“IPCC – How not to compare temperatures“. Anthony Watts posts Frank Lansner’s complaint on a favorite subject, alleged flaws in how the IPCC compares temperatures. Frank says that “there are numerous issues discussed intensely when it comes to IPCC-illustrations of historic temperatures” but other than the standard check list of groundless accusations and fabrications, his case really boils down to this: the IPCC is using too many temperature data sources and is averaging them!
So suddenly the problem isn’t lack of data, now it’s too much data. I also have to note with some astonishment that this is a complaint about how the IPCC illustrates historic temperature trends, not about the statistical trends that they are attempting to illustrate.
Sorry Ben, and I’m not sure how good your math background is, as it seems to be not that flash, but long scale averaging DOES introduce bias when combined with shorter scale averages.
[You understand that this “problem” is only about the summary illustrations in the IPCC reports don’t you? The real statistical trends are generated from the unsmoothed data. The WUWT posturing in the criticized post could actually be useful if they took the time to accurately demonstrate it instead of just accusing the published scientists of “doing it wrong”. There’s usually a reason why they don’t though… – Ben]
Sure that may be the case when looking at recent trends. But the smoothing (and/or resolution) used for reconstructions masks the true variation that occurred during that time, and gives a completely false picture when spliced onto more recent data that has a shorter averaging time/resolution. The point that is made in the post is entirely valid in my view. It is these mixed data series that are used to convey the messages. One only has to recall the Gore Hokey stick, that still gets promoted as being the real deal, to realize how powerful these incorrect presentations are.
[Ah, the Gore “Hokey” Stick. My advice is to drop that subject, you’re just revealing your truculence. I don’t think Gore created it, and it has never in fact been legitimately challenged. Your raising of it does hold to consistent denialist theme though – assaults on particular charts rather than on the evidence they are depicting.]
[Regarding smoothing, I agree that for most scientific purposes showing the full data behind the trends is better. But it seems to me that arguing against it is a tactic to exclude as many of the most recent years as possible and hence “hide the incline”… – Ben]
And still just as Hokey :)
Terry, Ben’s right. This post, like almost everything at WUWT, shortchanges the scientific process to get at Anthony’s pre-determined answer. Whatever merits may exist in Lansner’s analysis, you’re not going to reveal them by bypassing a careful review of the original study (in this case, studies) and having your own work peer reviewed. It’s revealing that the one time Anthony was given an opportunity to participate in that process (the Menne et. al. paper on station data), he declined.