What does green mean? Modelling herbivore foodscapes and resource tracking in forested environments

The forage maturation hypothesis predicts that herbivores should select plants at intermediate growth stages because they strike a balance between biomass and nutritional quality. Methods for identifying this key stage of plant phenology commonly rely on satellite-based greenness indices and their derivatives. By combining such indices with fine-scale movement data, ecologists have uncovered many novel behaviours and life-history strategies, including 'green-wave surfing' by herbivores. Yet, few studies have evaluated the degree to which greenness indices accurately reflect variation in forage characteristics or their selection by herbivores in forested habitats. We studied relationships among remotely sensed greenness indices, on-the-ground measures of forage biomass and quality and resource selection by white-tailed deer (Odocoileus virginianus), elk (Cervus canadensis) and moose (Alces alces) in the coniferous forests of northern Idaho, USA. We hypothesized that remote-sensing methods for measuring the green wave (i.e. the phenological progression of understory food plants) are not useful in coniferous forests because variation in greenness is driven mainly by conifers. We also hypothesized that patterns of forage selection by large herbivores are not detectable in the forest using greenness indices. We tested predictions derived from these hypotheses using on-the-ground measures of forage biomass and quality and more than 10 years of movement data from GPS collars. Remotely sensed greenness indices explained <2.5% of variation in forage biomass (adjusted R² < 0.024), which was more readily predicted by variation in tree canopy cover, topography and climate (total biomass: adjusted R² = 0.313; high-quality biomass: adjusted R² = 0.174). Moreover, with only one exception (elk selected pixels where the instantaneous rate of green-up was near its peak), greenness indices were among the least useful predictors of selection by large herbivores. Indeed, although all three species selected forage biomass, elk and moose simultaneously avoided greener pixels. Our results indicate that greenness indices are of limited value for predicting variation in forage resources or their selection by large herbivores in forested environments. Despite their intuitive appeal and ease of accessibility, such indices cannot always be substituted for on-the-ground measurements of plant traits that directly influence nutrient intake by herbivores.