VOL. 79, NO. 16 JOURNAL OF GEOPHYSICAL RESEARCH JUNE I, 1974
VLF Wave Injection Into the MagnetosphereFrom Siple Station, Antarctica
R. A. HELLIWELL AND J.P. KATSUFRAKIS
RadioscienceLaboratory, Stanford University,Stanford, California 94305
Radio signalsin the 1.5- to 16-kHz range transmittedfrom Siple Station, Antarctica (L = 4), are used to control wave-particleinteractions in the magnetosphere.Observations at the conjugatepoint show signalgrowth and triggeredemissions including risers, failers, and hooks. Growth ratesof the order of 100 dB/s and total gainsup to 30 dB are observed.Triggered emissions may be cut off or havetheir fre- quencyrate of changealtered by Siplepulses. Similar effectsoccur in associationwith harmonicsfrom the Canadianpower system. The observedtemporal growth is predictedby a recentlyproposed feedback model of cyclotroninteraction. Plasma instability theories predicting only spatialgrowth are not in ac- cord with these observations.Possible applications include study of nonlinear plasma instability phenomena,diagnostic measurements of energeticparticles trapped in the magnetosphere,modification of the ionosphere through control of precipitation, and VLF communication through the magnetosphere.
This is a report of the first resultsof a new experimentto plasma instability. To obtain further experimental data, a control wave-particleinteractions in the magnetosphere,us- transmitter was neededthat could operate at lower frequen- ing VLF waves.Coherent whistlermode signalswere injected cies and whose power, pulse length, and frequency could into the magnetospherein the 1.5- to 16.0-kHz range from a easily be varied. 100-kW transmitter located at Siple Station, Antarctica. The The next step was constructionof a 33.6-km-long horizon- output signalswere observedat the Sipleconjugate point near tal dipole at Byrd Station (L = 7.25) in 1965 [Guy, 1966]. This Roberval, Quebec. The signalsfollowed field-aligned ducts antenna was shared by a VLF step frequency sounder near L = 4 (seesketch in Figure la). Among the new resultsis operated by the University of Washington and a VLF the observationthat the signal receivedat Roberval generally magnetospheresounder operated by StanfordUniversity. The showsexponential growth as a function of time, with growth dipole was laid on the ice to lower its Q and henceincrease its rates of the order of 100 dB/s. Total power gains of up to 3 bandwidth. Soundings of the D and E regions were orders of magnitude (30 dB) have been observed.Following successfullycarried out over the freq'uencyrange 3-30 kHz the exponentialgrowth phase,emissions at new frequencies [Helmset al., 1968]. Whistler modetransmissions in the range are usually triggered, in the form of narrow band rising or 6-9 kHz were observed by the Stanford University VLF falling tones.A by-productof theseinvestigations is the find- receiveron the Ogo 4 satelliteas it flew over the transmitter in ing that VLF radiation at harmonics of the commercial the upper part of the F region. However, the signalswere too power systemappears to be present in the magnetosphere, weak to excite detectablewhistler mode signalsin the con- modifying the spectrum of the triggered emissions. jugate hemisphere.One reason was the low efficiencyof the In the following sectionof this paper we review briefly the low Q antenna and another was its high latitude. This high previotisexperimental work on VLF waveinjection into the latitude meant that signals from the transmitter would first magnetosphere.In the next three sectionswe describethe have to travel about 1000 km in the 1ossyearth ionosphere sitingof Siple Station, the apparatus,and someof the first ex- wave guide to reach the field lines on which propagation to perimental results.In the last sectionwe discussthese results the conjugatehemisphere was commonly observed. and some possibleapplications. SIPLE STATION PREVIOUS WORK From the Byrd experimentsit was clear that the transmitter Man-made whistler mode signals were first detected at would haveto be movedto a lower geomagneticlatitude. The Cape Horn from Station NSS on 15.5 kHz, located at An- desired site characteristics were (1) proximity to the napolis, Maryland [Helliwell and Gehrels, 1958]. Later dis- plasmapause(L • 4), (2) a stable ice sheet at least 2000 m crete VLF emissionswere observedto be triggeredby Morse thick, and (3) a conjugatepoint readily accessibleto year- code transmissions from Stations NPG on 18.6 kHz and round manned operation. A site satisfyingthese criteria was NAA on 14.7kHz [Helliwell et al., 1964].Nearly all triggering found at 76øSand 84øW (L = 4.1) and was occupiedin the was causedby the Morse dashes.When the dots did trigger, austral summerof 1969. It was named Siple Station; its con- the resultingemissions were alwaysweak falling tones [Lasch, jugate point is near Roberval, Quebec(48øN, 73øW). 1969]. Emissiontriggering by the long (• 1 s) pulsesfrom a With the aid of impedancemeasurements made on a test relatively low power (• 100 W) Omega transmitter operating dipole erectedabove the snow during the 1969-1970 austral on 10.2 kHz was also observed[Kimura, 1968]. Observations summer a 21.2-km-long operational antenna was designed during a drifting duct event showed that triggering was [Raghuramet al., 1974]. It was constructedduring the austral enhancedwhen the transmitter frequencywas near one half summers of 1970-1971 and 1971-1972. The measured reso- the equatorialgyrofrequency [Carpenter, 1968]. Theseresults nant frequency is 5.1 kHz, close to half the minimum could not be interpreted in terms of existing theories of gyrofrequency(•6 kHz) on the field line through Siple Sta- tion. The bandwidthis about 2 kHz. The radiationefficiency Copyright ¸ 1974 by the American GeophysicalUnion. at 6 kHz is estimatedto be 4%. The averageelevation of the 2511 2512 HELLIWELLAND KATSUFRAKIS.' BRIEF REPORT
(a) (b)
WWV PROGRAMMER RECE I V ER AND •LOOP CLOCK il TAPE PREAMP RECORDER
RECEIVER
ß
2O
(c) GEOMAGNETIC EQUATOR TRANSMITTER
ELEVATED DI POLE 2O 17000krn 21.2 krn _ I
I00kW COMPUTER TELETYPE 6O AMP'ß i KEYER FREQUENCY I ! STANDARD Fig.1. (a)Sketch ofmagnetospheric pathfrom Siple Station (SI), Antarctica, toRoberval, Quebec (RO); HB is Halley Bay.(b) RobervalVLF receiverblock diagram. (c) SipleVLF transmitterblock diagram. antennaabove the snowsurface was initially 6 m, with passiveexperiments were also installed,including whistler provisionfor periodicallyraising the antenna to compensateand VLF emissionrecordings (Stanford University), magnetic for an annual snow accumulation of 1-2 m. field intensity(Bell Laboratories),magnetic pulsations A year-roundmanned facility, including the 100-kW high- (Universityof Minnesotaand University of NewHampshire), efficiencysolid state VLF transmitterfrom Byrd Station, was and ionosphericabsorption (University of SanDiego). constructedduring the australsummer of 1972-1973.Various Operationof the passiveexperiments began in early1973. A RO 23 JUNE 73t
o.o
'::;;;:;::;?•.•"•••••••••:••;...•.•:•:•:::•.;•..::•:•::.•.:•;•:•.•:•.•...•:•:•:):•5:•?.•+•+?•"•'"•'•••'•'••••••'•'"" ":'::•:-'•[.:•".•:-•?•:•.•:•::-::•-•:•:--':':::•:•-...... :½..•:...... :...... •;•::•;•...... •:•:::•-:•;;•:'•-•¾ ...... - ...... : .+.•:"•cy--•-•-...;.: .-.:•
' ' ß...... -- *:'""-•-•..... g. .*--',•...•.:'•:•h.-•.-- .. .."•,.• ..... ::.'•.':.•:• ....:':..'.:2' ..... --• .... -" .• •: X":..Y'•'..::'?• .•? "•:':'":•.?•75...:•:•?•.•*:',.' •'5•;.'•- ..•'•'A • '•:•.4:•;•. '-'- -"' ' :'•:...... '::';;;•¾}:'•'•;';•5 •x:,,..,..:'•.....'>. :"•"5.,•::•.•,•:>•...... '::" ' ...... +•'•'' •."•,...• ...... "'"• ' ' '...... •"".g•'" ...... :.'.:•''...,•:;;2"::•' :':' ..... :%" . :, ..':' •?•...:;•.:•'"'";":'"':"'":'" ...... '"'"'"'":::':' <:. ""'•'"'"'"'""':""::'-- •. . .-...... :':'"..... •.:.•" ..... :•:.... :• '::•' '•'...... '...... :.. "'-:-•- ":•':•':,-:-.-. "':':"::':::':":'":"...... -...... •,'•..:.....'•.... ::,::•.•...... ::•"• .. :,':...... ß...... -.:. :.'. .... :...... L7•:'::..'...... '. ::::..•..... ::•:":•...:':...:...' :: .--•::."-:'..:,:' ß ..':.':'.:'.:"•:5-'--... :'•..'..-•-..:..::.' ...... ::::. '..:....:• I ...... t...... :• ...... i...... •...... i ...... i ':i" •:•...... :• '"'"I ...... i .....• ...... :• '• ! ..... i.... i i I ...... ! 0 -'10 •O 30 sec Fig.2. Variablepulse length sequence received atRoberval. Lower panel shows the spectrum, andupper panel shows theamplitude ina 130-Hz bandwidth centered on5.5 kHz. Pulse lengths vary from 50 to 400 ms in 50-ms steps asis in- dicatedbythe numbers between panels. Atwo-hop whistler, with echoes at•3 kHz,appears at8-10 sand originates inthe sfericat 4.2 s. A strongwell-defined two-hop whistler component extending upto 5.5 kHz is seen at about 8.2 s and cor- respondsto the one-hopdelay of the Siplepulses. HELLIWELL AND KATSUFRAKIS:BRIEF REPORT 2513
Fig. 3. Pulseenvelope of the 150-, 200-, and 250-mspulses in the second18-s sequence shown in Figure2. Solid bars representthe transmittedpulses corrected in time for the one-hoptravel time of 2.02 s.
The VLF transmitter was placed in servicein April 1973 and found that the period around local sunriseis often active. A has since operated an averageof 6 hours a day on different wide variety of spectralforms hasbeen observed in-the signals frequencies,with various modulation patterns and power stimulatedby the Siple transmissions,including risers, failers, levels. Most transmissionshave been confined to the period hooks, and interactions with other signals.Whistlers may on 0800-1400 UT, with occasional transmissions during occasionshow much strongertraces at frequenciesabove the 0000--0800and 1600-2000 UT. The fixed frequenciesused so transmitter frequency than at those below. Some of these far are 1.5, 2.5, 3.5, 4.0, 4.5, 5.0, 5.04, 5.5, 5.52, 7.1, 7.6, 14, phenomenaare discussedin the following paragraphs.More and 16 kHz, togetherwith variousfrequency sweeps. The out- details will be given on these topics in later papers. put can be frequencymodulated (FM) or keyedon and off. A When the Siple transmitter is first turned on for a daily run, wide range of FM wave forms can be programed,including the response of the magnetosphere is immediate. There square waves and sawtooth ramps. Power output can be appearsto be no buildup time requiredto 'prime' the plasma. varied continuouslyfrom zero to maximum with the aid of a The first transmitted pulse exhibits the same amplification motor-driven variable autotransformer. Values between 0.4 and triggering as later pulses. and 100 kW have been employed up to date. In September One of the first experiments performed with the Siple 1973the keyingof the transmitterwas placedunder control of transmitter was a study of the dot-dash anomaly mentioned a minicomputer,so that unmannedoperation over the full 24 above. A series of pulses of varying duration was trans- hours could be achieved. A simplified block diagram of the mitted at constant power. The pulses were formed by transmitter is shown in Figure lc. switchingthe frequ,ency between 5.0 and 5.5 kHz. The pulse The main observationpoint for the whistler mode signals length was varied from 50 to 400 ms in 50-ms steps. Five from the Siple transmitter is Roberval, Quebec. A block pulseswere transmitted at each length, giving this variable diagram of the receiver is shown in Figure lb. Continuous pulse length (VPL) sequencea total duration of 18 s. broad band tape recordings are made during all VLF Two typical examplesof the 18-sVPL program as recorded transmissions.from Siple Station. Synoptic recordings are at Roberval are shown in Figure 2. The spectrum (lower made for 2 min everyhour throughoutthe year, for studiesof panel) showssignificant activity only on 5.5 kHz, although the plasmapause;whistler propagation, and VLF emissions. thereis a strongisolated falling tone at about 33 s triggeredat The tapesare sentweekly to StanfordUniversity for spectrum 5.0 kHz. (Some of the other sequences(not illustrated) ß analysis. Quick-look results are used to guide the future programing of the transmitter. Accurate timing of dB transmitted and receivedpulses is establishedby frequency 40• 1147:48UT 23 JUNE1975 standards at Siple and Roberval. These standards are calibrated regularly against Station WWV and are kept within a few milliseconds of UT. Siple whistlermode signalshave beenobserved by the VLF receiverson various satellites,including Imp 6 and Explorer 45 (D. Gurnett and R. Anderson, personal communication, 1973) and Isis 2 (F. Palmer, personalcommunication, 1973). Subionospheresignals from Siple have been observed 2O at Halley Bay, Antarctica (K. Bullough, personal com- munication, 1973), and at Dunedin, New Zealand (R. Dowden, personal communication, 1973). I0 z,• 128dB/sec PRELIMINARY RESULTS Am•Ai •150,200, 250,400mspulses The occurrenceof detectableSiple signalsat Roberval is highlyvariable. They are seldompresent when there is natural O0 40 80 120 160 200 240 280 320 360 400ms emissionactivity on the same frequency.Their appearance FiB. 4. ^verage amplitudesof output pulsesas a functionof time seemsto dependon the prior existenceof substormactivity. after the start of the receivedpulse. Straisht line fit 8ivesexponential Although diurnal trendshave yet to be established,we have 8rowth rate of 128 dB/s. Total 8rowth is/I,, - /I• = 30 dB. 2514 HELLIWELL AND KATSUFRAKIS:BRIEF REPORT
kHz RO ---.:-.a...• •';.• .... -A'• '• .... .,:-•-:-•-.."•k-:•:...... :•:•:•.-.• •:.•;•:.'•:•:•:•:• •..:....:"'."':.'•* 6.0 •*"'•.3-... '•' :' '*•:-::*...... •.'---.-?::"".-:.• •'-::' • • ..... -...... --'..:-.,.C• ...... •'•:-. .,' :.;.•. • .....,t•. -:..• ::,..•: ,-;:•...... >*•.'•..: ..':•;•;•:.:, -':::.-•:,•....•- •,::•:.?.,•F.•;.•:;.•:.. , -,.:..,' :':%•. •:T:r'[t:•.•.;•..., ...... *-:".:'---." '%•--.•.•'•*:,d-:*.-t:':•' ..'::*:E:"' :*"..--.• '•.-"' '";' .;"• 5.0
i312 UT
--%:-::.}.•::.:,.: ::--: ., ,: .:;...... ,.•.•:,, -•,•::.%..?...... ;....:•-:-:•--,.,;' ....:.•..... • '..;•... ,% .- .•.,...•;..-.-.-•...,:
',:• -•. -. . '-*."•:,•;,...... :a.,...... :,'--,-•-: .-•: - '-..:•..-:::.:-.. ,•.:, -:.•, .•.•::::',•;•;•;.,..-'•::...•-•-..;•,. 5;0-*':,'';•}'??""-'-':*:' 4** -'-.' ..... *.•,;*-.."*- ' *-'• '...... * '•'...... •' ;•:/f':•" , ,.I ' .;:.E -.
kHz RO 19 JULY 73., !3i I: 25UT
1 i I i I i 0 I 2 3 4 5 sec Fig. 6. Hooks triggeredat 5.0 kHz showinflections and reversalsin slopeat powerline harmonics,defined by the horizontal lines. HELLIWELL AND KATSUFRAKIS.' BRIEF REPORT 2515
06 AUG 7.• 1116'10 UT
..:-:•...... '::: ..,::--.... •:.:::::.: :.,: ::../.•:. :...... •:':'".::.• :: .•:...•?.:::-'.: ':::::•:':•:'"'....'r'•:-::?...... ":'.'•';":.*: --•:'-•;:: ...... :.....*-' ...... :.•:::/:: ---•..:: ... :;.::. '-•:-.... ;:: ...... 4 ...... -.... -:: -,:..:.: -- .-._ ---...:?•,.::.':•••/•-•••••:•"•.....-•::,-.-:'"' "• ....:...."*':'*-:::'•::-•x ...... -::...•::-,,....'= ....-:....-:.-=:.:.•-,:;--:-:.-: .... I I I o 5 I.O TIME (see) Fig.7. Two l-s pulsesare transmitted 5 s apartat 4.5 kHz.Similar emissions are excited on two paths separated in group delay by about 0.7 s.
showedmore activityon 5.0 kHz.) The relativeamplitude in a length shown in the upper part of Figure 2 implies a cor- 130-Hzband centered on 5.5 kHz is shownin theupper panel. respondingtemporal growth of eachpulse. An expandedver- It is clear that the intensityof both the whistlermode signal sion of a portion of the secondVPL sequenceof Figure 2 is and its associatedemission increases with pulselength. Except given in Figure 3 to showthe growth of eachpulse. The solid for the 50-ms pulses,which are too weak to be seen,each set bars belowthe trace representthe pulseinput to the growth of five pulsesis clearlydefined. For lengthsless than 300 ms region.The leadingedge of the input pulsedoes not coincide the peak.intensity approximately doubles each time the pulse exactlywith the minimum in the receivedsignal because there length is increasedby 50 ms. This increasecorresponds to a is still some signal left from the emissiontriggered by the
growthrate of 120 dB/s. For lengthsof 350 and 400 ms the precedingpulse. If one begins,. at the minimumamplitude, it is peak intensityis roughly constant,and thus it is indicatedthat seenthat the output pulsegrows exponentially with time until saturation has been reached. the input pulseterminates. The amplitudemay then increasea The pattern of triggeredemissions is directlyrelated to the little before it beginsto fluctuateand decay. pulse length, as can be seen from Figure 2. Triggered The amplitudesof each output pulsewere scaledover the emissionsfirst appear at the ends of the 150-mspulses. All period of the input pulse by using the 150-, 200-, 250-, and emissionsfirst rise in frequency;they may then fall or rise 400-ms pulses.(The 300- and 350-ms pulseswere omitted dependingon the length of the triggeringpulse. For pulse becausetheir amplitudeswere contaminatedby previously lengthsof 250ms and less the terminal part of the emissions is triggeredemissions falling within the 130-Hz passband.)The a falling tone, whereasfor pulse lengthsof 300 and 400 ms averagesof these amplitudesare plotted on a logarithmic they are mainly risers.The upperpanel showsamplitude fluc- scalein Figure 4 as a function of time with respectto their tuations, some of which appear to relate to preceding leadingedges. The leading edge of therec•eived pulse could be whistlersand ASE's (artificially stimulatedemissions). Thus seenon the recordsonly occasionallywhen the sferic levels the third 250-mspulse in the first 18-ssequence is reducedat were low. However, its time of occurrencewas reliably es- the time of arrival of the two-hop whistler. Other reduced timated from the path delay and the known pulsetransmis- amplitudescan be associatedwith unusuallystrong preceding sion schedule.The estimatedinitial intensityof the received risers. pulseis plotted as At on Figure 4. Betweenabout 80 and 200 The systematicincrease in pulsepeak amplitudewith pulse ms the growth is seento be purelyexponential, the rate being
khz ro 16,AUG 73 I 103"17.5 UT
• ...... :.• :;.•.•::.-. ..:•...•:.• ...... •::,• ...... :::: ..•.:. • - ...;:::•.::-:::-•-.•-•.._:•-•-•:.•:•::•..--•::;-.•:.::.:..•::::.•-.:::, .,.,•:;::'-•.::•:,.:.:: ...-,*.•.:..-.:.: ...... :•,:•---:•: ::'::;•.:;j• ::,:.."'/'"•-'.•:• .."ß ...... -.$:...... :::.::-- ::.•...'--..:;'.-'"':*' ...•;..• ...... ':'.:::'-:".:;•.::•b":-•?•:...... :,::..: .•:.....•.<.... :* -:::::::'*•::'-'"'-'::-'"*'•':"*'--:*:: *:::"::'':?'""•-.:,• ..-::'*:' ::'-'""*:':-•;•--:--*:•, ...... '...... :2...... •:c-:-'•:::•,:' ':? ,:•" •--..•:•:::.:::.'.-::?:*,: ::':::•'.'..c.",74-•: .;"':/•:--... 5 . "-'e:,...... :.:,:.":-::,'*::.::: ,.:.:..::::- ,. :::-.:**..,-":'-: ....::,.---•..,<--:-...:.: ....::'•,-:...... ::..--...... ,:. :;..";'"•'•-::.**::--:c• ...... :,:<:::*-:- '" 5' '2 ...... :::%:: ':-..:':'*';:'""..,,'..:.;. :'::;:::":.:"':'"....::.'""'....'/':'....= '".•.."" ...'.....,"...'*"X:.:" ::...:::. '":•'":'""'•::•:;.'..;'"":.*;•: ':'-:".-'...... :":'"--,:,'"::...... :,:..::-:: * ...... :::--:::.'::....:,.•?...:-,----?:::-;'": ::::":'"*•:'""*' .:: ::-:-:•**:-:::, "'::'* ...... '...... -::":'•:-.:'•5'::'•":•:....:--*¾ ':':•:•;*:•: ': ..,:: -•' •' •-:::•:'" .:. '...... "•-. '"}::;:...... '"'*":""•'"•½0•' '---:'"":; ....::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: :.:,:- - '•::•-•>.•'.:•::,.. :--:.-,-:::-':•.:: :.. -•:- ..... - -::. : .?:'.":S:
4• ::::::::::::::::::::::::::::::::...... ,.,..:.:: :,: ...... ? ::, ,:.:---<' ' .... "*":*'"':*...... ':'*'"".....
:•: 0 5 .•:'...... 'TI ME (sec) Fi•. 8. At 5.04kHz the•rst •v½pulses ar• 1• msin l•n•th separatedby 6• ms(the puls• l•n•th on5.52 kHz). They do not tri•cr. Th• last•v• pulsesar• 200ms 1on•, also separated by 6• ms,and they tri•r fallin• toneson at l•astfour separatepaths. 2516 HELLIWELL AND KATSUFRAKIS.'BRIEF REPORT
kHz .RO -03 AU'6 73 t305:59 UT
...... :"-.,. ,:'..i•-..:-• - ' ' -•.'•.•.•:•-•!•...... :.•:::.-•/.:i• •*'•:•, '-. :;'•: --i::::i:,.--!'.!:•,-•:-•!.'.-•f' •;'.; : •.• :..".'-..::-'•:. , ...... -.•'•;....• •. ..•;.r:•...:•;:.:'?•:..,•:?•__.:•f•:•.•:,:;-;•.•:• •-...•...... --•--:--':;":L'.• ..•.. :,.::/,•::•.•..•:.•: •".-•.•t:.• ':-.•/•:._.:•...:....:..•--;:•:•:•. •:;..•: '"Z:•tt....q•. .:•... :...... '•. :..•:..•:•:.½-•...• . ..•.-.-.., ...... • :-?½.:•;•...•-. •:•..,.:%•: >..•-.•.• •.• ,.. ß . •...... ,...... •,0' '"...... •"'"•' " '•'• •'?' ""<' "• •' "" '..:...... '•:'.." .•..'.-. '-z;.'" '•.•'• ,.• •' •. - " •-'-"•: .'.-.•':'-- "•"-" '"'• '..•:;•"•":' •...... I I i ! 'i.... 0 5 I0 15 20 TIME (see)- Fig. 9. Pulsesof 1 s weretransmitted every 5 s on 4.5 kHz. In thecenter oF the record a whistler-triggeredfalling tone is entrained by the Siple pulse. 128 dB/s (14.7 Np/s). Beyond 200 ms the growth levels off, line radiation might act to 'entrain' the echoingwave packets, the saturationlevel A mat 30 dB abovethe estimatedinput in- causingeach new emissionto start at a particular power line tensity being reached. harmonic. Since discrete VLF emissions are known to induce Alternate transmissionon two frequencieshas led to an in- electronprecipitation [Rosenberg et al., 1971],it followsthat teresting observation, illustrated in Figure 5. The ASE's if power line radiationexists in the magnetosphere,it may triggered at 5500 Hz are frequently cut off by the signal 5000 affect the terrestrial radiation belts. Further tests of this Hz. The hook near the right sideof the middlepanel is cut off hypothesisare clearly needed before it can be accepted when its frequencyreturns to the startingfrequency, where it without doubt. encountersthe beginningof the next pulse. Among other featuresof interest are the following: Emissioncutoff or slope changemay occur at frequencies Multipath/tSE's. Figures 7 and 8 show the triggeringof not connected with known transmitters. This effect had been similar emissionson separatepaths. When this happens,it is observed in spectra of triggered emissions[e.g., Helliwell, possiblefor the spectraltrace of an emissionon one path to 1965, Figure 7-62b], but it had not been explained.An ex- crossthe traceof a signalor emissionon anotherpath. An ex- ample from the presentexperiment is shown in Figure 6, in ample of this crossingeffect appearsin the middle panel of which two triggeredhooks are observedto inflect or reverse Figure 5, at 5.5 kHz. The first emissionfirst rises, then falls, slopeat variousfrequencies. These frequencies are associated crossing5.5 kHz after the triggeringpulse with the shorter with the horizontal lines on the record, which are harmonics delay has terminated. However, the triggeringpulse on the from the Canadian power distribution system.We offer as a path with the longer delay is still present and therefore possibleexplanation that harmonic radiation from the Cana- appearsto crossthe falling tone. Usually, however,the traces dian power grid is amplifiedin the magnetosphereand affects of emissionsgenerated on the same path do not crossone emissionsin exactly the same manner as the Siple pulses. another. When the frequencyof an emissionapproaches an amplified Entrainment. At about 12 s in Figure 9 a strongwhistler- power line harmonic, the emissionis cut off or the slope is triggeredfalling tone appearsto be 'captured' or entrainedby modified.Support for this interpretationwas found in a coin- the Siple pulse.The total signalremains strong but fluctuates cident increasein the intensitiesof certain power line har- about the transmittedfrequency of 4.5 kHz. Then when the monics observedat Eights Station and its conjugate,Quebec transmittedpulse terminates, a strongrising tone is initiated. City, on October 17, 1963, at 2029 UT. This casewill be dis- The two pulsesbefore and the two pulsesafter this event, cussedin a later paper. however, are seento be relatively weak. Acceptance of the power system radiation hypothesis Pulsations. Figure l0 showsa key down transmissionof would help to explain a puzzling characteristicof periodic 10-s duration, in which ASE's are triggered at slightly less emissions.This is their tendencyto trigger repeatedlyat the than l-s intervals.After eachASE the carrier is suppressed,as same frequency[Helliwell, 1965]. In this situationthe power in the caseof Figure 2. The effectappears similar to the sup- kHz RO 23 OCT 73 1600' 18 UT ...... •-:.,• :•.•-•...... ,,.. , ..- •.....•4'• .•.•- * ,..• • • ....• •:.. , ! ,' ,' ...... ,' , :.'.;:'...... •;:.'.•..-•:...... :•:•..•.•; .:.:i•: ...... '.. :•:• .: . -•: . '"'•' • ...... :•,•.½.• .... • .•-- r ....•.'• ...... ½...... •'•...... • ..• .• .:.:.':•.. •-:•': $¾•:'-•-'::..-.-:' 3.5. ,,, . --."-.,.'::' -...::-..: ...., ':,'...----':- -- ..... •;::•;7..' ...... ' ...... ;: .....•.:-.•:-;•:.".-;.'7 ..... ' ...... - :"•- • • •'•••••'•••••• •2'• ...... -.... '...... -.•':•. •' •L•,'2,%' , .'-- •,.%,•.•.:-•..._Z:.(• %•.::;•%;• '• '•••••;•' %%•7;•'?•2%:;.•:•%.....:'".:-:7:' ...... ;".'"7'*'.:::• ':"•''* '•'••c•½:' ...... •7•L•__,-:½.,•:?•'"•27%'"•. 'T:Z' ½ :':'%" ...... 2•'• •'•'.•"••.; ...... /•:•{.•,.;' ..';.•%;;%• % •...•½--::-•.•-:.:: • • -•--7-:--•--...... :,::--•...... -.- - •...•.: ...•:••.-7.•'-••--- .:c•.½.½:-, :..-•...% ...... •...... •:•-- •-•...•..:-• :::--•"•%;.•.•7•-",.•-•- •. -•:•. r...-.-.:•.•?.,•::•: ...... • .... i; .:'•" -...... •...... '•g•:':'"•'•::'• •.,..•.•..•.•. .:,•"•'";•:• ;? -•.: •.•.. ."•:•%•.-• ...... • •:• ...... I I I I 0 5 I0 15 TIME (sec) Fig. 10. A 10-s key down transmissionon 3.5 kHz pulsatesin the form of risers,occurring regularly with spacings averagingslightly less than I s. After eachriser the signalfirst growsat the transmittedfrequency before breaking into the next riser. HELLIWELLAND KATSUFRAKIS: BRIEFREPORT 2517 kHz RO 15 SEP 73 1213:27 UT ...... ::.-,'•. "' ...... ---:;:•--: ',•..+½./' *:• :'...... • -...... ',,;½%:",---.';½-.•.... -'- :;::.•:$'•:,-:'•--t:.??: ..-:½. •:..;%•- ...... ----...&;&,.-...• -...:•...... ,•,.;:;;':•½.':•--,:--.•...... - (.."?'•½:;.;{-)%,.:•,•...... •½,. ':•:<<-; ...... :•-.-..-'.•;:--.:.:-: :, ...... -,., -.... ?'.--.: ...... :-.-, ...... p.. ",.. ½-2..½:½-(.,. .:-...... ½• :,-•::':':;'-•:½ ?'¾-,:--..- --•:,; ::-;-...½•"'.,--•--?.:•:- .:•... ,:.-,...--•.: ...... :.•:.•.• :....-:----. :,:;;..:'% ...... -.½:,;::•--:--":•."--•'•;.-.t-'----•;. ".'•-'-'•-. '....•;.• ...... • '"'•'•-:,-, :'•.-:½:-'"':• .....'• ...... -...... •-,-½ •:-:-½' .•-.----•:•'" .•-.':• ...... •'•,:' ..... ,'•*-."•½-•:-•*;.- -.-:--:-' :•;.•;•.;,,..:.•:.:.."%'-•'."•--:•:,:' :•.'.;.;•:•.•::"•'•-•,'" •....-,-•::•W.½'•-•',-.'-'.. ';'-"½'•-; .•:<'•-t'-':': :.:"•-'v.•- -,,--.• ...... ,...*•--*•.-,.½:½-:,,.•?;.'"-"•;?: .•:- :;:%,.-:%;:...... ;.,• .½:":.,z...... ,:".;, .:3;9 ';•-'-.; ?'L:;- •'"•*'½'r:•.;?,-•,/"•'*'½:..SP,-.- ;-....?:(.4,.5..% %;:.•½".. :•,,½'-:::•;• •:.".,-,;4 V---.•.7• '::,•.;;•'k,-' '?½•.,•-•-.....:.:&; z2';?;.-',:,,,'-:".'.-•..-'-. ..•;•;;:;::,'c%•-,:?,•-:,•,•"-:-½: '.?•::.•,e-...,;..:?;?.?-,-•,,:•.,.; .... ½;"•'-•4,:;::•.½:'½-.-'•-:½•:•--':'.---:,.--'•.:'•'.½':•?..-' ':½...... ½½-:•:;,--::'-': .....?•-.•';-'--.½.::•::.:.-½,•:.•c-•:-.-'•:•2,; .':•.;;f'..,:S::•.:e½.:--•...... •;•?:.-:...:.'.-;•-'.:•::?½..,.v;,•4::½•.,.•,:.•.•:-•;•.:;•4;•:•..•..•-:,.•:½ '-:•::t(h;.:;:½'";'•:•-:.--'.½::-.-½ <:::•,,.,;,½ß...... - ...... ::•..-;;•::,..,?&,:.-½½',•:..e•.¾•: '- ;- '• *"'-" *-'-•"-.:.-.•'::½":'::½-!•;; •-:,.:";:?!•-:::'-•*-':"':•E?f'"' 2 ' ""*•&• '•'••• '••;'""•'• '•:••'•;• ...... "•'••'::'z•'-::"::'z•'•' "•'"••'••••'•--•'•""•' I..... !-' I' I 0 5 I0 15 T I ME (sec) Fig. 11. Simultaneoustriggering on 2.5-'and3.5-kHz FSK (frequencyshift keying) program. At 3.5 kHz thefirst four pulsesare 800 ms long,the next 10 are 100ms long,the next 10 are 200 ms long, and the last 10 are400 mslong. At 2.5 kHz, two adjacent800-ms pulses trigger risers, beginning at I and 2.6 s, respectively.These two risersare eachfollowed by three-hopand five-hopechoes. The first and fourthpulses of the 400-msgroup at 2.5 kHz triggerhooks that alsoecho. pressioneffect seenon three-phaseperiodic emissions[Brice, transmitted from the ground. We have shown that the ex- 1965]. It appears to differ from the pulsationsobserved on ponentialgrowth rate can be obtained from a singlepulse. If NAA key down transmissions[Bell and Helliwell, 1971] that we accept the interaction model referencedhere, the growth were limited to the carrier frequency. rate is a direct measureof the differential particle flux at the Frequencyrange. Frequenciesfrom 1.5 to 16.0 kHz have resonant velocity. been transmitted from Siple, whereas signals have been Another application of these experimentsis in VLF com- detectedat Roberval only between2.5 and 7.6 kHz. An exam- munication. This experiment has shown that a sufficiently ple of triggeringat 2.5 and 3.5 kHz is shownin Figure 11. The longcoherent wave train canbe amplifiedas much as 30 dB, a riserstriggered at 2.5 k Hz producewhistler mode echoesand 3 order of magnitudeincrease in effectiveradiated power thus are the lowest frequencyemissions so far observedfrom the being given. It is of coursenecessary that the signal crossthe Siple transmitter. equatorialplane to be amplified,the explanationgiven above being assumed.Thus a satellitetransmitting to the ground in DISCUSSION AND CONCLUSIONS the northern hemispherewould have to be located in the The observationof temporal growth, as shownin Figure 4, southern hemisphere.It is also necessaryto recognize that can be explainedby a recently proposedmodel of cyclotron amplification has been observed so far only with ducted interaction with feedback included [Helliwell and Crystal, propagation. The properties of nonductedamplification re- 1973]. Most publishedtheories of cyclotroninteraction in the main to, be investigated. magnetosphereomit the effect of feedback and derive a Acknoddedgments. We want to expressour appreciationto the spatial growth rate. According to such theories [e.g., office of Polar Programs of the National ScienceFoundation and to Liemohn, 1967] all parts of a finite wave train should exhibit the Naval Support Force, Antarctica, for establishingSiple Station equal growth. Our observationsshow instead that the input and making it possibleto carry out the researchreported in this note. wave triggers an oscillation that grows exponentially with Special thanks are due to Evans Paschaland William Trabucco, who time until the input signalterminates or until the gain reaches carried out all the VLF transmissionsfrom Siple Station and to Lionel Lefebvre, who recorded the data at Roberval, Canada. In addition we •30 dB. would like to thank Frank Palmer of the Communications Research An important question remaining to be answered is Centerof Canadawho reported the Isis 2 resultsand Don Gurnett whetherthe maximumoscillation amplitude observed in these and RogerAnderson who reportedthe Explorer45 and Imp 6 results. experimentsis the maximum that can be obtained from the Appreciation is expressedto Paul Smith and Bob Hoffman of God- dard SpaceFlight Center who arrangedfor the tracking of Explorer magnetosphere.Another questionis the mechanismby which 45. The authorsappreciate the counseland assistanceprovided by our externally excited signals, such as the Siple pulses and the colleaguesat Stanford University. The researchreported in this note power line harmonic radiation, can affect'emissions and was supported by the National ScienceFoundation Office of Polar whistlers.Not one of theseresults has beenpredicted, or even Programsunder grant GV-28840X and the AtmosphericSciences Sec- tion under grant GA-32590X. The transmitter was provided by the explained,by analyticalplasma instability theories. Therefore Office of Naval Research under contract N-0012. further magnetosphericVLF wave injection experimentsare neededfor the benefit of both magnetosphericphysics and laboratory plasma physics. TheEditor thanks R. E. Barringtonand the late N.M. Bficefor An expectedby-product of emissionstimulation by the Si- their assistancein evaluating this report. ple signalsis modification of the energyand pitch anglesof REFERENCES the interacting particles. Modification experiments with natural whistlermode waves[Rosenberg et al., 1971;Helliwell Bell, T. F-, and R. A. Helliwell, Pulsationphenomena observed in et al., 1973]have producedX rays and enhancedionization in long-duration VLF whistler mode signals,J. Geophys.Res., 76, 8414, 1971. the D region. In future Siple transmitter experimentssuch Brice, N.M., Multiphase periodic very low frequency emissions,._ precipitationeffects will be sought.It may then be possibleto Radio Sci., 69D(2), 257, 1965. control X-ray production, ionization, and light emissionsin Carpenter, D. L., Ducted whistler mode propagation in the the io.nosphere. New typesof quantitativeexperiments on the magnetosphere;A half-gyrofrequencyupper intensity cutoff and some associatedwave growth phenomena, J. Geophys.Res., 73, ionosphereas well as the magnetospherecould then be per- 2919, 1968. formed. Diagnostic measurementsof the trapped energetic Guy,A. W.,The characteristics oflong antennas buried in polar ice, electron fluxes may now be possible, using VLF signals Tech. Rep. 101, Univ. of Wash., Seattle, 1966. 2518 HELLIWELL AND KATSUFRAKIS: BRIEF REPORT Helliwell,R. A., Whistlersand Relaied Ionospheric Phenomena, pp. Kimura, I., Triggeringof VLF magnetosphericnoise by a low-power 276, 304, Stanford University Press, Stanford, Calif., 1965. (• 100 W) transmitter,J. Geophys.Res., 73, 445, 1968. Helliwell, R. A., and T. L. Crystal,A feedbackmodel of cyclotronin- Lasch, S., Unique features of VLF noise triggered in the teraction between whistler mode waves and energeticelectrons in magnetosphereby Morse codedots from NAA, J. Geophys.Res., the magnetosphere,J. Geophys.Res., 78, 7357, 1973. 74, 1856, 1969. Helliwell, R. A., and E. Gehrels,Observations of magneto-ionicduct Liemohn, H. B., Cyclotronresonance amplification of VLF and ULF propagationusing man-made signals of very low frequency,Proc. whistlers,J. Geophys.Res., 72, 39, 1967. IEEE, 46, 785, 1958. Raghuram, R., R. L. Smith, and T. F. Bell, VLF Antarctic antenna: Helliwell, R. A., J. Katsufrakis, M. Trimpi, and N. Brice, Artificially Impedanceand efficiency,IEEE Trans. AntennasPropagat., in stimulated VLF radiation from the ionosphere,J. Geophys.Res., press, March 1974. 69, 2391, 1964. Rosenberg,T. J., R. A. Helliwell, and J.P. Katsufrakis, Electron Helliwell, R. A., J.P. Katsufrakis, and M. L. Trimpi, Whistler- precipitationassociated with discretevery low frequencyemissions, inducedamplitude perturbation in VLF propagation,J. Geophys. J. Geophys.Res., 76, 8445, 1971. Res., 78, 4679, 1973. Helms, W. J., H. M. Swarm, and D. K'. Reynolds,An experimental study of the polar lower ionosphereusing very low frequency (Received December 14, 1973; waves, Tech.Rep. 123, Univ. of Wash., Seattle, 1968. acceptedFebruary 15, 1974.)