Solar and Cosmic Rays
It’s the Sun!
We’ve established that the world has warmed, but what if it isn’t caused by increasing greenhouse gas concentrations. What if there was some other factor causing the warming, like, the Sun, for example. Or perhaps galactic cosmic rays?
Martian heat wave
Perhaps the most glib claim made by skeptics on this topic is “Mars is warming.” i.e. there are no SUVs on Mars, so it must be the Sun. In reality, the scientists who study Mars would be the first to explain why its temperature fluctuates.
The short term temperature changes on Mars are driven by giant dust storms that can grow so large they encircle the entire planet. During these storms, the temperature of the atmosphere can increase over 30 degrees Celsius due to the heat absorbing properties of the dust.
The temperature during other periods depends greatly on how much lighter dust is covering darker rock, which alters the planet’s albedo. The darker the object, the more energy it absorbs. As it’s blown off, the dark rock is exposed and the planet warms. Below shows a composite image (mosaic) of Mars from the pictures taken during the Viking mission in the late ‘70s and again from images taken by the Mars Global Surveyor (MGS) between 1999 and 2001.
The more recent image is darker, and thus Mars is warmer.
Much of this albedo change depends on whatever the dustiness happens to be at the time the pictures were taken.
The graphs show the change in the albedo from the Viking mosaic (higher) to the MGS mosaic (lower). On the right shows the Martian albedo at other times, and we can see that the darkness of the planet changes literally with the wind.
Over longer time spans, Mars has its own version of the Milankovitch cycles, which cause the poles to gain and lose ice leading to further albedo changes (see section 7 for more on Milankovitch cycles). None of the observed changes in Martian temperature in recent decades, or other planets for that matter, have been attributed to the Sun for the same reason that the changes in earth’s temperature have not been attributed to the Sun.
Forcings since 1750
It goes without saying, or at least it should, that climate scientists do not ignore the Sun. This chart shows the estimated contribution of different forcings.
The small bar near the bottom labeled “solar irradiance” represents the estimated warming caused by changes in the Sun since 1750. This contribution is significantly less than what was estimated in the TAR, and current research shows an even smaller contribution.
The black brackets on each forcing indicate the level of uncertainty. Also, you can see that not only have we caused warming, but we’ve caused cooling as well, through our emissions of aerosols and our land use changes. However, the net effect is one of warming.
Sunspots and the Maunder Minimum
Much of what we know about the Sun’s past activity comes from the counting of sunspots. Sunspots are relatively cool areas of the Sun’s surface that form as the result of magnetic fluctuations. The Sun goes through periods of high and low activity on an 11 year cycle. We have spotty records of sunspot activity during the 1600s, but better records have been kept since about 1749.
The most unusual period of sunspot activity on record occurred from the mid 1600s to the early 1700s, when virtually no sunspots were observed. This period of time is called the Maunder Minimum, named after Edward Maunder, the English astronomer who first noticed the lack of sunspots in the historical records. The Maunder Minimum occurred during the “little ice age,” a relatively cold period for much of the Northern Hemisphere (see section 5). For skeptics, this is the primary piece of evidence that changes in the Sun are the driving force behind climate change, even today.
However, the same computer models that tell us that current temperatures are increasing due to anthropogenic causes are likewise able to reproduce what is known of the climate during the little ice age using a combination of solar and volcanic causes. The strength of solar activity affects the strength of high latitude westerly winds. It also affects sea surface temperatures, which in turn change the strength of ocean currents, especially the North Atlantic Oscillation which has strong climatic consequences for Europe and North America. 
No solar increase
For the last three decades, we do not need to rely on proxy evidence to establish that there has been no increase in solar activity. Since then, we’ve been measuring the sun’s output with satellites. This is called Total Solar Irradiance (TSI). There are two major analyses of satellite measurements. PMOD (top) and ACRIM (bottom)
Like the satellite analyses of tropospheric temperature (see section 4), creating a time series of TSI is difficult because no single satellite has been in operation the entire time. Readings from the various instruments must be calibrated and then patched together. That is the cause for the slight differences between the two analyses. When you remove the fluctuations caused by the solar cycle, ACRIM shows more variation than PMOD, but regardless, there is no appreciable trend in either of them. 
The final graph shows the smoothed surface temperature increase in the NASA GISS and Hadley/CRU analyses over the same time (see section 4).
It is in the last three decades that scientists tell us human caused global warming emerged from natural variability, and it is also the last three decades that most conclusively show us that solar activity has not changed.
TSI through time
There have been many attempts to estimate the TSI of the past using the sunspot counts and other measurements as proxies. As mentioned, scientists have been reducing their estimates of the sun’s variability. This graph compares three reconstructions of TSI.
The green line from Hoyt and Schatten is an older reconstruction and this is the version skeptics like the most because it shows a lot of variation over the centuries. More recent versions, such as the blue line from Lean and the red line from Svalgaard show less and virtually no variation in TSI, other than the 11 year cycle.
TSI reconstructions are used in climate models to calculate the effect that changes in solar activity have on climate. Below shows the various radiative forcings calculated by the latest version of the NASA climate model.
Solar irradiance is the wavy orange line. This model uses the Lean TSI reconstruction from 2000, which is the middle of the road reconstruction, and is generally considered to show too much variation based on the latest science. You can see that even if we were to halve the change in solar irradiance, it wouldn’t matter much compared to the combined size of the other forcings.
The search for the solar cycle
One implication of reduced variability is that it puts increasing emphasis on the solar cycle itself. That is, the difference between low and high solar activity is simply the difference between the peaks and valleys of the solar cycle. Under this thinking, the Maunder Minimum would simply be an extended period of time where solar activity equaled that during the low points of the solar cycle.
Translating the peaks and valleys of the solar cycle into observable temperature fluctuations at the Earth’s surface is a difficult task because there is significant noise in the temperature record that obscures the activity of the sun. Perhaps the most credible of these efforts is shown in the graph, with TSI represented by the dashed line, and the corresponding temperature fluctuations in black.
According to this analysis, the solar cycle is responsible for temperature fluctuations of about 0.1 to 0.3 degrees. When we add this to the influence of the El Nino/Southern Oscillation (ENSO), volcanic eruptions, and to lesser natural variations, the combination temporarily builds upon or counteracts warming due to the steady increase of greenhouse gases. (see section 4 for more on the influence of natural factors).
“Temperature matches solar activity exactly!”
You may have seen a graph showing temperature and solar activity matching almost exactly. This work was first published in 1991 by Friis-Christensen and Lassen (FCL), and later by Lassen and Friis-Christensen (LFC). They examined the length of sunspot cycles, which are usually about 11 years, but can vary from shorter to longer periods. This graphic shows the sunspot cycle lengths since the year 1700, with the solid circles indicating the length from minimum to minimum (m-m) and the open circles representing the maximum to maximum (M-M).
FCL used sunspot cycle length as a proxy for solar output, which is a controversial assumption. Because the data is so noisy, their results are heavily filtered to provide a smoother, more comprehensible graph. This is shown in the next graph, comparing temperature of the northern hemisphere to the filtered sunspot cycle length.
This shows a very strong correlation, with the implication that the Sun is driving the temperature increase, not man.
However, there were mathematical errors in the original paper, as shown in the final graph.
The filtering of the most recent sunspot cycle lengths was done incorrectly, and when you fix this problem, the recent increase disappears. This is important because it is the last three decades that we are most interested in.
A 400 year correlation?
LFC published a paper in 1995 extending the sunspot cycle length/temperature correlation back more than 400 years. They used an early temperature reconstruction of the Northern Hemisphere from 1979 for their comparison. This graph shows the LFC results.
A strong correlation between sunspot cycle length and temperature is evident. There is a gap during the 1600s due to the Maunder Minimum. Since there were virtually no sunspots during that time, the sunspot cycle length cannot be calculated.
More graphs, more problems
The version of the graph presented by LFC shifts the sunspot cycle length upward, strengthening the correlation with recent temperatures at the expense of earlier years. This makes it appear that sunspot cycle length explains the most recent rise in temperatures better than any other time in the graph.
In addition, there are problems with the temperature reconstruction. Like the hockey stick, it uses a variety of proxies to derive temperature up to 1880 (see section 5), and since then, uses instrumental readings from thermometers. LFC said they used an 11 year moving average of temperature, but their version (red) was more heavily smoothed than the actual 11 year moving average (dashed). From 1880, rather than using the original temperature analysis from the 1979 paper, they used the analysis of Hadley/CRU (blue). But this analysis is on a different scale and shifted about 1 tenth of a degree downward. This is the difference between the blue and dashed lines.
After you correct these problems and properly scale the solar cycle length to achieve the best fit with temperatures prior to 1850 (when greenhouse gas concentrations began to markedly rise), modern temperatures strongly diverge from the solar cycle length as shown in the final graph.
If there is any correlation, it fell apart in the 20th Century.
Good enough for The Swindle
All of this was still good enough for The Great Global Warming Swindle, which featured LFC’s chart of temperature and sunspot cycle length. Except The Great Global Warming Swindle did them one better. Whoever created the graphics for the documentary took it upon themselves to connect the early data points to the rest of the data, bridging the early 1600s to the 1700s, and they decided that the sunspot cycle length should exactly follow temperature.
This error was removed from subsequent airings.
It came from outer space
If there is no direct solar explanation for the recent temperature increase, what about an indirect cause?
Covering this scenario are cosmic rays. Or specifically, galactic cosmic rays. This is radiation from outer space. At times of high activity, the Sun’s magnetic field deflects more cosmic rays from reaching the earth. The theory is that cosmic rays enhance the formation of clouds by creating ions — in this case, electrically charged air molecules. The ions clump together forming aerosols, or suspended particles. Aerosols serve as cloud nuclei, and thus, clouds are formed. According to the theory, less cosmic rays means less cloud cover and warmer tempeatures because less energy is reflected. This theory, called “cosmoclimatology,” has been championed most notably by Henrik Svensmark, director of the Centre for Sun-Climate Research at the Danish Space Research Institute.
The effect of clouds on surface temperatures depends on their type and altitude. Clouds act to both reflect and trap energy. Thin, high altitude clouds are believed to cause net warming by trapping more energy than they reflect. Thick, high altitude clouds, on the other hand, do the opposite. Thin, low altitude clouds cause cooling, but thick, low altitude clouds cause warming. However, the net effect of all cloud cover is believed to be one of cooling.
Cosmic rays and total cloud cover
At first, it was alleged that total cloud cover and cosmic rays were correlated as shown in the first graph.
However, that depended on the combination of two incompatible cloud datasets. When the comparison was made with proper data, shown next, the correlation fell apart.
The proponents of cosmoclimatology now allege that the correlation exists between cosmic rays and low level clouds.
Solar irradiance vs. cosmic rays
There is some evidence that cosmic rays do play a part in the formation of clouds. However, there is also evidence that changes in solar irradiance changes cloud cover. For example, increased solar irradiance leads to warmer sea surface temperatures which result in a reduction in cloud coverage. Because the Sun’s activity drives both the intensity of solar irradiance and the amount of cosmic rays reaching the Earth’s surface, it is difficult to sort out which is causing what. The graph shows how low cloud cover (green) is correlated with cosmic rays (black) and anti-correlated with solar irradiance (red). 
One way to separate these factors is to examine cloud cover by latitude. The magnetic field of the earth helps to deflect cosmic rays, but it is weaker near the poles. More cosmic rays penetrate the atmosphere near the poles than near the equator. Thus, clouds in higher latitudes should react more strongly to an increase in cosmic rays. By comparing the strength of the correlation between low level clouds and cosmic rays at different latitudes, it is possible to separate what is caused by cosmic rays, and what isn’t. One study concluded that no more than 23% of the change in cloud cover explainable by the solar cycle is caused by cosmic rays, with at least 77% caused by other solar causes.
Uncooperative cloud cover
Sometimes lost in this argument is whether the change in cloud cover is actually capable of explaining the observed warming. In order to do so, there must be a corresponding downward trend in low level cloud cover. Such a trend exists in satellite measurements, but there are problems with the satellite cloud dataset. The data that we have is strongly influenced by the angle at which clouds are viewed by satellites. Measurements of cloud cover from observers on the surface were conducted until 1997, and they correspond with satellite readings in the case of high level clouds, but are not consistent in the case of low level clouds.
From the surface, low level cloud cover has been observed to increase significantly, which, if true, would have caused cooling. From 1983 to the present, satellites have observed a decrease in low level cloud coverage, but this trend is not to be believed either.
The measurements of cloud cover, particularly low level cloud cover, are highly influenced by satellite viewing geometry. This distortion is obvious when you compare the satellite field of view to the changing cloud cover.
The red grid-boxes show those areas that most contribute to the change in the global mean cloud coverage. They strongly correspond to the edges of the field of view of the North American, European, and Japanese geostationary satellites, indicating that such changes are an artifact and not real. The map shows total cloud cover, but it is low level clouds that contribute the most to these artifacts.
Over the years, we’ve added more satellites. As more satellites have come online and the existing satellites are repositioned for better coverage, the result is that we are looking at clouds more directly, and thus they appear smaller and measured cloud cover is reduced. When a satellite failed, the measured cloud cover increased.
Satellite observations of cloud cover are highly sensitive to such changes, and the long term effect is a general decrease in the measured cloud cover that is either mostly or entirely spurious.
Uncooperative cosmic rays
Not even the Cosmic Rays themselves want to cooperate. The graph shows the running mean of the cosmic ray count from the Climax, Colorado neutron monitor (top), and the NASA GISS and Hadley/CRU temperature series for reference (below).
They bottomed out in the mid ‘80s, yet warming has continued.
The bottom line on solar
Although the estimated variability of the Sun has been reduced in recent years, indirect effects of changing solar activity have been postulated. The strength of westerly winds is affected, as is ocean circulation. Both of these have consequences for regional climate. In addition, solar irradiance and cosmic rays are believed to play a part in cloud formation which would amplify any change in solar output by changing the earth’s albedo.
All of this might be true, but there has been no significant change in the output of the sun since the mid 20th Century, and no corresponding change in the cosmic ray intensity to explain the rise in temperature. The decreasing trend in low level clouds is either mostly or entirely the result of measurement problems, and is incapable of explaining the warming.
And if you are still wondering about Mars, there are no clouds on Mars for cosmic rays to stimulate.
 (Phillips, 2001) Online here
 (Geissler, 2005) Online here
 (IPCC, 2007) Online here
 (Rohde) Online here
 (Shindell, Schmidt, Mann, Rind, & Waple, 2001) Online here
 (Fröhlich, 2008) Online here
 (Willson & Mordvinov, 2007)
 (tamino, 2007) Online here.
 (Lockwood & Frohlich, 2007) Online here.
 (Svalgaard, 2008) Online here
 (Sato, 2007) Online here
 (Camp & Tung, 2007) Online here
 (Lassen & Friis-Christensen, 1995) Abstract here
 (Friis-Christensen & Lassen, 1991) Abstract here
 (Damon & Laut, 2004) Online here
 (Lassen & Friis-Christensen, 1995) Abstract here
 (Laut, 2003) Online here
 (Groveman & Landsberg, 1979) Abstract here
 (Laut, 2003) Online here
 (Laut, 2003) Online here
 (Kristjansson, Staple, Kristiansen, & Kaas, 2002) Online here
 (Sloan & Wolfendale, 2008) Online here
 (Norris, 2007) Online here
 (Evan, Heidinger, & Vimont, 2007) Online here
 (Lockwood & Frohlich, 2007) Online here.
Sources cited in Solar and Cosmic Rays
Camp, C. D., & Tung, K. K. (2007). Surface warming by the solar cycle as revealed by the composite mean difference projection. Geophysical Research Letters , 34.
Damon, P. E., & Laut, P. (2004). Pattern of Strange Errors Plagues Solar Activity and Terrestrial Climate Data. EOS , 85, 370, 374.
Durkin, M. (Director). (2007). The Great Global Warming Swindle [Motion Picture].
Evan, A. T., Heidinger, A. K., & Vimont, D. J. (2007). Arguments against a physical long-term trend in global ISCCP cloud amounts. Geophysical Research Letters , 34.
Friis-Christensen, E., & Lassen, K. (1991, November 1). Length of the Solar Cycle: An Indicator of Solar Activity Closely Associated with Climate. Science , 698-700.
Fröhlich, C. (2008). Solar Constant: Construction of a Composite Total Solar Irradiance (TSI) Time Series from 1978 to present. Retrieved June 15, 2008, from Physikalisch-Meteorologisches Observatorium Davos: World Radiation Center: http://www.pmodwrc.ch/pmod.php?topic=tsi/composite/SolarConstant
Geissler, P. E. (2005). Three decades of Martian surface changes. Journal of Geophysical Research , 110.
Groveman, B., & Landsberg, H. (1979, October). Simulated northern hemisphere temperature departures 1579-1880. GeophysicalResearch Letters , 6 (10), pp. 767-769.
IPCC. (2007). Summary for Policymakers. In: Climate Change 2007: The Physical Science Basis, Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. (S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K. Averyt, et al., Eds.) Cambridge, UK and New York, NY, USA: Cambridge University Press.
Kristjansson, J. E., Staple, A., Kristiansen, J., & Kaas, E. (2002). A new look at possible connections between solar activity, clouds and climate. Geophysical Research Letters , 29 (23).
Lassen, K., & Friis-Christensen, E. (1995, July). Variability of the solar cycle length during the past five centuries and the apparent association with terrestrial climate. Journal of Atmospheric and Terrestrial Physics , 835-845.
Laut, P. (2003). Solar activity and terrestrial climate: an analysis of some purported correlations. Journal of Atmospheric and Solar-Terrestrial Physics , 65, pp. 801-812.
Lockwood, M., & Frohlich, C. (2007). Recent oppositely directed trends in solar climate forcings and teh global mean surface air temperatures. Proceedings of the Royal Society .
Norris, J. (2007). Observed interdecadal changes in cloudiness: real or spurious? In S. Broennimann, J. Luterbacher, T. Ewen, H. Diaz, R. Stolarski, & U. Neu (Eds.), Climate Variability and Extremes During the Past 100 Years (pp. 169-178). Springer.
Phillips, T. (2001, October 11). The Perfect Storm Strikes Mars. Retrieved June 15, 2008, from Science@NASA: http://science.nasa.gov/headlines/y2001/ast11oct_2.htm
Richardson, M. I., & Vasavada, A. (2007). Some Coolness on Martian Global Warming and Reflections on the Role of Surface Dust. American Geophysical Union, Fall Meeting 2007, abstract #P31D-05. San Francisco: American Geophysical Union.
Rive, N., & Friis-Christensen, E. (2007, April 27). NR+EFC Statement. Retrieved June 15, 2008, from Universitetet i Oslos: http://folk.uio.no/nathan/web/statement.html
Rohde, R. A. (n.d.). Image:Sunspot Numbers.png. Retrieved June 19, 2008, from Global Warming Art: http://www.globalwarmingart.com/wiki/Image:Sunspot_Numbers_png
Sato, M. (2007, October 9). Forcings in GISS Climate Model. Retrieved June 19, 2008, from Goddard Institute for Space Studies: http://data.giss.nasa.gov/modelforce/
Shindell, D. T., Schmidt, G. A., Mann, M. E., Rind, D., & Waple, A. (2001). Solar Forcings of Regional Climate Change During the Maunder Minimum. Science , 294, 2149-2152.
Sloan, T., & Wolfendale, A. (2008). Testing the proposed causal link between cosmic rays and cloud cover. Environmental Research Letters , 3.
Svalgaard, L. (2008). TSI (Reconstructions).xls. Retrieved June 16, 2008, from Leif and Vera Svalgaard’s Home Page: http://www.leif.org/research/TSI%20(Reconstructions).xls
tamino. (2007, July 24). PMOD vs ACRIM. Retrieved June 15, 2008, from Open Mind: http://tamino.wordpress.com/2007/07/24/pmod-vs-acrim/
Willson, R., & Mordvinov, A. (2007, September). ACRIM Composite TSI Time Series (Daily Means). Retrieved June 15, 2008, from ACRIM: Active Cavity Radiometer Irradiance Monitor: http://www.acrim.com/RESULTS/Earth%20Observatory/earth_obs_fig26.pdf