Diving deep into trouble: the role of foraging strategy and morphology in adapting to a changing environment

Rapid environmental change may require fur seals and sea lions to make longer or deeper dives to find resources. A systematic review of diving behaviour and physiology revealed that a foraging strategy (epipelagic vs mesopelagic or benthic) was more important in predicting dive behaviour than morphology (fur seal vs sea lion).


Introduction
Globally, loss of species in marine environments has been slower than in terrestrial systems, but it now appears to be accelerating rapidly (McCauley et al. 2015), with increasing consequences predicted for many trophic levels including top predators (Schumann et al., 2013). Humans have profoundly decreased the abundance of many marine fauna and reshaped the genetic structure of a number of marine animal populations through over-and selective harvesting, respectively (McCauley et al., 2015), as well as fundamentally changing the structure of many marine communities (Vergés et al., 2014). Climate change is exacerbating these effects with distribution ranges shifting markedly as temperature regimes alter (Hazen et al., 2012;Brown et al., 2015;Deutsch et al., 2015). These changes raise profound implications for top predators that may have to adapt foraging strategies to cope with shifting prey resources (Bakun et al., 2015;Sydeman et al., 2015;Carroll et al., 2016).
In general, predator foraging strategies account for the predictability and availability of food resources Kuhn 2011;Arthur et al., 2016). While most predators have sufficient inherent flexibility in their foraging strategies to adapt to moderate changes in their prey base (Grémillet and Charmantier, 2010), this capacity is limited by behavioural and physiological capacity. Air-breathing diving marine predators, such as marine mammals, face additional constraints imposed by their physiology and morphology (Schreer and Kovacs, 1997;McCafferty et al., 1998;Rosen et al., 2007). In the face of environmental uncertainty, there is significant potential for the foraging efforts and associated energetic demands of foraging marine predators to increase, particularly in relation to increased search or travel time (Kuhn 2011;Bestley et al., 2015). This is likely to be most severe for the smaller marine mammals such as female otariids (fur seals and sea lions), as adult females are small [30-120 kg except Steller sea lions (Eumetopias jubata)], and central place foragers while rearing offspring, and under suboptimal conditions must increase both individual dive duration and trip length Staniland et al., 2010;. Otariids inhabit temperate coastal waters of every continent and are semi-aquatic, high-trophiclevel predators. Foraging strategies used by otariids are predominantly pelagic (epipelagic or mesopelagic) or benthic (Costa et al., 2001;Costa et al., 2004;Gallagher et al., 2015). Otariids that forage in the epipelagic zone are typically found in open water (0-400 m) near the continental slope or off the shelf where small schooling prey (crustacea, fish and squid) tend to occur in dense patches that are spatially dynamic (Boyd and Murray, 2001;Harcourt et al., 2002;Robson et al., 2004). Otariids that forage benthically generally feed on larger prey that occur in lower densities but are more evenly distributed over the continental shelf in waters often, though not always, less than 200 m (Littnan and Arnould, 2007;Sigler et al., 2009;. When otariids forage in the epipelagic zone, foraging dives are short (typically less than 2 min) and shallow (less than 80 m) and are performed less than a third of the time that seals are at sea (e.g. Boyd and Duck, 1991;Trillmich and Kooyman, 2001;Beauplet et al., 2004). Benthic and mesopelagic foraging strategies are usually associated with longer dive durations, greater depths and a higher proportion of time at sea spent diving (TASD) (e.g. Costa and Gales 2000;Fowler et al., 2007;Spence-Bailey et al., 2007), much of which is spent at the bottom of the dive. As a consequence, benthic and mesopelagic divers are more likely to push their physiological limits than their pelagic diving counterparts . Benthic and mesopelagic divers are predicted to spend more time underwater during dives than epipelagic divers, potentially remaining beyond their estimated physiological capacity Arnould and Costa 2006;Gallagher et al., 2015).
The aerobic dive limit (ADL) is a measure of the duration of a dive during which an animal uses only aerobic metabolism (Davis 2014(Davis , 2019. Since Kooyman et al., (1980) introduced the concept of ADL, it has been used throughout the literature as a measure of the physiological capacity of diving animals. When animals dive beyond their ADL, they are forced to use the less efficient anaerobic metabolism, which results in an accumulation of blood lactate (Kooyman et al., 1983). Lactic acid build-up generally results in a less efficient foraging cycle due to the disproportionate increase in surface durations between dives (Kooyman et al., 1980;Davis and Kanatous, 1999;Davis and Williams, 2012). Measuring the ADL directly through post-dive circulating lactic acid concentrations is a difficult task and to date it has only been measured in two phocids, the Weddell seal (Leptonychotes weddelli) (Kooyman et al., 1980) and the Baikal seal (Phoca sibirica) (Ponganis et al., 1997c), one otariid, the California sea lion (Zalophus californianus) (Ponganis et al., 1997b), and one diving bird, the emperor penguin (Aptenodytes forsteri) (Ponganis et al., 1997a). Instead, an estimate of ADL is generally used, which is calculated from total available body oxygen stores and the rate at which these are theoretically depleted by metabolic processes (Costa et al., 2001). This is the calculated ADL (cADL) and represents the time that an air-breathing animal can theoretically dive relying solely on aerobic metabolism (Costa et al., 2001). Although it is unlikely that there is a hard 'switch' between aerobic and anaerobic metabolism, cADL provides a standardized comparative measure of the aerobic diving capacity of marine mammals (Gerlinsky et al., 2013). A limitation of using cADL is that it can only be as good as both the measures of metabolic rate and of O 2 stores for the different species (Butler, 2006), which may change seasonally (Villegas-Amtmann and Costa 2010).
Otariids notably have shorter dive durations and shallower dive depths than other similarly sized air-breathing diving mammals, a pattern that may be a product of physiological constraints related to their aerobic capacity that may explain which prey they target (Schreer et al., 2001) Boyd and Croxall, 1996), while the larger sea lions may exceed their cADL in 40-60% of dives Fowler et al., 2007). The body size both empirically and theoretically has a direct influence on diving capacity as metabolic rate scales to body mass 0.75 while oxygen stores scale with body mass 1.0 . This means that simply due to allometry, larger animals will have a lower mass-specific metabolism for a relatively constant proportion of oxygen storage capacity (Ponganis, 2015) and so larger animals can dive longer and deeper than small ones based on the body size alone (i.e. dive ability scales with body mass 1.25 ) (Boyd, 2002). Furthermore, fur seals and sea lions differ in thermoregulatory strategies with sea lions primarily relying on blubber for insulation (Liwanag et al., 2012b), while fur seals use layers of fur to stay warm (Liwanag et al., 2012a). Blubber is relatively incompressible and so maintains a more effective insulative capacity at depth compared to the fur seals' double fur layer, which relies on air bubbles trapped in the underfur for insulation (Liwanag et al., 2012b).
However, Costa et al., (2004) hypothesized that the foraging strategy (benthic and mesopelagic vs epipelagic) was the main driver of relative foraging effort rather than morphology (sea lion vs fur seal) per se. They reported that benthic and mesopelagic divers were far more likely to exceed their cADL, spend more time at sea diving and have greater dive durations Gallagher et al., 2015). They concluded that sea lions, as a result of being primarily benthic and/or mesopelagic divers, are likely operating near or at their physiological maximum Fowler et al., 2007;Villegas-Amtmann and Costa, 2010). Diving for periods longer than the cADL is correlated with reduced foraging efficiency (Boyd et al., 1997) and may represent a limitation in their scope for behavioural adaptation to changing ecosystems. The proportion of dives that exceed an animal's cADL can be used as a measure of the species' dive performance.
Foraging efficiency is generally considered in terms of the proportion of the dive cycle spent actively pursuing prey. Hence, for marine mammals, foraging efficiency increases by maximizing their bottom time (Austin et al., 2006), and this is directly influenced by vertical distance travelled (Kramer, 1988). Regardless of whether otariids use an epipelagic, mesopelagic or benthic foraging strategy, environmental changes may require them to switch to deeper or longer dives. When seals are compelled to repeatedly dive for longer periods (e.g. if in a rich prey patch), they may have the ability to accumulate oxygen debt and replenish their stores later (Horning, 2012). The ability to increase their foraging effort without impacting their foraging efficiency (through increased reliance on anaerobic metabolism) is limited by their physiological capacity. An alternate strategy is to alter their overall foraging strategy rather than increase foraging effort under challenging prey conditions. While this theoretically makes sense, there is very little evidence that fur seals switch foraging strategies during periods of low prey abundance (Boyd et al., 1994) and clear evidence that at least some benthic foragers continue to use the same foraging habitat even when the fish guild structure changes significantly (Lowther et al., 2013). This suggests that some species may be physiologically constrained in their capacity to respond to environmental challenges as identified in Costa et al., (2004) and Arnould and Costa (2006). These original studies only looked at a subset of extant otariids and were based on the limited evidence available produced at that time. Diving studies on marine mammals inherently suffer from small sample sizes and become more robust as sample sizes increase (Sequeira et al., 2019). Over the past 15 years, numerous new studies of dive behaviour using larger numbers of animals and more species have appeared. This provides an opportunity to revisit the earlier forecasts of Costa and colleagues. Therefore, the aim of this review is to (i) summarize physiological and diving parameters for otariid species with the most up-to-date literature, (ii) revisit the approach devised by Costa et al., (2004) and Arnould and Costa (2006) by comparing and contrasting the diving characteristics (dive performance and foraging efficiency) of fur seals and sea lions as epipelagic vs mesopelagic and benthic divers and (iii) identify those species that may be operating at or near their physiological maximum and the implications of this in a changing world.

Methods
The family Otariidae (colloquially known as 'otariids') consists of nine extant species of fur seal (Arctocephalinae) with one subspecies and six species of sea lion (Otariinae) (Wolf et al., 2007). Despite no clear phylogenetic distinction between sea lions and fur seals (Berta and Churchill, 2012), we differentiate and compare these groups within this study because of their distinct morphological differences. The primary distinction between the two groups is their size and thermoregulation strategy, where the smaller fur seals rely more on their fur while the larger sea lions rely more on blubber for insulation (Berta et al., 2005). Using an array of search platforms (Google Scholar and Macquarie University Library) and databases (Wiley Online and Elsevier), we collected information from 16 species of lactating female otariid (where available). Physiological parameters were collected first on as many species as possible. Physiological parameters included total body oxygen stores (TBOS), diving or field metabolic rate (DMR or FMR), cADL and average body mass. Behavioural diving parameters were then collected from published sources for these same species when possible. Behavioural diving parameters included average and maximum dive duration and depth, percentage of TASD and post-dive surface interval (Tables 1 and 2). Not all of the required parameters were available for all 16 species and in some cases the value was inferred from a species matched from phylogenetic relationships informed by morphology  (Churchill et al., 2014) and assumed to be equivalent .
Values of cADL are only reported for those species where either it had been previously calculated and published or only one required parameter was missing (either TBOS or DMRsee below). This enabled us to report or estimate the cADL for females of six fur seal and six sea lion species. We focussed on females as these were most likely to be physiologically constrained due to their smaller size and the impact that changes in foraging efficiency have on nursing offspring. There was insufficient information to include Guadalupe fur seals (Arctocephalus philippii townsendi), South American fur seals (Arctocephalus australis), Cape fur seals (Arctocephalus pusillus pusillus) or New Zealand fur seals (Arctocephalus forsteri) to calculate cADL.
While cADL is only an estimate of actual ADL (Gerlinsky et al., 2013), it remains the best standardized measure of diving abilities available (Butler, 2006). In this paper, cADL values were either taken directly from the literature as the author reported it or calculated from other available physiological data, specifically by dividing TBOS (mL O 2 kg −1 ) by the DMR (mL O 2 kg −1 min −1 ) of the animal (Costa et al., 2001): cADL and reported cADL differed significantly, and there were no reported cADL values for all species. Therefore, for consistency, only cADL values (those which were derived from the above formula) were used for analyses and reported cADL values are captured in the tables but not used.
TBOS are the combined oxygen stores in the lung (usually estimated as lung volume multiplied by the fraction of air in the lungs that is assumed to be oxygen at the start of a dive), blood (calculated from measures of blood volume, haemoglobin concentration and the haemoglobin oxygen binding capacity) and muscle (calculated from estimates of muscle mass, myoglobin concentration and the myoglobin oxygen binding capacity) (Lenfant et al., 1970). Insufficient data were available to provide an estimate of TBOS for Galapagos fur seal (Arctocephalus galapagoensis) and South American fur seal (Arctocephalus australis) so the estimate for Antarctic fur seal (Arctocephalus gazella) was used. All estimates were adjusted for mass and are detailed in Table 3. DMR is a measure of the energy that is expended during submerged activity. For this review, we prioritized DMR data that were directly measured via respirometry, and where this was not available, we used values estimated via doubly labelled water in the field (or FMR). Values estimated via respirometry are generally regarded as the 'gold-standard' of energy expenditure measurements and take into account only the energy expended during the dive (Rosen et al., 2016). FMR is an estimate of energy expenditure over an entire foraging trip, creating difficulties in extracting activityspecific measures of energy expenditure (Costa et al., 1989;Dalton et al., 2014). To more accurately capture diving effort, only at-sea FMR (i.e. excluding measures incorporating onland periods) was used as an estimate of DMR. Either method of estimating diving metabolism may lead to over-or underestimates of the true cADL; thus, we use both to allow for comparisons among species.
Where possible, estimates of DMR were taken from published material and, when necessary, converted into ml O 2 min −1 kg −1 . Any estimates of DMR that were reported as W kg −1 were converted into ml O 2 min −1 kg −1 using the following calculation: DMR was not available for two species [Juan Fernandez fur seals (Arctocephalus philippii philippii), Subantarctic fur seals (Arctocephalus tropicalis)] and, for these, DMR was calculated from an estimate from a similar species (Table 3) and adjusted for the body mass of the target group: where DMR = diving metabolic rate, BM = body mass, target = target species and similar = similar species.
Diving behavioural data were taken from published reports of wild otariid foraging. The average and maximum dive duration, the average and maximum dive depth, and proportion of time at sea diving (TASD), the number of seals it was calculated for and the standard deviation for each value were extracted for each species where it was available. TASD was not always recorded, but a diving rate (dives per hour) was available. We used the following to estimate TASD from other dive parameters: (6) where standard deviation was not available a crude estimate was made as Species were categorized as benthic and mesopelagic or epipelagic divers based on their primary mode of foraging    Kooyman and Gentry (1986); 14. average: Trillmich et al., (1986); 15. Thompson et al., (2003); 16. Kuhn and Costa (2014) Werner and Campagna (1995); 24. average: Costa and Gentry (1986); 25. Baylis et al., (2015a). Information has been collated from published reports of diving parameters for species listed. A southern distribution is represented by negative latitude and is an estimate only. N is the sample size, mass is the average mass of animals studied, depth and duration are the average typical depths and durations of otariids dive, max depth and max duration are the maximum depth and maximum duration recorded by any otariid and TASD is the percentage time at sea spent diving. Arnould and Costa 2006). Species that foraged primarily on demersal prey on the benthos were regarded as benthic foragers. Those that foraged in the deep pelagic zone (below 200 m) were classified as mesopelagic foragers and those that foraged in the upper pelagic zone (above 200 m), tracking their prey through migratory patterns, were classified as epipelagic foragers. Where research found multiple strategies within a single species, the information for each type of foraging mode was captured and analysed separately.

Statistical analysis
A key goal of the study was to estimate how often animals dived beyond their cADL. This cannot be determined from point estimates of dive duration (e.g. mean or maximum), so we simulated a series of dive durations for the 13 species-6 fur seals and 7 sea lions-with complete dive parameters available from the literature. Simulated distributions were created with the mean and max dive duration and the number of seals from which the parameters were derived. We simulated diving durations using negative binomial distributions using the MASS package in R (Ripley et al., 2013) and then calculated the percentage of those dives that exceeded the cADL. We first estimated theta using μ = mean dive duration, y = average number of dives on a single foraging trip, df = number of seals − 1, then we simulated the dive distribution of a given species using the function rnegbin from the MASS package with 20 000 simulations, with the estimated theta and the specified limits of lower = 0 and upper = max

Arnould and Hindell
; 15. Feldkamp et al., (1988). Information has been collated from published reports of diving parameters for species listed. A southern distribution is represented by negative latitude and is an estimate only. N is the sample size, mass is the average mass of animals studied, depth and duration are the average typical depths and durations otariids dive, max depth and duration are the maximum depth and duration recorded by any otariid and TASD is the percentage time at sea spent diving.
dive time. We ran the simulation for the estimated number of foraging trips conducted on average each year for each species dive behaviour group. Average number of foraging trips was estimated from average duration spent at sea (days) and average colony duration (days). For example, the average number of dives that exceed the cADL for the Australian fur seal is estimated by first calculating theta using an average dive time of 2.9 min for an average of 849 dives per foraging trip from 13 seals. Theta is then used to simulate the dive distribution for a given foraging trip with a maximum dive duration of 6.7 min, where the number of dives exceeding the cADL of 2.5 min. The simulation is repeated 59 times, the average number of foraging trips conducted per year. The final value used is the mean value of proportion of dives that exceed the cADL of the simulated foraging trips.
All analyses were conducted in R and, before any parametric testing was conducted, all relevant assumptions (i.e. homogeneity of variances and normality) were tested and met. The summarized data for each species were used rather than individual values as the data for individuals were rarely available, and it has been shown that the conclusions and effect sizes made from summary data are very similar to those made with individual data (Steinberg et al., 1997;Tudur Smith et al., 2016). Two-way analysis of variance (ANOVA) for summary data (Cohen 2002) were used to look for statistical differences and interactions between and within foraging mode (benthic and meso vs. epipelagic) and morphology group (sea lion vs. fur seal). To implement the ANOVAs, the mean, standard deviation and sample size were included in the formula and implemented using the function ind.twoway.second() in the R package rpshychi (Okumura and Okumura 2012).
Pearson's correlation tests were used to examine the relationship between all the response variables (mass, depth, duration, TASD, TBOS, DMR, cADL, dive performance and percentage of dives that exceed the cADL). Dive performance is measured as the ratio of mean dive duration to cADL (Arnould and Costa 2006).

Results
Values used to calculate the following results are in Tables 1-4 and the summary statistics derived related to the following results are in Table 5   Information has been collated from published reports of diving parameters for species listed. N is the sample size, mass is the average mass of animals studied, TBOS is the total body oxygen stores, which have either been estimated from other species (Estimate) or from blood samples (Bloods), cADL is the calculated ADL and FMR is the field metabolic rate, which has either been estimated from other species (Estimate), from doubly labelled water (DLW) or from respirometry (Respirometry).
TBOS varied both by morphology group (Table 5), with sea lions having larger TBOS than fur seals (x =48.74, SE = 3.34 ml O 2 min kg 0.75 vs x =45.54, SE = 3.10 ml O 2 min kg 0.75 ) and by foraging strategy with epipelagic divers having larger TBOS (x =47.86, SE = 3.14 ml O 2 min kg 0.75 vs x =45.68, SE = 3.22 ml O 2 min kg 0.75 ). No variation could be explained by the interaction between the two ( Table 5). A small amount of variation in DMR could be explained by foraging strategy (Table 5) Both foraging strategy, morphology and their interaction accounted for variation in dive depth and duration (Table 5).
Morphology accounted for a large amount of the variation in depth, while foraging strategy accounted for a large amount of variation in duration (Table 5). There was a large effect of morphology on TASD and a small interaction effect of morphology and foraging strategy (Table 5) (Table 5). The ratio was larger for benthic and mesopelagic divers (benthic and mesopelagic x= 1.13, SE = 0.11 vs epipelagic x = 0.76, SE = 0.07). The proportion of dives that exceed the cADL varied largely by foraging strategy (Table 5) where benthic and mesopelagic divers were far more likely to exceed the cADL (benthic and mesopelagic x = 40.75., SE = 6.16% vs epipelagic x = 26.90, SE = 3.13%).
There was no relationship between body mass and dive depth or dive duration (Pearson's correlation test: P > 0.05 for both tests). There was no significant correlation between

Review article
Conservation Physiology • Volume 8 2020  Information has been collated from published reports of diving parameters for species listed. N is the sample size, mass is the average mass of animals studied, TBOS is the total body oxygen stores, which have either been estimated from other species (Estimate) or from blood samples (Bloods), cADL is the calculated ADL and FMR is the field metabolic rate, which has either been estimated from other species (Estimate), from doubly labelled water (DLW) or from respirometry (Respirometry).
TBOS and dive depth or duration (P > 0.05). A positive correlation was found between dive performance (the ratio of dive duration to cADL) and dive depth (correlation coefficient = 0.46, P = 0.04; Fig. 2). Similarly, there was a positive correlation between percentage of dives that exceed the cADL and dive depth (correlation coefficient = 0.33, P = 0.18), where benthic and mesopelagic divers were more likely to exceed their cADL (Fig. 3).
Overall, the probability of exceeding the cADL on any given dive was 1.5 times more likely for benthic and mesopelagic diving animals than for an epipelagic diver. Dive durations of epipelagic divers were predominantly less than half the duration of their maximum dive time (Figs 3  and 4). For benthic and mesopelagic diving animals, the distributions were more normally distributed (Figs 3-5). Benthic diving Northern fur seals were most likely to exceed their cADL, diving beyond the estimated threshold on 62% of dives.

Comparison of multi-strategy animals
There were three species that foraged using both a benthic or mesopelagic strategy and an epipelagic strategy [Northern fur seals (Fig. 3), Southern sea lions and Galapagos sea lions (Fig. 4)]. Many of the benthic dives exceed the cADL for Northern fur seals (62%), while most pelagic and benthic dives were within the cADL for Southern sea lions (68-71%) and Galapagos sea lions (78-82%).

Discussion
Global changes to ocean ecosystems are predicted to result in range shifts for many otariid prey species Bakun et al., 2015). For predators, this means that they may need to increase their foraging effort in order to gain sufficient energy, as failing to meet energetic demands directly impacts survival (Boyd et al., 1994). Diving mammals may be required to forage deeper and for longer , and/or target larger demersal prey that are predicted to be less influenced by changes in the ocean ecosystem (Perry et al., 2005). It has been hypothesized that species that normally dive within their ADL can increase foraging effort with fewer consequences by drawing on oxygen reserves to pursue prey at depth (Boyd et al., 1994;Costa et al., 2004). However, those species already operating at or near their physiological maximum may not have a similar capacity to increase foraging effort.
By reviewing the physiological (DMR, TBOS and cADL) and behavioural (dive depth, duration and TASD) information of females from every extant otariid, we have shown that foraging mode (benthic and mesopelagic or epipelagic), rather than morphology (fur seal or sea lion), is more indicative of which species operate at or near their physiological maximum. Crucially, we reviewed abilities for females that are not only more likely to be physiologically constrained, but are the limiting sex, directly influencing the size of populations. We found that female otariids that dive using a benthic or mesopelagic strategy forage deeper and longer and exceed    Figure 1: Interaction plots for the means ± SE of diving (depth, duration, TASD) and physiological (TBOS, DMR, cADL, ratio and % dives exceeding the cADL) variables for foraging strategy (benthic and mesopelagic vs epipelagic) and morphology (fur seal-circles, sea lion-triangles) for female otariids their cADL 1.4 times as frequently as do epipelagic divers, regardless of whether they were a fur seal or sea lion. This may make the former more vulnerable to environmental changes that cause prey to move deeper or further offshore as their physiological scope for adaptation to diving deeper and longer is limited (and switching foraging strategies from benthic to epipelagic would not help them find prey). In contrast, as epipelagic divers generally do not currently exceed their cADL, they may have more scope to switch to a benthic foraging strategy; although if they do, the likelihood of exceeding their cADL substantially increases (Northern fur seal- Fig. 3). Therefore, switching foraging modes during periods of low prey abundance is not a risk-free solution to finding additional energy during these times.
Our study confirms that foraging strategy is a better predictor than morphology (fur seal vs sea lion) for identifying species operating near their physiological maximum and accords with earlier theorizing Arnould and Costa 2006). We found that foraging strategy accounted for variation in more behavioural parameters than morphology, as either the sole predictor of variation (cADL ratio and depth) or as the interaction between morphology and foraging strategy (dives exceeding cADL, duration and TASD). In fact, morphology could only directly account for variation in one behavioural parameter, depth. Benthic and mesopelagic divers were more likely to exceed their cADL on a given dive, as demonstrated by an average cADL ratio (cADL/dive duration) greater than one. They dived to deeper depths for longer and, despite having relatively low DMR's, also had smaller TBOS that limited their ability to dive longer within their cADLs (Fig. 1). Deeper dives were related to larger cADL ratios meaning that, as they dived deeper, they were more likely to exceed their cADLs. Benthic diving sea lions were most likely to dive for longer on average than their cADL (Fig. 2), while most of the epipelagic foragers dived on average for durations less or equal to their cADL.
The otariids that are currently operating closest to their physiological maximum are benthic and mesopelagic diving sea lions-the Australian sea lion, the New Zealand sea lion Chilvers et al., 2006;Fowler et al., 2007;Villegas-Amtmann and Costa 2010)  57% of their dives (Fig. 4). They have previously been identified as operating near their physiological maximum Chilvers et al., 2006;Fowler et al., 2007). The tendency to exceed their cADL is likely to be a contributing factor to the slow recovery of populations of Australian and New Zealand sea lions, which all routinely exceed their cADL (Boyd and Croxall 1996;Chilvers et al., 2006;Fowler et al., 2007).
All other sea lion species demonstrate mixed strategies (Fig. 4). Sea lions are typically much larger than fur seals (Tables 6 and 7), and larger animals theoretically have the ability to make longer and potentially deeper dives (Villegas-Amtmann and Costa 2010; Weise et al., 2010). However, we found that body mass was not correlated with the three behavioural parameters (TASD, dive depth and dive duration) nor the physiological parameters (DMR, cADL and TBOS). It is not surprising that neither TBOS or DMR were related to body mass, as we used scaled values in our analyses, which also permits us to identify trends that were independent of body mass. For example, compared to other sea lion species, Galapagos sea lions have the highest massspecific body oxygen stores [Villegas-Amtmann and Costa (2010), this study]. Galapagos sea lions are shown to have plasticity in their diving abilities linked to physiological (Villegas-Amtmann et al., 2008) and environmental (Jeglinski et al., 2015) differences. Individuals that specialize as benthic divers have higher TBOS than those classified as pelagic divers (Villegas-Amtmann and Costa 2010) and had different diet compositions as influenced by their foraging habitat (Jeglinski et al., 2015). Similarly, for Southern sea lions, the longest and deepest diving animals had significantly larger TBOS than the shallowest and shortest duration divers (Hückstädt et al., 2016) and the habitat they utilized also influenced their foraging behaviour (Baylis et al., 2015b). Studies on the California sea lion demonstrate that there was not a significant cost of using a benthic diving strategy as the at-sea FMR did not differ for a deep diving or mixed strategy McHuron et al., 2018). Perhaps because of these physiological adaptations, these benthic diving sea lions were no more likely to exceed their cADL than pelagic divers. Though, future work should investigate the complex linkages between foraging strategy, environmental gradients and physiological constraints that influence the adoptation and change of foraging strategies (Jeglinski et al., 2015).
The species least likely to be operating near their physiological maximum are the epipelagic diving fur seals. This includes the Antarctic, subantarctic, Galapagos, Guadalupe, Juan Fernandez and New Zealand fur seals. They are least likely to exceed their cADL (17-31% of dives, Fig. 3) and, with the exception of the Galapagos and Antarctic fur seal, have increasing or stable populations and are classified as least concern, despite all having been historically driven nearly to extinction by the fur trade (Table 6). Galapagos fur seal population dynamics are likely constrained by external forces, such as El Niño and fisheries, rather than physiological limitations related to foraging efficiency (Edgar et al., 2010). Epipelagic fur seals dive to shallower depths for shorter durations and spend less TASD than benthic or mesopelagic foragers. However, it is unclear whether all epipelagic divers could or would adopt a benthic or mesopelagic foraging strategy. For example, the Galapagos fur seal is unlikely to exceed its cADL on a given dive (26%). Even in times of increased competition, in years of limited food availability and mass starvation, there was no evidence that these fur seals switched foraging strategies (Horning and Trillmich 1997). When Antarctic fur seals switched from primarily short and shallow pelagic dives to deep and long mesopelagic dives during periods of prey shortages, they increased the probabilty of exceeding their cADL on any given dive from 13.6 to 35.2% (Boyd et al., 1994). The data available from the literature allowed us to test the impact of foraging strategy switching with one fur seal species-the Northern fur seal. There results demonstrate that the Norther fur seal exceeds its cADL when using a benthic diving strategy more often than not. Northern fur seal populations are currently in decline, seemingly in part related to changing prey distributions (Kuhn 2011). These environmental changes are likely to result in a change in Northern fur seal diving behaviour, pushing them to dive deeper and for longer (Kuhn et al., 2010), and if that is not available to them, increasing their trip durations (Georges et al., 2000a;Soto et al., 2006).
Grouping otariids by species and foraging strategy does limit how far we can draw conclusions from our data. For each combination of species and foraging strategy, a single data point representing their dive and physiological parameters must be selected to be representative, even though these may have been measured several times. We are also limited by the behavioural and physiological variables that we have included. Other behavioural variables such as time spent at sea, or physiological variables such as body fat or thermoneutral zones, may be important for developing a deeper understanding of the differences in species or foraging strategies. In addition, we did not explore the impacts of environmental variables such as oceanography (Jeglinski et al., 2015), climate change (Simmonds and Isaac 2007) or prey distributions (Horning and Trillmich 1999;Kuhn et al., 2015) that are known to influence these parameters. Future work should seek to include these important variables.
An inability by some populations to display plasticity results in other responses to food shortages such as increasing the duration of foraging trips, (Boyd et al., 1994) or spending more TASD (Georges et al., 2000b), rather than increasing the depth or duration of dives. However, there is little evidence that for epipelagic foragers an increase in foraging effort can compensate for large-scale environmental changes. Despite overall increases in foraging effort, the growth rate of Antarctic fur seal pups reflects the food availablity of the year, where low food availability corresponds to poor growth and overall lower survival of pups (Trillmich et al., 1991;Boyd et al., 1994;McCafferty et al., 1998). More widely, demonstrable environmental perturbations causing changes in fish abundance and latitudinal shifts in many of the ecosystems otariids inhabit are presenting them with new challenges, including prey scarcity and indirect resource competition (Bakun et al., 2015;Sydeman et al., 2015;Carroll et al., 2016). The adaptive capacity of marine mammals depends, in part, on their ability to change their diet and or foraging behaviour in the face of these challenges. Prey becoming scarcer and more patchily distributed means that mothers may need to forage further from their rookeries or dive deeper and longer to obtain enough food to survive and provide for their pups (Boyd et al., 1994). Foraging further from the rookery may have consequences on pup survival as the pups fast while mothers are at sea Harcourt et al., 2002), while increasing effort by diving deeper for longer is only an option if it is within the physiological capacity of the individual.
This study supports the theorizing of Arnould and Costa (2006) and Costa et al., (2004) that some species may already be operating at their physiological maximum and therefore do not have the capacity to further adapt their foraging behaviour to a changing ecosystem. Here, the original conclusions by Arnould and Costa (2006) and Costa et al., (2004) have been expanded, showing that switching from a pelagic to a benthic or mesopelagic foraging strategy significantly increases the likelihood of exceeding the cADL. Further, these results show that sea lions that switch strategies have physiological adaptations to do so. This study also provides further evidence that the Australian sea lion and New Zealand sea lion are also operating near their physiological limits. Where there is considerable variation in the experimental environment, variation among individuals combined with small sample sizes, it is useful to retest hypotheses to ensure that they stand up once additional studies add more species and increase the sample size. These are important findings to reevaluate with a rapidly changing climate that has increased the pressures these animals face, and the need for considered conservation measures are more urgent now than ever before.

15
We caution that species thought to be able to change foraging strategy may not be able to do so due to the high cost of deep diving to undertake benthic and mesopelagic foraging. In the face of rapidly changing coastal ecosystems, female otariids face increasing constraints due to the central place foraging requirement arising from their income breeding strategy, that is, the need to return to feed their pup. For those species not able to push their operating constraints further, there may be untoward population consequences.

Funding
This work was supported by a Macquarie University Research Excellence Support grant awarded to M. Ladds.