Solar Influence on Global and Regional Climates

Mike Lockwood 0 1 0 M. Lockwood RAL Space, Rutherford Appleton Laboratory , Harwell Campus, Chilton, Didcot, Oxfordshire OX11 0QX, UK 1 M. Lockwood (&) Department of Meteorology, University of Reading , P.O. Box 243, Earley Gate RG6 6BB, UK The literature relevant to how solar variability influences climate is vastbut much has been based on inadequate statistics and non-robust procedures. The common pitfalls are outlined in this review. The best estimates of the solar influence on the global mean air surface temperature show relatively small effects, compared with the response to anthropogenic changes (and broadly in line with their respective radiative forcings). However, the situation is more interesting when one looks at regional and season variations around the global means. In particular, recent research indicates that winters in Eurasia may have some dependence on the Sun, with more cold winters occurring when the solar activity is low. Advances in modelling ''top-down'' mechanisms, whereby stratospheric changes influence the underlying troposphere, offer promising explanations of the observed phenomena. In contrast, the suggested modulation of low-altitude clouds by galactic cosmic rays provides an increasingly inadequate explanation of observations. Solar electromagnetic radiation powers Earth's climate system and, consequently, it is often navely thought that changes in this solar output must be responsible for changes in Earth's climate. However, the huge thermal time constant of the outer part of the Sun limits the variability in its surface temperature, and hence its total power output, which is - dominated by visible and infrared emissions from the solar surface (the photosphere) (see review by Lockwood 2004). As a result, changes in solar power output on decadal, centennial and millennial timescales are limited to small changes in effective surface temperature (associated with magnetic fields) and potential, although as yet undetected, solar radius variations (see reviews by Solanki et al. 2005; Lockwood 2010). Larger percentage variations are seen in solar UV emissions (Lean et al. 1997) which arise from the lower solar atmosphere (the chromosphere) (Loukitcheva et al. 2009) and which influence the stratosphere in Earths middle atmosphere between about 10 and 50 km (see review by Gray et al. 2010). Even more variable are solar X-rays and extreme ultraviolet (EUV) emissions that originate in the upper solar atmosphere (the corona) and dominate the behaviour of Earths uppermost atmosphere (the thermosphere, above about 90-km altitude) (Le et al. 2011). In addition to these electromagnetic outputs, the Sun modulates energetic charged particle fluxes incident upon the Earth. Solar energetic particles (SEP) are emitted by solar flares and from the shock fronts that form ahead of super-sonic (and super-Alfvenic) ejections of material from the corona (Schwenn 2006). SEPs are incident upon Earths atmosphere in polar regions where they enhance the destruction of stratospheric ozone (Jackman et al. 2006; 2008). Even more energetic are galactic cosmic rays (GCRs). These particles are not generated by the Sun; rather, they originate at the shock fronts emanating from violent galactic events such as supernovae explosions. However, the expansion of the shielding solar magnetic field into interplanetary space results in the Sun modulating the number of GCRs reaching Earth (see, for example, review by Potgieter 2008). Air ions generated by GCRs enable Earths global electric (thunderstorm) circuit (Rycroft et al. 2008), and it has been proposed that they also modulate the formation of low-altitude clouds (Svensmark and Friis-Christensen 1997). The Sun also emits a continuous stream of low-energy charged particles called the solar wind (e.g., Marsch 2006). A small fraction of the solar wind energy incident on Earth is extracted by the geomagnetic field and deposited in the thermosphere at high latitudes (Cowley 1991; Thayer and Semeter 2004). This deposition changes the behaviour of the thermosphere globally (e.g., Fuller-Rowell et al. 2007), but this is an extremely low-density atmospheric layer, and there are no robust observations, nor any confirmed theory, that suggests these thermospheric variations are transmitted through the middle atmosphere to the troposphere below. Both electromagnetic and charged particle emissions from the Sun are known to vary over the decadal-scale solar magnetic activity cycle, as do GCR fluxes (see review by Lockwood 2004). But any effects on climate are much more significant for any variations over longer timescales. This review discusses and evaluates potential effects on Earths climate of variations in these solar emissions. Top-down mechanisms involve solar UV irradiance (or perhaps energetic particles) modulating stratospheric temperatures and winds which, in turn, may influence the underlying troposphere where Earths climate and weather reside. These contrast with bottom-up effects in which the total solar irradiance (TSI, dominated by the visible and near-IR) variations cause surface temperature changes and upward coupling to the troposphere. 2 Historical Perspective It is interesting to ponder what the author citation h-index of the astronomer Sir Frederick William Herschel (17381822) would be. Certainly, his 1801 paper (Herschel 1801), in which he speculated on a connection between sunspots and regional climate (for which he used the market price of wheat as a quantifiable proxy), has been cited a great many times and continues to be cited regularly today. Herschel appears to have been more aware of the limitations of his apparent correlation (writing this prediction ought not to be relied on by anyone) than many who have subsequently cited his paper. Indeed, Herschel himself notes limitations to his speculation that many subsequent studies of solar influence on climate have failed to adequately consider. Common pitfalls include the following: The potential for selection effects mean that one must ask, Would a reported correlation coefficient be as high if an equivalent data set were substituted? (For example, in Herschels case, was the behaviour of the price of barley similar to that of wheat? If not, why not?) Selection effects can often arise (unintentionally or otherwise) from the use of restricted data intervals and/or the choice of which parameters to compare. Selection effects are the major reason why the specific issues (2)(5) below are particular problems. Sunspots are only indirectly related to the solar outputs that are relevant to climate. (Herschel attempts to build an argument about enhanced emission from sunspots which we now know to be incorrect). The lack of suitable measurements of Earths global climate on long timescales leads to the use of indirect proxies (Herschel used the price of wheat), and there are many other factors, unrelated to climate, which can influence such proxies. The climate data us (...truncated)