Large sensory volumes enable Southern elephant seals to exploit sparse deep-sea prey

Abstract

The ability of echolocating toothed whales to detect and classify prey at long ranges enables efficient searching and stalking of sparse prey in these time-limited dives. However, nonecholocating deep-diving seals such as elephant seals appear to have much less sensory advantage over their prey. Both elephant seals and their prey rely on visual and hydrodynamic cues that may be detectable only at short ranges in the deep ocean, leading us to hypothesize that elephant seals must adopt a less efficient reactive mode of hunting that requires high prey densities. To test that hypothesis, we deployed high-resolution sonar and movement tags on 25 females to record simultaneous predator and prey behavior during foraging interactions. We demonstrate that elephant seals have a sensory advantage over their prey that allows them to potentially detect prey 5 to 10 s before striking. The corresponding prey detection ranges of 7 to 17 m enable stealthy approaches and prey-specific capture tactics. In comparison, prey react at a median range of 0.7 m, close to the neck extension range of striking elephant seals. Estimated search swathes of 150 to 900 m² explain how elephant seals can locate up to 2,000 prey while swimming more than 100 km per day. This efficient search capability allows elephant seals to subsist on prey densities that are consonant with the deep scattering layer resources estimated by hydroacoustic surveys but which are two orders of magnitude lower than the prey densities needed by a reactive hunter.

Keywords: biologging; deep-diving predators; mesopelagic zone; prey detection.

Fig. 1.

Prey detection by SES. (A and B) Tracks of seals tagged with high-resolution sonar and movement tags in (A) Peninsula Valdes, Argentina, and (B) Kerguelen Islands. Tracks are colored by daily counts of PrCAs inferred from distinctive acceleration transients. (C) Example of a dive profile with true PrCA and fake PrCA intervals. (D–F) Prey detection behavior during the 20 s preceding each prey capture attempt. Each line summarizes between 3,093 and 20,146 PrCAs for an individual seal (n = 25). (D) Proportion of PrCAs for each seal with above-threshold pointing angle changes in 1.6-s bins synchronized to the prey strike. Pointing angle changes comprise changes in heading and/or pitch, and the 1.6-s interval is approximately the duration of one swimming stroke. The threshold was chosen from a comparable number of randomly selected fake PrCAs for each seal (dashed red lines). A high proportion of pointing angle changes 5 to 8 s before the strike indicate the likely detection distance. This is supported by the swimming behavior prior to the strike (E and F). Being denser than seawater, tagged seals glide on descents and stroke on ascents. When seals are ascending 20 s before a PrCA (E), swimming activity, parameterized by the root-mean-squared of the lateral acceleration averaged over 1.6-s bins, is initially high but seals often switch to gliding 4 to 8 s before the strike. Conversely, if seals are initially descending (F), they begin stroking just before the strike. In both cases, gliding prior to the strike increases stealth. See ESM for individual animal data.

Fig. 2.

Female elephant seals employ different capture tactics for individual and schooling prey. (A and B) Echogram displaying individual (A) or schooling prey (B) insonified over successive sonar pings. The horizontal axis shows the time relative to PrCA start in seconds, and the vertical axis shows the distance from the sonar tag in meters. The color scale indicates echo-to-noise ratio (ENR) on a dB scale. (C and D) Pitch (rotation around the left-right axis), roll (rotation around the longitudinal axis), and heading (rotation around the dorso-ventral axis) angles showing prey-dependent approach tactics, illustrated by a schematic representation of SES orientation. When approaching individual prey, SES generally maintained a fixed horizontal posture, while when approaching schooling prey, some SES consistently made an upward backflip maneuver.

Fig. 3.

Ecological significance of early prey detection. (A) Schematic representation of a female Southern elephant seal with estimated prey detection range, scaled with the seal body size. (B) Required prey density to achieve the same prey encounter rate with varying detection distance. The dotted and dashed lines show the estimated detection distances for a reactive forager and for a SES, respectively. The gray shaded region represents the apparent prey density of SES based on observed attack rates and travel distance; details of the calculation are given in the Materials and Methods section “Search volumes”. The blue arrow shows the estimated mesopelagic biomass obtained by dividing the total predicted biomass (4) by the approximate volume of the mesopelagic. Only a portion of this biomass is suitable SES prey. The red arrow shows the minimum biomass density needed if SES had a prey detection range of 0.7 m. While SES attack rates are consistent with the predicted mesopelagic resources, a reactive forager would require two orders of magnitude higher prey density to achieve a similar rate. (C) and (D) Daily averaged apparent prey densities for each seal, i.e., attack rate per (25 m)³ searched, are relatively constant along their far-ranging foraging tracks, indicating a stable widely distributed food resource.