Genetic differentiation and local adaptation of dominant grass along the climate gradient of the Midwest: Implications for climate change

Background/Question/Methods

Big bluestem is an ecologically dominant grass in Central US grasslands. With wide distribution across a precipitation gradient (400-1200mm yr^-1), we expect ecotypic variation in drought tolerance and local adaptation. Understanding ecotypic variation will help predict how a dominant prairie grass may respond to current and predicted future climate change. Because genetic diversity may be critical for predicting a species’ ability to adjust to climate change, we assessed genetic diversity and population differentiation.  We investigated phenotypic variation, genetic diversity and transcriptional variation using reciprocal gardens across the Midwest gradient (wet to dry) in Carbondale, Illinois, Manhattan and Hays, KS and a site in Colby, KS (to test tolerance in even drier areas). Three ecotypes (each comprised of seeds collected from populations in Central KS, Eastern KS, and Illinois) were reciprocally planted in seeded communities with competitors. We measured phenotypic variation in drought tolerance across ecotypes and sites in the reciprocal gardens from 2010-2014. Additionally, we expanded our study to 44 geographically distributed populations across the central US grasslands to model current and predicted big bluestem distribution under climate change.

Results/Conclusions

A phenotypic cline in drought tolerance of ecotypes was observed with plant cover showing significant ecotype, site, and interaction effects. Central KS (CKS) and Southern Illinois (SIL) ecotypes had greatest cover in their home environment, indicating local adaptation. Ecotype-specific outlier loci were identified as under diversifying selection and related to seasonal temperature and precipitation. While our populations are as much as 1100 km separated from each other, “isolation by environment” played a more important role than “isolation by distance” in population differentiation. CKS and SIL ecotypes also varied in their gene expression profiling, with the SIL ecotype differentially upregulating stress genes. On a broad geographic scale, MAXENT species distribution modeling indicates the current center of dominance in Kansas will shift to the northern Midwest and southern Canada, under scenarios of doubling of CO₂ by 2070. Our results provide insights and predictions about the extent to which a dominant prairie grass may respond to current and future climates, via plasticity or genetic adaptive variation and may also guide prairie restoration strategies.