E. M. Shishkina and O. A. Troshichev
Arctic and Antarctic Research Institute, St. Petersburg, Russia
The influence of IMF and magnetic activity on the auroral oval has been founded on the basis of auroral imaging data [Akasofu et al., 1973; Feldstein and Starkov, 1967; Holzworth and Meng, 1984; Lassen, 1963] and has subsequently been confirmed by spacecraft measurements of auroral precipitating particles [Gussenhoven et al., 1981; Hardy et al., 1981, 1985, 1989; Kamide and Winningham, 1984; Makita et al., 1983, 1985, 1988]. The auroral oval displaces equatorward and expands when the IMF is southward or a substorm is in progress, and it displaces poleward when the IMF turns northward. In spite of a large amount of observations, the statistical relationship between oval boundaries and geomagnetic activity or IMF parameters was examined in only a few studies: Hardy et al. [1985, 1989] presented statistical models of auroral electron and ion precipitation as a function of the Kp index; Makita et al. [1983, 1985] examined variations of the oval boundaries as a function of the AE index and Bz ; and Holzworth and Meng  showed the dependence of the auroral oval on the IMF Bz and By components.
The question on the poleward boundary of the oval under conditions
of a northward IMF remains open up to now. The polar cap is usually
identified with the
Knowledge of the influence of the IMF
By component on the auroral oval
configuration is sparse. According to
Holzworth and Meng ,
the center of the auroral circle shifts duskward in
the southern hemisphere when
By changes from negative to strongly
positive. An analogous result was obtained by
Troshichev et al. [1996b]
for a circle representing the poleward boundary of the
hard auroral precipitation (HAP): the center of the HAP circle in
the northern hemisphere moves toward dawn (dusk) with
By > 0 (By < 0) ; the HAP circle
in the southern hemisphere shifts in the
opposite direction. Another conception was put forward by
Elphinstone et al. ,
who performed a case study of
the auroral oval configuration on the basis of auroral images from
the Viking spacecraft. According to
Elphinstone et al. ,
By component inflates either dawnside or duskside of
The aim of this study is a statistical treatment of the effects of the IMF azimuthal component on configuration and position of the auroral oval determined by precipitating particles. To reveal these effects, we have to take into account a stronger influence of the IMF Bz component and magnetic activity. Three auroral oval boundaries are examined in our analysis: the poleward boundary of the regular auroral oval, the poleward boundary of hard auroral precipitation (HAP boundary), and the equatorward boundary.
According to Troshichev et al. [1996a], the poleward boundary of the auroral oval is easily identified for all conditions by a decrease of the diffuse background ion and electron precipitation and, correspondingly, by a decline in averaged normalized ion and electron fluxes. The method proposed by Troshichev et al. [1996a] has been slightly modified in this analysis. The electron and ion number fluxes are converted to a 0.5o running mean, which is normalized by the flux peak observed for each ascending and descending node. Then latitudes at which the normalized and averaged fluxes drop to the 0.10, 0.15, or 0.20 level are determined for ions. If one of these three latitudes agrees with the drop in the normalized and averaged electron flux, that latitude is taken as a poleward boundary of the auroral oval. If no correspondence between the electron and ion boundaries is reached, the crossing is excluded from the study. The latter can occur when the spacecraft does not reach sufficiently high latitude to enter the polar cap. The poleward boundary identified in such a way is equivalent to the poleward edge of the main auroral oval in the Newell et al.  scheme, where this boundary is marked by a spatially sharp dropoff in flux energy by a factor of at least 4 down to the levels below that typical of the auroral oval. Although this dramatic dropoff does not always occur at precisely the same location for electrons and ions, in practice both species indicate the same boundary as b5 in Newell et al. .
The poleward boundary of hard auroral precipitation (HAP boundary)
has been identified by
Troshichev et al. [1996b]
as a boundary at which the averaged and normalized flux of electrons
Ee > 0.45 keV drops below the 0.1 level. Our analysis
that the position of the HAP boundary is practically invariant when
the threshold level varies from
E > 0.45 keV to
E > 3 keV. As a result, in this study we choose the
more representative threshold level of
Ee > 1 keV as
to identify the boundary of the hard auroral
precipitation. This boundary is evidently
different to the "transition
Makita et al. ,
at which the intensity
The equatorward boundary of the auroral oval is determined here in a
different manner for dawn and dusk sectors in accordance with
peculiarities of ion and electron precipitation: the dawn boundary
is determined by a drop of the averaged and normalized electron flux
below the 0.1 level, the equatorward dusk boundary is determined by a
drop of the ion flux below the same level. This boundary is close
This combined automated algorithm has been applied to particle data
obtained on board the DMSP F8 and F9 spacecraft in February-April
1990. Analysis has been carried out separately for the northern and
southern hemispheres for geomagnetic latitudes above 60o. If
equatorward or poleward boundaries could not be determined, these
orbits were deleted from the statistical analysis. The equatorward,
HAP, and poleward boundaries of the precipitation region were
determined for every available magnetic local time (MLT). The
has been taken as a measure of magnetic activity in the auroral
zone. In the former studies
[Makita et al., 1983, 1988]
parameterization of the oval boundaries was fulfilled by
using the hourly
AE index or the average
Bz component taken 1 hour
Makita et al. [1983, 1988]
showed extremely large dispersion of data. That is why in our
analysis we have tried to estimate the role of the delay time for
Bz component. With this purpose, we have calculated
wind parameters for the following time intervals preceding the
moment of the oval crossings: 10-30 min ( t
In order to make the average precipitation patterns
statistically more significant, we combine the precipitation data
To take into account the effects of magnetic activity and
Bz component, we calculated averages for
The dependency of the oval boundary latitudes on the IMF vertical component is less prominent, if values of the AE index are arbitrary (Figure 3). The correlation coefficients do not exceed 0.5 in any MLT sector, and the rate of the latitude decrease with negative Bz is rather small: in the northern hemisphere the boundaries shift by 0.23-0.27o in the dawn sector and 0.16o in the dusk sector as the Bz component decreases by 1 nT. It means that the auroral oval is displaced equatorward by no more than 3o when Bz changes from zero to -10 nT. Similar results are obtained for the southern hemisphere. It should be noted that the dependence of the boundary positions on Bz seems to be identical for negative and positive values of Bz (see Figure 3); that is, the sign of the Bz gradient is effective irrespective of the Bz sign. In the dawn, dusk, and noon sectors the influence of the Bz component looks similar to the AE index effect: the auroral oval displaces as a whole with change of Bz , with a maximum rate near noon (coefficient of regression K = 0.36 ). The poleward and HAP boundaries of the oval near midnight are practically insensitive to changes in the Bz component: the correlation between latitudes of these boundaries and the Bz component is close to zero here.
The calculated regression relationships between latitudes of the oval boundaries and AE/Bz values made it possible to exclude AE/Bz effects in every observation of the oval. With this aim, we multiplied the appropriate values of AE and Bz by corresponding regression coefficients and subtracted the quantities obtained from observed oval latitudes. The oval boundaries, extrapolated in such a manner to the level AE = 0 and Bz= 0 , have been examined as a function of the By component. Analysis has been performed separately for the northern and southern hemispheres (Figures 4 and 5).
Whereas the equatorward boundary is hardly affected by the IMF
azimuthal component, evidence of the
By effect is found for
poleward and HAP boundaries. Although the correlation in this
is insignificant as well, a certain regularity can be derived by
comparison of the
By effects in the northern and southern
hemispheres for opposite signs of
By . The results presented in Figure 4
for the northern hemisphere show that poleward and HAP boundaries
tend to displace equatorward in the dawn sector and poleward in the
dusk sector, when
By changes direction from negative to positive.
The opposite tendency is typical of the southern hemisphere (Figure 5),
where the boundaries shift toward higher latitudes in the dawn
sector and toward lower latitudes in the dusk sector when going
By < 0 to
By > 0 . In spite of a large dispersion, the
regression coefficients, describing the relationship between boundaries
By , turned out to be almost the same for the poleward and
boundaries in each MLT sector, implying that both boundaries
displace in coordination while
By changes. Such a regularity can
be observed, if the region of soft auroral precipitation (that is,
the region confined between poleward and HAP boundaries) displaces
as a whole along the
The average global
characteristics of precipitating electrons and ions have been
Hardy et al. [1985, 1989],
using the data from the DMSP F6 and F7 satellites.
The global patterns of
the integral number flux and integral energy flux, presented in
their study for seven levels of planetary magnetic activity index
Kp , make it possible to detect the configuration of the region of
precipitating auroral electrons and ions. Statistical analysis of
the dependence of the auroral oval boundary shifts on geomagnetic
activity and interplanetary magnetic field was carried out by
Makita et al. [1983, 1988].
They used a method
of boundary identification different from that in our analysis,
namely, the equatorward and poleward boundaries of the oval were
determined as the locations where the electron energy and number
flux rises or drops noticeably from the background level. To
identify the boundary between the hard and soft regions of electron
precipitation (transition boundary),
Makita et al. [1983, 1988]
used the intensity in the
The positions of the equatorward boundary derived by Hardy et al.  and by us are in good agreement for both quiet ( Kp = 1 or AE = 50 nT) and disturbed (Kp = 4 or AE = 500 nT) conditions. The results for the poleward boundary are also quite consistent. It is evident that the use of precipitating ions forms the basis for agreement between these two patterns. The noon sector presents an exception, where our determination provides lower latitudes of the oval than those of Hardy et al. , since we used only the data from the southern hemisphere in this MLT sector. The Makita et al. [1983, 1988] patterns regularly provide higher latitudes of the equatorward and poleward boundaries than our pattern. The latitudes of the poleward boundary in quiet conditions turned out to be the most inconsistent in our analysis and in that of Makita et al.  owing to quite different methods of identification of this boundary. But the good agreement holds between the HAP boundary latitudes, identified by the flux of precipitating electrons with Ee > 1 keV in our analysis, and the transition boundary, identified by electrons with Ee > 2.35 keV by Makita et al. [1983, 1988].
It should be noted that the scatter plots of the oval boundaries
Makita et al. [1983, 1988]
demonstrated large dispersion of data. Putting into operation a new
method of identification of the poleward and equatorward boundaries
of the oval, derived from the
equatorward boundary of the auroral precipitation that we determined
is adequate to the "zero-energy" convection boundary in the
Newell et al. 
Galperin and Feldstein 
schemes, where the electron and ion
The poleward boundary of the auroral precipitation under conditions of a southward IMF is naturally identified with the boundary between closed magnetic field lines and open field lines connected directly with the IMF. For strongly northward IMF, the poleward boundary of the auroral oval can be regarded [Troshichev, 1990] as the mapping of a separatrix dividing earthward and antiearthward plasma flows in the distant plasma sheet. The results of the global MHD magnetospheric simulations under northward IMF conditions [Fedder and Lyon, 1995] show that the flow inside the magnetosphere is split into predominantly sunward flow at distances earthward of X = -90 RE and mainly tailward flow further down the magnetotail. These theoretical results find support in the GEOTAIL measurements in the distant tail [Nishida et al., 1994].
The hard auroral precipitation boundary is the boundary that
divides the hard and soft zones of auroral electron precipitation.
This boundary is often regarded as the ionospheric footprint of the
boundary between the central plasma sheet and the boundary
plasma sheet. Examining simultaneous ion drift and particle
measurements from the DMSP spacecraft,
Troshichev et al. [1996b]
found that the
Only the precipitating electron and related auroral distribution was
usually examined when analyzing the oval configuration. Taking into
account auroral ions, the equatorward and poleward boundaries of
the oval, as well as the HAP boundary, turned out to resemble a
circle after exclusion of
Bz/AE effects (see Figure 6). The circular
form of the HAP boundary has been shown by
Troshichev et al. [1996b],
using independent data on the ionospheric
Holzworth and Meng 
that the poleward boundary has a
circular form. Our results, based on a
good statistical set of data, support this conclusion. Our results
also show that the HAP boundary
boundary. Thus we come to an inference that the auroral oval is
This inference is in obvious conflict with concepts of
Lassen et al.  and
Elphinstone et al. 
teardrop-shaped (or pear- shaped) form of the regular polar cap.
Elphinstone et al. ,
the influence of the
By component would result in inflation of the premidnight
segments of the polar cap, and therefore the maximum distortion of
the oval poleward boundary would be seen in these very MLT sectors.
By component influence does not show such regularity.
It is meaningful that the
By component substantially affects
locations of the poleward and HAP boundaries, but not the
equatorward boundary. Assuming that the HAP boundary corresponds to
the inner edge of the boundary layer plasma sheet and that the equatorward
boundary corresponds to the
The poleward and equatorward boundaries of the auroral oval and the hard auroral precipitation boundary have been determined by an automated algorithm with allowance made for auroral ions. The results of the analysis show that all three boundaries are influenced by magnetic activity ( AE index) to a greater extent than by variations of the IMF vertical component. When the AE index and southward IMF increase, the statistical auroral oval in the noon, dawn, and dusk sectors displaces equatorward, keeping its width. The auroral oval in the midnight sector expands equatorward with magnetic activity increase mainly at the cost of an appropriate shift of the equatorward boundary.
The boundaries of the statistical auroral oval turned out to resemble circles after exclusion of the AE/Bz effects, the centers of the circles being inconsistent with each other and located along the noon-midnight meridian. Turning of the IMF By component from a negative to a positive direction produces displacement of the poleward and HAP boundaries toward dawn (dusk) in the northern (southern) hemisphere, whereas the equatorward boundary is not affected by the azimuthal IMF. The displacement is maximum in the dawn/dusk sectors and minimum in the nighttime sector. The polar cap looks like a circle, and no signatures of a teardrop- or pear-shaped polar cap are founded. Under the influence of the IMF By component, the polar cap is displaced as a whole toward dawn or dusk, depending on the sign of By .
Akasofu, S.-I., P. D. Perreault, F. Yasuhara, and C.-I. Meng, Auroral substorms and the interplanetary magnetic field, J. Geophys. Res., 78, 7490, 1973.
Cowley, S. W. H., Magnetospheric asymmetries associated with the
Cowley, S. W. H., and W. J. Hughes, Observations of an IMF sector effect in the Y magnetic field component at geostationary orbit, Planet. Space Sci., 31, 73, 1983.
Elphinstone, R. D., K. Jankowska, J. S. Murphree, and L. L. Cogger, The configuration of the auroral distribution for interplanetary magnetic field Bz component, 1, IMF Bx and By dependencies as observed by the Viking satellite, J. Geophys. Res., 95, 5791, 1990.
Elphinstone, R. D., et al., The double oval UV auroral distribution, 1, Implications for the mapping of auroral arcs, J. Geophys. Res., 100, 12,075, 1995.
Fedder, J. A., and J. G. Lyon, The Earth's magnetosphere is 165
Feldstein, Ya. I., and V. G. Starkov, Dynamics of auroral belt and polar geomagnetic disturbances, Planet. Space Sci., 15, 209, 1967.
Galperin, Yu. I., and Ya. A. Feldstein, Mapping of the precipitation regions to the plasma sheet, J. Geomagn. Geoelectr., 48, 857, 1996.
Galperin, Yu. I., et al., Diffuse auroral zone in the evening and near midnight sectors, Kosm. Issled., 15, 421, 1977.
Gorney, D. J., D. S. Evans, M. S. Gussenhoven, and P. Mizera, A
Gussenhoven, M. S., Extremely high latitude auroras, J. Geophys. Res., 87, 2401, 1982.
Gussenhoven, M. S., D. A. Hardy, and W. J. Burke, DMSP/F2 electron observations of equatorward auroral boundaries and their relationship to magnetospheric fields, J. Geophys. Res., 86, 768, 1981.
Gusev, M. G., and O. A. Troshichev, Hook-shaped arcs in the dayside polar cap and their relation to the IMF, Planet. Space Sci., 34, 489, 1986.
Gusev, M. G., and O. A. Troshichev, Simultaneous ground based observations of polar cap arcs and spacecraft measurements of particle precipitation, J. Atmos. Terr. Phys., 54, 1573, 1992.
Hardy, D. A., W. J. Burke, M. S. Gussenhoven, N. H. Heinman, and E. Holeman, DMSP/F2 electron observations of equatorward auroral boundaries and their relations to the solar wind velocity and the north-south component of the interplanetary magnetic field, J. Geophys. Res., 86, 9961, 1981.
Hardy, D. A., M. S. Gussenhoven, and E. Holeman, A statistical model of auroral electron precipitation, J. Geophys. Res., 90, 4229, 1985.
Hardy, D. A., M. S. Gussenhoven, and D. Brautigam, A statistical model of auroral ion precipitation, J. Geophys. Res., 94, 370, 1989.
Hoffman, R. A., R. A. Heelis, and J. S. Prasad, A
Holzworth, R. H., and C.-I. Meng, Auroral boundary variations and the interplanetary magnetic field, Planet. Space Sci., 32, 25, 1984.
Hones, E. W., J. D. Craven, L. A. Frank, D. S. Evans, and P. T. Newell, The horse-collar aurora: A frequent pattern of aurora in quiet times, Geophys. Res. Lett., 16, 37, 1989.
Ismail, S., and C.-I. Meng, A classification of polar cap auroral arcs, Planet. Space Sci., 30, 319, 1982.
Ismail, S., D. D. Wallis, and L. L. Cogger, Characteristics of
Kamide, Y., and J. D. Winningham, A statistical study of the instantaneous nightside auroral oval: The equatorward boundary of electron precipitation as observed by the ISIS 1 and 2 satellites, J. Geophys. Res., 89, 7441, 1984.
Lassen, K., Geographical distribution and temporal variation of polar aurorae, Det Danske Meteorol. Inst. Medd., 16, 1963.
Lassen, K., and C. Danielsen, Quiet time pattern of auroral arcs for different directions of the interplanetary magnetic field in the Y-Z plane, J. Geophys. Res., 83, 5277, 1978.
Lassen, K., C. Danielsen, and C.-I. Meng, Quiet-time average auroral configuration, Planet. Space Sci., 36, 791, 1988.
Lui, A. T. Y., et al., Special features of a substorm during high solar wind dynamic pressure, J. Geophys. Res., 100, 19,095, 1995.
Makita, K., C.-I. Meng, and S.-I. Akasofu, The shift of the auroral
electron precipitation boundaries in the
Makita, K., C.-I. Meng, and S.-I. Akasofu, Temporal and spatial variations of the polar cap dimension inferred from the precipitation boundaries, J. Geophys. Res., 90, 2744, 1985.
Makita, K., C.-I. Meng, and S.-I. Akasofu, Latitudinal electron precipitation patterns during large and small IMF magnitudes for northward IMF conditions, J. Geophys. Res., 93, 97, 1988.
Murphree, J. S., et al., Characteristics of optical emission and particle precipitation in the polar cap arcs, Planet. Space Sci., 31, 161, 1983.
Newell, P. T., D. G. Sibeck, and C.-I. Meng, Penetration of the interplanetary magnetic field By and magnetosheath plasma into the magnetosphere: Implications for the predominant magnetopause merging site, J. Geophys. Res., 100, 235, 1995.
Newell, P. T., Ya. I. Feldstein, Yu. I. Galperin, and C.-I. Meng, The morphology of nightside precipitation, J. Geophys. Res., 101, 10,737, 1996.
Nishida, A., et al., Structure of the neutral sheet in the distant tail ( X = -210 RE ) in geomagnetically quiet times, Geophys. Res. Lett., 21, 2951, 1994.
Troshichev, O. A., Global dynamics of the magnetosphere for northward IMF conditions, J. Atmos. Terr. Phys., 52, 1135, 1990.
Troshichev, O. A., E. M. Shishkina, C.-I. Meng, and P. T. Newell, Identification of the poleward boundary of the auroral oval using characteristics of ion precipitation, J. Geophys. Res., 101, 5035, 1996a.
Troshichev, O. A., E. M. Shishkina, G. Lu, and A. D. Richmond, The relationship of the ionospheric convection reversal to the hard auroral precipitation boundary, J. Geophys. Res., 101, 15,423, 1996b.
Valladares, C. E., H. C. Carlson, and K. Fukui, Interplanetary
magnetic field dependency of stable