Correlation Dimensions of Local, Short-Term Climatic Attractors from Observations and a Global Climate Model

View More View Less
  • 1 Atmospheric, Environment, Service, Canadian Climate Centre, Downsview, Ontario, Canada
  • | 2 Department of Physics, McLennan Physical Laboratory, University of Toronto, Toronto, Ontario, Canada
© Get Permissions
Restricted access

Abstract

Correlation dimensions ν were estimated for 500-mb geopotential heights over eastern North America as predicted by the Atmospheric Environment Service-Canadian Climate Centre general circulation climate model (GCM). A value of 6.0 to 7.0 was obtained, which agrees to within experimental error with a previously estimated value determined from measured data. This agreement is encouraging for it may suggest that on short time scales, the GCM is capable of producing fluctuations on a regional scale that resemble those of the real atmosphere. For western Europe, however, observed and GCM time series of 500-mb geopotential heights yielded values for ν of about 8.0 and 6.0, respectively, but display similar power spectra. In contrast, over the tropical Pacific, ν from GCM data either does not exist or exceeds 11. These results show that topologies of local climate attractors vary over the globe, thus making it difficult to speak of a global attractor.

Abstract

Correlation dimensions ν were estimated for 500-mb geopotential heights over eastern North America as predicted by the Atmospheric Environment Service-Canadian Climate Centre general circulation climate model (GCM). A value of 6.0 to 7.0 was obtained, which agrees to within experimental error with a previously estimated value determined from measured data. This agreement is encouraging for it may suggest that on short time scales, the GCM is capable of producing fluctuations on a regional scale that resemble those of the real atmosphere. For western Europe, however, observed and GCM time series of 500-mb geopotential heights yielded values for ν of about 8.0 and 6.0, respectively, but display similar power spectra. In contrast, over the tropical Pacific, ν from GCM data either does not exist or exceeds 11. These results show that topologies of local climate attractors vary over the globe, thus making it difficult to speak of a global attractor.

Save