Several simple indices of surface air temperature patterns are used to describe global climate variability and change. The indices include the land-ocean temperature contrast, the interhemispheric contrast, the meridional gradient, and the magnitude of the seasonal cycle, as well as the global-mean temperature. These indices retain the main features of the fingerprints of greenhouse climate change but are easier to interpret. Also, they are associated with dynamical factors determining the large-scale atmospheric circulation. For natural climate variations, they contain information independent of the global-mean temperature.

The behaviour of the indices is investigated using global observational data for the period 1881-1999, paleoclimate reconstructions from proxy data for 1700-1990, and long control and anthropogenic climate change simulations with five different coupled ocean-atmosphere climate models. Comparison of the variability and correlation structure of these indices from model simulations with observational data provides a simple but thorough evaluation of the model simulation of global-scale surface temperature variability. This is a progress report on a CMIP sub-project.

The variability of all the indices on interannual and decadal timescales from the control model simulations compares well with detrended observational data and with the proxy data for 1700-1900. Hence, the simulation of global-scale surface temperature variability in these climate models is surprisingly good. The observed trends over the last 40 years in all the indices are consistent with model simulations of anthropogenic climate change and are unlikely to have occurred due to natural climate variations, apart from the interhemispheric contrast. Hence, these indices provide simple and easy-to-expain evidence of a human impact on global climate.