Sink or Swim: Strategies for Cost-Efficient Diving by Marine Mammals

See allHide authors and affiliations

Science  07 Apr 2000:
Vol. 288, Issue 5463, pp. 133-136
DOI: 10.1126/science.288.5463.133


Locomotor activity by diving marine mammals is accomplished while breath-holding and often exceeds predicted aerobic capacities. Video sequences of freely diving seals and whales wearing submersible cameras reveal a behavioral strategy that improves energetic efficiency in these animals. Prolonged gliding (greater than 78% descent duration) occurred during dives exceeding 80 meters in depth. Gliding was attributed to buoyancy changes with lung compression at depth. By modifying locomotor patterns to take advantage of these physical changes, Weddell seals realized a 9.2 to 59.6% reduction in diving energetic costs. This energy-conserving strategy allows marine mammals to increase aerobic dive duration and achieve remarkable depths despite limited oxygen availability when submerged.

Swimming is energetically expensive for mammals and results in transport costs that are 2 to 23 times the levels predicted for fish (1, 2). To reduce these costs, marine mammals have developed a wide variety of energy-conserving swimming behaviors. Adherence to a narrow range of routine transit speeds (3, 4), wave-riding (5), and porpoising (6) decrease the amount of energy expended when pinnipeds and cetaceans move near the water surface. Although these energy-conserving strategies are especially beneficial during underwater activity, when access to ambient oxygen is limited, two of the behaviors, porpoising and wave-riding, cannot be used when the animal is submerged. In view of this, it has been assumed that marine mammals swim constantly at cost-efficient routine speeds during diving (3, 4). Indeed, the routine speeds of many freely diving marine mammals fall within a relatively narrow range (7, 8). A paradox arises when metabolic rates are assigned to these swimming speeds. Calculations based on measured speeds during diving and metabolic rates for bottlenose dolphins swimming near the water surface predict that the animals would be unable to complete a 200-m-deep dive using aerobic metabolic pathways. Yet, dolphins perform these dives with only small changes in postdive plasma lactate concentrations, where elevated levels would indicate a transition to anaerobic metabolism (9). Similar discrepancies between predicted aerobic capabilities and diving performance have been reported for a wide variety of marine birds and mammals (10, 11). The mechanisms by which these divers resolve the apparent conflict between the energetic demands of swimming and the conservation of limited oxygen stores during submergence are not understood (12, 13). Metabolic depression (14) and regional heterothermy (15) associated with cardiovascular changes during diving have been suggested, although the influence of activity level has not been assessed. This has been due in part to the difficulty of observing and monitoring diving animals at depth.

Here, we monitored locomotor behavior during diving with video cameras carried by free-ranging cetaceans and pinnipeds. Unlike other instruments placed on marine mammals in which behavior has been inferred from time-depth records or velocity profiles (16), video images permit direct observation of swimming periods, stroke frequency, and glide sequences. Coupled with time-depth recorders, these new tools allowed us to assess the locomotor strategies used by marine mammals throughout their dives.

Subjects for this study included three adult Weddell seals (Leptonychotes weddellii, body mass = 393 ± 2 kg) diving from an isolated ice hole in McMurdo Sound, Antarctica (17), a juvenile northern elephant seal (Mirounga angustirostris, 263 kg) freely diving in Monterey Bay, California (17), an adult bottlenose dolphin (Tursiops truncatus, 177 kg) trained to dive to submerged targets offshore of San Diego, California (18), and an adult blue whale (Balaenoptera musculus, estimated mass = 100 tons) traveling offshore of northern California along Cordell Bank (19). Each animal carried a submersible video system with a camera facing either forward to record movements of the head or backward to record propulsive movements of the flukes or hind flippers. Data loggers simultaneously monitored duration and depth of dives.

Instrumented animals were free to perform sequential dives in open water or, in the case of the Weddell seals, below the frozen sea ice. The video system and instrumentation were retrieved when the animals hauled out (Weddell seals, elephant seal) or returned to a trainer (dolphin), or the package was detached by a release mechanism (blue whale). Swimming mode, relative stroke amplitude, stroke frequency, and gliding periods were determined for each video sequence, using a motion analysis system (Peak Performance, Englewood, Colorado). These data were then matched to duration and depth of the associated dive (20). To assess the effect of changes in locomotor pattern on energetic costs, we measured postdive oxygen consumption of instrumented Weddell seals breathing into a metabolic hood (21,22).

Despite independent evolution of swimming in cetaceans and pinnipeds, and differences in body size and propulsive mechanisms, we found a similar sequence of locomotor gaits during diving for the four species examined (Fig. 1). Diving descents began with 30 to 200 s of continuous stroking that was followed by a marked, prolonged period of gliding to maximum depth. Gliding began at similar depths (86 ± 10 m,n = 3 species) and continued to the bottom of the dive for the seals and dolphin, although maximum dive depths ranged from 115 to 385 m. The blue whale began gliding at comparatively shallower depths (18 ± 1 m, n = 3 dives) during dives of 36 to 88 m in depth. Descent rate during the glide varied little among the three smaller species (1.1 ± 0.1 m s–1, n = 3 species), whereas the blue whale descended considerably slower at 0.3 to 0.4 m s–1. The absolute duration of stroking or gliding sequences depended on maximum depth and dive duration. Deep divers (the phocid seals) showed the longest absolute glide periods. Maximum glide duration was 6.0 min for the juvenile elephant seal descending to nearly 400 m and 6.2 min for an adult Weddell seal descending to 540 m.

Figure 1

Locomotor activity of four species of diving marine mammal. Representative deep dives are presented for the (A) Weddell seal (maximum depth = 385 m), (B) northern elephant seal (333 m), (C) bottlenose dolphin (115 m), and (D) blue whale (84 m). Each curve represents dive depth in relation to time elapsed during the dive. Color of the line corresponds to stroking (black) and gliding (red) periods. Stroking periods include both continuous stroking and stroke-and-glide activities. Note the prolonged gliding period during descent for each species.

Initial ascent of each dive was characterized by sequential, large-amplitude strokes. The range of frequencies during steady stroking on initial ascent was 60 to 110 strokes min–1(1.0 to 1.8 Hz) for the three smaller species (dolphin, elephant seal, Weddell seal). In comparison, the range of stroke frequencies was one-tenth of this range (6 to 10 strokes min–1; 0.1 to 0.2 Hz) for the massive blue whale. Ascent rate during the period of constant stroking was 1.0 ± 0.2 m s–1 for all four species examined.

Following the period of continuous stroking, the animals switched to stroke-and-glide swimming for the remainder of the ascent except for a short (<100 s) glide to the surface. Only the Weddell seals did not glide the final 10 m to the surface, a behavior that was undoubtedly influenced by the presence of the sea ice and the maneuvering required to reach the isolated ice hole.

The similarity in locomotor behaviors for these four species is striking given the ranges of body sizes and propulsive mechanisms. Both cetaceans use dorsoventral undulations of a lunate tail for propulsion (23). The two pinniped species swim with alternate lateral sweeps of paired hind flippers in which the posterior half of the body may flex (24).

Passive gliding by the seals and dolphin began at nearly identical depths, suggesting that changes in hydrostatic pressure and buoyant forces prompted the incorporation of prolonged glide sequences during descent. Previous studies have shown that bottlenose dolphins (25) and elephant seals (26) modify ascent and descent rates during deep dives in response to changes in buoyancy. In dolphins diving to 100 m, the magnitude of the buoyant force changes from positive (+24.3 N) near the water surface to negative (–25.7 N) at depths exceeding 67 m (25). These changes are attributed to the gradual collapse of the lungs and a decrease in lung volume that occur with increasing hydrostatic pressure during descent. Complete collapse of the alveoli occurs once dolphins have reached pressures equivalent to 65 to 70 m in depth (27, 28). Likewise, the morphological structure of the respiratory system of elephant seals and Weddell seals indicates the capacity for collapse that may affect buoyancy during the course of a dive (29, 30). Because compression of the air spaces decreases the volume of the animal without a change in mass, buoyancy decreases on descent. When the downward force of negative buoyancy exceeds drag forces, the animal may glide passively during descent, thereby avoiding the energetic costs associated with active stroking.

As might be expected, dive depth, and therefore distance traveled, affects the percentage of time available for gliding. The percentage of time spent gliding during descent increased significantly (n = 53, r 2 = 0.70,P < 0.001) and nonlinearly with increasing dive depth (Fig. 2). This percentage ranged from 10 to 63% for shallow dives of less than 100 m and reached a plateau of 82 ± 2% (n = 21) for deep dives exceeding 200 m. All deep dives were by the phocid seals. Blue whales also showed extended gliding sequences that exceeded 78% of the descent period for dives to 88 m.

Figure 2

Percentage glide time during descent in relation to dive depth for four species of marine mammal. Each point represents an individual dive. The data were described by the nonlinear function, percentage glide time =Embedded Image(n = 53,r 2 = 0.70, P < 0.001). Except for the dolphins, the range of depths was determined by the free-ranging behavior of the instrumented animals.

A reduction in locomotor effort afforded by gliding should be manifested as a decrease in energetic cost. This has been reported for short-duration glides associated with intermittent (stroke-and-glide) locomotion in fish (31) and diving birds (32,33). We found a similar result for Weddell seals that incorporated prolonged glides during diving (Fig. 3). Oxygen consumption during the recovery periods of individual dives was measured for three adult, free-ranging seals wearing video instrumentation (21, 22). Two groups of dives covering similar distances but differing in gliding and swimming behaviors were compared (34). Dives incorporating gliding during descent resulted in a 9.2 to 59.6% (mean = 27.8 ± 5.5%,n = 10) reduction in recovery oxygen consumption compared with dives using stroke-and-glide or continuous swimming. In general, greater savings occurred with deep dives, which is consistent with the increase in the proportion of time gliding with depth (Fig. 2). In view of these results, there appears to be a significant energetic advantage to gliding rather than swimming on descent by marine mammals.

Figure 3

Recovery oxygen consumption of gliding dives in relation to stroking dives for free-ranging Weddell seals (34). Each point represents a gliding dive paired with a stroking dive of equal distance traveled (±60 m) for an individual seal. Total distance traveled ranged from 354 to 3614 m, which resulted in the range of energetic costs. The thin line through the origin represents the line of equality for the cost of gliding dives and stroking dives. The thick solid line denotes the least-squares linear regression through the data points. Dives incorporating prolonged gliding were consistently less costly than stroking dives of similar distance, as described by glide cost = 0.88 stroke cost – 7.30 (n = 10,r 2 = 0.91, P < 0.001). Consequently, all paired dives fell below the line of equality.

The value of the energetic savings is demonstrated by examining the effect on the oxygen reserves of the diving seal. A 400-kg Weddell seal stores 87 ml of O2 per kg of body weight (ml O2 kg–1) in its lungs, blood and muscle to support aerobic metabolism while submerged (10, 11). An average energetic savings of 27.8% (Fig. 3) due to prolonged gliding represents 24.2 ml O2 kg–1. The metabolic rate of Weddell seals during rest or low levels of underwater activity was 3.2 ml O2 kg–1 min–1. At this metabolic rate, the oxygen saved by gliding allows the seal to extend its aerobic dive time by 7.5 min (24.2 ml O2kg−1 divided by 3.2 ml O2 kg–1min–1) assuming the same level of activity. This additional time represents 38% of the routine 20-min dive duration of free-ranging Weddell seals. The energetic savings could make the difference between completing a dive aerobically or relying on anaerobic metabolism with the coincident disadvantages associated with the accumulation of lactate and prolonged recovery (10, 11). For marine mammals that are hunting, these savings may increase the overall efficiency of foraging. During traveling, the energetic savings when submerged may reduce the cost of long-distance migrations.

The ability of marine mammals to take advantage of physical changes at depth permits the conservation of limited oxygen stores during submergence. These results provide insight into the means by which diving marine mammals resolve the conflict between the energetic demands of swimming and the need for energy conservation during submergence. Prolonged gliding behavior by diving marine mammals appears to be a general phenomenon, irrespective of the method of propulsion and size of the animal. Even the largest mammalian diver, the blue whale, displays this behavior.

  • * To whom correspondence should be addressed. E-mail: williams{at}


View Abstract

Navigate This Article