Fortunately, exceptional mathematical fluency is required in only a few disciplines, such as particle physics, astrophysics and information theory. Far more important throughout the rest of science is the ability to form concepts, during which the researcher conjures images and processes by intuition.

http://online.wsj.com/article/SB10001424127887323611604578398943650327184.html?mod=WSJ_LifeStyle_Lifestyle_5

Their methods smoothed away all the high frequency data until one sudden last uptick… Comparing apples to orchards, indeed.

Mark

Is the problem you are addressing that they try to make the final value appear much more unprecedented than it really is, by comparing its unsmoothed value (or equivalently, Mannian endpegged value) to the previous smoothed values, while presenting both as “smoothed”? (A comparison of apples to orchards, so to speak!)

Yes, new level of unprecedentedness can be obtained with MR:

See also the TAR figure above, http://climateaudit.org/2011/01/07/uc-on-mannian-smoothing/#comment-252233

It is quite a trick to use MR and not to mention it in the caption.

it is from this mail, http://www.eastangliaemails.com/emails.php?eid=373&filename=1066166844.txt

So, “smooth is effected beyond a single filter width because I chose to use IIR filter” would be more accurate and compact description.

UC
Posted Mar 4, 2011 at 3:08 PM |
….
In the text it is said “This warmth is, however, dwarfed by late 20th century warmth which is observed to be unprecedented at least as far back as AD 200″. For such conclusion the minimum roughness is clearly a wrong smoothing method, as one can see from the videos above.

But in their series, isn’t the terminal (c 1998?) unsmoothed annual value higher than all the previous unsmoothed annual point estimates, so that their statement is true in that sense (aside from the issue of confidence intervals)?

Is the problem you are addressing that they try to make the final value appear much more unprecedented than it really is, by comparing its unsmoothed value (or equivalently, Mannian endpegged value) to the previous smoothed values, while presenting both as “smoothed”? (A comparison of apples to orchards, so to speak!)

By the way, you may notice that the smooth is effected beyond a single filter width of
the boundary. That’s because of spectral leakage, which is unavoidable (though minimized
by e.g. multiple-taper methods).

Where is this taken from? Also, is he speaking of a time-domain process? If so, he doesn’t know what he is talking about.

“Spectral leakage” is a frequency domain issue and it has nothing to do with filtering. It is merely what happens when a signal is not bin centered in an FFT, i.e., when a signal does not have an integer number of cycles and you perform an FFT. There’s even a Wikipedia article on it.

If he’s using an IIR, then yes, that is why the response is longer than the “filter width.” In fact, it is infinite, hence the use of the word in the phrase infinite impulse response. An FIR would not do this because it has a finite impulse response. You are absolutely correct, UC. Mann doesn’t know what he’s talking about.

Mark

1066166844:

By the way, you may notice that the smooth is effected beyond a single filter width of
the boundary. That’s because of spectral leakage, which is unavoidable (though minimized
by e.g. multiple-taper methods).

This can be seen from the videos clearly. But, Butterworth is IIR, what if one uses FIR? Mark T / some other filtering expert, can you explain to layman what Mann talks about here?

I’m hoping this provides some food for thought/discussion, esp. for purposes of IPCC…
mike

For CA readers it is clear why Minimum Roughness acts this way, see for example RomanM’s comment in http://climateaudit.org/2008/09/02/uc-on-mann-smoothing/#comment-160990 (some figures are missing there, will try to update).

Very interesting, as usual, UC! Unfortunately, the figures in RomanM’s comment on the background 9/2/08 thread must have gotten lost in the post-Climategate transition to the new WordPress system. However, I’m not able to edit into the comments to see if I can recover the link unless I’m the post author.