Letter to the editor Electric field fluctuations (25–35 min) in the midnight dip equatorial ionosphere

Measurements with a HF Doppler sounder at Kodaikanal (10.2°N, 77.5°E, geomagnetic latitude 0.8°N) showed conspicuous quasi-periodic fluctuations (period 25/35 min) in F region vertical plasma drift, Vz in the interval 0047/0210 IST on the night of 23/24 December, 1991 (Ap = 14, Kp < 4−). The fluctuations in F region vertical drift are found to be coherent with variations in Bz (north-south) component of interplanetary magnetic field (IMF), in geomagnetic H/X components at high-mid latitude locations both in the sunlit and dark hemispheres and near the dayside dip equator, suggestive of DP2 origin. But the polarity of the electric field fluctuations at the midnight dip equator (eastward) is the same as the dayside equator inferred from magnetic variations, contrary to what is expected of equatorial DP2. The origin of the coherent occurrence of equatorial electric field fluctuations in the DP2 range of the same sign in the day and night hemispheres is unclear and merits further investigations.


Introduction
DP2 is a global ionosphere equivalent current system proposed to account for quasi-periodic (QP)¯uctuations in the geomagnetic ®eld that occur coherently at high latitudes and dayside dip equator and in the B z (north-south) component of interplanetary magnetic ®eld, IMF (Nishida et al., 1966;Nishida, 1968a, b). The twin-vortex DP2 current system is thought to be driven by the IMF-controlled magnetospheric convection electric ®eld. The recent case study of Kikuchi et al. (1996) using EISCAT radar observations showed that DP2 magnetic¯uctuations at auroral latitudes are due to ionosphere Hall current caused by the convection electric ®eld. Their work further suggests that the magnetospheric electric ®eld penetrates almost instantaneously through the polar ionosphere to the equatorial ionosphere and drives the current system responsible for equatorial DP2. The noteworthy enhancement of DP2 amplitude at dayside dip equator compared to low latitudes is believed to be due to the Pedersen current ampli®ed by the Cowling eect.
Most of the earlier studies of equatorial DP2 are limited to the dayside and there is an acute paucity of information on DP2 at the evening and nightside dip equator. Detailed evaluation of the local time dependence of the characteristics of equatorial DP2 will help advance our understanding of the substorm phenomenon as a whole (Rostoker, 1993) and of related equatorial ionospheric disturbances (see, Fejer, 1997 and references therein). We have very recently found evidence for DP2 electric ®eld¯uctuations (period 25 min) at the duskside dip equator which supports the view that the magnetospheric electric ®eld responsible for DP2 penetrates to equatorial ionosphere on the duskside as on the dayside and leads to electric ®eld perturbations of the same polarity (eastward) as on the dayside (Abdu et al., 1998). We present and discuss the characteristics of an event of electric ®eld variations in the DP2 period range (25±35 min) evidenced in the midnight dip equatorial ionosphere.

Results and discussion
A HF Doppler sounder is regularly operated on 4 MHz at Kodaikanal, India (10.2°N, 77.5°E, geomagnetic latitude 0.8°N) to monitor the Doppler velocity, V d of F region re¯ections at normal incidence in the eveningnight hours. V d represents the F region vertical plasma drift, V z (=V d /2) due to zonal electric ®elds at this location over the local time period mentioned. The details of the experimental set-up were published elsewhere (Sastri et al., 1985). The HF Doppler data of Kodaikanal have proved useful, in addition to other investigations, in studies of equatorial electric ®elds associated with transient geophysical phenomena like storm sudden commencements (ssc), sudden impulses (si) and substorms (Sastri et al., 1993(Sastri et al., , 1997. Doppler measurements made on the night of 23/24 December 1991 (A p = 14, K p < 4 ) ) showed a distinct train of quasi-periodic variations in V z just after local midnight as can be seen from raw V z data (one minute resolution) presented in Fig. 1. The presence of three distinct cycles of QP¯uctuations (labelled 1 to 3, period 25±35 min) over the interval 0047±0210 IST (1917± 2040 UT) is obvious from the ®gure. We have explored whether this disturbance in F region vertical plasma drift is a signature of DP2 activity by analysing Kodaikanal V z data in conjunction with geomagnetic ®eld variations elsewhere and with B z component of IMF. Figure 2 shows the temporal variation over the interval 1800±2116 UT of the B z component of IMF (IMP-8 satellite), F region V z at Kodaikanal and geomagnetic H/X component at Ancon, Peru (12.08°S, 77.02°W, geomagnetic latitude 1.47°), Fort Churchill (59°N, 94°W), Cambridge Bay (69°N, 105°W) and the IMAGE station, Masi (69.46°N, 23.7°E). Note that the data presented are 3-min running means of the various parameters at one minute resolution. This is done to suppress the short-period (<1 min)¯uctuations and highlight the longer period ones in the DP2 range. Further, a time shift of 17 min is introduced for IMF B z to allow for the delay time between B z¯u ctuations and their eects at ground level. IMP-8 satellite was a radial distance of 39.4 R e over the time period studied here. The delay of 17 min which is visually assessed to be the optimum is therefore reasonable and appropriate (see Nishida, 1968b for details of delay time between changes in IMF and ground level magnetic ®eld). Over the time interval under consideration, Cambridge Bay, Fort Churchill and Ancon were in the noon-afternoon sector, while Masi was in the dusk sector and Kodaikanal in the midnight sector. It is evident from Fig. 2 (Nishida, 1971). The unfortunate data gap at Ancon till 1946 UT precluded estimation of the amplitudes of the ®rst two peaks near the dayside dip equator. The magnitude of the third peak around 2025 UT is 2.6 nT which is reasonable keeping in view that it corresponds to a time of rapidly decreasing electrojet strength. It is to be noted here that the amplitude of DP2 magnetic variations in the dayside hemisphere (a) decreases rapidly with decrease in latitude but increases again near the dip equator and (b) varies with local time to be highest around noon compared to other times at the dip equator (Nishida, 1986b;Kikuchi et al., 1996).
The amplitude of the three peaks (1 to 3 in Fig. 2) in F region V z at Kodaikanal estimated by the same procedure is 2 m/s (0.08 mV/m), 7.4 m/s (0.3 mV/m) and 8.6 m/s (0.34 mV/m) in that order. This suggests a tendency for the amplitude to increase towards the early morning side. Note that the ambient vertical plasma drift of equatorial F region is generally downward (westward electric ®eld) in the nighttime. The virtual height of the bottomside F region over Kodaikanal to which the 4 MHz Doppler observations correspond to was in the range 265±285 km prior to and during the interval of QP¯uctuations (1800±2030 UT). At altitudes less than 300 km the chemical recombination induces an apparent upward drift, V c that makes the true downward plasma drift. This is the reason for the absence of an apparent downward drift over Kodaikanal prior and during the QP¯uctuations. Corrections for chemical loss eects are not made for V z data because they are basically a d.c eect and¯uctuations with periods <40 min of speci®c interest here are unlikely to be in¯uenced by the altitude-dependent changes in the chemical loss or in the ionization scale height (Sastri, 1995). The sense of the perturbation peaks in V z is to be positive (upward drift) in view of their close temporal association with IMF B z and magnetic variations elsewhere (see Fig. 2).
The coherence of the QP¯uctuations in V z at Kodaikanal with IMF B z and magnetic variations at the various stations is evaluated through correlation analysis. The correlation coecient between V z and IMF B z for the period 1915±2038 UT covering the three cycles is )0.671 and for the subperiod 1939±2038 UT covering the last two cycles (2 and 3) is )0.667. The corresponding values for X-comp at high latitude stations on the dayside, at Fort Churchill (Cambridge Bay) are 0.805 and 0.915 ()0.723 and )0.84) in that order. For Masi the values are )0.541 and )0.688. All the correlation coecients are statistically signi®cant con®rming the temporal coherence of the QP variations in F region V z at Kodaikanal with magnetic¯uctuations at high latitudes on the dayside as well as nightside. The correlation coecient between V z and H-component at Ancon is 0.43 (statistically signi®cant) for the period 1947±2040 UT for which data are available. The correlation coecient increased to 0.897 when the H-®eld is detrended for the diurnal variation. This excellent temporal association can clearly be seen from Fig. 3. But the most noteworthy feature in Figs. 2 and 3 is that the QP¯uctuations at the two dip equatorial stations bear an anti-phase relationship with IMF B z such that the southward turning of B z is associated with an increase in H-®eld (increase of electrojet strength) on the dayside (Ancon) and a predominantly upward F region plasma drift (eastward electric ®eld) on the nightside (Kodaikanal). The ®rst of the features is in agreement with the earlier work of Nishida (1986b) while the second one is the new fact brought to light by the current study.
It is known that DP2 activity generally develops during the growth phase of the substorm and persists even during the expansion phase (Clauer and Kamide, 1985). Careful scrutiny of the one-minute resolution data of the auroral electrojet index, AE (courtesy T. Iyemori) showed that auroral activity was low (AE < 200 nT) during the time interval of DP2 type QP¯uctuations. An increase in AE index indicative of substorm activity began only from 2105 UT which peaked (570 nT) at 2130 UT. The QP variations in F region vertical plasma drift at Kodaikanal as well as in geomagnetic ®eld at the various stations studied here are therefore unlikely to be eected by the longer period (2±3 h) disturbances associated with substorms.
DP2 type magnetic and electric ®eld¯uctuations are one of a class of disturbances that originate from globalscale current systems set up by sources in the polar regions, the other being the storm sudden commencements (ssc) and sudden impulses (si). The detailed case study of Kikuchi et al. (1996) demonstrated that the transient component of the electric ®eld set up by region 1 and 2 ®eld-aligned currents (FAC)¯owing in and out of the polar ionosphere is responsible for DP2 at auroral latitudes (see Fig. 9 of their paper). The coherent occurrence of magnetic¯uctuations all over the globe with a conspicious enhancement near the dayside dip equator that characterises DP2 activity implies that the usually prevalent shielding of the low latitude ionosphere from polar cap electric ®elds by region 2 FAC breaks down facilitating the low-latitude penetration of polar electric ®elds. Based on analysis of magnetometer data, Araki (1977) proposed that the simultaneous appearance of the preliminary impulse (PI) of storm sudden commencement (ssc) at the dayside dip equator and at high latitudes is due to the extension to the dayside equator of DP2 type ionospheric current system driven by the dusk-to-dawn electric ®eld imposed on the polar ionosphere. Theoretical studies showed that the Earthionosphere wave guide facilitates the instantaneous transmission of a suddenly imposed polar electric ®eld to the dip equator as a zeroth-order transverse magnetic (TM) electromagnetic wave (Kikuchi et al., 1978;Kikuchi and Araki, 1979). The subsequent numerical simulations of the two-dimensional current system by Tsunomura and Araki (1984) lent support to the wave guide model and provided an insight into the latitudinal pro®le and local time pattern of penetration electric ®elds and associated currents. Very recently, Tsunomura (1999) obtained a realistic solution of the polar-originating current system including the equatorial enhancement which showed that one of the non-diagonal terms of the conductivity tensor in the dip equatorial region signi®cantly in¯uences the equatorial electric ®elds. For the initial phase representing an enhancement in magnetospheric convection brought by a southward turning of IMF B z , the new simulations show the equatorial zonal electric ®eld to be eastward between 06 and 22 LT with a broad morning maximum between 08 and 09 LT and an evening maximum at 19 LT. The electric ®eld is westward at other times with a broad maximum in the presunrise hours (see Fig. 5 of their paper).
The polarity in H-®eld (increase of electrojet strength) of the QP¯uctuations at Ancon near the afternoon dip equator, their small amplitude and neagative correlation with IMF B z are in accordance with the model results of Tsunomura (1999) as well as of Senior and Blanc (1984) and Tsunomura and Araki (1984). All of these models consistently show that the penetration electric ®eld for the initial phase is eastward during daytime and changes in magnitude to attain low values in the afternoon around 16 LT. The sign (upward drift indicative of eastward electric ®eld) of the QP uctuations in F region vertical plasma drift at Kodaikanal near the postmidnight dip equator, is however, inconsistent with the model simulations all of which show a westward electric ®eld in the midnight-sunrise sector (see Fig. 5 of Tsunomura, 1999). In otherwords, the present observations show the simultaneous presence of QP¯uctuations of the same polarity in parameters representative of equatorial zonal electric ®elds in the afternoon and postmidnight sectors, instead of opposite polarity expected from the current understanding of equatorial DP2. The tendency for the magnitude of the QP¯uctuations in F region V z at Kodaikanal to increase towards the early morning side is nevertheless in agreement with the local time dependence of penetration electric ®elds in the midnight-sunrise period. In our knowledge, there is only one documented event of similar quasi-periodic electric ®eld¯uctuations of eastward polarity at the nightside (22-04 LT) dip equator. This event of February 17±18, 1976, studied by Gonzales et al. (1979) as well as Earle and Kelley (1987), is characterised by¯uctuations with signi®cant power at »1 h period in IMF B z as well as in electric ®elds at auroral and dip equatorial stations. But this event too does not seem to be of DP2 origin because the day-tonight polarity of the equatorial electric ®elds (eastward on nightside and westward on dayside) is at variance with the theoretical pattern of DP2 electric ®elds.
In conclusion, the present case study brought to light that quasi-periodic¯uctuations in the DP2 range (period 25±35 min) do occur in F region vertical plasma drift (zonal electric ®eld) near the postmidnight dip equator coherent with magnetic¯uctuations near the dayside dip equator and auroral/subauroral locations both in the sunlit and dark hemispheres. The electric ®eld¯uctuations do not, however, seem to be signatures of DP2 activity since they exhibit the same eastward polarity as those seen at the dayside equator, although they possess most of the other known characteristics of DP2. The origin of the distinct electric ®eld variations in the midnight dip equatorial ionosphere reported here is enigmatic and merits further studies.