## Frankfurter Allgemeine

The coverage of the hockey-stick story continues in Germany. Frankfurter Allgemeine has a lengthy article. Here’s a link. . You can get an idea of the contents through a Babelfish processing (www.altavista.com), especially if you are used to the different word orders in German.

1. John A
Posted Mar 19, 2005 at 12:40 PM | Permalink

Interestingly (or not) the Frankfurter Allgemeine Zeitung displays the wrong graph for the text. The text refers to MBH99 but the graph is the longer “Hockey Stick II” of Jones and Mann (2003)

Oops!

Let me ask a dumb statistical question: Looking at the Jones and Mann graph, all of the variation is entirely within the error bars. Quite apart from the likelihood of being able to measure temperature using tree rings in the 3rd Century to + or – 0.2C accuracy, does the displayed variation have any statistical significance when it’s entirely within the error bars?

2. Michael Ballantine
Posted Mar 19, 2005 at 3:07 PM | Permalink

John A, In the fields of Engineering where we deal with hard data and easily verifiable facts, the observed variation is so small compared to the error bars that it would be dismissed as artificial artifacts of noise with aliasing of the signal being a huge problem.
For the variations to have any meaning at all, the raw data must have a resolution that is one significant digit finer than what you want to display. If you want to say 1 degree then your data only needs to have a resolution of 1 degrees. If you want to say 1.0 degrees your data must have a resolution of 0.1 degrees or your statement of 1.0 has no meaning. If you want to say 1.06 degrees then your data MUST have a resolution of 0.03 degrees. If you want to compare results from different sites your instruments must all agree to the same ridiculous level across the entire measurement range tha is used.
I challenge anyone reading this to show proof of 10 or more temperature measuring instruments that can meet this critera outside of a well controlled laboratory. 0.5 degrees of agreement in the field would be incredibly accurate. 1 degree is more likely with the absolute accuracy getting worse as we go back in time. Most fluid displacement thermometers are only accurate at 0.0C and 100C (at sea level). Variations in the middle of +/- 2 degrees are common.
So, if all of the instrumental temperature data from the field correlates to 0.5 degrees or worse then anyone claiming accurate variations in the massaged data of less than 0.5 degrees is either lying or deluded.

3. John A
Posted Mar 19, 2005 at 6:43 PM | Permalink

Re: #2

Of course, an error bar of + or – 2C put the entire variation of “global mean temperature” over the last 150 years as statistically insignificant..but there’s plenty of belief that it must be true.

4. Louis Hissink
Posted Mar 19, 2005 at 10:19 PM | Permalink

John,

True – the GMT does seem to be statistically insignificant but it raises a new problem.

We “know” they were growing grapes in England, and Greenland was as we knew it to be etc.

So how did these regions become colder is climate is a global thing, although in terms of latitude, there are zones of different climates.

I have an idea that might explain it, but it is too novel at the moment.

I prefer to adopt Hannes Alfven’s approach, wait a few more years before saying out loudly, and I need some data from my library in Sydney (I am in Perth and moving, slowly, back to Perth) but when Mann et al reckon the MWP was a localised effect, they may have been right, but for the wrong reasons.

It is almost as if the UK and Greenland were in warmer latitudes (no shifting plates) during the MWP and then something caused them to be ing the current latitudes.

I would suggest one play around with a common office “globe” and engage in some “De Bono” thinking exercises.

5. Hans Erren
Posted Mar 21, 2005 at 8:14 AM | Permalink

RTPW?

6. John A.
Posted Mar 27, 2005 at 2:22 PM | Permalink

Louis,

All that time down under is causing too much pressure to the brain…