Originally published in:
Eos, Transactions, American Geophysical Union, Vol. 76, No. 8,
February 21, 1995, Page 75, 83
Two earlier articles published in the December 6 and December 27, 1994 issues of Eos outlined the history of solar-terrestrial research from Sabine's founding discovery in 1852 through Chapman and Ferraro's landmark paper in 1931. This article deals with the development of the field from 1930 to the present with a focus on the identification of the solar sources of the two basic types of geomagnetic storms: recurrent and sporadic.
In a series of three papers during 1928-1929, W.M.H. Greaves and H.W. Newton provided a key insight to the Sun-magnetic storm puzzle. Following in the footsteps of Ellis and Maunder, they compared Greenwich solar and magnetic observations, using data from 1874 through 1927. In a now familiar pattern, they grouped the magnetic activity according to size. They confirmed Maunder's finding of an association between the great storms and large sunspots near disk center. They noted, as had previous authors, that the larger storms were those most likely to be preceded by sharp onsets or "sudden commencements." Their breakthrough discovery, however, resulted from an examination of the recurrence patterns of storms as a function of their sizes (Figure 1). This figure clearly showed that 27-day recurrence is primarily a property of the smaller storms and that the larger storms occur sporadically.
As Greaves and Newton pointed out, this behavior was unexpected. "At first sight it is rather paradoxical that the storms of larger range should not exhibit the recurrence tendency, for it is precisely these storms which seem to show a connection with sunspots larger than the average, and in view of the marked tendency of the larger spots to persist for more than one rotation, it might have been expected that the largest storms would show the recurrence tendency best.... there is seen to be a marked antithesis between the tendency to recurrence and the tendency for storms to be associated with spots."
Greaves and Newton stopped short, however, from hypothesizing two separate types of storms, one associated with spot regions and one with quiet regions on the solar disk. They did suggest that the streams associated with great and small storms were fundamentally different; those associated with the great storms were generally short-lived and did not survive a solar rotation in many cases.
A detailed investigation by Bartels in 1932 took the work of Greaves and Newton a step further and accelerated the breakdown of the exclusive coupling between spot regions on the Sun (either active or dormant) and magnetic storms that had held for 80 years following Sabine's initial discovery. Bartels analyzed 27-day recurrences of storms and quiet periods for the years 1906-1931 by means of stacked 27-day plots of the magnetic C index. In this representation, sequences of storms and quiet periods were immediately recognizable as vertical columns of like symbols. Bartels noted, "If the time T of passage from the Sun to the Earth would be constant for all corpuscular streams, then our diagram could be conceived of as a chart of the Sun, indicating the heliographic longitude of the active regions on the Sun-which we shall call here M regions."
Bartels pointed out that while sequences of storms could last for more than 10 rotations, sunspot groups only very rarely had life times as long as 5 months. He noted that especially long sequences of storms tended to occur near the end of the sunspot cycle and called attention to a sequence of storms in 1923 that persisted through times when "for several weeks in succession, not a single sunspot was visible!" Conversely spot-groups had been observed crossing the Sun's central meridian without causing magnetic storms. Bartels concluded, "The identification of the M regions with sunspots or other solar phenomena is possible in some cases only, while in many cases the M regions lead, so to say, an independent life."
The work of Greaves and Newton and Bartels led to the realization [Bartels, 1940; Allen, 1944] that there were two basic types of magnetic storms: sporadic and recurrent. The large sporadic storms were clearly associated with sunspots while recurrent storms, which tended to be smaller, originated in M regions, whose solar signature, if any, remained to be determined. There were other key differences. Sporadic storms tended to be preceded by sudden commencements and occurred most frequently near solar maximum. Recurrent storms typically had gradual onsets and favored the post-maximum period.
Bartels' work and the evocative nomenclature he introduced initiated a 40-year search for solar manifestations of M regions. The basic approach used to identify M region signatures was to take a visible solar feature characteristic of a center of activity-for example, a sunspot group, filament, or coronal enhancement as observed with the Lyot coronagraph-and to use the time of central meridian passage (CMP) of this phenomenon as a time marker in a superposed-epoch study of magnetic activity. Despite the clear indications given by Greaves and Newton ("marked antithesis") and Bartels ("not a single sunspot") of the true nature of M regions, relatively few studies investigated regions lying outside of active centers or their remnants as candidate M regions. Bell and Glazer [1957] provide a notable exception to this rule.
The various statistical studies revealed a minimum in magnetic activity approximately 3 days after CMP of active region features and a maximum approximately 6 days after. Because the speed of the recurrent corpuscular streams was not known, these results were interpreted in diametrically opposite ways. One school, of which Mustel' was the most prominent advocate, argued that the active regions were the source of the streams arriving at Earth at CMP + 6 days to give rise to recurrent storms; the other school, led by Allen and Saemundsson, postulated that active regions were related to the magnetic minimum at CMP + 3 days and that the maximum at CMP + 6 days was due to M regions lying outside of active regions. By the early 1970s, the preponderance of evidence from in situ space observations and theory [see Hundhausen, 1972] supported the Allen and Saemundsson viewpoint.
The line of research that began with Birkeland's "corpuscular hypothesis" and Maunder's 27-day recurrent streams culminated in the late 1950s and early 1960s with Parker's prediction of a continuously outflowing "solar wind" and its subsequent detection by in situ explorations of the interplanetary medium by early spacecraft. Parker's theory followed from the interpretation by Grotian and Edlen about 1940 of the coronal emission-lines in terms of a million-degree corona; Chapman's analysis showing that a static corona of 106 K must extend beyond the orbit of the Earth; and Biermann's suggestion that comet tails pointed away from the Sun because of a continuous outflow of solar particles. Thus it was realized that the corpuscular streams, either transient or corotating, that were associated with magnetic activity were components of a more general continual outward expansion of the corona.
In a key development of the early spaceflight era particularly relevant to magnetic storms, it was hypothesized by Dungey and verified by in situ observations that coupling between interplanetary disturbances and the Earth's magnetic field would be most efficient when the interplanetary field contained a component that was anti-aligned with the Earth's dipole field. This effect was subsequently used by Russell and McPherron to explain the semiannual variation of geomagnetic activity.
Despite indications going back to Carrington and echoed through the years by Tacchini, Fitzgerald, Chapman, and others that the influence of sunspots on terrestrial magnetism was felt only sporadically when disturbances or eruptions took place in sunspot regions, the first focused study of the relationship between bright chromospheric eruptions (solar flares) and magnetic storms did not appear in the literature until 1931. In that year Hale published a paper subtitled "Solar Eruptions and Their Apparent Terrestrial Effects" in The Astrophysical Joumal in which he gave an anecdotal review of cases in which a flare on the Sun was followed by a magnetic storm. In all there were fewer than 10 good associations, including the Carrington event of 1859, the flare observed with the spectroheliograph in 1892, the events of 1908 and 1909 that provided support for the corpuscular hypothesis, and other later events.
Fewer than 10 events seems a relatively small number for a 70-year
period. The problem was basically instrumental; the spectroheliograph
was a cumbersome instrument that was not well-suited
for observing rapidly changing phenomena such as flares. Thus
during the 1920s, Hale developed the spectrohelioscope, an
instrument that worked on the same principle as the
spectroheliograph but which allowed for the first time the visual
observation of the solar disk at selected wavelengths. The
resultant observations of the rich complexity of solar activity
were compelling. Hale described a large flare observed with the
new instrument in 1926 as "the most remarkable solar phenomenon I
have ever seen"-this coming after 40 years of observing the Sun.
As his last contribution to solar-terrestrial physics, Hale made arrangements to have spectrohelioscopes distributed to observatories around the world. At his urging, a worldwide flare patrol was instituted under the auspices of the International Astronomical Union. The results were reported regularly in the Quarterly Bulletin of Solar Activity beginning in 1934. The patrol resulted in an explosion of knowledge about solar flares during the 1930s and 1940s. One early result was the use of the spectrohelioscope to identify flares as the causes of the sudden ionospheric disturbances (SIDs) associated with flare shortwavelength emission. Thus in 1937 Bartels was able to correctly explain the prompt-a type of SID referred to as a magnetic crochet-and delayed, or magnetic storm, geomagnetic responses of the 1859 eruptive flare
In 1943, observations from the flare patrol also enabled Newton (Figure 2) to give statistical weight to the relationship between flares and storms indicated by Hale's 1931 paper. Newton examined the magnetic response of 37 large (class 3+) solar flares and found that over 80% of such flares located within +45° longitude of disk center were followed by storms. This gave an indication of the size of transient streams, and the symmetric distribution of points east and west of the equator (Figure 3) indicated that the transient streams propagated radially. The average delay for unambiguous cases of great storms was ~26 hr, identical to that found by Hale for a smaller sample; for the greatest storms the delay was ~20 hr. Newton's result also provided a ready explanation for the case in which large sunspots were observed transiting the disk without an accompanying storm. "... the reason is to be attributed to the absence of a corpuscular stream at that time rather than to the Earth's just missing a narrow but continuously ejected stream..."
In 1950, Kahn demonstrated that the radiation pressure mechanism generally invoked by advocates of the corpuscular hypothesis was inadequate to expel transient streams from the Sun. The nature of the propulsion mechanism of transient streams and the acceleration mechanism for recurrent high speed streams remain open questions to this day. In situ observations in the 1960s showed that transit times of 40-60 hours were the norm for transient streams and that the flarestorm delays of about 1 day reported by Hale and Newton referred only to a small number of exceptional events.
In a remarkable demonstration of the power of new types of observations, space borne instruments in the early 1970s helped to solve one long-standing problem of solar-terrestrial physics and initiated a paradigm shift in another area. Coronal holes, observed as dark rents in the solar atmosphere in images of the EUV and X ray corona by OSO spacecraft and rocket flights, were rapidly identified as the long-sought M regions- the sources of high-speed wind streams and the 27-day recurrent geomagnetic storms. The observation of coronal mass ejections (CMEs) by OSO 7 and Skylab provided a coronal counterpart to chromospheric flares.
The significance of CMEs for storms was immediately recognized, but there were surprises. CMEs appeared to be better associated with eruptive prominences than with flares; timing studies indicated that CME "liftoff" often preceded flares; and the scale sizes of CMEs were much larger than flares. Most flares, even some big flares, lacked associated CMEs. The energy of CMEs was comparable to that estimated for large flares alone and impressive CMEs could have relatively unimpressive flare signatures. These facts were deduced from analyses of CMEs observed with coronagraphs on Skylab and later space missions. Comparisons of CME observations with interplanetary shocks, solar energetic particle events, and geomagnetic storms resulted in a paradigm shift, still in progress, from flares to CMEs as the key solar phenomenon for nonrecurrent interplanetary disturbances [Kahler, 1992; Gosling, 1993]. In a direct and early application of this new paradigm to geomagnetic storms, Joselyn and McIntosh showed in 1981 that disappearing solar filaments outside of solar active regions (an eruptive "nonflare" phenomenon), long suspected by Newton and others as possible sources of storms, could, in fact, be associated with fairly large storms.
This review [including Cliver, 1994a,b] has addressed the first ~140 years of research on the relationship between solar activity and geomagnetic storms from the solar side of the question. From that vantage point, the history is one of intermittent progress toward definition of the geoeffective solar phenomena. Thus sunspots, which provided the initial evidence for a solar-terrestrial connection, were eventually discarded in favor of flares and complemented by M regions. In turn, flares and M regions were displaced by CMEs and coronal holes, respectively, although this last statement is an oversimplification that ignores ongoing debates.
Much remains to be done. Basic questions regarding the propulsion of CMEs and the acceleration of the high-speed wind streams from coronal holes remain open. Attempts to identify signatures of CMEs in the interplanetary medium have proven to be controversial, and the relative efficacies of the mechanisms through which solar wind energy is coupled to the magnetosphere are undetermined. Research on these various topics is progressing with an up-tempo version of the same scientific minuet of forward, backward, and sideways steps that characterized the search for the solar sources of geomagnetic storms. Lord Kelvin's challenge of 100 years ago still has resonance, "The more marvelous, and for the present inexplicable, all these subjects are, the more exciting becomes the pursuit of investigations which must, sooner or later, reward those who persevere in this work."
Allen, C. W., Relation between magnetic storms and solar activity, Mon. Not. R. Astron Soc., 104, 13, 1944.
Bartels J., Terrestrial-magnetic activity and its relations to solar phenomena, Terr. Magn. Atmos. Electr., 37, 1, 1932.
Bartels, J., Solar radiation and geomagnetism, Terr. Magn. Atmos. Electr., 45, 339, 1940.
Bell, B., and H. Glazer, Geomagnetism and the emission-line corona, 1950-1953, Smithson. Cont Ashrophys., 2, 51, 1957.
Cliver, E. W., Solar activity and geomagnetic storms: The first 40 years, Eos, 75, 569, 1994a.
Cliver, E. W., Solar activity and geomagnetic storms: The corpuscular hypothesis, Eos 75, 609, 1994b.
Gosling, J. T., The solar flare myth, J. Geophys. Res., 98, 18, 937, 1993.
Greaves, W. M. H., and H. W. Newton, On the recurence of magnetic storms, Mon. Not. R. Astron Soc., 89, 641, 1929.
Hundhausen, A. J., Coronal Expansion and Solar Wind, Springer-Verlag, New York, 1972.
Kahler, S. W., Solar flares and coronal mass ejections, Ann. Rev. Astron. Astrophys., 30, 113, 1992.