ABSTRACT Elk ( Cervus elaphus) in the western United States are an economically and socially valuable wildlife species. They have featured species status for federal land management planning; hence, considerable modeling focused on habitat evaluation and land management planning has been undertaken for elk. The extent to which these and other habitat models for large ungulates account for influences of nutritional resources varies greatly, probably because of varying recognition of the importance of nutrition and uncertainty about how to measure and model nutrition. Our primary goals were to 1) develop greater understanding of how habitat conditions influence foraging dynamics and nutrition of elk in summer and autumn; and 2) illustrate an ecological framework for evaluating and predicting nutritional resources so that nutritional needs of elk can be integrated within landscape-scale plans, population models, and habitat evaluation models. We evaluated foraging responses of elk to clearcut logging and commercial thinning, forest succession, and season across ecological site potentials. We also identified the extent to which plant communities satisfied nutritional requirements of lactating female elk and their calves. Our study was conducted in the temperate rainforests of the Pacific Northwest on industrial and public timberlands. We evaluated relations between habitat conditions and elk nutrition in plant communities representing a range in stand age and ecological conditions at 3 study areas, 1 near the Canadian border in the north Cascades Mountains (Nooksack), 1 in the Coast range southwest of Olympia, Washington (Willapa Hills), and the third in the central Cascades near Springfield, Oregon (Springfield), from late June to November, 2000-2002. In 98-143 macroplots per study area, we measured forage abundance by plant species, digestible energy content by plant life-form group, and forest overstory. In a subset of these macroplots (∼30 per study area), we held 4 tame lactating elk with calves in electrified pens ( n = 15-25 adult elk per year), and sampled activity budgets, dietary composition, forage selection, and other measures of foraging behavior; dietary digestible energy (DE; kcal/g) and protein (DP; %) levels; and intake rates of these nutrients. In 15 of these pens, we held elk for extended periods (13-21 days) to monitor changes in body fat of adults and growth of calves. We developed equations to predict dietary DE and DP and per-minute intake rates of each in a nutrition prediction model that reflected vegetation attributes and ecological site influences. Total abundance of forage in the western hemlock series after clearcut logging in low to moderate elevations (≤1,000 m) ranged from a peak of 3,000-4,500 kg/ha in 5- to 10-year-old stands to 100-300 kg/ha in 20- to 50-year-old stands with only moderate increases through late succession. Patterns were similar in higher elevation forests (1,000-1,800 m), although peaks and troughs in forage abundance developed more slowly. Deciduous shrubs, forbs, and graminoids were abundant in early seral stages after stand disturbance, but these were rapidly replaced by shade-tolerant evergreen shrubs and ferns as conifer overstories closed 15-20 years later in low-elevation forest zones, and 20-40 years later in high-elevation zones. Digestible energy within plant life-form groups generally declined with season and with advancing succession, increased with elevation, and was highest in forbs and deciduous shrubs and lowest in evergreen shrubs and shade-tolerant ferns. Levels of DE in elk diets exhibited a strong asymptotic relation with abundance (kg/ha) of plant species that were eaten in proportions equal to or greater than availability (i.e., accepted species). Marked declines in dietary DE occurred in stands containing <400 kg/ha to 500 kg/ha of accepted species, largely because elk began to increase consumption of avoided species, and these typically contained low levels of DE. The asymptotic pattern was generally consistent among seasons, study areas, and habitat types (potential natural vegetation categories), although the asymptote averaged 10-12% greater in high- versus low-elevation forests. Abundance of accepted species in early seral stands averaged 7-10 times that in mid and late seral stages, and dietary DE levels varied accordingly. Dietary DE was little influenced by thinning in 20- to 60-year-old stands. In contrast, levels of dietary DP were unrelated to forage composition and abundance of accepted or avoided species, and varied little between low and high-elevation forests. Dietary DP increased with overstory canopy cover, was higher in thinned and hardwood stands, particularly those hardwood stands with saturated soils in late summer, declined with season, and was lowest in the driest forest communities in our study. Overall, soil moisture regime and season accounted for the majority of variation in dietary DP. Relations between nutrient intake rate and vegetation conditions varied among study areas and habitat types. Nevertheless, elk maintained about double the intake rate of DE in early seral stages versus closed-canopy forests. Intake rate of DP was similar between early seral versus closed-canopy forests, despite modestly lower dietary DP in early seral stages. Protein intake rate was greater in thinned and hardwood-riparian stands. In early seral stages, dietary DE typically met the requirement of 2.7 kcal/g of ingested forage (necessary to maintain body fat levels of lactating elk in summer) in the low-elevation forest zones and exceeded that level in high-elevation forest zones. In closed-canopy forests, dietary DE averaged below requirements, markedly so in low-elevation forests (2.25-2.5 kcal/g) and moderately so in high-elevation forests (2.4-2.65 kcal/g). Evidence of deficiencies based on DE intake rate was greater, averaging about 50% of requirements (28 kcal/min; 21,000 kcal/day) in closed-canopy forests and 80% of requirements in early seral stages. In contrast, dietary DP and DP intake rates generally approached or exceeded estimated requirements (6.8% DP; 380 g/day) in many habitat types that we sampled, with the greatest potential for deficient DP intake rates in relatively dry, low-elevation forests. Body fat dynamics and growth of calves confirmed nutritional deficiencies suggested by our data on DE intake. Adult elk lost body fat during all trials at rates generally in accordance with expectations at the dietary DE levels they consumed, and rate of change in body fat was inversely related to abundance of accepted species. Calves grew at about half the rate of which they are capable (1 kg/day) if summer nutrition is sufficient. Daily calf growth was positively related to their mother's dietary DE and protein intake levels. Elk compensated for limited foraging options in many plant communities via several behavioral strategies. Selection was generally strong for plants with higher DE levels, where selected species composed nearly 5 times more of the diet than did species that elk avoided, yet avoided species were 10 times more abundant. As abundance of accepted species declined below approximately 400 kg/ha, elk increased intake of avoided species. This strategy delayed declines in per-minute forage and DE intake rate as long as abundance of accepted species remained above roughly 200 kg/ha, despite declining dietary DE levels apparent at <400 kg/ha to 500 kg/ha of accepted species. Elk traveled faster while foraging to compensate for plant communities with very low abundance of total forage, increased bite rate as bite mass declined, increased time spent feeding at night in pens with low abundance of total forage or relatively low dietary DE levels, and increased rumination time particularly as dietary fiber levels increased. Dietary DE, DP, and intake rates of these nutrients therefore were robust to substantial variation in overall forage quality and quantity. Nevertheless, these strategies were insufficient to compensate for low abundance of high-quality forage typically present under closed forest canopies. Our nutrition model included nonlinear and multiple regression equations to predict 1) dietary DE (kcal/g of ingested forage), based primarily on abundance of accepted species ( r2 = 0.49-0.62); and 2) dietary DP (% of ingested forage), based primarily on abundance of accepted species, overstory canopy cover, and site characteristics intended to index soil moisture ( r2 = 0.60). Additional equations to predict intake rates per minute included the same covariates, but the variance explained was modestly lower (DE intake: r2 = 0.43; DP intake: r2 = 0.45-0.54). With these equations, we created nutrition-succession profiles to illustrate dietary DE and DP intake dynamics across the successional sequence for each habitat type and study area. These profiles may serve as inputs for spatially explicit maps of nutritional resources for elk. Because they were developed using nutrition data from foraging elk, they should help alleviate much of the uncertainty arising from proxy variables often used as indices of nutritional resources. Our data demonstrated that nutritional resources in forests of western Oregon and Washington are generally deficient for lactating elk in summer and early autumn. They provided evidence that inadequate nutritional resources are largely responsible for low body fat in autumn and reduced pregnancy rates reported for many elk herds in the Pacific Northwest. Our data also illustrated that nutritional value of habitats is highly variable depending on ecological context, disturbance, and succession. Thus, how, if, and where forested elk habitats are managed can greatly influence the nutritional suitability of an area. Finally, our data indicate a considerable need for integrating nutritional assessments in landscape planning processes where maintaining abundant and productive elk populations is one of several forest management goals in the Pacific Northwest. © 2016 The Wildlife Society. [ABSTRACT FROM AUTHOR]