Seascapes of fear and competition shape regional seabird movement ecology

The study investigates how non-consumptive predator effects (fear) and competition with fisheries shape the spatial ecology of Cape gannets in the Benguela upwelling ecosystem off South Africa. Building on the Seascape of Fear (SoF) concept—spatial variation in prey perception of predation risk—the authors ask whether fear of Cape fur seals and overlap with fisheries alter gannet space use at regional scales and over diel cycles. They predict that gannets will avoid seals, especially at night when risk is highest, inducing a diel spatial shift toward safer offshore waters. During daytime, when predation risk is lower but foraging needs are pressing, they expect gannets to balance risk and resource acquisition, potentially leading to overlap with fisheries targeting small pelagic fish.

The paper situates its research within the growing field of seascape ecology and extends concepts from terrestrial landscapes of fear to marine environments. Prior work has shown fear effects can modify prey behavior, induce migrations, and cascade to influence population dynamics and ecosystems. Despite increasing interest, marine applications of the landscape/seascape of fear framework remain comparatively rare, with most studies focused on coastal reefs, seabirds, and large predatory fishes. The authors note that temporal dynamics of fear have gained attention, yet seminal seascape ecology syntheses have underrepresented fear effects, highlighting a gap this study addresses by evaluating SoF at a regional scale in an upwelling system.

Study area and period: Western Cape, South Africa; chick-rearing Cape gannets from Malgas Island (33.09°S, 17.93°E) tracked during October–November from 2008 to 2015. Gannet tracking: 197 adult Cape gannets equipped with GPS loggers (Technosmart; ~65 g; sampling intervals 30 s to 2 min). Birds were captured at the colony using a pole with loop; devices attached to the lower back with waterproof tape. Tracks were processed to account for missing positions; behaviors (foraging vs resting) identified from speed-based criteria with smoothing over adjacent steps. Nighttime tactics classified as (i) return to colony at night or (ii) rest at sea. Seal tracking: 25 Cape fur seals Argos-tracked from three colonies during September–November across years overlapping and preceding gannet tracking (e.g., 2003–2004; 2012–2014). Tags included Argos-linked platform transmitters (e.g., Spot-6, ST6, ST20). Sampled ~6800 daytime and ~650 nighttime Argos locations. Standard capture and anesthesia procedures were used; permits and ethics approvals obtained. Seascape of fear proxy (seal occurrence): For each seal colony, resource selection functions (RSFs) were fitted separately for daytime and nighttime as proxies of seal occurrence at sea. RSFs compared used Argos positions (>500 m from colony) to 5× more random locations sampled within the 95% kernel UD of each colony. GLMMs with binomial error included environmental predictors (bathymetry, bathymetric slope, distance to colony), random intercept for seal ID with large fixed variance, random slopes for predictors; continuous predictors centered/scaled; VIF < 2. Model robustness assessed via leave-one-out cross-validation using rank correlations of predicted occurrence. Fisheries data: Vessel log-book records (2008–2015) provided annual catch distributions on a 10×10 nautical mile grid for fleets targeting small pelagics (anchovy Engraulis capensis/Engraulis encrasicolus and sardine Sardinops sagax; purse-seiners) and demersal hake (trawlers). Data from the Department of Forestry, Fisheries and the Environment (Marine Resource Management). Daytime foraging habitat selection: For GPS-tracked gannets, daytime foraging locations were contrasted with simulated available locations (track-based nulls respecting trip constraints from the colony). A binomial-choice modeling framework (generalized additive/mixed models) incorporated habitat, individual variability, and movement tactic. Models examined relationships with seal occurrence (RSF index), fisheries catches (purse-seiners, trawlers), and sea surface temperature (SST; Aqua MODIS, 4-km). Mixed-effects logistic regression assessed the probability of daytime foraging vs nighttime resting for birds that stayed at sea at night (n ≈ 142), including individual random intercepts; model selection used AICc; collinearity low (VIF < 2). Analyses conducted in R using mgcv, lme4, and related packages. Summary metrics: Reported nighttime rest behavior proportions and basic descriptive statistics of daytime habitat (distance to shore/colony, SST, seal occurrence, purse-seiner and trawler catches). Ethics approvals and permits were obtained for fieldwork.

• Movement tactics and diel behavior: Two nighttime tactics were observed. About 28% of gannets returned to the colony at night after evening foraging, while 72% remained at sea. Of nighttime resting, 76% occurred at the sea surface. Birds that remained at sea shifted offshore at night to areas with lower seal occurrence, with an average spatial shift of 13.1 km between daytime foraging and nighttime resting locations.
• Seascape of fear: Seal occurrence decreased with distance from shore similarly by day and night. Gannets selected daytime foraging areas that reduced co-occurrence with seals, consistent with prediction 1. At night, birds mitigated risk by either returning to the colony or moving offshore; the offshore shift reduced predation risk by approximately 25%.
• Day vs night trade-offs: During the day, gannets ventured into areas with higher seal occurrence than at night but still preferentially foraged where seal risk was lower. The probability of nighttime resting (vs daytime foraging) increased as seal occurrence decreased, except in the highest-risk areas where foraging still predominated.
• Thermal effects: Increasing SST initially had similar effects on foraging vs resting, but at high SST values, resting became more likely than foraging, implying a small thermoregulatory benefit that was outweighed by predation risk on average.
• Fisheries interactions: Gannets competed with fisheries during the day. They were more likely to forage in areas with higher purse-seiner catches of small pelagics and less likely to forage in areas with high trawler (hake) catches. On average, daytime foraging areas had higher purse-seiner catches and lower trawler catches than nighttime resting areas; however, at the daytime scale, foraging habitat selection did not rely solely on fisheries metrics.
• Descriptive habitat metrics (daytime, 197 birds): Distance to shore mean 24.5 km (range 0.0001–62.7 km); distance to colony mean 5.8 km (0.0001–14.5 km); SST mean 22.6°C (10.9–30.9°C); seal occurrence index mean 2.69 (1.00–4.77); purse-seiner catch mean 2.60 tons/h (0–5.52); trawler catch mean 1.42 tons/h (0–6.32). Nighttime proportion of time resting at sea: mean 0.7383 (min 0.5274, max 0.9542, SD 0.0955).

Findings validate that both fear (from Cape fur seals) and competition (with fisheries) jointly shape the spatial ecology and diel movement tactics of Cape gannets in the Benguela upwelling ecosystem. By day, gannets accept higher risk to access resources, yet still select foraging areas that minimize seal co-occurrence. By night—when gannets are most vulnerable—they either return to the colony or shift offshore to reduce overlap with seals, effectively lowering predation risk. These behavioral adjustments underscore the importance of integrating non-consumptive predator effects into marine spatial ecology and management. The study also highlights the complexity of trophic interactions (e.g., seals, seabirds, sharks, orcas) and the potential for ecological feedbacks in a system already stressed by climate change and overfishing. Incorporating SoF dynamics can improve understanding of seabird space use and aid in predicting responses to environmental change and human activities.

This study provides regional-scale evidence that fear and competition structure the movement ecology of Cape gannets. Gannets exhibit diel tactics that reduce overlap with predators at night and balance risk with foraging needs by day, including competitive interactions with fisheries. These insights extend the Seascape of Fear framework to a dynamic upwelling ecosystem and have direct implications for conservation planning. The authors recommend that marine protected area (MPA) design and broader marine management incorporate fear-driven spatial dynamics, alongside foraging habitat needs, especially under concurrent pressures from climate change and overfishing. Future research should integrate multi-predator networks and temporal dynamics of risk to better predict spatiotemporal distributions and population-level consequences.

Potential limitations include: (1) temporal mismatch and variability in seal tracking (data spanning different years and colonies) relative to gannet tracking; (2) Argos location uncertainty for seals vs higher-resolution GPS for gannets; (3) inference of seal occurrence via RSFs as a proxy for risk rather than direct predation events; (4) reliance on fisheries logbooks and gridded catch as proxies for resource competition; and (5) behavioral classification from movement data with assumptions about speed thresholds. While cross-validation and careful modeling were applied, these factors may affect precision and generalizability.