The Dagger In The Heart? Maybe….. A Remedial Explanation Of Marcott’s HS Blade ……. Mikey? What’s That About A Dagger?

 

This is a fairly lengthy post, but I think it’s an easy read.  For people advanced in graphing and smoothing, you can skip some of mid parts.  It’s intended for laymen to come to a better understanding. 

 

This?    LOL, more like this. 

Mikey Mann is babbling again!  From Tom Nelson……

Twitter / MichaelEMann

Watching the pro #climatechange #denial smear artists spinning desperately in wake of new article (http://www.nytimes.com/2013/03/08/science/earth/global-temperatures-highest-in-4000-years-study-says.html?_r=1& …). Dagger in heart?

  Uhhmm, no Mikey, It’s done skeptics worlds of good!  You guys were pretty boring lately.  I mean sure, there’s ton’s of material to poke fun, still, but it gets monotonous.  Well, to tell the truth the hockey stick idiocy is monotonous, but, at least this is a newish twist. 

In a post yesterday, Hank did some heavy lifting for us.  I thought I’d try some light lifting.  Dr. Curry referenced Hank’s post over at Climate Etc.  So, I wandered over to see what was being stated.  Rud Istvan was the main writer of the post.  It’s an interesting and accurate take.  Here’s the part I wish to focus on……

The paper contains a comparison in Figure 1B to a version of the TAR hockey stick chart in which the Medieval Warm Period (MWP) disappeared, and ‘Mike’s Nature trick’ was used to ‘hide the [recent proxy] decline’ by pasting in thermometer temperatures. The paper found good agreement of their reconstruction with Mann et al.’s previous reconstructions:  “indistiguishable within uncertainty”.

Istvan illustrates a graph from the SI, which Hank and I discussed in our email exchanges. 

 

He then goes on making the same comparison Hank did with Ljungqvist’s work. 

Now, I’m never really particularly interested in the distant past other than having the well known warm and cool periods being accurately represented relative to what’s happening today.  And, that’s the objection to this paper.  They started with proxies and ended with thermometer readings.  Where, exactly, they got the thermometer readings is a question.  In the SI they mention Mann like a gazillion times and attach a curious label to his work….. “(CRU-EIV)”.  And, no, I haven’t bothered to run down exactly what that means.  Did the CRU adopt Mann’s lunacy?  IDK.  But, here’s the bombshell in what Istvan is writing about, IMHO. 

How has the MWP almost disappeared again, just in time to perhaps go missing in IPCC AR5? Science’ supplemental information says the average resolution of the 73 paleoclimate series is 160 years, and the median is 120. The proxy selection was deliberately weighted toward ‘low frequency’ resolution, since the entire Holocene was being assessed. Figure S18c  (below) shows there is no statistically valid resolution to the combined proxy set for anything less than 300-year periods. …. In other words, for periods less than three hundred years, white noise in is white noise out (no matter whether the Monte Carlo sampling interval is 20 or 120 years) while for periods over 2000 years the output is about 90% ‘valid’ signal.] The paper itself said, “…our temperature stack does not fully resolve variability at periods shorter than 2000 years…”

Is that what they did?  Did they resort to the graphing equivalent of high school parlor tricks?  Let me see if I can explain this for us common lay people.  The period of time at resolution smooths out variances.  I’ll use graphs from the WFT site to illustrate.  We can pick a time period, any time period to show this.  For our purposes, we’ll pick a small time period and then expand it.  Let’s look at HadCrut3’s rendition of earth’s temps for the period 1950-1970.  I’ve circled some of the extremes shown in this monthly resolution graph…….

See all of those noisy jagged lines?  Note where they are at in relation to the scale on the left and the scale on the bottom.  I drew some lines for reference.  Let’s say we want to smooth this out from a monthly resolution to an annual resolution.  That is, we want the annual mean and not the monthly.  We’ll end up with a graph like so……..

Hey!!!  Wait!!! What happened to the scaling?!?!?!?!??!?!? 

While the timeline on the bottom remained the same, The scale on the left changed!!!  Look at the first circled part on each graph.  Notice that on the first graph the temperature anomaly was less than -0.5 C.  But, now, it’s just barely less than -0.3C.  And, the same happens with the extremes at the top!  The very top extreme was 0.2C, and now it’s 0.0C!!!  And this is how it is.  The longer the time period viewed, the less extremes we’ll see.  And this is what is wrong (among other things) with Marcott’s HS graph. 

It doesn’t matter what he spliced on the end of the graph, and as Hank demonstrated yesterday, he did splice something on there.  If it was Mann’s absurdities or HadCru’s temps it matters not one whit.  Recall what Istvan quoted from the paper’s SI itself and his comments…… 

……the average resolution of the 73 paleoclimate series is 160 years, and the median is 120. …… there is no statistically valid resolution to the combined proxy set for anything less than 300-year periods ……  The paper itself said, “…our temperature stack does not fully resolve variability at periods shorter than 2000 years…”

Whatever Marcott spliced on the end of the graph, we do absolutely know it wasn’t at a 2000 year resolution, it wasn’t even at a 300 year resolution.  Nor was it at 120 year resolution.  We know this because we hardly have 150 years of temp data altogether!  So, what does the spliced time period look like?  Let’s look to see when the “blade” of the hockey stick appears.  (the huge jump at the end of the graph).  For this purpose I believe figure S4a is the best to determine this…….

The description of the graph is describe in the SI…..

Fig. S4: Temperature reconstructions separated by method. (a) 5×5 degree weighted temperature envelope (1-σ) of the jack-knifed global temperature anomaly (30% removed light gray fill; 50% removed dark gray fill), RegEM infilled anomaly (light purple line), standard temperature anomaly (blue line) and Mann et al.’s(2) global temperature CRU-EIV composite (darkest gray).

The anomaly used, “Temperature anomaly is from the CE 1961-1990 average.”

Now, I drew a straight green line to help with the timing.  Given that their “zero” is 1950, I’d say the uptick starts at about 1900, and we should note that in this graph, the blue line does not extend past 1950.  That’s more bit of chicanery itself but for another post.  So, what does our temp anomalies say about 1900 to 1950?  Now, remember, we want to put this at the same meaningful resolution as the rest of the graph for an apples to apples comparison.  But, we can’t!!!    We only have 50 years to deal with.  So clearly we can’t do the 2000, 300, or even the 150 year resolution!  So, let’s go get enough years to show the 1900-1950 interval with proper smoothing…. or as proper as we can get.  Now, remember the scaling at the left……

This is temp data from 1875-1975 with 50 year smoothing.  We have an increase of about 0.25 C.  Now, scroll back up to fig S4a.  That’s not what they did.  Their increase is 0.8C.  Where did they get that?  Maybe the just got confused and spliced the entire temp record on there.

Even here we see a raise of about 0.45 C at 50 years, which isn’t close to their 0.8.  Where did they get the 0.8 C raise? 

Here’s what I think they did.  From their SI……

We chose a 20-year time step in part to facilitate comparison with 91 the high-resolution temperature reconstructions of the past millennium.

But, recall, their proxies don’t have that resolution.  Which is what you would expect from proxies dating back thousands of years.  You can’t look at a piece of sediment and say, “oh look!  It averaged 55.7 degrees F in 1050!”.  But, that’s what our mischievous friends did at the end of their graph.  I give you a 0.8C rise in temps from 1900 with a 20 year smoothing. 

Here’s what the same temps and time period looks like at even 100 year resolution…..

Yes, a whooping 0.11 C increase.  And even that is too high resolution for the graph. 

To illustrate what this would look like instead of having that silly blade, I’ll erase the blade, and draw reverence lines for the start of the blade and the end with the proper rise at only 100 year resolution (which is smaller on average than most of their proxies used.) 

Mikey?  What was that about a dagger?