OUP user menu

Hibernation and seasonal fasting in bears: the energetic costs and consequences for polar bears

Charles T. Robbins, Claudia Lopez-Alfaro, Karyn D. Rode, Øivind Tøien, O. Lynne Nelson
DOI: http://dx.doi.org/10.1644/11-MAMM-A-406.1 1493-1503 First published online: 17 December 2012

Abstract

Global warming has the potential to reduce arctic sea ice and thereby increase the length of summer-fall fasting when polar bears (Ursus maritimus) lose access to most marine mammals. To evaluate the consequences of such changes, we compared the cost of fasting by polar bears with hibernation by brown bears (U. arctos), American black bears (U. americanus), and polar bears and made projections about tissue reserves polar bears will need to survive and reproduce as fasts become longer. Hibernating polar bears expend energy at the same rate per unit mass as do brown bears and black bears. However, daily mass losses, energy expenditures, and the losses of lean mass are much higher in fasting, active polar bears than in hibernating bears. The average pregnant polar bear living around Hudson Bay during the 1980s and 1990s could fast for 10.0 ± 2.3 months (X̄ ± SD), and the average lactating female with cubs born during the preceding winter could fast for 4.2 ±1.9 months. Thus, some pregnant or lactating females with lower levels of body fat content were already approaching or beyond the constraint of being able to produce cubs and survive the required 8 months of fasting if producing new offspring or 4 months if accompanied by older offspring. Pregnant or lactating females and their dependent offspring have the most tenuous future as global warming occurs. Thus, we predict a significant reduction in productivity with even modest increases in global warming for polar bears living in the very southern part of their range and are concerned about more northern populations depending on their ability to accumulate increasing amounts of fat.

Key words
  • bear
  • black bear
  • brown bear
  • fasting
  • hibernation
  • polar bear

Prolonged seasonal fasts during periods of food scarcity are an important part of the life strategy of many animals, including Holarctic bears (Atkinson et al. 1996), many rodents (Drew et al. 2007), and many marine mammals and birds (Verrier et al. 2009). Brown bears (Ursus arctos) and American black bears (U. americanus) fast while denned during the winter. Denning or hibernation in these species minimizes activity, mass and energy loss, and protein catabolism (Atkinson et al. 1996; Barboza et al. 1997; Harlow et al. 2002). The most extensive seasonal fast in polar bears (U. maritimus) occurs during the summer and fall in the more southern portions of their range, where sea ice disappears seasonally and bears lose access to most marine mammals (Gleason and Rode 2009; Ramsay and Stirling 1988; Rode et al. 2010a). Polar bears do not den during this time, although activity can be greatly reduced (Lunn and Stirling 1985). Pregnant females, however, must subsequently den during the winter to produce cubs. Thus, the combination of fasting during the summer and fall, often referred to as “walking hibernation,” and denning in the winter by pregnant female polar bears can mean going without food for as long as 8 months (Atkinson and Ramsay 1995; Ramsay and Stirling 1988).

Global warming is increasing the duration of ice-free conditions in many areas (e.g., Hudson Bay and southern Beaufort Sea—Amstrup et al. 2010; Stirling and Parkinson 2006). The 120-day fast that was typical of males and nonpregnant female polar bears living around Hudson Bay during the 1980s is expected to increase to 180 days as warming occurs (Molnár et al. 2010). Molnár et al. (2010) estimated that 3% of the adult male population at Hudson Bay dies with a 120-day fast, but that mortality increases to 28% of the population if fasting extends to 180 days. This potential change in polar bear ecology that will both increase the length of fasting and decrease the time to accumulate tissue reserves dictates the need to understand better the energy and matter costs for polar bears (Hunter et al. 2010; Rode et al. 2010a; Stirling et al. 2011).

Several studies have provided significant insight into the costs of fasting and hibernation in bears (Atkinson et al. 1996; Atkinson and Ramsay 1995; Farley and Robbins 1995; Harlow et al. 2002; Maxwell et al. 1988; Polischuk et al. 2002; Tøien et al. 2011; Watts et al. 1987). Unfortunately, costs estimated from these studies either have not agreed or are not easily compared. For example, the average energy costs measured recently during mid-hibernation in black bears (Tøien et al. 2011) are much lower than what would be estimated by the mass losses summarized by Farley and Robbins (1995) for hibernating brown bears and black bears. Similarly, although energy conservation is assumed to occur to some degree in fasting polar bears on the basis of evidence of protein conservation (Derocher et al. 1990; Harlow et al. 2002) and similarities in blood chemistries with hibernating brown bears and black bears (Nelson et al. 1983), no one has compared the costs of hibernation in brown bears, black bears, and polar bears with the summer–fall fasting of polar bears.

We hypothesized that hibernation costs for brown bears, black bears, and polar bears will be similar when measured under comparable conditions and expressed on a per unit mass basis, but those costs will be less than that for summer–fall fasting by polar bears. Because of the potentially elevated costs for fasting by polar bears relative to that occurring during hibernation, global warming and decreasing sea ice may have a much greater detrimental effect on polar bears than would extending hibernation in other bears that do not have a summer–fall fast. Consequently, we compare the costs of hibernation in brown bears, black bears, and polar bears, compare those costs with summer–fall fasting in polar bears, and make projections about tissue reserves that polar bears will need to survive and reproduce as fasts become longer.

Materials and Methods

Housing and diets for captive brown bears.—We used 10 captive brown bears 1–23 years old and housed at the Washington State University Bear Research, Education, and Conservation Center in 2003–2011. The cessation of feeding and foraging started on October 21 when all bears were closed into their individual dens and runs, and feeding resumed in mid- to late March. We chose these times because appetites and, therefore, the amount of food offered by us had dropped significantly by late October. When brown bears were deprived of food during the active season, they showed very obvious signs of discomfort (i.e., growling, roaring, and extreme aggression). When food was removed in late October, the bears exhibited silence and indifference. Similarly, bears had become increasingly active by late February when daytime temperatures were noticeably warmer. Thus, both October and March behaviors indicated that the bears were going into and coming out of hibernation (Nelson and Robbins 2010).

Bears hibernated either singly or in pairs in individual 3-m × 3-m concrete dens and were given a bale of straw to form a bed. Drinking water was always available. Dens were connected to 3-m × 6-m outside, concrete-floored runs. The doors between the dens and outside runs were left open so each den was illuminated by natural light. Bears were never closed into their dens and always had access to the runs, but they could not go into the large 0.56-ha exercise yard during hibernation (Rode et al. 2001). The overall den area was closed to human entry during hibernation to minimize disturbance except during infrequent entries for weighing.

Animals used in this study were cared for in accordance with past and current animal care and use guidelines approved by the American Society of Mammalogists (Sikes et al. 2011). This project was approved by the Washington State University Institutional Animal Care and Use Committee under protocol #03054-003.

Measuring activity patterns in hibernating brown bears.—Activity patterns were monitored and used as an indicator of when maximum energy conservation was occurring during hibernation. Cameras (MGB600 high-resolution black and white with QLAV2 2.6–6-mm lens in V28CC housing; Silent Witness, Surrey, British Columbia, Canada) and infrared illuminators (Silent Witness model SWIR-24 volt AC) were installed in each den and outside run. The cameras were connected to a computer and video screen (X240 16-channel high-definition digital recorder with 500 gb hard drive; Open Eye, Spokane, Washington) distant to the den that permitted storage and replay of previously recorded images. Activity was defined as either lying or other (e.g., standing, walking, grooming, or rearranging bedding material).

Measuring mass and estimating energy lost during hibernation or fasting in captive brown bears, American black bears, and polar bears.—At the start and after various stages of hibernation in captive brown bears, we measured body mass of each bear using electronic scales weighing to the nearest 0.5 kg. Although composition of the mass lost in the captive brown bears was not measured during this study, previous studies on these animals indicated that mass lost between December and February (i.e., hibernation) was solely fat (Barboza et al. 1997). Bear hibernation is known for its minimal to no urination and defecation and, therefore, protein conservation when fat stores are plentiful (Atkinson et al. 1996; Barboza et al. 1997; Fedorov et al. 2009; Maxwell et al. 1988; Tøien et al. 2011). Similarly, because drinking water was always available and bears were observed drinking, dehydration should not have occurred. Thus, the mass lost was multiplied by 9.4 kcal/g (i.e., fat) to estimate energy loss. We did not estimate energy losses during the transitions between the active and hibernating states because of the unknown composition of that loss.

We do not know of any published values on the masses of wild brown bears during hibernation that could be compared with values generated with the captive bears. Wild brown bears are not handled during hibernation in North America, and the few brown bears that have been captured, weighed, and body composition determined in the fall and spring are either not immediately going into hibernation or have been out of the den for some unknown period of time. Thus, a direct comparison of mass losses between captive and wild brown bears is not possible currently.

Wild black bears have been weighed frequently during hibernation (summarized by Harlow et al. 2002). However, mass losses are variable and often very high relative to what would be predicted if fat was the primary tissue being catabolized. These problems also may occur in captivity when bears are not provided access to water or are relatively lean at the beginning of hibernation. For example, the caloric equivalents of the mass loss in captive polar bears (Watts et al. 1987), brown bears (Watts and Jonkel 1988), and black bears (Watts and Cuyler 1988) during hibernation ranged from 3.2 to 8.9 kcal/g (Watts 1990), which suggests significant water loss due to dehydration or lean mass catabolism at the lower caloric equivalents (Maxwell et al. 1988). The various studies by Watts and colleagues indicated that water or snow were provided ad libitum before denning, but the den was sealed during hibernation and there was no indication that water or snow were subsequently available. We have observed dehydration in our captive brown bears if not provided access to water. Although it has been assumed that water losses during hibernation are balanced by water produced from fat catabolism (Nelson et al. 1983), modest dehydration may be a natural part of hibernation in wild bears. Thus, many of the mass loss values on wild and captive bears probably include more than fat catabolism, which makes using those values problematic in estimating the cost of hibernation.

However, energy expenditures during mid-hibernation by 2 captive brown bears, 7 black bears, and 3 polar bears have been measured using indirect calorimetry (Tøien et al. 2011; Watts and Cuyler 1988; Watts and Jonkel 1988; Watts et al. 1987). These values would not be affected by modest changes in water balance. However, the energy expenditures generally reported have been the least or minimum observed metabolic rates measured when no sound or movement was occurring in the den (Tøien et al. 2011; Watts and Cuyler 1988; Watts and Jonkel 1988; Watts et al. 1987). Nevertheless, they are useful as a baseline for estimating what hibernation metabolic rates would be when measured over longer periods of time that might include costs due to thermoregulation and other activities that occur during hibernation, such as grooming, sitting, standing, or rearranging bedding material.

Limited data on masses and characteristics of mass lost during ice-free (i.e., fasting) conditions for wild polar bears living in northeastern Manitoba, i.e., Hudson Bay, are available (Arnould and Ramsay 1994; Atkinson et al. 1996; Atkinson and Ramsay 1995; Polischuk et al. 2002). Those data on wild polar bears provide the opportunity to compare mass losses, energy costs, and efficiencies of fasting and hibernation with those for hibernation measured on captive brown bears, black bears, and polar bears. Losses of fat and protein in male polar bears and females producing twins were multiplied by 9.4 kcal/g fat and 5.64 kcal/g protein to estimate energy loss (Atkinson et al. 1996). Because polar bears remain active during the summer–fall fast, they have the opportunity to consume terrestrial foods (Dyck and Ermias 2009; Dyck and Romberg 2007), although such ingestion is considered by many authors to contribute minimally to the total energy budget of most polar bears (Rode et al. 2010b; Schliebe et al. 2008). Nevertheless, energetic costs estimated from tissue loss for polar bears are net costs that would be the difference between total energy expenditure minus ingested energy. To provide a common baseline for comparison, we compared brown bear, black bear, and polar bear energy expenditures during hibernation or fasting to the allometric equation of basal metabolic rates (y = 61.9x0.77, where y is kcal/day and x is animal mass in kg—data from McNab 2008) measured on 61 species of carnivores, ranging from 77-g least weasels (Mustela nivalis) to 388-kg Weddell seals (Leptonychotes weddelli). Although McNab (2008) used metabolic rates measured during hibernation in polar bears, brown bears, and black bears in his original allometric equation, those metabolic rates were not used in the above equation.

Estimating survival times and consequences of increased polar bear fasting.—During prolonged fasting, animals go through several phases that are defined in part by the tissues being mobilized to meet their energy requirements (Caloin 2004; McCue 2010). Phase I lasts for only a few hours to days as animals transition from the fed to fasting states by decreasing energy and protein metabolism and using residual food in the gastrointestinal tract and liver and muscle glycogen to meet energy needs. Phase II, which is the longest in obese animals, is characterized primarily by fat utilization and relatively minimal loss of protein. As fat stores are depleted, the animal begins phase III, in which protein utilization increases. This phase, if prolonged, will result in death as the animal exhausts critical structural and metabolic proteins. Even though starved animals often have <1% body fat at death, protein depletion is the immediate cause of death (Caloin 2004).

To estimate survival times and critical stores of fat necessary for various lengths of fasting, we developed a deterministic model with a daily time-step (Fig. 1). The model was programmed in Stella Modeling and Simulation Software 8.1. Inputs were initial body mass, initial body fat content, and reproductive status. Initial masses ranged from 100 to 500 kg in 100-kg units, and initial body fat contents for each mass ranged from 10% to 50% in 10% units to cover the range of masses and fat contents occurring in subadult and adult polar bears (Atkinson et al. 1996; Atkinson and Ramsay 1995; Polischuk et al. 2002). At the beginning of each simulation, the mass-specific energy cost was estimated on day 1 for summer–fall fasting by males and barren females or the combined summer–fall fasting and winter hibernation–lactation by adult females producing cubs. These estimated energy expenditures include all energetic costs incurred while polar bears fast or hibernate and produce cubs because they are based on tissue and energy losses observed in wild polar bears (Atkinson et al. 1996; Atkinson and Ramsay 1995; Polischuk et al. 2002). The daily energy and mass costs were subtracted from the starting amount of fat and lean mass available on that day to estimate the next day's body mass, energy expenditure, and available tissue reserves. This iteration process of calculating daily energy expenditure and subtracting it from available energy reserves was repeated until all available reserves were exhausted.

Fig. 1

Diagram of the model used to estimate survival time and mass loss of fasting and hibernating–lactating polar bears. On the basis of the inputs of body mass, body composition, and the composition and energy content of the mass loss, daily energy expenditure and the amount of mass lost were calculated. A new body mass and available tissue energy were calculated at the start of the next day and the process repeated at a daily time-step until all available fat and lean mass were lost and the hypothetical bear died or, if pregnant, failed to implant embryos.

We assumed that all fat and 30% of the lean mass (or protein) were available and would be exhausted at death. The amount of lean mass lost at death can vary between species and with level of initial obesity (Cherel et al. 1992). Because it has never been measured in polar bears, we used 30% as a reasonable, conservative estimate on the basis of studies in other animals (Caloin 2004). Any error created by under- or overestimating the amount of lean mass that can be lost before death over a prolonged fast will be relatively small in that fat (9.4 kcal/g) has ∼7 times more energy/g than does lean or fat-free mass (1.25 kcal/g) that is composed of 22.2% protein (5.64 kcal/g), 73.4% water, and 4.4% minerals (Atkinson and Ramsay 1995; Farley and Robbins 1994; Harlow et al. 2002; Molnár et al. 2009). Our inclusion of lean mass loss by fasting polar bears differs from the very efficient protein conservation that occurs during hibernation in relatively fat brown bears and black bears (Barboza et al. 1997; Harlow et al. 2002; Lohuis et al. 2007). However, black bears will mobilize lean mass as fat reserves are depleted (Maxwell et al. 1988), and polar bears lose significant lean mass during summer–fall fasting (Atkinson et al. 1996; Atkinson and Ramsay 1995).

We also used the model to estimate the number of days of fasting that would be required to reduce the body fat content from higher levels to 20%, which likely approximates the body fat content at which pregnant females will not implant developing embryos (Robbins et al. 2012). This estimate may be particularly important to populations where starvation is less of an issue, but fasts are long enough to reduce female condition below critical thresholds for implantation or adequate milk production necessary for survival of dependent offspring (Derocher et al. 2011; Ferguson et al. 2000; Schliebe et al. 2008).

Although tissue losses and therefore energetic costs of fasting have been reported for multiple age and sex classes of polar bears (i.e., subadult males <8 years old, adult males >8 years old, and pregnant females >4 years old), it has not been reported for lactating females that produced cubs the preceding winter (Arnould and Ramsay 1994; Atkinson et al. 1996; Atkinson and Ramsay 1995; Polischuk et al. 2002). This group is particularly important because at Hudson Bay, Foxe Basin, Baffin Bay, and Davis Strait, these bears have the least amount of time between when they go back out on the ice to hunt seals at the end of the winter and when they come back on land during the summer. Thus, this group of females in populations that must fast annually in successive, ice-free summers and falls may be the most likely to experience the earliest effects of global warming. Because of the limited data on this group of bears, we estimated their minimal energetic cost as the sum of the basal fasting rate observed in nonreproductively active polar bears of the same mass plus the energy content of the milk that they produced (Arnould and Ramsay 1994).

Results

Activity, mass, and energy characteristics.—Activity by all fasting, captive brown bears declined after feeding ceased in late October and reached a minimal level during mid-December, January, and early February (Fig. 2a). During those 9 weeks, the bears were inactive for 98.2% ± 0.5% (X̄ ± 1 SD) of the 24-h day, or active for approximately 24 min/day. Lactating females were even less active after birth in early January, as they did not stand during the first 3 weeks postpartum, and stood for only 7 ± 6 min/day during the next 5 weeks. The greatly reduced movement by lactating females presumably occurs because continuous maternal contact may be essential for neonatal survival when born in dark snow- and ice-covered dens. Cubs are born with minimal hair, cannot crawl or walk, and do not have functioning eyes until at least 40 days when their eyelids open. Consequently, movement by lactating females that threatened to dislodge their neonates elicited screaming by the neonates and maternal attention. Lactating females exited hibernation approximately 2 weeks later than did nonlactating bears.

Fig. 2

a) The relative proportion of the time that captive brown bears (Ursus arctos) spend lying in the den during hibernation once deprived of food at the Washington State University Bear Research, Education, and Conservation Center. b) Mass lost in captive brown bears during the transition between active and hibernating states (late October to mid-December and mid-February to first feeding), during hibernation when little activity occurred (mid-December to mid-February), and during the total time including the transitions and hibernation. c) Energy lost or expended by captive brown bears during hibernation estimated from the mass loss multiplied by the caloric content of fat (i.e., 9.4 kcal/g) relative to the allometric equation for basal metabolic rates of carnivores (data from McNab 2008). Energy loss during transitioning into and out of hibernation could not be estimated because the composition of the mass loss was unknown.

The daily mass lost by brown bears during the transition periods of either decreasing or increasing activity ranged from 2 (large bears) to 4 (small bears) times higher than during actual hibernation (Fig. 2b). Thus, the total mass lost during the transition and hibernation is a weighted average of these 2 different processes. Estimated daily energy expenditure for hibernating brown bears ranged from 30% of that predicted by the allometric equation for basal metabolic rates in carnivores for a 25-kg bear to 62% for a 300-kg bear (Fig. 2c).

However, the least-observed metabolic rates of hibernating black bears, brown bears, and polar bears are even less than the average hibernating metabolic rates. For example, the regressions for the calorimetrically measured least-observed metabolic rates (Fig. 3) and the metabolic rates of brown bears estimated from the caloric content of the weight loss (Fig. 2c) have virtually identical exponents (1.09 and 1.06), but the regression constants (4.8 and 7.4) differ by 54%. That difference, which is based on an interspecific comparison across studies, is very similar to the intraspecific difference of 52% between the least-observed metabolic rate and the average metabolic rate observed in hibernating black bears (data of Tøien et al. 2011). Thus, the average energetic cost of hibernation when measured over longer periods of time is likely ∼50% higher than the least-observed metabolic rates reported by several authors (Tøien et al. 2011; Watts and Cuyler 1988; Watts and Jonkel 1988; Watts et al. 1987) and can be estimated by the equation: y = 7.2x1.09 where x is body mass (kg) and y is energy expenditure (kcal/day).

Fig. 3

Least-observed metabolic rates in hibernating polar bears, brown bears, and black bears when no sound or movement occurred in the den (data from Tøien et al. 2011; Watts and Jonkel 1988; Watts and Cuyler 1988; Watts et al. 1987) relative to the allometric equation for basal metabolic rates of carnivores (data from McNab 2008).

Fasting, nonhibernating polar bears weighing from 200 to 500 kg lost ∼1.6 times more mass per day than predicted for brown bears of the same mass during the transition phase and 2.4 to 3.2 times more mass per day than brown bears during hibernation (Figs. 2b and 4a). The energy lost by fasting, wild polar bears is very similar to the allometric equation for carnivore basal metabolic rates (Fig. 4b) and, therefore, from 1.7 (500-kg bear) to 2.2 (200-kg) times higher than the least-observed metabolic rates of hibernating bears of the same mass (Fig. 3). The fat content of the mass lost by male polar bears ranged from 27% to 92% (49% ± 20%). The fat content of the mass lost by fasting, relatively fat female polar bears with twins was independent of the initial fat content and averaged 64% ± 5% (data of Atkinson et al. 1996; Atkinson and Ramsay 1995; Polischuk et al. 2002). The additional energy lost per day by females giving birth to and nursing twins averaged 19% ± 11% greater than that lost by males and, presumably, barren females of the same mass (Fig. 4b).

Fig. 4

Mass and energy lost by subadult and adult male polar bears (□; U. maritimus) and pregnant female polar bears (○) fasting during ice-free conditions at Hudson Bay (data from Atkinson and Ramsay 1995; Atkinson et al. 1996; Polischuk et al. 2002) in comparison with mass lost by hibernating brown bears and brown bears transitioning between active and hibernating states. Estimates of energy loss are based on the caloric content of the mass loss. Regressions for energy loss are compared with the allometric equation for basal metabolic rates of carnivores (data from McNab 2008).

Model estimates of survival.—Estimated survival was longer in larger polar bears than smaller ones at the same body fat content because of their decreasing energy requirement per kilogram (Figs. 4b and 5). Polar bears could theoretically fast without fat (i.e., depend solely on lean mass with a maximum allowable loss of 30%) for 24 to 35 days depending on initial body size and level of productivity. Thus, each 1% of lean mass lost would provide enough energy for 1 to 1.4 days of fasting, whereas each 1% increase or decrease in body fat content at the start of the fast provided enough energy for 6 to 9 days of fasting depending on initial body size and level of productivity (Fig. 5d).

Fig. 5

a) Estimated number of days that males and barren females and b) pregnant and lactating polar bears (U. maritimus) of various initial weights and body fat contents can survive while fasting. The lines for bears of various masses in each graph are for a) adult and subadult males and b) pregnant females that cover the range of observed masses that occur in these sex and age groups in the Hudson Bay population (data from Arnould and Ramsay 1994; Atkinson and Ramsay 1995; Atkinson et al. 1996; Polischuk et al. 2002). The data points for various ages and sexes of polar bears are based on the weight and body fat content observed in polar bears captured at Hudson Bay during the 1980s and 1990s. The data point for lactating females with offspring born the previous winter is below the survival curves for pregnant females because of the larger amounts of milk produced for older offspring. c) Estimated number of days that pregnant females of various initial weights and body fat contents can fast before body fat content is reduced to 20%, which likely approximates the body fat content at which pregnant females will not implant developing embryos (Robbins et al. 2012). d) Change in the estimated survival time as the initial body fat content increases or decreases by 1% at the start of fasting.

Discussion

Hibernation in brown bears, black bears, and polar bears is a very different process from summer–fall fasting in polar bears. Daily mass losses, energy expenditures, and the losses of lean mass are much higher in fasting, active polar bears than in hibernating bears. This rate of mass loss in fasting polar bears is even higher than that in brown bears transitioning between active and hibernating states. Estimated metabolic rates of fasting polar bears are characteristic of mammalian basal metabolic rates in general (McNab 2008) and are not characteristic of the level of energy conservation that occurs during hibernation. Thus, we recommend that the term “walking hibernation” (Nelson et al. 1983) not be used to describe summer–fall fasting in polar bears as it implies greater energy and protein conservation than actually occurs. Similarly, although polar bears may be able to consume some terrestrial food during ice-free conditions (Smith et al. 2010), the net cost of remaining active is still quite high relative to bears that hibernate.

The increased energetic cost of polar bears producing twins during hibernation relative to nonreproducing bears (19% ± 11% higher) is less than the difference occurring in brown bears and black bears (41% to 95% higher; Farley and Robbins 1995; Harlow et al. 2002). This may be due to 2 causes: the estimates of Farley and Robbins (1995) were for only the last 60 days of hibernation when lactation was occurring and for 120 days for Harlow et al. (2002) as compared with the longer intervals (i.e., 149 to 218 days) that included more non-lactation, less costly times over which the polar bear measurements were made; and female polar bears may be able to offset much of the additional cost of milk production by denning and, thereby, lowering the overall energy expenditure from the higher fasting levels characteristic of the summer–fall fast to lower hibernation levels. Indeed, the least-observed metabolic rates of denned polar bears do not differ from similar measurements on brown bears and black bears when expressed as a function of body mass (Fig. 3). Thus, although denning by pregnant polar bears provides a protected, warmer environment for birth and growth of the altricial cubs (Derocher et al. 2011), it also conserves energy (i.e., fat) that may be essential if cub production follows a prolonged fast.

The average adult male polar bear handled on the shores of Hudson Bay during the early ice-free period in the 1980s and 1990s (Atkinson et al. 1996; Atkinson and Ramsay 1995) had enough tissue reserves to fast for 8.0 ± 2.6 months (Fig. 5a). Even though adult males had a lower initial body fat content than pregnant females, their larger size and lack of reproductive effort are advantageous in reducing their daily energy expenditure relative to their reserves. Thus, few adult males would have had a problem surviving the 4-month fast that occurred in the 1980s, as also indicated by the estimated mortality of 3% of Molnár et al. (2010). However, we estimate that as many as 16% of the adult males would die if fasting lasted for 5.4 months (i.e., percent of the population below 1 SD of the mean estimated survival time). Thus, the current estimate of survival by adult males and those of Molnár et al. (2010) of 28% mortality with a 6-month fast as global warming occurs are in surprising agreement when one considers the very different models used (Molnár et al. 2009).

The average pregnant female during the same time period could have survived for 10.0 ± 2.3 months, and the average lactating female with cubs born during the preceding winter could have survived for 4.2 ±1.9 months (Fig. 5b). This difference is largely due to the lower body fat content of lactating females and the larger amounts of milk produced for older offspring. Nonetheless, some pregnant or lactating females with low levels of body fat content were already approaching or beyond the constraint of being able to produce cubs and survive the required 8 months to produce new offspring or 4 months if accompanied by older offspring. The additional cost of milk production by females accompanied by older offspring born during the preceding winter (2,607 ± 526 kcal day−1 female−1; data of Arnould and Ramsay 1994) is significant and predicted to reduce their survival time by at least 1/3 relative to nonproductive bears. Of course, we would expect both pregnant and lactating females without adequate reserves not to produce cubs, to reduce their milk production, or to abandon their cubs before jeopardizing their own survival (Regehr et al. 2010; Robbins et al. 2012). Although not producing cubs or abandoning cubs has obvious consequences, reducing milk production will reduce cub growth and size and thereby also reduce their survival (Derocher and Stirling 1996, 1998; Robbins et al. 2012).

Critical thresholds of body fat content for pregnant females to produce cubs relative to the length of fasting can be estimated with the current model. For example, Robbins et al. (2012) found that bears must have a body fat content >20% at the time of denning to produce cubs. We estimate that a pregnant female would have to have >34% body fat at the start of a 4-month summer–fall fast to have at least 20% body fat remaining when entering the maternity den (Fig. 5c). However, females that had much higher levels of body fat than 20% at den entrance produced larger cubs that had a better chance of survival than smaller cubs (Derocher and Stirling 1996; Robbins et al. 2012). Thus, pregnant polar bears at Hudson Bay relocated the following spring with cubs had an average body fat content of 42% ± 9% during the preceding summer–fall (Atkinson and Ramsay 1995). More northern populations that must adapt rapidly to global warming and the requisite summer–fall fasting that may not have occurred previously face a difficult challenge if females cannot accumulate such high levels of body fat.

Mortality of cubs between spring and summer or autumn of their 1st year in the Hudson Bay population during the 1980s averaged 47%, with lighter females more likely to lose their offspring than heavier females (Derocher and Stirling 1996, 1998; Ramsay and Stirling 1988). Survival of dependent offspring and subadults between 1984 and 2004 decreased during years when ice breakup and fasting began earlier by even a few days (Regehr et al. 2007). Thus, pregnant or lactating females and particularly their dependent offspring have the most tenuous future as global warming occurs. On the basis of the current analyses, significant portions of the Hudson Bay population were already approaching metabolic limitations to fasting in the 1980s and 1990s. Therefore, we predict a significant reduction in productivity with even modest increases in global warming and ice-free duration for polar bears living in the most southern portion of their range, and are concerned about more northern populations, depending on their ability to accumulate the increasing amounts of fat required to withstand longer summer–fall fasts before producing large, robust cubs (Hunter et al. 2010; Regehr et al. 2007, 2010).

Acknowledgments

Funding was provided by the Interagency Grizzly Bear Committee, United States Fish and Wildlife Service, the Raili Korkka Brown Bear Endowment, Bear Research and Conservation Endowment, and Nutritional Ecology Endowment at Washington State University, and the United States Army Medical Research and Material Command award W81XWH0920134. The findings and conclusions in this article are those of the authors and do not necessarily represent the views of the United States Fish and Wildlife Service.

Footnotes

  • Associate Editor was Roger A. Powell.

Literature Cited

View Abstract