Lowand mid-latitude ionospheric electric fields during the January 1984 GISMOS campaign

Transcription

1 Utah State University From the SelectedWorks of Bela G. Fejer March 1, 1990 Lowand mid-latitude ionospheric electric fields during the January 1984 GISMOS campaign Bela G. Fejer, Utah State University M. C. Kelley C. Senior O. de la Beaujardiere J. A. Holt, et al. Available at:

2 JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 95, NO. A3, PAGES , MARCH 1, 1990 Low- and Mid-Latitude Ionospheric Electric Fields During the January 1984 GISMOS Campaign B. G. FEJER,1M. C. KELLEY,2 C. SENIOR,30. DE LA BEAUJARDIERE,4 J. A. HOLT,5 C. A. TEPLEY,6 R. BURNSIDE,6 M. A. ABDU,7 J. H. A. SOBRAL, 7 R. F. WOODMAN, 8 Y. KAMIDE,9 AND R. LEPPING10 This paper examines detail the electrical coupling between the high-, middle-, and lowlatitude ionospheres during January 17-19, 1984, using interplanetary and high-latitude magnetic field data together with F region plasma drift measurements from the EISCAT, Sondre Stromfjord, Millstone Hill, Saint-Santin, Arecibo, and Jicamarca incoherent scatter radars. We study the penetration of both the zonal and meridional electric field components of high-latitude origin into the low-latitude and the equatorial ionospheres. In the dusk sector, a large perturbation of the zonal equatorial electric field was observed in the absence of similar changes at low and middle latitudes in the same longitudinal sector. The observations in the postmidnight sector are used to compare the longitudinal variation of the zonal perturbation electric field with predictions made from global convection models. Our resultshow that the meridional electric field perturbations are considerably more attenuated with decreasing latitude than the zonal fluctuations. As a result, we conclude that variations in the meridional electric field at low latitudes are largely due to dynamo effects. These observations are used to show that the global convection models reproduce a number of characteristics of low-latitude and equatorial electric fields associated with changes in the polar cap potential drop. In addition, we highlight several areas where there is still substantial disagreement between the electric field data and the theoretical results. 1. INTRODUCTION There is now extensive evidence that during periods of strong geomagnetic activity the middle- and low-latitude electric fields and currents can depart appreciably from their quiet-day patterns, which are driven by tidal winds. The effects of high-latitude current systems at lower latitudes include (1) simultaneous electric field perturbations over a large range of latitudes and longitudes [e.g., Nishida, 1968; Blanc, 1978, 1983; Gonzales et al., 1979, 1983; Kelley et al., 1979; Reddy et al., 1979; Wand and Evans, 1981; Fejer, 1981, 1986; Somayajulu et al., 1987] and (2) wind disturbances which have latitude-dependent time delays and create (3) storm time dynamo electric fields [e.g., Blanc and Richmond, 1980; Fejer et al., 1983]. 1 Center for Atmospheric and Space Sciences, Utah State University, Logan. 2 School of Electrical Engineering, Cornell University, Ithaca, New York. 3 Centre National d'etudes des Tdldcommunications/Centre de Recherch en Physique de 1' Environnment, St.-Maur des Fosses, France. 4 Radio Physics Laboratory, SRI International, Menlo Park, California. 5 Haystack Observatory, Massachusetts Institute of Technology, Westford. 6 Arecibo Observatory, Arecibo, Puerto Rico. 7 Instituto de Pesquisas Espaciais, S5o Jose dos Campos, S5o Paulo, Brazil. 8 Instituto Geofisico del Peru, Lima. 9 Kyoto Sangyo University, Kita-ku, Kyoto, Japan. 10 Goddard Space Flight Center, Greenbelt, Maryland. Copyright 1990 by the American Geophysical Union. Paper number 89JA /90/89JA Large-scale magnetospheric electric fields imposed on the high-latitude ionosphere [e.g., Kamide, 1988] can be transmitted almost instantly to the middle, low, and equatorial regions through the Earth-ionosphere waveguide [e.g., Kikuchi et al., 1978]. Incoherent scatter radar observations have been used extensively to determine the average electric field patterns at middle and low latitudes for different levels of magnetic activity [e.g.,blanc, 1983; Wand and Evans, 1981; Ganguly et al., 1987], and to measure the extent of the magnetospheric electric field effects on the plasmasphere [e.g., Blanc, 1978; Gonzales et al., 1979; 1983; Fejer, 1986; Earle and Kelley, 1987]. Several numerical studies have examined the global distribution of ionospheric electric fields and currents during both geomagnetically quiet and disturbed conditions by solving the continuity equation for a twodimensional spherical shell for given high-latitude potential or current distributions, and for different longitude- and latitudedependent ionospheric conductivities [e.g., Nopper and Carovillano, 1978; Kamide and Matsushita, 1981; Takeda and Maeda, 1982; Tsunomura and Araki, 1984; Senior and Blanc, 1987; Spiro et al., 1988]. Detailed measurements of ionospheric electric fields and currents have been made recently during the Incoherent Scatter World Day campaigns. These coordinated studies include observations with the EISCAT (European Incoherent Scatter), Sondre Stromfjord, Millstone Hill, Saint-Santin, Arecibo, and Jicamarca radars and from several other ground-based and in situ probes. The GISMOS (Global Ionospheric Simultaneous Measurements of Substorms) campaign was designed to measure in detail the ionospheric and magnetospheric effects associated with substorms and to provide detailed data sets for testin global thermospheric, ionospheric and magnetospheric models. The instrumentation is described by de la Beaujardiere [1987]. Some of the high-latitude convection electric fields and currents during the January 16-19, 1984 GISMOS campaign have been discussed by de la Beaujardiere et al. [1987a,b], Alcayd et al. [1986], Senior and Blanc [1987], and Richmond et al. [1988]. In this study we concentrate on F region plasma drift measurements from incoherent scatteradars to study the relationship between lowand mid-latitude ionospheric electric fields and the high- 2367

3 2368 FEJER ET AL.: LOW- AND MID-LATITUDE ELECTRIC FIELDS latitude electric fields and currents. These observations are then compared in detail with numerical studies of global electric fields. 2. ELECTRIC FIELD RESULTS In this section we present the electric field measurements made by the EISCAT, Sondre Stromfjord, Millstone Hill, Saint-Santin, Arecibo, and Jicamarca radars, and also some ionosonde and magnetic field data obtained during the January 1984 GISMOS campaign. We study most closely the equatorial, middle-latitude, and low-latitude results, but will also introduce the high-latitud electric and magnetic field data to determine their effects on the electric field perturbations (relative to the quiet-time values) at lower latitudes Radar Operating Modes During this campaign, extensive incoherent scatter radar measurements were performed [e.g., de la Beaujardiere et al., 1987a,b] in the F region (between about 200 and 500 km in altitude). We will discuss only the cross-field ion plasma drift measurements which are related to the perpendicular component of the ionospheric electric field through Va = E x B/B2. The EISCAT radar (69.6 ø N, 19.2 ø W) operated with the CP3 mode elevation scans in the magnetic meridian with a time resolution of 30 min, and a latitudinal coverage from 61 ø to 72 ø invariant latitude [Alcayd et al., 1986; Senior and Blanc, 1987]. The Sondre Stromfjord radar (67.0 ø N, 51.0 ø W) measurements were made between about 68 ø and 80 ø invariant latitude with a cycle time resolution of about 17 min [de la Beaujardiere et al., 1987a]. The operating mode consisted of measurements at nine fixed positions followed by a continuous elevation scan from north to south in the magnetic meridian. The Millstone Hill radar (42.6 ø N, 71.5 ø W) was operated on an 180 ø azimuthal scan at an elevation of 4 ø [Holt et al., 1984; 1985] coveting the invariant latitudes between 55 ø and 65 ø with a time resolution of about 30 min. The Arecibo radar (18 ø N, 67 ø W; magnetic dip 50 ø N) scanned in azimuth over 360 ø at a fixed elevation of about 75 ø, with a cycle time of about 16 min [e.g., Behnke and Harper, 1973]. The Arecibo drift measurements assume constant and spatially uniform drifts over the whole scan. The Saint-Santin radar (44.6 ø N, 2.2 ø E) measured only the zonal component of the F region plasma drift at an invariant latitude of 44 ø [e.g., Blanc and Amayenc, 1979]. The Jicamarca radar (12.0 ø S, 76.9 ø W; magnetic dip 2 ø N) was pointed perpendicular to the Earth's magnetic field and measured only the F region vertical plasma drift [Woodman, 1970] Equatorial Observations and the Relationship With the IMF and High-Latitude Convection Measurements The equatorial plasma drift measurements at Jicamarca provide a very accurate measure of the ionospheric electric fields when the radar beam is pointed perpendicular to the Earth's magnetic field in the F region. Over Jicamarcan upward plasma drift of 40 rn/s corresponds to an eastward electric field of about 1 mv/m. The accuracy of these measurements is about 2 m/s, corresponding to a zonal electric field of about 0.05 mv/m, and the integration time is typically 5 min. [e.g., Woodman, 1970]. The high accuracy is also a result of the small geophysical noise level on the vertical velocity component [Gonzales et al., 1983]. The variation of this electric field component with magnetic activity has been studied so extensively that a considerable knowledge base exists [Fejer et al., 1979; Gonzales et al., 1979; 1983; Kelley et al., 1979; Fejer, 1981, 1986; Spiro et al., 1988]. We therefore concentrate first on the equatorial zonal electric field measurements. Figure 1 shows the auroral electrojet indices, the north-south component of the IMF measured by the IMP-J satellite, the horizontal component of the magnetic field over Huancayo, Peru (magnetic dip 1.2øN), and the Jicamarca zonal electric fields during a 48-hour period. The Jicamarca data show a large eastward electric field perturbation relative to the quiettime pattern starting at about 1020 UT on January 18 when Bz turned sharply northward. Then a very unfortunate data gap occurred, after which Bz was steadily northward for over three hours. Eastward electric field perturbations are commonly observed in this time sector over Jicamarca following sudden northward IMF changes that occur after about an hour or longer of southward IMF Bz [e.g., Kelley et al., 1979; Fejer, 1986]. Substorm activity is often curtailed when Bz turns northward suddenly. Both of these characteristics are exhibited in the UT period. 5OO I0 I0 0 IOO O JANUARY 1984 O O O0 Fig. 1. Auroral electrojet indices, north-south component of the interplanetary magnetic field, horizontal component of the magnetic field in Huancayo, and eastward electric field over Jicamarca during January 18-19, The smooth curves denote the average quiet-time variations, and the small circles indicate local midnight. The Sondre Stromfjord radar data, shown in Figure 2, and in more detail by de la Beaujardieret al. [1988], confirm this interpretation. When Bz turned southward at around 0700 UT, tl e east-west velocity grew steadily more eastward, indicating a strengthening of the sunward convection in the prenoon auroral oval. At about 1020 UT, when Bz abruptly turned positive, the convection over Sondre Stromfjord began to decay with roughly a one-hour time constant. The time period during which the southward auroral oval electric field (eastwardrift) decreased to zero matches exactly the eastward perturbation Jicamarca. The event is very much reminiscent of that of April 13, 1978 [Gonzales et al., 1979], in which a northward IMF turning generated a simultaneous "bite-out" in the auroral zone convection in the Alaskan sector and an eastward electric field turning at Jicamarca. The other characteristics of the high-latitudelectric field data shown in Figure 2 will be discussed later. Uo'[

4 FEJER ET AL.' LOW- AND MID-LATITUDE ELECTRIC FIELDS OO OO ( ' 18-'19 JANOAR' ICj84 ' IMF ElSCAT v, VN MILLSTONE HILL v.!,, - - _ i i i i i i i O0 12 O Fig. 2. Interplanetary and auroral electrojet magnetic field data, and high-latitude eastward and northward F region plasma drifts observed with the EISCAT, Sondre Stromfjord, and Millstone Hill radars during a 48-hour period. The electric field measurements were made at invariant latitudes of 66 ø, 74 ø, and 65 ø, respectively. The small circles indicate local midnight. The subsequent electric field pattern over Jicamarca shown in Figure 1 closely followed the quiet-time variation up to about 1700 UT on January 18 when, shortly after a southward Bz turning, substorm activity began once again. At this time a small-amplitude westward electric field perturbation was recorded in the Jicamarca data. This effect is seen even more clearly in the Huancayo magnetic field data, which also indicate the presence of a westward electric field perturbation. The Jicamarca electric field perturbation was westward up to about 2400 UT (1900 LT). The westward electric field perturbation relative to the quiet-time pattern during this period might not have been due to the high-latitude substorm activity since there are considerable variations in the quiet-time patterns particularly in the afternoon and evening sectors. In fact, the afternoon electric fields on 17 January (not presented here), which was a quiet day, were also more westward than the quiet-time pattern (smooth curve) shown in Figure 1. Daytime westward electric field perturbations lasting several hours were also present over Jicamarca and Huancayo about 24 hours later. It is important to note that these perturbations (16-24 UT on January 18 and UT on January 19) last for many hours. Long-lasting and slowly varying equatorial electric field perturbations have been discussed by Fejer et al. [ 1983] in the context of a disturbance dynamo. Therefore, the rapid variations we see in these data (e.g., UT on January 18, UT, UT, UT on January 19) may be due to direct penetration of electric fields while the slow variations are likely due to dynamo effects. A very large eastward electric field perturbation was observed over Jicamarca at about 00 UT on January 19 coincident with the onset of substorm activity as indicated by the auroral magnetograms. This eastward electric field U.T. perturbation ceased suddenly at about 0040 UT as Bz turned north-ward, and at the same time as the AL index decreased. Very large and longer-lasting eastward electric field perturbations were also seen in the postmidnight sector later on the same night. For isolated substorms these postmidnight equatorial perturbations are almost always associated with sudden northward Bz turnings and with onsets of substorm recovery phases. However, the relationship between the Jicamarca electric field perturbations and changes in the highlatitude ionosphere between about UT is not clear since IMF data from the IMP-J satellite were not available for most of this period. The prenoon (12-17 UT) drifts on January 19 were fairly close to the quiet-time values but, as noted above, the afternoon drifts were clearly below the quiettime pattern. The sudden impulses observed on the Huancayo magnetometer data at about 1800 UT on January 18 and at 1600 UT on January 19 are essentially absent on the electric field data. It seems unlikely that this absence is due entirely to the averaging of the electric field data; it is more likely that the magnetometeresponded to a magnetosphericurrent system not mirrored in the ionosphericurrents. Although the IMF data are very spotty, we have some help from the high-latitude incoherent scatteradar data in Figure 2. The period of transition in Bz from north to south, followed by seven hours of average Bz southward (16-24 UT on January 18), has been studied in some detail by de la Beaujardier et al. [1987a,b, 1988]. They show that the reversal in the meridional field at Sondre Stromfjord at 1800 UT was due to the equatorward motion of the convection reversal, not to a change in the electric field (convection) intensity. It corresponds to the time when, directly overhead of Sondre Stromfjord, the sun-ward convection shifted to antisunward without any significant change in the magnitude of convection. (The convection had intensified earlier at about 1600 UT.) Indeed, no large perturbation in the low-latitude electric field was observed. On the other hand, the event observed at Jicamarca near 00 UT on January 19 is somewhat mysterious. It begins at the time of a substorm onset and ends with a sharp Bz northward excursion. Unfortunately the IMF data end abruptly at this time. We can say only that the period during which the AE index was large ended shortly thereafter but it is not clear whether the end was triggered by Bz or whether it was related to processes internal to the magnetosphere and associated with substorm activity. The high-latitude electric fields decayed within an hour or so of midnight UT but were more or less unchange during that period. Figure 3 shows the Jicamarca (12.0 ø S, 76.9 ø W, magnetic dip 2 ø N) vertical plasma drifts, the variation of h'f (virtual height of the bottomside F layer) over Fortaleza, Brazil (38 ø W, 4 ø S, magnetic dip 3.5 ø S), the horizontal magnetic field data over Trivandrum, India (1 ø S, ø E, geomagnetic), and the Dst index during the most active 24-hour period of this campaign. Bittencourt and Abdu [1981] have shown that near the geomagnetic equator, when h'f is above about 300 km, the temporal variation of this parameter provides a good measure of the vertical plasma drift velocity (i.e., of the zonal electric field). The typical rise in h'f, starting at about 2000 UT (1700 LT over Fortaleza), corresponds to the prereversal enhancement of the upward plasma drift [e.g., Woodman, 1970; Fejer, 1981] which is driven by F region dynamo electric fields [e.g., Rishbeth, 1971; Heelis et al., 1974]. The enhancement ended at about 2200 UT since h'f was virtually constant for two hours. The west-ward electric fields observed over Jicamarca in the afternoon sector did not extend as far as Fortaleza, about 40 ø east of Jicamarca, as is usually the case for disturbance dynamo electric field effects [e.g., Fejer et al., 1983; Fejer, 1986]. The large eastward electric field perturbation starting at about UT over Jicamarca was well correlated with the sudden rise in h'f at Fortaleza. Therefore, this eastward electric field perturbation, which started with the onset of substorm activity, as indicated

5 2370 FEJER ET AL.' LOW- AND MID-LATITUDE ELECTRIC FIELDS m/s o' km V z J..AMARCA h'f FORTALEZA AH TRIVANDRUM I!! I JANUARY 1984 I Figure 4 shows the variation of the upward/northward plasma drift velocity at Jicamarca, Arecibo, and Millstone Hill, and the high-latitude and interplanetary magnetic field data during a 48-hour period. The Arecibo and Millstone Hill data were averaged for 1 hour and 1.5 hours, respectively, to reduce the statistical fluctuations. The typical error bars for the data from these stations are about 5 m/s and 10 m/s, respectively. The three radars are nearly in the same local time sector. The smooth curves superposed on the Millstone Hill local data correspond to the averagempirical quiet-time winter solstice model for Kp = 0 [Wand and Evans, 1981], whereas the Arecibo quiet-time pattern corresponds to the local winter average for Kp _< 2+ [Ganguly et al., 1987]. Over Arecibo and Millstone Hill an eastward electric field of 1 mv/m corresponds to a northward drift, perpendicular to the magnetic field, of about 27 and 20.6 m/s, respectively. nt 20- DST I ' ' ' I ' ' ' I ' ' UT O Fig. 3. F region vertical plasma drifts over Jicamarca, h'f over Fortaleza, Brazil, horizontal component of the magnetic field over Trivandrum, India, and the Dst index during the central 24 hours of the 48-hour period illustrated in Figures 1 and 2. The small circles indicate local midnight A:64' o- - o LLSTONE H LL - 2 =55" > by the high-latitude magnetic field indices (see Figure 1), 40- occurred simultaneously over a range of at least 40 ø in longitude at the equator. V -40- Figure 3 also shows a series of simultaneous perturbations 0 0 in the Jicamarca plasma drift velocity, the height h'f over Fortaleza, and the magnetic field data from Trivandrum starting at about 0530 UT on January 19. The first pertur- 40- ARECI80 O- ' 0 bation had a small amplitude in the Jicamarca vertical drift data, but is clearly seen in the data from the other two -40- equatorial stations. The large upward drift velocity perturbation over Jicamarca about one hour later (at about UT) was correlated with a simultaneous decrease in the horizontal component of the magnetic field over Trivandrum which corresponds to a westward electric field perturbation. These data illustrate the local time dependence of the low latitude electric field perturbations of high-latitude origin [Fejer, 1986]. Note, for example, that the Jicamarca vertical O0 08 I 6 O O0 drift perturbations on the nightside between 07 and 08 UT (02-03 LT) had a magnitude of about 50 rn/s (corresponding to an eastward electric field perturbation of about 1.2 mv/m). On the other hand, the amplitude of the perturbation on the dayside was less than about one half of the maximum daytime Fig. 4. Interplanetary and auroral electrojet magnetic field data together with the zonal electric field data at four different latitudes in the American sector. The smooth curves represent the quiet-time patterns, and the small circles indicate the local midnight. AH amplitude (when the magnetospheric current contribution, indicated by Dst, is removed), corresponding to a westward electric field perturbation of less than 0.25 mv/m. Due to the great degree of variability in the Millstone Hill and Arecibo data sets there are only two periods when 2.3. Relationship Between the Zonal Electric FieM Perturbations at Middle and Low Latitudes something significant can be said. The first is UT on January 19 when very large perturbations were seen at Jicamarcand Fortaleza but not over Millstone Hill (A = 55 ø) or Arecibo. This is significant (and mysterious) since the For clarity, we will discuss the zonal electric field (i.e., upward/northward drifts) and south-ward electric fields (eastwar drifts) separately. As mentioned previously, electric magnitude and spikelike nature of the event should have made it detectable. No such perturbation was observed even on the original Arecibo data obtained with an integration time of 16 field measurements at mid-latitudes were made with the min. The equatorial electric field perturbation occurreduring Millstone Hill, Arecibo, and Saint-Santin radars. The aperiod of large westward electric field north of Millstone Hill Millstone Hill radar also provided the high-latitude (A = 64 ø) (A = 64ø). The other time period worth discussing observationshown in Figure 2. We will now compare these measurements with the Jicamarca results. UT on January 19 when eastward perturbations were detected at Millstone Hill, Arecibo, and Jicamarca. Electric field JICAMAR O 0

6 ß FEJER ET AL.' LOW- AND MID-LATITUDE ELECTRIC FIELDS 2371 perturbations were also inferred from the Fortaleza and Trivandmm data during this period (see Figure 3). Only two other latitudinal studies of this type exist [Gonzales et al., 1979; Gonzales et al., 1983]. They presenthree events with large eastward perturbations observed at Jicamarca simultaneous with changes at either Millstone Hill (April 13-14, 1978) or at Arecibo (October 10-11, 1980). As mentioned above, the Arecibo and Millstone Hill data were averaged for about 1 hour in Figure 4. The Arecibo data show typically large fluctuations, making the unambiguous evaluation of the electric field perturbations of high-latitude origin quite difficult. This is particularly the case in the postmidnight period, where there is large variability in the plasma drifts even during relatively quiet conditions. For example, very large northward velocities were observed at about 0300 LT (0700 UT) on January 17 (not shown here), at a time when only modest activity was observed at high and equatorialatitudes. Earle and Kelley [ 1987] have pointed out that gravity wave activity may provide a source of electric fields with periods of a few hours which compete with geomagnetic sources. To summarize, the results shown here indicate that the electric field perturbations observed at the equator in the postmidnight sector and associated with the onset of substorm recovery phases, possibly triggered by sudden northward Bz turnings, are associated with similar changes at low and mid latitudes, at least for the same meridional sector. The latitudinal variation of these zonal electric field perturbations will be examined later. In the evening sector, there seems to be no mid- and low-latitude zonal electric field perturbation associated with the eastward electric field changes, observed at the equator in the American sector during a time of a large increase in substorm activity, followed quickly by a rapid recovery Meridional Electric Fields The latitudinal variation of meridional electric fields has not been studied in detail so far. Figure 5 shows the zonal plasma drifts (north-south electric fields) during the January period for Millstone Hill, Saint-Santin, and Arecibo. Zonal plasma drifts were not measured over Jicamarca during this period. However, it is well known that the equatorial zonal plasma drifts (vertical electric fields) are considerably less affected by magnetic activity than the vertical plasma drifts [Woodman, 1972; Fejer et al., 1981; 1985; Gonzales et al., 1983]. Conversely, this velocity component was measured over Saint- Santin using the procedure described by Blanc and Amayenc [1979]. The smooth curves correspond to the quiet-time empirical curves for Kp = 0 at Millstone Hill [Wand and Evans, 1981] and for Kp _< 2+ for Saint-Santin [Blanc and Amayenc, 1979]. Figure 5 shows the clear predominance of westward drift velocity perturbations. At Millstone Hill, the observations are below the average curve for nearly the entire two-day period. This is particularly surprising for January 18 when the largest Kp values were 2+ during UT and 3+ during UT. According to the empirical model of Wand and Evans [1981] for winter, a southward electric field perturbation of about 6 mv/m at about 23 UT would be typical of a period of Kp - 5. With only a brief exception, the Arecibo and Saint- Santin measurements were also westward relative to the quiet curves. The exception was a Saint-Santin observation of a large eastward electric field perturbation between about 16 and 20 UT on January 18, which had no counterpart in the American sector. The Saint-Santin and Millstone Hill observations were discussed in detail by Senior and Blanc [1987]. The Millstone Hill and Arecibo zonal drift for the entire three-day period is shown in Figure 6. Note that the daily variation in the Arecibo westward drift did not change -10 5OO' I, -50 MILLSTONE HILL :500 A=64 ß / _ A /'"'X..._ - 0 HILL --' MILLSTONE 60 A =55 ø _J o / ø }- -3 LU -60 -Z o -120 i i i i i i i i i i O O O0 U.T. Fig. 5. Interplanetary and auroral magnetic field data and southward ionospheric electric field data at four different latitudes. The smooth curves denote the quiet-time average variations, and the small circles indicate local midnight JANUARY i ß i ß! ß i ß i ß i ß i i i ß MILLSTONE; HILL o - AREClBO o - ß i ' 1 ' i. ' i ' i ' i ß i ' i i ' 16 O O O O0 Fig. 6. Eastward F region drift velocity for Millstone Hill (A = 55 ø) and Arecibo (A = 33ø). appreciably during this entire period even though there were large variations in the Millstone Hill drifts, particularly near the dusk sector ( UT). This suggests that the lowlatitude zonal plasma drifts were primarily due to thermospheric dynamo electric fields rather than to the penetration of meridional electric fields from high latitudes. The Arecibo zonal drift measurements during January are shown in a common plot in Figure 7. Of these three days, January 18 was the quietest and January 19 was the most disturbed. Notice that the main effect of the increased magnetic activity seems to be the occurrence of increased U.T.

7 2372 FEJER ET AL.' LOW- AND MID-LATITUDE ELECTRIC FIELDS i i i,! i i i i i i i i I E 4 "'. J""'. r ', t o JZ\-X m / \ \ /_.,"? : ß -,.,,----ø w t ' ,9 jo.uory 1-12o/,,,,,,,,,,, / Fig. 7. Eastw d plasma ft the 3-day GISMOS period. LOCAl,. TIME velocity from Arecibo during westward drifts at earlier local times in the postmidnight sector. The data shown in Figure 7 are in good agreement with the results by Ganguly et al. [1987] for 2+ < Kp < 4. These authors suggested that disturbance dynamo effects are the main source for mefidional electric field perturbations over Ancibo during active periods. Therefore, the amplitudes of meridional electric field perturbations associated with enhancements of the high-latitude ionospheric currents decrease significantly with latitude. At Ancibo these electric field (zonal drift) perturbations have small amplitudes relative to both the quiet-time and the disturbance dynamo values, and at equatorial latitudes these perturbations are usually undetected Summary of Experimental Results 1. As reported before, large zonal electric field perturbations wen observed simultaneously over a large range of latitudes and longitudes. The zonal electric field perturbations are westward during the day and eastward at night. The amplitudes are largest in the postmidnight sector when magnetic activity is decreasing. 2. In an entirely new nsult, a large eastward electric field perturbation observed over Jicamarca and Fortaleza in the dusk sector seemed unrelated to similar changes at higher latitudes, at least in the American sector. 3. Meridional electric field perturbations observed simultaneously at high and middle latitudes as a result of convection electric fields during active periods seem to have very small amplitudes in the low-latitude ionosphere. 4. The main effect of magnetic activity on the zonal plasma drift at low latitudes occurred in the postmidnight sector, a few hours later than at mid-latitudes, and led to increased westward plasma drifts. On these occasions, the shift from eastward to westward drifts occurred earlier than during quiet periods. 5. The zonal and north-south electric field perturbations at low and equatorialatitudes do not occur simultaneously and have different latitudinal and longitudinal dependence and driving mechanisms. The zonal electric field perturbations occur when then are large changes in convection and penetrate efficiently down to the equatorial ionosphen. The high- and mid-latitude meridional electric field perturbationseem to be caused mostly by the penetration of convection electric fields. At low latitudes disturbance dynamo effects seem to be an important source of meridional electric field perturbations. 3. DISCUSSION The penetration of high-latitudelectric fields to middle and low latitudes has been examined in a number of theoretical and numerical studies [Wolf, 1970; Nopper and Carovillano, 1978; Nisbet, 1978; Maekawa, 1980; Kamide and Matsushita, 1981; Spiro et al., 1981; 1988; Senior and Blanc, 1984; 1987] based on the convection model approach introduced by Vasyliunas [1970]. The global ionospheric currents and electric fields are obtained by applying either an electrostatic potential drop across the polar cap or field-aligned currents on a two-dimensional (thin shell) ionosphen, with assumed models of ionospheri conductance. In this section we will compare our observations with previous results and with pndictions from numerical models. The models to be examined in more detail hen are the Senior- Blanc semiempiricalinear convection model [Senior and Blanc, 1984], the Rice University convection model (RCM) [e.g., Wolf et al., 1986; Spiro et al., 1988], and the model described by Tsunomura and Araki [1984]. The Rice model has recently been extended to low latitudes by introducing nalistic equatorial boundary conditions and a denser grid at low latitudes, and it has been used to examine neutral wind effects on the low-latitude electric fields [e.g., Spiro et al., 1988]. The work of Tsunomura and Araki [1984] was concerned mainly in explaining the pnliminary impulse of geomagnetic sudden commencement. However, similar models, using the field-aligned currents and realistic ionospheri conductivities, have been used by a number of other authors to study the middle- and low-latitude ionospheric naction to sudden changes in the high-latitude current system [e.g., Maeda and Maekawa, 1973; Kamide and Matsushita, 1981] Daily Variation of the Perturbation Electric Fields Figure 8 shows the variations of the equatorial zonal electric fields as a function of local time obtained from the three numerical studies mentioned above. The results from the Senior-Blanc and from the Rice models are the initial response corresponding to a sudden incnase in the polar cap potential drop by 100 kv. The pattern from the Tsunomura-Araki model, derived with the assumption that the field-aligned currents do not affect the conductivity in the source region, was obtained by multiplying their electric field nsponse curve by a factor of 10 to correspond to a magnetic field perturbation of 100 nt at the dayside equator. We have also changed the sign of electric field pattern derived by Tsunomurand Araki [1984] to be consistent with an increase in the polar cap potential drop. Figure 8 shows that the thne models give essentially the same initial time response for the equatorial electric field nsulting from a sudden increase in the polar cap potential drop (or in the high-latitude field-aligned currents) with eastward perturbations in the dayside and eveningside and westward perturbations in the morningside. The eastward electric field perturbations occur over a larger time sector as a 1 i i i 1 i i i i i i t INITIAL TIME RESPONSE A o:lookv SENIOR-BL,NC [ 1984 ] TSUNOMURAIARAKI [ 1984]... SPIRO ET AL. [1988] / '.,.. /. I 04 0 I I o 24 LOCAL TIME Fig. 8. Initial time zonal electric fields resulting from a sudden incnase in the polar cap potential drop by 100 kv.

8 FEJER ET AL.' LOW- AND MID-LATITUDE ELECTRIC FIELDS 2373 result of the daytime conductivity enhancement. Conductivity effects are also responsible for most of the differences in the electric field amplitudes obtained by the different models. In particular, the very small eastward electric fields in the dusk sector in the RCM pattern are a result of the very large assumed values for the field-line integrated Pedersen conductivity, which has the effect of shorting out the equatorial electric field perturbations. The model electric field patterns shown in Figure 8 show that in the evening sector the smallest eastward electric fields are given by the Senior-Blanc model. As time evolves after the sudden increase in the polar cap potential drop, this initial electric field response decays as a result of the readjustment of the shielding currents in the inner magnetosphere. The time constant for this shielding effect varies from about 10 min. to 1 hour, depending on the plasma sheet temperatures. For a sudden decrease, not increase, in the polar cap potential drop, the electric field perturbations would have the opposite signs, that is, westward in the dayside and eveningside and eastward in the morningside. To compare the electric field perturbations with the model results shown in Figure 8, we need to examine initially the time variation of the polar cap potential drop. Figure 9 shows IMF Bz data and two models for the polar cap potential drop. The first one, calculated by Senior and Blanc [1987] and used for their simulation of the high- and mid-latitude convection during January 18-19, 1984, was determined from the IMF data using the empirical formula derived by Reiff and Luhmann [ 1986]. In this case, the time delay between changes in Bz and in the polar cap potential drop was chosen as 30 min based on the variations in the Ds[, AU; and AL indices. The second o curve corresponds to the difference between the potential maxima and minima derived by Richmond et al. [1988] for the GISMOS period. They followed the procedure developed by Richmond and Kamide [1988] which relies on simultaneous use of ground-based magnetic field observations, electric fields measured by radars, and conductivities to determine self-consistent electrodynamical patterns at high latitudes. Upward and downward pointing arrows indicate the starting times of some of the equatorial electric field perturbations either observed at Jicamarca or inferred from the Huancayo, Fortaleza, and Trivandrum data. We will compare these perturbations with the electric field model results taking into account the polar cap potential variations in Figure 9. We should keep in mind that increases (decreases) in the polar cap potential drop should result in eastward (westward) electric field perturbations in the daytime sector and westward (eastward) in the nighttime sector. Figure 9 shows that the increase in southward Bz starting at about 1630 UT resulted in gradual increases in the polar cap potential drop. This increase in ti)o did not produce simultaneous eastward electric field perturbations (see Figure 1). The numerical models suggest that the low-latitude response to a gradual increase in the potential drop over an hour or two is considerably less intense than shown in Figure 8, in agreement with the absence of corresponding electric field perturbations in the equatorial data. In contrasto the usually small effects associated with south-ward Bz changes and with increases in the polar cap potential drop, even small decreases in the polar cap potential drop associated with large and sudde northward Bz changes often result in large electric field perturbations (eastward during the day and westward at night) at Jicamarca [Gonzales et al., 1979; Fejer et al., 1979; Fejer, 1986]. At resent, this asymmetric response of the equatorial electric eld to changes in Bz and in the polar cap potential drop remains unexplained. The other major difficulty for the global convection model is the observation of longer-lasting electric field perturbations associated with sudden decreases than with increases the polar cap potential drop. An example of a longlasting eastward electric field perturbation is seen in the Jicamarca data starting at 0820 on January 19. Recently, it has been suggested that neutral wind effects might explain this!! i! i!!!!! i i IMFB z -I0 - to0 (:I:)o/ "T= SENIOR ANO BLANC [ ! I! II I! I I I I 1 I I I I I I I I I O I0 12 Fig. 9. North-south component of the interplanetary magnetic field and estimated polar cap potential drop profiles during part of the most active January 1984 GISMOS period. The upward and downward pointing arrows indicate the starting times of equatorial zonal electric field perturbations consistent with increases and decreases in the polar cap potential drop, respectively. asymmetric response and the long-lasting perturbations associated with sudden decrease in o and with sudden northward Bz changes [e.g., Spiro et al., 1988]. The very large eastward electric field perturbation observed at Jicamarca and Fortaleza starting at about UT ( LT) on January 18 (see Figure 3) occurred at nearly the same time as the large decrease in the AL index. Senior and Blanc [ 1987] pointed out that this intensification of the westward electrojet occurred in the dawn sector following the ring current injection event between 23 and 00 UT. This large eastward electric field perturbation cannot be explained by the convection models as an initial time response to an increase in the polar cap potential drop if we assume the potential drop derived from the solar wind and IMF data by Senior and Blanc [ 1987], but it would be consistent with the pattern derived by Richmond et al. [1988], which shows a large increase in the polar cap potential drop. These results show the importance of realistic estimates of o for model calculations. As mentioned previously, however, this equatorial electric field enhancement was not associated with corresponding effects over Arecibo and Millstone Hill although large west-ward fields were observed north of Millstone Hill (A = 64 ø) during this period. The equatorial electric field enhancementstopped very abruptly at about 0040 UT (1940 LT over Jicamarca) at the same time as Bz turned northward and the polar cap potential drop decreased. The o derived by Senior and Blanc [1987] and shown in Figure 10 would predict a westward electric field perturbation U.T. J-- SENIOR-BLANC [ { INITIAL TIME RESPONSE ^ / 1-,,,o:,OO,V l\ a.-- ol.=56' IX -I z ' ' ca '- m l=o ß ca. I / 4 tam I.. -8 O IZ 16 ZO 24 LOCAL TIME Fig. 10. Initial time responses for a polar cap potential drop of 100 kv obtained by the Senior-Blanc model at the invariant latitude of Millstone Hill and at the equator.

9 2374 FEJER ET AL.: LOW- AND MID-LATITUDE ELECTRIC FIELDS at about 0100, about half an hour later than actually observed. Here again the more detailed ti)o calculations of Richmond et al. [1988] result in better agreement with our results. Richmond et al. [1988] suggested that the total potential drop across the polar cap responds rapidly to changes in the IMF, but the high-latitude current system usually responds with a time lag. Our results indicate that zonal electric field perturbations the equator are better correlated with changes in the polar cap potential drop than with variations in the intensity of the high-latitude currents. These results illustrate both the progress in the simulation studies and their present limitations. The responses of the electric field to simple changes in the polar cap potential drop are well reproduced by the models. However, this is not the only high-latitude parameter that should have a strong effect on electric field penetration even if we neglect neutral wind effects. Ionospheric conductivity, for example, plays an important role in the amplitude of the electric field perturbations and in the shielding time constant [Jaggi and Wolf, 1973; Siscoe, 1982]. Large longitudinal conductivity gradients in the dusk sector might explain the absence of an eastward electric field perturbation over Millstone Hill and Arecibo at about 2400 UT on January 18. There are other effects such as changes in the radius of the polar cap (corresponding to increase or decrease of the magnetic flux in the tail lobe), and the release of plasma down in the tail occurring in substorms which must also cause electric field penetration effects (R. Wolf, private communication, 1988). These last two effects have not been modeled realistically so far. The east-west drifts observed with the EISCAT, Scandinavian Twin Auroral Radar Experiment (STARE), Millstone Hill, and Saint-Santin radars during the January 1984 campaign were compared with the predictions of the Senior-Blanc model [Senior and Blanc, 1987]. They showed good agreement at high-latitude stations, but there were some significant disagreements at the middle latitudes particularly for the period between about 1800 and 2100 UT where the model predicted a northward electric field of about 3.5 mv/m for Saint-Santin, and the observations showed a southward electric field of about 4 mv/m. It was suggested that this disagreement could have been caused by a local electric field source. As mentioned previously, meridional convection electric field perturbations usually do not seem to penetrate efficiently into the low-latitude ionosphere Latitudinal Variation of the Perturbation Electric Fields Mozer [1970] showed that, assuming a centered dipole geomagnetic field with equipotential field lines, the zonal and meridional ionospheric projections of an electric field independent of the radial distance in the equatorial plane would vary as L3/2 and 2L(L-3/4)l/2, respectively. A similar expression for the latitudinal variation of the electrostatic potential was used by Volland [1975]. We use these simple formulas only as a general guide since there is no theoretical reason to expect a uniform electric field in the magnetospheric equatorial plane. So far, very few model studies have examined the longitudinal variation of perturbation electric fields of magnetospheric origin. Senior and Blanc [ 1984] calculated the zonal electric field initial time responses at Millstone Hill and at the equator for a sudden increase in the polar cap potential drop of 100 kv. Their results are shown in Figure 10. Here we have used a vertical scale proportional to L3/2 so that electric field perturbations following this mapping would have identical amplitudes as measured in, say, centimeters. This figure shows that the predicted ratio between the equatorial and the Millstone Hill electric field perturbations is largest (comparable to L3t2) at about 04 LT and smallest in the dusk to midnight period. The simulations with the RCM also indicate that the largest low-latitude electric field perturbations are found in the late postmidnight period [Spiro et al., 1981]. Gonzales et al. [1980] showed zonal electric field measurements in the range L = using whistler duct drifts and Millstone Hill incoherent scatter observations, consistent with the theoretically predicted mapping. Earle and Kelley [1987], using spectral analysis of zonal electric field data from Chatanika (L = 5.51), Arecibo, and Jicamarca over the range 1-10 cycles/hour during two active days, observed that the variations between Chatanika and Arecibo were consistent with the L3/2 mapping. On the other hand, the amplitude ratio between Jicamarcand Arecibo was about 75% lower than expected, assuming the same mapping. Blanc et al. [ 1980], using 8 hours of simultaneous observations, obtained a ratio of from Malvern (L = 2.6) and Saint-Santin (L = 1.8), which is steeper than the variation corresponding to a uniform dawn-to-dusk electric field in the equatorial plane. Their data suggest that the shielding mechanism is still operating between L = 2.6 and L = 1.8. Figure 11 shows the latitudinal variation of the mid- and low-latitude zonal electric field obtained by subtracting the quiet-time electric field patterns from the measured values presented in Figure 4. Here we again use a vertical scale proportional to L3t2 so that electric perturbations following this mapping would have the same amplitudes (say, in centimeters) at different latitudes. No clear latitudinal variation can be determined for the first day, partly due to the uncertainty of Millstone Hill and Arecibo measurements. However, the perturbation electric field starting at about 04 UT on January 19 seems to follow the L3/2 mapping of Mozer and the calculations of Senior and Blanc [ 1984] for a sudden decrease in the polarpotential drop, probably associated with a sudden northward Bz turning. Assuming the theoretical mapping shown in Figure 10, the large perturbation shortly after 00 UT on January 19 would correspond to an eastward electric field of about 6 mv/m over Millstone Hill and about 2 mv/m over Arecibo. In spite of the relatively long integration time used by the Millstone Hill radar, an electric field perturbation of this magnitude would have produced a considerably larger effect than what was observed. As mentioned previously, the Arecibo observations were actually made with an integration time of about 16 min and would, therefore, also have detected a much larger effect if this perturbation had occurred overhead. The results above indicate that two important points must be..e JANUARY!984 MILLSTONE HILL '-4 ARECIBO. ' / I O O O0 U.T. Fig. 11. Zonal electric field perturbations obtained by subtracting the quiet-time patterns from the measured electric field.

10 b'ejer ET AL.: LOW- AND MID-LATITUDE ELECTRIC FIELDS 2375 kept in mind when using radar data to study the latitudinal variation of zonal electric field perturbations of high-latitude origin. First, even when observations are made in nearly the same longitudinal sector, not all perturbations are present at all sites. Second, the ratio of electric field perturbations middle and low or equatorial latitudes is strongly local time dependent. Several radar studies have shown that high-latitude northsouth electric field perturbations are not usually detected at the magnetic equator [e.g., Woodman, 1972; Gonzales et al., 1983; Fejer, 1986]. The variation in the north-south component of magnetospheric electric fields between high and middle latitudes was studied using simultaneous observations from STARE ( ø geomagnetic latitude) and Saint-Santin [Mazaudier et al., 1984; 1987]. In the first of these studies, the north-south electric fields inferred from the radar measurements and from ground-based magnetic field data near local noon were compared with the predictions from the Senior- Blanc model. The theoretical results were found to be in good agreement with the experimental profiles provided that conductivities were normalized using the STARE electric field and the corresponding magnetic field data. In this case, the initial time response for the north-south electric fields was estimated to decrease by a factor of about 4.2 between 60 ø and 40 ø in excellent agreement with the mapping expected for a constant electric field in the equatorial plane. Mazaudier et al. [1984, 1987] pointed out that the east-west electric field component is not very well reproduced by the Senior-Blanc model, at least for the noon time sector. The simulation underestimates the ratio between the zonal and northward components at high latitudes but overestimates it at Saint- Santin. Kamide and Matsushita [ 1981] have examined the electrical coupling between high and low latitudes by solving numerically the steady current continuity equation with fieldaligned currents and anisotropiconductivities. Their results indicate the north-south electric field perturbations decay more rapidly with latitude in the evening than in the morning sector as a result of the longitudinal asymmetry in gradient of ionospheri conductivity. In the simplest case, for uniform ionosphericonductivities, their computations near subauroral latitudes indicate a decrease in the north-south electric field proportional to L1.4. For realistic conductivity distributions, the electric field decrease with latitude is larger in the evening sector than in the morning sector. For a typical substorm the latitudinal variations were found to be proportional to/52.8 and L1.8, in the evening and morning sectors, respectively. Furthermore, these variations increase with the intensity of the field-aligned currents in the equatorward side of the auroral oval. If we use the Millstone Hill measurements at the two invariant latitudes shown in Figure 6, we get a southward electric field perturbation proportional to about/52.2 in the dusk sector. This would give an amplitude of about 1 mv/m at the latitude of Arecibo, which is within the experimental error. These results indicate that the north-south component at middle and low latitudes is smaller than that predicted by the homogeneous conductivity case, in agreement with the computations of Kamide and Matsushita [1981]. From our data, we cannot estimate the latitudinal decrease for the northsouth electric field in the morning sector. Additional studies of the latitudinal variation of the northsouth component of the ionospheric electric field were performed by Spiro et al. [1981] and Tsunomurand Araki [1984]. The results from these studies cannot be related directly to our observations, but both are consistent with the experimental data and previous theoretical results indicating that the zonal electric field perturbations penetrate more efficiently into the low-latitude ionosphere than the meridional electric fields. However, more detailed simulations are necessary to better understand this electric field mapping process Thermospheric Effects Blanc and Richmond [1980] showed that changes in the global thermospheric circulation from storm-time Joule heating can result in significant changes in the middle- and low-latitude ionospheric electric fields. The occurrence of disturbance dynamo electric field effects with a latitude-dependent time delay of a few hours or more [Blanc and Richmond, 1980] has been detected at middle, low, and equatorialatitudes [Blanc, 1983; Fejer et al., 1983; Mazaudier et al., 1985; Ganguly et al., 1987]. The mid- and low-latitude electric fields predicted by the disturbance dynamo model are poleward at all local times [Blanc and Richmond, 1980]. On the other hand, at very low latitudes the daytime drifts have a small equatorward component. Disturbance dynamo zonal electric fields are westward during the day and eastward at night, that is, have the same sign as the perturbation electric fields associated with a decrease in convection. As a result, it is often difficult to separate the electric field perturbations due to these two processes. Although we have not done a detailed study of disturbance dynamo effects in this paper, it is possible that this mechanism was responsible for the slowly varying departures from the quiet-time patterns over both Arecibo and Jicamarca. For example, this explanation is consistent with Jicamarca zonal electric field data between about UT on January 18 and UT on January 19, and also with the large postmidnight Arecibo westwar drifts. 4. CONCLUSIONS We have examined an extensive data set of ionospheric electric field measurements that show very clear examples of the perturbation of the mid-latitude, low-latitude, and equatorial electric field associated with the high-latitude current systems. The equatorial observations, which have excellent time resolution, confirmed that the zonal electric field perturbations driving the F region vertical plasma drifts and the equatorial electrojet occur simultaneously on a global scale. These IMF-driven processes affect simultaneously the high-, middle-, and low-latitude ionosphere. As reported previously, we have seen that the largest zonal electric field perturbations occur in the midnight-dawn sector. We also show that large eastward electric field perturbations can occur in the equatorial region over a longitudinal range of at least 40 ø with no corresponding signatures in the same longitudinal sector at middle and low latitudes. The low and equatorial ionospheres are considerably better shielded from meridional electric field perturbations than from zonal electric field fluctuations. The low-latitude meridional component is probably also affected by disturbance dynamo processes. We have compared in detail our data with results from a number of global convection models. However, there are great difficulties that make this comparison far from trivial. From the experimental aspect these are, for example, the separation of magnetospheric electric field effects from the (quiet or disturbance) dynamo variation, the accuracy, the relatively long integration times, and the absence of significant longitudinal coverage from the middle- and low-latitude data. On the other hand, the model inputs such as the polar cap potential drop, ionospheri conductivities, etc., are highly idealized. Nevertheless, the numerical models can explain several aspects of the data. They predict the correct daily variation of the electric field perturbations for increases and decreases in the polar cap potential drop, and the occurrence of the largest low-latitude zonal electric field perturbations in the midnight-dawn sector. The latitudinal variation of the zonal electric field perturbations is also in reasonably good agreement with the observations. The models cannot explain the highly asymmetric response of the zonal electric field perturbations to sudden increases and decreases in the potential drop. It is also not completely clear why the equatorial zonal

11 2376 FEJER ET AL.: LOW- AND MID-LATITUDE ELECTRIC FIELDS electric field perturbations associated with decreases in the polar cap potential drop triggered by sudden northward Bz changes last considerably longer than those associated with sudden increases in convection. The longitudinal variation in the meridional electric field perturbations particularly at low and equatorial latitudes is not well explained by the models either. There are also a number of processes that have not been examined in detail as far as the low-latitude electric field perturbations are concerned. They include thermospheric effects, substorm-generated electric fields, the effect of gravity waves, of disturbance dynamo electric fields, IMF By effects, longitudinal effects, the relationship between the drift velocity perturbations parallel and perpendicular to the Earth's magnetic field, and others. The extensive experimental and modeling studies using coordinated World Day observation should hopefully provide some clues for these questions. Acknowledgments. We thank R. Wolf for very helpful comments. The work at Utah State University was supported by the Aeronomy Program, Division of Atmospheric Sciences, of the National Science Foundation, through grant ATM , and through DOD Air Force grant F C The research at Cornell University was sponsored by the National Science Foundation through grant ATM The Jicamarca Observatory is operated by the Instituto Geofisico del Peru with support from the National Science Foundation. The Arecibo radar is operated by Cornell University under contract with the National Science Foundation. The Saint-Santin radar was operated with financial support from the Centre National de la Recherche Scientifique and the Centre National d' Etude des T l communications. The Millstone Hill and Sondre Stromfjord radars are supported by the National Science Foundation. The EISCAT Scientific Association is supported by Centre National de la Recherche Scientifique of France, Suomen Akatemia of Finland, Max-Planck-Gesellschaft of Germany, Norges Almenvitenskapelige Forskningsr id of Norway, Naturvetenskapliga Forskningsr det of Sweden, and the Science and Engineering Research Council of the United Kingdom. The editor thanks D. R. Moorcroft and A.D. Richmond for their assistance in evaluating this paper. REFERENCES Alcayd6, D., G. Caudal, and J. Fontanari, Convection electric fields and electrostatic potential over 61 ø < A < 72 ø invariant latitude observed with the European Incoherent Scatter facility, 1, Initial results, J. Geophys. Res., 91, 233, Behnke, R. A., and R. M. Harper, Vector measurements of F region ion transport at Arecibo, J. Geophys. Res., 78, 8222, Bittencourt, J. A., and M. A. Abdu, A theoretical comparison between apparent and real ionization drift velocities in the equatorial F region, J. Geophys. Res., 86, 2451, Blanc, M., Midlatitude convection effects and their relation to ring current development,geophys. Res.Lett., 5,203,1978. Blanc, M., Magnetosphericonvection effects at midlatitudes, 1, Saint-Santin observations, J. Geophys. Res., 88, 211, Blanc, M., and P. Amayenc, Seasonal variations of the ionospheric E x B drifts above Saint-Santin on quiet days, J. Geophys. Res., 84, 2691, Blanc, M., and A.D. Richmond, The ionospheric disturbance dynamo, J. Geophys. Res., 85, 1669, Blanc, M., D. P. Thomas, and P. J. S. Williams, On the latitude variation of the magnetospheric electric fields at midlatitudes, J. Atmos. Terr. Phys., 42, 407, de la Beaujardiere, O., GISMOS program and workshop, Eos Trans. AGU, 68, 483, de la Beaujardiere, O., D. S. Evans, and Y. Kamide, The transition from quiet to active conditions: GISMOS ionospheric observations in the early afternoon sector, in Quantitative Modeling of Magnetosphere - Ionosphere Coupling Processes, edited by Y. Kamide and R. A. Wolf, p. 77, Kyoto Sangyo University, Kyoto, 1987a. de la Beaujardiere, O., D. S. Evans, Y. Kamide, and R. P. Lepping, Response of auroral oval precipitation and magnetosphericonvection to changes in the interplanetary magnetic field, Ann. Geophys., 5, 519, 1987b. de la Beaujardiere, O., D. S. Evans, Y. Kamide, and R. Lepping, Electric fields and particle precipitation during the substorm of January 18, 1984, in Modeling Magnetospheric Plasma, Geophys. Monogr. Ser., vol. 44, edited by T. E. Moore and J. H. Waite, p. 165, A G U, Washington, D.C., Earle, G. D., and M. C. Kelley, Spectral studies of the sources of ionospheric electric fields, J. Geophys. Res., 92, 213, Fejer, B. G., The equatorial ionospheric electric fields. A review, J. Atmos. Terr. Phys., 43, 377, Fejer, B. G., Equatorial ionospheric electric field associated with magnetospheric disturbances, in Solar-Wind Magnetosphere Coupling, edited by Y. Kamide and J. A. Slavin, p. 519, Terra Scientific, Tokyo, Fejer, B. G., C. A. Gonzales, D. T. Farley, M. C. Kelley, and R. F. Woodman, Equatorial electric field during magnetically disturbed periods, 1, The effect of the interplanetary magnetic field, J. Geophys. Res., 84, 5792, Fejer, B. G., M. F. Larsen, and D. T. Farley, Equatorial disturbance dynamo electric fields, Geophys. Res. Lett., 10, 537, Fejer, B. G., E. Kudeki, and D. T. Farley, Equatorial region zonal plasma drifts, J. Geophys. Res., 90, 12249, Ganguly, S., R. A. Behnke, and B. A. Emery, Average electric field behavior in the ionosphere above Arecibo, J. Geophys. Res., 92, 1199, Gonzales, C. A., M. C. Kelley, B. G. Fejer, J. F. Vickrey, and R. F. Woodman, Equatorial electric fields during magnetically disturbed periods, 2, Implications of simultaneous auroral and equatorial measurements, J. Geophys. Res., 84, 5803, Gonzales, C. A., M. C. Kelley, D. L. Carpenter, T. R. Miller, and R. H. Wand, Simultaneous measurements of ionospheric and magnetospheric electric fields in the outer plasmasphere, Geophys. Res. Lett., 7, 517, Gonzales, C. A., M. C. Kelley, R. A. Behnke, J. F. Vickrey, R. Wand, and J. Holt, On the latitudinal variations of the ionospheric electric field during magnetospheric disturbances, J. Geophys. Res., 88, 9135, Heelis, R. A., P. C. Kendall, R. J. Moffet, D. W. Windle, and R. H. Rishbeth, Electrical coupling of the E and F regions and its effects on F-region drifts and winds, Planet. Space Sci., 22, 743, Holt, J. M., R. H. Wand, and J. V. Evans, Millstone Hill measurements on 26 February, 1979 during the solar eclipse and formation of the midday F region trough, J. Atmos. Terr. Phys., 46, 251, Holt, J. M., J. V. Evans, W. L. Oliver, and R. H. Wand, Millstone Hill observations of ionosphericonvection, in Physics of Space Plasmas, pp , Scientific Publishers, Cambridge, Mass., Jaggi, R. K., and R. A. Wolf, Self-consistent calculation of the motion of a sheet of ions in the magnetosphere, J. Geophys. Res., 78, 2852, Kamide, Y., Recent issues in studies of magnetosphere ionosphere coupling, J. Geomagn. Geoelectr., 40, 131, 1988.

12 FEJER ET AL.: LOW- AND MID-LATITUDE ELECTRIC FIELDS 2377 Kamide, Y., and S. Matsushita, Penetration of high-latitude electric fields into low latitudes, J. Atrnos. Terr. Phys., 43, 411, Kelley, M. C., B. G. Fejer, and C. A. Gonzales, Anomalous low latitude electric fields associated with a northward interplanetary magnetic field, Geophys. Res. Lett., 6, 301, Kikuchi, T., T. Araki, H. Maeda, and K. Maekawa, Transmission of polar electric fields to the equator, Nature, 273, 650, Maeda, H., and K. Maekawa, A numerical study of polar ionosphericurrents, Planet. Space Sci., 21, 1287, Maekawa, K., Global distribution of electric fields and currents in the ionosphere produced by field aligned currents of magnetospheric origin, J. Atrnos. Terr. Phys., 42, 673, Mazaudier, C., M. Blanc, E. Nielsen, and M. Zi, Latitudinal profile of the magnetosphericonvection electric field at ionospheric altitudes from a chain of magnetic and radar data, J. Geophys. Res., 89, 375, Mazaudier, C., R. Bernard, and S. V. Venkateswaran, Saint- Sanfin radar observations of lower thermospheric storms, J. Geophys. Res., 90, 2885, Mazaudier, C., C. Senior, and E. Nielsen, Global convection electric field and current: Comparisons between model's predictions and data from STARE, Saint-Santin, and magnetometers, J. Geophys. Res., 92, 5991, Mozer, F. S., Electric field mapping in the ionosphere at the equatorial plane, Planet. Space Sci., 18, 259, Nisbet, J. S., M. J. Miller, and L. A. Carpenter, Currents and electric fields in the ionosphere due to field-aligned auroral currents, J. Geophys. Res., 83, 2647, Nishida, Geomagnetic DP-2 fluctuations and associated magnetospheric phenomena, J. Geophys. Res., 73, 1975, Nopper, R. W., and R. L. Carovillano, Polar equatorial coupling during magnetically active periods, Geophys. Res. Lett., 5, 699, Reddy, C. A., V. V. Somayajulu, and C. V. Devasia, Global scale electrodynamicoupling of the auroral and equatorial dynamo regions, J. Atmos. Terr. Phys., 41, 189, Reiff, P. H., and J. G. Luhmann, Solar wind control of the polar cap voltage, in Solar-Wind Magnetosphere Coupling, edited by Y. Kamide and J. A. Slavin, p. 453, Terra Scientific, Tokyo, Richmond, A.D., and Y. Kamide, Mapping electrodynamic features of the high-latitude ionosphere from localized observations: Technique, J. Geophys. Res., 93, 5741, Richmond, A.D., Y. Kamide, B. H. Ahn, S.-I. Akasofu, D. Alcayde, M. Blanc, O. de la Beaujardiere, J. C. Foster, E. Friis-Christensen, T. J. Fuller-Rowell, J. M. Holt, D. Knipp, H. W. Kroehl, R. P. Lepping, R. J. Pellinen, C. Senior, and A. N. Zaitsev, Mapping electrodynamic features of the high-latitude ionosphere from localized observations: Combined incoherent scatter radar and magnetometer measurements for January 18-19, 1984, J. Geophys. Res., 93, 5760, Rishbeth, H., Polarization fields produced by winds in the equatorial F-region, Planet. Space Sci., 19, 357, Senior, C., and M. Blanc, On the control of magnetospheric convection by the spatial distribution of ionospheric conductivities, J. Geophys. Res., 89, 261, Senior, C., and M. Blanc, Convection in the inner magnetosphere: Model predictions and data, Ann. Geophys., 5, 405, Siscoe, G. L., Energy coupling between regions 1 and 2 Birkeland current systems, J. Geophys. Res.,87,5124,1982. Somayajulu, V. V., C. V. Devasia, C. A. Reddy, and K. S. Viswanathan, Penetration of magnetosphericonvective electric field to the equatorial ionosphere during the substorm of March 22, 1979, Geophys. Res. Lett., 14, 876, Spiro, R. W., M. Harel, R. A. Wolf, and P. H. Reiff, Quantitative simulation of a magnetospheric substorm, 3, Plasmaspheric electric fields and evolution of the plasmapause, J. Geophys. Res., 86, 2261, Spiro, R. W., R. A. Wolf, and B. G. Fejer, Penetration of high latitude electric field effects to low latitudes during SUNDIAL 1984, Ann. Geophys., 6, 39, Takeda, M., and H. Maeda, Three-dimensional structure of ionosphericurrents produced by field-aligned currents, J. Atmos. Terr. Phys., 44, 694, Tsunomura, S., and T. Araki, Numerical analysis of equatorial enhancement of geomagnetic sudden commencement, Planet. Space Sci., 32, 599, Vasyliunas, V. M., Mathematical models of magnetospheric convection and its coupling to the ionosphere, in Particles and Fields in the Magnetosphere, edited by B. McCormac, p. 60, D. Reidel, Hingham, Mass., Volland, H., Models of global electric fields within the magnetosphere, Ann. Geophys., 31,154, Wand, R., and J. V. Evans, The penetration of convection electric fields to the latitude of Millstone Hill ( A= 56ø), J. Geophys. Res., 86, 5809, Wolf, R. A., Effects of ionospheric conductivity on convective flow of plasma in the magnetosphere, J. Geophys. Res., 75, 4677, Wolf, R. A., R. W. Spiro, and G. A. Mantjoukis, Theoretical comments on the nature of the plasmapause, Adv. Space Res., 6, 177, Woodman, R. F., Vertical velocities and east-west electric fields at the magnetic equator, J. Geophys. Res., 75, 6249, Woodman, R. F., East-west ionospheric drifts at the magnetic equator, Space Res., 12,969, M. A. Abdu and J. H. A. Sobral, INPE, Cp 515, S o Jos6 dos Campos, S o Paulo, Brazil. R. Burnside and C. A. Tepley, Arecibo Observatory, P.O. Box 995, Arecibo, PR O. de la Beaujardiere, Radio Physics Laboratory, SRI International, 333 Ravenswood Ave., Menlo Park, CA B. G. Fejer, CASS, Utah State University, Logan, UT J. A. Holt, Haystack Observatory, Massachusetts Institute of Technology, Westford, MA Y. Kamide, Faculty of Science, Kyoto Sangyo University, Kamigamo, Kita-ku, Kyoto 603, Japan. M. C. Kelley, School of Electrical Engineering, Cornell University, Ithaca, NY R. Lepping, NASA/Goddard Space Flight Center, Greenbelt, MD C. Senior, CNET/CRPE, 4, Avenue de Neptune, F St.-Maur des Fosses, CEDEX, France. R. F. Woodman, Instituto Geofisico del Peru, Apartado 374, Lima 100, Peru. (Received October 26, 1988; revised July 24, 1989; accepted July 26, 1989.)