JUNE 20TH 2006


The climate usually behaves in an unpredictable and irregular fashion. Weather forecasters do their best to discover the influence of the seasons, the air and sea currents and anything else they can monitor. They have models, but they have to be continually tested and adjusted, and they are still often wrong..
One thing the climate does not do is to change consistently in a linear fashion; either up or down. Sequences of climate observations are almost always rirregular and subject to unexpected changes.
This makes it difficult to argue that the climate can be overwhelmingly controlled by a single, steadily increasing influence; the rise in greenhouse gases. 
Everybody nowadays is familiar with the technique of linear regression, which uses the method of least squares to draw a straight line through any sequence of numbers plotted against time. This used to be a difficult and time-consuming process, but now it  is available  on every "scientific" calculator, merely by pressing a few buttons,  and it is a popular facility on every spreadsheet. Everything can be linearised and converted into a "trend".
The beauty of irregular sequenecs, such as climate observations, is that you can obtain almost any "trend" you wish, purely by choosing the right starting point and the right ending point. The IPCC and climate scientists revel in this manipulative process.
For example the  unreliable global durface temperatuure record is highly irregular, on several time scales.
If you take the annual time scale and the smoothing of the Jones record beloved of the IPCC you find
* A temperature fall from 1855 to 1860
* A rise from `1860 to 1880
*a fall from 1880 to 1910
*A rise from 1910 to 1942
* A fall from 1942 to 1950
* A slow rise from 1950 to 1978
* a rise from 1978 to 1998
* A fall from 1998 to 2005
They routinely select which one of these suits them for particular arguments. They never go back to 1855 because it looks better if they take 1900 to 1998, namely a "century" They do not care they the early part could not conceivably have been influneced by an increase in greenhouse gases, and they do not try to explain why the temperature should fall from 1942 to 1950.
Their statement about "warming observed over the last fifty years" was carefully chosen to avoid the fall in temperature in the surface record from 1942 to 1950, and to eliminate from consideration the two lower troposphere observations, the radiosondes, which began in 1958, and the satellites, which began in 1978.
Since the rise in  temperature from 1950 to 1978 was not very impressive, they shifted ground, and chose 1978 to 2005 for their more recent claims, which is much more impressive, provided you ride over the 1998n peak, and ignore the fact that it is falling after this. They ignore the low or absent temperature rises of the previous periods. It is as if the greenhouse effect only began in 1978.
The latest NOAA Report, which I have dissected in the following paper on the NZ Climate Science website

plays ducks and drakes with the lower troposphere global temperature records by drawing a linear trend which includes the large El Niño peak in 1998. Because this peak is higher in the lower troposphjere than on the surface, they can use it to claim, that the "trends" are the same in both sets of records. As I point out, if you choose a different climate sequence for each which avoids the 1998 El Niño; 1958 to 1997 for the radiosondes amd 1978 to 1997 for the staellites, the "trend" in the lower troposphere suddenly drops to zero, whereas the "trend" on the surface remains high, so it cannot be due to greenhouse forcing.

Another use of linear agression was mentioned in my last newsletter, with sea level records. If you study the actual sea level  records on the website

you will find that for most parts of the world recent sea  level records show little change. This is particularly true of the Pacific, and includes Tuivalu and New Zealand. But the global warmers will have none of it. They conceal the actual records and make use of often fragmentary past records to claim that the sea level is rising. As with the lower troposphere temperatures they can also make use of the 1998 El Niño which gave a convenient low peak.  But the  usual "trend" chosen  is usually over the last 100 years.despite the unlikely influence of greenhouse gases at the beginning.and the fact that few of the records are even slightly linear.

 They try another trick. Your pocket calculator provides a measure of the accuracy of a linear trend with the "standard deviation". Most people do not know that this is only valid if the data are distributed on a Gaussian or "bell" curve, and even if they are, one standard deviation gives you only two chances in three that the true value falls between these limits. Proper scientists usually quote two standard deviations. which gives a one in 20 chance that the true value is within the limits.

Sea level scientists always use only one standard deviation. the data are so irregular that two standrad deviations would show up their "trends" as ridiculous

Yet another example is ocean heat. They  have no hesittation  in drawing a linear "trend" through a record which is plainly periodic. The beginning of the curve, in 1955, is the bottom of a trough. It goes up to a peak in 1980 and then down again to 1988, Then up, so in 2004 we probably have another peak. But a linear "trend" shows that there is a large rise since 1955, but ignores the fact that the current figure is noit much hgher than that for 1980. The periodic character is almost certainly associated with the El Niño phenomenon, but the spurios "trend"  has to be blamed on greenhouse gases even when there is no temperature rise up aloft..

The trouble is, as soon as the current fall in temperature is discovered, they will soon be switching from greenhouse warming to their favourite topic of the 1970s, the coming ice age.


 Vincent Gray
75 Silverstream Road

Crofton Downs
Wellington 6004
New Zealand
Phone/Fax 064 4 9735939
"It's not the things you don't know that fool you.
It's the things you do know that aint so"
Josh Billings