A Test of Corrections for Extraneous Signals
in Gridded Surface Temperature Data
Ross McKitrick, University of Guelph Department of Economics
Patrick J. Michaels, University of Virginia Department of Environmental Sciences
Climate Research 26(2):159-173 (2004)
Monthly surface temperature records from 1979 to 2000 were obtained from 218 individual stations in 93 countries and a linear trend coefficient determined for each site. This vector of trends was regressed on measures of local climate, as well as indicators of local economic activity (income, GDP growth rates, coal use) and data quality. The spatial pattern of trends is shown to be significantly correlated with non-climatic factors, including economic activity and sociopolitical characteristics of the region. The analysis is then repeated on the corresponding IPCC gridded data, and very similar correlations appear, despite previous attempts to remove non-climatic effects. The socioeconomic effects in the data are shown to add up to a net warming bias, although more precise estimation of its magnitude will require further research.
NOTICE OF CORRECTION:
There was a small error in the calculation of regression coefficients in our paper. Our conclusions were not affected by this problem, and corrected command files and output files are posted below. A correction notice published in CR Vol 27(3) is available which explains the changes to the published tables of results.
DATA and CODE:
Note: If you want to download and use these files on Unix, remove the ".txt" extensions from the filename.
Go up to Publications and Papers Page
Return to Ross McKitrick's home page