Meet the new climate temperature data, not the same as the old climate data.
Paul Homewood reports
the decision of Dr. James Hanson, director of the Goddard Institute for Space Studies (GISS), to rely on a different source for ocean surface temperatures - ERSST (NOAA's Extended Reconstructed Sea Surface Temperature) instead of the "Hadley/Reynolds" (evidently referring to Hadley Centre for Climate Prediction and Research
and NOAA's Richard W. Reynolds
Dr. Reto Ruedy responded to Homewood's inquiries, stating that the data changes "are well within [the] margin of error," and that "all our basic data (land and ocean) com[ing] from the same source...simplifies the description of what we are doing." Homewood states that GISS had "not given any reason for abandoning Hadley/Reynolds, or suggested there are errors or inaccuracies with it," and believes the change is outright data manipulation.
Homewood also discovered that original temperature records often do not match those recorded in the 2008 publication of the GISS historical data - meaning that there have been two sets of changes.
Homewood also blogged on this at Watt's Up With That
. Note that the graph shows the differences between GISS 2008 and GISS 2013 (brought about by the switch to ERSST), and not the differences between GISS 2008 and the original records.