Clouds Haven’t Behaved the Way the IPCC Or the Models Say

Spread the love

Michael Jonas

I have been working for quite a while on cloud data, and have now had a paper published on the behaviour of clouds, which challenges the way clouds are interpreted by the IPCC and in the climate models. The IPCC indicate that clouds (1) provide a positive feedback to (CO2-driven) climate change, and (2) have reacted to aerosols with an increasing cooling effect. The IPCC logic is essentially that man-made CO2 helped by #1 has been so powerful that it has more than overcome #2. My paper argues that clouds behaved largely independently of CO2 and had as much of a warming effect as CO2, maybe much more. As far as I can tell from searching the scientific literature, this analysis has not been presented before.


The patterns of behaviour of clouds, both for cloud area and cloud optical thickness, are studied over the period of available data, 1983 to 2017. There was a decrease in cloud cover over the study period, while global surface temperatures increased. The patterns of clouds and temperature indicate that the cloud cover decrease could not have been caused by the increased surface temperature. The clear implication is that the decrease in global cloud area must have been caused by some other unspecified factor, and was not caused directly or indirectly by CO2. Evaluation of the changes in clouds and CO2 over the study period indicate that this unspecified factor had as much positive impact as the increase in CO2, with respect to the amount of radiation reaching the surface (radiative forcing), and possibly a much larger positive impact. The climate models, which have zero or negative cloud impact on radiative forcing independently from CO2, need to take this into account in order to avoid over-estimating the influence of CO2.

The paper covers a fair amount of ground, and is open source so that everyone can read it, here. ‘Full text article in PDF‘ opens the full paper.

First, I should point out that whereas scientific papers tend to go into ever more intricate detail, this study stays at a much higher level, dealing only with global numbers or with global numbers split by sea and land. In other words, I am looking at the ‘forest’, while many scientific papers tend to look only at the ‘trees’. That’s particularly true of the climate models, which try to build a picture of the ‘forest’ from all the individual ‘trees’ – a futile approach in my view, since no ‘tree’ can be predicted more than a few days into the future.

A brief summary of the paper follows (NB. This is only a summary, if you want to disagree with it then please do so by disagreeing with the paper not with the summary):-

Available cloud data runs from mid 1983 to mid 2017 (34 years), so that is the paper’s study period. Global cloud cover went down over this period, while global surface temperature went up [Figure 1]. But if you look at cloud and temperature patterns over shorter time scales (a few months), warmer temperatures cause more cloud, not less [Figures 2, 3, 4]. The reducing cloud cover could therefore not have been caused by increasing temperature. The cloud behaviour must therefore have been independent of the temperature, ie, independent of man-made CO2.

The paper goes on to evaluate this independent cloud behaviour [Figures 5, 6]. Instrumental in this, is that cloud area decreased, while cloud optical thickness increased. But cloud area decreased very similarly over sea and land, while cloud thickness increased much more over sea than over land [Figures 7, 8]. The most likely explanation is that the increase in thickness over the sea was not caused directly by the unspecified factor mentioned in the abstract, but was a reaction to higher temperatures.

The conclusion is that the independent warming effect of clouds was comparable to the effect of man-made CO2, and possibly much larger. The IPCC and the models contain no provision for this [Figure 9]. If the models do take this into account, then it can help them to avoid running too ‘hot’.

Not stated explicitly in the paper (maybe it should have been), although it is implied, is that if the models do take this cloud behaviour into account, then the amount of warming from man-made CO2 necessarily becomes a lot smaller when the models are tuned against observed temperature. Hence the future effect of man-made CO2 becomes a lot smaller.

Also omitted from the paper is any attempt to identify the unspecified factor causing cloud cover to decrease. The reason is simply that I hadn’t studied it. I could possibly have mentioned things like cosmic rays or solar UV (ultra-violet) as possible causes, linking them to the IPCC reports which mention them, but I chose not to. It would only have been pure speculation.

The paper does state explicitly that the models’ projections are already known to be highly unreliable (citing the reasons given in the IPCC reports), so there is no excuse on that score for not adding in a provision for independent cloud behaviour.

Those who like to see error-bars on all numbers will be disappointed. The reason is that the calculations are all very approximate anyway, so error-bars would be meaningless. The final conclusions, “similar impact” and “possibly a much larger impact” are not numbers anyway.

Note: I have written before on WUWT about cloud feedback:

That study was restricted to ocean areas only. This paper starts with global data in order to make it more relevant for comparison with the IPCC re radiative forcing.

via Watts Up With That?

June 1, 2022