Using dynamic foodscape models to assess bottom-up constraints on population performance of herbivores

Resource heterogeneity governs a multitude of ecological processes, but the mechanisms by which heterogeneity influences population performance are not fully resolved. Because optimizing behavior is challenging in heterogeneous landscapes, individual variation in foraging and movement strategies is common, and understanding the consequences of that variation is one of the most pressing challenges in modern ecology. In theory, such consequences should be modulated at least in part by nutrition, which directly influences discretionary energy available for growth and reproduction. We developed a series of linked dynamic models for predicting (1) spatiotemporal variation in the foodscapes available to seven distinct populations of elk (Cervus canadensis) in Idaho, USA, and (2) variation in pregnancy rates among those populations as a function of foodscape use and availability. Foodscape models, which predicted variation in suitable forage biomass (biomass of forage that met or exceeded requirements of female elk at peak lactation), generally performed well, with adjusted R² values ranging from 0.34 to 0.51. Patterns of foodscape use differed among populations and years, with some populations showing selection for the foodscape and others exhibiting indifference or even avoidance of high-quality forage resources. Pregnancy rates ranged from 66% to 100%, and our top model relating pregnancy to metrics of forage availability explained 41% of the variation among 20 elk population-years. Our top model relating pregnancy to foodscape use by elk explained 57% of the variation in pregnancy rates among 12 population-years. Pregnancy rates were influenced more strongly by heterogeneity in foodscape use and availability than by differences in mean or maximum suitable biomass among populations. Our results suggest that population performance of elk was modulated both by the availability of high-quality forage and by factors that constrained use of the foodscape by elk. The dynamic modeling approach we developed for linking nutritional resources to herbivore performance is generalizable to many other species and systems and can be used by wildlife managers to assess whether herbivore populations might be limited by bottom-up factors.