Citation
Decameter-wavelength radio observations of the planets in 1962

Material Information

Title:
Decameter-wavelength radio observations of the planets in 1962
Creator:
Lebo, George Robert, 1937- ( Dissertant )
Smith, Alex G. ( Thesis advisor )
Carr, Thomas D. ( Reviewer )
Dunnam, F. E. ( Reviewer )
Omer, Guy C. ( Reviewer )
Blake, R. G. ( Reviewer )
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Copyright Date:
1964
Language:
English
Physical Description:
xv, 214 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Apparitions ( jstor )
Calibration ( jstor )
Electromagnetic noise ( jstor )
Flux density ( jstor )
Histograms ( jstor )
Index numbers ( jstor )
Jupiter ( jstor )
Longitude ( jstor )
Solar flares ( jstor )
Sunspots ( jstor )
Dissertations, Academic -- Physics -- UF
Physics thesis Ph. D
Planets -- Observations ( lcsh )
Radio astronomy ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Abstract:
The decameter radio emission from the planet Jupiter, first observed accidentally by Burke and Franklin in 1955 (l), has been monitored by the radio astronomy group at the University of Florida since 1957. Such extensive "coverage" can be justified merely by noting the existence of numerous conflicting theories purporting to explain this strange electromagnetic phenomenon (2), (3), {h), (5), {6), (7). These theories will be discussed in a later chapter. The results of the 1962 observations are offered here as experimental evidence to aid in deciding which, if any, of these theories is correct. This analysis must, of course, be considered as a continuation of the analyses from previous years by T. D. Carr (8), N. E, Chatterton (9), and N. F. Six (10). Details of data-taking, execution of the watch, and data analysis that were not changed for the I962 apparition (observing season) are not discussed here, since N. F. Six (10) covered these in his dissertation. During the 1962 apparition observations were also made of Mars, Venus, and Saturn. --INSTRUMENTATION-- The 1962 observations were made from two sites: one at the University of Florida campus in Gainesville, Florida, and the other near Maipu, Chile, about twenty miles from Santiago. At the Chile station, observations of Jupiter were made in I962 at frequencies of 5, 10, 15, 16, 18, 22.2, and 27.6 Mc/s, while simultaneously observations were being made at frequencies of 15, 18, 22.2, and 27.6 Mc/s in Florida. At each frequency an antenna was connected to a commercial communications receiver and two pen recorders. One of these, a low-speed recorder, ran during the entire listening time at a chart speed of six inches per hour; the other, a high-speed recorder, ran only when Jovian activity was received. The chart speed of the latter recorder was 5 mm/sec, 25 mm/sec, or 125 mm/sec, depending on which speed was needed to resolve the pulses. The data analysis, as discussed in this paper, primarily involves data taken on the low-speed Texas recorders, except at 15 Mc/s (Chile) and 27.6 Mc/s (Chile) for which the low-speed data were taken on Esterline-Angus recorders.
Thesis:
Thesis - University of Florida.
Bibliography:
Bibliography: leaves 210-213.
General Note:
Manuscript copy.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
022266676 ( AlephBibNum )
13554772 ( OCLC )
ACZ2000 ( NOTIS )

Downloads

This item has the following downloads:


Full Text











DECAMETER-WAVELENGTH RADIO

OBSERVATIONS OF THE PLANETS

IN 1962












By
GEORGE ROBERT LEBO











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
August, 1964













ACGIO 7LEEJDGIEITS


The author wishes to express deepest gratitude to his commit-

tee chairman and adviser Dr. Alex G. Smith for his capable and patient

guidance. Without his help this project would not have been possible.

Thanks are also due Dr. Thomas D. Carr for many helpful suggestions

and discussions, and Drs. F. E. Dunnyam, G. C. Omer, and R. G. Blake for

serving on his committee.

The data in this thesis could not have been collected by one per-

son. The author wishes to thank the following staff members for carry-

ing out routine observations during the 1962 apparition:

T. Anderson R. J. Leacock

J. Aparici J. Levy

W. F. Block E. J. Lindsey

H. Bollhagen J. ay

G. W. Brown W. Mock

I. Cain C. II. Olsson

T. D. Carr I. Shever

I.. L. Fagerlin I'. F. Six, Jr.

S. Gulkis A. G. Smith

A. T. Jusick C. F. Tiberi

G. Walls.

R. R. Hayward, C. A. Arlington, and S. L. Danielson proved to

be of great assistance in reducing the data, and W. W. Richardson de-

serves highest commendation for his untiring work on the drawings.










The expert editorial and typing services rendered by Mrs. Ruth Pierce

lightened the publication burden immeasurably.

The author also wishes to thank R. A. Smith and the rest of

the staff at the University of Florida Computing Center for their ex-

pert advice in writing the computer programs. Without the use of the

I. B. M. 709 facility at the University of Florida Computing Center, the

data analysis in this thesis could not have been undertaken.

The writer expresses thanks to Dr. S. S. Ballard, head of the

Physics department, for supporting him in the form of teaching assist-

ships and to the U. S. Army Research Office Durham, the National

Science Foundation, and the Office of Naval Research for their aid in

the form of research assistantships.

The author is indeed grateful to his parents and to his wife's

parents for their encouragement.

The author's wife deserves more than appreciation for her pa-

tience, encouragement, and hard work during six years of school life.

Her devotion and understanding helped as nothing else could. It is to

her that this thesis is dedicated.


iii













TABLE OF COITEIITS


Page


ACKI uEL ITS . . . . . . . . . . . . ii

LIST OF TABLES . . . . . . . . .. . .. vii

LIST OF FIGURES ................ ...... .viii

Chapter

I. IIITRODUCTIOl. . . . . . . . . . . .. 1

Instrumentation . . . . . . . . .. 1

II. CALIBRATIII PROCEDURES. . . . . . . . . 4

Easic Concepts . . . . . . . . .. 5
The Florida Calibrator . . . . . . . 7
The Chile Calibrator . . . . . . . . 7
Calibration Using the Chile Calibrator . . . 9
Use of the Calibration . . . . . . ... 9
Receiver Characteristics . . . . . ... .16

III. ANALYTICAL PROCEDURES. ..... . . . . . 21

Data Reduction ...... . . . . . ..

Data Reduction as Performed in 1957-i901
and Continued in 192 "Old Analysis" . 21
Data Reduction as Started in 1961 by Brown
and Continued in 1962 "ilew Analsis" . . 23

Jupiter Programs for the I.B.l-M. 709. .. . . ..

The "Old Progrsam." .. . . . . . ... 26
The "liew Program" ................ 2d

IV. PROBABILITY STUDIES . . . . . . . ... 37

Gross Statistics . . . . . . . . . 37









TABLE 'OF COClITEiTS (Continued)


Page

Probabilit' of CObserving Jupiter Radiation
as a Function of Jo-:.ian Longitude . . . .

LDta from Individual Charnnels . . . .. C
erged Dt . . . . . . . . 45
Irrift Studies of Source A . . . . .. 62
.Drift Studies of Source B . . . . . .
Possible Correelation of Radio SoLurce Drift
with Red Spot ETift. . . . . ... 7

Frobabilit' Velrsus Hour Argie . . . . . ..
Frobability Versus Uni-.ersal Time . . . . . 2
Probabiiit-' Versus Time From Sunset . . . . .
Probabiiit' Versus The Time From Sunrise. . . . 89
Frobabil-;'-i -ersus J.upter Elonr-atioi . . . . 9

V. IiTEliSIT STUIES. . . . . . . . ..... .. 97

Interisit, Distribution. . . . . . . .. 97
Gross Statistics. . . . . . ... . .. 101
Determination of Calibration- System Reliability . 105
Intei-nsity Versus System III Longitude . .. . ...
Hour Angle Studies .............. ... 115
Dil.ui'rnal Studie ... .. . . . .... . . 117
Time-From-Suni set Studies. . . . . . . . .
Ti'me-From-Sunrise Studies . . . . . . .. 12
Elongation Studies. . . . . . .... .. . 121
Search for Periodic Recur-renjces .. . . . ... 12

VI. SCLAC ATD GECPFhYSiCLL CFRELLLTICOIS WITH JOC'TAI
l- SSI !ii. . . . . . . . . . . . . 132
Log-em Ef.fects. . . . . . . . . . 32
Short-ITerm Correlations . . . . . . . . 13

Correction of the 196i A-nasis . . . .. 135
Changes in the Solar Flare Program. . . . 143
Chree A-na-isis Results of the 1962 Solar
Correlation Studies. . . . . . . 147
Concl1usEions and Suggestions for Further
Studies . . . . . . . ... 1 5

VII. DECAUL iTE--UAVE-ElIGTH OESERiVAIC'IS OF SATUFil, I-M.RS,
AiD' VEHUS Iii 1962 AlD 1963. . . . . . ... 192


Saturn. . . . . . . . . . .


192









TABLE OF COiTE1I1TS (Continued)


Page

Introduction and InstrLumentation . . . 192
Eata Reduction and Analysis . . . . . 193

Venus and I-Mrs. . . . . . . . . ... 197

VIII. SUTR. ARY AIFD R' ARKS Oil THE COI2PAF.ISOI0 OF THE
EXPER,~IEiTAL RESULTS WITH EISTIllG THEORIES ... . 202

Theories as to the Origin of Jupiter's Decametric
Radiation . . . . . . . . . . . 202

Summary. . . . . . . . . . . . 204

LIST OF REFEREJCES . . ................... .... .210

BIOGRAPHICAL SKETCH . . . . . . . . ... . . 214













LIST OF TABLES


able


T


vii


Page

3

38

39

60


112


1. Equipment Description . . . . . . . . .

2. Gross Statistics . . . . . . . . . . .

3. Activity Correlation Study . . . . . . . . .

4. Source Position as a Function of Frequency . . . . .

5. Criteria for Determining Reliability of the Averages
in Figs. 47 and 48 . . . . . . . . . .

6. Days of Peak Activity used in the Chree Analysis of
the 18 Mc/s Data in 1962 . . . . . . . . .

7. Days of no Activity and Days of Peak Activity in
Jovian Longitude Regions from 00 to 900 and 900 to 1900
for the 18 Mc/s Florida Data in 1962 . . . . . .

8. Days of Peak 18 Mc/s Activity in Jovian Longitude
Regions from 1900 to 2900 and 2900 to 3600 and
Days of Peak 27.6 Mc/s Data in 1962 . . . . . .

9. Gross Statistics for the Saturn Observations in 1962 . .

10. Pulses Labelled "Possible Saturn" recorded in 1962 . . .

11. Gross Statistics for the Mars and Venus Observations
in 1962 . . . . . . . . . . . .


172



173

194

195


200













LIST OF FIGURES


Figure Page

1. cSiple noise generator based on two Sylvania 5722 noise
diodes. . . . . . . . . . .... .

2. Amplifier used with the calibration system at the Chile
station during 1962 .. . . . . . . .. 8

3. Block diagrams of calibrator Systems "A," "B," and "C"
used at the Chile station in 1962 ............ 10

4. Examples of Chile calibrations. Top: Systems "A," "B," and
"C" calibrations at the Chile station (Calibrations using
syste:.;s A and B were not normally this complete.) Center:
Florida calibration. Bottom Systems "A," "B," and "C"
calibrations as they were n:-io.ally performed at the Chile
station . . . . . . . . ... . . . 11

5. Visualization of the gain calculation using the Chile cali-
brator system . . . . . . . . . . 12

6. Receiver-recorder response curves used as an aid to calcu-
late the gain of the amplifier in the Chile calibration
systems (100 chart deflection units is full scale deflec-
tion) . . . . . . . . .. . ... . .. .15

7. Systems "B" and "C" calibrations made during the watch to
determine the noise contributed by the Chile amplifier.. 17

Receiver-recorder response curves at 15.0 Mc/s and 22.2
Ic.s using C. .lins receivers and Texas recorders (100
deflection units represent full scale deflection) . .. 18

9. iHarriarlund and R;odhe-Scharz response c.i..es at 10.0 and
5.0 lc's, respectivelyl. .................. 19

10. Average probability of observ.inrg activity at the Chile
Station in 196. . . . . . . . . .

11. FProbability hiistogra! for the 1I.0 Ic "s data taken at
the Florida station in 1962, plotted in Polar coordi-
natess . . . . . . . . . . .. ... 3


viii









LIST OF FIGURES (Continued)


Figure Page

12. Probability histograms for data taken at 15.0, 18.0,
22.2, and 27.6 Mc/s at the Florida station in 1962. . 44

13. Probability histograms for data taken at 5.0, 10.0,
15.0, 16.0, 18.0, 22.2, and 27.6 Mc/s at the Chile
station in 1962 . . . . . . . . ... ... 46

14. Smoothed, merged probability histograms for data taken
at 10.0 Mc/s. . . . . . . . .. . 48

15. Smoothed, merged probability histograms for data taken
at 15.0 Mc/s. . . . . . . . .. . 49

16. Smoothed, merged probability histograms for data taken
at 18.0 Mc/s. . . . . . . . .. .. . .51

17. Smoothed, merged probability histograms for data taken
at 22.2 Mc/s. . . . . . . . .. . 53

18. Smoothed, merged probability histograms for data taken
at 27.6 Mc/s. . . . . . . . .. .. . .54

19. Smoothed, merged histograms for all the frequencies moni-
tored, using only 1962 data . . . . . .... 56

20. Smoothed, merged histograms for all the frequencies moni-
tored, using all the data recorded since observations
began in 1957 . . . . . . . . . . 58

21. Variation of source position with frequency (The dashed
lines provide an alternate selection of source posi-
tions. The point labelled 7 may not be source C.). . 59

22. Probability histograms for the 1960, 1961, and 1962
data taken at the Florida station . . . . .... 6

23. Yearly variation of the position of the 18 Mc/s source A . 64

24. Variation of the position of the 18 Mc/s source A since
1960 using partial apparition histograms. . . . ... 66

25. Probability histograms for the 18 Mc/s data taken at the
Chile station in 1962 using the "old" (TIII = 91 55" 29335)
and the "new" (TIII = 9h 55m 30970) periods . . .. 70









LIST OF FIGURES (Continued)


Figure Page

26. The 1957-1962 smoothed, merged 18 1.c/s histogram.
using the "old" (TI1 = 9h 551 29s35) period until
1960 and the "new" (T'III = 9 55" 3070) period af-
ter that date. . . . . . . . . . ... 71

27. il I1e/s source B drift studies using 1961 and 1962 data 73

28. 18 lIc/s source B drift studies using the merged 1961-
1962 data . . . . . . . . . . . 75

29. Time variation of the System II longitude of Jupiter's
Great Red Spot . . . . . . .... . . 76
9
30. Time variation of the longitude of Jupiter's Great
Red Spot for a period TpS which maintains the Red
Spot longitude constant from 1945-1958 compared
with Fig. 23 . . . . . . . . . . . 7.

31. Time variation of the longitude of Jupiter's Great
Red Spot since 1835, using a special period T = 9h 55m
37158. Eased largely on Peek (16) . . . .... .. 79

32. Probability versus hour angle histograms for the data
taken at the Florida station in 1962 . . . . .

33. Frobability versus hour angle histograms for the data
taken at the Chile station in 1962 . . . . . .

34. Probability versus Universal Time histograms for the
data taken at the Florida station in 1962. . . .. 84

35. Probability versus Universal Time histograms for the
data taken at the Chile station in 1962 . . . ... 85

36. Probability versus time-frmn-sunset histograms for
data taken at the Florida station in 1962. . . . 87

37. Probability versus time-from-sunset histograms for
the data taken at the Chile station in 1962. . . ..

38. Probability versus timee-from-sunrise histograms for
the data taken at the Florida station in 19'2. . ... 90

39. Probability versus time-from-sunrise histograms for
the data taken at the Chile station in 1962. . . . 91









LIST OF FIGURES (Continued)


Figure Page

40. Smoothed probability versus time-from-sunrise histo-
grams for the data taken at 18.0 Mc/s in 1962. . . ... 93

41. Variation of the average monthly probability with
time and Jupiter elongation, for data taken at the
Florida station in 1962. .... . . . . . 95

42. Time variation of the average monthly probability
for data taken at the Chile station in 1962. . . ... 96

43. Intensity distribution for the 18.0 Mc/s data taken
at the Florida station in 1962 . . . . . .. . 100

44. Frequency variation of the average flux density, SA . .. 103

45. Frequency variation of the average flux density;, SL .... 104

46. Comparison of the S versus Systemi III histograms
for the 18.0 Mc/s data using calibrations ("with
cal") with those using the cosmic noise level as
a reference ("no cal") . . . . . . . . . 108

47. Smoothed SA -:ers..s System III histograms for data
taken at the Florida station in 1962 (The dashed
lines represent regions of poor statistics.) .. .. 1.09

S4. Smoothed SA versus System III histograms for data
taken at the Chile station in 1962 . . . . . . 111

49. Smoothed SL versus System III histograms for data
taken at the Florida station in 1962 . . . . . 113

50. Smoothed SL versus System III histograms for data
taken at the Chile station in 1962 . . . . ... 114

51. Smoothed SL and SA versus hour angle histograms
using the 18.0 Mc/s data taken in 1962. . . . 116

52. Smoothed SL and S. versus Universal Time histograms
using the ld.0 Mc/s data taken in 1962 . . . . .. 119

53. Smoothed SL and SA versus time-from-sunset histograms
using the 18.0 Mc/s data taken in 1962 . . . . . 120

54. Smoothed SL and SA versus time-from -sz.urise histograms
using the 18.0 Mc/s data taken in 1962 . . . . .. .122









LIST OF FIGURES (Continued)


Figure Page

55. Variation of the monthly average activity index
rate with time and Jupiter elongation, for the
Florida data taken in 19i62. .. . ... . . 12

56. Time variation of the monthil,- average activity
index rate for the Chile data taken in 1962 ..... .. 125

57. Iaily activity index rate for the 5.0 and 10.0 Mc/s
data taken in 1962. . . . . . . . . 127

58. Daily activity index rate for the 15.0 IMc/s data
taken in 1962 .......... . . . 125

59. Daily activity index rate for the l,.O lc/s data
taken in 1962 . . . .... . . . . . 129

60. raily activity index: rate for the data taken at
22.2 and 27.6 Mc/s in i1962. . . . . . . 130

61. Comparison of the width of source A and of the average
probability of emission with SLinspot number . ... 133

62. Comparison of Jupiter's solar latitude and the
apparition average activity index rate with sunsrot
number. . . . . . . . . . . . . 3

63. Chiee analysis of the solar flare activity index in
,groups one, two, and three, using the 20 peak days
of 18.0 I-ic/s Jupiter emission at the Florida station
in i9'i . . .. . . . . . . . 13

64. Regions on the solar disk as viewed from Jupiter as-
signing flares to groups one, two, or three ..... 13

65. Chree analysis of solar flare activity index in groups
one, two, and three, using the 20 peak days of 27.6
MIc./s Jupiter emission at the Florida station in 1961. i4

66. Chriee analysis of solar flare activity index in groups
one. two, and three, using the 20 peak days of 18 Iic/s
Jupiter emission at the Florida station during the
three months around opposition in 1961 . . . ... 142

o6. Chrlee analysis of solar flare activity index in groups
one, two, and three, using the 20 peak days of 18 IMc/s
Juriter emission at the Chile station in 1960 .. 144


.xii









LIST OF FIGURES (Continued)


Figure Page

68. Chree analysis of the solar flare activity index in
groups one, two, and three, using the 20 days of
peak activity of the 18 Mc/s Jupiter emission at
the Florida station in 1962 . . . . . . ... 149

69. Chree analysis of the number of flares in groups
one, two, and three, using the 20 days of peak
Jupiter activity at 18 Mc/s at the Florida
station in 1962. . . . . . . . . . ... 151

70. Chree analysis of the 2800 Mc/s solar flux, the
Zurich provisional sunspot number (RZ), the geo-
magnetic index (Ap), and the Jovian daily activity
index rate using the 20 days of peak Jupiter activity
at 18 Mc/s at the Florida station in 1962. . . .. 153

71. Chree analysis of the solar flare activity index
in groups one, two, and three, using the 21 days
of peak activity of the 18 Mc/s Jupiter emission
at the Chile station in 1962 . . . . . .... .155

72. Chree analysis of the number of flares in groups
one, two, and three, using the 21 days of peak
Jupiter activity at 18 Mc/s at the Chile station
in 1962. . . . ... . . . . . . . 157

73. Chree analysis of the 2800 Mc/s solar flux, the
Zurich provisional sunspot number (RZ), the geo-
magnetic index (Ap), and the Jovian daily activity
index rate, using the 21 days of peak Jupiter activity
at 1o [c's at the Chile station in 1962 .. ... . 5

74. ChLree analysis of the solar flare activity inde:.:
excludingg subflares) in groups one, two, and three,
using the 2'?. days of peak activity: of the i.d iLIc/s
Jupiter emission at the Florida station in 1962. ... i61

75. Chree anai,-Lis of the rn.,mber of flares (excluding
subftares) in groups ore, two, and three, using the
2., das:; of peak Jupiter activ.it: at the Florida
station in 1962. . . . . . . . . . .. 161

76. Chree analysis of the solar flare activity, inde: in
groups one, two, and three, using the 2'. da; s of peak
Jovian activity (Florida plus Chile) monitored by:
both stations at 1 Mc/s during the three months
around opposition in 1962. . . . . . .... 3163









LIST OF FIGURES (Continued)


Fi guire

7. Chree anal,-sis of the number of .solar flares in groups
one, two, and three, using the 20 days of peak Jo-
vian activity (Florida plus Chile) monitc.red by b.th
stations at, 1 I-l'c/s during the three montlis around
opposition in 1962. . . . . . . . . .


Page


7S. Chree analysis of the 2500 i.c,'s so lar flu., the Zurich
provisional sunspot number iF;), the geomagnetic in-
de:. (Ap), and the Jovian daily activity index: rate,
using the 20 days of peak Jupiter activity (Florida
plus Chile) monitored by both stations at 18 I-ic s
during the three months around opposition in 1962 .. 1.66

79. The solar wind and the interplanetary magnetic
field. After Parker (25) . . . . . ..... 68

,o. Chree analysis of the 20C0 [lic/s solar flZu', the
Zurich provisional s'urspot number (RZ), and geo-
magnetic index (Ap), and the Jovian daily activity
index rateusing the 29 days of "no activity" at
both stations during the si.: months around opposi-
tion in 1')62 ...................... 170

:l. Cti'ee analysis of the 2:00 I-ic/s solar f l.:, the geo-
n.-gnetic index t.'p the suni'pot nLumiber R., and tlhe
solar flare activity indices for regions one, two,
and three, using the 11 days of peak Jupiter ac-
tivity received in the IO to 90r i'.I interval on
the 1 -; lc/'s, Florida channel in 19o2. . . . . 174


:2. Chree analysis of tlhe 2500 e.cZ s olar flux:, the geo-
magnetic index Ap, the sunispot number R', and the
solar flare activity indices for regions one, two,
and three, using the 17 rpeaI days of Jupiter activity
received in the 9.30 to 190' A interval on tihe
l ic' zs,. Florida channel in 192. . . . . .


LS. Chree analysis of the 2b00 I.-c s solar flu-:, the geo-
magnetic index Ap, tlhe sunrspot number ?R, and the
solar flare activity indices for regions one, two,
and three, using the 19 days of peak Jupiter acti-
vity received in the 190' to 2900 A II interval
on the 18 Il-c/s, Florida channel in 1962 . . .. .


164


S 175


170









LIST OF FIGURES (Continued)


Figure Page

S3. Cnree analysis of the 2500 .Mc/s solar flu.:x, the
geomagnetic index Ap, the sunspot number R7,
and the solar flare activity indices for regions
one, two, and three, using the 21 days of peak
Jupiter activity received in the 2900 to 3600 A-
interval on the 18 Mc/s, Florida channel in 1962. . 177

85. Double correlation studies of the 2800 nIc/s solar
flux Chree analysis with the Chree analyses of the
solar flare activity indices in regions one, two,
and three, using the 20 days of peak Jupiter acti-
vity monitored at 18 Mc/s by the Florida station
in 1962 ............... . . . 179

86. Double correlation studies of the geomagnetic in-
dex Chree analysis witn the Chree analyses of
the solar activity indices in regions one, two,
and three, using the 20 days of peak Jupiter ac-
tivity monitored at 18 Mc/s by the Florida sta-
tion in 1962 . . . . . . . .... .... 1

7. Cihree analysis of the solar activity indices (FX1,
EX2, and FX:3), using the 19 days of peak 27.6
i.c/s Jupiter activity received at the Florida sta-
tion in 1962 . . . . . . . . ... . 182

38. Chree analysis of the solar flare numbers (Fi1, F1i2,
and Fli), using the 19 days of peak 27.6 IMc/s Ju-
piter activity received at the Florida station in
1962 . . . . . . . . ... ...... .183

89. Chree analysis of the 23j800 Mc/s solar flu.:, the
geomagnetic index Ap, the sunspot number RB, and
the Jovian activity index rate, using the 19 days
of peak 27.6 Mc/s Jupiter activity received at
the Florida station in 1962.. . . . . . . 18

90. Location of Saturn bursts in activity longitude
systems plot ed for rotational periods of
1'h 5m.6 and 11h 57m., where the longitude
was assumed to be zero at Oh U.T. on May 10, 1962 . 198













CHAPTER I

IIITRODUCTIO1


The decameter radio emission from the planet Jupiter, first ob-

served accidentally, by, Burke and Frankrlin in 1955 (1), has been moni-

tored by the radio astronomy group at the University of Florida since

1957.

Such extensive "coverage" can be justified merely by, noting

the existence of numerous conflicting theories purporting to explain

this strange electromagnetic phenomenon (2), (3), (4), (5), (6), (7).

These theories rill be discussed in a later chapter. The results of

the 1962 observations are offered here as experimental evidence to aid

in deciding which, if any, of these theories is correct. This analy-

sis must, of course, be considered as a continuation of the analyses

from previous years by T. D. Carr (o), II. E. Chatterton (9), and II. F.

Six (10). Details of data-taking, execution of the watch, and data

analysis that were not changed for the 1962 apparition (observing sea-

son) are not discussed here, since II. F. Six (10) covered these in his

dissertation.

During the 1962 apparition observations were also made of I-ars,

Venus, and Saturn.


IIISTRUI-Eii'TATiOII

The 1962 observations were made from two sites: one at the

University of Florida campus in Gainesville, Florida, and the other near










I-Iaipu, Chile, about twenty miles from Santiago. At the Chile station,

observations of Jupiter were made in 1962 at frequencies of 5, 10, 15,

16, 18, 22.2, and 27 .6 .Ic/s, while simultaneously observations were

being made at frequencies of 15, 18, 22.2, and 27.6 Mc/s in Florida.

At each frequency an antenna was connected to a commercial communica-

tions receiver and two pen recorders. One of these, a low-speed re-

corder, ran during the entire listening time at a chart speed of six

inches per hour; the other, a high-speed recorder, ran only when Jovian

activity ;-as received. The chart speed of the latter recorder was

5 mm/sec, 25 mm/sec, or 125 rnmm/sec, depending on which speed was needed

to resolve the pulses. The data analysis, as discussed in this paper,

primarily involves data taken on the low-speed Texas recorders, except

at 15 I.c/s (Chile) and 27.6 MIc/s (Chile) for which the low-speed data

were taken on Esterline-Angus recorders.

Pertinent information regarding the antennas and their receivers

is given in Table 1.















V VVVVVvVVVVV


V3 (.3

.-i *,-i


00
0U


000
- .-I


r.I
r- r00

00
0D L)


o La in 0 0 0 cI i o [(- 0
--I -Tr [r-- U-\ b'\ 0 0 C' U\ r-i
L-,\_T u- 'L\ %O \D \C, %% mn ,- -I C-)


-r -- r-i\ O Lr\ L.-\ CO i\ r- [*-


4-.

1u



(.3 0
'0

n0
fiO


.) 11)
30 "0

00 cd
00 P
i"
O PQ
(tf


III
III
III
I I I


.3)
*,4

,-
o


0)


r -'
E
*r-1




S0- I
..-4 -I


* .-4



CSd I

a)a)


bO bO
..-- ..-4
Zt

Lid Li


t> tiU ti) tit tu
.0. .
-.-4 0 -1 *
ccxxx
00000 AA
L> U G) U 0
iG .1
'*i !- '*i 'H '-


Li --4*Ht C
Sc: CO o )
L%\ .i) 0
Hi L'I rJ cy' *)


LiO

*.-I (
O -.--

'0


Hr *H


'.~ 1:4
H -
C) C)7
ri i~

~-,4~)-1


rdrd



-3) .L) P-, P
QU-,z.,


lU\ 0 LI\ U-\ \10 CO CO CO
rlrllHrrlrl


0JOJLJ

01010
imiU(


' o

-I


.1


4,



*
0
..-1I
0, 0
MO


.,

-4






-4





.-,




0





- O






4-.0

00 G
,>0

S 0 C) H.

) p .,- 0
0 0 aj

0*+


l l


t(o





-4 4-)


cd
*,- 0



0






U >-I

COj
* 0
ul' I-l


1-1 f-












CHAPTER II

CALIBRATION PROCEDURES


The analyses of preceding apparitions were used predominately

for probability studies. If, on the other hand, accurate intensity

studies are to be made, the received signal intensity must be known as

a function of time. By dividing the activity periods into small inter-

vals and determining the intensity for each interval, this time depen-

dence can be established, as was done by G. W. Brown (11) and T. D. Carr

et. al. (12). Calibrators were installed at each station for just such

a purpose. At the end of every watch each channel which had recorded

activity was calibrated with a noise generator composed basically of two

Sylvania 5722 noise diodes connected in parallel (Fig. 1). This calibra-

tion was necessary each time radiation was.received, since the gain of

a receiver might drift from day to day. .------

SRECEIVER

I I-


S5722 5722 R R

I

B+ J_ ,
0 0- L--------

VARIAC

Fig. 1.--Simple noise generator based on two Sylvania
5722 noise diodes






5

BASIC COUiCEPTI

llyquist's Lai,


P = kT Af, (1)


where k is Boltmiann's constant (1.38 x 1,'-2 joules / K), T is the re-

sistor temperature in degrees Kelvin, and \f is the range of frequen-

cies over which the noise power P is to be measured, gives the noise

power produced by a resistor at a temperature T. What would the resis-

tor temperature have to be in order to yield sufficient noise power to

calibrate a typical Jovian pulse?

Consider the flux density 3 (power per unit area per unit of

bandwidth) as being received by an antenna of effective area A connected

to a receiver of bandwidth Af. The power then can be written as


P = 1/2 SA Af, (2)


where the 1/2 is introduced because the antenna is sensitive to only one

of the two linearly polarized components of the radiation. Equation (1)

with Equation (2) leads to:


T = 1/2 _- t (3)
k


Substituting 720 mru for A and 10 -21w m-2cps for S (a typical value for

Jovian decameter activity at 13 i.ic/s), one finds T to be about 2.6 1.: IC K.

Since a resistor temperature this high is obviously out of the question,

a calibrator based on noise diodes is used (Fig. 1).

If I is the d.c. component of the plate current, i the r.m.s.









fluctuation of the plate current, and R the terminating resistance of

of a Sylvania 5722 noise diode operating over a frequency range Zf,

then by Schottky's Law:



i V2el Af, and (4)



p 9 R=- eIA (5)
22


In the above, e is the electronic charge (1.6 x 10-9 could ) and i/2

instead of i is used in Equation (5) since only one-half of the a.c.

plate current reaches the receiver. The other half is dissipated in

the terminating resistor h. Using Equations (2) and (5), one arrives

at:


I = SA/eF. (6)


If one substitutes the same -.alues for S, A, and e that were used in

Equation (3), along with h = 50fl, I becomes 90 m.a. The maximum d.c.

plate current that can be used with the Sylvania 5722 noise diode is

35 m.a. Even with two diodes connected in parallel (I = 70 m.a.), one

would not be able to obtain a calibraticn deflection equal to the deflec-

-2 -2 -l
tion resulting from a flux. density of 10 w m cps oteworthy, also,

is the fact that radiation from the giant planet often exceeds the cited

flux density, especially at frequencies below 18 Ic/'s. At 5 Mc/'s, pulses

have been recorded vhich correspond to calibration currents as high as

70 ampsi Such currents are certainly not readily created by connecting

additional noise diodes in parallel if each diode supplies no more than









35 m.a. Since higher-current noise diodes are not available, the only

recourse is either to amplify the noise generator signal or to attenu-

ate the signal from Jupiter. The first alternative \as tried at the

Chile station during the 1962 apparition.


THE FLOIFaIA CALIEFATOR

Fig. 1 sufficiently describes the simple calibrator used at the

Florida station. Regardless of how limited this system proved to be,

it \as a step toward better calibration and a still better system is

being used during the 1963 apparition.


THE CHILE CALIEBATOR

In Chile. the first improvement was to replace the plate resis-

tor R, shown in Fig. 1, with a higher value (- 450fO). 'This was done

simply by inserting a 375-1 resistor in series with the original 75nI

plate resistor (the receivers were tuned to 75 Sf input impedance in

Chile instead of the C50! input impedance used in Florida). The re-

sulting calibrator-to-receiver mismatch i.'s corrected by connecting the

two with a coaxial cable cut to an appropriate length. This procedure

made possible an equivalent noise generator gain (K) of about si-:. With

the additional resistor and the appropriate matching cable inserted, the

calibration rwas latelled "high range," while, .with the noise generator

alone, the calibration was labelled "low range." The calibrator in

"high range" still generated insufficient noise power to allow calibra-

tion of the stronger Jovian pulses.

An amplifier (Fig. 2) -was, therefore, inserted with the noise

generator in "high range," thus producing a calibration system known as







RESISTORS OHMS
CAPACITORS- Mfd


6800 Pf
6800 Pf


SIZED FOR A RANGE
SIZED FOR A RANGE


OF 15-20 MC/S
OF 5-10 MC/S


V2


INPUT

75 1
OHMS


.025


L2


.002 -


+290 V.


+155V. REG.


VI V2


.


AM PLIFI ER


OF CHILE CALIBRATOR


Fig. 2.--Amplifier used with the
tem at the Chile station during 1962


calibration sys-


6SG7
6SG7


V,
V2
C,
C,


470


TPUT


75
HMS


LI









"System C" (see Fig. 3). This made possible pen deflections correspond-

ing to nearly all of the pulses of Jovian origin. Obviously, for satis-

factor- calibration, the equivalent current Ie (that current through a

noise diode which would produce the given pen deflection) had to be

Known, since the flux density S is fouaid by inserting Ieq into:


S = eRIeq, (7)
A


an expression arrived at merely by rewriting Equation (6). In order to

find the equivalent current, the dependence of the amplifier gain (G)

on the calibration current (I) had to be measured as described below.

As a further modification of the system, an attenuator of known im-

pedance wa s inserted into "System C," thus producing what was known as

"System A" (Fig. 3). The noise generator alone in "lou range" was known

as "System B."


CALIBRAT'OII USIIG T[E CHILE CALIBRATOR

After each %watch the receiver gain was increased, so that full

scale deflections could be obtained with "System B": calibrations were

then made using Systems "A" and "B." Fig. 4 shows these calibrations

as they appeared on the Texas records. iote that the calibrator used

in Florida corresponded to "System B" in Chile. However, the Florida

calibrator was used during the watch; hence,it %was used at lower receiver

gain than that used for the "System E" calibrations in Chile.


USE OF THE CALIBRATIO1I

Fig. 5 illustrates the procedure used for finding the gain (G)










SYSTEM A

HIGH
RANGE
.............. -- 7-

NOISE AMPLIFIER ATTENUATOR RECEIVER RECORDER
GENERATOR




SYSTEM B


LOW
RANGE

NOISE RECEIVER RECORDER
GENERATOR




SYSTEM C

HIGH
RANGE
14 1 I 1 I I R
NOISE AMPLIFIER RECEIVER RECORDER
GENERATOR
Fig. S .--Block diagrams of calibrator Systems "," "E," and "C"
used at the Chile station in 1962












I.-

II-


-* ""t : : _t,. m
-, ,*

T _
.. .. PO__- r- .,i ---


II

'I-1-


T-
I 2

i 4


-.-


Y L L ... .. _, . ..


- r- 0 -
t. -: J| .-.. __

. .. ..








Fig.. ).--E:-anilea of Chile :alibrationp. Top: S.- teams "A:" "B:"
and "C" calibtrti.-rns at the Chile station (CalibratlionS uing S, stems
A an d B ..ere not nr;.,ii:all, t-his ':o.mplete.). Center: Flori'il calibra-
tionr. Eottomi S-.'stemics "A1, "BE" and "C" ca vibration, as the, ':ere
noroimall., perfonTied at the Chile station


I- .
~i~








SYSTEM A
HIGH


I I 1 I I --
NOISE I AMPLIFIER ATTENUATOR
GENERATOR

SYSTEM B

LOW
I 1 RANGE I i


RECEIVER


SAME E L
REFLECTION


I I_ I_ A _&
NOISE BI RECEIVER RECORDER
GENERATOR
PA ,G + PN
w PB,


.SYSTEM A
HIGH
RANGE
I-- 1> -E- -----I-I -- I
__ ^\ I -- s i i ^ -
NOISE A2AMPLIFIER ATTENUATOR RECEIVER REC.
GENERATOR

.SYSTEM B

LOW DEFLECTION
RANGE


NOISE "2 RECEIVER RECORDER
GENERATOR P G + P
A2 N
= P,


Fig. 5.--Visualization of the gain calculation using the Chile
calibrator system








of the system. G equ-ls the product of the amplifier gain and the

"high range" over the "low range" gain (K). P,..... are the re-

spective noise powers created by the noise generator with currents

S......I ("System A"), P.....F are the respective noise powers

created by the noise generator with currents I .....IBi ("System B"),

Pi is the noise power inherent in the amplifier, and CC is 0 the attertLa-

tion coefficient of the attenuator. 'With current I, in "System A" and
1
1BE in "System B," a deflection @i is obtained for each system; thus:


l G(P + P P1 (8)
OC

By changing Ii to I12 and I t I t to obtain a deflection "p for each

system, the relation

P, G(FQ ) p. FE
`2 k 2 + i B2 (9)
CX


results. Subtraction of Equation (9) from Equation (') leads to:


P C(P ) P1 G ( r) = C( P (i0)
A 1 i A-


But, from Equation (5) we see that the power P. is proportional to the

current Ii: hence Equation (10) can be -written:


IA, GIA) G(IAG = C (IB I)* (


If G(I-1 ) = G(I )=G( IA- an appro:ximation that is reasonable,

if IAl IA is small, Equation (ii) reduces to:










G 1 2 1 ()
S / ,- 1I,'



Fig. o is a plot of the 15 i.Ic/s "System A" and "System E" calibrations

shozi'-n in the top record of Fig. 4. From it can be found values for I'Ai,

IE1 ,-.. and IB.' Equation (12') yields G(1:.6 i..a. ) -b if the values

used are: A1. = 43.7 m.a., I1 = 15.b 1:1.a., IAs = 29.,0 m.a., and IB =

10.0 m.a., as selected in Fig. 6.

Unforti.uately, most of the "S.yste:m 1" and "Systemi B" calibra-

tions appeared as sho.-n in the bottom record of Fig. 4. Hence, it .-as
Jl I-+
iripossible to' draw the c.arves needed to calculate ( + U .Values

for I' 'A, I I 1 anrd I r.might then be selected as follows:



I = I = 0 = min + E ~' p min

I = IE-, = ,", = ,' r:in
(13)
I = = i :1m.a. '.' = ma::

I = B1 = suiif'ficient current to = ma::,
produce p ma::


'..here E.. represents the small additional deflection due to the noise in-

herent in the amplifier. Using these values in Equation (12), one finds:


I,
( ) -B 4)
'A


where the additional subscripts have been dropped. Is this value of

gain a sufficient representation of the actual gain to justify its use

for cali'tration?

j study of the curves C(IAi) versus IA, obtained from the few





15







< SYSTEM A
50- -60
SYSTEM B "
-0) /


25- /50
LLI / LuI




20- 40=
I-I


/
15- B30


o 0
-0 /ao -


0r /



0 I I 0
5- 10 -10




o ----- ^---i--------- '--L o
0 20 40 60 80
CHART DEFLECTION

Fig. 6.--Receiv'er-riecorder respon-se cuLes used as
an aid to calculate the gain of the amplifier in the Chile
calibration systems (100 cart deflection units is full
scale deflection)









calibrations having enough points to ailov.r curves to be drawn, sho'.ed

that -G(i. ) scar'cei;, ever varied from (i1. /2) by more than 20 per cent.

Since slight antenna-to-receiver mismatch, antenna pattern change vrith

hour angle, etc., account for variations of this magnitude, this 20 per

cent gain variation should not be too objectionable. Also, the amypli-

fier gain variations from da-y to day were small (less than 10 per cent);

hence, amplifier gain calibrations ..ere made only veekly. We shall

drop the ir. io the n atior. the fuictic.nal dependei-ce of C on IA remembering

that C really means G(-= i2).

In. order to find the actual equi.'valent current (leq) of the

amplified noise po'.er, one need merely to calculate:


iq = G 1i + %i, (15)


'here ..as the curr~-nr.t in "System C," G -..'s the gain, and i x.'as the

"equivalent current" necessary to produce the noise over present due

to the amplifier. i -w.'as easily read from calibrations involving "Sys-

tem C" and "Systerm B" made at the same receiver gain, as sho-.m in Fig. 7.

Ini this case !l is a little greater than 60 m.a. Unfortunately, the am-

plifier noise at 1., 22.2', and 27.6 [Ics .-as equal to. or greater than,

that due to the cosmic backgroL.und noise. This excessive noise -was

naturally a disturbing so-rce of error.


RFECEI V CThl ACTEI ST CS

Also needed for accurate calibrations iuere the response cu-.-'es

(deflection versus voltage) of each receiver and recorder combination.

Figs. d and 9 sho..' t-,pical response cur-ves for the Collins, Harmmairlud,.







































o
'0


'04-








.0.+<


,-4
I ."




.4-, .-I





















, *., ,. I







,- 4
h -' -
Ic' >!'

~
















22.2 MC/S


GALACTIC AVERAGE


.1 .2 .5 .4
INPUT VOLTAGE IN pV


15.0 MC/S




CTIC AVERAGE


.2
INPUT


.4 .G .8 1.0 1.2
VOLTAGE IN !V


Fig. 5.--Receiv.er--rcorder response curves at
15.0 Mc/s and 22.2 I/' s using Collins receivers and
Te::as recorders (100C deflection units represent full
scale deflection)


40

30

20




















10.0 4 C/S


AVERAGE


50 100
POWER IN U


NIT


150 200
S OF I019 WATTS


5.0 M C/S


CTIC AVERAGE


15 30 45 60 75 90


POWER IN UNITS OF


10-9 WATTS


Fig. 9.--Hainmmar1lmd and Eodhe-Schwarz response
cirves at 10.0 and 5.0 I-Ic/s, respectively


1

I-
z
=)


120


100




0.

JO
LI
2
20









and Rodhe and Schwarz receivers. Note that the Rodhe and Schwarz re-

ceiver (operated at 5 Mc/s in Chile) is the only receiver with a linear

response if power is plotted versus deflection. The rest of the re-

ceivers show a linear response if voltage is plotted versus deflection,

or if power is plotted versus the square of the deflection. As might

be expected, the response of a given receiver changes with frequency.

For the most part, the Collins receivers showed a linear response (de-

flection versus voltage) for all frequencies, if only deflections above

the galactic level were considered (see Fig. 8).














CHAPTER III

A1IALTICAL PROCEDURES


ILTA REDUCTIOII

The data -.ere analyzed with t..wo ends in mind:

A. To continue analyses started in 1957 and used each

year since;

B. To start a new-1 ana lsis that -w-ill provide better in-

formation about the intensity.


La.ta Reduction as Performed in 1957-1961 and
Continued in 1':- "Old Anralysis

In his dissertation, ii. F. Six (i10) explained this analysis in

detail. For that reason, procedures used in the 1957-1-961 data analysis

..ill merely, be listed and defined.

Records of all frequencies were reduced simultaneously, :ith con-

stant reference being made to the logs kept b' the observers. Since iden-

tification of the radiation rested originally, with the observer, the logs

and siutad eous rec ;ds assisted the -erson reducing the data in check-

ing the validity of the observer's decisions. On the whole, the ot-

servers seemed to be quite accurate.

Information taken from the records was:


Li stening Feriod

The listening period .as that period of time (U.T.) -.ihen the









receiver was working, Jupiter was in the antenna beam, and interference

was not excessive. The beam limits of the tracking antennas were de-

fined to be five degrees above each horizon. The beam limits of the

fixed antennas were:


1. 5B (Chile) + 4 hours from the meridian

2. 27C (Chile) + 4 hours from the meridian

3. 22P (Chile) + 3 hours from the meridian

4. 15C (Florida) + 4 hours from the meridian

5. 22P (Florida) + 3 hours from the meridian.


Activity Period (Storm)

A storm was that period of activity not interrupted by more than

8.3 minutes (five degrees of Jupiter rotation) of no activity. Note: The

listening period was divided so that each activity period had a corres-

ponding listening period.


Date (Month, Day, and fear)

A day was defined to begin at noon on day X and to end at noon on

day X + 1. The date recorded was X + 1.


Intensity

If D was the pen deflection of a Jupiter pulse and G was the pen

deflection of the galactic background, then:


Intensity = K (D2.) (16)
G


where the indicated average was performed for the three highest Jupiter






23


peaks in any given storm. If there were less than three peaks during a

sto, the average s perfoed ohe .erae perfthe existing peaks. K is the

normaliziation constant defined below.


ilonralization Constant

The normalization constant \.as that constant which normaliizes

the galactic background to the galactic background of our first observa-

tions in 1957. This constant is necessary. since, as Jupiter changes its

right ascension (two hours per year), the portion of the galax:., that is

monitored also changes.


Coin[Lients

Co:rnments were remarks made by the data reducer as to equipment

condition, calibration, listening conditions, antenna direction, etc.

As a continuation of work already- performed, this analysis .as

necessary. but it is easy. to see that in order to impro-.e our grasp on

the intensity:- we must di-.ide the storms into small time intervals. G.

U. Erouwn (11) of the Florida group first did this by di-.iding the storms

of the 1961 apparition into ten-minute intervals. Extending his work, we

divided the sto-rms of the 1'62 apparition into five-minute intervals, as

shown in the next section.


Data Reduction as Started in i'19l by Erou'.
and Continued in 194o2 "iiew nalysis

Information taken from the records for the "new analysis" can be

listed briefly, as follows:


Date

The definition of the date \.as not changed from that used in the

"old analysis."









Listening Period

The listening period was that time spanned by the Universal

Time at the beginning and end of effective listening as defined for the

"old analysis." Again, it should be pointed out that the listening

period is split up so that there is a listening period for each activity

period.


Activity Period

The times of Jovian activity were divided into five-minute in-

tervals. Any remaining time less than five minutes in duration was

considered to be a separate interval.


Intensity

The pen deflection for the highest value of Jupiter activity,

and the pen deflection for the average galactic background were read

for each activity interval on the record.


Jupiter Transit

The time of Jupiter transit was looked up in the American

Ephemerus and Nautical Almanac. Since the stations were not at the cen-

ter of their respective time zones, corrections had to be made. The cor-

rected Jupiter transit time (local standard time) was recorded in the log

by the observer.


Pulse Quality

The antennas used by our stations have relatively broad beams,

so noise from other sources comes into the receiver. Our identification

of Jupiter activity depends mainly on the judgment of the observer, who

with experience can become quite proficient at identifying Jovian pulses.






25


The interference most difficult to distinguish from Jupiter radiation

is that caused by a weak radio station. The observer logs each time .

Jovian activity, is recorded and estimates the certainty of his identi-

fication of the activity,. With logs in hand and records taken simul-

taneously at the Chile and Florida stations, the person reducing the

data decides whether the activity in any' interval was "definite,"

"possible," or "dubious." By the term "definite" is meant that listen-

ing conditions were good, the antenna -as pointed toi-wrd Jupiter, ac-

tivity was recorded at the other station if it -a.-as listening -ith good

conditions, the observers labelled it "Jupiter," etc. If any, one of

the conditions named above was not met, the data reader might have

chosen to label the activity "possible," depending upon his confidence

in the observer. Pulses getting a "dubious" label were often short and

isolated or buried in outside interference.


Calibration

The Florida and Chile calibrations were different, so the in-

formation needed to calculate the flux: density of a pulse would also be

different.

1. Florida.--The records 'were calibrated directly with noise

diodes.

Lumbers read from the records were:

a. each calibrator current value

b. the corresponding deflection

c. the total number of calibrator current

values used in any one calibration.









2. Chile.--The records were calibrated using calibration

Systems "A," "B," and "C" as described in Chapter II.

Information read from the records included:

a. each "System A" current value (one per calibra-

tion)

b. each "System B" current value (one per calibra-

tion)

c. each "System C" current value

d. the corresponding pen deflection for each

"System C" current value

e. the total number of "System C" current value

per calibration.


JUPITER PROGRAMS FOR THE I.B.M. 709

As has been indicated already, the 1962 Jupiter data analysis

took two forms. We "shall call the program used in the 1957-1961 analy-

sis the "old program," and the one used with the "five-minute-interval"

analysis the "new program." The data described in the preceding sec-

tion were punched into two decks of cards, one to be used with the "old

program," the other to be used with the "new program."


The "Old Program"

N. F. Six (10) gave a complete description of the "old program"

in his dissertation. We will, therefore, be content with a list of the

input and output.


Input

The following information was punched into cards to be used with









the "old program":

1. station

2. date (month/day/year)

3. beginning and end of listening period (U.T.)

4. beginning and end of activity period (U.T.)

5. intensity of Jovian radiation (average of the three

highest peaks)

6. normalization constant

7. Durich provisional relative sunspot number

8. System IT longitude at 0h U.T.

9. Julian day number.


Output

The following information ias printed out by the computer:

1-8. all of the input except the Julian day number

9. beginning and end of the listening period (System III

longitude, Jupiter)

10. beginning and end of the activity period (System III

longitude)

11. daily activity index

12. daily activity index rate

13. monthly average activity index rate

14. yearly average activity index rate

15. monthly average sunspot number

16. probability histogram table.

This program also stores the daily activity index rate and the daily

su-nspot number on a tape to be used with the solar flare program.









MIention should also be made that, for the 1963 studies now

under w\ay, the sunspot number is not read into the "old program," and

that the intensity (average of the three highest peaks) is calculated

by the computer--not by hand as before.


ThIe "liew Program"

Since the studies proposed for the "five-minute-interval analy-

sis" were far afield from those used in the "old program," a completely

different program wa\s written. However, parts of the "old program"

were incorporated into the "new program" to save computer de-bugging

time. Again, we will divide our discussion into two parts: input and out-

put. VWiere terms are identical with those in the "old program," this fact

i:ill be noted and further discussion vill be eliminated.


Inrut

The total number of input parameters in the "new program" ex-

ceeded that of the "old program." Therefore, to save space on the cards,

parameters that did not change throughout the apparition were read only

once from a single card at the beginning of the execution of the program.

These parameters were;

1. average effective area of the antenna (taken from

0. W. Bromn's thesis (11) or calculated using in-

formation from p. 16h, American Radio Relay League

(A.R.R.L.) Antenna Book)

2. transmission coefficient (b) of the transmission line

(assumed to be about .75)(Since the noise power to the

receiver has actually been attenuated by this factor,

Equation (7) should be written:









S elR ) (17
bA


3. equivalent current (I) of the noise inherent in

the calibrator amplifier

4. attenuation coefficient of the attenuator used in

"System A"

5. Julian day number of the first day of the appari-

tion

6. station

7. frequency in [.Ic/s

8. equivalent current of the average galactic back-

ground noise for the entire apparition (used as a

calibration reference when no calibration was per-

formed on a given watch)

9. year.


Input parameters read on each of the succeeding cards included:

10. month and day

11. beginning and end of the listening period (U.T.)

12. beginning and end of the activity period (U.T.)

(never longer than five minutes in duration)

13. deflection (G) due to the galactic background radia-

tion

14. deflection (D) due to the highest Jupiter peak

in the activity period

15. quality of the pulse (definite, possible, or

dubious)









16. time of transit (standard time at a given station)

17. time of snriise (standard time)

18. time of sunset (standard time)

19. AOh U. T. System II longitude of the central
II

meridian of Jupiter at Ch U. T. of the date (X + 1)

in question

20. number (ii) of calibration points

21. "S:ystem A" calibration current (A -.alue of one was

entered in the case of Florida data which had no

calibrator amplifier, and zero was entered in case

no calibration was performed.)

22. "System B" calibration current (Zero was entered

in the case of Florida data, or if no calibration

was performed.)

23. each calibration current -value ("System C" in the

case of Chile)

24. the pen deflection corresponding to each calibra-

tion current value.


COutput


.What did the computer do with these input data? It used them to

calculate the output information that is listed below. It also printed

out most of the input data along with the calculated parameters to guide

the person reading the output. In fact, all of the input parameters

listed above were included as output except numbers 5, 6, 17, S1, 20,

21, and 22. Output information calculated by the computer under the









control of the "new program" was:

25. equivalent current of D

26. equivalent current of G

27. flux density (S) of Jovian peak

28. beginning and end of listening period (System III

longitude)

29. beginning and end of activity period (System III

longitude)

30. activity index (the product of the flux density

and the activity time in minutes)

31. beginning and end of listening period (local hour

angle)

3. beginning and end of activity period (local hour

angle)

33. daily activity index rate ( i activity index
total listening time

where i runs over all of the activity periods of any

given day)

3L-37. daily, activity index rate for activity in each of the

System III longitude intervals: C0 to 92o (no source),

900 to 190 (source B), 190 to 2900 (source A),

2900 to 3600 (source C) (The term "source" will be ex-

plained later.)

38. monthly activity index rate for all activity in a given

month

39-O2. monthly activity index rate for activity in the longi-

tude intervals listed in numbers 34-37






32

43. yearly activity index rate for all activity in the

entire apparition

44-47. yearly activity index rate for activity in the longi-

tude intervals listed in numbers 34-37

48. daily listening time in minutes

49. monthly listening time in minutes

50. yearly listening time in minutes

All of the above output .as listed for each data card with daily,

monthly, and yearly averages being written only uhen they ,jere needed;

i.e., daily average at the end of a day, etc. After all of the output

data had been listed, several histogram tables were generated. each of

which contained all of the parameters in numbers 51-55 listed as a func-

tion of one of the parameters in numbers 56-60. Every table was divided

into ti-hree sub-tables, uhicih jer'e generated using different restricTions

on the data. One sub-table was calculated using definite data only,

another using definite and possible data, and the third using all the

data.

51. probability of observing Jovian radiation

52. average flu>: density (SA) (averaged over those

flue-: densities actually detected)

53. average flux density (SL) (averaged over all

intervals, i.e., letting the flux: density equal

zero uhen no activity was received)

54. total number of intervals during uxhich activity

was received

55. total number of intervals during which listening

was recorded









56. System III longitude on Jupiter in five-degree in-

tervals

57. local hour angle of Jupiter in five-degree inter-

vals

58. Universal Time in ten-minute intervals

59. time with respect to sunset (one hour before to

seven hours after) in ten-minute intervals

60. time with respect to sunrise (seven hours before

to one hour after) in ten-minute intervals..

Some of the output needs discussion. Equation (17) gives the

flux density of a pulse -wihich has an equivalent current I. Since the

activity from the giant planet is superimposed upon the cosmic radio

noise, the deflection is not entirely due to Jovian radiation. There-

fore, Equation (17) must be modified to give actual Jupiter flux den-

sities. If SC is the flux density of the cosmic noise and Sj-j is the

flux density of the cosmic noise plus Jupiter noise, then Sj, the flux

density of the Jupiter noise, can be written:


Sj = Sc SG. (i8)


Since the equivalent current is proportional to the flux density,


Ij = (1-+ IG. (19)


Equation (16) becomes:


SJ = (I +- IG). (20)
bA

Another problem encountered at this point was that of determining









I,+J and IG. These were found by interpolating between the values ac-

tually appearing in the calibration. This interpolation was performed

with the characteristics of a given receiver in mind; i.e., if the de-

flection varied directly as the voltage, and thus as the square root

of the flux density, this non-linearity was considered.

On some of the records calibrations were missing or, for sane

reason, were .unreliable. When this occurred, the average equivalent

current of the cosmic noise for the entire apparition, which was part

of the input data, was used as a reference. .Etrapolation to the Ju-

piter peak was made by using either Equation (21) or Equation (22).

Equation (21) was used for each frequency except 5 Mc/s, for which

Equation (22) was used because of the characteristics of the Rodhe and

Schwarz receiver.


c - c 2 (21)


eI R
j D (22)
bA C


If any part of a five-degree Jovian longitude interval emitted

radiation, the entire interval was credited with having emitted radia-

tion. This practice could create problems. Consider this example.

Jupiter's System III longitude is divided into five-degree intervals,

each of which we monitor for about 8.3 minutes as it rotates by the

"central meridian," the "central meridian" being that meridian bisecting

the visible disk. We shall assume that from 0553.0 to 0606.3 U.T. a

given ( XIi) interval was on the central meridian. Assume also that










a storm vass in progress and that iwe had divided our five-minute ac-

tivity intervals in this manner: 0555.0 to 0600.0, 0600.0 to 0605.0,

and 0605.0 to 0610.0. A portion of each of these activity periods is

included in the 0558.0 to 06o6.3 time interval associated with the

five-degree rotation of Jupiter. But the computer credits each AIII

interval with a count each time any portion of an activity period falls

within it. In this example the \III interval in question received three

counts for one storm. It is easy to see how the five-minute intervals

could be chosen so that only tuo count-s were credited to this III in-

terval. With the "old program" this multiple counting vas eliminated

by requiring that two storms be separated 8.3 minutes in time. But in

the "nev program" one activity period (five-minute interval) often fol-

lows ruilediately after another, and thus it is possible for multiple

counting to occur. Therefore, the "new program" allowed a five-degree

longitude interval no more than one activity or listening count per day.

Another provision would have to be made if a watch included more than

one rotation of Jupiter.

After all of the output had been calculated, an attempt vas made

to get more reliable flut: density values by performing an iteration. As

mentioned above, an intensity versus hour angle histogram vas plotted.

Since the effective area of an antenna and the ionosopheric atteniia-

tion of a pulse would change with hour angle, the assumption of a con-

stant effective area in Equation (20) might lead to erroneous results.

Therefore, each pulse ras normalized by multiplying its flux density by

the ratio of the average flux density (SA) received at the meridian to

the average flu;.: density (SA) received at the hour angle of the pulse.









This iteration was performed three times; once using histograms plotted

with definite data, once with definite and possible data, and once with

all the data. After inspection of the intensity versus hour angle his-

tograms, which showed very little structure, these iterations were

omitted in the interest of saving I.B.M. 709 machine time.

Mention should be made, before discussing any results, that the

calibrations were made quite some time after the Jupiter storms, and

hence at different receiver gains, since the gains changed with time.

Furthermore, the calibrator's impedance was not always well matched to

the receiver's impedance. These two facts were indicated by unusually

great changes in the equivalent current of the cosmic noise level during

a watch (nearly one order of magnitude in some cases), and by the change

in the equivalent current of the cosmic noise level from one part of the

apparition to another. Since Jupiter was well away from the galactic

center, gross changes in the cosmic noise background would not be ex-

pected. Therefore, the data were also processed using the cosmic noise

level as a reference throughout the entire apparition. Equation (21)

was then used for calculating flux densities. When this was done the

data were labelled "no cal." When the calibration was used, the data

were labelled "with cal." Later, when the intensity histograms are

studied, they will be so labelled.














CHAPTER IV

PROBABILITY STUDIES


Since the 1962 data were of better quality than any recorded

at the Florida and Chile observatories up to that time, a more mean-

ingful analysis was possible. The studies which had been carried out

for the 1957-1961 data dealt mainly with the probability of observing

radiation from Jupiter. The extension of this work is important;

hence, the probability studies are presented as the first of our analy-

ses.


GROSS STATISTICS

Table 2 shows the extent of the observing season, the number

of hours of effective listening, and the number of hours of Jovian ac-

tivity.

It is striking that the number of hours of activity reaches a

maximum at 18 Mc/s. The decline toward the higher frequencies can be

explained by the fact that Jupiter radiates less at these frequencies.

The low-frequency decline is probably due to increased ionospheric re-

fraction and scattering of the low frequencies.

Of interest, also, is a study of the correlation of noise

storms between the two stations ("storm" will be defined in the section

of this chapter involving data reduction). For each frequency, Table 3

shows the total number of storms, days of activity, simultaneous storms,









TABLE 2

GROSS STATISTICS


total effective
freq. observing season listening total activity
in tIc/s station beginning end time hr. time hr.

5 Chile 6/10/62 11/26/'2 205 2
10 Chile 5/8/62 12/7 /62 382
15 Chile 4/ /6 1/20/62 1099 166
15 Florida 6/17/62 1/25/63 625 45
16 Chile 7/26/62 12/21/62 516* --
18 Chile 4/1/62 12/20/62 1192 197
18 Florida 3/20/62 2/26/63 080 165
22 Chile 5/ 1/62 12/20/62 --- -
22 Florida 6/6/62 3//63 108Cr< 67
27 Chile 8/21/62 11/27/62 90 6
27 Florida 3/20/62 3/6/63 944 25


.Polarimeter Hlew program not
E-E rata from both 22Y and 22P
-i-* Possible and definite data


used so total time not calculated


days for which simultaneous storms were recorded, and storms received at

one station that "should have been," but were not, received at the other.

By "should have been" is meant that if both stations were listening with

good conditions, simultaneous activity "should have been" expected at

both stations. One can see that for the frequencies using corner re-

flectors (15C, Florida, and 27C, Chile) the correlation was the poorest.

However, mention should be made that all of the ten storms heard on 15B

Chile, that "should have been" heard on 15C, Florida, were recorded be-

fore August 23. Until this date the impedance of 15C was 141 + j33 ohms,

when it should have been 50 + JO ohms. This unintentional antenna-to-

receiver mismatch undoubtedly caused a loss of sensitivity. The other












4-'
S
Cu

a)

-4 r. '2r
aa C C
0E 0 o L c,




CJ .,I z L


U 0 CCO -
4 f r, N C
' -- ,- 0 r U

*-i B *-1 i-l Cuj Cd .-4


c,
C-



0










4J
'-o













oj






OO
0





,a


4-1

















Oj
4->





0
4,
00
O





>1


















03
'o
4-j










'0


,-.



+03
L2)
'-





SC)





03






CO
r,


00 o'l


( 0a) 0






< i 3 OC
C-r u)
Cu


If\ ,\ 0 u-









'1-



"- ['- -r '
Cun H
Cu i -i


UI\ 1 -4 -X LP\
-a C rC D Cu






uI -r 0 'Do 0
-r rni ON
CtJ


I -


0)



E-I

H








H
E-i

<;


4O- 0 -1 --















0 0 0 -1


I
I
I
I












I
I
I
I






I
I
I
I






I
I
I
I









frequencies show incredibly good correlation. All in all, simulta-

neous listening was recorded for 6 months at 15 MIc/s, 9 months at

18 Mc/s, 7 months at 22.2 Mc/s, and 3 months at 27.6 Ilc/s.

Fig. 10 shows the average probability of observing Jovian ra-

diation plotted versus frequency. Every point except that at 27.6 Mc/s

was taken from the Chile data. (The Florida 27.6 MIc/s radiometer was

more reliable than that at Chile.) One can easily see the tendency of

the probability to peak around 10 Mc/s. However, this was not the fre-

quency at which the greatest number of hours of activity occurred. The

difference was probably caused by increased ionospheric absorption, and

hence poorer listening conditions, at 10 Mc/s as compared to 18 IMc/s,

where the total activity time reached its peak. Iote should also be

made that the 15, 18, and 22.2 Mc/s data at Florida yielded average

probabilities lower than those at Chile for the same frequencies. This

discrepancy was probably due to the fact that the Florida observing sea-

son extended closer to the time of conjunction than did the Chile sea-

son. Some of the Florida observations were made when Jupiter radiation

had to pass through the solar corona, and hence suffered attenuation.

The only frequency for which the average probability at Florida ex-

ceeded that at Chile was 27.6 Mc/s. The 27.6 IMc/s antenna in Chile

(a corner reflector) did not work well, and thus little activity was

recorded on that channel.


PROBABILITY OF OBSE.'EVIiG JUPITER FADIATIOII
AS A FUNCTIOi OF JOVIAJI LOGITUDE


Eata from Individual Channels

One of the first major characteristics discovered in the


















X POSSIBLE AND DEFINITE DATA

0 DEFINITE DATA ONLY


25


FREQUENCY,

Fig. 10.--Average probability
Chile station in 1962


MEGACYCLES

of observing activity at the


0

cs .20
cc

0
0
o


LL.
o


o .15

>-
I-
_J


o .10
o
0.
0.
U


Qr .05
w


10


50









decameter radiation from Jupiter was that the probability of its occur-

rence depended on which Jovian longitude was on the "central meridian."

The histogram in Fig. 11 shows that four portions of the planet exhibit

a relatively high probability of emission. These portions are called

sources B1, B2, A, and C, as indicated in the figure. Sources BI and

B2 are often thought of as only one source, "source B," but as can be

seen in the following histograms, the bifurcation seems to be a perma-

nent feature. However, unless a special study is being made of sources

B1 and B2, they will be referred to collectively as "source B."

Let us take a closer look at the sources for each frequency.

Fig. 12 shows the probability histograms for the 18, 22.2, and 27.6

Mc/s, possible and definite data, and the 15 Mc/s histogram for the

definite data only, as recorded at the Florida station in 1962. As

has already been mentioned, the 15 Mc/s channel was in something less

than perfect operating condition for part of the apparition; hence,

we see less source structure than might be expected. The peak at 1150

is source B, the peak at 2400 is source A and the peak at 3200 is

source C. Note that the shapes of sources A and C change with fre-

quency.

For 18 Mc/s, sources B (1050 to 1700), A (2200 to 280"), ard

C (2900 to 3450) are all quite well defined. Source B has a weakly-

defined dip at 1300, thus forming sources B1 and B2. Another feature,

of which we will keep track, is the general null between 20 and 55 .

Sources B, A, and C are evident at 22.2 Mc/s, though source C is quite

poorly defined. For the 27.6 Mc/s histogram, sources B and A are

evident, but source C becomes quite diffuse.

































C)



,1



*r





0 C0
, - '

















CT





1 ,
1
..-I















'1' ,.







FLORIDA


00 90 1800 2700
LONGITUDE, SYSTEM IT


360


Fig. 12.--Probability histograms for data taken at
15.0, 18.0, 22.2, and 27.6 Mc/s at the Florida station in
1962


1962









Fig. 13 shows the probability histograms for 5, 10, 15, 16,

18, 22.2, and 27.6 NM/s in Chile. The peaks are less evident, since

the vertical scale had to be reduced to get all of the graphs in one

figure. Note that at 15, 16, and 18 Mc/s source A also seems to be bi-

furcated.

What general trends can we see in Figs. 12 and 13? Easily no-

ticed are the facts that source A shifts to higher longitudes for lower

frequencies (at least up to a point) and that the width of Source A in-

creases with a decrease in frenquency. Which peak is source A on the

5 Mc/s histogram might be debatable. If we choose the general null

at 2100 on the 5 Mc/s histogram to correspond with the one mentioned

earlier, the peak at 2600 corresponds with source B and the peak at

3400 might correspond with source A. Another interesting feature is

that the ratio of the source C maximum to the source A maximum in-

creases as the frequency decreases.


Merged Data

Before making further studies of Figs. 12 and 13, let us look

at the composite or merged histograms for those frequencies which were

monitored by both stations. The merging process is described in de-

tail by II. F. Six (10), and to maintain continuity in ou- analysis,

these merged histograms will be presented in much the same manner as

he presented them. Therefore, we used the three-point smoothing tech-

nique in which the probability for each five-degree interval is aver-

aged with those just preceding and following it. These points are

then joined to form the smoothed histogram. While the smoothing proc-

ess may smear out fine-structure effects, it also rids the histograms








CHILE 1962
627.6 MC/S'-_ S DE F.

.4 22.2 MC/S

.2 POS. AND DE.


0

y" 4 18.0 MC/S
|w POS. AND DEF.


0
0 -

S.4 16.0 MC/S
S POS. AND DEE -


"-i 0
> 0 ---
Es
-j
m0
< 15.0 M C/S -
o .2- PC S. AND DEE



.2-
10.0 MC/S POS. fLnrr
.2 _n-J^Lr AND DEF
o---------------

5.0 MC/S DEF.
.2-


00 900 1800 2700 3600
LONGITUDE, SYSTEM II
Fig. 13.--Probability histograms for data taken at
5.0, 10.0, 15.0, 16.0, 18.0, 22.2, and 27.6 Mc/s at the
Chile station in 1962










of much of their roughness by eliminating bumps due to minor statis-

tical variations. The merging also tends to minimize bad effects

caused by local conditions, such as equipment failure, atmospheric

disturbances or an unattentive observer, and makes possible a contin-

uous run of observations not attainable at one station.

Fig. 14 shows histograms for the 10 Mc/s Chile data. The 1962

histogram is shown at the top of the figure and the 1960-1962 merged

histogram is shown at the bottom. The sources are labelled as II. F.

Six (10) chose them in 1961. We will later see that another choice of

labels is possible. Source A is less prominent in the 1962 histogram

than in the merged histogram. Also noticeable is a small but definite

peak at a longitude of about 1400. The fact that it is shifted to

about 1430 in the 1962 histogram will be discussed later.

Fig. 15 shows the 15 Mc/s histograms for the Florida-Chile merge

for 1961 and 1962, the Florida-Chile merge for 1962 only, the Florida

data for 1962, and the Chile data for 1962. In the bottom plot each

source is clearly defined and we can see a very slight hint of a bifur-

cation of source B. The 1962 merge shows both B1 and B2 and a defi-

nite splitting of source A--this dofublet structure of source A is ap-

parent also in the 16 and 18' Mc/s Chile data. The plot for the Chile

data shows the source A double structure still more clearly, while the

plot for the Florida data shows little identifiable structure in this

region. Noticeable, however, is the marked difference in the position

of the source B peak in the Florida and Chile histograms. It is hard

to argue this shift out of existence by the fact that the impedance of

one antenna was not correct, unless source E shifts during the year.










CHILE.
1962


10.0 MC


CHILE


1960-1962


I0.0 MC


00 900 1800


LONGITUDE,


SYSTEM Ir


Fig. 14.--Smoothed, merged probability histograms
for data taken at 10.0 Mc/s


.20





w
.10
Z2'0
0


c)
0


>-
-J



o.20
a.


.101


2700


3600
















0 -.3 0


, -A .20

Sz CHILE '

/ 1962 .10
o 15.0 MC

.30 -0
o CHILE- FLORIDA
20 1962
>1.20- 15.0 M C
I-

S.J10

m
o
a.A

.30- CHILE C -
FLORIDA
1961-1962
.20- 15.0 M C






00 90 180 2700 3600
LONGITUDE, SYSTEM II

Fig. 15.--Smoothed, merged probability histograms for
data taken at 15.0 Mc/s









If source B does vary its position during the year, this apparent

shift could be explained by the fact that the impedance was incorrect

for only the first part of the year. Since the data from the first

part of the apparition were therefore probably incomplete, a shift

such as the one seen in these two plots might then be expected. This

study will be pursued later in detail. We also see that the source B

peak on the Florida histogram is skewed to lower longitudes. We will

find that the source B peak is generally skewed to higher longitudes.

Since 1957, 18 Mc/s observations have been made by at least

one of our stations, making this frequency that for which we have the

most data. Fig. 16 shows the 1957-1962 merge, the 1962 merge, and the

individual histograms for 1962. IHotice the difference in source struc-

ture between the Chile and Florida histograms. Source B2 is better de-

fined on the Florida plot than on the Chile plot, but source B as a

whole seems to be easier to identify on the Chile plot. Again, Chile

shows the slight splitting of source A that vas noticed for 15 Mc/s,

and the Florida plot also shows source A to be skewed slightly on the

high-longitude side. Also of interest is the fact that sources B and

C in Florida seem to be shifted to slightly lower longitudes than the

same sources in Chile, while source A seems to be located at exactly

the same longitude. As with the 15 Mc/s histograms, the 1962 merged

histograms show the sources to be shifted to slightly higher longi-

tudes than those on the 1957-1962 merged histograms. The 1962 merged

histogram shows the basic source structure, including bifurcations,

very clearly, while the 1957-1962 merged histogram shows the sources

without the bifurcations. The reason for this smearing is probably

















FLORIDA
1962


.20






z 0
Ll



0
.40
LL
0

>.30

- .
m .20
0C
0
a. 10


1962
18.0 MC


CHILE- FLORIDA
1957-1962
18.0 MC


0'
00


1.10


3600


900 1800 2700
LONGITUDE, SYSTEM 1I


Fig. 16.--Smoothed, merged probability histograms
for data taken at 18.0 Mc/s


.30


F CHILE- FLORIDA


.10









an apparent source shift to higher longitudes or, effectively, to an

increase in the System III rotational period. This change in period,

which apparently began in 1960, is also the reason that the 1962

merged sources are shifted to higher longitudes than those of the all-

time merge. A special section will be devoted to a study of this ap-

parent change of period.

The 22.2 Mc/s smoothed histograms are shown in Fig. 17. Again,

we see that in the top two histograms the structure of source B is dif-

ferent at each station and that corresponding sources seem to lie at

about the same longitudes. Note that, as for the 15 Mc/s Florida

source B peak, the 22.2 Mc/s Chile source B peak is skewed to lower long-

itudes. While the anomaly on source A was on the high-longitude side

for 18 Mc/s, it appears on the low-longitude side at 22.2 Mc/s. In the

merged histograms the bifurcation of source B is less evident. In fact,

if any splitting of B is to be mentioned in connection with the 1962

merged histogram, the source should be referred to as a triplet, not a

doublet. Again, the shift of source A to higher longitudes in the 1962

merged histogram is noticed when it is compared with the 1958-1962

merged histogram. While differences in source structure between the

Florida and Chile histograms were also noted by N. F. Six (10) in 1961,

these differences were not the same as we see in 1962; hence, we will

not try to draw conclusions from any but the most obvious details.

Even though the 27.6 Mc/s data are scant, 1962 is the first

year that we have had even this much activity at this frequency. Fig.

18 shows the smoothed histograms for 27.6 I-Mc/s. We notice again the










.20


.10


.50


.30

CHILE
r 1962 20
cr
D 22.2 MC
C., .10
0

o 0

A
S.30- CHILE-FLORIDA
^- B1962

< .20- 22.2 MC

SB
10 -




.20- CHILE- FLORIDA
1957-1962

.10- 22.2 MC


0 1
00 900 1800 2700 3600
LONGITUDE, SYSTEM II

Fig. 17.--Smoothed, merged probability histograms
for data taken at 22.2 MTc/s


























































900 18
LONGITUDE,


0


0 2700

SYSTEM III


Fig. 18.--Smoothed, merged probability histograms
for data taken at 27.6 Mc/s


.05


0

u
O
o
z
LJ
cc.05
C.=

0
0
0


>.-
I-
-J
- .05

0

0


.05


00


3600









shift of source B on the Chile histogram with respect to source B on

the Florida histogram. As with the 15 IMc/s histograms in Fig. 16,

we can argue the possibility of this difference in peak longitude if

source E shifts during the year, since the 27.6 Mc/s channel in

Florida operated during the entire apparition, while the 27.6 M1,c/s

channel in Chile did not begin monitoring Jupiter until August 21,

1962. Since opposition was on August 31, 1962, nearly all of the

before-opposition data were missed in Chile. This argument still seems

weak in the face of the large shift (approximately 20 degrees). -When

we compare the two individual histograms with the 1962 merged histo-

gram, we see that the apparent shift of the Chile data scarcely ap-

pears. This is probably due to the fact that the Florida listening

time was so much longer than that at Chile, thus minimizing the impor-

tance of the Chile peak.

Let us look at all of the 1962 smoothed, merged histograms to-

gether, as shown in Fig. 19, to see if any characteristics of the

sources change with frequency. One marked feature is the appearance

of the anomaly on the high-longitude side of source A at 18 Mc/s, its

increase at 16 Mc/s, and finally, its smearing of sources A and C at

15 Mc/s. Also noticeable is the decrease in the ratio of the source

A maximum to the source C maximum with frequency.

A real problem arises when an attempt is made to locate sources

A, B, and C on the 5 and 10 Mc/s histograms. In his dissertation, II. F.

Six (10) assumed the sources to be as shown in the top set of letters

on the 10 Mc/s histogram. Credence is given to this choice by noticing

that it maintains the source separations fairly well. However, a

stretch of the imagination is required to keep the source spacings







CHILE-FLORIDA


90 1800 2700
LONGITUDE, SYSTEM Ir


Fig. 19.--Smoothed, merged histograms for all the
frequencies monitored, using only 1962 data


3600


1962









constant on the 5 Mc/s histogram. Another problem that arises, if we

assume the top set of labels, is that the height of source B seems ex-

cessive on the 10 [Mc/s histogram in light of the fact that its maximum

is less than half that of source A on the 22.2, i3, 16, and 15 Mc/s his-

tograms. For 10 I-cl/s the bottom set of labels provides for a drift to

lower longitudes with decreasing frequency at frequencies below.16 Mc/s.

Table 4 lists the positions of the sources as they occur at different

frequencies. O.nly one choice for each source was made for frequencies

of 15 Mc/s and above.

When all of rhe data that have been taken at either station are

merged and smoothed we get the histograms shown in Fig. 20. The graph

has been labelled "Chile-Florida 1957-62" but not all of the frequencies

have been monitored over this entire period. The 5 IMc/s histogram repre-

sents only the 1962 Chile data, the 10 Mc/s histogram represents the

1960-1962 Chile data, the 15 IMic/s histogram represents the 1961-1962

Florida-Chile data, the i6 Mc/s histogram represents the 1960 and 1962

Chile data, the 1i [Mc/s histogram represents the 1957-1962 Florida data

and the 1960-i162 Chile data, the 22.2 MIc/s histogram represents the

1958-1962 Florida data and the 1960-1962 Chile data and the 27.6 Mc/s

histogram represents the 1958-1962 Florida data and the 1962 Chile data.

The 10 and 15 Mc/s high-longitude source is better defined in

Fig. 20 than is its counterpart in Fig. 19, while, as should be ex-

pected, anomalies like the one on the high-longitude side of the 15

Mic/s source A in Fig. 19 are less noticeable in Fig. 20. The positions

of the sources, however, are about the same in both figures.

Fig. 21 shows a plot of the values in Table U. The dashed lines







CHILE-FLORIDA 1957-62


360


LONGITUDE,


SYSTEM =m


Fig. 20.--Smoothed, merged histograms for all the fre-
quencies monitored, using all the data recorded since ob-
servations began in 1957




















A C


25






20





15!


i'


10 -


/
/
/
I


I I I


S. I


3600 1200 2400 3600 1200 2400
LONGITUDE, SYSTEM IT


Fig. 21.--Variation of source position with frequency (The
dashed lines provide an alternate selection of source positions.
The point labelled 7 may not be source C.)


0


1 I


2400


1 1 r I 'I I I


30 -


5k>4
& .0'





,p.1











SOURCE POSITIOII


TABLE 4

AS A FE-UCTIOi OF FREQUENCY


Source A Source C Source B

27.6 232 10 2 + + 5 128 + 50
O C C
0 0 c
2.2 2 37 + 5 31 125

18 250 + 5 3l + 3 140 + 10

18 Mc/s 252 + 325 + 50 148 + 10

15 Nc/s 260 + 150 327 + 7 140 + 10


322 + 6o + 8 20o7 12
0 0
5 1c/s o-- 107 i 277 + 10
107 + 15 277 + 10 ---
107 15 ---- 277 + 10



(odd numbers) provide for various patterns of source shift to lower long-

itudes sith decreasing frequency, and the solid lines (even numbers) show

the possible identifications of the sources for shifts to higher longi-

tudes with decreasing frequency. ilote that since there are only two ma-

jor sources at 5 Mc/s, it takes only two lines to describe the 5 Mc/s

sources. Redundant points are plotted for 5 and 10 Mc/s, since the scale

has been folded at 360 to allow continuous lines to be drawn through the

sources. If we assume that all of the sources shift in one direction as

the frequency is decreased from 15 to 10 Mc/s, we can choose the sources

to be represented either by lines numbered one, three, and five, or by

lines numbered two, four, and six. However, we see that one of the

sources disappears as the frequency is decreased from 10 to 5 Mc/s.









(One of the sources presumably either decreases in intensity or merges

with another source as the frequency decreases below 10 Mc/s.) We see

that even if we restrict ourselves to the assumption that all sources

shift in the same direction, there are four pairs of points in Fig. 21

that can represent the 5 Mc/s sources. The lines running through these

pairs of points are numbered one and three, three and five, two and four,

and four and six. notice that source A is present in all of the choices

and that either source B or source C is the other.

It is the opinion of the author that source C is one of the

sources at 5 I-c/s, in keeping with the prediction by Field (5) that the

effects of source C should become more noticeable at low frequencies.

Also more palatable to the author is the shift to lower longitudes with

decreasing frequency, since this allows sources A and C to continue to

merge as they do at 10 M-c/s. Therefore, the author's choice is repre-

sented by lines one and three. The 1200 source is presumably a combi-

nation of sources A and C and the 2710 source is source B. However, the

shift of the null between sources C and E seems to dictate a shift to

higher longitudes. Presumably, this dilemma could be resolved by ana-

lyzing the polarization at 5 and 10 M1/s, for both Field (5) and Carr

(2), maintain that radiation from source C comes from Jupiter's southern

hemisphere, and hence that it is polarized in the opposite sense (left-

hand elliptical) to that of sources A and B, which they believe to

originate in the northern hemisphere and which are of right-hand polar-

ization. Samuel Gulkis (13) and C. H. Earrow (14) have found that the

polarization of source C at 16 MIc/s is predominantly left-hand polarized,

a phenomenon not so clear at higher frequencies. Presumably, then, it









should be more evident at lower frequencies. The 5 Mc/s polarimeter

to be built at the Florida station will be used to identify the polar-

ization of these sources.

Looking again at Figs. 19 and 20, one can see that the source

width seems to increase as the frequency decreases, as discussed by

N. F. Six (10). Identification of the sources for the frequencies be-

low 16 Mc/s, however, makes quantitative analysis difficult and super-

fluous.


Drift Studies of Source A

In 1961 Carr et. al. (15) established that the rotational period
h m s
of the radio sources on Jupiter was 9 55 29.35, a value that closely

agreed with the period obtained by Douglas (16)--9 55 29.37. Upon

analysis of the 18 Mc/s Florida probability histograms for 1960-1962, as

shown in Fig. 22, even a casual glance shows that source A seems to be

drifting toward higher longitudes with time. So unexpected was this

drift that all preceding calculations were checked and re-checked. The

1963 data for 18 Mc/s were reduced, analyzed, and plotted. They, too,

showed the apparent drift.

In order to determine whether or not the drift was just a

scattering of the points, the source A position was plotted for each

year beginning in 1951, as shown in Fig. 23. The bars indicate estimated

errors as determined by the person analyzing the data. The two points

close together at the top of the graph were obtained by dividing the

1963 data into two parts (one hundred days before opposition and one

-hundred days after opposition) to get partial apparition histograms.








18.0 MC/S


o90
LONGITUDE,


30 2700
SYSTEM I


Fig. 22.--Probability histograms for the 1960, 1961,
and 1962 data taken at the Florida station


360


FLORIDA











a *_ 1 B


1964.0 1-


1962.0


1960. 0


9h55m30so52 --


.1-O-

_C


1958.0 -


1956.0


1954.0




1952.0


'-0--


-9-h55m29s35


a I


260


a I


2000


2200


2400


LONGITUDE OF SOURCE A,

Fig. 23.--Yearly variation of the
18 Mc/s source A


A r (1957.0)

position of the


280.


__ I __


I


fl


I


. I


I









At the time of this writing the data had not been analyzed for the

entire 1963 apparition, since observations were still being made.

Fig. 23 clearly indicates that source A did begin to drift at some

time early in 1960.

Further, each apparition was divided into tree parts, begin-

ning with the 1960 apparition, probability histograms were plotted for

each part and each station, the histograms were "shuffled" in an effort

to eliminate bias introduced by the person anaj,-zing the data, and the

source A position was plotted versus time as shown in Fig. 24. Tie

points labelled 1 and 2 in the figure were highly uncertain due to the

poor statistics of their respective histograms and were not included in

the least-squares calculation, even though they probably, would have con-

tributed cancelling effects. From the least-squares line the apparent

drift was calculated to be 10.45 degrees per year or, if e-pressed as

that change of rotational period necessary to maintain the sources at

their o960 longitude, plus 1.17 seconds. This makes the new System

III rotational period 9h 51m '52. The calculations above were made

using the relation:






where AT is the difference in seconds between rotational periods,

A A is the shift between longitude systems in degrees/year, and T

is the mean rotational period in hours. If T is near the radio period

(9.925 hours), then Equation (23) reduces to:


LT = 0.112;.AA.


(24)



























-0



!00










*0






O I-
0 .
OZ <
9_ ro


O)

L-






0
3 -p



a m



0 -1
O> O



oo



0 0
< oCd




*H


O--



o o
C cd




0 .3
I W


o 0 0 0 0
'r- 6
CD CO (D (O D
O' 0) 0) O) G)










The sources are probably linked with the magnetic field of the

planet; therefore, their rotational period is the same as that of the

planet itself. Since it seems unlikely that the rotational period of

the planet has changed, then the sources have apparently changed their

positions. How could this happen? Suppose that one of the sources is

located at Jovian System III longitude (X) and that this longitude

does not change. However, for radiation to reach the earth, it may not

be necessary for longitude X to be that which lies on the "central me-

dirian." if the interplanetary magnetic field is curved, as indeed

it may be (see Chapter VI T then Jovian radiation may be bent by it.

For bending of electromagnetic waves we also need electron density

gradients. For this study we assume that the necessary electron den-

sit, gradients are present. Tf the rays are bent, then some System

III longitude, which w_ e w-.ill call Y, will be on the "central meridian."

But if, due to the change in solar activity dur ing the sunspot cycle,

the interplanetary magnetic field shape is changed, thus changing the

degree of bending of Jovian radiation, then not i', but some other longi-

tude 2, will be on the "central meridian." But the longitude region

which is on the 'central meridian" when we receive activity is that

which we reckon to be the active region, so we see an apparent (1Y-Z)

drift in the position of the source located at X. Carr (17) has sug-

gested -hat the observed shift of the emission pattern, toward higher

longitudes at lower frequencies, results from a tendency of Jovian

magnetic field lines to cu'rve west.ward with an increase in radial

distance, due to a drag on the rotating field. Since Jupiter's

magnetospheric electron density presumably also changes with sunspot






68


activity, it could also be responsible for the temporal change in

source position. It should be pointed out that, if the interplanetary

medium bends Jovian radio emission, the positions of the visible disk

and the radio disk would not coincide. However, interferometer meas-

urements show that they do coincide so the magnetospheric bending seems

more likely. It is difficult at the moment to see how this could ac-

count for the sudden change of source position shown in Fig. 23, or

the fact that once the source position did change, it continued to

change in a linear manner as shown in Fig. 24. In addition to this,

the fact that Jupiter's Great Red Spot also began to drift in its

longitude system (discussed later in this chapter) weakens our argu-

ment even more, for electromagnetic i.aves in the optical spectrum are

not influenced appreciably by the interplanetary medium. ii.eertheless,

if an eleven-year period could be found, although a wild stretch of

the imagination is needed to find one in Fig. 23, then an obvious con-

clusion would be that the drift is caused by the change in solar ac-

tivity. UnfortLnately, there are insufficient data to determine this

now, but if the source A position starts to move to lower longitudes

in the ne:.:t few ;,ears, a variation of the source longit-ude with the

su.nspot cycle may, be established. Interesting also would be a stud,

of the drift of the sources others than so.u-ce A at 18 McI/s, and of all

the sources at frequencies other thar, 18 Mci/s. Robert Hay,-.ard is cur-

rently making just such a study at the University of Florida.

Using the partial apparition histograms, a subjective study

.-as made of the source structure to determine whether or not it

ch-anged in a regular manner during the year. As far as could be









determined, the sources maintained their relative intensities quite

well during the apparition, as should be expected, since Jupiter does

not know where the earth happens to be. Further study of the drift

of source B will be discussed in the next section.

In Fig. 25 we see how the change in rotational period moves

the source positions. The rotational period w.-as calculated to be

h 55m U.'T from a sk-etched curve, dra-.-n before the 1963 points were

put on the graph in Figs. 23 and 24. This period compares quite well

with the period of 9h 55m 3'.-,52, ich .-as finally reached by using

the least-squares curv.e. Therefore, the histo.gram for the "new." period

(T = 9h 55 30'70) in Fig. 25 should be representative of the ac-

tual "new" period (Ti = 9h 5 50.752). Is source A better defined for

the "new" 9 55 30.7 period than for the "old" period? 1o, if anything,

the opposite is true. In fact, none of the histogramrs improved in de-

tail when the period was changed. The 'widthn of source A at 22.2 I-Ic/s

in Chile actually became greater when the 9h ; 30.7 period .as used.

Carrying this study further, let us e-xaniine the 1957-1962 merge

for i c/s using the 9h 55" ',:,.72 period, assuming, as sho.mn in Fig.

23, that the period abruptly changed to this value from h 55 29:35 on

January 1, 1o60. The smroothed, merged (1957-i962), 1 IIc/s histograms

for the "old" and the "new" periods are sho.un in Fig. 26. The peaks

of all the sources in the histogram using the "new" period are higher

and the sources are sharper than those in the histogram using the "old"

period. liote that sources C and EB become more pronounced for the "new"

period. This is the only frequency for which so noticeable an improve-

ment was effected by the period change, which makes sense when one





70
CHILE 1962 POSSIBLE AND DEFINITE


900
LON


1800 2700
GITUDE, SYSTEM TII


3600


Fig. 25.--Probability histograms for the 18 Mc/s data taken
at the Chile station in 1962 using the "old"(TT I = 9h 55m 29?35)
and the "new" (III = 9h 55m 30T70) periods



















































00 900 18
LONGITUDE,

Fig. 26.--The 1957-1962
tograms using the "old" (TTIi
1960 and the "new" (Ti11 = 9'
date


Oo 2700 3600
SYSTEM I=

smoothed, merged 1~ M.c/s his-
= 1h 55m 2935) period until
55m 30T70) period after that









remembers that this is the frequency that w as used to calculate the

source drift. Eobert HayoNrrd of the University of Florida is examin-

ing the source drift for different frequencies to determine whether

or not the source drift varies with frequency.


Drift Studies of Source B

In 1963, Warwick (4) suggested that source B drifts to lower

longitudes during an apparition and he showed graphically, using his

1960 and 1961 data, that during the time from 140 days before opposi-

tion to 40 days after opposition the source B center shifted from ap-

proximately 132, to 1000 ( 1ii,' 1957). Since this .ias the first time

such a drift had been proposed, a search of the 1961 and 1962 data is

a.-rranted to see if this effect exists in our data. We do not include

our 19yS data, since splitting it intc several parts wouldd result in

unreliable statistics, and since Warwick's conclusions were drawn

mainly from the 1961 data.

Fig. 27 shows the variation of source B in System III longi-

tude as calculated from the 1961 and 1962 partial apparition histo-

grams, using the "old" period (91 55rm 29735). Ho error bars are shown,

even though the best values should be regarded as accurate to no bet-

ter than + 30. For both years source B- apparently drifts to lower

longitudes mntil opposition, and then drifts back to higher longitudes.

Source B2 seems to maintain its longitude quite well throughout the

apparition. The center of source E, as a whole, seems to behave some-

w:hat like source B. Mention should be made that the position of

source B -ras chosen as midi.ay between the minima at both ends of the













S h- -


-- O B2
x . . .. x J


0


.X


CENTER
OF B





Bi


f' .. -


MERGED


FLA.-CHILE, 1961


600 -0- MERGED FLA.-CHILE, 1962

A


-10O


-80


80


160


MEAN EPOCH W.R.T. OPPOSITION IN
DAYS

Fig. 27.--i8 I /'s source B drift studies using
and 1962 data


1600


1400-


fr-
Lo

.. 1200
#, .


100ooo


800 -


191


S


I









source. This is the most expedient, but perhaps not the most reliable,

method for choosing the center of this weakly-defined, skewed source.

The centers of the sources B1 and B2 were chosen by sketching gaussian

curves which approximated the sources, and choosing the centers of the

curves.

In order to get more extensive data, the "before," "around,"

and "after" opposition partial histograms were merged for 1961-1962.

This was done for both the "old" period (9h 55m 29S35) and the "new"

period (9 55 30.70). Studies similar to those in Fig. 27 were made

and these are shown in Fig. 28. The probable errors of these points

are less than those in Fig. 26. We see the "eyebrow" motion of source

B1 for either period, the irregular motion of source B2, and again,

the "eyebrow" motion for the entire source. At any rate, our data

fail to show the extensive yearly drift of source B to lower longitudes

that was seen by Warwick (4). Incidentally, Warwick's data do show the

apparent long-time source shift to higher longitudes that was mentioned

in the preceding section.


Possible Correlation of Radio Source
Drift with Red Spot Drift

The University of Florida group compared the rotational period

of the radio sources with the rotational periods exhibited by various

Jovian optical prominences as early as 1957 (18). At the suggestion of

A. G. Smith we again conducted such a study, this time to compare pos-

sible changes of the Red Spot rotational period with the observed change

in the rotational period of the radio sources.

The System II longitude of the Red Spot is shown in Fig. 29














MERGED FLORIDA-CH!LE, 1961-1962


1600





I-0o.- B2


O- 2' 0'
S1200'
> CENTER OF B
/ x


100 "
o



800




600--0- "NEW" PERIOD (9 55 m 3070)
---X-- "OLD" PERIOD (9h55m 29 35)

-160 -80 0 30 160
MEAN EPOCH W.R.T. OPPOSITION IN DAYS

Fig.26.--18 iIc "s soouce E drift studies using the
merged i961-1962 data














MOTION OF RED SPOT
1965- IN SYSTEM 1
0
Ty= 9h 55m 40%632 o

O
1960-
6T= i.71S


1955 -




1950- #


/O T=0.E98S

1945 -



200 2500 3000 3500 40'
LONGITUDE, SYSTEM 7
F'ig. 29.--Tne -.variation of the System II lor ni-
tude of Jupiter' s Creat Red Spot


Ld
,>Y









as a function of time. The solid line represents a drift of 7.750/ .yJear,

or if expressed as an effective period change, 0.393 seconds. In about

195 the Red Spot started to drift 14.,,/year in System II. In order

to examine this change of drift more closely,, we plotted its longitude

for a period of T =9 t= 1 551(30 (System "Red Spot"), whichh was

arrived at by adding 0.898 seconds to the System II period, TI =

9 5m ,63o2. The graph shown in Fig. 3-0 resulted. (The plot in Fig.

23 is also shorn in Fig. 30.) As can be seen in the graph, the Red Spot

started to shift to higher longitudes about one year earlier than did

the radio source. The Red Spot started to shift 9^/y.ear in 1959 and the

radio source started to shift 10.~ /year in 1960, in their respective

longitude systems. The fact that the slopes of the tw'o drift lines are

so nearly, equal and that they.' match so closely in time suggests that

the drift motions of the Fed Spot and the radio source, in their respec-

tive longitude systems, must be related in some way.

In an effort to find the best representative period of the Red

Spot, Peaek (19) tab.iulated its longitude in System II for many years.

Using the tabulated data, he arrived at a best period of T = h 3 .

Fig. 31 shows the longitude of the Red Spot versus time using Peek's

period, assuming that the longitude in this special system w.as -26. in

189. Equation (2 ) gives the longitude in the special system as a

fu action of A- (System II longitude) and t.


A= T 26. + ,+?62 t, (25)


where t = (time in years 19kh) : 365.25,-/398.. (26)


It is interesting to note the difference between the Fed Spot period










DRIFT

RED SPOT


STUDIES

RADIO SOURCE


2100 2300
SPECIAL LONGITUDE
SYSTEM WITH
Ts= 9h 55m41530


220 2400 2600
LONGITUDE, SYSTEM
I (1957)
Ti 9h55m29 35


Fig. 30.--Time variation of the longitude of Jupiter' s
Great Red Spot for a period T;S which maintains the Red Spot
lonpgitude constant from 1945-1958 compared rith Fig. 23








196


195


LONG-TERM
OF GREAT
SPOT


DRIFT
RED


T = 9h 55m 37s58


1860


1S50


1340


-480' -240' 0' 240" 480- 720-
LONGITUDE
Fig. 31.--Timie variation of the longitude of Jupiter's Great
Red Spot since 1835, using a special period T = 9 5m 37958.
Eased largely on Peek (16)









h m s h m s
(9 55 37.58) used by Peek (19) and the period (9 55 41.530) ar-

rived at by using the data since 1945. This suggests that if the Red

Spot drift is related to the radio source drift, determination of a

period which maintains the radio source longitudes constant may be im-

possible.


PROBABILITY VERSUS HOUR ANGLE

The 1962 analysis initiated several new studies, one of which

was an examination of the variation of Jupiter radiation with hour

angle. The computer tabulated several parameters with each five de-

grees of hour angle from -90 to 90 (east to west), but the only param-

eter that we will discuss here will be the probability of Jovian emis-

sion. These histograms also can be loosely construed as being proba-

bility versus elongation studies, since most of the activity received

at hour angles between -900 and 0 occurred prior to opposition and

most of the activity received at hour angles between 0 and 90 oc-

curred after opposition, although it is easy to see that this 1ras not,

by any means, arlays the case. U-eren any interval iras listened to less

than 30 times no points were plotted; hence, the sharp cutoff on each

side of the histograms.

Fig. 32 shois the probability versus hour angle histograms for

the Florida data. We can see the effect of the impedance mismatch

early in the apparition for the 15 lic's data by the fact that the his-

togramr is skewied to the west. It is very interesting to note that both

the 18 iic/s and the 22.2 i.-c/s histograms reach maxima at approximately

3,0 east. Perhaps this is caused by the fact that the ionosphere is








1962 DEFINITE


-80 -60 -40 -20
HOUR ANGLE


0 20 40 60
IN DEGREES


Fig. 32.--Probabilityr -ersus hour angle histograms
for the data taken at the Florida station in 1962


80


DATA


FLORIDA









much clearer after midnight than it is before midnight, thus making

observations before opposition more favorable. The 27.6 Mc/s histo-

gram also shows a tendency to peak in the east.

Fig. 33 she .s the probability versus hour angle histograms

for the Chile data. The 16 and 22.2 Mc/s histograms are missing,

since the "new" analysis was not performed for the polarimeter data.

The 5 Mc/s histogram shows the beam width of the fixed broadside an-

tenna. The rest of the histograms again show that the probability

peaks in the east. Note, also, the dip at 50 west on the 5 and 10

Mc/s histograms. These are the only histograms on which this occurs.

It appears that some special phenomenon is causing an attenuation of

the low frequencies incident at 50 west. We shall examine future

data for this effect. The 15 and 18 Mc/s histograms show slight minor

peaks between 400 and 500 west. The 27.6 Mc/s histogram is not in-

cluded, since the data at that frequency were very sketchy.

Another possible explanation of the skewed nature of the histo-

grams is that the 00 to 900 intervals recorded many more listening

counts than did the -900 to 00 intervals, thus including more bad lis-

tening time, i.e., time close to sunset.


PROBABILITY VERSUS UNIVERSAL TIME

Since the dependence of the probability on Universal Time has

not been studied, it seems worthwhile that we should include it here.

Figs. 34 and 35 show Florida and Chile histograms for just such stud-

ies. The statistics become poor at each extreme on the curves, since

the listening time decreased for times close to sunrise (1200 U.T.)

and sunset (2300 U.T.).








CHILE 1962 DEFINITE


10.0 MC/S


5.0 MC/S


-80 -60 -40
HOUR


-20
ANGLE


20 40 60
IN DEGREES


Fig. 33.--Probability,- versus hour angle histograms
for the data taken at the C:ile station in 1962


80


DATA







FLORIDA


27.6 M C/S


22.2 MC/S


18.0 MC/S


15.0 MC/S


POS. AND DEF


POS. AND DEF


DEF


DEE


0100 0300 0500
UNIVERSAL


0700 0900
TIME


Fig. 34.--Probability versus Universal Time histograms
for the data taken at the Florida station in 1962


2300


1100


1962







CHILE 1962


POS. AND
DEF.


18.0 MC/S


DEF.


15.0 MC/S


POS. AND


5.0 MC/S


DEF.


0300 0500 0700 0800 1100
UNIVERSAL TIME


Fig. 35.--Probability versus Universal Time histograms for
the data taken at the Chile station in 1962


.04
.02


.2 F


. ^sr-


DEF


000
0100




Full Text

PAGE 1

DECAMETER-WAVELENGTH RADIO OBSERVATIONS OF THE PLANETS IN 1962 By GEORGE ROBERT LEBO A DISSERTATION PRESENTED TO THE GRADUATE COUNOL OF THE UNIVERSITY OF FLORIDA TN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA August, 1964

PAGE 2

ACKNOWLEDGMENTS The author wishes to express deepest gratitude to his committee chairman and adviser Dr. Alex G. Smith for his capable and patient guidance. Without his help this project would not have been possible. Thanks are also due Dr. Thomas D. Carr for many helpful suggestions and discussions, and Drs. F. E. Bunnam, G. C. Omer, and R. G. Blake for serving on his committee. The data in this thesis could not have been collected by one person. The author wishes to thank the following staff members for carrying out routine observations dijiring the 1962 apparition: T. Anderson J. Aparici W. F. Block H. Bollhagen G. W. Brown W. Cain T. D, Carr M. L. Eager lin S. Gulkis A. T. Jusick R. J, Leacock J. Levy E. J. Lindsey J. May W. Mock C. N. Olsson I . Shever R. F. Six, Jr. A. G. Smith C. F. Tiberi G. Walls. R. R. Hayward, C. A. Arlington, and S. L. Danielson proved to be of great assistance in reducing the data, and W. W. Richardson deserves highest commendation for his untiring work on the drawings. 11

PAGE 3

The expert editorial and typing services rendered by Mrs. Ruth Pierce lightened the puhlication burden immeasurably. The author also wishes to thank R. A. Smith and the rest of the staff at the University of Florida Computing Center for their expert advice in writing the computer programs. Without the use of the I. B. M. 709 facility at the University of Florida Computing Center, the data analysis in this thesis could not have been undertaken. The writer expresses thanks to I)r. S. S. Ballard, head of the Physics department, for supporting him in the form of teaching assistships and to the U. S. Army Research Office Durham, the National Science Foundation, and the Office of Naval Research for their aid in the form of research assistant ships. The author is indeed grateful to his parents and to his wife's parents for their encouragement. The author's wife deserves more than appreciation for her patience, encouragement, and hard work during six years of school life. Her devotion and understanding helped as nothing else could. It is to her that this thesis is dedicated. Ill

PAGE 4

TABLE OF CONTENTS ACKNOWLEDGMENTS ^^ LIST OF TABLES ^^^ LIST OF FIGURES '^^^^ Chapter I. INTROnJCTION -^ Instr\xnentation II. CALIBRATION PROCEDURES ^ Basic Concepts ^ The Florida Calihrator ^ The Chile Calibrator ' Calibration Using the Chile Calibrator 9 Use of the Calibration 9 Receiver Characteristics III. ANALYTICAL PROCEDURES ^^ 21 Data Reduction Data Reduction as Performed in 1957-I96I and Continued in I962 "Old Analysis" .... 21 Data Reduction as Started in I961 by Brown and Continued in I962 "New Anal^^rsis" .... 23 Jupiter Programs for the I.B.M. 709 ^^ of. The "Old Program" • ^g The "New Program" • IV. PROBABILITy STUDIES 37 •37 Gross Statistics -" j-v

PAGE 5

TABLE OF CONTENTS (Continued) Page Erobability of Observing Jupiter Radiation as a Function of Jovian Longitude ^0 Data from Individual Channels ^0 Merged Data ^^5 Drift Studies of Source A 62 Drift Studies of Source B 72 Possible Correlation of Radio Source Drift with Red Spot Drift TU Probability Versus Hour Angle 80 Probability Versus Universal Time 82 Probability Versus Time From Sunset 86 Probability Versus The Time From Sunrise , . 89 Probabil-ity Versus Jupiter Elongation ........ 92 V. INTENSITY STUDIES 97 Intensity Distribution 97 Gross Statistics 101 Determination of Calibration System Reliability . . ' . I05 Intensity Versus System III Longitude IO8 Hour Angle Studies 115 Diurnal Studies 117 Time-From-Sunset Studies ' II8 Time-From-Sunrise Studies 121 Elongation Studies 121 Search for Periodic Recurrences 126 VI. SOLAR AND GEOPHTSICAL CORRELATIONS WITH JOVIAJS EMISSION 132 Long-Term Effects 132 Short-Term Correlations 135 Correction of the I96I Analysis 135 Changes in the Solar Flare Program 1^3 Chree Analysis Results of the I962 Solar Correlation Studies 1^7 Conclusions and Suggestions for Further Studies 185 VII. DECAMETER-WAVELENGTH OBSERVATIONS OF SATURN, MARS, AND VENUS IN I962 AND I963 192 Saturn 192

PAGE 6

TABLE OF CONTENTS (Continued) I^ge Introduction and Instrumentation 192 Data Reduction and Analysis 193 Venus and Mars 197 yill. SUMMARY MD REMARKS OK THE COMPARISON OF THE EXPERIMENTAL RESULTS WITH EXTSTmC THEORIES 202 Theories as to the Origin of Jupiter's Decametric Radiation 202 20l+ Summary LIST OF REFERENCES 210 BIOGRAPHICAL SKETCH 2l4 VI

PAGE 7

LIST OF TABLES TalDle ^ge 1. Equipment Description 3 2. Gross Statistics 38 3. Activity Correlation Study 39 k. Source Position as a Function of Frequency 60 5. Criteria for Determining RelialDility of the Averages in Figs. h'J and kQ 112 6. Days of Peak Activity used in the Chree Analysis of the 18 Mc/s Data in I962 1^8 7. Days of no Activity and Days of Peak Activity in Jovian Longitude Regions from 0° to 90° and 90° to 190° for the 18 Mc/s Florida Data in I962 172 8. Days of Peak I8 Mc/s Activity in Jovian Longitude Regions from 190° to 290° and 290° to 360° and Days of Peak 27.6 Mc/s Data in I962 173 9. Gross Statistics for the Saturn Ohservations in I962 .... 19^ 10. Pulses Labelled "Possible Saturn" recorded in I962 195 11. Gross Statistics for the Mars and Venus Observations in 1962 200 Vll

PAGE 8

LIST OF FIGURES Figure j^g^ 1. Simple noise generator based on two Sylvania 5722 noise diodes i^ 2. ALiplifier used with the calibration system at the Chile station d-aring I962 S ZBlock diagrams of calibrator Systems "A," "B," and "C" used at the Chile station in I962 10 h. Examples of Chile calibrations. Top: Systems "A/' "B," and "C" calibrations at the Chile station (Calibrations using syste:.-s A and B were not normally this complete.) Center: Florida calibration. Bottom Systems "A/' "B," and "C" calibrations as they were nonnally performed at the Chile station yL 5. Visualization of the gain calculation using the Chile calibrator system 12 6. Receiver-recorder response curves used as an aid to calculate the gain of the amplifier in the Chile calibration systems (lOO chart deflection units is full scale deflsction) ]_5 7. Systems "B" and "C" calibrations made during the watch to determine the noise contributed by the Chile amplifier. . I7 8. Receiver -recorder response curves at 15.O Mc/s and 22.2 Mc/s using Ct _lins receivers and Texas recorders (lOO deflection units represent full scale deflection) .... 18 9. Hammarlund and Rodhe-Schwarz response curves at 10.0 and 5.0 Mc/s, respectively I9 10. Average probability of observing activity at the Chile Station in I962 kl 11. Probability histogram for the iB.O Mc/s data taken at the Florida station in I962, plotted in Polar coordinates ^2 Vlll

PAGE 9

LIST OF FIGURES (Continued) Figure Page 12. ErotalDility histograms for data taken at 15-0, I8.O, 22.2, and 27.6 Mc/s at the Florida station in I962. ... kk 13. Prohahility histograms for data taken at 5.O, 10.0, 15.0, 16.0, 18.0, 22.2, and 27.6 Mc/s at the Chile station in I962 h6 ik. Smoothed, merged probability histograms for data taken at 10.0 Mc/s U8 15. Smoothed, merged probability histograms for data taken at 15.0 Mc/s 49 16. Smoothed, merged probability histograms for data taken at 18.0 Mc/s 51 17. Smoothed, merged probability histograms for data taken at 22.2 Mc/s 53 18. Smoothed, merged probability histograms for data taken at 27.6 Mc/s 5^ 19. Smoothed, merged histograms for all the frequencies monitored, using only I962 data ^6 20. Smoothed, merged histograms for all the frequencies monitored, using all the data recorded since observations began in 1957 58 21. Variation of source position with frequency (The dashed lines provide an alternate selection of source positions. The point labelled 7 niay not be source C). . . . 59 22. Probability histograms for the i960, I961, and 1962 data taken at the Florida station 63 23. Yearly variation of the position of the 18 Mc/s source A . . 6k 2k. Variation of the position of the I8 Mc/s source A since i960 using partial apparition histograms 66 25. Probability histograms for the I8 Mc/s data taken at the Chile station in I962 using the "old" (Tjjj = 9^ 55™ 29?35) and the "new" (Tjjj = 9^ 55"^ 30?70) periods 70 IX

PAGE 10

LIST OF FIGURES (Continued) Figure I^ge 26. The 1957-1962 smoothed^ merged 18 Mc/s histogram^. using the "old" (Titi = 9^ 55"^ 29?35) period ui.til i960 and the "new" (Tjjj = 9^ 55"^ 30?70) period after that date 71 27. 18 Mc/s source B drift studies using 196I and 1962 data . . 73 28. 18 Mc/s source B drift studies using the merged I96I1962 data 75 29. Time variation of the System II longitude of Jupiter's Great Red Spot 76 30. Time variation of the longitude of Jupiter's Great Red Spot for a period Tpg which maintains the Red Spot longitude constant from 19^5-1958 compared with Fig. 23 78 -m 31. Time variation of the longitude of Jupiter's Great Red Spot since 1835^ using a special period T = 9^ 55 37?58. Based largely on Peek (16) 79 32. Erobahility versus hour angle histograms for the data taken at the Florida station in I962 Bl 33. EroTsability versus hour angle histograms for the data taken at the Chile station in I962 83 3ij-. Erohahility versus Universal Time histograms for the data taken at the Florida station in I962. 8i<35. Probability versus Universal Time histograms for the data taken at the Chile station in I962 85 36. Probability versus time-from-sunset histograms for data taken at the Florida station in I962 87 37. Probability versus time-from-sunset histograms for the data taken at the Chile station in I962 88 38. Probability versus time-from-sunrise histograms for the data taken at the Florida station in I962 90 39. Probability versus time-from-sunrise histograms for the data taken at the Chile station in I962 91

PAGE 11

LIST OF FIGUEES (Continued) Figure ^^e UO. Smoothed probability versus time-from-sunrise histograms for the data taken at I8.O Mc/s in I962 93 kl. Variation of the average monthly probability vith time and Jupiter elongation, for data taken at the Florida station in I962 95 k2. Time variation of the average monthly probability for data taken at the Chile station in I962 9^ k^. Intensity distribution for the I8.O Mc/s data taken at the Florida station in I962 100 kk. Frequency variation of the average fl\ix; density, S^^ 103 45. Frequency variation of the average flux density, S^ -^^^ U6. Comparison of the S^ versus System III histograms for the 18.0 Mc/s data using calibrations ("with cal" ) with those using the cosmic noise level as a reference ("no cal") IO8 47. Smoothed S^ versus System III histograms for data taken at the Florida station in I962 (The dashed lines represent regions of poor statistics.) 109 k8. Smoothed S^ versus System III histograms for data taken at the Chile station in I962 Ill ^4-9. Smoothed Sl versus System III histograms for data taken at the Florida station in I962 113 50. Smoothed Sl versus System III histograms for data taken at the Chile station in I962 11^ 51. Smoothed Sl and Sa versus hour angle histograms using the I8.O Mc/s data taken in I962 II6 52. Smoothed Sl and S^ versus Universal Time histograms using the 18.O Mc/s data taken in I962 119 53. Smoothed Sl and S^ versus time-from-sunset histograms using the I8.O Mc/s data taken in I962 120 54. Smoothed Sl and Sa versus time-from-sunrise histograms using the I8.O Mc/s data taken in I962 122 XI

PAGE 12

LIST OF FIGURES (Continued) Figure Page 55. Variation of the monthly average activity index rate with time and Jupiter elongation^ for the Florida data taken in I962. . '' 124 56. Time variation of the monthly average activity index rate for the Chile data taken in 1962 125 57. I^ily activity index rate for the 5,0 and 10.0 Mc/s data taken in I962 12? 58. Daily activity index rate for the I5.O Mc/s data taken in I962 128 59. Eaily activity index rate for the I8.O Mc/s data taken in I962 1^9 60. Daily activity index rate for the data taken at 22,2 and 27.6 Mc/s in I962 130 61. Comparison of the width of source A and of the average probability of emission with sunspot number 133 62, Comparison of Jupiter's solar latitude and the apparition average activity index rate with sunspot number 136 63. Chree analysis of the solar flare activity index in groups one^ two^ and three^ using the 20 peak days of 18,0 Mc/s Jupiter emission at the Florida station in 1961 138 6h. Regions on the solar disk as viewed from Jupiter assigning flares to groups one, two, or three 139 65. Chree analysis of solar flare activity index in groups one, two, and three, using the 20 peak days of 27.6 Mc/s Jupiter emission at the Florida station in I961. . l^J-l 66. Chree analysis of solar flare activity index in groups one, two, and three, using the 20 peak days of I8 Mc/s Jupiter emission at the Florida station during the three months around opposition in I96I. 1^2 67. Chree analysis of solar flare activity index in groups one, two, and three, using the 20 peak days of I8 Mc/s Jupiter emission at the Chile station in I96O ihk Xll

PAGE 13

LIST OF FIGURES (Continued) Figure Ps.ge 68. Chree analysis of the solar flare activity index in groups one, two, and three, using the 20 days of peak activity of the l8 Mc/s Jupiter emission at the Florida station in I962 1^9 69. Chree analysis of the number of flares in groups one, two, and three, using the 20 days of peak Jupiter activity at I8 Mc/s at the Florida station in I962 151 70. Chree analysis of the 28OO Mc/s solar flux, the Zurich provisional sunspot number {B.-^) , the geomagnetic index (Ap), and the Jovian daily activity index rate using the 20 days of peak Jupiter activity at 18 Mc/s at the Florida station in I962 153 71. Chree analysis of the solar flare activity index in groups one, two, and three, using the 21 days of peak activity of the I8 Mc/s Jupiter emission at the Chile station in I962 155 72. Chree analysis of the number of flares in groups one, two, and three, using the 21 days of peak Jupiter activity at I8 Mc/s at the Chile station in 1962 157 73. Chree analysis of the 28OO Mc/s solar flux, the Zurich provisional sunspot number (Rz)^ "the geomagnetic index (Ap), and the Jovian daily activity index rate, using the 21 days of peak Jupiter activity at 18 Mc/s at the Chile station in I962 I58 jh. Chree analysis of the solar flare activity index (excluding subf lares) in groups one, two, and three, using the 20 days of peak activity of the I8 Mc/s Jupiter emission at the Florida station in I962 160 75. Chree analysis of the number of flares (excluding subf lares) in groups one, two, and three, using the 20 days of peak Jupiter activity at the Florida station in I962 I6I 76. Chree analysis of the solar flare activity index in groups one, two, and three, using the 20 days of peak Jovian activity (Florida plus Chile) monitored by both stations at I8 Mc/s during the three months around opposition in I962 163 -Xlll

PAGE 14

LIST OF FIGUEES (Continued) Figure i^ge 77. Chrae analysis of the numlDer of solar flares in groups one J twO;, and three^ using the 20 days of peak Jovian activity (Florida plus Chile) monitored by both stations at 18 Mc/s during the three months around opposition in I962 l6k 78. Chree analysis of the 28OO Mc/s solar flux^ the Zurich provisional sunspot number (Rz) > "the geomagnetic index (Ap);, and the Jovian daily activity index rate^ using the 20 days of peak Jupiter activity (Florida plus Chile) monitored by both stations at 18 Mc/s during the three months around opposition in I962 . . . I66 79» The solar wind and the interplanetary magnetic field. After Parker (25) I68 80. Chree analysis of the 28OO Mc/s solar flux, the Zurich provisional sunspot number (Rz), and geomagnetic index (Ap), and the Jovian daily activity index rate, using the 29 days of "no activity" at both stations during the six months around opposition in 1962 170 81. Chree analysis of the 28OO Mc/s solar flux, the geomagnetic index Ap, the sunspot number R^, and the solar flare activity indices for regions one, two, and three, using the 11 days of peak Jupiter activity received in the 0° to 90° ^III interval on the 18 Mc/s, Florida channel in 1962 I74 82. Chree analysis of the 2800 Mc/s solar flux, the geomagnetic index Ap, the sunspot number R^, and the solar flare activity indices for regions one, two, and three, using the I7 peak days of Jupiter activity received in the 90° to 190° A^-jj interval on the 18 Mc/s, Florida channel in 19o2 I75 83. Chree analysis of the 2800 Mc/s solar flux, the geomagnetic index Ap, the sunspot n-umber Rz, and the solar flare activity indices for regions one, two, and three, using the I9 days of peak Jupiter activity received in the 190° to 290° Xjij interval on the 18 Mc/s, Florida channel in 1962 I76 XIV

PAGE 15

LIST OF FIGURES (Continued) Figure Page 8^. Chree analysis of the 2800 Mc/s solar flux, the geomagnetic index Ap, the sunspot number B.^, and the solar flare activity indices for regions one, two, and three, using the 21 days of peak Jupiter activity received in the 290° to 360° Ajj-r interval on the I8 Mc/s, Florida channel in I962. . . . 177 85. Double correlation studies of the 2300 Mc/s solar fliox Chree analysis with the Chree analyses of the solar flare activity indices in regions one, two, and three, using the 20 days of peak Jupiter activity monitored at I8 Mc/s by the Florida station in 1962 179 86. Double correlation studies of the geomagnetic index Chree analysis with the Chree analyses of the solar activity indices in regions one, two, and three, using the 20 days of peak Jupiter activity monitored at 18 Mc/s by the Florida station in 1962 180 87. Chree analysis of the solar activity indices (FXl, FX2, and FX3), using the I9 days of peak 27.6 Mc/s Jupiter activity received at the Florida station in 1962 182 88. Chree analysis of the solar flare numbers (FWl, FN2, and FNS), using the 19 days of peak 27.6 Mc/s Jupiter activity received at the Florida station in 1962 183 89. Chree analysis of the 28OO Mc/s solar flux, the geomagnetic index Ap, the sunspot number Ez, and the Jovian activity index rate, using the 19 days of peak 27.6 Mc/s Jupiter activity received at the Florida station in I962 l84 90. Location of Saturn bursts in activity longitude systems plo't ' ed for rotational periods of 10^ k^'^.6 and 11^ 57°^. 8, where the longitude was assumed to be zero at 0^ U.T. on May 10, I962 . . I98 XV

PAGE 16

CHAPTER I IKTROrUCTIOW The decameter radio emission from the planet Jupiter, first observed accidentally by Burke and Franklin in 1955 (l), has been monitored by the radio astronomy group at the University of Florida since 1957. Such extensive "coverage" can be justified merely by noting the existence of numerous conflicting theories purporting to explain this strange electromagnetic phenomenon (2), (3), {h), (5), {6), (7). These theories will be discussed in a later chapter. The results of the 1962 observations are offered here as experimental evidence to aid in deciding which, if any, of these theories is correct. This analysis must, of course, be considered as a continuation of the analyses from previous years by T. D. Carr (8), N. E, Chatterton (9), and N. F. Six (10). Details of data-taking, execution of the watch, and data analysis that were not changed for the I962 apparition (observing season) are not discussed here, since N. F. Six (lO) covered these in his dissertation. During the I962 apparition observations were also made of Mars, Venus, and Saturn. INSTRUMENTATION The 1962 observations were made from two sites: one at the University of Florida campus in Gainesville, Florida, and the other near

PAGE 17

Maipu, Chile, about twenty miles from Santiago. At the Chile station, observations of Jupiter were made in I962 at frequencies of 5, 10, 15, 16, 18, 22.2, and 27.6 Mc/s, while simultaneously observations were being made at frequencies of 15, 18, 22.2, and 27.6 Mc/s in Florida. At each frequency an antenna was connected to a commercial communications receiver and two pen recorders. One of these, a low-speed recorder, ran during the entire listening time at a chart speed of six inches per hour; the other, a high-speed recorder, ran only when Jovian activity was received. The chart speed of the latter recorder was 5 mm/sec, 25 mm/sec, or 125 mm/sec, depending on which speed was needed to resolve the pulses. The data analysis, as discussed in this paper, primarily involves data taken on the low-speed Texas recorders, except at 15 Mc/s (Chile) and 27.6 Mc/s (Chile) for which the low-speed data were taken on Esterline-Angus recorders. Pertinent information regarding the antennas and their receivers is given in Table 1.

PAGE 19

CHAPTER II CALIBRATION PROCEDURES The analyses of preceding apparitions were used predominately for probability studies. If, on the other hand, accurate intensity studies are to "be made, the received signal intensity must "be known as a function of time. By dividing the activity periods into small intervals and determining the intensity for each interval, this time dependence can be established, as vas done by G. W. Brown (ll) and T. D. Carr et. al . (12). Calibrators were installed at each station for just such a purpose. At the end of every watch each channel which had recorded activity was calibrated with a noise generator composed basically of two Sylvania 5722 noise diodes connected in parallel (Fig. l). This calibration was necessary each time radiation was received, since the gain of a receiver might drift from day to day. RECEIVER R VARIAC Fig. 1. --Simple noise generator based on two Sylvania 5722 noise diodes

PAGE 20

BASIC CONCEPTS Nyquist's Lav, P = kT Af, (l) where k is Boltzmann's constant (I.38 x 10"^^ joules /°K), T is the resistor temperature in degrees Kelvin, and Af is the range of frequencies over which the noise power P is to be measured, gives the noise power produced by a resistor at a temperature T. What would the resistor temperature have to "be in order to yield sufficient noise power to calibrate a typical Jovian pulse? Consider the tlux density S (power per unit area per unit of bandwidth) as being received by an antenna of effective area A connected to a receiver of bandwidth Af . The power then can be written as P = 1/2 SA Af, (2) where the l/2 is introduced because the antenna is sensitive to only one of the two linearly polarized components of the radiation. Equation (l) with Equation (2) leads to: T = 1/2 Si. . (3) k Substituting 720 m for A and lo" "^w m"^cps""^ for S (a typical value for Jovian decameter activity at 18 Mc/s), one finds T to be about 2.6 x 10 °K. Since a resistor temperature this high is obviously out of the question, a calibrator based on noise diodes is used (Fig. l). If I is the d.c. component of the plate current, i the r ~.s.

PAGE 21

fluctuation of the plate current, and R the terminating resistance of of a Sylvania 5722 noise diode operating over a frequency range Af, then by Schottky's Law: i = V2el Af, and (^) M -•^^ p_ eIR Af . (5) In the above, e is the electronic charge (1.6 x 10"^^ coul. ) and i/2 instead of i is used in Equation (5) since only one-half of the a.c. plate current reaches the receiver. The other half is dissipated in the terminating resistor R. Using Equations (2) and (5), one arrives at: I = SA/eR. (5) If one substitutes the same values for S, A, and e that vere used in Equation (3); along with R = 50^1, I becomes 90 m.a. The maxmum d.c. plate current that can be used with the Sylvania 5722 noise diode is 35 m.a. Even with two diodes connected in parallel (l = 70 m.a.), one would not be able to obtain a calibration deflection equal to the deflection resulting from a flux density of lO'^^w m'^cps"^. Noteworthy, also, is the fact that radiation from the giant planet often exceeds the cited flux density, especially at frequencies below 18 Mc/s. At 5 Mc/s, pulses have been recorded which correspond to calibration currents as high as 70 amps I Such currents are certainly not readily created by connecting additional noise diodes in parallel if each diode supplies no more than

PAGE 22

7 35 m.a. Since higher -current noise diodes are not available, the only recourse is either to amplify the noise generator signal or to attenuate the signal from Jupiter. The first alternative was tried at the Chile station during the I962 apparition. THE FLORIDA CALIBRATOR Fig. 1 sufficiently describes the simple calibrator used at the Florida station. Regardless of how limited this system proved to be, it was a step toward better calibration and a still better system is being used during the 1963 apparition. THE CHILE CALIBRATOR In Chile, the first improvement was to replace the plate resistor R, shown in Fig. 1, with a higher value (S 450^). This was done simply by inserting a 375A resistor in series with the original 75J^ plate resistor (the receivers were tuned to 75 SI input impedance in Chile instead of the 50iT. input impedance used in Florida). The resulting calibrator-to-receiver mismatch was corrected by connecting the two with a coaxial cable cut to an appropriate length. This procedure made possible an equivalent noise generator gain (K) of about six. With the additional resistor and the appropriate matching cable inserted, the calibration was labelled "high range," while, with the noise generator alone, the calibration was labelled "low range." The calibrator in "high range" still generated insufficient noise power to allow calibration of the stronger Jovian pulses. An amplifier (Fig. 2) was, therefore, inserted with the noise generator in "high range," thus producing a calibration system known as

PAGE 23

V, 6SG7 Vg 6SG7 RESISTORS -OHMS CAPACITORSMfd C, 6800 Pf C2 6800 Pf L, SIZED FOR A RANGE OF 15-20 MC/S L2 SIZED FOR A RANGE OF 5-10 MC/S ;NPU J_OUTPUT ^ i\ ::::: > .002 • + 29 V. • +I55V. REG. O-vfiilfiQiLrj AMPLIFIER OF CHILE CALIBRATOR Fig. 2 . —Amplifier used with the calihration system at the Chile station during I962

PAGE 24

"System C" (see Fig. 3)' This made possible pen deflections corresponding to nearly all of the pulses of Jovian origin. Ohviously, for satisfactory calibration^ the equivalent current I (that current through a eq noise diode which would produce the given pen deflection) had to be known, since the flux density S is found by inserting 1^^ into: S = -ffsi, (7) A an expression arrived at merely by rewriting Equation (6). In order to find the equivalent current, the dependence of the amplifier gain (G) on the calibration current (l) had to be measured as described below. As a further modification of the system, an atten'uator of known impedance was inserted into "System C," thus producing what was known as "System A" (Fig. 3). The noise generator alone in "low range" was known as "System B." CALIBRATION USING THE CHILE CALIBRATOR After each watch the receiver gain was increased, so that full scale deflections could be obtained with "System B" : calibrations were then made using Systems "A" and "B." Fig. h shows these calibrations as they appeared on the Texas records. Note that the calibrator used in Florida corresponded to "System B" in Chile. However, the Florida calibrator was used during the watch; hence, it was used at lower receiver gain than that used for the "System B" calibrations in Chile. USE OF THE CALIBRATION Fig. 5 illustrates the procedure used for finding the gain (g)

PAGE 25

10 SYSTEM A HIGH RANGE NOISE AMPLIFIER GENERATOR ATTENUATOR RECEIVER RECORDER SYSTEM B LOW RANGE NOISE GENERATOR RECEIVER RECORDER SYSTEM C HIGH RANGE J NOISE GENERATOR AMPLIFIER RECEIVER RECORDER Fig. 3--Block diagrams of calibrator Systems "A/' "B/' and "C" used at the Chile station in I962

PAGE 26

11 'ir ""• •""'"tlliiiililtliniiiiltniltiiiininini.i.f lit. ,, Fig. h, — Examples of Chile calibrations. Top: Systems "A/' "B/' and "C" calibrations at the Chile station (Calibrations using systems A and B were not normally this complete.). Center: Florida calibration. Bottom Systems "A/' "B/' and "C" calibrations as they were normally performed at the Chile station

PAGE 27

12 SYSTEM A HIGH RANGE l5H=3 NOISE '"I AMPLIFIER ATTENUATOR RECEIVER GENERATOR SAME Tl PEFLECTION LOW RANGE SYSTEM B NOISE B, GENERATOR RECEIVER %'^ -^ ""n = p, 3, REC. RECORDER SYSTEM A HIGH RANGE -i-A NOISE ''^2amPLIFIER ATTENUATOR RECEIVER GENERATOR SYSTEM B LOW RANGE J-' NOISE GENERATOR RECEIVER Pa G -^ Pm Ag N = P B, SAME DEFLECTION REC. RECORDER Fig. 5' — Visualization of the gain calculation using the Chile calihrator system

PAGE 28

13 of the system. G equals the product of the amplifier gain and the "high range" over the "low range" gain (K). P. P». are the respective noise powers created by the noise generator with cuirrents I. I» ("System A"), P-n P-g. are the respective noise powers created Toy the noise generator with currents I-g I__ ("System B" ) , P^ is the noise power inherent in the amplifier^ and OC is the attenuation coefficient of the attenuator. With current I„ in "System A" and 1-Q in "System B/' a deflection 0-, is obtained for each system; thus: Pa, g(PaJ ^P^ ^ Pb, ^ (8) OC By changing I_^ to I_^ and I-g to I-o to obtain a deflection 02 ^°^ each system, the relation Pa ^(Pa ) P Pb OC results. Subtraction of Equation (9) from Equation (8) leads to: Pa. g^% ) X ^(^a,) = ^ (^Bi PBg)(10) But, from Eq-oation (5) we see that the power P. is proportional to the current Ij_, hence Equation (lO) can be written: IA-l GdA,) lAg G^Ag) = ^ ^^\ ^Bg)(11) /Ia lAp\ If G(I_^ ) = G(I, )sG[ — ^^-^ -})^^ approximation that is reasonable, if Ia Ia is small. Equation (ll) reduces to:

PAGE 29

Ik lAi + IAj (12) Fig. 6 is a plot of the 15 Mc/s "System A" and "System B" calibrations shown in the top record of Fig. h. From it can he found values for Ij^-, , I-n , '^Ao' ^^^ -^Bo' Equation ( 12 ) yields G(36 m.a.) = 28 if the values used are: Ij^n = ^3-7 m.a.;, Ig^ = 15.8 m.a.^ Ij^p = 29. m,a._, and Ig^ = 10.6 m.a.^ as selected in Fig. 6. Unfortunately^ most of the "System A" and "System B" calibrations appeared as shoi-m in the bottom record of Fig. h, Hence^ it vas impossible to draw the curves needed to calculate I j ' Values for I^ ; I_^ , It, , and I-n might then be selected as follows: I IA2 = I = Ib2 I = I_^ = 70 m.a. I = 1-Q sufficient current to produce max min + €jj = min min max max. (13) where € represents the small additional deflection due to the noise inherent in the amplifier. Using these values in Equation (12), one finds: G ^A OC ^B Ia (iM where the additional subscripts have been dropped. Is this value of gain a sufficient representation of the actual gain to justify its use for calibration? A study of the curves G(Ij^) versus 1^, obtained from the few

PAGE 30

15 30 CO 25LU CO >05 201•21 UJ or 3 15 o < 10 C3 < O — SYSTEM A SYSTEM B ^, 60 50 20 40 60 CHART DEFLECTION UJ \>O) 40^ •z. LU cr q: 30 3 o h20 < q: CQ < 10 80 Fig. 6. — Receiver-recorder response curves used as an aid to calculate the gain of the amplifier in the Chile calihration systems ( 100 chart deflection \xnits is full scale deflection)

PAGE 31

16 calibrations having enough points to allow curves to be dravn^ showed that G(I^) scarcely ever varied from G(Ij^/2) by more than 20 per cent. Since slight antenna-to-receiver mismatch^ antenna pattern change with hour angle, etc.;, account for variations of this magnitude, this 20 per cent gain variation should not be too objectionable. Also, the amplifier gain variations from day to day were small (less than 10 per cent); hence, amplifier gain calibrations were made only weekly. We shall drop the indication of the functional dependence of G on I_^ remembering that G really means G(^A/2). In order to find the actual equivalent c\irrent (!„„) o^ '^'^^ et^ amplified noise power, one need merely to calculate: Ieq=^Ic + %^ (15) where I^^ was the current in "System C," G was the gain, and I^^ was the "equivalent current" necessary to produce the noise power present due to the amplifier. I^^ was easily read from calibrations involving "System C" and "System B," made at the same receiver gain, as shown in Fig. 7. In this case 1-^ is a little greater than 6o m.a. Unfortunately, the amplifier noise at l8, 22.2, and 27.6 Mc/s was equal to, or greater than, that due to the cosmic background noise. This excessive noise was nat'orally a disturbing source of error. RECEIVER CHARACTERISTICS Also needed for accurate calibrations were the response curves (deflection versus voltage) of each receiver and recorder combination. Pigs. 8 and 9 show typical response curves for the Collins, Hammarlund,

PAGE 32

17

PAGE 33

18 2 2.2 MC/S :— GALACTIC AVERAGE .1 .2 .3 .4 INPUT VOLTAGE IN pV o H 30o UJ 15.0 MC/S 'Y-" <— GALACTIC AVERAGE .2 .4 .6 .8 1.0 1.2 INPUT VOLTAGE IN pV Fig. 8. — Receiver-recorder response curves at 15.0 Mc/s and 22.2 Mc/s using Collins receivers and Texas recorders (lOO deflection units represent full scale deflection)

PAGE 34

19 lO CJ. 10.0 MC/S •-
PAGE 35

20 and Rodhe and Schvarz receivers. Note that the Rodhe and Schvarz receiver (operated at 5 Mc/s in Chile) is the only receiver with a linear response if power is plotted versus deflection. The rest of the receivers show a linear response if voltage is plotted versus deflection, or if power is plotted versus the square of the deflection. As might "be expected, the response of a given receiver changes with frequency. For the most part, the Collins receivers showed a linear response (deflection versus voltage) for all frequencies, if only deflections above the galactic level were considered (see Fig. 8).

PAGE 36

CHAPTER III ANALYTICAL PROCEDURES LATA RErUCTIOH The data were analyzed with two ends in mind: A. To continue analyses started in 1957 ^-nd used each year since; B. To start a new analysis that will provide better information about the intensity. Data Reduction as Perforraed in 1957-1961 and Continued in I962 "Old Analysis" In his dissertation, N. F. Six (lO) explained this analysis in detail. For that reason;, procedures used in the 1957-19^1 data analysis will merely be listed and defined. Records of all frequencies were reduced simultaneously, with constant reference being made to the logs kept by the observers. Since identification of the radiation rested originally with the observer, the logs and simultaneous rec rds assisted the person reducing the data in checking the validity of the observer's decisions. On the whole, the observers seemed to be quite accurate. Information taken from the records was: Listening Period The listening period was that period of time (U.T.) when the 21

PAGE 37

22 receiver was working, Jupiter was in the antenna beam, and interference v-as not excessive. The heam limits of the tracking antennas were defined to he five degrees above each horizon. The beam limits of the fixed antennas were : 1. 5B (Chile) + k hours from the meridian 2. 27c (Chile) 4k hours from the meridian 3. 22P (Chile) + 3 hours from the meridian h. 15c (Florida) + k hours from the meridian 5. 22P (Florida) + 3 hours from the meridian. Activity Period ( Storm ) A storra was that period of activity not interrupted by more than 8.3 minutes (five degrees of Jupiter rotation) of no activity. Note: The listening period was divided so that each activity period had a corresponding listening period. Date (Month, Day, and /ear ) A day was defined to begin at noon on day X and to end at noon on day X + 1. The date recorded was X + 1. Intensity If D was the pen deflection of a Jupiter pulse and G was the pen deflection of the galactic background, then: Intensity = K ^^l2) ' (16) G where the indicated average was performed for the three highest Jupiter

PAGE 38

23 peaks in any given storm. If there were less than three peaks during a storm.;, the average vas performed over the existing peaks. K is the normalization constant defined below. JTormalization Constant The normalization constant was that constant which normalizes the galactic hackgroiind to the galactic background of our first observations in 1957This constant is necessary since, as Jupiter changes its right ascension (two hours per year), the portion of the galaxy that is monitored also changes. Comments Comments were remarks made by the data reducer as to equipment condition, calibration, listening conditions, antenna direction, etc. As a continuation of work already performed, this analysis was necessary but it is easy to see that in order to improve our grasp on the intensity we must divide the storms into small time intervals. G. W. Brown (ll) of the Florida group first did this by dividing the storms of the 1961 apparition into ten-minute intervals. Extending his work, we divided the storms of the 1962 apparition into five-minute intervals, as shown in the next section. Data Reduction as Started in 1961 by Brown and Continued in I962 "Kew Analysis " Information taken from the records for the "new analysis" can be listed briefly as follows: Date The definition of the date was not changed from that used in the "old analysis."

PAGE 39

2k Listening Period The listening period vas that time spanned by the Universal Time at the beginning and end of effective listening as defined for the "old analysis." Again, it should be pointed out that the listening period is split up so that there is a listening period for each activity period. Activity Period The times of Jovian activity were divided into five-minute intervals. Any remaining time less than five minutes in duration was considered to be a separate interval. Intensity The pen deflection for the highest value of Jupiter activity, and the pen deflection for the average galactic background were read for each activity interval on the record. Jupiter Transit The time of Jupiter transit was looked up in the American Ephemerus and Nautical Almanac . Since the stations were not at the center of their respective time zones, corrections had to be made. The corrected Jupiter transit time (local standard time) was recorded in the log by the observer. Pulse Quality The antennas used by our stations have relatively broad beams, so noise from other sources comes into the receiver. Our identification of Jupiter activity depends mainly on the judgment of the observer, who with experience can become quite proficient at identifying Jovian pulses.

PAGE 40

25 The interference most difficult to distinguish from Jupiter radiation is that caused by a weak radio station. The observer logs each time . Jovian activity is recorded and estimates the certainty of his identification of the activity. With logs in hand and records taken simultaneously at the Chile and Florida stations, the person reducing the data decides whether the activity in any interval was "definite/' "possible," or "dubious." By the term "definite" is meant that listening conditions were good, the antenna was pointed toward Jupiter, activity was recorded at the other station if it was listening with good conditions, the observers labelled it "Jupiter," etc. If any one of the conditions named above was not met, the data reader might have chosen to label the activity "possible," depending upon his confidence in the observer. Pulses getting a "dubious" label were often short and isolated or buried in outside interference. Calibration The Florida and Chile calibrations were different, so the information needed to calculate the flux density of a pulse would also be different . 1. Florida. — The records were calibrated directly with noise diodes. Numbers read from the records were: a. each calibrator current value b. the corresponding deflection c. the total number of calibrator current values used in any one calibration.

PAGE 41

26 2. Chile. — The records were calibrated using calibration Systems "A," "B," and "C" as described in Chapter II. Information read from the records included: a. each "System A" current value (one per calibration) b. each "System B" current value (one per calibration) c. each "System C" current value d. the corresponding pen deflection for each "System C" current value e. the total number of "System C" current value per calibration. JUPITER PROGRAMS FOR THE I.B.M. 7 09 As has been indicated already^ the 1962 Jupiter data analysis took two forms. We shall call the program used in the I957-I961 analysis the "old program/' and the one used with the "five-minute-interval" analysis the "new program." The data described in the preceding section were punched into, two decks of cards^ one to be used with the "old program/' the other to be used with the "new program." The "Old Program " N. F. Six (10) gave a complete description of the "old program" in his dissertation. We will, therefore, be content with a list of the input and output. Input The following information was punched into cards to be used with

PAGE 42

27 the "old program" : 1. station 2. date (month/day /year) 3. beginning and end of listening period (U.T. ) h. beginning and end of activity period (U.T. ) 5. intensity of Jovian radiation (average of the three highest peaks) 6. normalization constant 7. Zurich provisional relative sunspot number 8. System I J longitude at 0^ U.T. 9. Julian day number. Output The following information was printed out by the computer: 1-8. all of the input except the Julian day number 9. beginning and end of the listening period (System III longitude , Jupiter ) 10. beginning and end of the activity period (System III longitude ) 11. daily activity index 12. daily activity index rate 13. monthly average activity index rate ik. yearly average activity index rate 15. monthly average sunspot number 16. probability histogram table. This program also stores the daily activity index rate and the daily sunspot number on a tape to be used with the solar flare program.

PAGE 43

2S Mention should also be made that, for the 1963 studies now under \^y, the stmspot number is not read into the "old program/' and that the intensity (average of the three highest peaks) is calculated by the computer--not by hand as before. The "New Program" Since the studies proposed for the "five -minute -interval analysis" were far afield from those used in the "old program/' a completely different program was written. However , parts of the "old program" were incorporated into the "new program" to save computer de-bugging time. Again^ we will divide our discussion into two parts: input and output. Where terms are identical with those in the "old program," this fact will be noted and further discussion will be eliminated. Input The total number of input parameters in the "new program" exceeded that of the "old program." Therefore, to save space on the cards, parameters that did notchange throughout the apparition were read only once from a single card at the beginning of the execution of the program. These parameters were: 1. average effective area of the antenna (taken from G. W. Brown's thesis (ll) or calculated using information from p. I6U, American Radio Relay League (A.R.R.L.) Antenna Book ) 2. transmission coefficient (b) of the transmission line (assumed to be about .75 )( Since the noise power to the receiver has actually been attenuated by this factor. Equation (7) should be written:

PAGE 44

29 £I5_.) (17) s = 3. equivalent current ( I^) of the noise inherent in ^ the calibrator amplifier k. attenuation coefficient of the attenuator used in "System A" 5. Julian day number of the first day of the apparition 6. station 7. frequency in Mc/s 8. equivalent current of the average galactic background noise for the entire apparition (used as a calibration reference when no calibration was performed on a given watch) 9 . year . Input parameters read on each of the succeeding cards included: 10. month and day 11. beginning and end of the listening period (U.T.) 12. beginning and end of the activity period (U.T.) (never longer than five minutes in duration) 13. deflection (g) due to the galactic background radiation ik. deflection (D) due to the highest Jupiter peak in the activity period 15. quality of the pulse (definite, possible, or dubious)

PAGE 45

30 16. time of transit (standard time at a given station) 17. time of sunrise (standard time) 18. time of sunset (standard time) 19. -xO^ U. T. System II longitude of the central -^11 meridian of Jupiter at 0^ U. T. of the date (X + l) in question 20. number (W) of calibration points 21. "System A" calibration current (A value of one was entered in the case of Florida data which had no calibrator amplifier, and zero was entered in case no calibration was performed. ) 22. "System B" calibration current (Zero was entered in the case of Florida data, or if no calibration was performed.) 23. each calibration current value ("System C" in the case of Chile) 2^4. the pen deflection corresponding to each calibration current value. Output What did the computer do with these input data? It used them to calculate the output information that is listed below. It also printed out most of the input data along with the calculated parameters to guide the person reading the output. In fact, all of the input parameters listed above were included as output except numbers 5^ 8^ 16, YJ , I8, 20, 21, and 22. Output information calculated by the computer under the

PAGE 46

31 control of the "new program" vas: 25. equivalent current of D 26. equivalent current of G 27. flux density (S) of Jovian peak 28. beginning and end of listening period (System III longitude ) 29. "beginning and end of activity period (System III longitude ) 30. activity index (the product of the flux density and the activity time in minutes) 31. beginning and end of listening period (local hoiur angle) 32. beginning and end of activity period (local hour angle ) z 33. daily activity index rate ( j activity index total listening time where i runs over all of the activity periods of any given day) 3i)— 37. daily activity index rate for activity in each of the System III longitude intervals: 0° to 90° (no source), 90° to 190° (source B), 190° to 290° (source A), 290° to 360° (source C) (The term "source" will be explained later . ) 38. monthly activity index rate for all activity in a given month 39-ii-2. monthly activity index rate for activity in the longitude intervals listed in numbers 3^-37

PAGE 47

32 43. yearly activity index rate for all activity in the entire apparition kh-k'J . yearly activity index rate for activity in the longitude intervals listed in nianbers 3^-37 kQ. daily listening time in minutes ij-9. monthly listening time in minutes 50. yearly listening time in minutes All of the above output was listed for each data card with daily, monthly, and yearly averages "being written only when they were needed; i.e., daily average at the end of a day, etc. After all of the output data had been listed, several histogram tables were generated, each of which contained all of the parameters in nunbers 51-55 listed as a function of one of the parameters in numbers 56-60. Every table was divided into three sub-tables, which were generated using different restrictions on the data. One sub-table was calculated using definite data only, another using definite and possible data, and the third using all the data. 51. probability of observing Jovian radiation 52. average fl\ix density (S^) (averaged over those flux densities actually detected) 53. average flux density (S^) (averaged over all intervals, i.e., letting the flux density equal zero when no activity was received) 54. total number of intervals during which activity was received 55. total number of intervals during which listening was recorded

PAGE 48

33 56. System III longitude on Jupiter in five-degree intervals 57. local hoirr angle of Jupiter in five-degree intervals 58. Universal Time in ten-minute intervals 59time with respect to sunset (one hoirr before to seven hoiirs after) in ten-minute intervals 60. time with respect to sunrise (seven hours "before to one hour after) in ten-minute intervals^) Some of the output needs discussion. Equation (I7) gives the flux density of a pulse which has an equivalent current I. Since the activity from the giant planet is superimposed upon the cosmic radio noise^ the deflection is not entirely due to Jovian radiation. Therefore, Equation (17) must he modified to give actual Jupiter flux densities. If Sq. is the flux density of the cosmic noise and Sq^ is the flux density of the cosmic noise plus Jupiter noise, then Sj, the flux density of the Jupiter noise, can he written: Sj=Sg^^-Sg(18) Since the equivalent current is proportional to the flux density, ^j = ^aw Iq(19) Equation (16) becomes: Sj = (Io^-Iq)(20) hA Another problem encountered at this point was that of determining

PAGE 49

3^ Iq_^j and Iq. These were found by interpolating between the values actually appearing in the calibration. This interpolation was performed with the characteristics of a given receiver in mind; i.e., if the deflection varied directly as the voltage, and thus as the square root of the flux density, this non-linearity was considered. On some of the records calibrations were missing or, for seme reason, were unreliable. When this occurred, the average equivalent current of the cosmic noise for the entire apparition, which was part of the input data, was used as a reference. Extrapol5,tion to the Jupiter peak was made by using either Equation (2l) or Equation (22). Equation (2l) was used for each frequency except 5 Mc/s, for which Equation (22) was used because of the characteristics of the Rodhe and Schwarz receiver. Sj = bA \G' (f-) <-^ el R , ^ Sj---^(— -1] (22) bA ^ G If any part of a five-degree Jovian longitude interval emitted radiation, the entire interval was credited with having emitted radiation. This practice could create problems. Consider this example. Jupiter's System III longitude is divided into five-degree intervals, each of which we monitor for about 8.3 minutes as it rotates by the "central meridian," the "central meridian" being that meridian bisecting the visible disk. We shall assume that from O558.O to O606.3 U.T. a given ( Aj-j--j.) interval was on the central meridian. Assume also that

PAGE 50

35 a storm was in progress and that we had divided oior five-minute activity intervals in this manner: 0555.O to O60O.O, O60O.O to 0605.O, and 0605.0 to 0610.0. A portion of each of these activity periods is included in the O558.O to O606.3 time interval associated with the five-degree rotation of Jupiter. But the computer credits each A-j-jj interval with a count each time any portion of an activity period falls within it. In this example the Ajjj interval in question received three counts for one storm. It is easy to see how the five-minute intervals could be chosen so that only two counts were credited to this A-j-jj interval. With the "old program" this multiple counting was eliminated "by requiring that two storms he separated 8.3 minutes in time. But in the "new program" one activity period (five-minute interval) often follows immediately after another ^ and thus it is possible for multiple counting to occur. Therefore, the "new program" allowed a five-degree longitude interval no more than one activity or listening count per day. Another provision would have to he made if a watch included more than one rotation of Jupiter. After all of the output had been calculated, an attempt was made to get more reliable flux density values by performing an iteration. As mentioned above, an intensity versus hour angle histogram was plotted. Since the effective area of an antenna and the ionosopheric attenuation of a pulse would change with hour angle, the assumption of a constant effective area in Equation (20) might lead to erroneous results. Therefore, each pulse was normalized by multiplying its flux density by the ratio of the average flux density (S^) received at the meridian to the average flux density (Sa) received at the hovo: angle of the pulse.

PAGE 51

36 This iteration was performed three times; once using histograms plotted vith definite data, once with definite and possible data, and once with all the data. After inspection of the intensity versus hour angle histograms, which showed very little struct-ure, these iterations were omitted in the interest of saving I.B.M. 709 machine time. Mention should he made, before discussing any results, that the calibrations were made quite some time after the Jupiter storms, and hence at different receiver gains, since the gains changed with time. Furthenaore, the calibrator's impedance was not always well matched to the receiver's impedance. These two facts were indicated by unusually great changes in the equivalent current of the cosmic noise level during a watch (nearly one order of magnitude in some cases), and by the change in the equivalent current of the cosmic noise level from one part of the apparition to another. Since Jupiter was well away from the galactic center, gross changes in the cosmic noise background would not be expected. Therefore, the data were also processed using the cosmic noise level as a reference throughout the entire apparition. Equation (2l) was then used for calculating flux densities. When this was done the data were labelled "no cal." When the calibration was used, the data were labelled "with cal." Later, when the intensity histograms are studied, they will be so labelled.

PAGE 52

37 CHAPTER IV PROBABILITY STUDIES Since the I962 data were of better quality than any recorded at the Florida and Chile ohservatories up to that time, a more meaningful analysis vas possible. The studies which had been carried out for the 1957-1961 data dealt mainly with the probability of observing radiation from Jupiter. The extension of this work is important; hence, the probability studies are presented as the first of our analyses. GROSS STATISTICS Table 2 shows the extent of the observing season, the number of hours of effective listening, and the number of hours of Jovian activity. It is striking that the number of hours of activity reaches a maximum at 18 Mc/s. The decline toward the higher frequencies can be explained by the fact that Jupiter radiates less at these frequencies. The low-frequency decline is probably due to increased ionospheric refraction and scattering of the low frequencies. Of interest, also, is a study of the correlation of noise storms between the two stations ("storm" will be defined in the section of this chapter involving data reduction). For each frequency. Table 3 shows the total number of storms, days of activity, simultaneous storms.

PAGE 53

38 TABLE 2 GROSS STATISTICS freq. in Mc/s station total effective observing season listening beginning end time hr. total activity time hr. *** 5

PAGE 54

39 on H O CQ o o cd -a H fn O -i -p 03 O -p CQ Hi •H O o -p CQ 0) iH •H O -P cci CQ !=! fn O 4J 02 P O -P fa •H o rH CQ >> 05 T3 P o O 02 03 o -p 02 CQ cd •H O CQ [in O -P 02 a o •H a CM rH C O •H x: o o 00 en CO +3 o . C -P c u cd OJ 0) (D r-S ?H O OO cj> -p cy T3 CT) rH ^ rH cd -P -P rH Cd C P^ S 3 I CM I I o o OJ OO CO u +3 Cd o o O CO o OJ OJ c:> OJ o VD CTn c-

PAGE 55

ko frequencies show incredibly good correlation. All in all, simultaneous listening was recorded for 6 months at 15 Mc/s, 9 months at 18 Mc/s, 7 months at 22.2 Mc/s, and 3 months at 27.6 Mc/s. Fig. 10 shows the average probability of observing Jovian radiation plotted versus frequency. Every point except that at 27.6 Mc/s was taken from the Chile data. (The Florida 27.6 Mc/s radiometer was more reliable than that at Chile.) One can easily see the tendency of the probability to peak around 10 Mc/s. However, this was not the frequency at which the greatest n-umber of hours of activity occurred. The difference was probably caused by Increased ionospheric absorption, and hence poorer listening conditions, at 10 Mc/s as compared to I8 Mc/s, where the total activity time reached its peak. Note should also be made that the I5, 18, and 22.2 Mc/s data at Florida yielded average probabilities lower than those at Chile for the same frequencies. This discrepancy was probably due to the fact that the Florida observing season extended closer to the time of conjvmction than did the Chile season. Some of the Florida observations were made when Jupiter radiation had to pass through the solar corona, and hence suffered attenuation. The only frequency for which the average probability at Florida exceeded that at Chile was 27.6 Mc/s. The 27.6 Mc/s antenna in Chile (a corner reflector) did not work well, and thus little activity was recorded on that channel. PROBABILITY OF OBSERVING JUFITER RADIATION AS A FUNCTION OF JOVIAN LONGITUDE Data from Individual Channels One of the first major characteristics discovered in the

PAGE 56

41 X POSSIBLE AND DEFINITE DATA O DEFINITE DATA ONLY 10 15 20 25 FREQUENCY, MEGACYCLES 30 Fig. 10. --Average proTaability of observing activity at the Chile station in 1962

PAGE 57

k2. decameter radiation from Jupiter vas that the probability of its occurrence depended on which Jovian longitude vas on the "central meridian." The histogram in Fig. 11 shows that four portions of the planet exhibit a relatively high probability of emission. These portions are called sources B-,, Bp, A, and C, as indicated in the figure. Sources B and Bp are often thought of as only one source^ "source B," but as can be seen in the following histograms, the bifurcation seems to be a permanent feature. However, unless a special study is being made of sources B]_ and B2, they will be referred to collectively as "source B." Let us take a closer look at the sources for each frequency. Fig. 12 shows the probability histograms for the I8, 22.2, and 27.6 Mc/s, possible and definite data, and the 15 Mc/s histogram for the definite data only, as recorded at the Florida station in I962. As has already been mentioned, the 15 Mc/s channel was in something less than perfect operating condition for part of the apparition; hence, we see less source struct\;ire than might be expected. The peak at 115 is source B, the peak at 2^0 is source A and the peak at 320 is source C. Note that the shapes of sources A and C change with frequency. For 18 Mc/s, sources B (l05° to 170°), A (220° to 280°), and C (290 to 3^5°) are all quite well defined. So\:irce B has a weaklydefined dip at 130°, thus forming sources B-]_ and BgAnother feature, o o of which we will keep track, is the general null between 20 and 55 • Sources B, A, and C are evident at 22.2 Mc/s, though source C is quite poorly defined. For the 27.6 Mc/s histogram, sources B and A are evident, but source C becomes quite diffuse.

PAGE 58

h3 < O cr o o -J 00 u. ~ CO I 1h O O o u cd o CM C! •iH ^ -p ON o ' ^ c: ,-1 c? •H

PAGE 59

i^4 FLORIDA 1962 T 90^ 180** 270" LONGITUDE, SYSTEM HI 360 Fig. 12. — Probability histograms for data taken at 15.0, 18.0 J 22.2, and 27.6 Mc/s at the Florida station in 1962

PAGE 60

i^5 Fig. 13 shows the probability histograms for 5, 10^ I5, 16, 18, 22.2^ and 27.6 Mc/s in Chile. The peaks are less evident, since the vertical scale had to be reduced to get all of the graphs in one figure. Note that at 15, 16, and 18 Mc/s source A also seems to be bifurcated. What general trends can we see in Figs. 12 and 13? Easily noticed are the facts that source A shifts to higher longitudes for lower frequencies (at least up to a point) and that the width of Source A increases with a decrease in frenquency. Which peak is so\;irce A on the 5 Mc/s histogram might be debatable. If we choose the general null at 210 on the 5 Mc/s histogram to correspond with the one mentioned earlier, the peak at 260 corresponds with source B and the peak at 3^0° might correspond with source A. Another interesting feature is that the ratio of the source C maxim\mi to the source A maximum increases as the frequency decreases. Merged lata Before making further studies of Figs. 12 and 13, let us look at the composite or merged histograms for those frequencies which were monitored by both stations. The merging process is described in detail by N. F. Six (lO), and to maintain continuity in our analysis, these merged histograms will be presented in much the same manner as he presented them. Therefore, we used the three-point smoothing technique in which the probability for each five-degree interval is averaged with those Just preceding and following it. These points are then joined to form the smoothed histogram. While the smoothing process may smear out fine -structure effects, it also rids the histograms

PAGE 61

h6 CHILE 1962 "27.6 MC/S T ZLu: DER 22.2 MC/S POS. AND DEF. LlI o q: DC 3 O O o >cn < m o a. .2h •T^ 18.0 MC/S POS. AND DEF. .4 16.0 MC/S POS. AND DEF \rvrxr-n^ .2.2 .2 15.0 MC/S PCS. AND DEF r 10.0 MC/S POS. AND DEF, DEF ,5.0 MC/S [L t.^.. n J 90' 180" 270' LONGITUDE, SYSTEM Itt 360 Fig. 13. — Probability histograms for data taken at 5.0, 10.0, 15.0, 16.0, 18.0, 22.2, and 27.6 Mc/s at the Chile station in I962

PAGE 62

h7 of much of their roughness "by eliminating bumps due to minor statistical variations. The merging also tends to minimize bad effects caused by local conditions, such as equipment failure, atmospheric disturbances or an unattentive observer, and makes possible a continuous run of observations not attainable at one station. Fig. ik shows histograms for the 10 Mc/s Chile data. The I962 histogram is shown at the top of the figure and the I96O-I962 merged histogram is shown at the bottom. The sources are labelled as K. F. Six (10) chose them in 1961. We will later see that another choice of labels is possible. Source A is less prominent in the I962 histogram than in the merged histogram. Also noticeable is a small but definite peak at a longitude of about lifO°. The fact that it is shifted to about l48° in the I962 histogram will be discussed later. Fig. 15 shows the 15 Mc/s histograms for the Florida-Chile merge for 1961 and 1962, the Florida-Chile merge for 1962 only, the Florida data for I962, and the Chile data for I962. In the bottom plot each source is clearly defined and we can see a very slight hint of a bifurcation of source B. The I962 merge shows both B-|_ and B^ and a definite splitting of source A--this doublet structure of source A is apparent also in the I6 and I8 Mc/s Chile data. The plot for the Chile data shows the source A doublet structure still more clearly, while the plot for the Florida data shows little identifiable structure in this region. Noticeable, however, is the marked difference in the position of the source B peak in the Florida and Chile histograms. It is hard to argue this shift out of existence by the fact that the impedance of one antenna was not correct, unless source B shifts during the year.

PAGE 63

U8 .30 >i< CO O.20 a. CHILE 1962 10.0 M C CHILE 1960-1962 ± ± 90<* 180® 270<* LONGITUDE, SYSTEM HI 360* Fig. lU. — Smoothed, merged probability histograms for data taken at 10.0 Mc/s

PAGE 64

k9 .20 FLORIDA 1962 15.0 M C 90® 180® 270® LONGITUDE, SYSTEM nX 360< Fig. 15. — Smoothed, merged probability histograms for data taken at 15.O Mc/s

PAGE 65

50 If source B does vary its position during the year, this apparent shift could "be explained by the fact that the impedance was incorrect for only the first part of the year. Since the data from the first part of the apparition vere therefore probably incomplete, a shift such as the one seen in these two plots might then be expected. This study will be pursued later in detail. We also see that the source B peak on the Florida histogram is skewed to lower longitudes. We will find that the source B peak is generally skewed to higher longitudes. Since 1957, l8 Mc/s observations have been made by at least one of our stations, making this frequency that for which we have the most data. Fig. l6 shows the 1957-1962 merge, the 1962 merge, and the individual histograms for 1962. Notice the difference in source structure between the Chile and Florida histograms. Source Bp is better defined on the Florida plot than on the Chile plot, but source B as a whole seems to be easier to identify on the Chile plot. Again, Chile shows the slight splitting of source A that was noticed for 15 Mc/s, and the Florida plot also shows source A to be skewed slightly on the highlongitude side. Also of interest is the fact that sources B and C in Florida seem to be shifted to slightly lower longitudes than the same sources in Chile, while source A seems to be located at exactly the same longitude. As with the I5 Mc/s histograms, the I962 merged histograms show the sources to be shifted to slightly higher longitudes than those on the 1957-1962 merged histograms. The I962 merged histogram shows the basic source structure, including bifurcations, very clearly, while the I957-1962 merged histogram shows the sources without the bifurcations. The reason for this smearing is probably

PAGE 66

51 90® 180® 270® LONGITUDE, SYSTEM m 360< Fig. l6.— Smoothed^ merged probability histograms for data taken at l8.0 Mc/s

PAGE 67

52 an apparent soiorce shift to higher longitudes or, effectively, to an Increase in the System III rotational period. This change in period, which apparently began in I96O, is also the reason that the 1962 merged sources are shifted to higher longitudes than those of the alltime merge. A special section will he devoted to a study of this apparent change of period. The 22.2 Mc/s smoothed histograms are shown in Fig. I7. Again, we see that in the top two histograms the structure of source B is different at each station and that corresponding sources seem to lie at ahout the same longitudes. Note that, as for the 15 Mc/s Florida source B peak, the 22.2 Mc/s Chile source B peak is skewed to lower longitudes. While the anomaly on source A was on the highlongitude side for 18 Mc/s, it appears on the lowlongitude side at 22.2 Mc/s. In the merged histograms the bifurcation of source B is less evident. In fact, if any splitting of B is to be mentioned in connection with the I962 merged histogram, the source should be referred to as a triplet, not a doublet. Again, the shift of source A to higher longitudes in the I962 merged histogram is noticed when it is compared with the I958-I962 merged histogram. While differences in source structure between the Florida and Chile histograms were also noted by W. F. Six (lO) in I961, these differences were not the same as we see in I962; hence, we will not try to draw conclusions from any but the most obvious details. Even though the 27.6 Mc/s data are scant, 1962 is the first year that we have had even this much activity at this frequency. Fig. 18 shows the smoothed histograms for 27.6 Mc/s. We notice again the

PAGE 68

53 .20 .10 LU O z LJ CH d z> o o o u. o ^ .30 r— < .20 OQ O ^ .10 FLORIDA 1962 2 2.2 MC .20,iOCHILEFLORIDA 1962 22.2 M C CHILEFLORIDA 1957-1962 2 2.2 MC 90° I80<* 270** LONGITUDE, SYSTEM HI 360* Fig. 17. — Smoothed^ merged probability histograms for data taken at 22.2 Mc/s

PAGE 69

'^h .05 T • FLORIDA 1962 27.6 MC LU O LU •^.os o o o u. o > 5.05 CD < m o q: a. .05 CHILE I 962 27.6 MC CHILEFLORIDA 1962 27.6 MC CHILE FLORIDA .1958-1962 27.6 MC 90® 180° 270® LONGITUDE, SYSTEM HI 360' Fig. l8. — Smoothed, merged probatility histograms for data taken at 27.6 Mc/s

PAGE 70

55 shift of source B on the Chile histogram with respect to source B on the Florida histogram. As with the 15 Mc/s histograms in Fig. l6, we can argue the possibility of this difference in peak longitude if source B shifts during the year, since the 27.6 Mc/s channel in Florida operated during the entire apparition, while the 27.6 Mc/s channel in Chile did not begin monitoring Jupiter until August 21, 1962. Since opposition was on August 31^ 19^2, nearly all of the bef ore-opposition data were missed in Chile. This argument still seems weak in the face of the large shift (approximately 20 degrees). When we compare the two individual histograms with the I962 merged histogram, we see that the apparent shift of the Chile data scarcely appears. This is probably due to the fact that the Florida listening time was so much longer than that at Chile, thus minimizing the importance of the Chile peak. Let us look at all of the I962 smoothed, merged histograms together, as shown in Fig. 19, to see if any characteristics of the 'sources change with frequency. One marked feature is the appearance of the anomaly on the high-longitude side of source A at I8 Mc/s, its increase at 16 Mc/s, and finally, its smearing of sources A and C at 15 Mc/s. Also noticeable is the decrease in the ratio of the source A maximum to the source C maximum with frequency. A real problem arises when an attempt is made to locate sources A, B, and C on the 5 and 10 Mc/s histograms. In his dissertation, N. F. Six (10) assumed the sources to be as shown in the top set of letters on the 10 Mc/s histogram. Credence is given to this choice by noticing that it maintains the source separations fairly well. However, a stretch of the imagination is required to keep the source spacings

PAGE 71

56 CHILE-FLORIDA 1962 90<> 180° 270° LONGITUDE, SYSTEM ni 360° Fig. 19. — Smoothed, merged histograms for all the frequencies monitored, using only I962 data

PAGE 72

57 constant on the 5 Mc/s histogram. Another prohlem that arises, if we assume the top set of lahels, is that the height of source B seems excessive on the 10 Mc/s histogram in light of the fact that its maximum is less than half that of source A on the 22.2, l8, l6, and 15 Mc/s histograms. For 10 Mc/s the bottom set of labels provides for a drift to lower longitudes with decreasing frequency at frequencies below. l6 Mc/s. Table k lists the positions of the soiorces as they occur at different frequencies. Only one choice for each source was made for frequencies of 15 Mc/s and above. When all of -^he data that have been taken at either station are merged and smoothed we get the histograms shown in Fig. 20. The graph has been labelled "Chile -Florida 1957-62" but not all of the frequencies have been monitored over this entire period. The 5 Mc/s histogram represents only the I962 Chile data, the 10 Mc/s histogram represents the 1960-1962 Chile data, the 15 Mc/s histogram represents the 196I-I962 Florida-Chile data, the 16 Mc/s histogram represents the 196O and I962 Chile data, the 18 Mc/s histogram represents the I957-I962 Florida data and the 196O-I962 Chile data, the 22.2 Mc/s histogram represents the 1958-1962 Florida data and the 196O-I962 Chile data and the 27.6 Mc/s histogram represents the 1958-I962 Florida data and the I962 Chile data. The 10 and 15 Mc/s high-longitude source is better defined in Fig. 20 than is its counterpart in Fig. 19, while, as should be expected, anomalies like the one on the highlongitude side of the I8 Mc/s source A in Fig. 19 are less noticeable in Fig. 20. The positions of the sources, however, are about the same in both figures. Fig. 21 shows a plot of the values in Table k. The dashed lines

PAGE 73

58 CHILEFLORIDA 1957-62 90** 180** 270** LONGITUDE, SYSTEM HI 360* Fig. 20.— Smoothed, merged histograms for all the frequencies monitored, using all the data recorded since observations "began in 1957

PAGE 74

59 30i w 25h -J O >o < o 201 UJ >o 15UJ o UJ u10! 5ttA ± 240® 560<^ 120° 240° 360° 120° LONGITUDE, SYSTEM 331 24 Fig. 21. — Variation of source position with frequency (The dashed lines provide an alternate selection of source positions. The point labelled 7 may not he source C. )

PAGE 75

60 TABLE k SOURCE POSITION AS A FUNCTION OF FREQUENCY Source A Source C Source B 27.6

PAGE 76

61 (One of the sources presumably either decreases in intensity or merges with another source as the frequency decreases "below 10 Mc/s. ) We see that even if we restrict ourselves to the assumption that all sources shift in the same direction, there are four pairs of points in Fig. 21 that can represent the 5 Mc/s sources. The lines running through these pairs of points are numbered one and three, three and five, two and four, and four and six. Notice that source A is present in all of the choices and that either source B or source C is the other. It is the opinion of the author that soirrce C is one of the sources at 5 Mc/s, in keeping with the prediction by Field (5) that the effects of source C should become more noticeable at low frequencies. Also more palatable to the author is the shift to lower longitudes with decreasing frequency, since this allows sources A and C to continue to merge as they do at 10 Mc/s. Therefore, the author's choice is represented by lines one and three. The 120° source is presumably a combination of sources A and C and the 270° source is source B. However, the shift of the null between sources C and B seems to dictate a shift to higher longitudes. Presumably, this dilemma could be resolved by analyzing the polarization at 5 and 10 Mc/s, for both Field (5) and Carr (2), maintain that radiation from source C comes from Jupiter's southern hemisphere, and hence that it is polarized in the opposite sense (lefthand elliptical) to that of sources A and B, which they believe to originate in the northern hemisphere and which are of right-hand polarization. Samuel GuUcis (13) and C. H. Barrow (ik) have found that the polarization of source C at I6 Mc/s is predominantly left-hand polarized, a phenomenon not so clear at higher frequencies. Presumably, then, it

PAGE 77

62 should be more evident at lower frequencies. The 5 Mc/s polarimeter to be built at the Florida station will be used to identify the polarization of these sources. Looking again at Figs. 19 and 20, one can see that the soiirce width seems to increase as the frequency decreases^ as discussed by W. F. Six (lO). Identification of the sources for the frequencies below l6 Mc/s J however, makes quantitative analysis difficult and superfluous . Drift Studies of Source A In 1961 Carr et. al . (15) established that the rotational period of the radio sources on Jupiter was 9 55 29.35^ a value that closely agreed with the period obtained by Douglas (16) — 9 55 29.37. Upon analysis of the I8 Mc/s Florida probability histograms for I96O-I962, as shown in Fig. 22, even a casual glance shows that soiirce A seems to be drifting toward higher longitudes with time. So unexpected was this drift that all preceding calculations were checked and re-checked. The 1963 data for I8 Mc/s were reduced, analyzed, and plotted. They, too, showed the apparent drift. In order to determine whether or not the drift was just a scattering of the points, the source A position was plotted for each year beginning in 1951^ as shown in Fig. 23. The bars indicate estimated errors as determined by the person analyzing the data. The two points close together at the top of the graph were obtained by dividing the 1963 data into two parts (one hundred days before opposition and one hundred days after opposition) to get partial apparition histograms.

PAGE 78

63 FLORIDA 8.0 MC/S r 90' ISO' 270 LONGITUDE, SYSTEM HI 360* Fig. 22.— ProToalDility histograms for the I960, I96I, and 1962 data taken at the Florida station

PAGE 79

1964.0 1932.0 1960.0 < 1958.0 LU >1956.0 I954.0h 1952.0 6h / / 9^55"'30^.52 ^ KH hO-J h-OH H-O—l HOH ^-9*^5 5"^2 9?35 V I 200** 220** 240** 2 6 0** 280** LONGITUDE OF SOURCE A. Ain:(1957.0) Fig. 23, — Yearly variation of the position of the 18 Mc/s source A

PAGE 80

. ^5 At the time of this writing the data had not heen analyzed for the entire 19^3 apparition, since observations were still "being made. Fig. 23 clearly indicates that source A did begin to drift at some time early in I960. Further, each apparition was divided into three parts, beginning with the i960 apparition, probability histograms were plotted for each part and each station, the histograms were "shuffled" in an effort to eliminate bias introduced by the person analyzing the data, and the source A position was plotted versus time as shown in Fig. 2U. The points labelled 1 and 2 in the figure were highly uncertain due to the poor statistics of their respective histograms and were not included in the leastsquares calculation, even though they probably would have contributed cancelling effects. From the least-squares line the apparent drift was calculated to be 10.45 degrees per year or, if expressed as that change of rotational period necessary to maintain the sources at their i960 longitude, plus I.I7 seconds. This makes the new System III rotational period 9 55^ 30.52. The calculations above were made using the relation: AT=^^T^ (23)' 877 where AT is the difference in seconds between rotational periods, <^A is the shift between longitude systems in degrees/year, and T is the mean rotational period in hours. If T is near the radio period (9.925 hours), then Equation (23) reduces to: At = 0.1123AA. i2k)

PAGE 81

66 0) o o o CO 02 +3 o c o •H ^ -p o o -p •H CO O P< -P •H I I C •H W • ^ OJ o VD • ON bO 1-1 •H |i< 0) O a •H CO yV3A

PAGE 82

6? The soiorces are probably linked with the magnetic field of the planet; therefore, their rotational period is the same as that of the planet itself. Since it seems ijnlikely that the rotational period of the planet has changed, then the sources have apparently changed their positions. How could this happen? Suppose that one of the sources is located at Jovian System III longitude (X) and that this longitude does not change. However, for radiation to reach the earth, it may not be necessary for longitude X to be that which lies on the "central medirian." If the interplanetary magnetic field is curved, as indeed it may be (see Chapter VI ), then Jovian radiation may be bent by it. For bending of electromagnetic waves we also need electron density gradients. For this study we assume that the necessary electron density gradients are present. If the rays are bent, then some System III longitude, which we will call Y, will be on the "central meridian." But if, due to the change in solar activity during the sunspot cycle, the interplanetary magnetic field shape is changed, thus changing the degree of bending of Jovian radiation, then not Y, but some other longitude Z, will be on the "central meridian." But the longitude region which is on the "central meridian" when we receive activity is that which we reckon to be the active region, so we see an apparent (Y-Z) drift in the position of the source located at X. Carr (17) has suggested that the observed shift of the emission pattern, toward higher longitudes at lower frequencies, results from a tendency of Jovian magnetic field lines to curve westward with an increase in radial distance, due to a drag on the rotating field. Since Jupiter's magnetospheric electron density presumably also changes with sunspot

PAGE 83

68 activity; it could also be responsible for the temporal change in source position. It should be pointed out that, if the interplanetary medium bends Jovian radio emission, the positions of the visible disk and the radio disk would not coincide. However, interferometer measurements show that they do coincide so the magnetospheric bending seems more likely. It is difficult at the moment to see how this could account for the sudden change of source position shown in Fig. 23, or the fact that once the source position did change, it continued to change in a linear manner as shown in Fig. 24. In addition to this, the fact that Jupiter's Great Bed Spot also began to drift in its longitude system (discussed later in this chapter) weakens our argument even more, for electromagnetic waves in the optical spectrum are not influenced appreciably by the interplanetary medium. Nevertheless, if an eleven-year period could be found, although a wild stretch of the imagination is needed to find one in Fig. 23, then an obvious conclusion would be that the drift is caused by the change in solar activity. Unfortunately, there are insufficient data to determine this now, but if the source A postion starts to move to lower longitudes in the next few years, a variation of the source longitude with the sunspot cycle may be established. Interesting also would be a study of the drift of the sources other than source A at l8 Mc/s, and of all the sources at frequencies other than l8 Mc/s. Robert Hayward is currently making just such a study at the University of Florida. Using the partial apparition histograms, a subjective study was made of the source structure to determine whether or not it changed in a regular manner during the year. As far as could be

PAGE 84

69 determined, the sources maintained their relative intensities quite well during the apparition, as should be expected, since Jupiter does not know where the earth happens to he. Further study of the drift of soiarce B will he discussed in the next section. In Fig. 25 we see how the change in rotational period moves the source positions. The rotational period was calculated to he -I o 9 55"^ 30.70 from a sketched curve, drawn before the I963 points were put on the graph in Figs. 23 and 2U. This period compares quite well with the period of 9 55 30-52, which was finally reached by using the least-squares curve. Therefore, the histogram for the "new" period (T-J--J-J = 9^ 55™ 30?70) in Fig. 25 should be representative of the actual "new" period (T-j-jj = 9^ 55"^ 30?52). Is source A better defined for h m s the "new" 9 55 30.70 perxod than for the "old" period? No, if anything, the opposite is true. In fact, none of the histograms improved in detail when the period was changed. The width of source A at 22.2 Mc/s in Chile actually became greater when the 9 5p 30. 70 period was used. Carrying this study further, let us examine the 1957-19^2 merge for 18 Mc/s using the 9 55^^ 30.52 period, assuming, as shown in Fig. 23, that the period abruptly changed to this value from 9 55°^ 29-35 on January 1, i960. The smoothed, merged (1957-1962), 18 Mc/s histograms for the "old" and the "new" periods are shown in Fig. 26, The peaks of all the sources in the histogram using the "new" period are higher and the sources are sharper than those in the histogram using the "old" period. Note that sources C and B]_ become more pronounced for the "new" period. This is the only frequency for which so noticeable an improvement was effected by the period change, which makes sense when one

PAGE 85

70 CHILE 1962 POSSIBLE AND DEFINITE lij o .5 .4 .3 .21t o: o o > il .5< g r .3.2rT T T T^=S*^ 55"^ 2 9? 35 18.0 MC J^ /I4 U A iE£ Tjj.: a*" SS^SO'.TO 18.0 MC /I / 1^ Jin ± 90° 180° 270° LONGITUDE, SYSTEM HZ 360< Fig. 25. — Probability histograms for the 18 Mc/s data taken at the Chile station in I962 using the "old" (Tjjj = 9^ 55"^ 29?35) and the "new" (Tjjj = 9^ 55^ 30?70) periods'

PAGE 86

71 FLORIDACHIL E l/i/57-3/1/63 n > I « I r T T= 9*^ 55"^29!35 IS.O MC/S 90° 180° 270° LONGITUDE, SYSTEM lEC 360* Fig. 26. --The 1957-1962 smoothed, merged l8 Mc/s histograms using the "old" {Tjjj = 9^ 55^ 29?35) period until i960 and the "new" (Tux = 9^ 55^ 30?70) period after that date

PAGE 87

72 remembers that this is the frequency that was used to calculate the source drift. Robert Hayward of the University of Florida is examining the source drift for different frequencies to determine whether or not the source drift varies with frequency. Drift Studies of Source B In 19^3; Warwick (4) suggested that source B drifts to lower longitudes during an apparition and he showed graphically^ using his i960 and 1961 data^ that during the tii.ie from l40 days before opposition to ^0 days after opposition the so\rrce B center shifted from approximately 132° to 100° ( Ajjj;, 1957)Since this was the first time such a drift had been proposed, a search of the I961 and 1962 data is warranted to see if this effect exists in our data. We do not include our i960 data, since splitting it into several parts would result in unreliable statistics, and since Warwick's conclusions were drawn mainly from the I961 data. Fig. 27 shows the variation of soiirce B in System III longitiide as calculated from the I961 and 1962 partial apparition histograms, using the "old" period (9^ 55"^ 29.35). No error bars are shown, even though the best values should be regarded as accurate to no better than + 3°' For both years source B-, apparently drifts to lower longitudes until opposition, and then drifts back to higher longitudes. Source B2 seems to maintain its longitude quite well throughout the apparition. The center of source B, as a whole, seems to behave somewhat like source B-, . Mention should be made that the position of source B was chosen as midway between the minima at both ends of the

PAGE 88

73 I60< 140* <5> " I20<> J hi g I00< o 80< h 60* OXB: •X J CENTER OF B B. '*-s^' X— MERGED FLA. -CHILE, 1961 _ — -0 MERGED FLA.-CHILE, r962 \ 160 -89 80 160 MEAN cPOCH VAR.T. OPPOSITION IN DAYS Fig. 27. — 18 Mc/s source B drift studies using I961 and 1962 data

PAGE 89

7^ soixrce. This is the most expedient, but perhaps not the most reliable, method for choosing the center of this weakly-defined, skewed source. The centers of the sources B-, and B were chosen by sketching gaussian curves which approximated the sources, and choosing the centers of the curves. In order to get more extensive data, the "before," "around," and "after" opposition partial histograms were merged for 196I-I962. This was done for both the "old" period (9^ 55"^ 29f35) and the "new" , h m s V period (9 55 30. 70}. Studies similar to those in Fig. 27 were made and these are shown in Fig. 28. The probable errors of these points are less than those in Fig. 26. We see the "eyebrow" motion of source B-j^ for either period, the irregular motion of source B , and again, the "eyebrow" motion for the entire source. At any rate, our data fail to show the extensive yearly drift of source B to lower longitudes that was seen by Warwick (4). Incidentally, Warwick's data do show the apparent long-time source shift to higher longitudes that was mentioned in the preceding section. Possible Correlation of Radio Source Drift with Red Spot Drift The University of Florida group compared the rotational period of the radio sources with the rotational periods exhibited by various Jovian optical prominences as early as 1957 (18). At the suggestion of A. G. Smith we again conducted such a study, this time to compare possible changes of the Red Spot rotational period with the observed change in the rotational period of the radio sources. The System II longitude of the Red Spot is shown in Fig. 29 '

PAGE 90

75 MERGED FLORIDA-CHILE, 1961-1962 1 J -T" 160 oL I40< H UJ > CO o H o I20< 100®'1 30< 60' CENTER OF B .--X "x^-O— "MEV;" PERIOD O'' 5 5"™ 30^70) OLD" PERIOD (9^55'" 29^35 ) -J 1 S « ' ' -160 -80 30 160 MEAN EPOCH W.R.T. OPPOSITION IN DAYS Fig. 28. — 18 Mc/s source B drift studies using the merged 196I-I962 data ^

PAGE 91

76 I965h r I jMOTION OF RED SPOT IN SYSTEM 31 Tjix= 9" 55"* 40!632 Tl 19 se2 1955 i950h 1945AT=0.898S -= r\0 200' 250 300 350' 40^ LONGITUDE, SYSTEM IE Fig. 29. — Time variation of the System II longitude of Jupiter ' s Great Red Spot

PAGE 92

77 as a function of time. The solid line represents a drift of 7-75°/ year^ or if expressed as an effective period change, O.898 seconds. In about 1958 the Red Spot started to drift lh.8°/yea.T in System II. In order to examine this change of drift more closely, ve plotted its longitude for a period of T-^^ = 9^ 55"^ ^1^530 (System "Red Spot"), which was arrived at by adding O.898 seconds to the System II period, T^.= 9 5p 40.632. The graph shown in Fig. 30 resulted. (The plot in Fig. 23 is also shown in Fig. 30) As can be seen in the graph, the Red Spot started to shift to higher longitudes about one year earlier than did the radio source. The Red Spot started to shift ^^/jear in 1959 s^nd the radio source started to shift 10.^5 /year in 19^0, in their respective longitude systems. The fact that the slopes of the two drift lines are so nearly equal and that they match so closely in time suggests that the drift motions of the Red Spot and the radio source, in their respective longitude systems, must be related in some way. In an effort to find the best representative period of the Red Spot, Peek (19) tabulated its longitude in System II for many years. Using the tabulated data, he arrived at a best period of T = 9 55 37 •58Fig. 31 shows the longitude of the Red Spot versus time using Peek's period, assuming that the longitude in this special system was -264 in 1894. Equation (25) gives the longitude in the special system as a function of Aj-r (System II longitude) and t. A = Ajj 26h°3 + 28?62 t, (25) where t = (time in years 1894) x 365.25/398.88. (26) It is interesting to note the difference between the Red Spot period

PAGE 93

78 DRIFT STUDIES RED SPOT RADIO SOURCE A 1964 T — » — r 9VYEAR d19601956^1 I0.457YEAR I C— H 1 1 2I0*» 230** 220*» 240** 260** SPECIAL LONGITUDE LONGITUDE, SYSTEM SYSTEM WITH lEE (1957) Tj,3= 9h55'"4l?530 T^j^ 9^55'"29!35 Fig. 30. — Time variatior of the longitude of Jupiter's Great Red Spot for a period Trs which maintains the Red Spot longitude constant from 19^1-5-1958 compared with Fig. 23

PAGE 94

19 L I9S0 I950h 1940 1930^!92ol^ I9l0h < ^ i900^ , . l39Ch H < ^ i880!i870h 18601350^ 1840 ;LONG-TERM DRIFT OF GREAT RED SPOT -430** -240** 240' LONGITUDE 480** 720** J Fig. 31— Time variation of the longitude of Jupiter's Great Red Spot since I835, using a special period T = 9^ 55°^ 37?58. Based largely on Peek (I6)

PAGE 95

80 , h m s h m s (9 55 37.58) used by Peek (19) and the period (9 55 i+1.530) arrived at by using the data since 19^5This suggests that if the Red Spot drift is related to the radio source drift, determination of a period which maintains the radio source longitudes constant may he impossible. PROBABILITY VERSUS HOUR ANGLE The 1962 analysis initiated several new studies, one of which was an examination of the variation of Jupiter radiation with hour angle. The computer tabulated several parameters with each five degrees of hour angle from -90 to 90 (east to west), but the only parameter that we will discuss here will be the probability of Jovian emission. These histograms also can be loosely construed as being probability versus elongation studies, since most of the activity received at hour angles between -90 and 0° occurred prior to opposition and most of the activity received at hour angles between 0° and 90° occurred after opposition, although it is easy to see that this was not, by any means, always the case. When any interval was listened to less than 30 times no points were plotted; hence, the sharp cutoff on each side of the histograms. Fig. 32 shows the probability versus hour angle histograms for the Florida data. We can see the effect of the impedance mismatch early in the apparition for the I5 Mc/s data by the fact that the histogram is skewed to the west. It is very interesting to note that both the 18 Mc/s and the 22.2 Mc/s histograms reach maxima at approximately 30 east. Perhaps this is caused by the fact that the ionosphere is

PAGE 96

81 FLORI DA 1962 DEFINITE DATA T -1 ] J 1 r^ — ^,-80 -60 -40 -20 20 40 60 80 HOUR ANGLE IN DEGREES Fig. 32. — Pro'ba'bility versus hour angle histograms for the data taken at the Florida station in I962

PAGE 97

82 much clearer after midnight than it is "before midnight, thus making observations "before opposition more favora"ble. The 27.6 Mc/s histogram also shows a tendency to peak in the east. Fig. 33 she ,3 the pro"bability versus hour angle histograms for the Chile data. The l6 and 22.2 Mc/s histograms are missing, since the "new" analysis was not performed for the polarimeter data. The 5 Mc/s histogram shows the "beam width of the fixed broadside antenna. The rest of the histograms again show that the probability peaks in the east. Note, also, the dip at 5 west on the 5 a-^d 10 Mc/s histograms. These are the only histograms on which this occurs. It appears that some special phenomenon is causing an attenuation of the low frequencies incident at 5 west. We shall examine future data for this effect. The 15 and l8 Mc/s histograms show slight minor peaks between ko and 50 west. The 27.6 Mc/s histogram is not included, since the data at that frequency were very sketchy. Another possible explanation of the skewed nature of the histograms is that the to 90° intervals recorded many more listening counts than did the -90 to intervals, thus including more bad listening time, i.e., time close to sunset. PROBABILITY VERSUS imiVERSAL TIME Since the dependence of the probability on Universal Time has not been studied, it seems worthwhile that we should include it here. Figs. 3^ and 35 show Florida and Chile histograms for just such studies. The statistics become poor at each extreme on the curves, since the listening time decreased for times close to sunrise (1200 U.T.) and sunset (230OU.T.).

PAGE 98

83 CHILE 1962 DEFINITE DATA T s r 18.0 MC/S -80 -SO -40 -20 20 40 60 80 HOUR ANGLE IN DEGREES Fig33— ProlsaTDility versus hour angle histograms for the data taken at the Chile station in I962

PAGE 99

8k FLORIDA 1962 -J p27.6 MC/S i > r POS. AND DER ^ .1 J^—^ Jl 2 2.2 MC/S POS. AND DEF. VVV^iHi^ A^sJkT I 8.0 MC/S DER .2!< .1 CD ' O cc Q.2 .1 ^^ J lyxnr^jA 15.0 MC/S DER qU L»JlDxi^ ' ' ' 2300 0100 OoOO 0500 0700 0900 1100 UNIVERSAL TIME Fig. 3^— Probability versus Universal Time histograms for the data taken at the Florida station in 1962

PAGE 100

85 CHILE 1962 T 1 1 r 27.6 MC/S .04 .02 .3 o 2 .1 UJ O O p o •' >. _J CQ .2 < CQ I cr Qli /\ ri IE 18.0 MC/S 15.0 MC/S _ 10.0 MC/S i£iL n r .4h .3 .2 .1 5.0 MC/S n POS. AND DEF. DEF. POS. AND DEF DER _LIL r^^uh I 1 0100 0300 0500 0700 0800 UNIVERSAL TIME ilOO Fig' 35' — Probability versus Universal Time histograms for the data taken at the Chile station in 1962

PAGE 101

86 Notice that in Fig. 3U both the 15 and 18 Mc/s histograms show small peaks at 0220, 0620, and O9IO and a dip between O7OO and O8OO. The 22.2 Mc/s histogram shows a very slight dip at OT^J-O, but the 27.6 Mc/s histogram shows different structiore altogether, with peaks around 0330 and 0730 and nulls around O53O and 092O. When we look at the Chile histograms in Fig. 35 we see that about the only tendency duplicated for any frequency at both stations is the slight dip at 0800 on the Mc/s histograms. The 5 and 10 Mc/s histograms in Fig. 35 show definite structure but, unfortunately, they have no counterpart in Florida. Frcm Figs. 3^ and 35 it seems safe to conclude that if any diurnal variation exists in the probability of receiving Jovian radio emission, it is peculiar to each station, not world-wide. PROBABILITY VERSUS TIME FROM SUNSET The part of the ionosphere presented toward the sun is quite different from that on the night-time side of the earth. Since the sun does affect the electron density profile, and since the electron density governs the transmission characteristics of the ionosphere, studies of Jovian emission for different solar positions might be instructive. Figs. 36 and 37 show the respective probability versus time-from-sunset histograms for the Florida and the Chile stations. All of the plots (except possibly the one for 15 Mc/s) in Fig. 36 show a tendency for the probability to increase as time from sunset increases. In the 15 Mc/s plot the probability is nearly constant from fo\ir to seven hovirs past sunset. This general increase is understood when one recognizes that the transmission characteristics of the ionosphere improve with time after simset. Scarcely any

PAGE 102

87 FLORIDA 1962 — r— J 1 s 1 r 27.6 MC/S POS. AND DEF. 22.2 MC/S POS. AND DEF 18.0 MC/S POS. AND DER I 2 3 4 5 6 TIME FROM SUNSET IN HOURS Fig. 36. — ProlDaMlity versus time-from-sunset histograms for data taken at the Florida station in I962

PAGE 103

88 n 1 27.6 MC/S CHILE 1962 T 1 r1 1 — ~~T r POS. AND DEF. 18.0 MC/S POS. AND DER 15.0 MC/S POS. AND DEF. ^^-^ 10.0 MC/S DEF. .2 .1 5.0 MC/S DEF. JL JL J. J. 12 3 4 5 6 TIME FROM SUNSET IN HOURS Fig. 37. — Probability versus time-frcm-sunset histograms for the data taken at the Chile station in I962

PAGE 104

89 listening time close to sunset vas regarded as effective unless a Jovian storm was in progress, since listening conditions were unusually bad at this time. This practice probably is the cause of the 18 Mc/s peak that occurs near sunset. In Fig. 37 we again see the tendency of the probability to increase with time from sunset. The 15 and 18 Mc/s plots both show a minimum at 3 hours 50 minutes past s\inset and a slight maximim just preceding it. The Florida 18 Mc/s histogram shows the same structiire, but it is displaced one hour further from sunset. The 22.2 Mc/s histogram also shows the minimum, but not the maximum. The only real characteristic of these histograms seems to be the increase of probability with time from sunset. PROBABILITY VERSUS THE TIME FROM SUKRISE The studies that were begun in the last section were continued to examine the variation of the probability with time from sunrise. Figs. 38 and 39 show the respective probability versus time-from-sunrise histograms for the Florida and Chile data. In Fig. 38 we see a dip between 2 and 3 hours before sunrise on all frequencies. It is interesting to note that this dip shows up at the same time relative to sunrise for the 10, 15 and I8 Mc/s histograms for the Chile data in Fig. 39' The 5 Mc/s histogram shows a definite peak between 3 hours and 5 hours before sunrise, but the scarcity of data at this frequency makes one shy about attributing any significance to it, other than to say that apparently the ionosphere was most transparent to radiation of that frequency at that time. This

PAGE 105

90 FLORIDA 1962 ! 1 r POS. AND DEF. 27.6 MC/S T 22.2 MC/S POS. AND DEF. 18.0 MC/S POS. AND DEF. .1 15.0 MC/S DEF .6 -5 -4 -3 -2 -I TIME FROM SUNRISE IN HOURS Fig. 38.--i^oba'bility versus time -fromsunrise histograms for the data taken at the Florida station in 1962

PAGE 106

91 •6 -5 -4 -3 -2 -I TIME FROM SUNRISE IN HOURS Fig39— Erotability versus time-from-sunrise histograms for the data taken at the Chile station in I962

PAGE 107

92 peak can be seen in less prominence on the 10, 15, and l8 Mc/s histograms. It also appears in ahout the same place on the Florida 15, l8, and 22.2 Mc/s histograms. If the histograms in Figs. 38 and 39 are thought of as continuations of those in Figs. 36 and 37, the continued increase of prohahility with time from sunset can be seen. This continued increase can be explained by the fact that the ionosphere continues to "clear up" until shortly before sunrise. Fig. ho displays the smoothed 18 Mc/s histograms from Figs. 38 and 39. If the small peak at 3 hoirrs before sunrise is neglected, we find that the two main Chile peaks occur respectively at k hours before sunrise and at 50 minutes before sunrise, while the corresponding Florida peaks occur at k hours 20 minutes before siinrise and at 50 minutes before sunrise. The minimum separating them occurs at 2 hours 35 minutes before sunrise for data from either station. None of these prominences is separated by the U6-minute time separation of the Florida and Chile stations. Note should be made that Smith (20) noticed this "sunrise effect" in the 1957-1958 data. It will be interesting to see if this structure is present in the I963 data. PROBABILITY VERSUS JUPITER ELONGATION In his dissertation, N. F. Six (lO) discussed the variation of Jovian average activity index rate with Jupiter elongation. He noticed that for several frequencies in I96I, the average activity index rate decreased around opposition (an elongation of I80 ). In an effort to observe this same effect in the probability of occurrence, the average

PAGE 108

93 18.0 MC/S POSSIBLE AND DEFINITE DATA -6 -5 -4 -3 -2 -I TIME FROM SUNRISE IN HOURS Fig. kO. — Smoothed protalDility versus time-from-sunrise histograms for the data taken at l8.0 Mc/s in I962

PAGE 109

9h monthly probability was calculated for each of the I962 channels. Fig. 4l compares the time variation of Jovian probability for the Florida data with the time variation of Jovian elongation. The slight dip in September that is seen on all but the I5 Mc/s curve is located close to opposition (maximum, elongation), which occurred August 31, 1962. One might argue that, since the I5 Mc/s data were unreliable during the first part of the apparition, the null was washed out for this frequency. However, the Chile data, shown in Fig. k2, fail to show this dip. At best the 22.2, 18, 15, and 10 Mc/s data show a change of slope in August, But the listening conditions after opposition in Chile were becoming worse; the Chilean summer was coming on, and Jupiter was moving into western hour angles. For these reasons, the null may have been washed out at the Chile station, while itwas observed at the Florida station, which had better post-opposition listening conditions, since this period came at the onset of the Florida winter. Note, however, that the effect of poorer post-opposition listening (western hour angles) predominates on all channels. This effect is seen in the hour angle plots (see Figs. 32 and 33 )• At this point it is hard to say whether the null is real or sp-urious. This observer believes that it may be real and that it may be due to attenuation of the incoming radiation by the turbulent wake in the earth's magneto sphere, caused by the solar wind. Dr. A. G. Smith has suggested that observation of radio stars through the wake would be a test of this hypothesis.

PAGE 110

95 FLORIDA f — i — I — I I I r^ MAR. MAY JULY SEPT. NOV. 1962 TIME JAN. r.lAR. 1963 Fia; 1+1 —Variation of the average monthly probability with time and Jupiter elongation, for data taken at the Florida station in 1962

PAGE 111

96 .Oo.1 := r .Ir CHILE 1962-3 —T j — i — r 27.6 MC/S T T T 22.2 MC/S 18.0 MC/S 15.0 MC/S 10.0 MC/S J L 1 MAR. MAY 5.0 MC/S I I SEPT. NOV. TIME 1 1 _ J JAN. MAR. 1963 Fig. k2. — Time variation of the average monthly probahility for data taken at the Chile station in I962

PAGE 112

CHAPTER V IKTMSITY STUDIES UTTEHSITY DISTRIBUTION The analysis previously carried out by K. F. Six (lO) included studies in which the average of the three highest peaks per storm was used as the measure of Jovian activity. Since a storm often lasted one or two hours^ the three highest peaks in a storm were^ at "best, an upper hound of the actual average intensity. Another disadvantage of the old analysis was that the use of Equation (16) amounted to assuming that the receiver response varied linearly with power (compare with Equation (22)) instead of linearly with voltage (compare with Equation (2l)). An improvement in the flux density analysis is, therefore, to "be desired. In his master's thesis G. W. Brown (ll) calculated the flux density for each ten-minute interval of Jovian activity in much the same manner as we have described in Chapter III. He found that and that Sn = KS (27) N N S^-^ X (S-,)j=^Z S., (28) ^ N j=.l ^ W j=l "^ where S]_ is the average flux density for one interval, § is the peak 97

PAGE 113

98 flux density for that interval, K is a constant, Sg is the average flux density for a season, and where j runs over all the activity intervals in the season. However, his calculation of K was made using measurements taken from one-minute intervals. Since he measured S for each ten-minute interval, T. D. Carr (l2) calculated K using measurements made from ten -minute intervals . The peak flux density (s), on the average, is less for the smaller intervals, hut the value of His independent of the interval size; therefore, K must change if S does. For the one-minute intervals, K was found to be O.lOij-, and for the tenminute intervals, it was O.O39. We have seen that K relates the average flux density (S-,) to the peak flux density (S) for a given interval (see Equation (26)). Its value then depends on the distribution and duration of the sporadic hursts, both of which may change with time or frequency. Brown (11) examined the possible variation of K with frequency, and found that it changed very little in the range from 10 to 27.6 Mc/s. The possible time (year-to-year) variation of K has not yet been determined except by a more or less casual impression of the observers that they thought the noise bursts sometimes to be different in character in 1962 than they had been before. Long "rollers" (pulses many seconds in length), which first appeared in I962, and a great abundance of very short, popping pulses led the observers to this conclusion. However, determination of K involves determination of the fraction of time during an interval that the square of the pen deflection exceeds some va^ne X^. This fraction can be regarded as a

PAGE 114

99 2 distribution function F(X ). It can only be found using the fast (5 mm/sec) Brush records which are needed to resolve the noise bursts. Since our study does not include reduction of Brush records, the time dependence of K will not be pursued here, but it should be recognized as a problem well worth studying. 0^xc intensity studies will deal only with the peak flux density in each five -minute interval. The program computed flux density averages in two ways: by averaging over activity intervals only, and by averaging over all intervals, reckoning the flux density to be zero when no activity was recorded. These averages will be respectively labelled S^ and Sj^; both will be studied thoroughly. Carr (2l) suggested that the average taken over all intervals, (S-r ) might not be correct, since the assumption that the flux density was zero for all intervals of no activity was not necessarily a valid one. The intensity may have been below the sensitivity of our radiometers. His point, therefore, was well taken, so an intensity distribution analysis, like the one performed by Brown (ll), vas made for three frequencies. Fig. 43 shows the variation .with S of the probability of observing intensities greater than S (designated by F(S)) for l8 Mc/s, Florida. The area under the curve represents the actual average of the flux density over all intervals, as can be seen from the expression: S3 = /"f(s) ds, (29)

PAGE 115

100

PAGE 116

101 which was derived by Brown (ll). Numerical integration of Equation (29) by using Fig. k2 leads to an average fl-ux density of 3.93 x -22 -2 -1 10 w m cps , This value differs by only 2.72 per cent from the -22 -2 -1 value 4.04 X 10 w m cps" which was calculated by assuming that the flux density was zero when no activity was recorded. In fact, the latter is greater than the integrated value, which is indeed surprising in light of the fa> . that in the average computed from the distribution function we effectively attributed non-zero values to the nonactivity flux densities. This discrepancy was probably caused by the slight inaccuracy of the numerical integration, which was performed by hand. The Florida 22.2 and 27.6 Mc/s data were also analyzed in the same manner. The two averages at 22.2 Mc/s differed by 1.39 pe^ cent, and the averages at 27.6 Mc/s by O.63 per cent. In light of the relatively small error incurred by using the assumption that the flux density is zero for non-activity periods, the averages calculated by the computer will be used. GROSS STATISTICS The frequency dependence of the yearly average probability of emission was shown in Fig. 10. An obvious extension of that study is to determine the frequency dependence of the yearly average flux density. As has been pointed out, however, this average can be performed in two ways: by averaging over all intervals, reckoning those of no activity to have zero flux density, or by averaging over only those intervals for which activity was received. It will be recalled that the two kinds of averages were designated S^^ and S^, respectively.

PAGE 117

102 Fig. hk shows the average flux density (S^) plotted versus frequency. The spurious point for 15 Mc/s, Florida, arises because only the most intense pulses were recorded on this unreliable channel. It is Interesting to note that, like the probability, the flux density peaks around 10 Mc/s and tails off at higher and lower frequencies. The decrease on the low frequency side may be caused by ionospheric attenuation, but the decrease on the high frequency side is real. The other average flux density (S^ ) is plotted versus frequency in Fig. ^5. The same basic shape as that in Fig. kk is shown. This should be expected, since S is the product of S and the averL A age probability (see Equation 30), both of which show this same frequency dependence. 21 S N 21 S W N ill _ i i i activity (30) Listening Activity Listening S S^ Probability We have no reason for believing that the curve in Fig. U5 should be unchanged from year to year. In fact, if it behaves like curves of other parameters, a time variation would be expected. Nonetheless, if it does not change, we notice that the value of K = O.lOt, which was used by Bro>/n (ll), makes Fig. h^ compare very closely with the average intensity versus frequency curve plotted by Carr et. al . (12) for the 1961 data. But, as mentioned previously, the observers felt that the pulse structure had changed from 1961 to I962. Whether or not this change was sufficient to change K effectively is debatable,

PAGE 118

103 ^I2| ^ FLORIDA O CHILE NO GAL. 10 15 20 FREQUENCY IN M C/S 25 Fig. kh. — Frequency variation of the average flux density, S A

PAGE 119

104 O FLORIDA X CHILE NO GAL. 10 15 FREQUENCY IN 20 MC/S Fig. 45. — Frequency variation of the average flux density;, Sj^

PAGE 120

105 It is the opinion of Dr. Carr that this change was sufficient to make a change in K, while Dr. A. G. Smith thinks not. If we assume that K = 0.1 and that the average flux density decreases linearly -15 2 to zero at 1 Mc/s, the mean integrated flux density is 2 x 10 w m and;, if Jupiter were an isotropic radiator, the total power emitted would he ahout 1.4 x 10 watts, the sarae order of magnitude arrived at by Carr (12). Jupiter is, in fact, not an isotropic radiator, hut "beams its radiation in a given direction often referred to as the cone of radiation so our estimate should be multiplied by the ratio of the solid angle formed by the radiation cone to kTC . This estimate in fact depends on the value of K, the spectrum below 5 Mc/s, and the directivity of Jovian emission. DETEEMIKATION OF CALIBEATION SYSTEM RELIABILITY Since the equivalent current of the cosmic radio noise was one of the output parameters in the "new" program, it was possible to check the reliability of the calibration. If the values of the equivalent current of the cosmic noise level changed much during one watch, the calibration may not have been giving correct flux densities, since one would expect fairly constant cosmic radiation over the length of a watch, especially when tracking antennas were used. The program output was checked and the equivalent current of the cosmic noise was found. to vary greatly (up to a factor of five) over the period of the watch. This variation, nearly always a decrease in time, was caused by a change in receiver gain, which started high and c ded low. The calibrations were nearly always made

PAGE 121

io6 at the end of the watch, sO;, if a storm had been recorded for higher receiver gain earlier in the watch, the apparent cosmic and Jupiter flux densities far exceed those measured close to the time of the calibration. This effect on the average flux density can be seen in Fig. k6, where the average flux density (S. ) is plotted versus Jovian System III longitude. Bfote that the Chile, "with cal." smoothed histogram seems to be a magnification of the "no cal." histogram. This discrepancy was probably caused not only by the fact that the calibrations were made near the end of the watch, but also by the temperamental Chile calib-"T.tor amplifier, which often changed its characteristics, making equivalent current calculations difficult. Jupiter's right ascension was about seven hours from the galactic center and was moving further away from it during the year; therefore, the galactic background did not change much (less than 10 per cent) due to the two-hour displacement of Jupiter during the year. At our frequencies the galactic background noise varies slowly for right ascensions removed more than four hoiirs from the galactic center. The galactic background noise, therefore, provided a convenient reference with which to compare the Jovian pulses, and it was used in the analysis that follows. The inherent advantages in using the galactic background noise as a calibration reference are: first, the change in receiver gain makes no difference in the calculated intensity, since the gain changes the same amount for the galactic background and for Jovian -nulses; and second, changes in ionospheric attenuation make little difference, since both signals are attenuated nearly the same amount. There is one pitfall in using the galactic

PAGE 122

107 < 4 CHILE NO CAL. VSO ISO 270 360 LONGITUDE, SYSTEM Hi (1957) Fig. hG — Coraparison of the S versus System III histograms for the l3.0 Mc/s data using^calibrations ("with cal") with those using the cosmic noise level as a reference ("no cal'^ )

PAGE 123

108 background noise as a calibration reference at frequencies just above the critical frequency of the ionosphere. Our antennas have wide beams so the galactic background noise is due to rays coming from a wide portion of th^ sky. But f = f ^ sec Qq, (31) e being the largest zenith angle at which radiation of frequency f will penetrate the ionosphere when its critical frequency is f^. Equation (31) shows that decreases as f^ increases, f remaining constant, so the cone of galactic radiation reaching the antenna changes as f^ changes from night to night, whereas no pronounced effect would occur for the point soujrce, Jupiter. Unfortunately, our calibrations were also less reliable at lower frequencies. However, the ionosphere does not appreciably change the galactic background radiation at frequencies well above f^. imMSITI VERSUS SYSTE-I III LONGITUDE The studies that had previously been made using the Florida and Chile data dealt primarily with probability of occurrence. However, R. L. Dowden (22) has made some total power versus System III longitude measurements for 10.1 Mc/s data taken in Tasmania. Since our five-minute-interval analysis lent itself to this same type of study, it was only natural that we also should examine the variation of Jupiter noise power with longitude. Fig. •'+7 shows the smoothed Florida histograms for the averag flux density (S^) versus System III longitude. The dashed lines :e

PAGE 124

109 FLORIDA 1962 90** 130^ 270** LONGITUDE SYSTEM HI 360" Fig. 47. — Smoothed S versus System III histograms for data taken at the Florida station in 1962 (The dashed lines represent regions of poor statistics.)

PAGE 125

110 represent portions of the curves for which poor statistics prevailed. However, the criterion used for determining whether or not there were poor statistics varied with frequency, since the amount of data for each frequency was not the same. If the same criterion were imposed for the 5 Mc/s data (see Fig. ii8) that was used on the l8 Mc/s data, the entire 5 Mc/s curve would have heen judged uiireliable. If the number of events that was used to determine any given point on the smoothed histogram was less than that recorded in Table 5, the statistics were judged, rather arbitrarily, to be poor. The histograms in Fig. i+T show a marked tendency to peak at longitudes attributed to source B in the probability of emission histograms. More will be said about this peak when theories are discussed. We ^also note peaks at 350° for the 22.2 Mc/s data, and at 5° for the 27.6 and I8 Mc/s data. These latter are all in the range of longitudes that were judged to have unreliable statistics, so it would probably be wise not to attribute much significance to them at this time. If they continue to exist for subsequent apparitions, then their importance will have to be reconsidered. Fig. ^8 shows the smoothed average flux density (S^) versus longitude for the Chile data. The 15 and I8 Mc/s histograms show the same tendency to peak at source B longitudes that the Florida channels showed. However, the 5 and 10 Mc/s channels were those for which the source identification was \mclear. If the 10 Mc/s peak at about 50° is source B, then the source position must shift to lower longitude with frequency and the dashed curve number one on Fig. 21 must be chosen. But, when we continue this curve to 270° on the 5 Mc/s

PAGE 126

Ill CHILE 1962 90** ISO** 270** LONGITUDE SYSTEM HI 360' Fig. 48 — Smoothed S. versus System III histograms for data taken at the Chile station in I962

PAGE 127

112 histogram we see that no peak occurs for source B. The 10 Mc/s peak at 130° and the 5 Mc/s peak at 220° occur in regions of poor statistics, hence they will not be discussed further. TABLE 5 CRITEEllA FOR DETEEVUNIKG RELIABILITY OF THE AVERAGES IE FIGURES 47 AND 48 FLORIDA

PAGE 128

113 FLORIDA I9S2 90" 130° 270** 3S0 LOrJGITUDE, SYSTEM lEC Fig. h^. — Smoothed S-j^ versus System III histograms for data taken at the Florida station in I962

PAGE 129

llU CHILE 1962 i T 90*' ISO" 270' LONGITUDE, SYSTEM HI 360 Fig. 50— Smoothed St versus System III histograms for data taken at the Chile station in I962

PAGE 130

115 Note that nearly every probalDility source can "be identified in each of the average flux density (S-^) histograms. Our 10 Mc/s histogram slightly resembles Dowden's 10.1 Mc/s total power histogram. HOUR AI\'GLE STUDIES Rays incident at hour angles near the horizon have a greater path in the ionosphere than do rays incident at hour angles near the zenith. This difference in ionospheric path length should show up on an intensity versus hour angle plot as an attenuation of signals incident at large hour angles. Fig. 51 shows the smoothed average flux density (S^ and S^) versus hour angle histograms for l8 Mc/s data from "both stations. The dashed lines on the S^ versus hour angle graphs indicate regions where three-point smoothing averages included less than eight events. The effective areas of the antennas (i8E at Chile and l8 Y at Florida) used to ohtain the data in Fig. 51 were assumed to he constant with hour angle, since they were both tracking antennas. This is a poor assumption for broadside antennas. Even though they are phase switched so that they track the planet the cross sectional area of the broadside array as seen from the planet changes with hour angle. The reader should bear this in mind when reading the S. and Sy versus hour angle histograms for data taken on broadside antennas. The assumption of constant effective area is quite good for tracking yagiSj that is if they are mounted high enough (two wavelengths) off the ground so that changing ground plane effects do not become a factor.

PAGE 131

116 18.0 MC/S DEr|f\JITE DATA H

PAGE 132

117 The curves showing S^ versus hour angle have only one common peak, which is located at -70 , and this is the region of poorest statistics. The most interesting single feature, however, is the marked peak from +20° to +6o° shown on the Chile plot. Remember that in Fig. 33 the peak on the I8 Mc/s probability versus hour angle histograms occurred in the east. Whereas the probability peak in the east could be justified by better pre-opposition listening conditions (especially good since this was during Chile's winter), no such argument canbe used for a peak in the west. The author can find no obvious explanation for it, other than that it is spurious, since the Florida S^ data fail to show such a pronounced peak for these hour angles. The Sl versus hour angle plots at the top of Fig. 51 show the peak in the east that was mentioned in the previous paragraph. The Florida and Chile peaks at +50° may be significant, but the fact that the major Florida peak located at -30° lies about ten degrees to the east of the major Chile peak suggests that we are really seeing a peak due to some other phenomenon, and that the peak separation is caused by the longitude separation of the stations. From the above considerations we conclude that if either Sj\^ or Sl shows hour angle dependence, other than the expected cutoff at large plus or minus hour angles, it must be for one station only, even though the S-jplots show an identical peak at +50 . DIURML STUDIES Since the major peaks on the S-r versus hour angle histograms

PAGE 133

118 were displaced by a time corresponding to the longitude separation of the two stations, perhaps these major peaks actually are identical if S^ is plotted versus Universal Time. Fig. 52 shows the smoothed S^ and S. versus Universal Time histograms for the I8 Mc/s data from both stations. The dashed lines retain the same meaning that they had in the hour angle plots. The S;^ versus Universal Time histograms show that if the Florida graph is shifted one hour to the right, both stations will have corresponding peaks at O33O, 0530, and II30. However, this displacement is what would be expected if the peaks were caused by an hour angle dependence. But since we saw that there was no structure common to both stations on the hour angle plots in Fig. 5I, it is probably safe to assume that none of the peaks is significant. The S-j^ versus Universal Time plots at the top of Fig. 52 show only one common peak, which is at 09OO. In view of the random location of other peaks of the same size, the location of these peaks at the same Universal Time is probably accidental. The hour angle plots suggested that the S^ peaks may have been temporal in nature, but this was not confirmed by the Universal Time studies. TIME-FROM-SUNSET STUDIES Fig. 53 shows smoothed 18 Mc/s, S. and Sy versus time-fromsunset histograms for data from both stations. The dashed lines again indicate averages computed for less than eight events. The S^ versus time-from-sunset histograms each show two main peaks, which can be made to coincide if the Florida peaks are shifted

PAGE 134

119 18.0 UC/S — rDEFINITE DATA — i — T" T I brV-i — -' 0200 0400 OSOO 0800 1000 1200 UNIVERSAL TIME Fig. 52. —Smoothed Sj. and S^ versus Universal Time tograms using the I8.O Mc/s data taken in I962 his-

PAGE 135

120 18.0 MC/S DEFINITE DATA 12 3 4 5 6. T!^^»E FROTvl SUNSET IN HOURS Fig53 — Smoothed S and S versus time-from-sunset histograms ^ -ing the l8.0 Mu/s data taken in 1962

PAGE 136

121 h^ minutes^ more or less, toward sunset. One of these peaks (the Chile peak at 1 hour 50 minutes) is located in the region of poor statistics. The S-r versus tirae-fromsunset histograms displayed at the top of Fig. 53 show little significant structure other than the tendency to increase with time from sunset. This was also seen on the prohability histograms in Figs. 36 and JJ . TIME-FROM-SUJffilSE STUDIES Continuing our search for possible spatial or temporal parameters that will yield simultaneous Florida and Chile average fl\r!c density peaks, we plot the smoothed S^ and Sj^ versus time-fromsunrise histograms for the I8 Mc/s Florida and Chile data, as seen in Fig. 5^Again, most of the Florida and Chile peaks can be made to coincide, this time by shifting the Florida peaksone hoiKtoward sunrise. The sun rises about one hour later in Florida than in Chile, so this separation iraplies that the peaks were probably at the same Universal Times. But since no peaks common to both stations showed up on the Universal Time histograms in Fig. 52, we again must assume that these peaks are just statistical fluctuations. Note that the Chile histograms for both S^ and S^ show a marked peak about six hours before sunrise. ELOHGATIOH STUDIES A natural continuation of the probability versus elongation study is an intensity versus elongation study. The parameter that we

PAGE 137

122 18.0 MC/S DEFIMITE DATA >-

PAGE 138

123 chose as a measiore of the intensity is the activity index rate. This parameter vas chosen hecause N. F. Six (lO) used it in his 1961 elongation studies^ in which he detected a tendency for the monthly average activity index rate to decrease around opposition. Fig. 55 shows the monthly average activity index rate for the Florida data plotted versus time, and the elongation of Jupiter plotted versus time. All of the graphS;, except that for 15 Mc/s, show a dip in September, a tendency also noticed on the probahility plots. But as with the average flux density (S^ ), the activity index rate depends directly on the probability, for we recall that the activity index rate 4(flux density), x (activity tiiae). is defined as: — ^ This is total listening intervals slightly different from the average flux density (S^^), which is defined 4-(flux density), x K., as: — where in both cases i runs over all total lisxenint intervals of the activity intervals. One reduces to the other if all of the intervals are of the same length, but we recall that if an interval was less than five minutes long, it was reckoned as a complete five-minute interval in the S. and S^ averages. At any rate, both S^ and the activity index rate involve the ratio of activity time to listening time, which can be recognized as the probability. To further check the behavior of the activity index rate with time, we construct the plots in Fig. 56 for the Chile data. The dip at opposition is noticeably absent for each frequency except 18 Mc/s, where it is very broad. The most striking feature of the plots in both Figs. 55 and 56 is the peak located in June for every frequency

PAGE 139

124 < 27.S MC/S '^^ YEARLY AVERAGE o O < liJ o U > < o . 55 10 ' YEARLY "" AVERAGE I YEARLY AVERAGE =0. 22.2 MC/S 13.0 MC/3 JULY SEPT. 1962 TK^! Fig55 ---Variation of the monthly average activity index rate with time and Jupiter elongation, for the Florida datataken in I962

PAGE 140

125 .10 'g^ .051 9. .2 I r YEARLY AVERAGE CHILE I R i 1 I ! r 27.6 MC/S -? 18.0 MC/S t YEARLY H AVERAGE 15.0 MC/S YE'ARLY AVERAGE 10.0 MC/S YEARLY AVERAGE i».ni..iiii.. n 5.0 MC/S H ^^^-YEARLY AVERAGE » <^__» MAR. I^JiAY JULY SEPT. NOV. yAil 1962 1963 TIME Fig. 56. — Time variation of the monthly average activity index rate for the Chile data taken in 1962

PAGE 141

126 except 27.6 Mc/s, Florida^ and 10 Mc/s^ Chile. The 10 Mc/s channel in Chile vas not on the air in June^ so the only frequency actually not showing a peak in June is 27.6 Mc/s^ Florida. Wotahly;, this is the frequency that shows the most pronounced dip at opposition^ hut comparison with the I961, 27.6 Mc/s elongation plot by W. F. Six (lO) shows that there is no resemblance whatsoever. Attention should he called to the fact that flux densities in our plots have not been normalized to the mean Earth-Jupiter distance of five astronomical units^ as was done by W. F. Six (lO) for the I961 data. Taking cognizance of the width of the dip at opposition in the Florida data^, and of its non-existence in the Chile data^ one must be careful about attributing too much significance to any theories put forth to explain it. SEARCH FOR PERIODIC RECURRMCES Observers have noticed that when Jupiter activity occurs^ it often continues for several days, which are then followed by several relatively quiet days. Carr et. al . (18) and Six (lO) found evidence that 18 Mc/s Jupiter activity occurs in an eight-day cycle. However, in the analysis of the i960 and 1961 data by Six, only the I961 data showed this periodic nature. In order to make a search for this eightday period and other periods, such as the 27-day rotational period of the s'on, and in order to present a compilation of our data, we plot in Figs. 57 through 60, respectively, the daily activity index rates for the 5 and 10 Mc/s Chile data, the 15 Mc/s Florida and Chile data, the 18 Mc/s Florida and Chile data, and the 22.2 and 27.6 Mc/s Florida data. The I6 and 22.2 Mc/s Chile data have not been plotted, since

PAGE 142

127 OJ vo -p n3 n3 to o O d H c o LTN (U ^ -P fn O Ch (D nJ CD
PAGE 143

8 CO Q. O CM I CVJ 1 O 3 2 I< ^ ^ fi UJ O Q >I< < 0<± 2 APR -^ I I I P 20 30 10 20 OBER NOVEMBER 15. c

PAGE 145

128 POSSIBLE AND DEFINITE DATA-15.0 MC/S. 1962 8 Q. O CM I I o c^ 2 UJ I< FLORIDA / 15 iJ_I IL Ll) O Q < 4 < o a-^ CHILE J. i n I..I iL 10 20 30 APRIL 20 30 MAY 10 20 30 JUNE 20 30 10 20 30 10 20 30 10 20 30 10 20 JULY AUGUST SEPTEMBER OCTOBER NOVEMBER Fig. 53.— Dally aotlv't. index rate fcr tue 15-0 Mc/b data taken In 1^/32

PAGE 146

t"4 CO Q. O CVJ I CVi 'o I< X Q 3 > 2 H o < < i LL_L Ll -J I'M H 1 3( 10 20 30 10 20 30 10 20 30 NOVEMBER DECEMBER JANUARY

PAGE 147

129 POSSIBLE AND DEFIN TE DATA 18. C MC/S, 1962 -r T r 1 r 1 r "1 r 1 1 \ r FLORIDA 1± ' III ill I II lillll'l ! ! Ii 1 r III, ill! IllliiliUl III I ! I ml LIL IN -II' I r -u-j lU 1 r J liJJ H I 30 CHILE 10 20 APRIL 30 a I 10 iili J_l illll|lllllll llllll|lllll X ijlL d 20 30 MAY 10 20 JUNE 30 10 20 30 JULV 10 2C 30 10 20 30 10 20 30 AUGUST SEPTEMBER OCTOBER I 10 T i T r 20 30 10 20 30 NOVEMBER DECEMBER — I 1 r10 20 30 JANUARY ^'^5S'.— I>-iljactiity Ino-ix -atfoi tue 13.' data taXeii in . /'.2 •fc/i /

PAGE 149

130 _ 4 en Q. o I 2\b UJ I I r I I I POSSIBLE AND DEFINITE DATA 1962 -1 1 I I 1 r T 1 r I I I FLORIDA 22.2 MC/S III i I I .1. .. . I I 1 < 18 Q z 14 >> 10 Io < 81> < Q 4L 2 FLORIDA 27.6 MC/S '• J. + f "1 I I I I I 1 1 1 I I I I I I I I I 1 1 1 1 1 1 r 10 20 30 10 20 30 10 20 30 10 20 30 10 20 30 10 20 30 10 20 30 10 20 30 10 20 30 MAY JUNE JULY AUGUST SEPTEMBER OCTOBER NOVEMBER DECEMBER JANUARY Fig. 60. — Daily activity index rate for the data taken at 22.2 and 27.6 Mc/s in 1962

PAGE 150

131 we did not reduce the data from the polar imeters for use with the new program. All of our activity index rate studies up to this point have heen made using output information from the new program, and thus it would be confusing to change to the old program analysis for this particular study. The Chile, 27-6 Mc/s data were also omitted since they were taken over a two-month period which yielded only nine days of activity. Figs. 57 through 60 fail to show any immediately obvious periodic structure. Further examination was made of the l8 Mc/s data received during the two months around opposition at the Chile station. This channel was chosen since it was the one analyzed by W. F. Six (lO) in 1961. The three-point running average of this data was plotted, but no periodic nature was detected. However, the smoothed plot showed, as we can see from several of the plots in Figs. 57 through 60, that Jovian activity generally occurs several days in succession. In the next chapter we will make a more systematic search for a periodic recurrence of Jupiter activity.

PAGE 151

CHAPTER VI SOLAR AMD GEOPHYSICAL CORRELATIONS WITH JOVIAN EMISSION In their respective studies N. F. Six (lO), J. W. Warwick {k), T. D. Carr (23), and A. J. Plourde (24) have found the sporadic decametric emission from the giant planet to be correlated in some vay with solar or geophysical activity. However, they did not all find the same types of correlation. In this chapter we will examine the 1962 data in hope of finding some type of correlation or, at least, some ideas for further studies. LONG-TERM EFFECTS It is well known that the geomagnetic index (Ap), cosmic ray intensity, auroral activity, and other terrestrial parameters are affected by the long-term variation of solar activity known as the sunspot cycle. It is not unlikely, then, that the sun might cause a variation of parameters describing events which occur elsewhere in the solar system. Fig. 61 shows the variation of two Jovian parameters, source width and average probability of observing Jovian radiation, with sunspot number. We see that the width of the I8 Mc/s source follows the inverse of the sunspot cycle very closely, while the average probabilities reach their minima about one year after sunspot maximum. 132

PAGE 152

133 -i — 3 955 1956 1957 I SSS'l 959'l 960'l961 'l962'l963' YEAR Fig. 6l. — Comparison of the width of soiorce A and of the average probability of emission with sunspot number

PAGE 153

13^ Note, also, that the small hump on the l8 Mc/s prohability plot occurs about one year later than the one on the source width curve. Six (lO) suggested that this source width dependence on solar activity might be caused by variable focusing originating in Jupiter's ionosphere. This focusing may also originate in the interplanetary medium. Even though this medium is very tenuous, the fact that it is a magnetoionic medium of great extent leads one to believe that it is capable of focusing electromagnetic waves. Nov, the electron density in Jupiter's ionosphere is probably directly dependent on solar activity. However, since Jupiter's ionosphere is so large, it seems likely that its electron density does not change simultaneously with changes in solar activity, but that it exhibits some delay. We see this same effect in the daily variation of the terrestrial ionospheric electron density, the electron density being lowest just before sunrise. If this same delay occurs on a longer time scale, and if changes of the Jovian ionospheric electron density are responsible for varying the probability of emission, then we should expect the displacement of the probability and solar activity curves seen in Fig. 6l. However, since the interplanetary medium is so tenuous, we should not expect so .:iarked a delay between changes in solar activity and changes in the interplanetary electron density. Therefore, if, as has been suggested, the change in source width arises because of changes in the interplanetary electron density, we should not see a displacement of features on the source width curve with respect to those on the solar activity curve. In

PAGE 154

135 Fig. 61 we see that features on the source width curve do line up with those in the solar activity curve. That the apparition average activity index rate (old analysis) varies inversely with the sunspot cycle can be seen in Fig. 62. The magnitude of the solar latitude of the sub-Jovian point is plotted at the bottom of Fig. 62 in the interest of maintaining continuity with the work begim by N. F. Six (lO) in I96I. A. G. Smith (25) suggested that in the event that solar flares spew particles radially away from the sun in narrow streams the solar latitude of the subJovian point may be important. It is hard to believe that the small variation of the magnitude of the solar latitude of the sub-Jovian point (it varies from to 8° during a Jovian year) should give rise to important effects. However^ since the intensity of solar cosmic rays reaching the earth has been discovered to change with the solar latitude of the sub -terrestrial pointy this study had a firm basis. If any latitude changes are responsible for these variations^ they should be the changes in latitude of the active regions^ which occur at latitudes ranging from + 30° to +5° during the sunspot cycle. It will be difficult to separate effects caused by latitude changes of the solar active regions from those caused by over -all activity changes, since they vary with nearly the same period. Smith has suggested that a study be made of a combination of the two effects. SHORT-TERM CORREIATIOWS Correction of the I96I Analysis The program that was used for correlating the I960 and 1961

PAGE 155

136 I9S3 Fig. 62. — Comparison of Jupiter's solar latitude and the apparition average activity index rate with sunspot numher

PAGE 156

137 solar and Jovian activity indices was described by Six (lO) in his dissertation. However, he did not write the program^ and evidently a lack of conimunication between him and the programmer resulted in a program error which deleted all pre-opposition solar activity calculations. All of his geomagnetic index (Ap), activity index rate, and sunspot number calculations were nevertheless correct. Fig. 63 shows FXl, FX2, and FX3 as determined by the Chree analysis of the 20 days of greatest activity for the I8 Mc/s, Florida data of 1961. FXl is the solar flare activity index for flares occurring on that part of the solar disk bounded on the north and south by the +30 latitudes and bounded on the east and west by the +30° longitudes (region one), the longitude of the sub-Jovian point being chosen as zero. FX2 is the flare activity index for flares occurring in the region boimded by the central meridian and the eastern limb as seen from Jupiter, and by the same latitudes described for FXl (region two). FX3 is the flare activity index for flares occurring in that half of the solar disk visible from Jupiter that was not included in FX2 (region three). Fig. 6k shows a sketch of these regions as found on page 1^7 of Six's (lO) dissertation. Figs. 65, 66, and 67 show the respective solar correlation studies for the 27.6 Mc/s data taken at Florida in I961, the I8 Mc/s Florida data taken during three months around opposition in 19^1, and the I8 Mc/s Chile data of i960. Plots of flare numbers were also made in the I96I analysis, but they are not included here in the interest of saving space.

PAGE 157

138 8.0 MC/S FLORIDA 1961 to 080|6oL 40 20 __j I i » .1,, I I L ' -!5 -io -II -9 -7 -5 -3-113 DAY NUMBER IN RELATION TO THE PEAK DAY Fig. 63. — Chree analysis of the solar flare activity index in groups one, two, and three, using the 20 peak days of 18.0 Mc/s Jupiter emission at the Florida station in I961

PAGE 158

139 \^<{(r^\ GROUP I GROUP 2 CENTRAL MERIDIAM OF TME SUN AS SEEN FROM JUPITER MERIDIANS Fig. 6k. — Regions on the solar disk as viewed from Jupiter assigning flares to groups one, two, or .three

PAGE 159

In Fig. 63 the outstanding feature on the FXl curve is the peak at -8 days. Other prominent features are the broad dip around -11 days and the dip at -k days. The FX2 curve shows a major peak at -3 days^ another at -1 day^ and another at -7 days. The major dip falls at -9 days. The FX3 plot shows maxima at -8, -6, and +1 days and a hroad minimum around -3 days. Since the 27.6 Mc/s emission occurred with less total probability than the I8 Mc/s emission;, solar events causing 27.6 Mc/s Jovian emission should be either greater in prominence or different in nature than those triggering 18 Mc/s emission. Fig. 65 shows FXl, FX2, and FX3 for the 27.6 Mc/s data taken at Florida in I961. The peak at -2 days on the FXl curve is sufficiently high to be conclusive, but the peak at +k days raises some doubt regarding the importance of the -2 day peak. However, the prominence of the latter is sufficient to warrant 27.6 Mc/s studies for other years. FX2 shows peaks at +1 and +3 days and FX3 shows peaks at -2, 0, and +h days, the -2 day and +k day peaks probably having arisen from the same flares that produced the -2 day and +k day peaks on FXl, since the flare regions used for FXl and FX3 overlap. Since Jupiter radiation may have been triggered by flares that could not have been seen from the earth when Jupiter was far from opposition, 20 days of peak activity were chosen from the three months around opposition. The Chree analysis of these 20 days of peak activity received on the I8 Mc/s channel at the Florida station in I961 is shown in Fig. 66. We see major peaks at -8, -6, and -+4 days, and

PAGE 160

lul 27.6 MC/S FLORIDA IS6{ Q L...^J_J>_..J.....J._.t , » t „? 1 1 \ i 1 , 1 ' I „' „ t -15 -r3 -U -9 -7 -5 -3-1 13 DAY NUMBER IN RELATION TO THE PEAK DAY Fig. 65.--Chree analysis of solar flare activity index in groups one, two, and three, using the 20 peak days of 27.6 Mc/s Jupiter emission at the Florida station in I961

PAGE 161

lU2 18.0 UC/S FLORIDA 1961 3 MONTHS AROUND OPPOSITION -15 -13 -II Fig. 66. — Chree analysis of solar flare activity index in groups one^ two, and tliree, using the 20 peak days of l8 Mc/s Jupiter emission at the Florida station during the three months around opposition in I96I

PAGE 162

1^3 minima around -12 and -3 days on the FXl plot, FX2 registers peaks at -3 and -1 days^ while FX3 shows a major peak at -6 days and a "broad dip around -2 days. As suggested hy Carr (2) the lack of solar activity on days heiore Jupiter emission may be required so that guiding ducts can be formed in Jupiter's ionosphere, or, as suggested by A. G. Smith (25), this quiet time may be needed to clear the interplanetary medium. Galactic cosmic rays are more numerous when the Sim is quiet. Fig. 67 shows that for the I96O, I8 Mc/s Chile data, FXl posesses major peaks at -1 and -I3 days and a broad minimum at -11 days. FX2 shows peaks at -1, -5, -7, and -I3 days and a dip at -10 days, while FX3 shows peaks at -8 and -14 days and a broad minimum at -11 days. Reviewing Figs. 63, 65, 66, and 67, we see that the only feature repeated on all of the 18 Mc/s, FXl plots is the general null at -11 days. This feature also appears in the Chile data for three months around opposition in I96O, a plot not shown here. The only feature repeated for FX2 is a peak at -1 day, and the main attraction on the FX3 curves is a peak at -8 days. However, none of these general features is evident in the 27.6 Mc/s figure. We will compare these general features with those exhibited by the 1962 data. Changes in the Solar Flare Program As was mentioned in the first section of this chapter, the solar flare program contained an error that caused deletion of all preopposition data. The grouping of flares into regions one, two, and

PAGE 163

Ikk 18.0 MC/S CHILE I960 "j — I — j — ! — J — i — i — ! — r3 -!5 -13 -II -9 -7 -5 -3 -I I DAY IV3UMEER IN RELATION TO THE PEAK DAY Fig. 67.— Chree analysis of solar flare activity index in groups one, two, and three, using the 20 peak days of I8 Mc/s Jupiter emission at the Chile station in I96O

PAGE 164

1^5 three, in order to obtain the flare indices FXl, FX2, and FX3, involved the solar longitude of the flare and the difference beween the right ascension of the sun and that of Jupiter. Using this information, the longitude displacement ( 3 ) of the flare from the solar subJovian meridian was found. But the computer did not always take the difference between the right ascensions of Jupiter and the sun in module 360°, as was necessary if flares were to be correctly grouped in sections one, two, and three. The fact that the computer calculated to be greater than 360° for all pre-opposition flares caused the program to exclude them from each of these sections. Effecting this correction was the most significant change made. However, four other changes were made. The first, only a matter of convenience, allowed analysis of any nimber of peak days, whereas 20 had to be chosen before. It also allowed the deck for peak days chosen from the entire apparition, the deck for peak days chosen around opposition, the deck for days of no activity, and any other decks chosen for one frequency, to be processed in one computer run. The second change involved decks of peak days chosen for each of the Jovian longitude regions, 0° to 90° (no source), 90° to 190° (source B), 190° to 290° (source A), and 290° to 360° (source C). In the Chree analysis of peak days chosen for a given Jovian region, the parameters FXl, FX2, K(3. FNl, FN2, FW3. Ap, R^. J^. and 1280O (activity index of the solar 2800 Mc/s flux), were averaged over each twoday period. (For the Chree analysis studies making use of activity from the entire planet averages were calculated for each day. ) For example, the regular solar flare program listed the average of each of the parameters for every one of the 15 days before a peak day to h days

PAGE 165

1^6 after peak emission vas received^ whereas the revised program listed two-day averages on -15, -13, -H +3 days. The averages were performed in this manner to insure that for each average value listed we had had an opportunity to ohserve the selected region of the planet. A single average length (five-hour) watch enabled us to see only half of the planet. As the reader may recall^, some of the Chree analyses in the first section of this chapter showed peaks at -15 days and +h days. E. N. Parker (26) has recently suggested that the particles in the solar wind may take up to three weeks to get to Jupiter. If these particles are indeed responsible for triggering Jovian emission, then the -15 day limit on the solar flare program is too small. For this reason, the solar flare program was modified so that the Chree analysis was performed for a time interval of -35 days to +k days. With this new range we must add the restriction that no peak days can be chosen which lie within 35 days of the beginning of the apparition, or within k days of the end of the apparition. The final change in the solar flare program can be seen in the list of calculated parameters already mentioned in this section. In addition to the parameters calculated in the I961 analysis, the 1962 analysis calculates the Chree analysis averages for the 2800 Mc/s solar flux. The daily values of solar flux at 2800 Mc/s as taken at Ottowa, Canada, were fed into the solar flare program along with the input parameters already described by Six (lO). The program listed the Chree analysis of the 2800 Mc/s solar fl\;uc with the rest of the Chree analysis output.

PAGE 166

147 Chree Analysis Results of the I962 Solar Correlation Studies A preliminary Chree analysis \ira.s made of the 20 days of peak activity (see Table 6) that were recorded on the I8 Mc/s channel at the Florida station in I962. This analysis showed^ as we saw in seme of the previous Chree analysis curves, that peaks appeared to develop around -l4 or -15 days, but the fact that the analysis was performed only on days numbering from -15 to +h made it impossible to see the parts of the peaks lying on days numbered before -15 . For this reason, and because of the previously-cited statement by E. N. Parker, we have extended the Chree analysis to include days numbering back to -35 days. Tnis number was chosen in order to include one complete solar rotation (27 days) so that any variations having this period would also be visible. C hree Analysis of the 18 Mc/s Jovian Emission Kc'lzored at the Florida Station in I962 Fig. 68 shows the Chree analysis of solar flare activity in regions one, two, and three for the 20 days of peak activity monitored on the 18 Mc/s;, Florida channel. The FXl plot shows peaks at -8, -15, -24, and -30 days, minima at -3U, -25, -21, and -12 days and a broad dip from -5 to +5 days. The FX2 plot shows peaks at -8, -22, and -27 days and minima at -31^ -25^ -9^ and -h days. The FX3 plot shows peaks at -9, -15, -24, and -30 days and minima at -34, -26, -11, and +2 days. We notice that the peaks and dips of the FXl and FX3 curves fall on about the same days. This would suggest that flares located

PAGE 167

143 TABLE 6 DAYS OF PEAK ACTIVITY USED IN THE CHREE ANALYSIS OF THE 18 Mc/s DATA IK I962 Days of peak activity 18 Mc/s Florida^ 1962 Days of peak activity 18 Mc/s Chile, 1962 Days of peak activity that resulted when the activity index rates from Florida and Chile were added May

PAGE 168

li+9 20 PEAK DAYS1 8.0 MC/S. FLORIDA I9S2 -'™^^^-'^•"-jiiU^lim ;;^^-'^"»"" 5'-imtm,tB-r—r^trm-^ ^-.— ...^ 1 -33-29 -25 -21 -17 -13 -9 -5 -I +3 DAY NUMBER IN RELATION TO THE PEAK DAY Fig. 68. — Chree analysis of the solar flare activity index in groups one, two, and thi'ee, using the 20 days of peak activity of the l8 Mc/s Jupiter emission at the Florida station in 1962

PAGE 169

150 in the sols-r longitude region from to 30° west (that portion of the sun common to regions one and three) are primarily responsible for the structure on the FXl and FX3 plots. We can notice an important feature at this point. The peaks on the FXl and FX3 plots occur at intervals of ahout eight days^ with the first one lying at ahout -8 days. We also see that the first FX2 peak occurs at -8 days. If any one flare was responsible for the peaks shown in Fig. 68, they should disappear when a Chree analysis is made of the number of flares. In doing thiS;, we weight each flare equally. Chree analysis of the flare number in regions one, two, and three is shown in Fig. 69. The peak days that were used in Fig. 68 were also used for the studies in Figs. 69 and 70. We see that the -8 day peak of FHl corresponds to that at -8 days on the FXl plot, but we find that two of the FXl peaks, -2k and -30 days, lie in regions of low flare number on the FNl plot. The F1I2 plot shows about the same structure as we saw in the FX2 plot. The FW3 plot also shows about the same structure as we saw in FX3, but a new peak has cropped up at -I7 days. An examination of the listing of the solar flare program showed that on September 1, I962, the sols.r flare activity index exceeded 1100 in regions one and three. This value is nearly twice that of any other flare index occ\nrring in regions one and three during the rest of the apparition. An examination of the original source of the data. Part B , Solar Geophysical lata , KBS, Boulder, Colorado, showed that two flares recorded by an observing station in Athens, Greece, gave rise to this high activity index. In both cases these

PAGE 170

151 20 PEAK DAYS-18.0 IVIC/S. FLORIDA 1962 f-T-n-^i — r-j — t i i I > t < I -33 -29 -25 -21 -17 -13 -9 -5 -I +3 DAY NUMBER IN RELATION TO THE PEAK DAY Fig. 69. — Chree analysis of the number of flares in groups one; twO; and three using the 20 days of peak Jupiter activity at 18 Mc/s at the Florida station in 19o2

PAGE 171

152 same flares were observed "by four other stations, "but none of them credited the flares with as high an importance number or as long a duration as was reported by the Athens station. However, when we reduced the data, we picked the duration and the importance number of a given flare in the following way: the beginning time was chosen to be the earliest time that any station observed it; the ending time was chosen to be th latest time that any station observed it; and the liaportance number was chosen to be the largest importance number attributed to it by any station. Thus, if two stations reported different flare durations and importance nijmbers, we chose those leading to the highest flare activity index. This practice may not lead to as reliable a solar flare activity index as would be obtained if some way of weighting the data were used. Fig. 70 shows the Chree analysis of the 280O Mc/s solar flux, the Zurich provisional sunspot number {^^) , the geomagnetic index (Ap), and the Jovian daily activity index rate. The 28OO Mc/s flux and R^ each show a peak at -8 days, but each also shows a greater peak at -I9 days. It is interesting to note that the time difference between the two major dips on the 28OO Mc/s plot is about 28 days, the rotational period of the solar regions located at latitudes of about +50 . The Ap curve shows a maximum at -28 days and the Jovian daily activity index rate curve shows maxima at and -32 days, the former, of course, being due to the fact that the 20 peak days were by definition chosen so that the activity was maximum. It is interesting to note that the effect seen on the 28OO Mc/s curve could be the superposition of a peak at -8 days with a

PAGE 172

153 20 PEAK DAYS-18.0 (VIC/S. FLORIDA 1962 i t I -33 -29 -25 -21 -17 -13 -9 -5 -I -^3 DAY MUiViBER IN RELATION 70 THE PEAK DAY Fig, 70.--Chree analysis of the 2800 Mc/s solar flu:-:^ the Zurich provisional sunspot number (H^), the geomagnetic index (Ap)^ and the Jovian daily activity index rate using the 20 days of peak Jupiter activity at l8 Mc/s at the Florida station in I962

PAGE 173

154 iDroad 27-day peak caused "by the solar rotation. Since the 28OO Mc/s fliix varies with the solar rotation^ one should determine if the Jovian activity index rate is greatest when a given solar longitude is facing Jupiter. However;, no 27-day cycle appeared on the Jovian daily activity index rate plots in Chapter V. AlsO;, in the case of the flare activity indiceS;, one could argue that the peak at -8 days was caused "by the 1 ansit time of flare produced particles from the sun to Jupiter, hut in the case of the 28OO Mc/s flux no such argument can he made, since electromagnetic radiation travels from the sun to Jupiter in ahout kO minutes. However, if particle emission occurred simultaneously with the 2800 Mc/s flux, then an eight-day lag would have meaning. Chree Analysis of the I8 Mc/s Jovian Emission Monitored at the Chile Station in 19o2 Fig. 71 shows the Chree analysis of FXl, FX2, and FX3 using the 21 days of peak Jupiter activity (see Tahle 6) as monitored at I8 Mc/s by the Chile station in I962. Fig. 72 shows the Chree analysis of FKL, FN2, and FN3 and Fig. 73 shows the Chree analysis of the 28OO Mc/s solar flux, R^, Ap, and the Jupiter daily activity index rate. FXl shows maxima at -1 and -I7 days and minima at -5 to -7 and -31 days. Note that the dips lie nearly one solar rotational period (27 days) apart. FX2 shows peaks at -13 and -I8 days and a broad minimum between -23 and -35 (iays. FK3 shows several peaks and minima, with the greatest peak lying at -17 days and the smallest minimum at -32 days. The only feature that any of these share with the Florida data is a minimura around -30 days. 1X2 does have a common peak at

PAGE 174

155 I9S2 ^ <-Olr L^L=J-.J:™=!^.. L.»L«i ' I I > t t r, I '! >iii ' -33 -29 -25 -21 -17 -13 -9 -5 -! 4-3 .DAY F!Ufl^3ER IW R2:LATI0i\J TO THE PEAK HAY Fig. 71— Chree analysis of the solar flare activi';/ index in groups one, two^ and fnree^ using the 21 days of peak activity of the l8 Mc/s Jupiter emission at the Chile station in 19o2

PAGE 175

156 -20 days, but for the most part, the Florida and Chile FXl, FX2, and FX3 plots possess very little common structure. This fact casts some doubt on any conclusions that are dravn from the Chree analysis of the Florida data, and not the Chile data. In Fig. 72 we see that the only structure on the FKl plot is a dip at -31 days. FW2 shows peaks at -6, -8, -13, and -20 days and a major dip from -26 to -28 days. FK3 shows peaks at -3, -lU, -19, and -29 days and a major dip at -2U days. The FN2 peaks at -6, -8, and -20 days, the FW3 peak at -29 days, and the FN3 minimum at -24 days are the only features shared by the Florida and Chile Chree analyses. In Fig. 73 we see the same broad 2800 Mc/s peak that was seen on the Florida plot. We also see the smaller peak at -8 days that was evident on the T^-lorida plot. On the R^ plot the saiae -I8 day maximum that we saw on the Florida plot is evident, but the additional structure that occurred on the Florida R^ graph is not seen on the Chile graph. No definite structure of any kind is present on the Ap Chree analysis curve, but the Jupiter daily activity index rate plot shows peaks at (the expected peak), -7, -1^, and -32 days, the -7 and -32 day peaks being in common with the Florida daily activity index rate curve. If one stretches his imagination, he can see the eight-day period cropping up in the Chile daily activity index rate curve. It seems alarming that the Florida and Chile data should yield such different Chree analysis curves, but remember that the Florida and Chile observing seasons were quite different (the Florida observing season extended both before and after the Chile season);

PAGE 176

157 21 PEAK DAYS-I8MC/S. CHILE 1962 -33 -29 -25 -21 -17-13 -9 -5 -I +3 DAY NUMBER IN RELATION TO THE PEAK DAY Fig: 72, — Chree analysis of the n\amber of flares in groups one, two, and three, using the 21 days of peak Jupiter activity at l8 Mc/s at the Chile station in 1962

PAGE 177

158 21 PEAK DAYS-I8.Q MC/S, CHILE 1962 ' I i I « I • i » — 1 • I • t i k — -33 -29 -25 -21 -17 -13 -9 -5 -I +3 DAY NUMBER IN RELATION TO THE PEAK DAY Fig. 73.— Chree analysis of the 2800 Mc/s solar flux, the Zurich provisional sunspot num'ber (Rz)^ the geomagnetic index (Ap), and the Jovian daily activity index rate, using the 21 days of peak Jupiter activity at l8 Mc/s at the Chile station in 1962

PAGE 178

159 hence^ different solar flare populations were sampled. I do not believe, however, that this should cause so great a difference between the Chree analyses performed for the two stations. We later discovered that the April 30th peak day fell 30 days frcm the beginning of the apparition. This violates our restriction which requires peak days to be at least 35 days from the beginning of the apparition, so the -30 to -35 day points on the Chile Chree analyses in Figs. 71, 72, and 73 may not be reliable. Chree Analysis of the Florida l8 Mc/s Data using a Solar Flare Deck which has had the Sub-flare Cards Removed In Figs, jk and 75 an effort is made to "clean up" the Chree analysis curves by deleting sub-flares from the input deck to the solar flare program. Fig. 7^1 shows FXl, FX2, and FX3 for the same peak days (see Table 6) as were used for the Florida data in Figs. 68, 69, and 70. If one compares the plots in Fig. jk with the plots in Fig. 68 that were obtained including the sub-flares, he sees that the structure is nearly identical and that deleting the sub-flares lowered the background of the flare activity index about five units. The relative heights of the peaks on the FX2 curve did change a little, however. Fig. 75 shows the FNl, FW2, and FN3 curves obtained by using the same peak days 'see Table 6) as were used for the Florida data in Fig. 69. As we noticed for the flare activity index curves (FXl, FX2, and FX3), the location of the peaks on the FNl and FN'2 ciorves changes very little when sub-flares are deleted. The peaks, however, become sharper and better defined when the sub-flares are deleted, but at

PAGE 179

i6o NO SUBFLARE 20 PEAK DAYS-18.0 MC/S. FLORIDA 1962 -33 -29 -25 -21 -17 -13 -9 -5 -I +3 DAY NUMBER IN RELATION TO THE PEAK DAY Fig. 7U. — Chree analysis of the solar flare activity index (excluding rS'ubf lares) in groups one, two, and three, using the 20 days of peak activity of the l8 Mc/s Jupiter emission at the Florida station in I962

PAGE 180

16 1 NO SUBFLARE 20 PEAK D AYS 18.0 MC/S , FLORIDA 1962 * » ' 1 t i ' i 1 i 1 t 1 -33 -29 -25 -21 -17 -13 -9 -5 -I +3 DAY NUMBER IN RELATION TO THE PEAK DAY Fig. 75.— Chree analysis of. the number of flares (excluding subflares) in groups one, two, and three, using the 20 days of peak Jupiter activity at the Florida station in 1962

PAGE 181

l62 the same time the structure on the FN3 plot is washed out. It seems safe to conclude that sub-flares play an insignificant role in producing structure on the Chree analysis curves, and that they could "be eliminated from further solar flare studies. Chree Analysis of the l8 Mc/s Jovian Emission Dui^in.j the Three Months Around Opposition in 1962 In the 1961 Chree analysis of the solar flare data Six (lO) also made a study of peak days chosen around opposition. The peak days for the I962 studies around opposition were chosen after the Florida and Chile daily activity index rates had "been summed. As was done for the 1961 "around opposition" analysis, we chose the peak days only from those days lying within I.5 months of opposition. These peak days are listed in Tahle 6. Fig. 76 shows the Chree analysis of FXl, FX2, and FX3 using the 20 days of peak Jovian activity that occurred during the three months around opposition. The FXl curve shows peaks occurring at -3, -8, -17, -23, and -30 days. These peaks do not fall at exactly eightday intervals, but the periodicity is evident. The FX2 curve shows surprisingly little structure, but the FX3 plot shows the same peaks that we saw in FXl. In fact, the peaks in the FX3 plot are sharper than those in FXl. Recall that the FXl and FX3 plots which were made for the Florida data, and which are shown in Fig. 68, showed about the same structure. Remember, also, that the FX3 peaks were sharper than the FXl peaks in that case. Fig. 77 shows the Chree analysis of FWl, FEE, and FK3 for the

PAGE 182

163 20 PEAK DAYS. CHILE-FLORIDA, 1962 18.0 MC/S, 3 MONTHS AROUND OPPOSITION -33 -29 -25 -21 -17 -13 -9 DAY NUMBER IN RELATION TO -5 -I +3 THE PEAK DAY Fig, 76.--Chree analysis of the solar flare activity index in groups one, two, and three, using the 20 days of peak Jovian activity (Florida plus Chile) monitored by both stations at 18 Mc/s during the three months around opposition in 1962

PAGE 183

161+ 20 PEAK DAYS CHILE-FLORIDA. 1962 18.0 MC/S. 3 MONTHS AROUND OPPOSITION ' I • i — « I ' — k < I k — I — I — j — s — I — I — s — J — r ] -33 -29 -25 -21 -17 -13 -9 -5 DAY NUMBER IN RELATION TO THE -I PEA" ks +3 DAY Fig. 77---Cliree analysis of the number of solar flares in groups one, two, and three, using the 20 days of peak Jovian activity (Florida plus Chile) monitored by both stations at 18 Mc/s during the three months around opposition in I962

PAGE 184

165 same peak days that were used for the studies in Fig. 76. The FNl peaks occior at about the same places as they occurred on the FXl plot, FM2 shows little structure, but the structure that we saw on the FX3 graph again shows up in FN3. Also, as with the FXl, FX2, and FX3 studies, we see that the FNl, FN2, and FN3 analyses that were performed for the peak days around opposition show more periodic structure than do their counterparts in the Chree analyses that were performed for the peak days of the entire apparition. This leads us to conclude that the "around opposition" peak days should be used in future studies. However, the studies that involved region three (that solar region lying between the central meridian and the western limb as seen from Jupiter) indicate that flares occurring in this solar region may also be significant in stimulating Jovian emission, since FX3 and FN3 show a regular structure that is absent on the FX2 and FN2 plots. If they are significant, then it seems that information might be gained by performing a Chree analysis of only the peak days occurring after opposition. However, when this study was subsequently carried out, it failed to show anything new. Fig. 78 shows the Chree analysis plots of the 280O Mc/s solar flux, B.^, Ap, and the Jovian daily activity index rate that were obtained by using the same 20 peak days around opposition that were used in Figs. 76 and 77. The Jupiter activity curve shows what might be construed as an eight-day recurrence, although it cannot be positively established from the curve. The geomagnetic index (Ap) curve shows a peak at -5 days. This suggests that particles producing magnetic activity on the earth arrive at Jupiter about five days later.

PAGE 185

166 20 PEAK DAYS. CHILEFLORIDA 1962 18.0 MC/S, 3 MONTHS AROUND OPPOSITION T -33 -29 -25 -21 -17 -13 -9 -5 -I +3 DAY NUMBER IN RELATION TO THE PEAK DAY Fig. 78.— Chree analysis of the 2800 Mc/s solar tlux, the Zurich provisional sunspot numher (R^,)^ '^^^ geomagnetic index (Ap), and the Jovian daily activity index rate, using the 20 days of peak Jupiter activity (Florida plus Chile) monitored "by both stations at l8 Mc/s during the three months around opposition in I962

PAGE 186

16? giving a sixn-toJupiter transit time of about 6.5 days. If the particles followed a straight path^ they would need a velocity of about 1^00 km/sec. However, if they, like solar cosmic rays, follow the spiral interplanetary magnetic field (see Fig. 79), their sun-toJupiter path is greater than five astronomical units and their average velocity is greater than li^•00 km/sec. The radial lines in the figure show the direction of the 300 km/sec solar wind. If we assume that the 300 km/sec solar wind particles are responsible for the spiral interplanetary magnetic field and that they continue uninhibited to Jupiter's orbit, the solar magnetic field pulled along with them will have formed a spiral which twists about ^+65°, since it takes a 30O Ian/sec particle about 35 days (I.3 solar rotational periods) to traverse the five astronomical unit sun-toJupiter distance. However, our Chree analysis did not show a peak at -35 days. Two possible explanations for this are: 1. The 300 km/sec solar wind particles are not responsible for exciting Jovian emission. 2. The 300 km/sec solar wind particles do not maintain a constant velocity during their trip from the sun to Jupiter. The fact that our "around opposition" Chree analysis shows more structure than does our "all apparition" Chree analysis suggests that the particles traversed a more-or-less straight line from the sun to Jupiter. Our Chree analysis of the geomagnetic index (Ap) especially shows this, for the Ap peak at -5 days suggests a 1.25 day sun-toearth trip, a value which corresponds fairly well with lag times

PAGE 187

168 SOLAR V^IND AND INTERPLANETARY ^^AGNETIC FIELD EARTH S ORBIT SOLAR WIND /^-MAGNETIC ' FIELD LINES Fig" 19' — The solar vind and the interplanetary magnetic field. After Parker (25).

PAGE 188

l69 observed "by Carr et. al . (15) in i960. HoweVer^ if the particles which trigger Jovian activity do follow the spiral, assuming for the moment that the particles in question do not create the interplanetary magnetic field, then they would have been excluded from our Chree analysis, their origin being at solar longitudes not visible from Jupiter. (Remember that our Chree analysis dear only with flares visible from Jupiter.) Solar cosmic rays follow this spiral (27), so it is not unlikely that the particles responsible for stimulating Jupiter noise storms also follow this trajectory. Note that their travel time need not be 35 days, as was the case with the 300 km/sec solar wind. A study of flares occurring at solar longitudes not visible from Jupiter might prove enlightening. Chree Analysis of the lays on which Wo I8 Mc/s Activity was Received at Either Station Fig. 80 shows the Chree analysis of the 2800 Mc/s solar flux, R2, Ap, and Jupiter's daily activity index rate, using the 29 days for which neither station recorded Jupiter activity during the six months around opposition in I962. The only curve that shows much structure is that of the daily activity index rate. We see the null at days and peaks at -5, -12, -20, and -28 days. The eight-day cycle that we have mentioned before is clearly present here also. It is interesting to note that the shape of the 280O Mc/s solar flux curve is nearly identical with its counterpart on the Chree analysis curves using days of peak activity. The R^ and Ap plots yield very little information. The parameters, FXl, FX2, FX3, FNl, FN2, and FW3 were

PAGE 189

170 29 DAYS OF NO ACT. AT EITHER STATION 8 MC/S, 6 MONTHS AROUND OPPOSITION, 1962 » — s — j — = '— -3 3 -29 -25 DAY NUMBER -21 -17 -13 -9 -5 IN RELATION TO THE -I PEAK +3 DAY Fic^ 80.— Chree analysis of the 2800 Mc/s solar flux, the Zurich provisional sunspot number (R^), the geomagnetic index (Ap), and the Jovian daily activity index rate, using the 29 days of "no activity" at hoth stations during the six months around opposition in 19^2

PAGE 190

171 also analyzed using these 29 days of no activity, but nothing significant vas noted. Chree Analysis of the l8 Mc/s Emission Originating in Each Source The probability histograms showed that a large amount of radiation is emitted from some Jovian longitudes, while very little radiation is emitted : om others. In choosing our days of peak activity, we did not weight the daily activity index rates for a given longitude according to the probability of receiving emission from that longitude. An approximation to this, however, is to perform Chree analyses for each region of Jupiter which contains a source. Consequently, we divided Jupiter into the 0° to 90° (no source), 90° to 190° (source B), 190° to 290° (source A), and 290° to 360° (source C) regions which we have already mentioned. (The Jupiter program had listed a daily activity index rate for each region.) For the 0° to 90° region we chose 11 days of peak activity (see Table 7), for the source B region we chose 17 days (see Table 7), for the source A region we chose I9 days (see Table 8), and for the source C region we chose 21 days (see Table 8). For each of these longitude regions we performed a Chree analysis of the 2800 Mc/s solar flux, Ap, R^, FXl, FX2, and FX3, using the previously discussed days of peak I8 Mc/s Jupiter activity monitored at the Florida station in I962. These analyses are shown for the 0° to 90° region, the 90° to 190° region, the 190° to 290° region, and the 290° to 360° region in Figs. 81 through 8^+, respectively. Kote that, as has already been mentioned, these Chree analyses were performed using two-day averages.

PAGE 191

172 TABLE 7 DAYS OF WO ACTIVITY AND DAYS OF PEAK ACTIVITY IN JOVIAN LONGITUDE REGIONS FROM 0° TO 90° AND 90° TO 190° FOR THE 18 Mc/s FLORIDA DATA IN I962 Days of no activity at either station, JuneNov., I8 Mc/s 1962 June

PAGE 192

173 TABLE 8 LAYS OF PEAK l8 Mc/s ACTIVITY IN JOVIAN LONGITUDE REGIONS FROM 190° to 290° AND 290° to 360° AND DAYS OF PEAK 27.6 Mc/s DATA IN I962 Days of peak activity recorded in the III Days of peak activity recorded in the region lying "bet-ween 190*3 and 290° (source A) III region lying between 290° and 360° (source C) Days of peak activity, 27.6 Mc/s Florida, 1962 June

PAGE 193

Yjk II PEAK DAYS 0-90"* 18.0 MC/S, FLORIDA 1962 -33 -29 -25 DAY NUMBER II -21 -17 -13 -S -5 -I I RELATION TO THE PEAK ^-3 DAY Fig. 81.— Chree analysis of the 2800 Mc/s solar flux, the geomagnetic index Ap, the sunspot number R^, and the solar flare activity indices for regions one, two, and three, using the 11 days of peak Jupiter activity received in the 0° to 90° A TTT interval on the l8 Mc/s, Florida channel in 1962

PAGE 194

175 17 PEAK DAYS. 90-190 18.0 MC/S, FLORIDA 1962 -(—I — j — i — I — r -3c -31 -27 -23 -19 -15 DAY NUfv:BER IM RELATION -il -7 -3 -l-l AE PiZAK DAY Fig. 82.--Chree analysis of the 2800 Mc/s solar flux^ the geomagrietic index Ap, the sunspot number E„, and the solar flare activity indices for regions one, two, and three., using the 17 peak days of Jupiter activity received in the 90° to 190*^^^^interval on the l8 Mc/s, Florida channel in I962 ^

PAGE 195

uoO 176 19 PEAK DAYS, 190-290* 18.0 MC/S. FLORIDA 1962 J * J * I • t -35 -31 -27 DAY NUMBER -23 -19 -15 -II -7 -3 IN RELATION TO THE PEAK -i-l DAY Fig, 83.— Chree analysis of the 2800 Mc/s solar flux, the geomagnetic index Ap, the sunspot numher E^, and the solar flare activity indices for regions one, two, and three ,-. using the 19 days of peak Jupiter activity received in the 190° to 290° A interval on the l8 Mc/s, Florida channel in 19o2 III

PAGE 196

177 21 PEAK DAYS, 290**360** 18.0 MC/S. FLORIDA 1962 1 — ! — T -35 -31 -27-23 -19 -15 -II -7 DAY NUMBER IN RELATION TO THE -3 +1 PEAK DAY Fig. 84. — Chree analysis of the 2£00 Mc/s solar flux, the geomagnetic index Ap, the sunspot number Ro.^ and the solar flare activity indices for regions one, two, and three, using the 21 days of peak Jupiter activity received in the 290° to 3^0° A-r-rx interval on the l8 Mc/s, Florida channel in I962

PAGE 197

178 The only really striking feature in Fig. 8l is the FXl peak at -29 days. FXl and FX3 both show peaks at -I7 days in Fig. 82. Fig. 83 yields little that can he construed as heing of significance, hut FXl and FX3 hoth show peaks at -23 days in Fig. dk. Therefore, it appears that radiation from each region of Jupiter is stimulated by solar particles of different velocity, or else the peaks that have just been mentioned are of no significance. The latter choice seems to be more reasonable. Double Correlation Studies At the suggestion of Dr. A. G. Smith, we made a multiple correlation study of some of the Chree analyses. We used the same days of peak Jovian activity which were used for the I8 Mc/s, Florida data in Figs. 68, 69, and 70. The correlation studies were made using the 2800 Mc/s flujx Chree analysis with the Chree analysis of FXl, FX2, and FX3 and using the Chree analysis of R^ with FXl, FX2, and FX3. In Figs. 85 and 86 we see the respective products of I28OO with FXl, FX2, and FX3 and of R^ with FXl, FX2, and FX3 where: 12800' = 12800 84.00 (32a) FXl' = FXl 5-00 (32b) FX2' = FX2 3.00 (32c) FX3' = FX3 lU.OO (32d) ^Z' " -^Z ~ 28.00 (32e) The values for I28OO, FXl, FX2, FX3, and R^ in Equations (32a) through (32e) were taken from the Chree analysis curves in Figs. 68-7O. The

PAGE 198

179 DOUBLE CORRELATION STUDIES 20 PEAK DAYS-13.0 ^y3C/S. FLORIDA 1962 Q K'f^^f , ' > -33 -29 -25 -21 -17 -13 -9 -5 -I DAY NUMBER IN RELATION TO THE PEAK -^3 DAY Fig. 85. — Double correlation studies of the 2800 Mc/s solar flux Chree analysis with the Chree analyses of the solar flare activity indices in regions one, two, and three, using the 20 days of peak Jupiter activity monitored at iB Mc/s by the Florida station in 1962

PAGE 199

l8o DOUBLE CORRELATION STUDIES 20 PEAK DAYS-! 8.0 yC/S, FLORIDA 1962 800fHUMBc -21 -17 -!3 -9 -5 -I -i-o IN RELATION TO THE PEAK DAY Fig. 86. — Double correlation studies of the geomagnetic index Chree analysis with the Chree analyses of the solar activity indices in regions one, two, and three, using the 20 days of peak Jupiter activity monitored at l8 Mc/s by the Florida station in I962

PAGE 200

I8i definitions in Equations (32a) through (32e) were chosen so that at some point on its respective Chree analysis curve, each primed variable took on a value close to zero. The double correlation curves in Fig, 85 all show about the same structL-.'e— a peak around -8 or -9 days, a broad general rise from -Ik to -25 days, and another smaller peak at -28 or -29 days. In Fig. 86 we see about the same curves that appeared in Fig. 85, except that in Fig. 86 the ratio of the height of the -8 peak to that of the large general peak is greater than it ms in Fig. 85. Also, in Fig. 86 we see that the -28 day peak is much weaker than it was in Fig. 85. As we have already suggested, the large general peak may be a manifestation of the 27-day solar rotational period (both the 28OO Mc/s solar flux and R^ vary with the sun's rotation), while the -8 day peak may be the sunto-Jupiter travel time of flare-produced particles. Chree Analysis of the 27.6 Mc/s Jumter Bnission Monitored at the Florida Station in Tg&g Up to now, all of our 1962 Chree analyses have been made using 18 Mc/s data. But the 1961, 27.6 Mc/s Ch^ee analysis of K[l showed a rather large peak at -2 days. In order to pursue this study further, we made a Chree analysis of FXl, FX2, FX3 (see Fig. 87), FNl, FTJ2, FN3, (see Fig. 88), 2800 Mc/s solar flux, R^, Ap, and the Jovian daily activity index rate (see Fig. 89), using the 19 days of peak Jovian activity at 27.6 Mc/s monitored by the Florida station during I962. The only outstanding peak on any of the curves in Figs. 87 and 88 is the -3 day peak on the FTT3 plot. In Fig. 89 we see peaks at -21 and -28

PAGE 201

182 19 PEAK DAYS-27.6 MC/S, FLORIDA 1962 I 1 i ^ — I — ! — I — I — I — r-T — s — ! — i — I — « — I — s — r 80 -33 -29 -25 -21 -17 -13 -9 -5 -I +3 DAY NUMBER IN RELATION TO THE PEAK DAY Fig. 87. — Chree analysis of the solar activity indices (FXl, FX2, and FX3), using the 19 days of peak 27.6 Mc/s Jupiter activity received at the Florida station in I962

PAGE 202

183 19 PEAK DAYS -27.6 MC/S. FLORIDA 19 62 i * I » I • I • 1 • 1 < 1 1 i 1 -33 -29 -25 -21 -17 -13 -9 -5 -I +3 DAY NUMBER IN RELATION TO THE PEAK DAY Fig. 88.— Chree analysis of the solar flare numbers (FNl, FN2, and FW3), using the 19 days of peak 27.6 Mc/s Jupiter activity received at the Florida station in 1962

PAGE 203

18U 19 PEAK DAYS2 7.6 MC/S . FLORIDA 1962 ? — i — 8 — r >

PAGE 204

185 days on the Ap plot_, and we see a peak at -7 days on the Jupiter activity plot. This latter peak is prohahly a manifestation of the 8-day period that we have seen in many of our studies. Additional studies were made hut are not included here. The "29 days of no activity" Cliree analysis plots were subtracted from the "three months around opposition" Chree analysis plots in an attempt to bring out additional information. But^ since the 29 days of no activity were chosen from the six months around opposition, such a difference may not have been significant, for the flare populations utilized by the two Chree analyses were different. Also, old program Chree analyses were made of peak days chosen according to daily activity index (^ S^ t;^) S'^^ according to the peak intensity i (S) on a given day, where S is the old program intensity, t is the time of a storm, and i is an index which runs over all the storms on one day. These last two studies were fruitless. Conclusions and Suggestions for Further Studies Of the preceding studies, those that were obtained from the days of peak Jovian activity received during the three months around opposition show the most significant structure. This is presumably the case for the Ap Chree analysis because the solar particles which reach Jupiter pass close enough to the earth to cause magnetic dist-urbance only at times around opposition. The solar flare Chree analyses are improved for "around opposition" analyses because the flares that were seen by observers on the earth, and hence were entered into our program, were very nearly the same that could be

PAGE 205

186 seen from Jupiter^ whereas this was not the case for flares recorded at times far removed from opposition. The Chree analysis using the solar flare deck minus the suhflare cards yielded no significant change. The douhle correlation studies show Chree analysis peaks at -8 days and at -20 days. These are each probably caused by a different phenomenon. In summary^ our analysis shows that: 1. Flare activity occurring in regions one and three peaks about eight days before we receive Jovian emission. Other peaks show up at about -l£, -24, and -32 days; i.e., the solar activity plots themselves show an eight-day period. Perhaps this eight-day period is just the recurrence of a few prominent solar flares in the Chree analysis caused by an eight -day recurrence in Jovian activity. But the eight-day cycle shown in the Chree analysis of the Jovian activity index rate is not nearly as evident as it is in FXl and FX3. Perhaps not one flare but a number spaced at eight-day intervals are needed to trigger Jovian radiation. In nature resonances nearly always produce more noticeable effects than do single events. A succession of flare events could, as in the forced harmonic oscillator, pump energy into solar particles in such a way that they reach Jupiter. An eight-day delay between flare and Jovian activity has been detected in earlier work by members of the University

PAGE 206

187 of Florida radio astronomy group, but to my knowledge no one has noticed this eight-day recurrence in flare activity itself. Studies will certainly be made to see if this shows up in future data. 2. The Chree analysis of the number of flares occurring in regions one and three shows about the same variation, though less pronounced, as is seen in the solar activity studies. 3. Studies of the flare activity and flare number in region two show that flares in this region are apparently not responsible for triggering Jovian emission. h. The s'onspot number studies and the 2800 Mc/s flux studies show that they each have a peak about eight days before we receive Jupiter activity. This peak is superposed on a larger peak that occurs at about 20 days before Jupiter is active. The larger peak is very broad and appears to be a manifestation of the 27-day solar rotational period. 5. The geomagnetic index (Ap) peaks about four or five days prior to Jupiter activity. This is probably caused by particles which have a 24to 30-hour sun-to-earth travel time. This travel time agrees quite well with that commonly attributed to particles which produce geomagnetic disturbances. 6. Jupiter activity itself seems to occur several days in succession. Some of our studies show that a group of

PAGE 207

188 active days occurs aloout every eighth day. All of the foregoing conclusions could not have been drawn from each of the analyses. The Chile analysis, for example, showed very few of the efi-cts mentioned ahove. For this reason, these conclusions should not be regarded as dogmatic statements of fact. They are, however, indications which encourage one to continue this search for correlation of solar activity with emission from our giant neighbor. Our analyses possess inherent weaknesses, some of which can be eliminated and some of which we must simply put up with. Listed below are studies which might circumvent some of the removable weaknesses: 1. Pick the days of peak Jovian activity according to a weighted daily activity index rate; i.e., one in which the daily activity index rate is multiplied by the inverse of the probability of receiving Jovian emission from those Jovian longitudes. This would correct any discrepancies introduced by the fact that, since on the average watch we observe less than half of the planet, we observe on different dates longitudes which radiate with different probabilities. 2. Assume that the particles responsible for Jovian emission travel along an Archimedes spiral such as the one shown in Fig. 79. The velocity of the solar wind producing the interplanetary magnetic field determines the

PAGE 208

189 tightness of this spiral. (The particles which induce Jovian emission need not be the same as those which produce the interplanetary magnetic field.) For a given solar wind velocity^ hence a spiral of given curvature, there is only one solar longitude from which particles may originate such that they travel along the spiral and reach Jupiter. We could then perform a Chree analysis using flares lying only at these solar longitudes. Such Chree analyses could be made for spirals of different tightness; i.e., spirals caused by different solar wind velocities. The analysis showing the most structure would indicate the probable particle path, hence the spiral shape. Since the spiral shape depends on the solar wind velocity, we have an indirect method _for determining the solar wind velocity. One problem exists. According to calculations by Parker (28), these spirals can be interrupted by a solar blast wave which propogates radially outward. The azimuthal component of the interplanetary magnetic field is about four times larger in regions near the shock front than it is in the undisturbed regions. The shock front, according to P&rker, forms a wall that is impervious to extragalactic cosmic rays, in this manner accounting for the occurrence of Forbusch decreases. Since this shock front is a discontinuity in the interplanetary magnetic field, a particle following the field would, upon

PAGE 209

190 encountering the discontinuity, have a rather illdefined path. This has "been' recently discussed by Parker (29). This would make our proposed study more difficult to execute. Perhaps the increased presence of these shock fronts around the time of sunspot maximum impedes or eliminates from the stream particles which stimulate activity on Jupiter. If sO;, this accounts for our inverse long-time correlation of Jupiter activity with the sunspot cycle. This also could acco;int for the fact that we see a gap in solar activity preceding Jovian emission. 3. Some of the Chree analyses showed a broad peak at about -20 days and a dip at days. If this dip is really connected with Jovian emission, the peak should reappear at +20days. For this reason, an extension of the Chree analysis to +35 days might be useful. k. We have already discussed the manner in which we chose the flare importance number and the flare duration, and we have mentioned a case where the flare activity index seemed to be excessive upon comparison with other data listed in H.B.S. Geophysical lata . Part B. Perhaps an average of the reports in the data should be made before flare importance numbers and diirations are entered into the program. 5. Correlation has been found between the solar wind

PAGE 210

191 velocity and the geomagnetic index^ the 27-day solar rotational period ;, and M-region activity (29), This leads us to "believe that the activity of M-region magnetic storms should also he included in our Chree analysis.

PAGE 211

CHAPTER VII DECAMETER-WAVELENGTH OBSERVATIONS OF SATURN, MARS, AND VENUS IN I962 AND I963 SATURN Introduction and Instrumentation Since Saturn can essentially be regarded as Jupiter's twin, it is not unreasonable that we should suspect that it, too, emits radio noise at decameter wavelengths. This possibility inspired the University of Florida group to begin observations of Saturn in 1957In i960 they observed several bursts which seemed to be Saturnian in origin, but doubt as to this being their actual origin was cast by the fact that, since Jupiter and Saturn were separated only one hour in right ascension, Jupiter was also in the antenna beam. In 1961 the two planets moved through conjunction, and in 1962 they were again separated by about one hour in right ascension. Therefore, for the reason stated above, possible Saturnian radiation was again very difficult to distinguish from that originating on Jupiter. Nonetheless, observations were carried on. The antenna designated l8S at the Florida station and that designated 18E at the Chile station were used primarily for these observations. The characteristics of these antennas and their respective receivers are listed in Table 1. The l8S antenna was used during the entire Florida observing season (June 9 "to November 6), while iBW 192

PAGE 212

193 was used to observe Saturn for the first 13 days of the Chile observing season (April 1 to November 2o) , after which 18E was used. The Chile season was longer because Saturn was at about -l6° declination during most of the 1962 season. Data Reduction and Analysis The data were reduced with three ends in mind: 1. To determine the number of hours that Saturn was monitored and the listening conditions that prevailed during these times. 2. To determine the flux density of each pulse suspected to be Saturnian in origin. 3. To determine if any period (possibly close to the 10^ 14m equatorial period) can be found in the times of these "possible Saturn" pulses. The observer on duty was responsible for designating the quality of the listening conditions each half hour diiring the watch. These designations were made using the nimbers zero to five^ where low numbers represented poor listening conditions and high numbers represented relatively good conditions. Using this information^ we tabulated the number of respective listening hours for times when the prevailing listening conditions were designated as three;, four, and five. Listening conditions were reckoned to be good only if Saturn was at lease one hour above the horizon and in the antenna beam, and only if the receiver gain was such that the pen deflection of the cosmic background noise was at least one-fourth of full scale on the Texas

PAGE 213

19^ recorder. The total listening times satisfying these criteria are shown in Table 9TABLE 9 GROSS STATISTICS OF SATUM OBSERVATIONS IE I962 Frequency Listening time in hoirrs Miniraura detectable signal Quality 3 Quality h Qimlity 5 listening listening listening conditions conditions conditions Total i8E (Chile) 18.0 IBs (Florida) 10.0 12.5 35-5 492 .5 ^92 . 5 10 -22 65.0 111.0 2 X 10 -22 In Table 10 we see information pertaining to each pulse that could possibly have originated on Saturn. We see that foirr of the six "possible Saturn" pulses recorded on l8S occurred at western hour angles. At these hour angles l8S possessed higher gain than did 18Y; thus; even though l8S was not directed exactly toward Jupiter, Jovian pulses that were received when Jupiter was near the western horizon were often more pronounced on l8S than they were on the Jupiter channel (i8y). Since l8S and 18Y were both five-element steerable yagis, why did we observe this difference in gain? The answer probably lies either in the fact that, located just west of i8Y was a corner reflector which could have screened signals originating in the west, or in the fact that i8Y was mounted on a shorter pole than was l8S, thereby suffering more undesirable ground plane effects. At any rate,since

PAGE 214

195 CM VD ON iH M n K O o o

PAGE 215

196 the gain of l8S was greater than that of 18Y when they were pointed at western hour angles, pulses which originated in the west and which were labelled "possible Saturn" may actually have originated on Jupiter. Note, also that we never received simultaneous "possible Saturn" pulses at both stations. In order to check the hypothesis that all of our "possible Saturn" pulses were really Jovian in origin, we calculated the Jovian System III longitude that was on the central meridian at the time of each pulse. All but three fell in the region of some source, but none occurred at the center of a source, nor did any occur in source A. Since none of these pulses came from Jupiter's major source, and since they were all received with greater amplitudes on the Saturn radiometers, we will continue oiir study on the assumption that these pulses did indeed originate on Saturn. Assume that Saturn, like Jupiter, radiates l8 Mc/s signals primarily from one portion of the planet. Can we find any period close to the 10^ 1^"^ rotational period of Saturn's equatorial region which will maintain constant the Saturnian longitudes of the pulses listed in Table 10? We shall refer to the longitude system on Saturn as Ag. A of each of the bursts listed _in Table 10 was computed for S rotational periods of 10^ if, 10^ 20^ 10^ 2^'' 11 Oo"". A plot of A. versus tLme from May 10, 1962, was drawn for each period. We will assimie that the zero longitude of our Saturnian longitude system faced the earth at 0^ U.T. on May 10, I962. The plot for the 10^ kf period showed that, with one exception, the longitude of each

PAGE 216

197 burst lay approximately along a straight line. When this period was changed to 10 4$ .G, all but one of the pulses fell in the 295°
PAGE 217

198 SATURN 1962 1 J—, , r—

PAGE 218

199 From January 11, 19o3^ "to March 23, 1963, we monitored both Ifers and Venus from the Florida station. During this time;i the earthto-Venus distance varied from O.58 to 1.11 astronomical units, while the earth-to-Mars distance varied from 0-73 "to O.89 astronomical units. In Tahle 11 we have listed information pertaining to these observations. Listening conditions were considered to he effective only if they were labelled "five" by the observer. In 1956, Kraus (32) reported that he had observed 11-meter radio emission from the planet Venus but later, in I96O, he reported that what he had thought to be Venus ian radiation may have been terrestrial interference (33)The University o'f Florida group (l4) conducted observations of Venus at its morning elongations in 1958 and i960. (Venus was also at a morning elongation during the I963 studies described here.) These earlier observations detected nothing that • could have been construed as Venusian radiation; thus it is not surprising that we received no radiation from the planet early in 1963. Notice that Table 11 shows that we monitored tJars under acceptable conditions for many more hours than we did Venus. This is understandable, for Mars was transiting the meridian between midnight and sunrise, during the time of good listening conditions. Therefore, it vas not unusual for one to experience four or five hours of near-perfect listening. Still we detected no radio emission from Mars. It seems unlikely that any definite decametric radiation will be detected from these planets unless an interferometer is used, since, at these wavelengths, only an interferometer has sufficient directivity to identify the weak radiation from these sources if they radiate

PAGE 219

200 TABLE 11 GROSS STATISTICS FOR THE MARS Affl) VENUS OBSERVATIONS IB I962 MARS

PAGE 220

201 at all. It should "be mentioned that Mars has already been monitored by a 22.2 Mc/s interferometer at the Florida station with negative results.

PAGE 221

coraCHAPTER Till SlMaRY AKD RKvii5-RKS OK THE COI'IPARISOE OF THE EXPERIMENTAL RESULTS WITH E^ISTIBlG THEORIES THEORIES AS TO THE ORIGIN OF JUPITER'S DECAMETRIC RADIATION It is interesting to note that since the discovery of the strong^ sporadic decametric radio noise from the planet Jupiter "by Burke and Franklin (l) in 1955. no less than 15 theories (not all plete) have heen advanced as an explanation of its origin. These the ories are listed below: 1. Shain (3^) 195^ spark discharges 2. Zhelezniakov (35) 1958 plasma oscillations set up hy turbulence in Jupiter's atmosphere 3. Gardner and Shain (36) 1958 ) plasma oscillations ) set up by volcanic h. C-allet (37) 1958 ) shock waves 5. Carr (38) 1959 cyclotron radiation of solar electrons spiraling in Jupiter's magnetic field 6. Sagan and Miller (39) 19^0 chemical explosions of acetylene in the Jovian atmosphere 7. Field (-UO) i960 spark discharges set off by high voltages induced by the slipping of Jupiter's magnetic field through its atmosphere 8. Strom (3), (^1) 1962 focusing of distant radio stars by Jupiter's ionosphere 202

PAGE 222

203 9Smith, A, G. (ii2) 19o2 cyclotron radiation frora electron bundles oscillating between mirror points in Jupiter's magnetic field 10. Warwick (k) 1962 reflected Cerenkov radiation caused by high energy particles which originate two to three Jupiter radii above Jupiter and spiral in toi-/ard the surface . 11. Field (5) 1902 -via-ve amplification of pole\-;ard-directed waves by equatorward-directed electrons 12. Six (10) 1962 cyclotron radiation caused by coherent motion of a group of electrons with emission at the local gyrofrequency 13. Carr (2) 19o2 guiding of auroral zone-produced radiation by field-aligned ionic ducts 1^1. Ellis (6) 1962 doppler shifted cyclotron radiation from electrons spiraling in Jupiter's magnetic field 15. Chang (7) 1963 amplified whistler radiation The majority of the above named authors does not claim to have complete theories. Most of the theories were put forth simply to provide a possible mechanism which could give rise to decameter radiation. As more experimental data were collected, however, attempts were made to refine the theories. Of the above-mentioned theories, only those put forth by Warwick, Field (5), Carr (2), and Ellis make an attempt to explain the following characteristics: the radiation mechanism, the longitude dependence, the spectral characteristics, and the polarization. Of these characteristics, probably the most difficult

PAGE 223

204 to explain is the longitude dependence. Neither a dipole nor a quadrupole magnetic field can readily explain three sources. We will point out that each of the theories hy Warwick, Field, Carr, and Ellis falls do-tm at some point when compared with the data. SUMVLAEY The 1962 Jupiter observations hy the University of Florida radio astronomy group, along with those of the years 1957-1961, iiave provided a profusion of information as to the nature of the Jovian decameter radiation. The analyses which preceded thisone yielded the following conclusions: 1. Radio noise emanates from three distinct longitude regions (sources A, B, and C). Field's theory (5) fails to acco'ant for this. Source B shows a definite bifurcation and is sometimes referred to as sources B-j_ and Bg. The longitude referred to is the System III longitude system, which rotates with a period of 9 55 29.35. 2. The position of these sources shifts to higher longitudes with decreasing frequency. 3. The sources are better defined at the higher frequencies. h. Decameter radiation from Jupiter cuts off not far above a frequency of 30 Mc/s. 5. The ratio of the magnitude of source A to that of source C increases with frequency.

PAGE 224

205 6. Jovian activity occurs in eight-day cycles. (This vas not evident in all of the data.) 7Jovian activity increases with elongation. 8. Jovian activity dips when Jupiter is near opposition. 9. The average probability^ the' source width, and the yearly activity index rate all show inverse correlations with the sunspot cycle. 10. Ho definite short-term correlation of the solar flare index and Jupiter activity had been found, but some three, four, and five day delays had been found between geomagnetic activity (Ap) and Jovian noise storms. 11. Warwick (k) found that the center frequency of a Jovian noise storm changes with time. The center frequency of noise storms originating at source B longitudes drifts up the spectrum while the center frequency of noise storms originating at source A longitudes drifts to lower frequencies. Carr's theory (2), however, predicts that positive drifts should be seen for storms occurring on the low longitude side of each source and that negative drifts should be seen on the high longitude sides. Nearly all of the above characteristics were verified by the 1962 analysis, but there were some exceptions. There are only two sources at 5 Mc/s. This bears out a prediction by Ellis (6). The positions of the sources were found to change drastically at frequencies below 15 Mc/s. It is the opinion of the author that this change

PAGE 225

206 is really a shift tc lover longitudes with decreasing frequency below 15 Mc/s, although this is not established. Field predicts that so\irce C should increase in magnitude with decreasing frequency and that it is left-hand elliptically polarized. If this is the case, the problem of identifying source C at 5 Mc/s should easily be settled by data from the 5Mc/spolarimeter now being built at the University of Florida. Another of the characteristics shown in the 1961, but not the 19^2, data was the dip of Jovian activity around opposition. The 1962 analysis dealt with both the probability of emission and the average flux density. Recall that from the probability studies in Chapter IV, we found the 5 Mc/s histogram to exhibit only two sources. Warwick (4) fails to acco\mt for this change in source structure with frequency. We also found that source B does not drift in the manner predicted by Warwick. The most exciting new discovery was that the rotational period of Jupiter seems to have changed in I96O. The author has proposed that what was actually observed was an apparent gradual change of source position brought on by changing conditions in the interplanetary medium or in Jupiter's magnet osphere, due to the variation of solar activity during the sunspot cycle. This argument was weakened considerably by the fact that the rotational period of Jupiter's Great Red Spot was also found to change, and by the fact that the onset of this gradml change seems to have appeared abruptly. Presumably, interplanetary or magnetospheric changes should not affect electromagnetic radiation at optical wavelengths. Hence, the mystery remains.

PAGE 226

207 The probability of observing Jupiter radiation was also examined as a function of ho'or angle^ Universal Tixc^e, time from sunrise, and tirae from sunset. We found the probability to peak at eastern hour angles and at hours between midnight and sunrise, and we saw the effects of the beam patterns of the fixed arrays. The probability, as is to be expected, does not depend appreciably on the Universal Time. Flux density histograms were constructed. Two averages vere calculated: average over listening time (S^), and average over activity time {Sa) • % was found to peak at longitudes attributed to source B in the probability studies. This contradicts Ellis' prediction that, at a given longitude, the intensity is proportional to the 1.3 power of the number of bursts. The hour angle studies of S^ showed that the most intense bursts occurred at western hour angles, a phenomenon for which the author has no explanation. The S^ histograms showed the same characteristics that could have been obtained by taking the product of the probability histograms and the S^ histograms. The solar flare studies were continued, and a new study was made of possible correlation of the 28OO Mc/s solar flux with Jovian emission. An error in the 1961 analysis was corrected, but nothing more significant was found after the change. The 1962 data, hov/ever, showed a definite eight-day delay between flare activity in the central and western portions of the solar disk (as seen from Jupiter) and Jovian activity. This eight-day delay was also seen in the 2800 Mc/s studies and, to some extent, in the sunspot studies. A five-day delay was found between the times of geomagnetic and Jovian activity.

PAGE 227

208 In addition to this^ a twenty-day delay was seen for most of the parameters. This last is presumahly a manisfe station of the solar rotational period. These effects were exhibited raost noticeahly in the studies conducted for activity received around opposition. Unfortunately; the validity of these conclusicns is weakened by the fact that few of these features showed up in studies using the Chile data. The author has suggested some sampling procedures that should improve these results. Of the theories that have been proposed^ most include good features. The author believes that there is a good possibility that the focusing mechanism is that described by Carr (2). Gulkis (13) has suggested that a combination of the theories proposed by Carr and Ellis (6) may lead to an accurate explanation of the observed decametric radiation from Jupiter. He proposes that doppler shifted cyclotron radiation is the mechanism by which the radiation is created and that it is focused to\^ra.rd the earth by field aligned ionic ducts. This appears to be very promising. In 1962 we detected several 18 Mc/s bursts that could have originated on Saturn. Using the assumption that Saturn radiates primarily from a single longitude^ we calculated a rotational period of 10''^ 4/'^.6. This is close to the 10 l4 period commonly attributed to the equatorial regions of Satiirn;, but differs quite markedly from 11^ 57"^. 8 period derived by Carr et. al . (15) from the I96O data. In light of this difference;, these conclusions should not be granted much significance. The 19o2 observations of Venus and JVIars were

PAGE 228

209 reviewed. It was concluded that neither Venus nor Mars emitted de_92 -2 -T caiueter radiation stronger than 2 x 10 " w m cps ^. We have shown that, unlike the case of Jupiter's decimetric radiation^ much work has yet to be done "before we will be able to explain completely the mechanism responsible for Jupiter's decametric emission. None of the theories seems to explain all of the observations. Therefore, it is imperative that we continue observations of our giant neighbor, Jupiter.

PAGE 229

LIST OF REFERENCES 1. Burke^ B. F. and Franklin, K. L. , "Observations of a Variable Radio Source Associated with the Planet Jupiter," Journal of Geophysical Research 6o_, 213 (1955). 2. Carr, T. D., "The Possible Role of Field-Aligned Ducts in the Escape of Decameter Radiation from Jupiter," proceedings of the discussion meeting relative to the planet Jupiter, held at MSA Goddard Institute for Space Studies, New York, October, 1962 (in press). 3. Strom, S. E. and K. M. , "A Possible Mechanism for Jovian Decameter Bursts," Astronomical Journal 67, 121 (1962). k. Warwick, J. W., "Dynamic Spectra of Jupiter's Decametric Emission, 1961," Astrophysical Journal 137, Ul (1963). 5. Field, G. B. , "Jupiter's Radio Emission," preprint of paper to be included in the proceedings of the discussion meeting relative to the planet Jupiter, held at HASA Goddard Institute for Space Studies, New York, October, 1962 (in press). 6. Ellis, G. R. A., and McCulloch, P. M. , "The Decametric Radio Emissions of Jupiter," Australian Joiarnal of Physics 16, 379 (1963). 7. Chang, D. B., "Amplified Whistlers as the source of Jupiter's Sporatic Decameter Radiation," Astrophysical Journal 138 , 1231 (1963). 8. Carr, T. D., Studies of Radio Frequency Radiations from the Planets (Ph.D. Dissertation, University of Florida, I958). 9. Chatterton, N. E., Spectral Characteristics of the Radio-Frequency Outbursts of the Planet Jupiter (Ph.D. Dissertation, University of Florida, I96I). 10. Six, N. F,, Jr., Analysis of the Decameter-Wavelength Radio Emis sion from the Planet Jupiter (Ph.D. Dissertation, University of Florida, I963}. 11. Brown, G. W. , Statistical Studies of the Decameter Radiation from Jupiter in I96I (Master's thesis. University of Florida, 1963). 210

PAGE 230

211 LIST OF REFERENCES (Continued) 12. Carr, T. D. , Brown, G. W. , Smith, A. G. , Higgens, C. S., Bollhagen, H., May, J., and Levy, J,, "Spectral Distribution of Decametric Radiation from Jupiter in 1961" ( Astrophysical Journal, in press). 13. Private communication with Sam Gulkis. lU. Barrow, C. H. , "Recent Radio Observations of Jupiter at Florida State University," paper delivered to the Physical Sciences Section of the twenty-eighth annual meeting of the Florida Academy of Sciences, Winter Park, Florida, March, 1962. 15. Carr, T. D. , Smith, A. G., Bollliagen, H., Six, N. F. , Jr., and Chatterton, N. E., "Recent Decameter -Wave-Length Observations Jupiter, Saturn, and Venus," Astrophysical Journal 13U, 105 (1961). 16. Douglas, J. W. , "A Uniform Statistical Analysis of Jovian Decameter Radiation, I95O-60," Astronomical Journal 65, J+87 (196O, abstract). 17. Carr, T. D. , "Guidance of Jupiter's Decameter Radiation by the Planetary Magnetic Field," Transactions American Geophysical Union kk. 92 (1963). . — 18. Carr, T. D. , Smith, A. G., Pepple, R., and Barrow, C. H. , "I8Megacycle Observations of Jupiter in 1957," Astrophysical Journal 127, 27^+ (1958). 19. Peek, B. M. , The Planet Jupiter (Macmillan Company, New York, I958), p. l46. 20. Smith, A. G. and Carr, T. D. , "Radio-Frequency Observations of the Planets in I957-I958," Astrophysical Journal 130, 6hl (1959). 21. Private communication with T. D. Carr. 22. Dowden, R. L., "Polarization Measurements of Jupiter Radio Bursts at 10.1 Mc/s," Australian Journal of Physics 16, 398 (1963). 23. Carr, T. D, , Smith, A. G., and Bollhagen, H. , "Evidence for the Solar Corpuscular Origin of the Decameter -Wavelength Radiation from Jupiter/' Physical Review Letters ^, kid (1960). 24. Plourde, A. J., Statistical Investigation of the Occurrence of Koise Emission from the Planet Jupiter (M. S. Thesis. University of Florida, 196O}. '

PAGE 231

212 LIST OF REFERENCES (Continued) 25. Smith, A. G., Carr, T. D. , SiX;, N. F., Jr., "Results of Recent Decameter-Wavelength Observations of Jupiter/' La Physique Pes Pla net es (institute d'Astrophysique, Cointe-Sclessin, Belgique, 26. Birker, E. K., "The Solar Wind/' Scientific American 210 , 66 (April, I96U). 27. McCracken, K. G., "The Cosmic-Ray Flare Effect," Joiurnal of Geophysical Research 67, kh^ (1962). 28. Parker, E. W., Interplanetary Dynamical Processes (interscience Publishers, University of Rochester, Rochester, New York, 1963)* p. 131. 29. I^rker, E. W., "The Scattering of Charged Particles by Magnetic Irregularities," Journal of Geophysical Research 69^ 1755 {'i-96h) . 30. o.-.yder, C. W. , Neugebauer, M. , and Rao, U. R., "The Solar Wind Velocity and Its Correlation with Cosmic-Ray Variations and with Solar and Geomagnetic Activity," Journal of Geophysical Research 68, 6361 (1963). 31. Smith, A. G., and Carr, T. D., Radio Exploration of the Planetary System (D. Van Nostrand Co., Inc., 196^), P. 52. 32. Kraus, J. D. , "Impulsive Radio Signals from the Planet Venus," Nature Y[d, 33 (1956). 33. , "Apparent Radio Radiation at 11-m. Wavelength from Venus," Nature I86 , h62 (i960). 3i+. Shain, C. A., "I8.3 Mc/s Radiation from Jupiter," Australian Journal of Physics 9^6l (1956). 35. Zhelezniakov, V. V., "On the Theory of the Sporadic Radio Emission from Jupiter," Soviet Astronomy -Astronomical Journal 2_, 206 (1958). 36. Gardner, E. F. and Shain, C. A., "Further Observations of Radio Emission from he Planet Jupiter," Australian Journal of Physics 11; 55 (1958). 37. Smith, A. G., "Radio Spectrum of Jupiter," Science 13^ , 58? (September, 1961). 38. Carr, T. D., "Radiofrequency Emission from the Planet Jupiter," report before the December I958 meeting of the American Astronomical Society, Astronomical Journal 64, 39 (1959^ abstract).

PAGE 232

213 LIST OF REFERENCES (Continued) ^^' ^^S;.'?* "-"£ ^'J^^^^S. L., "Molecular Synthesis in Simulated Reducing Planetary Atmospheres," Astronomical Journal 65 499 (I90O, abstract). ~^' ^^ 40. Field G. B "Source of Radiation from Jupiter at Decimeter Wavelengths: 2 Cyclotron Radiation by Trapped Electrons/' Journal of Geophysical Research 65,1661 (i960). 41. Strom, S E and K. M. , "A Possible Explanation for Jovian Decameter i3ursts, Astrophysical Journal 136, 307 (1962). 42. Sraith A. G., Carr, T. D., and Six, n. F., Jr., "Non-Thermal Radiation from Jupiter in the Decameter Wavelength Range," Proceedmgs of tne Third S.ymposiu in on Engineering Aspects o f Mag netohy' arod.ynamics (Gordon and Br P«..h_ I... v^^v (in pre -c) ' ' •

PAGE 233

BIOGRAPHICAL SIffiTCH George Robert Lebo was born September ZJ , 1931, in Chadron, Nebraska. He attended various grade schools in Nebraska, South Dakota, and Colorado before moving to Wheaton, Illinois, where he \i3.s graduated from the community high school in 1955He remained in Wheaton to attend Wheat on College, a small liberal arts school, where he received a Bachelor of Science Degree with a ma: or in physics and a minor in mathematics in 1959During the sumrr.er following his junior and senior years he was employed as a student aide at Argonne National Laboratory in Lemont, Illinois. He then moved to Champaign, Illinois, where he attended the University of Illinois, receiving a Masters of Science Degree with a major in physics and a minor in mathematics in I96O. After receiving his M. S. he moved to Gainesville, Florida, where he attended the University of Florida. In 1964 he completed the requirements for the Doctor of Philosophy Degree with a major in physics and a minor in mathematics. He is a char't^r member of the Wheaton chapter of Sigma Pi Sigraa. 21k

PAGE 234

This dissertation was prepared under the direction of the chairman of the candidate's supervisory committee and has heen approved by all memhers of that committee. It was submitted to the Dean of the College of Arts ana Sciences and to the Graduate Council, and was approved as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August, 1964 Dean, CoilDean, Graduate School Chairman

PAGE 235

A 9^ zilQ


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 E8C2BQXL4_S8PJGR INGEST_TIME 2017-07-19T20:43:36Z PACKAGE UF00097933_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES