Appeared in Reviews of Geophysics, 27, 1989, p. 103-114.
Polar Magnetic StormsOne additional piece of the picture deserves mention: magnetic disturbances associated with the aurora, like those observed by Hiorter and Celsius. Such disturbances are far more intense, rapid, and frequent than magnetic storms observed at low and middle latitudes. Birkeland studied them in 1902-1903 using a network of four stations--in Norway and on Iceland, Spitzbergen (Svalbard), and Novaya Zemlya [Birkeland, 1908, 1913; Boström 1968]--and concluded that there existed a distinct type of magnetic storm, the "elementary polar magnetic storm" with a typical time scale of less than an hour, associated with the aurora and with electric currents which descended along auroral field lines and flowed horizontally along auroral arcs.
Birkeland died in 1918; his work was not followed up for many years, and in the decades that followed, relatively few magnetic studies were performed at high latitudes. Chapman did not believe that Birkeland's "polar storms" existed. He realized that they were much shorter than the nonpolar storms with which he was familiar, and in his encyclopedic two-volume treatise on geomagnetism, jointly written with Julius Bartels [Chapman and Bartels, 1940] (see also Chapman ), he suggested that Birkeland's events were probably just isolated phases of magnetic storms. He noted there that "a great magnetic storm is a unitary phenomenon, going through regular phases" and maintained that Birkeland's polar storms ". . . seem to be clearly part of a single phenomenon, waxing and waning in unison with the non-polar disturbance field."
The name "substorm," coined by Chapman for this phenomenon about 20 years later, reflected that attitude, though by then Chapman must have begun to realize the importance of Birkeland's early observations [Akasofu, 1970, p. 603] (see also Siscoe ). We now view substorms as impulsive acceleration events, quite possibly terrestrial analogs of solar flares.
By the 1950s this realization was slowly forming, and there was considerable interest in "magnetic bays," large magnetic disturbances in the auroral zone which would nowadays be classified as substorms [Silsbee and Vestine, 1942]. Currents flowing into the auroral zone and out of it, however, were observed only in the late 19608, and their global pattern was first mapped in 1974 [Zmuda and Armstrong, 1974; Iijima and Potemra, 1976]: they are now known as Birkeland currents [Schield et al., 1969, p. 247]. Contrary to Birkeland's interpretation, much of the horizontal part of their circuit, in the ionosphere, flows not along auroral arcs but perpendicular to them, for by a quirk of electrodynamics [Fukushima, 1969, 1976] the main circuit produces only a weak magnetic signature on the ground. What Birkeland observed was mostly the signature of an associated Hall current, the "auroral electrojet" which parallels auroral arcs.
The picture changed considerably after 1957, the start of the International Geophysical Year (IGY). The IGY was an international effort which included the launch of the first artificial satellites, and it formed a natural transition in the history of magnetospheric physics.
The implications of that transition are best appreciated in the context of other research on our physical environment. The surface of the Earth, the oceans, and atmosphere are completely accessible and can be directly studied, even experimented upon: in the jargon of Earth observation from space, we have "ground truth." The realm of the astrophysicist, on the contrary, can only be sensed remotely and imperfectly, and the amount of information we can ever hope to receive from it is severely limited [Harwit, 1981]. By necessity our explanations of astrophysical phenomena are laced with guesswork, and in many cases (e.g., the origin of cosmic rays) it is quite likely that such guesses will never find convincing confirmation.
Magnetospheric physics stands halfway between those extremes. Until the IGY it was very much like astrophysics: the magnetosphere could only be sensed remotely, and much of what was believed about it was merely intelligent guesswork. Then came artificial satellites and provided some "ground truth," and it is interesting to compare what they revealed with what was believed beforehand.
Many important magnetospheric features had indeed been inferred before spacecraft were available, but in almost every case some important detail was missing or wrong. The Chapman-Ferraro cavity was predicted as a temporary rather than permanent feature, and the same was true for the radiation belt. Alfvén's convection contained a nucleus of truth, but electric field effects supplemented rather than supplanted the Chapman-Ferraro picture, and the convection which they produced was found to flow from the tail sunward, opposite to its direction in Alfvén's theory. Birkeland's auroral currents did exist, but their configuration was not the one predicted. The existence and importance of the magnetospheric tail generally went unsuspected, and so did the existence of parallel electric fields along auroral arcs, although Alfvén later developed the theory of quasi-neutral equilibria, relevant to such fields. All this underscores the essential role of in situ observations: one can only speculate how much of this might be paralleled in astrophysics.