Did Romm And Mann Give The Game Away?

image

Source is Hank’s fun flippedSmile

And

Does It Matter If Marcott Spliced Or Not?

After a few posts on Marcott, I’ve been content to let others dig deeper into this…. this…. I’ve no words for it, other than it’s a complete debacle for alleged science, and a validation of skeptics in more ways than I can count. 

I didn’t make too much of these things when I saw them, and I’m not sure what to make of them now ……  other than either way you slice it, it’s hysterical. 

Let’s back up for a minute.  The other day, I wrote my dagger post, which attempted to explain how graphics can easily be manipulated, and that Marcott had done just that.  The post was meant to discuss the resolution and frequency issues.  I was also convinced that they spliced something on to the end of the their proxy graph.  But, that was before it was revealed that they’ve engaged in time travel, move proxy dates up and down the time line. 

In a very quick graphing, I posted this, yesterday……

image

All I did was sort the temp values from low to high, which, of course, moves the things up and down the time line.  I kept the dates with the values to illustrate this.

I did this too show that splicing isn’t necessary when time and resolution mean nothing. 

Dirk at WUWT makes this point…….

It looks like Marcott and Shakun tried to re-date proxies until the resulting curve resembles Mann’s reconstruction (and M&S forgot or didn’t care that since 1850 Mann’s is CRUTEM).

MAYBE M&S have even optimized their proxy re-dating algorithmically to minimize the difference to Mann. And so they have ended up with their bizarre proxy-re-dating (and proxy culling before 0 BP == 1950 AD).

In other words, maybe they fiddled with their proxy data until they matched Mann’s hockey stick. 

But, this comment was in a thread which discussed a comment from Steve Mac’s Climate Audit from Jean S……

Hah! There is some additional fun in Marcott’s main plot (Figure 1A). Mann’s hockey stick there is the global EIV-CRU from Mann et al. (2008), which means that there is no actual reconstruction post 1850, since it’s the Reg-EM produced EIV reconstruction! So they have now essentially “grafted the thermometer record onto” Mann’s reconstruction. To his credit, Mann has always been careful to plot the post 1850 part in EIV reconstructions in a different color. He is actually explicitly warning in his data description spreadsheet that the values for 1850-2006 are instrumental data.

So in Marcott et al Fig 1A we have a comparision in the interval 1850-1950 between their reconstruction (uptick) and Crutem3 (LAND only) (annual?) intrumental record (no uptick). But that’s not all, folks! See the associated uncertainties … Mann et al (2008) uncertainties (which seem to match in the plot to those given in the spreadsheet, i.e., 2 sigma, whereas Marcott et al uncertainties are 1 sigma) are naturally calculated only up to 1849 (as there is no actual reconstruction afterwards), but in the Figure 1A they continue all the way to the end. Where did those 1850-2006 uncertainties come from?

Jean even provides a close up of one of the graphs I put in my post…..

image

Now, this is all very entertaining.  Did they splice?  It doesn’t matter because if they didn’t splice the thermometer record on it, as Dirk states, they could simply manipulate their own proxy data to be just like the thermometer record used by Mann. 

image

So, it’s a difference without a distinction.  They either spliced or they manipulated until the graph was exactly like the instrumental record. 

Which brings me to this curious exchange.  The day before Jean saw these interesting details, Anthony had this post up.

Monday Mirthiness – Watch the genesis (and retraction) of a smear

Anthony was having a bit of fun with Joe Romm……

image

The events are described by Anthony…..

I got his back almost immediately from Romm at 6:45AM PST:

Now you are denying the instrumental record, too?

This made me laugh, because neither Romm’s graph, nor Marcott’s, has the instrumental record in it, only Marcott’s reconstructed temperature and Romm’s red line “projected” add on. Plus, as McIntyre points out, Marcott et al did NOT splice on the instrumental record:

Even more curious is that Mike Mann entered the discussion…….

image

image

Now, this sent Anthony in a different direction.  We all know Anthony doesn’t deny the instrumental record.  He puts a lot more stock in it than I do.  So, he conversed with Romm in protest of the characterization that he would deny the instrumental record. 

Mann did post this shortly after the exchange between Romm and Anthony……

image

Remember, it is a distinction without a difference as to whether or not actual data was spliced or the data was manipulated to replicate the instrumental record. 

How is it that Romm and perhaps Mann (we don’t know exactly what Romm relayed to Mann), immediately believed that the instrumental record is part of the the Marcott hockey stick?  Read the middle tweet by Mann.   Rolling on the floor laughing

Clowns. 

About these ads
This entry was posted in Climate. Bookmark the permalink.

23 Responses to Did Romm And Mann Give The Game Away?

  1. gator69 says:

    I wondered about this when I read the exchange on Watt’s site. These guys really are not all that bright.

  2. leftinbrooklyn says:

    So according to Mann, Watts isn’t a DenyingDenialistDenier, he’s just a DenyingDenialist. Glad to see Mann maintaining some scientific objectivity, after all.

  3. philjourdan says:

    Well, mann has the ears and nose for a clown. Now that he is acting the part, who can tell the difference?

  4. DirkH says:

    Sounds like they concocted a scheme to emulate the thermometer record without using thermometer data as visible input (but it was an input during the optimization process; the cost function was the difference between thermometer record and shape of the M&S stick).

    They were trying to be more cunning than last time but came up with something unjustifiable and flimsy.

  5. omnologos says:

    I can easily imagine reviewer Mann stating that the paper had to show the uptick otherwise the authors would be denounced as denialists. Hence the difference with the graphs in the dissertation.

  6. Latitude says:

    ..you would think that someone would have noticed a 7-8 degree increase in temps

    wouldn’t you?

  7. HankH says:

    James, you’re right. It doesn’t matter if the hockey stick resulted from time shifting of the proxy dates with selective truncation or splicing in the instrument record. The first produces a statistical artifact that’s not in the data. The second introduces a signal that isn’t in the data. Both are capable of producing a higher frequency signal in otherwise low frequency data if not controlled for through low pass filtering.

  8. Correct me if I’m wrong.

    If Marcott’s graph takes low resolution data and smooths it so that it is spreads out over 400 years that means any data showing an increase or decrease that lies on either end of the graph is going to stick out like a sore thumb (or a hockey stick blade) as it has nothing to be smoothed with.

    • suyts says:

      Right you are my friend! See here for a bit of a remedial explanation. Here’s a graph with the resolution as near as I could get with the thermometer record attached as it should look like. http://suyts.files.wordpress.com/2013/03/image_thumb181.png?w=629&h=376

      It’s from a post I wrote a few days ago, linked up at the top of the post.

      Welcome Justin.

      • Thanks for the welcome.

        Marcott’s 2011 thesis invalidates the splicing of the data to his graph.

        from page 52 of his thesis.
        “However, our temperature stack only can be used to address changes at the low
        frequency (>100 yrs) because of the resolution of the datasets that comprise the
        global stack and therefore cannot be used to compare with the annual and decadal
        variability of higher-frequency records.”

        So I’m going under the assumption that he did not splice the data and that the hockey blade is an artifact of Marcott’s methodology. I’m just questioning if I understand the effects of his methodology correctly.

        The lack of smoothing at the end of the graph plus the redating of the proxies (as explained by Steve McIntyre) to selectively pull them to the end of the graph seems to be all that is needed to manufacture your very own hockey stick.

      • suyts says:

        Yes, it isn’t necessary to splice data when you can manipulate the data to look exactly like the splice.

      • HankH says:

        I understand that Marcott also truncated one or more late dated proxy datasets to end on a high note.

    • DirkH says:

      Justin, if they had used the same number of proxies over the entire interval, then the averaging effect would have been the same over the entire interval, and an increase in the end would have been a legitimate signal – just like a decrease would have been. For instance if ALL proxies synchronously drop, we could get a severe drop of the sum signal and it would be legit.

      It is by reducing the number of proxies near the end that they change the power spectrum from a highly lowpassed one to a spectrum that contains undampened high frequencies near the end, as less averaging takes place.

      Of course, I count such a hidden change of signal composition as scientific fraud. Marcott himself answered to Steve Mac that he KNOWS that the uptick is “not robust” – meaning he knows full well what he and Shakun have down – ergo fraud, not simple oversight.

  9. tckev says:

    I still think a 1,500 year rolling average on temperatures would be more enlightening!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s