Torturing Climate Statistics  
John L. Daly
1 Dec 2001

It is said that if you torture statistics hard enough, they will tell you anything you want them to.

The growing gap between the (warming) surface record of global temperature and the (neutral) temperature of the lower troposphere (i.e. lower atmosphere from 1 km to 8 km altitude) obtained from satellites, has put claims of recent global warming into considerable doubt, and also throws doubt on the soundness of the models that predict future warming. Most of the surface warming is actually caused by urban heat islands and bad maintenance and geographical spread of ground-based thermometers in developing countries. The satellites prove it, since they do not find the `global warming’ so earnestly sought for by the industry.

First, there was an inconclusive report from the National Academy of Sciences, and now we have a new paper titled "Accounting for the Effects of Volcanoes and ENSO in Comparisons of Modeled and Observed Temperature Trends," in the November edition of the Journal of Geophysical Research - Atmospheres. Authors representing LLL, NCAR, GISS, CRU and the Max Planck Institute all contributed to this paper, which was merely a statistical dissection of the temperature records from the surface and satellites, comparing them with the usual `models’ - (there’s always a `model' in there somewhere).

Quite why a statistical analysis should be authored by 5 major institutions is unusual in itself. A single author could have done it easily on an ordinary PC. This sponsorship overkill suggests a wider political purpose for the paper.  Politics aside, the paper says essentially that the 1979-1999 satellite record would have shown a stronger warming, more consistent with the surface record, - and more consistent with the models - had it not been for the intervention of two volcanic eruptions forcing down temperatures.

`We wuz robbed' is the between-the-lines message.

This claim ignores three important factors. First, the lower troposphere, being transparent to sunlight, cannot be warmed or cooled except in response to changes in temperature at the surface. In other words, if the troposphere was cooled by the volcanoes, it could only have happened through surface cooling first, followed by the troposphere. This being the case, why does the surface record not show the same level of cooling? Second, volcanoes big enough to impact on climate are part of the normal climatic background anyway. They happen on average about once per decade, the last one being Pinatubo in 1992. Third, while volcanoes do indeed push down temperatures temporarily, El Niño events do the reverse and force up temperatures for a time.

Since 1979, we have had two major volcanoes, two major El Niño events, plus two minor El Niñoes. In other words, these natural events over 22 years are largely self-cancelling, leaving a satellite record giving us a reasonable representation of the real climate of the lower atmosphere - i.e. a global climate with little or no warming.

Since volcanoes and El Niñoes are part and parcel of the normal climate background, it is scientific nonsense to conjure up a statistical scheme where these periodic events are wished out of existence in order to prop up a bogus claim that warming really did happen after all. Some governments are already wise to these political manoevres by over-funded greenhouse institutions, which may explain the reluctance of many, including Australia, to ratify the Kyoto Protocol whose whole rationale is based on these dubious claims.