1966, Vol. 14, pp. 587 to 595.
THE AURORAL OVAL, THE AURORAL SUBSTORM, AND THEIR RELATIONS WITH THE INTERNAL STRUCTURE OF THE MAGNETOSPHERE S.-I. AKASOFU*
Geophysical Institute, University of Alaska, College, Alaska (Received
15 February 1966)
Abstract-If the neutral sheet is the source region of aurora1 particles, as suggested by Piddington, they flow in a thin layer just outside the outer boundary of the trapping region from the neutral sheet towards the Earth and produce an oval-shape glow on the polar ionosphere, which we identify here as the aurora1 oval. It is shown that the growth and decay of the ring current and the neutral sheet current during geomagnetic storms change the internal structure of the magnetosphere and that it is manifested by changes of the aurora1 oval, in particular the equatorward shift of the oval during a large main phase decrease and a rapid expansion and the subsequent contraction of the oval width during the aurora1 substorm. 1. INTRODUCTION
Recent studies of the aurora1 substorm and the polar electrojet have revealed that these two fundamental polar phenomena are a manifestation of interactions between the magnetospheric plasma near the outer boundary of the outer radiation belt and the neutral atmosphere underneath. (l) Two kinds of interaction, dynamical and atomic, occur in a narrow belt along the intersection curve between the outer boundary of the outer belt (or more This belt has an oval shape and its center precisely, the trapping region) and the ionosphere. is located a few degrees away from the dipole pole toward the dark hemisphere, and is called the aurora1 ov~l.~~)Both interactions occur most violently in the ionsospheric E-region of the oval which is a rather thin transition region between the magnetospheric plasma and the neutral atmosphere. There, an intense polar electrojet flows, and a significant portion of the aurora1 light is produced. The average location of the oval obtained by Feldsteinc2) is shown in Fig. 1. Khorosheva’3) has confirmed that at a particular instant, auroras tend to lie along the oval. Roughly speaking, the oval is fixed with respect to the Sun, and the Earth rotates under the oval. Here, the difference between the aurora1 ovaI and the aurora1 zone should be clearly recognized. The aurora1 zone (which is approximately the dipole latitude circle of 67”) is simply the locus of the midnight point of the oval, where active auroras are most frequently seen. Figure 1 shows also the iso-intensity contour of the flux = (104/cm2 set) of trapped elecSince the trons of energies greater than 40 keV, obtained by Frank, Van Allen and Craven.t4) flux of the electron decreases rapidly to the poleward side of this contour, this contour should lie close to the intersection line between the ionosphere and the outer boundary of the outer radiation belt (or the trapping region). Their agreement is remarkable and must provide an important clue to understand aurora1 phenomena. In fact, we propose here that the aurora1 oval is a belt which lies just to the poleward side of the intersection curve between the trapping region and the ionosphere and thus is directly connected to the neutral sheet. This has already been suggested by Piddington. (5) Obviously, however, this conjecture cannot be justified from Fig. 1 which is a combination of the two rough statistical results.
FLUX = IO’ (Cm’-Sec)e’ TRAPPED
FIG. 1. LOCATIONOFTHE AURORALOVALOBTAINED BY FELDSTEIN(~'AND OFTHEISO-INTENSITY (THE FLUX = 104/cmesec) CONTOUR LINI~OF THE TRAPPED ELECTRONS OF ENERGIES GREATER THAN 40 keV.f41
FIG. 2. SCHEMATIC DRAWING (THE NOON-MIDNIGHT MERIDIAN CROSS-SECTION) OF THE STRUCTUREOFTHEMAGNETOSPHEREAND OFTHE AURORALOVAL. NOTETHATTHEGREATDAY-NIGHT ASYMMETRY OF THE TRAPPING REGION IS PROJECTED ON THE IONOSPHERE AND IS SEEN AS THE ECCENTRICITY OF THE AURORAL OVAL.
However, Fritz and GurnetP) have observed intense fluxes of low energy (auroral) electrons just outside the outer boundary of the trapping region in the midnight sector, supporting such a conjecture. Figure 2 shows schematically the proposed relationship between the aurora1 oval and the internal structure of the magnetosphere. In Fig. 2 the trapping region has been defined as the region where energetic electrons (E > 40 keV) can execute their motions by approximately conserving the three adiabatic invariants.“) Beyond the trapping region, the field lines are greatly stretched toward the antisolar direction along the equatorial plane. This is caused by an electric current system in the equatorial plane of the tail region, the tail current (or the neutral sheet current). W) In the equatorial plane, the transition from the trapping region to the tail region occurs where the magnetic field produced by the neutral sheet current (denoted by the N field) is comparable to the magnetic field (G) originating in the solid Earth, (lo) plus the magnetic field produced by current systems located in the trapping region, such as the ring current (R). In the tail region, N varies little with the
AND THE MAGNETOSPHERE
geocentric distance. (11) In this paper, the terms ‘close’ and ‘open’ are also used as defined by Piddington.(5s8) The ‘closed’ field lines are those that cross the equatorial plane in the trapping region of the magnetosphere, and the ‘opened’ field lines are those that close in the magnetospheric tail or may be connected to interplanetary magnetic fields. During geomagentic storms, the aurora1 oval undergoes two basic types of change. The first is the enlargement of the oval as a whole, and both the poleward and the equatorward boundaries of the oval belt shift equatorwards. The second is repeated expansions and subsequent contractions of the width of the oval, particularly in the midnight sector. Undoubtedly, such changes of the oval are a manifestation of large-scale changes of the internal structure of the magnetosphere. The purpose of this paper is to infer the changes of the internal structure of the magnetosphere and resulting changes of the aurora1 oval as a result of the growth of the ring current and the neutral sheet [email protected]
,s) 2. CHANGING
When the Sun is very quiet, the oval contracts from its average location to dipole colatitude 15” or less in the midnight sector and becomes very faint or even invisible,(12s13) However, even weak geomagnetic acitivity (E k, cu 10) is associated with the expansion of the oval from its quiet time location to its average location, so that its midnight radius increases to dp co-lat 23” (that of the aurora1 zone) in the midnight sector.‘12)
ARCS LATITUDE- DSt(H)
0 FIG. 3. RELATION BETWEEN ARCS DURING GEOMAGNETIC
THE LOWER LIMITOFLATITUDE STORMS AND THE MAGNITUDE
ATTAINED BYQUIET MAINPHASEDECREASE @St).
During an intense geomagnetic storm, the oval expands greatly, at times as far as dipole co-lat 40” or dp lat 50” during extremely intense storms. Akasofu and Chapman’14) have shown that the dipole latitude of the midnight portion of the oval depends on the magnitude of the main phase decrease (the Dst values); their diagram is reproduced here as Fig. 3. As expected from Fig. 2, the equatorward shift of the aurora1 oval is associated with the ‘squashing’ of the trapping region toward the equatorial plane. A drastic equatorward shift of the outer boundary of the outer radiation belt during intense magnetic storms was first observed by Maehlum and O‘Brien. (15) Since then a large number of observations have been made, particularly by Williams and Palmer(lQ and Ness and Williams.(17) Williams
and Palmeros) showed that the day-night asymmetry (the eccentricity) of the outer belt (electrons of energies greater than 40 keV) becomes more obvious as the &-index is increased. In addition to the changes in its size and eccentricity, the oval belt rapidly repeats expansions and subsequent contractions of its width, particularly on the dark sector. The expansion and the contraction occur during the expansive phase and the recovery phase of the aurora1 substorm, respectively. In the midnight sector, during the expansive phase, quiet aurora1 arcs lying in the narrow oval become bright first and advance rapidly polewards, resulting in an explosive expansion of the oval width.([email protected]
‘) This initiates a planetary-scale activity of auroras that lie in the other sector of the oval. In the evening sector of the oval, a surge-type motion of auroras (the westward traveling surge) is generated by the expansion which travels along pre-existing arcs, namely along the oval towards the evening sector and sometimes to as far as the afternoon sector. In the morning sector, particularly to the equatorward side of the oval, arcs disintegrate into patches, and the resulting patches drift rapidly eastward. In this way, the whole aurora1 system in the oval is progressively activated from its midnight portion. The first expansive phase occurs in an explosive manner and lasts only for 5 - 30 min, but the recovery phase progresses much more slowly and lasts for 1 - 3 hr. The lifetime of this transient phenomenon, the aurora1 substorm, is of order 1 - 3 hr. 3. CHANGING INTERNAI, STRUCTURE OF THE MAGNETOSPHERE
The internal structure of the magnetosphere is expected to change greatly during geomagnetic storms when both the ring current and the neutral sheet current grow and vary. The major part of the ring current is now inferred to be located at about a geocentric distance of r, = 3a (a = the earth radius). t20) In this section we show that the growth and decay of the ring current and the neutral sheet current plays an important role in determining the location of the aurora1 oval in the midnight sector. During geomagnetic storms, the magnetic field intensity B at a point in the magnetosphere can be expressed by the combination of the magnetic field originating in the solid Earth (denoted by G) and the magnetic fields generated by the magnetospheric boundary current (C > 0), the ring current (R $ 0), the neutral sheet current (N 9 0), and the polar electro-jet (P 3 0). We note, however, that BE G on the Earth’s surface and Brsi G + C + R + Nin the equatoriaI plane of the magnetosphere. t21) Further, the C field becomes important only when both the R and Nfields are small, so that it is not taken into account. The first approximation (B _N G) gives rise to an error of less than 1 per cent in the following magnetic flux calculation. The consequence of the second approximation will be discussed later. Consider two equatorial radial lines (OA, OB) separated by longitude k$ dl from the Sun-Earth line in the midnight sector; the Earth’s center is denoted by 0. The two radial lines intersect the boundary of the trapping region at A and B, respectively, and also the Earth’s surface at A’ and B’, respectively. Denoting the dipole pole N, consider also a half sectorial area NA’B’ on the Earth’s surface; let A” and B” be the intersection points of the lines NA’ and NB’ with a dipole latitude circle 4. At a point in the equatorial fan-shaped area (bounded by the two radial lines, the solid Earth and the boundary of the trapping region, AA’B’B),
W-,)N We) + We) + Nr,)
and the total magnetic flux Fe across the area AA’B’B is given by I;=/BdS=f/;Br,dr,dl
AND THE MAGNETOSPHERE
where dS denotes an elementary area and rT the geocentric distance of the trapping boundary in the midnight sector. For a given set of the values of G(r,), R(r,) and N(r,), Fe can be estimated. The latitude C$of the intersection between the ionosphere and the outer boundary of the outer belt can then be determined by using the magnetic conservation theorem(22~23) by equating
with F,, where dS’ denotes an elementary area in the half sectorial area NA’B’, B, the radial component of B on the Earth’s surface and F, the magnetic flux from the area A’B’B”A” on
FIG. 4. RELATIONBETWEENT~IEMINIMUMLATITUDEOPENEDBYTHE N-R INTERACXONANDTHE INTENSITY OF THE RING CURRENT FIELD(R) FOR DIFFERENT INTENSITIESOF THE NFBTRAL SHEET FIELD (N).
the Earth’s surface. The latitude thus determined gives the minimum latitude of the field lines which are opened by the growth of both the ring current and the neutral sheet current or the latitude which is exposed to energetic plasma particles from the neutral sheet (see Fig. 2). Figure 4 shows the latitude rj as a function of the ring current intensity ]RI, taking the intensity of the neutral sheet current field INI as the parameter. The function G(r,) is taken to be a dipole field, namely Bo/(re/u)3, with B,, = O-32 G. The parameters which specifiy R(r,) are given by the geocentric distance of the center line of the ring current belt (reo = 1.5a), the two constants determing the particle distribution inside and outside r&g1 = 2.990 and g, = 0.499, respectively) and the constant determining the pitch-angle distribution of the ring current particles (CC= 2-O).W) The kinetic energy density of the ring current is expressed in terms of a measurable quantity R at the equator on the Earth’s surface (r, = a). The function iV(r,) cannot be determined accurately unless the two dimensional configuration of the neutral sheet is known. Taylor and Hones (lo) took a linear decrease of N(r,) from the trapping boundary to the Earth’s center by assuming that the tail current exists beyond the
trapping region in the night sector. But since the neutral sheet is likely to extend to the day sector, surrounding the trapping region (Fig. 2), the gradient of N toward the Earth may be less than theirs. In this paper, N is assumed to be constant; the results are not, however, greatly affected by the functional form of N(r,) because Gfr,) increases rapidly toward the Earth. Figure 4 shows that the neutral sheet of intensity INI = 50 y alone can open the field lines beyond dp lat 70”, but the growth of the ring current increases considerably the opening efficiency; for the ring current of intensity IRl = 300 y, the field lines beyond dp lat 53” 37’ can be opened and thus this latitude can be exposed to the hot plasma from the neutral sheet. For an extreme case of the ring current intensity IR] = 300 y and the neutral sheet intensity INI = 90 y, dp lat 47“ 30’ can be exposed to the hot plasma from the neutral sheet. This latitude is little less than the minimum dp latitude attained by quiet aurora1 arcs during the IGY (Fig. 3). The expansion of the oval from the quiet time location (>dp lat 74’) to the average location (dp lat 67’) may also be caused by other processes such as those proposed by Taylor and Hones(i”) or by William and Mead. (=) However, for the expansion from its average location, the growth of both R and N seems to play a major role. As mentioned already, we have ignored C(r,) in our calculation, and it is not known accurately for the ma~etosphere with an extended tail. However, if Mead’s calculation can be used as a first approximation, it is fairly uniform over an extensive region beyond r, = 7a. Therefore, we may regard the parameter N to be the combination of N and C, namely @I + jCl>; note, however, N < 0 and C > 0 at r, < reT. Therefore, when accurate values of C become available, Fig. 4 can be used without redrawing the curves by replacing INI by (IiVI+ ICI). For a quiet situation, Cis of the order + 10 y. Inside re = 7a, it becomes less important since G(r,) increases rapidly toward the Earth. Besides the equatorward expansion of the aurora1 oval as a whole, the poleward boundary of the oval belt repeats poleward motions and subsequent equatorward motions, namely the aurora1 substorm. Figure 5 shows the poleward motion of the poleward boundary of the oval during an exceptionally intense explosive phase of one of the aurora1 substorms recorded during the great geomagnetic storm of February 11, 1958. The expansion began almost simultaneously over the entire region in the dark sector along dp lat 48” - 50”, and the poleward boundary (where brightest aurora1 bands were seen) reached as high as dp lat 71” at the maximum epoch of the substorm; the region swept by the poleward boundary was covered by patchy auroras. Suppose that the intensity of the neutral sheet is suddenly decreased. As is clear from Fig. 4, a decrease of the intensity of the neutral sheet must indicate an increase of the minimum open latitude r#. In terms of the configuration of the magnetic field lines, this corresponds to a change of the field lines from a greatly stretched situation to a more or less contracted dipolar situation. For the ring current intensity IRI = lOOy, a decrease of the neutral sheet intensity from INI = 507 to 107 can cause the closure of the field lines which anchor between dp lat 57” and 68”. Therefore, as soon as the neutral sheet intensity begins to decrease, the region which is exposed to the neutral sheet shifts polewards from dp lat 57” to 68”. In fact, during geomagnetic storms of a medium intensity (1~1N 100 - 150 y), the aurora1 oval descends from its average location (dp lat 67”) to about dp lat 57” - 60’; then during aurora1 substorms, the poleward boundary of the oval moves rapidly to about dp lat 70°,(1**1g)Brightest aurora1 bands are seen at the advancing boundary, and the region swept by such bands is covered by irregular bands or patches. The substorm shown in
OF THE WIDTH OF THE GREAT
AND THE MAGNETOSPHERE
OF THE AURORAL OVAL DURING ONE OF THE AURORAL GEOMAGNETIC STORM OF FEBRUARY 11, 1958.
Fig. 5 could be explained if the intensity of the neutral sheet decreases from INI = 80 y to INI = 10 y (or a little less) and if IRI is of order 300 y. In fact, such a drastic substorm occurs only during greatest geomagnetic storms.(26) It is interesting to note in this connection that Behannon and Nesst2’) observed sharp decreases of the magnetic field intensity in the tail region of the magnetosphere, when intense polar magnetic substorms were observed at College. After reaching the nothernmost latitude (in the Northern Hemisphere), the bright bands begin to return towards their initial location, the recovery phase. This phase proceeds much more slowly and lasts from 1 - 3 hr. Since the region which is exposed to the neutral sheet is shifting towards lower latitudes, the neutral sheet current must be increasing its intensity during this phase. 5
In constructing Fig. 4, we have used the flux conservation theorem and have not considered motions of the magnetospheric plasma associated with the changes of the ring current and the neutral sheet current intensity. Clearly, in the equatorial plane, the magnetospheric plasma will have radial motions with the magnetic field lines which are stretched or contracted by the growth and decay of the neutral sheet current and the ring current. In other words, the changing configuration of the geomagnetic field lines outside the trapping region does not ‘propagate’ with velocity of light, but with veIocity of hydroma~etic waves VA; the plasma in the neutral sheet moves also with this speed. Since N is almost constant outside the trapping region, we may take INI N 50 y (= 5 x lO+ G) as an example. If the number density of the plasma is of order 10/cm3, V’ = N/2/(47rp) N 250 km/set; here p denotes the mass density of the plasma. Therefore, for the field line, which is stretched to a distance of 2Oa from the trapping boundary, it will take of the order of 20a/Va N 250 set to restore its more or less dipolar situation.
Although far more detailed studies on the physics of the neutral sheet and the ring current are needed, it is likely that the growth of the ring current and the neutral sheet undoubtedly plays a vital role in determining the internal structure of the magnetosphere and the auroral oval during geomagnetic storms. In turn, the changing auroral oval can be considered to be a manifestation of the changing internal structure of the magnetosphere. A simultaneous study of the auroral substorm occurring in the ionosphere and magnetic variations in the tail region is needed for a further detailed study. We have not discussed specifically the mechanism that causes the decrease of the tail field during the substorm. One of the possible mechanisms is the tearing instability in the neutral sheet which can convert the magnetic energy to thermal energy thereby proving both the energization of the plasma and the decrease of the magnetic [email protected]
@ Acknowledgements-I would like to express my thanks to Dr. S. Chapman and Dr. J. A. Van Allen for their discussions during the preparation of this paper. I would also like to thank Dr. J. H. Piddington, Dr. R. N. Dewitt, and a referee for their critical comments on an early version of the manuscript. The work reported here is supported in part by grants from the National Aeronautics and Space Adminis~ation (NsG 201-62 and NsG 233-62) and the National Science Foundation (GP 2721). REFRRRNCES 1. S.-I. AKASOFU, S. CHAPMANand S.-C. MENG, J. atmos. ferr. Phys. 27, 1275 (196.5). Y. I. FELJXTEIN,Geomag. & Aeronomy 3, 183 (1963). 0. V. &IOROSHEVA,Geomag. & Aeronomy 2,696 (1962). L. A. FRANK, J. A. VAN ALLENand J. D. CRAVEN,J. geophys. Res. 69, 3155 (1964).
2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
J. H. PIDDINGTON,Planet. Space Sci. 13,565 (1965). T. A. Fnrrz and D. A. GURNE’IT,J. geophys. Res. 70,2485 (1965). L. A. FRANK and J. A. VAN ALLEN,Research ingeophysics, Vol. I, pp. 161-187 (Ed. H. Odishaw) (1964). J. H. PIDDINGTON,J. geophys. Res. 65,93 (1960). See also Planet. Space Sci. 11,1277 (1963). W. I. AxFORD, H. E. PETXXEK and G. L. Srscoe, J. geophys. Res. 70,123l (1965). H. E. TAYLORand E. W. HONES,J.geophys. Res. 70,3605 (1965). N. F. Nsss, J. geophys. Res. 70,2989 (1965). S.-I. AKA~OFU,J. utmos. ferr. Phys. 26, 1167 (1964). W. J. STRINGER,A. B. BELONand S.-I. AK~~OFU,.T. atmos. terr. Phys. 27, 1039 (1965). S.-I. AKASOFUand S. CHAPMAN,J. atmos. terr. Phys. 25,9 (1963). B. MAENLUMand B. J. O’BRIEN,J. geophys. Res. 68,997 (1963). D. J. WILLIAMSand W. F. PALM!ZR, J. geophys. Res. 70,557 (1965). N. F. NESSand D. J. WILLIAMS,Pub. Goddard Space Fight Center, NASA, Jury (1965). S.-I. AKASOFU,Planet. Space Sci. 12, 273 (1964).
19. 20. 21. 22. 23. 24. 25. 26.
S.-I. AKASOFU, Space Sci. Rev. 4,498 (1965). L. J. CAHILL and D. H. BAILEY, Trans. Am.geophys. Un. 46, 116 (1965). S. CHAPMANand J. BARTELS, Geomagnetism. Oxford University Press (1940). A. J. DEFILER and R. KARPLUS, J. geophys. Res. 66,2289 (1961). S.-I. AKASOFU, Space Sci. Rev. 2,91 (1963). D. J. WILLIAMSand G. D. MEAD, J. geophys. Res. 70, 3017 (1965). G. D. MEAD, J.geophys. Res. 69,llSl (1964.) S.-I. AKA~~FU and S. CHAPMAN,J. atmos. terr. Phys. 24, 735, (1962). 27. K. W. BEHANNONand N. F. NESS, Pub. Goddard Space Flight Center, NASA, October (1965). 28. B. COPPI, G. LAVAL and R. PJSTTAL, International Center for Theoretical Physics Pub., Trieste (1965). Pew&we-Ecnn
HetiTpaJIbHbItiIIJIaCTRBJIReTCH CJIOeM IIOpOW~eHEiR BbI3BaHHbIX CeBepH~IM CmKneM SacTssq, KaK nonaraeT &~~AHH~TOH,TO OHH nporenaror B TOHKOM cnoe IIpJIMO CHapyNi BHeIIIHem rpaHHIJbI patiOHa 3aXBaTa I13 HeiTpanbHOrO IlJlaCTa II0 HanpaBneH~H,K3eMJre~CO3~aroTcBe~eH~eOBaJIbHOt(POpMhIHanOJIFIpHOm~oaoc~epe, OIIO3HaBaeMOrOHaMHBAaHHOti pa6oTeB BHAe BbI3BaHHOrO UOJISIPEbIMCIWIHUl3MOB&ilI&. Ynasbrsaercn, 9~0 pocr II paman KonbqeBoro ToKa, a TaKxe TOK KePTpanbKoro IIJIaCTa,BO BpeMR reOMarHnTHbIX 6ypb,n3MeHmoT BHyTpeHHIOEO CTpyKTypy [email protected]
~OKa3bIBaeTCFIIl3MeHeHIlFIMIlBbI3BaHHOrO EOJIHpHbIM CEiRHHeMOBaJIa, B gaCTHOCTH CABIWOM OBaJIa B CTOpOHy 3KBaTOpa BO BpeMH KpyIIHOrO y6bIBaHm r~aBao~~a3bI~6bIcTporopa3lu~peH~~,a3aTeMcoKpa~eH~~~~p~~bIoBa~a~nep~o~ BbI3BaHHbIXIIOJIRpHbJM CLlRHHeM 6ypb.