Edited by William H. Schlesinger, Cary Institute of Ecosystem Studies, Millbrook, NY, and approved November 26, 2014 (received for review May 30, 2014)
Methodological constraints can limit our ability to quantify potential impacts of climate warming. We assessed the consistency of three approaches in estimating warming effects on plant community composition: manipulative warming experiments, repeat sampling under ambient temperature change (monitoring), and space-for-time substitution. The three approaches showed agreement in the direction of change (an increase in the relative abundance of species with a warmer thermal niche), but differed in the magnitude of change estimated. Experimental and monitoring approaches were similar in magnitude, whereas space-for-time comparisons indicated a much stronger response. These results suggest that all three approaches are valid, but experimental warming and long-term monitoring are best suited for forecasting impacts over the coming decades.
Inference about future climate change impacts typically relies on one of three approaches: manipulative experiments, historical comparisons (broadly defined to include monitoring the response to ambient climate fluctuations using repeat sampling of plots, dendroecology, and paleoecology techniques), and space-for-time substitutions derived from sampling along environmental gradients. Potential limitations of all three approaches are recognized. Here we address the congruence among these three main approaches by comparing the degree to which tundra plant community composition changes (i) in response to in situ experimental warming, (ii) with interannual variability in summer temperature within sites, and (iii) over spatial gradients in summer temperature. We analyzed changes in plant community composition from repeat sampling (85 plant communities in 28 regions) and experimental warming studies (28 experiments in 14 regions) throughout arctic and alpine North America and Europe. Increases in the relative abundance of species with a warmer thermal niche were observed in response to warmer summer temperatures using all three methods; however, effect sizes were greater over broad-scale spatial gradients relative to either temporal variability in summer temperature within a site or summer temperature increases induced by experimental warming. The effect sizes for change over time within a site and with experimental warming were nearly identical. These results support the view that inferences based on space-for-time substitution overestimate the magnitude of responses to contemporary climate warming, because spatial gradients reflect long-term processes. In contrast, in situ experimental warming and monitoring approaches yield consistent estimates of the magnitude of response of plant communities to climate warming.
Author contributions: S.C.E. designed research; S.C.E., G.H.R.H., R.D.H., A.M.F., W.A.G., L.H., A.H., I.S.J., J.C.J., E.L., B.M., U.M., I.H.M.-S., S.F.O., C.R., C.E.T., and M.D.W. performed research; S.C.E. analyzed data; and S.C.E., G.H.R.H., and R.D.H. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
Data deposition: Plant abundance data from the unmanipulated and experimentally warmed plots have been deposited in the Polar Data Catalogue at www.polardata.ca (CCIN ref. no. 10786).
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1410088112/-/DCSupplemental.
