Falsification tests of climate theories
Do clouds disappear when cosmic rays get weaker?
or “Don’t you worry, my dear, we’ve seen no tigers”
“The Sun makes fantastic natural experiments” Henrik Svensmark says, “that allow us to test our ideas about its effects on the Earth’s climate.” Most dramatic are the events called Forbush decreases. Ejections of gas from the Sun, carrying magnetic fields, can suddenly cut the influx of cosmic rays coming to the Earth from exploded stars.
According to the Svensmark hypothesis, cosmic rays seed the formation of low clouds, so there should be a reduction in the Earth’s low cloud cover in the aftermath of a Forbush decrease. During the past few years there have been repeated attempts to declare the hypothesis falsified, when various teams failed to find the expected decrease in the low cloud cover.
One morning in April 2008, I woke up to find that since midnight the BBC had spread all around the world the news that British physicists had more or less destroyed the Svensmark hypothesis. Violating a basic principle of objective reporting, the broadcasts went out before Svensmark himself had a chance to comment.
By lunchtime he and I had done our best to limit the damage – and the deception of the public – in brief radio and TV interviews. A remark from Svensmark went belatedly onto the BBC website, that the critic it quoted had “simply failed to understand how cosmic rays work on clouds”.
Two years later, critics still don’t understand it. But they go on telling the tale that Forbush decreases have no important effect on clouds, and the media go on echoing them. When Svensmark and his colleagues published in August 2009 a report that showed very clear effects, and explained why others had failed to see them, the BBC and almost everyone else ignored it. But not the scientific critics, who returned to the fray in December 2009 and February 2010.
In my dictionary, kibosh means “kill off finally”. Svensmark commented to me, “It’s crazy. They’re making all the same mistakes that we did ten years ago, when we were first looking at the Forbush decreases.”
Scientific journals and smart journalists usually give only passing attention to negative results. Once you’ve confirmed that the Moon is not made of green cheese, that’s enough, thank you. But endless repetitions about Forbush make them the chief Popperian-style effort to falsify the Svensmark hypothesis, which is known to be the strongest challenger to the assumptions of the man-made global warming hypothesis.
2008: Sloan & Wolfendale + Kristjánsson v. Svensmark
The stir in April 2008 followed a report by Terry Sloan (Lancaster) and Arnold Wolfendale (Durham) in Environmental Research Letters. Their paper included a discussion of clouds at different latitudes, but I’ll leave that till later and focus on their main point about the Forbush decreases. They noted previous suggestions of an observable effect on clouds but concluded there was none.
Sloan & Wolfendale considered whether 24 Forbush decreases during 1984–2005 gave rise to changes in low cloud cover as recorded by the International Satellite Cloud Climatology Project (ISCCP). This project pools visible and infra-red observations of clouds by geostationary and polar-orbiting satellites. In most cases (black circles in the figure below) Sloan & Wolfendale compared the globally averaged low cloud cover in the month when the Forbush decrease occurred with the average of the three preceding months. For some large events (open squares) the timescales were shorter – clouds for one week after the event compared with those for 14 days before it.
A reader unaware of the nuances of the subject might look at the Sloan & Wolfendale figure, see that on average there was no change in low cloudiness, and say, “Test failed.” Indeed, some people thought Svensmark was just being stubborn not to admit there and then that his theory was dead. Wasn’t Sir Arnold Wolfendale an ex-Astronomer Royal and the UK’s top expert on cosmic rays? Hadn’t he presided over talks at CERN in Geneva about experimental tests of the Svensmark hypothesis?
Soon afterwards, Jón Egill Kristjánsson (Oslo) reported another test of the Forbush effect. He and his colleagues used four types of cloud data from the Moderate Resolution Imaging Spectroradiometer (MODIS) microwave instrument on NASA’s Terra and Aqua satellites and looked for the impact of 22 Forbush decreases. “No statistically significant correlations,” they said, “were found between any of the four cloud parameters and [galactic cosmic rays].”
The BBC showed this graph, saying it cast “further doubt on the notion that cosmic rays are a major influence on the Earth’s climate”. With the cloud amount apparently increasing after the cosmic rays went to a minimum, a fair-minded onlooker might well say, “There you go, the Svensmark hypothesis has failed the test again.”
But back in Copenhagen, Svensmark had been re-investigating Forbush decreases for himself, together with his young colleagues Torsten Bondo and Jacob Svensmark. He not only found the impacts but knew exactly why neither Sloan & Wolfendale nor Kristjánsson & Co. could do so.
These investigators, he said, were like explorers returning from a jungle and reporting, “We saw no tigers, so there can’t be any there.” Clouds change a lot from day to day for purely meteorological reasons, whatever the cosmic rays are doing. Those variations can hide the cosmic rays’ effects as easily as undergrowth conceals the camouflaged coat of a tiger. Only by knowing how to watch for a tiger will you have much chance of spotting it before it eats you.
2009: Svensmark, Bondo & Svensmark
With the right tracking skills, the Copenhagen team confirmed all their expectations about the Forbush decreases. The first chance to try to put Sloan right came in October 2008, at a scientific conference in Oslo, where Svensmark showed some slides. But getting papers published in scientific journals has never been easy for his team, and another ten months were to pass before the Forbush paper finally appeared in Geophysical Research Letters in August 2009. It told of huge impacts of the Forbush decreases on clouds and on the aerosols that seed them.
Compare the Kristjánsson graph above with four of the Danes’ below, and you’ll see a quite different picture. The first of them shows a temporary shortage of fine aerosols, chemical specks in the air that normally grow until water vapour can condense on them, so seeding the liquid water droplets of low-level clouds. The remaining three graphs display the observable loss of the clouds that would have been seeded if the aerosols had survived to do their job. Three different kinds of satellite observations tell the same story.
In preparing for this successful hunt, the team’s first task had been to calculate the effects of many solar outbursts on the Earth’s space environment. They were then able to identify and put in rank order 26 Forbush events since 1987 that caused the largest reductions in the energetic cosmic rays that affect the air at low altitudes. The results of laboratory experiments by Svensmark’s team already gave them a clear idea of what the observable chain of events in the atmosphere should be, following those Forbush decreases.
Cosmic rays continuously promote the formation of micro-clusters of sulphuric acid and water molecules, but initially these are far too small to be detectable by remote observation. After growing routinely over a number of days the invisible specks floating in the air influence the normal colour of sunlight as seen from the ground, by scattering away its violet light. Conversely, a shortage of fine aerosols after a shortage of cosmic rays should make the Sun appear abnormally bright in at the violet end of the spectrum.
Ground-based stations of the world-wide AERONET programme monitor subtle changes in the colour of sunlight and, bingo, the Sun’s violet light intensified after the strongest Forbush events. Interpreted as a loss of fine aerosols, these fell to a minimum about five days after the lowest counts of cosmic rays.
Next, the impact of the Forbush decreases on clouds should become apparent. To trace them, the three sources of data used by Svensmark, Bondo and Svensmark were independent of one another. First, the Special Sensor Microwave/Imager (SSM/I) instruments of the US Defense Meteorological Satellite Program cover the world every three days, measuring the liquid water content of clouds over the oceans.
From the MODIS data (also used by Kristjánsson) the Copenhagen team took the liquid water cloud fraction (LWCF). Thirdly, the ISCCP data were of the same kind as those used by Sloan & Wolfendale, namely low clouds over the oceans detected by infra-red.
As the graphs above show, all of these observational data sets showed much the same pattern of events after the strongest Forbush decreases since 1998, namely a decrease in liquid water clouds that reached its lowest point six to nine days after the mimimum count of cosmic rays.
This was longer than some other experts expected. For example Sloan & Wolfendale stopped looking for any effect seven days after their largest events (less, really, because they started their count sooner). But knowing the wide range of opinions among aerosol chemists about the rate of growth of cloud condensation nuclei, Svensmark himself was not surprised by the result – just glad to be able to give those experts useful information about how small aerosols behave in the Earth’s atmosphere.
As for the magnitude of the impact on cloud cover, it was huge. A 7 % decrease in cloud water seen by SSM/I translates into 3 billion tonnes of liquid water vanishing from the sky. The water remains there in vapour form, but unlike cloud droplets it does not block sunlight trying to warm the ocean. After the same five Forbush decreases, the extent of liquid-water clouds measured by MODIS fell on average by 4 %, while ISCCP showed 5 % less cloud below 3200 metres over the ocean.
The four observational data series start from different years, and the wish to compare the aerosols and clouds fixed the time-frame for the results shown already, because AERONET was the Johnny-come-lately. But as the team had identified 26 strong events going back to 1987, they were able to plot their impacts throughout the life-span of each set of observations, as compared with the relative strength of each event.
The blue lines are statistical fits to all of the plotted data points. In all four cases the the impact increases as the decrease in cosmic rays gets bigger, so for the Svensmark hypothesis, test passed.
These graphs also show why other trackers failed to spot the tigers. Sloan and Wolfendale used the ISCCP data, for which the uncertainties (black error bars in the graphs) are particularly large. What’s more, most of their data points relate to monthly averages of cloudiness (ISCCP D2), which means that any Forbush effect lasting a week or so is likely to be smothered by three other weeks of quasi-random and unrelated changes in cloudiness.
For MODIS the error bars are smaller, but while Svensmark, Bondo and Svensmark selected only 13 Forbush events in the period 2000–2007, Kristjánsson et al. used about 22. “As a result,” the Danes concluded, “their data were dominated by weak events that would be plotted to the left of our data … in a region where uncertainties due to variations in meteorology are much greater than the [Forbush decrease] signal.”
From solar activity to cosmic ray ionization to aerosols and liquid-water clouds, a causal chain appears to operate on a global scale. Those were the closing words of the report by Svensmark, Bondo & Svensmark in Geophysical Research Letters. Confident that his hypothesis had survived this important falsification test, Henrik Svensmark hoped that any open-minded physicist should be quite impressed by the results.
2009-2010: the fight goes on
Open-mindedness is often in short supply in climate physics and Svensmark trod, not for the first time, on the toes of the supporters of the man-made global warming hypothesis. A loss of 4 or 5 % of low clouds may not sound very much, but strong Forbush decreases briefly boost the sunlight reaching the oceans by about 2 watts per square metre – equivalent to all the global warming during the 20th Century.
Although they are too short-lived to have a lasting effect on the climate, the Forbush decreases dramatize the cosmic climate mechanism that works more patiently during the 11-year solar cycle. When the Sun becomes more active, the decline in low-altitude cosmic radiation is greater than that seen in most Forbush events, and the loss of low cloud cover persists for long enough to warm the world. That explains the alternations of warming and cooling seen in the lower atmosphere and in the oceans during solar cycles. And the overall increase in solar activity during the 20th Century implies a loss of low clouds sufficient to explain most of the “global warming”.
So the critics have been keen to continue the battle and to say that Svensmark, Bondo and Svensmark’s Forbush results must be wrong. First back in the fray, in December 2009, was Arnold Wolfendale, this time in association with geographers at Sussex, Benjamin Laken (lead author) and Dominic Kniveton. For some reason they concentrated their attention on just one of the four data sets used by the Danes, the MODIS low cloud fraction LCF. These were their main complaints.
- They thought the timescale concerning aerosol growth into cloud-condensation nuclei was unbelievable. Amazingly, Laken, Wolfendale and Kniveton relied on the say-so of a particle physicist at CERN, that it would be two days at the most, compared with the six to nine days observed. This is a subject on which Svensmark and his colleague Martin Enghoff have published a lengthy scientific review and they know there is a much wider range of opinion among experts about aerosol growth rates than Wolfendale’s hit-men imagine.
- The five strongest events used by Svensmark and Co. were said to be “incoherent”, with only the second strongest showing the effect claimed, and that for extraneous reasons. Surprisingly neither the journal’s editors nor their referees asked for any statistical warranty for these arm-waving statements.Incoherent? Judge for yourself.
I happen to have to hand four of the five individual events used by the Danish team before conflation. They show cloud water, not low cloud fraction, but they are good examples of what the individual events look like.
All are untidier than the conflated graphs, where random-looking meteorological “noise” tends to cancel out. Is there anything particularly odd about that second strongest event (second from the left above) as Laken, Wolfendale and Kniveton suggested? Well, there was a previous dip in cloud water unrelated to the cosmic ray changes, but so there was in other cases illustrated. In every case shown, the cloud water (blue) was lower than usual 5-10 days after the cosmic ray minimum (red).
Another falsification attempt came In February 2010. “Sudden cosmic ray decreases: No change of global cloud cover” is its title, once again in Geophysical Research Letters. The lead author is Jasa Calogovic of the Hvar Observatory in Croatia, but it is the work of a Swiss-German collaboration of scientists from the University of Bern and EWAG led by Frank Arnold from the Max Planck Institute for Nuclear Physics in Heidelberg. It inspired the “Forbush Puts Kibosh on Theory” headline quoted near the outset.
At the risk of discourtesy to the distinguished authors, I can report that Svensmark laughed when he read the paper from Arnold’s group. Where his own team studied three different satellite data sets and 26 Forbush decreases, Arnold’s took just one data set (ISCCP) and only six events – those ranking 4th, 10th to 13th, and 26th, in Svensmark, Bondo and Svensmark’s assessment of effects on cosmic rays reaching the lower atmosphere. In the Danes’ plot of all their ISCCP results (see above) all but one of the Swiss-German selection have “strengths” between 33 % and 69 %, where any reduction in clouds is similar to the uncertainty. “Of course they couldn’t see anything,” Svensmark said to me.
Bringing the Earth’s magnetism into the story
Both Wolfendale’s and Arnold’s teams repeated a different complaint going back to Sloan and Wolfendale in 2008, about clouds at different latitudes. In promising to return to it later, I was saving the neatest rebuttal till last. Here, for example, is how Laken, Wolfendale and Kniveton expressed their concern.
“An analysis of the latitudinal distribution of the [low cloud fraction] variations reveals that this decrease is predominately located at mid to low latitudes, whereas if the phenomenon were related to variations in the [cosmic ray] flux it should be predominately located at high latitudes.”
What this is all about is the influence of the Earth’s magnetic field on the influx of cosmic rays. As a shield of sorts, it works much better in the tropics than towards the magnetic poles. So with more cosmic rays coming in at higher latitudes, and varying more, shouldn’t there be a bigger effect on clouds there, than at at low latitudes?
The argument was most clearly illustrated by Sloan & Wolfendale, with a diagram that related not to sudden Forbush decreases but to long-term variations in cosmic rays over Solar Cycle 22 (1983-96) and to an earlier report on their effects on clouds by Nigel Marsh and Henrik Svensmark in 2000. I’ve replaced rather technical labels on the axes of the graph with simpler ones of my own.
The line NM is the variation in cosmic rays at different latitudes, which Sloan & Wolfendale say should be followed by any variation in low cloud cover if the Svensmark effect is real. The various symbols show the actual variations in the clouds according to an analysis of ISCCP data, which are obviously missing the target line completely.
Test failed? Not a bit of it. Here is the same graph with a red line added by Svensmark, showing how he computes that the cloud effect should vary with latitude. It fits the Sloan & Wolfendale data surprisingly well, when you remember how “noisy” the ISCCP data are.
Test passed. We have come to the nub of the misconception, where the critics haven’t grasped an elementary point about Svensmark’s physics. For ten years he has said the clouds most affected by cosmic rays are low clouds. So the cosmic rays that matter are charged particles (mainly muons, heavy electrons) that penetrate low into the atmosphere. They’re generated mostly by very energetic protons from the Galaxy on which the Earth’s magnetic field has little influence. Hence the much reduced slope of the red curve, compared with Sloan & Wolfendale’s NM slope.
NM stands for neutron monitors, and there’s the blunder. Neutrons are very handy for showing changes in cosmic ray intensities, whether in a Forbush decrease or during a solar cycle. But as high-school students know, neutrons are uncharged. They don’t ionize the air or affect the clouds. To rely on neutrons to tell you what the clouds should do is as rash as expecting tigers to have smoke coming out of their heads.
The overriding importance of the muons is the reason why Svensmark’s team went to so much trouble to compute the ionization of the lower air for many Forbush decreases. None of the critics cited here used the same reckoning to lead them to the effects on clouds. This is what lay behind Svensmark’s remark to the BBC back in 2008, that Sloan “simply failed to understand how cosmic rays work on clouds”.
That said, it’s hard to avoid the impression that Svensmark’s critics don’t want to understand the physics, or to see the Forbush effects on clouds, because Mother Nature is not being politically correct. When the Sloan & Wolfendale report came out in 2008, the Institute of Physics in London put out a press release saying it refuted a TV programme, The Great Global Warming Swindle (which briefly mentioned the Svensmark hypothesis). In a press release from Lancaster University, Sloan said that seeing that documentary provoked his research, although he claimed he was “open-minded” about what he’d find. As for Wolfendale, he has criticized the editor of Astronomy & Geophysics for daring to publish an article by Svensmark in 2007.
And these two go on nagging at the issue, most recently in the CERN Courier in February 2010. In company with Anatoly Erlykin of Moscow’s Lebedev Physical Institute, they introduce a new argument, about an apparent decline in low clouds in recent years. (That will be the subject of a later “Falsification” note.) The authors also make rather rambling excursions through the possible roles of cosmic rays in lightning and the origin of life on Earth – perhaps in case the reiteration yet again of their shaky argument about geomagnetic latitudes might set their readers yawning.
“Don’t you worry, my dear, we’ve seen no tigers.” With reassurances from big-cat experts like these, would you leave your gun at home when going into the jungle? Or stake the well-being of the Earth’s people, who have to cope with the ever-changing climate, on the belief that the Sun’s variations, amplified by the cosmic rays, don’t matter any more?
A last word from Svensmark, in a comment to a Swiss journalist when the report by Arnold’s group came out in February 2010: “The case for cosmic rays and their impact on clouds looks better than ever.”
T. Sloan and A.W. Wolfendale, Environ. Res. Lett., 3, 024001, 2008
J. E. Kristjánsson et al. Atmos. Chem. Phys., 8, 7373–7387, 2008
H. Svensmark, T. Bondo and J. Svensmark, Geophys. Res. Lett., Vol. 36, L15101, 2009
Laken, B., A. Wolfendale, and D. Kniveton (2009), Geophys. Res. Lett., 36, L23803, 2009
M.A.B. Enghoff and H. Svensmark, Atmos. Chem. Phys., 8, 4911-4923, 2008
J. Calogovic et al. Geophys. Res. Lett., 37, L03802, 2010
N. Marsh and H. Svensmark, Phys. Rev. Lett., 85, 5004, 2000
A. Erlykin, T. Sloan, A. Wolfendale, CERN Courier, 24 February 2010
DTU press release on the Svensmark, Bondo & Svensmark paper 2009 http://www.alphagalileo.org/ViewItem.aspx?ItemId=59890&CultureCode=en