Sexual dimorphism does not translate into foraging or trophic niche partitioning in Peruvian boobies (Sula variegata)

Intraspecific competition can lead to sexual segregation of diets or foraging behaviors in seabirds, and in some species the resulting niche partitioning is facilitated by sexual dimorphism. However, environmental stochasticity can mediate intraspecific competition and thus the extent of sex-based partitioning. The Peruvian booby (Sula variegata) is a sexually dimorphic seabird endemic to the Humboldt Current System (HCS), a highly variable environment due to El Niño Southern Oscillation. To determine the extent of sexual partitioning in this species, we quantified the foraging and trophic niches of breeding Peruvian boobies at Isla Guañape Norte, Peru in two years with different oceanographic conditions and nesting propensity. Morphometrics, GPS-tracked foraging behaviors, diets via regurgitates, and isotopic niches were compared between sexes and years where sample sizes permitted. Although females were larger and in better body condition than males, breeding Peruvian boobies in our study did not exhibit sex-specific foraging or isotopic niche partitioning and had few differences in diet. Anchoveta (Engraulis ringens) dominated diets in both years, reflecting Peruvian boobies’ dependence on this prey. Overall, while oceanographic conditions in 2016 were unfavorable enough to reduce nesting propensity, these effects did not qualitatively translate to foraging or dietary niche partitioning between the sexes for those individuals who opted to breed. In combination, our results suggest weak intraspecific competition during our study period, and highlight how the foraging strategies of Peruvian boobies have adapted to the variable environmental conditions found in the HCS.

Methods
We sampled Peruvian boobies from Isla Guañape Norte (8.545°S, 78.964°W), an island located approximately 9 km offshore of northern Peru, in December 2016 and in November 2019. According to the Índice Costero El Niño, birds were experiencing weak El Niño conditions (0.47) in December 2016, and neutral conditions (-0.40) in November 2019.

Morphometrics measurement collection and processing

We sampled Peruvian boobies in December 2016 and in November 2019. Breeding adults were captured at the nest soon after dawn (06:00–08:00) using a monofilament lasso attached to a 5-meter telescopic pole. In 2016, adults selected for capture (14 females and 4 males) had eggs and/or chicks less than 7 days old. In 2019, the selected adults (20 females and 12 males) had nests containing chicks between 3–8 weeks old and represented 2% of individuals found breeding on the study plot. Both parents were present at the nest during capture, ensuring that chicks or eggs were not left unattended. Sex, weight (g), breeding condition, culmen length (mm), tarsus length (mm), and wing chord (mm, naturally arched) were recorded for each captured bird. Sex was assigned from observed vocalizations at the nest.

We generated a body condition index (BCI) by performing a principal components analysis (PCA) on the culmen, tarsus, and wing chord measurements of all individuals, then regressing the first principal component (PC1, with an eigenvalue of 2.258 explaining 75.3% of variance) of each individual against body mass, and finally assigning the residuals of this regression as the body condition index. Three birds were excluded from this index due to missing measurements.

GPS tagging and processing

We tagged a subset of captured breeding birds with a GPS device, which was attached to the central tail feathers using Tesa® 4651 waterproof tape. In total, we successfully tracked 7 females and 3 males in 2016 (15 foraging trips total), and 20 females and 12 males in 2019 (60 trips). Tagged birds had their head feathers marked using PAINTSTIK® livestock markers (LA-CO Industries, Inc.). In 2016, both i-gotU GT 600 (30 grams) and GyPSy-5 (14 grams) GPS devices were used, set to record fixes every 2–10 seconds, and waterproofed by placing each device inside a condom and sealing it inside a heat-sealed polypropylene plastic bag (1 gram). In 2019, Axy-Trek Marine GPS devices weighing 32 grams (Technosmart Europe S.r.l., Rome, Italy) were used to tag birds and set to record GPS locations and dive depth in continuous mode every 1 second. GPS weight was 2% of the weight of the lightest tagged bird (1200 g). Birds were recaptured late afternoon (15:00–18:00) on the day of tagging.

After retrieval and download of the GPS tracks, the data were recovered from the devices, examined in ArcGIS, and processed prior to analysis. In total, 10 individuals in 2016 and 32 individuals in 2019 yielded usable tracks. As birds were all nesting at the colony, for each track, we first removed all points inside of a 300 meter radius from a point designated as the colony center. We then noted departures and returns for each foraging trip. If a single individual left and returned to the colony center multiple times, each unique segment between a departure/return pair of points, as bounded by start and end time, was defined as separate trips. After confirmation of foraging trips, we then visually checked for GPS data such as non-foraging trip activity (bathing behavior near colony immediately after capture and tagging of the individual) and GPS errors (single outlier points). These data points were removed from tracks either manually or using a similar radius approach. The cleaned data were imported into R, and each trip by each individual was assigned a unique trip identifier. GPS points for each trip were exported. This data was used to further calculate foraging trip metrics, EMbC, and kernel densities as described in the paper.

For foraging trip metrics, we calculated total trip distance, maximum distance from colony, and trip duration in using the sp (v.1.4-5; 55), adehabitatLT (v.0.3.27; 56), and lubridate (v1.7.9; 57) packages in R (v4.2.2; 58). Specifically, we determined the duration of each trip by calculating the absolute date/time difference between the earliest departure point and the latest return point. We determined maximum distance from colony by specifying the latitude/longitude coordinates of the colony, calculating the distance to colony from each GPS point (km), and taking the maximum value returned for each bird and each trip. We determined total trip distance by calculating the distance between successive points via taking the difference between the latitude/longitude coordinates of each pair of points and totaling up these distances for each individual trip.

We characterized at-sea behaviors of boobies using the Expectation Maximization binary Clustering (EMbC) algorithm (v2.0.4; 59). It uses two input variables (speed and turning angle) to determine and assign one of four behaviors to sets of velocity/turn pairs in the movement data.

Tissue collection for stable isotopes (δ13C and δ15N)

We collected 0.5 ml of blood from the tarsal vein of all captured breeding birds. We preserved collected blood in vials with 1 ml of 99.9% ethanol from a common source.

Peruvian booby whole blood samples were dried and homogenized using a mortar and pestle. Approximately 0.6 mg of each sample was then loaded into tin cups and flash-combusted using a Costech ECS4010 elemental analyzer. These samples were analyzed for carbon and nitrogen stable isotopes (δ13C and δ15N) using an interfaced Thermo Delta XP continuous flow stable isotope ratio mass spectrometer. Raw δ values were normalized on a two-point scale using glutamic acid reference materials with low and high values (i.e., USGS-40 (δ13C = −26.4‰, δ15N = −4.5‰) and USGS-41 (δ13C = 37.6‰, δ15N = 47.6‰)). Sample precision based on repeated sample and reference material was 0.1‰ and 0.2‰ for δ13C and δ15N, respectively. Stable isotope ratios are expressed in δ notation in per mil units (‰). The Rstandard values were based on the Vienna PeeDee Belemnite (VPDB) for δ13C and atmospheric N2 for δ15N.

Diet collection and processing

Stomach samples were obtained either through induced regurgitation of captured boobies upon return from a feeding trip or at random from either the periphery of the nesting sites or beaches. In 2016, sampling occurred from December 15th to the 22nd, and we induced regurgitation of randomly selected birds when they returned to the nest (n = 82). Unfortunately, the sex of birds was not recorded in 2016. In 2019, sampling occurred from November 16th to December 8th, and we induced regurgitation of sexed birds that were equipped with GPS loggers immediately upon their return from a feeding trip (n = 25). Whole prey in undigested samples were sorted, identified, and measured. After the identification of prey items, we assessed the frequency of occurrence of each prey species, diet composition by species (percent mass), total regurgitate mass, number of prey items, and individual prey size (total length in mm). Total fish length was measured from the snout to the end of the tail fin. Intact fish were assessed by direct measurement (accuracy ±1 mm), while measurements for partially digested Engraulis ringens were obtained by digitally measuring an intact sagittal otolith with a Zeiss Stemi SV 6 dissecting microscope with a digital SPOT RT camera attachment and Image-Pro Plus software (Opelco; accuracy ± 0.01 mm). Initial E. ringens length was estimated using the equation fish length (cm) = 0.798 + 3.33 (otolith length [mm]).