Sustained Disruption of Narwhal Habitat Use and Behavior in the Presence of Arctic Killer Whales

Introduction

• Predators influence prey behavior, but electronic tracking data in marine environments rarely consider predator effects.
• A study was conducted in Admiralty Inlet, Eastern Canadian Arctic, tracking killer whales (N = 1 family group) and narwhals (N = 7) using satellite telemetry.
• The study analyzed movement data using a switching-state space model and mixed effects models.
• The presence of killer whales significantly alters narwhal behavior and distribution.

Key Findings

• Narwhals moved closer to shore (within ~100 km of killer whales), presumably for safety.
• Under predation threat, narwhal movement patterns were more likely transiting (highly autocorrelated), while resident behavior was more likely in the absence of threat.
• Effects persisted for 10 days, the duration of killer whale presence.

Significance

• Killer whales can reshape Arctic marine mammal distributions and behavior due to sea ice reduction and increased sightings.
• Predators can strongly affect movement behavior of tracked marine animals.
• Understanding predator effects is crucial, potentially more so than resource distribution or bottom-up drivers.

Consumptive vs. Nonconsumptive Effects

• Consumptive effects (density-mediated effects): Mortality when predators kill and consume prey, controlling populations and restructuring ecosystems through trophic cascades (1–3).
• Nonconsumptive effects (trait-mediated effects): Altered prey behavior and space use due to perceived predation risk, leading to decreased fitness (loss of foraging areas, disrupted social structure, increased energy expenditure, stress, decreased reproduction) (3–7).
• Nonconsumptive effects are sublethal but can impact many individuals, potentially exceeding consumptive effects (8, 10–12).
• Electronic telemetry tracking tags show carnivores affect prey species’ space and habitat selection in terrestrial systems (13–15).
• These nonconsumptive effects can negatively impact population dynamics (10, 13, 11) and potentially lead to trophic cascades (16, 17, 3).
• However, the strength of nonconsumptive effects to cause trophic cascades is debated (18–20).

Challenges in Marine Systems

• Marine environments are harder to observe, and tracked animals move over larger scales.
• It makes measuring predator density difficult.
• Analyses focus on habitat preference, resource distribution, and oceanographic controls (23–25).
• Ignoring predator effects can bias inferences about habitat preference.
• Nonconsumptive effects (e.g., lost foraging) could be misattributed to changes in productivity.

Research Approach

• Synchronously tracked killer whales (Orcinus orca) and narwhals (Monodon monoceros) in the Eastern Canadian Arctic (ECA).
• Demonstrated persistent interaction with killer whales induces changes in narwhal behavior and habitat use.
• Previous findings show killer whales elicit antipredator responses in other marine mammals (6, 27–32), but these are limited to immediate responses or simulated encounters.
• This study shows behavioral changes extend beyond discrete predation events.

Visual Representation

• Fig. 1: Map of tracking data after sSSM fitting; numbers indicate days since killer whale tag deployment.
• Red and blue indicate sSSM-inferred behavior (resident/transit) for narwhals.
• Green indicates the tracked killer whale.
• Fig. 2: Behavioral time series for three tracked narwhals.
• Colors indicate behavioral state (red=resident, blue=transit).
• Black lines: movement vector; orange lines: distance from killer whale; green lines: killer whale displacement vector (if <15 km).

Behavioral Changes and Press Disturbance

• Behavioral changes extend beyond predation events, with altered behavior and habitat use persisting.
• Narwhal behavior returned to normal after killer whales left.
• This suggests killer whales act as a press disturbance: persistent change while present, rapid recovery after removal.

Relevance to Arctic Ecosystems

• Arctic summer sea ice cover is declining (33, 34), affecting trophic levels (35–37) and ice-dependent species (38–43).
• Ice degradation allows more open water species access (44, 45).
• Killer whales are historically limited but now have increased access to areas like Davis Strait and Lancaster Sound (46), and are observed annually in Hudson Bay (47).
• Arctic warming may bring new predation threats, reshaping marine mammal distributions.

Experimental Setup

• Argos tracking tags were deployed on killer whales and narwhals in Admiralty Inlet in 2009 (72.55◦ N, 86.28◦ W).
• Admiralty Inlet: a long (300 km), wide (50 km), deep (>1,000 m) fjord (Fig. 1).
• One killer whale tracked from August 15, 2009, representing a group of 12–20 individuals.
• Seven narwhals tracked contemporaneously (August 15–18, 2009); shared Admiralty Inlet for ~10 days.
• Killer whale departure on August 28, 2009, allowed comparison of behavior under predation risk (exposure period) vs. after departure (postexposure period).

Analytical Methods

• Tracking data analyzed with a behavioral-state switching-state space model (sSSM) (49–51) and mixed-effects models.
• sSSM discriminated two behavioral states in narwhal tracks: transit (highly autocorrelated) and resident (negatively/nonautocorrelated).
• Killer whale movement not clearly discriminated into two states due to patrolling behavior.
• Narwhals did not move as a cohesive social unit, showing fission–fusion dynamics (52, 53).
• Movement patterns and reactions to killer whale presence were not correlated across individuals via group movement dynamics (54).

Habitat Use

• Habitat use differed strongly between exposure and postexposure periods (Fig. 3).
• Exposure period: narwhals constrained to a narrow band near shore (within 500 m).
• Postexposure period: narwhals moved offshore (4–10 km from coastlines), avoiding areas
• Depth preference mirrored this pattern:

Statistical Models

• Mixed effects model fits: killer whale presence, distance to killer whales, and behavioral state significantly predicted distance from shore (Table 1).
• Exposure category was the most predictive.
• Additive model included behavioral state, exposure category, and distance to killer whales.

Key Parameters

• Exposure resulted in habitat use close to shore.
• Transit behavior was associated with being close to shore.
• As distance to killer whales increased, narwhals moved farther from shore.

Movement Behavior

• GLMMs indicate killer whale presence significantly affected narwhal behavior (Table 2).
• Depth, distance to shore, exposure category, and distance to killer whales predicted behavioral state.
• Killer whale exposure increased probability of being in transit state, as did shallower water.
• Interaction parameter: narwhals exposed to killer whales in deep water were more likely in transit state.
• Turn angles were straighter during exposure period, more turns near 180◦ postexposure.
• Significant difference (P < 0.001) using Watson two-sample test (55).
• Dive behavior: deeper dives about 10% more frequently, shorter dives by about 25 s (14%).

Nonconsumptive Predator Effects

• Killer whale call playback experiments and direct observations show strong evasive responses (6, 27, 30, 31).
• Narwhals initiate evasive behaviors during/after killer whale attacks (26, 57–59).
• This study shows behavioral changes extend beyond predation events, persisting for the duration of killer whale presence.
• These are nonconsumptive predator effects: altered behavior and habitat usage.

Marine vs Terrestrial Systems

• Telemetry data show nonconsumptive effects in terrestrial species (8, 13–15).
• Data showing nonconsumptive effects in open marine systems are lacking.
• Simultaneous tracking efforts have yielded mixed results (60–62).
• Standard survey methods show changes in prey distribution due to predatory sharks (21, 22, 63, 12, 64).

Relevance to Ecosystems

• Changes in behavior caused by whale-watching vessels were associated with modest but significant increases in energetic costs, but disproportionately larger losses in foraging opportunities (8).
• Nonconsumptive effects could be more important than consumptive effects (56).

Population Impact

• Changes in narwhal habitat use/behavior likely affect all narwhals in Admiralty Inlet (~35,000 individuals) (70, 71).
• Small nonconsumptive effects across a large population can have an appreciable impact.
• Seals, narwhal, and beluga are mesopredators; increased killer whale densities can elicit structural changes to Arctic ecosystems (3, 72, 73).

Implications for Telemetry Data

• Killer whales are globally distributed predators (77–79); large predatory sharks exist worldwide (80).
• Predator location is key to understanding prey movements.
• Interpretation of movement data rarely includes exact predator positions or densities.
• This may lead to incorrect inference about an animal’s biology and management advice.

Management Considerations

• Tracking data are used for marine management (23, 85–88).
• Researchers/managers need to consider how predators affect space use.
• Analyses might erroneously infer areas used as refugia from predators to be preferred areas.

Contextual Factors

• Killer whale predators share habitat with Arctic prey for 1–2 months, but most marine mammals are free from predation for most of the year.
• In other areas (North Pacific, North Atlantic), killer whales/sharks have a perennial presence, provoking large nonconsumptive effects (89).

Alternative Explanations

• Nutritional stress, reproductive failure, and starvation are often attributed to changes in primary/secondary production (90–92).
• These symptoms can also result from increased predator presence in foraging areas.
• Without understanding nonconsumptive effects, conclusions about movement behavior or demographic parameters may be misleading.

Methods

• sSSMs used to estimate locations and infer behavior (49–51, 93).
• Mixed effects models and GLMMs compared habitat preference and behavioral state of tracked narwhals during exposure and postexposure periods (50).
• Habitat parameters: distance to shoreline and water depth.
• Differences in dive depth, duration, and time at depth were compared between exposure categories using GLMMs.