In vertebrates, patterns of resource utilization change throughout development according to age- and or size-specific abilities and requirements. Thus, interspecific competition affects different age classes differently.
Adults of sympatric species often show distinct foraging niche segregation, but juvenile resource use might overlap with adult competitors of similar body size. Resultant negative effects on juveniles can have important consequences for population dynamics, yet such interactions have received little attention in studies of mammalian communities.
Using GPS tracking devices, time-depth recorders and stable isotope data, we compared diving depth, activity time, trophic position and foraging habitat characteristics to investigate foraging niche overlap between similar-sized sympatric juvenile Galapagos sea lions (Zalophus wollebaeki) and adult Galapagos fur seals (Arctocephalus galapagoensis) and compared each group with much larger-bodied adult Galapagos sea lions.
We found little indication for direct competition but a complex pattern of foraging niche segregation: juvenile sea lions and adult fur seals dived to shallow depths at night, but foraged in different habitats with limited spatial overlap. Conversely, juvenile and adult sea lions employed different foraging patterns, but their foraging areas overlapped almost completely.
Consistency of foraging habitat characteristics between juvenile and adult sea lions suggests that avoidance of competition may be important in shaping foraging habitat utilization. Resultant specialization on a limited habitat could contribute to low sea lion numbers that contrast with high fur seal abundance. Our data suggest that exploitation by multiple predators within spatially restricted foraging ranges of juveniles might negatively impact juvenile foraging success and ultimately influence population dynamics.
Competition for food resources is a crucial factor in shaping the coexistence of species and the structure of ecological communities (Begon, Townsend & Harper 2006). Competition influences the energy budget of foragers by, for example, reducing food intake rate (Shealer & Burger 1993; Rutten et al. 2010) and can negatively affect growth, reproduction and survival of individuals (Cody & Diamond 1975; Eccard & Ylönen 2003). At the population level, competition can reduce the abundance of the less competitive species (Kennedy & Strange 1986; Creel & Creel 1996; Wauters, Lurz & Gurnell 2000) or even lead to displacement through competitive exclusion (Hardin 1960).
Ecologically similar species can coexist when foragers partition the available resources and specialize within their foraging niches. Such niche separation can involve some or all niche dimensions and can be achieved by consuming different types of food, foraging in different areas or at different times (Schoener 1968; Abbott, Abbott & Grant 1977; Alatalo et al. 1987; Barlow et al. 2002; Wilson 2010).
Morphometric differences and body size are often indicators of resource partitioning because they are connected to specific foraging abilities and foraging strategies (Grant & Grant 1982; Basset & Angelis 2007; Alexandrou et al. 2011). For example, the size of morphological structures such as beaks or teeth can constrain effective handling of specific types of prey (Boag & Grant 1984; Aguirre et al. 2003; La Croix et al. 2011). Body size can also set physiological limits to a foraging strategy. For example, diving depth and duration are limited by the allometric relationship between body size and oxygen storage capacity (Kooyman 1989; Schreer, Kovacs & O'Hara Hines 2001; Mori 2002; Weise, Harvey & Costa 2010), which restricts the exploitable three-dimensional foraging habitat. Therefore, sympatric species of similar body size with similar morphological and physiological abilities are likely to have overlapping foraging niches leading to competition for resources.
A large body of literature exists on foraging niche separation and resource partitioning between adult foragers of sympatric species (Cody & Diamond 1975; Schoener 1983; Begon, Townsend & Harper 2006). However, competition and patterns of foraging niche separation between different developmental stages of sympatric species have received considerably less attention (Polis 1984), especially in mammalian systems. In size-structured species, ontogenetic niche shifts lead to differences in resource use of juvenile and adult stages (Werner & Gilliam 1984; Polis 1984; Wikelski, Gall & Trillmich 1993; Wikelski & Trillmich 1994; Adams 1996; Subalusky, Fitzgerald & Smith 2009). This reduces intraspecific competition, but increases the potential for interspecific competition between juveniles of large-bodied species and adults of small-bodied species (Polis 1984). At the onset of independent foraging, juvenile animals are inefficient foragers, constrained by incomplete physiological and morphological maturation and lack of experience (Recher & Recher 1969; Goss-Custard & Dit Durell 1987; Weathers & Sullivan 1989, 1991; Sol et al. 1998; Daunt et al. 2007; Thornton 2008). Therefore, juvenile animals are often poor competitors, which can lead to impaired health, reduced growth, delayed maturation and reproduction and, in extreme cases, reduced juvenile survival (Sol et al. 1998). Stage-specific foraging niche overlap can therefore negatively affect the demography of the larger species by reducing juvenile recruitment.
Sympatric sea lion and fur seal species are an excellent model system for the investigation of interspecific foraging competition between adults and juveniles because juvenile sea lions and adult fur seals have a similar body size, while adult sea lions are about 2–3 times larger (Riedmann 1990). Sympatric Galapagos sea lions (Zalophus wollebaeki) and Galapagos fur seals (Arctocephalus galapagoensis) provide a particularly suitable example because sea lion juveniles are constrained by slow growth and variable environmental conditions (Mueller et al. 2011) and exhibit one of the slowest developmental rates to independent foraging among pinnipeds (Jeglinski et al. 2012). Any further obstacle to their foraging success is therefore likely to have a severely negative effect. Despite overall similar population numbers, 95% of the Galapagos fur seal population is found around the western islands Fernandina and Isabela, whereas sea lion colonies occur throughout the whole archipelago with only 5% of the population found in the west (Fig. 1). Galapagos fur seals outnumber Galapagos sea lions in sympatric breeding colonies approximately fivefold (J. W. E. Jeglinski, unpublished data), and fur seal population numbers seem to be increasing, whereas sea lion numbers seem to be declining (F. Trillmich, unpublished data), a showcase of the global situation of sympatric fur seal sea lion populations (Arnould & Costa 2006). The little information available indicates dietary niche segregation and differences in foraging strategy between sympatric adult Galapagos sea lions and adult fur seals (Kooyman & Trillmich 1986a,b; Horning & Trillmich 1997; Dellinger & Trillmich 1999; Paéz-Rosas et al. 2012), similar to a distinct niche segregation between adults of other sympatric sea lion and fur seal species (Antonelis, Stewart & Perryman 1990; Goldsworthy & Page 2007; Franco-Trecu et al. 2012; Waite et al. 2012). This precludes direct competition between adults as a possible cause for differences in population numbers. Despite the obvious similarity in body size, no study to date has investigated foraging competition between juvenile sea lions and adult fur seals.
Hence, the aim of this study was to investigate stage-specific foraging niche overlap between Galapagos sea lions and Galapagos fur seals. Foraging niches are ‘n-dimensional hypervolumes’ (Hutchinson 1957) which are inherently complex and difficult to measure (Newsome et al. 2007), especially in cryptically foraging marine mammals. Recent technological advances in ecological methodology aid attempts to address this complexity: biologging tools such as time-depth recorder (TDRs) and GPS tracking devices can sample detailed data on foraging behaviour and have been successfully used to investigate overlap in temporal activity and in both horizontal and vertical space use (e.g. Masello et al. 2010; Wilson 2010). Stable isotope analysis is a powerful tool to delineate the ‘isotopic niche’ of animals by making use of the fact that consumer tissues integrate information on resource use and foraging habitat (Bearhop et al. 2004; Newsome et al. 2007). While nitrogen isotope ratios increase in a predictable way along trophic chains, allowing comparisons of consumer trophic position (e.g. DeNiro & Epstein 1981; Minagawa & Wada 1984; Post 2002; Newsome, Clementz & Koch 2010), carbon isotope ratios change minimally with trophic position but mirror baseline ecosystem signatures and provide information on foraging strategies and locations (DeNiro & Epstein 1978; Hobson 1999; Kelly 2000). Further, variation in stable isotope signatures can be an effective indicator of trophic niche width as such variation reflects the range of isotopically different prey items that are consumed and of isotopically distinct foraging locations that are visited (Bearhop et al. 2004; Newsome et al. 2007).
Here, we combine data from GPS tracking devices and TDRs with stable isotope data to investigate the overlap of spatial, temporal and dietary niche dimensions. We predict: (i) that juvenile sea lions and adult fur seals use similar foraging strategies and foraging areas and thus exhibit foraging niche overlap. We hypothesize that this indication of foraging competition could disadvantage juvenile sea lions and might ultimately help to explain the large disparity in population sizes of sympatric sea lions and fur seals. In line with previous findings, we expect to find differences in foraging strategies and foraging areas indicative of; (ii) intraspecific foraging niche separation between juvenile and adult sea lions; and (iii) interspecific foraging niche separation between adults of both species.
Materials and methods
This study was carried out in the breeding season 2009 (2nd of October – 5th of November 2009) at Cabo Douglas, Fernandina Island, Galapagos (0·18° S, 91°39 W). Galapagos sea lions and fur seals breed sympatrically in this colony, and average census numbers were 42 ± 11 and 215 ± 25 animals for sea lions and fur seals, respectively (mean ± SD; n = 3, censuses performed at the onset, peak and end of the breeding season from October–December 2009).
To establish a cohort of known age sea lion juveniles, 32 pups (< 2 months) were marked by clipping numbers into the fur of their lower back and by inserting passive transponders (Trovan; Euro ID Identification Systems GmbH, Weilerswist, Germany) subcutaneously above their shoulder blades in October 2008. In March 2009, 25 of these animals were recaptured and tagged on the trailing edge of both front flippers with numbered blue plastic tags (tag size 0; Allflex Europe Ltd., Hawick, UK).
We deployed TDRs equipped with Fastloc GPS to study the diving behaviour and spatial distribution of sea lions and fur seals. GPS–TDRs were deployed on seven 12-month-old juvenile sea lions, 11 adult fur seal females and ten adult sea lion females. Adult females of both species were suckling young pups (< 2 weeks). We chose 12-month-old juveniles at the onset of independent foraging and adult females with young pups, because these age classes would potentially be most affected by foraging competition.
All animals were captured with hoop nets with a small muzzle opening to allow the animals to breathe freely and were manually restrained. Adult animals were marked with passive transponders and flipper tags (adult sea lions) or individual fur clippings (adult fur seals). Body mass was recorded using a digital scale hanging from a tripod (Kern HUS 300 K 100; Kern & Son GmbH, Balingen Germany; precision 0·2 kg, max 100 kg), and morphometric measurements of standard length (SL), fore-flipper length and width (FL, FW) and axillary girth were taken (see details in Jeglinski et al. 2010, 2012).
Instrument sets of TDR–GPS (MK10 or MK9; Wildlife Computers, WA, Redmond, USA, and Sirtrack Fastloc GPS; Sirtrack, New Havelock, New Zealand) and VHF radiotransmitters were attached to the dorsal fur on the midline of the back of the study animals just below the shoulder blades using fast setting epoxy (Araldite 2012; Huntsman Advanced Materials, Basel, Switzerland). TDRs were programmed to sample time and depth every 2 s, with a depth resolution ± accuracy of 0·5 m ± 1%. Fastloc GPS was set to acquire a position every 15 min. Mean GPS error has been estimated to 36 m (Costa et al. 2010). Sets of instruments (TDR, GPS and VHF) weighed between 0·3% and 0·7% of the animals' body mass. For TDR and VHF retrieval, animals were recaptured after about 2 weeks and instruments were removed. One fur seal and one sea lion female could not be recaptured, one adult sea lion record was lost due to instrument failure, and one juvenile sea lion was only recaptured in December 2009. This long record was cropped to end on the latest recapture date of the regular field period (5th of November 2009) to be comparable with the other deployments. All study animals were monitored in a subsequent field season and showed no physical impact or behavioural abnormalities as a consequence of instrument deployment.
Stable isotope sampling and analysis
During TDR–GPS deployment, skin samples from front or back flippers of the study animals were collected with a sterile leather punch, air-dried and stored in sterile plastic vials. Pinniped skin provides a short-term integration of dietary isotopes (2 months; Kurle & Worthy 2002). We extracted lipids from skin samples following Dobush, Ankney & Krementz (1985) and Kurle & Worthy (2002) (details in Jeglinski et al. 2012). Isotope ratios were measured with an Isotope Ratio Mass Spectrometer (Isoprime, Elementar, Hanau, Germany). Carbon and nitrogen isotope ratios are reported relative to VPDBee and atmospheric air, respectively, and expressed as δ-notation in [‰]:
where Rsample and Rstandard are the ratios of the heavier to the lighter isotope of the sample and the standard, respectively, and X is the element (C or N) considered. Samples were standardized to IAEA-N2, IAEA-CH-6 (International Atomic Energy Agency, Vienna, Austria) and an IVA protein standard. Repeated measurement precision was 0·12 ‰ and 0·14 ‰ for δ13C and δ15N, respectively.
Dive and spatial analysis
Zero offset correction and dive analysis were performed with a purpose written MATLAB toolbox (IKNOS, Y. Tremblay, unpublished data), and dive summary files were created using a minimum depth of 5 m and minimum duration of 12 s. Travelling dives were not discarded from the dive analysis because there was no objective criterion to distinguish between shallow foraging dives and shallow travelling dives. Trips to sea (here defined as exceeding 45 min wet) were determined based on wet/dry sensor data of TDRs using a custom written MATLAB function (P. W. Robinson, unpublished data). GPS positions were decoded using the DAP processor (Wildlife Computers) and filtered using a speed filter (3 km h−1 for juveniles; 6 km h−1 for adults) to remove erroneous points (Tremblay et al. 2006; Kuhn et al. 2010). Filtering retained c. 80% of GPS positions for juvenile and adult sea lions and adult fur seals, respectively. GPS data were split into separate trips based on wet dry data and interpolated using a hermite spline (Tremblay et al. 2006). A land avoidance algorithm was applied to interpolated tracks to adjust positions that were on land to nearby water positions (Götz et al. 2012). Each dive was associated with a GPS location using a time-based linear interpolation between track points. For subsequent analyses, data were converted into the Universal Transverse Mercator (UTM) coordinate system.
Spatial analyses were performed using ArcMap (ArcGIS Desktop, release 9·2 and 10; Environmental Systems Research Institute, Redlands, CA, USA). Distance of dive locations to the coast was calculated from a detailed coastline shape file of the Galapagos Islands, created from a high-resolution Blue Marble image (500 m pixel−1, Blue Marble Next Generation, Stöckli, R., NASA Earth Observatory, downloaded on 2nd July 2011). Bathymetrical depth and slope at each dive location were calculated using a fine-scale bathymetrical map based on multibeam and side-scan sonar data sampled by Geist et al. (2006) (resolution of 250 m/grid cell, vertical accuracy 10–20 m, accessed through the Marine Geoscience Data System (MGDS), Lamont–Doherty Earth Observatory, Columbia University, NY, USA, assembled using GeoMappApp 3·0·0). An index of benthic distance was determined as a modification of Simmons et al. (2007), where each dive to within 20 m of the sea floor was defined as benthic and all other dives as pelagic. The buffer zone of 20 m to the sea floor was added to account for measurement error of bathymetry and GPS locations.
Bearings of each dive location to the colony were determined, Watson's test of uniformity (R package circular) of mean individual bearings was used to test the directionality of foraging locations within groups, and Watson's two-sample test was used to test for differences in orientation between groups. To identify the prominent haul-out locations, we created 250-m buffers around each haul-out point (start and end locations of trips). Overlapping buffer areas were merged and assigned an individual haul-out identifier, and the frequency of utilization of each haul-out location was calculated per group.
Size and overlap of foraging ranges were determined by constructing utilization distributions (UDs) from dive positions of juvenile and adult sea lions and adult fur seals, standardized for the number of animals in each group. We calculated fixed Gaussian kernel density estimates (R package adehabitatHR), based on 200-m2 grid cells and a scaled reference smoothing factor (0·6× reference bandwidth) to prevent over-smoothing (Kie et al. 2010). We used a land mask to omit foraging range area on land. In accordance with most studies, 90% and 50% isopleths were chosen to represent the extent of the foraging area and the core foraging range, respectively (Kernohan, Gitzen & Millspaugh 2001), and the percentage of overlap was calculated separately for overall and core foraging ranges.
Statistical comparisons were performed using R. 2.15.0 (R development Core Team 2010) and spss 17.0. We used t-tests to compare variables representing different niche dimensions between the three groups and corrected for multiple testing using a Bonferroni correction. Data were checked for normality visually and using the Shapiro–Wilk test and for homogeneity of variance using the Fligner–Kileen test. Non-normally distributed data were log-transformed, and proportions were arcsine-transformed. Nonparametric tests were used when transformation did not result in a normal distribution. A stepwise canonical discriminant function analysis was used to investigate whether the three groups could be separated using variables representing different niche dimensions and to identify which variables most effectively separated the three groups. Each function represented a linear combination of the predictor variables utilizing only the combination of variables that results in the smallest Wilk's lambda value, thereby maximizing the between-group differences. Correct group classification was verified using leave-one-out cross validation. Five of the variables representing niche dimensions were used to account for the relatively small sample sizes within groups (Table 1). The significance level was set at P <0·05, and data are reported as mean ± SE if not specified otherwise.
Table 1. Structure matrix showing standardized discriminant function coefficients and pooled within-group correlations between discriminating variable and standardized canonical discriminant functions. The discriminating variable in italics was omitted during the stepwise procedure because it did not contribute to increase the distance between groups
Distance to coast (km)
Mean slope inclination (%)
Time spent diving at night (%)
Mean diving depth (m)
Instruments were deployed for 12·7 ± 0·7 days yielding records for seven juvenile sea lions, 10 adult fur seals and eight adult sea lions. Individual records consisted on average of 1269 ± 186 dives and 264 ± 42 GPS positions (Table 2).
Table 2. Summary statistics of variables representing different niche dimensions and deployment details of juvenile sea lions, adult fur seals and adult sea lions sampled in autumn 2009 in the Cabo Douglas colony, Fernandina
Variables representing niche dimension
Juvenile sea lions (n = 7)
Adult fur seals (n = 10)
Adult sea lions (n = 8)
Values are given as mean ± SE. Significant test results of group comparisons are indicated by numbers (1significantly different from juvenile sea lions, 2significantly different from adult fur seals and 3significantly different from adult sea lions), significance level was set to P <0·05. Test statistics are given in the supplementary material Table S1. Bottom time was calculated as the duration spent between 80% and 100% of the maximum depth of each dive.
Morphometric characteristics as indicators of foraging niche separation
Juvenile sea lions and adult fur seals were very similar in body size: both groups weighed c. 30 kg and were on average 104 cm long with front flippers that were 21 cm long (Table 2, Appendix Table S1 for details on statistical analysis of all further comparisons). In contrast, adult sea lions were twice as heavy and long and had longer front flippers (Table 2).
Vertical niche dimension: overlap of foraging depth and foraging strategy?
As morphological similarities suggested, juvenile sea lions and adult fur seals utilized the vertical space underwater in a very similar way, diving to shallow depth during both day and night (Fig. 2, Table 2). In contrast, adult sea lions used a different part of the water column during the day and dived on average three times deeper than juvenile sea lions and adult fur seals. These differences were not consistent at night, when mean depth of all three groups overlapped considerably (Fig. 2, Table 2). Fur seals dived pelagically only, whereas sea lions employed a mixed strategy with c. 19% and 25% of dives classified as benthic (juvenile and adult sea lions, respectively, Table 2). Maximum dive depth and duration indicated major differences in physiological abilities between all three groups: juvenile sea lions dived deeper and for longer than adult fur seals (Table 2); the deepest and longest dive was 234 m and 8·1 min, in contrast to 97 m and 5·4 min recorded for adult fur seals. Adult sea lion depth maxima exceeded juvenile sea lion and adult fur seals depth maxima two- and fourfold, respectively, reaching a depth of 584 m, the deepest recorded for this species (Table 2).
Temporal niche dimension: overlap in foraging time?
Marine foragers that use similar areas within the water column can minimize competition by adopting different temporal patterns of foraging activity. However, both juvenile sea lions and adult fur seals dived predominantly at night (Fig. 3, Table 2). Diving activity of both groups peaked at similar times, in the first and last hours of the night after dusk and before dawn (Fig. 3 18:00–21:00 and 02:00–4:00 h). Adult fur seals were exclusively nocturnal, while juvenile sea lions exhibited a small proportion of diving activity during the day (13·5% of the time spent diving, Fig. 3, Table 2). Both groups can be characterized as shallow diving, night active foragers that overlap considerably in these two niche dimensions. In contrast, adult sea lions showed no clear diel pattern: they dived evenly at day and night (Fig. 3, Table 2), and their foraging activity peaked at dusk (17:00–19:00 h) and dawn (4:00–7:00 h), earlier and later than that of both other groups.
Dietary niche dimension: overlap of trophic position?
Similarities in temporal distribution of foraging activity and vertical distribution in the water column suggest that juvenile sea lions and adult fur seals might exploit similar types of prey, which would be indicated by similar stable nitrogen isotope values. However, δ15N of juvenile sea lions were enriched above adult fur seal values by 1·2‰, suggesting a difference in trophic position (Fig. 4, Table 2). All but one juvenile sea lion were observed suckling during our study, suggesting that juvenile stable nitrogen signatures were at least partially derived from maternal milk. Juvenile sea lion δ15N were depleted by 1·5‰ compared with adult sea lions (Fig. 4, Table 2). As feeding on maternal milk results in enriched offspring δ15N values (Jenkins et al. 2001), this supports the idea that juvenile sea lions contributed to their diet by independently acquiring prey of a low trophic level. Stable nitrogen isotope ratios of adult sea lions were enriched by 2·7‰ above adult fur seal values, clearly segregating adult sea lions and fur seals into different trophic niches (Fig. 4, Table 2). Variances in δ15N as a measure of trophic niche width did not differ between any of the three groups (F-test, all P >0·05).
Horizontal niche dimension: foraging trips, haul-out locations and characteristics of foraging habitat
The spatial distribution of foraging movements of juvenile sea lions was very different from that of adult fur seals: Juvenile sea lions performed short trips of 5·1 ± 0·4 h in close vicinity to the study colony, ranging on average within a radius of 2·0 ± 0·7 km (maximum straight line distance of 14·9 km, Fig. 6, Table 2). Dive locations were distributed in a predominately easterly direction from the colony (Watson's test, U² = 0·30. P <0·01, Fig. 6). Juvenile sea lions used 8 different haul-out locations with only 33% of haul outs in Cabo Douglas where they were originally captured (Fig. 6). In contrast, adult fur seals performed long and far-ranging trips (18·8 ± 0·5 h, average straight line distance of 20·6 ± 2·6 km, maximum 68·9 km), heading exclusively towards the west and north-west (Watson's test, U² = 0·70, P <0·01, Fig. 6, Table 2). Adult fur seals used only 4 different haul-out locations and hauled out predominantly at the study colony (94·6%). Adult sea lion trips were short (8·5 ± 0·5 h), but adult sea lions ranged further from the colony than juvenile sea lions (mean straight line distance: 12·6 ± 3·9 km, maximum: 50·5 km). Foraging movements of adult sea lions were mostly directed to the east (Watson's test, U² = 0·19, P <0·05), similar to those of juvenile sea lions (Fig. 6, Table 2). Adult sea lions used 32 different haul-out locations along the shore of Fernandina and Isabela Island, and only 55·8% of these haul outs were located in the Cabo Douglas colony (Fig. 6).
Foraging habitat characteristics of juvenile and adult sea lions were very similar: juvenile and adult sea lions foraged in the immediate vicinity of the coast, spending 100% and 93·3% of the time diving within 5 km to the coast (Fig. 5). Both groups dived mostly on the shelf and at its edge at locations of steep inclinations of the sea floor. (Fig. 6, Table 2). In contrast, fur seals foraged offshore and in deep waters at locations with an even sea floor, spending 89·3% of the time diving beyond a distance of 5 km and performing 90·2% of their dives in water deeper than 2000 m (Figs. 5 and 6, Table 2). Unexpectedly, stable carbon isotope signatures did not reflect the different distances of foraging locations to the coast (all comparisons, P >0·05). Strong upwelling near the coast of Fernandina might confound the distinction of coastal vs. offshore δ¹³C signatures of the prey. However, variances of adult sea lion δ¹³C values were larger than those of juvenile sea lions (F₇,₆= 18·18, P <0·01) and adult fur seals (F₇,₉ = 7·14, P <0·01), indicating a higher variability in foraging locations or foraging mode, which is well supported by the broader range of their diving depth (Fig. 2) and the spatial distribution of their dive locations (Fig. 6).
Foraging range size and overlap
Foraging ranges of juvenile sea lions were small, equivalent to small proportions of the area that adult females of both species used (Fig. 6, 5% and 15% of adult fur seal and adult sea lion foraging ranges, respectively). Adult fur seals had the largest foraging range, exceeding that of adult sea lions threefold (Fig. 6). Juvenile sea lion foraging ranges overlapped partially with areas used by adult fur seals (Fig. 6, 35·9% of 90% contour and 21·8% of 50% contour) and almost completely with adult sea lion foraging ranges (Fig. 6, 99·9% of the 90% contour and 97·8% of the 50% contour). For adults, the overlap with juvenile sea lions represented a small fraction of their respective foraging ranges (1·8% and 0·67% for adult fur seals and 13·8% and 12·7% for adult sea lions, 90% and 50% contour, respectively) and was located in close vicinity of the colony. Similarly, the overlap between adults was small in comparison with the extent of their respective foraging ranges and was restricted to the immediate area around the colony (Fig. 6, 4·7% and 2·2% for adult fur seals and 12·8% and 9·1% for adult sea lions, 90% and 50% contour, respectively). In all cases, overlapping foraging areas close to the colony encompassed areas of high foraging intensity (50% kernel density estimate, Fig. 6).
Intra- and interspecific foraging niche separation
A canonical discriminant function analysis was conducted to investigate whether foraging niches of juvenile and adult sea lions and adult fur seals could be differentiated based on predictor variables that represent different niche dimensions. Mean diving depth did not increase the distance between groups and was excluded during the stepwise process of creating the discriminating functions. Discriminant function analysis confirmed the niche segregation of juvenile sea lions, adult fur seals and adult sea lions based on two significant canonical functions determined by the remaining four variables (Table 1, Wilk's lambda = 0·008, P <0·01 and Wilk's lambda = 0·229, P <0·01 for function I and II, respectively). Juvenile and adult sea lions were clearly distinguished from fur seals based on distance to coast and slope inclination, indicating pronounced species-specific foraging habitat segregation (function I, explaining 88·9% of the variance; Table 1, Fig. 7). Further, juvenile and adult sea lions segregated into two distinct groups based on δ15N, distance to coast and proportion of time spent diving at night, indicating an ontogenetic shift in foraging strategy and location (function II, explaining 11·1% of the variation; Table 1, Fig. 7). Leave-one-out cross validation assigned 92% of animals correctly to their respective groups. One adult sea lion was misclassified as juvenile sea lion, while all fur seals were correctly assigned, stressing the high degree of segregation between the two species.
We studied the foraging niche overlap of sympatric predator species, where juveniles of the larger species are morphologically similar to adults of the smaller species and are therefore likely to compete for similar resources. Combining stable isotope and biologging data, we have demonstrated a complex system of niche segregation. We found similarities in morphology and foraging strategy but, in contrast to our prediction, limited overlap in foraging habitat, which likely reduces direct competition between juvenile sea lions and adult fur seals.
Niche differentiation between juvenile sea lions and adult fur seals
Despite similarities in several niche dimensions (diving depth, activity period), juvenile sea lions and adult fur seals exploited spatially distinct habitats with very different characteristics (Figs. 6 and 7). Given the available data, trophic niche overlap for juvenile sea lions and adult fur seals was difficult to decipher. Depleted stable nitrogen isotope values of milk-consuming juveniles compared with adult sea lions suggest that juveniles fed to a considerable extent on prey of low trophic level, similar to adult fur seals (Clarke & Trillmich 1980; Dellinger & Trillmich 1999). Unfortunately, snapshot dietary information from scats, regurgitations or stomach contents could not be sampled during this study, and there are no published data on juvenile diet to clarify dietary niche overlap. Stable carbon isotope ratios of juvenile sea lions and adult fur seals were very similar and did not reflect differences in foraging location (coastal vs. offshore) that our GPS data revealed. These findings stress the importance of integrating isotopic data of consumers within the ecological context of their prey and their foraging habitat when studying isotopic niches (Bearhop et al. 2004; Newsome et al. 2007), but also highlight the importance of using additional methods such as biologging.
GPS tracking data showed that juvenile sea lions and adult fur seals clearly differed in the horizontal distribution of their foraging locations, which reduced overall foraging niche overlap. Schoener (1968) describes a similar pattern of reduced niche overlap between different age classes of Anolis lizard species with similar body size, which forage at different heights within a vertically structured habitat. In contrast, high interspecific niche overlap between grey and red squirrels in Britain (although foraging niche overlap between juvenile and adults has not been directly investigated) seems to have a stronger impact on juvenile animals, leading to reduced juvenile recruitment and a reduction in red squirrel densities (Wauters, Lurz & Gurnell 2000; Gurnell et al. 2004). Even subtle differences along a number of niche axes may lead to ‘apparent separation’ between coexisting members of an ecological community (Wilson 2010). Indeed, cases of obvious niche overlap and consequently high competition between groups might be most evident in recently perturbed ecological communities into which a novel competitor has been introduced (Wauters, Lurz & Gurnell 2000; Gurnell et al. 2004; Juliano, Lounibos & O'Meara 2004) or in which changes in the density of a competing species have occurred (Gustafsson 1987). In established ecological communities, even morphologically similar species or age groups appear to adjust aspects of their foraging niche within their physiological and morphological limits (Schoener 1968; Wilson 2010) to avoid effects of direct competition or else they could not co-occur due to competitive exclusion (Hardin 1960).
In our example, the small body size of juvenile sea lions and resultant limited oxygen storage capacity likely constrain the exploitable underwater space and the duration of underwater activity (Kooyman 1989; Horning & Trillmich 1997; Schreer, Kovacs & O'Hara Hines 2001). The use of the horizontal space is less directly affected by such physiological limitations, and juvenile sea lions are able to adjust to the presence of a competing forager by segregating in the horizontal niche dimension. Due to the unique positioning of our study colony, similar travelling distances to either deep water or shelf habitat along the coast enabled juvenile sea lions to reach either habitat with approximately similar travelling effort. This stands in an interesting contrast with other sympatric sea lion and fur seal species, where extensive shelf areas around the study colony likely constrain the habitat choice of juvenile sea lions (Goldsworthy & Page 2007; Franco-Trecu et al. 2012; Waite et al. 2012). Juvenile foraging habitat selection in our study is not confounded by differential habitat accessibility. Therefore, competition avoidance could be a factor responsible for juvenile foraging habitat choice.
Foraging habitat overlap between juvenile and adult sea lions
As expected, foraging niches of juvenile and adult sea lions differed in almost all investigated dimensions, and discriminant function analysis effectively separated juvenile and adult sea lions based on trophic position, foraging location and foraging activity time. This ontogenetic niche shift (Werner & Gilliam 1984; Jeglinski et al. 2012) likely reflects age-specific differences in physiological and morphological foraging abilities (Wikelsky, Gall & Trillmich 1993; Adams 1996; Subalusky, Fitzgerald & Smith 2009) and a resulting difference in the availability and effective exploitability of specific resources. However, while adult sea lions foraged at a broader range of locations, an aspect also indicated by the larger variance of stable carbon isotope signatures, overall foraging habitat characteristics were very similar to those of juveniles. Given that adult females appear not to show juveniles where to forage (Fowler, Costa & Arnould 2007), foraging habitat selection developed at an early age may be maintained in adulthood, raising interesting questions about the mechanisms shaping habitat preferences (Davis & Stamps 2004) and the role of competing species.
Nonetheless, competition might occur despite apparent ontogenetic niche shifts when juveniles and adults target different age groups of the same prey or prey items that are mobile along niche axis (Wilson 2010). This calls for cautious interpretation of segregation along niche axis as complete absence of competition and resultant negative effects.
Adult fur seals and sea lions show distinct foraging niche segregation
Adult sea lions differed from fur seals along most niche axes, in line with our expectations and with previous evidence (Kooyman & Trillmich 1986a,b; Dellinger & Trillmich 1999; Merlen 2000; Paéz-Rosas et al. 2012). Shallow diving activity at night suggested similarities in fur seal foraging behaviour, but stable nitrogen isotope data taken together with published dietary records (Clarke & Trillmich 1980; Dellinger & Trillmich 1999) suggest that adults of both species are targeting different prey. Sea lions fed primarily on small pelagic and bathydemersal schooling fish of the families Engraulidae and Chlorophtalmidae, while fur seals fed on organisms of the deep scattering layer, mainly myctophids and squid (Clarke & Trillmich 1980; Dellinger & Trillmich 1999). Ideally, a comparison of the isotopic niches of consumers would be supported by a more detailed analysis including isotope values of primary prey, that is, by using mixing models (Phillipps & Gregg 2001; Bearhop et al. 2004; Newsome et al. 2007). While isotopic data from primary prey are not available from the study site (but see Paéz-Rosas et al. 2012 for a summary of possible prey, sampled in the whole archipelago and elsewhere), the consistency between stable nitrogen and carbon isotope data with the distribution of foraging locations and published records of diet supports our interpretation that Galapagos sea lions and fur seals employ distinct foraging strategies and present a clear-cut case of foraging niche segregation.
Our findings support evidence of niche segregation between adults of different sympatric sea lion and fur seal species (Goldsworthy & Page 2007; Franco-Trecu et al. 2012; Waite et al. 2012). Here, we broadened the approach of previous studies by including multiple niche dimensions and extended the scope of previous work by demonstrating that foraging habitat segregation, possibly induced by competition avoidance, is already evident at the onset of independent foraging.
Whether competition avoidance plays an important role in shaping habitat preferences can be tested by comparing the results of our study with the behaviour of sea lions that are allopatric to fur seals in the centre of the archipelago. The removal of a possible competitor generally leads to an increase in niche width of the remaining species (Alatalo et al. 1987), so that larger variances of stable isotope signatures and a higher variability in foraging habitat could be expected. Such questions are of particular ecological and evolutionary interest in Galapagos sea lions, because the western population constitutes a genetically and morphologically distinct cluster (Wolf et al. 2008), and it is not clear how genetic differentiation can develop in such a mobile species. Our results suggest that habitat specialization, possibly induced by adaptations to a sympatric competitor at an early age, might be a mechanism. A comparison of foraging strategies and habitat preferences of juvenile and adult sea lions of the western and central cluster could therefore contribute to our understanding of the role of a competitor in shaping foraging habitat choice and the emergence of limited gene flow between ecologically different groups of a species.
Foraging on the brink: are western Galapagos sea lions constrained to a spatially limited habitat?
Overall, the complex pattern of foraging niche segregation suggests that direct competition between similar-sized juvenile sea lions and adult fur seals or between adults of both species is not responsible for the small population size of western Galapagos sea lions. Equally, competition for terrestrial space in breeding colonies seems unlikely: both species breed in low-density colonies, and there is an extensive unoccupied coastal habitat available. Further, Galapagos fur seals and sea lions display a certain degree of fine-scale breeding habitat segregation within colonies, with sea lions hauling out preferably on sandy beaches and fur seals concentrating on rocky habitat (J. W. E. Jeglinski, personal observation).
As an alternative, habitat specialization and differences in the availability of suitable habitat might contribute to the disparity in population numbers: in our study area, suitable sea lion foraging habitat covered a very limited area compared with the extensive epipelagic habitat of Galapagos fur seals (Fig. 1), which was also demonstrated by the difference in size and shape of sea lion and fur seal foraging ranges (Fig. 6). Accordingly, differences in the carrying capacity of sea lion and fur seal habitat might be an important factor creating the observed differences in population size, as has been suggested for other fur seal–sea lion systems elsewhere (Costa & Gales 2003).
Overlap of foraging ranges: Indirect effects of exploitative competition
Direct competition is not the only ecological interaction that can have asymmetric effects on different groups within an ecological community. Barlow et al. (2002) suggest that adult Antarctic fur seals (Arctocephalus gazella) that are able to perform long-distance foraging trips have a competitive advantage over much smaller sympatric macaroni penguins (Eudyptes chrysolophus), which might explain decreasing population size of penguins, whereas fur seal numbers are increasing. In a similar fashion, Ainley, Ballard & Dugger (2006) conclude that communal exploitation of Adélie penguins (Pygoscelis adeliae), minke whales (Balaenoptera acutorostrata) and orcas (Orcinus orca) altered the availability of krill in the form of trophic cascades, in this case leading to increased foraging trip durations and changes in penguin diet. In central place foragers such as seabirds, increased foraging pressure of multiple foragers close to the colony leads, over time, to the formation of a ‘Storer – Ashmole's halo’ around the colony, where either prey is depleted or it has migrated to inaccessible depths to avoid predation (Storer 1952; Ashmole 1963; Birt et al. 1987; Lewis et al. 2001; Elliott et al. 2009). This likely occurs in a similar way in central place foraging pinnipeds such as sea lions and fur seals. Large foraging ranges of adult sea lions and fur seals indicated that the supply of appropriate prey close to the colony was not sufficient and that long-distance trips were necessary to increase foraging efficiency (Lewis et al. 2001; Elliott et al. 2009). In contrast, the restricted juvenile foraging area suggested that such adjustments were not possible (Fig. 6), which is likely a result of energetic constraints linked to slow swimming speeds typical for juvenile marine mammals (Horning & Trillmich 1997; Noren, Biedenbach & Edwards 2006; Fowler, Costa & Arnould 2007). These constraints implied that juvenile foraging ranges overlapped extensively with those of adult fur seals and sea lions, while the area of overlap was proportionally small for adult foragers. By directly depleting the same food resource or interfering through their presence with the juveniles' ability to obtain prey, additional foragers are likely to decrease the quality of juvenile foraging habitat and diminish juvenile foraging success (Goss-Custard & Dit Durell 1987; Sol et al. 1998). These negative effects on juveniles might contribute to low sea lion population numbers in the western Galapagos archipelago. Similar effects likely apply to juvenile foragers that share foraging locations with multiple groups of foraging species and can be expected to increase as the complexity of such foraging communities increases. Our findings support calls for refining comparisons below the species level in analyses of niche overlap in ecological communities (Polis 1984) and for a cautious interpretation of apparent absence of competition (Wilson 2010).
Investigating multiple niche dimensions using stable isotope and biologging data, we demonstrated a complex system of niche segregation between juvenile sea lions, adult fur seals and adult sea lions. Foraging habitat specialization was the most important factor differentiating the two species. Our study addressed the developmental aspect of niche differentiation between two closely related species and has demonstrated that habitat specialization is evident at the onset of independent foraging, thus extending our understanding of competitive interactions. The availability of foraging habitat might limit the population of western Galapagos sea lions, but we suggest additional effects through multi predator exploitation that may reduce habitat quality in the foraging range of juvenile Galapagos sea lions. Furthermore, our results point towards early habitat learning as a possible mechanism for ecologically derived genetic differentiation.
The present study was performed under the permits No PC-11-08 and PC-043-09 of the National Park Service, Galapagos and financed by the German Science Foundation (DFG, grant TR 105/19-1) and National Geographic (grant No 8682-09). DPC and KTG were supported by the E & P Sound and Marine Life Joint Industry Project of the International Association of Oil and Gas Producers (JIP2207-23) and the Office of Naval Research. We received highly appreciated material sponsorship by Panasonic, Ortlieb, Zarges and Huntsmann Advanced Materials. The Charles Darwin Research Station and the National Park Service provided logistic support during fieldwork. We thank David Anchundia, Enzo Garcia Bartholomei, Sara Maxwell and Maria Szphegyi for their help in the field. Barbara Teichner and Elke Hippauf helped with laboratory work. Patrick Robinson assisted invaluably in spatial data analysis. Paddy Brock kindly corrected the English, and Oliver Krüger and three anonymous reviewers gave helpful comments on the manuscript.