Best known for its studies of the fundamental constituents of matter, the CERN particle-physics laboratory in Geneva is now also being used to study the climate. Researchers in the CLOUD collaboration have released the first results from their experiment designed to mimic conditions in the Earth's atmosphere. By firing beams of particles from the lab's Proton Synchrotron accelerator into a gas-filled chamber, they have discovered that cosmic rays could have a role to play in climate by enhancing the production of potentially cloud-seeding aerosols. Describing their findings in this week's Nature, the team has also found that our current understanding of the chemistry of these aerosols is inadequate and that manmade pollution could have a larger role in their formation than previously thought.

Aerosols are tiny liquid or solid particles suspended in the atmosphere that can warm or cool the climate directly by absorbing or scattering radiation. They can also act as surfaces on which water vapour condenses, leading to the formation of cloud droplets and so tending to cool the planet. Around half of all cloud droplets are thought to form on aerosols that are injected directly into the atmosphere, such as dust particles, sea spray or pollution from the burning of biomass. The other 50% form on aerosols that are produced by the clustering of molecules of trace gases found in the atmosphere. However, it is not well understood exactly how this clustering takes place and precisely which kinds of molecules are involved.

There has also been much debate about the possible role of cosmic rays in the formation of these aerosols. Henrik Svensmark of the National Space Institute in Copenhagen and colleagues hypothesize that the ions that are formed as (charged) cosmic rays pass through the atmosphere act as a kind of glue that makes it easier for molecules to stick together and form aerosols. This hypothesis has proved controversial because it suggests a role for solar variation, as well as human emissions of greenhouse gases, in climate change – the idea being that the stronger the Sun's magnetic field, the more cosmic rays are deflected away from the Earth, resulting in the formation of fewer clouds and so a warmer Earth, with a weaker solar magnetism having the opposite effect.

Cloud in a canister

The CLOUD collaboration, an international group led by CERN's Jasper Kirkby, was set up to settle the question of whether or not there is a link between cosmic rays and climate. The experiment, which has been running since the end of 2009, consists of a 3 m-diameter stainless steel chamber containing humidified ultra-pure air and selected trace gases, which is placed in the path of a charged-pion beam that simulates ionizing cosmic rays. By varying the concentrations of the trace gases, adjusting the temperature and humidity inside the chamber, turning the beam on and off, and then measuring the concentration of aerosols inside small samples removed from the chamber, Kirkby and colleagues can establish how changing atmospheric conditions affect the rate of aerosol production. The fact that they can do this very precisely and with extremely low levels of contaminants means that they can make much cleaner and more controlled measurements than is possible in the real atmosphere.

To their surprise, the researchers found that when simulating the atmosphere just a kilometre above the Earth's surface, sulphuric acid, water and ammonia – the components generally believed to initiate aerosol production – were not on their own enough to generate the quantities of aerosols observed in the real atmosphere, falling short by a factor of up to a thousand, even when the pion beam was switched on. They conclude that other molecules must also play a role, and say that an organic compound or compounds are most likely.

As Kirkby explains, if the missing substance is manmade, then human pollution could be having a larger cooling effect than is currently believed (emissions of sulphur dioxide are already known to generate the sulphuric acid that is vital for aerosol production). Otherwise, says Kirkby, if the missing substance comes from a natural source, the finding could imply the existence of a new climate feedback mechanism (possibly, he adds, higher temperatures increasing organic emissions from trees).

However, when simulating the atmosphere higher up, the researchers found a stronger cosmic-ray effect. They discovered that at altitudes of 5 km or more, where temperatures are below –25 °C, sulphuric acid and water can readily form stable aerosols of a few nanometres across and that cosmic rays can increase the rate of aerosol production by a factor of 10 or more.

More detailed experiments required

Svensmark welcomes the new results, claiming that they confirm research carried out by his own group, including a study published earlier this year showing how an electron beam enhanced production of clusters inside a cloud chamber. He acknowledges that the link between cosmic rays and cloud formation will not be proved until aerosols that are large enough to act as condensation surfaces are studied in the lab, but believes that his group has already found strong evidence for the link in the form of significant negative correlations between cloud cover and solar storms (which reduce atmospheric ionization). "Of course, there are many things to explore," he says, "but I think that the cosmic-ray/cloud-seeding hypothesis is converging with reality."

I think that the cosmic-ray/ cloud-seeding hypothesis is converging with reality Henrik Svensmark

Jeffrey Pierce, an atmospheric scientist at Dalhousie University in Canada, however, is more cautious. Modelling carried out by his group shows that a 10–20% variation in atmospheric-ion concentrations, roughly the variation associated with solar storms or across a solar cycle, produces less than a 1% change in the concentration of cloud condensation nuclei, with the diminishing returns resulting from more aerosols having to share a given quantity of molecular raw material and aerosols merging with one another. "This change is very likely too small to explain the effect on clouds reported by Svensmark," he says. "We must continue to explore other potential physical connections between cosmic rays and clouds."

Kirkby shares Pierce's caution. He argues that CLOUD's results "say nothing about cosmic-ray effects on clouds" because the aerosols produced in the experiment are far too small to seed clouds. But he adds that the collaboration will have some "interesting new results" to present later this year regarding the role of organic molecules in aerosol formation. "What is needed now to settle this question are precise, quantitative measurements," he adds.

Jasper Kirky describes the broad aims of the CLOUD experiment.