Ethon checks out the air conditioning. . .

UPDATE: As Boris says in the comments:

Paul is right on here, which explains why the denialist side is so interested in pushing this nonsense. It doesn't matter if the trends aren't changed at all by microsite issues, doesn't matter at all that satellite data matches the surface trends extremely well, even for the US region. What matters is that we have a propaganda tool to mislead the public, the same way we mislead the public with the increases of high altitude glaciers, the lag between temp and CO2 in the ice record, erasing Hansen's graphs, and on and on and on and on...

UPDATE: BCL has some rather good stuff on the ways of the righteous and scientifical or at least how to do the job right. Curiously, very similar to what has been recommended here.

Ethon had been looking over the real estate ads at surface stations and having seen some promising properties he went out west checking out the Class 5 stations for one with a nice nesting place, A/C and perhaps a hibachi to cook the daily liver on. On the way back he dropped in at NOAA and ran into TC Peterson in the cafeteria. Tom was happily munching on his chopped liver sandwich (believe me folks it is ALL chopped liver).

Ethon and TC fell to nattering about the seminal paper by Davey and Pielke which, with little exaggeration, could be pointed to as the seed for all the heavy breathing. They did a bit of surfing and indeed did find the smoking BBQ at the defunct Climate Science blog in a comment from Roger Sr.

Thanks Dave for your comment. Their adjustments do not address the issue of whether unrecognized systematic biases are still retained due to poor microclimate exposure and its change over time. Our recommendation is that each station used to construct the USA and global land-surface temperature data record be photographed, in the manner presented in http://blue.atmos.colostate.edu/publications/pdf/R-274.pdf. As other examples of sites, we are compiling photographs on our web site (e.g., see http://ccc.atmos.colostate.edu/Alaskacoopsites.php) (with more to come).

Davey and Pielke took their Brownies and photographed some stations out in the mountain west. Out of the 10 stations that posed, six were in the USHCN network, two were "good" and four "bad". TC, being essentially the source of all homogenizations guy at NOAA (don't take this too literally) had examined the photos that Davey and Pielke relied on and looked at the data. As he said

Essentially there are two competing hypotheses about the effects of poor siting that yield very different predictions. The first hypothesis is that homogeneity adjustment methodologies would account for changes to locations with poor siting. If the homogeneity adjustments are appropriately accounting for all artificial changes at the stations, then an adjusted temperature time series from the poorly sited stations should be very similar to the time series from the stations with good siting. The trends from the poorly sited stations may be a little higher or a little lower, but they should still be about the same. This hypothesis would, of course, also hold if poor siting did not cause a bias in the original data and the homogenization did not introduce any biases. The second hypothesis is that poor current station siting produces an artificial bias in the temperature record that is not being addressed by homogeneity adjustments. While Davey and Pielke suggested that poor siting–induced bias could be positive or negative,

a point that appears to have been lost at Surface Stations although not Rabett Run as attested by some mouth foaming comments to our Cool Station of the Day series

the underlying concern about the effects of potential siting biases is whether a significant portion of the recent warming indicated by the U.S. and global temperature record could be due to this bias rather than climate change. Therefore, the second hypothesis predicts that homogeneity-adjusted temperature trends at the poorly sited station would be significantly different than the temperature trends at the stations with good siting, and that these differences would most likely be that the poorly sited stations are warming relative to nearby stations with good siting

Peterson ran the numbers, and in the words of Anthony Watt

But hey, they can "fix" the problem with math and adjustments to the temperature record.

which, of course is what Peterson found: that the trends for the good, the bad and the not very nice looking were the same after homogenization.
Now, of course, our friends in Boulder were not about to take this without some harumphing

I assumed the Peterson article would also be published with a Reply from Christopher Davey and I. However, despite my requests to permit us to prepare a Reply to the Peterson article, it was decided that there was new information in the Peterson article. My request was refused. I was written that
“In the case of your 2005 article, Jeff Rosenfeld felt that since your work raised significant (though potentially justified) criticism of an observing network that the entire scientific community relies upon and would impact the public confidence in those networks, that a companion comment was appropriate to provide additional perspective. This does not appear to be the case with Peterson’s current article, which is simply providing scientific evidence to clarify arguments for alternative hypotheses.” [Jeff Rosenfeld is Editor-in-Chief of the Bulletin of the American Meterological Society].
Since the Peterson article claims to resolve the problem, yet we have serious issues with his contribution, it would seem that the same approach of two articles would have been permitted. Nonetheless, this was not allowed. This imbalance in the ability to present climate science viewpoints unfortunately permeates the scientific literature including that of the Bulletin of the American Meteorological Society (BAMS).
We have, therefore, written an article for BAMS in response to the Peterson article, and it is authored and titled
Pielke Sr., R.A, C. Davey, J. Angel, O. Bliss, M. Cai, N. Doesken, S. Fall, K. Gallo, R. Hale, K.G. Hubbard, H. Li, X. Lin, J. Nielsen-Gammon, D. Niyogi, and S. Raman, 2006: Documentation of bias associated with surface temperature measurement sites. Bull. Amer. Meteor. Soc., submitted. [it should not yet be cited or reproduced as it is currently under review; comments to us on the manuscript, however, are welcome].

Eli, being a RTFR kinda Rabett went and RTFR, which has been published now. The interesting part is the conclusion

As Davey and Pielke (2005) documented and Peterson (2006) acknowledges, several USHCN stations are poorly sited or have siting conditions that change over time. These deficiencies in the observations should be rectified at the source, that is, by correcting the location and then ensuring high-quality data that are locally and, in aggregate, regionally representative. Station micrometeorology produces complex effects on surface temperatures, however, and, as we show in this paper, attempting to correct the errors with existing adjustment methods artificially forces toward regional representativeness and cannot be expected to recover all of the trend information that would have been obtained locally from a well-sited station.

Translated from the we refuse to admit we were wrong this means that homogeneity adjustments do recover regional (and thus continental) and global trends BUT, of course local information is lost and it would be better to have better stations. Let the perfect be the enemy of the useful and tally-ho.