Global climate change is projected to cause droughts of increased frequency, duration and severity in North America. Grasslands are dynamic ecosystems that could be strongly impacted by these precipitation changes. North American grasslands exist on a west-to-east productivity gradient, with productivity primarily determined by precipitation. Regions that differ in productivity may show divergent responses to drought. Comparing drought sensitivity between regions is difficult because many drought experiments are small in scale and methods are not standardized between experiments. EDGE (Extreme Drought in Grasslands Experiment) used identical infrastructure to impose rainfall reductions at six sites spanning desert grassland, shortgrass steppe, mixed grass prairie, and tallgrass prairie. A chronic drought treatment consisting of 66% rainfall reduction throughout the growing season was imposed for four years.
Results/Conclusions
Plant species composition differed significantly between chronic drought and control treatments at all six sites by the third year of treatment, according to permanova analysis (P<0.05). In the third year of treatment, the magnitude of the effect of drought (expressed as Cohen’s D) on relative cover of the dominant species at the two driest sites was approximately 2x the effect size at the wettest site. In the fourth year, the magnitude of the drought effect at one of the two driest sites was over 13x the effect at the wettest site. Nonlinear regression showed a negative relationship between mean annual precipitation and the effect of drought on the most sensitive species at each site. Drought caused declines (or crashes in the case of the two driest sites) in the cover of the dominant C4 grasses, leading to booms of C3 co-dominants and/or ephemeral annual species. This effect was largest at the three driest sites, each of which was strongly dominated (66% to 81% of total plant cover) by a single C4 grass species. Sites with intermediate precipitation levels also showed a decline in the most abundant C4 grasses in the drought treatment, but those sites were co-dominated by C4 and C3 graminoids, such that declines in C4 grasses had smaller effects on community composition. Only the wettest site showed no decline in the dominant C4 grass under chronic drought. Our results indicate that grasslands with lower mean annual precipitation may be particularly vulnerable to extreme drought under future climate change.