Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.izmiran.rssi.ru/magnetism/Feldstein_2012/PDF/356.pdf Äàòà èçìåíåíèÿ: Wed Mar 11 16:39:10 2009 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 21:08:54 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: magnetic north

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 99, NO. A7, PAGES 13,251-13,262, JULY 1, 1994

Electric potential patterns in the northern and southern polar regions parameterized by the interplanetary magnetic field
V. O. Papitashvili, B. A. Belov,D. S. Faermark,Ya. I. Feldstein,
S. A. Golyshev,L. I. Gromova, and A. E. Levitin
Institute of Terrestrial Magnetism, Ionosphere,and Radio Wave Propagation, Troitsk, Moscow Region, Russia

Abstract.

Electric potential patterns have been obtained from the IZMIRAN

electrodynamic model(IZMEM) for the northernand southern polar regions during
summer, winter, and equinox. The model is derived from a large quantity of

high-latitude ground-based geomagnetic data (above4- 57Ü corrected geomagnetic latitude) at all magnetic local time hours. A linearregression analysis technique
has been used to obtain the quantitative responseof each magnetic observatory to

changes interplanetarymagnetic of field (IMF) components. Sinceno ionospheric
conductivitymodel exists specifically the southernpolar region, the statistical for

modelof Wallis and Budzinski (1981) has beenappliedin both hemispheres. A cross-polar "background" potentialof ·35 kV, derivedby Reiff et al. (1981), is
usedto calibrate IZMEM's potential patterns. The model'sresponses changes to in the IMF By and B· components analyzedto obtain a set of "elementary" are
convection patterns in both polar regionsfor eachseason the year. Asymmetry in of the potential pattern geometryin both hemispheres can be attributed either to the influence of the "northern" ionosphericconductivity model which was applied to the southernpolar region,or to somenatural phenomena.The modeledbackground

cross-polar potentialfor the condition whenB· = By = 0 is foundto be ·37 kV.
Averagevaluesof the modeledpotential drop causedby each nanoteslaof the IMF are the following:~14 kV for southward B·; · -4 kV for northwardB·; and · 4-4.5 kV for By components. The latter is not applicable the "dawn-dusk" to potential drop; it may be applied across cuspregiononly. Nevertheless, combinationof the a the background and elementarypotential patternsin the casestudiesgivesa certain estimationof the cross-polar potential drop, which may be stronglydistorted during

time of largeBy. It is concluded IZMEM provides that realistic convection patterns
parameterizedby the IMF componentdirectionsand magnitudesand may be used to provide routine estimatesof convectionpatterns and electric potentials if IMF
data are available.

Introduction

form characterof ionospheric conductivity,novel numerical methods have been developedto resolvethe that relate the magneticfield perturbations Interpretation of ground-basedgeomagneticpertur- equations electric field bationsby meansof ionospheric electricfieldsand cur- on the Earth's surfaceto the ionospheric 1977;Mishinet al., 1980;Kamide rents beganwith the fundamentalworksof K. Birke- potential[Faermark, of set land, S. Chapman, and H. Alfven. Since then Kern et al., 1981].The evolution ideas forthby Chapin [1966] founda relationbetween scalar the current func- man and Alfven has been presented detail by FeldsteinandLevitin[1986].Considerable progress been has tion obtainedfrom divergence-free portionof the steady madesince that time in inferringthree-dimensional curcurrent distribution in a thin spherical layer and the rent systems combining ground-based by the magnetic densityof the field-aligned currents the caseof a uniin electric potential and form ionospheric conductivity. Because the nonuni- data and modeledionospheric of fields.The assimilative mappingof ionospheric electro· Also at SpacePhysics Research Laboratory,Universityof
Michigan, Ann Arbor, Michigan. Copyright 1994 by the American GeophysicalUxfion.
Paper number 94JA00822.

dynamics (AMIE) technique [Richmond Kamide, and 1988], which is a further development the KRM of method[Karnide al., 1981],hasbegun be widely et to used the analysis electrodynamic of thehighfor of state latitudeionosphere Knipp et ai., 1991]. [e.g.,
Empirical modelsof high-latitude electric fields for differentorientations the IMF havebeendeveloped of

0148-0227 [94J A-00822$05.00 [94

13,251


13,252

PAPITASHVILI

ET AL.: ELECTRIC

POTENTIAL

PATTERNS

from satellite observations [e.g., Heppner,1977; Hep- the high-latitude electric potential in both hemispheres pner and Maynard, 1987; Hairston and Heelis, 1990]. parameterized by the IMF conditions, and show the Subsequent synoptic studies radarmeasurements of [see IZMEM's ability to model electric potential patterns
Foster, 1983; AIcayde el al., 1986; de la Beaujardiere for any moment of time when the IMF data are availel al., 1991]have yieldedresultsthat are in substan- able. The numerical scheme also estimates ionospheric current tial agreement with statistical satellite models. Most electric fields associated with three-dimensional of the ground-based magnetometer studies that are systems. The model has a transportable FORTRAN basedon the KRM and later the AMIE techniques also code and may be run on a personal computer. showa good agreement with satellite and radar models

There are also a number of reports representing the theoreticalapproaches the modelingof high-latitude to convection patterns[seeHeelisel ai., 1982;Clauerand Friis-Chrislensen,1988; Blomberg and Marklund, 1991, and references therein]. At the end of the 1970s,numericaltechniques which are similar to the KRM were developed independently by two scientificgroupsin Russia:the IZMIRAN electrial Magnetism,Ionosphere, RadioWave Propagaand tion, Troitsk, Moscow Region[Belovet al., 1977],and the technique inversion the magnetograms of of (TIM) at SibIZMIR, Irkutsk [Mishin et al., 1977]. We refer the readerto the originalwork of Mishin et al. [1980]
the IZMEM method only.

[e.g.,Friis-Chrislensen ai., 1985;Knippel al., 1993]. Model Description el
We postulate that the magnetosphere-ionosphere coupling link can be considered a black box, which acas

ceptschanges the IMF andsolarwindplasma of (SW) parameters (Bx, By, Bz, velocity anddensity as V, n)
an input signal, and inducesground-based geomagnetic perturbations as an output signal. This approachhas already been used in others works, in particular, those

trodynamic model(IZMEM) at the Instituteof Terres- employingthe linear predictionanalysis[see Clauer, 1986,and references therein]. A numberof interplanetary parametersare knownto be associated with mag-

netospheric interaction[Levitinet al., 1982]. For example, there is much evidence in the literature show-

ing impactof the IMF By and Bz components the on for details of the TIM technique. Here we shall describe
AMIE, KRM, and TIM techniques analyzegroundbasedmagneticfield perturbationsderivedfor an event with a durationfrom a few hoursto a few days. They require a selectionof nearby "magneticallyquiet" period which is a most subjectivepart of the analysesbecause the IMF conditions might be considerably differentfor "disturbed" and "quiet" periods. These techniques use alsospherical harmonic expansion process to initial geomagnetic data and coversizable gapsbetween magnetic observatories, especiallyat high latitudes. Therefore the expansion uncertainties muchhigher,for examare ple, in the Arctic oceanor Antarctic region than for
Canada or Scandinavia.

magnetic field at the Earth's surface. The division of Bz into negativeand positive valuesmay representdis-

turbed and quiet geomagnetic conditions, respectively, thougha northward Bz can inducea strongpolar cap currents as well. The IMF B· component has been foundto showlittle correlationwith geomagnetic variations[Maezawa, 1976;Levitinet ai., 1982;Troshichev,

1982].Therefore cancompute regression we the coefficientsKBx but may disregardtheir contributionto the

model.We havetried a number SW parameters of (velocity V, density n, temperature T, and someof their

combinations) find a better correlation to with groundbased dataandconcluded V2 andnV2 show that significantcorrelations. The V · term may,in part, represent
"quasi-viscous" interaction of the solar wind plasma with Earth's magnetosphere; nV · is proportional the
to dynamic pressureof the solar wind.

At IZMIRAN, a regression analysis wasusedto study geomagnetic variations causedby changes the interin planetary magnetic field(IMF). Initial results havebeen publishedin Russianin the proceedings IZMIRAN of and Geomagnelism Aeronomy and [e.g.,Afonina el al.,

1980](seeotherreferences the workby Levitinet al. IMF parameter: in [1982] andFeldslein al. [1984]).These et investigations

We use a regression model where regression coefficientsrelate any ground-based geomagnetic field component,for example,H, to changes the corresponding of

haveutilized geomagnetic data from the northernpolar i zz Hi = Kk.·,B· + Kji.y By+ K}iB B + HO (1) i regiononly. However, usingmagnetic observations from the southern polarcap[Mansurov al., 1981],wehave The freeterm of equation canbe expanded the el (1) for applied the same regression techniqueto the Antarc- solar wind parameters:

tic data [Papitashvili al., 1983,1989, 1990]. This el

approachprovidesa parameterization geomagnetic of variationsby the IMF, and the ionospheric electrodynamics may then be defined. The IZMEM does not require a selectionof magneticallyquiet period or use of the spherical harmonicexpansion. Thesedistinguish the IZMEM from other techniques. The purpose of this paper is to describebriefly a numericalschemeof the IZMEM, presenta summary of resultsobtainedin the form of average patternsof

uo' = .%v, v?+

n,

+ uoo'

(2)

HereK·/ are regression coefficients i = 1,...,24, for
wherei is universal time (UT) hour; HO is a residi ual part of (1) for the average conditions solarwind of

(n = 4 cm V = 450 kin/s); HO0 represents -a, i geomagnetic variations which are free of the 1MF and SW

impact(we shallomit indexi further). In this paper
we shall consider modelparameterized the IMF the by


PAPITASHVILI

ET AL.' ELECTRIC

POTENTIAL

PATTERNS
Thule

13,253

only and refer the readerto the work by Levitin el al.

[1982]and Papitashvili al. [1990]wherethe solar et
wind parametersare considered. The total hourly mean valuesof the IMF and groundbasedgeomagnetic data for each season the year of

KZBy Bz>0 SummerBz<0 Ey0 6 12 18 MLT 6 12 18 24 n
I I I I I I I I

(summer, winter,and equinox) bothnorthern and and
southernpolar regionsabove·4-570 correctedgeo-

magnetic (CGM) latitudehavebeenused the regresin
sion analyses. The regression geomagnetic model for the northernpolar regionhasbeendeveloped usingthe

datafrom 15 magnetic observatories 1968-1969 in [Levitin el al., 1982].The northernmost stationwasThule (· = 86.2Ü). The samemodelfor the southern hemisphere beenobtainedfromthe 1978-1980and 1983has 1984data (21 magnetic observatories autonomous and magnetometers). the datafromfourstations Here poleward of-85 o latitude were available and the southern-

1967

1966

1965

most station was at · = -89.1 o [Papitashvili al., el 1990].
The arrays of the IMF and geomagneticdata were subjectedto regression analyses eachof 24 UT hours for of eachdayoverthe entireseason the year (120 days). of The resultant magneticlocal time (MLT) daily variation of regressioncoefficientsKH and AH0 around daily mean value H0 were obtained. These results have been comparedfor the same hourly mean valuesof the IMF and geomagneticdata, and IMF valuesone hour

110nT

Figure 2. Diurnal variationof KZBy regression coefficient at Thule during 1965-1970. [After Papitashvili, 19821 .
ahead of the ground-baseddata. A better correlation
was obtained when the same hourly mean values were compared. With this model we assumethat ground-basedgeo-

laTI, nT

Thul½ Sununer IATI,nT

//

magneticdisturbances proportionalto variationsof are the IMF componentsand there are a variety of physical mechanisms that provide links that transfer energy from the solar wind plasmato the high-latitude magnetosphere and ionosphere. The assumedlinearity was

studied and confirmedfor the Bz component[Papilashvili el al., 1981; Troshichev, 1982]. Figure 1, for example, shows a dependenceof the total horizontal
I I
I

component perturbation, /kT- (K·Bz + K·,oz)·/9', at
Bz,nT
11 - 13 MLT

10
5 - 07 MLT

100

100

Thule for negative and positive IMF B· and for different MLT hours during summer. Dashed lines show a regression between AT and Bz for two separatearrays of original data: for B· < 0, and B· > 0. Short segments representa regression betweenground-based data and the IMF within binned(by 2-6 nT) intervals of B·. All of these segments follow the corresponding dashedregression lines. The solar cycle variation has been studied by Papi-

lashviii[1982].Figure2 shows example the KZBy an of
-10 10

diurnal variations for the Z componentat Thule dur-

-10 17- 19 MLT

10

Bz, nT
23 - 01 MLT

ing the summerfor eachyear separately (1965-1970),
and for all of the original data combined into one array correctedfor the secularvariation. The magnitude and shapeof the variationsare similar for eachyear. Therefore a regressionmodel of geomagneticvariations can be used over the entire solar activity cycle. These two

studies (linearityandsolarcycle)havebeencarriedout for the polar cap stationsonly (Thule, Godhavn,Vostok, and Mirny), and they should extended future be in
Figure 1. Linear regression dependence an intensity for the auroral and subauroralmagneticobservatories. of of geomagnetic field horizontalcomponent perturbation The "regression modeling" approachhas severaladat Thule againstthe negativeand positiveIMF Bz val- vantages: total values geomagnetic (1) of field compoues. [After Papitashvili al., 1981]. el nents are usedin the analysis,and there is no subjective


13,254

PAPITASHVILI

ET AL.: ELECTRIC

POTENTIAT

PATTERNS

selection a perturbationbaseline; the technique of (2) usesmany measurements made by a limited amount of magnetic observatories different at localtimesdue to the Earth's rotation, therefore,24 valuesof Kn are foundfor eachobservatory; only an interpolation (3) of Ks alongmeridians required, is insteadof spherical harmonicexpansion; only the IMF valuesare (4)

the IMF components a givenMLT hour;½bz ½b· for and are the solutions (4) for eachset of regression of coefficientsand correspond changes a given parameter to of

(electricpotential, ionospheric currents,etc.) on the IMF 1-nT step;·0 is the solutionof (4) for the setsof freeterm in (1) duringdifferent conditions the IMF in (e.g.,negative positiveBz and B·). or required to model geomagneticvariations, and then The ·0 term in (5) represents "background" the poelectrodynamic parameterscan be obtained using the tentiM, which exists in the ionosphereduring average IZMEM during all three seasons the year in both conditionsin the solar wind, that is, "viscous"convecof tion according Reiff et al. [1981].The otherterms to northern and southernpolar regions. This regression model of geomagneticvariations is of the (5) represent the "elementary convection cells" used as an input for numerical solution of the second- at high latitudes causedby the IMF components. A combination of these elementary cells for definite conorderpartial differential equation [Faermark, 1977]:

vx

- vx x

ditions in the IMF will reproduce a typical convection pattern observed satellites and radars over the polar by

Here iselectrostatic ½b potential - 0 at T - :t=57 ·2 allowsone to constructa quantitative model of the iono(· Ü), is a tensorof nonuniformionospheric conductivity,n· is spheric electrodynamics. a unit radial vector, and 9 is an equivalentcurrent funcThe IZMEM electrodynamic parameters have been tion, uniquelyrelated to geomagnetic perturbationson obtained initially using the ionosphericconductivity the Earth's surface. A definition of the current function
in the IZMEM method is similar to that in the work by

regions.Therefore,as the (1) describes basicstruca ture of the high-latitude geomagnetic variations, (5) the

Kamideet al. [1981].By analogy with the AMIE technique [Richmondand Kamide, 1988]the IZMEM approachrepresents regression a mappingof ionospheric electrodynamics (RMIE). Equation(3) may be rewrittenin spherical coordinates0 (colatitude)and A (eastlongitude)[Feldstein andLevitin,1986]:

distributions derived 1700UT and 0500UT (north for magnetic poleat localnoonand midnightrespectively).
To avoid the UT dependenceof the IZMEM output,

the averaged ionospheric conductivity distribution (betweentwo derived 1700UT and 0500UT) hasbeen for
used to compute the entire set of electrodynamic parameters. This averagingassumes that the geographic and geomagnetic poles are coincident. In this casethe difference betweenthe cross-polar potentialsfor the 'av-

eraged'and UT-dependentmodelsis about ~25- 30%,

-

sin0Es· +sin· 0q EH·

but the electric potential distributions are very similar. The followinganalysisand figuresare presented the for

+
= sin·· · sinO a 0A (4)

UTaveraged model.
While we have used a commonly adopted division of

the year into seasons (May-August:northern summer
and southern winter, November-February: northern winter and southern summer; March, April, September, and October are equinoctial months for both hemi-

geomagnetic latitude and one MLT hour. Sinceno iono- the solar UV ionosphericconductivity will be better utilized. spheric conductivitymodels exist specifically for the southern polar region,the particle precipitation statis-

it to where andEli-areheight-integrated andPed- spheres), is alsopossible bin the databy the Earth's EH Hall ersen ionospheric conductivities specified a gridof 1Ü dipole tilt and developthe UT-dependent model where on

tical conductivity model WallisandBudzinski of [1981] Convection Patterns Parameterized by
and the solar UV conductivitymodel of Robinsonand the IMF

Vondrak [1984] usedfor both hemispheres. are
The distributions of electric potential can be determinedand parameterized a superposition the IMF as of
related terms:

The IZMEM modeloutputs (electricpotential,electric and magneticfields, ion convection velocity,etc.)
have been already comparedwith the satellite and radar

ú (·, MLT, B·, B·) = ·(·,

MLT) As,

measurements Beiovet ai., 1984;Dremukhina [e.g., et ai., 1985; Papitashvili and Clauer, 1993]. The calcu-

+ ·,(·, MLT)Als·+ (I)o(·, MLT)

(5) lated valuesof electrodynamic parameters are usually

Here9·isa corrected geomagnetic latitude;MLT ismagnetic local time; · may represent electricpotential,as well as electricand magnetic fields,ionospheric (Hall and Pedersen) field-aligned and currents, Jouleheator

smallerthan the measured data. Geomagnetic disturbances obtained from the regression model(1) havenot been separatedinto the internal and external variations and extendedupward on the ionospheric level. Therefore an amplification factor should be derived from the

with data. For example,Feidingrate;As, andAls·aredimensionless amplitudes comparison experimental of


PAPITASHVILI

ET AL.: ELECTRIC

POTENTIAL

PATTERNS

13,255

steinandLevitin[1986]compared OGO 6 observationsclockwisein the positive vortex. These particular patas [Heppner, 1972] with themodeled electric fields ob- ternsfor Bz = By = 0 would be classified viscous and
the potential "...remainswhen tained the factor of 3.0. Another approachwill be used cellsbecause background

to calculateamplification factorsin this paper. (Note the mergingtheory predictszero, and this residualmay (closemodel) that all of the followingfigurespresentthe uncorrected logically be attributed to non-merging [Rei. modeloutputs,that is, without the useof amplification transferprocesses..." ff et al., 1981,p. 7645]. The "standard" two-cell convectionis well developed factor.)
in both hemispheres;the vortices occupy almost the

Modeled Electric Potential for Bz - By - 0
The ground-based geomagnetic data have been initially divided into two arrays according to the sign of Bz, that is, for the disturbed (Bz < 0) and quiet (Bz > 0) statesof the magnetosphere. Thesearrays have been subjected to regression analysesseparately. Thereforetwo setsof all termsin the equation(1) have

same areas for each season. There is a little asymme-

try in the location of antisunward transpolar fio;vs in both hemispheres, but they are directed generallyfrom

0900-1000MLT to 2100-2200MLT. Reiff et al, [1981] havesketched possible the flowoverthe entirepolar cap; however, Reiff andBurch[1985] haverelnoved flow this
for the zero IMF condition. We think that the convec-

been obtained. The free term in both casesshowsgethe omagneticdisturbances remained in the polar regions whenthe II·[F Bz = By = 0 because magnetosphere will be closed completely and viscousinteraction can when the IMF equals zero. The electric potential distributions inferred from the free terms should be similar

tion flow acrossthe centerof polar cap may exist even

takeplace 0vcrthecntirc magnct0pausc surface, For
very similar to our "northern summer" pattern in Fig-

et for the disturbed and for the quiet conditions because example,Friis-ChriMensen al. [1985]haveobtained the electricpotentialdistribution zero!MF, whichis for the separatearrays are subsets from the gencralpopulation.

patFigures3 showselectricpotential distributions for ure 3. It is particularly interestinghow convection
the disturbed conditions. Similar distributions were ob-

terns are similar in both hemispheres,especiallysince

tainedfor the quiet conditions (not shown).Thesepat- they were obtained from the data of different years, and terns may alsobe considered "convcction as patterns" even different solar activity cycles. Modeledcross-polar potential(the MAX-MIN value) bccause ionospheric plasmamovesalongthe equipotena dependence. Averaged(betweendistial lines: clockwisein the negative vortex and anti- shows seasonal
1· Northern 1· Northern
Northern ·inter
60'

06

}lax Min

lB
3.7 -5.3

06 lB
00 MLT
Southern

Max B.3 Min-10.0

06 lB
00 MLT
1· Southern

Max Min

lB 5 -9.3

00 MLT
Southern
Winter

.*/' '._·ummer
o
18 06 lB

./'..·quinox
/ÜÜ4
06 8

-60'

O6

Max 7.3 Min -6.4

O0 MLT

Max Ó.· Min -3.9

O0 MLT

a Min

87 -9.6

oo

MLT

Figure 3. Modeledbackground (freeof the IMF impact)electricpotentialdistributions over the'northern andsouthern polarregions disturbed < 0) conditions. for (Bz Polarplotsare in the corrected geomagnetic latitude - magnetic local time coordinates. Contourintervalsequal
1.5 kV.


13,256

PAPITASHVILI

ET AL.: ELECTRIC

POTENTIAL

PATTERNS

Table 1. Amplified Electric Potential in Kilovolts Modeled by the IZMEM to Changesof the Interplanetary Magnetic Field 1-nT Step
B· ·0 B· ·0 By · 0 By ·0

Season
Summer

Hemisphere
northern southern

!*
3.2 3.0

Bz = By = 0
30.4 37.8 14.1 10.5

4-Bu
-3.2 -4.5

Bz < 0
5.4 -4.8

B· > 0
5.4 -5.1

B· < 0
-5.4 6.0

B· > 0
-5.4 6.6

Equinox

northern
southern

2.3
2.7 1.5 2.5

41.4
30.5 40.5 40.5

15.4
13.0 13.0 15.8

-2.5
-2.4 -4.8 -5.8

4.6
-0.5 4.8 -4.8

4.4
-0.5 2.7 -3.8

-4.6
2.7 -4.8 7.8

-4.4
3.0 -2.7 7.0

Winter

northern southern

Average

potential

drop

36.8 q- 5.1

13.6 q- 1.9

-3.9 q- 1.4

4-3.9 4- 1.7

4-5.0 4- 1.7

*Amplification factors f are calculated from the cross-polarbackground electric potential ~35 kV, derived by Reiff et

al. [1981].

turbed quietconditions)values summer, and for equinox,capsduringsummer and equinox; however, is disit and winter are 9.5, 18.0, and 27.1 kV respectively placed towards afternoon in winter. During summer in the northernhemisphere, and 12.6, 11.3, and 16.2 kV both dawn and dusk convectionvortices are approxin the southern one.The electric potential significantlyimately equal; the dawn cell is larger in both hemiincreases the northernhemisphere in from summerto spheresduring equinoxand winter. winter;thisis not the case the southern for hemisphere. The cross-polar (dawn-dusk) potential inferred from We haveno explanation this: perhaps is a natural IZMEM is slightlylargerin the northernhemisphere: for it phenomenon the northernionospheric or conductivity summer kV (north),3.5kV (south); 4.4 equinox kV 6.7
causes this discrepancy.

background cross-polar potential,and the potentialval- Batch [1985]. The sunward convection flow is directed uesfor different IMF components. Average cross-polarfrom midnight dayside to overthe near-pole region. potential 36.8 kV (after amplification) the IMF The main finding here is that the reverseconvection of for Bz = By = 0 conditionis guaranteedto be close~35 cellsare spreading overthe entire polar region;they kV derived Reiff et al. [1981]. by are not limited to the high-latitude polar cap only as suggested Reiff andBurch by [1985].Friis-Chrislensen Modeled Electric Potential for B· < 0 et al. [1985] haveobtained alsothe expanded reverse Figure4 shows standard the two-cell convection pat- convection down to 70o latitude, but their resultsdo ternscaused the southward by IMF (Bz < 0), which not reveal the sunward convection flow near the noon wouldbe classified "merging"cellsin the modelof meridian. as The reverse convection in our model extends over auReiff andBurch[1985].Modeled patterns verysimare ilar in both hemispheres antisunward with transpolar roral latitudes and forms a third convection cell situated convection flow whichis directedgenerally from 1000- near midnight during equinox and winter. The flow 1100MLT to 2000-2200MLT [Friis-Christensen al., concentrates et near midnightat ~ 4-770latitude during 1985]. This flow moves through center polar the southern the of equinox and northern winter. This may

of ~35 4. 10 kV obtained Reiff el al. [1981] by from a 1-nT step of the southward IMF is 13.6 kV. This the low-altitude satelliteobservations. Recalling that is comparable with the valueinferredby Reiff et al. the IZMEM derives smaller values than actual measure- [1981]: seeequation (b(kV)= 324- 14.7 for B, < 7 B, ments,we may divide the background potentialby the nT in their Table 1. As these merging convection cells averagedmodeled values and then fit the obtained ra- encompass viscous the cells,the total cross-polar potios by a straightline against division the year tentiMwill increase the of correspondingly.typicalvalue A of into seasons (four months each). This procedure the cross-polar repotential for the IMF Bz = -5 nT and sultsin the amplification factors: summer (north), averagesolar wind conditionswill be around 100 kV in 3.2 3.0 (south); equinox (N), 2.7 (S); winter1.5 (N), accordance with the numbers from Table 1. 2.3 2.5 (S). Thesenumbers be used furtheramplican for ficationof modeled (for eachIMF parameter) values Modeled Electric Potential for B· > 0 until moreinformation about ionospheric conductivity Figure5 shows 'reverse' the two-cell convection patin the southernpolar regionwill be collected.Table 1 ternsfor the northward IMF (B· > 0), whichwould shows amplification the factors, corresponding modeled be classified "lobe"cellsin the modelof Reiff and as

If we assume that the viscous cross-polar electric po- amplification these of values the corresponding by factential is not changingwith the season,we can com- tor fromTablei results the comparable in cross-polar parethe modeled values with the background potential potential.The average cross-polar potential change for

(N), 4.8kV (S);winter kV (N), 6.3kV (S).However, 8.7


PAPITASHVILI

ET AL.:

ELECTRIC

POTENTIAL

PATTERNS
12

13,257
Northern
Winter

12

Northern

12

Northern

60'

18

06

18

06

06

Max 2.4 Min -2.0

O0 MLT
12 Southern

Max 4.6 · Min -2.1

O0 MLT
12
Southern

73 Min -1.4

00

MLT

12

Southern

Equinox
-60'

,

Winter
-60'

18

06

18

06

18

06

Max

Min -1,5

2.0

O0 MLT

Max

Min -O,B

4.0 ·

Max

4.9

O0 MLT

Min

- 1.4

O0 MLT

Figure Modeled 4. electric potential distributions southward (Bz- -1 nT) comforthe IMF
ponent.Contour intervals equal0.5 kV.

12

Northern

12

Northern

12

Northern

mer 0 o
184
06 18 · ',

····. ·· Equinox

Winter

·
Max Min 1.0 -0.3 O0 MLT

·koo '
18

,
06

Max Min

0.8 -0.5

O0 MLT

06 Ma
Min -1.8

O0 MLT

12

Southern
Summer

12

Southern

12

Southern

60*

.o,·inox 60' ·:".., ':
18

Winter -60'
18 06

18·

06

06

Max

Min

-0.7

O0 MLT

Max Min

1.2 -0.9

00 MLT

Min

-1.2

00 MLT

Figure 5. Modeled electric potential distributions the northward for IMF (Bz - +1 nT) component. Contourintervalsequal 0.25 kV.


13,258
be a manifestation

PAPITASHVILI ET AL.' ELECTRIC POTENTIAL PATTERNS
of substorms which occurred dur-

ing the quiet (Bz > 0) conditions not shown.This are ing northward (notethat UT hours Bz with substorm's representsthe ionosphericcurrent system termed DPY
activity were not excludedfrom geomagneticdata sub-

by Friis-Christenscn Wilhjelm and [1975]. TheDPY

jected to regression analyses). current system produced me. is by ridional electric fields The reverse Convectionis caused by "dusk-dawn" whosedawn-duSk potential differenceis almost ze.ro electric fields. Because the MAX and MIN values on the [Feldstein Levitin,1986]. Therefore and only'theelecfigure may showextreme points over the entire polar re- tric potential acrossthe polar cuspsmust be·;onsidered gion, the dusk-dawnpotential 'across daysideportion for the IZMEM calibration a purposes (equat0rward ionoof the polar cap must be taken into accountseparately: spheric electricfield is assumed positive). as

summer -1.0 kV (north),-1.5 kV (south); equinox Convection bothFigures and7 is in an agreement in 6 -1.1 kV (N),-0.9 kV (S); winter-3.2 kV (N),-2.3 kV withtheSvalgaard-Mansurov which dffect, indicates a
factor from Table ! results in the comparable poten-

.

(S). Amplification thesevalues the correspondingdifferent direction of flows in the northern and south. of by ern polar capsfor the samesignof By.. The central tial drops.The average cross-polar potentialchange for vortex in the northernhemisphere well developed is dura 1-nT step of the northward IMF is -3.9 kV. There ing all seasons the year;its centershiftswith a seaof
is still a recognizable seasonalasymmetry: the "sum- son from magnetic to 850 pole latitude along 1100 the mer north" potential equals -3.2 kV, and the "winter MLT meridian. This movement more complicated is in

south"one equals,.-5.8 kV. The other pair ("summer the southern hemisphere. northern The nightside posisouth"and "winternorth") havecloser values:-4.5 kV tive (negative Figure7) vortexbecomes in largerduring and -4.8 kV. Reiff et al. [1981] theorized that they may equinox and winter. The southern potentialpatterns
overestimatethe ~35 kV backgroundpotential drop by are more structured. The near-cuspDPY currentsare as much as 5 kV becauseof the presence an embed- well developed of during all serous except the southern

dedreverse convection (i.e., Bz > 0) in the analyzed equinox. nightside cell The :"DPY-type" currents concendata. This value is comparablewith our -3.9 kV per
+1 nT of Bz.

trate during equinoxand winter near 800 -850 in both hemispheres.

The "cross-cusp" potential(Figure6) inferred from the IZMEM for the.IMF positiveBy duringthe disFigures 6 and 7 showthe single-cellconvection pat- turbedconditions summer kV (north),-1.6 kV are 1.7 ternsfor the IMF azimuthalBy component duringthe (south),equinox kV (N), -0.2 kV (S), winter3.2 2.0 disturbed(Bz < 0) conditions.Similarpatternsdur- kV (N), -1.9 kV (S). The same values during quiet the
Modeled Electric Potential for By
12 Nor[bern Summer
60*

12

Northern

12

Northern

/·quinox
18 06 18

Winter
06

18

06

Max Min

0.1 -1.7

O0 MLT
12 Southern
Summer

lain

-2.0

O0 lILT

Max Min Southern

1.8 -3.0

O0'IALT
12 Southern

12

./···quinox
06 18 06 t8

Winter
06

Max Min

1.5 · -0.5

Max

3.1

O0 lILT

lain

-0.8

O0 la·LT

Min

-0.7

O0 lILT

positive azimuthal (By - +1 nT) component. IMF Contour intervals equal kV. 0.3

Figure6. Modeled electric potential distributions disturbed < 0) conditions the during (Bz for


PAPITASItVILI
12 Northern

ET AL.: ELECTRIC
12

POTENTIAL
Northern

PATTERNS
12

13,259
Northern

Summer
60 Ü

Equinox
60 Ü

'·

Winter

18

06

18

06

18

06

Max Min

1.7 -0.1

Max

2.0

00 MLT

Min

-0.7

00 hiLT

Max Min

30 -1.8

O0 MLT

12

Southern

12

Southern

12

Southern

Summer

Equinox
-60 Ü

,?*
18

-60 Ü

,_.-··%xWinter
?":..\
18
06

06

18

06

Max Min

0.4 -1.5

00 hiLT

Max Min

1.7 -0.1

00 hiLT

Max Min

2.9 -1.9

00

MLT

Figure 7. The sameelectricpotentialdistributions on Figure6 but for the negative as azimuthal

IMF (By --1

nT). Contour intervals equal0.3 kV.
of HeppnerandMaynard[1987] HairstonandHeelis or [1990]. Eachof theseelementary convection repcells
resents a separate component of the IMF interaction

conditions (not shown)are summer1.7 kV (N),-1.7 kV (S), equinox kV (N),-0.2 kV (S), winter1.8kV 1.9 (N),-1.5 kV (S).
The same cross-cusp potential in the northern hemi-

with the magnetosphere its ionospheric and manifesta-

of sphere negative for IMF By areopposite that in Fig- tion. The final pattern is a superposition the elemento ure 6 becausethe northern geomagnetic data were not

tary cellsfor any givensituation in the IMF. Heppner

two-cell conseparated the regression for analyses the signof By. andMaynard[1987]statedthat standard by The southern cross-cusp potential (Figure7) for the dis- vection pattern can be rotationally twisted clockwise
strong, and the sign and turbed conditions are summer 2.0 kV, equinox 1.0 kV, when the IMF Bz becomes of role in the disand winter 3.1 kV. Samevaluesfor the quiet conditions magnitude By will play a significant

(not shown)are summer kV, equinox kV, and 2.2 1.1
winter 2.8 kV. Table 1 presentsall these values after amplification. It seemsthat a division of the original geomagnetic data accordingto the IMF disturbed and quiet conditions plays a lesserrole than a division in accordance

tortion.

Their

model shows a sunward convection

in

the daysidepolar cap as a deformationof the two-cell pattern. After a massiveanalysis of satellite data the authors concludedthat "...the nightsidedilemmasthat plaguethree- and four-cellmodelsdesigned explain to sunwardconvection polar regions in under +Bz condi-

with the IMF By sign. The average cross-cusp potentim change a 1-nT stepof the IMF By component for equals 4-5.0 kV for positiveBy, and 4-3.9kV for negative By, that is, the IMF By > 0 component causes largergeomagnetic disturbances than By < 0. On the
average,the IMF azimuthal component produces ~4.5

tionsdo not appear..."[Heppner Maynard,1987,p. and 4467]. A similarconclusion: "...the clockwise rotation of the potentialpatternwith increasing (positive) Bz
while the conductivity remains constant..." was made

by Feidstein Levitin[1986, 1170] and p. fromthe analysisof ground-based geomagnetic data only. The results

kV of the cross-cusp potential changes 1 nT of By. presentedhere confirm these conclusions. to in
The asymmetryin geometryof convection patternsin both hemispheres may be causedby an applicationof Discussion the northernionospheric conductivitymodel, and perhapsalsoby a natural "north-south" asymmetryin the The elementaryconvection patterns presented this electricpotential. The southern in geomagnetic pole, for paper should not be compareddirectly with existing example,is locatedasymmetrically againstthe northempirical convectionmodels, for example, the models ern one;southernpolar magneticlocal midnightoccurs


13,260

PAPITASHVILI ET AL.: ELECTRIC POTENTIAL PATTERNS

Finally, the IZMEM model may be usedin comparison with the results of other modelingtechniques.For spheric conductivity should slightlylargerfor "sum- example, the AMIE technique usesan initial distribube mer australis" than for "summer borealis". The latter tion of electric potential, then ground-based geomagis not applicable for equinoctial months and for win- netic data and additional satellite and radar measureters in both hemispheres.For example, we found the ments are subjected to the analysis. We believe that "northernbackground" potentialsmallerduringsum- AMIE techniquewill provide a good output if a good mer than the "southern background" (30.4kV and ground-basedcoverageis achieved. However, Arctic one 37.8 kV respectively; Table 1), but both potentials oceanand Antarctic large land masswill neverbe propsee areequalduringwinter.Onecanseefromequation (3) erly coveredwith geomagneticobservations.So, there for that the southernsummerpotentialshouldbe larger is a problemof getting good global coverage AMIE if the smaller(due to the influence the Earth's or- modeling. The IZMEM model can providean initial esof bit ellipticity)northernsummer ionospheric conductiv- timation of global potential distributionsas a valuable ity is applied to the southern hemisphere. The rel- input for the AMIE technique at least when the IMF ative perihelion-aphelion difference in the Sun-Earth data are available. The IZMEM model may analyze geomagnetic distance is about 3 %, which means that the corre- particular eventswhen the ground-based sponding illumination (andhence conductivity) polar data collectionis impossibleor significantlydelayed. of The IZMEM model has also a "now-casting" capaionospheres should differ by about 6%. bility providing realistic convectionpatterns over the The IZMEM can successfully model magnetic and electric fields measuredby satellites,for example, Cos- entire polar regions. Since there are low-altitude satel-

at 1530 UT - only 10 hours past the local midnight at the north geomagneticpole at 0530 UT. Ellipticity of the Earth's orbit may also enhance the asymmetry of the polar cap potential distributionsbecausethe planet is in its perihelionin the beginningof January and the southernpolar cap is lighted better. Thereforethe iono-

during this period, the model can not be applied. The
latter conclusion is not firm in all cases and we continue

to investigatecapabilitiesof the IZMEM to model such
short-term events.

(e.g., DMSP data), the mos 184 [Belovet al., 1984], Magsat [Dremukhina lite plasmadrift measurements et al., 1985], OGO 6 [Feldstein and Levitin, 1986], DE electric potential inferred from these data can be compotential pro2 [Drernukhina al., 1990]. It was found that the pared with a set of modeled cross-polar et IZMEM output valuesare smaller,in general,than mea- files along satellite trajectories, which may represent
a number of situations in the IMF; for example, electric potential patterns can be modeled in advance for

The modeledprofile similar to the observed data gives an approximateestimation of the IMF valuesduring In conclusion, summarizeaverage we valuesof electric the satellite pass, and, therefore, the IZMEM can re(now-cast) entire convection the patternsover potential causedby 1-nT step in the IMF as they are produce modeled by IZMEM with the use of amplification fac- both polar regions. Another example is that if the IMF (e.g., from tors: ~14 kV for the IMF southward Bz, ~-4 kV for data will be availablean hour in advance the L1 satellite),the IZMEM canpredictthe potential the IMF northward Bz, and ~ 4-4.5 kV for the IMF azimuthalBy components.The average background pattern configurationand magnitude in both polar recross-polarpotential for the zero IMF is ~37 kV. A gions. The IZMEM exhibits these capabilitiesfor the combinationof the elementarycellsallowsone to model large-scale,quasi-steadyevents, but the model cannot quantitatively the convectionpatterns for different con- forecastmagneticsubstorm. ditions in the IMF. For example, a combinationof the Acknowledgments. We would like to thank Robert viscousand mergingcellsfor B, - -5 nT gives 107 Clauer for valuable comments, discussion, and help with kV of the dawn-dusk potential drop. This value may preparation of the paper. Marie Cooper has provided asfirmed or corrected in future studies.

surementsby factor ~3.0 - 3.5. Similar values are obtained for the summer seasonin this paper and the correspondingamplification factors are estimated for equinox and winter. These numbers should be con-

each1-nT stepbetween -30 to +30 nT of Bz and By.

change drastically strongBy. The background for po- sistancein programmingon IDL. This work has been sup-

tential will be erodedduringpositiveB,, but it will not become zero for Bz = 7 nT becausethe corresponding electric field developsin the daysidepolar cap only. Large reversecross-polar potential ~80 kV may develop for Bz = +10 nT and By - 4-10nT. The IZMEM model can be used to investigatethe particular case studiesbut it is limited to specifying large-scale, quasi-steady events. While it worksfine for hourly mean values, it can also be used for studying time-varying phenomena; in that case a proper time delaybetweenthe IMF changes and their manifestation
on the Earth's surface should be taken into account.

ported by the Russian Foundation for Fundamental Research under the grant 93-05-8722. One of the authors

(V.O.P.) hasbeensupported the University Michigan by of
under funding from the National ScienceFoundationgrant ATM-9106958. V.O.P. especially thanks both refereesfor their assistance the evaluationof the paper'sclarity and in language.
The Editor thanks two referees for their assistance in eval-

uating this paper.

References
Afonina, R. G., B. A. Belov, A. E. Levitin, M. Yu. Markova, Ya. I. Feldstein, and M. V. Fiskina, Connection of components of the interplanetary magnetic field with field-

The "geomagnetic activity history" during last 20-40 min is alsoimportant, for example,if a substormoccurs


PAPITASHVILI

ET AL.: ELECTRIC

POTENTIAL

PATTERNS

·3,26·

aligned currents in high-latitudes of the northern hemisphere, Geomagn. Acton., oe0,1073, 1980. Alcayde, D., G. Caudal, and J. Fontanari, Convection electric fields and electrostatic potential over 61Ü < A < 71Ü invariant latitude results, J. Geophys. Res., 91, 233, 1986. de la Beaujardiere, O., D. Alcayde, J. Fontanari, and C. Leger,Seasonal dependence high-latitudeelectricfields, of J. Geophys.Res., 96, 5723, 1991. Belov, B. A., R. G. Afonina, A. E. Levitin, and Ya. I. Feldstein,The effectsof interplanetary magneticfield orientation on geomagneticfield at the northern polar cap

Heppner, J.P., Empirical models of high-latitude electric fields, J. Geophys.Res., 8oe,1115, 1977. Heppner,J.P., and N. C. Maynard, Empirical high-latitude electric field models, J. Geophys.Res., 9oe,4467, 1987. Kamide, Y., A.D. Richmond, and S. Matsushita, Estimation of ionosphericelectric fields, ionosphericcurrents and field-alignedcurrentsfrom groundmagneticrecords, J. Geophys.Res., 86, 801, 1981. Kern, J. W., Analysis of polar magnetic storms, J. Geo. magn. Geoelectr., 18, 125, 1966. Knipp, D. J., A.D. Richmond, B. Emery, N. U. Crooker, O. de la Beaujardiere, D. Evans, and H. Kroehl, Iono(in Russian), Variations MagneticField and the A uin of rora, edited by Ya. I. Feldstein and A. E. Levitin, p. 15, sphere convection responseto changing IMF direction, IZMIRAN, Moscow, 1977. Geophys.Res. Lett., 18, 721, 1991. Belov, B. A., Yu. I. Galperin, L. V. Zinin, A. E. Levitin, Knipp, D. J., et al., Ionospheric convection response slow, to R. G. Afonina, and Ya. I. Feldstein, Plasma convection strong variationsin a northwardinterplanetary magnetic in the polar ionosphere: A comparison of the satellite field: A casestudy for January 14, 1988, J. Geophys.Res.,
Cosmos-184 measurements and the model related to the

98, 19273, 1993.

IMF vector, Kosm. lssled., oeoe, 201, 1984. Levitin, A. E., R. G. Afonina, B. A. Belov, and Ya. I. FeldBlomberg,L. G., and G. T. Marklund, High-latitude convecstein, Geomagneticvariationsand field-alignedcurrentsat tion patterns for various large-scalefield-aligned current northern high-latitudes and their relations to solar wind configurations, Geophys.Res. Lctt., 18, 717, 1991. parameters, Phil. Trans. R. Soc. London Ser. A, 253, 1982. Clauer, C. R., The techniqueof linear prediction filters applied to studies of solar wind-magnetospherecoupling, Maezawa,K., Magnetospheric convection inducedby the in Solar Wind-Magnetosphere Coupling, edited by Y. interplanetarymagneticfield: Quantitative analysisusing Kamide and J. A. Slavin, p. 39, Terra Scientific, Tokyo, polar cap magnetic records, J. Geophys.Res., 81, 2289,
1986.

1976.

Clauer, C. R., and E. Friis-Christensen,High-latitude day- Mansurov, S. M., O. A. Troshichev, A. N. Zaitzev, V. O. side electric fields and currents during strong northward Papitashvili, G. A. Timofeev, and M. A. Kandibolotskaya, interplanetary magnetic field: Observations and model Characteristicsof magnetic disturbances producedby the simulation, J. Geophys.Res., 93, 2749, 1988. IMF in the southernpolar cap, Geomagn.Aeron., Eng]. Dremukhina, L. A., Ya. I. Feldstein, and A. E. Levitin, Transl., oe1,428, 1981. Model calculationsof currents and magnetic fields along Mishin, V. M., M. I. Matveer, G. B. Shpynev, A.D. Bazara Magsat trajectory, J. Geophys.Res., 90, 6657, 1985. zhapov, and V. K. Kompanets, Preliminary results of Dremukhina, L. A., A. E. Levitin, and Ya. I. Feldstein, the three-dimensional current system calculations from Convectionin high latitudes in periodsBz > O, Geomagn. the ground-based geomagnetic disturbances Russian), (in Aeron., Engl. Transl., 30, 160, 1990. lssled. Geomagn. Aeron. Fiz. Solntsa,J3, 14, 1977. Faermark, D. S., A restoration of 3-dimensionalcurrent sys- Mishin, V. M., A.D. Bazarzhapov,and G. B. Shpynev,Electemsin high-latitudesby the useof ground-based geomagtric fields and currents in the Earth's magnetosphere, in: netic observations,Geomagn. Aeron., Engl. Transl., 17, Dynamics of the Magnetosphere, edited by S.-I. Akasofu, 114, 1977. pp. 249-268, D. Reidel, Norwell, Mass., 1980: Feldstein,Ya. I., A. E. Levitin, D. S. Faermark, R. G. Afon- Papitashvili,V. O., Relationshipbetweengeomagnetic variina, B. A. Belov, and V. Yu. Gaidukov, Electric fields and ations in the polar cap and the interplanetarymagnetic potentials in the high-latitude ionosphere different sitfor field during the solar activity cycle, Geomagn. Aeron., uations in the interplanetary space, Planet. SpaceSci., Engl. Transl., oeoe, 130, 1982. 3oe, 907, 1984. Papitashvili, V. O., and C. R. Clauer, Comparison and calFeldstein, Ya. I., and A. E. Levitin, Solar wind control of ibration of the modeledhigh-latitude convectionpatterns electricfields and currentsin the ionosphere,J. Geomagn. with observationsfrom the January 27-29; 1992 GEM Geoelectr., 38, 1143, 1986. campaign, Eos Trans. A GU, 7J, 254, 1993. Foster, J. C., An empirical electric field model derived from Papitashvili, V. O., O. A. Troshichev, and A. N. Zaitzev, Chatanika radar data, J. Geophys.Res., 88, 981, 1983. Linear dependence the intensity of geomagnetic of variaFriis-Christensen,E., and J. Wilhjelm, Polax'cap currents tionsin the polar regionon the magnitudesof the southern oeor differentdirections the interplanetarymagnetic of field and northern components the interplanetary magnetic of in the Y- Z plane, J. Geophys. Res., 80, 1248, 1975. field, Geomagn. Aeron., Engl. Transl., oe1,565,'1981. Friis-Christensen, E., Y. Kamide, A.D. Richmond, and S. Papitashvili, V. O., A. N. Zaitzev, and Ya. I. Feldstein, Magnetic disturbancesgenerated by the interplanetary Matsushita, Interplanetary magneticfield controlof highlatitude electric field and currents determined from Greenmagneticfield in the southernpolar cap during the summer, Geomagn. Aeron., Engl. Trans]., oe3,506, 1983. land magnetometer data, J. Geophys. Res., 90, 1325, 1985. Papitashvili, V. O., B. A. Belov, and L. I. Gromova, Fieldaligned currents and convectionpatterns in the southern Hairston, M. R., and R. A. Heelis, Model of high-latitude polar cap during stable northward, southward, and azionosphericconvectionpattern during southward interimuthal IMF, IEEE Trans. Plasma Sci., 17, 167, 1989. planetary magneticfield usingDE 2 data, J. Geophys. Res., Papitashvili, V. O., Ya. I. Feldstein, A. E. Levitin, B. A. 95, 2333, 1990. Belov, L. I. Groinova, and T. E. Valchuk, Equivalent ionoHeelis, R. A., J. K. Lowell, and R. W. Spiro, A model of sphericcurrents aboveAntarctica during the austral sumthe high-latitude ionospheric convectionpattern, J. Geomer, Antarct. Sci., oe,67, 1990. phys. Res., 87, 6339, 1982. Heppner, J.P., Polar cap electric field directions related Papitashvili, V. O., B. A. Belov, Ya. I. Feldstein, S. A. Golyshev, L. I. Gromova, and A. E. Levitin, Model of to the interplanetary magnetic field direction, J. Geo. relationsbetweenthe solar wind and IMF parametersand phys. Res., 77, 4877, 1972.
ú


13,262

PAPITASHVILI

ET AL.: ELECTRIC

POTENTIAL

PATTERNS

geomagnetic variations at high-latitudes, in Proceedings o] Solar-Terr. Predict. IV Workshop, Ottawa, Canada, May 18-oeoe, 199oe, vol. 3, edited by J. Hruska, M. A. Shea, D. F. Smart, and G. Heckman, p. 229, National Oceanic and Atmospheric Administration, Boulder, Colo., 1993. Reiff, P. H., and J. L. Burch, IMF By-dependent plasmaflow and Birkeland currentsin the daysidemagnetosphere, 2, A global model for northward and southwardIMF, J. Geo. phys. Res.., 90, 1595, 1985. Reiff, P. H., R. W. Spiro, and T. W. Hill, Dependence of polar cap potential drop on interplanetary parameters, J. Geophys.Res., 86, 7639, 1981. Richmond, A.D., and Y. Kamide, Mapping electrodynamic features of the high-latitude ionosphere from localized observations: Technique, J. Geophys.Res., 93, 5741, 1988. Robinson, R. M., and R. R. Vondrak, Measurement of E

region.ionization and conductivity produced by solar illumination at high latitudes, J. Geophys.Res., 89, 3951,
1984.

Troshichev, O. A., Polar magnetic disturbancesand fieldaligned currents, SpaceSci. Rev., 3oe,275, 1982. Wallis, D. D., and E. E. Budzinski, Empirical models of height-integrated conductivities, J. Geophys. Res., 86,
125, 1981.

V. O. Papitashvili, B. A. Belov, D. S. Faermark, Ya. I. Feldstein, S. A. Golyshev, L. I. Gromova, and A. E. Levitin, IZMIRAN, Troitsk, Moscow Region, 142092, Russia.

(Received August24, 1993;revisedMarch 21, 1994; accepted March23, 1994.)