06 November 2011

Real Climate Walks Back

At Real Climate Stefan Rahmstorf has a non-responsive response to my earlier critique of the RC11 paper on the Russian heat wave. While he does not address any of my substantive critiques, he does walk back his earlier statements about attribution (video interpretation shown above;-).

Here is a quick rejoinder to the new post:

1. Red Herring - Stefan claims that he is accused of:
"hyping up the number of [heat] records." 
With this, he is introducing a red herring as his idealized ball and urn stuff is probability 101. It is not the paper's mathematics that are problematic, but its argument and conclusions.

2. Avoidance - By following the red herring, Stefan ignores the four cherry picks that I discussed in my critique: Linear trend, station, data set, and non-linear trend. But you don't have to believe me -- NOAA scientists have put up a devastating critique of RC11 here -- a web site that I see Stefan failed to acknowledge, link to or offer a response to.

3. Walk Back - Ultimately, Stefan takes a big step back from their original strong statement about attribution and how it impeached earlier research, now writing
"Our statistical approach nevertheless is not in itself an attribution study" 
 This is a big difference from what they had earlier claimed:
"With this conclusion we contradict an earlier paper by Dole et al. (2011), who put the Moscow heat record down to natural variability." 
Dole et al. was actually an attribution study. Logically, if RC11 "is not itself an attribution study," then it can hardly contradict the conclusions of Dole et al. on attribution.

Now Stefan seems to agree:
This method does not say anything about the physical cause of the trend process – e.g., whether the post-1980 Moscow warming is due to solar cycles, an urban heat island or greenhouse gases. Other evidence – beyond our simple time-series analysis – has to be consulted to resolve such questions.
So, Dole et al., while certainly not the last word on this subject, are not in fact contradicted by RC11.

Thus,we are all in agreement!


  1. Over at Real Climate Rahmstorf walks back further:

    "This got much bigger than expected since during the review process the Dole et al. paper came out, and the reviewers asked us to elaborate on the Moscow data - that is how this became a much larger aspect of the paper than originally intended."

    I sure hope that he is not trying to imply that he was unaware of the Dole et al. paper until the review process.

  2. Roger,

    why do you say,

    'I sure hope that he is not trying to imply that he was unaware of the Dole et al. paper until the review process.'

    Because he should of?

  3. There also seems to be some disagreement about what the NOAA analysis is trying to show. Stefan claims,

    'NOAA has in a recent analysis of linear trends confirmed this non-linear nature of the climatic change in the Moscow data: for different time periods, their graph shows intervals of significant warming trends as well as cooling trends.'

    In that comment, he links to one of the plots you put up in your second post on the topic. But I thought such a plot showed the total temperature anomaly for a given time period as a line. That is, the slope of the linear trend times the amount of time in the series.

    There shouldn't be any non-linearities that one can infer from that graph. Just that some periods are noisier than others.

    He quotes seems to make it seems as though that graph shows the actual time series itself, which ,as I understood it, it does not.

    Is there something fuzzy going on here, other than in my head?


  4. -4-maxwell

    I suggest reading what the NOAA folks have written:


    Have a look at Figs 4 and 5 especially (and related discussion) here:


  5. The whole "non-linear" thing is just an abuse of language. This stuff is non-linear in the same way that a drinking glass is non-linear.

    When you hear "non-linear trend," you think of logarithms, polynomials, exponentials, or something along those lines. You don't think, "moving average." Do you?

  6. Matt,

    I'm having the same issue with the use of 'non-linear' is this context. It seems like the time series that Roger points to from the NOAA analysis are linear trends with noise. The noise is large enough that a statistically significant non-zero trend does not come out of the data.

    But noise is the data does not equal a 'non-linear' trend, or sets of 'non-linear' trends. That's really where my confusion on the language used here comes from.

  7. -7-maxwell

    It seems pretty clear that the use of the term "non-linear trend" to describe a simple 30 point smoothing routine is highly misleading.

    Your interpretation of the data seems accurate enough.