Long-term research and the future of grassland ecosystems: What have we learned?

Background/Question/Methods: All ecosystems are dynamic in space and time, but grasslands are among the most responsive, temporally and spatially, to changes in disturbance regimes, resource availability and climate. Forecast alterations in climatic means, variability and extremes will influence resource availability within all ecosystems, both directly and indirectly, and long-term experiments coupled with observations can provide unique insight into how the structure and function of ecosystems may be altered in the future.  We have synthesized results from several long-term experiments and observational data sets to assess how grasslands, as diverse and globally important ecosystems, will respond to expected alterations in resources with future climate change. We focus on water availability because of the key role of precipitation in determining the dynamics of most grassland ecosystems temporally and spatially. 

Results/Conclusions: Based on results from long-term experiments (15-20 years) in which water availability has been manipulated in native grasslands, coupled with long-term climatic and ecological data (aboveground net primary production, ANPP), ecosystem function is expected to respond dramatically to some types of climate change, while demonstrating surprising resilience to others. Similarly, across gradients from arid to mesic grasslands, sensitivity to climate change may be quite variable.  We conclude that long-term experiments are essential for assessing how chronic changes in resources may impact ecosystem function, particularly if responses are driven by community change. However, long-term observations are equally important for providing system context regarding the nature and magnitude of responses to both chronic and episodic (extreme) aspects of forecast climate change. While the role of long-term studies in advancing ecological understanding of the consequences of climate change is clear, networks of experiments coupled with long-term observations are needed to evaluate ecological responses at scales commensurate with the key drivers. At this macro-scale, a consistent definition of ecosystem sensitivity will be required.