More sloppiness (polite term) found in the latest pride of "Nature" magazine



"Nature" magazine (2012, 484, 49–54) recently announced with great fanfare that if you ignored those pesky ice-core records from Antarctica, then the last deglaciation was PRECEDED by a CO2 rise -- which is necessary if you are to claim that CO2 CAUSED the warming. The paper by Shakun et al. is headed “Global warming preceded by increasing carbon dioxide concentrations during the last deglaciation”

I immediately pointed out the belief in magic needed to come to that conclusion but other skeptics have been looking at the claim too.

So a quick layman's summary of three other bit of trickery found in the "Nature" article. Don't rely on my summary as anything more than an introduction, though.

In his first post on the subject, Willis Eschenbach pointed out that the authors hid some very embarassing stuff by using averages. When you look at the individual temperature proxies that the authors used, you see that they show vastly different times at which CO2 levels peaked. So it becomes impossible to say what the sequence was. Some records suggest that warming came first and others say that a CO2 rise came first. So the whole Shakun claim collapses in a heap.

In his second post Eschenbach shows that the authors cut off the more recent end of the data record -- Where CO2 rises but temperatures FALL! That's pure fakery and dishonesty.

And Piers Corbyn points out that Antarctica is where temperature changes are first seen (it has 91% of the earth's glacial mass) and that the Arctic follows the Antarctic but with a lag of several thousand years. Changes that began in the Antarctic take a long while to percolate through to the Arctic. So the emphasis given by the "Nature" authors to the Arctic is misplaced and they are in fact missing the main game.

Finally, there has just appeared a paper from IPCC reviewer Vincent Gray which points to the the multiple violations of standard statistical assumptions in the paper. The results reported in the paper just cannot be accepted as statistically significant, meaning that they could be due to chance alone:
The Second graph plots the extent of the lag of temperature behind CO2 against the length of the lag over 20.000 and 10,000 years and shows that in the Southern Hemisphere the lag is in the opposite direction, namely CO2 lags temperature. The lag of temperature against CO2 happens only in the Northern Hemisphere, and there seems to be a generally smaller lag in the opposite direction in the Southern Hemisphere.

BUT the small print in the caption beneath the graph states that the confidence intervals given are one standard deviation about the mean. It has been conventional in the statistical and the scientific literature to use two standard deviations for confidence limits, which give the 95% limits in which the true figure may lie. The use of limits of only one standard deviation is a device frequently used by the IPCC and its supporters to give a spurious impression of accuracy, as it includes only 68% of the possible range of the true figure

In this case, for the Globe, the figure 460±340 means that there is a 16% chance that the true figure may be less than 120 and a 2.5% chance that the figure may be less than -220.

For the Southern Hemisphere the figure is -620±660, which means there ia a 16% chance that it is greater than +40 and a 2.5% chance in that it may be greater than +700. For the Northern Hemisphere the figure is 720±330, which means that there is a 16% chance in that the figure may be less than+390 and a 2.5% chance that it may be less than -60

These figures apply when and only when there is a large number of all the samples and they all fit the normal curve closely. Any deviation from these requirements means that the chance that the two sets of figures are not significantly different increases

The study claims to give a global cover. The location of the samples is shown in Figure 1

They appear to show a balanced coverage between the Northern and Southern Hemispheres, but otherwise the sample is grossly unrepresentative of the earth’s surface. The only places that are firmly on land are those in Antarctica. Only three or four are from the ocean and the rest appear to be from coastal sites. The apparent difference between the Northern and Southern Hemispheres may merely be a reflection of this gross sampling deficiency.

The many inaccuracies involved in all of the measurements, combined with the poor sampling and the evident attempt to cover them up by quoting misleading confidence limits leads inevitably to the conclusion that this paper has failed to show a genuine global lag between carbon dioxide and temperature over the Pleistocene, in either direction to a significant degree of accuracy

If there's no such thing as a happy Greenie, it also seems that there is no such thing as an honest Warmist

No comments:

Post a Comment

All comments containing Chinese characters will not be published as I do not understand them