Grasslands of the central US are a critical national natural resource, accounting for $8 billion annually in agricultural production in Kansas alone (NASS 2018). These grasslands are dominated by one of the most palatable forage grasses, Big Bluestem or Andropogon gerardii (Dwyer 1961, Boe et al. 2004), which comprises up to 70-80% of aboveground biomass (Rogler 1944). In this proposal, we would investigate the genetic and phenotypic responses of A. gerardii to drought across the Great Plains and construct a predictive spatial model for forecasting future responses to drought. Owing to the foundational role of A. gerardii in Midwestern agriculture, this work is critical to understanding the basis of a significant portion of this region's agricultural production. Our research thus addresses priority areas by examining processes in grasslands that determine productivity and resilience to drought. Andropogon gerardii grows across a strong precipitation gradient ranging from 400 to 1200 mm of annual rainfall spanning the Great Plains. Changes in amount and timing of precipitation are critical abiotic stressors determining the future productivity and sustainability of these grasslands. Indeed, western US grasslands regularly experience drought (NOAA 2012) and severe "megadroughts" are expected to become increasingly common (Cook et al. 2010). Understanding how ecotypic and genetic variation of a dominant forage grass will respond to increased temperature and drought is essential to preserving rangeland productivity and informing prairie conservation.Our overall goal is to understand and predict changes in genotype and phenotype of A. gerardii under anticipated changes in temperature and drought. This proposal builds on and extends our USDA-funded research that uses reciprocal gardens arranged along a rainfall gradient extending from Carbondale, Illinois to Colby, KS. This research platform has enabled us to uncover ecological, physiological, and genetic responses of A. gerardii to drought within realistic grassland communities (Olson et al. 2013, Caudle et al 2014, Mendola et al. 2016, Johnson et al. 2015, Galliart et al. 2019). This proposal extends the research to multiple populations of A. gerardii across the Great Plains, and uses phenotypic and genetic data from these sites as input for a novel predictive spatial model. New to our approach are expanded research platforms that include rainout shelters over reciprocal gardens as a model for experimental drought; complemented by a characterization of phenotypes, genotypes, and transcriptomes from 44 geographically diverse populations across the range of A. gerardii in the Great Plains (CO to MI, OK to ND); and drought simulations under controlled greenhouse conditions. Taken together, the expanded research platform would generalize our previous work on the effects of drought on agricultural production to the entirety of the Great Plains .To accomplish our goal, we address the following objectives: Objective 1: Characterize phenotypes and genotypes from 44 geographically diverse populations across the Great Plains rainfall and temperature gradient (including range edges). This is a complement to our well-characterized ecotypes from reciprocal gardens from a narrower cross-section from KS to IL. Objective 2: Characterize differences in gene expression to drought in two sets of experiments: a) from phenotypically divergent dry and wet ecotypes in response to experimental rainfall drought manipulation in the field using long-term reciprocal gardens; and b) in greenhouse drought manipulations using populations from across the geographic range, including populations from climate extremes. Objective 3: Integrate transcriptome expression and genotypes for biomass, and ecophysiological responses of divergent ecotypes to altered rainfall. We will then use these genetic and phenotypic variables as input to a novel phenotypic-genotypic distribution model. Objective 4: Develop a novel, joint phenotypic-genotypic distribution model to predict current and future distribution of genotypes and phenotypes. The model will provide critical information for range managers for anticipating the effects of future growing conditions on grassland function.