In recent years, the idea that the climate is driven by clouds and cosmic rays has received plenty of attention. Interest in the idea was prompted by a Danish physicist named Henrik Svensmark, who first suggested it in the late 1990s. Using satellite data on cloud coverage, which became available with the establishment of the International Satellite Cloud Climatology Project in 1983, Svensmark found a correlation between lower troposphere cloud cover and the 11-year solar cycle.
He proposed that cosmic rays initiate the formation of aerosols in the lower atmosphere that then form condensation nuclei for cloud droplets, increasing cloud formation from water vapor. Since low-level clouds increase Earth’s albedo (the amount of incoming solar radiation that is reflected back into space), more clouds mean cooler temperatures. Svensmark claimed that this mechanism was responsible for virtually every climatic event in Earth history, from ice ages to the Faint Young Sun paradox to Snowball Earth to our current warming trend. Needless to say, this would overturn decades of climate research.
Cosmic “rays” are actually energetic subatomic particles. The solar wind shields the Earth from many of the cosmic rays coming from elsewhere in the Milky Way, so the number of rays that reach the Earth is modulated by variations in solar activity, such as the well-known 11-year solar cycle.
Early work by Svensmark and a group at CERN (we recently covered their initial results) has indicated that charged particles like cosmic rays can cause molecules of sulfuric acid, water, or other vapors to combine and form aerosols (particles about 1 nanometer in diameter). This provides a potential link between cosmic rays and cloud formation.
Note the word “potential.” There are a number of things that need to happen before those aerosols can affect cloud cover. They must increase in mass about 100,000-fold before they’re the size of the condensation nuclei that facilitate water droplet formation in clouds. The CLOUD experiment at CERN may eventually be able to provide insight into that process; in the meantime, other studies have examined it using atmospheric models that simulate aerosol processes.
These studies have indicated that the number of cloud condensation nuclei is not very sensitive to the nucleation of aerosols by cosmic rays. There are a few reasons for this. First, there are many other sources of aerosols (including particles in sea spray and anthropogenic emissions), so the total change in aerosols due to fluctuations in cosmic-ray-induced nucleation is not as significant as it might otherwise be. Second, the aerosols are competing with each other to condense a limited supply of vapor, meaning that any increase in the total cloud condensation nuclei is limited.
In addition, most aerosols collide and combine with other particles long before reaching the size of a cloud condensation nucleus—a process called “coagulation.” Increasing the number of aerosols increases the frequency of these collisions, again dampening the effect on the number of condensation nuclei.
In model simulations, a 15 percent increase in cosmic rays (which is about the variation in one 11-year solar cycle) leads to an increase in condensation nuclei of less than 0.2 percent. Even assuming that cosmic rays could have a significant effect on cloud condensation nuclei, it remains to be shown that this would, in fact, account for the observed fluctuations in global low cloud cover.
Rays, clouds, and climate
Since the direct connection between cosmic rays and clouds remains tenuous, the case for cosmic ray control of climate has been made primarily on correlations between global low cloud cover and solar activity. Historically, some climate events can be correlated with changes in solar activity, but there has been no long-term trend in cosmic rays to accompany the temperature rise of the last few decades.
The apparent correlation between rays and cloud cover described by Svensmark has been strong enough to maintain interest in the hypothesis; however, other analyses have suggested that the correlation may be weaker than Svensmark suggested. And, as one researcher noted in Nature, “Because the climate displays a multitude of cycles on almost all timescales, detection of a correlation among climate variables usually meets with initial and healthy skepticism.” That’s especially true for correlations between data covering short periods of time.
In the latest analysis, researchers from Purdue University put together the most up-to-date cloud and cosmic ray data (their paper is currently in press in the Journal of Climate), and the update is noteworthy. The correlation that had been observed between low cloud cover and cosmic rays between 1983 and 2004 did not continue in the four years following. (Data from 2009-2010 is not yet available.) When cosmic rays increased as the solar cycle reached a low, cloud cover did not follow. In fact, it decreased instead—at a time of unusually high cosmic rays.
The authors write, “It is concluded that the observational results presented, showing several years of disconnect between GCRs and lower troposphere global cloudiness, add additional concern to the cosmic ray-cloud connection hypothesis. In fact, this has been done in the most dramatic way with the measurement of record high levels of GCRs during the deep, extended quiet period of [solar] cycle 23-24, which is accompanied by record low levels of lower troposphere global cloudiness.”
The authors are quick to point out that there are general concerns about the reliability of satellite measurements of low cloud cover, as it may not be possible to completely separate the effects of higher-level clouds. Nevertheless, the marked divergence of the two datasets is likely to be a hot topic as the research moves forward and new data continues to roll in. One thing we know for certain about Svensmark’s cosmic ray hypothesis, though: it attracts a lot of attention.