the evening. Notably, while the majority of aggregations, especially in high number, were observed at

Predator activity as a function of prey distribution