<b>7.4.5 Impact of Cosmic Rays on Aerosols and Clouds</b>
High solar activity leads to variations in the strength and three-dimensional structure of the heliosphere,which reduces the flux of galactic cosmic rays (GCR) impinging upon the Earth’s atmosphere by increasing the deflection of low energy GCR. As GCR is the primary source of atmospheric ionization, it has been suggested that GCR may act to amplify relative small variations in solar activity into climatologically significant effects (Ney, 1959), via a hypothesised relationship between ionization and cloudiness (e.g.,Dickinson, 1975; Kirkby, 2007). There have been many studies aiming to test this hypothesis since AR4, which fall in two categories: i) studies that seek to establish a causal relationship between cosmic rays and aerosols/clouds by looking at correlations between the two quantities on timescales of days to decades, andii) studies that test through observations or modelling one of the physical mechanisms that have been put forward. We assess these two categories of studies in the next two sections.
<b>7.4.5.1 Correlations Between Cosmic Rays and Properties of Aerosols and Clouds</b>
Many empirical relationships have been reported between GCR or cosmogenic isotope 1 archives and some aspects of the climate system (e.g., Bond et al., 2001; Dengel et al., 2009; Ram and Stolz, 1999). The forcing from changes in total solar irradiance alone does not seem to account for these observations, implying the existence of an amplifying mechanism such as the hypothesized GCR-cloud link. We focus here on observed relationships between GCR and aerosol and cloud properties. Such relationships have focused on decadal variations in GCR induced by the 11-year solar cycle, shorter variations associated with the quasi-periodic oscillation in solar activity centred on 1.68 years or sudden and large variations known as Forbush decrease events. It should be noted that GCR co-vary with other solar parameters such as solar and UV irradiance, which makes any attribution of cloud changes to GCR problematic (Laken et al., 2011).
Some studies have shown co-variation between GCR and low-level cloud cover using global satellite dataover periods of typically 5–10 years (Marsh and Svensmark, 2000; Pallé Bagó and Butler, 2000). Such correlations have not proved to be robust when extending the time period under consideration (Agee et al.,2012), restricting the analysis to particular cloud types (Kernthaler et al., 1999) or locations (Udelhofen and Cess, 2001; Usoskin and Kovaltsov, 2008). The purported correlations have also been attributed to ENSO variability (Farrar, 2000; Laken et al., 2012) and artefacts of the satellite data cannot be ruled out (Pallé,2005). Statistically significant, but weak, correlations between diffuse fraction and cosmic rays have beenfound at some locations in the UK over the 1951 to 2000 period (Harrison and Stephenson, 2006). Harrison (2008) also found a unique 1.68-year periodicity in surface radiation for two different UK sites between 1978 and 1990, potentially indicative of a cosmic ray effect. Svensmark et al. (2009) found large global reductions in the aerosol Ångström exponent from AERONET, liquid water path from SSM/I, and cloud cover from MODIS and ISCCP after large Forbush decreases, but these results were not corroborated by other studies who found no statistically significant links between GCR and clouds at the global scale (Calogovic et al., 2010; Kristjánsson et al., 2008; Laken and Calogovic, 2011). Although some studies found small but significant positive correlations between GCR and high- and mid-altitude clouds (Laken et al.,2010; Rohs et al., 2010), these variations were very weak, and the results were highly sensitive to how the Forbush events were selected and composited (Laken et al., 2009).
<b>7.4.5.2 Physical Mechanisms Linking Cosmic Rays to Cloudiness</b>
The most widely studied mechanism proposed to explain the possible link between GCR and cloudiness isthe “ion-aerosol clear air” mechanism, in which atmospheric ions produced by GCR facilitate aerosolnucleation and growth ultimately impacting CCN concentrations and cloud properties (Carslaw et al., 2002;Usoskin and Kovaltsov, 2008). The variability of atmospheric ionization rates due to GCR changes can beconsidered relatively well quantified (Bazilevskaya et al., 2008), whereas resulting changes in aerosolnucleation rates are very poorly known (Enghoff and Svensmark, 2008; Kazil et al., 2008). The CosmicsLeaving OUtdoor Droplets (CLOUD) experiment at CERN indicates that GCR-induced ionization enhanceswater–sulphuric acid nucleation in the middle and upper troposphere, but is very unlikely to give asignificant contribution to nucleation taking place in the continental boundary layer (Kirkby et al., 2011).Field measurements qualitatively support this view but cannot provide any firm conclusion on the role ofions due to the scarcity and other limitations of free-troposphere measurements (Arnold, 2006; Mirme et al.,2010), and due to difficulties in separating GCR-induced nucleation from other nucleation pathways incontinental boundary layers (Hirsikko et al., 2011). If strong enough, the signal from GCR-inducednucleation should be detectable at the Earth’s surface because a big fraction of CCN in the global boundarylayer is expected to originate from nucleation taking place in the free troposphere (Merikanto et al., 2009).Based on surface aerosol measurements at one site, Kulmala et al. (2010) found no connection between GCRand new particle formation or any other aerosol property over a solar cycle (1996–2008). Our understandingof the “ion-aerosol clear air” mechanism as a whole relies on a few model investigations that simulate GCRchanges over a solar cycle (Kazil et al., 2012; Pierce and Adams, 2009a; Snow-Kropla et al., 2011) or duringstrong Forbush decreases (Bondo et al., 2010; Snow-Kropla et al., 2011). Although all model studies found adetectable connection between GCR variations and either CCN changes or column aerosol properties, theresponse appears to be too weak to cause a significant radiative effect because GCR are unable to effectivelyraise CCN and droplet concentrations (Kazil et al., 2012).
A second pathway linking GCR to cloudiness has been proposed through the global electric circuit (GEC). Asmall direct current is able to flow vertically between the ionosphere (maintained at approximately 250 kVby thunderstorms and electrified clouds) and the Earth’s surface over fair-weather regions because of GCRinduced atmospheric ionization. Charges can accumulate at the upper and lower cloud boundaries as a resultof the effective scavenging of ions by cloud droplets (Tinsley, 2000). This creates conductivity gradients atthe cloud edges (Nicoll and Harrison, 2010), and may influence droplet-droplet collision (Khain et al., 2004),cloud droplet-particle collisions (Tinsley, 2000), and cloud droplet formation processes (Harrison andAmbaum, 2008). These microphysical effects may potentially influence cloud properties both directly andindirectly. Although Harrison and Ambaum (2010) observed a small reduction in downward LW radiationwhich they associated with variations in surface current density, supporting observations are extremelylimited. Our current understanding of the relationship between cloud properties and the GEC remains verylow, and there is no evidence yet that associated cloud processes could be of climatic significance.
<b>7.4.5.3 Synthesis</b>
Although there is some evidence that ionization from cosmic rays may enhance aerosol nucleation in the free troposphere, there is medium evidence and high agreement that the cosmic ray ionization mechanism is too weak to influence global concentrations of CCN or their change over the last century or during a solar cyclein any climatically significant way. The lack of trend in the cosmic ray intensity over the last 50 years (Ageeet al., 2012; McCracken and Beer, 2007) provides another strong argument against the hypothesis of a major contribution of cosmic rays to ongoing climate change.