Unexpected decadal density-dependent shifts in California sea lion size, morphology, and foraging niche

Many marine mammal populations are recovering after long


In brief
Valenzuela-Toro et al. show that contrary to predictions, male California sea lions increased rather than decreased their average body size over a 46-year population recovery period. Although selective ecological dynamics can shield populations from boosted densitydependent competition, global warming may reduce their capacity to overcome it.

SUMMARY
Many marine mammal populations are recovering after long eras of exploitation. 1,2 To what degree densitydependent body size declines in recovering species reflect a general response to increased resource competition is unknown. We examined skull size (as a proxy for body size), skull morphology, and foraging dynamics of the top marine predator, the California sea lion (Zalophus californianus), which have been steadily increasing over the last few decades and have approached or reached their carrying capacity in southern California. 3 We show that, contrary to predictions, male California sea lions increased rather than decreased their average body size over a 46-year (1962-2008) recovery period. Larger males had proportionally longer oral cavities and more powerful bite strength, and their foraging niche expanded. Females between 1983 and 2007 maintained stable skull dimensions, but their isotopic niche was broader than contemporary males. Increased male body size is compatible with an intensification of density-dependent sexual selection for larger and more competitive individuals concurrent with an expanding foraging niche. High foraging variability among females would explain their body size stability during decades of population recovery. We demonstrate that body size reduction is not the universal response to population recovery in marine mammals and show that selective ecological dynamics could contribute to protecting populations against the increased density-dependent intraspecific competition. However, prey shifts associated with climate change will likely prevent California sea lions (and other marine mammals) from attaining these ecological dynamics, augmenting their vulnerability to resource competition and diminishing their capacity to overcome it.

RESULTS
A negative relationship between body size and population density has been observed in mammals. 4 As populations approach carrying capacity, intra-specific competition for food increases, affecting the animal's growth and driving decreasing body size. 5 We test whether density-dependent body size reduction is the general response to population recovery. California sea lions (Zalophus californianus) are the most abundant and rapidly growing marine mammal species in the eastern North Pacific, so body size declines are anticipated. While variable population dynamics have been recorded in the Gulf of California, 3 during the last five decades, the California sea lions' population has steadily increased in the Channel Islands (Southern California), with recent expansions of rookery sites to central California, which are currently experiencing exponential growth 3,6,7 (Figures 1A and S1-S3). Disentangling how California sea lions' body size, morphology, and foraging have shifted during their recovery is vital for designing appropriate conservation strategies moving forward. Here, we conducted the first decadal-scale study of density-dependent morphological and ecological shifts in the California sea lion. We recorded skeletal and isotopic measurements from museum specimens from central and northern California to test the prediction that California sea lions underwent a density-dependent reduction in body size during recovery. We examined if and how resource competition influenced the growth and foraging of females and males (over a 24and 46-year period, respectively) as their population increased.
Shifts in adult body size and functional skull traits over time We recorded six linear measurements of the skulls of adult female (n = 66) and male (n = 273) California sea lion skeletons collected from the coastlines of central and northern California ( Figure 2A; Table S1). Females remain near the breeding colony and are considered residents of the small central California colonies based on this species' life history and historical population dynamics. In contrast, males are transient and likely stranded during their lengthy post-breeding foraging trip from much larger breeding sites on the Channel Islands (STAR Methods). We calculated four morphological indices (detailed in Table S1) associated with feeding and physical display performance. We found that skull length (condylobasal length, CBL) and standard body length (SL) were significantly correlated for females (Pearson's r = 0.43, p < 0.001) and males (Spearman's r = 0.39, p < 0.001). We, therefore, used skull length as a proxy for body size. 15 The ratio between the CBL and SL of both sexes did not change over the period (females, p = 0.81; males, p = 0.41); consequently, changes in the skull size over time are expected to be associated with changes in the standard length. The average CBL (females, 232.10 ± 6.28 mm; males, 286.60 ± 8.81 mm) and the maximum (females, 246.06 mm; males, 311.0 mm) and minimum (females, 218.88 mm; males, 268.05 mm) CBLs for both sexes were within the range of the adult size reported by previous studies for this species. 16 We hypothesized that as their population increased, intraspecific competition would lead to a decline in adult body size for female and male California sea lions, as observed in other pinniped species. Contrary to this prediction, we found that male adult California sea lions' skull dimensions (including our skull length) significantly increased between 1962 and 2008 ( Figure 2B . Dashed lines indicate the local polynomial regression line; colored shadows represent their 95% confidence intervals. Black dots represent the historical population of female and male California sea lions obtained from counts at rookeries and haulouts in southern California. 8,9 Vertical gray rectangles represent moderate and strong warm anomalies in the eastern North Pacific after 1972. Data from NOAA. 10 (B) Distribution of principal California sea lions' rookeries along the North Pacific (circles 1-5) and the Gulf of California (circles 6 and 7). Colored lines represent the post-breeding foraging trips of females (purple) and males (yellow) from the Channel Island rookeries. (C) Reconstructed abundance of two of the historically preferred prey species of California sea lions, the Pacific sardine and northern anchovy, along the western coast of the United States between 1950 and 2014. Data from Dunstan et al. 11 Percentages indicate the proportion of these species in California sea lions' diet over time from Stewart et al., 12 Weise and Harvey, 13 (yellow circles, 1962-1982; yellow diamonds, 1983-2006). Note that the upper temporal limit for the stable isotope analysis of males is shorter than for their morphometric analysis. Ellipses represent the sizecorrected standard ellipse area (SEA C ), including 95% of the data. The yellow dashed and continuous ellipses represent the SEA C of males collected between 1962 and 1982 and 1983 and 2006, respectively. Purple ellipse represents the SEA C of females.
(D) Violin plots of the isotopic niche space (Bayesian standard ellipse area) for male and female California sea lions during the same periods indicated in (C). Inset boxplots represent the median (horizontal line), inter-quartile range (rectangle), 95% range (vertical lines), and outliers (black dots). Letters on top represent significant differences between groups (p < 0.001). Table 1). Likewise, metrics of skull function, such as the skull shape index (SSI; an indicator of the masticatory muscle insertion area) and relative mastoid width (an indicator of neck muscle insertion area), significantly increased over this 46-year interval (Table 1). These results are consistent with adult male California sea lions evolving more robust skulls with enhanced abilities to engage in physical encounters. Our dataset's temporal record for females was shorter ) and restricted to resident animals from central California. Females did not show changes in skull dimensions or functional traits during this period. While the number of females at the central California breeding sites increased at the end of this interval (in the early 2000s), there is no evidence that they were at carrying capacity by 2007 ( Figure S3). In contrast, males stranding on the central California coast during this time interval  significantly increased their skull length and the dimensions of their mouth, including their mechanical advantage (an indicator of bite force) and their relative palatal length (size-corrected length of the oral cavity; Table 1). These changes occurred contemporaneously with the regional increase of California sea lion population, approaching their carrying capacity ( Figure S2).

Decadal shift in foraging ecology
We tested whether California sea lions shifted their foraging and habitat preferences during this period of population recovery by measuring the carbon (d 13 C) and nitrogen (d 15 N) isotope values of bone collagen for a subset of adult females (n = 67) and males (n = 104), encompassing a period between 1983 and 2007, and between 1962 and 2006, respectively ( Figure 2C). Note that the time frame record for males analyzed for stable isotopes was slightly briefer than for the morphometric analysis. We found significant differences between the standard ellipse area of females and males (c 2 (2) = 982.1, p < 0.001) ( Table 2). Pairwise comparisons indicated that males collected between 1962 and 1982 had significantly smaller isotopic niches than males collected between 1983 and 2006 (p < 0.001) and females (p < 0.001) (Figure 2D). We could not evaluate long-term variation in the female isotopic niche due to the shorter time frame of our female dataset ). Yet the isotopic niche breadth of females is significantly larger than that of contemporary males (p < 0.001; Figures 2C and 2D), indicating that females have greater foraging variability than contemporaneous males.
The selected generalized linear models (GLMs) to account for the d 13 C values of females and males only explained a low percentage of their variance and included the nonsignificant effect of morphological indices (Table S2). The selected GLMs explained 33% and 15% of the variance in the d 15 N values of females and males, respectively. The SSI (t = 2.11; p = 0.040) and the year of collection (t = À2.73; p = 0.0087) had a significant effect on the d 15 N values in females, whereas the year of collection had a significant effect on males (t = À3.22; p = 0.0019).

DISCUSSION
Over the past 50 years, legal protections have facilitated the recovery of many marine mammal populations after centuries of human exploitation. 1 Several populations of pinnipeds (seals, fur seals, sea lions, and walruses) have shown a marked recovery, reaching and even surpassing historical baselines and becoming iconic emblems of conservation success. 2 While population recovery for many marine mammal species has been a long-term conservation goal, marine predator recoveries can affect the structure and function of food webs by exerting topdown pressure on prey, sometimes producing new conservation challenges. 2 Predicting the long-term effects of marine mammal recovery on food webs and how those webs will respond to climate change requires a mechanistic understanding of the ecological dynamics experienced as populations increase over time.
As marine mammal populations increase and approach their carrying capacities, density-dependent limiting factors such as intraspecific competition for resources will intensify, affecting traits such as body size. In marine mammals, body size increases with food availability during growth. 4 During population recovery, increased competition for resources is thought to drive declines in adult body size, as observed in northern fur seals, South American sea lions, and harbor seals. 17-20 These examples suggest density-dependent body size reduction is a general response to pinniped population growth. Yet pinnipeds display distinctive species-and sex-specific life-history strategies that influence the response to density-dependent pressures. The extent to which body size reduction is a stereotypic response to population growth and how shifting body size might relate to other fundamental traits of individuals, such as their foraging ecology and reproduction, remained unknown.
Contrary to our predictions, we found that over a 46-year (1962-2008) period, during which the male California sea lion population was growing and approaching carrying capacity Non-parametric Spearman's correlation between year of collection and morphological variables is described in Figure 2A and Table S1. Years between parentheses indicate the temporal range of collection of the individuals analyzed. r (rho), Spearman's correlation coefficient; p, significance level. a Significant correlation ll OPEN ACCESS ( Figures 1A and S2), adult males significantly increased rather than decreased their average body size, as deduced from increases in skull length (and other skull dimensions; Figure 2B). Male sea lions collected between 1983 and 2008 had larger and broader skulls, enhanced bite force, and an expanded isotopic niche width ( Figures 2C and 2D). Our sample of females from central California was not at carrying capacity, and female skull dimensions remained stable, but their isotopic niche was conspicuously larger than that of contemporaneous males.
These results indicate that body size decrease is not the universal response to increased competition associated with population recovery in marine predators. Moreover, we show that male California sea lions have expanded their foraging niche over recent decades and that females display unexpectedly high foraging variability. The morphometric response of females and males suggests that food limitation did not adversely affect sea lions' body size growth during the decades over which their population increased. This shouldn't be surprising for females, as their increased density did not approach carrying capacity in central California, but it is surprising for males, whose population numbers had plateaued. This unexpected increase in the body size and change in the functional traits (i.e., mechanical advantage, relative mastoid width, relative palatal length, and SSI) of male sea lions suggest that other density-dependent factors may have influenced their morphological and ecological dynamics. California sea lions are polygynous breeders that annually congregate at reproductive sites. As their population recovered, the density of males at breeding sites increased ( Figure S2), and male-male confrontations for territorial control presumably intensified. 21,22 Body size influences male competitiveness through aggressive displays 23,24 and impacts attendance patterns during the breeding season. Larger individuals can endure extended fasting spans, allowing them to stay and defend territories for longer periods, increasing their reproductive success. 25,26 Consequently, as the population recovers, larger and more competitive individuals with higher fasting ability would be favored by increased sexual selection in the breeding colonies.
Bite force and neck mobility are also relevant to male-male confrontations and reproductive success. 27,28 Therefore, if sexual selection intensified, a positive selection of biting force and neck mobility would be anticipated. Indeed, our results show that adult male California sea lions have developed broader insertion areas for the muscles associated with biting (e.g., the temporalis muscle as indicated by an increased SSI) and in neck lateral flexion and rotation (e.g., sternocleidomastoid muscle as indicated from relative mastoid width) between 1962 and 2008 (Table 1). Long-term assessments of sexual selection are not available for this species; however, future decadal-scale studies on relative baculum size, 29 reproductive steroid hormones, 30 or genetic paternity 31 in museum specimens might provide ways to test and quantify the occurrence of densitydependent sexual selection in adult male California sea lions associated with population recovery.
Larger individuals could, in theory, forage further, longer, and deeper with proportionally lower energy expenditure than smaller individuals. 32-35 Therefore, positive selection for more efficient foraging may have reinforced the trend toward larger size. While increased body size results in higher absolute energy requirements, 36 adult males have expanded their isotopic niche, hinting at the exploitation of more diverse resources during their post-breeding foraging trips (Figures 2C and 2D). Some individuals from 1983 to 2006 had lower d 15 N values than earlier males, which is consistent with a northward extension of their foraging trips and the consumption of higher latitude (with lower isotope baseline) prey. 37 This aligns with sightings of male California sea lions in southeastern Alaska in recent decades. 38 Moreover, the increase in oral cavity length of adult males would enhance their ability to capture and handle larger prey. 39 This is further supported by some contemporary adult males displaying higher maximum d 15 N values, likely indicating preferential consumption of higher trophic-level prey.
While other studies have documented significant body size reductions of female sea lions during population recovery (e.g., Drago et al. 18 ), adult female California sea lion size remained stable. The spatial distribution of our sample and the distinct population dynamics observed in sea lions' breeding rookeries along the California coast likely influence these results. Unlike males, which have occupied haul-out sites along central California for several decades, 8,40 female California sea lions were relatively rare until the mid-2010s, when new breeding sites were established at Añ o Nuevo and the Farallon Islands 3 ( Figure S3). Information on the feeding ecology of females inhabiting these sites is not available. However, female California sea lions are central place foragers and conduct short foraging trips near their breeding sites, although exceptional sightings of females from the Channel Islands foraging in Monterey Bay in central California have been documented. 41 The degree of foraging overlap between females and males in this area is unknown, yet some spatial segregation is expected to be associated with body size differences. 42 These observations suggest that female sea lions inhabiting central California were rare during the study period . They likely experience low-density-dependent intraspecific competition for resources, contributing to their body size stability. We found that adult females have greater variability in their foraging and dietary preferences (as characterized by the isotopic niche) than contemporary adult males. The high foraging variability among females aligns with studies showing that they are highly efficient foragers 42 with diverse foraging strategies. 43 The mechanisms underlying these observations are unclear; however, the association between the SSI and the d 15 N values hints that skull morphology might contribute to niche breadth by favoring foraging specialization that lessens intraspecific competition among females. Additional decadal-scale studies of female breeding rookeries in the Channel Islands ( Figure S1), which have presumably reached their carrying capacity (Figure S2), are required to test the occurrence (or not) of densitydependent shifts in body size and to further explore the effect of skull morphology on foraging dynamics.
Although dietary data are not available for our entire study period , reports from the 1980s to the early 2000s show that California sea lions' diet was diverse but consistently dominated by energy-rich pelagic species such as Pacific sardines (Sardinops sagax) and northern anchovies (Engraulis mordax), 12,13 mirroring fluctuations in their regional abundance ( Figure 1C). Following the collapse of these fisheries in the 2010s, 44,45 however, California sea lions' diet shifted toward a higher contribution of less energy-dense prey 14 (Figure 1C), coinciding with lower pup recruitment 46 ( Figure 1A). Likewise, depletions of pelagic fish during El Niñ o events ( Figure 1C) have negatively affected California sea lions' energy budget, reproduction, and even their immune response (e.g., Costa et al. and DeRango et al. 47,48 ). Together, these observations demonstrate that the availability of pelagic prey is critical for California sea lions' population dynamics, sustaining their recovery and their decadescale, density-dependent shifts in body size, morphology, and foraging niche.
It is uncertain how the recovery of marine mammals will be affected as changes in prey abundance and distribution associated with climate further develop. Climate models predict widespread shifts for the California Current system because of global warming. Among other consequences, models forecast an increase in the frequency and intensity of El Niñ o-like warming conditions 49,50 and an overall reduction of body size and dispersal ability and poleward shift of pelagic fishes. 51,52 These new conditions will likely impact marine predators, such as California sea lions, increasing their foraging effort and depressing their energy budget by decreasing foraging performance. Indeed, the energetic and ecological tradeoffs resulting from density-dependent selection pressures like those presented here may be unattainable, lowering their capability to overcome resource competition and reducing species' carrying capacity, leading to a steep population decrease.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

METHOD DETAILS Specimens
We measured skeletal material from specimens at the Ornithology and Mammalogy Collection at the California Academy of Sciences. We surveyed 339 adult California sea lion skulls collected between 1962 and 2008 for males (n = 273) and between 1983 and 2007 for females (n = 66). We selected specimens based on the preservation of skull morphology, geographic origin, year of collection, and relative age. Sex was determined when collected and confirmed by the presence of the baculum when available. We only included skeletal materials collected from central and northern California (encompassing the coastline between Del Norte to Monterey counties, but mainly (> 85%) from Añ o Nuevo State Park, the San Francisco seashore area, and the Point Reyes National Seashore to avoid the confounding effect of geographical variation among populations. 54,55 Because of their sexually dimorphic life history and variable population dynamics across their geographic range, stranded female, and male sea lions in central and northern California between the 1960s and the late 2000s would chronicle different density-dependent dynamics.
During the study period , the male California sea lion population was dominated by individuals breeding on the Channel Islands ( Figure S2). Consequently, male skeletal remains found in central and northern California from the early 1960s to the late 1970s, would correspond predominantly to transient individuals stranded during their post-breeding foraging trip from or toward their reproductive sites in Southern California.
As new breeding colonies were established on Añ o Nuevo and the Farallon Islands in the late 1970s to early 1980s 56 ( Figure S3), an increasing number of resident individuals (i.e., breeding in central California) might have contributed to the local death assemblage (and our dataset). However, the scale of breeding in the region was dramatically smaller than on the Channel Islands for much of our study interval (compare Figures S2 and S3). Thus, male skeletal remains chronicle the population and ecological dynamics experienced by sea lions in the Channel Islands and, more recently, in central California. Males presumably experienced rising densitydependent intraspecific competition for resources and space as their population in the Channel Islands expanded ( Figure S2), which was likely intensified further as the number of resident males in central California began to rise.
Long-term monitoring indicates that females remained in very low densities in central California until the late 2000s when large scale breeding sites were established in this area ( Figure S3). These observations, combined with the fact that females do not conduct long-distance foraging trips and remain in their breeding rookeries all year round indicate that female' skeletal remains collected between 1983 and 2007 in central California primarily correspond to resident individuals with a minor contribution of individuals from the Channel Islands. Limited information of the females now residing in central California is available, and critical aspects of their foraging ecology such as their preferred foraging locations and diet remain unknown. Yet, lactating females tend to be central place foragers, undertaking short foraging trips near their breeding colony, 25,43 which hints that resident female may be exploiting foraging grounds near Añ o Nuevo and the Farallon Islands. If so, resident females in central California likely encountered low density-dependent intraspecific competition for resources during the study period. While it is unknown to what extent foraging preferences between resident females and males might overlap, body size differences between them are significant, suggesting that some type of spatial foraging segregation occurred, including targeting different prey sizes or the exploitation of distinct depths in the water column, as has been observed in other pinniped species. 57 Age estimation Various methods have been used to age pinniped skulls. The count of annual growth layer groups in sectioned canine teeth is one of the most reliable, 58,59 but this method is destructive. Relative age measures, such as the cranial suture age or Suture Index (SI), have been developed. 54,60 SI is based on the observed degree of fusion of nine cranial sutures (occipito-parietal, squamosoparietal, interparietal, interfrontal, coronal, basioccipito-basisphenoid, maxillary, basisphenoid-presphenoid, and premaxilla-maxillary), 60 which ranges from 1 (unfused suture) to 4 (completely fused suture). The SI is calculated as the sum of the degree of fusion, with physically adult skulls ranging between 19 and 36. 54,60 The validity of this approach has been tested for several taxa (e.g., Kahle et al. 61 ), including seals and otariids for which a strong positive correlation between SI and age exist. [61][62][63] Previous studies of cranial sutures in California sea lions have shown a correlation between chronological age and SI, with 15-yearold males having all cranial sutures fully closed (SI > 30). 64 However, extensive variability in the cranial suture pattern across carnivorans suggests a strong influence on feeding ecology. 65 Furthermore, male otariids, including California sea lions, display biphasic development with secondary growth in structures associated with male-male combat (and territorial dominance) well after physical maturity is reached. 16 Despite being physically and reproductively mature, males that fail to maintain territories do not develop the complete secondary sexual characteristics, including the complete suture fusion of the facial skeleton. 7,62 Moreover, the fusion of premaxilla-maxillary suture is delayed in otariids, resulting in underestimating the SI relative to counts of the annual growth layer groups. 63 Male California sea lions reach their maximum skull length, zygomatic breadth, mastoid breadth, and sagittal crest when ten years old, corresponding to an SI of 20. 16,64 Similarly, females reach their maximum skull length at SI equal to 17-19. 16 For this study, we included male specimens whose SI was equal to or greater than 20. Likewise, we included female skulls whose SI was equal to or greater than 19. In addition, other qualitative features were evaluated to confirm the physical maturity of the specimens, including the absence of deciduous teeth, the display of dental wear, the prominence of the sagittal crest, and the relative ossification of the tympanic bulla and the dentaries. 66,67 explanatory variables for each sex. We verified that the correlation between explanatory variables was %0.7 through the Pearson correlation coefficient using the packages corrplot 75 and ggcorrplot. 76 We ran models for d 13 C and d 15 N, employing a gamma distribution with an inverse link function and a gaussian distribution for the absolute d 13 C and d 15 N values, respectively. We ranked models based on their Akaike's Information Criterion (AIC) using the package AICcmodavg. 77 The models with the lowest AIC values were considered to best fit. 78 Changes in the isotopic niche through time The population of California sea lions has sustained growth over at least the last six decades (Figures 1, S2, and S3). To investigate the relationship between population increase, foraging, and habitat preferences of California sea lions through time, we categorized individuals analyzed for stable isotope values into groups depending on their sex and the year they were collected. We defined three groups: (1) males collected between 1962 and 1982, (2) males collected between 1983 and 2006 (note that this group's upper temporal limit is lower than those included in the morphometric analysis), and (3) females collected between 1983 and 2007. These groups chronicle distinct density-dependent ecological pressures based on their sex-specific life history and population dynamics across their range. Males belonging to groups (1) and (2) broadly correspond to transient individuals stranded during their postbreeding foraging trips from breeding sites in the Channel Islands. Group (2) would also include a number of individuals breeding at Añ o Nuevo and the Farallon Islands in central California consistent with the recent establishment of breeding rookeries in the region ( Figure S3). These differences imply that individuals from group (1) chronicle the effects of increasing density-dependent ecological pressures resulting from California sea lions' steady population increase in the Channel Islands ( Figure S2). Individuals from group (2) document the effects of intensified density-dependent competition resulting from the accelerated population increase observed in the Channel Island between the 1980s and the late 2000s (the period during which they approached and presumably reached their carrying capacity) ( Figure S2) strengthened by the consistent population expansion of males residing in central California ( Figure S3). Conversely, females (group 3) primarily represent residents from central California, which remained in low density during the study period, experiencing potentially low levels of intraspecific competition.
To estimate isotopic niche width (a proxy for an ecological niche) between groups in the population, we used Bayesian multivariate ellipse-based metrics implemented in the package SIBER. 79 We calculated the standard ellipse areas considering 95% of the data corrected for small sample size (SEA C ) and Bayesian standard ellipses (SEA B ) for each group. We estimated SEA B using 5 chains of 10,000 iterations, a burning of 1,000, and a thinning of 10. We compared each group's posterior probabilities of SEA B by using a Kruskal-Wallis analysis of the variance, followed by a Dunn test with Bonferroni correction for multiple comparisons using the package FSA 80 in R.