Ionospheric scintillation observations from SHAR

Ionospheric scintillation observations from SHAR H. CHANDRA, G. D. VYAS, H. S. S. SINHA, R. N. MICRAand S. PRAKASH Physical Research Laboratory, Navrangpura, Ahmedabad 380009. India (Rrceiced in.final,fimn

18Murch 199 I )

Abstract-VHF intensity scintillations associated with equatorial spread-F were recorded at SHAR (dip 13”N). India, using the 136.1 MHz beacon from the geostationary satellite ETS-2. The observations were made during the periods 12 March-26 April 1986 and later from 15 September to 5 October 1988. The maximum hourly occurrence frequency of scintillations was about 50% during March 1986and 30% during April 1986. It increased to about 85% during September-October 1988. The onset time of scintillations was generally between 20 and 21 h IST (77.5’E) during 1986 and between 1900 and 1930 h IST during 1988. Scintillations lasted for a longer duration of night in 1988 than in 1986. Data for a few nights were analysed to study the temporal variation of the scintillation index S,, computed at close time intervals. Wave-like fluctuations with a dominant periodicity of about 1h and another periodicity of about 20 min were seen. Power spectra were studied for 225 data samples of 100 s duration each for the night of 45 October 1988. The spectral index &P(f) ~,f~-“f ranged between 2 and 6.8 with a mean value of 4. The temporal variation of the spectral index @) also showed a periodic fluctuation. The higher the S, index the steeper were the spectral slopes.

1. INTRODUCTION

Ionospheric irregularities covering a wide range of scale sizes are associated with the phenomenon of equatorial spread-F. A number of remote groundbased and in-situ techniques as well as theoretical work are used to study this phenomenon. Recent reviews (FEJER and KELLEY, 1980; OSSAKOW, 1981; KELLEY and MCCLURE, 1981) cover these aspects in detail. Recording of the amplitude fluctuations (scintillations) of the radio beacons transmitted from a geostationary satellite provides a simple way of continuously monitoring the sub-kilometer scale size irregularities associated with equatorial spread-F. Work on equatorial scintillations has been reviewed by BASUand BASU (1981). A great deal of work on equatorial scintillations comes from the observations in the American zone. In the Indian zone detailed studies ofequatorial scintillations began with the multifrequency scintillations recorded during the period August l975-July 1976 when the ATS-6 satellite was positioned at 34% Scintillation observations at 40, 140 and 360 MHz were made for the period October 1975-August 1976 at Ootacamund (dip 6”N) by the Physical Research Laboratory (RASTOGIet al., 1977). Night-time scintillations recorded at Ootacamund were associated with equatorial spread-F while the daytime (strong) scintillations were associated with blanketing type E, (CHANDRAet al., 1979). The basic features of scintillations at Ootacamund have been reported by RASTOCX et a/. (1982) which showed equi-

noctial maxima. Multifrequen~y scintillations were also recorded at Thumba (dip O.BjS) (M~RTHY Edal., 1979). Later, with the availability of the radio beacon at 136.1 MHz from the Japanese satellite ETS-2, spaced receiver scintillation recordings were made at Tiruchirapalli (dip 4.8”N) during the year 1978 by the Physical Research Laboratory. The occurrence frequency of scintillations was noticed to be more during April than during July and December [CHANDRA e? al. (1989)]. The irregularities were found to move eastward with a velocity ranging from about 200 to 300 m/s in the post-sunset period and decreasing to about 100 m/s around 02 h. Away from the magnetic equator detailed observations have been made at Calcutta, near the anomaly crest region, which show equinoctial maxima during high sunspot years (DAS GUPTA et al., 1981). RASTOGI et al. (1989) have compared scintillations recorded at Bombay with spread-F at Ahmedabad and observed that both show nlaximum occurrence during winter and minimum occurrence during the summer months for the low solar activity period. Recently scintillations are being recorded at a number of locations in India; however, there exist gaps between stations close to the magnetic equator and stations close to the anomaly crest. A receiver to record 136. I MHz beacon from ETS-2 was set up at SHAR (dip 13”N) by PRL and the system was operated during the period I2 March26 April 1986. The recordings were made on a variable speed paper chart recorder. The primary aim of the experiment 167

was to make

co-ordinated

scintillation

168

H. CHANDRA et

observations along with the rocket-borne in-situ measurements of the electron density irregularities. Unfortunately the RH-560 rocket carrying the Langmuir probe failed in its performance. The scintillation observations recorded at SHAR during this period have been analysed for the occurrence frequency of scintillations. On a few nights the scintillations were recorded with a faster chart speed. The scintillation index S,, which is the normalized r.m.s. value of the intensity fluctuations (BRIGGSand PARKIN, 1963)and is a measure of the fluctuations in the electron density, has been computed on these occasions at close intervals of time. Scintillation observations were made again during September-October 1988 when a RH560 rocket, instrumented with a Langmuir probe and electric field double probes, was flown from SHAR for in-situ measurements of the fluctuations in electron density and electric fields associated with equatorial spread-F. The new digital ionosonde at SHAR was operated every I.5 min and on near-continuous mode for some of the nights. This was the first time that simultaneous rocket-borne in-situ and scintillation observations have been made in the Indian zone. The scintillation data were recorded digitally on audiocassettes using a newly developed microprocessorbased system. Data were digitized at a rate of 10 Hz and recorded on blocks of 100 s duration each. The sampling at 10 Hz has permitted power spectra1 estimations up to a frequency of 5 Hz. The scintillation observations during the two campaigns are presented in this paper. In-situ observations and their comparison with scintillation data will be reported later.

al. 061

SHAR

ET.??-2

1364MHz 1

lEl....l,~.......l~..~~.. IS

20

. . 25

I

3OMAR,l966

Fig. I, The occurrence of scintillations at SHAR (full lines) and of spread-Fat Kodaikanal (dashed lines) shown for each day of the period of observations during March-April 1986.

fact that Kodaikanal lies about 400 km south from the sub-ionospheric point viewed from SHAR. Since the phenomenon of equatorial spread-F is associated with the field-aligned irregularities, the altitude to which the F-region rises near the magnetic equator will determine the N-S extent of the latitude belt over which the irregularities will be seen. An examination of the variations of the minimum virtual height of the F-layer (h/F) at Kodaikanal was therefore attempted. Figure 2 shows one such example to illustrate the differences in h’F variations over Kodaikanal on

2.REWLTS

The occurrence of scintillations recorded at SHAR (sub-ionospheric point at 300 km being 82.3% I3.5”N) during night hours (18-06 ET) of the period of observations for March-April 1986 are shown in Fig. 1. The full lines in the figure show the time duration for which scintillations were observed. For comparison the occurrence of spread-Fat the nearest ionosonde location Kodaikanal (latitude 10”N) obtained from the published hourly ,I;,F2 tabulations are also marked with dashed lines. In this figure symbol ‘C’ denotes no data due to equipment failure or no observations made. In general there is a good agreement between the occurrence of spread-F at Kodaikanal and that of scintillations observed from SHAR. However, there are a few nights when spread-F was observed at Kodaikanal but there were no scintillations noticed at SHAR. This could be due to the

I Y

CL -c

Fig. 2. The variations of the minimum virtual height of Flayer (h’fl over Koda~kanal for spread-F nights (i) IO-1 I April 1986 when s~intiliations were present over SHAR. (ii) lb13 April 1986 when scintillations were absent over SHAR.

Ionospheric

scintillation

observations

from SHAR

169

There is a very good correspondence presence of scintillations and spread-F this period of observations.

between the over SHAR in

Temporal variation of scintillation index

lid 15

1

1

20

25

During 1986 scintillations were recorded with a fast speed chart for a few nights. The data on these occasions were digitized every 2.5 s and analysed to compute the strength of scintillations as characterized by the index S,. A sample length of 6 min was used to compute S, values. Figure 4 shows an example of the variations of S, with time shown for the night of 18-19 March 1986. The S, values fluctuated between 0.3 and 0.7 with a dominant periodicity of 67 min and another periodicity of 20 min as shown by the power spectrum in Fig. 5. If these periodicities in S, are related to large-scale wave type structures then the horizontal wavelengths can be estimated provided the drift velocity is known. Since 100 m/s is a typical value

I

I

30

05

SEPTEMBER

OCTOBER

1968

Fig. 3. The occurrence of scintillations (full lines) and spreadF (dashed lines) over SHAR shown for each day of observations during September-October 1988.

spread-F days with scintillations present or absent at SHAR. On the night of IO-11 April 1986 both spread-F and scintillations were present but during the night of 12-13 April 1986 scintillations were absent at SHAR even though spread-Foccurred at Kodaikanal. The base of the F-layer (h’F) on l&l 1 April rose rapidly in the post-sunset period reaching an altitude of 340 km at 1930 h. The h’F was 350 km at 2230 h but decreased to 260 km at 0130 h. In contrast the h’F remained around 260 km from 1830 to 2130 h on 12 April. The h’F, however, rose beyond 300 km around midnight. Figure 3 shows the occurrence of scintillations and spread-F at SHAR during the period of 1.5 September-5 October 1988. The occurrence of scintillation had increased to nearly 85% for this epoch. The onset time of scintillations at around 1900-1930 h is earlier by at least 1 h as compared to 1985. Further the duration of the scintillations over the night is longer with scintillations present almost up to 0530 h.

d

SHAR

5

3 x

0.0I 0.5

0.0

I 1.0

FREQUENCY

Fig. 5. Power spectrum

1.5

IN mtiz

obtained from the time variations S, in Fig. 4.

18-19 MARCH 1986

1 8

0.5 -

5 Ii F z s:

1 0.0.

-

1

2300

I 0000

I HOUR

Fig. 4. Time variation

of the scintillation

I

0100

0200

I 0300

IST

index S, at 136.1 MHz recorded 18-19 March 1986.

over SHAR during

the night of

of

H. CHANDRA

170

SHAR

el

al

WAR

4 OCT. 1988

4 OCT. 1988

0

1932 HR

1851 HRS

-8o-I

90

l /,CONFIDENCE LIMIT

I

0.01

I

1

0.1

1.0

FREQ.

(a)

IN

Fig. 6. Typical examples

I

HZ

POWER

(b)

I

0.1

I.0 FREQ.

of the power spectra obtained from scintillation data recorded the night of 4 October 1988. (a) 1815 h, (b) 1932 h.

of the drift of irregularities as measured by spaced receiver scintillations, these periodicities would correspond to wavelengths of 400 and 120 km, respectively. Such large-scale horizontal structures associated with equatorial spread-F are known to occur with signatures in backscatter radar maps and in the h’F (minimum virtual height of F-layer) variations. It may be pointed here that from earlier scintillation observations made at Tiruchirapalli, CHANDRA et al. (1989) have reported the variation of the scintillation index S,, for a few nights during February 1978. An examination of the S, values presented by them also shows fluctuations with a period of 60-90 min.

3.

0.01

I

IN

HZ

over SHAR during

two examples of the power spectra obtained from the scintillation data for the night of 4 October 1988. The spectral index values computed by fitting a straight line between 0.2 and 3 Hz (50&33 m for an assumed drift of 100 m/s) are 5.5 and 3.94 for the two examples with S, values of 0.25 and 0.34, respectively. About 225 samples were analysed for power spectra during this night. Figure 7 shows the histograms of the number of occurrences of the spectral index values. The values range from 2.0 to 6.8 with a mean value of 4.07 and a median value of 3.96. The mean value of 4 corresponds to a value of 3 for a one-dimensional in-situ comparison. In Fig. 8 are shown the temporal variations of the

SPECTRA

The data during 1988, recorded digitally at 10 Hz sampling rate, have permitted power spectral studies up to 5 Hz. In terms of the spatial scale this corresponds to 20 m for an assumed drift speed of 100 m/s. Thus the useful power spectral range is from about 800 m (Fresnel zone size) down to 20 m. The power spectra were computed using an FFT routine. A data length of 100 s was used and after removing the mean, auto-correlation functions were calculated up to 128 lags. A 512 point FFT was used for computing power spectra. The frequency resolution of the spectra is 0.08 Hz. However, the usefulness of the spectra at the lower frequency end is determined by the Fresnel zone size which turns out to be 0.1 Hz for an assumed drift velocity of 100 m/s, and 0.2 Hz for an assumed drift velocity of 200 m/s. Figure 6 shows

Fig. 7. Histograms of occurrence of the spectral index p obtained from scintillation data recorded over SHAR during the night of 4-5 October 1988.

Ionospheric I

v

I

,

SHAR 4-iOCTOBER 1988

3;

I

scintillation =

- 50

t:

-4.0

;

0.4-

-3.0

u a L

0.3-

-: VI

O.Z-

FLIGHT

1 1.8

4

20

I

t ,

l,ME I 00

22

HOUR

I 02

I S 1

Fig. 8. Time variations of the scintillation index S, and the spectral index value p obtained from the scintillation data recorded over SHAR during the night of 4-5 October 1988.

S, index and the exponent valuep for 4 October 1988. It is interesting to note that there are oscillations in both the parameters. It appears from the variation of S, and spectral index that higher S, values are accompanied by a steeper spectral slope in power spectra. This is illustrated in Fig. 9 where mean spectral index values are plotted for different groups of S,. The mean spectral index value varies from 3.30 for S, values in the range of 0.20-0.25 to 4.70 for Sq values in the range 0.450.50. There is evidence of the spectra1 index value approaching a saturation value at higher S4 values. 4. DtSCUSSlON

The mean spectral index value of 4 is close to the results obtained at other locations. BASU and WHITNEY (1983) have reported a mean value of 3.5 for scintillations observed at Ancon, near the magnetic equator. These values were obtained in the spatial scale range of 700-100 m. The spectral indices ranged

Fig. 9. Variation of the average value of the spectral index p with S, index. The numbers indicate the number of data points over each S, grouping.

observations

from SHAR

171

between 3 and 3.5 for weak scintillations. BRAMLEY and BROWNING (1978) reported a mean spectral index value of 4.4 for a mid-latitude station, while KERSLEY and CHANDRA (1984) obtained a mean value of 3.58 for the scintillations associated with the equatorward edge of the high-latitude scintillation region. The spectral index value of 4 obtained here from the temporal spectra of scintillations would correspond to a onedimensional spectral index value of 3. In-situ observations obtained from rocket-borne and satelliteborne measurements in the intermediate scale of wavelengths (20 km-100 m) give a mean spectral index close to 2.5. BASU et al. reported a spectral index value of 2.8 for scale sizes less than 1 km and of 1.5 for scale sizes greater than 1 km from in-situ data obtained from AE-E satellite. This type of break around 1 km has been confirmed in the recent CONDOR campaign also, with an index value of 3.0 for smaller scale sizes (KELLEY et al., 1986). The present results are for irregularities between about a few hundred of meters down to a few tens of meters and hence are in agreement with the in-situ observations. PRAKASH et al. (1991) have studied the power spectra of in-situ electron density fluctuations An&r, measured using a rocket-borne Langmuir probe flown from SHAR during well-developed spread-F. The spectral index values in the scale size range 20-200 m varied between - 1.5 and -4.6, which would correspond to the scintillation spectral index (p) values from 2.5 to 5.6. The present results with p values generally between 2.4 and 6.0 are in close agreement with these in-situ observations from SHAR. Further the spectral index values were seen to increase with the spectral power (An,./n,,)* of the irregularities. PRAKASHet al. (199 1) have described the relationship between the spectral index and spectral power by a Gaussian function which gives a maximum value of -4.425 for spectral index. This would correspond to a maximum p value of 5.425, which is in close agreement with the results shown here. Since larger (AN,)’ would contribute to stronger scintillations the present results showing steeper spectral slopes with increasing S, index are also consistent with the in-situ results. The increase in the spectral index with scintillation index is also consistent with similar results reported for equatorial scintillations at Ancon and for high latitude scintillations. The scintillation index varied between 0.2 and 0.5 only in the present data set. However, an examination of the results presented by BASU and WHITNEY (1983) shows that this steepening of spectral slopes is basically seen up to an S4 value of about 0.54.6 only, beyond which spectral index remains independent of S,. RINO (1979a,b) has developed phase screen models for both weak and strong scatter cases. The irregularities are

H. CHANDRAet al.

172

characterized by a power law spectral function of the form C,q-(2Y+ ‘1 where C, is the three-dimensional turbulent strength of the electron density irregularities. The power spectral form of the phase scintillations is described by 7”:f-p, where p = 2v and T is a function of the radio wavelength, layer thickness, zenith angle, anisotropy parameters of the irregularities, radio propagation angle and the relative satellite irregularity motion. The scintillation index S, for such a power spectral form of irregularities is given by S4 K

$3+P)!4

or

Xf-‘3+P)i4,

This shows that the frequency dependence of scintillation index itself varies with the spectral dependence of irregularities. For spectral index values varying between 3 and 5 here the frequency dependence of S, would vary from f-_ '5to f- *. However, the frequency exponent (N) values obtained from multifrequency scintillations at Ootacamund were always less than 1.5 (RASTW et al., 1990). However, this discrepancy has been explained in terms of the two

component power law form, wherein the spectral slope for the larger wavelengths (lower frequency roll oh) would determine the frequency dependence of S, (FRANKE et al., 1984). What is a more interesting finding here is the periodic fluctuations in both S, as well as in the spectral index P. The spectral index values undergo periodic fluctuations ranging from 3 to 5 in a time of about 1 h. What causes this change is not known. Whether this is a consequence of changes in the altitude of the F-layer or in the velocity caused by changes in the electric field is to be ascertained. Simultaneous velocity measurements by spaced receiver scintillations and phase path measurements would surely help in dete~ining this. Besides, there could be a possibility of fluctuations in neutral winds, as all these parameters control the growth rate of irregularities. There is a need to measure various parameters simultaneously for a better understanding of the irregularities associated with spread-Fand scintillations. Ackt?o~l~~ge?~ents--The authors are grateful to the Range Director, IREX, SHAR, for providing the facilities to record the scintillations a2 SHAR.

REFERENCES BASUSA. and BASUSu. BASUSA. and WHITNEYH. E. BRAMLEYE. N. and BROWNINGR. BRIGGSB. H. and PARKIN1. A. CUAN~RA H., PATEL V. P., DE~HPANDEM. R. and VYAS G. D. CHANDRAH., VATS H. O., DESHPANDE M. R., RASTOGI R. G., SETHIAG., SASTRYJ. H. and MURTHYB. S. DAS GUPTA A., MAITRAA. and BASUSA. FEJERB. G. and KELLEYM. C. FRANKE S. J., LIU C. H. and FANG D. J. KELLEYM. C., LABELLEJ., KUDEKIE., FUER B. G., BASU SA., B&U Su., BAKER K. D., HANUISEC., ARGO P., W#~~AN SWERTZW. E., FARLEYD. T. and MERIWE~ER J. W. JR. KELLEYM. C. and MCCLURE1. P. KERSLEYL. and CHANDRAH. M~~RTHY K. K., REDDYC. R. and MURTHYB. V. K. OSSAKOWS. L. PRAKASHS., PAL S. and CHANDRAH. RASTOGIR. G., CHANDRAH. and DESHPANDE M. R. RASTOGIR. G., DFSHPANUE M. R., VATS 0. H., DAVIESK., GRUFF R. N. and JONESJ. E. RA~TOGIR. G., KOPARKAKP. Q., CHANDRAH. and DE~HPAWDE M. R. RASTOGIR. G., KOPARKARP. V.. CHANDKAH. and VYASG. D. RINOC. L. RINOC. L.

1981 1983 1978 1963 1989

J. atmos. terr. Phys. 43, 473. Radio Sri. 18, 263. J. atmos. terr. Phys. 40, 1247. J. atmos. terr. Phys. 25, 339. Ind. J. Radio @ace Fhys. 18, 125.

1979

Ann. Geophys. 35, 145.

1981 1980 1984 1986

Radio Sci. 16, 1455. Rea. Geophys. 18,401. Radio Sci. 19.695. J. geophys. Rk. 91, 5487.

1981 1984 1979 1981 1991 1982 1977

J. atmos. terr. Phys. 43,427. J. otmos. terr. Phys. 46,667. J. atmos. terr. Phys. 41, 123. J. atmos. terr. Phys. 43,437. J. atmos. terr. Phys. 53, 917. Ind. J. Radio Space Phys. 11, 240. Pramana 8. 1.

1990

J. nimos. terr. Phys.. 52, 69.

1989

Amt. Geophys. 7, 28 I

1979a 1979b

Radio Sci. 14, 1135. Radio Sci. 14, 1147.