Interpretation Bias

It is amazing how people’s view of the strength of evidence changes when it confirms their biases. Due to interpretation bias, one should approach bold statements with some skepticism. The social sciences tell us that interpretive bias will be found particularly in those sciences with particular societal interests, or a dominant ideology — such as global warming.

In 2001 a team of UK-based scientists published evidence which they said proved unequivocally that global warming is real.
According to a BBC news article, comparing data obtained from two satellites which orbited the Earth 27 years apart, they claimed to have found significantly less radiation escaping into space than previously.

In the article by Harries et al. in Nature (pdf here) they claimed:

Our results provide direct experimental evidence for a significant increase in the Earth’s greenhouse effect that is consistent with concerns over radiative forcing of climate.

This claim was repeated in 2007 in Chapter 2 of the latest IPCC report, the AR4
Working Group I Report “The Physical Science Basis”.

Harries et al. (2001) analysed spectra of the outgoing
longwave radiation as measured by two satellites in 1970 and
1997 over the tropical Pacific Ocean. The reduced brightness
temperature observed in the spectral regions of many of the
greenhouse gases is experimental evidence for an increase
in the Earth’s greenhouse effect. In particular, the spectral
signatures were large for CO2 and CH4.

Was this evidence conclusive as claimed? At least one researcher thought not. E. Raschke at the GKSS Research Center, University of Hamburg, Germany cited Harries et al. 2001 article and said:

Several greenhouse gases, which are in part or entirely produced by human activities, have accumulated in the atmosphere since approximately the middle of the 19th century. They are assumed to have an additional greenhouse effect causing a further increase of atmospheric temperatures near the ground and a decrease in the layers above approximately 15 km altitude. The currently observed near-surface warming over nearly the entire globe is already considered by a large fraction of our society to be result of this additional greenhouse effect. Complete justification of this assumption is, however, not yet possible, because there are still too many unknowns in our knowledge of participating processes and in our modeling capabilities.

When I searched the 2007 IPCC document for Raschke’s name the result was “No instances found.”. The number of citations for Raschke’s paper is zero, compared with 39 for Harries’ 2001 Nature article. Clearly, skepticism does not pay.

Delving further into the claims of Harries et al. 2001, John Daly posted a balanced review, pointing out an error which he claims incidentally resulted in an erratum being published by Harries. Overall he argues that the effect observed by Harries et al. was very small, that CO2 gives only weak
indications in this study, well within the range of instrument error between two very different instruments separated by technologies 27 years apart, as shown on the figure below.

Even if we accept the claims of statistical significance of the two gases identified as showing the greatest effect, namely methane and the CFCs, methane has stopped increasing and CFC is in decline. Also, ozone decreased during the period, but the absorption line deepened. Something doesn’t make sense here?

I always like to read articles by the same author published both before and after the one in question. Here in the Journal of Climate Harries summarized his previous finding in a somewhat more precise form:

A recent comparison between data taken by two different satellite instruments, the Interferometric Monitor of Greenhouse Gases (IMG) that flew in 1997 and the Infrared Interferometer Spectrometer (IRIS) that flew in 1970, showed evidence of a change in the clear-sky greenhouse radiative forcing due to the increase in greenhouse gas concentrations between those years.

The phrase “evidence of a change in the clear-sky greenhouse radiative forcing” is a much weaker claim than the previous “experimental evidence for a significant increase in the Earth’s greenhouse effect”. After all, they did not perform an experiment in the usually understood sense, nor did they directly measure the Earth’s greenhouse effect, which relates to radiation across the whole infrared spectrum, from the surface to top of atmosphere.

Another paper in 2004 reveals more uncertainties, particularly in the earlier instrumentation:

The results suggest that while the sampling pattern of the IRIS instrument is sufficiently well distributed and dense to generate monthly regional mean brightness temperatures that are within 1.5 K of the true all-sky values, the IMG sampling is too sparse and yields results that differ from the true case by up to 6.0 K. Under cloud-free conditions the agreement with the true field for both instruments improves to within a few tenths of a kelvin. Comparisons with the observed IMG–IRIS difference spectra show that these uncertainties due to sampling presently limit the conclusions that can be drawn about climatically significant feedback processes.

In addition, it seems that in this paper comparing three satellites spectra, an increase in methane was found even between observations when methane was not increasing. They also highlight an inaccuracy in the MODTRAN spectroscopic model. This suggests the only really significant result of Harries et al. 2001 at all, the deepened methane line, could have been an artifact.


I first became interested in this paper from reading the IPCC AR4 report where Harries et al. 2001 is cited as prima facie evidence for greenhouse caused warming. If that were then case, would it not be sensible for another group to have followed up the Harries et al. 2001 study just to verify their claims? I suspect the uncertainty is still underestimated, as the result is achieved by the subtraction of two series of large numbers of variable scaling.

A simple progression through the literature reveals challenges to their claims (Raschke), and dilution the original bold claims. While initial enthusiasm is perhaps understandable, which is why bold claims should be viewed with skepticism, the IPCC report six years later sticks with the original claims, and ignores the subsequent challenges and clarifications. This reveals their bias for their preferred ideology — global warming due to increases in greenhouse gases — and lack of concern with quality of evidence issues. A little background reading reveals that the results of Harries et al. 2001 are an example of weak evidence accompanied by strong interpretive bias.


Harries, J.E., H.E. Brindley, P.J. Sagoo, and R.J. Bantges, 2001: Increases in greenhouse forcing inferred from the outgoing longwave radiation spectra of the Earth in 1970 and 1997. Nature, 410, 355–357.

Is the additional greenhouse effect already evident in the current climate?, Journal Fresenius’ Journal of Analytical Chemistry, Volume 371, Number 6 / November, 2001, Pages 791-797

H. E. Brindley and J. E. Harries, Journal of Climate, Article: pp. 3820–3833
Observations of the Infrared Outgoing Spectrum of the Earth from Space: The Effects of Temporal and Spatial Sampling