There is an excellent article by Joseph D’Aleo at Pajamas Media titled
It very effectively summarizes a number of major issues with the quality of the land portion of the long-term surface temperature trend record that was used in the 2007 IPCC report, and is being assumed as robust at the current Copenhagen meeting.
I recommend this article for anyone who wants to see how really bad this temperature data is with respect to its application to the quantitative assessment of long-term surface temperature trends.