Understanding the Variable Sun
Changes in the energy from the Sun potentially could influence global change directly by modifying the Earth's surface temperature (Chapter 2) and by creating and destroying atmosphere ozone at variable rates (Chapter 3). Solar variability may also influence global change indirectly, by modifying the middle atmosphere, which is connected chemically, dynamically, and radiatively with the troposphere/biosphere (Chapter 3). In the upper layers of the Earth's atmosphere, and in the geospace environment, solar variations cause dramatic changes that are critical for understanding the processes within those regions, although the extent to which these changes couple to lower atmospheric layers (Chapters 4 and 5) is uncertain.
Observations over the past decade have provided an exciting perspective on how the Sun's energy inputs to the Earth change with time. In this period were obtained the first long term records from space of the solar radiative energy inputs to the Earth that are critical for studying solar influences on global change: total solar irradiance and the solar UV spectral irradiance, as well as fluxes of energetic protons and electrons. Ground based measurements were also made of solar observables closely related to the energy inputs measured from space. Physical associations between open field regions on the Sun, high speed solar wind streams, coronal mass ejections, and geomagnetic activity were established through
a variety of space missions. Taken together, these observations have revealed new insights into how solar magnetic activity modulates terrestrial solar energy inputs and how magnetized plasma from the Sun evolves as its flows to the Earth. These observations have established beyond doubt that the Sun's energy output varies continuously on all observed time scales.
Predicting, understanding, and monitoring global change are the ultimate objectives of the USGCRP (Chapter 1). Yet contemporary measurements of solar energy inputs alone reveal little about future solar variability nor of past solar variations that might have influenced the paleoclimate record, which is the focus of the Earth System History science element of the USGCRP. To begin to understand how the Sun varied in the past and how it might vary in the future, we must first understand why the Sun varies at all.
The fundamental physical processes that generate the variations observed in solar energy production are associated with the 22-year magnetic cycle of the Sun. The sunspot number time series remains the principal historical indicator of this cycle, and it is shown in Figure 6.1. This is the record of solar activity that was compared with the 14C and temperature time series in Figure 1.3 and with surface temperature anomalies in Figure 2.4. Recent monitoring from space indicates that both the total solar irradiance (Figure 2.1) and the UV irradiances (Figure 3.2) increase near the peak of the sunspot cycle and decrease during times of few sunspots. Likewise, the flow of energy, plasma, and magnetic fields from the Sun into the Earth's environment depends on the magnetic cycle. Fundamental to understanding the Sun's behavior as a variable star is understanding how variations in its emitted energy are generated from the magnetic activity cycle.
Origins of Solar Variability
The 22-year magnetic cycle of the Sun manifests itself as the familiar 11-year sunspot cycle, the 22-year cycle being simply two 11-year cycles having reversed magnetic field polarities. Physically, the sunspot cycle is a roughly periodic emergence, approximately every 11.1 years, of strong magnetic flux tubes at the solar surface in the form of sunspots. More
generally, the solar activity cycle pertains to the periodic emergence ofmagnetic flux that generates not just sunspots, which are dark, but avariety of phenomena, especially bright regions known as plages andfaculae that radiate strongly at UV and EUV wavelengths. The darksunspots and bright plages and faculae modify the radiation from the solardisk, thereby generating the variations observed by spaceborne solarradiometers. Also, changes in the Sun's magnetic field topology, due to
both flux tube emergence and latitudinally differential rotation of the solar atmosphere, generate field configurations that lead to transients such as solar flares and coronal mass ejections, and longer lived features such as coronal holes. These latter phenomena affect the Earth through input of high energy particles and plasma into the geospace environment.
Solar flares and solar global oscillations are prominent examples of solar energy variations on time scales of seconds to hours. Flares, although very energetic, occur over sufficiently small fractions of the solar hemisphere that even the largest and most energetic of them do not enhance total solar irradiance more than a hundredth of a percent. Nevertheless, enhancements in high energy emissions, such as EUV and X-rays, can be dramatic. The solar eruptions with which flares are associated also produce significant fluxes of energetic particles (electrons and protons and other nuclei). The enhanced EUV and particle fluxes from flares and coronal mass ejections can significantly alter the ionization state of the Earth's middle and upper atmosphere. At times when many eruptions occur in succession, the effects may persist for many months, as demonstrated by the semicontinuous solar proton events during 1989–1990 (Reid et al., 1991), discussed in Chapter 3.
Like flares, global oscillations of the Sun, such as the five minute p-mode oscillations, have minimal effect on total solar irradiance on the order of 3 parts per million of the total flux (e.g., Hudson, 1987). The significance of the solar oscillations is that they provide a unique technique for sensing physical properties and variations of the solar interior (Leibacher et al., 1985; Gough and Toomre, 1991), thus providing crucial insight into the mechanisms of the solar cycle.
The Sun's 27-day rotation modulates the steady radiative and plasma outputs from the Sun because at times (especially near activity maxima) the sources of these outputs are localized in narrow heliocentric longitudinal bands and are sufficiently long-lived that they reappear on successive rotations (see Lean, 1987, for examples), Rotational modulation is clearly seen in the total solar irradiance, with a rather complex waveform that reflects the competing effects of dark spots and bright faculae. Although 27-day periodicity is attributable to the Sun's rotation, no physical models explain the emergence, evolution, and decay of the magnetic active regions on the solar surface that cause changes in the amplitude and phase of this cycle. Since the 27-day rotation modulates solar irradiance by changing
the amount of active region emission seen at the Earth, observations and investigations of this periodicity aid in understanding the origins of the geomagnetic and photon flux variations that are important for the terrestrial environment (e.g., Lean, 1987).
Nor is the emergence of magnetic activity with an 11-year periodicity properly understood. Conceptually, a dynamo process is thought to underlie the Sun's magnetic cycle (e.g., DeLuca and Gilman, 1991), and models have been constructed to provide insight into the specific basic solar characteristics that must change to produce the periodic magnetic flux tube emergence that is responsible for solar energy input variations to the Earth. These models are necessarily theoretical, based on the internal interactions of a plasma rotating differentially within a convective envelope. No widely accepted solar dynamo model can reproduce all of the observed features of the solar cycle. A particular problem is the models' inconsistency with helioseismologic observations of the internal rotation of the Sun (e.g., Leibacher et al., 1985), an essential component of any model of flux emergence.
Even though the 11-year solar activity cycle appears to be the fundamental cause of changes in terrestrially sensed energy on decadal time scales, the possibility that long term changes may also be occurring in the Sun because of non magnetic mechanisms cannot be ignored. Short (five-minute) p-mode oscillations arising from a natural cavity resonance inside the Sun are well established. These oscillations are a property of the global Sun, a star, as opposed to localized sources of fluctuation such as sunspots and plages/faculae. Other solar features, such as ephemeral magnetic regions and the chromospheric grains seen in Ca IIK, contribute to the energy output and are widely distributed over the Sun.
Over times much longer than the 11-year activity cycle, there is the possibility of superimposed slow secular change in the Sun's energy inputs to the Earth. The Maunder Minimum (Eddy, 1976), seen clearly in the sunspot record in Figure 6.1, is evidence for such change and the only example in the contemporary solar record. The absence of sunspots and possible supression of the 11-year cycle during this extended period suggests that the flux emergence process may have stopped completely for several decades. Relative to the present, the Sun's internal circulation may have been substantially different (Eddy et al., 1976), and its diameter larger (Nesme-Ribes et al., 1993). Indirect evidence for many similar
episodes in solar behavior comes from the radiocarbon and auroral records (Eddy, 1976), but only very recently has it been recognized that substantial decreases in the Sun's radiative output might have accompanied these episodes (White et al., 1992; Lean et al., 1992a). The 14 C data also suggest that solar activity was very high in the twelfth century, an epoch corresponding to the Medieval Warm Period of approximately 300 years (Figure 1.3). Eddy (1976, 1977) suggests that the total irradiance may follow the envelope of the sunspot cycle curve, waxing and waning on century time scales. These findings indeed suggest a relationship between the Sun and the climate, but a solar variability model that describes the Sun's energy inputs to Earth at times in the past has yet to be developed.
Because the Sun is apparently a normal star, insights into solar variability can be gleaned from observations of Sun-like phenomena in stars with mass, age, and rotation rates similar to the Sun's (Baliunas, 1991). In particular, comparative solar and stellar Ca II emission measurements indicate that the activity levels of the contemporary Sun correspond to the highest levels observed in other stars (White et al., 1992). These observations also suggest that times of arrested activity, as exemplified on the Sun by the Maunder Minimum, may be common in other stars, and that such times appear to be accompanied by reduced energy output (Baliunas and Jastrow, 1990; Lean et al., 1992a).
Relationship Between Solar Surface
and Energy From the Sun-as-a-Star
To better understand the relationships between changes on the Sun and changes in the Earth's atmosphere, it is important to understand how measurements made at the Earth at 1 astronomical unit (AU) project back to structures on the solar surface.
The transmission path of solar radiative input to the Earth is line-of-sight. Active regions (sunspots, plages and faculae, filaments, coronal holes, etc) modify the local intensity of the solar surface, resulting in an inhomogeneous solar disk whose radiation field is variable in both
time and direction. As shown in Figure 6.2, these surface features evolve continuously throughout the solar activity cycle, and they have the largest effect near times of activity maxima. The radiative energy input to the Earth is the integral of the radiance from the entire solar hemisphere visible at the Earth--that is, the irradiance. As a result of this integration, the interpretation of variations observed in the total and spectral solar irradiances involves constructing irradiance time series by combining the contributions from sunspots, plages, the bright magnetic network, and internetwork regions. In this way, variations in both total and ultraviolet spectral irradiance can be traced to the changing populations of active regions on the solar disk.
Ground based observatories measure the position, brightness, and area of sunspots, plages, and faculae daily (e.g., Beck and Chapman, 1993). These data have been combined in simple empirical models to reconstruct the measured irradiances (Cook et al., 1980; Lean et al., 1982; Foukal and Lean, 1988; Willson and Hudson, 1991). Comparisons of the model calculations with the measured irradiances show that the principal contributors to radiative variability are sunspots, plages/faculae, and chromospheric network, both on solar rotation time scales as well as over the decadal scale of the solar cycle. Residuals between the measurements and reconstructions are analyzed for the possibility that they arise from experimental error, incorrect assumptions in the models, or missing energy in storage deep in the Sun.
In making the connection between fluctuations in irradiances and solar surface inhomogeneities, two types of ground based data have been crucial. One is photometric data on sunspots, faculae, and plages, and the other is spectral irradiance data; both are measured in appropriate regions of the solar spectrum. The photometric data provide records of the individual active regions that generate irradiance variations (Figure 6.2). Sunspots are identified most clearly in white-light solar spectroheliograms, whereas images in the Ca II K line remain the principal source of photometric data on plages and the network, features that are also seen clearly in solar images in other spectral lines, such as the He I 1083 nm line. The spectral irradiance (Sun-as-a-star) data measure the integrated effects of these active regions on layers of the Sun from which the total and UV radiations are also emitted. Figure 6.3 shows three of the most informative solar records: the Sun-as-a-star chromospheric Ca II and He I indices (White
and Livingston, 1981; Harvey, 1984; Livingston et al., 1988; White et al.,1990) and the sunspot blocking. In this regard, the solar 10.7 cm radioflux is also important, especially in an historical context, since it was theonly Sun-as-a-star indicator measured during solar cycles 19 and 20.
Figure 6.4 shows a comparison of the total solar irradiance measured by ACRIM I (on SMM) and by Nimbus 7/ERB with models developed from ground based data for the respective data sets. The rotational modulation data during 1982 illustrate that the day-to-day variation that arise from the competing effects of dark sunspots and bright faculae are well reproduced by the Foukal and Lean (1990) model. Much of the longer term variability during solar cycle 21 and the ascending phase of cycle 22 is also reproduced by this model, with the exception of the first years of the record. The longer term solar cycle changes occur because of a brightness component in addition to the sunspots and the brightest faculae associated with magnetic active regions. While the existence of this additional 11-year variability component has not been verified by direct observation, it is thought to reside, at least in part, in the network of bright emission that surrounds the large active regions (Foukal and Lean, 1988; Foukal et al., 1991). It may also have a global (i.e., non magnetic) component (Kuhn et al., 1988; Kuhn and Libbrecht, 1991). Discrepancies between the measurements and the model during 1979–1980 may be instrumental in origin. Ifnot, they raise the possibility of a variability component acting over time scales longer than the 11-year cycle (Lee III et al., 1994). The utility of these models, and the need to resolve discrepancies with the measurements, emphasize the need for high precision image data in the interpretation of irradiance time series in the future.
Statistical comparisons indicate that the Ca II and He I indices provide reconstructions of solar UV irradiance variations superior to those afforded by the 10.7 cm radio flux, over both solar rotation and solar cycle time scales (Barth et al., 1990; Bachmann and White, 1994). Nevertheless, the 10.7 cm flux has been used extensively for the past few decades as a proxy for EUV and UV irradiance variations in terrestrial applications. For example, essentially all analyses of ozone data for the purpose of extracting long term trends have used the 10.7 cm flux, in lieu of UV irradiance data, to account for solar forcing of ozone changes (Stolarski et al., 1991; Hood and McCormack, 1992; Randel; and Cobb, 1994; Reinsel
et al., 1994b) (Chapter 3). With the availability of ground based data that better reflect the processes that generate solar radiative output variations, empirical models that predict these variations can be improved. Knowledge of the relationship between the 10.7 cm radio flux and solar radiative output nevertheless remains important for its historical relevance.
Plasma and Particles
Plasma from the solar corona flows radially outward from the Sun and impacts the Earth as the solar wind. Frozen into this plasma is the interplanetary magnetic field (the extended field of the Sun). Solar rotation, combined with the outward motion of the highly conducting plasma, winds this field into a spiral pattern and produces high speed and low speed streams. High speed solar wind streams (velocities of 700 to 850 km/sec) originate in regions where the magnetic field lines are open and connect to the interplanetary field; in contrast, the slow streams (velocities of 300 to 400 km/sec) come from regions above sunspot complexes where the magnetic field is closed to the solar surface. Shocks are formed when a high speed stream overtakes a low speed stream. Coronal mass ejections periodically disrupt the quasistationary pattern of high and low speed flows from coronal holes and streamers. The projection of solar wind profiles at the Earth back to their origins on the Sun can therefore be ambiguous, which emphasizes, again, the importance of knowledge of solar magnetic field structure in determining the time profile of solar energy input at 1 AU.
Solar energetic particles carried to the Earth by the solar wind interact with the Earth's magnetic field in ways that depend on the spectral distribution of their energy. The particles spiral along terrestrial magnetic field lines, entering atmosphere primarily in the auroral zones surrounding the two geomagnetic poles (Figure 3.3). Projection of this complex field interaction backward to the Sun is aided by solar images in the visible spectrum that locate the sites of flares that are frequently associated with the eruptive events that are the sources of these particles. At times of high solar activity, however, the occurrence of numerous eruptions makes identification of the source of the particle flux more difficult.
The diffuse cosmic ray flux that produces 14C in the Earth's atmosphere (see Chapters 3 and 5) is modulated by solar activity on passing through the heliosphere (e.g., Lopate and Simpson, 1991). The particle flux at the Earth is highest when solar activity is at a minimum, so that the naturally archived 14C record contains the signature of past solar variability excursions. Interpreting the 14C record depends on understanding this modulation process as well as other effects, such as changes in magnetic field strength and in the terrestrial carbon cycle (Stuiver and Braziunas, 1993). Until there is a solid appreciation of the connection between the cosmogenic isotope variations and the modulation of the energy output from the Sun by magnetic active region phenomena (sunspots, plages, network), it will be difficult to construct reliable quantitative estimates of the strengths of extrema in solar energy, such as during the Maunder Minimum type episodes that appear to have occurred commonly (every 200 years or so) in the past.
Requirements For Improved Understanding
Since relatively reliable knowledge of solar radiative output variations exists for only the most recent 11-year solar cycle, much will be learned by the continuation and improvement of the total and UV spectral irradiance observations from space and of the ground based indicators of the various solar processes that cause and reflect their variations. The required ground based data are primarily white light images of sunspots and the Ca II and He I indices and images. The images (preferably of a few arc-seconds resolution or better, with accurate photometric calibration) are necessary not only for the formal construction of irradiance time series from ground based surrogates, but also to allow fundamental descriptions of spots, plages, faculae, and network as they evolve in time. Continuous daily measurements are needed at least for the duration of the Sun's magnetic cycle of 22 years, and more ideally for 100 years or longer, since the 88-year Gleissberg cycle may represent the dominant forcing factor.
The basic process that leads to variations in solar energy input to the Earth is the emergence and evolution of magnetic flux tubes in the solar atmosphere. The principal observational needs are for better photometric and positional data pertaining to the flux tubes that form sunspots, plages, faculae, the magnetic network and their motions on the solar disk. Theoretical models are needed to understand the basic energetics in these flux tubes and how the magnetic field interacts with solar convection to change the energy transport. Since this affects the brightness of magnetic structures throughout the solar spectrum, it bears directly on interpretation of the full disk observations.
An important tool in interpreting measurements of solar radiative output is the comparison of the measurements with synthesized solar spectra (Kurucz, 1991; Mitchell and Livingston, 1991; Avrett, 1991). Theoretical reconstructions of the solar spectrum, although based on one-dimensional thermodynamic and structure models, include millions of solar spectral lines from EUV to infrared wavelengths and provide insight into the spectral composition of the solar irradiance. At present it is not certain how the spectral energy distribution of the total irradiance changes over the solar cycle. The theory of the formation of the solar spectrum also establishes physical connections between different wavelengths, providing insight into the extent that variations at one wavelength mimic variations at another. Consideration of such connections may lead to more efficient observational programs in the future.
Magnetic flux tube emergence is thought to be associated with a dynamo lying deep inside the Sun. The goal of stellar dynamo models is to reproduce the periodic variation of flux tube emergence seen on the Sun and also to show the systematic latitude and polarity variations that occur as spots move from high latitudes toward the equator as the cycle progresses, with the polarity of sunspot pairs and the polar fields reversing from one cycle to the next. The necessary empirical boundary conditions for such models come from detailed knowledge of magnetic flux tubes, their distribution in both time and position on the Sun, and evolution of their energy transport seen in observations of radiative output. The most important constraints lie in the differential rotation structure of the solar interior and the properties of the interface between the convection zone and the Sun's radiative core, where the solar dynamo is thought to lie. Helioseismic observations continue to play a crucial role in studying the
dynamo because they provide a way to probe the solar interior, but they will illuminate the dynamo problem only if measurements extend over solar cycle time scales.
In the context of global change, understanding the past behavior of the Sun may well be essential for unraveling the paleoclimate record. Coincidences between climate change in the twelfth and seventeenth centuries and changes in 14C as a proxy for solar activity (corroborated by the 10Be record) suggest that there may be threshold levels of high and low solar activity at which the Sun begins to play a significant role in changing global climate (see Chapter 2). Knowledge of physical conditions on the Sun at these extrema is primitive because modern scientific observations have all been made during an era of high solar activity. Galileo's and others' discovery of sunspots at the beginning of the seventeenth century came at a time to document an ensuing period of low solar activity. Without this documentation, the possible role of the Sun in the Little Ice Age might still not be appreciated. Given the meager and often discontinuous evidence for solar variability in the past, inference of the physical state of the Sun at times of extrema is speculative but necessary. The first steps can be taken with empirical models now available (e.g., Foukal and Lean, 1990; White et al., 1992; Lean et al., 1992a; Hoyt and Schatten, 1993; Nesme-Ribes et al., 1993), but the credibility of such research would be strengthened immeasurably through development of a successful physical model of the solar cycle.
An expanded view of solar variability is provided by knowledge of cyclic behavior in other stars, afforded by long term measurements of stellar Ca II emissions in Sun-like stars. Baliunas and Jastrow (1990) present stellar cycle data that may indicate the presence of Maunder Minimum-type episodes in one-third of the observed stars, but their sample (13 stars) is so small that this conclusion must be regarded as speculative at this time. It is, however, consistent with an independent result from analysis of the radiocarbon data by Damon (1977). Furthermore, the distribution of Ca II emission exhibited by the Sun-like stars does appear to be consistent with the range of Ca II K emission seen in the present-day Sun (White et al., 1992). By assuming that during the Maunder Minimum,
the mix of active region emission that generates the solar Ca II irradiance was somewhat different than is currently seen in the Sun, the radiative output of the Sun can be estimated for the Maunder Minimum (Figure 2.3).
Current ability to predict solar activity is at best primitive. Statistical methods predict sunspot numbers and the 10.7 cm radio flux 12 months in the future with moderate success. There are also precursor methods that predict the strength of the next solar cycle from the behavior of polar structure on the Sun and geomagnetic activity in the declining phase of the current cycle (e.g., Schatten and Pesnell, 1993; Thompson, 1993). But there is limited physical understanding of why these precursor methods should be appropriate except that the magnetic fields and corona near the solar poles change near solar maximum and hence may herald the onset of the new cycle before the next generation of sunspots appears.
On century time scales, the periodicities of 11 and 88 years identified in the sunspot record, together with the 208 year periodicity found in the 14C record, provide limited guidance to future solar behavior, such as the occurrence of the next Maunder Minimum. The time span of solar measurement is simply too short for reliable prediction of solar extrema occurring sporadically every 200 years or so. Nevertheless, it has been speculated that the concatenations of the 208 and 88 year periods may have contributed to generally increasing solar activity levels during the twentieth century, with maximum activity predicted to occur during the first half of the twenty-first century (Damon and Sonnet, 1991).
Predictive capability will be substantially improved when a complete understanding is obtained of the mechanisms within the solar atmosphere that produce the emitted radiation, and form sunspots and plages. Predicting the Sun-as-a-star energy quantities needed for global change studies will ultimately require development of a theory for the solar dynamo that can accommodate known solar behavior. Yet, the very nature of solar variability, whether driven by an internal chronometer (Dicke, 1978) or by stochastic or chaotic processes (Mundt et al., 1991; Morfill et al., 1991; Kremliovsky, 1994), remains elusive. Solar activity levels may well defy reliable prediction in the near future.