Under a Variable Sun


From 93 million miles away, the churning ball of gas that is our Sun seems little more dynamic than a light bulb—a steady, always-on, really powerful light bulb. Sure, in the winter it seems dimmer and less likely to linger in the sky, but somewhere in middle school or junior high, we learn that it’s not the Sun that changes, but the Earth. As the planet circles the Sun, the tilt in Earth’s axis tips one hemisphere toward the Sun and the other away. The seasons may change, but the Sun does not. A closer inspection, however, reveals that the Sun is more variable than our everyday experience would lead us to believe.

Since the 1600s European astronomers have kept track of dark spots that emerge and travel across the surface of the Sun, and then fade over days, weeks, or months. These sunspots are accompanied by smaller bright areas called faculae. Driven by a dynamo beneath the Sun’s surface, these magnetic features rise and fall over an 11-year cycle. Times of increased sunspot activity are called solar maximums, and quieter times are called minimums. In their continued effort to understand the Sun, solar physicists of the 21st century are using satellites to observe the amount of energy reaching the Earth and how it changes.


Page 2


Not surprisingly, the subject is controversial. After all, the energy from the Sun is the fundamental driving force of climate. As people around the world struggle to make difficult decisions in the face of climate change, any evidence that the Sun could have something to do with global warming might encourage us to scale back our efforts to reduce greenhouse gases. Then again, in the face of something we can’t control, controlling what we can might become more important. With so many social, environmental, and economic decisions depending on climate change research, it’s no wonder that interpreting the satellite observations of the Sun’s brightness has become so controversial.


The sun, once assumed to be perfect and unchanging, was first shown to be variable with the discovery of sunspots in 1610. The number of these dark, turbulent regions increases during times of high solar activity, when the sun emits more energy than normal. (Image courtesy the Royal Swedish Academy of Sciences Institute for Solar Physics)


Graph of Sunspot Number, 1750 to 2003


On one side of the controversy is the work of solar physicist Richard Willson of Columbia University in New York. Willson is the principal scientific investigator for a series of NASA satellite sensors called Active Cavity Radiometer Irradiance Monitors (ACRIM) whose observations of the total amount of solar radiation reaching the outskirts of the Earth’s atmosphere, called total solar irradiance, span much of the last two decades: ACRIM1 from 1980-1989, ACRIM2 from 1991-2001, and ACRIM3, from 2000 to the present. Willson says his work with ACRIM and a handful of other sensors shows not only that the total solar irradiance varies over the 11-year solar cycle, but that it has crept upward between the last two solar minimums. It’s this latter claim that has sparked disagreement within the solar research community.

  Continuous records of sunspots go back more than 250 years, providing a history of solar activity. These records reveal the 11-year solar cycle, when the sun’s energy output rises and falls. Sunspots alone, however, don’t reveal exactly how much energy is emitted by the sun. To measure total solar irradiance, scientists must use satellite sensors, which are above the Earth’s atmosphere. (Graph by Robert Simmon, based on data from the National Geophysical Data Center)


Is the Sun Brighter or Not?


In 1997 and again in 2003, Willson published papers in which he described the construction of a data set of satellite observations of total solar irradiance. Though mostly based on the results from the three ACRIM missions, Willson also spliced into the data set the results from two other missions: NASA’s NIMBUS 7-ERB, launched in 1978; and SOHO-VIRGO, a European Space Agency-NASA mission launched in 1995. Data from these two missions were necessary to fill in some time gaps in the ACRIM record, but it is this splicing that makes the results so controversial.

  Page 1Page 3

Graph of Willson's Reconstructed Total Solar Irradiance Data


“The problem is that no one sensor has collected data continuously over this time period, and so to make a long-term dataset, we have to splice together the results from different instruments, each with its own accuracy and reliability issues, only some of which we are able to account for,” says solar physicist Judith Lean from the U.S. Naval Research Laboratory in Washington, D.C.

  Satellite records of total solar irradiance show a far more detailed view of the solar cycle than a count of sunspots alone. This graph represents data collected by multiple instruments launched over the past 25 years. These sensors provide a way to more precisely estimate total solar irradiance, which is about 1,368 Watts per square meter. (Graph courtesy Richard Willson, Columbia University)

Such joining of data sets can be tricky. While all six of the Sun-monitoring instruments launched between 1978 and the present show the same general trend in total solar irradiance—going up at the solar maximum and falling off during solar minimum—none of them has recorded the same absolute value. During the solar maximum of 1980, NIMBUS 7-ERB recorded the total solar irradiance as about 1,374 watts of energy per square meter, while ACRIM1 measured 1,369. At the following solar minimum, NIMBUS7-ERB recorded values around 1,371, while ACRIM1 recorded 1,367. Scientists can be confident that the general trend is for solar output to increase at solar maximum and decrease at solar minimum, but determining the absolute value at any of those times is a lot more troublesome.


Graph of multi-instrument Total Solar Irradiance Measurements, 1978 to 2003


Willson says one of the strengths of his data set is that he used the original results of the satellite sensors as they were reported in scientific journals and are archived by the experiment’s original science teams. He also standardized all the measurements to match up with what he thinks is the most accurate sensor—ACRIM3. When he put all the data together this way, Willson said that the total solar irradiance had increased 0.047 percent during the 9.75 years between the minima of cycles 21 and 22. In other words, when the Sun’s output dipped down during the solar minimum in 1996, it didn’t dip as far down as it had in 1986.

  Unfortunately, total solar irradiance measurements made by different instruments don’t agree with one another. The magnitude of change from one moment to the next is nearly equal, but the absolute measurement of solar irradiance differs by up to 0.7 percent. This doesn’t sound like much, but the change in solar irradiance from solar maximum to solar minimum is only about 0.2 percent. The situation is even more complicated because the datasets from different instruments don’t always overlap, making comparisons difficult. (Graph courtesy Richard Willson, Columbia University)

Graph Comparing Willson's Results with Lean and Frolich's Results


Not everyone agrees with how Willson put the data together. Claus Fröhlich, principal investigator for the VIRGO sensor on the SOHO mission, and Judith Lean made their own solar irradiance data set. One big difference between their approach and Willson’s is that they corrected the original results from NIMBUS 7-ERB for what they claim (in papers published in 1998 and 2002) are known sources of error caused by changes in the instrument over time. As an example, Lean points out that the first results from NIMBUS 7-ERB show some clearly anomalous behavior that couldn’t possibly be real—such as a sharp spike in recorded measurements right at the start of the mission. NIMBUS 7 didn’t have any self-checking capability of its own to help scientists correct for this unusual behavior, but when the VIRGO sensor came along, Fröhlich noticed that it behaved very much like NIMBUS 7 right after launch—also showing a spike in total solar irradiance at the start of the mission. Fröhlich and Lean corrected the NIMBUS-7 results partly based on the similarities between those two sensors and partly on additional comparisons with a third data set: NASA’s ERBS-ERBE, launched in 1984.


Integrating the conflicting satellite measurements into one consistent data set is as much art as it is science—the data sets of two research groups disagree. The data compiled by Richard Willson and Alexander Mordvinov (blue line) show an increase in solar irradiance between the past two solar minima (in 1986 and 1996), while Claus Frölich and Judith Lean’s data (red line) show no difference in solar irradiance over the same time period. (Graph courtesy Judith Lean, Naval Research Laboratory)


Can Digging into the Details Settle this Debate?


Each scientist points out the problems associated with the other’s approach. “Adjusting other people’s published results is a questionable practice,” says Willson. “You can’t make any modifications to the data sets without a thorough understanding of the instrument and its calibration history. The best time-proven approach for getting the most out of a set of data is the peer-reviewed publication process. The original science teams for the experiments have more information about the experiments than any other individual or group can ever acquire.”

He also draws attention to the crucial issue that between mid-1989 and late 1991 (the ‘ACRIM gap’), ERBS recorded generally decreasing values of total solar irradiance while the Nimbus7/ERB results were increasing. “One of the most significant and universally accepted findings from satellite observations to date is that TSI is directly proportional to solar activity levels, ”says Willson. “The ‘ACRIM gap’ was a period of rapidly increasing solar activity levels, and TSI should have been rapidly increasing as well. The Nimbus7/ERB results are consistent with this finding whereas the ERBS results are not.” Using ERBE data to confirm that the correction made to the NIMBUS 7 data is accurate only makes sense if you are sure the ERBE data themselves are accurate—and Willson says we can’t be sure they are.

  Page 2Page 4

Graph Comparing Nimbus 7/ERB and ERBS/ERBE Data from 1989 to 2001


ERBE mission scientist Robert B. Lee, III dismisses that claim. “You have to look at the data set as a whole, not just bits and pieces of it. The ERBS/ERBE Solar Monitor is still making precise TSI measurements after more than 19 years of operations. During this period, the ERBS/ERBE solar monitor experienced no instrument malfunctions which affected the resulting TSI measurements,” he says. According to Lee, the high level of variability in the sensor’s measurements between 1989-1992 reflected real variability in solar irradiance during that time, not instrument problems. Lee offers what seems to be a reasonable approach. “The only way this will be resolved is for all of us to turn over our data to an independent source like the National Institute of Standards and Technology and let them investigate and make their own conclusion,” concludes Lee.

For her part, Lean says that the original results from NIMBUS 7 couldn’t be corrected because the technology and expertise to do so weren’t available at that time. As evidence that their combined data set is the correct way to assemble the satellite data, Lean compares the results of the data set Fröhlich and she put together to a model of solar activity that uses observable features of the Sun, namely the sunspots and faculae that are the primary cause of total solar irradiance variation, as substitutes (proxies) for total solar irradiance. These proxies are necessary because the atmosphere absorbs so much of the Sun’s total output that we can’t measure it accurately from the ground. These models also help us extrapolate what the total solar irradiance might have been like in the past, since sunspot records go back to the 1700s. Lean points out how well their composite compares to the solar models, which give no suggestion of an increase in total solar irradiance between the solar minima in 1986 and 1996.

  Although the trends measured by different satellite total solar irradiance instruments usually agree, there are exceptions. During the critical period between the death of ACRIM 1 and launch of ACRIM 2, measurements from the Nimbus-7 satellite (green line) showed an increase in solar irradiance, while the Earth Radiation Budget Satellite (blue line) showed a decrease in total solar irradiance from July 1989 through December 1991. Sorting out these differences will likely resolve questions about the sun’s energy output. (Graph by Robert Simmon, based on data from the National Geophysical Data Center)

Graph Comparing Modelled Total Solar Irradiance Data with Willson and Lean's Results


The deviation of his data set from the solar models doesn’t trouble Willson, who says that while the solar proxy models are useful for trying to describe solar activity back in time before scientists had actual observations of total solar irradiance , “the models are not competitive in accuracy or precision with even the worst total solar irradiance satellite observations.”

Joe Gurman, US Project Scientist for SOHO, isn’t so quick to dismiss the models. “These models have done a very good job in simulating the solar activity over recent solar cycles as well as more pronounced, long-term changes in solar output that can be linked to historically documented changes in climate,” he says. “If you accept Willson’s conclusion that total solar irradiance has increased over the last two solar cycles, then not only do you have to explain why the models don’t show it, but you also have to explain why no other single instrument shows a similar increase, and why none of our other solar indicators, like total magnetic flux or ultraviolet light output, has shown a similar increase.” Although there has been a lot of evidence suggesting that magnetic activity is the primary cause of observable solar variability, Gurman says that there is still a lot we don’t know about the Sun. It is always possible there is some physical process that influences solar irradiance that we don’t know about yet.

That possibility doesn’t seem like such a stretch to astrophysicist Willie Soon of the Harvard-Smithsonian Center for Astrophysics. Soon says that, over the years, his study of stars similar to the Sun suggests that there is no physical reason why changes in total solar irradiance couldn’t be real. Other Sun-like stars have been observed to vary over longer cycles. He is excited about the possibility that Willson’s results suggest a longer term solar variation that is yet to be explained. Even this is controversial, however; other physicists think that the Sun is not as much like other stars as we once thought and that there is no evidence to date that anything other than the magnetic activity that we already know about is influencing solar irradiance.

Hugh Hudson, a solar physicist with University of California in Berkeley, agrees the door should be kept open to the possibility of solar variability such as that claimed by Willson, but says he’s not sure that the measurements collected to date are precise enough to support those claims. “Still,” he says, “models and theory will only take you so far, and at some point, you may have to suspend your disbelief if the data require it.” Like several others in the relatively small solar irradiance research community, Hudson seems to think it’s just too soon to tell.


Computer models based on sunspots and faculae on the sun’s surface (black line) suggest that there was no increase in solar irradiance from 1986 to 1996. These models are useful in the study of historical solar energy output, but they may not be as reliable as more direct satellite measurements. (Graph Courtesy Judith Lean, Naval Research Laboratory)


Why is such a small increase such a big deal?


Climate change scientists may be willing to let the solar physicists split hairs, however, at least for now. “Whether the Sun’s output has increased by 0.05 percent per decade over the past two solar cycles makes little difference to climate,” says Dr. Gerald North, first Study Scientist for NASA’s Tropical Rainfall Measuring Mission, and a climate modeler with Texas A&M. “With up to 95 percent confidence we can say that changes in solar output of 0.1 percent over the 11-year solar cycle produces a maximum increase in temperature of two hundredths of a degree (Kelvin), hardly significant in itself.”

  Page 3

Photograph of Sunset

Climate modeler Jim Hansen, of NASA’s Goddard Institute for Space Studies, agrees that the climate impact of .05 percent per decade would be practically non-existent if it were only maintained for one decade. “If such a small change were followed by no further change or a decrease, it’s not important,” he says. “But if that rate of change were maintained for a century, it would be a change of 0.5 percent, which would be very important. A half of a percent change in solar output could raise temperatures, eventually, about three-quarters of a degree Celsius, which, coincidentally, roughly equals the observed warming in the past century,” says Hansen. The apparent coincidence is no smoking gun, however. Because of their great heat storage capacity, the Earth’s oceans would buffer any increase in the Sun’s output for a long time. “Nevertheless, the potential is there for the Sun to be a significant player in the climate game, at least over the long term,” says Hansen, “which is why we need to keep studying the issue.”

The controversy caused by the uncertainty over the accuracy and reliability of various sensors underscores the need for overlapping observations that can be used for cross comparison. In pursuit of that objective, NASA launched its Solar Radiation and Climate Experiment (SORCE) in January 2003. Using a suite of four radiometers that will measure more characteristics of the Sun’s output than have ever been observed before, the mission will add its solar observations to the ones currently being made by SOHO-VIRGO and ACRIM3. “Previous sensors would experience drifts in measurements caused by environmental influences on the sensor—such as a change in temperature of the whole spacecraft [and not just the sensor itself] when it was exposed to the Sun,” explains Robert Cahalan, SORCE Project Scientist. SORCE will use a new measurement approach to filter out this background noise; it will collect observations in short, regular pulses rather than continuously—not giving background noise a chance to interfere with the signal. It will also be the first mission to measure irradiance at the individual wavelengths that account for 95 percent of the total solar energy, rather than as one lump sum.

Until more observations are collected, however, we are left with controversy and the dependence of difficult environmental decisions on our imperfect understanding of the Sun and its influence on climate. The question of whether there is an overall upward trend in the Sun’s output over the last two decades becomes more than a scientific debate when it steps out from the pages of research journals and into the world where individuals and societies are facing difficult decisions about how to respond to climate change.

When asked how he felt about the possibility that his results might be used as justification for not doing anything to reduce greenhouse gas emissions, Willson said, “It would be just as wrong to take this one result and use it as a justification for doing nothing as it is wrong to force costly and difficult changes for greenhouse gas reductions per the Kyoto Accords, whose justification using the Intergovernmental Panel on Climate Change reports was more political science than real science.”

The potential for the findings to be used such a way is something Lean has considered. “The fact that some people could use Willson’s results as an excuse to do nothing about greenhouse gas emissions is one reason we felt we needed to look at the data ourselves,” says Lean. “Since so much is riding on whether current climate change is natural or human-driven, it’s important that people hear that many in the scientific community don’t believe there is any significant long-term increase in solar output during the last 20 years.”

  • Willson, R.C., and A.V. Mordvinov. (2003) Secular total solar irradiance trend during solar cycles 21–23. Geophysical Research Letters, 30 (5), 1199.
  • Frohlich, C., and J. Lean. (2002) Solar irradiance variability and climate. Astronomische Nachrichten 323: 203–212.

Although small changes in total solar irradiance may be insignifcant in the short term, if the sun continues to change it could impact the Earth’s climate, perhaps amplifying or counteracting global warming. Because of this, the scientific debate over solar variability has spilled over into politics. (Photograph copyright Joe Klein, SkyChasers.net)