G. A. Mikhailova, Yu. M. Mikhailov, and O. V. Kapustina
Institute of Terrestrial Magnetism, Ionosphere, and Radio Wave Propagation, Troitsk, Moscow Region, Russia
A typhoon as a meteorological phenomenon is a powerful revolving air whirlwind which is formed in the tropical zone mainly over the sea surface. Its diameter reaches ~500 km, and the revolving storm moves along the sea surface with a velocity of up to 15 km h -1. In the revolving center the so-called eye is formed, in which the air pressure is depleted, the sky over the eye being clear [Gidrometeoizdat, 1991]. The aerospace surveys showed [Sharkov, 1997] that a dense cloud structure up to several hundred kilometers wide and 12-15 km high is developed at the eye wall. This cloudiness is the source of strong thunderstorms and downpours, which, as a rule, accompany typhoons. The electromagnetic signals of the ULF-VLF range (8 Hz to 30 kHz) emitted (so-called whistlers) are able to propagate along a short distance into the topside ionosphere where they are registered on board satellites as partially dispersed whistlers (WH). Hence it follows that the intensity and spectral variations of these signals may serve as a basis of the electromagnetic method of typhoon studies in addition to the existing methods which use meteorological and geostationary satellites [Sharkov, 1997].
As far as we know, there is only one work showing information on observations of the ULF-VLF electromagnetic disturbances during the typhoon in the Caribbean region [Sobolev and Mikhailov, 1998]. The E component of the electromagnetic field was registered on board the Cosmos 1809 and Intercosmos 24 satellites in the broad frequency band (0.07-20 kHz) at h 600-1500 km. The information receiving center was situated in Havana ( j = 23.13 o N, l = 82.3 o W). The main result of this work is that the wideband discrete signals have the same character of frequency variation of the maximum intensity with time in the spectrograms as the partially dispersed WH generated by ordinary lightning discharges. The second result is that there is a considerable difference in the WH occurrence frequency from the background frequency (~10 min -1 in the daytime) and a similar difference between the daytime and nighttime values (5 and 60 s -1, respectively) during a typhoon. The strong difference between the daytime and nighttime occurrence frequencies shows that the WH source is in the Earth's atmosphere, because in the daytime the ionospheric D layer strongly absorbs the ULF-VLF electromagnetic waves. The high lightning activity accompanying typhoons provides the anomalously high WH succession frequency. Sobolev and Mikhailov  presented no quantitative evaluations of the intensity of the electric field component.
Apart from the Caribbean region, there are several more regions of typhoon observations over the World Ocean aquatory. In the Pacific Ocean, where, for example, in 1990, there occurred 59 typhoons out of 91 observed over all regions, is the most active of typhoons [Pokrovskaya et al., 1993].
The preliminary results of the study of spatial and spectral characteristics of the ULF-VLF electric field absolute values in the topside ionosphere over typhoons in the Pacific Ocean in the daytime in September 1990 (maximum of the typhoon season activity in this region [Pokrovskaya et al., 1993]) are presented below.
According to the Pokrovskaya et al.  catalogue, the typhoons over the Pacific Ocean were observed on September 3-9 and 10-20, 1990, with their maximum development stage in its northwestern part. The wind velocity in the vortex center on these days exceeded 33 m s -1.
The records on board the Intercosmos 24 satellite were used in this paper. The satellite had the orbit with a perigee of ~500 km, apogee of ~2500 km, and inclination of ~82.6o. The revolution period around Earth was T 115.8 min. Apart from the broadband receiver [Sobolev and Mikhailov, 1998] the narrowband filters at the frequencies shown in Table 1 were installed at the satellite. The filters covered the ELF ( f < 30 Hz), ULF (30 Hz < f < 3 kHz), and VLF (3 kHz < f < 30 kHz) frequency ranges. The reading time of the filters was 40.96 s in the ZAP4 regime and 5.12 s in the ZAP3 regime. Under partial dispersed WH duration of 30-150 ms at heights of h 500-1000 km [Kapustina et al., 1981] this reading frequency means that the noise background level along the satellite orbit and not an individual WH spectrum was measured similar to measurements of the atmospheric noise level created by the whistlers near the terrestrial surface. The detector of the electromagnetic field electric component, the data of which are considered below, was oriented along the satellite velocity vector. The channel calibration allowed us to measure field values at the satellite level. The minimum values in various channels were slightly different but were on the average ~2 m V m -1 . The maximum values were ~80-100 m V m -1. Table 2 shows the list of the satellite orbits (41) near the equator in the daytime. Thirty of them were over the Pacific Ocean water area ( l = 120o-280o E). Table 2 also shows the field values at the frequencies of two maxima in the WH spectrum ( f1 = 8 Hz and f2 = 225 Hz) and the corresponding parameters of the orbit. One can see in Table 2 that in September 1990 the satellite passed over the equator in the Pacific Ocean at height of ~1500-2500 km at 1200-1300 LT. Keep in mind that the typhoons in Pacific Ocean stormed on September 3-9 and 10-20.
Figure 1 shows a series of the flights on September 3 (Figure 1a) and 4 (Figure 1b), where the field intensity splashes are shown by bright color and the red line corresponds to the L shell at the height of 2400 km. It is typical that under the flight over South America and Africa (the centers of global thunderstorm activity) the signal level at the satellite was several m V m -1 at the 225 Hz frequency (for example, the orbits 4222-4223, 4231-4232, and 4309-4310, see Table 2). As soon as the satellite was shifted in space to Pacific Ocean longitude, the electric field intensity increased strongly (by a factor of more than 20) and remained high until the eastern shores of the Asia continent, where the signal level decreased sharply (the orbits 4227-4228 and 4313-4314). Similar results were observed at all satellite orbits over Pacific Ocean in active typhoon period (see Table 2). It also follows from Table 2 that in the active typhoon period the electric field intensity at the frequency of f2 = 225 Hz exceeded the upper limit of the receiver channel dynamic range ( 80 m V m -1 ), whereas in the quiet period (September 21-29) the signal intensity at this frequency was several m V m -1. In the active period the field intensity at the frequency of f1 = 8 Hz varied from one orbit to another within 15-45 m V m -1, exceeding the background value in the absence of typhoons by a factor of 3-9. On September 18 the signal intensity at f2 = 225 Hz increased as compared with the background but stayed within the receiving channel dynamic range. The second peculiarity of the phenomenon is that the intense emission band 5o-8o wide along latitude drifted along the L cover ( L 1.3-1.4 ) from the east to the west following the motion of the active typhoon region over the ocean.
Figure 2 shows two examples of electric field spectral distribution at l = 118o at the 4229-4230 orbit on September 3 (outside the Pacific Ocean aquatory) and at l = 235o at the 4224-4225 orbit (the narrow splashes at f = 9 and 15 kHz were caused by the service marks). For the ULF filters the scale of the field amplitude was 50 m V m -1, and at f = 9 and 15 kHz the scale was 12.5 m V m -1. For the 4229-4230 orbit the scale was the same (12.5 m V m -1 ) at all frequencies. Outside the Pacific Ocean aquatory in the daytime the signal level was maximum and equal to 12.5 m V m -1 at f = 225 Hz, was slightly lower at f = 150 Hz, and was almost absent (~2 m V m -1 ) at frequencies below 50 Hz. Such spectral distribution is typical for the partially dispersed WH observed in the daytime at middle latitudes [Kapustina et al., 1981]. The spectral distribution with intensity maximum at f = 225 Hz remained at the 4224-4225 orbit, but the maximum magnitude exceeded the upper limit of the receiver channel dynamic range (was above 80 m V m -1 ). Moreover, the spectral components increased at frequencies below 50 Hz and, particularly, at a frequency of 8 Hz. At 9 and 15 kHz the signal level was below the receiving equipment sensibility.
The spectral distribution of the electric field with maximum at f2= 225 Hz, and the absence of the emission at frequencies f = 9 and 15 kHz agrees qualitatively with the theoretical calculations of the attenuation function of the ULF-VLF wave whistler mode through the ionospheric D layer [Aksenov, 1966; Aksenov and Lishin, 1967]. At middle latitudes the attenuation coefficient (defined as the ratio of the energy fluxes of the passed and incident waves) is maximum at f = 200-300 Hz and equal to 0.15. In the daytime the attenuation coefficients decrease rapidly with geomagnetic latitude decrease [Aksenov and Nazarova, 1972], so no ELF electromagnetic emission has been observed in the topside ionosphere in the usual daytime conditions [Mikhailov et al., 1999]. Moreover, studying the emission distribution at f = 3.2 kHz in the topside ionosphere on board the Ariel 4 satellite at geographic latitudes 30o, Hayakawa  noted that the maximum intensity was observed over Africa ( l = 30o E) and Southeast Asia ( l =110o-140o ) at night. The emission intensity is very low between these regions particularly over Pacific and Atlantic Oceans. In the daytime the emission is absent at all latitudes and longitudes. The anomalous electric fields we observed in the ULF-VLF ranges in the daytime in the tropical zone contradict these experimental data and the theoretical ideas on the whistler mode properties of the ULF-VLF wave passage through the ionospheric D region. This result makes it possible to suppose that either these field sources (lightning discharges accompanying typhoons) are superpowerful ones as compared with the discharges even in the global centers of thunderstorm activity (South America, Africa, and Southeast Asia) or the lower ionosphere properties reducing the ULF-wave attenuation under their passage into the topside ionosphere are significantly changed over typhoons. The following facts favor the former suggestion. According to the data of meteorological observations, an increased thunderstorm activity is actually detected under tropical disturbances of the typhoon type [Gidrometeoizdat, 1991]. Moreover, the broadband observations of partially dispersed WH over the typhoon in the Caribbean Sea showed anomalously high occurrence frequency of the ULF signals with increased intensity as compared with the background values in this region [Sobolev and Mikhailov, 1998]. Currently, neither the theory nor the experimental data provide an answer to the question of whether the lower ionosphere parameters change. Further statistical studies of the electric field distribution over typhoons in various regions of the Global Ocean in various geophysical conditions are needed.
1. Anomalously high electric fields are detected for the first time in the daytime conditions at h = 1500-2500 km during typhoon development period over the Pacific Ocean aquatory ( l = 120o-280o E, j = 7o S-15o N). This result contradicts the daytime measurements of the electric fields of thunderstorm discharges on board satellites and theoretical ideas on the properties of the attenuation coefficient of the ULF-VLF waves propagation through the lower ionosphere.
2. Two maxima at f1 = 8 Hz and f2 = 225 Hz are found in the spectral distribution of the electromagnetic field E component. The first maximum intensity varied strongly from one orbit to another. The second maximum intensity often exceeded the upper limit of the receiver channel dynamic range (~80 m V m -1 ). The signal level at the frequencies of 9.6 and 15 kHz was below the channel sensibility threshold (~2 m V m -1 ). Such spectral distribution of the field E component agrees qualitatively with the theory of the ULF-VLF wave whistler mode propagation through the lower ionosphere.
Aksenov, V. I., Passage of electromagnetic waves of ultralow frequency through the ionospheric plasma, Radiotekh. Elektron. (in Russian), 11 (6), 1030, 1966.
Aksenov, V. I., and I. V. Lishin, Ion influence on the propagation of ULF electromagnetic waves through the ionosphere, Radiotekh. Elektron. (in Russian), 12 (4), 703, 1967.
Aksenov, V. I., and M. V. Nazarova, Numerical solution of the problem on ULF electromagnetic wave propagation though the lower ionosphere, Radiotekh. Elektron. (in Russian), 13 (2), 72, 1972.
Gidrometeoizdat, Atmosphere (in Russian), p. 144, 1991.
Hayakawa, M., Satellite observation of low latitude VLF radio noise and their association with thunderstorms, J. Geomagn. Geoelectr., 41, 537, 1989.
Kapustina, O. V., G. A. Mikhailova, and Yu. M. Mikhailov, Features of the amplitude spectra of the whistlers at a frequency below 1 kHz (the Intercosmos 5 and Intercosmos 14 satellites), Geomagn. Aeron. (in Russian), 21 (3), 457, 1981.
Mikhailov, Yu. M., G. A. Mikhailova, and O. V. Kapustina, ULF-ELF electromagnetic emissions over the fault in Kangaroo Valley of India and their relation with radon emanation (Intercosmos 24 satellite data), J. Tech. Phys., 40 (1), 317, 1999.
Pokrovskaya, I. V., E. A. Sharkov, I. N. Klepikov, and I. A. Karaseva, Catalogue and database of global cyclogenesis during 1988-1992, Preprint 1869 (in Russian), 29 pp., Inst. of Space Res., Moscow, 1993.
Sharkov, E. A., Aerospace studies of tropical cyclones, Issled. Zemli Kosmosa, 6, 87, 1997 (in Russian).
Sobolev, Ya. P., and Yu. M. Mikhailov, Typhoons in Caribbean region registered on satellites in VLF band, in Proceedings of Second Indo-Russian Symposium on Nature and Variations of the Geomagnetic Field, Moscow, July 28-31, 1997, Geomagnetic Field and Ionosphere Variations, edited by Kh. Kanonidi and Yu. Ruzhin, p. 230, Inst. of Terr. Magnetism, Ionosphere, and Radio Wave Prop., Moscow, 1998.