Individuality counts: A new comprehensive approach to foraging strategies of a tropical marine predator

Abstract

Foraging strategies are of great ecological interest, as they have a strong impact on the fitness of an individual and can affect its ability to cope with a changing environment. Recent studies on foraging strategies show a higher complexity than previously thought due to intraspecific variability. To reliably identify foraging strategies and describe the different foraging niches they allow individual animals to realize, high-resolution multivariate approaches which consider individual variation are required. Here we dive into the foraging strategies of Galápagos sea lions (Zalophus wollebaeki), a tropical predator confronted with substantial annual variation in sea surface temperature. This affects prey abundance, and El Niño events, expected to become more frequent and severe with climate change, are known to have dramatic effects on sea lions. This study used high-resolution measures of depth, GPS position and acceleration collected from 39 lactating sea lion females to analyze their foraging strategies at an unprecedented level of detail using a novel combination of automated broken stick algorithm, hierarchical cluster analysis and individually fitted multivariate hidden Markov models. We found three distinct foraging strategies (pelagic, benthic, and night divers), which differed in their horizontal, vertical and temporal distribution, most likely corresponding to different prey species, and allowed us to formulate hypotheses with regard to adaptive values under different environmental scenarios. We demonstrate the advantages of our multivariate approach and inclusion of individual variation to reliably gain a deeper understanding of the adaptive value and ecological relevance of foraging strategies of marine predators in dynamic environments.

Keywords: Broken stick algorithm; Conservation; Galápagos sea lion; Hidden markov models; Individual differences.

Fig. 1

Dive profile analyzed with the broken stick algorithm. Blue segments represent sections with a vertical sinuosity index above 0.9, indicating transit, while red segments represent sections with a vertical sinuosity index below 0.9, indicating foraging episodes

Fig. 2

Dendrogram of 177 foraging trips clustered with a hierarchical cluster analysis (Euclidean distance, Ward’s method) into three clusters (cluster 1 = 62 trips, cluster 2 = 60 trips, cluster 3 = 55 trips)

Fig. 3

Visual comparison of the dive variables (a–e) between the three identified clusters (Cl.1, Cl.2, Cl.3) from the hierarchical cluster analysis through violin plots (cluster 1: n = 62, cluster 2: n = 60, cluster 3: n = 55)

Fig. 4

Group summary of pelagic divers’ HMM results, comparing the three states by dive variables (a–d), their time distribution over 24 h (f) (12 animals with 31 foraging trips), as well as an example of the spatial distribution of states for one pelagic diver (e) (id1719)

Fig. 5

Group summary of benthic divers’ HMM results, comparing the two states by dive variables (a–d), their time distribution over 24 h (f) (12 animals with 57 foraging trips), as well as an example of the spatial distribution of states for one benthic diver (e) (id127€)

Fig. 6

Group summary of night divers’ HMM results, comparing the three states by dive variables (a–d), their time distribution over 24 h (f) (7 animals with 33 foraging trips), as well as an example of the spatial distribution of states for one night diver (e) (id1638)

Fig. 7

Comparison between benthic, pelagic, and night divers with regard to their distribution of dives with different depths over 24 h (left) and their spatial distribution of dives classified as foraging based on the HHMs (benthic divers (blue): n = 12, pelagic divers (green): n = 12, night divers (red): n = 10) on a bathymetric map (black = 300 m, white = 0 m, lines at 10 m intervals). In the right lower corner, a cutout of the bathymetric map is presented to better visualize shallow areas