I got to thinking from discussions with Steve Mosher and others at Climate Audit about just what sort of adjustments might be made to a place with a good record and little in the way of station moves, and that had changed little in 100 years.
One place came to mind that I’ve visited; Cedarville, CA
It has it’s issues, such as the Stevenson Screen being encroached upon by things being built around it, like a new concrete pad for the Forest Service office.
But as the town goes, it has changed very little in 100 years. There is no Interstate highway nearby, its off essentially in the middle of nowhere by itself, a self contained agrarian community, mostly hay farmers. The small main street has many of the original buildings from 100+ years ago:
This is the sort of view (below) you can see to the east of the town, open land as far as the eye can see. It is definitely rural.
Image above from the Surprise Valley website
Here is a Google earth view. The weather station is at the far north end of town.
According to US census data:
As of the census of 2000, there were 849 people, 381 households, and 249 families residing in the ZCTA of 96104. The population density was 3.2 per sq mi. There were 457 housing units at an average density of 1.6/sq mi.
There are only 3 missing years all the way back to 1894, so it seemed like a good candidate. 1894 is a partial record, 1915 and 1957 are also partials with not enough data to complete the yearly average.
So this seemed like a really good candidate to test for what adjustments GISS might make to it.
I plotted the GISS provided USHCN data and the homogenized data from the GISTEMP website and the graph for that is shown below:
Note the missing year at 1957, 1915 should have not plotted also, but for some reason my program insists on doing so.
But the important thing is that once again, GISS has made the past colder and the present is unchanged. Even more odd, the far past prior to 1900 is adjusted upwards, warmer.
Yet it is the recent past to present when the most change has occurred in Cedarville, such as the addition of a concrete pad to the Forest Service building, etc. If anything, pre 1900 data should likely be colder because exposure of thermometers was not standardized until after the US Weather Bureau was formed in 1892.
They seem backwards. Why does GISS do this? We’ll find the answer.