Here’s another strange CRU gridcell in which warm early values of Huhehaote aren’t used.
Here is a plot of the HAdCRU3 gridcell values (heavy black points) against 13 individual station anomaly series. The early record comes from Huhhehaote which, along with 9 other stations), has no recorded values after 1990. Post-1990 values come from Erenhot and two Russian stations, Sainshand and Zanyn-Uud. What is the reason why the early values for Huhehaote aren’t used? I’m sure that the reasons are “good” reasons, but have these “good” reasons been consistently applied? Early station values are used in other stations.
Here is a comparison of the NOAA gridded version (black) to HAdCRU3 (red). The Huhehaote values seem to be good enough for NOAA to use, but not for CRU. Brohan et al 2006 casts no light on the matter.
Again, while I’m sure that CRU had a “good” reason for not using the early data for this gridcell, the effect of not using this data, as so often with seemingly arbitrary individual decisions, is to increase the amount of measured 20th century warming.
The explanation is not the CRU is shy about using disconnected early data. Here’s another Chinese gridcell in which cool early 20th century station values are used. The heavy black is HadCru3 in this plot.