Click to embiggen and read the title text. From xkcd.
Every time someone writes a post about random walks, I feel somewhat motivated to write a little program to generate a few (with adjustable parameters, of course). Anyway, Tamino’s recent discussion motivated me enough to write one.
You can see, for example, the difference between a proper random walk:
And a walk that’s been bounded on one side:
The code is written in .NET (C#) and licensed under the GPL. The binaries are built for the .NET Framework 2.0, so you’ll need that.
And, of course, because it’s the only reason why anyone clicked on this link:
Cross-posted on Young Australian Skeptics
One of the most hyped emails from the Climategate hack was this one, sent by Phil Jones:
I’ve just completed Mike’s Nature trick of adding in the real temps to each series for the last 20 years (ie from 1981 onwards) amd from 1961 for Keith’s to hide the decline.
“Proxy records” for temperatures provide a method measuring temperature without having to have a convenient thermometer around. Many natural phenomena occur at different rates depending on climate conditions, and these differences can be observed in, for example, rocks, ice cores, corals and trees. Obviously, Earth has had a climate for a few billion years, rather than the few hundred years for which we’ve had thermometers. Accordingly, if you want to understand long-term climate you’ll need proxies. Tree rings are one common proxy record for spring-summer temperature, and, in general, closely match other proxy records and (when available) the instrumental (thermometer) record.
Until the late 20th century. The “decline” refers to the “divergence problem”, documented in Briffa et. al. (1998). Basically, during recent years, the tree ring proxy record has been diverging away from the instrumental record, as shown in the figure on right.
There are several variables which can influence tree growth. At extremely high or low latitudes, temperatures are typically a major factor. If each of these other variables is held constant, changes in temperature will be echoed in the tree rings. However, if another of those variables starts changing, the tree ring trends will no longer reflect the temperature trends.
It seems that, late in the previous century, another of these variables started changing. Several hypotheses as to which variable(s) are changing have been made, and are discussed, for example, in D’Arrigo et. al. (2008).
So what was “Mike’s Nature trick”? The “trick” was used in Mann, Bradley & Hughes (1998). Basically, it involves a diagram of the Northern Hemisphere temperature record from 1610-1995, the “NH” portion of figure 7.
From 1610-1980, they use the tree ring record. However, from 1981, the proxy record diverges away from the instrumental data, and so they use the instrumental data for that period. If you look closely (click to embiggen), you’ll notice that the last part of the diagram is drawn differently – the proxy record (1610-1980) is a dashed line, the instrumental record (1981-1995) is a dotted line. This is explained clearly in the diagram’s caption:
‘NH’, reconstructed NH temperature series from 1610–1980, updated with instrumental data from 1981–95.
And that’s it.
So “Mike’s Nature trick” consisted of a legitimate way of displaying the most accurate available data, clearly documented in Mike’s Nature paper.
Here are a few other relevant posts discussing this (feel free to let me know about any other good posts):
- Greenfyre: “Mike’s Nature trick … to hide the decline.”
- Island of Doubt: Hacked emails, tree-ring proxies and blogospheric confusion
- Pharyngula: Why climatologists used the tree-ring data ‘trick’
- RealClimate: The CRU hack
- Skeptical Science: What do the hacked CRU emails tell us?
D’Arrigo, Rosanne; Wilson, Rob; Liepert, Beate; Cherubini, Paolo (2008). On the ‘Divergence Problem’ in Northern Forests: A review of the tree-ring evidence and possible causes. Global and Planetary Change (Elsevier) 60: 289-305. doi:10.1016/j.gloplacha.2007.03.004 [fulltext]
March 10, 2010 at 1:23 am Comments (3)