The detection of anthropogenic climate change in observations and the validation of climate models both rely on understanding natural climate variability. To evaluate internal climate variability, we apply spectral analysis to time series of surface air temperature (SAT) from nine coupled general circulation model (GCM) simulations, three recent global paleotemperature reconstructions, and Northern Hemisphere (NH) instrumental records. Our comparison is focused on the NH due to the greater spatial and temporal coverage and validation of the available NH temperature reconstructions. The paleotemperature reconstructions capture the general magnitude of NH climate variability, but not the precise variance and specific spatial, temporal, or periodic signals demonstrated in the instrumental record. The models achieved varying degrees of success for each measure of variability analyzed, with none of the models consistently capturing the appropriate variability. In general, the models performed best in the analysis of combined mean annual land and marine variability.
Bibliographical noteFunding Information:
Funding for LCS and JLB from the David and Lucile Packard Foundation is gratefully acknowledged. JLB would like to thank Curt Covey, Ron Stouffer, Michael Mann, and Keith Briffa for useful discussion of data and methods. The authors also thank everyone who contributed proxy data and model results to this study as well as the two anonymous reviewers for insightful suggestions.
Copyright 2017 Elsevier B.V., All rights reserved.
- Climate model
- Climate variability
- Spectral analysis