Anthony Watts has a couple more stations in hand (Electra and Fallon) about which I’ll comment later today. The detailed examination of USHCN stations fits into the prior debate initiated by the Davey and Pielke 2005 survey of sites in eastern Colorado.
The original Davey and Pielke article was accompanied by an official response from Vose, Karl et al in which one of their primary defences was that Pielke et al had merely found problems in one area and had not proved that there was a problem over the entire network.
If one views the Davey and Pielke study as a type of audit spot check, this aspect of the Karl and Vose response is both obfuscating and appalling and should have raised alarm bells in any properly administered system.
Pielke Sr gives a little history to his article at his blog here. According to him, the AMS took three years to process the article, in part to accommodate the simultaneous publication of an official response from Vose, Karl et al, the editor of BAMS explaining:
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.
The Vose, Karl et al 2005 response, entitled Comments on “Microclimate Exposures of Surface-Based Weather Stations” , BAMS, 2005 is online here.
They begin by re-iterating the mantra that USHCN is a “high-quality” network:
USHCN is a high-quality subset of the much larger U.S. Cooperative Observing Network, stations having been originally chosen based upon factors such as their spatial coverage, record length, data completeness, and historical stability (i.e., the number of changes in station location, instrumentation, and observing practice). In addition, the recommendations of the nation’s State Climatologists, whom were often knowledgeable of site exposures, figured prominently in the selection of USHCN stations (Quinlan et al. 1987).
They don’t mention anything about barbecues or incinerators. They then proceed to defend themselves on the basis that Davey and Pielke have only shown problems in eastern Colorado:
Furthermore, their analysis was a static assessment of site exposures over a relatively small part of the country, an area within which station exposures varied considerably. In other words, their results do not show that a large number of USHCN stations have a comparable exposure problem…
Responsibility for quality control rests with the agency responsible for the network. It wasn’t Pielke’s job to survey the entire country. Pielke did a spot check and reported that the claims that the network was “high-quality” in this area were false. Consider Pielke’s article a type of audit. Business auditors do spot checks all the time. They don’t have to explain why they picked one invoice or one contract to examine. If they find problems with an invoice or contract, they will expand the scope of the examination. Can you imagine the response if a company told an auditor: well, you’ve just found one fraudulent contract, you haven’t shown that a large number of contracts are fraudulent. While this may well be the hope of the company, once this type of problem comes to light, it has to be investigated. You’d have thought that this would have prompted a program to re-assess the USHCN network, but Karl appears to have taken no action whatever to institute any additional quality control on the USHCN network.
Karl agreed that “documentation” could be “improved” and viewed actual photographs as an “ideal” situation but, even though sites are supposedly inspected, the idea of actual photographs was laughed off as a kind of utopian fantasy:
the USHCN database could definitely benefit from improved site exposure documentation. Under ideal conditions, this new documentation would meet the high standards set forth by Davey and Pielke (2005). Until such metadata become available, however, we encourage the users of USHCN—particularly those interested in relatively small study areas, such as eastern Colorado— to review Cooperative Station Reports (i.e., B-44 forms), which are available online from the National Climatic Data Center. Although most station histories no longer contain site sketches, these forms still contain a plethora of information about each station, including numerical values of azimuth, range, and elevation (which with some effort could be used to create pseudosketches of site exposure). We also encourage users to contact State Climatologists and the National Weather Service for additional information about exposures around individual stations.
Their one valid point – a point re-stated in Peterson 2006 – was that Davey and Pielke 2005 did not show a connection between poor siting and differing trends – something that Peterson 2006 attributed (incorrectly in my opinion) to the success of their homogenization adjustments. In the eastern Colorado situation, none of the sites, even the bad sites, had modern values in excess of those of the 1930s. It’s the sites with strong trends that really need to be examined. In addition, because USHCN2 uses Reno as a type case, the current site census in the Reno-California area will give material for a check on the Reno calculation.
One question about the Karl article. They say that the B-44 forms are online. I spent some time searching and couldn’t find them. I found one example of what a B-44 form looked like in a larger reportand show it below (it includes good directions in this case, by the way.) The very specific directions show that a site inspector has been there (and presumably other sites). This suggests that it is neither impossible nor utopian for USHCN inspectors to photograph the site installations and the failure to require this is presumably due to flaccid administration by Karl and his associates.