Radar studies of mid-latitude ionospheric plasma drifts

Radar studies of mid-latitude ionospheric plasma drifts
L. Scherliess, B. G. Fejer, J. Holt, L. Goncharenko, C. Amory-Mazaudier, M.
J. Buonsanto
To cite this version:
L. Scherliess, B. G. Fejer, J. Holt, L. Goncharenko, C. Amory-Mazaudier, et al.. Radar studies of
mid-latitude ionospheric plasma drifts. Journal of Geophysical Research, American Geophysical
Union (AGU), 2001, pp.1771-1783. <hal-00979238>
HAL Id: hal-00979238
http://hal.upmc.fr/hal-00979238
Submitted on 15 Apr 2014
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destin´ee au d´epˆot et `a la diffusion de documents
scientifiques de niveau recherche, publi´es ou non,
´emanant des ´etablissements d’enseignement et de
recherche fran¸cais ou ´etrangers, des laboratoires
publics ou priv´es.
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 106, NO. A2, PAGES 1771-1783, FEBRUARY
Radar studies of midlatitude
1, 2001
ionospheric plasma drifts
L. Scherliess
•, B. G. Fejer•, J. Holt2, L. Goncharenko
2,
C. Amory-Mazaudier
a, and M. J. Buonsanto
2,4
Abstract.
We use incoherent
scatter
radar
measurements
from Millstone
Hill
and
Saint Santin to study the midlatitude F regionelectrodynamicplasma drifts during
geomagneticallyquiet and active periods. We present initially a local time, season,
and solar flux dependent analytical model of the quiet time zonal and meridional
ExB drifts over these stations. We discuss,for the first time, the Saint Santin drift
patterns during solar maximum. We have used these quiet time models to extract
the geomagneticperturbation drifts which were modeled •s • function of the time
history of the auroral electrojet indices. Our results illustrate the evolution of the
disturbance drifts driven by the combinedeffectsof prompt penetration •nd longer
lasting perturbation electric fields. The meridion•l electrodynamic disturbance
drifts have largest amplitudes in the midnight-noon sector. The zonal drifts •re
predominantly westward, with largest amplitudes in the dusk-midnight sector •nd,
followinga decreasein the high-latitude convection,they dec•y more slowlyth•n the
meridional drifts. The prompt penetration •nd steady state zonal disturbance drifts
derived from radar measurements•re in good agreement with results obtained from
both the ion drift meter data on board the DynamicsExplorer 2 (DE 2) satellite
•nd
f?om the Rice Convection
1.
Introduction
Model.
The midlatitude ionospheric plasma drifts can be simultaneously driven by several electric field sources.
The two most important ones are the ionosphericwind
dynamo and the direct penetration of high-latitude convection electric fields into the plasmasphere. The ionospheric wind dynamo generateselectric fields through
the dynamo action of the thermospheric wind circulation driven by solar heating and during geomagneticactive periods, also by thermospheric disturbance winds
driven by enhanced Joule heating in the auroral re-
Over the last two decades, incoherent scatter radar
observationshave been used extensively to study the
midlatitude ionosphericelectrodynamic(ExB) plasma
drifts during geomagnetically
quiet and disturbedperiods. Blanc and Amayenc [1979] have used incoherent scatter radar observationsduring geomagnetically
quiet days over Saint Santin during the 1973- 1975
solar minimum period to derive an analytical seasonal
model of their quiet time F region plasma drifts. This
model included a steady drift componentand the first
four diurnal harmonicoscillations. Wand [1981] has
used Millstone Hill observations during May 1976 to
November 1977 to analytically model the upper mid1987; Mazaudier and Venkateswaran, 1990; Scherliess latitude solar minimum electrodynamic plasma drifts
gions [Blanc and Richmond, 1980; Mazaudier et al.,
and Fejer, 1997]. Prompt penetrationelectric fields for geomagneticallyquiet and disturbed periods. The
can affect the midlatitude plasma drifts nearly instan-
disturbance
taneouslyduring transient periodsof insufficientshield-
Kp index. Buonsantoet al. [1993]haveuseda significantly larger databaseof Millstone Hill plasma drift
ing by plasmasheetcurrents[Jaggiand Wolf,19731and
drifts
were modeled
as a function
of the
also during periodsof strongsteady auroral convection. observationscomprising73 experimentsfrom February
1984 to February 1992 and determined the averageseasonal quiet time plasma drift patterns for both solar
•Center for Atmospheric and Space Sciences,Utah State
University, Logan.
2HaystackObservatory,MassachusettsInstitute of Technology, Westford.
aCentre
d'Etude
des Environnements
Terrestre
minimum
and maximum
conditions.
The
characteris-
tics of equatorial, low-latitude, and midlatitude plasma
drifts
measured
with
incoherent
scatter
radar
were re-
viewedby Richmond[1995]. More recently,Buonsanto
and Witasse[1999]presentedan updatedstudy of the
local time, season,solar cycle, and geomagneticactivity
dependent climatologiesof the Millstone Hill F region
plasma drifts and thermosphericwinds. In this study,
magneticactivity effectswere again accountedfor as a
et Plane-
taires, Saint-Maur-de-Fosses, France.
4DeceasedOctober 20, 1999.
Copyright2001 by the AmericanGeophysicalUnion.
Paper nulnbcr 2000JA000229.
0148-0227/01/2000JA000229509.00
function of the Kp index.
1771
1772
SCHERLIESS
ET AL.'
MIDLATITUDE
A detailed study of the middle- and low-latitude F
region zonal plasma drifts measured by the Ion Drift
Meter (IDM) on boardthe polar orbitingDynamicsExplorer2 (DE 2) satelliteduringthe 1981-1983solarmax-
PLASMA
DRIFTS
three directions and with the assumption of a uniform
velocityfield [e.g., Buonsantoet al., 1993; Buonsanto
and Witasse,1999]. The measurements
in our database
correspondto an altitude of 300 kin, have a typical er-
imum periodwaspresentedby Heelisand Coley[1992]. ror of •08 m/s, and a cycletime of •-l hour. We will use
They derived the average latitudinal plasma drift pat- here drift componentsperpendicular to the Earth magterns during geomagneticallyquiet and disturbed peri- netic field. In the F region over Millstone Hill, a plasma
ods correspondingto Kp _(2.0 and Kp_)3.0, respectively. velocityof 22 m/s corresponds
to an electricfield of 1
They could not derive the seasonaldependenceof these
drifts
since the DE 2 measurements
cal time
had season and lo-
interlocked.
Recently,Fejer and $cherliess[1995]haveintroduced
The Saint Santin incoherentscatterradar site (45øN,
2øE, 40ø magnetic latitude) operated originally on a
bistatic
mode with
a transmitter
in Saint Santin
and
a new methodology which was able to separate,for the a receiver at a distance of 302 km in Nancay. In 1973,
first time, the effects of different electric field sources two additional receiverswere installed in Monpazier and
on the plasma drifts during geomagnetic active peri- Mende (•-100 km from the transmittersite), allowing
ods. $cherliessand Fejer [1998]and Fejer and $cher- for the determination of three-dimensional ion drift vecliess[1998]have usedthis methodologyto analytically tors at a typical heightof •0300km [Baueret al., 1974].
errorwasof the orderof 5-10m/s and
model the temporal and latitudinal variations of the The measurement
low- and middle-latitude F region zonal perturbation the integration time was •030 min. The Saint Santin
drifts measured by the DE 2 satellite during geomagnetically disturbed periods as a function of the time
history of the hourly averaged AE indices. They determined the drift patterns owing to transient prompt
penetration electric fields and of longer lasting distur-
radar operated until 1986. Over Saint Santin, a plasma
velocityof 25 m/s in the F regioncorresponds
to an
electricfield of 1 mV/m.
We have used 15 years of Millstone Hill incoherent
scatter radar observations from 1978 to 1992, comprisbances(time constantslongerthan a few hours)because ing more than 4500 hours of plasma drift observations,
of the combined effects of ionosphericdisturbance dy- which have been recently reanalyzed. We have also
namo electric fields and leakageof high-latitude electric used 14 years of plasma drift observations from the
Saint Santin incoherent
fields to lower latitudes.
We have used extensive
incoherent
scatter radar ob-
which were extracted
scatter radar from 1973 to 1986
from the National
Center for At-
mospheric Research Coupling, Energetics and Dynam-
servationsfrom Millstone Hill and Saint Santin to study
midlatitude F region zonal and meridional plasma drifts
during geomagnetic quiet and active conditions. We
initially present a comprehensivesummary of the local
time, seasonal, and solar cycle variations of the quiet
icsof AtmosphericRegions(NCAR CEDAR) database.
We discarded Saint Santin late afternoon and nighttime drift observations(1600-0700 LT) obtained be-
time drifts over these two midlatitude
expectedlylarge nighttime drifts (westwardvelocities
stations.
This in-
tween September 1981 and March 1982 owing to un-
quiet
cludes,for the first time, the seasonaldependenceof the exceeding200 m/s, evenduring geomagnetically
average solar maximum plasma drifts over Saint Santin. periods). The reasonfor theselarge drifts is unknown
The main focus of this work is the study of the tempo- to us. The average values of the solar decimetric flux
ral evolution of the midlatitude plasma drifts during indices for the Millstone Hill and Saint Santin data sets
geomagneticactive periods using the methodologyin- are 140 and 125, respectively. The number and dates
troducedby Fejer and $cherliess[1995].We haveused of the Millstone Hill and Saint Santin experiments used
our quiet time drift model to determine the disturbance
in this study can be obtained from the NCAR CEDAR
drifts which were then modeled
database.
as a function
of the AE
index.
In the following sections,we will first briefly review
the radar measurement techniques and describe our
data analysis. Then we present our quiet time model,
describe the storm time dependent disturbance drift
patterns and compare them with results from earlier
studies.
2. Measurement
Analysis
Techniques and Data
2.1.
Quiet Time Analysis
We have used these data to determine empirical analytical modelsfor both the quiet time and disturbance
plasma drifts over these midlatitude stations. For our
quiet time plasma drift models, we have included only
data with a current 3-hour magnetic index Kp<3.0
(Kp -•1.8). We will discusslater the implicationsof
this relatively relaxed quiet time criterion.
The electrodynamiccomponentsof the plasma drifts
(perpendicular/eastward
and perpendicular/northward
The MillstoneHill radar (42.6øN,288.5øE,Apex magto B) were modeledas:
netic latitude 54ø) measuresthe line-of-sightplasma
drift velocity from the Doppler shift of the 440 MHz
backscatter signal. The three-dimensionaldrift velocity over the radar site is obtained from measurements in
8
6
v(t,d,Sa)- Z • ai,kfkNi,4(t)
i=1
k=l
(1)
SCHERLIESS
ET AL.:
MIDLATITUDE
with
PLASMA
DRIFTS
1773
equatorward edge of the shielding layer is poleward of
the radar
fl = 1
May-August
f2 = 1
fa = 1
November-February
March-April, September-October
f4 -- f• (Sa - 140)
f5 = f2 ($a- 140)
f6 = fs ($a- 140)
Here t denoteslocal time, d is day of the year, and $• is
the daily 10.7 cm solar flux index. Outside their above
definedlimits, the functionsf• to f3 are set to zero. The
site.
We further
minimized
the inclusion
of
auroral drifts, and also of unusuallylarge drifts (associatedwith subauroralion drifts (SAIDs), for example)
into our database, by limiting the magnitudes of the
perturbation drifts to values smaller than 150 and 80
m/s for the zonal and perpendicular/northward
drift
component, respectively.
Following the procedure describedby Scherliessand
Fejer [1998], we have characterizedthe level of enhanced high latitude geomagneticactivity, using modified hourly averagedauroral electrojet (JEd) indices,
with AEd = AE- 160 nT. For the time period prior to
July 1988, we have used the standard AE index based
coefficients
ai,1 to ai,3 determinethe basicseasonal
patternsof the electrodynamic
driftsfor a solarflux of 140, on measurements from 12 auroral stations and, for the
and the coefficients
ai,4 to ai,6 representthe linearvari- later period, a provisional index based on 11 stations.
ations of these drifts with the solar flux. The local time
For the secondhalf of 1988 and also for 1989 (with
dependenceis describedby eight normalized cubic-B the exception of March) AE indicesare not available,
splinesof order four with equallyspacednodesat 0, 3, and, consequently, data from this period have not been
6, ..., 21 hours[e.g.,Fejer and$cherliess,
1997].These used in our perturbation drift analysis. The AE in-
basicfunctionsare well suitedto describeplasmadrift
variations up to terdiurnal oscillations. The model coef-
dex has been empirically related to both the polar cap
crosspotential and to the hemispheric high-latitude en-
ficientshavebeendeterminedby weightedleastsquares ergy input [Jhn et al., 1983, 1992]. This allowsus to
fits to the individual data sets, with each individual compare our empirical disturbance patterns with results
observationweightedby the inversesquareof the mea- from global convection models.
surementerror. However,in order not to overemphasize
For each radar station the ionospheric perturbation
measurementswith relatively small errorbars, which do drift components were expressedas
not necessarilyrepresentthe averagequiet time pattern,
6
we have set errorbarssmallerthan one half of the dayto-day variability, obtained by the standard deviation
v(t)- • [ai,iAAE(t
- 30rain)+
i--1
of binned data presentedbelow, equal to this value.
We have observed that variations of this criterion do
ai,2AAE(t - 90 rain) +
not changeour model resultsin any significantway and
ai,3AEd(1 - 3 hour) +
that evenan unweightedfit leadsto essentiallythe same
(2)
ai,aAEd(q- 9 hour) I Ni,•(t).
drift patterns.
Finally, a simple linear interpolation schemeover a The first two ternas under the summation resemble the
range of 4- 15 days was employedfor the transition be- parametrizationusedby Scherliessand Fejev [1998]to
tween our 4-month seasons.This interpolation provides accountfor prompt penetration drifts related to changes
a reasonably realistic transition between seasons.
2.2.
Storm Time Analysis
As mentioned above, our quiet time model drifts correspond to an average Kp index of •-1.8. In the next
step, we have determined the perturbation drift patterns during geomagneticactive periods and their storm
time evolution. To extract the perturbation drifts out
of our data set, we have subtracted the seasonand solar cycledependentaveragequiet time drifts, usingour
new empirical quiet time models, from the measured
in the auroral current systemswith averagetime delays
of 30 and 90 min: respectively. To account for longer
lasting drift perturbations, which have been parame-
terized by $cherliessand Fejer [1998]by an averageof
AE• indices over the past nine hours, we have instead
used two separate parameters, i.e., the averageof AE•
indices over the past 1 - 3 hours and the average over
the past 4- 9 hours. This parametrization significantly
improves the temporal evolution of the longer lasting
perturbations and provides a better representationof
the perturbation drifts during the initial hours of geodrifts (for a detailed descriptionseeFejer and Scherliess magneticactive periods. After severalhoursof auroral
[1997]). The resultingperturbationsare due to storm activity, however,this parametrization givesessentially
time generatedelectric fields and alsoto the day-to-day the same results as presented by $cherliess and Fejer
[1998]. The limitations of the use of the AE• indices
variability of the ionosphericdynamo electric fields.
Sinceduring large nighttime storm periods, Millstone in our analysiswere discussedby Fejer and $cherliess
Hill can be in the auroral zone, we have excludedplasma [1997,1998]. Our empiricalmodels,basedon the AE•
drift observationswhen the maximum Kp index over the index, do not take into accountsomepotentially imporpreceding period of 6 hours was above 5.0. This crite- tant processesand averageout the effectsof a number
rion restricts our Millstone Hill storm time analysisto of ionosphericand magnetosphericprocessesthat could
moderately disturbed periods, i.e., to periods when the play important roles on the magnitude and phase of
1774
SCHERLIESS
ET AL'
MIDLATITUDE
PLASMA DRIFTS
MILLSTONE HILL
MAY-AUG
MAR-APR
60
Kp<_3
SEP-OCT
NOV-FEB
Sa=180
-
.. Sa=190
40
-
Sa=200
20
.• •
0
-20
_
"
•
,
T T
_
_
-40
__
40
Sa=85
_
Sa=85
2O
-20
_
111]
'1 '111
_
.,
_
,._.
-40
I IIIIIIIIIIIII111
00
04
08
i
IIIIII
12
16
20
24
00
04
08
12
LOCAL
16
20
24
00
04
08
12
16
24
20
TIME
Figure 1. Quiet time averageperpendicular
northwardplasmadrifts overMillstoneHill for
low and highsolarflux conditions.The
solidcurvesindicatethe velocitypatternsobtainedfrom
our empirical model. Here $a denotesthe averagedecimetricsolarflux index.
the electric field perturbations as well as on the shielding time constants. For example, results presentedby
3.1.
Quiet Time Electrodynamic
Drifts
Fosteret al. [1986]suggest
that IMF By mightcause
Figure I showsthe average seasonalquiet time pat-
Millstone Hill. A comprehensivestudy of these additional processesrequires a significantly larger database
than currently available.
In summary, we have developed individual empirical analytical models for the Millstone Hill and Saint
Santin electrodynamicalplasma drifts for geomagnetically quiet and disturbed periods. These models include
the seasonaland solar cycle variations of the quiet time
drifts as well as short-term and longer lasting perturbation drifts during geomagneticallydisturbed periods.
Combined, these models can describethe averagemidlatitude plasma drifts during various geomagneticcon-
ponent over Millstone Hill for low and high solar flux
plasmadrift comlarge changes in the perturbation electric fields over ternsfor the perpendicular/northward
Results
and
Discussion
In this section we will discussbriefly our solar minimum and solar maximum average quiet time drift patterns obtained
models.
Then
from our Millstone
we will
describe
The
data
have been divided
into
summer
(May-August),equinox(March, April, September,October), and winter (November-February).
The average
decimetric
solar flux indices are •85
for the low solar
flux intervals and 180, 200, and 190, for summer, winter, and equinox high solar flux periods, respectively.
The standard deviation of our averagedrifts is between
•20 and 40 m/s. The averagenumberof samplesper
bin is •20, resulting in standard errors of the mean
between•5 and 10 m/s. Figure I indicatesthat the
perpendicular/meridionaldrifts are northward in the
early morning and prenoon sector and southward at
the later times, with a return to the northward direction in the late afternoon during low solar flux summer
and equinox conditions. The northward morning drifts
ditions.
3.
conditions.
Hill and Saint Santin
our disturbance
drift
patterns and compare them with other observationsand
also with predictionsfrom global convectionmodels.
reachmaximum valuesbetween•20 and 30 m/s, and
the afternoon
southward
drifts
have maximum
values
of •10 m/s. Although, large solar cyclevariationsare
observedin the perpendicular/northwarddrift component (e.g., equinoxearly morning), no significantsystematic effects are evident. The results in Figure I are
SCHERLIESS
ET AL.' MIDLATITUDE
PLASMA
DRIFTS
1775
solarflux data. The averageKp index is again • 1.8 with
ported by Buonsantoet al. [1993]and Buonsantoand slightlylarger valuesduring summerand equinoxhigh
Witasse[1999],whodiscussed
their main characteristics solar flux conditions. The Saint Santin drifts resemble
and large day-to-day variability. The latter work also many of the featuresobservedover Millstone Hill, with
in good agreement with the average drift patterns re-
reported differences between the vernal and autumnal
drift patterns, which were most pronouncedin the early
morning to noon period. These effectsare not being accounted for in our current study.
The thick solid lines in Figure i represent our quiet
time model predictions, which have been calculated us-
northward drifts in the prenoon sector and southward
drifts in the afternoon. In general, the solar minimum
resultsshownin Figure 2 exhibit strongsemidiurnaloscillations and closely resemblethe solar minimum pat-
ing (1) for eachdata point and binnedand averagedthe
terns presentedearlier by Blanc and Amayenc[1979]
who used 3 years of drift observations.However,they
reportedlarge downward/southward
drifts during De-
same way as the observations. As expected, our model
results are in good agreement with the average drift
spondingresultsindicateupward/northwarddrifts.
patternswith typical fluctuationsof --•1-5m/s. Systematic discrepanciesbetween model and data are observed
only during low solar flux winter afternoon conditions
when the empirical model underestimates the equator-
ward/downwarddrift velocities. These differencesare
largely due to our use of a linear solar cycle variation
cember solsticenighttime periods, whereasour correThe Saint Santin northward drifts showstrong semidiurnal patterns near solar minimum and diurnal type
variations during solar maximum. Saint Santin solar maximum drifts have not been studied earlier, and
therefore ours is the first model representation for these
drifts. Figure 2 indicates that from -•0800 to 1400
LT, the equinoctial drifts do not changesignificantly
Figure 2 showsthe averageperpendicular/northward with the phase of the solar cycle, whereasthe solsti-
in our empirical model.
quiet time plasma drifts over Saint Santin for solar minimum and maximum conditions and the corresponding
empirical model results. The average solar flux values
for summer, winter, and equinox are •85, 85, and 75 for
the low solar flux data and 180,205, and 180 for the high
tial data show opposite variations with an increasein
the northward drifts in the summer data and a decrease
in the winter. The relatively large southwarddrifts in
the morningfrom -•0400 to 0800 LT and in the afternoon from -•1400 to 1800 LT drastically decreasewith
Saint Santin
- MAR-APR
Kp<3
NOV-FEB
MAY-AUG
•
60-SEP-OCT
40:
Sa=180
Sa=180
•
_
Sa=205
_
20
_
0
T
_ -.
_
/
.-
__
-20
_
-40
40
_
,
Sa=75
_
Sa:85
20
Sa=85
_
7-
--
-20
_
_
40
-
-
_
IIit11111111111tll
00
04
08
12
16
20
24
00
04
08
12
LOCAL
16
20
24
00
04
TI ME
Figure 2. Same as Figure 1, but for Saint Santin.
08
12
16
20
24
1776
SCHERLIESS
ET AL.'
MIDLATITUDE
PLASMA
MILLSTONE HILL
6O --
MAR-APR
-
SEP-OCT
-
.,
ß
40 TM
-
Kp<3
MAY-AUG
-
-
-
-
NOV- FEB
-
-
-
Sa=180
'"•Sa=190
-
20 TM
DRIFTS
Sa-200
-
ß
'-
_
,. •/•-i•
-2O
•
-4O
__
_
._
ß
i
--
_
_
,.
..
..
..
_
__
_
.
ß
40 TM
Sa-85
-
-
Sa--85
"
--
Sa=85
20 -l
,
_
,,
.,
.
-2O
.....
./
-4O
..
,
__
.,
I II
00
I[
04
111
08
II
IIII
12
I II
16
II
I111
20
I I I I I I I I I I I I I I I I I I"1 I.I I I
24
0
04
OR
12
LOCAL
Figure 3.
1R
20
IIIII
:1. 0 (
IIIIIIIIIIIIIIII
04
08
II
12
16
20
2,
TI ME
Quiet time averagezonal plasmadrifts over Millstone Hill for low and high solar flux
conditions.
increasing solar flux during equinox and summer and
even reverse toward
the northward
direction
in the June
solsticemorning sector.
It is not clear if the December solstice drift pattern
is entirely realistic or biased by instrumental offsets.
First of all, the December solsticedaytime and nighttime drifts are significantly different from the equinoctial values. In addition, as pointed out by Takami et
flux values given for the northward component. The
zonal drifts over Millstone Hill are westward at night
with maximumvaluesof •60 m/s in winterand 40 m/s
in summer. During daytime, and particularly around
noon, the zonal drifts have eastward amplitudes with
largestmagnitudes(up to 35 m/s) duringequinox.During summer, the average zonal drifts are always westward.
The
standard
deviation
is between
•25
and 50
al. [1996], the nighttime Saint Santin winter drifts m/s, indicatinglarge quiet time variability. We will
shouldbe strongly affectedby the correspondingconjugate summer ionosphere,which lies at a geographiclatitude of 31øS. Therefore we would expect a closesimilarity between the Saint Santin December solsticeand the
show later that some of this variability was introduced
by the choice of our quiet time binning criterion. The
thick solid lines in Figure 3 represent again the predictions of our quiet time model. Generally, the zonal
Shigaraki (35øN dip latitude) June solsticenighttime plasma drifts at Millstone Hill tend to be more westward
drifts measuredby the MU radar [Takamiet al., 1996]. at night and more eastwardduring the day with increasHowever, the June solstice solar maximum nighttime ing solar flux, although, systematic variations are relaaverage drifts measured by this radar show southward tively small (of the orderof 10 m/s). Theseresultsare
drift velocitiesof only •5 - 10 m/s, which would sug- also consistentwith thosefrom earlier studies[Buonsanto et al., 1993;Buonsantoand Witasse,1999].
gest a DC bias. Clearly, our solar maximum December
solstice Saint Santin
drifts should be considered with a
Figure 4 showsthe average seasonalsolar minimum
ahd maximum quiet time zonal drift patterns over Saint
high degree of caution.
The seasonallyaveraged Millstone Hill eastward drift Santin. The average solar flux indices are the same as
pattern for geomagneticallyquiet conditionsare shown for Figure 2. The solar minimum drift patterns are in
in Figure 3. The data are again shown for low and good agreementwith the resultsof Blanc and Amayenc
high solar flux conditions with the same averagesolar [1979],with the exceptionof the summerearly night-
SCHERLIESS
ET
AL.'
MIDLATITUDE
PLASMA
Saint Santin
-
MAR-APR
"-
SEP-OCT
80 -l
1777
Kp_<3
MAY-AUG
-
-
NOV_ FEB
-
'
_
60
I
--
40
20
-
DRIFTS
_
Sa=180 ....
::
?',,,•,•.
_
__
_ ....
-20
-40
-
ß
..
-60
..
z
..
o
N
40
_
-
Sa=75
_
"
-
_
Sa=85
_
-
Sa=85
_
_
-
_
-
20
_
_
_
_
..
-20
ß
-40
-60
-I I I I I I I ] I I I I I I I I I I I I I I 1-
lJlllllllllllllllltJlll
00
04
08
12
16
20
24
00
04
08
12
L¸CAL
Figure 4.
reversal
20
24
-I I I I I I I'1'1 I I I I I I I I I I I J I I00
04
08
12
16
20
24
TI ME
Same as Figure 3, but for Saint Santin.
time drifts for which their study indicated a westward
drift
16
which is not seen in our results based on
hemisphere,which is located at a significantlylarger geographiclatitude (45øS),largelydeterminesthe Arecibo
a significantly larger data set. The Saint Santin zonal
drifts vary significantlywith the phase of the solar cy-
nighttime plasma drifts [Fcj½r, 1993; Ta]carai½t al.,
1996]. The conjugatepoint of Saint Santin, however,
cle. The nighttime drifts increasefrom •20-30 m/s to lies at a lowerlatitude (18øS)than this station. There•40-80 m/s betweenour low and high flux conditions. fore, as expected, the solar cycle variations of the Saint
Figure 5 shows the combined winter and equinox average solar cycle variation in the early nighttime period. In this case, we have used solar flux bins of 30
units for indices between 60 and 210 and a single bin
for larger values. The vertical bars indicate the standard deviation in each bin, and the solid line is our
empirical model representation. These drifts increase
linearly with increasingsolarflux at a rate of •40 m/s
per 100 flux units. The solar cycle variations during
summer are •50•
kOSaintSantin
September-Apri
•
'G'
100
I---'
Empirical
Model
.
smaller.
Zonal plasma drift observationsat Jicamarca (12øS
aA,
Arecibo(18øNgeographic,
magneticlatitude30øN)[F½jet, 1993],and from the IDM onboardthe DE 2 satellite
[Sch½rliess,
1997] indicate increasingsolar cyclevariations with decreasinglatitude, in good agreementwith
the observed solar cycle dependence at Millstone Hill
and Saint Santin. It is interestingto note that solarcycle variations at Arecibo are smallest during December
solstice,when the ionosphericdynamo in the conjugate
50
%o"i'b'o' "i56'"i5&"'i&&"
Solar
5i6'";>6
FIux
Figure 5.
Solar cycle variation of the equinox and
winter early night Saint Santin zonal drifts.
1778
SCHERLIESS
i
100-
i
i
i
i
i
i
[
i
[
i
i
i
i
i
i
ET AL.: MIDLATITUDE
i
i
i
[
i
1
PLASMA
i
--
_
-
Millstone
Hill
4020-
-
.....
-
Kp(0-6hrs)=l.8
Kp(0-6hrs)=0.7
i , , , ito,t1[t,2 , i , , it3,t4it5, , i ,tl6
-
•-500
-
LU
3001ø
8060-
DRIFTS
100•-
_-
-3
--
'
,
00
'
, I , , , ;, ' I" ,
0'3
06
09
, ,
1'2
"
15
Storm-Time (Hours)
0
Figure 7.
-20-
Idealized scenarioof the variation of the
AE index used in this work.
40-
_
_
2øfTTT
II!TTT
quiet drifts (Kp = 0.7) correctedfor seasonaland solar
cycle effects. Under the same conditions, the amplitudes of the Saint
-40
..................
02
06
10
Local
14
Time
Santin
zonal disturbance
drifts
about a factor of two smaller than over Millstone
18
are
Hill.
The effectsof the steady state leakageof high-latitude
zonal electric fields, which drive perpendicular northward disturbancedrifts, are essentiallynegligiblefor Kp
22
Figure 6.
(top) Yearly averageMillstone Hill perpendicular eastward drifts for two levels of magnetic
activity. (bottom) Averagezonal drifts by subtracting
the drifts in Figure 6; top. The solid curve indicates
the pattern from our disturbance model.
Santin zonal drifts are smallest in the summer, when
they are determined mostly by the dynamo action in
the local hemisphere.
The quiet time model described above allowed us to
determine the perturbation drifts which were used in
the storm time study presented in section 3.2. As we
will see, high-latitude electric fields affect the lowerlatitude plasma drifts with decreasingamplitudes toward the equator, even under quasi-stationary condi-
< 3 even over Millstone
3.2.
Midlatitude
Hill.
Disturbance
Plasma
Drift
Patterns
Figure 7 shows an idealized storm scenariowith an
increaseof the AE index by 300 nT above our quiet
time value of 130 nT during a period of 9 hours. Figure
8 presentsthe Millstone Hill disturbancedrift patterns
for the storm times defined in Figure 7. The data and
the scatterbarswereobtainedby binningand averaging
the data for the conditionsin Figure 7 and smoothedby
a three-point running average. The averagevaluesfor
the parametersdefinedin (2) are givenin Table 1. The
solid curvespresent the results from our empirical analytical modelobtainedfrom (2). It is important to note
tions. Therefore,for a givenquiet time criterion(based that our data binning can only approximately reproon the Kp index, for example), we expect our quiet duce the idealized storm scenario and that it does not
time drifts to have latitudinally increasingcontributions completely separate the responsesof the prompt penedue to electric fields of high-latitude origin. Figure 6 tration and disturbance dynamo processes.In spite of
shows that the zonal disturbance drifts over Millstone
these limitations, the binned data is generally in good
Hill havesignificantamplitudesevenfor Kp _<3 (Kp = agreement with the results from our simultaneousmul1.8). In this caseFigure 6 (top) was obtainedby bin- tiparameter fitting procedureand, as we will seelater,
ning the yearly averaged Millstone Hill zonal drifts for
also with theoretical patterns obtained from the Rice
geomagnetically
quiet (average6-hourKp<_3.0)and ex- ConvectionModel (RCM).
Figure 8 indicates that at time to, following an intremely quiet (average6-hour Kp_<l.3) conditions,and
Figure 6 (bottom) showsthe differenceof the two curves creasein the AE index by 400 nT, the prompt penetraand the results
obtained
from
our disturbance
model.
The scatterbars indicate the variability of the extremely
tion zonal electric fields drive perpendicular northward
drifts during the day and larger amplitude southward
Table 1. Average Storm Time Parameters
Storm Time
to
t•
t2
t3
t4
t5
t6
AAE(t- 30 min) (nT)
AAE(t- 90 min) (nT)
AEd(1-3 hour) (nT)
AEd(4-9 hour) (nT)
243
48
24
60
-33
209
149
79
-58
-27
293
89
-34
-29
287
275
-225
46
271
126
-41
-230
249
141
11
-7
30
165
SCHERLIESS ET AL' MIDLATITUDE
Millstone
PLASMA DRIFTS
1779
Hill 1978-92
_
2O
10
0
-10
[
[
[
[
I
[
]
[
]
I
[ _
3O
z 15
0
-15
Z .,,•• o
z
_..
10
0
-10
z
-
t1/••,.•, • 15
:
.....
_
0
-15
t
15
0
-15
10
0
-10
"--...__;
t
10
0
-10
15
10
0
-10
15
0
-15
• ,.
ß(
t5
...-15
0
-15
10
0
-10
;:..
)(
•
,
,,
10
0
-10
15
0
-15
z
_
]
02
Figure 8.
06
10 14 18
Local Time
22
02
]
•
06
t
I
I
[
]
[
10 14 18
Local Time
I
22
Averageperpendicular
northwardandeastward
disturbance
driftsoverMillstone
Hill for the conditionsand storm times shownin Figure 7. The solid lines indicate the patterns
from our disturbancemodel,and the dashedlinesat stormtimes t2 and t3 wereobtainedusing
the parametrization
of $cherliess
andFejer[1998].The scatterbarsdenotethe standarderror
of the means.
1780
SCHERLIESS ET AL.' MIDLATITUDE
PLASMA DRIFTS
driftsat nightwith a maximumvalueof •-25 m/s at
- '
-
0400 LT. The initial time responseof the meridional
electric fields generatesa small eastward disturbance
drift in the earlymorning-noon
periodandlargerwestwarddrifts at otherlocaltimes. The largestinitial time
westwarddrift perturbationis •40 m/s, and it occurs
near dusk. As stormtime increases,the disturbance
drifts initially decreasein amplitude and shift to later
local times but do not changemuch after •-2 hoursof
continuousmagneticactivity, as shownby the patterns
at storm times t2 and t3.
Stormtime t4 (Figure8) showsthe disturbance
patternsresultingfrom the suddendecrease
in the AE by
400 nT. In this case,the nighttime perturbationdrifts
' ' ' i , , , , i , , , , i , , , , i i , , _
Radar Model
AAE=400nT
80-- ..... DE-2Model
-
_
_
60--
_
Millstone
Hill
-
40--
_
_
20:
..-----•ial Time
Response
-_
-20 /
•.
_
•)
20 -
E
0
--
t0+90min
•.,
_
• -20
__
-
_
_
t'-
o
U
-
--
Saint Santin
-
-
_
_
changefromthe downward/southward
directionto up20Initial Time Response
ward/northward,
andtheprenoon
downward/southward
. .._...• ;::-,_.-._
driftsarefurtherincreased.
Themostnoticeable
changes
in the zonaldriftsoccurbeforenoonwhenthey become
20to+90min
westward. These perturbation drifts are again due to
the combinedeffectsof largetransientpromptpenetration drifts (with oppositesignto thosefor storm time
to in Figure 8) and longerlastingtime-delayeddistur'0'2' ' '0'6' ' '1•0 14• ' ' 1'8'• '2'2'
bances.The agreementbetweenthe averageddata and
Local Time
our model predictions(solidline) is very good,but a
phaseshiftof roughly1-2 hoursin the promptpenetra- Figure 9.
Comparisonof the prompt penetration
tion responsetowardlater local timesfor both compo- zonal disturbancedrift patterns derived from the Mill-
_
_
_
-
_
_
_
_
nents would make the agreementevenbetter. It is inter-
estingto notethat Blanc[1983]hasnotedthat a similar
phaseshiftwould also bring the Saint Santin drift data
in closeragreementwith his model results.
Onehour after the decrease
of the AE index,at storm
stone Hill and Saint Santin radar measurements and
from DE 2 observations after an increase in the AE in-
dex by 400 nT. The thin and thick dashed lines in the
Saint Santin part indicate the DE 2 resultsobtained for
A=35ø and A=45ø, respectively.
time t5, the meridionalperturbationdrifts have significantlysmalleramplitudes. The zonaldrifts, how- It is important to note that the DE 2 results were deever, still exhibit large nighttime westwardperturba- rivedfrom longitudinallyaveragedmeasurements.
Figtions, and small eastward perturbations around sunrise. ure 9 (bottom) comparesthe Saint Santin resultsand
Storm time t6 (Figure 8) showsthat after 6 hoursof the correspondingDE 2 results. Fejer and $cherliess
quieting, the perpendicular/northwarddrift perturba- [1998]pointedout that in the postmidnight-morning
tions have largely returned to their quiet time level, sector,the DE 2 initial time disturbancedrifts change
but the zonaldrifts still displaylargewestwardpertur- noticeablyfor invariantlatitudesfrom45ø to 35ø. Figbationsin the eveningsector,whichare not fully cap- ure 9 showsthat surprisinglythe Saint Santin initial
tured by our current model. This suggeststhat other time disturbancepattern is in better agreementwith
disturbancemechanismsor drift perturbationswith dif- the DE 2 resultsfor an averageinvariantlatitude of 35ø
ferenttimescales
than considered
in ourmodelmightbe thanfor A = 45ø. FejerandScherliess
[1998]suggested
operating during the recovery phase.
that the large changeof the DE 2 initial zonal distur-
The smaller database of drift measurements from
Saint Santin did not allow us to determine their storm-
bancedrift patternsfrom 45ø to 35ø was due in part
to the largeruncertainties
resultingfromthe largede-
time dependence
in asmuchdetail asgivenin Figure8.
creasein the numberof postmidnightmeasurements
at
However, as will be shown below, we were still able to lowerlatitudes.The SaintSantinresultssuggest
that
capture the basic featuresof their prompt penetration this argument is questionable. The t0+90 rain DE 2
and longer lasting disturbancedrift patterns.
andSaintSantindisturbance
drift patternsarein good
$cherliessand Fejer [1998]and Fejer and$cherliess agreement,althoughthe radar data showlarger west[1998]have used extensiveion drift meter data from ward drifts in the duskto early night period.
the DE 2 satellite to determine the middle- and low-
latitude ionosphericzonal disturbancedrift patterns.
Figure 9 (top) showsthe good agreementbetweenthe
Millstone Hill prompt penetration zonal drifts at storm
times t0+30 rain and t0+90 rain and the corresponding DE 2 drifts for an averageinvariant latitude of 55ø.
Figure 10 showsthe Millstone Hill and Saint Santin
a•d DE 2 disturbancepatterns after an increasein the
AE index by 400 nT over a period of 9 hours. The
radar drifts are again consistentwith the satellitedrifts
but havesystematically
largermagnitudes,
in particular
duringthe duskto early nighttimehours.However,we
SCHERLIESS ET AL.- MIDLATITUDE
[
[
!
!
]
!
[
,
[
i
,
,
!
!
]
60- 40-
-
!
,
Radar
AEd(1-9hrs)-400
nT
20-
!
Millstone
,
i
!
PLASMA DRIFTS
1781
Figure11 showsthe comparison
of the MillstoneHill
and SaintSantinpromptpenetrationdrift patternsand
,
Models
--- DE-2Models
the RCM results for an increase in the polar cap poten-
Hill
_
tial of 33 kV. These RCM patterns, which correspond
0
to run C of Spiroet al. [1988],werediscussed
in detail
by Fejer et al. [1990]and comparedto DE 2 model
resultsby Fejer and Scherliess[1998]. Figure 11 in-
-20
-40
dicates that the initial time responseof the Millstone
Saint
Santi
2O
0
_
-20
-40
....
02
]06....
110
....
Local
14
[ ....18
t ,2,2 ,
Time
Hill driftsare in goodagreement
with the RCM results.
Notice that the drift amplitudesdecreaseand shift to
later local times as the plasma sheet inner edge adjuststo the new crosspolar cappotential.The Saint
Santinperpendicular
meridionaldisturbance
driftsare
in generalagreementwith the modelresults,whereas
the zonal drifts near dusk are noticeably larger. Fe-
Figure 10. Sameas Figure 9, but for the long-lasting jet and $cherliess[1998]havepointedout that these
RCM quasisteadystate drifts underestimate
the DE 2
drifts.
zonal disturbance
zonal disturbance drifts for latitudes A _< 45 ø consistent with our results. The results above indicate that
have to reiterate here that the DE 2 results represent
the radar and DE 2 perturbation drift patterns are in
longitudinallyaveragedpatternsand haveseasonand reasonablygoodagreement.They are alsogenerally
local time lockedtogether (the early nighttime period, consistentwith the predictionsfrom the RCM, as well
as from other global convectionmodels.
for example,corresponds
to equinoctialconditions).
4. Comparison With Theoretical
Models
5. Summary and Conclusions
We have used incoherent scatter radar observations
from Millstone Hill and Saint Santin to study the char-
The penetrationof high-latitudeelectricfieldsto middle, low, and equatoriallatitudeshas alsobeen investi- acteristicsof the averagemidlatitude quiet time and
gatedusingseveralglobalconvection
models[e.g.,Spiro disturbanceelectrodynamicplasma drifts. Our quiet
e• al., 1981, 1988; Senior and Blanc, 1984; Zakharove• time Millstone Hill drift patterns are in good agreeal., 1989; Fejer et al., 1990; Peymirat and Fontaine, ment with earlier studies based on smaller databases.
wereusedto
1994;Peymirat,1998; Tsunomura,1999]. Thesemod- The Saint Santinquiet time measurements
variationsof
els calculatethe global ionosphericelectricfields and determine,for the first time, the seasonal
currentsby solvingthe continuityequationfor iono- these drifts for both low and high solar flux conditions.
sphericcurrentsfor a givenhigh-latitudeelectrostatic The Saint Santin eveningzonal drifts increasewith sothe corresponding
MillstoneHill drifts
potential or the field-alignedcurrent distributionon a lar flux, whereas
two-dimensional
(thin shell)ionosphere
with giveniono- are essentiallysolar flux independent.
We have modeled the Millstone Hill and Saint Santin
sphericconductances.The middle- and low-latitude
perturbationelectricfieldsobtainedfromdifferentmod- disturbancedrifts obtainedby removingfrom eachmeaels are similar sincethey depend mostly on the high- surement the season and solar flux dependent quiet
latitude potential penetratingto lower latitudes and time values. These stormtime dependent perturbation
on the ionosphericconductances
and not on the details drifts were usedto study and model,for the first time,
of magnetospheric
processes.
The mostcomprehensivethe temporalevolutionof the promptpenetrationand
theoreticalprompt penetrationresultshave been pro- longerlastingzonaland meridionalE xB disturbance
videdby the RCM whichaccounts
for the coupledelec- drifts. The radar disturbance patterns are in good
trodynamicsof the innermagnetosphere
andionosphere agreementwith resultsderivedfrom zonaldrift mea[e.g.,Wolfet al., 1986;Spiroet al., 1988]. The local surementsby the DE 2 satellite. This, and alsoearlier
time and latitude zonal and meridional perturbation
studies,alsoindicatesthat the RCM, as well as other
electricfields predictedby the RCM were studiedby detailed convectionmodels, can reproduce the average
Fejer et al. [1990].We cancomparetheir resultswith disturbancedrift patterns with reasonableaccuracy.
The present empirical model providesa relatively
the radar disturbance drift patterns presented above by
simple
description
of the time dependentevolutionof
usingthe empirical relationshipbetweenthe polar cap
potential drop and the AE index derivedby Ahnet al.
[1992],i.e., •(kV):
the midlatitude disturbance drifts based on the history
36 + 0.082AE•2(nT),whereAE•2 of the AE index. A copy of the current model can be
is the auroral electrojet index using 12 stations. In this obtained from the authors. The major challengefor the
of the large
case,a changein the AE index by 400 nT corresponds next few yearslies in the understanding
to a changein the crosspolar cap potentialof •33 kV. variability of these disturbancepatterns.
1782
SCHERLIESS
ET AL.'
Radar
]
[
]
[
I
i
]
[
MIDLATITUDE
Model
[
PLASMA
DRIFTS
....... RCM
I
[
[
'
]
I
]
[
[
[
I
[
_- AAE=400nT
_- AAE=400nT
_-
-
_-
_-
Millstone Hill
-
MillstoneHill
Init al Time Response
40
-
20 •-
,,' ,,'
-,...
InitialTimeResponse 60
: \
30
0
-20 '"
'•
'
"
-......
-30
to+60min-
to+60m'n
"",,..,,'"•
30
0
-30
-
Saint Santin
_-
Saint Santin
_
_
•.
•_. 10--
-
Initial
Time
Response;,
F
:/'-"',,
oL
-10
'.
,,'/
',
;
'.
-
-
-
,' "-•',
',.',_
'.'
"'--:•'.--/' "••
, ,:
Initial
Time
Response
- 15
0
',•
_•' /
'"'"
'-'-15
to+60min
•
", '-
0 ..,•_
_.,"•-; "/'••-
-
-10
02
06
10
Local
14
Time
18
22
02
06
10
Local
14
Time
18
15
0
-15
22
Figure 11. Comparison
of the disturbance
drift patternsobtainedfromour MillstoneHill and
SaintSantinmodels
withresults
fromtheRiceConvection
Model(RCM)foran increase
in the
polarcappotentialdropby 33 kV. Herethe RCM initial time responses
correspond
to a time
(to+ e)immediately
afterthechange
in • (short-dashed
line)and10minlater(long-dashed
line).
SCHERLIESS
Acknowledgments.
ET
AL.:
MIDLATITUDE
This work was supported by the
Aeronomy Program, Division of Atmospheric Sciences,of
the National ScienceFoundation through grants ATM-97146
77 and
ATM-9731704
and the
National
Aeronautics
and
SpaceAdministration through grant NAG5-4469. The Saint
Santin data were obtained through the CEDAR database,
which is supported by the National ScienceFoundation. We
also thank WDC-C2 Kyoto AE index service for providing
us with the preliminary AE indices.
Janet
G.
Luhmann
thanks
Oliver
G. Witasse
and W.
Robin Coley for their assistencein evaluating this paper.
References
Ahn, B.-H., S.-I. Akasofu,and Y. Kamide, The Jouleheat
productionrate and the particle energyinjectionrate as
a functionof the geomagnetic
indicesAE and AL, J. Geophys. Res., 88, 6275-6287, 1983.
Ahn, B.-H., Y. Kamide,H. W. Kroehl,and D. J. Gorney,
Crosspolarpotentialdifference,
auroralelectrojetindices,
and solar wind parameters,J. Geophys.Res., 97, 13451352, 1992.
Bauer, P., P. Waldteufel, and C. Vialle, The French multistatic incoherentscatter facility, Radio Sci., 9, 77-83,
1974.
Blanc, M., Magnetosphericconvectionelectric fields at midlatitudes, 3, Theoretical derivation of the disturbance con-
vectionpattern in the plasmasphere,J. Geophys.Res., 33,
235-251, 1983.
Blanc, M., and P. Amayenc, Seasonalvariations of the iono-
sphericExB drifts aboveSaint Santin on quiet days, J.
Geophys.Res., 8•, 2691-2704, 1979.
Blanc, M., and A.D. Richmond, The ionosphericdisturbancedynamo, J. Geophys.Res., 85, 1669-1688,1980.
Buonsanto, M. J., and O. G. Witasse, An updated clima-
PLASMA
DRIFTS
1783
latitudes observedby Dynamics Explorer 2, J. Geophys.
Res., 97, 19,461-19,469, 1992.
Jaggi, R. K., and R. A. Wolf, Self-consistentcalculation of
the motion of a sheet of ions in the magnetosphere,J.
Geophys. Res., 78, 2852-2866, 1973.
Mazaudier, C. A., and S. V. Venkateswaran,Delayed ionospheric effects of the geomagnetic storms of March 22,
1979 studied by the sixth co-ordinated data analysiswork-
shop (CDAW-6), Ann. Geophys.,8, 511-518, 1990.
Mazaudier, C. A., A.D. Richmond, and D. Brinkman, On
thermosphericwinds producedby auroral heating during
magnetic storms and associated dynamo electric fields,
Ann. Geophys.,5, 443-448, 1987.
Peymirat, C., A magnetosphere-thermosphere-ionosphere
electrodynamics general circulation model, J. Geophys.
Res., 103, 17,467-17,477, 1998.
Peymirat, C., and D. Fontaine, Numerical simulation of
magnetospheric convection including the effects of field
aligned currents and electron precipitation, J. Geophys.
Res., 99, 11,155-11,176, 1994.
Richmond, A.D., Ionosphericelectrodynamics,in Handbook
of Atmospheric Electrodynamics,vol. 2, edited by H. Volland, pp. 249-290, CRC Press, 1995.
Scherliess,L., Empirical studiesof ionosphericelectric fields,
Ph.D. dissertation, Utah State Univ., Logan, 1997.
Scherliess,L., and B. G. Fejer, Storm time dependenceof
equatorial dynamo zonal electric fields, J. Geophys. Res.,
102, 24,037-24,046, 1997.
Scherliess, L., and B. G. Fejer, Satellite studies of mid- and
low-latitude ionospheric disturbance zonal plasma drifts,
Geophys.Res. Lett., 25, 1503-1506, 1998.
Senior, C., and M. Blanc, On the control of magnetospheric
convection by the spatial distribution of ionosphericcon-
ductivities, J. Geophys. Res., 89, 261-284, 1984.
Spiro, R. W., M. Harel, R. A. Wolf, and P. H. Reiff, Quantitative simulation of a magnetospheric substorm, 3, Plastology of thermosphericneutral wind and F regionion
maspheric electric fields and evolution of the plasmapause,
drifts oboveMillstoneHill, J. Geophys.Res., 10•, 24,675J. Geophys. Res., 86, 2261-2272, 1981.
24,687, 1999.
Spiro, R. W., R. A. Wolf, and B. G. Fejer, Penetration of
Buonsanto, M. J., M. E. Hagan, J. E. Salah, and B. G.
high-latitude electric field effects to low latitudes during
Fejer, Solarcyclevariationsin Fregion electrodynamics
at
SUNDIAL 1984, Ann. Geophys.,6, 39-50, 1988.
MillstoneHill, Y. Geophys.Res., 98, 15,677-15,683,1993.
Takami, T., W. L. Oliver, A.D. Richmond, and S. Fukao,
Fejer, B. G., F region plasma drifts over Arecibo: Solar
Ionospheric drift similarities at magnetic conjugate and
cycle,seasonal,and magneticactivity effects,J. Geophys.
nonconjugate locations, J. Geophys. Res., 101, 15,773Res., 98, 13,645-13,652, 1993.
Fejer, B. G., The electrodynamicsof the low-latitude iono-
15,782, 1996.
Tsunomura, S., Numerical analysisof global ionosphericcursphere: Recent resultsand future challenges,J. Atmos.
rent system including the effect of equatorial enhanceSolar Terr. Phys., 59, 1465-1482, 1997.
ment, Ann. Geophys., 17, 692-706, 1999.
Fejer, B. G., and L. Scherliess,Time dependentresponse
Wand, R. W., A model representation of the ionospheric
of equatorialionospheric
electricfieldsto magnetospheric
disturbances,Geophys.Res. Lett., 22, 851-854, 1995.
Fejer, B. G., and L. Scherliess,Empirical models of storm
time equatorialzonalelectricfields,J. Geophys.Res., 102,
24,047-24,056, 1997.
Fejer, B. G., and L. Scherliess,
Mid- and low-latitudeprompt.
penetrationionosphericzonalplasmadrifts, Geophys.Res.
Left., 25, 3071-3074, 1998.
Fejer, B. G., R. W. Spiro, R. A. Wolf, and J. C. Foster,
Latitudinal variationof perturbationelectricfieldsduring
magnetically disturbaed periods: 1986 SUNDIAL obser-
vations and model results, Ann. Geophys.,8, 441-454,
1990.
Fejer, B. G., E. R. dePaula, S. A. Gonzales,and R. F. Woodman, Average vertical and zonal F region plasma drifts
electric field over Millstone Hill (A = 56ø), J. Geophys.
Res., 86, 5801-5808, 1981.
Wolf, R. A., R. W. Spiro, and G. A. Mantjoukis, Theoretical
comments on the nature of the plasmapause, Adv. Space
Res., 6, 177-186, 1986.
Zakharov, V. E., M. A. Nikitin, and O. A. Stairnov, Re-
sponseof low-latitude electric field to the action of magnetic source, Geomagn. Aeron., 29, 344-349, 1989.
C. Amory-Mazaudier, Centre d'Etude des Environnements Terrestre et Planetaires, 94107 Saint-Maur-deFosses, France.
M. J. Buonsanto,L. Goncharenko,and J. Holt, Haystack
Observatory,Massachusetts
Institute of Technology,West-
overJicamarca,
J. Geophys.
Res.,96,13,901-13,906,
1991. ford, MA 01886.
Foster,J. C., J. M. Holt, R. G. Musgrove,
and D. S.
B. G. Fejer and L. Scherliess,Center for Atmosphericand
Evans,Solarwinddependencies
of high-latitude
convec- Space Sciences,Utah State University, Logan, UT 84322tionandprecipitation,
in SolarWindMagnetosphere
Cou- 4405. ([email protected],
[email protected])
pling,editedbyY. KamideandJ. A. Slavin,pp. 477-494,
Terra Sci., Tokyo, 1986.
Heelis, R. A., and W. R. Coley, East-west ion drifts at mid-
(ReceivedJune 9, 2000; revisedAugust 11, 2000;
acceptedAugust 14, 2000.)

Download Report