Skip to content

The recent debate about the resignation of the editor of the journal Remote Sensing because of his allowing the publication of a controversial paper on cloud influences on climate by Spencer and Baswell has been a widespread talking point in the past week or so. So has the swiftly accepted rebuttal by Dessler published in a different journal.

I don’t want to comment on the politics, but on some aspects of the science. It shows that, once again, what some commentators are saying is based on assumptions about what has actually been found out. Time and time again in the climate change debate we see a bold statement exchanged and passed on only to see that when one looks at the scientific basis of that assertion we see it’s more shaky than some realise.

Concerning the effect of clouds on climate change the equation being considered is a very simple one. Changes in the global temperature are due to inputs by clouds and the oceans minus what is lost by radiation to space. In his analysis of the factors in this equation Dessler says he is using data from observations to quantify them rather than making assumptions about what they might be.

Dessler finds that the value for the oceans is 10 w per sq m and the loss is 0.5 w per s m. The term for the cloud contribution comes from his previous paper in 2010. He estimates this as 0.5 w per sq m. Thus he concludes that clouds are unimportant and that the fundamental driver of the climate of the past ten years has been the ocean. We have looked at this paper before and concluded that it rests on unimpressive statistics.

Looking back at our analysis carried out in December 2010 “Clouds From Both Sides.” Here is his diagram. Click on the image to enlarge.


The researchers looked at changes in global cloud cover during the EL Nino’s and La Nina’s of the past decade and, after removing the effect of H2O feedbacks, say they found a small positive feedback, but then added that they cannot exclude the possibility of a small negative feedback.

The slope of the fitted line is the strength of the cloud feedback. Its gradient determines whether the feedback is positive or negative. It’s clear that the result is equivocal and that the small slope calculated depends upon only a handful of outlier datapoints out of a total 120! Technically Dessler found that the answer to the question I poised at the start of this article; Is the reflectance of clouds a function of temperature, to be yes and gives the result 0.54 +/- 0.74 Watts per sq metre.

Let me be a little more forthright than I was in the December article. The analysis of the diagram is nonsense. Sure you can calculate a regression line through the scattered data points, that’s just number crunching, but honestly just look at the diagram and look at the scatter of the data points. Who would conclude that it is justifiable to consider that they represent a straightforward linear relationship? For me, the paper falls down here and should not have been published, or allowed to conclude such definite statements from such ambiguous data.

At Climate Audit Steve Mcintyre noticed the same thing. Climate Audit reanalyses the same data introducing a four-month time lag (with is said to be a better fit to the data) and actually reverses the cloud feedback coefficient at a greater (though still not that impressive) level of statistical analysis.

Analysing A Decade Of No Temperature Change

Another aspect of the recent analysis concerns me. Dessler concludes that the fundamental driver of the climate in the past ten years has been the ocean through its El Nino and La Nina effects.

El Nino and La Nina are short-term effects and are, as far as we know, unaffected by the global warming of the past three decades (the last IPCC report said that global warming seen prior to 1960 is natural and not due to anthropogenic effects.) In essence by saying that the major climatic effects of the past decade have been the occurrence of these natural phenomenon is actually a statement that anthropogenic effects have been unimportant.

The choice of the period 2000 to 2010 for the data is interesting. If clouds have been causing climate change then they have not been doing a very good job in the past decade as we now know that the global average temperature hasn’t changed at all during that period. Looking for climate change this way in the past ten years would not find any!

We are often told that ten years is too short a period to discern climatic effects. Perhaps it is but that is not to say there is not significant information in global temperature data at the decade level. In this case it seems that whilst ten years is too short to see climatic effects it is not to short to deduce what are climate’s main influences. I hope I am not alone in seeing this inconsistency.

Clouds and their effect on climate is one of the most pressing problems in climate science and a great limit to our ability to make predictions.

I don’t think the recent kerfuffle over the cloud papers has advanced science. Too many people have adopted a scientific stance that is too firm and not supported by the data, and used that to political ends. There is a line in the version of the recent Dessler paper I read that says Spencer and Braswell “argued that reality is reversed.” This is a regrettable thing to put in a scientific paper. Not only does it beg the question the paper is trying to address, it is unnecessary and ungentlemanly. How are the numerous uncritical commentators and advocates that crusade such a paper to behave if scientists themselves cannot do better?