To date most assessments of species’ biogeographic responses to climate change employ species distribution models (SDMs) which are constructed on the assumption that each species responds as a homogenous unit to changing environmental conditions. However, field studies demonstrate that local adaptation may be the rule, rather than the exception, making this assumption untenable. Here we extend the SDM framework to predict the phenotypic response of a dominant prairie grass of the North American Central Grasslands, Andropogon gerardii (Big Bluestem). We use machine-learning algorithms to predict the present and future distributions of the species and individual plant biomass, height, leaf width, and blade chlorophyll content as a function of current and future climate. Data permutation was used to estimate the relative importance of each predictor. Data for species-level models was obtained from georeferenced natural history museum collections. Data for phenotype distribution modeling was acquired from greenhouse-grown plants grown from seeds sourced from 33 populations spanning across ~1000-km of the Great Plains’ gradient in precipitation and temperature.
Depending on the emissions scenario, the SDM predicted that by the 2070s the core of the species’ distribution will increase in areal coverage by ~15 to 45% but shift from the middle of the Central Grasslands up to ~800 km northeastward to the Great Lakes region. At the same time, the PDMs predicted that in the current core biomass and height will decline by ~60% and leaf width by 40%. Chlorophyll content displayed a much shallower gradient across the current range and a much weaker response to climate change. The SDM was most sensitive to maximum temperature of the hottest month while the PDMs were most sensitive to mean annual precipitation. Our results portend a dramatic shift in the role of this species from a current dominant that represents up to 80% of productivity to a dwarfed phenotype now found in the short-grass prairie along the margin of the species’ current range. Owing to the long lifespan of A. gerardii (~50 yr) and its role as a “foundation” species in ecological restoration and a major component of livestock forage, anticipating these shifts will be important for ensuring ecosystem services as climate changes.