First, 74% of all U.S. stations are adjusted, while only 37% of ROW stations are adjusted

This is a valid concern about the methodology used to assign the meta-data to the stations, and I know this has been written about here many times before.
Thanks for the links!

http://www.climateaudit.org/?p=2815

http://www.climateaudit.org/?p=2815#comment-219608

http://www.climateaudit.org/?cat=50

what I found was another one of those counter-intuitive downward homogenization adjustments

Is the GISTEMP adjustment mechanism not documented? It certainly appears that the source code is available, and from what I understand, it’s very much possible for there to be positive temperature adjustments at some stations for some time periods using their adjustment mechanism (which doesn’t seem to have a “hey, let’s make 1930 colder” step).

As to the technique they use, my understanding is that they are applying an error adjustment based on the fact that climate variation measurements that are close in proximity are highly correlated. And on the whole, the entire homogeneization technique was designed to adjust for the UHI and it should have a net result of reducing the severity of the global heating trend, not increasing it.

It would be fine to question the adjustment technique, but unfortunately that wasn’t the topic of this post, in fact I’m not sure what the scientific point was, but I’m not impressed that someone who I think knows better would go around pointing out the counter-intuitive nature of the results of the homogeneization technique in an attempt to imply to the layman that it is at best an incorrect thing to do and at worst, [snip]

I’m not an expert, and the early downward adjustment certainly baffled me at first, until I researched it. And judging by the comments, nearly all of the commenters have been suckered in by this trick as well.

Got to eliminate (or at least reduce) those warm 1930s.

Why the wild speculation about whether arbitrary adjustments were made? If you have all the evidence available to prove whether something was done incorrectly, why don’t you just prove it!?

Let me summarize: People are going to the effort of driving to take pictures of all of these stations (a laudable effort indeed!). But instead of doing the small amount of incremental work to actually analyze whether the adjustment technique was correctly applied, the blogger tries to make a big deal out of the “counter-intuitive” nature of the adjustment. There’s something fishy going on at this blog.

As long as we’re speculating wildly, let’s try this one: I’ll speculate that in the critical analysis of these data adjustments, the bloggers _have_ actually run the software themselves and _already_ know that the data adjustments were applied correctly to this station’s data but chose to let people infer that they might not have been.

I may have to do the work myself: There have been references made to the standard adjustment procedures applied by GISS. I can’t seem to find the canonical source for this information, can someone provide a link?

<blockquote>4) No mechanism has been proposed or used by GISS for microsite bias adjustments as we see in this example.</blockquote>

And that’s the point, isn’t it? Not only the point for the specific issue at hand but for the microsite issue writ large.

I have run the numbers of the 502 observed stations, according to their minimum estimated CRN violation ratings. And I will note further that Anthony Watts is VERY lenient when it comes to those violations; it seems to me from the photos the actual violations are worse than he says.

The aggregate measure of all stations measured is a 2 degrees C warm bias.

1.) This does NOT include UHI.

2.) Most of these violations occurred or worsened AFTER 1980, mainly because of the MMTS switchover, but also because of exurban creep bringing with it both waste heat and massive heat sink violations as defined in the CRN handbook.

It is obvious that the CRN ratings estimates are not in serious error as has been suggested on other blogs: we have all seen the jumps in temperatures as the sites have been changed for the worse.

Not only that, but FILENET spreads the bias around, thus masking it. A

And UHI effect is calculated off of often badly compromised rural stations thus resulting in a severe lowballing of UHI.

Call it “positive adjustment feedback”, if you will.

That is a novel idea except it does not apply to this station. The adjustment suggests that this station artificially warmed from 1890 to 1920 then artificially cooled from 1930 to 2000. There is nothing in the metadata or in the surrounding urban environment to support such an adjustment. This does however, provide evidence that the 1930′s were artificially cooled in order to eliminate a natural signal from the record.