Truemors
Image Credit: Ross McKitrick, et al.
Original Caption:
Figure 6. Top panel: PC1 of the post-1400 NOAMER tree ring network, calculated by MBH98 using short-segment standardization. Second panel: simple mean of proxies. Third panel: PC1 using standard software without short-segment standardization. Bottom panel: Unreported PC1 calculated by MBH after censoring Graybill-Idso high-altitude series. All normalized to 1902-1980.
Explanation:
Panel 1 is essentially a representation of the 'hockey stick graph' algorithm applied to data sets including a stand of bristlecone pines at Sheep Mountain, California (which were overemphasized in the graph). These [bristlecone pine] trees were studied because of their pattern of cambial dieback. They all turned out to exhibit a 20th century growth spurt that has not been fully explained, but is likely to be at least in part due to CO2 fertilization and is known not to be a temperature signal since it does not match nearby temperature records. The original authors (and others) have stressed that they are not proper climate proxies.
Panel 2 uses a much simpler algorithm to standardize guesses at temperature variations, witout specifically emphasizing the bristlecones.
Panel 3 uses standard software without the short-segment standardization used in creating the 'hockey stick graph.'
Panel 4 Shows how the 'hockey stick graph' would have looked using its algorithm, but excluding the anomalous bristlecones not thought to be an indicator of local temperatures.
The last 3 panels are roughly in agreement. The first panel appears to be an anomaly due to undue weight having been given to anomalous data.
(What is the ‘Hockey Stick’ Debate About?)
http://www.climateaudit.org/index.php?p=166
http://www.climatechangeissues.com/files/PDF/conf05mckitrick.pdf
Comments (0)