Back to INAG Homepage Back to UAG-104 Contents page
This page was created by Phil Wilkinson


Shun-rong Zhang Xin-yu Huang Yuan-zhi Su

Wuhan Institute of Physics, Chinese Academy of Sciences,

P.O.Box 72003, Wuhan 430072, P. R. CHINA


With the help of servo model, plasma drifts near F2 peak derived from ionosonde data for 1986, 1987 (low solar activity) and 1982 (high solar activity) are obtained over 4 East Asian stations. The general pattern of the drift is "W"-shaped for the diurnal curve, and superimposed with collapse at about 3LT, which is extremely obvious for March 1982. Possibly, the ionospheric F2 layer dynamo is responsible for the collapse of the drift. A case study of the vertical drift spectra is also carried out. The results show obvious 2.1-day oscillation over the low latitude station, Okinawa, and it seems to provide evidence for the modulation of the electric field by planetary waves, which may eventually modify the oscillation of the ionospheric characteristics.


Current investigations indicate that existing ionosonde data obtained with relatively low costs and good time and geographical coverage can be used efficiently to estimate the global features of the thermospheric wind within the ionospheric F2 region. This research was originated from the study of wind effect on the ionospheric F2 peak. According to theory, for example, servo model (Rishbeth, 1967; Rishbeth et al, 1978), a vertical plasma drift approximately modifies the balance height, established by the balance between the photochemical loss and diffusion. Two major approaches to deriving neutral wind were developed by Buonsanto (1986), Buonsanto et al (1989) and Miller et al (1986) respectively. The method of Buonsanto et al uses the complete formulae suggested by Rishbeth et al (1978). In this approach, besides ionospheric data, namely foF2 and hmF2, the neutral density and temperature and also O+ - O collision frequency are needed so that appropriate balance heights can be obtained. The other approach expresses the changes in the F2 layer peak in terms of a linear wind-induced drift, and employs a theoretical ionospheric model, from which the linear coefficient can be obtained at various neutral winds by a linear regression. Following this approach, several empirical formulae for the linear coefficients have been established for the equatorial, mid-latitude and high latitude regions (Forbes et al, 1992, and the references therein). In this paper, the Buonsanto et al method is used.

Comparing mid-latitudes, which is the case Buonsanto and other researchers have focused on (Buonsanto et al, 1989; Buonsanto, 1990), the lower mid-latitude is a much more complicated case, where E x B drift becomes more important for the vertical motion of ionization while wind-induced drift still plays a significant role. It is impossible, of course, to discriminate between each component, if there is no further information available. However, by using data from several stations, with different latitude along approximately the same longitude, the relative importance may be determined. Accordingly, we will be able to gain insight into some ionospheric phenomena, as will be discussed in this paper, for the chain of 4 East Asia ionosonde stations. The derived vertical drifts will be given as a function of month, season and solar activity over these locations.

Ionospheric oscillations characterised by the fluctuations in foF2 and hmF2 is another interesting subject. The few observations available for studying plasma drift, seem insufficient to reveal the mechanism of the ionospheric oscillation. The great advantage of derived vertical drifts from ionosonde data is that it makes it possible to obtain the variation for any time scale depending only upon the availability of the ionospheric data. We thus display a case study of the spectral structure obtained with an FFT for the vertical drift in a full month (December 1981) for a mid-latitude and a low latitude station, with emphasis on the tidal and 2-day wave components. This is different from Buonsanto's work (1991), where a Fourier decomposition of the mean daily variations for each month concerned at mid-latitudes is carried out.

Table 1 Locations of Ionosonde Stations



Geographic Latitude.

Geographic Longitude.

Geomagnetic Latitude



45.4 N

141.7 E

35.3 N



31.2 N

130.6 E

20.4 N



30.5 N

114.4 E

19.2 N



26.3 N

127.8 E

15.3 N


Method and Data

The basic equation is the so-called "servo equation" (Rishbeth et al, 1978), and the related parameters are from the paper of Buonsanto et al (1989), except that the factor of O+ - O collision frequency is assumed to be 1 instead of 1.7. This technique suffers from the shortcoming of assuming an equilibrium situation in the ionosphere for the period between sunrise and pre-noon, which might be unrealistic, as has been recently pointed out, Titheridge (1993). According to Titheridge's model study the servo balance height differs from the calculated one with more complete processes included, implying the occurrence of a non-equilibrium ionosphere. However, further work should be done to estimate the extent to which this off-equilibrium behaviour may cause significant errors in the derived vertical drift. The position of the stations is listed in table 1.

Morphology of Monthly Mean Drifts

Month-to-month and seasonal features for low solar activity

Fig 1 shows the local time vertical drifts for a full year over Wuchang. "W" shaped or its variant curves can be found. Generally, double peaks, a pre-noon one and an afternoon one occur at 8-10h and 15-18h, while 4 types can be identified and each of them corresponds to a different season. In summer, only a pre-noon peak occurs, and in autumn an additional afternoon one occurs with small amplitude. The fluctuations dominate the winter curves. For spring, double peaks with nearly the same amplitude are clear. For the other 3 Japanese stations, similar variations can be found. These results are approximately identical to the MU radar observations of the meridional wind in trend (Oliver et al, 1990). It should be noted that the evening enhancement of the vertical drift, which is frequently observed near the low latitudes and equatorial areas (Goel et al, 1990, and the references therein), is not obvious here, even at Okinawa, whose geomagnetic latitude is 15.3; instead, the collapse is at dawn (about 3LT or later) followed by the increasing at 5-6LT of vertical drifts.

Latitudinal variations of abnormal drifts

Yamagawa and Okinawa have similar longitude but differ by 5 degrees in latitude, while Yamagawa and Wuchang have similar latitude but differ by 14 degrees in longitude. From Fig. 2(a) and (b), the variation for the stations with similar longitude, Yamagawa and Okinawa, is similar, and differs from Wuchang, suggesting a longitudinal dependence of the vertical drift. Since these 3 stations are located at the northern crest of the equatorial ionization anomaly (EIA), the longitudinal effects on the EIA may have to be considered, as Wakkanai is another pattern not only in amplitude but in phase, and this may be due to the obvious gap of the geographical position. It should also be noted that there exists a large downward drift at about 09h for the stations with similar longitude (Fig 2(a), (b)). This may be partly explained by the non-equilibrium in the ionosphere (Titheridge, 1993). The result for Wuhang is somewhat different at this time, suggesting a possible longitudinal difference for the ionospheric equilibrium condition.

The collapse is more noticeable in December and March under low solar activity, as shown in Fig. 2(b) and (c). The abnormal drift is much more obvious for high solar activity, as shown in Fig.3. When the collapse can be observed in all months concerned, of which March is particularly dominant, the maximum of downward drifts appears at 2LT for the first half of the year and 3LT for the other half. This behaviour becomes more pronounced, with decreasing latitude, and may an E x B contribution. These characters coincide with the evening increase in vertical drifts over some low latitudes and equatorial areas (Goel et al, 1990 and the references therein). It seems reasonable to assume a similar mechanism for the morning "collapse" as for the evening "enhancement", providing the sufficient loads (resistances) between E and F region along the line of force and also between two sides of the terminator occur under appropriate geophysical conditions. In this context, the polarized electric field may contribute to the collapse at dawn , which is driven by the east-west neutral wind and coupled with the ionospheric E region along the magnetic force lines, and is the widely accepted interpretation of the evening enhancement (Rishbeth, 1981).

Spectra of Vertical Drifts: A Case Study

A case study is carried out so as to gain an insight into the spectral structure for the vertical plasma drifts. Two sets of data are selected in the period of December 1-31,1982. For this month, the median of sunspot number R is 138, Ap index is 10, Ottawa 10.7cm solar emission flux is 201 s.f.u., and the daily Ap and F107 were obtained from the Solar Geophysical Data. The sampling interval is 1 point per hour, and the width of the Hamming window is 1024 hours.

foF2 and hmF2 Spectra

For the foF2 spectra at the 2 stations (not shown here), 24, 12, 8 and 6h components can be found with the strongest peak at 24h. The amplitude decreases as the period becomes short for Wakkanai, but it remains nearly unchanged for Okinawa. As for the hmF2 spectra (not shown here), it is much more abundant at Okinawa than at Wakkanai. Besides the tidal components, a weak peak at about 48h can be seen for Okinawa. The above results are summarised as table 2.

Spectra for Vertical Drifts

We pay special attention to the drift spectrum. At Wakkanai, as shown in Fig.4 (also see table 2), peaks at 6h and 12h with similar amplitude are stronger than at 8h; and the 24h component is the strongest. Components with periods greater than 24h are not obvious. At Okinawa, the strongest peak is at 8h, and the peak at 6h is also very strong; diurnal and semi-diurnal components are nearly the same in amplitude. Most important of all, a 52h (2.1 days) component can be seen. Deduced drifts contain the contribution of wind-and E x B induced drifts, however, firstly, the mid-latitude station Wakkanai does not show a 2-day component, instead, the low latitude Okinawa does; secondly, current observations in the thermosphere do not reveal the similar neutral wind oscillation with a planetary wave period of 2.1 days, therefore, it seems reasonable to ascribe the 2-d oscillation of the vertical drift to the contribution of E x B.



Table 2 Periods and Normalised Amplitude


Period (hours)
















































Note: normalised amplitude = amplitude / Max. amplitude

As it is well known that there exists in the middle atmosphere a west-ward planetary wave frequently observed in summer and with the period of 2.1d. Chen (1992) suggested that the electric field modulated by this 2-day wave in the ionospheric dynamo region may couple into the F2 layer along the force line, and may produce the 2-day oscillation of the ionospheric characteristics through E x B drift to lift the F2 layer plasma and diffusion along the force line above the equatorial anomaly regions. If the electric field modulated by planetary waves is true, then the E x B drift, with 2-day period, should be valid on a global scale, however the ionospheric movement in the vertical direction induced by this drift may only be pronounced at the region where the dip is small, corresponding to low latitudes and equatorial areas. Our results indicate that the modulated field is likely to be present even in winter, and the related drift may act as a local source to produce the 52h component in the spectrum for the low latitude station Okinawa.


From the above discussions, we conclude that, (1) the general pattern of vertical drift induced by wind and E x B is "W"-shaped in the ionospheric F2 region over these 4 East Asia stations; and the collapse at about 3LT on the diurnal drift curve is obvious, and it is extremely marked over Okinawa for March under high solar activity conditions. This phenomenon may be caused by the polarized electric field within the F2 region; (2) a case study for the spectral structure of vertical drifts shows that the 2.1-day component, besides the usual tidal component, is very clear over Okinawa, and it supports the assumption that a modulated electric field by planetary wave may actually emerge and result in the 2.1-day oscillation of the ionospheric characteristics.


Thanks are due to Japanese Radio Research Laboratory, Ministry of Posts and Telecommunications for providing the Japanese ionosonde data. The NSSDC of Goddard Space Flight Center is greatly acknowledged for furnishing us with MSIS86, HWM90 and IRI90 models.


Banks P. M. and G. Kockarts 1973 Aeronomy, Academic Press, New York

Buonsanto, M.J. 1986 .J. Atmos. Terr. Phys, 48, 365-373

Buonsanto M. J. 1990 J. Atmos. Terr. Phys., 52 223-240

Buonsanto M. J. 1991 J. Geophys. Res., 96, 3711-3724

Buonsanto M. J., J.E. Salah, K. L. Miller 1989 J. Geophys. Res., 94, 987-997

Chen P. R. 1992 J. Geophys. Res., 97, 6343-6357

Farley D. T., E. Bonelli, B. G. Fejer 1986 J. Geophys. Res., 91,13723-13728

Forbes J. M., D. N. Anderson, I. Batista 1992 Adv. Space Res., 6, (6)293-(6)301

Goel M. K., S. S. Singh and B. C. N. Rao 1990 J. Geophys. Res., 95, 6237-6246

Miller K. L. and P. G. Richards 1986 J. Geophys. Res., 91,4531-4535

Rishbeth H., S. Ganguly and J. C. G. Walker 1978 J. Atmos. Terr. Phys., 40, 767-784

Rishbeth H. 1981 J. Atmos. Terr. Phys., 43, 387

Rishbeth H. 1967 J. Atmos. Terr. Phys., 29, 225-238

Oliver, W. L., S. Fukao, T. Takam et al 1990 J. Geophys. Res. 95, 7683-7692

Titheridge J. E. , 1993 J. Atmos. Terr. Phys, 48, 365-373.