The Climate Research Unit or CRU at the University of East Anglia is one of a few places where weather data from around the world are gathered and analysed to produce the iconic graphs of rising global temperatures. The 3rd edition of their data product, called CRUTEM3 (i.e. CRU TEMperature data version 3) had the curious feature that from 1998 until 2012 there was a long “hiatus” during which no upward trend was observed. This pause caused no end of annoyance to the alarmists. Then to their relief CRU version 4 came out and the hiatus vanished, with 2005 and 2010 edging ever-so-slightly higher than 1998. At the time, people assumed the reason was CRUTEM4 added in records from thousands more locations. But UK blogger Clive Best looked closely and found a different explanation: the CRU had changed the weather station values themselves, so even if they still used CRUTEM3 the hiatus would have disappeared. Science is so much easier when you can change the data to fit the theory.

The ever-changeable nature of climate data must make life easier for climate alarmists. We noted previously that satellite data warming trends depend heavily on a decision about whether or not to use the last few years of records from a dying satellite. Since the various teams can't agree, you get a choice of how big a trend you end up with.

The surface station record is even messier, and the ease with which downward trends become upward ones with every new revision gives any fair-minded observer considerable skepticism about the purity of the data. Blogger Tony Heller at realclimatescience.com has documented startling changes in historical temperatures that always have the effect of cooling the past and warming up the present.

When the CRU put out their 4th edition, the station count went up from about 4100 weather stations to about 7800. What Best showed was that the original 4100 records used in CRUTEM3 looked the same in the CRUTEM4 version up to about 1998 but from then onwards they had been nudged upward by just enough to eliminate the hiatus. It is remarkable that meteorologists knew how to record temperatures up to 1998 but then afterwards kept getting it systematically wrong. And oh so convenient that the experts at the CRU were available to fix the problem.