Monday, February 22, 2010

Re-tracing the steps of the Climate guys

Even as the edifice of global warming theory crumbles, many lay people who have been agreeing with it, even advocating for it at parties, are falling back on a position like this: I don't understand the science, but I can't see how a group of reputable scientists could all have been dishonest, in exactly the same way, over a period of years.

I'll give in to the snark for a moment and say this is a bit like a defender of the Catholic Church, not so long ago, saying it is unthinkable that priests, in any significant number, would abuse children, given their vows, etc. Or better: it is unthinkable that the Church would cover up such crimes once they were discovered, move the same bad priest from one innocent parish to another, etc.

But I digress. If you make statements that are wrong, repeatedly, for twenty years, it seems that you have displayed either dishonesty or ignorance, or a mixture of the two. When there is a big moral and political debate (if you don't agree with us, you're a bad person), then "ignorance" can certainly include a kind of moral zeal that prevents you from looking carefully at evidence.

It’s increasingly clear that 20th-century temperature data, which should be the most solid information available to climate scientists, is suspect. Weather stations are sometimes included, and sometimes not. It was necessary to subject data to certain processes—for example, to allow for the urban heat island (UHI) effect—but as data was “homogenized,” it became very difficult for anyone studying the published results to trace the connection between actual readings and published results. Willis Eschenbach has made a number of investigations of individual weather monitoring stations. How do temperatures from a given station, or a group of stations, show up in the official, never-to-be questioned data sets with trends? Is there any way of figuring out how "trends" were generated, given the raw data that is available? (See also here).

Now he has looked at Anchorage Alaska. Since World War II Anchorage has grown immensely, so any use of temperature data from there would have to account for the UHI effect--heat caused by buildings and other human artifacts, not by climate. GISS does indeed show Anchorage "homogenized"--not only not to be warming, but actually to be cooling, presumably to counteract the warming that would be real but deceptive. Oddly, the decrease appears in a graph in stair-step fashion. Then Willis looks at the closest rural station, since GISS says one way to correct for problems at an urban station is to blend results with the nearest rural one. Bizarrely, Matanuska is shown as having a huge temperature increase beginning in 1970.

Somehow the climate guys have told the computers to homogenize, and in some cases to correct for artificial warming. But somehow their programming instructions have also caused warming to be introduced in a bizarre and arbitrary way. The overall result generally shows "warming trends," but based on what?

If temperature data were all “homogenized” to make it warmer, that would look like a conspiracy to conceal the truth—flat-out lying. Instead, we have a more human situation: some results made warmer, some cooler, and some just making no sense compared to actual thermometer readings. Yet the warming trend is the one that is published, and constantly repeated by the media.

It seems most likely that climate scientists began with a somewhat rudimentary knowledge of computer programming and statistical methods. They began entering temperature data into some program or programs, to see what trends resulted. They probably produced a number of results that they have never revealed publicly—some of them showing non-warming, some of them probably showing cooling, some of them showing obviously anomalous or ridiculous results, such as a very warm place presented as cool, or whatever.

Eventually they started to produce warming trends. How exactly did they do that? Did the programs and methods that produced warming trends pass tests of statistical significance, moreso than other results that were rejected? Could the results be reproduced? These are the questions that Steve McIntyre has always focussed on, to his lasting credit.

The lying/clueless question becomes more urgent once these “warming” results are used as the basis of “peer-reviewed” papers (reviewed by very few peers), and ultimately as the basis of IPCC reports. To what extent did the climate scientists provide disclaimers that their science was still new, and conclusions were tentative? To what extent were they aware that their published warming trends were shaky at best? To what extent did they deliberately make it difficult for anyone to check their work? As time went on, it would seem that there was more and more dishonesty, as compared to moral zeal combined with clumsiness, and rushing to get into print and get more grant money and acclaim.

I’m impressed at the moments in the CRU e-mails when the big shots admit, to a few friends, that they don’t know much about what causes temperature to be as it is for a period of time. See here and here.

No comments:

Post a Comment