Abstract: Difficulties quantifying in situ prey patch quality have limited our understanding of how marine predators respond to variation within and between patches, and throughout their foraging range. In the present study, animal-borne video, GPS, accelerometer and dive behaviour data loggers were used to investigate the fine-scale foraging behaviour of Macaroni penguins (Eudyptes chrysolophus) in response to prey type, patch density and temporal variation in diving behaviour. Individuals mainly dived during the day and utilised two strategies, targeting different prey types. Subantarctic krill (Euphausia vallentini) were consumed during deep dives, while small soft-bodied fish were captured on shallow dives or during the ascent phase of deep dives. Despite breeding in large colonies individuals seemed to be solitary foragers and did not engage with conspecifics in coordinated behaviour as seen in other group foraging penguin species. This potentially reflects the high abundance and low manoeuvrability of krill. Video data were used to validate prey capture signals in accelerometer data and a Support Vector Machine learning algorithm was developed to identify prey captures that occurred throughout the entire foraging trip. Prey capture rates indicated that Macaroni penguins continued to forage beyond the optimal give up time. However, bout-scale analysis revealed individuals terminated diving behaviour for reasons other than patch quality. These findings indicate that individuals make complex foraging decisions in relation to their proximate environment over multiple spatio-temporal scales.

Abstract: Fossorial locomotion is often considered as the most energetically costly of all terrestrial locomotion. Small arctic rodents, such as lemmings, dig tunnels not only in the soil but also through the snowpack, which is present for over 8 months of the year. Lemmings typically dig in the softest snow layer called the depth hoar but with climate change, melt-freeze and rain-on-snow (ROS) events are expected to increase in the Arctic, leading to a higher frequency of hardened snowpacks. We assessed the impacts of snow hardness on the locomotion of two lemming species showing different morphological adaptations for digging. We hypothesized that an increase in snow hardness would (1) decrease lemming performance and (2) increase their effort while digging, but those responses would differ between lemming species. We exposed four brown lemmings (Lemmus trimucronatus) and three collared lemmings (Dicrostonyx groenlandicus) to snow of different hardness (soft, hard, and ROS) during 30-min trials (n = 63 trials) in a cold room and filmed their behavior. We found that the digging speed and tunnel length of both species decreased with snow hardness and density, underlining the critical role of snow properties in affecting lemming digging performance. During the ROS trials, time spent digging by lemmings increased considerably and they also started using their incisors to help break the hard snow, validating our second hypothesis. Overall, digging performance was higher in collared lemmings, the species showing more morphological adaptations to digging, than in brown lemmings. We conclude that the digging performance of lemming is highly dependent on snowpack hardness and that the anticipated increase in ROS events may pose a critical energetic challenge for arctic rodent populations.