Finally! The missing link between exploding stars, clouds and climate on Earth

By Henrik Svensmark and Nir shaviv

Our new results published today in nature communications provide the last piece of a long studied puzzle. We finally found the actual physical mechanism linking between atmospheric ionization and the formation of cloud condensation nuclei. Thus, we now understand the complete physical picture linking solar activity and our galactic environment (which govern the flux of cosmic rays ionizing the atmosphere) to climate here on Earth though changes in the cloud characteristics. In short, as small aerosols grow to become cloud condensation nuclei, they grow faster under higher background ionization rates. Consequently, they have a higher chance of surviving the growth without being eaten by larger aerosols. This effect was calculated theoretically and measured in a specially designed experiment conducted at the Danish Space Research Institute at the Danish Technical University, together with our colleagues Martin Andreas Bødker Enghoff and Jacob Svensmark.


It has long been known that solar variations appear to have a large effect on climate. This was already suggested by William Herschel over 200 years ago. Over the past several decades, more empirical evidence have unequivocally demonstrated the existence of such a link, as exemplified in the examples in the box below.

The fact that the ocean sea level changes with solar activity (see Box 1 above) clearly demonstrates that there is a link between solar activity climate, but it can be used to quantify the solar climate link and show that it is very large. In fact, this “calorimetric” measurement of the solar radiative forcing is about 1 to 1.5 W/m2 over the solar cycle, compared with the 0.1-0.2 W/m2 change expected from just changes in the solar irradiance. This means that a mechanism amplifying solar activity should be operating—the sun has a much larger effect on climate than can be naively expected from just changes in the solar output.  

Over the years, a couple of mechanisms were suggested to explain the large solar climate link. However, one particular mechanism has accumulated a significant amount of evidence in its support. The mechanism is that of solar wind modulation of the cosmic rays, which govern the amount of atmospheric ionization, and which in turn affect the formation of cloud condensation nuclei and therefore how much light do the clouds reflect back to space, as we now explain.

Cosmic Rays are high energy particles originating from supernova remnants. These particles diffuse through the Milky Way. When they reach the solar system they can diffuse into the inner parts (where Earth is) but lose some energy along the way as they interact with the solar wind. Here on Earth they are responsible for most of the ionization in the Troposphere (the lower 10-20 km of the atmosphere where most of the “weather” takes place). We now know that this ionization plays a role in the formation of cloud condensation nuclei (CCNs). The latter are small (typically 50nm or larger) aerosols upon which water vapor can condense when saturation (i.e., 100% humidity) is reached in the atmosphere. Since the properties of clouds, such as their lifetime and reflectivity, depends on the number of CCNs, changing the CCNs formation rate will impact Earth’s energy balance.

The full link is therefore as follows: A more active sun implies a lower CR flux reaching Earth and with it, lower ionization. This in turn implies that fewer cloud condensation nuclei are produced such that the clouds that later form live shorter lives and are less white, thereby allowing more solar radiation to pass through and warm our planet.


Figure 5: The link between solar activity and climate: A more active sun reduces the amount of cosmic rays coming from supernovae around us in the galaxy. The cosmic rays are the dominant source of atmospheric ionization. It turns out that these ions play an important role in (a) increasing the nucleation of small condensation nuclei (a few nm) and (b) increasing the growth rate of the condensation nuclei (which is the effect just published). The larger growth rates imply that they are less likely to stick to pre-existing aerosols and thus have a larger chance of reaching the sizes of cloud condensation nuclei (CCNs, typically > 50 nm in diameter). Thus, a more active sun decreases the formation of CCNs, making the clouds less white, reflecting less sunlight and therefore warming Earth.  

Until today we had just empirical results which demonstrate that this link is indeed taking place. The main results are summarized in Box 2 below. In particular, we have seen correlations between solar activity and cloud cover variations, as well as between cosmic ray flux variations arising from changes in our galactic environment and long term climate change using geological data.   

The first suggestion for an actual physical mechanism was that ions increase the nucleation of small (2-3 nm sized) aerosols called condensation nuclei (CNs). The idea is that small clusters of sulfuric acid and water (the main building blocks of small aerosols) are much more stable if they are charged. That is, the charge allows the aerosols to grow from a very small (few molecule) cluster to a small stable CN without breaking apart in the process. This effect was first seen in our lab (Svensmark 2006). The effect was seen again in the CLOUD experiment running at CERN (Kirkby 2011). Later experiments have shown that ions accelerate also other nucleation routes in which the small clusters are stabilized by a third molecule (such as Ammonia). That is, ions play a dominant role in accelerating almost all nucleation routes (as long as the total nucleation rate is lower than the ion formation rate).

Figure 7: The Ion induced nucleation effect measured in the lab. Left: The first demonstration in our SKY experiment showing that increased ionization increases the nucleation of small aerosols (typically 3 nm in size). Right: Corroboration of the results in the CLOUD experiment at CERN.

In the meantime, a number of research groups aimed at testing the idea that cosmic ray ionization could help the formation of cloud condensation nuclei (CCN). This was done by using large global circulation models coupled with aerosol physics. The idea was to see if an added number of small aerosols would grow into more CCNs. All of the numerical models gave the result that the small aerosols were lost before they could become large enough, leading to the conclusion that the effect of cosmic rays on the number of CCN over a solar cycle was insignificant (e.g., Pirece and Adams 2009). This could also be explained analytically (Smith et al. 2016). It was therefore proclaimed that the theory was dead.  

Given the empirical evidence, it was clear to us that a link must be present, even if the ion induced nucleation mechanism itself is insufficient to explain the link. Thus, our response was to address the same question without using models but instead to test it experimentally. Therefore, in 2012 we tested if small nucleated aerosols could grow into CCN in our laboratory and discovered that without ions present, the response to increased nucleation was severely damped, just like the above-mentioned models; however with ions present, all the extra nucleated particles grew to CCN sizes, in contrast to the numerical model results (Svensmark et al. 2013). So, experiments contradicted the models. The logical conclusion was that some unknown ion-mechanism is operating, helping the growth. 

Figure 8: Left: When injecting small aerosols, the relative increase decreases with aerosol size because as aerosols grow they tend to coagulate with larger aerosols. Right: However, when increasing the ionization in the chamber, not only are more aerosols nucleated, the relative increase survives to larger sizes implying that some mechanism is increasing the survivability of the aerosols as they grow. 

Following the experimental results showing that increased ionization does indeed increase the number of large CCNs, the natural question to ask was whether these results were caused by the particular experimental conditions—perhaps this mechanism does not work in the real atmosphere. It is therefore fortunate that our Sun carries out natural experiments with the whole Earth.

On rare occasions, “explosions” on the Sun called coronal mass ejections result in a plasma cloud passing the Earth, with the effect that the cosmic rays flux decreases suddenly and remains low for about a week. Such events, with a significant reduction in the cosmic ray flux, are called Forbush decreases, and are ideal to test the link between cosmic rays and clouds. Finding the strongest Forbush decreases and using three independent cloud satellite datasets and one dataset for aerosols, we clearly found a response to Forbush decreases. These results validated the whole chain from solar activity, to cosmic rays, to aerosols (CCN), and finally to clouds, in Earth’s atmosphere (Svensmark et al 2009Svensmark et al. 2016). 

Figure 9: The average effect of the 5 strongest Forbush decreases in the 1987-2007 period on cloud properties. Plotted in red is the reduction in the cosmic ray flux following “gusts” in the solar wind (from Coronal Mass Ejections). In black we see the reduction in aerosols over the oceans and three different cloud parameters from three different datasets (Svensmark et al 2009). These results provide an in situ demonstration of the effect of cosmic rays on aerosols and cloud properties.

With the accumulating empirical and experimental evidence, it was clear that atmospheric ionization is playing a role in the generation of the aerosols needed for cloud formation, however, the exact mechanism proved to be elusive. For this reason, we decided to setup another laboratory experiment mimicking conditions found in the real atmosphere and study how atmospheric ions may be affecting the production of CCNs. This also led us to look for alternative mechanisms which will increase the survivability of the CNs as they growth. Indeed, after several years of research, one was found.

The discovery

A little more than 2 years ago, we made the realization that charge will play a role in accelerating the growth rate of small aerosols. When more ions are present in the atmosphere, more of them end up sitting on sulfuric acid clusters of a few molecules. Moreover, the charge makes the sulfuric acid clusters stick to the growing aerosols much faster, as we explain in the box below. Since faster growing aerosols have lower chances of coagulating with larger aerosols, more of the growing aerosols can then survive to reach larger sizes. In other words, when the ionization rate is higher, more CCNs can are formed. 


After realizing that this effect should be taking place we did two things. First, we calculated how large it should be and found that for the typical conditions present in the pristine air above oceans, in which the typical sulfuric acid density is a few 106 molecules/cm3, the ions accelerate the growth by typically 1 to 4%. However, because the number of aerosols surviving the growth is exponentially small (typically several e-folds), the relative change in the CCN density is a few times larger still (by the number of e-folds in the exponential damping to be precise). Thus, over the solar cycle (which changes the tropospheric ionization by typically 20%), we expect a several percent variation in the CCN density and with it, the cloud properties, as is observed.  

The second thing we did was go to the lab and design an experiment in which we can see this effect taking place (and also validate our theoretical calculations). This is not trivial because the effect is larger for lower sulfuric acid levels (as a larger percentage of the molecules would be charged). However we cannot measure at very low sulfuric acid levels because the aerosols then grow very slowly such that they stick to the chamber walls before their growth can be reliably measured. This forced us to measure at high sulfuric acid levels for which the effect is smaller. This posed a formidable technological challenge. To overcome this, we designed an experiment which can keep relatively stable conditions over long periods (up to several weeks at a time) during which we could automatically increase or decrease the ionization rate at the chamber. This allowed us to collect a large amount of data and get high quality signals (e.g., see fig. 11 in the box below).

We found that aerosols indeed grow faster when the ionization rate is higher, totally consistent with the theoretical predictions (as can be seen in fig. 12 in the box below). This allows them to survive the growth period without coagulating with larger aerosols.

So, what do the results imply? Until now we had significant amount of empirical evidence which demonstrated that cosmic rays affect climate, but we didn't have the actual underlying physical mechanism pinned down. Now we have. It means that we not only see the existence of a link, we now understand it. Thus, if the solar activity climate link was until now ignored under the pretext that it cannot be real, this will have to change. But perhaps more interestingly, it also explains how long term variations in our galactic environment end up affecting our climate over geological time scales.