JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 102, NO. A1, PAGES 417-424 JANUARY 1, 1997
High electron temperature associatedwith the prereversal enhancementin the equatorial ionosphere K.-I. Oyama,1 M. A. Abdu,2 N. Balan,1'3 G. J. Bailey,4 S. Watanabe,5 T. Takahashi,5 E. R. de Paula,2 I. S. Batista,2 F. Isoda,6 and H. Oya5
Abstract. High electrontemperature in the equatorialionization anomaly region detectedfirst by Kyokkosatellite and later observedfrequently by Hinotori satelliteare foundto be closelyassociated with the ionizationcrests of the anomaly.This phenomenon,called the equatorialelectron temperature anomaly, is found to occur predominantlyin the equinoctialmonths and to becomeenhanced with the increasein solaractivity. It is mainlya nighttimephenomenon and showsmaximum temperature enhancementat around2100 LT. Using a theoreticalmodel, a mechanismfor its occurrenceis presented.The mechanismis basedon the plasmatransport in the evening equatorialionosphere resulting from the sunsetelectrodynamical processes.
1. Introduction equatorialionosphere sunset electrodynamics is suggestedfor the observedanomalous Te enhancement. The Japanesesatellite Kyokko, launchedin 1978, encoun- tered regionsof high electron temperatureTe over the low- latitudeionosphere during several of its eccentrictrajectories. 2. Morphology of the Hot Region In somecases the temperatureenhancement, with respectto Figure la showsthe electrondensity, and Figure lb shows the expectednighttime backgroundtemperature, exceeded the electrontemperature measured by Hinotori on pass3599. 2000 K [Oyamaand Schlegel,1984]. The region of the en- The two shadedregions in Figure lb representenhancements hancedelectron temperature extended from magneticlatitude in T• of •600 K and •1000 K at magneticlatitudes 8 ø and -9 ø, -30 ø to 7ø . The temperatureenhancement showed maximum respectively,with respectto the backgroundT• values(repre- valuesin the 300ø-310ø geographiclongitude and -30 ø to -8 ø sentedby a dashedcurve). These regionscoincide with the geographiclatitude region.Further, it may be noted that this electrondensity enhancements of the equatorialanomaly (Fig- electrontemperature enhancement occurred in the region ure la). The latitudinalstructure of the equatorialanomaly is wherethe electrondensity was also high. About 16 casesof T e characterizedby a troughat the magneticequator with crests enhancementwere observedout of all the Kyokkopasses. It on eitherside. This feature is reflectedclearly in the latitudinal was also noted that all these Te enhancementsoccurred in variationof T•. The equatorialelectron temperature anomaly (EETA) is thus associatedwith the equatorial ionization March 1978.Although this was a new finding,a detailedin- anomaly(EIA). For a detailed descriptionof the EIA, see vestigationcould not be carried out due to the difficultyin Hanson and Moffet [1966], Moffet [1979], Walkerand Chan identifyingits mechanismand the availabilityof only a small number of observations. [1989],and Balan and Bailey[1995]. The T• enhancementof 1500 K with respectto the backgroundat 0500 LT is a sunrise In 1981the Hinotori satellitewas placed into a nearlyequa- associatedphenomenon. torial (with inclinationof 31ø ) circularorbit at an altitude of about600 km. This satellitealso recorded the phenomenonof 2.1. Global Distribution of the Hot Region the Te enhancement.During its 17 monthsof operation,Hi- The locationsof enhancedT e with respectto the normally notorifrequently observed this phenomenon. The observations expectedbackground were identified in a large number of describedin this paper clearlydemonstrate the associationof Hinotori passesin the evening-premidnightlocal time sector. the Te enhancementwith the electrondensity enhancement of The locationsso identifiedare marked by triangleson the the evening-nighttimeequatorial ionization anomaly. Using a trajectorysegments of the differentpasses shown in Figure2. theoretical model, a mechanism that invokes the role of the The magneticequator is shownby a dashedcurve. As Figure2 showsthe hot regionsare distributedpredominantly at mag- qnstituteof Spaceand Astronautical Science, Sagamihara, Japan. neticlatitudes away and on eitherside of the magneticequator, 2NationalInstitute for SpaceResearch, Sao Jose dos Campos, Bra- not on it. Though not shownseparately, these locationsare zil. 3Onleave from the Departmentof Physics,University of Kerala, alsothe locationsof the electrondensity enhancements of the Trivandrum, India. EIA crests.The reasonfor the relativelypoor data statisticsin 4AppliedMathematics Section, School of Mathematics,University the longituderegion of 130ø-180øis due to the fact that during of Sheffield,Sheffield, England. the timewhen the satellitewas located in thislongitude region SFacultyof Science,Tohoku University, Sendai, Japan. the satellitetracking center (Kagoshima) was recoveringthe 6Facultyof Education,Yokohama National University, Yokohama, Japan. data storedfrom other longitudes. Copyright1997 by the AmericanGeophysical Union. 2.2. SeasonalDependence Paper number 96JA02705. The Te enhancementswhich exceed 200 K were separated 0148-0227/97/96JA-02705509.00 and are plotted in Figure 3a at the latitudesof their occur-
417 418 OYAMA ET AL.: EQUATORIAL ELECTRON TEMPERATURE ANOMALY
DATE ----81/10/20 REV"- 3599
1
3000-
2000-
LU LU!-- 1001 UT 3:50 4:0 4:10 4:20 4:30 4:40 4:50 5:0 5:10 5:20 5 30 LT 14:12 16:44 18:55 22:9 23:51 2:43 5:11 7:20 9:38 12:21 15:9 LON 155.5 191.0 221.2 252.3 290.0 330.7 5.2 35.0 66.8 105.2 144.6 I_AT 29.3 16.5 -2.2 -20.2 -30.7 -27.6 -12.9 6.2 23.1 31.2 25.8 MLAT 21.3 14.6 2.0 -11.2 -19.5 -19.0 -10.0 3.0 14.7 20.1 16.5 ALT 624.7 611.5 592.5 575.5 567.4 571.6 586.4 605.6 621.4 627.4 621.2 Figure1. (top) The electrondensity. (bottom) Electron temperature measured by Hinotoriduring pass 3599.The twoequatorial ionization anomaly (EIA) peakscentered around the magnetic equator at 0410UT coveringthe localtime interval1700-2300 LT are associatedwith the enhancedelectron temperatures (shadedregions) with respect to the backgroundtemperature and the IRI representation(dashed curve) [Bilitza,1990]. The T e enhancementaround 0511 LT is the sunriseeffect. rencesagainst the dayof the year(top) togetherwith the daily the largerenhancements are concentratedin the equinoctial solar flux at 10.7 cm (bottom).The daysare countedfrom monthsof March and April (aroundday 90), Septemberand January1, 1981; the Hinotori data startedon February22, October(around day 270), and againin March and April 1981, and continued until June 11, 1982. The sizes of the (aroundday 450) of thefollowing year. To seethe solar activity trianglescorrespond to the degreeof departureof the Te value dependenceon the Te enhancement,•Te wasplotted against from its nighttimebackground value (•T e); the larger the thesolar radio flux at 10.7cm in Figure3b. A faintdependence triangle,the largerthe T e enhancement.It is clearlyseen that of the Te enhancementappears to existin dayswhen the solar
mo ø _
! I I I I I I I II I I I I I I ! I I 10.00 36.00 72.00 108.00 144.00 180.00 216.00 252.00 288.00 324.00 360.00 LONGITUDE Figure2. Locationsmarked in solidlines of enhancedelectron temperature along Hinotori trajectory segments,the maximum value of Te beingindicated by triangles.The dashedcurve is themagnetic equator. OYAMA ET AL.: EQUATORIAL ELECTRON TEMPERATURE ANOMALY 419
0•1 I I I I I I I I I I I I I I I I I I 0 60 120 180 240 300 360 420 480 540 600 DAY NUMBER Figure 3a. (top) Occurrenceof hot plasmaregions marked at geomagneticlatitude versus day number startingfrom January 1, 1981(1 monthand 20 daysbefore the Hinotorimission), till the endof the Hinotori dataon June 11, 1982. Increasing sizes of theplotted triangles indicate increasing Te deviation (up to 1200K). (bottom)The solarF•o.7 cm flux diurnalaverages for the sameperiod.
flux exceedsabout 150 units. In fact, the seasonaldistribution 2.3. AssociationWith Ground-BasedEIA Diagnostics of the A Te shownin Figure3a is somewhatmodulated by solar fluxvalues greater than 150 unitswith AT e generallyincreas- CachoeiraPaulista (45øW, 22.5øS) in Brazil is situatedvery ing as the solarflux increases.This modulationis to be spe- closeto the EIA crestlocation. Ionograms obtained from rou- ciallynoted for the casesof increasingflux at arounddays 345 tine soundingsover this station were studied when Hinotori and 400, whichproduced enhanced T e valuesduring Decem- flew over Brazil. Although the region surveyedby the ion- ber and February,exceptions from the equinoctialoccurrence osondeis not exactlythe sameas that of the satellitepasses, pattern. their proximity(in somecases by field line connection)could
1400
12.00
1 ooo
800
.
600 : : ß • ß ß • e• 400 ...... !...... ';'..... ','"'i";'"," ...... ß...... ':...... !...... -: ,..': 'i."l '. ,. ' ! : 2OO ...... i ...... ;",&'"'"•'"'""i ...... ' ...... t ...... ,--- ...... ß...... !...... -: : 0 .... ' .... ' .... ' .... J .... ; 1 O0 1 50 2.00 2.50 300 350 F10.7 Figure 3b. Departureof Te fromits nighttime background (/•Te) versussolar F•o.7 cm flUX are plotted in the time period 1900-2100 LT. 420 OYAMA ET AL.: EQUATORIAL ELECTRON TEMPERATURE ANOMALY
F- r- --40
_
_
-- ß ß
600
500E_ ß . 400• -.1500
' •_ ,1 ooo 200:300 ¾ ' ' r • II"' ••500 :- I I ' II1:] •oo- I , , • ,,,,, I , , II,,, I I I I .,I I I I I iI, I I I I I I I I I I o 0 12 00 12 00 12 00 12 00 12 00 12 00 LOCAL TIME(45øW)
Figure 4. (top) Dst valuesfrom May 26 to 31, 1981, for which ionosphericparameters over Cachoeira Paulista(45øW, 22.5øS) in Brazilare plotted middle panel for foF2 andbottom panel for hpF 2 (a represen- tativeparameter for the F 2 layerpeak height).The solidlines represent quiet day meanvalues, and the dots are valuesfor eachday. (bottom) The electrontemperature enhancements observed by Hinotori satellitein the vicinityregions of CachoeiraPaulista with scaleon the right-handside. make a comparisonwith ionosondedata very valuablein the eveningvertical drift enhancementsand the formation of the way of identifyingthe phenomenon.In Figure4 are plottedthe postsunsetequatorial anomaly on all thesedays, which there- F layerpeak density (critical frequency, foF2) andpeak height fore supportsthe associationbetween the EIA and the EETA. (hpF2) parametersfor a few consecutivedays of Hinotori The valuesof Dst index for these daysare also shownin the passes.The daytime equatorial anomalyproduces maximum top panelof Figure4 to showthat/x Te increaseis not directly foF2 valuesover CachoeiraPaulista at around 1700-1800 LT. related to the geomagneticdisturbance. Followed by a short-liveddecrease in the electron density, there is a resurgenceof the EIA startingat approximately2000 2.4. Local Time Variation of EETA LT. ThehpF 2 values(bottom) that decrease in the afternoon In Figure 5 the maximumvalue of /XTe on each day is hours show increasesagain starting at 1800 LT. These in- plotted againstlocal time from 1800 LT to 0200 LT. There creasesare producedby the well knownprereversal enhance- existsa significantscatter in the data.However, a tendencyfor ment of the eveningequatorial electric field [see,e.g., Wood- man, 1970;Fejer et al., 1991].This electricfield enhancementis believed to be producedby the interactionof the eastward 1500 blowingthermospheric wind with the Pedersonconductivity longitudinalgradient that existsacross the sunsetterminator [see,e.g., Rishbeth, 1971; Heelis et al., 1974;Farley et al., 1986; Batistaet al., 1986].The electricfield enhancementhas earlier z /x,Te to increasefrom 1800 LT, to reach a maximumvalue near the conjugateionosphere before being thermalized[Mantas et 2100 LT, and then to decreaseto low valuesby midnightcan be al., 1978].In the photoelectronheating code, backscattering is noted.The risingpart of A T e is due to the continuinginfluence included self-consistentlyand pitch angle diffusion is taken of the daytimetemperature into the nighttimein sucha way care of by assumingthat a fraction of the photoelectronsis that the Te, with respectto an otherwisedecaying nighttime trappedand that all the energyof the trappedflux is deposited (background)temperature, shows up as an increasingfunction in the plasmasphere.However, photoelectrontrapping is not of time. The modifiedtemperature itself startsits decayfrom usedin the presentstudy as the studyis for nighttimeconditions. 2100 LT onwardresulting in the decreasingpart of ATe. Max- Figures6a, 6b, and 6c showthe latitudinalvariations of the imum Te of the EETA increasesas the associatedmaximum model T e at 1900 LT, 2000 LT, and 2100 LT, respectively.At electron densityincreases. each local time the variations are shown for five different altitudes(500, 650, 800, 950, 1100km). At 1900LT (Figure6a) the latitudinal variation of T e showsa minimum around the 3. Mechanism equator. This minimum,which dependson altitude, is caused 3.1. Theoretical Modeling mainly by the strongprereversal upward E x B drift which The model values of electron temperature are calculated raisescomparatively cold plasma from the bottomsideiono- using the Sheffield University plasmasphere-ionosphere sphere to the topside ionospherein the form of a forward model, SUPIM [Baileyet al., 1993, 1996; Bailey and Balan, plasmafountain [Balan and Bailey,1995]. At 2000 LT (Figure 1996]. In this model, coupled time-dependentequations of 6b), the variation of T e showscrests on either side of the continuity,momentum, and energybalance are solvedalong equator at altitudesabove 800 km. At later times, as seen in closedmagnetic field lines to give values for the concentra- Figure 6c, the crestsappear at lower altitudesand disappearat tions,field-aligned fluxes, and temperaturesof the O +, H +, higher altitudes;note the appearanceof small crestsat 650 km He+, N•-, O•-, and NO + ions,and the electronsat a discrete and the disappearanceof the crestsat 1100 km at 2100 LT set of points along the field lines. For the presentstudy, the (Figure 6c). At eachaltitude the T e crestsare found to fall on model equationsare solvedalong 112 eccentric-dipolemag- the same magneticline of force, and they are found to occur netic field linesdistributed with apexaltitude between150 and slightlypoleward of the ionization crests.The T e crestsarise 4000 km; this givesa reasonable24-hour distributionof model from the combinedeffects of the eveningprereversal upward T e valuesbetween 200 and 1500 km altitude and _+30ø mag- E x B drift, postreversaldownward E x B drift, and nighttime netic latitude. The model calculationsare for magnetically cooling.The downwardE x B drift, which bringshot plasma quiet (Ap = 4) equinoctialconditions (day 264) at high solar from high altitudes and latitudes in the form of a reverse activity(Flo.7 -- 238 and Flo.7A = 210) and for the longitude plasmafountain [Balan and Bailey, 1995;Balan et al., 1996], of Jicamarca(77øW). causesthe electrontemperature to increasein the regionof the The E x B drift and neutral wind velocities are included in plasma fountain. Outside this region, electron temperature the model calculations. The vertical E x B drift velocities are decreasesdue to nighttime cooling. Consequently,Te crests obtained from the measurements made at Jicamarca under occur on either side of the equator before the equatorial T e equinoctialconditions at high solaractivity [Fejer et at., 1991]. minimum (Figure 6a) disappears.The model values of ion The drift velocityundergoes a prereversalupward strengthen- temperatures(not shown)are found to be very nearlyequal to ing with peakvalue of about45 m s-• at 1845LT and the the electrontemperature; the ion temperaturesare alsofound velocityreverses downward at about 1930 LT. The Jicamarca to exhibit the featuresshown in Figures6a-6c. drift velocitiesare used for magneticfield lines with apex altitude less than 450 km. For field lines with apex altitude 3.2. Comparison BetweenTheory and Observation greater than 3000 km, the drift velocity is taken to be zero. In Figures 6a, 6b, and 6c the mean T e values which were Linear interpolationis usedat intermediateapex altitudes. The measuredby Hinotori satellite in the time periods 1830-1930 drift velocityvaries along the geomagneticfield linesin accor- LT, 1930-2030 LT, and 2030-2130 LT for F•o.7 > 250 are also dancewith the field line geometry[Bailey and Batan, 1996]. plotted asrepresentative values at 1900LT, 2000 LT, and 2100 The zonal componentof the E x B drift velocityis neglected LT, respectively.At 1900 LT, T e showsa dip at around the sinceit has only negligibleeffect on electrondensity profiles equator and at 2000 LT and 2100 LT, T e peaks appear at [Anderson,1981]. The neutral wind velocityin the magnetic around _+15 geomagneticlatitude. A comparisonof the obser- meridian, which also varies with altitude and latitude, is deter- vations and model resultsat 2100 LT showsthat the model Te mined from the meridionaland zonal wind velocitiesgiven by peak at 650 km is locatedabout 20-3 ø poleward of the observed the HWM90 thermosphericwind model [Hedin et at., 1991]. peaks.The magnitudeof the modelpeak T e almostagrees with The concentrationsand temperaturesof the neutral gasesare the observation.Although the model doesnot fully reproduce taken from the MSIS86 thermosphericmodel [Hedin, 1987] the observations,the phenomenoncan be explainedas result- and the solar EUV fluxes from the EUV94 solar EUV flux ing from the plasma flows and plasma cooling during the model(a revisedversion of the EUV91 solarEUV flux model eveninghours. [Tobiska,1991]). The flows of plasma during eveninghours as given by the The photoelectronheating rates are calculatedby the pho- model havebeen presentedand explainedby Balan and Bailey toelectroncode developedby Richardsand coworkers[Rich- [1995]andBalan et al. [1996].The plasmaflows at 1900LT and ards and Torr, 1988; Torr et at., 1990]. This code, which has 2100 LT are shownin Figures7a and 7b, respectively.During been incorporatedinto SUPIM, usesthe two-streamapproxi- the prereversalperiod, the upwardE x B drift raisescompar- mation methodof Banksand Nagy [1970] and includesinter- atively cold plasma from the bottomsideionosphere to the hemispherictransport of photoelectrons.Photoelectrons heat topsideionosphere (Figure 7a); this causeslow T e aroundthe the thermal electronsthrough Coulomb collisions,magnetic equator as shownin Figure 6a. After the drift reversesfrom trappingdue to pitch anglediffusion, and backscatteringfrom upward to downward,plasma from higher altitudesand lati- 422 OYAMA ET AL.: EQUATORIAL ELECTRON TEMPERATURE ANOMALY tudes flows to lower altitudes and latitudes in the form of a reverseplasma fountain (Figure 7b). Near the equator(about 3000 _10 ø in latitude) the temperatureis still low becausemost of 2800 the upliftedcold plasma drifts down, while at around10ø-20 ø in 2600 latitude most of the plasmaflows down from higher altitude 2400 and therefore T e is higher than near the equator.At latitudes 22OO outsidethe reach of hot plasma,where plasmais transported 2000 horizontallyfrom higherlatitudes but is still at sloweraltitudes 1800 than at latitudes10ø-20 ø, the electrontemperature is low and 1600 decreasesdue to nighttime cooling.Accordingly, the T e ob- 1400 servedby a satellitewhich crosses the magneticfield lineswith high T e showstwo peaks,one in the north hemisphereand the 1200 other in southhemisphere (see Figure 1). IOO0 8OO 600 4. Discussion 400 The morphologyof the EETA discussedin section2 seems 2OO to be well explained on the basis of the plasma transport 0 I I I patternin the eveningequatorial ionosphere [Balan and Bailey, 30 25 20 15 10 5 0 -5 -10 -15 -20 -25 -30 1995]. It is dearly shownthat the location of the T e peaks Latitude (de9) coincidewith the latitudinalposition of the EIA crests,partic- ularly in the equinoctialmonths of March-April and Septem- 3000 ber-October(Figure 4). The thermosphericzonal wind (blow- 2800 ing awayfrom the subsolarpoint) is generallystrongest in the 2600 eveninghours of thesemonths [Hedin et al., 1991]. This situ- 2400 ation is conduciveto the large amplitudesof the evening(pre- 220O reversal) zonal electric field (vertical drift) enhancement 2000 [Batistaet al., 1986;Farley et al., 1986]as well as of the vertical 18O0 electric field (zonal plasma drift), thus contributingto the 1600 mechanismfor the enhancedelectron temperature as dis- 1400 cussedin section3. It is also interestingto note that the solar 1200 F lo.7cm flUX controlsboth the vertical and zonal equatorial 1000 plasmadrifts in the eveninghours as discussedby Fejer et al. 8OO [1991].For example,an increaseof solarflux from 100 to 200 units could result in 100% increasein the yearly averaged 6OO vertical drift and by as much as 50% in the zonal plasmadrift 4OO velocity. Therefore a large increasein the 10.7-cm solar flux 2OO shouldproduce, as is evidentfrom the EETA mechanismde- 0 I I I I I I i i I i i scribedabove, hot plasmain the ionizationanomaly region. 30 25 20 15 10 5 0 -5 -10 -15 -20 -25 -30 The eventscentered around days 320, 340, and 400, that do not L@ri rude (de9) fall in the equinoctialmonths, are goodexamples of this effect. The factorswhich might be consideredto explainthe quanti- 3000 tative differencebetween the theory and observation,espe- 2800 cially in height, might be the backgrounddensity and the alti- 2600 tude variation of the upward plasmadrift velocity.Another 2400 factor which might be taken into accountis the neutral tem- 2200 perature. Raghvaraoet al. [1991] reported that an eastward 2000 neutral wind is deceleratedby ion-neutraldrag and the bulk 1800 motion of the wind is convertedto the kinetic temperatureof 1600 neutralgases of about50 ø . Accordingly,the energyloss rate of 1400 thermalelectrons to the neutralgas is reducedwhich causes T e 1200 to be higherthan thosewithout an enhancementin neutralgas 1000 temperature.Also, Sridharanet al. [1991] have found a high 8OO 600 4OO Figure 6. (opposite)Theoretical model values of T e plotted 2OO againstgeomagnetic latitude at five different altitudes(solid 0 I I I I I I i I I I I curve, 500 km; dotted curve, 650 km; short dashed curve, 800 30 25 20 1ES 10 5 0 -5 -10 -1ES -20 -25 -30 km; dash-dottedcurve, 950 km; and long dashedcurve, 1100 km) for (a) 1900 LT, (b) 2000 LT, and (c) 2100 LT. Solid La•:i ::ude (de9) circlesshow mean T e valuesobserved by Hinotori for Flo.7 > 250 duringthe time periodsof 1830-1930 LT, 1930-2030 LT, and 2030-2130 LT. OYAMA ET AL.: EQUATORIAL ELECTRON TEMPERATURE ANOMALY 423 a 2200 quantitatively,simultaneous measurements of the electronand ion temperaturesand densities,vertical and horizontal plasma 2000 drifts,polarization electric field, neutral wind and ion compo- sition in the 400-1000 km altitudes are needed. Such measure- 1800 mentscould be accomplishedby a multiple tethered satellite 1600 missionfor whichfeasibility studies are beingcarried out. • 14oo Acknowledgments.We would like to expressour sincerethanks to 1200 the satellitelaunching staff of ISAS and the related institutions.We are alsograteful to late W. B. Hansonfor his continuousencouragement. looo M. A. Abdu wishesto expresshis thanks to the Sao Paulo State 8oo Foundationfor promotionof research(FAPESP) for the support receivedtowards INPE-ISAS collaborations.N. Balan expresseshis 600 gratitudeto ISAS for the awardof a fellowship. The Editor would like to thank K. Schlegeland T. Tanaka for their 400 assistancein evaluatingthis paper. 30 25 20 15 10 5 0 -5 -10-15 20 25-30 References La:,:ude (de9) Abdu, M. A., I. S. Batista,G. O. Walker, J. H. A. Sobral,N. B. Trivedi, and E. R. de Paula, Equatorial ionosphericelectric field during magnetosphericdisturbances: Local time/longitudedependence from recent EITS campaigns,J. Atmos. Terr.Phys., 57, 1065-1083, 1995. 2000 Anderson,D. N., Modelingthe ambient,low latitudeF regioniono- 1800 sphere•A review,J. Atmos. Terr.Phys., 43, 753-762, 1981. Bailey,G. J., and N. Balan,A low-latitudeionosphere-plasmasphere 1600 model,in STEPHand Book,edited by R. W. Schunk,p. 173,Utah State Univ., Logan, 1996. • 1400 Bailey, G. J., R. Sellek, and Y. Rippeth, A modellingstudy of the '-' equatorialtopside ionosphere, Ann. Geophys.,11,263-272, 1993. • 1200 Bailey,G. J., N. Balan, and Y. Z. Su, The SheffieldUniversity plas- masphere-ionospheremodel•A review,J. Atmos. Terr. Phys.,in 5, 1000 press,1996. Balan,N., and G. J. Bailey,Equatorial plasma fountain and its effects: 800 Possibilityof an additionallayer, J. Geophys.Res., 100, 21421-21432, 1995. 600 Balan, N., G. J. Bailey, M. A. Abdu, K. I. Oyama, P. G. Richards, J. MacDougall,and I. S. Batista,Equatorial plasma fountain and its 4oo effectsover three locations:Evidence for an additionallayer, the F3 I -•1 I I --'•1•-I--- • I - • I - -- layer,J. Geophys.Res., in press,1996. Banks, P.M., and A. F. Nagy, Concerningthe influenceof elastic 30 25 20 15 10 5 0 -5 -t0-15-20-25-30 scatteringupon photoelectrontransport and escape,J. Geophys. l_a•:• •:ude (deg) Res., 75, 1902-1910, 1970. Batista,I. S., M. A. Abdu, and J. A. Bittencourt,Equatorial F-region Figm'e ?. (a) Vector plasma fluxes at 1900 LT calculated verticalplasma drift: Seasonaland longitudinalasymmetries in the underthe samecondition as for Figure 6. The fluxesare given American sector,J. Geophys.Res., 91, 12055-12064, 1986. in log•oscale. The minimumvector length (zero length)cor- Bilitza,D., Internationalreference ionosphere 1990, URSI-COSPAR, respondsto plasmaflux less than 1 x 108cm -2 s-•. Positive Rep.NSSDC/WDC-A-R&S, 90-22, Natl. SpaceSci. Data Cent., Lan- latitude is northward.(b) Same as Figure 7a but at 2100 LT. ham, Md., 1990. Farley,D. T., E. Bonelli,B. G. Fejer, and M. F. Larsen,The prerever- sal enhancementof the zonal electricfield in the equatorialiono- sphere,J. Geophys.Res., 91, 13723-13728,1986. neutral temperaturepeak at 2100 LT in Indian longitudere- Fejer, B. G., E. R. de Paula, S. A. Ganguly,and R. F. Woodman, Averagevertical and zonalplasma drifts over Jicamarca, J. Geophys. gion. As yet, a reasonfor its occurrencehas not been estab- Res., 96, 13901-13906, 1991. lished. Hanson,W. B., and R. J. Moffet, Ionizationtransport in the equatorial F region,J. Geophys.Res., 71, 5559-5572, 1966. Hedin, A. E., MSIS-86 thermosphericmodel, J. Geophys.Res., 92, 5. Conclusions 4649-4662, 1987. Hedin, A. E., et al., Revisedglobal model of thermosphericwinds The high electrontemperature in the night time equatorial usingsatellite and ground-basedobservations, J. Geophys.Res., 96, ionospherediscovered by the Japanesesatellite Kyokko in 7657-7688, 1991. 1978was confirmed by the Hinotori satellitelaunched in 1981. Heelis, R. A., P. C. Kendall, R. J. Moffet, D. W. Windle, and A detailed study on the morphologicalfeatures of the hot H. Rishbeth,Electrical coupling of the E- and F-regions and its effectson F-regiondrifts and winds,Planet. Space Sci., 22, 743-756, plasmaregion demonstrates that theseregions are closelyas- 1974. sociatedwith the well known equatorialionization anomaly Mantas, G. P., H. C. Carlson, and V. B. Wickwar, Photoelectron flux during the postsunsethours. The hot plasma?egions occur buildupin the plasmasphere,J. Geophys.Res., 83, 1-15, 1978. predominantlyin the equinoctialmonths and showsignificant Moffet, R. J., The equatorialanomaly in the electrondistribution of the terrestrialF-region, Fundam. CosmicPhys., 4, 313-391, 1979. dependenceon the intensityof solar 10.7 cm radio flux. The Oyama,K.-I., and K. Schlegel,Anomalous electron temperature above presentlyused theoretical ionosphere-plasmaspheremodels the South American Magnetic Anomaly, Planet. Space Sci., 32, can predict the EETA, qualitatively.However, to understand 1513-1522, 1984. 424 OYAMA ET AL.: EQUATORIAL ELECTRON TEMPERATURE ANOMALY Raghavarao,R., L. E. Wharton,N. W. Spencer,H. G. Mayr, and L. H. Woodman, R. F., Vertical drift velocitiesand east-westelectric fields at Brace,An equatorialtemperature and wind anomaly (ETWA), Geo- the magneticequator, J. Geophys.Res., 75, 6249-6259, 1970. phys.Res. Lett., 18, 1193-1196,1991. Richards,P. G., and D. G. Torr, Ratios of photoelectronto EUV M. A. Abdu,I. S. Batista,and E. R. de Paula,National Institute for ionizationrates for aeronomicstudies, J. Geophys.Res., 93, 4060- SpaceResearch, INPE 12.225,Sao Jose dos Campos, Brazil. 4066, 1988. G. J. Bailey,Applied Mathematics Section, School of Mathematics, Rishbeth,H., Polarizationfields produced by winds in the equatorialF Universityof Sheffield,Sheffield S3 7RH, England. region,Planet. Space Sci., 19, 357-367, 1971. N. Balan and K.-I. Oyama, Institute of Spaceand Astronautical Sridharan,R., S. Gurubaron, R. Raghavarao,and R. Suhasini,Co- Science,Yoshinodai, Sagamihara 229, Japan. ordinatedthermospheric and F regionmeasurements from low lat- F. Isoda,Faculty of Education,Yokohama National University, Ho- itudes,J. Atmos. Terr.Phys., 53, 515-519, 1991. dogaya,Yokohama 240, Japan. Tobiska, W. K., Revised solar extreme ultraviolet flux model, J. Atmos. H. Oya, T. Takahashi,and S. Watanabe,Faculty of Science,Tohoku Terr.Phys., 53, 1005-1017, 1991. University,Aoba, Sendai 980-77, Japan. Torr, M. R., D. G. Torr, P. G. Richards,and S. P. Yung, Mid- and low-latitudemodel of thermospheric emissions, 1,O+(2P) 7320 • andN2(2P ) 3371•, J. Geophys.Res., 95, 21147-21168, 1990. Walker, G. O., and H. F. Chan, Computer simulationof seasonal variationof the ionosphericequatorial anomaly in East Asia under (ReceivedDecember 27, 1995;revised June 7, 1996; solarminimum condition, J. Atmos. Terr.Phys., 51, 953-974, 1989. acceptedJune 14, 1996.)