1Nutrition influences most aspects of animal ecology: juvenile growth rates and adult mass gain, body condition, probability of pregnancy, over-winter survival, timing of parturition, and neonatal birth mass and survival. We provide an overview among ungulates of the extent of these influences resulting from interactions among bioenergetics, foraging, and nutritional demands.
2 Body condition of an animal is the integrator of nutritional intake and demands, affecting both survival and reproduction. The deposition and mobilization of body fat and body protein vary with physiological requirements and environmental conditions as species use dietary income and body stores to integrate the profits of summer and the demands of winter. Results from our simulation model and uncertainty analysis of the influence of body mass and changes in body composition of Rangifer over winter indicate that percent body fat rather than body mass in early winter is most important in determining whether animals die, live without reproducing, or live and reproduce. Animal responses are also sensitive to rates of change in body protein. Depending on timing of calving and maternal reserves, seasonal habitats vary in their nutritional value for the production of offspring.
3For free-ranging animals, life is a balance among numerous ecological factors, including nutritional requirements, nutritional resources to meet those demands, and intra- and inter-specific interactions. Predation effects on population demography may mask nutritional limitations of habitat. We suggest that over the long term of life histories, ungulates use seasonal strategies that minimize the maximum detriment, and that the basis for most strategies is nutritional.
Nutritional ecology is the science of relating an animal to its environment through nutritional interactions, whether those environments are natural ones or managed ones. Although the concept of nutrition is inherent in animal well-being, the implications of nutrition to management and conservation often are not realized. For free-ranging ungulates, nutritional interactions occur at the level of the individual, the population, and the ecosystem. Simplistically, individuals with access to high nutritional food resources often attain larger body sizes and better condition than individuals for which nutrition is inadequate, with subsequent influences on survival and reproduction. Wildlife populations with access to high-quality food often reach higher numbers than when limited by food. In contrast, severe nutritional limitation may provide selection pressure for the evolution of small size (Simard et al. 2008). At large ecosystem scales, population dynamics and feeding behaviour can stabilize or destabilize plant community structure (Hobbs 1996). In cases where ungulates limit the abundance and diversity of food supplies, there may be cascading effects that result in the reduction of other guilds or even the extirpation of other species including carnivores (Hazebroek et al. 1994; Focardi et al. 2000; Berger et al. 2001; Allombert et al. 2005; Côté 2005).
Considerable research efforts have been directed towards nutritional ecology of ungulates to define the relationships between the environment and the use of energy and nutrients by individuals and populations. Our nutritional knowledge base is broadest for cervids, such as arctic Rangifer tarandus (reindeer and caribou) and for north-temperate Odocoileus spp. (white-tailed deer, O. virginianus; mule deer, O. hemionus hemionus; black-tailed deer, O. h. sitkensis), Cervus elaphus (European red deer and North American elk), and Alces alces (moose). In addition to studies with specific nutritional emphasis, concepts of nutritional ecology are often embedded within studies of behavioural and population ecology, including long-term studies on red deer, roe deer (Capreolus capreolus), mountain goats (Oreamnos americanus), bighorn sheep (Ovis canadensis), chamois (Rupicapra rupicapra), and ibex (Capra ibex). In this review we present the strength of evidence for nutritional effects on survival and reproduction, and provide an overview of the interactions among bioenergetics, foraging, and nutrition in the field of ungulate ecology. We give examples from a variety of species, and place emphasis on Rangifer because that species is circumpolar, usually subjected to extreme seasonality, and likely to be most affected by large changes in climate.
Survival and reproduction depend on environmental constraints and the ability of animals to meet those demands with food and shelter. The need by management agencies and conservation biologists to define limiting factors for ungulate populations has led to numerous collaborations with physiological ecologists. Controlled mechanistic studies using captive or tamed animals at small scales define what animals are physically capable of doing, and thus provide insights for why free-ranging animals use their environments as they do. Field studies describe the strategies that animals use on different landscapes. Together, these two genres of study enable a better understanding of the physiological and behavioural adaptations of ungulates and the nutritional value of habitats.
We explore the following perspectives in our review and by providing a model for reproduction in Rangifer:
1Nutrition is the basis for annual production of ungulates and other animals.
2Nutrition has more than one currency (see Boggs 2009; Raubenheimer et al. 2009). Energy is a common currency for survival, but protein and other cellular components such as water and minerals may drive reproduction.
3Environmental change alters the amount and quality of food in the time available for production.
Nutritional constraints: biological states and environmental limitations
energy costs and intake
Nutritional constraints for free-ranging ungulates typically have been assumed to be energetic ones, largely because of environmental limitations induced by seasonally changing environments. Metabolic rates for the maintenance of body mass by northern ungulates are generally high compared to other ungulates (Hudson & Christopherson 1985; Hudson & Haigh 2002). In the arctic where food is abundant only over a very short season, higher metabolic rates enable tissue synthesis for growth and replenishment of body reserves during the short summer window (e.g., Lawler & White 2003). The lower energy requirements for maintenance by white-tailed deer in southern environments compared to more northern conspecifics serve as an adaptation to semi-arid environments with limited net primary productivity (Strickland et al. 2005). It is unknown whether seasonal cycles in basal metabolism, when corrected for previous intake, occur in all species (Mautz et al. 1992). When they do occur, it is unclear whether they are driven by photoperiod-induced appetite or by requirements for maintenance of different tissues and secretions because appetite and metabolic rate are linked to different hormones as well as influenced by the environment (Hudson & Haigh 2002).
Across species, lactation is the biological state when daily energy costs are highest for females. Energy requirements for female ungulates increase 65–215% during the first month post partum (Oftedal 1985; Robbins 1993). Before lactation, >90% of the energy requirements for gestating females occurs during the last trimester, and these costs are almost 50% higher for pregnant than non-pregnant animals (e.g., Pekins et al. 1998). Hence, highest energy costs for females occur from late winter to mid summer. For male ungulates, mass-specific seasonal energy requirements are typically highest during the autumn breeding period when animals are most active and when foraging is usually reduced, resulting in mass loss (e.g., bighorn sheep: Pelletier 2005; Himalayan tahr (Hemitragus jemlahicus): Forsyth et al. 2005; red deer: Yoccoz et al. 2002; Rangifer: Barboza et al. 2004). The consequences of reproductive effort were reviewed by Mysterud et al. (2004).
Variable and harsh weather augments seasonal energetic costs, with variable impacts depending on age structure of the population (Coulson et al. 2000) and potentially strong cohort effects (Fritz & Loison 2006). A weather-sensitivity hypothesis related to rain, strong wind, and low temperatures explained habitat segregation by male and female red deer (Conradt et al. 2000); the higher weather sensitivity of males due to higher energy losses relative to intake rates and the depletion of body reserves during the breeding season may make males more susceptible to the loss of forest cover. Rainfall during cool summer temperatures increased energy costs for black-tailed deer by 28% after 5 h in the rain (Parker 1989), and winter rainfall was a significant predictor of body condition in white-tailed deer (Garroway & Broders 2005). Although they are typically only small increases in metabolic cost, constant supplemental thermoregulatory expenditures can pose cumulative over-winter energetic drains, compounded with effects on daily foraging behaviours (e.g., feral goats (Capra hircus): Shi et al. 2003; kudus (Tragelaphus strepsiceros): Owen-Smith 2002). Managers have used lower and upper limits of thermoneutrality to define the importance of vegetative cover as thermal cover from heat and cold. Of note, the long-time prescriptions for retaining forested stands as thermal cover from temperature and wind for elk populations are now being revisited in light of the contrasting benefits of solar heat gain by wintering animals in non-cover areas (Cook et al. 1998). Increased energetic demands associated with cold temperatures, wind and deep snow are the long-standing impetus for the management of deer yards as shelter.
In regions with snowfall, snow depths directly influence the choice of traditional wintering areas, where energy costs are usually lower and food availability is higher (e.g., Sabine et al. 2002). Energetic costs of movement increase exponentially depending on sinking depth of the animal, doubling at 60% of brisket height (Hudson & Haigh 2002) and reaching three to eight times the cost of locomotion without snow depending on snow density for both cervids (Parker et al. 1984; Fancy & White 1987) and bovids (Dailey & Hobbs 1989). Snow depth is the primary influence on body condition of wintering white-tailed deer (Garroway & Broders 2005) and plays a significant role in energy balance of Alaskan black-tailed deer (Parker et al. 1999). Delayed snowmelt increases winter mortality of deer and bighorn sheep (Dumont et al. 2000; Jacobson et al. 2004) if body reserves are depleted long before new plant growth resumes. Following winters with deep snows, caribou give birth to smaller calves (Adams 2005). Consequently, the cumulative energetic effects of snow over consecutive winters can negatively influence population demography (e.g., Patterson & Power 2002).
Metabolic demands necessitate dietary consequences. For many ungulates characterized by sexual dimorphism, absolute metabolic costs are higher for larger males and relative (per kg) energy costs differ between sexes depending on season, resulting in intra-specific dietary differences (reviewed by Pérez-Barbería et al. 2008). Barboza & Bowyer (2000, 2001) proposed a gastrocentric hypothesis for cervids that explains seasonal differences in diet selection and habitat segregation among reproductive females, non-reproductive females, and males. Because reproductive females have higher nutritional demands during gestation and lactation, the requirements for higher dietary minima favour segregation from non-reproductive females and males in late winter and summer, but not during autumn. Reproductive females also increase feeding times in both winter and summer compared to non-reproductive animals. Males, with larger ruminal capacity and longer retention of forages, can subsist on the lowest quality diets. This nutritional basis for segregation of the sexes indicates that differences in foraging behaviour may be a consequence of different metabolic demands rather than social constraints or competitive exclusion.
In environments with prominent seasonal changes, food resources are commonly limited during dormant seasons. Dietary quantity and quality are highly variable, with significant declines in digestible nutrients during the winter or dry season. Consequently, highest intakes of digestible nutrients by herbivores occur in summer or rainy seasons. Seasonal patterns of intake generally coincide with seasonal patterns of reproduction and maintenance (Owen-Smith 2002). Ungulates are able to discriminate between feeding patches on the basis of quantity and quality of food, which has major implications for time budgets and nutritional status (Langvatn & Hanley 1993). Metabolic and nutritional requirements may preclude animals from feeding in areas with low forage abundance or low nutritive value (Cook 2002).
The mechanics of how animals forage for nutritional gain over different temporal and spatial scales, and the physiological and physical factors that influence intake rates have been defined by numerous reductionist studies. Across species, maximum short-term intake rates scale with body mass and correspond with scaling of metabolic rates (Shipley et al. 1994). Rates of food intake in relation to food abundance (the functional response of an animal to its nutritional environment) vary directly with bite size and indirectly with fibrousness of the food; they depend further on digestive constraints, gut morphology, and interactions with plant chemistry (Shipley et al. 1999; see also Torregrossa & Dearing 2009). Food availability is the main constraint on daily intake at low food abundance, whereas digestive processing capacity limits intake at high food abundance when food quality is low or metabolic demands are very high (Owen-Smith 2002). At high food abundance and high quality, surplus energy intake beyond immediate metabolic needs may be stored. At intermediate food abundance, animals adjust intakes in relation to nutritional values. Food abundance from the animal's perspective includes only available foods that it will consume, and which subsequently support metabolism, growth, and reproduction.
Ungulates use physiological and behavioural mechanisms to accommodate seasonal variation in both their nutritional requirements and the nutritional value of habitats. Presumably to conserve energy, cervids in north-temperate regions reduce intake rates voluntarily in winter (corresponding with natural declines in food availability), even when provided with ad libitum access to foods (e.g., black-tailed deer: Parker et al. 1993; white-tailed deer: Taillon et al. 2006; elk: Hudson & Haigh 2002; moose: Schwartz & Renecker 1998; Rangifer: Barboza & Parker 2008). Interestingly, white-tailed deer on low-quality diets reduced food intake less than deer on higher quality food, potentially to compensate for the lower nutrient content of the diet (Taillon et al. 2006). Dietary breadth was constrained for white-tailed deer by low forage quality as well as by mobility in snow (Dumont et al. 2005). For black-tailed deer, the processing of lower quality food in coastal environments in winter resulted in more time spent ruminating and fewer foraging bouts; deer did not increase time spent foraging to compensate for decreasing dietary quality, perhaps to avoid the high energy costs of movement in snow (Parker et al. 1999). They also reduced dietary breadth with decreasing nutritional value of available forages. In contrast, during the adverse dry season in tropical savanna, kudus spent more time active and increased both foraging time and dietary breadth to compensate for declines in forage quality (Owen-Smith 1994). Resource heterogeneity, duration of the dormant season, and the rate of decline in forage quality all affect the seasonal cycle of intake (Illius 2006). Because of interactions between the abundance and value of food resources, the quantity, quality and composition of the diet also vary with changes in population density (Nicholson et al. 2006).
Ultimately, nutrient intake, which depends on bite size and the digestible nutrient content of those bites, in relation to nutritional requirements, provides a critical link between food resources and animal performance (Parker et al. 1996). Bite sizes taken during foraging are a small-scale process that can have large consequences. The nutritional influence of numerous bites compounds over a foraging bout, a day, and a season to affect growth, survival, and reproduction (reviewed by Shipley 2007). Even small differences in food value can have large influences on animal performance through multiplier effects (White 1983). More than 75% of diet selection by red deer could be attributed to maximizing energy intake during winter and spring (van Wieren 1996), and daily rates of energy intake explained selection of feeding patches (Wilmshurst & Fryxell 1995). As a consequence of snow in winter and the increased demands of lactation in summer, availability of digestible energy was the greatest nutritional limiting factor for black-tailed deer in Alaska (Parker et al. 1999). Similarly, white-tailed deer in Quebec consistently preferred diets that were highest in digestible energy content in winter (Berteaux et al. 1998), presumably reflecting physiological needs. Energy intake during summer strongly affects body mass gain, including deposition of both body fat and body protein (Allaye Chan-McLeod et al. 1994). Energy balance on a year-round basis is generally more sensitive to variations in energy intake than in energy costs (Fancy 1986; Hobbs 1989), and consequently, intake rates drive what is possible in terms of body mass and condition of the animal. Yet animals probably have less control over maximum energy intake, which is largely dependent on plant growth, than their energy expenditures for activity.
Nutritional constraints for ungulates may be more than just energetic constraints. In addition to short-term elemental or chemical needs that influence movements and distribution of ungulates (Ayotte et al. 2006, 2008), there is increasing evidence that protein constraints may be an important nutritional challenge. Protein requirements are typically highest during body growth, which usually coincides with highest forage protein. Neonatal growth of red deer depends on milk protein, and protein to fat ratios in milk are highly correlated with birth mass (Landete-Castillejos et al. 2001, 2003). Lactating caribou allocate mostly surplus protein not used for the replacement of maternal protein to milk production (Allaye Chan-McLeod et al. 1994, 1999). Male reindeer that lose 23% of body protein during the breeding period and incur potentially more protein losses over winter must rely on spring and summer forage to replenish these protein stores (Barboza et al. 2004). Short-term protein intake rates in summer influence the selection of food patches (Langvatn & Hanley 1993).
Similar to energy demands, high-protein demands also can occur when the nutrient content of food resources is low. Protein requirements increase during foetal growth, particularly in late winter when 80% of foetal mass is deposited (Robbins & Robbins 1979; Robbins 1993) and forage protein is lowest of the year. During lactation, protein requirements may increase 110–130% in Rangifer (Barboza & Parker 2008), with high early-lactation requirements often occurring before the new growth of plants in spring. For reindeer and caribou that consume large quantities of low-protein lichens during a long winter followed by lactation demands, protein balance can be negative for 7 months of the year (Gerhart et al. 1996). Lichens, with their low protein content, are relatively high in digestible energy content, and the ratio of digestible energy to protein is much higher in lichens than in forage species consumed by other cervids during winter (Parker et al. 2005). To minimize excretory nitrogen losses and spare the use of body protein, Rangifer employs mechanisms such as urea recycling and oxidizing nitrogen from dietary protein (Barboza & Parker 2006, 2008). The extent to which protein is a limiting factor (possibly in addition to energy) has not been researched in detail for many northern wintering ungulates. Northern populations of moose consume winter forages that are near the limits of adequate protein content to support maintenance or reproductive requirements (D.E. Spalinger, unpublished; Schwartz & Renecker 1998). There is also some evidence from sapling fertilization experiments suggesting that white-tailed deer discern differences in protein content of individual plants and increase foraging rates on browse species with highest protein content during winter (Tripler et al. 2002).
Animal condition: mass, body fat, body protein
Body condition of an animal is the integrator of its location-specific energy and protein demands and its food intake, and the potential driver of demographic variation. Male ungulates are typically in best condition in autumn before the breeding season and in worst condition in late winter. By comparison, timing of highest and lowest body condition is delayed for reproductive females. Females are typically in best condition at the onset of winter when their nutritional demands are lowest and in worst condition 2–3 weeks after parturition in spring. Body size and condition, as indicators of habitat and weather (Hobbs 1989), have direct consequences to reproduction and population dynamics. The probability of conceiving and carrying a foetus to term is determined primarily by summer conditions and autumn body mass (e.g., caribou: Cameron et al. 1993; Gerhart et al. 1996; Cook et al. 2004a). Timing of parturition, birth mass, and early survival of offspring are closely linked to winter and spring nutrition (e.g., reindeer: Skoogland 1989; Dall's sheep (O. dalli): Rachlow & Bowyer 1991; bison (Bison bison): Berger 1992; moose: Keech et al. 2000; elk: Cook 2002). Physiological indicators reflecting changes in condition, food intake, and stress were reviewed by Parker (2003).
Nutritional resources available in summer and autumn are used by juveniles to increase the likelihood of attaining a body mass that enables them to survive winter. Because of their smaller size, limited body reserves, and relatively higher metabolic demands, juveniles are most susceptible to harsh conditions. Juvenile survival, which determines recruitment, is commonly the key factor in the dynamics of their populations (Gaillard et al. 1998, 2000; Coulson et al. 2001). Elk calves with access to higher nutritional levels reach larger body sizes; and the larger the body mass at the beginning of winter, the more days they survive (Cook et al. 2004a). Body mass at the beginning of winter was also the best predictor of over-winter survival by white-tailed deer fawns (Taillon et al. 2006). In areas with winter supplemental feeding programs to reduce loss of body mass, higher dietary intake may increase nutritional status by reducing endogenous tissue catabolism (Tarr & Pekins 2002; Page & Underwood 2006); and the heat increment from additional feeding may compensate for much of the energetic costs of thermoregulation associated with winter severity (Jensen et al. 1999). Pettorelli et al. (2003) suggested that the presence of key preferred plant species in maternal home ranges shapes winter body mass of roe deer fawns; others also have noted the overwhelming importance of habitat quality in determining body mass of juveniles (Côté & Festa-Bianchet 2001; Ericssen et al. 2002; Kjellander et al. 2006). Juvenile size, whether influenced by variation in environmental conditions or density-induced nutritional limitations, also affects age at first reproduction (e.g., Solberg & Saether 1994; Gaillard et al. 2000) and subsequent adult body mass (e.g., Pettorelli et al. 2002). Given that compensatory growth is rare (e.g., Feder et al. 2008, but see Toïgo et al. 2006), animals that are born small or have restricted early growth rates often remain compromised as smaller adults that are likely to produce fewer offspring over their lifetimes (Gaillard et al. 2003).
Regain of the body mass lost during winter is critical for adults, which typically mobilize body reserves because of reduced forage resources. Losses in body mass for northern ungulates over winter commonly range from 15–30%, with highest absolute and proportional losses incurred by the largest animals (Parker et al. 1993; Allaye Chan-McLeod et al. 1999; Hudson & Haigh 2002; Festa-Bianchet & Côté 2008). The timing for when mass regain occurs has important implications for critical habitats. Female caribou lose body mass for approximately 3 weeks following calving because of the high costs of lactation (Parker et al. 1990). They subsequently put on significant body mass during the summer period, emphasizing the importance of good summer habitats to regain body condition before winter. This strategy contrasts with another Arctic ungulate, the muskox (Ovibos moschatus), which after calving maintains that body mass (without losing additional mass) through the summer, and then relies on autumn habitats to regain mass and condition (Parker et al. 1990). The nutritional contribution of summer versus autumn habitats, therefore, may differ among species even within similar regions. Heavier females are more likely to reproduce and to produce offspring earlier than lighter females (e.g., Cameron et al. 1993; Gerhart et al. 1997; Adams & Dale 1998; Festa-Bianchet et al. 1998). There also may be nutritional consequences of adult body sizes to sex of the offspring. Reindeer mothers with high body mass in autumn are more likely to produce male calves the following spring (Holand et al. 2006); male calves, which are usually heavier at birth, require greater maternal investment.
Body fat is the major energy reserve of the body. The year-round body fat cycle for ungulates in northern environments where energy costs exceed forage energy in winter was aptly described by Mautz (1978). Carry-over or temporally lagged effects of previous nutritional deprivation may ultimately affect pregnancy rates if animals are unable to replenish reserves following severe winters or successive years of producing young (Fig. 1). Garroway & Broders (2007) noted that the winter severity of one year before gestation (not the winter during which gestation took place) reduced the probability of producing a foetus. This suggests an adaptive ability to divert energy away from reproduction as a consequence of environmental constraints. Others have posed a ‘selfish cow’ explanation (Russell et al. 1993, 2005) in which mature females conserve maternal condition for self maintenance and future reproduction at the expense of allocating already depleted resources to potentially smaller offspring that have low chances of survival (Clutton-Brock et al. 1989). Similarly, if producing and supporting a neonate drains maternal body stores excessively, the next pregnancy may be compromised (Cameron 1994; Gerhart et al. 1997; Cook 2002; Hudson & Haigh 2002) (Fig. 1). Not reproducing every year may better ensure survival and higher lifetime reproduction (Festa-Bianchet & Côté 2008).
Autumn body fat levels depend largely on summer-autumn nutrition. Even with high energetic costs of lactation during summer causing body fat levels of females to be 50% less than non-lactating animals, lactating females can accrue as much fat by mid-autumn as non-lactating animals, assuming adequate summer forage quality (e.g., Cook et al. 2004a). Failure to breed is a function of low body fat levels in elk (Cook et al. 2004a, 2004b), caribou (Gerhart et al. 1996; Gustine et al. 2007), and moose (Heard et al. 1997). Body fat thresholds indicating sufficient condition for pregnancy have been proposed for elk (>5%, Cook et al. 2004a) and caribou (6–7·8%, Crête et al. 1993; Ouellet et al. 1997). Elk females with the most body fat breed at earlier dates in autumn (Cook et al. 2004a). Higher fat levels in moose increase the likelihood of pregnancy, twinning, and larger calves (Heard et al. 1997; Testa & Adams 1998; Keech et al. 2000). From an adaptive standpoint, young of the year accumulate larger fat reserves in areas with severe winter conditions, particularly in northern environments (Lesage et al. 2001; Kjellander et al. 2006). The maximum fat reserve accumulated in autumn by fawns, when coupled with the forage energy available during winter, may set the northern limits of white-tailed deer range in North America (Lesage et al. 2001). Juvenile mountain goats living in harsh mountain environments may accumulate energy reserves at the expense of body growth (Festa-Bianchet & Côté 2008).
Winter and spring body fat levels buffer the effects of declining food supplies when energetic demands can not be met by foraging alone (Parker et al. 1999). With low fat levels, there may be increased incidence of embryonic mortality in some species (caribou: Russell et al. 1998), although there is strong selection to maintain pregnancy at almost any cost in other species (elk: Cook et al. 2002). Differences in body condition thresholds for aborting the foetus may reflect the likelihood of recovering condition during gestation; long winters limit recovery of body condition lost in early winter and favour a greater incidence of abortion in caribou. Under adverse conditions females in poor condition may extend gestation length to potentially match calving date with maximum plant production (red deer: Garcia et al. 2006). The interaction between nutritional condition and available food resources also may influence timing of movements to spring ranges. Winter feeding programs that increase nutritional status can induce behavioural effects of postponed migration by prolonging nutrient availability of natural forages (mule deer: Peterson & Messmer 2007).
More recently, some studies have addressed the role of protein in body condition (DelGiudice et al. 2001). To meet energy demands, mobilization of body protein may be necessary to supplement mobilization of body fat by malnourished animals. Rates of protein depletion are usually less than fat depletion (Parker et al. 1993; Barboza & Parker 2006), but tissue wasting may increase significantly when fat stores are depleted. Reindeer are in negative protein balance in winter when >46% of excreted urea-nitrogen originates from body tissue (Barboza & Parker 2006). To meet protein demands, dietary protein usually is used before body protein, but intake of very low-protein foods may necessitate the use of additional body protein, and possibly impact foetal growth and neonatal mass. In Rangifer, calf mass at birth is correlated with maternal protein reserves (Allaye Chan 1991). Those reserves are mobilized during pregnancy and early lactation (Barboza & Parker 2008), and the replacement of maternal protein reserves in summer becomes critical for future foetal investment.
Barboza & Parker (2008) provided evidence that the resilience of Rangifer populations to changing environments may be influenced by their ability to alter timing and allocation of body protein to reproduction. Both caribou and reindeer rely on body protein for most of foetal growth and for calf growth during the first month of lactation. On one end of a continuum for body capital and dietary income, reindeer with large fat reserves and relatively sedentary behaviour are able to spare the use of body protein to meet energetic demands during winter. Because they calve approximately one month earlier than caribou, they must rely on body stores (capital) that were put down in the autumn to produce a calf. On the other end of this continuum, caribou that migrate and calve later in spring closer to spring green-up can use more income from the diet at calving grounds that have predictable timing of plant growth for calf production. Hence, depending on the timing of calving, autumn habitats versus spring habitats vary in their nutritional value for the production of offspring. Across species, Moen et al. (2006) posited that most large herbivores in arctic and alpine areas are closer to the capital breeder end of the continuum. Body reserves in capital breeders serve as insurance against unforeseen conditions during winter and early spring (Fauchald et al. 2004). Capital breeders such as bighorn sheep and Soay sheep (O. aries) provide initial post-natal care from body reserves and are less affected by temporal mismatches between vegetation green-up and birth date (Durant et al. 2005). Roe deer are the ultimate income breeder, which does not accumulate body reserves or change condition substantively throughout the year and which times birthing to match a less variable spring green-up (Andersen et al. 2000).
Energy versus protein: influence on animal response
Researchers have tended to concentrate on energy costs, available forage energy, and body fat more than protein demands, forage protein intake, and body protein, but the two nutritional currencies are clearly linked in defining the role of nutrition in population dynamics. High-energy high-protein diets of spring and summer allow ungulates to regain mass and condition, replenishing the endogenous reserves that are necessary for over-winter survival and foetal growth (Fig. 1). These diets also support the high energetic costs of lactation and the protein demands for neonatal growth. Under poor nutritional conditions, foetal growth and milk production may be impaired (Oftedal 1985). Winter diets, even with low digestible energy and/or protein content, reduce the extent of mobilization of body reserves, allowing animals to survive winter deficits. Experimental approaches using matched diets indicate that cervids may be able to discriminate between dietary energy and protein (Berteaux et al. 1998). Consequently, selection that assumes some nutritional wisdom would be expected to vary with biological state and a constantly changing environment. Seasonal losses and gains in body condition are probably less variable than intakes and diet, and more representative of an animal's physiological needs.
Changes in body fat and body protein enable different animal responses. Over winter, the mobilization of fat and protein stores acquired primarily in summer determines whether animals (i) die, (ii) live without reproducing, or (iii) live and reproduce. We posed the question: does body mass, fat, or protein have the greatest influence on animal response? Population trends result as a consequence of variation around condition and rates of change in condition. To assess the relative importance of that variation, we developed a simulation model using peak body mass and associated body composition, and the changes in mass and condition of adult females over winter. The rates of change are a quantification of how animals integrate their energy losses and nutritional gains.
For this exercise, we used data for Rangifer, a species that is documented to be limited by energy or protein constraints, or both (data from Barboza & Parker 2008). We generated normal distributions (based on means and SD) for initial body mass, percent body fat and protein, and rates of change in body fat and protein per unit metabolic mass to adjust for allometric differences in demand (Table 1). For each run of the model, we randomly sampled from these five distributions using 1000 simulations to predict animal response. Feasible ratios of body fat to body protein were set between 0·123 and 2·058 for a plausible animal (data from Barboza & Parker 2008). We assumed a minimum of 6% body fat for reproduction (Ouellet et al. 1997). We used the following criteria for animal response at the end of a 23-week period in winter (from peak body mass to parturition): (i) Animals died if either final body fat was <3% or final body protein was <65% of initial body protein. Minimum possible fat content was 3% based on the composition of young Rangifer (Gerhart et al. 1996). The labile protein store was 35% of peak body protein based on the seasonal gain of body protein in adult females (Barboza & Parker 2006). (ii) Animals lived if body fat was ≥3% and final body protein was ≥65% of initial body protein. (iii) Animals reproduced if the protein ≥65% of initial body protein was >0·67 kg protein, as estimated from Gerhart et al. (1996) for a minimum-sized viable Rangifer calf weighing 3·9 kg (Skoogland 1989; Adams 2005). For both caribou and reindeer, we ran the model 25 times (1000 simulations each) and then summarized the resulting outcomes (Table 1). Given these parameters and their inherent variation, the model predicted higher incidence of successful reproduction and lower mortality by caribou than reindeer. The capital strategy of reindeer relies on using stores of both fat and protein for both survival and reproduction. Conversely, the income strategy of caribou is associated with low rates of fat and protein loss from stores of body fat and protein that are similar to reindeer. Severe winters increase the rates of fat and protein loss and therefore increase the likelihood of death or reproductive failure in both caribou and reindeer. Changes in food intake, however, could partially compensate for high rates of energy and protein loss. The model responses are driven by variation around the parameters, and may not reflect conditions experienced within a particular herd of free-ranging Rangifer or the ability of the animal to vary food intake. In the model, high mortality of reindeer reflects the importance of restoring body tissue from diet for capital breeders with high rates of protein and fat loss.
Table 1. Parameters (mean ± SD) used in the simulation model and uncertainty analysis of the importance of peak body mass, and changes in body fat and protein reserves during winter by Rangifer to animal responses at the end of winter. Rates of change in body fat and protein differed between reproductive and non-reproductive caribou and reindeer
We then evaluated the effects of the input parameters (Table 1) on the predicted animal response. These uncertainty analyses assessed the relative influence of initial body mass and composition, and the changes in body fat and protein on the three animal responses across the range of variation of all input parameters (Latin hypercube design, Swartzman & Kaluzny 1987). For each analysis we considered each of the five input parameters to be uniformly distributed between the mean ± 2 SD. We divided each distribution into 1000 equal intervals, and for each of 1000 iterations, we randomly sampled each parameter without replacement. The input values for each parameter and predicted responses were saved. After ranking the parameters, we regressed the dependent variable (animal response) against the independently selected ranked parameters. To remove the potential effects of the other parameters in the regression, we calculated the relative partial sum of squares (RPSS, SAS Institute Inc. 2005) (Bartell et al. 1986; Swartzman & Kaluzny 1987), and then calculated the percent contribution of each parameter (its RPSS) as a percentage of the sum of the RPSS. Parameters with high contributions to the total RPSS are those that had a relatively higher contribution to modelled animal response.
Peak body mass explained little variation in the modelled responses over winter for adult female caribou or reindeer. Rather, percent body fat had the greatest influence as a single parameter in determining whether animals died, lived without reproducing, or lived and reproduced at the end of the 23-week period (Table 1). In caribou, animal response was also sensitive to percent body protein and to rates of protein change. In reindeer, the rates of change in protein and fat together explained more variation in animal response than percent body fat. These results underscore the importance of fat and fat dynamics, which have been well studied and incorporated in specific models of energy balance (e.g., Hudson & White 1985; Hobbs 1989; Russell et al. 1993; Moen et al. 1998; Parker et al. 1999; Russell et al. 2005). However, the results also indicate that protein may contribute substantively to animal response. Interestingly, when we ran simulations that increased the mean rates of both protein and fat loss of caribou to those of reindeer, there was greater sensitivity to protein loss (28·2% vs.17·7 % and less sensitivity to initial fat content (39·9% vs. 61·8%). Protein stores therefore become more important as winter severity increases for animals such as caribou that use smaller body stores of fat and rely more on dietary income for reproduction. Increasing the range of body fat content in caribou to that of reindeer reduced the sensitivity to the rate of protein loss, as body fat spares body protein. For juvenile animals without extensive body fat and protein reserves, body mass associated with body size directs survival (e.g., Toïgo et al. 2006).
Minimizing the maximum detriment to free-ranging ungulates
For free-ranging ungulates, life is a balance among numerous ecological factors that include nutritional requirements, availability of nutrients to meet those needs, and intra- and inter-specific interactions. Constraints to resource gain in a habitat may be digestive, metabolic, thermal, or the risk of mortality (Owen-Smith 2002). Animals may select forages defended by toxic plant chemicals if the nutrient content outweighs the negative effects (McArthur et al. 1993). Animals may alter foraging patterns to avoid thermal costs or forage during heat and cold stress if the nutrient gains from food resources outweigh the costs (Dussault et al. 2004; Maloney et al. 2005; Hay et al. 2008). Animals may forego the best foraging opportunities to avoid predation risk, particularly when it is associated with reproduction (e.g., Festa-Bianchet 1988; Poole et al. 2007). Field biologists strive to define what and when is most critical to sustain populations. We suggest that over the long term of life histories, animal strategies tend to minimize the maximum detriment to fitness, and that the underlying foundation of the detriment is nutritional.
In winter, the maximum detriment to the individual and the population would be to have low body stores that reduce adult survival or foetal development. Consequently, animal strategies should minimize energy costs and the loss of protein stores. Numerous studies have shown that ungulates reduce activity in winter (and corresponding increased energetic costs in snow and in very cold temperatures) when food quantity and quality are limited (e.g., Cook 2002; van Oort et al. 2007) and daily food intake can decline by 70% compared to summer (Parker et al. 1999). Under extreme conditions, animals partition the use of reserves down to the minimum by eating and resting, with minimal other activity. This reduced mobilization of fat reserves to meet energetic demands also spares the use of body protein. In spring, the maximum detriment would be unsuccessful calving or failed parturition. Predation pressure is usually biased towards juveniles in most ungulate populations (Linnell et al. 1995; Fritz & Loison 2006). Consequently, animals should avoid predation risk in areas where large carnivores are present and maximize intake of high-quality forage (digestible energy and protein). Caribou, for example, commonly avoid areas of high vegetation biomass if those areas are associated with high predation risk, and may use topography to increase segregation from predators (Barten et al. 2001; Griffith et al. 2002). Parturient caribou forage selectively though in areas of relatively high vegetation quality to meet the nutritional requirements of lactation (Gustine et al. 2006). In summer, the maximum detriment would be mass and condition regains by adults that are insufficient for breeding and growth rates of calves that are too low for over-winter survival. Animals must maximize intake to avoid compromising these gains. For black-tailed deer, intake rates in summer compared to winter were four times higher for digestible energy and 10 times higher for digestible protein (Parker et al. 1999). Adult females had intake rates per kilogram that were twice as high as males at the end of summer following high lactation demands. Red deer and reindeer in Scandinavia use a diversity of altitudes and aspects to continually access high-quality (protein and energy) forage, resulting in larger autumn body mass (Albon &Langvatn 1992; Mysterud et al. 2001), and minimizing the potential consequence of lower mass regains during summer. By following snowmelt patterns to higher altitudes, animals access high-quality emerging shoots ‘in spring conditions as long as possible during the summer’ (Moen et al. 2006). Similarly, Stone's sheep (O. dalli stonei) track a phenological gradient of high forage protein by moving up in elevation as the growing season progresses (Walker et al. 2006). In autumn, the maximum detriment would be if breeding was not successful or if body reserves were not sufficient for over-winter survival. Animals should continue to maximize intake where possible, and begin to decrease energy costs. Females often allocate more time than males to foraging during the breeding season, but males increase their foraging times significantly post-breeding (Bunnell & Gillingham 1985) even though they may not always recover the body mass they lost during breeding before winter (Barboza et al. 2004). Given these seasonal strategies, animals should select for high nutritional value in spring, summer, and autumn before the dormant season. From an applied perspective, habitats should be managed or conserved to provide the widest window of nutrient gain between spring and autumn for both genders.
Free-ranging ungulates typically show flexibility or plasticity within seasonal strategy. At large spatial and temporal scales, there is often more than one way for animals to use the landscape and get to the same endpoint of having sufficient body condition to survive and reproduce. Rettie & Messier (2000) proposed that the factors with the greatest potential to limit individual fitness are those that influence large-scale selection. They also noted that there may be a hierarchy of limiting factors. As an example, woodland caribou (R. t. caribou) selected calving areas that tended to be at higher elevations and steeper than the landscape around them, have relatively low risk of predation by wolves (Canis lupus) compared to the areas around them, and have relatively low vegetation biomass (Gustine et al. 2006). The calving sites that caribou chose within the general calving areas had lower risk of grizzly bear (Ursus arctos) predation and relatively high vegetation quality despite low biomass. Caribou subsequently moved to summering areas with higher vegetation quantity, thereby maximizing nutrient intake. From Rettie & Messier's perspective, predation affecting large-scale selection may be the proximate factor most limiting individual fitness, as observed for numerous woodland caribou populations (e.g., Chowns & Gates 2004). Within hierarchical selection, nutritional attributes defined the choice of calving sites and summer habitats. From the strategy of minimizing the maximum detriment, both factors (nutrition and predation) are significant. Similarly, barren ground caribou (R. t. granti) move to calving grounds on the Arctic Coastal Plain of Alaska in response to predictable green-up of high-quality vegetation and reduced predation risk; and move away from calving grounds to summering areas with higher forage biomass (Griffith et al. 2002).
Nutrition matters to population health even if predation is a limiting factor (e.g., Brown & Mallory 2007; Brown et al. 2007). Poor nutrition contributes to the high rate of predator-caused mortality for juveniles (Mech 2007). Maternal nutrition during winter may also predispose neonates to early death if body condition or foetal development is compromised (Roffe 1993). Furthermore, predation on neonates may result in increased body condition of adult females. Females that are relieved of the energetic constraints of lactation can regain mass quickly to conceive the next fall (Allaye Chan-McLeod et al. 1999). Predator removal may benefit short-term recruitment, but higher gestational and lactational demands on adult females can ultimately result in lower pregnancy rates and recruitment for the population if the nutritional value of habitats is lacking.
Conclusion and perspectives
Food and nutrient availability are the ultimate causes of the reproductive cycle of ungulates (Pekins et al. 1998), but there are complex interactions between internal physiological regulators and the external environment (Schwartz & Renecker 1998). The cycle is both restricted and adaptive, with dates of conception and parturition in synchrony with cycles of body reserves and the onset of spring green-up. Late winter and early spring may be a critical time in the bioenergetics of ungulates when body stores are depleted and when nutritional demands increase for gestation and lactation. Although bottlenecks to individual survival and population growth are associated with the dormant season for many large herbivores and there has been emphasis placed on the importance of resources used then (Illius 2006), those resources are not necessarily the key factor determining population size. Nutrient availability in summer drives replenishment of body reserves and subsequent reproductive success (Fig. 1). Because reserves are not deposited without limit, of course food resources in winter are also important. Our model describing differences in the reproductive strategies of two Rangifer subspecies synthesizes the interactions between body fat representing the profits of summer and rates of body protein and fat loss reflecting winter severity. Some researchers have reported that digestible energy in summer regulates populations because of its influence on condition and probability of reproduction (elk: Cook et al. 2004a); others have shown that protein constrains reproduction (Rangifer: Barboza & Parker 2008). Differences in energy and nutrient demands between reproductive and non-reproductive animals lead to differences in the timing of deposition and mobilization of body tissues (Allaye Chan-McLeod et al. 1999). Hence, nutrient partitioning and allocation strategies vary among ungulates given different physiological demands and environmental conditions, including habitat, topography, weather and predation.
As environmental conditions change, survival and reproduction will depend on whether habitats can meet animal demands. Climatic fluctuations influence demography through direct effects of snow depth and hardness, and energetic demands on winter survival and foetal development (Post & Stenseth 1998, 1999; Forchhammer et al. 2001; Patterson & Power 2002; Moen et al. 2006) and through indirect effects of increased vulnerability to predation in deep snow (Post & Stenseth 1998, 1999; DelGiudice et al. 2002). Climatic changes to precipitation regimes may alter the timing of both spring and fall movements (Sabine et al. 2002) and the onset of spring green-up (Pettorelli et al. 2005), which subsequently affect maternal condition and juvenile growth rates. Inter-annual variation in plant phenology caused by climatic variability particularly in northern, arctic, alpine, and mountainous environments may induce variation in the timing of parturition (Post et al. 2003), number of calves born (Post & Forchhammer 2008), and growth and survival of juveniles (Weladji & Holand 2003; Pettorelli et al. 2007). To forecast the effects of long-term climate or anthropogenic changes on timing, duration, and abundance of resources (Durant et al. 2005) and the responses of ungulates to those changes, it is important to understand how different species use dietary income and body stores to integrate the profits of summer and the demands of winter.
Authors acknowledge insightful discussions with A.W. Hawley on minimizing the maximum detriment and helpful comments from anonymous referees.