4. Discussion

[15]  Let us consider the revealed morphological features from the point of view of their possible origin. Unlike middle latitudes where Q disturbances seem to be due to different processes depending on conditions (MDL), the F2 -layer variations at the geomagnetic equator can be mainly related to vertical EB plasma drifts [Leschinskaya and Mikhailov, 1984; Mikhailov et al., 1996], and this strongly simplifies the analysis. It should be noted, however, that some morphological results may be due to the method of the Q disturbances extraction used. For instance, contrary to the midlatitude case it was found that negative Q disturbances were more numerous compared to positive ones (Figure 3). A 27-day running median used in our analysis bears the effects of positive F2 -layer disturbances which dominate in the equatorial F2 region [Adeniyi, 1986]. Therefore any normal quiet day is perceived as a negative Q disturbance, while the F2 layer should be strongly modified to mark a given day as a positive Q disturbance event. Thus the situation is not symmetric with respect to the two types of disturbances. We have tried to avoid this effect dealing only with large ( d N_mF2 > 40% ) %%% deviations, but such asymmetry however takes place in our results and it can be at least partly attributed to the method used. On the other hand, morphological differences between positive and negative Q disturbances really exist. It appears in diurnal and seasonal variations of their occurrence (Figures 4 and 5) as well as in their amplitudes (Figure 6) and this may be related to a great extent to EB vertical drift variations.

[16]  During nighttime hours when vertical drift V_z is downward [Fejer et al., 1991; Scherliess and Fejer, 1999; Woodman, 1970], Q disturbances are due to recombination of O⁺ ions presenting the main ion in the F2 region. The stronger downward V_z the lower h_mF2, and so the higher recombination rate which increases exponentially being proportional to linear loss coefficient b = g₁ [ N₂] + g₂ [ O₂]. This effect is clearly seen in Figure 1 where large h_mF2 deviations from the median result in large negative and positive N_mF2 deviations. It should be stressed that the effect is totally due to the vertical EB plasma drift rather than thermospheric winds as the magnetic inclination I=0 at the geomagnetic equator. In principle, any noticable variations of neutral composition are not expected in the equatorial thermosphere under quiet conditions, but this question needs a special analysis for the periods of Q disturbances. Therefore here all nighttime morphological results will be considered in the framework of the vertical EB drift/recombination mechanism.

[17]  Figures 4 and 6 show that negative Q disturbances at solar minimum are observed for the whole night, while they cover only the postmidnight period at solar maximum. The maximal disturbance amplitudes are also larger at solar minimum. These differences can be related to the peculiarities of the observed vertical drifts. According to Fejer et al. [1991, Figure 1], vertical drifts during the dusk (1600-2000 LT) period at solar minimum are systematically lower compared to those at solar maximum being around zero or even negative. Another peculiarity of the drifts under solar minimum compared to solar maximum is their large day-to-day variability [Woodman et al., 1977] especially during nighttime hours [Fejer et al., 1979]. Therefore large negative drifts are very probable in the night at solar minimum, and this can contribute to the observed occurrence and amplitude distributions for negative disturbances under solar minimum (Figures 4a and 6a). Negative disturbances at solar maximum are seen to cluster in the postmidnight LT sector only and they are less in number compared to solar minimum. This may be related to the following reasons. The earlier discussed recombination effect depends on the linear loss coefficient b variations when F2 layer is shifted by vertical drifts. The thermospheric temperature is much higher at solar maximum compared to solar minimum. For instance, according to the thermospheric model MSIS 86 [Hedin, 1987] at the location of Huancayo and 0000 LT T _ex =1060 K under solar maximum ( F_10.7 =200 ) and T _ex =676 K at solar minimum ( F_10.7 =70 ). This temperature difference results in corresponding neutral scale height ( H=kT/mg ) difference for [N₂] and [O₂] height distributions. Therefore the same F2 -layer shift in altitude due to vertical drift gives larger changes in b (and in N_mF2 ) at solar minimum when neutral scale height is less. Similarly, according to MSIS 86 at high solar activity T _ex decreases from 1264 K (1900 LT) to 985 K (0600 LT). Therefore the early morning hours are more preferable for negative disturbances to appear. Moreover, average negative drifts are large in the postmidnight LT sector at solar maximum [Fejer et al., 1991], and this implies that individual (for a particular day) V_z values may be even larger, thus increasing the probability for a negative disturbance to occur.

[18]  Another peculiarity of the vertical drift variations, the evening prereversal upsurge of V_z, which is large at solar maximum and is practically absent at solar minimum [e.g., Fejer et al., 1991], may also contribute to the discussed solar activity differences in the disturbance occurrence. Because of strong upward V_z at solar maximum more plasma is uplifted and stored in the topside ionosphere (in the magnetic tubes of force). Any nighttime increase of the downward V_z (after the time of V_z reversal) is accompanied by plasma influx to the equatorial F2 region from above, which compensates to some extent the recombination losses. The low-latitude plasma tube content is not large [Carpenter and Park, 1973], and this process may be efficient only for some evening hours after the V_z reversal. When by the second part of the night the electron content of the magnetic tubes is depleted, the loss process begins to dominate in nighttime F2 region. This is another reason for negative disturbances to appear only in the early morning LT sector during solar maximum (Figures 4 and 6). At solar minimum, on one hand, the background magnetic tube electron content is relatively small, on the other hand, the prereversal plasma uplift is practically absent [Fejer et al., 1991]. Therefore the plasma influx to the evening F2 region due to downward V_z is not sufficient to cover the recombination losses and large negative disturbances are seen to occur starting from the evening hours (Figures 4 and 6).

[19]  Positive nighttime Q disturbances correspond to low downward V_z and are due to low recombination losses as F2 layer is located high enough in this case (see Figure 1d). Being determined with respect to the running N_mF2 median, the maximal positive deviations are read in the early morning (presunrise) hours when the median N_mF2 values are the lowest (see Figure 1). This explains the occurrence of positive disturbances in the postmidnight LT sector (Figure 6). Also, in the end, relatively small total number of negative and positive disturbances revealed at solar maximum may be also related to higher stability of vertical drift variations observed under solar maximum conditions [Woodman et al., 1977].

[20]  Daytime negative and positive Q disturbances are not numerous (Figures 4 and 6 and Tables 2 and 3), however, the daytime peak in the occurrence of positive disturbances is well pronounced both at Huancayo and Kodaikanal (Figure 4), the amplitude of both type of disturbances being practically independent of solar activity level (Figure 6). The formation of the equatorial daytime F2 region is also strongly controlled by vertical EB plasma drifts which normally are upward with a velocity of about 20 m s^-1 [Fejer et al., 1991; Scherliess and Fejer, 1999]. The main mechanism of N_mF2 changes is the plasma uplift from the F2 -region heights with its subsequent transfer along magnetic lines apart from the geomagnetic equator, so-called fountain effect [Hanson and Moffett, 1966]. Therefore an increase of V_z (eastward electric field) results in an N_mF2 decrease, while a decrease of V_z helps plasma to store at the F2 -region heights thus increasing N_mF2. Because of this upward plasma drift the F2 layer is lifted from the region of strong recombination and its effects are not that crucial as in the nighttime F2 region. However, in case of strong westward electric field (strong downward plasma drift) the daytime F2 layer may be moved downward to the strong recombination area with a subsequent N_mF2 negative disturbance effect [Mikhailov and Leschinskaya, 1991]. However, normally daytime EB drifts vary staying positive and N_mF2 changes are controlled by the efficiency of plasma removal from the F2 region due to this upward drift.

[21]  Therefore daytime positive Q disturbances (Figure 4) result from V_z decrease followed by h_mF2 lowering clearly seen in Figure 2 on 16 February 1966. The effect may be related to a well-known equatorial phenomenon, counter electrojet [Mayaud, 1977; Rastogi et al., 1992; Vyas et al., 1979] when normal zonal E_y electric field is decreased or even reversed. An example of strong counter electrojet observed in the Indian sector on a quiet ( Ap=3 ) day 21 July 1976 is given by Rastogi et al. [1992]. A decrease in E_y is clearly manifested in D H diurnal variations as well as in the F2 -layer parameter changes. As a result of V_z decrease, f_oF2 at Kodaikanal is elevated with respect to a control (also quiet with Ap=4 ) day 22 July 1976, h'F is decreased, and TEC in the crest zone (Ahmedabad) is also decreased, the latter is due to a weakening of the fountain effect. Some cases of V_z direct observations at Jicamarca for the days of counter electrojet are given by Woodman et al. [1977]. In four presented cases the observed daytime V_z becomes negative for some hours.

[22]  The other morphological feature of daytime Q disturbances: the independence of the amplitude on solar activity (Figure 6) may be also related with the daytime vertical drifts which are essentially independent of solar activity [Fejer et al., 1991, 1995; Scherliess and Fejer, 1999]. A comparison of Q disturbance occurrences in the American and Indian sectors (Figure 4) shows general similarity in diurnal variations, but the variations are shifted to earlier hours at Kodaikanal. Besides, the nighttime peaks in the occurrence are broader at Kodaikanal than at Huancayo, practically covering all dark hours. In accordance with the earlier discussed mechanisms these differences should reflect the longitudinal differences in the vertical drift variations. Although the global equatorial F region vertical drift model by Scherliess and Fejer [1999] demonstrates longitudinal variations in the vertical drifts, the differences between Indian and American sectors are not that large to explain the revealed differences in the Q disturbances occurrence. Relatively large differences between vertical drift variations at Trivandrum and Jicamarca are reported by Hari and Krishna Murthy [1995] but for high solar activity only when Q disturbances are rarely observed. Therefore the problem of longitudinal differences in the Q disturbance occurrence needs a special analysis.

[23]  Annual variations of the nighttime negative and positive Q disturbance occurrence at Huancayo exhibit well-pronounced winter peaks (Figure 5). This annual variation pattern may be also explained by vertical drift seasonal variations. Jicamarca observations [Fejer et al., 1991, Figure 1] show that winter prereversal upward vertical drifts are the smallest in the year and the prereversal upsurge V_z is completely absent at solar minimum. In accordance with the earlier discussed mechanism one may expect a very modest plasma influx from above to the nighttime F2 region after the V_z reversal. This plasma influx seems to be insufficient to cover the recombination losses when F2 layer is shifted to lower heights by an increased negative V_z. Large number of negative disturbances at solar minimum (Figure 6) seems to confirm this explanation. The other factor discussed earlier (winter/summer neutral temperature difference) is small ( 20 K) and cannot much contribute to the seasonal difference in the disturbance occurrence.

[24]  The seasonal peculiarities in the vertical drift variations should reflect in the N_mF2 median values as well. Indeed, the International Telecommunications Union [1997] global empirical monthly median model gives the lowest N_mF2 early morning values in winter (June-July). This model prediction was checked using real monthly medians observed at Huancayo for some selected years when R₁₂ index was relatively stable during the whole year. The model annual N_mF2 variations were confirmed with one exception; there is an additional minimum in January taking place for some years. This allows us to explain annual variations for the nighttime positive Q disturbance occurrence (Figure 5c). As it was mentioned earlier positive deviations are read with respect to the running median N_mF2 values which are the lowest in winter. The results of our analysis also give some percent of events in January (Figure 5), which obviously are related to the additional January minimum in median N_mF2 values.

[25]  The situation with the annual variations of Q disturbances in the Indian sector is less obvious (Figure 5a, 5c, and 5e) and seasonal variations are not pronounced. It looks as if there are seasonal and nonseasonal components and their contributions vary with the station location. Mayaud [1977] has drawn to such a conclusion analyzing the effect of counter electrojet at Huancayo and Kodaikanal. As it was mentioned earlier daytime positive Q disturbances may be related to counter electrojet when zonal electric field is decreased or even reversed. The daytime (afternoon) counter electrojet occurrence exhibits a well-pronounced summer peak at Huancayo, while at Kodaikanal the occurrence distribution is broad covering all seasons [Mayaud, 1977, Figure 8]. However, Vyas et al. [1979] give the maximal daytime counter electrojet occurrence in summer for the Indian sector as well. Mayaud suggests that there exists two sources for the counter electrojet events. A source centered on the month of January at any longitude and strongly modulated by the Moon and a source centered on the solstice in each hemisphere. The two sources overlap at Huancayo and we have a well-pronounced summer peak for the daytime positive Q disturbance occurrence (Figure 5e).