Skip to content

Link Between Low Sun and Low Temperatures?

Dr David Whitehouse

The media have hailed this research paper as demonstrating a link, for England at least, between low solar activity and lower winter temperatures. The authors were seemingly interested in seeing if there was a link between the recent European low winter temperatures and the low solar activity we are experiencing. It was nice to see scientists change their mind as previously one of them had dismissed any link between recent low solar activity and low temperatures.

To be fair though there are subtle differences here. This paper suggests a link between slightly colder winters occurring than the concurrent trend in winter temperatures when solar activity is low. Although it uses data from the coincident period of low solar activity and low temperatures (the Little Ice Age) in the 17th century, it does not seek to explain the coldness of the MM in terms of low solar activity.

The researchers first take the HadCRUT3v northern hemisphere global temperature series which goes back to about 1860. They extend this back to 1660 using data from a series of papers analyzing tree rings, stalagmites and other proxy data. They then look at the Central England Temperature data from 1660 to the present day extracting the average December-January-February temperature. Looking at this graph (their fig 2b) there does not seem to be any long-term trend. Click on diagram to enlarge it.


They use the open solar flux measurement of solar activity which by using historic observations of geomagnetic effects can be reconstructed back to about 1700. Finally, they “detrend” their Central England Winter temperature dataset. Now its time to plot one variable against another, their figure 3, click on diagram to enlarge it.


It’s obvious that the plot of Central England temperatures against solar activity shows nothing (Fig 3a). A plot of Central England Temperature vs the HadCrRUT3v northern hemisphere also shows little. The regional data varies greatly with respect to the hemispheric mean (fig 3b). The authors again “detrend” this data (note that red dots are more recent years) in a further statistical adjustment on the highly scattered data.

The key diagram in the paper is Fig 3c which shows the detrended mean temperatures plotted against the open solar flux (Fs). Looking at the diagram the authors admit the obvious that any solar effect on the severity of middle England winters is subtle. Note that the dark blue dots are from the period of the Maunder Minimum and they span the entire range of temperatures meaning that during the MM there were relatively warm winters and relatively cold winters.

The authors say there is an absence of points from the lower right of the diagram, ie, high solar activity and cooler winters. Although that’s true there are not a lot of points in the lower left (low solar activity – cooler winters) either. There is little here.

If it’s not obvious by eye then statistics are called for, though I can’t help remembering what the great physicist Ernest Rutherford had to say about statistics, if you need statistics to analyse your experiment you should have done a better experiment.

The authors use some fairly simple statistical measures, well known to science undergraduates, the Mann-Whitney and Kolmogorov-Smirnov tests of significance and find that there is a “highly significant” difference between the bottom left and bottom right of the diagram, ie that there is a correlation between low solar activity (as measured by the open solar flux index) and cooler than average winters.

The problem with this is that it depends upon a very few data points. Of the 350 winters displayed on the graph about five are to be found in the low sun-cold winter region. It’s an easy matter to take away a handful of winters, (about one per cent with one in the MM), redo the statistical calculations and conclude, as was obvious by just looking at the diagram, that there is no meaningful correlation between low solar activity and cooler than average winters. In summary, great conclusions are being made on the basis of a few datapoints.

It’s somewhat of a surprise then to read at the conclusion of the paper that the authors not only say that their research says that their results not only “strongly” (my quotemarks) supports the low solar- colder winters conclusion but also the active sun-warmer winter correlation as well. Those few data points have a lot to answer for! That is the essence of the paper.

Why the supposed correlation of cool winters and low solar activity holds is then commented upon drawing upon speculations from other papers. There is nothing new here. There have been lots of papers in recent years discussing the link between solar activity and the various ways it could influence weather systems in the northern hemisphere. The media, having only looked at the press release associated with this paper, go for the solar-induced blocking of warm westerly winds reaching Europe in winter months. Of course this is pure speculation.

The media have not picked up on one other explanation mentioned as a possibility in the paper. “Enhanced cooling through an increase in maritime clouds may have resulted from the cosmic ray flux increase.” In other words, low solar activity provides less of a shield against cosmic rays with strike the earth more abundantly and cause cloudyness which reflects sunlight. It’s an ironic conclusion given what some solar researchers have had to say about this controversial hypothesis.

It is claimed that this research shows that as solar activity remains low at the start of solar cycle 24 there will be more severe winters like the several we have experienced recently despite what the trend in the global average temperature might be. If this is true it’s a very weak effect and I don’t think it provides any explanation for the severe winters we have experienced recently, and it has only weak predictive power to foretell cold winters in the future.