Temporal changes in net energy balance of animals strongly influence fitness; consequently, natural selection should favor behaviors that increase net energy balance by buffering individuals against negative effects of environmental variation. The relative importance of behavioral responses to climate‐induced variation in costs vs. supplies of energy, however, is uncertain, as is the degree to which such responses are mediated by current stores of energy. We evaluated relationships among behavior, nutritional condition (i.e., energy state), and spatiotemporal variation in costs vs. supplies of energy available to a large‐bodied endotherm, the North American elk (Cervus elaphus), occupying two ecosystems with contrasting climates and energy landscapes: a temperate, montane forest and an arid, high‐elevation desert. We hypothesized that during spring through autumn, behavioral responses to the energy landscape would be both context dependent (i.e., would vary as a function of the contrasting environmental conditions experienced by elk in the forest vs. the desert), and state dependent (i.e., would vary as a function of the energy balance of an individual). We tested several predictions derived from that hypothesis by combining output from a biophysical model of the thermal environment with data on forage quality, animal locations, and nutritional condition of individuals. At the population level, elk in the desert selected areas that reduced costs of thermoregulation over those that provided the highest‐quality forage. In the forest, however, costs imposed by the thermal environment were less pronounced, and elk selected areas that increased access to high‐quality forage over those that reduced costs of thermoregulation. At the individual level, nutritional condition did not influence strength of selection for low‐cost areas or high‐quality forage among elk in the forest. In the desert, however, strength of selection for low‐cost areas (but not forage quality) was state dependent; individuals in the poorest condition at the end of winter showed the strongest selection for areas that reduced costs of thermoregulation during spring and summer, and also expended the least amount of energy on locomotion. Our results highlight the importance of understanding the roles of behavior and nutritional condition in buffering endotherms against direct and indirect effects of climate on fitness.

How climate constrains species' distributions through time and space is an important question in the context of conservation planning for climate change. Despite increasing awareness of the need to incorporate mechanism into species distribution models (SDMs), mechanistic modeling of endotherm distributions remains limited in this literature. Using the American pika (Ochotona princeps) as an example, we present a framework whereby mechanism can be incorporated into endotherm SDMs. Pika distribution has repeatedly been found to be constrained by warm temperatures, so we used Niche Mapper, a mechanistic heat-balance model, to convert macroclimate data to pika-specific surface activity time in summer across the western United States. We then explored the difference between using a macroclimate predictor (summer temperature) and using a mechanistic predictor (predicted surface activity time) in SDMs. Both approaches accurately predicted pika presences in current and past climate regimes. However, the activity models predicted 8-19% less habitat loss in response to annual temperature increases of ~3-5 °C predicted in the region by 2070, suggesting that pikas may be able to buffer some climate change effects through behavioral thermoregulation that can be captured by mechanistic modeling. Incorporating mechanism added value to the modeling by providing increased confidence in areas where different modeling approaches agreed and providing a range of outcomes in areas of disagreement. It also provided a more proximate variable relating animal distribution to climate, allowing investigations into how unique habitat characteristics and intraspecific phenotypic variation may allow pikas to exist in areas outside those predicted by generic SDMs. Only a small number of easily obtainable data are required to parameterize this mechanistic model for any endotherm, and its use can improve SDM predictions by explicitly modeling a widely applicable direct physiological effect: climate-imposed restrictions on activity. This more complete understanding is necessary to inform climate adaptation actions, management strategies, and conservation plans.

An accurate estimate for orangutan nest decay time is a crucial factor in commonly used methods for estimating orangutan population size. Decay rates are known to vary, but the decay process and, thus, the temporal and spatial variation in decay time are poorly understood. We used established line-transect methodology to survey orangutan nests in a lowland forest in East Kalimantan, Indonesia, and monitored the decay of 663 nests over 20 months. Using Markov chain analysis we calculated a decay time of 602 days, which is significantly longer than times found in other studies. Based on this, we recalculated the orangutan density estimate for a site in East Kalimantan; the resulting density is much lower than previous estimates (previous estimates were 3-8 times higher than our recalculated density). Our data suggest that short-term studies where decay times are determined using matrix mathematics may produce unreliable decay times. Our findings have implications for other parts of the orangutan range where population estimates are based on potentially unreliable nest decay rate estimates, and we recommend that for various parts of the orangutan range census estimates be reexamined. Considering the high variation in decay rates there is a need to move away from using single-number decay time estimates and, preferably, to test methods that do not rely on nest decay times as alternatives for rapid assessments of orangutan habitat for conservation in Borneo.

When possible, many species will shift in elevation or latitude in response to rising temperatures. However, before such shifts occur, individuals will first tolerate environmental change and then modify their behavior to maintain heat balance. Behavioral thermoregulation allows animals a range of climatic tolerances and makes predicting geographic responses under future warming scenarios challenging. Because behavioral modification may reduce an individual's fecundity by, for example, limiting foraging time and thus caloric intake, we must consider the range of behavioral options available for thermoregulation to accurately predict climate change impacts on individual species. To date, few studies have identified mechanistic links between an organism's daily activities and the need to thermoregulate. We used a biophysical model, Niche Mapper, to mechanistically model microclimate conditions and thermoregulatory behavior for a temperature‐sensitive mammal, the American pika (Ochotona princeps). Niche Mapper accurately simulated microclimate conditions, as well as empirical metabolic chamber data for a range of fur properties, animal sizes, and environmental parameters. Niche Mapper predicted pikas would be behaviorally constrained because of the need to thermoregulate during the hottest times of the day. We also showed that pikas at low elevations could receive energetic benefits by being smaller in size and maintaining summer pelage during longer stretches of the active season under a future warming scenario. We observed pika behavior for 288 h in Glacier National Park, Montana, and thermally characterized their rocky, montane environment. We found that pikas were most active when temperatures were cooler, and at sites characterized by high elevations and north‐facing slopes. Pikas became significantly less active across a suite of behaviors in the field when temperatures surpassed 20°C, which supported a metabolic threshold predicted by Niche Mapper. In general, mechanistic predictions and empirical observations were congruent. This research is unique in providing both an empirical and mechanistic description of the effects of temperature on a mammalian sentinel of climate change, the American pika. Our results suggest that previously underinvestigated characteristics, specifically fur properties and body size, may play critical roles in pika populations' response to climate change. We also demonstrate the potential importance of considering behavioral thermoregulation and microclimate variability when predicting animal responses to climate change.

Animals may partially overcome environmental constraints on fitness by behaviorally adjusting their exposure to costs and supplies of energy. Few studies, however, have linked spatiotemporal variation in the energy landscape to behaviorally mediated measures of performance that ostensibly influence individual fitness. We hypothesized that strength of selection by North American elk (Cervus elaphus) for areas that reduced costs of thermoregulation and activity, and increased access to high-quality forage, would influence four energetically mediated traits related to fitness: birth mass of young, nutritional condition of adult females at the onset of winter, change in nutritional condition of females between spring and winter, and neonatal survival. We used a biophysical model to map spatiotemporally explicit costs of thermoregulation and activity experienced by elk in a heterogeneous landscape. We then combined model predictions with data on forage characteristics, animal locations, nutritional condition, and mass and survival of young to evaluate behaviorally mediated effects of the energy landscape on fitness-related traits. During spring, when high-quality forage was abundant, female elk that consistently selected low-cost areas before parturition gave birth to larger young than less-selective individuals, and birth mass had a strong, positive influence on probability of survival. As forage quality declined during autumn, however, lactating females that consistently selected the highest quality forage available accrued more fat and entered winter in better condition than less-selective individuals. Results of our study highlight the importance of understanding the dynamic nature of energy landscapes experienced by free-ranging animals.