Measures have been taken to curtail commercial fishing of walleye pollock (Theragra chalcogramma) in Alaska in an attempt to stop the decline of its endangered population of Steller sea lions (Eumetopias jubatus). But our night-time observations of these mammals in Prince William Sound using infrared scanning technology, combined with acoustic surveillance of their prey's behaviour, reveal that the sea lions feed exclusively on Pacific herring (Clupea pallasi), which are less abundant than pollock but are found closer to the surface at night.

Food limitation is the principal factor in the decline of Steller sea lion populations1,2,3,4. This decline could be explained by competition with commercial fisheries, as it has coincided with the growth of the pollock-fishing industry, which has become one of the largest fisheries in the world, or it could be related to a change in predator–prey relationships, possibly driven by ocean climate shifts5. Central to the uncertainty surrounding the drop in the numbers of Steller sea lions is a lack of observational data on their foraging ecology. There is no quantitative information available that directly relates the foraging behaviour of these animals to the abundance of prey species.

During the winter period, nutritional stress is high. Sonar surveys6,7 of the abundance and distribution of adult Pacific herring and walleye pollock in winter have been made in Prince William Sound in Alaska since the early 1990s8. Steller sea lions were seen during the day near herring schools, but as no foraging activity was detectable, the significance of this co-occurrence was questionable.

We complemented our sonar surveys during March 2000 with infrared scanning of the Steller sea lions. This technology, which is widely used in night-time military operations and surveillance, enabled us to monitor the animals' activity during the hours of darkness. Our system had a 27° × 18° field of view and a spectral response of 7–14 μm.

The estimated herring biomass in Prince William Sound in the sonar survey of March 2000 was 7,281 metric tonnes (95% confidence interval, 5,898–8,664). The estimate of pollock biomass was 28,277 metric tons (95% confidence interval, 26,034–30,420). Despite the much greater abundance of pollock, the infrared system revealed that foraging by Steller sea lions was exclusively on herring and was conducted only at night. Foraging activity was intense on dense herring schools (Fig. 1). Steller sea lions were often observed swimming side by side in a row of 50 or more individuals along the edges of a school, suggesting that they were herding the herring. Humpback whales and seabirds were also seen to be feeding alongside the sea lions (Fig. 1). By contrast, no sea lions were coincident with pollock schools.

Figure 1: Location of groups (pods) of Steller sea lions around herring schools.
figure 1

a, Combined acoustic and infrared sensors reveal sea lions and birds located on the surface above the herring school at night in Rocky Bay, Prince William Sound (March 2000). b, Infrared video image showing a line of Steller sea lions and a humpback whale on the sea surface above a school of herring.

Full size image

The sonar records revealed herring schools at depths of 10–35 m at night, but deeper during the day. Walleye pollock, on the other hand, remained at depths of over 100 m during both day and night. Pollock schools were also found in less protected regions and were further offshore. Although Steller sea lions are capable of dives exceeding 250 m (ref. 9), the more accessible distribution of herring at night may be the primary factor in the foraging behaviour of the sea lions. This distribution of herring is characteristic during an extended overwintering period in the North Gulf of Alaska.

Our results indicate that the dependence of Steller sea lions on herring as prey has been underestimated. The infrared scanning technology that has led us to this conclusion should also help in the evaluation of night-time foraging behaviour of other marine mammals and seabirds, with its remarkable ability to detect individual fish flipping on the sea surface at a distance of 5–30 m, as well as sea lions, whales and birds at over 100 m.