Assessing Ecosystem Sensitivity to Chronic Resource Alterations: A Synthesis of Long-term Experiments

The scale, scope and pace of change occurring in ecological systems today and forecast for the future are by all accounts unprecedented in human history. The global scale of this change means that all ecosystems are, and will continue to be, impacted to some extent by alterations in atmospheric CO2, temperature, precipitation and nitrogen deposition. One apect that all of these global changes have in common is that they will result, either directly or indirectly, in chronic alterations in resource levels for ecosystems. Forecasting how any particular ecosystem will respond to these chronic alterations key resources (N, water, etc.) requires knowing (1) the magnitude of the change (e.g., change in N or precipitation), and (2) the sensitivity of the ecosystem to a given change in a resource, including identifying if there are potential thresholds of chronic resource change that may trigger rapid, non-linear ecological responses. At present, we know much more about how resources are likely to change than the sensitivity of ecosystems to these forecast changes. We developed a framework to describe four possible ecosystem responses to global change factors: (1) no response; (2) transient responses; (3) directional responses; (4) lagged responses (which could occur within either transient or directional responses). With data from a wide range of LTER long-term experiments in which resources have been altered chronically, as well as data from other long-term experiments, we tested the relative responses of ecosystems to chronic changes in key resources. We then examined whether ecosystem responses resulted from changes in community composition (due to species turnover or immigration) as predicted by the Hierarchical Response Framework (Smith et al 2009). We examined data from 28 chronic resource manipulation experiments with a total of 58 community types. Across these community types, we identified the temporal ANPP responses to 145 global change treatments. We found that approximately half of the global change treatments had an effect on ANPP (49%). Of those, most responses were lagged or directional (26% and 21%, respectively), with very few transient responses (6%). Community shifts only correlated with ecosystem response in 5 of the 71 observed ANPP responses. Overall, we found that many ecosystems will respond to chronic global change drivers, however these responses will likely vary by ecosystem type. Long-term experiments are important for determining ecosystem responses to global change drivers, as many responses are likely to be lagged. However, once a response is observed, it is likely that it will be maintained, as few ecosystem responses in this study were found to be transient. Finally, our understanding of what triggers shifts in ecosystem responses to global change drivers is still limited and future studies examining these triggers are warranted.