Russ Steele, who’s been working closely with Anthony Watts and who also deserves much credit for this recent initiative, reports on Ukiah CA at his blog here
Ukiah is in the same gridcell as Petaluma. Russ shows a picture of the weather station which is on grass – there’s a tree nearby, but no incinerators, barbecues, air conditioner exhaust, MIG fighters, asphalt; it’s not mounted on a concrete slab. According to Russ, The curator says that the station has not changed location since 1931 (this may not be totally correct as the CDIAC file shows a couple of short moves). The CDIAC file shows only one small change in observation time: at 5 pm from 1892 to 1974 and at 6 pm thereafter, so that there is a negligible potential bias in time-of-observation adjustment.
It sounds like the sort of station that should be used a framework for analysis of nearby stations, such as Petaluma. So let’s see what the adjustment jockeys do. You’d better sit down first.
First here is a plot of all seven versions of Ukiah.
Now for the adjustments. Please note that I’ve reversed the orientation in these adjustments relative to my earlier graphics. I’m now showing adjusted minus raw (rather than raw minus adjusted). The reason is that this shows the contribution to HS-ness from the adjustment without having to flip the series over in your head. As you see the adjustments to the Ukiah series (which on the face of it is a site that meets WMO standards a lot better than Petaluma or Fresno airport) are very large and, in all three networks (top three rows) amount to about 1 degree between the 1930s and the present day.
As I mentioned before, I’m starting to try to find a pattern in the GISS raw data and it is seriously weird here. The GISS raw data is equal to the USHCN raw version from Jan 2000 on; it is equal to the USHCN tobs version from Aug 1987 to Dec 1999 and it is virtually equal to the USHCN adjusted version (filnet) prior to August 1987. I’m having trouble thinking of an objective criterion which could arrive at this particular pattern and it’s not the sort of thing that one normally thinks of looking for, that’s for sure.
Here is the plot showing the USHCN adjustments, between the TOBS and station history adjustments. There is negligibe TOBS adjustment (and there has been negligible change in observation time.) The USHCN adjustment has reduced summer temperatures in the 1930s by nearly 2 deg C relative to modern temperatures. A substantial jump change was introduced in 1989. MMTS appears to be introduced in 1984 but there is no contemporary change in adjustments.