I added a tab page about GISTEMP which has more detail on the status of ccc-gistemp. Of note from that page:

It is our opinion that the GISTEMP code performs substantially as documented in Hansen, J.E., and S. Lebedeff, 1987: Global trends of measured surface air temperature. J. Geophys. Res., 92, 13345-13372., the GISTEMP documentation, and other papers describing updates to the procedure.

10 Responses to “GISTEMP tab”

  1. Nick.Barnes Says:

    I’m not sure the project can be said to have an opinion – or even to have read all the code or the papers. I have read it all and I agree with David, so probably the “we” could be taken as David and myself.

  2. drj Says:

    It is our aim of course, to make this clear, and at the moment, it’s far from clear.

  3. Mesa Says:


    So, there is a big debate about the adjustments. A great idea would be to look at the distribution (statisitics) of all the adjustments, say by decade. Ie what is the mean (or median) adjustment per decade. This would give an excellent idea of what part of the overall trend is due to the adjustments, etc

  4. Richard Brooksby Says:

    How about a “code” tab that gets me to the code? Given this is about code, it’s remarkably hard to get there from here.

  5. Nick.Barnes Says:

    The code for GISTEMP was available from the GISTEMP tab. But here’s a new tab to go straight there.

  6. Michael Young Says:

    Hmm. Interesting, but I’m afraid the About page doesn’t really tell me what this is about or what it is supposed to do.

  7. D. Robinson Says:


    Having read your goals and thinking I understand them leaves me wondering if you understand skeptic’s complaints about GISStemp.

    However good or bad Gisstemp CODE is, it is highly doubtful that there is some rounding, loop or variable type problem the elimination of which makes the whole 20th century warming trend disappear. Besides, the other datasets from CRU and NOAA don’t run the same code so if the code was really the concern GISS has a built in ‘defense’.

    When I think about GISS or any of the others possibly exaggerating 20th century warming, it’s the poor classification of rural vs. urban, inadequate or no UHI corrections, poor records of station moves, step wise changes in station data, poor sampling where temperature stations can be smeared much too far, man’s impact on even rural stations, and the massive fallout of stations since the mid 70′s that is the problem.

    When us skeptics complain about GISS, it is a complaint about their entire process. Improving the code is a noble goal, but your still going to have the gridding problems, rural vs. urban, stepwise changes etc. This is an example of a problem with GISS, A significant positive anomaly in a country with no temp stations since 1990:

    If you are not try to improve the input data, which is awful, well garbage in = ….

  8. Nick.Barnes Says:

    D. Robinson @ 7:

    One thing that probably unites us is a sense that this aspect of climate science is vital: it is important to get it right, and to be seen to get it right. I am not a scientist: I am a software professional, so CCC is what I can do towards that.

    There have been many criticisms of and complaints about GISTEMP. The questions that you raise about source data quality, and algorithm choice, are among them. Those are really questions about the science, and I am not a scientist, so I can’t answer them. However, I can help others to answer them because at the end of the day GISTEMP is source code: questions about the algorithm, or about source data choice, are impossible to answer – and often hard to intelligently ask – without understanding the source code. CCC can reveal the algorithms, and thereby hopefully expose them to constructive criticism.

    Other criticisms, right from the original GISTEMP source code release, have focussed on the quality and clarity of the code. That distracts us from the essential debate we should have about the science and its results. And furthermore, it’s something I can fix.

    Whether these criticisms come from “skeptics” is irrelevant to the CCC project. I have also criticized GISTEMP code myself. I don’t think it is necessarily the best global temperature anomaly dataset (the JMA algorithm, for instance, is considerably simpler). But it is very visible and somewhat influential: improving the public understanding of it might, I hope, help all of us come to better decisions about it.

    Looking briefly at that ChiefIO link, this CCC blog isn’t really the forum for discussing “the Bolivia effect”. There may or may not be a problem with getting Bolivian data into GHCN; I don’t know and it’s not relevant to CCC. The 1200km correlation radius was described, with some supporting data, in Hansen and Lebedeff 1987. Note that this is a correlation radius for anomalies, not for temperatures. GISTEMP deals in anomalies, not temperatures, so the relative temperatures of Peruvian beaches, the Amazon rainforest, or the Bolivian Andes, are irrelevant (something known to anyone who has any interest in GISTEMP). Certainly some mountain regions show a great deal of warming. The same goes for the Arctic.

    There is a basic problem in trying to construct a global anomaly record: a lot of the globe doesn’t have a good temperature record. GISTEMP uses the anomaly correlation to address this. Other datasets take different approaches. CCC does not take a view on the best approach, but we hope to provide tools to help people make such a choice.

  9. D. Robinson Says:

    You know I’ve never quite grasped the implication by GISS supporters that the anomaly will filter out problems with the dataset. If there is an artificial trend in the data caused by UHI/population growth, a latitude or elevation shift in thermometers, or thermometer fallout how exactly will the anomaly filter this out?

    Yes it may geographical problems, but not data problems.

    The net adjustments for all datasets increase the trend. This implies that TOBS adjustments demand a larger positive adjustment to the trend than should be required to correct population growth and UHI which should be cause for any analytical person to raise an eyebrow.

    Nick, run your own correlation test between between station count and the net temperature corrections made to the GHCN temp record since mid-century. This IMO shows that NOAA makes insufficient UHI adjustments, CRU makes none and these trends match GISS. You read the blogs, you hopefully look at the data, do you feel UHI is adequately adjusted for?

  10. Nick.Barnes Says:

    I don’t think that measuring the anomaly will filter any of those things out: I’m just saying that the observed anomaly correlations should allow scientists to infer anomaly histories for areas with more sparse, infrequent, or unreliable instrument records.

    As for the UHI, I don’t know whether the GISTEMP peri-urban adjustment is adequate, and that’s not a judgement for CCC to make, although our work might well help others to examine it. I don’t even really know the peri-urban adjustment algorithms: I, personally, haven’t yet worked on that part of the code. You could ask Paul Ollis.

    I do recollect, from elsewhere, that surface temperature reconstructions performed after omitting all urban stations have more-or-less the same warming signal. And the tropospheric temperature datasets from satellites are similar again.

    But even that much is off-topic here.

Leave a Reply