Citation
Spectrum and origin of the Jovian radio burst structure

Material Information

Title:
Spectrum and origin of the Jovian radio burst structure
Creator:
Block, Wilbur Frank, 1930-
Publication Date:
Language:
English
Physical Description:
x, 106 l. : illus. ; 28 cm.

Subjects

Subjects / Keywords:
Dissertations, Academic -- Physics -- UF
Jupiter (Planet) ( lcsh )
Physics thesis Ph. D
Radio astronomy ( lcsh )
City of Gainesville ( local )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis--University of Florida, 1965.
Bibliography:
Bibliography: l. 101-104.
General Note:
Manuscript copy.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
022276225 ( ALEPH )
13589885 ( OCLC )

Downloads

This item has the following downloads:


Full Text
SPECTRUM AND ORIGIN OF THE JOVIAN
RADIO BURST STRUCTURE
By
WILBUR FRANK BLOCK

A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
April, 1965




To my wife Jo Ann
and
Our children Herman Karl
Mary Edna Johh Beal
Grace Ellen




ACKNOWLEDGMENTS

As is usually the case for fledgling Doctors of Philosophy in physics earning their wings through research in radio astronomy at the University of Florida, this dissertation represents an endproduct of the work of a team of research workers without whose Significant contributions this study could never have been accomplished.
Dr. A. G. Smith., as chairman of the author's supervisory committee, provided the impetus for this work. His intense interest and uncommon enthusiasm forresearch have provided the author a worthy example. Likewise, Dr. T. D. Carr, a member of the committee, in numerous consultations was always ready and able to resolve puzzling questions., and pose a few more! Dr. D. C. Swanson's encouragement and conversations have proved invaluable. The other committee members, Dr. T. A. Scott and Dr. R. G. Blake., were particularly helpful during the author's preparation for the qualifying examinations., always having time and desire to discuss and interpret concepts troublesome to the writer.
Dr. S. S. Ballard., Chairman of the Department of Physics and Astronomy, has been most generous in terms of personal consultation, encouragement, and arranging financial assistance. Mr. Harold J. Huber of Orlando, Florida, was instrumental in obtaining for the writer the Research Task Award of the Martin Company (Orlando,, Florida). The author is deeply indebted to the Martin Company for this assistance during the early part of his doctoral studies.
iii




The technical staff of the Department of Physics and Astronomy have always been willing to help the author to the full extent of their capabilities. Mr. Ralph Warren's extraordinary ability to make any type of gadgetry function contributed especially to the satellite tracking project. One cannot say enough good things to describe Mr. Hans W. Schrader whose practical knowledge of physics, machine work, photographic techniques, and electronics are supplemented by his personal integrity and spirit of cooperation.
The writer enjoyed many enlightening conversations with his colleagues, particularly Dr. G. R. Lebo, Dr. N. F. Six, Dr. A. T. Jusick, Dr. Samuel Gulkis, Frank Tiberi, Jorge May, and C. N. Olsson. Dr. Lebo personally performed the Chree analyses utilizing the University's computing facility.
The incomparable W. W. Richardson assisted in setting up the artificial satellite observation channel, performed data analysis, prepared photographs and slides, and did his usual excellent job in preparing the drawings.
W. A. Morton contributed directly to the experiment in which artificial satellite signals were compared with Jupiter storms. He helped set up the equipment, record and analyze the data, and plan the presentation and interpretation of the results. In addition, he contributed significantly to the Jovian continuum search and to the Jovian burst morphology study.
M. L. Fagerlin constructed the log-periodic antenna and has been primarily responsible for this systems maintenance.




Other staff members who have contributed to this work in some way include Mrs. Lee Potzner, R. J. Leacock, T. Anderson, G. W. Brown, W. Cain, E. J. Lindsey, W. Mock, I. Shever, G. Walls, W. Greenman, C. Arlington, P. Trescott, R. Hayward, M. Evans, K. Williams, N. Chotas,
J. Moeller, D. Smoleny, and our Chilean colleagues, H. Bollhagen and J. Levy.
The writer thanks his parents, Dr. and Mrs. Herman H. Block,
for constant encouragement and financial assistance without which this work would certainly never have been completed.
The author's brother, David Block, and his friend, Thomas B. Elfe, have offered unusually stimulating advice and encouragement.
The professional touch of. Mrs. Thomas Larrick in editing and typing the manuscript made a joy of the burdensome task of assembling the work in final form. To work with such people as she certainly
renews one's faith in humanity.
To all these fine people the author wishes to express his most sincere thanks.
Gratitude is also expressed to the sponsors of the University of Florida radio astronomy projects: the National Aeronautics and Space Administration, the National Science Foundation, the Office of Naval Research, and the United States Army Research Office at Durham.




TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS ........................ iii
LIST OF TABLES . .......... . . . .... vii
LIST OF FIGURES . . . . ................ viii
Chapter
I. INTRODUCTION . . .......... . . . .. 1
II. THE COLLECTION AND ANALYSIS OF JOVIAN
DECA=ETRIC SPECTRA . . . . ........ .. 3
1. Background ............ .............. 3
2. Description of Jovian Spectra . .......... 6
3. Frequency Dependence of Ionospheric Faraday
Rotation from Jovian Decametric Spectra ..... 29
4. Comparison with Other Workers' Data ... ....... 38
5. Spurious Spectra ..................42
III. MORPHOLOGY AND ATMOSPHERIC MODIFICATION .... ......... SO
1. The Search for a Jovian Continuum ........ 50
2. Jovian Burst Morphology .............60
a. Introduction ............. . 60
b. The Jupiter Pulse Character Index . .. 61
c. Diurnal Variations ..... ............. 63
d. Seasonal Variations ..... ............. .69
e. Frequency Dependence ................. 75
f. Altitude Dependence ................. 75
g. Geomagnetic Activity Dependence ...... ... 78
h. Solar Activity Correlation ............ 78
i. System III Longitude Dependence ...... ...s
j. Dependence on the Elongation of Jupiterts Moon lo ...... ............. 85
3z A Comparison of Jovian Radio Bursts with
Artificial Satellite Signals at Appulse . .. 89
IV. CONCLUSION . . . . . ........... 99
LIST OF REFERENCES . .................... 101
BIOGRAPHICAL SKETCH ............. ............ 105




LIST OF TABLES

Table Page
1. Definitions of Pulse Character and Scintillation
Indices ....... .................. . . . 62
2. The General Comparison of Jovian Pulse Character
with Artificial Satellite Signal Scintillations
Near Appulse ................. *. 97




LIST OF FIGURES

Figure Page
1. Block diagram of the University of Florida radio
spectrograph ....... ...................... 4
2. Examples of Jupiter bursts of different character
index C . ... . . . . .......7
3. The spectrum of slow Jupiter (September 26, 1963) . 9
4. Frequency variation of Faraday rotation on slow
Jupiter spectra (October 11, 1963).. . . . 11
5. Normal Jupiter spectrum (August 26, 1963) ... ....... 12
6. Normal Jupiter spectrum (August 26, 1963) . . . 14
7. Normal Jupiter activity at 16 Mc/s from the highspeed oscillograph record ... .......... . . 15
8. The comparison of the time variation of the amplitude
of a spectral spike at 18 Mc/s with the 18 Mc/s
oscillograph record . . . . . . 16
9. A frequency-drifting pulse in the spectrum of mixed
fast and normal Jupiter (September 13, 1963).... 18
10. The spectrum of mixed fast and normal Jupiter
(September 13, 1963) . .. . . ...... 20
11. Spectra of fast Jupiter (November 5, 1963) . . . 21
12. Spectra of fast Jupiter (November 5, 1963) . . . 22
13. Spectra of fast Jupiter (November 5, 1963) . . .. 23
14. Spectra of fast Jupiter (September 26, 1963) . . 25
15. Spectra of fast Jupiter (September 26, 1963) . . 26
16. Spectra of fast Jupiter (September 26, 1963) . .. 27
17. Logarithmic plot of rate of change of Faraday rotation
spectral fringes as a function of frequency . . 32

viii




LIST OF FIGURES (Continued)

Figure Page
18. Terrestrial ionospheric Faraday rotation as a function
of frequency for Gainesville, Florida, 0520 Universal Time, October 11, 1963 .... ................ 34
19. Comparison of S-66 signals at 20 and 40 Mc/s . . . 36
20. Jovian decametric spectra as recorded at the High
Altitude Observatory at Boulder, Colorado [19] .... 39
21. Various types of spectral interference observed with
the University of Florida radio spectrograph ....... 45
22. Spectral response of the University of Florida radio
spectrograph to broad-band noise ........... 47
23. A 22.2 Mc/s Jupiter storm of mixed normal swishes and
continuum . .... 52
24. A 22.2 Mc/s Jupiter storm of swishes with no continuum 55
25. A 22.2 Mc/s Jupiter storm superimposed on the
Cassiopeia A continuum ............... 56
26. The diurnal variation of ionospheric scintillations . 64
27. The diurnal variation of Jovian burst fine structure
(Florida, 1962) . . . . ............ 65
28. The diurnal variation of Jovian burst fine structure
(Chile, 1962) ..................... 66
29. The diurnal variation of Jovian burst fine structure
(Florida, 1963) ............... .......... 67
30. The diurnal variation of Jovian burst fine structure
(Chile, 1963) ......................... . . 68
31. The seasonal and altitude variations of radio star
scintillations ..................... 70
32. The seasonal variation of Jovian burst fine structure
(Florida, 1962) . . . . . . . . . 71
33. The seasonal variation of Jovian burst fine structure
(Chile, 1962) . . . . . ........... 72
34. The seasonal variation of Jovian burst fine structure
(Florida, 1963) . . . ........ . . 73




LIST OF FIGURES (Continued)

Figure Page
35. The seasonal variation of Jovian burst fine structure
(Chile, 1963) . . . ............... 74
36. Frequency dependence of radio star scintillations [42] 76
37. Frequency dependence of Jovian burst fine structure . 77
38. Altitude dependence of Jovian burst fine structure . 79
39. Geomagnetic activity influence on Jovian burst fine
structure . . . . . . . . . 80
40. Chree analysis of sunspot number ............ 82
41. Chree analysis of FX1, FX2, FX3, C, A (all dimensionp
less), and the 2800 Mc/s solar flux (10-22 w/m 2/cps) 83
42. System III longitude dependence of Jovian burst fine
structure . ................... . 86
43. The influence of Jupiter's moon Io on Jovian burst
fine structure . . . . .. 87
44. Extreme examples of scintillation index S . . . . 92
45. A direct comparison of Jovian burst structure with
satellite signal scintillations . . . .... 93
46. The dependence of the scintillation index S on
scintillation frequency .............. 94
47. Scatter diagrams of Jovian pulse character C versus
scintillation index S ................ 96




CHAPTER I

INTRODUCTION
Jupiter's sporadic decametric radio emissions, discovered by
Burke and Franklin [1] in 1955, have been eagerly recorded and systematically analyzed by University of Florida physicists since 1957. Following the inspiring leadership of A. G. Smith, these scientists are attempting to compile information sufficient to suggest a suitable physical model for the origin of the no-longer strange Jovian signals.
Successive dissertations of T. D. Carr [2], N. E. Chatterton [31, N. F. Six [41, and G. R. Lebo [51, have dealt primarily with the gross statistics of the experimental data. The purpose of this work is to amplify the previous papers through a consideration of the "fine structures so to speak, of the reams of data which have been collected at Biven's Bank. Thus, where Six [4] and Lebo [5] discussed exhaustively Jupiter's radiation probability and intensity, here will be treated the complementary material: the morphology of the Jovian bursts in the time and frequency domains.
Specifically, the dynamic spectrum analyzer, set up by Chatterton [3] in 1960, has been improved through replacement of the cumbersome, inefficient, immobile rhombic antenna by an automatic tracking, broad-band, log-periodic structure [61. The miles of film produced by this instrument since Chatterton's publication of 1961 have been




analyzed for spectral characteristics of Jovian emission and are reported later in Chapter II. The 22 Mc/s interferometer has been used to search for a radiation continuum emanating from Jupiter, the results appearing in Sec. III.1. Likewise, the high-speed oscillograph records have been examined for pulse duration characteristics as discussed in Sec. 111.2. Finally an attempt has been made to learn the effect of the Earth's ionosphere and magnetic field on the Jovian signals by a study of the signals received from a beacon satellite, S-66, orbiting outside the terrestrial ionosphere, as reported in Sec. 111.3.
Possibilities of the contrary notwithstanding, it is certainly hoped that the information documented here will speed the evolution of a suitable explanation for the origin of JupiterTs decametric radio bursts.




CHAPTER II

THE COLLECTION AND ANALYSIS OF
JOVIAN DECAMETRIC SPECTRA
1. Background
N. E. Chatterton, as part of his doctoral dissertation research, set up the University of Florida radio spectrograph in 1959, first observing Jovian decametric spectra on February 5, 1960. His dissertation [3] included many photographic reproductions of typical spectra observed through April 27, 1961, along with exhaustive descriptions of their characteristic features. Available at the time were spectra contained in some 2100 feet of 16 millimeter movie film. Chatterton noted particularly that bandwidths of the bursts ranged from 0.1 to 3 Mc/s, and that individual pulse shapes ranged from symmetric to nondescript. The tendency of the individual pulses to bifurcate was pointed out, and the general peaking or fringing fine structure was examined in detail. The continuing interest of the Florida group in Jupiter's spectrum is well known [7, 8, 9, 10].
A portion of the present work includes the extension of the spectral collection and the accompanying analysis. Much of the new spectral data have been taken on an improved version of the original Florida spectrograph. Figure 1 is a block diagram of the equipment. During observations Jovian radio signals incident on the log-periodic




LOG-PERIODIC
ANTENNA

SPECTRUM ANALYZER

Fig. 1.--Block diagram of the University of Florida radio spectrograph.




antenna are preamplified, then displayed as an intensity versus frequency plot on the cathode ray oscilloscope tube of a commercial spectrum analyzer (Panoramic Model SPA-3). At the discretion of the observer, the center frequency may be chosen between 13 and 23 Mc/s, the spectral sweep width between 0.25 and 4 Mc/s. Desirable data are recorded with time of occurrence on 16 millimeter movie film (Bolex Model 16R Camera, Eastman Kodak Tri-X Negative Film) at the rate of
12 frames per second. The principal modification to the original system was the replacing of the fixed rhombic antenna with an automatictracking log-periodic dipole antenna constructed by M. L. Fagerlin [6]. The reasons for changing antennas included increased observation time, improved efficiency, broader system bandwidth, and the determination of whether spurious ringing in the rhombic antenna system caused the peaking observed on Jovian spectra. The timer was changed from a minute counter to a clock displaying hours, minutes, and seconds. The framing rate of the camera was increased from 4 to 12 frames per second to help preserve quick-changing spectral effects.
Figures 3 through 16 include reproductions of characteristic examples of the collection which now contains approximately 8700 feet of 16 millimeter movie film, including data through December 21, 1963. Since Chatterton's study some new types of Jovian spectra have been identified, and thelarger data sample available now enables one to more satisfactorily classify the spectra. In Sec. 11.2 the spectral collection examples will be described in detail. Then in Sec. 11.3




the peaking or fringing phenomenon will be analyzed. A comparison with other workers' results follows in Sec. 11.4. Finally, Sec. 11.5 contains a discussion of some troublesome spurious spectral effects.
2. Description of Jovian Snectra
The discussion of the features of Jovian spectra is expedited
by the following brief description of the audio envelope of these radio signals. Jupiter's decametric radio storms appear to a terrestrial radio telescope as groups of pulses of electromagnetic energy, spaced randomly in time. The length in time of the individual pulses ranges from the order of milliseconds to tens of seconds. For analytic purposes the storms have been arbitrarily classified as "fast," "normal," or'blow,1 depending on whether the bulk of the pulses in a storm are of the order of approximately 0.1 second and shorter, 1 second, or several seconds in duration, respectively. These classes of Jupiter pulses are often referred to as "popping" or "spitting Jupiter," "swishes," and "long rollers." In Fig. 2, after Mock [11], typical high-speed oscillograph records of each type of Jupiter storm are shown. Slow pulses appear in the top strip, the center strip is a recording of normal Jupiter, and the lowest record depicts fast Jupiter activity.
The characteristic names are derived from the sound of the
storms as heard over the loudspeakers of the observatory. The normal swishes sound somewhat like ocean waves breaking on a distant beach. The fast pulses sound similar to the popping of popcorn in a covered pan. The slow rollers are detected as a slow change in the intensity of the hissing sound due to the galactic background noise and are




AUG. 27, 1962
-* I SEC.
MAY 24,1961 JONE 7, 1962

Fig. 2.--Examples of Jupiter Top: Slow Jupiter, C = 1. Center: C = 3.

bursts of different character index C. Normal Jupiter, C = 2. Bottom: Fast Jupiter,

CHILE 10 MC/S
X
J, 8J




sometimes scarcely audible, since the rate of change of intensity is so small. In Sec. III.2.b a character index is assigned to each class of pulse to lend organization to a parametric study of the Jovian burst morphology in the time domain. For the present the qualitative nomenclature suffices.
Figure 3 is a good example of slow Jupiter. In this and all succeeding spectral displays, time begins at the top left frame, progressing down to the end of the column, thence to the top of the next columnn,and so forth. This particular burst is centered at 19.8 Mc/s, with a sweep width of 3 Mc/s, and it was observed at 0239 Eastern Standard Time on September 26, 1963, lasting some 15 seconds. Successive frames shown here are spaced 1.5 seconds apart, except for the final frame, which has expanded sweep width to resolve possible fine structure details absent in the spectra filmed at the normal sweep width of 3 Mc/s. The activity is classified as slow Jupiter because the changes in intensity at any particular frequency are gradual, some features maintaining stability over several seconds. Thus there are four peaks which can be traced easily through the first six frames spanning 9 seconds. Other fringes may be traced correspondingly throughout the long sequence. Additional visible features, not the sole property of slow Jupiter, are the obvious peaking tendencies or fringes approximately
0.4 c/s wide and an apparent drift of the center frequency of the burst from 20.5 Mc/s to 19.0 Mc/s. The fringes are attributed to Faraday rotation of the Jovian signals in the terrestrial ionosphere and are discussed in detail in Sec. 11.3. The cause of the center frequency




(0) (b) (c)
UIm/I

I I I i
18.3 19.8 21.3
(Mc/s)

I a I I I a I a I
18.3 19.8 21.3 18.3 19.8 21.3
(Mc/s) (Mc/9)

I a I I 19.3 19.8 20.3
(Mc/s)

Fig. 5.--The spectrum of slow Jupiter (September 26, 1963).




drift is unknown. The sharp spike at 21.6 Mc/s is of instrumental origin, this local oscillator trouble being described in Sec. II.S. The final frame of column (c) at expanded sweep width of 1 Mc/s does not reveal any additional hy-perfine structure.
Figure 4 shows an unusual broad-band Jupiter storm observed
October 11, 1963. During the period 0015 to 0040 Eastern Standard Time slow rollers were recorded at frequencies ranging from 12 to 22 Mc/s. Single isolated frames are shown to demonstrate the broad-bandedness of the storm. In Fig. 4(a) the center frequency is 13.5 Mc/s, sweep width 3 Mc/s. The rapid fall-off of signal level near 12 Mc/s is due to the receiving band width of the spectrograph's log-periodic antenna. Faraday rotation fringes are clearly evident, crowding more closely together towards the lower frequency end of the spectrum. Figure 4(b),
(c), and (d) show slow Jupiter at center frequencies 15, 16, and 19 Nc/s, respectively, with sweep width 5 Mc/s. In each case the Faraday fringes become wider toward higher frequencies. The spectra shown as Fig. 4(e) and (f) show the slow Jupiter centered at frequencies 18 and 17.2 Mc/s, respectively, with the sweep width expanded to 1 Mc/s in hopes of finding hyperfine structure detail. Nothing uniform was observed, only the finely spaced irregular noise pulses of the galactic background.
Figure 5 is a display of normal Jupiter activity at 18 Mc/s and sweep width 3 Mc/s as observed from 0300 to 0301 Eastern Standard Time on August 26, 1963. Frame spacing is 1 1/2 seconds and the activity lasts some 28 seconds. While superficially this spectrum appears




I i I i I 12 13.5 15
(Mc/s)

13.5 15 16.5 16.5 18 19.5
(Mc/s) (Mc/s)

(e)

I p I .
17.5 19 20.5
(M c/s)

(f)

17.5 18 18.5 16.7 17.2 17.7

(Mc/s)

(M c/s)

Fig. 4.--Frequency variation on slow Jupiter spectra (October 11,

of Faraday rotation 1963).




- ')I um
I I I
16.5 18 19.5
(Mc/s)

I I I I
16.5 18 19.5
(M C/s)

Fig. 5.--Normal Jupiter spectrum (August 26, 1965).




similar to that of Fig. 3, close examination reveals that frame-toframe changes in fine structure are more noticeable in Fig. 5 than in
Fig. 3. This is the mark of normal Jupiter spectra: fine structure changes occurring in the order of 1/2 to 1 second. Note again the Faraday fringes and the apparent frequency drift from high to low. The lonesome spike appearing sporadically at'18.7 Mc/s is attributed to radio station interference. An interesting apparent low-frequency cut-off occurs at 17.3 Mc/s in frames 4, 5. and 6, column (a), gradually changing to a high frequency cut-off at approximately 17.7 Mc/s in frames 3, 4., and 5 of column (c).
Figure 6 is another example of normal swishy Jupiter pulses.
Frame spacing is now only 1/2 second for this activity of 0259 to 0300 Eastern Standard Time, August 26, 1963, centered at 18 Mc/s with sweep width 3 Mc/s. Two distinct spectral spikes at 17.6 and 18.0 Mc/s are seen to grow and decay. The 18.0 Mc/s peak was monitored as Jupiter activity and recorded through channel 18Y on the high-speed Brush oscillograph. Figure 7 is a reproduction of this record. Figure 8 is a plot of the amplitude versus time for the pulse as observed on the spectrograph (solid line) and the Brush record (dashed line). If we accept the observers judgment in identifying the Brush record pulse as Jovian in origin, the correlation of the two curves in Fig. 8 confirms the narrow spike at 18.0 Mc/s in Fig. 6 to be the spectrum of the pulse
on the Brush record, and hence to be of Jovian origin. The similar pulse at 17.6 Mc/s should likewise be due to Jupiter. However, the noticeably more narrow spike at 18.7 Mc/s is probably due to a radio




I I i I
16.5 18 19.5
(Mc/s)

I I i I I i I i I
16.5 18 19.5 16.5 18 19.5
(M c/s) (M c/s)

Fig. 6.--Normal Jupiter spectrum (August 26, 1963).




Fig. 7.--Normal Jupiter activity at 18 Mc/s from the high-speed oscillograph record.




---- OSCILLOGRAPH
-SPECTRAL PEAK

18 Mc/s

Il~
( '- \,~I

Al
- 1~
'I I'
I

F-:
0
0'
PO
0

, I I I

0 2
TIME (SECS)

Fig.
amplitude of oscillograph

8.--The comparison of the time variation of the a spectral spike at 18 Mc/s with the 18 Mc/s record.

16.0

12.0V-

10.0 8.0

6.0-

5.0'
5

I I

3

58

I I

14.0 f-




station. Note this same spike at 18.7 Mc/s in Fig. 5. Figure 6 is identified with normal Jupiter because intensity changes occur in the order of 1 second. Isolated spectral spikes like these, with bandwidths of approximately 0.1 Mc/s, are relatively rare.
During a period of mixed normal and fast Jupiter pulses the
data in Fig. 9 were taken, in which a pulse was observed to drift uniformly in frequency. The time was 0203 Eastern Standard Time, September 15, 1963, with center frequency 18 Mc/s, sweep width 1 Mc/s. Successive frames are spaced at intervals of 0.5 seconds. Fast Jupiter is evident at 17.3 Mc/s in the first three frames of Fig. 9(a). Note the rapid frame-to-frame changes. Similar fast activity is evident at approximately 18.5 Mc/s in the last three frames of Fig. 9(a). Better examples of fast Jupiter appear in later figures. These examples are explicitly mentioned here to emphasize that sometimes more than one
kind of Jupiter activity appears simultaneously. Thus, in the third frame of Fig. 9(a), a swishy pulse develops at approximately 18.06 Mc/s, which lasts throughout the sequence, a time of 8 seconds. This pulse changes amplitude principally in the manner of normal Jupiter. However, some fast changes at the second, third, fourth, fifth, and sixth frames of Fig. 9(c) show the pulse also to possess fast Jupiter characteristics. Another highly unusual feature of the pulse is its uniform drift in frequency from 18.06 to 17.95 Mc/s in 7 seconds, a drift rate of approximately 0.007 Mc/s per second. The tendency of Jovian storms to drift in frequency has long been known, but this is the first clear example of a specific Jovian pulse's drifting in frequency. It is noted




(c)

U

U

U

I I I I i I I
17.5 18 18.5 17.5 18 18.5
(Mc/s) (M c/s)

I l i i i 175 18 18.5
(M c/s)

Fig. 9.--A frequency-drifting pulse in the spectrum of mixed fast and normal Jupiter (September 13, 1963).




that the bandwidths of the Jovian pulses shown in Fig. 9 vary but are in the order of 0.1 Mc/s.
Figure 10 shows more mixed normal and fast Jupiter observed 0214 to 0216 Eastern Standard Time, September 13, 1963. Frames are spaced 1/2 second apart. The sequence in Fig. 10(a) is centered at 18 Mc/s, with sweep width 3 Mc/s. Figures 10(b) and 10(c) are both centered at 22.5 Mc/s with 1.0 Nc/s sweep width. While the normal Jupiter character is evident in the slowly changing envelope of each burst, the fast Jupiter features appear in sharp frame-to-frame variations of the individual pulse shape and the ragged, ill-defined tops of the pulses. This latter characteristic is seen especially at 17.2 Mc/s in frames 2 and 3 of Fig. 10(a), and again at 22.9 Nc/s in frames 2, 3, and 4 of Fig. 10(b). Clearly, then, fast and normal Jupiter can occur simultaneously. Aural monitoring agrees with this conclusion, in that observers plainly hear pops superimposed on swishes.
Figures 11, 12, 13, 14, 15, and 16 are examples of fast Jupiter pulses at various frequencies and sweep widths. Figures 11, 12, and 13 are taken from data of November S, 1963. and each consists of three strips of four successive frames spaced at 1/12 second. The center frequency is 18 Mc/s, sweep width 3 Mc/s. Fast Jupiter spectra are characterizedby the extremely rapid changes in pulse shape from frame to frame. For example, the first and fourth frames of the sequence in Fig. 11(a) show little activity, but the second and third frames show high intensity, fast Jupiter bursts at 18 Mc/s. Similarly, note the change in appearance of the spectral sequence in the first three frames




(b)

U

m 1L lm
I i I i I,
16.5 18 19.5
(Mc/s)

U
I I I I
22 22.5 23
(Mc/s)

I I i I 22 22.5 23
(Mc/s)

Fig. 10.--The spectrum of Jupiter (September 13, 1963).

mixed fast and normal




j I I I I i I
16.5 18 19.5 16.5 18 19.5
(Mc/s) (Mc/s)

I I i
16.5 18 19.5
(Mc/s)

Fig. 11.--Spectra of fast Jupiter (November 5, 1963).




(a) (b) (c)
I i l I I I I I I I
16.5 1 8 19.5 16.5 18 19.5 16.5 18 19.5
(M c/s) (M c/s) (M c/s)
Fig. 12.--Spectra of fast Jupiter (November 5, 1965).




(a) (b) (c)
i i l l 1 I I i lI i
16.5 I 8 19.5 16.5 18 19.5 16.5 18 19.5
(M c/6) (M C/s) (M c/s)
Fig. 13.--Spectra of fast Jupiter (November 5, 1963).




of Fig. 11(b). Then in Fig. 11(c) a fast, large amplitude pulse appears
at approximately 18.4 Mc/s in the second frame only, missing the first and third frames. Recalling that normal Jupiter swishes are produced by the small changes occurring over the 1/2 second frame spacing of Figs. 5 and 6, it is not surprising that these large changes in the spectrum over intervals of 1A2 second as in Fig. 11 accompany popping, spitting, or cracking noises over the loudspeakers at the observatory. More examples of these spectra of fast Jupiter follow.
In Fig. 12(a) the second frame alone shows an isolated burst at
18.5 Mc/s. The best similar example of fast Jupiter in Fig. 12(b) is also the second frame, which shows an intense burst of 1 Mc/s band width, there being scarcely any activity in the preceding and succeeding frames. Figure 12(c) shows what might be a fast Jupiter burst moving in frequency from low to high. The bursts in the first, second, and third frames exhibit the ragged-top character of fast pulses and are centered successively at approximately 18.0, 18.2, and 18.4 Mc/s. Figure 15 shows similar fast Jupiter bursts. Intense activity not shown in adjoining frames may be seen in the second and third frames of Fig. 15(a) near 18 Mc/s, the second and fourth frames of Fig. 13(b) near 18 Mc/s, and in the third frame of Fig. 15(c) just below 18 Mc/s.
In Figs. 14, 15, and 16 features similar to those of Figs. 1l, 12, and 13 are shown with expanded sweep. Each of these figures has three columns of three successive frames spaced 1/12 second apart. Figure 14(a) shows fast Jupiter in the second frame at 18.5 and 18.6 Mc/s, as does the second frame of Fig. 14(b) at 18.1 Mc/s. Figure 14(c) shows




179 18.4 18.9
(Mc/s)

I I I I I I IL
17.9 18.4 18.9 17.9 18.4 18.9
(Mc/s) (Mc/s)

Fig. 14.--Spectra of fast Jupiter (September 26, 1963).




(a) (b) (c)
i l I l i .I 1. I a I
179 18.4 18.9 179 18.4 18.9 179 18.4 18.9
(Mc/s) (Mc/s) (Mc/s)
Fig. 15.--Spectra of fast Jupiter (September 26, 1965).




(a) (b) (c)

OWN
I I
18.75 19 19.25
(Mc/s)

I i I I 1 I l i l l
18.75 19 19.25 18.75 19 19.25
(Mc/s) (Mc/s)

Fig. 16.--Spectra of fast Jupiter (September 26, 1963).




a fast burst at 18.1 Mc/s in the second frame. Figure 15(a) contains fast Jupiter at 18.1 Mc/s in the first frame, 18.05 and 18.6 Mc/s in the second frame, and 18.1 Mc/s in the third frame. Figs. 15(b) and 15(c) show fast Jupiter bursts at 18.4 Mc/s, which change shape from frame to frame. In Fig. 16 the sweep width is expanded to 0.5 Mc/s. The second frame of Fig. 16(a) has a fast Jupiter pulse at 19.1 Mc/s. Figures 16(b) and 16(c) feature a fast-changing burst at 18.75 Mc/s. Generally speaking, the major distinction of fast Jupiter spectra is the absence of the uniformly spaced Faraday fringes observed in both normal and slow Jupiter spectra.
Before leaving spectra of fast Jupiter, it should be mentioned
that apparently the sweep time of.the spectrograph's oscilloscope, which is approximately 30 milliseconds, does not limit the bandwidth of the spectra, because if this were so, during Jupiter activity the center frequency of the burst would move about the scope in random fashion, depending on the beam position at the time of reception of a Jovian burst. This is not the case, as shown by Figs. ll, 12, 13, 14, 15, and 16 where, although the Jupiter pulses are sporadic in time, they tend to center about a narrow frequency band in each sequence. In Figs. ll, 12, and 15, where all data were taken between 0125 and 0127 Eastern
Standard Time, November 5, 1963, the activity all centers near 18 Mc/s. In Figs. 14 and 15, for 0346-0347 Eastern Standard Time, September 26, 1963, the activity centers about 18.4 Mc/s; while the fast Jovian spectra of 0336 Eastern Standard Time, September 26, 1963, in Fig. 16 center mainly near 18.75 Mc/s.




In summary, Jovian spectra exhibit characteristics which can
be used to identify the time morphology of the radiation. Slow Jupiter observed thus far exhibits regular Faraday fringes with relatively smooth tops. Normal Jupiter likewise shows either Faraday fringes with more ragged appearing tops, or isolated narrow-band spikes of timevarying amplitude. Fast Jupiter does not show regular Faraday fringing and is highly irregular in pulse shape, with washed-out, ragged patches revealing the fast-changing spectral burst structure.
3. Frequency Dependence of Ionospheric Faraday
Rotation from Jovian Decametric Spectra
For many years research workers have been using a simplified model of the rotation of the plane of polarization of radio waves in a magnetoionic medium (Faraday rotation) to interpret such phenomena
as lunar radio echoes (12, 13, 14), artificial satellite signal fading [15, 16, 17, 18], and Jovian decametric spectral fringes [19]. Browne et al. [13] have shown that for the quasi-longitudinal approximation and no ray path splitting of the ordinary and extraordinary modes, the Faraday rotation of plane polarized electromagnetic waves traveling
through the ionosphere is inversely proportional to the frequency squared. Other workers have used these results in the form
K h
=7 H cos 9 sec x'f N dh (I)
f 0
where K is a constant, f is frequency, H is the strength of the terrestrial magnetic field, e is the angle between the ray and H, x is the




angle between the ray and the vertical, N is the electron density along the ray path, and joh N dh is the total electron content along the path.
Browne et al. calculated the ionospheric electron content using experimental values of 0 at approximately 120 megacycles per second for radio echoes from the moon [13]. Warwick and Dulk [19] measured from Jovian spectral data, calculated the electron content of the ionosphere and extrapolated the Faraday rotations to Jupiter to deduce the initial orientation of the Jovian radiation polarization ellipse and the nature
of the mechanism of this radiation. In each case the inverse square frequency dependence of Eq. 1 is assumed valid and some special method must be used to assure the absolute determination of C2, which through direct measurement can usually be measured only to within some undetermined integral number of rotations.
It is the purpose of this section to present a method for the determination of an unambiguous 0 directly. from Jovian spectra and to show that the frequency dependence of the Faraday rotation of decameterwavelength radio waves through the terrestrial ionosphere can be measured experimentally by using Jupiterls decametric spectra. An independent check of this result, using the signals from the beacon satellite
S-66(BE-B), is included.
Recall Fig. 3, a reproduction of a typical Jovian burst
recorded on the Florida spectrograph. The center frequency is 19.8 Mc/s.
The well-defined fringes have been attributed to Faraday rotation occurring in the Earth's ionosphere.. Warwick and Dulk [19] used Faraday fringes appearing on Jovian spectral records of a somewhat different




nature to calculate the terrestrial ionospheric electron content, assuming the validity of Eq. 1 and arguing that all of the observed Faraday rotation occurred in the ionosphere.
The Florida data are such that the frequency dependence of Faraday rotation in the inosphere can be determined experimentally; that is, the validity of Eq. 1 can be tested. For example, data from radiation emitted by Jupiter's source C ("II = 310, Warwick's "late source") between 0514 and 0526 Universal Time, October 11, 1963, can be used to perform the following calculation. Since ., the absolute number of Faraday rotations, cannot be determined directly from spectral records, the differential rotation rate with respect to frequency, was measured and plotted as a function of frequency using the least squares method. This curve appears as the solid line in Fig. 17. For comparison purposes the dashed line in Fig. 17 is the theoretical curve, having a slope of -3 as determined through differentiation of Eq. 1. Note that for this plot the dimensions of the ordinate are rotations per Mc/s. In the Florida raw data, such a: Fig. 3, each fringe represents one-half of a Faraday rotation because the signal and antenna polarizations are parallel twice per Faraday rotation. The resulting experimental curve has a slope m 2.7 and an ordinate intercept c = 3.9. The equation of the line is
log m log f + c, (2)
6f=
which can be written
log log (kf) (3)




C)J 1
7I
o5
..J
C)
>4
J
LEAST SQUARES w FIT
.SLOPE = -2.7 ~2O 1
Ld2
(D
THEORETICAL
o SLOPE =-3.0
01
0.
10 20 30 40
FREQUENCY (MC/S)
Fig. 17.--Logarithmic plot of rate of change of Faraday rotation spectral fringes as a function of frequency. One
Faraday "revolution" corresponds to 2 fringes in Fig. 5.




where k 10c = 7950, using logarithms to the base 10. It immediately follows that
7: (4)
which can be integrated from 0 to f, resulting in
- k+1 m -l. (5)
Inserting the experimentally-determined k and m produces the equation
S=7950f f--1.7
1.7 -4675 f -.7 (6)
Thus, the frequency dependence of the Faraday rotation of radio signals passing through the terrestrial ionosphere has been experimentally
determined for the period of the observations. The solid line in Fig. 18 is a plot of Eq. 6, Faraday rotation (!2) versus frequency (f).
Note that as the frequency is varied from 13 to 20 Mc/s, the Faraday rotation ranges from 60 to 29 rotations. The dashed line in Fig. 18 represents the theoretical inverse square frequency dependence and ranges from 55 rotations at 13 Mc/s to 25 rotations at 20 Mc/s.
These results are preliminary, and continued study of Jovian decametric spectra may refine the apparent inverse 1.7-power frequency dependence of ionospheric Faraday rotation for sub-20 Mc/s electromagnetic waves. The present calculation was made from data not recorded for this specific purpose and the significance of the difference between the measured index of 1.7 and the theoretical value of




54
60
FARADAY ROTATION
FREQUENCY DEPENDENCE
55 -x
"-- .. THEORETICAL EXPERIMENTAL
50
o
zS
45-
ok
40 -.
35
- 30
HX
% x
25
10 15 20
FREQUENCY (MC/S)
Fig. 18.--Terrestrial ionospheric Faraday rotation as a function of frequency for Gainesville, Florida, 0520 Universal Time, October 11, 1963.




2.0 is at present uncertain. However, the method and the results are of current interest. Additional data taken for radiation traveling through the ionosphere at different angles with respect to the Earthts magnetic field will determine the range of validity of the quasilongitudinal approximation in the derivation of Eq. 1. Future Jovian spectral data will be gathered with these specific goals in mind.
As an independent check of the preceding results, the 20.005 and 40.010 Mc/s beacons aboard S-66(BE-B), an artificial satellite orbiting outside the ionosphere, have been utilized in a comparison determination of the frequency dependence of ionospheric Faraday rotation for this region of the spectrum. The form of the frequency dependence is expected to be
= G fm (7)
where G is a parameter involving various quantities noted in Eq. 1, and m is again a constant to be determined experimentally. Differentiation of Eq. 7 with respect to the time, t, yields
Tt f-mTG (8)
where Q may be approximated as the ratio L and measured from simulayt b At
taneous oscillograph recordings of the Faraday rotation modulated signals of the 20.005 and 40,010 Mc/s beacons of S-66. Figure 19 shows such a record of data taken at the University of Florida on November 21, 1964. Measuring (At)20 and (At)40 as the times for A = 1/2 Faraday rotation for the 20.005 and 40.010 Mc/s signals, respectively,




Fig. 19.--Comparison of S-66 signals at 20 and 40 Mc/s.




Eq. 8 becomes

1(t)2 2.05'
2(At )20 T
Rewriting Eq. 9 for the 40.010 1c/s case and dividing the two results in
(At)40 )40.010 '
(At20 20.005 =2.(10)
Therefore
log[
M = ( 11)
log 2
Preliminary determinations of m using this technique with the data shown in Fig. 19 give m = 2.08 in agreement with the theoretical value of m = 2 used in Eq. 1, within the accuracy of the measurements. The difference between this value and the value of 1.7 as calculated from the Jovian spectra may be a real effect, since the ionospheric critical frequency lies just below the range of the spectral data, and certainly, for frequencies lower than the critical frequency, there is no transmission, hence no Faraday rotation, hence no frequency dependence. It is plausible that the theoretical inverse square frequency dependence is valid for frequencies above approximately 20 Mc/s, but that for the range from 20 Mc/s down to the critical frequency a more rigorous theory is required for purposes such as the precise calculation of ionospheric electron content.




4. Comparison with Other Workers' Data

Other than the Florida spectral data discussed in this work and Chatterton's dissertation [3], the only comprehensive treatment of Jupiterts decametric spectra is that of Warwick of the High Altitude Observatory at Boulder, Colorado [19, 20]. A typical example of the Boulder data appears as Fig. 20, where the ordinate is frequency and the abscissa is Universal Time. This storm occurred on September 13, 1963, and the instrument sweeps from 7.6 to 41 Mc/s every 1.3 seconds. The diagonal white streaks washing out strips of data approximately 0.5 Mc/s wide are interference fringes due to the antenna system. The faint alternate black and gray horizontal bands are caused by Faraday rotation fading and correspond to the Faraday fringes observed in the Florida spectra shown and analyzed above in Secs. 11.2 and 11.3. At times the storm shown in Fig. 20 appears to extend from 13 to 36 Mc/s with a bandwidth of some 23 Mc/s. Other isolated bursts have bandwidths of less than 1 Mc/s. Compare Fig. 20 with Fig. 9, shown previously as an example of the Florida data. The time of the Florida data is 0203 Eastern Standard Time, which is 0703 Universal Time, and the Florida record thus represents an enlarged view of the faint speck appearing at 18 Mc/s and 0703 U.T. in Fig. 20.
The Florida spectrograph is obviously superior for recording fine structure characteristics of less than 1 Mc/s bandwidth, but it cannot record the features involving bandwidths exceeding 4 Mc/s. This comparison, then, is limited to fine structure effects of extent less than 4 Mc/s. Specifically, Warwick claims to have recognized




2 14
16I618-1
I
20
242832-" 3640-

0630

0640
0640

0650

0700

Fig. 20.--Jovian decametric spectra as recorded at the High Altitude Observatory at Boulder, Colorado [19].

0710




events of bandwidths as low as 0.2 Mc/s [211, although these bursts are extremely difficult to recognize in his published data. Now, bandwidths of this order are easily recognized in the Florida data. In
Fig. 9, for example, the moving pulse near 18 Mc/s has a bandwidth of approximately 0.05 Mc/s, roughly an order of magnitude less than the
smallest reported by the Boulder group. Thus far Florida workers have not recognized any Jovian spectral spikes of bandwidth lower than 0.05 Mc/s. The rare narrow spikes such as were seen in Fig. 6 have never been filmed with an expanded sweep, and from data such as Fig. 6 it is
determined that this type of normal Jupiter has a bandwidth of somewhat less than 0.1 Mc/s.
The only other specific statement which Warwick makes concerning the spectral bandwidth is that a rough positive correlation exists
between pulse duration and spectral bandwidth. Examination of the Florida data shows that this is not always the case. For example, normal Jupiter activity exhibiting bandwidths ranging from 0.05 to 3 or 4 Mc/s has been observed at Gainesville. Fast Jupiter has ranged in bandwidth from 0.05 to 1.5 Mc/s. Indeed, the storm shown in Fig. 20 has at times a bandwidth of 23 Mc/s. This activity was also recorded on the Florida spectrograph and consisted primarily of fast Jupiter. The Boulder equipment was apparently too insensitive to fine structure fluctuations in time to show this characteristic of the storm.
Riihimaa of the University of Helsinki, Finland, is studying fine structure of the Jovian decametric spectra using an instrument which produces data similar to Warwickts, but of much higher resolution.




In a brief preliminary report appearing in a letter to Nature [22], Riihimaa states that the Finnish spectrograph sweeps the range 18 to 20 Mc/s with a repetition frequency of 10 c/s. The resulting data lack the overall bandwidth of the Boulder data, but they do resolve the fine structure details, though not as well as the Florida data. Riihimaa shows typical Jovian spectra recorded by the Finnish spectrograph but postpones conclusions to a detailed analysis of his results which is in preparation. Thus, at this writing there is no basis for a general comparison of the Finnish data with those of Boulder and Florida.
Apparently workers in the realm of Jovian decametric spectra
require equipment capable of presenting the best of the spectral characteristics seen in the Boulder, Florida, and Finnish data before the nature of these spectra can be completely understood. The Boulder spectograph sacrifices fine structure detail to present excellent broad-band characteristics of the spectra in time. The Florida data excel in recording fine structure details, but lose the broad-band characteristics. The Finnish data seem to be superior for systematically recording the slow changes in time of spectral fine structure at the expense of the loss of some amplitude versus frequency detail.




5. Spurious Spectra
The observer of Jovian spectra that are displayed on the
University of Florida radio spectrograph must be aware of the characteristics of the many spurious spectra which frequently appear in order that the desired data be correctly interpreted. There follows a description of these worrisome phenomena.
Among the superfluous spectra produced, lightning interference is the only type having natural origin. It is easily identified by correlating the sharp static sounds from a properly tuned receiver with the sudden rises and drops of the entire spectral trace on the oscilloscope. The trace sometimes shows discontinuities at points where the lightning crack happens to catch the sweeping beam, and these points are located at random frequencies. This characteristic aids in the distinguishing of spitting Jupiter from lightning since, as was noted in Sec. 11.2, the Jupiter spitting bursts tend to localize at particular frequencies for many successive sweeps of the electron beam
of the analyzer. Strong Jupiter storms may be recognized through intermittent lightning interference.
A particularly bothersome effect has been that called in the Florida Radio Observatory Log (February 10, 1964), "Strange Interference,'t "Strange X," and "Blow-Up Phenomenon," deriving these names from the strange quality of producing no audio noise in the observatory receivers (Log, December 20-21, 1963), yet appearing more intense than most Jupiter bursts observed on the analyzer oscilloscope. At first this signal was thought to be due to internal equipment




malfunction, but this assumption proved erroneous. Another panoramic spectrum analyzer was set up to extend the range of our normal equipment (Log, February 10, 1964). The second system displayed the spectrum from 24 to 34 Nc/s. The intermittent strange interference was found to be present simultaneously in both systems, and furthermore cut off abruptly at approximately 30 Nc/s. Moreover, the interference was found to be sensitive to antenna setting, peaking for a logperiodic antenna setting of 550o West (Log, December 20-21, 1963). (The antenna was set for declination -100). Additionally, disconnection of the antenna always stopped the interference (Log, January 10, 1964). The conclusion was that this interference probably originated at the University of Florida Medical Center, entering our spectrograph through the back lobe of thie log-periodic antenna. The appearance of the signal's spectrum is shown in Fig. 21(a) and (b), where two different characteristics can be seen. In Fig. 21(a) the trace resembles a modulated, distorted sinusoid, while in Fig. 21(b) a square wave envelope is more evident. This type of interference, when present, blocks our spectral equipment over its entire bandwidth, rendering Jovian spectral analysis impossible, since the entire trace jumps on and off the screen and changes rapidly between the two types shown in Fig. 21(a) and (b), sometimes more frequently than once per second.
Another category of interference has been labeled as "buzz"? by
the observers (Log, October 14 -15, .1963),, and is shown in Fig. 21(c) and (d). While this spe ctrum, closely res embles the Faraday rotation feature of the Jovian spectra previously discussed in Sec. 1=.3, there




is no possibility of mistaking its regular audio buzz (somewhat like the sound a small electric motor with brushes produces in an AC-DC
radio) for the randomly-spaced Jupiter swishes or spits. During periods of buzz interference spectra of Jupiter storms cannot be satisfactorily analyzed.
Occasionally a swept-frequency jammer (Log, June 14, 1963) and (June 15, 1963), blocks the spectrum analyzer as shown in Fig. 21(e) and (f).' The audio of this particular type of interference sometimes appears on separate channels as a short buzz while the square (in frequency) wave front sweeps back and forth across the oscilloscope face. There should be no chance of confusing these short bursts of buzz with
the random Jovian bursts. At other times the jamimer completely blocks one or more channels simultaneously. In general, Jovian spectra observed under jamming conditions are unintelligible.
Previous to the identification of the Faraday rotation fringes
in Jovian spectra it was feared that the regular peaking observed in the spectra might be due to periodic gain irregularities [3].
Chatterton correctly believed this not to be the case [311. In an effort to settle this question, a study was made in which continuous
adpulsed noise was fed into the Pan system at various points (L~og, December 7-8, 1963). Figure 22 shows the results of this experiment. Figure 22(a) is the spectrum of the galactic background incident on
the log-periodic antenna. Periodic gain fluctuations are not evident under these conditions. For Fig. 22(b) the antenna and preamplifier were disconnected (see Fig. 1, the system block diagram) and broad-band




20 21.5 23 20.5 22 23.5
(Mc/s) (Mc/s)

(a) Strange
Interference

(b) Strange
Interference

I I I
20.5 22 23.5 (Mc/s)
(c) Buzz

J i 1 1 Ia l s i I
17.5 18 18.5 16.5 18 19.5
(Mc/s) (Mc/s)

(d) Buzz

(e) Jammer

20 21.5 23 (M c/s)
(f) Jammer

Fig. 21.--Various types of spectral interference
observed with the University of Florida radio spectrograph.




noise generated in a type 5722 noise diode was introduced directly into the spectrum analyzer. With maximum sensitivity, the noise scarcely broadened the oscilloscope trace. Figure 22(c), (d), and
(e) show the results of introducing unmodulated noise into the preamplifier with the antenna disconnected, with a center frequency of 18 Mc/s and different sweep widths of 5, 1, and 0.1 Mc/s, respectively.
The spectra all look the same and exhibit no fringing or other features of Jovian spectra. With different center frequencies over the range of the spectrograph the results were the same. Figure 22(f), (g), (h), and
(i) show the results of introducing modulated broad-band noise from the type 5722 diode into the preamplifier as before, with the antenna disconnected. The steep-front modulation was produced by mechanically switching the noise generator on and off several times per second in hopes of not observing any internal ringing subject to shock excitation. The noise spectra appear on only a portion of each trace because the "on" time of the noise generator was less than the cathode ray tube sweep time (approximately 1/30 second). Again at the various center frequencies and sweep widths available to the system, no features of Jovian spectra were synthesized through shock excitation and no spurious spectral features appeared.
It must be pointed out that for a time the spectrum analyzer was out of alignment and occasionally an internal oscillation would produce a spurious peak on the -oscilloscope. The effect was random and intermittent. Such a peak is the left-most peak in Fig. 21(c) at approximately 20.6 Mc/s. Note the characteristic skip in the base




I I ,a I I a I 1 I
16.5 18 19.5 16.5 18 19.5
(a) Background (b) Noise, No
Preamplifier

I I I lt l
175 18 18.5 1795 18 18.05

(d) Noise

(e) Noise

I I a 9. 16.5 18 19.5

(c) Noise

I I p I
16.5 18 19.5
(f) Pulsed Noise

17.5 18 18.5 20.5 22 23.5

(g) Pulsed Noise

(h) Pulsed Noise

21.5 22 22.5
(i)Pulsed Noise

Fig. 22.--Spectral response of the University of Florida radio spectrograph to broad-band noise.




line, which is higher to the right of the break than to the left. While it is not obvious in Fig. 21(c), the leading edge of this peak
is characteristically almost vertical and the trailing edge on the right markedly less steep. An experienced ,-bsarver can easily reconize this troub Moreover,, realignment of the equipment has removed
this source of confusion and routine periodic maintena-.Ice should prevent its recurrence.
Radio stations' spectral spikes are another class of spurious signals appearing in our spectra. These are usually identifiable by their narrow width. A good example of a spectral line due to a radio station appears at 18.5 Nc/s in Fig. 6. Notice the narrowness of the station spike in contrast to the pair of Jupiter peaks at 16.75 and 18.0 Xvc/s, recalling that these two peaks are representative of the slimmest Jupiter spectral peaks identified at the University of Florida Radio Observatory (Sec. 11.2). Thus most radio stations' spectra are recognized on the basis of bandwidth alone. In cases of doubt the observer should set a receiver on the exact frequency of a spike, using a grid dip meter, listening for Jovian or station audio characteristics from which to judge the origin of the spectral peak.
In summary, all the characteristics observed presently on the University of Florida radio spectrograph are attributed to external sources. The characteristics of the Jovian spectra are documented in Sec. 11.2. The spurious spectra which completely disrupt observations are strong lightning, closely-spaced stations, the inaudible strange interference, the swept-frequency jammer, and strong buzz. Troublesome




49
effects which partially impair observations but can be recognized by an experienced observer include mild atmospherics due to distant lightning, occasional mild radio station activity., and an internal ringing due to misalignment of the spectrum analyzer.




CHAPTER III

MORPHOLOGY AND ATMOSPHERIC MODIFICATION
1. The Search for a Jovian Continuum
The sporadic, bursty nature of decameter wavelength Jovian
radio emissions has been described above in Sec. 11.2, and of course, by other research workers. The question remains whether there exists a continuum of radiation of flux density lower than that of the normal storms. In an attempt to resolve this question, a 22.2 Mc/s phaseswitching interferometer was established at Bivents Bank in 1961 [23, 24, 251. This observation channel, called 22B, should be capable of detecting flux densities as low as 1 x 10-23 w/m 2/cps [25]. The instrument is steerable in declination only, and hence is effective only within a period of approximately one and one-half hours before and after meridian transit of the celestial object under observation. The theory and operation of the phase-switching interferometer have been treated elsewhere [23, 24, 25, 26, 273, and will not be repeated here, where the interpretation of the Florida data is discussed.
Watson observed radio sources Cygnus A and Cassiopeia A using the 22B antenna and plotted smoothed curves [24] of the response of the system to these two sources at declinations 400 361 and 580 32', respectively. The maximum fringe occurs at the time of antenna boresight transit, approximately 10 minutes after meridian transit as




explained by Watson [24]. The fringe spacing near transit is inversely proportional to the cosine of the declination of the source producing the fringes [27]. Therefore, the fringes due to a source of known declination may be extrapolated to produce an expected fringe pattern for a source of different declination. The fringe pattern for Cygnus A in Fig. 19 of Watson's thesis [24] was contracted.in time, point-bypoint, in the ratio of the cosine of the declination of Cygnus A to the cosine of Jupiterts declination to produce the predicted response of 22B to a radio source at Jupiter. Such a contracted fringe appears as the smooth curve at the bottom of Fig. 23. This is the predicted response of 22B for a 22.2 Mc/s source at declination 4.10, Jupiter's declination on September 26, 1963. Similar fringes were constructed as required by the varying declination of Jupiter and compared with all of the 22B records made to date to determine whether or not fringes due to continuous radiation from Jupiter were present. The results of this analysis follow.
Channel 22B was operated extensively during the 1961 apparition of Jupiter. Effective watch was held 154 nights, the equipment being out of order on 36 nights when regular watch was observed. During the effective observations on 22B, Jupiter storms of 22.2 Mc/s were observed 31 times within one and one-half hours of transit on the observatory's other 22.2 Mc/s equipment. Thirteen of these storms were recognized on 22B and consisted solely of sporadic bursts. Never during these 31 storms, or at any other time during the apparition, was there any evidence on the 22B record of fringes such as would be




F0300o

20200 JT 0190En-s C
......22 Y 15 m in -- -

-- J0200 0100 -4
pp ~JUPITER. s

UNIVERSITY of FLORIDA UNIVERSITY of FLORIDA

26SEPT. 1963 PRINCIPAL LOBE

PREDICTED RESPONSE of 22B for SOURCE at DECLINATION 4.10

Fig. 23.--A 22.2 Mc/s Jupiter storm of mixed normal swishes and continuum.




produced by a 22.2 Mc/s radio source at Jupiter, whose declination ranged from 180 to 210 South during the apparition [28]. During each of the 154 effective watches, fringes from Cassiopeia A were observed despite the antenna's being directed toward declination 150 South. (Recall that Cassiopeia A's declination is 580 32t North [26]). The flux density of Cassiopeia A at 22.2 Mc/s is approximately 4.6 x 10-22 w/m 2/cps [25].
For the 1962 apparition the phase-switching circuitry was used for another experiment and no 22B data were recorded.
Channel 22B was used effectively during 67 watches of the 1963 apparition. During an additional 67 watches interspersed throughout the apparition, the equipment was turned on but was not functioning properly. In eight instances fringes due to radio sources other than Jupiter were observed and identified. During'ten of the effective watches Jupiter storms at 22.2 Mc/s were observed on other 22.2 Mc/s equipment. Seven of these storms were observed as sporadic'bursts on 22B, with some tendency toward fringing. However, in three instances the 22B record clearly'exhibited fringes as predicted for continuous 22.2 Mc/s Jovian radiation.
Figures 23, 24, and 25 contain typical examples of 22B data with other 22.2 Mc/s records for comparison. In Fig. 23, utilizing data for September 26, 1963, the upper record shows Jupiter activity at 22.2 Mc/s on channel 22Y, which features an automatic tracking Yagi antenna. The center record is data taken simultaneously on 22B at 22.2 Mc/s. Note between 0007 and 0235 E.S.T. the regular fringes which




compare favorably in spacing with those of the lower trace, the predicted response of 22B to a radio continuum at Jupiter. The comparison is made using time correlations of maxima and minima. Superposition of a tracing of the lower curve on the 22B record in the center shows near coincidence of the maxima and minima, especially near antenna boresight transit, which occurs later than meridian transit as noted previously. For some reason the westward offset of the 22B central lobe from the meridian now measures 0.60 in contrast to Watson's measurement of 1.9.0 This change is probably due to electrical path length changes introduced between the receiver and the East and/or West arrays of 22B during system modifications made since 1960, when Watson's data were taken.
Apparently, in addition to the normal swishes recorded on 22Y during this intense Jupiter storm, there was emitted continuous radiation at 22.2 Mc/s as shown by the fringes on the 22B record. The time constant of 22B was set at 10 seconds for most of the storm, and while no high speed oscillograph record was made at 22.2 Mc/s, the simultaneous Brush records at other frequencies fail to show pulses of this length. Additionally, when the time constant of 22B was set at 20 seconds between 0046 and 0052 E.S.T., the fringe pattern continued, despite the fact that this time was very close to the time for a null in the fringe pattern when the instrument was relatively insensitive. Final evidence of this continuous 22.2 Mc/s radiation is two periods during the storm when the swishes on22Y stopped but the fringes of 22B continued. In Fig. 23 note the two lulls in the storm on 22Y at




It?), I ~ I I
I5mirL-4

10000

~Ji-~j~4G0QO
()

.1

UNIVERSITY of FLORIDA
28 SEPT. 1963

Fig. 24.--A 22.2 Mc/s Jupiter storm of swishes with no continuum.




'2300 EST

"A" TRANSIT
L L

Fig. 25.--A 22.2 Mc/s Jupiter storm superimposed on the Cassiopeia A continuum.

10000

'2200




0026 to 0033 E.S.T. and 0133 to 0152 E.S.T., while the fringe pattern on 22B seems to continue during these periods, although perhaps reduced in intensity. Therefore, Fig. 23 apparently displays a Jupiter storm of combined sporadic and continuous 22.2 Mc/s radiation. Douglas and
Smith have suggested the existence of this continuum [29].
Figure 24 shows an instance of strong sporadic Jovian activity on 22Y occurring approximately 1 hour after transit during the watch of September 28, 1963. Again at the top of the figure is the 22Y record, at the center is 22B, and at the bottom the predicted fringes for Jovian radiation. Note that the storm of September 26, 1963, shown in Fig, 23 produced fringes detected 1 hour 20 minutes before and after
transit. If continuous Jupiter radiation within the sensitivity of 22B had been present along with the normal Jovian swishes on September 28, 1965, Fig. 24 should show signs of fringes from approximately 0200 to 0230 on the center trace, which is the 22B record. No such fringes are evident. The conclusion is that no continuous 22.2 Mc/s flux from Jupiter exceeded 1 x 10-2 w/m2 cps during this storm, at least while Jupiter was in the beam of 22B.
Figure 25 also shows a Jupiter storm observed on 22Y which
apparently did not have an accompanying continuum, the records having been produced October 21-22, 1963. The trace at the bottom of the figure is the response of 22B to Cassiopeia A, after Watson [24]. The top record is a Jupiter storm observed on 22Y. The center record shows fringes observed on 22B. Note the good match of these fringes to the known response of 22B to Cassiopeia A. While sporadic Jupiter bursts




appeared on 22B,, there is no evidence of fringes corresponding to the pattern predicted for radiation from Jupiter. Specifically, the group of four intense but separated peaks beginning on the 22B record at 2335 E.S.T. should form a smooth maximum, if the radiation is continuous in time. Clearly this is not the case, the high peaks being spaced at intervals of approximately 1 minute. The absence of the 22.2 Mvc/s continuum is suggested by this data, the fringe due to Cassiopeia A being clearly evident despite the strong simultaneous Jovian storm.
Unfortunately, most of the inoperative periods of 22B came near the end of the 1963 apparition and indicated the aging of the equipment. The system would not function a t all for 16 trial watches
during the 1964 apparition. The circuitry had become highly sensitive to temperature and power fluctuations. The Florida sun,, humidity, and an occasional stampede of Brahma cattle had taken their toll of the antenna masts and fittings and 22B was shut down for repairs. The antenna is being rebuilt and a new, more stable phase-switching circuit is planned for future observations with the system. This is necessary since our competent engineer, Gr. F. Walls, had learned to maintain 22B in operative condition only at the cost of daily attention.
The data gathered from 22B support the conclusion that Jupiter does not continuously emit 22.2 Nc/s radiation exceeding 1 x 10C23 w/m 2/cps in flux density. The three instances of a Jovian continuum accompanying strong sporadic Jupiter activity, however,, suggest that Jupiter storms at times consist of a combination of sporadic bursts with an underlying continuum. This result confirms and amplifies the




findings of Stone, Alexander, and Erickson [30], who used data from the University of Maryland 26.3 Mc/s Christiansen-type array with a Ryle-Vonberg radiometer to conclude that Jupiter radio storms consist of at least two distinct components, characterized by different flux densities. At 26.3 Mc/s these flux densities were "just above" 10-23 w/m2/cps and 9 x 10-25 w/m2/cps, the stronger component corresponding to the radiation commonly observed by most other workers. At the time of their publication Stone et al. had not considered any fine structure details. It appears that the presently reported underlying continuum is in fact the weaker component reported by Stone et al. and that this component of Jove's radio emissions either is a continuum or is composed of bursts longer than the order of 10 seconds, the time constant of the Florida interferometer when the pertinent data were taken. Perhaps the characteristics of the two components are not resolved in the Maryland data due to the relatively short integration time used, namely, 2 seconds. The sensitivity of the Maryland system is 5 x 10-24 w/m 2/cps, a factor of 2 better than Florida's 10-23 w/m2/cps. Florida's 22B obviously is marginal for further examination of the weaker component, and the sensitivity of 22B should be improved in order that this study be pursued with promise. Then of course, as Stone et al. suggest, the experiment should be extended to other frequencies.




2. Jovian Burst Morphology
a. Introduction
Kraus [31], Gallet [32], Gardner and Shain [33], A. G. Smith et al. [34], H. J. Smith et al. [35, 36], and Mock and A. G. Smith [11], have all studied the fine structure of the Jovian storms. Generally there is agreement that the terrestrial ionosphere modulates the Jovian signals and is in part responsible for some of their features
as observed in Earth. In particular, A. G. Smith et al. [34] concluded that the overall burstiness of the radiation is in fact a scintillation caused by fluctuations occurring in the ionosphere. Then Mock and Smith [11], through an extensive statistical analysis of the durations of normal Jovian pulses received simultaneously in Florida and Chile, concluded that the pulses constituting the fine structure of the bursts came from the same population and that, therefore, this feature of the
fine structure either originated in the Jovian environment or was due to a world-wide ionospheric effect of some indeterminate nature. Less complete data suggested this same conclusion for the slow and fast Jupiter pulses.
In this section is reported a study, the purpose of which is
to determine the origin of the basic Jovian pulse structure, apparently a choice among the terrestrial ionosphere, the interplanetary medium,
and the Jovian environment. The primary method of the analysis is to examine the variation of the character of the pulses with parameters such as time of year, time of day, frequency, altitude of the planet, and geomagnetic activity. IXow, radio star scintillations are apparently




ionospheric in origin, and their behavior in terms of the cited variables has been examined extensively by others [26, 37, 36, 39, 40,141].
Similar scintillations of artificial satellite signals are well documented [42]. It is expected that if the basic Jovian pulse character is an ionospheric scintillation, this phenomenon will exhibit the same parametric variations as radio star and satellite scintillations. Lack of -such correlation will indicate some other source for this fine structure, probably the Jovian environment, the alternative proposed in Mock's thesis [11]. The comparison of Jupiter's basic pulse structure and ionospheric scintillation phenomena follows.
b. The Jupiter Pulse Character Index
The discussion is facilitated by the definition of the Jupiter
pulse character index C. This index is described in Table 1(a). For storms consisting of slow Jupiter pulses lasting a few seconds or longer, C is defined as 1. Normal Jupiter pulses of approximately
1 second duration are assigned the index value C = 2. Fast Jupiter pulses of 0.1 second or less duration have the value C = 3. Reference to Fig. 2 of Sec. 11.2, along with Table 1 of this section, should
provide the reader with a suitable interpretation of C. In the actual record reduction, nonintegral values of C were used for cases of unclear average pulse length. For example, sometimes normal Jupiter has fast Jupiter superimposed. The record reducer uses his judgment in assigning C for these cases the value 2 1/4, 2 1/2, or 2 3/4, depending on the relative amount of each type of pulse that is present. The possibility of confusing normal Jupiter (C = 2), with mixed slow




TABLE 1
DEFINITIONS OF PULSE CHARACTER AND SC INTILLATION INDICES
(a) Jupiter Pulse Character Index C, Value of C Description
1 Slow Jupiter pulses a few seconds or longer
2 Normal Jupiter pulses approximately 1 second
5 Fast Jupiter pulses 0.1 second or less
(b) Satellite Signal Scintillation Index S Value of S Description
0 No irregular scintillation. Faraday rotation fading
is regular.
0.5 Scintillations do not exceed 50 per cent modulation.
Faraday rotation fading is regular.
1.0 Scintillations exceed So per cent modulation. Faraday
rotation fading clearly evident.
1.5 Some evidence of regular fading. Scintillations
approach 100 per dent modulation.
2.0 Extreme scintillations. No regular fading observable.




(C = 1), and fast (C = 3) pulses is slight. In all the data examined, slow and fast Jupiter were only rarely recognized to occur simultaneously.
Other workers have devised scintillation indices for radio star and satellite signals. Those cited will not be redefined here. It will simply be noted that the scintillation index in each cited example ranges from 0 for no scintillations to some arbitrary positive
value for maximum observed scintillations. Lawrence and Martin's [42] scintillation index for satellite signals is used later in Sec. 111.3., and is described in Table l(b).
c. Diurnal Variations
Radio star and satellite signal scintillations are significantly
more pronounced near local midnight than at other times of day. [37, 42]. For example, Fig. 26 shows the typical diurnal variations observed in these scintillations. The upper plot, Fig. 26(a), shows Hewish's scintillation index versus local time for observations of radio stars [37].
The relatively sharp maximum near midnight, falling off rapidly towards sunrise and sunset, is indeed striking. Less striking, but present, is the maximum near local midnight in-Fig. 26(b), a plot of Lawrence and Martin's satellite scintillation index versus local time [42]. Inspection of the next four figures, Figs. 27, 28, 29, and 30, reveals no
such maxima. Each of these figures is a plot of the Jovian pulse character (defined above) versus local time, as observed at 18 Mc/s on the high-speed oscillograph records taken at Gainesville, Florida, and Maipu, Chile. Fig. 27, representing the 18 Mc/s data from Florida for




14 18 22 2 6 10 14
LOCAL TIME (H)

(a) RADIO

STARS

I I I

LOCAL. TIME (H)

(b) SATELLITE

SIGNALS

Fig. 26.--The diurnal variation of ionospheric scintillations.

z
0 j=.05
5~O ..J
-J
C,




FLORIDA 1962 18 MC/S
w
C
0
ILi
w
18 20 22 24 2 4 6 8
LOCAL TIME (H)
Fig. 27.--The diurnal variation of Jovian burst fine structure (Florida, 1962).




3 -

21-

18 20 22 24 2 4 6
LOCAL TIME (H)
Fig. 28.--The diurnal variation of Jovian burst fine structure (Chile, 1962).

CHILE 1962
18 MC/S

I

, !

!

!




FLORIDA 1963 18 MC/S
2_ -- _- - ,_ - --

! t

20

22

24 2
LOCAL TIME (H)

Fig. 29.--The diurnal variation of Jovian burst fine structure (Florida, 1963).

I
I I

!




CHILE 1963
18 MC/S
F ---'-L--,L _

-I -

24 2
LOCAL TIME (H)

Fig. 50.--The diurnal variation of Jovian burst fine structure (Chile, 1965).

3r-

2F

, I .I

20

22

. I

I

.I

I




the entire 1962 apparition, has a marked minimum near midnight. Figure 28, the 18 Mc/s data taken in Chile in 1962, shows some tendency towards a minimum shortly after midnight, at 0200 local time. Figure 29 shows a broad weak minimum near midnight, this data representing the 18 Mc/s Jupiter storms recorded at Florida during the 1963 apparition. Figure 30, 18 Mc/s, Chile, 1963, again shows some tendency towards a minimum at 0200.
Now, admittedly, the tendencies toward minima were perhaps
inconclusive in these figures, but there certainly was no evidence of a maximum as observed by Lawrence and Martin in Fig. 26(b); and certainly nothing in these Jovian pulse character plots corresponds to the sharp increase in radio star scintillations near local midnight in Fig. 26(a). The data represented in Figs. 27, 28, 29, and 30 are, therefore, interpreted as evidence that the Jovian pulse character is not an ionospheric scintillation phenomenon. d. Seasonal Variations
The comparison drawn in this section parallels the preceding discussion. Figure 31(a) shows the seasonal variation of radio star scintillations as observed by Koster and Wright [39]. Note that maximum scintillation activity occurs near the autumnal equinox, while a lesser peak occurs near the vernal equinox. Lawrence and Martin observed the autumnal equinox peak in satellite scintillations, though not the springtime peak [42]. Figures 32, 33, 34, and 35 are seasonal plots of the Jovian pulse character index of the 18 Mc/s data for the respective apparitions Florida 1962, Chile 1962, Florida 1963, and




J F MIAso'Jl'A'SOi'NoiJt
(a) The Seasonal Variation of Scintillation

0
90

700 500 300
Altitude of the Source

(b) Altitude Variation of Scintillation
Fig. 31.--The seasonal and altitude variations of radio star scintillations.

70

.10
X

C
0
O
0
(,)

.05 r




FLORIDA
I 8

40
CC
0
I
I
0
wf2 I
> ~---OPPOSIT! ON
MAR MAY JULY SEP NOV JAN
Fig. 32.--The seasonal variation of Jovian burst fine structure (Florida, 1962).

1962 NMC/ S




CHILE 1962
18 MC/S

2_

---OPPOSITION

MAY

JUL

T --r----r 1-- I i --I F I

SEP

NOV

JAN

Fig. 35.--The seasonal variation of Jovian burst fine structure (Chile, 1962).

MAR




FLORIDA 1963
18 MC/S

I-OPPOSTON
I
I
I
1
----OPPOSITI ON

___ I

-[--------7 -I- I I T I I I I I I

JUL

SEP

NOV

JAN -

Fig. 54.--The seasonal variation of Jovian burst fine structure (Florida, 1963).

3r

MAR

MAY




CHILE 1963
18 MC/S

21-

MAY

JUL

-OPPOSITION" II

SEP

NOV

Fig. 35.--The seasonal variation of Jovian burst fine structure (Chile, 1965).

MAR

JAN




Chile 1963. None of these plots exhibits the anticipated vernal equinox peak. In fact, the only distinctive feature common to even three
of the four plots is the November or December peak in Figs. 32, 33, and 34. Here is more evidence that, since Jovian pulse character does not exhibit the seasonal variations of radio star and satellite scintillations, Jupiter's burst fine structure is not a scintillation phenomenon.
e. Frequency Dependence
Radio star and satellite scintillation exhibits an inverse
relationship with frequency as shown, for example, by Aarons [411 and Briggs and Parkin [361. Figure 36, after Aarons [41], shows this relationship for scintillation on signals from Cygnus A. Note in Fig. 36 the increased scintillations as frequency decreases from 1200 to 108 Nc/s. The Florida and Chile data for the 1962 and 1963 apparitions, plotted as average pulse character versus frequency in Fig. 37, do not follow such a relationship. In fact, from 15 Nc/s to 27.6 Nc/s the index actually increases, instead of decreasing as scintillation phenomena do. Once again, there is evidence that the Jovian burst fine structure is not a scintillation phenomenon. The Florida and Chile 1962 and 1963 data, plotted separately, yield curves similar in character to Fig. 37, and are not shown here.
f. Altitude Dependence
Hewish [37] showed that radio star scintillations increase with decreasing altitude of the source. Satellite scintillations behave similarly, as implied by Lawrence and Martin [421]. But once again,




2
C)
z
22 MC ZERO RESET
400 MC
- 12 00 N4C
2315 2310 2305
EST
Fig. 56.--Frequency dependence of radio star scintillations [42].




3r-

FLORIDA-CHILE

1962-1963

N1

2F

FREQUENCY

30

20
(Mc/s)

Fig. 37.--Frequency dependence of Jovian burst fine structure.




Jovian pulse character does not exhibit this behavior. Figure 31(b)
is a plot, due to Hewish [37], of radio star scintillations versus altitude. Figure 38 is the companion plot of Jovian pulse character
versus altitude for the 18 Mc/s Florida data of 1962. Clearly, the Jupiter fine structure once again fails to match a known characteristic of the scintillation phenomenon.
g. Geomagnetic Activity Dependence
Hewish [371 and Brown and Lovell [26] cite the correlation
between radio star scintillations and geomagnetic activity. In times of increased geomagnetic activity, stronger scintillations are observed. Figure 39 is a plot of Jovian pulse character versus the Geomagnetic Activity Index K The pulse character actually appears to decrease
P
with increasing geomagnetic activity which is additional evidence that the pulse character is not an ionospheric scintillation. The dotted portion of the curve represents only 2 per cent of the data and is considered to be unreliable.
h. Solar Activity Correlation
Previously the Florida group and others [4, 5, 8] have noted some correlation between Jovian decametric emission and solar or geomagnetic activity. N. F. Six [4] and G. R. Lebo [5] have used a computer to examine the correlation between Jovian activity and the parameters FX1, FX2, FX35 FNl, FN2, FN3, SSN, Ap, and the solar 2800 Mc/s flux density. The FX's are solar flare activity indices for specific regions of the Sun [4]. The FN's are solar flare numbers for those




FLORIDA

18 MC/S

21

I0

20

300

ALTITUDE

OF JUPITER

Fig. 58.--Altitude dependence of Jovian burst fine structure.

I I

I I

40

50

1962

, I

I I




3r

w
!
0
i:
w
(I)
-j2
0
w
I(9
w3

FLORIDA 1962
18 MC/S

I
I
8
S
3
I
I
I
I
I

0 I 2 3 4 5 6
GEOMAGNETIC ACTIVITY Ii')EX Kp
Fig. 59.--Geomagnetic activity influence on Jovian burst fine structure.




same regions [4]. SSN is the Zurich sunspot number, and A is a standp
*ard geomagnetic activity index. G. R. Lebo modified the solar flare
program so that the Jovian pulse character index C could be examined for correlation with these same parameters. The computer performs
a Chree analysis of each of these variables, and of C, itself. Figures 40 and 41 present the results of this study for data of the 1962 apparition as recorded in Florida and Chile. Each separate curve in these
two figures represents the output of the computer, programmed for Chree analysis, and shows how the variable under inspection behaves, on the average, during the period 35 days before to 4 days after Jovian activity of character index C. The average is taken over the total number
of days of Jovian activity having a particular index C.
Figure 40 shows the results of the Chree analysis of sunspot number. The five curves represent average sunspot number as a function of days relative to Jovian activity for various C's and for the two stations listed. There were insufficient data for a meaningful study of C = 1 for Chile in 1962. The most interesting feature of this set of curves is the simultaneous appearance of strong peaks in
the two curves for C = 5, Florida and Chile, 1962, at -6 and -25 days. Similar simultaneous minor peaks occur at -51 days. Note the absence of these peaks in the two curves for C = 2. The data for C = 2 do show peaking tendencies, but not nearly so strongly as the C = 3 curves. Likewise the single plot for C = 1 exhibits peaks at 0,
-13, and -28 days. These data are much more sparse than those for the other curves, and no significance is attached to these peaks, especially in view of the height of the 0-day peak with respect to the others.




-29 -21 -13 -5 0 +3
DAYS RELATIVE TO JOVIAN ACTIVITY
Fig. 40.--Chree analysis of sunspot number.

40 35 30

40 35 30
35 30
35 30

35 30 25




C
x 25La. I 5
LL
10 ,
50
I.
1.0
-29 -21 -5 0 +
DAYS RELATIVE TO JOVIAN ACTIVITY
Fig. 41.--Chree analysis of FX1, FX2, FX3, C, A (all dimen-22 2 P
sionless), and the 2800 Mc/s solar flux (10 w/m /cps).




On the basis of the similarity between the Florida and Chile
sunspot number correlations for C = 3 and for C = 2, and the differences between the two pairs of curves, it is concluded that fast Jupiter (C = 3), does indeed depend in some way on sunspot number.
According to E. N. Parker [43], solar particle flux consists of at least two components, one reaching Jupiter in approximately 23 days, the other in 5 to 10 days. Clearly this information lends credence to the peaks observed for fast Jupiter (C = 3), at -23 days and -6 days. Furthermore, these data suggest that the Jovian pulse character originates at Jupiter, and that in fact, fast Jupiter is simply normal Jupiter pulses modulated in some way at Jupiter by turbulence introduced in the ever-present solar wind by sunspot activity (the 23-day
effect), and by the fast particles ejected in flares occurring at times of sunspot activity (the 6-day effect).
Figure 41 shows typical examples of the other correlations,
none of which shows peaks at approximately the same place, and one can
integrate the three FX curves visually and see that the result is essentially the same as the sunspot number correlation discussed above. The FX curves for C = 2 and C = 1 are not shown, since they behave similar to the FX curves for C = 3 in Fig. 2. This discussion applies equally to the FN curves, none of which are shown. Apparently no specific region of the Sun influences the formation of the basic Jovian pulse character.
Each Chree analysis curve for C resembles the C curve in Fig. 41, having only the expected main peak at 0 days. No cyclical effects are observed.




The peak at -5 days in the A data of Fig. 41 is considered
P
spurious, since it does not show up on the other A plots which, again, are not shown. The A data are considered to be inconclusive.
p
Likewise, the 2800 Mc/s flux correlation appears to be inconclusive, the curve in Fig. 41 being typical of each of the various plots. The slow decline from -25 days to +5 days is present, though not so evident, in the other data. Whether this decline is a real effect, is as yet uncertain.
i. System III Longitude Dependence
Curiously, the basic pulse structure seems to exhibit a slight System III longitude dependence, as Fig. 42 shows. In this plot appears Average Pulse Character versus System III Longitude, the data being that of Florida, 1962, at 18 Mc/s. Discounting the peaks and valleys outside the region 800 to 3100, wherein lies 85 per cent of all the data, it appears that Source B (800 to 1600) tends to produce more fast Jupiter than Sources A and C, located from 2000 to 2900. This is construed as positive evidence of the origin of the fine structure at the Jovian environment, for there appears here a correlation between this fine structure and the actual Jovian sources. j. Dependence on the Elongation of Jupiter's
Moon Io
Since Jupiterls activity is known to be related to the position of at least one of the Jovian moons, Io [44, 451, Fig. 43 was prepared as a plot of Jovian pulse character versus the departure of Io from geocentric superior conjunction. The data were for Florida, 1962, at




FLORIDA 1962
18 MC/S
X
WI
t1dt > 85 % OF DATA '
<
0I
I I
I I
I I
00 0 0 0 0 0 0
0 45 90 135 180 225 270 515 60
LONGITUDE, SYSTEM Il
Fig. 42.--System III longitude dependence of Jovian burst fine structure.




0
350

0 0 100 150 200 250 0
0 50 100 150 200 250 :500

DEPARTURE FROM

GEOCENTRIC SUPERIOR CONJUNCTION

Fig. 43.--The influence of Jupiter's moon Io on Jovian burst fine structure.




18 Mc/s. In considering only the shaded areas as significant data, there is seen to be some tendency towards faster Jupiter when Io is near '85. In the other favored position, 2400, there is a slight tendency towards slower Jupiter. Realizing that Io's position and the
Jovian System III longitude combine to influence the probability of observing intense Jupiter activity, one can only speculate as to which parameter, if either, causes differences in the fine structure. On the basis of this evidence, the longitude dependence is favored as a source of fine structure characteristics, since the variation of pulse character seems to be more clearly defined in the longitude plot, Fig. 42.




3. A Comparison of Jovian Radio Bursts with Artificial Satellite Signals at Appulse*
The successful launching of NASA's beacon satellite S-66
presented an opportunity to determine the extent of the influence of the terrestrial ionosphere on radio signals, and hence to supplement the data presented in Sec. 111.2, which indicate that the basic Jovian pulse structure originates either in the Jovian environment or in the interplanetary medium. The satellite orbits at approximately 600 miles altitude, well outside the ionosphere. Aboard are several CW radio frequency beacons, including one operating at 20.005 Mc/s. Since S-66 first achieved orbit on October 10, 1964, the University of Florida radio astronomers, as part of the regular decametric-wavelength Jupiter observations have recorded near-simultaneous signals from S-66 and Jupiter at appulse, in hopes of measuring the correlation between the basic Jovian pulse character and the ionospheric scintillations observed on the satellite signal.
The apparatus for the experiment is relatively simple. Two identical 20 Mc/s, 3-element Yagi antennas separately feed crystalcontrolled receivers. One of these channels, known as 20YJ, monitors Jupiter in the normal AM mode. The other, 20YS, monitors S-66, using single sideband reception to provide aural monitoring. Thus 20YS is set at, say, 20.0055 Mc/s (lower sideband), and the audio output of the receiver is a fading 500 c/s t6ne, easily heard by the observer.
This work has been supported under Proje.ct A01 of the University of Florida NASA Institutional Grant No. NAS G542.




The audio envelope of the Jupiter activity and S-66 signal beat tone are displayed side by side on a Brush dual-chanel oscillograph, generally operated at a chart speed of 5 mm/s. Channel 20YJ is detuned from the S-66 carrier to approximately 20.015 Mc/s, so that the Jupiter record will contain no interference from the beacon. In the absence of 20 Mc/s Jovian activity the other regular channels ranging from 10 to 53 Nc/s may be displayed on the Brush in place of 20YJ, permitting the observer to make use of any Jupiter storm available during an S-66 pass. In principle the effects of Jupiter activity superimposed on the S-66 record could be removed by some subtraction scheme. Thus far the development of this technique has not been necessary, since no records have been sufficiently good to contain this interference. Usually the Jupiter activity has either preceded or followed the S-66 pass by a few minutes, or the bandwidth of the Jupiter storm has not included the beacon carrier, and Jupiter-beacon interference has been no problem.
In the same manner as described in Sec. 111.2, the pulse character is described quantitatively by an index C which is assigned the value "Y" for slow Jupiter, "2" for normal Jupiter, and 11311 for fast Jupiter. Correspondingly, the S-66 signal scintillations are assigned an index S similar to that used by Martin and Lawrence [421 and ranging from 0 for no detectable scintillations to 2.0 for extreme scintillations. The two indices are described in Table 1 of Sec. 111.2. Figure 2, Sec. 11.2, shows examples of Jupiter pulses described by C = 1 (top), 2 (center), and 3 (bottom). Figures 44 and 45 illustrate S-66 signals with scintillations described by various S values.




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EQBXX3J11_DRY94G INGEST_TIME 2017-07-11T21:28:00Z PACKAGE AA00058576_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

SPECTRUM AND ORIGIN OF THE JOVIAN RADIO BURST STRUCTURE By WILBUR FRANK BLOCK A DISSERTATION PRESENTED TO THE GRADUATE COUNOL OF THE UNIVERSITY OF FLORIDA IN PAllTlAL FULFILLMENT OF THE REQUillEMENTS FOB. THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA April, 196,

PAGE 2

To my wife Jo Ann and Our children Herman Karl Mary Edna John Beal Grace Ellen

PAGE 3

ACKNOWLEDG.l'1ENTS As is usually the case for fledgling Doctors of Philosophy in physics earning their wings through research in radio astronomy at the University of Florida, this dissertation represents an end product of the work of a team of research workers without whose sig nificant contributions this study could never have been accomplished. Dr. A.G. Smith, as chairman of the author 1 s supervisory com mittee, provided the impetus for this work. His intense interest and uncommon enthusiasm for research have provided the author a worthy example. Likewise, Dr. T. D. Carr, a member of the committee, in numerous consultations was always ready and able to resolve puzzling questions, and pose a few morei Dr. D. C. Swanson's encouragement and conversations have proved invaluable. The other committee members, Dr. T. A. Scott and Dr. R. G. Blake, were particularly helpful during the author 7 s preparation for the qualifying examinations, always hav ing time and desire to discuss and interpret concepts troublesome to the writer. Dr. s. S. Ballard, Chairman of the Department of Physics and Astronomy, has been most generous in terms of personal consultation, encouragement, and arranging financial assistance. Mr. Harold J. Huber of Orlando, Florida, was instrumental in obtaining for the writer the Research Task Award of the Martin Company (Orlando, Florida). The author is deeply indebted to the Martin Company for this assistance during the early part of his doctoral studies. iii

PAGE 4

The technical staff of the Department of Physics and Astron omy have always been willing to help the author to the full extent of their capabilities. Mr. Ralph Warren's extraordinary ability to make any type of gadgetry function contributed especially to the satellite tracking project. One cannot say enough good things to describe Mr. Hans W. Schrader whose practical knowledge of physics, machine work, photographic techniques, and electronics are supplemented by his per sonal integrity and spirit of cooperation. The writer enjoyed many enlightening conversations with his colleagues, particularly Dr. G. R. Lebo, Dr. N. F. Six, Dr. A. T. Jusick, Dr. Samuel Gulkis, Frank Tiberi, Jorge May, and C. N. Olsson. Dr. Lebo personally performed the Chree analyses utilizing the Uni versity1s computing facility. The incomparable W.W. Richardson assisted in setting up the artificial satellite observation channel, performed data analysis, prepared photographs and slides, and did his usual excellent job in preparing the drawings. W. A. Morton contributed directly to the experiment in -which artificial satellite signals were compared with Jupiter storms. He helped set up the equipment, record and analyze the data, and plan the presentation and interpretation of the results. In addition, he con tributed significantly to the Jovian continuum search and to the Jovian burst morphology study. M. L. Fagerlin constructed the log-periodic antenna and has been primarily responsible for this system 1 s maintenance. iv

PAGE 5

Other staff members who have contributed to this work in some way include Mrs. Lee Potzner, R. J. Leacock, T. Anderson, G. W. Brown, W. Cain, E. J. Lindsey, W. Mock, I. Shever, G. Walls, W. Greenman, C. Arlington, P. Trescott, R. Hayward, M. Evans, K. Williams, N. Chotas, J. Moeller, D. Smoleny, and our Chilean colleagues, H. Bollhagen and J. Levy. The writer thanks his parents, Dr. and Mrs. Herman H. Block, for constant encouragement and financial assistance without which this work would certainly never have been completed. The author's brother, David Block, and his friend, Thomas B. Elfe, have offered unusually stimulating advice and encouragement. The professional touch of Mrs. Thomas Larrick in editing and typing the manuscript made a joy of the burdensome task of assembling the work in final form. To work with such people as she certainly renews one 1 s faith in humanity. To all these fine people the author wishes to express his most sincere thanks. Gratitude is also expressed to the sponsors of the University of Florida radio astronomy projects: the National Aeronautics and Space Administration, the National Science Foundation, the Office of Naval Research, and the United States Army Research Office at Durham. V

PAGE 6

TABLE OF CONTENTS ACKNOWLEDGMENTS LIST OF TABLES LIST OF FIGURES Chapter I. INTRODUCTION ...................... II. THE COLLECTION AND ANALYSIS OF JOVIAN DECAMETRIC SPECTRA . 1. 2. 3. 4 5. Backgr o und Description of Jovian Spectra Frequency Dependence of Ionospheric Faraday Rotation from Jovian DecQ~etric Spectra Comparison with Other Workersf Data Spurious Spectra III. MORPHOLOGY AND ATMOSPHERIC MODIFICATION IV. 1.. The Search for a Jovian Continuum 2. Jovian Burst Morphology . a. b. c. d. e. f. g. h. i. j. Introduction The Jupiter Pulse Character Index Diurnal Variations Seasonal Variations Frequency Dependence Altitude Dependence Geomagnetic Activity Dependence Solar Activity Correlation System III Longitude Dependence Dependence on the Elongation of Jupiter 1 s Moon Io 3~ A Comparison of Jovian Radio Bursts with Artificial Satellite Signals at Appulse CONCLUSION . . . . LIST OF REFERENCES . . BIOGRAPHICAL SKETCH. . vi Pa g e iii vii viii l 3 3 6 29 38 42 50 50 60 60 61 63 69 75 75 78 78 85 85 89 99 101 105

PAGE 7

LIST OF TABLES Table Page 1. Definitions of Pulse Character and Scintillation 2. Indices The General Comparison of Jovian Pulse Character with Artificial Satellite Signal Scintillations Near Appulse .. .. vii 62 97

PAGE 8

Figure 1. LIST OF FIGURES Block diagram of the University of Florida radio spectrograph 2. Examples of Jupiter bursts of different character 3. 4. index C .. .. The spectrum o f slow Jupiter (September 26, 1963) Frequency variation of Faraday rotation on slow Jupiter spectra (October 11, 1963) 5. Normal Jupiter spectrum (August 26, 1963) 6. Normal Jupiter spectrum (August 26, 1963) 7. Normal Jupiter activity at 18 Mc/s from the high. Pa g e 4 7 9 11 12 14 speed oscillograph record 15 8. The comparison of the time variation of the amplitude of a spectral spike at 18 Mc/s with the 18 M c/s oscillograph record 16 9. A frequency-drifting pulse in the spectrum of mixed fast and normal Jupiter (September 13, 1963) 10. The spectrum of mixed fast and normal Jupiter (September 13, 1963) 11. Spectra of fast Jupiter (November 5, 1963) 12. Spectra of fast Jupiter (November 5, 1963) 13. 14. Spectra of fast Jupiter (November 5, 1963) Spectra of fast Jupiter (September 26, 1963) 18 20 21 22 23 25 15. Spectra of fast Jupiter (September 26, 1963) 26 16. Spectra of fast Jupiter (Septem b e r 26, 1963) 27 17. Logarithmic plot of rate of change of Faraday rotation spectral fringes as a function of frequency 32 viii

PAGE 9

LIST OF FIGURES (Continued) Figure 18. Terrestrial ionospheric Faraday rotation as a function of frequency for Gainesville, Florida, 0520 UniverPage sal Time, October 11, 1965. 54 19. Comparison of S-66 signals at 20 and 40 Mc/s 20. Jovian decametric spectra as recorded at the High Altitude Observatory at Boulder, Colorado [19] 21. Various types of spectral interference observed with 56 59 the University of Florida radio spectrograph. 45 22. Spectral response of the University of Florida radio spectrograph to broad-band noise. 47 25. A 22.2 Mc/s Jupiter storm of mixed normal swishes and contirluum 52 24. A 22.2 Mc/s Jupiter storm of swishes with no continuum 55 25. A 22.2 Mc/s Jupiter storm superimposed on the Cassiopeia A continuum ........ 56 26. The diurnal variation of ionospheric scintillations 64 27. The diurnal variation of Jovian burst fine structure (Florida, 1962) 65 28. The diurnal variation of Jovian burst fine structure (Chile, 19 62) 66 29. The diurnal variation of Jovian burst fine structure (Florida, 1963) 67 30. 51. 32. 33. 34. The diurnal variation of Jovian burst fine structure (Chile, 1963) The seasonal and altitude variations of radio star scintillations The seasonal variation of Jovian burst fine structure (Florida, 1962) . . . . The seasonal variation of Jovian burst fine structure ( Chile, 1962) .. . . . . The seasonal variation of Jovian burst fine structure (Florida, 1965) ix 68 70 71 72 73

PAGE 10

LIST OF FIGURES (Continued) Figure Page 35. The se a sonal variation of Jovian burst fine structure ( Chile, 1965) 7 4 36. Frequency dependence of radio star scintillations [42] 76 37. Frequency dependence of Jovian burst fine structure 77 38. Altitude dependence of Jovian burst fine structure 79 39. Geomagnetic activity influence on Jovian burst fine structure 80 40. Chree analysis of sunspot number . . 41. 42. 45. Chree analysis of FJO.., FX2, FX3, C, A (all dimension p less), and the 2800 Mc/s so1ar flux (lo22 w/m 2 /cps) System III longitude dependence of Jovian burst fine structure . . The influence of Jupiteris moon Io on Jovian burst fine structure 44. Extreme examples of scintillation index S 45. A direct comparison of Jovian burst structure with satellite signal scintillations . . 46. The dependence of the scintillation index S' on scintillation frequency . . . 47. Scatter diagrams of Jovian pulse character C versus scintillation index S . . . . X 82 83 86 87 92 93 94 96

PAGE 11

Cf.[APTER I :INTRODUCTION Jupiter 1 s sporadic decametric radio emissions, discovered by Burke and Franklin [l] in 1955, have been eagerly recorded and syste m atically analyzed by University of Florida physicists since 1957. Fol lowing the inspiring leadership of A.G. Smith, these scientists are attempting to compile information sufficient to suggest a suitable physical model for the origin of the no-longer strange Jovian signals. Successive dissertations of T. D. Carr [2], N. E. Chatterton [3], N. F. Six [4], and G. R Lebo [5], have dealt primarily with the gross statistics of the experimental data. The purpose of this work is to amplify the previous papers through a consideration of the "fine structure," so to speak, of the reams of data which have been collected at Biven 1 s Bank. Thus, where Six [4] and Lebo [5] discussed exhaustively Jupiter 1 s radiation probability and intensity, here will be treated the complementary material: the morphology of the Jovian bursts in the time and frequency domains. Specifically, the dynamic spectrum analyzer, set up by Chatter ton [3] in 1960, has been improved through replacement of the cumber some, inefficient, immobile rhombic antenna by an automatic tracking, broad-band, log-periodic structure [6]. The miles of film produced by this instrument since Chatterton's publication of 1961 have been 1

PAGE 12

2 analyzed for spectral characteristics of Jovian emission and are reported later in Chapter II. The 22 Mc/s interferometer has been used to search for a radiation continuum emanating from Jupiter, the results appearing in Sec. III.l. Likewise, the high-speed oscillograph records have be e n examined for pulse duration characteristics as discussed in Sec. III.2. Finally an attempt has been made to learn the effect of the Earth 1 s ionosphere and magnetic field on the Jovian signals by a study of the signals received from a beacon satellite, S-66, orbiting outside the ter restrial ionosphere, as reported in Sec~ III.3. Possibilities of the contrary notwithstanding, it is certainly hoped that the information documented here will speed the evolution of a suitable explanation for the origin of Jupiter 1 s decametric radio bursts.

PAGE 13

CHAPTER II THE COLLECTION AND ANALYSIS OF JOVIAN DECA:METRIC SPECTRA 1. Background N. E. Chatterton, as part of his doctoral dissertation research, set up the University of Florida radio spectrograph in 1959, first observing Jovian decametric spectra on February 3, 1960. His disserta tion [3] included many photographic reproductions of typical spectra observed through April 27, 1961, along with exhaustive descriptions of their characteristic features. Available at the time were spectra con tained in some 2100 feet of 16 millimeter movie film. Chatterton noted particularly that bandwidths of the bursts ranged from 0.1 to 3 Mc/s, and that individual pulse shapes ranged from sym.~etric to nondescript. The tendency of the individual pulses to bifurcate was pointed out, and the general peaking or fringing fine structure was examined in detail. The continuing interest of the Florida group in Jupiter's spectrum is well known [7, S, 9, 10]. A portion of the present work includes the extension of the spectral collection and the accompanying analysis. Much of the new spectral data have been taken on an improved version of the original Florida spectrograph. Figure 1 is a block diagram of the equipment. During observations Jovian radio signals incident on the log-periodic

PAGE 14

4 LOGPERIOD! C ANTENNA PREAMPLIFIER SPECTRUM ANALYZER MOTION PICTURE CAMERA I CLOCK I Fig. L--Block diagram of the University of Florida radio spectrograph. \.

PAGE 15

5 antenna are preamplified, then displayed as an intensity versus fre quency plot on the cathode ray oscilloscope tube of a commercial spec trum analyzer (Panoramic Model SPA-3). At the discretion of the observer, the center frequency may be chosen between 13 and 23 Mc/s, the spectral sweep width between 0.25 and 4 Mc/s. Desirable data are recorded with time of occurrence on 16 millimeter movie film (Bolex Model 16R Camera, Eastman Kodak Tri-X Negative Film) at the rate of 12 frames per second. The principal modification to the original sys tem was the replacing of the fixed rhombic antenna with an automatic tracking log-periodic dipole antenna constructed by M. L. Fagerlin [6]. The reasons for.changing antennas included increased observation time, i..~proved efficiency, broader system bandwidth, and the determination of whether spurious ringing in the rhombic antenna system caused the peak ing observed on Jovian spectra. The timer was changed from a minute counter to a clock displaying hours, minutes, and seconds. The framing rate of the camera was increased from 4 to 12 frames per second to help preserve quick-changing spectral effects. Figures 3 through 16 include reproductions of characteristic examples of the collection which now contains approximately 8700 feet of 16 millimeter movie film, including data through December 21, 1963. Since Chatterton's study some new types of Jovian spectra have been identified, and the larger data sample available now enables one to more satisfactorily classify the spectra. In Sec. II.2 the spectral collection examples will be described in detail. Then in Sec. II.3

PAGE 16

6 the peaking or fringing phenomenon will be analyzed. A comparison with other workers' results follows in Sec. II.4. Finally, Sec. II.5 con tains a discussion of some troublesome spurious spectral effects. 2. Description of Jovian Sn e ctra The discussion of the features of Jovian spectra is expedited by the following brief description of the audio envelope of these radio signals. Jupiter 1 s decametric radio storms appear to a terrestrial radio telescope as groups of pulses of electromagnetic energy, spaced randomly in time. The length in time of the individual pulses ranges from the order of milliseconds to tens of seconds. For analytic purposes the storms have been arbitrarily classified as 11 fast, 11 11 normal, 11 or 1 lslow, 11 depending on whether the bulk of the pulses in a storm are of the order of approximately 0.1 second and shorter, 1 second, or several seconds in duration, respectively. These classes of Jupiter pulses are often referred to as 11 popping 11 or 11 spitting Jupiter, 11 llswishes, 11 and 11 long rollers. 11 In Fig. 2, after Mock [11], typical high-speed oscillograph records of each type of Jupiter storm are shown. Slow pulses appear in the top strip, the center strip is a recording of normal Jupiter, and the lowest record depicts fast Jupiter activity. The characteristic names are derived from the sound of the storms as heard over the loudspeakers of the observatory. The normal swishes sound somewhat like ocean waves breaking on a distant beach. The fast pulses sound similar to the popping of popcorn in a covered pan. The slow rollers are detected as a slow change in the intensity of the hissing sound due to the galactic background noise and are

PAGE 17

CHILE 10 MC/S X J X J CHILE 18.0 M-C I J Ol23.5 E ST -r AUG. 2 7, 1962 -,. J MAY 24 1 1961 JUNE 7, 1962 Fig. 2.--Examples of Jupiter burst s of different character index c Top: S l ow Jupiter, C = 1. Center: Norma l Jupiter, C = 2. Bottom : Fast Jupi ter, C = 3

PAGE 18

8 sometimes scarcely audible,since the rate of change of intensity is so small. In Sec. III.2.b a character index is assigned to each class of pulse to lend organization to a pa r ametric study of the Jovian burst morphology in the ti.me domain. For the present the qualitative nom e clature suffices. Figur e 3 is a good example of slow Jupiter. In this and all succeeding spectral displays, ti.me begins at t~e top left fr~~e, pro gressing down to the end of the column, thence to the top of the next column,and so forth. This particular burst is centered at 19.8 Mc/s, with a sweep width of 3 Mc/s, and it was observed at 0239 Eastern Stand ard Ti.me on September 26, 1963, lasting some 15 seconds. Successive frames shown here are spaced 1.5 seconds apart, except for the final frame, which has expanded sweep width to resolve possible fine struc ture details absent in the spectra filmed at the normal sweep width of 3 Mc/s. The activity is classified as slow Jupiter because the changes in intensity at any particular frequency are gradual, some features maintaining stability over several seconds. Thus there are four peaks which can be traced easily through the first six frames spanning 9 sec onds. Other fringes may be traced correspondingly throughout the long sequence. Additional visible features, not the sole property of slow Jupiter, are the obvious peaking tendencies or fringes approximately 0.4 Mc/s wide and an apparent drift of the center frequency of the burst from 20.5 Mc/s to 19~0 Mc/s. The fringes are attributed to Fara day rotation of the Jovian signals in the terrestrial ionosp~ere and are discussed in detail in Sec. II.3. The cause of the center frequency

PAGE 19

(0) 18.3 19.8 21.3 (MC/&) (b) 18.3 19.8 21.3 (M C/8) (c) I I I I l 18.3 19.8 21,3 (Mc/&) 19.3 19.8 20.3 (Mc/a) Fig. 3.--The spectrum of slow Jupiter (September 26, 1963). 9

PAGE 20

10 drift is unknown. The sharp spike at 21.6 Mc/sis of instrumental origin, this local oscillator trouble being described in Sec. II.5 The final frame of col1.L111I1 (c) at expanded sweep width of 1 Mc/s does not reveal any additional hyperfin~ structure. Figure 4 shows an unusual broad-band Jupiter storm observed October 11, 1963. During the period 0015 to 0040 Eastern Standard Time slow rollers were recorded at frequencies ranging from 12 to 22 Mc/s. Single isolated frames are shown to demonstrate the broad-bandedness of the storm. In Fig. 4(a) the center frequency is 13.5 Mc/s, sweep width 3 Mc/s. The rapid fall-off of signal level near 12 Mc/sis due to the receiving band width of the spectrograph 1 s log-periodic antenna. Faraday rotation fringes are clearly evident, crowding more closely together towards the lower frequency end of the spectrum. Figure 4(b), (c), and (d) show slow Jupiter at center frequencies 15, 18, and 19 Mc/s, respectively, with sweep width 3 Mc/s. In each case the Faraday fringes become wider toward higher frequencies. The spectra shown as Fig. 4(e) and (f) show the slow Jupiter centered at frequencies 18 and 17.2 Mc/s, respectively, with the sweep width expanded to 1 Mc/sin hopes of finding hyperfine structure detail. Nothing uniform was observed, only the finely spaced irregular noise pulses of the galactic background. Figure 5 is a display of normal Jupiter activity at 18 Mc/sand sweep width 3 Mc/s as observed from 0300 to 0301 Eastern Standard Time on August 26, 1963. Frame spacing is 11/2 seconds and the activity lasts some 28 seconds. While superficially this spectrum appears

PAGE 21

(a) 12 13.5 15 (Mc/s) (d) 17.5 19 20.5 (M c/s) ( b) 13.5 15 16.5 (Mc/s) (e) 17.5 18 18.5 (M c/s) (c) 16.5 18 19 5 (M c/s) ( f ) 16.7 17.2 17. 7 (M c/s) Fig. 4 .-Frequency variation of Farada y rotation on slo w Jupiter spectra (October 11, 19 6 3). 11

PAGE 22

(a) 16.5 18 19.5 (Mc/s) ( b) (c) 16.5 18 19.5 (Mc/s) Fig. 5.--Normal Jupiter spectrum (August 26, 1963). 12

PAGE 23

13 si r. i:i..lar to that of Fig. 3, close examination reveals that frame-to frame changes in fine structure are more noticeable in Fig. 5 t h an in Fig. 3. This is the mark of normal Jupiter spectra: fine structure changes occurring in the order of 1/2 to 1 second. Note again the Faraday fringes and the apparent frequency drift from high to low. The lonesome spike appearing sporadically at.18.7 Mc/sis attributed to radio station interference. An interesting apparent low-frequency cut-off occurs at 17.3 Mc/sin frames 4, 5, and 6, column ( a) gradu ally changing to a high frequency cut-off at approximately 17.7 Mc/s in frames 3, 4, and 5 of column (c). Figure 6 is another example of normal swishy Jupiter pulses. Frame spacing is now only 1/2 second for this activity of 0259 to 0300 Eastern Standard Time, August 26, 1963, centered at 18 Mc/s with sweep width 3 Mc/s. Two distinct spectral spikes at 17.6 and 18.0 Mc/s are seen to grow and decay. The 18.0 Mc/speak was monitored as Jupiter activity and recorded through channel 18Y on the high-speed Brush oscillograph. Figure 7 is a rep r oduction of this record. Fi~u.re 8 is a plot of the amplitude versus time for the pulse as observed on the spectrograph (solid line) and the Brush record (dashed line). If we accept the observer's judgment in identifying the Brush record pulse as Jovian in origin, the correlation of the two curves in Fig. 8 confirms the narrow spike at 18.0 Mc/sin Fig. 6 to be the spectru. ~ of the pulse on the Brush record, and hence to be of Jovian origin. The s~llilar pulse at 17.6 Mc/s should likewise be due to Jupiter. However, the noticeably more narrow spike at 18.7 Mc/sis probably due to a radio

PAGE 24

14 (a) (b) (c) 16.5 18 19.5 16.5 18 19.5 16.5 18 19.5 (Mc/s) (M c/s) (M c/s) Fig. 6.--Normal Jupiter spectrum (Au g ust 26, 1963).

PAGE 25

26 AUG. 1963 J Fig. 7.--Normal Jupiter activit y at 18 M c/s f r om the high-speed oscillograph record. 15

PAGE 26

16 is.or t ---OSC LLOGRAPH -SPECTRAL PE A K ij 16.0! fl i \v / ,\ I \ !!.. I D / \ I I I 14.0r Ii t li 1 1 12 .ol,_ _J I ,. I"\ I c... '--, \ .... I --, \ I f-I g I __ a.or r,') en 1,1 6,0j~ ..... (0 C\l I 0 0 1 ci 5.J i'I) I ::, O
PAGE 27

17 station. Note this same spike at 18.7 Mc/sin Fig. 5. Figure 6 is :identified with normal Jupiter because intensity changes occur in the order of 1 second. Isolated spectral spikes like these, with band widths of approximately 0 .1 Mc/s, are relatively rare. During a period of mixed normal and fast Jupiter pulses the data in Fig. 9 were taken, in which a pulse was observed to drift uni formly in frequency. The time was 0203 Eastern Standard Time, Septem ber 13, 1963, with center frequency 18 Mc/s, sweep width 1 Mc/s. Suc cessive frames are spaced at intervals of 0.5 seconds. Fast Jupiter is evident at 17.3 Mc/sin the first three frames of Fig. 9(a). Note the rapid frame-to-frame changes. Similar fast activity is evident at approximately 18.5 Mc/sin the last three frames of Fig. 9(a). Better examples of fast Jupiter appear in later figures. These examples are explicitly mentioned here to emphasize that sometimes more than one kind of Jupiter activity appears simultaneously. Thus, in the third frame of Fig. 9(a), a swishy pulse develops at approxL~ately 18.06 Mc/s, which lasts throughout the sequence., a time of 8 seconds. This pulse changes amplitude principally in the manner of normal Jupiter. However, some fast changes at the second, third, fourth, fifth, and sixth frames of Fig. 9(c) show the pulse also to possess fast Jupiter characteris tics 'mother highly unusual feature of the pulse is its uniform drift in frequency from 18.06 to 17.95 Mc/sin 7 seconds, a drift rate of approximately 0.007 Mc/s per second. The tendency of Jovian storms to drift in frequency has long been lmown, but this is the first clear example of a specific Jovian pulse 1 s drifting in frequency. It is noted

PAGE 28

(a} ( b) (c) 2 Q I I I 17.5 18 18.5 17.5 18 18.5 (Mc/s) (Mc/1) Fig 9 .-A frequency drifting pulse in the s~ect rum of mixed fast and n ormal Jupit er (Septe mbe r 13, 1963) 18

PAGE 29

19 that the bandwidths of the Jovi a n pulses shm,m in Fig. 9 vary but are in the order of 0.1 Mc/s. Figure 10 shows more mixed norm a l and fast Jupiter observed 0214 to 0216 Ea s tern St a ndard Time, September 13, 1963. Frames are spaced 1/2 second apart. The sequence in Fig. l0(a) is centered at 18 Mc/s, with sweep width 3 Mc/s. Figures l0(b) and l0(c) are both centered at 22.5 Mc/s with 1.0 Mc/s sweep width. While the normal Jupi ter character is evident in the slowly changing envelope of each burst, the fast Jupiter features appear in sharp frame-to-frame variations of the individual pulse shape and the ragged, ill-defined tops of t~e pul ses. This latter characteristic is seen especially at 17.2 Mc/sin frames 2 and 3 of Fig. l0(a), and again at 22.9 Mc/sin frames 2, 3, and 4 of Fig. l0(b). Clearly, then, fast and normal Jupiter can occur simultaneously. Aural monitoring agrees with this conclusion, in that observers plainly hear pops superimposed on swishes. Figures 11, 12, 13, 14, 15, and 16 are examples of fast Jupiter pulses at various frequencies and sweep widths. Figures 11, 12, and 13 are taken from data of November 5, 1963: and each consists of three strips of four successive frames spaced at 1/12 second. The center fre quency is 18 Mc/s, sweep width 3 Mc/s. Fast Jupiter spectra are charac terized by the extremely rapid changes in pulse shape from frame to frame. For example, the first and fourth frames of the sequence in Fig. ll(a) show little activity, but the second and third frames show high intensity, fast Jupiter bursts at 18 Mc/s. Similarly, note the change in appearance of the spectral sequence in the first three frames

PAGE 30

(a) 16.5 18 19.5 (Mc/s) ( b) 22 22.5 23 (Mc/s) (c) F ig. 1 0. -The spectrum of mixed fast and normal Jupiter (September 1 3, 1 963). 20

PAGE 31

(o) I I I I I 16.5 18 19.5 ( M c/s) (b) I I I 16.5 18 19.5 (Mc/s) (c) I I I I I 16.5 18 19.5 (M c/s) Fig. 11.--Spectra of fa st Jupiter (November 5, 1963). 2 1

PAGE 32

(a) I I I I 16.5 18 19.5 ( M c/s) (b) I I I I I 16.5 18 19.5 ( M c/s) (c) I I I I 16.5 18 19.5 (Mc/s) Fig. 12.-Spectra of fast Jupiter (November 5, 1963). 22

PAGE 33

(a) I I I I 16 5 18 19 5 (MC/&) ( b) I I I I I 16.5 18 19.5 (M C/1) (c) I I I I I 16.5 18 19.5 ( M c/s) Fig. 13.--Spectra of fast Jupiter (November S, 1963). 2 3

PAGE 34

24 of Fig. ll(b). Then in Fig. ll(c) a fast, large amplitude pulse appears at approximately 18.4 Mc/sin the second frame only, missing the first and third frames. Recalling that normal Jupiter sw"ishes are produced by the small cr.anges occurring over the 1/2 second fran1.e spacing of Figs. 5 and 6, it iB not surprising that these large changes in the spectrum over intervals of 1/12 second as in Fig. 11 accompany popping, spitting, or cracking noises over the loudspeakers at the observatory. More examples of these spectra of fast Jupiter follow. In Fig. 12(a) the second frame alone shows an isolated burst at 18.5 Mc/s. The best similar example of fast Jupiter in Fig. 12(b) is also the second frame, which shows an intense burst of l Mc/s band width, there being scarcely any activity in the preceding and succeeding frames. Figure 12(c) shows what might be a fast Jupiter burst moving in fre quency from low to high. The bursts in the first, second, and third frames exhibit the ragged-top character of fast pulses and are centered successively at approximately 18.0, 16.2, and 18.4 Mc/s. Figure 15 shows similar fast Jupiter bursts. Intense activity not shown in adjoin ing frames may be seen in the second and third frames of Fig. 13(a) near 18 Mc/s, the second and fourth frames of Fig. 15(b) near 18 Mc/s, and in the third frame of Fig. 15(c) just below 18 Mc/s. In Figs. 14, 15, and 16 f~atures similar to those of Figs. 11, 12, and 15 are shown with expanded sweep. Each of these figures has three columns of three successive frames spaced 1/12 second apart. Fig ure 14(a) shows fast Jupiter in the second frame at 18.5 and 18.6 Mc/s, as does the second frame of Fig. 14(b) at 18.1 Mc/s. Figure 14(c) shows

PAGE 35

(a) I I I 17 9 18.4 18.9 (Mc/1) ( b) I I I I 17.9 18 4 18.9 (Mc/s) (c) I I I I 17.9 18.4 18.9 ( M C/6) Fig. 14.--Spectra of fast Jupiter (September 26, 1965). 25

PAGE 36

(a) I I I I I 179 18.4 18.9 (Mc/1) ( b) I I I I L 179 18.4 18.9 (Mc/a) ( C) I t I I I 179 18.4 18.9 (MC/I) Fig. 15.--Spectra of fast Jupiter (September 26, 1963). 26

PAGE 37

(a) I I I I 18 75 I 9 19.25 ( M c/s) ( b) I I I 18.75 19 19.25 ( M C/S) ( C) I l I l 18.75 19 19.25 ( M C/s) Fig. 16.--Spectra of fast Jupiter (September 26, 1963). 2 7

PAGE 38

28 a fast burst at 18.l M.c/s in the second frame. Figure lS(a) contains fast Jupiter at 18.l M.c/s in the first frame, 18.05 and 18.6 Mc/sin the second frame, and 18.l Mc/sin the third frame. Figs. lS(b) and lS(c) show fast Jupiter bu.:-sts at 18.4 Mc/s, which change shape from frame to frame. In Fig. 16 the sweep width is expanded to 0.5 Mc/s. The second frame of Fig. 16(a) has a fast Jupiter pulse at 19.l Mc/s. Figures 16(b) and 16(c) feature a fast-changing burst at 18.75 Mc/s. Generally speaking, the major distinction of fast Jupiter spectra is the absence of the uniformly spaced Faraday fringes observed in both normal and slow Jupiter. spectra. Befora leaving spectra of fast Jupiter, it should be mentioned that apparently the sweep time of the spectrograph's oscilloscope, .mich is approximately 30 milliseconds, does not limit the bandwidth of the spectra, because if this were so, during Jupiter activity the center frequency of the burst would move about the scope in random fashion, depending on the beam position at the time of reception of a Jovian burst. This is not the case, as sho'Wil by Figs. ll, 12, 15, 14, 15, and 16 where, although the Jupiter pulses are sporadic in tilne, they tend to center about a narrow frequency band in each sequence. In Figs. ll, 12, and 15, where all data were taken between 0125 and 0127 Eastern Standard Time, November 5, 1965, the activity all centers near 18 Mc/s~ In Figs. 14 and 15, for 0546-0547 Eastern Standard Tilne, September 26, 1965, the activity centers about 18.4 Mc/s; while the fast Jovian spec tra of 0356 Eastern Standard Time, September 26, 1965, in Fig. 16 center mainly near 18.75 Mc/s.

PAGE 39

29 In summary, Jovian spectra exhibit characteristics which can be used to identify the time morphology of the radiation. Slow Jupiter observed thus far exhibits regular Faraday fringes with relatively smooth tops. Normal Jupiter likewise sho~~ either Faraday fringes with more ragged appearing tops, or isolated narrow-band spikes of tillle varying amplitude. Fast Jupiter does not show regular Faraday fring ing and is highly ir:.-egular in pulse shape, with washed-out, ragged patches revealing the fast-changing spectral burst structure. 5. Freauency Dependence of Ionospheric Faraday Rotation from Jovian Decametric Spectra For many years research workers have been using a simplified model of the rotation of the plane of polarization of radio waves in a magnetoionic medium (Faraday rotation) to interpret such phenomena as lu.~ar radio echoes (12, 15, 14], artificial satellite signal fading [15, 16, 17, 18], and Jovian decametric spectral fringes (19]. Browne et al. [15] have shown that for the quasi-longitudinal approximation and no ray path splitting of the ordinary and extraordinary modes, the Faraday rotation of plane polarized electromagnetic waves traveling through the ionosphere is inversely proportional to the frequency squared. Other workers have used these results in the form K ----h n .. 2 H cos 9 sec X / N dh f 0 (1) where K is a constant, f is frequency, His the strength of the terres trial magnetic field, 9 is the angle between the ray and H, xis the

PAGE 40

30 angle between the ray and the vertical, N is the electron density along the ray path, and 1 h N dh is the total electron content along the path. 0 Browne et al. calculated the ionospheric electron content using experi..~ental values of Oat approximately 120 megacycles per second for radio echoes from the moon [13]. Warwick and Dulk [19 J measured O from Jovian spectral data, calculatedthe electron content of the ionosphere and extrapolatedthe Faraday rotations to Jupiter to deduce the initial orientation of the Jovian radiation polarization ellipse and the nature of the mechanism of this radiation. In each case the inverse square freq~ency dependence of Eq. 1 is assumed valid and some special method must be used to assure the absolute determination of o, which through direct measurement can usually be measured only to within some undeter mined integral number of rota,tions. It is the purpose of this section to present a method for the determination of an unambiguous D directly from Jovian spectra and to show that the frequency d e pendence of the Faraday rotation of decameter wavelength radio waves through the terrestrial ionosphere can be meas ured experimentally by using Jupiter 1 s decametric spectra. An independ ent check of this result, using the signals from the beacon satellite S-66(BE-B), is included. Recall Fig. 3, a reproduction of a typical Jovian burst recorded on the Florida spectrograph. The center frequency is 19.8 Mc/s. The well-defined fringes have been attributed to Faraday rotation occur ring in the Earth 1 s ionosphere. Warwick and Dulk [19] used Faraday fringes appearing on Jovian spectral records of a somewhat different

PAGE 41

nature to calculate the terrestrial ionospheric electron content, assuming the validity of Eq. 1 and arguing that all of the ob s erved Faraday rotation occurred in the ionosphere. The Florida data are such that the frequency dependence of 31 Faraday rotation in the inosphere can be determined experimentally; that is, the validity of Eq. 1 can be tested. For example, data from radiation emitted by Jupiter 1 s source C (~II= 310, Warwick 1 s n1ate sourcen) between 05 1 4 and 0526 Universal Time, October 11, 1963, can be used to perform the followi.~g calculation. Since O, the absolute number of Faraday rotations, cannot be determined directly from spec tral records,~, the differential rotation rate with respect to fre quency, was measured and plotted as a function of frequency using the le a st squares method. This curve appears as the solid line in Fig. 17. For comparison purposes the dashed line in Fig. 17 is the theoretical curve, having a slope of -3 as det e rmined tm.~ough differentiation of Eq. 1. Note that for this plot the dim rn sions o f the ordinate are rotations per Mc/s. In the Flori d a raw data, such a :: Fig. 3, each fringe represents one-half of a Faraday rotation because the signal and antenna polarizations are parallel twice per Faraday rotation. The resulting experimental curve has a slope m = 2.7 and an ordinate intercept c = 3.9. The equation of the line is 1 1 f og af = m og + c, ( 2) which can be written log ~~= log (k.1") ( 3)

PAGE 42

LEJ.\ST SQUARES FIT \ \ THEORETICA.L 32 l i 7 SLOPE = -3.0 \ I ,b,. 1 .......,. ___ -1 ___ ......,.l ___ ,.._\ __ ....,_ ____ 1 40 10 20 FREQUENCY 30 ( M C/S) Fig, 17.--Logarithmic plot of rate of change of Faraday rotation spectral fringes as a function of frequency. One Faraday 11 revolution 11 corresponds to 2 fringes in Fig. 3.

PAGE 43

33 C where k 10 7950, using logarithms to the base 10. It im.~ediat e ly follows that k 11, (4) which can be integrated from Oto f, resulting in k:fil+l 0 = 1 m+ m ,/,. -1. ( 5) Inserting the experimentally-determined k and m produces the equation ( 6) Thus, the frequency dependence of the Faraday rotation of radio signals passing through the terrestrial ionosphere has been experimentally determined for the period of the observations. The solid line in Fig. 18 is a plot of Eq. 6, Faraday rotation (o) versus frequency (f). N ote that as the frequency is varied from 13 to 20 Mc/s, the Faraday rotation ranges from 60 to 29 rotations. The dashed line in Fig. 18 represents the the o retical inver s e s q uare frequ e ncy dependence and ranges from 55 rotations at 13 Mc/s to 23 rotations at 20 M c/s. These results are pr e liminary, and continued study of Jovian decametric spectra may refine the apparent inverse 1.7-power frequency dep e ndence of ionospheric Faraday rotation for sub-20 Mc/s electro magnetic waves. The present calculation was made from data not recorded for this specific purpose and the significance of the differ e nce bet w een the measured index of 1.7 and the theoretical value of

PAGE 44

6r FARADAY ROTATION SJ FREQUENCY DEPENDENCE \ 2{ I \ \ ---~--THEORETlCAL ,. \ EXPERIMENTAL \ x 50 \ \ \ en \ z 0 \ I\ :::> 45 \ _j 0 ji. > \ w 0::: \{ \ -., "'40\ 0 \ Ix,
PAGE 45

35 2.0 is at present uncertain. However, the method and the results are of current interest. Additional data taken for radiation traveling through the ionosphere at different angles with respect to the Earth's magnetic field wil l determine the range of validity of the quasi longitudinal approximation in the derivation of Eq. 1. Future Jovian spectral data will be gathered with these specific goals in mind. As an independent check of the preceding results, the 20.005 and 40.010 Mc/s beacons aboard S-66(BE-B), an artificial satellite orbiting outside the ionosphere, have been utilized in a comparison determination of the frequency dependence of ionospheric Faraday rota tion for this region of the spectrum. The form of the fre quency depend ence is eArpected to be 0 = G f-m ( 7) w h ere G is a parameter involving various quantities noted in Eq. 1, and mis again a constant to be determined experimentally. Differen tiation of Eq. 7 with respect to the time, t, yields oO f-m oG ot = ot ( 8) where~~ may be approximated as the ratio~~ and measured from simul taneous oscillograph recordings of the Faraday rotation modulated sig nals of the 20.005 and 40.010 Mc/s beacons of S-66. Figure 19 shows such a record of data taken at the University of Florida on November 21, 1964. Measuring (tt) 20 and (tt) 40 as the times for MJ = 1/2 Faraday rotation for the 20.005 and 40.010 Mc/s signals, respectively,

PAGE 46

Fig. 19.--Comparison of S-66 signals at 20 and 40 Mc/s.

PAGE 47

Eq. 8 becomes 1 = (20.005)-m oO at 37 (9) Rewriting Eq. 9 for the 40.010 Mc/s case and dividing the two results in (t,t)40 (40.010) m m (bt)20 = 20.005 = 2 (10) Therefore log [(bt) 40 l (tt)20 m = (11) log 2 Preliminary determinations of musing this technique with the data shown in Fig. 19 give m = 2.08 in agreement with the theoretical value of m = 2 used in Eq. 1, within the accuracy of the measurements. The difference between this value and the value of 1.7 as calculated from the Jovian spectra may be a real effect, since the ionospheric critical frequency lies just below the range of the spec~ral data, and certainly, for frequencies lower than the critical frequency, there is no transmission, hence no Faraday rotation, hence no frequency depend ence. It is plausible that the theoretical inverse square frequency dependence is valid for frequencies above approximately 20 Mc/s, but that for the range from 20 Mc/s down to the critical frequency a more rigorous theory is required for purposes such as the precise calculation of ionospheric electron content.

PAGE 48

38 4. Comparison w:i.th Other Workers I Data Other than the Florida spectral data discussed in this work and Chatterton 1 s dissertation [3], the only comprehensive treatment of Jupiter 1 s decametric spectra is that of Warwick of the High Altitude Observatory at Boulder, Colorado [19, 20]. A typical example of the Boulder data appears as Fig. 20, where the ordinate is frequency and the abscissa is Universal Time. This storm occurred on September 13, 1963, and the instrument sweeps from 7 .6 to 41 Mc/s every 1.3 seconds. The diagonal white streaks washing out strips of data approximately 0.5 Mc/s wide are interference fringes due to the antenna system. The faint alternate black and gray horizontal bands are caused by Faraday rotation fading and correspond to the Faraday fringes observed in the Flo;ida spectra sho"Wn and analyzed above in Secs. II.2 and II.3. At times the storm sho"Wn in Fig. 20 appears to extend from 13 to 36 Mc/s with a bandwidth of some 23 Mc/s. Other isolated bursts have bandwidths of less than 1 Mc/s. Compare Fig. 20 with Fig. 9, shown previously as an example of the Florida data. The time of the Florida data is 0203 Eastern Standard Time, which is 0703 Universal Time, and the Florida record thus represents an enlarged view of the faint speck appearing at 18 Mc/sand 0703 U.T. in Fig. 20. The Florida spectrograph is obviously superior for recording fine structure characteristics of less than 1 Mc/s bandwidth, but it cannot record the features involving bandwidths exceeding 4 Mc/s. This comparison, then, is limited to fine structure effects of extent less than 4 Mc/s. Sp e cifically, Warwick claims to have recognized

PAGE 49

1820G40... I 0630 I 0640 I 0650 I 0700 Fig. 20.--Jovian decametric spectra as recorded at the High Altitude Observatory at Boulder, Colorado [19]. I 0710

PAGE 50

40 events of bandwidths as low as 0.2 Mc/s [21], although these bursts are extremely difficult to recognize in his published dat 2 Now, band widths of this order are easily recognized in the Florida data. In Fig. 9, for example, the moving pulse near 18 Mc/s has a badwidth of approximately 0.05 Mc/s, roughly an order of magnitude less than the smallest reported by the Boulder group. Thus far Florida workers have not recognized any Jovian spectral spikes of bandwidth lower than 0.05 Mc/s. The rare narrow spikes such as were seen in Fig. 6 have never been filmed with an expanded sweep, and from data such as Fig. 6 it is determined that this type of normal Jupiter has a bandwidth of somewhat less than 0.1 Mc/s. The only other specific statement which Warwick makes concern ing the spectral bandwidth is that a rough positive correlation exists between pulse duration and spectral bandwidth. Examination of the Florida data shows that this is not always the case. For example, nor mal Jupiter activity exhibiting bandwidths ranging from 0.05 to 3 or 4 Mc/s has been observed at Gainesville. Fast Jupiter has ranged in bandwidth from 0.05 to 1.5 Mc/s. Indeed, the storm shown in Fig. 20 has at times a bandwidth of 23 Mc/s. This activity was also recorded on the Florida spectrograph and consisted primarily of fast Jupiter. T he Boulder equipment was apparently too insensitive to fine structure fluctuations in time to show this characteristic of the storm. Riihimaa of the University of Helsinki, Finland, is studying fine structure of the Jovian decametric spectra using an instrument which produces data similar to Warwick's, but of much higher resolution.

PAGE 51

4]_ In a brief preliminary report appearing in a l e tter to Nature [22], Riinimaa states that the Finnish spectrograph sweeps the range 18 to 20 Mc/s with a repetition frequency of 10 c/s. The resulting data lack the overall bandwidth of the Boulder data, but they do resolve the fine structure details, though not as well as the Florida data. Riihimaa shows typical Jovian spectra recorded by the Finnish spectro graph but postpones conclusions to a detailed analysis of his results wrj ch is in preparation. Thus, at this wri ti ng there is no basis for a genera l comparison of the Finnish data with those of Boulder and Florida. Apparently workers in the realm of Jovian decametric spectra require equipment capable of presenting the best of the spectral char acteristics seen in the Boulder, Florida, and Finnish data before the nature of these spectra can be completely understood. The Bo~lder spectograph sacrifices fine structure detail to present excellent broad-band characteristics of the spectra in time. The Florida data excel in recording fine structure details, but lose the broad-band characteristics. The Finnish data seem to be superior for systemati cally recording the slow changes in time of spectral fine structure at the expense of the loss of some amplitude versus frequency detail.

PAGE 52

4 2 5. Spurious Spectra The observer of Jovian spectra that are displayed on the University of Florida radio spectrograph must be aware of the char acteristics of the many spurious spectra w hi c h frequently appear in order that the desired data be correctly interpreted. There follows a description of these worrisome phenomena. Among the superfluous spectra produced, lightning interfer ence is the only type having natural or i gin. It is easily identified by correlating the sharp static sounds from a properly tuned receiver with the sudden rises and drops of the entire spectral trace on the oscilloscope. The trace sometimes shows discontinuities at points where the lightning crack happens to catch the sweeping beam, and these points are located at random frequencies. This characteristic aids in the distinguishing of spitting Jupiter from lightning since, as was noted in Sec. II.2, the Jupiter spitting bursts tend to localize at particular frequencies for many successive sweeps of the electron beam of the analyzer. Strong Jupiter storms may be recognized through intermittent lightning interference. A particularly bothersome effect has been that called in the Florida Radio Observatory Log (February 10, 1964), "Strange Interfer ence/ 11 Strange X, 11 and nBlow-Up Phenomenon, n deriving these names from the strange quality of producing no audio noise in the observa tory receivers (Log, December 20-21, 1963), yet appearing more intense than most Jupiter bursts observed on the analyzer oscilloscope. At first this signal was thought to be due to internal equipment

PAGE 53

4 3 malfunction, but this assumption proved erroneous. Another panoramic spectrum analyzer was set up to extend the range of our normal equip ment (Log, February 10, 1964). The second system displayed the spec trum from 24 to 34 Mc/s. The intermittent strange interference was found to be present si m ultaneously in both systems, and furthermore cut off abruptly at approximately 30 Mc/s. Moreover, the interference was found to be sensitive to antenna setting, peaking for a log periodic antenna setting of 55 West (Log, December 20-21, 1963). 0 (The antenna was set for declination -10 ). Additionally, disconnection of the antenna always stopped the interference (Log, January 10, 1964). The conclusion was that this interference probably originated at the University of Florida Medical Center, entering our spectrograph through the back lob~ of te log-periodic antenna. The appearance of the signal's spectrum is shown in Fig. 2l(a) and (b), where two differ ent characteristics can be seen. In Fig. 21(a) the trace resembles a modulated, distorted sinusoid, while in Fig. 2l(b) a square wave envelope is more evident. This type of interference, when present, blocks our spectral equipment over its entire bandwidth, rendering Jovian spectral analysis impossible, since the entire trace jumps on ?nd off the screen and changes rapidly between the two types shown in Fig. 2l(a) and (b), sometimes more frequently than once per second. Another category of interference has been labeled as 11 buzz 11 by the observers (Log, October 14-15, 1963), and is shown in Fig. 2l(c) and (d). While this spectrum closely resembles the Faraday rotation feature of the Jovian spectra previously discussed in Sec. II.3, there

PAGE 54

44 is no possibility of mistaking its regular audio buzz (somewhat like the sound a small electric motor with brushes produces in an AC-DC radio) for the randomly-spaced Jupiter swishes or spits. During per iods of buzz interference spectra of Jupiter storms cannot be satis factorily analyzed. Occasionally a swept-frequency jammer (Log, June 14, 1963) and (June 15, 1963), blocks the spectrum analyzer as shown in Fig. 2l(e) and (f); The audio of this particular type of interference sometimes appears on separate channels as a short buzz while the square (in fre quency) wave front sweeps back and forth across the oscilloscope face. There should be no chance of confusing these short bursts of buzz with the random Jovian bursts. At other times the jammer completely blocks one or more channels simultaneously. In general, Jovian spectra ob served under jamming conditions are unintelligible. Previous to the identification of the Faraday rotation fringes in Jovian spectra it was feared that the regular peaking observed in the spectra might be due to periodic gain irregularities [3]. Chatterton correctly believed this not to be the case [3]. In an effort to settle this question, a study was made in which continuous and pulsed noise was fed into the Pan system at various points (Log, December 7-8, 1963). Figure 22 shows the results of this experiment. Figure 22(a) is the spectrum of the galactic background incident on the log-periodic antenna. Periodic gain fluctuations are not evident under these conditions. For Fig. 22(b) the antenna and preamplifier were disconnected (see Fig. 1, the system block diagram) and broad-band

PAGE 55

I I I 20 21.5 23 (Meis) (a) Strange Interference 17.5 18 18.5 (M els) (d) Buzz 20.5 22 23.5 (M els) (b) Strange Interference 16.5 18 19.5 ( M c/s) (e) Jammer 20.5 22 23.5 (Meis) (e) Buzz 20 21.5 23 (Meis) (f) Jammer Fig. 21.--Various types of spectral interference o b s e rved with the University of Florida radio spectrograph. 45

PAGE 56

4 6 noise generated in a type 5722 noise diode was introduced directly into the spectrum analyzer. With max:ilnum sensitivity, the noise scarcely broadened the oscilloscope trace. Figure 22(c), (d), and (e) show the results of introducing unmodulated noise into the pre amplifier with the antenna disconnected, with a center frequency of 18 Mc/sand different sweep widths of 3, 1, and 0.l Mc/s, respectively. The spectra all look the same and exhibit no fringing or other features of Jovian spectra. With different center frequencies over the range of the spectrograph the results were the same. Figure 22(f), (g), (h), and (i) show the results of introducing modulated broad-band noise from the type 5722 diode into the preamplifier as before, with the antenna dis connected. The steep-front modulation was produced by mechanically switching the noise generator on and off several times per second in hopes of not observing any internal ringing subject to shock excitation. The noise spectra appear on only a portion of each trace because the llon 11 time of the noise generator was less than the cathode ray tube s w eep time (approximately 1/30 second). Again at the various center frequencies and sweep widths available to the system, no features of Jovian spectra were synthesized through shock excitation and no spur ious spectral features appeared. It must be pointed out that for a time the spectrum analyzer was out of alignment and occasionally an internal oscillation would produce a spurious peak on the oscilloscope. The effect was random and intermittent. Such a peak is the left-most peak in Fig. 2l(c) at approximately 20.6 Mc/s Note the characteristic skip in the base

PAGE 57

16.5 18 19.5 { a) Background 17.5 18 18.5 16 5 18 19.5 {b) Noise, No Preamplifier 17.95 18 18.05 16.5 18 19 5 (c) Noise 16.5 18 19 .5 {d)Noise (e)Noise (f)Pulsed Noise 17.5 18 18.5 20.5 22 23.5 21.5 22 22.5 (g) Pulsed Noise (h) Pulsed Noise li) Pulsed Noise Fig. 22.--Spectral response of the University of Florida radio spectrograph to broad-band noise. 47

PAGE 58

48 line, which is higher to the right of the break than to the left. While it is not obvious in Fig. 21(c), the leading edge of this peak is characteristically aL~ost vertical and the trailing edge on the right markedly less steep. An experienced ()bserver can easily reco g nize this troub Moreover, realigrunent of the equipment has removed this source of confusion and routine periodic maintena ~ ce should pre vent its recurrence. Radio stations 1 spectral spikes are another class of spurious signals appearing in our spectra. These are usually identifiable by their narrow width. A good example of a spectral line due to a radio station appears at 18.5 Mc/sin Fig. 6. Notice the narrowness of the station spike in contrast to the pair of Jupiter peaks at 16.75 and 18.0 Mc/s, recalling that these two peaks are representative of the slimmest Jupiter spectral peaks identified at the University of Florida Radio Observatory (Sec. II.2). Thus most radio stations 1 spectra are recog nized on the basis of bandwidth alone. In cases of doubt the observer should set a receiver on the exact frequency of a spike, using a grid dip meter, listening for Jovian or station audio characteristics from which to judge the origin of the spectral peak. In summary, all the characteristics observed presently on the University of Florida radio spectrograph are attributed to external sources. The characteristics of the Jovian spectra are documented in Sec. II.2. The spurious spectra which completely disrupt observations are strong lightning, closely-spaced stations, the inaudible strange interference, the swept-frequency jam.i~er, and strong buzz. Troublesome

PAGE 59

49 effects which partially impair observations but can be recognized by an expe r ienced observer include mild atmospherics due to distant light ning, occasional mild radio station activity, and an internal rin g ing due to misalignment of the spectrum analyzer.

PAGE 60

CHAPTER III MORPHOLOGY AND ATMOSPHERIC YlODIFICATION 1. The Search for a Jovian Continuum The sporadic, bursty nature of decameter wavelength Jovian radio emissions has been described above in Sec. II.2, and of course, by other research workers. The question remains whether there exists a continuum of radiation of flux density lower than that of the nor mal storms. In an attempt to resolve this question, a 22.2 Mc/s phase switching interferometer was established at Biven 1 s Bank in 1961 [25, 24, 25]. This observation channel, called 22B, should be capable of -25 / 2; detecting flux densities as low as 1 x 10 w m cps [25]. The instrument is steerable in declination only, and hence is effective only within a period of approximately one and one-half hours before and after meridian transit of the celestial object under observation. The theory and operation of the phase-switching interferometer have been treated elsewhere [25, 24, 25, 26, 27], and will not be repeated here, where the interpretation of the Florida data is discussed. Watson observed radio sources Cygnus A and Cassiopeia A using the 22B antenna and plotted smoothed curves [24] of the response of the system to these two sources at declinations 40 56 1 and 58 52 1 respectively. The maximum fringe occurs at the time of antenna bore sight transit, approximately 10 minutes after meridian transit as 50

PAGE 61

51 explained by Watson [24}. The fringe spacing near transit is inversely proportional to the cosine of the declination of the source producing the fringes [27]. Therefore, the fringes due to a source of lmown declination may be extrapolated to produce an expected fringe pattern for a source of different declination. The fringe pattern for Cygnus A in Fig. 19 of Watson's thesis [24] was contracted in time, point-by point, in the ratio of the cosine of the declination of Cygnus A to the cosine of Jupiter's declination to produce the predicted response of 22B to a radio source at Jupiter. Such a contracted fringe appears as the smooth curve at the bottom of Fig. 23. This is the predicted response of 22B for a 22.2 Mc/s source at declination 4.1, Jupiter's declination on September 26, 1963. Similar fringes were con~tructed as required by the varying declination of Jupiter and compared with all of the 22B records made to date to determine whether or not fringes due to continuous radiation from Jupiter were present. The results of this analysis follow. Channel 22B was operated extensively during the 1961 apparition of Jupiter. Effective watch was held 154 nights, the equipment being out of order on 36 nights when regular watch was observed. During the effective observations on 22B, Jupiter storms of 22.2 Mc/s were ob serv\;ld 31 times within one and one-half hours of transit on the observ atory's other 22.2 Mc/s equipment. Thirteen of these storms were recognized on 22B and consisted solely of sporadic bursts. Never dur ing these 31 storms, or at any other time during the apparition,was there any evidence on the 22B record of fringes such as would be

PAGE 62

';"' 0 000 1 I I ~"""'0""'3'""'0"' 0 ___, ~' "''""' """' '-~~ ~ ~ ~ -, --l~ 0 ~ 2~0~0 ~---~ 0100 E~ ""'"' ~~ -~ -, ... ~ u~, --1 00 00 I ~UPITER j ~; '~ ,.. 228 '-) -~ d :.RANSIT = ? .:j, I l ~111~1 ~! ~ j* \,~!,,~((?~~~~:~&~ ;~ t ~\ ,ilP~~I~ I 1.-t ~ '-':~ l I! ; L I ~:;~i .. UNIVERSITY of FLORIDA 26 SEPT. 1963 t ---.-----++--~ PRINCIPAL LOBE -----'----, ,,__ ~ PREDICTED RESPONSE of 2 2 B for SOURCE at DECLINATION 4. I 0 Fig. 23.--A 22.2 Mc/s Jupiter storm of mixed normal swishes and continuum.

PAGE 63

53 produced by a 22.2 Mc/s radio source at Jupiter, whose declination ranged from 18 to 21 South during the apparition [28]. During each of the 154 effective watches, fringes from Cassiopeia A were observed despite the antenna's being directed toward declination 15 South. (Recall that Cassiopeia A's declination is 58 32' North [26]). The flux density of Cassiopeia A at 22.2 Mc/sis approximately 4.6 x 1022 w/m 2 /cps [25]. For the 1962 apparition the phase-switching circuitry was used for another experiment and no 22B data were recorded. Channel 22B was used effectively during 67 watches of the 1963 apparition. During an additional 67 watches interspersed throughout the apparition, the equipment was turned on but was not functioning properly. In eight instances fringes due to radio sources other than Jupiter were observed and identified. During ten of the effective watches Jupiter storms at 22.2 Mc/s were observed on other 22.2 Mc/s equipment. Seven of these storms were observed as sporadic bursts on 22B, with some tendency toward fringing. However, in three instances the 22B record clearly.exhibited fringes as predicted for continuous 22.2 Mc/s Jovian radiation. Figures 23, 24, and 25 contain typical examples of 22B data with other 22.2 Mc/s records for comparison. In Fig. 23, utilizing data for September 26, 1963, the upper record shows Jupiter activity at 22.2 Mc/son channel 22Y, which features an automatic tracking Yagi antenna. The center record is data taken simultaneously on 22B at 22.2 Mc/s. Note between 0007 and 0235 E.S.T. the regular fringes which

PAGE 64

5 4 compare favorably in spacing with those of the lower trace, the pre dicted response of 22B to a radio continuum at Jupiter. The compar ison is made using time correlations of maxL~a and minima. Superposi tion of a tracin g of the lower curve on the 22B record in the center shows near coincidence of the maxima and minima, especially near an tenna boresight transit, which occurs later than meridian transit as noted previously. For some reason the westward offset of the 22B cen tral lobe from the meridian now measures 0.6 in contrast to Watson's 0 measurement of 1.9 This change is probably due to electrical path length changes introduced between the receiver and the East and/or West arrays of 22B during system modifications made since 1960, when Watson's data were taken. Apparently, in addition to the normal swishes recorded on 22Y during this intense Jupiter storm, there was emitted continuous radia tion at 22.2 Mc/s as shown by the fringes on the 22B record. The time constant of 22B was set at 10 seconds for most of the storm, and while no high speed oscillograph record was made at 22.2 Mc/s, the simultan eous Brush records at other frequencies fail to show pulses of this length. Additionally, when the time constant of 22B was set at 20 seconds between 0046 and 0052 E.S.T., the fringe pattern continued, despite the fact that this time was very close to the time for a null in the fringe pattern when the instrument was relatively insensitive. Final evidence of this continuous 22.2 Mc/s radiation is two periods during the storm when the swishes on_22Y stopped but the fringes of 22B continued. In Fig. 23 note the two lulls in the storm on 22Y at

PAGE 65

I I -'--} 0300 'U :U:0 200 u UNIVERSITY of FLORIDA 28 SEPT 1963 _, """"'tr" .. t JUPITER TRANSIT PRINCIPAL LOBE =H oou E.S J ,, u PREDICTED RESPONSE of 22 B for SOURCE at DECLINATION 4.1 F i g 24 .-A 22 2 M c /s J upiter s torm o f swi s hes with no continuum .cJOOD CJ7 CJ7

PAGE 66

(. I 0000 ---,,,--~ 2 ?,00 E S2 2 Y ~~1.. J,(,J.J I t I _-__r-0 ~ ,-f UNIVERSITY of FLORIDA 21-22 OCT 1963 RESPONSE of 2 2 B to CA S SIOPEIA "A" a t DECLINATION 58 5 JUPITER f TRANSIT / f 2 300 .. 1 2200 l~,h._ 22 v F= 1 i,' CA SS IOPEI A I A T RAN S IT if-. t \-_~ ,\ ":. i_ ~. "> 1 h < PRINCIPAL LOBE Fi g 2S.--A 22.2 Mc/s Jupiter stor m superi m posed on the C assiopeia A continuum. en (J)

PAGE 67

5 7 0026 to 0033 E.S.T. and 0133 to 0152 E.S.T., while the fringe patter n on 22B seems to continue during these periods, although perhaps reduced in intensity. Therefore, Fig. 23 apparently displays a Jupiter storm of combined sporadic and continuous 22.2 Mc/s radiation. Douglas and Smith have suggested the existence of this continuum [29]. Figure 24 shows an instance of strong sporadic Jovian activity on 22Y occurring approximately 1 hour after transit during the watch of September 28, 1963. Again at the top of the figure is the 22Y record, at the center is 22B, and at the bottom the predicted fringes for Jovian radiation. Note that the storm of September 26, 1963,shown in Fig. 23 produced fringes detected 1 hour 20 minutes before and after transit. If continuous Jupiter radiation within the sensitivity of 22B had been present along with the normal Jovian swishes on September 28, 1963, Fig. 24 should show signs of fringes from approximately 0200 to 0230 on the center trace, which is the 22B record. No such fringes are evident. The conclusion is that no cont i nuous 22.2 Mc/s flux from -23 ;. 2 Jupiter exceeded 1 x 10 w 1 m cps during this storm, at least while Jupiter was in the beam of 22B. Figure 25 also shows a Jupiter storm observed on 22Y which apparently did not have an accompanying continuum, the records having been produced October 21-22, 1963. The trace at the bottom of the fig ure is the response of 22B to Cassiopeia A, after Watson [24]. The top record is a Jupiter storm observed on 22Y. The center record shows fringes observed on 22B. Note the good match of these fringes to t h e lmown response of 22B to Cassiopeia A. While sporadic Jupiter bursts

PAGE 68

5 8 appeared on 22B, there is no evidence of fringes corresponding to the pattern predicted for radiation from Jupiter. Specifically, the group of four intense but separated peaks beginning on the 22B record at 2335 E.S.T. should form a smooth maximum, if the radiation is continuous in time. Clearly this is not the case, the high peaks being spaced at intervals of approximately 1 minute. The absence of the 22.2 Mc/s con tinuum is suggested by this data, the fringe due to Cassiopeia A being clearly evident despite the strong simultaneous Jovian storm. Unfortunately, most of the inoperative periods of 22B came near the end of the 196~ apparition and indicated the aging of the equipment. The system would not function at all for 16 trial watches during the 1964 apparition. The circuitry had become highly sensitive to temperature and power fluctuations. The Florida sun, humidity, and an occasional stampede of Brahma cattle had taken their toll of the antenna masts and fittings and 22B was shut down for repairs. The antenna is being rebuilt and a new, more stable phase-switching circuit is planned for future observations with the system. This is necessary since our competent engineer, G. F. Walls, had learned to maintain 22B in operative condition only at the cost of daily attention. The data gathered from 22B support the conclusion that Jupiter -23 does not continuously emit 22.2 Mc/s radiation exceeding 1 x 10 2 w/m /cps in flux density. The three instances of a Jovian continuum accompanying strong sporadic Jupiter activity, however, suggest that Jupiter storms at times consist of a combination of sporadic bursts with an underlying continuum. This result confirms and amplifies the

PAGE 69

5 9 findings of Stone, Alexander, and Erickson [30], who used data from the University of Maryland 26.3 Mc/s Christiansen-type array with a Ryle-Vonberg radiometer to conclude that Jupiter radio storms consist of at least two distinct components, characterized by different flux densities. At 26.3 Mc/s these flux densities were "just above" 1023 2 -23 2 w/m /cps and 9 x 10 w/m /cps, the stronger component corresponding to the radiation commonly observed by most other workers. At the time of their publication Stone et al. had not considered any fine struc ture details. It appears that the presently reported underlying con tinuum is in fact the we~ker component reported by Stone et al. and that this component of Jove's radio emissions either is a continuum or is composed of bursts longer than the order of 10 seconds, the time constant of the Florida interferometer when the pertinent data were taken. Perhaps the characteristics of the two components are not resolved in the Maryland data due to the relatively short integration time used, namely, 2 seconds. The sensitivity of the Maryland system -24 / 2; is 5 x 10 w m cps, a factor of 2 better than Florida 1 s 1025 2 w/m /cps. Floridats 22B obviously is marginal for further examination of the weaker component, and the sensitivity of 22B should be improved in order that this study be pursued with promise. Then of course, as Stone et al. suggest, the experiment should be extended to other frequencies.

PAGE 70

60 2. Jovian Burst Morphology a. Introduction Kraus [31], Gallet [32], Gardner and Shain [33], A.G. Smith et al. [34], H.J. Smith~[35, 36], and Mock and A.G. Smith [11], have all studied the fine structure of the Jovian storms. Gener ally there is agreement that the terrestrial ionosphere modulates the Jovian signals and is in part responsible for some o f their features as observed in Earth. In particular, A.G. Smith et al. [34] concluded that the overall burstiness of the radiation is in fact a scintillation caused by fluctuations occurring in the ionosphere. Then Mock and Smith [11], through an extensive statistical analysis of the durations of normal Jovian pulses received simultaneously in Florida and Chile, concluded that the pulses constituting the fine structure of the bursts came from the same population and that, therefore, this feature of the fine structure either originated in the Jovian environment or was due to a world-wide ionospheric effect of some indeterminate nature. Less complete data suggested this sa~e conclusion for the slow and fast Jupiter pulses. In this section is reported a study, the purpose of which is to determine the origin of the basic Jovian pulse structure, apparently a choice among the terrestrial ionosphere, the interplanetary medium, and the Jovian environment. The primary method of the analysis is to examine the variation of the character of the pulses with parameters such as time of year, time of day, f re quency, altitude of the planet, and geomagnetic activity. E ow, radio star scintillations are ap p arently

PAGE 71

6 1 ionospheric in origin, and their behavior in terms of the cited vari ables has been examined extensively by others [26, 37, 38, 39, 40, 41]. Similar scintillations of artificial satellite signals are well docu mented [42]. It is expected that if the basic Jovian pulse character is an ionospheric scintillation, this phenomenon will exhibit the same parametric variations as radio star and satellite scintillations. Lack of such correlation will indicate some other source for this fine struc ture, probably the Jovian environment, the alternative proposed in Mock 1 s thesis [11]. The comparison of Jupiter's basic pulse structure and ionospheric scintillation phenomena follows. b. The Jupiter Pulse Character Index The discussion is facilitated by the definition of the Jupiter pulse character index C. This index is described in Table l(a). For storms consisting of slow Jupiter pulses lasting a few sec ands or longer, C is defined as 1. Normal Jupiter pulses of approximately 1 second duration are assigned the index value C = 2. Fast Jupiter pulses of 0.1 second or less duration have the value C = 3. Refer ence to Fig. 2 of Sec. II.2, along with Table 1 of this section, should provide the reader with a suitable interpretation of C. In the actual record reduction, nonintegral values of C were used for cases of unclear average pulse length. For example, sometimes normal Jupiter has fast Jupiter superimposed. The record reducer uses his judgment in assign ing C for these cases the value 2 1/4, 2 1/2, or 2 3/4, depending on the relative amount of each type of pulse that is present. The possibility of confusing normal Jupiter (C = 2), with mixed slow

PAGE 72

TABLE 1 DEFINITIONS OF PULSE CHARACTER AND SCINTILLATION INDICES (a) Jupiter Pulse Character Index C Value of C 1 2 3 Description Slow Jupiter pulses a few seconds or long e r Normal Jupiter pulses approximately 1 second Fast Jupiter pulses 0.1 second or less (b) Satellite Signal Scintillation Index S Value of S 0 0.5 Description No irregular scintillation. Faraday rotation fading is regular. Scintillations do not exceed 50 per cent modulation. Faraday rotation fading is regular. 62 1.0 Scintillations exceed 50 per cent modulation. Faraday rotation fading clearly evident. 1.5 Some evidence of regular fading. Scintillations approach 100 per cent modulation. 2.0 Extreme scintillations. No regular fading observable.

PAGE 73

63 (C = 1), and fast (c = 5) pulses is slight. In all the data examined, slow and fast Jupiter were only rarely recognized to occur simultan eously. Other workers have devised scintillation indices for radio star and satellite signals. Those cited will not be redefined here. It will simply be noted that the scintillation index in each cited example ranges from O for no scintillations to some arbitrary positive value for maximum observed scintillations. Lawrence and Martin's [42] scintillation index for satellite signals is used later in Sec. III.5, and is described in Table l(b). c. Diurnal Variations Radio star and satellite signal scintillations are significantly more pronounced near local midnight than at other times of day [57, 42]. For example, Fig. 26 shows the typical diurnal variations observed in these scintillations. The upper plot, Fig. 26(a), shows Hewish 1 s scin tillation index versus local time for observations of radio stars [57]. The relatively sharp maximum near midnight, falling off rapidly towards sunrise and sunset, is indeed striking. Less striking, but present, is the maximum near local midnight in Fig. 26(b), a plot of Lawrence and Martin 1 s satellite scintillation index versus local time [42]. L~spec tion of the next four figures, Figs. 27, 28, 29, and 50, reveals no such maxima. Each of these figures is a plot of the Jovian pulse char acter (defined above) versus local time, as observed at 18 Mc/son the high-speed oscillograph records taken at Gainesville, Florida, and Maipu, Chile. Fig. 27, representing the 18 Mc/s data from Florida for

PAGE 74

X .,sr w a z .ro z 0 I<( _) .05 _) Iz (.) (/) 0 X .10 w a z z 0 1.05 <( _) _) 1z u I Y. 14 18 22 2 6 LOCAL TIME {H) {o) RADIO STARS 6 4 10 14 (/) o~-'-----------------------'-----18 24 6 12 LOCAL TIME {H) (b) SATELLITE SIGNALS Fig. 26.--The diurnal variation of ionospheric scintillations.

PAGE 75

w l<.> w Cl) :::> a. tJJ (9 <( 3 2 I FLORIDA 1962 18 MC/S """""__,._._,-.,,,J,....,m===-1 --~...!'--~~=-x/.,-=d..' -~=-=l. ==--'-"""""""'L~--w ==:.' =-==, ...\........m.,...J 18 20 22 24 2 4 6 8 LOCAL Tl ME (H} Fig. 27.--The diurnal variation of Jovian burst fine structure (Florida, 1962). (J') en

PAGE 76

3 w l(.) <( 0:: a.. w (!)
PAGE 77

3 FLORIDA 1963 oc 18 MC/S w I(.) <( oc <( :c (.) s l r w L I CJ) 2 L ... _J ::::> (L IJJ (.!) <( I I J .J ,.,,J. J t. = .J 18 20 22 24 2 4 6 8 LOCAL TIME (H) Fig. 29.--The diurnal variation of Jovian burst fine structure (Florida, 1963). (J) -.J

PAGE 78

0:: w lu <( 0:: <( :c u w (/) _J ::) a. w (!) <( w > <( 3 CHILE 1963 18 MC/S ..J _,J L-,________j 2 I 18 20 22 24 2 4 6 LOCAL TIME (H) Fig. 30.--The diurnal variation of Jovian burst fine structure (Chile, 1963). j 8 (J) m

PAGE 79

69 the entire 1962 apparition, has a marked mini~um near midnight. Figure 28, the 18 Mc/s data taken in Chile in 1962, sho~~ so me ten dency towards a minimum shortly after midnight, at 0200 local time. Figure 29 shows a broad weak minimum near midni gh t, t hi s data repre senting the 18 Mc/s Jupiter storms recorded at Florida during the 1963 apparition. Figure 30, 18 Mc/s, Chile, 1963, again shows some tendency towards a minimum at 0200. Now, admitted l y, the tendencies toward minima were perhaps inconclusive in these figures, but there certainly was no evidence of a maximum as observed by Lawrence and Martin in Fig. 26(b); and cer tainly nothing in these Jovian pulse character plots corresponds to the sharp increase in radio star scintillations near local midnight in Fig. 26(a). The data represented _in Figs. 27, 28, 29, and 30 are, therefore, interpreted as evidence that the Jovian pulse character is not an ionospheric scintillation phenomenon. d. Seasonal Variations The comparison drawn in this section parallels the prece9ing discussion. Figure 31(a) shows the seasonal variation of radio star scintillations as observed by Koster and Wright [39]. Note that max imum scintillation activity occurs near the autumnal equinox, while a lesser peak occurs near the vernal equinox. Lawrence and Martin observed the autumnal equinox peak in satellite scintillations, thou gh not the springtime peak [42]. Figures 32, 33, 34, and 35 are seasonal plots of the Jovian pulse character index of the 18 Mc/s data for the respective apparitions Florida 1962, Chile 1962, Florida 1963, and

PAGE 80

X Q) "I::, C: C 0 0 C 0 (/) )( Q) "O C J .05 Or-,,--r....,.--r--r....,..--r~-r---,-.....,...-,-...,.._ J F M A M J J A S 0 N D J (a) The Seasonal Variation of Scintillation .10 .05 0 C: 0 Cl') 0 "-------'--__., ______ .i..._~-900 Altitude of the Source (b) Altitude Variation of Scintillation Fig. 31.--The seasonal and altitude variations of radio star scintillations. 70

PAGE 81

3 FLORIDA 1962 0:: 18 MC/S l!J I(..) <.( er: <( :c (..) w Cf) 2 _J ::::) 0.. w (.!) <( 0: w OPPOSITION > I<
PAGE 82

3 CHILE 1962 18 MC/S w I(.) I <( I I I I I ~t='r = I i -= r=====r 1 MAR MAY JUL SEP NOV JAN Fig. 33,--The seasonal variation of Jovian burst fine structure (Chile, 1962).

PAGE 83

0: w I0
PAGE 84

3 w I(.) <( <( CHILE 1963 18 MC/S r=~-~---=1 ==~,"= 1 MAR MAY JUL I 1< OPPOSITIO N I I I I I I m l 1 = = r===7rw r== = -= -, SEP NOV JAN Fig, 35.--The seasonal variation of Jovian burst fine structure (Chile, 1963),

PAGE 85

75 Chile 1963. None of these plots exhibits the anticipated vernal equi nox peak. In fact, the only distinctive feature common to even three of the four plots is the November or December peak in Figs. 32, 33, and 34. Here is more evidence that, since Jovian pulse character does not exhibit the seasonal variations of radio star and satellite scintilla tions, Jupiter's burst fine structure is not a scintillation phenomenon. e. Frequency Dependence Radio star and satellite scintillation exhibits an inverse relationship with frequency as shown, for example, by Aarons [41] and Briggs and Parkin [38]. Figure 36, after Aarons (41], shows this relationship for scintillation on signals from Cygnus A. Note in Fig. 36 the increased SC?,-lltillations as frequency decreases from 1200 to 108 Mc/s. The Florida and Chile data for the 1962 and 1963 appar itions, plotted as average pulse character versus frequency in Fig. 37, do not follow such a relationship. In fact, from 15 Mc/s to 27.6 Mc/s the index actually increases, instead of decreasing as scintillation phenomena do. Once again, there is evidence that the Jovian burst fine structure is not a scintillation phenomenon. The Florida and Chile 1962 and 1963 data, plotted separately, yield curves similar in char acter to Fig. 37, and are not shown here. f. Altitude Dependence Hewish [37] showed that radio star scintillations increase with decreasing altitude of the source. Satellite scintillations behave similarly, as implied by Lawrence and Martin [42]. But once again,

PAGE 86

224 MC (.) RESET :;: 0 400 MC 0:: t.,_ en _J 1200 MC
PAGE 87

3 FLORIDA-CHILE 1962-1963 0:: w I(.) <( 0:: <( ::c (.) w (/) 2 _J ::> a.. w (9 <( w .> <( I ~=-==""""'"'==ch L== ,,,_=--~-l.._.-==,==e='-=='==-=====L===--~,,......J,=.,.,=====cl J 0 5 10 15 20 25 30 FREQUENCY (MC/S) Fig. 37,--Frequency dependence of Jov ian burst fine structure.

PAGE 88

7 8 Jovian pulse character does not exhibit this behavior. Figure 3l(b) is a plot, due to Hewish [37], of radio star sci.~tillations versus altitude. Figure 38 is the companion plot of Jovian pulse character versus altitude for the 18 Mc/s Florida data of 1962. Clearly, the Jupiter fine structure once again fails to match a known characteristic of the scintillation phenomenon. g. Geomagnetic Activity Dependence Hewish [57} and Brown and Lovell [26] cite the correlation between radio star scintillations and geomagnetic activity. In times of increased geomagnetic activity, stronger scintillations are observed. Figure 59 is a plot of Jovian pulse character versus the Geomagnetic Activity Index K. The pulse character actually appears to decrease p with increasing geomagnetic activity which is additional evidence that the pulse character is not an ionospheric scintillation. The dotted portion of the curve represents only 2 per cent of the data and is considered to be unreliable. h. Solar Activity Correlation Previously the Florida group and others [4, S, 8] have noted some correlation between Jovian decametric emission and solar or geo magnetic activity. N. F. Six [4] and G. R. Lebo [SJ have used a com puter to examine the correlation between Jovian activity and the para meters FXJ., FX2, FX5, FNl, FN2, FN3, SSN, A, and the solar 2800 Mc/s p flux density. The FX 1 s are solar flare activity indices for specific regions of the Sun [4]. The FN 1 s are solar flare numbers for those

PAGE 89

3 10 20 FLORIDA 1962 18 MC/S """00 ALTITUDE OF JUPITER 50 Fig. 38.--Altitude dependence of Jovian burst fine_structure. 79

PAGE 90

0::: w I(.) <.t
PAGE 91

81 same regions [4]. SSN is the Zurich sunspot number, and A is a stand p ard geomagnetic activity index. G. R. Lebo modified the solar f l are program so that the Jovian pulse character index C could be ex a mined for correlation with these same parameters. T h e computer perfo rm s a Chree analysis of each of these variables, and of C, itself. Figures 40 and 41 present the results of this study for data of the 1962 appar ition as recorded in Florida and Chile. Each separate curve in these two figures represents the output of the computer, programmed for Chree analysis, and shows how the variable under inspection behaves, on the average, during the period 35 days before to 4 days after Jovian activ ity of character index C. The average is taken over the total number of days of Jovian activity having a particular in d ex C. Figure 40 shows the results of the Chree analysis of sunspot number. The five curves represent average suns p ot number as a func tion of days relative to Jovian activity for various C 1 s and for the two stations listed. There were insufficient data for a meaningful study of C = 1 for Chile in 1962. The most interesting feature of this set of curves is the simultaneous appearance of strong peaks in the two curves for C = 3, Florida and Chile, 1962, at -6 and -23 da y s. Similar simultaneous minor peaks occur at -31 days. Note the absence of these peaks in the two curves for C = 2. The data for C = 2 do show peaking tendencies, but not nearly so strongly as the C = 3 curves. Likewise the single plot for C = 1 exhibits peaks at o, -13, and -28 days. These data are much more sparse than those f or t h e other curves, and no significance is attached to these peaks, especi ally in view of the height of the 0-day peak with respect to t h e others.

PAGE 92

82 40l 'C=31 ...,.., i -n I I ij CHILE 1962 I 1 ) 7 j ) -: C=3 1962 re 40 w 7 CD I -.:::351 J l 1-3 0 ~ 0 0... Cf) I z C=2 35 CHILE 1962 I FLORIDA 1962 30 C = I FLORIDA 1962 l 25 I l I L -29 -2 1 3 -5 0 +3 DAYS RELATIVE TO JOVIAN ACTIVITY Fig. 40.--Chree analysis of sunspot number.

PAGE 93

83 (\J X 25 1 -1.L 50 I ________ 2.0~~ u1.of6ce-r= DAYS RELATIVE TO JOV!AN ACTIVITY Fig. 41.--Chree analysis of FXl, FX2, FX3, C, A (all dimen-22 2 p sionless), and the 2800 Mc/s solar flux (10 w/m /cps).

PAGE 94

On the basis of the similarity between the Florida and Chile sunspot number correlations for C = 3 and for C = 2, and the differ ences between the two pairs of curves, it is concluded that fast Jupi ter (C = 3), does indeed depend in some way on sunspot number. According to E. N. Parker [43], solar particle flux consi s ts of at least two co mponents one reaching Jupiter in approximately 23 days, the other in 5 to 10 days. Clearly this i."1.formation lends credence to the peaks observed for fast Jupiter (C = 3), at -23 days and -6 days. Furthermore, these data suggest that the Jovian pulse character ori ginates at Jupiter, and that in fact, fast Jupiter is simply normal Jupiter pulses modulated in some way at Jupiter by turbulence intro duced in the ever-present solar w:L~d by sunspot activity (the 23-day effect), and by the fast particles ejected in flares occurring at times of sunspot activity (the 6-day effect). Figure 41 shows typical examples of the other correlations, none of which shows peaks at app r oximately th e same place, and one can integrate the three FX curves visually and see that the result is essen tially the same as the sunspot number correlation discussed above. The FX curves for C = 2 and C = 1 are not shown, since they behave sim ilar to the FX curves for C = 3 in Fig. 2. This discussion applies equally to the FN curves, none of which are sho-vm. Apparently r.o specific region of the Sun influences the formation of the basic Jovian pulse character. Each Chree analysis curve for C resembles the C curve in Fig. 41, having only the expected main peak at 0 days. No cyclical effects are observed.

PAGE 95

The peak at -5 days in the A data o f Fig. 41 is consid e red p spurious, since it does not show up on the other A plots wh ich, p 85 again, are not shown. The A data are consid e red to be inconclusiv e p Likewise, the 2800 Mc/s flux correlation appears to be inco n clusive, the curve in Fig. 41 being typical of each of the various plots. The slow decline from -23 days to +3 days is present, thou g h not so evi dent, in the other data. Whether this decline is a real effect, is as yet uncertain. i. System III Longitude Dependence Curiously, the basic pulse structure seems to exhibit a sli gh t System III longitude dependence, as Fig. 42 shows. In this plot appears Average Pulse Character versus System III Longitude, the data being that of Florida, 1962, at 18 Mc/s. Discounting the peaks and valleys outside the region 80 to 310, wherein lies 85 per cent of all the data, it appears that Source B (80 to 160) tends to produce more f t J t th S A d C 1 t d from 200 to 290. as up1 er an ources an oca e This is construed as positive evidence of the origin of the fine structure at the Jovian environment, for there appears here a correlation between this fine structure and the actual Jovian sources. j. Dependence on the Elongation of Jupit e r's Moon Io Since Jupiter7s activity is known to be related to the position of at least one of the Jovian moons, Io [44, 45], Fig. 4 3 w as pr e pared as a plot of Jovian pulse character versus the departure of Io fro m g eo centric superior conjunction. The data were for Florida, 1962, at

PAGE 96

a:: w 1(.) <( a:
PAGE 97

3 a:: w ..... u <( <( 0 50 0 100 FLORIDA 1962 18 MC/S 0 150 200 Io 0 250 0 300 DEPAHTURE FHOr v1 GEOCEf\JTRIC SUPERIOR CONJU~ JCTION Fig, 43.--The influ e nce of Jupiter's moon Io on Jovian burst fine struc tu r e 0 350 0) -J

PAGE 98

88 18 Mc/s. In considering only the shaded areas as significant data, there is seen to be some tendency towards faster Jupiter when Io i s near 85. In the other favored position, 2 40, there is a slight ten dency towards slower Jupiter. RGalizing that Io's position and the Jovian System III longitude combine to influence the probability of observing intense Jupiter activity, one can only speculate as to which parameter, if either, causes differences in the fine struct ure. On the basis of this evidence, the longitude dependence is favored as a source of fine structure characteristics, since the variation of pulse c harac ter seems to be more clearly defined in the longitude plot, Fig. 42.

PAGE 99

3. A Comparison of Jovian Rad-io B ur st s w i t h A rtifici a l Satellite Signals at A ppu lse '-~ 89 The successful launching of NASA's beacon satellite S-66 presented an opportunity to determine the extent of t he influe n ce o f the terrestrial ionosphere on radio signals, and hence to s~ ppl ement the data presented in Sec. III.2, which i n dicate that the basic Jovian pulse structure originates either in the Jovian environm e nt or in the interplanetary medium. The satellite orbits at approximately 6 0 0 miles altitude, well outside the ionosphere. Aboard are several C W radio frequency beacons, including one operating at 20.005 Mc/s. Since S6 6 first achieved orbit on October 10 1964, the University of Florida radio astronomers, as part of the regular decametric-wavelength Jupiter observations have recorded near-simultaneous signals from S-66 and Jupiter at appulse, in hopes of measuring the correlation bet w een the basic Jovian pulse character and the ionospheric scintillations ob served on the satellite signal. The apparatus for the experiment is relatively simple. Two identical 20 Mc/s, 3-element Yagi antennas separately feed crystal controlled receivers. One of these channels, known as 20YJ, monitors Jupiter in the normal AM mode. The other, 20YS, mo n itors S6 6, usin g single sideband reception to provide aural monitoring. Thus 2 0YS is set at, say, 20.0055 Mc/s (lower sideband), and the audio output of the receiver is a fading 500 c/s tone, easily heard by th e observer. This work has been supported under Proje ct AOl of t h e Un iversity of Florida NASA Institutional Grant No. NA S G5 4 2.

PAGE 100

90 The audio envelope of the Jupiter activity and S-66 signal beat tone are displayed side by side on a Brush dual-cha r...n.e l oscillograph, gen erally operated at a chart speed of 5 mrn/s. Channel 20YJ is detuned from the S-66 carrier to approximately 20.015 Mc/s, so that the Jupiter record will contain no interference from t ~e beacon. In the absence of 20 Mc/s Jovian activity the other regular channels ranging from 1 0 to 55 Mc/s may be displayed on the Brush in place of 20YJ, permitting the observer to make use of any Jupiter storm available during an S-66 pass. In principle the effects of Jupiter activity superimposed on the S-66 record could be removed by some subtraction scheme. Thus far the devel opment of this technique has not been necessary, since no records have been sufficiently good to contain this interference. Usually the Jupiter activity has either preceded or followed the S-66 pass by a few minutes, or the bandwidth of the Jupiter storm has not included the beacon carrier, and Jupiter-beacon interference has been no problem. In the same manner as described in Sec. III.2, the pulse char acter is described quantitatively by an index C which is assigned the value 11 1 11 for slow Jupiter, 11 2n for normal Jupiter, and 11 5 11 for fast Jupiter. Correspondingly, the S-66 signal scintillations are assigned an index S similar to that used by Martin and Lawrence [42] and ranging from 0 for no detectable scintillations to 2.0 for extreme scintilla tions. The two indices are described in Table 1 of Sec. III. 2 Figure 2, Sec. II.2, shows examples_of Jupiter pulses described by C = 1 (top), 2 (center), and 5 (bottom). Figures 44 and 45 illustrate S-66 signals with scintillations described by various S values.

PAGE 101

9 1 Figure 44 shows two portions of the dual-trace oscillo graph recording obtained during an S-66 pass on January 17, 1965. From the top, the four traces represent,respectiv ely, 20YJ near 1813 E.S.T., the S-66 signal near 1813, 20YJ near 1807, and S 66 near 1807. There was no detectable 20.015 Mc/s Jupiter activity during the time cov ered by the figure. The purpose of the figure is to show examples of satellite signal scintillations of different magnitude. From 1806:4 4 to 18 06:54, S = 1.5. Time progresses towards the right, as before. After 1 807, S = 2.0. Then, near 1813, S = 0. Figure 4 5, a similar record contin uous from 0213: 35 to 0215: 34 E .s. T. on November 2G, 19 64., co1 :i. tains examples of S = 1.5 near 0213:30, S = 1 from 0214:25 to 0214:50, and S = 0.5 after 0214:50~ Thus the figures contain exa mp les of all values of the scintillation index S describ ed in Table 1. The 22 Mc/s Jupiter activity shown in Fig. 45 will be discussed after the explanation of the presentation of the data. First note that the index C used to describe Jupiter 1 s storms is essentially a measure of frequency, while the scintilla tion index S is a measure of amplitude fluctuations. In justification of drawing a comparison between Jupit er activity and satellite scintillations using the defined indices, Fig. 46 is presented. Here is shown a scat ter diagram of Scintillation In dex S versus Scint illa tion Frequency for the data which will be used presently. The marked correspondence between Sand the Scintillation Frequency, the tendency for the points to line up along a diagonal line f ro m lower left to upper ri g ht, is evidence supporting the validity of the co mparison which follo w s.

PAGE 102

FLORIDA '-2 0. 015 MC .IS JllPlIEB FLORIDA 2--G. Ol-5--M 6-/ s J U PIT R I -t17JAN.1965 ROTATION FADING 17 JAN.1965 -+4-ISEC. t...AOJUST GAIN 1 807 EST Fi g 44 .-Ext re m e ex a m ples o f sci nt illa t i on i n d ex S At 1 8 13 E S T ., S = 0 At 1 807 E S T ., S = 2. 0 CD N

PAGE 103

~ Fig. 45 .-A direct co m parison of Jovi a n burst s t ructure with satellite signal sci n ti lla tion s. Near -a ppu lse time is 0215 E S T At 02 13: 30, scintillation index S = 1.5. At 0214, S = 1.0. At 0215, S = 0 .5. Jovian pulse character i ndex C = 1.0. w CJ1

PAGE 104

z en 0 >< _J LJJ ::!~ 1z 0 (j) 2:... X X X X X ){ X )<( 3) X )( SCI NTI LLATI ON FREQUENCY (C/S) Fig. 46. The dependence of the scintill ation index Son scintillation frequency. tO ,r,.

PAGE 105

95 Consider Fig. 47(a), a hypothetical scatter diagra m of C versus S. If the in dex C is truly representative of an io~1ospheric scintillation pheno men on, then simply tallying each appulse in the proper portion of Fig. 47(a) should tend to fill in the shaded area. For example, an instance of no observed satellite signal scintillation (S = O), should correspond to the long Jupiter pulses (c = 1). Like wise, strongly scintillated S-66 signals (S = 2), should correspond to the short Jupiter pulses (C = 3). A moderate amount of ionospheric scintillation (S = 1) should produce normal swishing Jupiter (c = 2). Table 2 lists the t wen ty observed a p pulses, with such perti nent data as date, minimum angle subtended by S-66 and Jupiter during the observation, time of appulse, time interval between the Jupiter activity and appulse, the S-66 20.005 Mc/s signal scintillation index S, the Jupiter pulse character index C, and the frequency of the observed Jupiter activity. The large_range of minimum angular subtence indicates the difficulty of performing the eA--periment and the quality of the statistics of the data. At best the data can only reliably sup port general qualitative conclusions. Figure 45 is an example of the data taken during this e :)er i ment. The upper trace is the record of the S-66 20.005 Mc/s signal as received on single sideband. The sideband mode is necessary for aural monitoring of the CW signaL The lo wer trace shows J upiter activity heard simultaneously at 22 Mc/s. Note th at the character index for the Jupiter activity is 1, while the scintillation in cax for the satellite signal is O.S at appulse, occurring just aft e r 0215 E.S.T. [~6].

PAGE 106

(_) ,. X w a z 3 1.5 I 2.6'1 SCINTILLATION INDEX S (a} HYPOTHETiCAL CASE: JOVIAN PULSE CHARACTER FOR ~ 1ED IN IONOSPHERE I XXX i I vv I XXX "/'I. XXX xx L~J I SCINTILLATION (b) ALL DATA X 1 INDEX S .J 31 XA i -,;-, --;:--i -' -~ 2 Xs 11 11 I Xe i 1 i i ij ~nJ I B 8 0.5 1 1.0 I 1.5 I 2.0 SCINTILLATION IND2X S (C) BEST DATA o m 1 'I 15 AC i Fig. 47.--Scatter di a grams of Jovian pulse character C versus scintilla t ion index S. 96

PAGE 107

97 TABLE 2 THE GENERAL COMPARISON OF JOVIAN ?ULSE CHAR.L\.CTER WITH ARTIFICIAL SATELLITE SIGNAL SCINTILLATIONS NEAR .APPULSE Minimum Time of Time Angle Jupiter Between Jupiter Date Between Activity Activity s C Frequency Jupiter (E.S.T.) and (Mc/s) and .Appulse S...-66 Begin End Oct. 20, 1964 47 0727 0734 7m o.s 2 20 Oct. 31, 1964 lo 0702 0703 +6m 2.0 1 20 Nov. 20, 1964 710 0211 0218 0 1.0 1 20 N ov. 27, 1964 43 0240 0240 +lh02m 1.5 1 20 Dec. 8, 1964 88 2313 2318 33m 2.0 3 18 Dec. 10, 1964 33 2327 2332 +Sm o.o 2 15 Dec. 16, 1964 35 2248 2259 0 0.5 2 20 Dec. 17, 1964 11 2328 0005 _gm 0.5 3 10 Dec. 19, 1964 19 0030 0051 42m 1.0 1 18,22 Dec. 31, 196 4 90 2100 2100 +3m o.o 2 20 Jan. 1, 1965 50 2311 2325 -lm 1.5 2 18 Jan. 5, 1965 56 2050 2115 +15m 0.5 2 18,22 Jan. 5, 1965 28 0 1905 1917 +28m 1. 0 2 18 Jan. 7, 1965 43 2033 2042 0 0.5 2 20 Jan. 8, 1965 50 2058 2237 0 0.5 3 10,15 Jan. 11, 1965 25 1926 2110 2"m 0 o.o 2 20 Jan. 11, 1965 55 1926 2110 0 0.5 3 10 Jan. 1 2, 1965 11 1942 1955 -1? o.o 2 10 Jan. 15, 1965 21 1919 2050 -15m o.o 2 10 Jan. 17, 1965 27 1920 1932 +lh06m o.o 2 18,20

PAGE 108

98 The regular Faraday fading is quite evident and there is superimposed scintillation of approximately 50 per cent modulation. Note ag a in that one would expect a character index of more than l w hen iono sp heric scintillations are present, if the basic Jupiter pulse character is ionospheric in origin. Figure 47(b) is a scatter diagram showing the correspondenc e between C and S for the series of observations bet w een October 10, 196 4 and January 31, 1965. Note that the results are not at all as expect e d (Fig. 27(a)), if the Jovian pulse structure were of ionos ph eric ori g in. Instead of grouping along the diagonal line from lower left to upper right, the tallies seem to do just the opposite, grouping about the diagonal from lower right to upper left, although the pattern is, to be sure, not precisely defined. However, this trend is maintained in Fig. 47(c)), in which appear only those tallies for the four instances when the minimum angular subtense of S-66 and Jupiter did not exce e d 11, and the time between Jupiter activity and appulse did not exceed 15 minutes. These characteristics appear at the right-hand side of this diagram. Since the data do not fall into place as sugg e sted in F ig. 4 7(a), the results are interpreted as implying t hat the basic Jovian pulse structure is not an ionospheric scintillation phenomenon, and that in fact, the basic Jovian pulse character is determined at Jupiter or in the interplanetary medium, in agreement with other pulse character studies discussed above in Sec. III.2.

PAGE 109

CHAPTER IV CONCLUSION The intent of the studies described in the preceding chapters was primarily to furnish new infor mat ion to aid th e theorists in their search for a suitable physical model of Jupiterls decametric radio emissions. Specifically, it was hoped that the influence of the ter restrial ionosphere on the Jovian signals would be determined in order to recognize those features of Jovian spectra and burst morphology which are due to ionospheric modification, and hence need to be ex plained as part of the nonthermal Jupiter radiation mechanism. The major feature impressed on the Jovian bursts by the ter restrial ionosphere seems to be the Faraday rotation seen clearly in our spectra. Thus a by-product o'f this work has been the calculation of the frequency dependence of ionospheric Faraday rotation in the 13 to 20 Mc/s region of the spectrum. The added ability to predict probable times of intense Jupiter activity through a knowledge of Io 1 s influence should aid in the future gathe.ring of spectra for such pur poses as refining and extending the Faraday rotation calculation. Thus at times of expected Jupiter storms, an extra observer might devote his time fully to the intelli gent collection of spectra, which, let the author assure you, is a full-time job A second observer wo uld provide other meaningful ~xperiments with a better chance for success. Through such efficient use of a special observer the spectra collection 99

PAGE 110

100 could presumably be improved significantly, so that a meaningful sta tistical analysis of bandwidths and frequency drifts of the spectral features would be feasible. The most significant contribution of this work is the extensive evidence indicating that the Jovian burst fine structure originates out side of the .Earth's ionosphere Hence, unless the occurrence of slow, normal, and fast Jupiter can be shown to be due to a modulation occur ring in the interplanetary medium, this phenomenon is an additional clue to the nature of the Jupiter nonthermal radiation mechanism. The author believes that the Jovian burst fine structure forms in the Jovian environment. .Although there is lmown turbulence at the boundary between the Earth 1 s magnetosphere and the solar particle flux, the interplanetary medium is rather tenuous and one finds it difficult to imagine its weak fields (approximately 5 x 10-S Gauss [ 47]),and low density fluxes (approximately 10/cm 5 [ 48]), creating modulations com parable to the steep-wavefront pops and sputterin g spits of fast Jupi ter. Indeed, such strong scintillation s have not been observed at Florida in the artificial satellite signals traversing the ionosphere. On the other hand, the correlation shown in Sec. III.2.h between fast Jupiter and the arrival at Jupiter of two different components of the solar particle flux suggests either that t he dumping of these particles into the Jovian environment imposes a modulation on the signals, or that these particles themselves are accelerated by Jupiter 1 s magnetic field to become the fundamental radiators of the Jovi a n storms, in which case the Jovian burst structure reflects the turbulence of the solar particle flux itself.

PAGE 111

LIST OF REFE RE NCES 1. Burke, B. F ., and Franklin K. L "Obs e r v atio n s o f a Va r iab l e Radio Source Associated with the Planet Ju p iter, 11 Journ a l of G e physical Research 60, 213 (1955). 2. 3. 4. s. 6. 7. 8. 9. 10. 11. 12. Carr, T. D., Studies of Radio Fre q u e ncy R adi a t i on s from t he Planets (Ph.D. Dissertation, University of Florida, 195 8) Chatterton, N. E., S ectral Characteristics of the Ra d i o Frequency Outbursts of the Planet Jupiter Ph~D. Dissertation, U ni ve rsi t y of Florida, 1961). Six, N. F., Jr., Anal sis of the D e ca m et e rW a ve le ngt h Radio Emis sion from the Planet Jupiter Ph.D. Diss e rtation, U nivers i ty o f Florida, 1963). Lebo, G. R., Decameter-Wavelen th Radio Obs ervation s of the Pl a nets in 1962 (Ph.D. Dissertation, University of Florida, 1 964 Fagerlin, M. L., Develo ment of a Lo -Period i c A nte nna fo r Decameter Wavelength Radio Astronomy M .S. Thesis, Universit y of F lo r i da 1963). Smith, A. G., 11 Radio Spectrum of Jupit e r, 11 Science 1 3 4 5 8 7 (1 96 1 ) Carr, T. D., Smith, A. G., Bollha g en, H., Six, N F., and Chatte r t o n N. E., 11 Recent Ilecameter-Wavelength Observ a tions o f J upite r, Saturn, and Venus," Astrophysical Journal 1 34 105 ( 1 96 1). Smith, A.G., Carr, T. D., and Chatterton, N. E., 11 Spectr a of t he Decameter Radio Bursts from Jupi te r, 11 Proc e edi ngs of the Twelfth International Astronautical C on gres s (Ac ad e m ic P r e s s I nc ., New York, 1963), p. 689. Carr, T. D., Brown, G. W., Smith, A .G., H i gg in s C. S ., Bo l lhagen, H., May, J., and Levy, J., 11 Spectral Distribution of Deca: m etri c Radiation from Jupiter in 1961, 11 A strophysical Jour na l 1 40 ( 2), 778 (1964). Mock, Willis, Jr., A Comparis o n o f Hi g hSpeed Re c o r din g s o f Jovian Radio Sign a ls Received i n Fl o r ida and Chil e (Maste r 1 s Thesis, University of Florida, 1 963 ). Murray, w. A. S ., and Hargreaves J. K ., 11 Lun a r Ra dio Echoes and the Faraday Effect in the Ionosphere, 11 Natur e 17 3 9 44 ( 1 954) 101

PAGE 112

13. 14. 15. 16 .. 17. 18. 19. 20. 2L 22. 23. 24. 25. 102 Browne, I. C., Evans, J V. Hargreaves, J K. and Murray, W A. S "Radio Echoes from the Moon," Proceedings of the Physical Society of London 69B, 901 (1956). Evans, J. V., "The Electron Cont en t of the Ionosphere,n Journal of Atmospheric and Terre s trial Physics 11, 259 (1957). Bowhill, S. A., "The Farad ay -rotation Rate of a Satellite Radio Signal, 11 Journal of Atmospheric and Te rr estrial Physics 1 3 175 (1958). Little, C. G., and Lawrence, R. s., "The Use of Polarization F a ding of Satellite Signals to Study the Electron Content and Irregu larities in the Ionosphere," Proceedin s of the First Int e national Space Science Symposium North -Holl and Publishing Co., Amsterdam, 1960), p. 340. Garriott, O. K., "Ionospheric Electron Content and Distribution Determined from Satellite Observations," Pro c eedings of the First Intern at ion al S ace and Science Symposium (North Holland Publishing Co., Amsterdam, 1960, p. 371. Blackband, W. T., "The Determination of Ionospheric Electron Con tent by Observation of Farad ay Fading," P r oceedin gs of t he Fi r st International Space Science Symposium (North-Holland Publishing Co., Amsterdam, 1960), p. 387. Warwick, J. W., and Dulk, G. A. "Farada y Rotation on Decametric Radio Emissions from Jupiter, 11 Science 380 (1964). Warwick, J. W ., 11 Dynamic Spectra of J up iter's Decametric Emission, 1961," Astrophysical Journal 137, 41 (1963). Warwick, J. W., and Gordon, M. A ., "Frequency and Polarization Studies of Jupiter's Decametric Emiss ion," 118th Meeting of the American Astronomical Society, 1 9 6 5 (Abstracts.) Riihimaa, J. J ., "High-resolution Spectral Observations of Jupiter 1 s Decametric Radio Em ission," Nature, 476 (1964). Perkins, W. H., A High Gain Radio-Frequency Interf er om ete r A n tenna for Plan eta r y Observations ( Master 1 s Thesis, Univer 3 ity o f Florida, 195 9) Watson, R. C., The Design and Construction of a Ph a se-S 1ii. tchin g Radio-Frequency In terferometer (Master's Thesis, U niversi t y of Florida, 1960). White J.E. Jr., Calibration and Applic 2 tion of a Rad i o Fre~~ency I~terfer~met er (Master's The sis, University of Florida, 1961 )

PAGE 113

26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 103 Brown, R. Hanbury, and Lovell, A. C. B., The EXPloration of Space by Radio (John Wiley and Sons, Inc., New Y or k 1957). Pawsey, J. L., and Bracewell, R. N., Radio As tronomy (Ox fo rd University Press, 195 4). U.S. Government Printing Office, The Ame.r ican Ephemeris and Nautical Almanac for the Yea r 1 96 1 (1959) Douglas, J. N., and Smith, H.J., 11 D e ca matr ic Rad iatio n f rom J up it er I. Synoptic Observations 1957-1961, 11 A str onomi cal Journal 68 163 (1963). Stone, R. G., Alexander, J. K., and Erickson, W. c., 11 Lo w -Lev e l Decameter Emissions from Jupiter,n Astrophysical Journ al 1 40 374 (1964). Kraus, J. D., "Some Observations of the Impulsive Radio Signals from Jupiter, 11 Astronomical Journal fil, 182 (1956). Gallet, R. M., nThe Results of the Observations of Jupiter 1 s Radio Emissions on 18 and 20 Mc/s in 1956 and 1957, 11 IRE '.::rans actions on Antennas and Propagation, 327 (July 1957, Abstr a ct). Gardner, F. F., and Shain, C. A., 11 Future Obs e rvations of Radio Emission from the Planet Jupiter, n Australian Journal of Physics 11, 55 (1958). Smith, A. G., Carr, T. D., Bollhagen, H., C ha tt e rton, N., and Six, N F., Jr., 11 Ionospheric Modification of t he Radio Emis sion from Jupiter, 11 Nature 187, 5 68 (1960). Smith, H. J ., Lasker, B. M., and Douglas, J. N., 11 Fin e Structure of Jupiter's 20 MC Noise Storms, 11 Astronomical Journ a l ._, 501 (1960). Douglas, J. N., and Smith, H.J., npresence and Correlation of Fine Structure in Jovian Decametric Radiation 11 Nature 1 92 741 (1961). Hewish, A., nThe Diffraction of Galactic Radio Waves as a Method of Investigating the Irregular Structure of the Ionosphere,n Proceedings of the Royal Society A214, 494 (195 2) Briggs, B. H., and Parkin, I. A;, 11 0n t he Variations of Radio Star and Satellite Scintillations with Zenith Ang le 11 Journ al of Atmospheric and Terrestrial Physic s 25 339 (1963).

PAGE 114

10 4 39. Koster, J. R., and Wrigh t R. W ., Rad io Star Scintillations and Associated Effects in Equatorial Regions, 11 Radio Astronom : ~c a l Studies of the A tmos phere Edited by Aarons, J. (North Holland Publishing Company, .Amsterdam, 1 963). 40. Booker, H. G., nThe Use of Radio Stars to Study Irregular Refrac tion of Ra d io Waves in the Ionosp here ," Proceedings of the IRE 46, 298 (1958). 41. Aarons, J nLow Angle Scintill at ions of Dis cr ete Sources, 11 Radio Astronomical and Satellite Studies of the Atmosph~, Edited by Aarons, J. (North-Holland Publishing Company, .Amsterdam, 1963). 42. Lawrence, J. D., Jr., and Martin, J. D., nDiurnal, Seasonal, Lati tudinal, and Height Variations of Satellite Scintillations,n Journal of Geophysical Research .2.., 1 293 (1 964). 43. Parker, E. N., nThe Solar Wind," Scientific .Ame rican 210, 66 (1964). 44. Bigg, E. K., llinfluence of the Satellite Io on Jupiter 1 s Decametric Emission," Nature 203, 1008 (1964). 45. Lebo, G. R., Smith, A.G., and Carr, T. D., ncorre l ation of the Longitudes of the Galilean Satellites with Ju piter 1 s Decametric Radiation,tt 118th Meeting of the American Astronomical Society, 1965. (Abstracts.) 46. Goddard Space Flight Center Orbit Bulletin for S-66 (BE-B), October, 1964 through January 1 965. 47. Coleman, P. J., Jr., Davis, Le verett, and Sennett, C. P., 11 Steady Component of the Inter p lanet ary Magne tic Field: Pioneer V, 11 Physical Review Letters, 43 (196 0) 48. Kellogg, P. J., 11 Flow of Plasma around t he Earth,n Journal of Geophysical Research 67, 3805 (196 3)

PAGE 115

BIOGRAPHICAL SKETCH Wilbur Frank Block was born January 2, 1930, in Jackson, Michiga~ to Herman H. and Edna Haft Block. He has t w o brothers and two sisters--Burton Peter, Eunice Margaret Fassbend er, David Herman, and Catherine Ellen Silas. He attended primary schools at Chesaning, Michigan, and Fountain City, Tennessee, and was graduated from Central High School at Fountain City in 1947. After attending Cornell University for t w o years, he transferred to the University of Florida whe re he received the degrees of Bachelor of Science and Master of Science in 1951 and 1953, respectively. In 1953 he married Jo Ann ~eal, the d aughte r o f Jo hn and Gladys Beal of Gainesville, Florida. The author then served as an instructor of physics at We stern Kentucky State Coll ege in Bowling Green, Kentucky, for three years (1953-1956). From 1956 t o 1958 he was employed by the Sperry Rand Electronic Tube Division of Gainesvill e Florida, whe re he assisted in the development of klystrons and magnetrons and stu died plasma oscillations in electron beams. From 1958 to 196 0 he served the Martin Aerospace Division at Orlando, Florid a, as a r adar systems engineer, devoting some time to laser research. Returning to the University of Florida in 1 960, the author pur sued the degree of Doctor of Philosoph y in physics for t wo years under a Martin Research Task Award. After about three additional years of 105

PAGE 116

10 6 study he is a candidate f o r this degree in Apr il, 1965. A part of his duties as a research assistant at the University of Florida consist ed of operating and maintaining a radio observ atory in the Andes at Huanta, Chile. A s senior author he has presented five papers to the F lo rida Academy of Sciences and was co-author of the Phipps-Bird awardwinning article in the 1963 Florida Academy of Sciences Quarterly Journal. Mr. Block is a member of Sigma Xi, the Ameri can Physical Soc iety, the American Geophysical Union, and the Florida Academy of Sciences. Wilbur and Jo Ann Block are members of the Episcopal C hurch and they have four children--Herman Karl, Mary Edna, John Beal, and Grace Ellen~

PAGE 117

This dissertation was prepared under the direc tion of the chairman of the candidate's supervisory committee and has been approved by all members of that committ ee It was submitted to the Dean of the College of Arts and Sciences and to the G~aduate Council, and was approved as partial fulfillment of the requirements for the degree of Doctor of Philosophy. April 24, 1965 Dean, Gradua te School Supervisory Committee: a/ii w t ~ Chairman I K C I / _,1 ,:1_ I~