Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.kosmofizika.ru/pdf1/ROSTOKER_JASTP99.pdf Äàòà èçìåíåíèÿ: Mon Oct 31 10:44:00 2005 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 00:57:51 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: vela satellite

\
PERGAMON
Journal of Atmospheric and Solar!Terrestrial Physics 50 "0888# 74ï099

The evolving concept of a magnetospheric substorm
Gordon Rostoker
Department of Physics\ University of Alberta\ Edmonton\ Alberta\ Canada T5G 1J0 Received 15 January 0887^ received in revised form 17 August 0887^ accepted 08 October 0887

Abstract A magnetospheric substorm is an episode of energy transport and dissipation in the Earth|s ionosphere and mag! netosphere which takes place in response to a time limited increase in energy input from the solar wind to the magnetosphere[ For the past few decades\ scientists have tried to understand the physical processes which take place that are responsible for the substorm disturbances of the geospace environment[ In this paper\ The development of the substorm concept is reviewed from its origins at the beginning of the 19th century to the present time[ The theoretical framework in which substorm physics is normally presented is then discussed\ and an outline is given of how that framework has changed in recent times[ This paper concludes by posing two questions which need to be answered if further progress is to be made in solving the substorm problem[ ÷ 0888 Elsevier Science Ltd[ All rights reserved[

0[ Introduction Since the beginning of the 19th century\ there has been an ever increasing interest in the origin of the aurora and the often large magnetic _eld perturbations that accompany displays of the the northern and southern lights[ Near the end of the 08th century\ it had already become clear that electromagnetic _elds were involved in the process whereby the auroras were created^ however\ it was the work of the great Norwegian scientist Kristian Birkeland that really marked the start of the modern age of solar!terrestrial studies using the techniques of mathematics and physics[ In his pioneering studies\ Birkeland "0897# recognized that signi_cant electrical currents ~owed in the upper atmosphere in the region of bright auroras[ He further correctly attributed the source of energy to ionized particles coming from the Sun and he carried out detailed modelling studies which dem! onstrated that the current system responsible for the observed magnetic perturbations was of the form shown in Fig[ 0[ The reader will recognize the form of what\ today\ is known as the substorm current wedge[ During the next three decades there was not a great deal of progress in understanding the way in which the Sun provided the energy for auroral disturbances[ In fact\

Tel[] ¦0!392!381!0950^ fax] ¦0!392!381!3145[ E!mail address] rostokerùspace[ualberta[ca[ "G[ Rostoker#

Fig[ 0[ Three dimensional current system proposed for the polar elementary storm by Birkeland "0897#[ The polar elementary storm was later named the polar magnetic substorm by Akasofu "0857# and\ to this day\ is thought to be associated with a three! dimensional current system of the type shown here[

S0253ï5715:88:, ! see front matter ÷ 0888 Elsevier Science Ltd[ All rights reserved PII] S 0 2 5 3 * 5 7 1 5 " 8 7 # 9 9 0 0 8 * 8


75

G[ Rostoker:Journal of Atmospheric and Solar!Terrestrial Physics 50 "0888# 74*099

Fig[ 2[ Magnetic perturbation pattern for high latitude magnetic disturbances obtained by Harang "0835#[ The evening "morning# sector eastward "westward# electrojet produces regions of posi! tive "negative# H!component[ These plots were made with hourly averaged data and hence are dominated by the low frequency component of substorm activity\ although not to the same extent as the equivalent current system of Silsbee and Vestine "cf Fig[ 1#[

Fig[ 1[ Equivalent current system for geomagnetic bays inferred from ground magnetometer data by Silsbee and Vestine "0831#[ The equivalent current vectors are 2!h averages and therefore pertain only to the low frequency component of substorm activity[

it was generally believe that space was a vacuum and that auroral outbursts only took place when ejecta from the Sun arrived at Earth orbit[ Even at the beginning of the third decade of the 19th century\ this concept was still prevalent when Chapman and Ferraro "0820# carried out their study of the interaction of a plasma stream with the terrestrial magnetic _eld[ Nonetheless\ the fact that plasma physics had developed to the stage that it could be used in studies of the SunïEarth environment was an important step forward which was to have a signi_cant impact several decades later[ During the 0829s and 39s\ further progress in under! standing the solar!terrestrial interaction came from ana! lysing data from irregularly distributed arrays of ground based magnetometers[ The work of Silsbee and Vestine "0831# began the era of the use of equivalent current systems to portray the two dimensional distribution of horizontal magnetic perturbation vectors measured at each observing site[ The idea here was that it was assumed that all currents ~owed only in the ionosphere and that a in_nite sheet current approximation could be used to evaluate the strength and direction of the current ~ow at each site[ Thus\ Fig[ 1 from Silsbee and Vestine was obtained by rotating the magnetic perturbation vectors by 89> and drawing the current con_guration that best _tted the data[ From this _gure it was inferred that there was a westward electrojet ~owing in the morning sector auroral region and a weaker eastward electrojet ~owing

in the afternoon sector[ A similar picture was obtained by Harang "0835#\ shown in Fig[ 2\ although here the magnetic perturbation data are shown and the reader is asked to infer that negative "positive# H!component perturbations re~ect the presence of a westward "east! ward# electrojet[ It is extremely important to recognize that the _gure of Harang was obtained using hourly averaged values of the magnetic perturbation vectors while Silsbee and Vestine used 2!h averaged values[ This will have important implications for substorm contro! versies\ as we shall see shortly[

1[ The question of the true nature of the substorm current system The work of Fukushima "0842# began the era of controversy as regards the character of the equivalent current systems which accompanied high latitude auroral activity[ He presented several examples of equivalent cur! rent systems taken at times of geomagnetic bay activity which did not conform to the two cell picture of Silsbee and Vestine and resembled more what one would expect from the three dimensional current system proposed by Birkeland "0897#[ Ultimately\ Sugiura and Heppner "0854# highlighted the problem by asking the question of whether or not geomagnetic bays were best described by a two!cell equivalent current system or a one cell system[ Around this time\ the term\ {substorm\| was coming into use after its introduction by S[ Chapman and S[!I[ Aka! sofu in the early 0859s and its use to describe the auroral breakups near midnight known to occur in conjunction with geomagnetic bays "viz the auroral substorm#[ So it was about this time that arguments began to be made in


G[ Rostoker:Journal of Atmospheric and Solar!Terrestrial Physics 50 "0888# 74*099

76

Fig[ 3[ Schematic diagrams of the two equivalent current systems proposed for the substorm disturbance "after Rostoker\ 0885#[ We now understand that there are disturbances during the time of a substorm\ one of which has a two cell equivalent current system and the other of which features the one cell system[ The one cell system is more appropriate for the shorter period component of the substorm disturbance[

terms of substorm properties rather than the properties of geomagnetic bays[ Sugiura and Heppner claimed that the substorm dis! turbance was best described by a two cell system con! ~icting with the argument by Akasofu et al[ "0855# that it was best described by a one cell system[ The question was answered by Rostoker "0858# who produced evidence to show that\ in fact\ disturbances featuring both types of equivalent current system co!existed during substorm activity[ Figure 3 shows schematically the two proposed equivalent current con_gurations\ while Fig[ 4 shows a set of three magnetograms from northern Scandinavia which demonstrate the coexistence of disturbances which are characterized by those equivalent current systems[ It turns out that the short lived "¨0!h time scale# dis! turbances are best represented by a one!cell equivalent current system while the longer lived "shown by the dashed curve on the Tromso magnetogram# disturbance á is best represented by the two!cell equivalent current system[ This is why the fact that the equivalent currents inferred by Silsbee and Vestine "0831# and later by Harang "0831# were of the two cell type[ Both analyses involved averaging of the data "2 h for Silsbee and Vestine and 0 h for Harang# which would produce results dominated by the lower frequency disturbance[ At the beginning of the 0869s\ substorm research received somewhat of a setback for two separate reasons[ The _rst reason related to the beginning of the serious use of indices of auroral zone activity to study individual events[ Davis and Sugiura "0855# introduced the AE index as a way to track the level of geomagnetic activity on a global basis[ The index involved establishing the upper "AU# and lower "AL# envelopes of the mag! netograms traces of the NorthïSouth components of the disturbance _eld from several stations distributed as uni!

Fig[ 4[ Magnetograms from the high latitude stations of Isfjord "63[4> corrected geomagnetic latitude CGL#\ Bjornoya "69[8> áá CGL# and Tromso "55[29> CGL# lying approximately along a á common geomagnetic meridian "after Rostoker\ 0885#[ The four disturbances visible between ½0729ï1129 UT would be called substorms by most researchers and indeed are characteristic of the polar elementary storm of Birkeland "0897# and the polar magnetic substorm of Akasofu "0857#[ The longer period dis! turbance on which the polar magnetic substorms are riders is best represented by a two cell equivalent current system[


77

G[ Rostoker:Journal of Atmospheric and Solar!Terrestrial Physics 50 "0888# 74*099

Fig[ 5[ Schematic diagram of the envelope of auroral oval H!component disturbances representing a typical AL index variation during a period of substorm activity "solid curve#[ Also shown is an H!component magnetogram from a midnight sector station whose data were used in the construction of this theoretical AL index "dotted line#[ In this case the peak disturbance of the substorm expansive phase portrayed here is slightly smaller than the contribution to the H!component from the directly driven activity which maximizes a a dawn sector station "shown by the dashed curve#[ The midnight sector substorm would not be detected using the AL index in this case[

formly around the world at auroral zone latitudes[ "The value of AE is the sum of the absolute values of AU and AL at any instant of time[# Although AE\ AU or AL are very useful for statistical studies of auroral zone dis! turbances\ researchers began to use the indices in the study of individual events primarily as an indicator of onset times of substorm expansive phases[ In doing so\ they were studying the combined signatures of dis! turbances described by the two!cell and one!cell equi! valent current systems in a situation where it was imposs! ible to decouple the contributions of the two systems[ Since expansive phase e}ects are best described by the one!cell equivalent current system\ increases in the two! cell current system could easily be misinterpreted as expansive phase onsets[ More importantly\ as can be seen from Fig[ 5\ an expansive phase onset could be missed completely if the maximum perturbation associated with that expansive phase was less than the maximum per! turbation associated with the two!cell equivalent current system[ Thus\ in the early years of the use of the AE\ AU and AL indices it was often not possible to establish from studies that utilized these indices which increases of the index were due to expansive phases of substorms and which were due to the disturbance which featured a two! cell equivalent current system[

A second reason for the setback came from the estab! lishment by McPherron et al[ "0862# of the substorm current wedge as a real "rather than equivalent# current system[ This approach succeeded in drawing attention to the possible physical mechanisms which might be respon! sible for the substorm expansive phase\ but had the unfor! tunate side e}ect of leaving to the side the other com! ponent of substorm activity\ viz\ the disturbance responsible for the two!cell equivalent current system[ However\ the concentration on the expansive phase dis! turbance that was best represented by the substorm cur! rent wedge led to an enhanced search for knowledge about the phenomenology of that component of sub! storm activity[ During this time it was discovered that the expansive phase current system was not a monolithic large scale current system\ but rather that it was a super! position of more local current elements which expanded in a step!wise fashion both poleward "Kisabeth and Ros! toker\ 0863# and westward "Wiens and Rostoker\ 0864# as the substorm developed[ The importance of the two cell equivalent current sys! tem re!emerged in the late 0869s in the work of Perreault and Akasofu "0867# who studied the response of the magnetosphere to solar wind input as quanti_ed by the o parameter de_ned by


G[ Rostoker:Journal of Atmospheric and Solar!Terrestrial Physics 50 "0888# 74*099
1 o 09vsB1 sin3"u:1#

78

where vs is the solar wind speed\ B is the magnitude of the interplanetary magnetic _eld "IMF#\ u is the polar angle measured from the northward geomagnetic axis of the component of the IMF normal to the SunïEarth line and 09 "½6 RE# is a constant with the dimensions of distance[ The comparison of o"t# with the energy dis! sipation in the magnetosphere!ionosphere system U"t# led Akasofu "0879# to argue that a large portion of the energy entering the magnetosphere from the solar wind was dissipated in what he called directly driven activity[ It turned out that his directly driven activity was mani! fested in the two!cell equivalent current system observed during substorm activity\ and once again some attention was paid to the large scale electrojet currents that ~ow in the auroral oval "i[e[\ the eastward electrojet in the eve! ning sector and the westward electrojet in the morning sector#[ Later\ Clauer et al[ "0872# used linear prediction _ltering to evaluate the impulse response of the mag! netosphere and from this work discovered that a sub! stantial fraction of the variance in the AL index "30)# was related to the direct input of solar wind energy "viz directly driven activity#[ From this research in the late 0869s and early 79s it was established that directly driven activity constituted a signi_cant portion of substorm energy dissipation[ However\ the ratio of energy dis! sipation in directly driven activity to that in the release process varies from event to event and from one time in a given event to another time in the same event[ Exactly what determines the proportions is not fully understood at this time[ To conclude this section\ we re~ect on the de_nitions of the various phases of substorm activity trying to relate the original de_nitions to how we presently view the substorm phenomenon[ The original de_nition of a sub! storm by Akasofu "0853# involved an expansive phase in which the auroras moved poleward and a recovery phase in which they drifted equatorward to their pre!expansive phase location[ Subsequently\ McPherron "0869# intro! duced the concept of a growth phase during which energy was stored in the magnetotail to be released sometime later in the expansive phase[ While the growth phase concept introduced by McPherron is accepted to this day\ the actual signatures "which were disturbances measured by ground based magnetometers# were not actually sig! natures of storage of energy in the tail but rather sig! natures of the growth of the directly driven electrojets[ Fortunately\ an increase in the energy input from the solar wind into the magnetosphere most often leads to both an increase in directly driven activity and storage of the energy in the magnetotail[ Thus the ground based magnetic signatures identi_ed by McPherron were indeed e}ective proxy measures for the storage of energy in the tail\ viz\ the growth phase[ Finally\ we should note that the concept of recovery phase as it was originally de_ned

by Akasofu "0853# really pertained to the local behaviour of the auroral arcs during a substorm "i[e[\ the devel! opment of the auroras in the _eld of view of an allsky camera# as shown in Fig[ 6[ Here recovery was under! stood as the period of time after the substorm disturbed region had expanded to its maximum poleward position till the arcs had drifted equatorward to their pre!sub! storm position[ In the more modern global view of the magnetospheric substorm shown in Fig[ 7\ one sees that growth is characterized by an equatorward expansion of the auroral oval while recovery is characterized by the contraction of the auroral oval to its original position "i[e[\ recovery on a global scale is characterized by pole! ward motion in contrast to recovery on a local scale which is characterized by equatorward motion#[ The auroral substorm as de_ned by Akasofu "0853# best describes the latitudinally and longitudinally localized regions of auroral enhancements that have surgelike form and which expand poleward and then die out in a relatively short time "½04ï29 min# compared to the lifetime of a typical magnetospheric substorm[ In fact\ on a global scale equatorward motion of arcs is more likely to be associated with growth in the sense of more energy enter! ing the magnetosphere from the solar wind and being stored in the magnetotail[

2[ Role of the magnetotail in the substorm In the previous section we have concentrated on the ionospheric signatures of substorm activity that domi! nated early research in this subject area[ In recent times\ research has concentrated on the physical mechanisms responsible for the various disturbances observed in the time frame of the magnetospheric substorm[ It is clear that the magnetic _eld lines threading the ionosphere "where auroral substorm disturbances are observed# map into the magnetotail[ Therefore\ a great deal of e}ort has been expended in trying to de_ne the behaviour of the particles and _elds in the near!Earth\ middle and distant magnetotail where the source regions for the observed ionospheric disturbances are likely to be located[ In this section we shall explore the evolution of the the near! Earth neutral line "NENL# model for substorms which is thought\ by most members of the substorm community\ to be the most likely framework in which to develop an understanding of the substorm process[ The present day model for substorms owes its origin to the proposal by Dungey "0850# that energy could be transferred from the interplanetary medium to the mag! netosphere through a process in which the solar wind magnetic _eld merged with the terrestrial magnetic _eld at the dayside magnetopause when there was a com! ponent of the IMF anti!parallel to the Earth|s magnetic _eld "i[e[\ when the IMF Bz component was approxi! mately southward#[ Following the dayside merging\ the


89

G[ Rostoker:Journal of Atmospheric and Solar!Terrestrial Physics 50 "0888# 74*099

Fig[ 6[ The auroral substorm as inferred from allsky camera data by Akasofu "0853#[ The expansive phase involves poleward motion of the region of auroral disturbance while the recovery phase involves equatorward motion of the remaining arc structures[

_eld lines would then be transported over the poles and would reconnect in the lee of the Earth with the recon! nected _eld lines contracting back towards the Earth as shown in Fig[ 8[ Camidge and Rostoker "0869# used IMP A and IMP B magnetometer data to show the response of the mag! netotail to substorm activity and reached the conclusion that substorms were associated with the formation of a neutral line normally somewhere beyond ½ -10 RE[ Rostoker and Camidge "0860# further concluded that

the disturbed region of the tail only occupied a limited azimuthal extent for each substorm intensi_cation[ Coroniti and Kennel "0861# explored the concept of the association of substorms with tail reconnection in more detail and reached the conclusion that there would nor! mally be an imbalance between the frontside merging rate and the reconnection rate in the magnetotail[ To begin with\ the frontside merging rate would exceed the tail reconnection rate and magnetic ~ux would pile up in the magnetotail[ From time to time there would be sud!


G[ Rostoker:Journal of Atmospheric and Solar!Terrestrial Physics 50 "0888# 74*099

80

Fig[ 7[ Global view of auroral oval evolution during a magnetospheric substorm "after Rostoker\ 0885#[ Storage of tail energy "i[e[\ growth phase# involves an expansion of the polar cap and a resultant equatorward shift of the auroral oval[ Release of tail energy "i[e[\ recovery phase# involves a poleward shift of the auroral oval poleward border[ The substorm expansive phase involves the appearance of localized bright auroral arc structures near midnight\ starting at the equatorward edge of the auroral oval and progressing poleward during the course of the event[ Expansive phase activity concludes with auroral arc activations at the poleward edge of the contracted oval[

Fig[ 8[ The recon_guration of magnetic _eld lines associated with frontside merging and nightside reconnection as proposed by Dungey "0850#[ In this original concept\ there is only one neutral line in the magnetotail[


81

G[ Rostoker:Journal of Atmospheric and Solar!Terrestrial Physics 50 "0888# 74*099

Fig[ 09[ The substorm current wedge "after McPherron et al[\ 0862#[ This _gure shows the substorm current system as a real "rather than equivalent# current system and suggests the physical processes which might go on in the tail during a time of substorm activity[

den bursts of reconnection in the tail during which the reconnection rate would exceed the frontside merging rate and magnetic ~ux would be returned to the dayside[ These bursts were thought to be the tail signatures of substorm expansive phases[ McPherron et al[ "0862# used OGO 4 satellite data taken in the magnetotail to support the Coroniti and Kennel idea of imbalance between dayside merging and nightside reconnection described above[ They further suggested that the formation of a neutral line relatively close to the Earth was associated with the formation of the substorm current wedge as shown in Fig[ 09\ with the reduction in crosstail current being due to shortcircuiting through the midnight sector ionosphere[ During the late 0859s and early 69s\ Hones and col! leagues at the Los Alamos National Laboratory carried out a series of studies of the characteristics of magnetotail particles using detectors aboard the VELA satellites whose ½07 RE circular orbits allowed them to sample regions inside ½-19 RE behind the Earth[ Those studies led to the suggestion by Hones "0865# that magnetic reconnection was occurring at almost all times in the distant magnetotail but that\ at the time of substorm onset\ a new neutral line formed closer to the Earth[ He termed this the near!Earth neutral line "NENL#\ and thus was born the NENL model of magnetospheric substorms[ The general character of substorm devel! opment in the magnetotail is shown in Fig[ 00[ The essence of the NENL paradigm as espoused in the ½19 years that passed since it was proposed can be summarized using the numbers beside each of the panels in Fig[ 00 that depict the evolution of the magnetotail during a substorm[ Panel 0 shows a quiescent magnetotail with reconnection occurring at some distant neutral line thought "at that time# to lie several tens of RE behind the

Earth[ An increase in energy input into the mag! netosphere sets in motion a series of events which alters the topology of the magnetotail[ The addition of energy to the magnetotail involves a thinning of the plasma sheet ultimately leading to the start of reconnection at a new near!Earth neutral line as shown in panel 1[ Recon! nection of closed _eld lines proceeds until\ as shown in panel 5\ the reconnection of open _eld lines commences[ That moment was viewed as marking the onset of the substorm expansive phase "the time from the start of increased energy input to the magnetosphere to the time at which open _eld lines begin to reconnecting being considered as the growth phase#[ A large blob of plasma "termed the plasmoid# was then disconnected from the Earth and began to move downtail at high speed leaving behind a thin plasma sheet "panels 6ï8#[ The recovery phase of the substorm was marked by a thickening plasma sheet\ the thickening proceeding downtail as recovery continues "panel 09#[ The presence of the ISEE 2 satellite in the distant magnetotail during the early 0879s provided an opportunity to study the charac! teristics of the {plasmoid| far downtail "cf Slavin et al[\ 0874#[ Figure 01 shows a {classic| plasmoid reported in a more detailed study by Slavin et al[ "0878#[ It carries with it the canonical signatures of a bipolar Bz magnetic _eld perturbation "_rst positive and then negative# and tail! ward plasma ~ow[ In the early 0889s\ as additional evidence began to be gathered during the Solar Terrestrial Energy Program "STEP# period using both ground based and satellite borne detectors\ a revised picture of the NENL model emerged[ For example\ both Lui "0880# and Rostoker "0880# pointed to key observational features in particle and _eld measurements made in space which either were inconsistent with the NENL model or which provided


G[ Rostoker:Journal of Atmospheric and Solar!Terrestrial Physics 50 "0888# 74*099

82

Fig[ 00[ The development of a magnetospheric substorm in the magnetotail according to the near!Earth neutral line hypothesis "after Hones\ 0873#[

Fig[ 01[ Magnetic _eld variations associated with a structure de_ned as a plasmoid which was detected ½199 RE behind the Earth by the ISEE 2 satellite "after Slavin et al[\ 0878#[ The variation in the u angle between ½9719ï9899 UT indicates a positive Bz perturbation followed by a negative perturbation[ The plasma ~ows tailward at ½399 km:s during this transient disturbance[ It is important to note that the magnetic _eld is strongest in the center of the structure contrary to what one would expect for a plasmoid as originally conceived but consistent with that expected for a ~ux rope[


83

G[ Rostoker:Journal of Atmospheric and Solar!Terrestrial Physics 50 "0888# 74*099

compelling evidence for the initiation of expansive phase onset at a near!Earth site[ However\ arguably the most important discovery in the early 0889s that impacted the NENL model was the de_nitive placement of the region of onset within the nightside magnetosphere in the equa! torial plane by Samson et al[ "0881a\ b#[ Until then\ researchers had thought that the near!Earth neutral line was threaded by magnetic _eld lines that mapped to the region of substorm expansive phase onset in the high latitude ionosphere[ "That explains why substorm expansive phase onset was considered to be marked by the start of reconnection of open "lobe# _eld lines\ since the energy required for the expansive phase ionospheric disturbance was thought to be supplied by the lobe mag! netic _eld through the reconnection process[# While some researchers "e[g[ Lopez et al[\ 0889^ Lui et al[\ 0880# had argued that expansive phase onset occurred in the near! Earth plasma sheet\ the NENL model in its original form continued to hold sway in the community until the obser! vations by Samson and colleagues made the NENL hypothesis in its original form untenable[ The _nal proof came from two di}erent sets of observations[ First of all Samson et al[ "0881a# were able to show a substorm expansive phase which was on _eld lines equatorward of a region of _eld line resonance that clearly lay on dipolar "or at the least quasi!dipolar# _eld lines characteristic of the inner edge of the plasma sheet[ In a second study\ Samson et al[ "0881b# showed that the breakup arcs of a substorm expansive phase were located in virtually the same volume of space as Hb emissions associated with precipitating energetic protons[ Since protons of that energy "tens of keV# are normally only found near the inner edge of the plasma sheet\ this implied that substorm onset must take place in that volume of space well earth! ward of the region where it was thought that a near! Earth neutral line might be formed "¨-01 RE#[ Baker et al[ "0885# recently presented a revised frame! work for the NENL hypothesis in which it is acknowl! edged that the substorm expansive phase onset lies well earthward of the near!Earth neutral line position[ Based on this change in the model\ they contend that substorm expansive phase onset actually starts at the time of recon! nection of closed _eld lines "cf panel 1 of Fig[ 00#[ Hence\ the expansive phase onset is no longer marked by the start of reconnection of open "lobe# _eld lines and accord! ingly the release of the plasmoid is no longer to be associ! ated with the substorm expansive phase onset[ A second aspect of the original NENL model "which is now being re!evaluated thanks to the better data sets acquired by satellites orbiting in the magnetotail during the STEP period# is the plasmoid itself[ Originally the plasmoid was viewed as a closed loop magnetic _eld structure containing hot plasma sheet particles[ However\ this view carried with it the presumption that the mag! netic _eld would be weakest in the center of the plasmoid[ Observations of most magnetic _eld structures described

as plasmoids do not conform to that view[ Even in the {classic| plasmoid shown in Fig[ 01\ it is clear that the magnetic _eld is in fact a maximum in the center of the plasmoid[ More recently this fact has been acknowledged and the combination of bipolar Bz and tailward ~ow is referred to more as a ~ux rope[ A ~ux rope involves a current ~owing across the tail within the magnetic _eld structure giving a crosstail core magnetic _eld compon! ent[ Introducing this concept has created more questions than it has answered[ If current ~ows across the tail\ does it close along the magnetopause boundary "and hence stretch across the entire tail# or does it close through _eld aligned currents which are connected by transverse currents ~owing in the high latitude ionosphere as sug! gested by Kivelson et al[ "0885#< This\ in turn\ begs the question of whether or not ~ux ropes are azimuthally localized giving them a property attributed to the original plasmoid[ If the ~ux rope:plasmoid is azimuthally local! ized\ then we must envision a channel of high speed anti! sunward ~ow near the center of the tail ~anked on both sides by slow speed earthward ~ow as shown in Fig[ 02[ This\ in turn speaks of two regions of very high velocity shear at the interfaces between earthward and tailward ~ow that must have associated with them _eld!aligned currents ~owing into and out of the ionosphere[ The resulting three dimensional current loop involving east! ward closure current in the ionosphere should be detect! able in the region just poleward of the substorm disturbed auroras in the midnight sector\ and identi_cation of such a current loop would be strong evidence for the proposed role of ~ux ropes:plasmoids in the substorm process[ If the NENL model is to explain substorms\ it must satisfy the observational constraints in the near!Earth plasma sheet[ Two of these pertain to the behaviour of the magnetic _eld in and around geostationary orbit[ The _rst relates to the radial current density pro_le near the inner edge of the crosstail current sheet[ It was found by Kaufmann "0876# that\ in order to reproduce the stret! ching of the near!Earth tail magnetic _eld near geo! stationary orbit characteristic of many substorm growth phases\ it was necessary to have an extremely high cros! stail current density "½299 mA:m# over a limited radial range[ Any substorm model has to be able to explain how this current density can gradually build up during the substorm growth phase[ The second constraint relates to the fact that\ a minute or two before the onset of an expansive phase there is a sudden explosive stretching of the tail magnetic _eld in the midnight sector "Ohtani et al[\ 0881# as shown in Fig[ 03[ Any substorm model must be able to explain this apparently unstable behaviour\ and the explanation must bring into play both the source of energy for the sudden enhancement of near!Earth cros! stail current and the the mechanism for its release in the expansive phase process[ At the present time\ some form of ballooning instability in the presence of either velocity shear "Voronkov et al[\ 0886# or a thin current sheet "A[


G[ Rostoker:Journal of Atmospheric and Solar!Terrestrial Physics 50 "0888# 74*099

84

Fig[ 02[ Plasma ~ows expected in the central plasma sheet "indicated by open circles# if there is a fast anti!sunward ~ow tailward of a near!Earth neutral line "after Rostoker\ 0880# during substorm expansive phase activity[ The velocity shear zones should be associated with _eld aligned currents which must close in the high latitude ionosphere[

Bhatacharjee\ private communication# appears attractive as a trigger for the substorm expansive phase[ In terms of how the revised NENL model explains the development of the expansive phase onset near the inner edge of the plasma sheet\ the recent observations by Ang! elopoulos et al[ "0881# provide the observational jus! ti_cation for the present view of this issue[ Angelopoulos and colleagues\ using AMPTE particle and _eld measure! ments in the region earthward of ½-08 RE\ established the presence of fast short lived earthward ~ows which they termed bursty bulk ~ows "BBFs#[ These ~ows typically last no more than a few seconds and feature speeds of hundreds of km:s against the slow background ~ow of a few tens of km:s[ These fast ~ows had also

been noted by Baumjohann et al[ "0878# using the same AMPTE data set\ and they established that almost all fast ~ows inside ½08 RE were earthward[ This obser! vation was extremely important in that it suggested that any near!Earth neutral line must normally lie outside the AMPTE apogee of ½-08 RE if\ indeed\ the source of the fast ~ows was reconnection at a neutral line[ The review by Baker et al[ "0885# of the revised version of the NENL model does recognize the fact that any near!Earth neutral line must lie outside ½-19 RE\ and in doing so calls for some way in which reconnection at that neutral line can contribute to the substorm expansive phase process[ Recent considerations of that matter "R[L[ McPherron\ private communication# reveal that NENL


85

G[ Rostoker:Journal of Atmospheric and Solar!Terrestrial Physics 50 "0888# 74*099

3[ Problems to be solved in the search for the correct framework in which to describe the substorm process In the _nal portion of this paper we point out a two areas in which further progress must be made before a satisfactory and minimally non!unique explanation of the substorm can be achieved[ Question Õ0] Is there more than one neutral line in the magnetotail\ and if so where are the two neutral lines located< To put this question in context\ one must recognize that the original Dungey version of the solar wind : magnetosphere interaction involved merging on the day! side and only one neutral line on the nightside located ½099 RE behind the Earth[ The proposal by Hones "0865# that there was a near!Earth neutral line activated during the substorm expansive phase was based largely on observations of anti!sunward plasma ~ow by the VELA satellites which orbited around -07 RE behind the Earth[ The anti!sunward ~ows were attributed to neutral lines which lay between the satellite and the Earth[ However\ it is now recognized that the near!Earth neutral line rarely lies that close to the Earth and hence the tailward ~ows which on which the NENL hypothesis was founded may owe their origin to some other e}ect[ Modern studies of the response of the magnetotail during substorm expansive phase activity "e[g[\ Nagai et al[\ 0887# seem to indicate that the near!Earth neutral line lies somewhere between ½-19 RE and -29 RE behind the Earth[ This is not far from the radial distance at which Frank and Paterson "0883# claim that the average ~ow in the plasma sheet changes from earthward to tail! ward[ One is then tempted to ask the question of whether or not there might be only one neutral line whose average position is much closer to the Earth than the original conjecture of Dungey would imply[ This question is made all the more complex by the claim of Nishida et al[ "0885# that there are two neutral lines\ one lying near the pos! ition at which Frank and Paterson claim that the average ~ow direction in the plasma sheet reverses and one located at ½-049 RE[ The distant neutral line position suggested by Nishida et al[ is well tailward of previous estimates of the location of that important region[ Thus there is enough uncertainty in proposed positions of the neutral line or neutral lines over the history of mag! netotail observations to make the question of whether or not there are one or two neutral lines active in the tail during substorms worth vigorously investigating with our increasingly better data sets[ Question 1] How can one map from the ionosphere to the magnetotail so as to be able to identify source regions in the magnetotail for auroral oval disturbances< One of the great di.culties in studying the structure and dynamics of the magnetotail centers on the fact that most observations of plasmas and _elds in that region of

Fig[ 03[ Explosive growth of the crosstail current detected by the AMPTE satellite ½1 RE outside geostationary orbit just in advance of a substorm expansive phase which takes place at ½0416]29 UT on 02 June 0874 "after Ohtani et al[\ 0881#[

proponents consider that the substorm expansive phase onset is caused by reconnection at a newly formed neutral line typically located between ½ 19ï29 RE downtail[ BBFs from this neutral line bring energy to the near! Earth plasma sheet quite rapidly in the form of kinetic energy of the earthward ~owing ions[ Braking of these BBFs results in the ~ow of the electric currents associated with the substorm current wedge "cf Shiokawa et al[\ 0886^ Slavin et al[\ 0886#[ In order to provide su.cient energy for the substorm process in the near!Earth plasma sheet\ a signi_cant volume of the center of the magneto! tail must be _lled with BBFs[ Future multi!satellite studies in the near!Earth plasma sheet will determine whether or not that energy constraint is met[


G[ Rostoker:Journal of Atmospheric and Solar!Terrestrial Physics 50 "0888# 74*099

86

Fig[ 04[ Magnetic perturbation in the EastïWest direction typi! cal of that which would have been measured by the polar orbiting satellite TRIAD as it crosses the auroral oval at ½799 km altitude near ½1099 MLT[ Magnetic perturbations of this mag! nitude are normal for moderately disturbed times "after Dono! van\ 0882#[

Fig[ 05[ Mapping of magnetic _eld lines from 58[14>N and 9199 LT in the ionosphere to the XY gsm equatorial plane for various strengths of Birkeland currents which would be detected at 799 km altitude[ The _eld!aligned currents would produce 9\ 51[4\ 014\ 076[4 and 149 nT\ and lead to ~ankward skewing of the magnetic _eld lines at the interface between Regions 0 and 1 currents as shown in the _gure[ A modest perturbation of ½149 nT at 799 km altitude would be associated with _eld!aligned currents capable of skewing the _eld lines by up to 6 RE for a _eld line with a neutral sheet crossing at ½04 RE behind the Earth "after Donovan\ 0882#[

space are made by a single satellite[ Since we are dealing with a magnetized plasma which is normally changing temporally while at the same time boundaries are moving with respect to a satellite\ it is extremely di.cult*if not impossible*to decouple the combined e}ects of spatial and temporal variations[ For this reason\ ground based observations of auroras and electric current systems "inferred from magnetic _eld variations# provide an important constraint for satellite observations in that they may permit the decoupling of the spatial and tem! poral variations to which the satellite detectors are exposed[ However\ to relate ground observations to sat! ellite observations\ we must have a model which allows us to map from the satellite position to the ionosphere following the magnetic _eld line on which the satellite is located[ Establishing the satellite {footprint| is a di.cult task which calls for an appropriate model of the Earth|s magnetic _eld including both the main _eld and the _elds due to currents ~owing in space[ While considerable e}ort has been expended in developing such models "e[g[\ Tsy! ganenko\ 0876\ 0878\ 0884#\ we still have a considerable way to go before we can be con_dent about mapping from the ionosphere to the magnetosphere[ The major block lies in taking into account the e}ects of _eld!aligned currents\ particularly those associated with the directly driven current system[ Figure 04 shows the magnetic perturbation from _eld!aligned currents measured by a satellite at ½799 km altitude on a NorthïSouth pass[ These currents extend from the ionosphere into the night! side magnetosphere where they must close through trans! verse currents that are driven by some sort of generator

process[ The important question to be answered is where\ in space\ this generator region is found[ The problem is that the magnetic perturbations of the _eld!aligned currents distort the background magnetic _eld between the current sheets\ so that those _eld lines can cross the equatorial plane at some considerable distance from where they might have crossed had the _eld!aligned cur! rents not been present[ Figure 05 shows the projection onto the XYgsm plane of magnetic _eld lines traced using a model developed by Donovan "0882# based on the T76 model of Tsyganenko "0876# and modi_ed by the inclusion of _eld!aligned current sheets whose e}ects are calculated in a self!consistent fashion[ It is clear that\ at a distance of ½04 RE behind the Earth\ _eld lines at the interface between the downward and upward currents can be skewed towards the ~anks of the magnetotail by as much as 6 RE for a Birkeland current system with a modest current density capable of producing a 149 nT magnetic perturbation at an altitude of ½799 km[ Thus a _eld line which one might have considered to cross the midplane near the center of the tail actually can cross the midplane close to the ~anks[ Choosing the wrong volume of space for the source region of the substorm expansive phase can have a negative impact on the task of trying to select the correct physical mechanism for explaining the disturbance[ Thus it is essential for a reliable magnetic _eld model to be developed which will permit information obtained by ground based instrument arrays to be used


87

G[ Rostoker:Journal of Atmospheric and Solar!Terrestrial Physics 50 "0888# 74*099

in placing constraints on satellite observations of vari! ations of particles and _elds in the magnetotail[

4[ Summary and conclusion We have attempted to review the properties of the magnetospheric substorm as measured by ground based magnetometers and have tried to relate in situ obser! vations of particles and _elds in the magnetotail to the substorm framework inferred from the ground based measurements[ In the _rst portion of this review we have tried to emphasize that there are two components of substorm activity] directly driven and storage:release[ Most of the attention over the past 29 years has been paid to the storage:release process\ for which the equi! valent three dimensional current system is the so!called substorm current wedge[ However\ directly driven activity accounts for a signi_cant portion of the energy dissipation in a substorm\ so it is rather surprising that most substorm models do not address in detail the physi! cal mechanism for directly driven activity nor the location of the volume of space in which the energy for this activity is made available[ Perhaps one of the few models address! ing this question is the renovated boundary layer dynam! ics model proposed by Rostoker "0885#^ however\ in the future researchers will no doubt try to _nd a description of directly driven activity in other frameworks "such as the NENL model#[ Another issue addressed in the _rst portion of this review was the terminology used to describe the so!called phases of a substorm[ Here the important point to note is that we can look at a substorm either on a global scale or on a local scale "cf Figs 6 and 7#[ On a local scale the recovery of an expansive phase intensi_cation involves the equatorward drift of auroral forms "following the poleward motion involved in the expansive phase#[ However\ on a global scale recovery actually involves poleward motion following the equ! atorward expansion of the auroral oval during the growth phase[ We expect a global poleward recovery to take place when the rate of energy input to the magnetosphere decreases to pre!substorm levels[ It is entirely likely that the {poleward leaps| of the aurora described by Hones "0874# simply represent the recovery from an expanded oval to a contracted oval con_guration associated with a reduction of energy input into the magnetosphere as can be e}ected by\ for example\ a northward turning of the IMF[ In the second part of this review\ we have looked at the evolution of the most well accepted framework in which researchers attempt to understand the substorm] the near!Earth neutral line "NENL# model[ We have shown that this model is in a state of change at the present time because of the realization that the onset of the sub! storm expansive phase does not occur at the time that lobe _eld lines begin to reconnect in the magnetotail\ as

previously thought[ One important consequence of this new development in the NENL hypothesis relates to the timing of plasmoids in the distant tail[ In its original form\ the NENL model featured the release of the plasmoid at the time when lobe _eld lines began to reconnect and the substorm expansive phase began[ Thus the size of plasmoid structures and speeds of propagation downtail were established assuming the plasmoid began to propa! gate at the time of expansive phase onset[ All these tim! ings must now be re!examined because the present ver! sion of the NENL model does not relate the release of the plasmoid to the substorm expansive phase onset[ A further question to be answered regarding the plasmoid relates to the fact that\ in almost all cases\ the structure has a strong core _eld making it more of a ~ux rope than a bubble of hot gas[ If\ indeed\ the structures which have been termed plasmoids in the past feature helical currents ~owing so as to generate a dawn!dusk core magnetic _eld within the structure\ then it is important to know how that helical current closes in the magnetosphereïiono! sphere system[ Almost certainly\ the solution to that question will be an important component to the overall problem of de_ning the role of the magnetotail in the substorm process[ Finally\ we have explained how chan! ges in the NENL model have demanded that some way be found to transport energy from the near!Earth neutral line to the region of space near the inner edge of the plasma sheet where substorm expansive phase activity is initiated[ At present\ the evolving view is that this trans! port is achieved by bursty bulk ~ows which are generated through reconnection at the near!Earth neutral line with some sort of braking mechanism close to Earth acting to convert the kinetic energy of the earthward ~owing ions to the electromagnetic energy for the substorm current wedge[ This new concept is less than a year old\ in terms of published work at the time of writing of this review\ and we can expect rapid evolution of the concept as the STEP data bases are further analysed in the next few years[ In this review\ we have not devoted attention to alter! native models for substorms[ The reader is directed to a series of papers in the June 0885 issue of the Journal of Geophysical Research for the most recent views of the proponents of the various models for substorms\ includ! ing the near!Earth neutral line model[

Acknowledgements I am grateful to my many substorm colleagues for discussions that have led\ over the years\ to the views expressed in this review[ This research was supported by the Natural Sciences and Engineering Research Council of Canada under Operating Grant OGP 4319[


G[ Rostoker:Journal of Atmospheric and Solar!Terrestrial Physics 50 "0888# 74*099

88

References
Akasofu\ S[!I[\ 0853[ The development of the auroral substorm[ Planet[ Space Sci[ 01\ 162[ Akasofu\ S[!I[\ 0857[ Polar and Magnetospheric Substorms[ D[ Reidel\ Norwell\ MA[ Akasofu\ S[!I[\ 0879[ What is a magnetospheric substorm< In] Akasofu\ S[!I[ "Ed[#[ Dynamics of the Magnetosphere[ D[ Reidel Publ[ Co[\ Dordrecht Holland\ pp[ 336ï359[ Akasofu\ S[!I[\ Chapman\ S[\ Meng\ C[!I[\ 0855[ The polar elec! trojet[ J[ Atmos[ Terr[ Phys[ 16\ 0164[ Angelopoulos\ V[W[ et al[\ 0881[ Bursty bulk ~ows in the inner central plasma sheet[ J[ Geophys[ Res[ 86\ 3916ï3929[ Baker\ D[N[\ Pulkkinen\ T[I[\ Angelopoulos\ V[\ Baumjohann\ W[\ McPherron\ R[L[\ 0885[ Neutral line model of substorms] Past results and present view[ J[ Geophys[ Res[ 090\ 01\864ï 02\909[ Baumjohann\ W[\ Paschmann\ G[\ Cattell\ C[A[\ 0878[ Average plasma properties in the central plasma sheet[ J[ Geophys[ Res[ 83\ 5486ï5595[ Birkeland\ K[\ 0897[ The Norwegian Aurora Polaris Expedition 0891ï0892\ 0\ section 0\ Aschhoug and Co[\ Christiania[ Camidge\ F[P[\ Rostoker\ G[\ 0869[ Magnetic _eld perturbations in the magnetotail associated with polar magnetic substorms[ Can[ J[ Phys[ 37\ 1991ï1909[ Chapman\ S[\ Ferraro\ V[C[A[\ 0821[ A new theory of magnetic storms[ Terr[ Mag[ Atmos[ Electr[ 25\ 66[ Clauer\ C[R[\ McPherron\ R[L[\ Searls\ C[\ 0872[ Solar wind control of the low!latitude asymmetric magnetic disturbance _eld[ J[ Geophys[ Res[ 77\ 1012[ Coroniti\ F[V[\ Kennel\ C[F[\ 0861[ Changes in magnetospheric con_guration during the substorm growth phase[ J[ Geophys[ Res[ 66\ 2250ï2269[ Davis\ T[N[\ Sugiura\ M[\ 0855[ Auroral electrojet activity index AE and its universal time variations[ J[ Geophys[ Res[ 60\ 674[ Donovan\ E[F[\ 0882[ Modeling the magnetic e}ects of _eld! aligned currents[ J[ Geophys[ Res[ 87\ 02\418ï02\432[ Dungey\ J[W[\ 0850[ Interplanetary magnetic _eld and the aur! oral zones[ Phys[ Rev[ Lett[ 5 \ 36\ 0850[ Frank\ L[A[\ Paterson\ W[R[\ 0883[ Survey of electron and ion bulk ~ows in the distant magnetotail with the Geotail space! craft[ Geophys[ Res[ Lett[ 10\ 1852ï1855[ Fukushima\ N[\ 0842[ Polar magnetic storms and geomagnetic bays[ J[ Fac[ Sci[ Tokyo Univ[ 7\ 182[ Harang\ L[\ 0835[ The mean _eld of disturbance of polar geo! magnetic storms[ Terr[ Mag[ Atmos[ Electr[ 40\ 242[ Hones\ E[W[ Jr[\ 0865[ The magnetotail] Its generation and dissipation[ In] Williams\ D[J[ "Ed[#[ Physics of Solar Plan! etary Environments[ Proceedings of the International Sym! posium on Solar!Terrestrial Physics[ American Geophysical Union\ p[ 447[ Hones\ E[W[ Jr[\ 0873[ Plasma sheet behaviour during substorms[ In] Jones\ E[W[ Jr[ "Ed[#[ Magnetic Reconnection in Space and Laboratory Plasmas[ American Geophysical Union Monograph 29\ Washington\ DC[\ p[ 067[ Hones\ E[W[ Jr[\ 0874[ The poleward leap of the auroral elec! trojet as seen in auroral images[ J[ Geophys[ Res[ 89\ 4222[ Kaufmann\ R[L[\ 0876[ Substorm currents] Growth phase and onset[ J[ Geophys[ Res[ 81\ 6360[ Kisabeth\ J[L[\ Rostoker\ G[\ 0863[ The expansive phase of

magnetospheric substorms 0[ Development of the auroral electrojets and auroral arc con_guration during a substorm[ J[ Geophys[ Res[ 68\ 861ï873[ Kivelson\ M[G[\ Khurana\ K[K[\ Walker\ R[J[\ Kepko\ L[\ Xu\ D[\ 0885[ Flux ropes\ interhemispheric conjugacy\ and mag! netospheric current closure[ J[ Geophys[ Res[ 090\ 16\230ï 16\249[ Lopez\ R[E[\ Luhr\ H[\ Anderson\ B[J[\ Newell\ P[T[\ McEntire\ ö R[W[\ 0889[ Multipoint observations of a small substorm[ J[ Geophys[ Res[ 84\ 07\786ï07\801[ Lui\ A[T[Y[\ 0880[ A synthesis model for magnetospheric sub! storms[ In] Kan\ J[R[\ Potemra\ T[A[\ Kokubun\ S[\ Iijima\ T[ "Eds[#[ Magnetospheric Substorms[ Amer[ Geophys[ Union\ Washington\ DC[ pp[ 50ï61[ Lui\ A[T[Y[\ Chang\ C[!L[\ Mankofsky\ A[\ Wong\ H[!K[\ Winske\ D[\ 0880[ A cross!_eld current instability for sub! storm expansions[ J[ Geophys[ Res[ 85\ 00\278ï00\390[ McPherron\ R[L[\ 0869[ Growth phase of magnetospheric sub! storms[ J[ Geophys[ Res[ 64\ 4481ï4488[ McPherron\ R[L[\ Russell\ C[T[\ Aubry\ M[P[\ 0862[ Satellite studies of magnetospheric substorms on August 04\ 0857 8[ Phenomenological model for substorms[ J[ Geophys[ Res[ 67\ 2020ï2038[ Nagai\ T[ et al[\ 0887[ Structure and dynamics of magnetic recon! nection for substorm onsets with GEOTAIL observations[ J[ Geophys[ Res[ 092\ 3308ï3339[ Nishida\ A[\ Mukai\ T[\ Yamamoto\ T[\ Saito\ Y[\ Kokubun\ S[\ 0885[ Magnetotail convection in geomagnetically active times 0[ Distance to the neutral lines[ J[ Geomag[ Geoelectr[ 37\ 378ï490[ Ohtani\ S[\ Takahashi\ K[\ Zanetti\ L[J[\ Potemra\ T[A[\ McEn! tire\ R[W[\ Iijima\ T[\ 0881[ Initial signatures of magnetic _eld and energetic particle ~uxes at tail recon_guration] explosive growth phase[ J[ Geophys[ Res[ 86\ 08\200ï08\213[ Perreault\ P[\ Akasofu\ S[!I[\ 0867[ A study of geomagnetic storms[ Geophys[ J[ Roy[ astron[ Soc[ 43\ 436[ Rostoker\ G[\ 0858[ Classi_cation of polar magnetic disturb! ances[ J[ Geophys Res[ 63\ 4050ï4057[ Rostoker\ G[\ 0880[ Overview of observations and models of auroral substorms[ In] Meng\ C[!I[\ Rycroft\ M[J[\ Frank\ L[A[ "Eds[#[ Auroral Physics[ Cambridge Univ[ Press\ New York[ pp[ 146ï160[ Rostoker\ G[\ 0880[ Some observational constraints for sub! storm models[ In] Kan\ J[R[\ Potemra\ T[A[\ Kokubun\ S[\ Iijima\ T[ "Eds[#[ Magnetospheric Substorms[ Amer[ Geophys[ Union\ Washington\ DC[ pp[ 50ï61[ Rostoker\ G[\ 0885[ Phenomenology and physics of mag! netospheric substorms[ J[ Geophys[ Res[ 090\ 01\844ï01\862[ Rostoker\ G[\ Camidge\ F[P[\ 0860[ The localized character of magnetotail ~uctuations during polar magnetic substorms[ J[ Geophys[ Res[ 65\ 5833ï5840[ Samson\ J[C[\ Wallis\ D[D[\ Hughes\ T[J[\ Creutzberg\ F[\ Ruo! honiemi\ J[M[\ Greenwald\ R[A[\ 0881a[ Substorm inten! si_cations and _eld line resonances in the nightside mag! netosphere[ J[ Geophys[ Res[ 86\ 7384ï7407[ Samson\ J[C[\ Lyons\ L[R[\ Newell\ P[T[\ Creutzberg\ F[\ Xu\ B[\ 0881b[ Proton aurora and substorm intensi_cations[ Geophys[ Res[ Lett[ 08\ 1056ï1069[ Shiokawa\ K[\ Baumjohann\ W[\ Haerendel\ G[\ 0886[ Braking of high!speed ~ows in the near!Earth tail[ Geophys[ Res[ Lett[ 13\ 0068ï0071[


099

G[ Rostoker:Journal of Atmospheric and Solar!Terrestrial Physics 50 "0888# 74*099 Tsyganenko\ N[A[\ 0876[ Global quantitative models of geo! magnetic _eld in the cislunar magnetosphere for di}erent dis! turbance levels[ Planet[ Space Sci[ 24\ 0236ï0247[ Tsyganenko\ N[A[\ 0878[ A magnetospheric magnetic _eld model with a warped tail current sheet[ Planet Space Sci[ 26\ 4ï19[ Tsyganenko\ N[A[\ 0882[ Modeling the Earth|s magnetospheric magnetic _eld con_ned within a realistic magnetopause[ J[ Geophys[ Res[ 099\ 4488ï4501[ Voronkov\ I[\ Rankin\ R[\ Frycz\ P[\ Tikhonchuk\ V[T[\ Samson\ J[C[\ 0886[ Coupling of shear ~ow and pressure gradient insta! bilities[ J[ Geophys[ Res[ 090\ 8528ï8549[ Wiens\ R[G[\ Rostoker\ G[\ 0864[ Characteristics of the devel! opment of the westward electrojet during the expansive phase of magnetospheric substorms[ J[ Geophys[ Res[ 79\ 1098ï1017[

Silsbee\ H[C[\ Vestine\ E[H[\ 0831[ Geomagnetic bays\ their occurrence frequency and current systems[ Terr[ Mag[ 36\ 084[ Slavin\ J[A[\ Smith\ E[J[\ Sibeck\ D[G[\ Baker\ D[N[\ Zwickl\ R[D[\ Akasofu\ S[!I[\ 0874[ An ISEE!2 study of average and substorm conditions in the distant magnetotail[ J[ Geophys[ Res[ 89\ 09\764ï09\784[ Slavin\ J[A[ et al[\ 0878[ CDAW 7 observations of plasmoid signatures in the geomagnetic tail] an assessment[ J[ Geophys[ Res[ 83\ 04\042ï04\064[ Slavin\ J[A[ et al[\ 0886[ WIND\ GEOTAIL\ and GOES 8 obser! vations of magnetic _eld dipolarization and bursty bulk ~ows in the near!tail[ Geophys[ Res[ Lett[ 13\ 860ï863[ Sugiura\ M[\ Heppner\ J[P[\ 0854[ The Earth|s magnetic _eld[ In] Hess\ W[N[ "Ed[#[ Introduction to Space Science[ Gordon and Breach\ New York[ p[ 4[