Global Weather Dataset being Systematically Corrupted

We will have to check everything these people do from now on.

For the past six days, several climate scientists have discovered an alarming trend: clear evidence of alteration of historical data at weather stations around the world, in order to support the contention of anthropogenic global warming (AGW).


The changes appear to affect the Global Historical Climate Network (GHCN), a project of the National Oceanographic and Atmospheric Administration’s National Climate Data Center. Note that this is the same agency that employs Dr. Eugene Wahl, who might be implicated in the research misconduct allegations made against Michael E. Mann at Penn State University.


Richard Keen at the University of Colorado was the first to notice the changes. On December 5, he published this report comparing his own research into the climate of Alaska with the official version of the UN Intergovernmental Panel for Climate Change (IPCC). He found no evidence of warming in Alaska over the past three decades, and no substantial difference in average temperature between 1935-1944 and the present time. Overall he found a warming trend of 0.69 Kelvin per century over the span of the twentieth century–while the GHCN dataset projects a warming trend of 2.83 K/century. (The Kelvin is the International System equivalent of a Celsius degree.)

Keen published his findings at icecap.us. They were little noticed until three days later, when Willis Eschenbach published this entry at WattsUpWithThat, containing an analysis of the raw data at the weather stations in and around Darwin Airport in Australia, and the data as adjusted by GHCN. Eschenbach charged that the Darwin data had been distorted, not adjusted, and suggested that the data ought to have been left alone, or else given one single adjustment, downward by 0.6 Kelvin, of the temperature record of 1941 and prior. The result of the official adjustments was a linear warming trend of 1.2 K/century, whereas the raw data showed a cooling trend of 0.7 K/century. Had the data received the single adjustment proposed by Eschenbach, the trend might have been nearly flat.
Read more.