<%BANNER%>

Approximating Ambient D-region Electron Densities using Dual-Beam HF Heating Experiments at the High-frequency Active Au...

MISSING IMAGE

Material Information

Title:
Approximating Ambient D-region Electron Densities using Dual-Beam HF Heating Experiments at the High-frequency Active Auroral Research Program (HAARP)
Physical Description:
1 online resource (125 p.)
Language:
english
Creator:
Agrawal, Divya
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Electrical and Computer Engineering
Committee Chair:
Moore, Robert C
Committee Members:
Lin, Jenshan
Uman, Martin A
Fitz-Coy, Norman G

Subjects

Subjects / Keywords:
density -- elf -- haarp -- heating -- ionosphere -- temperature -- vlf
Electrical and Computer Engineering -- Dissertations, Academic -- UF
Genre:
Electrical and Computer Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Dual-beam ELF/VLF wave generation experiments performed at the High-frequency Active Auroral Research Program (HAARP) HF transmitter in Gakona, Alaska are critically compared with the predictions of a newly developed ionospheric high frequency (HF) heating model that accounts for the simultaneous propagation and absorption of multiple HF beams. The dual-beam HF heating experiments presented herein consist of  two HF beams transmitting simultaneously: one amplitude modulated (AM) HF beam modulates the conductivity of the lower ionosphere in the extremely low frequency (ELF, 30~Hz to 3~kHz) and/or very low frequency (VLF, 3~kHz to 30~kHz) band while a second HF beam broadcasts a continuous waveform (CW) signal, modifying the efficiency of ELF/VLF conductivity modulation and thereby the efficiency of ELF/VLF wave generation. Ground-based experimental observations are used together with the predictions of the theoretical model to identify the property of the received ELF/VLF wave that is most sensitive to the effects of multi-beam HF heating, and that property is determined to be the ELF/VLF signal magnitude. The dependence of the generated ELF/VLF wave magnitude on several HF transmission parameters (HF power, HF frequency, and modulation waveform) is then experimentally measured and analyzed within the context of the multi-beam HF heating model.  For all cases studied, the received ELF/VLF wave magnitude as a function of transmission parameter is analyzed to identify the dependence on the ambient $D$-region electron density ($N_e$) and/or electron temperature ($T_e$), in turn identifying the HF transmission parameters that provide significant independent information regarding the ambient conditions of the $D$-region ionosphere. A theoretical analysis is performed to determine the conditions under which the effects of $N_e$ and $T_e$ can be decoupled, and the results of this analysis are applied to identify an electron density profile that can reproduce the unusually high level of ELF/VLF magnitude suppression observed on 25~July 2011. Finally, dual-beam ELF/VLF time of arrival (TOA) observations are analyzed and used to demonstrate that the dual-beam experiment is sensitive to structure within the $D$-region ionosphere.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Divya Agrawal.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Moore, Robert C.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0044876:00001

MISSING IMAGE

Material Information

Title:
Approximating Ambient D-region Electron Densities using Dual-Beam HF Heating Experiments at the High-frequency Active Auroral Research Program (HAARP)
Physical Description:
1 online resource (125 p.)
Language:
english
Creator:
Agrawal, Divya
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Electrical and Computer Engineering
Committee Chair:
Moore, Robert C
Committee Members:
Lin, Jenshan
Uman, Martin A
Fitz-Coy, Norman G

Subjects

Subjects / Keywords:
density -- elf -- haarp -- heating -- ionosphere -- temperature -- vlf
Electrical and Computer Engineering -- Dissertations, Academic -- UF
Genre:
Electrical and Computer Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Dual-beam ELF/VLF wave generation experiments performed at the High-frequency Active Auroral Research Program (HAARP) HF transmitter in Gakona, Alaska are critically compared with the predictions of a newly developed ionospheric high frequency (HF) heating model that accounts for the simultaneous propagation and absorption of multiple HF beams. The dual-beam HF heating experiments presented herein consist of  two HF beams transmitting simultaneously: one amplitude modulated (AM) HF beam modulates the conductivity of the lower ionosphere in the extremely low frequency (ELF, 30~Hz to 3~kHz) and/or very low frequency (VLF, 3~kHz to 30~kHz) band while a second HF beam broadcasts a continuous waveform (CW) signal, modifying the efficiency of ELF/VLF conductivity modulation and thereby the efficiency of ELF/VLF wave generation. Ground-based experimental observations are used together with the predictions of the theoretical model to identify the property of the received ELF/VLF wave that is most sensitive to the effects of multi-beam HF heating, and that property is determined to be the ELF/VLF signal magnitude. The dependence of the generated ELF/VLF wave magnitude on several HF transmission parameters (HF power, HF frequency, and modulation waveform) is then experimentally measured and analyzed within the context of the multi-beam HF heating model.  For all cases studied, the received ELF/VLF wave magnitude as a function of transmission parameter is analyzed to identify the dependence on the ambient $D$-region electron density ($N_e$) and/or electron temperature ($T_e$), in turn identifying the HF transmission parameters that provide significant independent information regarding the ambient conditions of the $D$-region ionosphere. A theoretical analysis is performed to determine the conditions under which the effects of $N_e$ and $T_e$ can be decoupled, and the results of this analysis are applied to identify an electron density profile that can reproduce the unusually high level of ELF/VLF magnitude suppression observed on 25~July 2011. Finally, dual-beam ELF/VLF time of arrival (TOA) observations are analyzed and used to demonstrate that the dual-beam experiment is sensitive to structure within the $D$-region ionosphere.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Divya Agrawal.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Moore, Robert C.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0044876:00001


This item has the following downloads:


Full Text

PAGE 1

APPROXIMATINGAMBIENTD-REGIONELECTRONDENSITIESUSINGDUAL-BEAMHFHEATINGEXPERIMENTSATTHEHIGH-FREQUENCYACTIVEAURORALRESEARCHPROGRAM(HAARP)ByDIVYAAGRAWALADISSERTATIONPRESENTEDTOTHEGRADUATESCHOOLOFTHEUNIVERSITYOFFLORIDAINPARTIALFULFILLMENTOFTHEREQUIREMENTSFORTHEDEGREEOFDOCTOROFPHILOSOPHYUNIVERSITYOFFLORIDA2013

PAGE 2

c2013DivyaAgrawal 2

PAGE 3

ThisdissertationisdedicatedtomyparentsRenuandSusheelandtomyhusbandAlexander 3

PAGE 4

ACKNOWLEDGMENTS IwouldliketothankmyadvisorDr.RobertMooreforprovidingmethewonderfulopportunitytoperformthisworkandforhispatienceandguidancethroughtheyears.Iamparticularlygratefultomycommitteemembers,Dr.JenshanLin,Dr.MartinUman,andDr.NormanFitz-Coyfortheirvaluablecommentsandsuggestions.Iextendspecialthankstonewandoldfriendsattheionosphericresearchlab,especiallyShujiFujimaruwhohasprovidedmeinvaluablesupportandmotivation,DanielKotovskyforhisvaluablediscussions,MichaelMitchellforbuildingallthehardware,andNealDupreeforhisexpertiseusingthehighperformancecomputingfacility.IalsowanttothankthenewestmembersofourgroupBrittanyFinchforherassistanceinprocessingsomedataandSydneyGreeneforbeingatruefriend.IwouldliketoextendmyappreciationtotheadministrativestaffoftheECEdepartment.ShannonChillingworth,ourgraduatestudentadvisorhasmademylifeeasieronsomanylevels,StephenieSparkman,EdwinaMcKayandAngelaPetringeloforeffortlesslyrunningthefrontofcewithawarmsmile.JanetHolman,ErlindaLane,KymMasonandJenniferFreemanformakingourtravelsandpurchasingsomucheasierandlessstressful.IwouldliketoalsothankRayMcClureandWaletaNewmanfortakingcareofallourpackagesandshipmentsandLaurieEdvardssonforhelpingmewithprintingandbindingmydissertation.Theyhaveallenrichedmygraduateschoolexperienceandbeenmypillarofsupport.ThisresearchwouldnothavebeensuccessfulwithoutthehelpofthecrewfromAlaska.IwouldespeciallyliketothanktheownersofChistochinaB&B,NormaandDoyleTrawfortakinggoodcareandcheckingonusandourreceiversite.JayScrimshawandDugh&JudyformindingourreceiversitesatOasisandParadiserespectively.Wewereabletogetphenomenaldatafromthesereceiversites,whichhasbeenextensivelyusedinthisresearch.IwouldalsoliketothanktheHAARPcrew,JayScrimshaw,MartyKarjalaandStephenforprovidingtheinexperiencedFlorida 4

PAGE 5

studentswithAlaskanweathergearandmentoringusonsomanylevelsandsharingtheirAlaskanexperienceswithus.Inaddition,IwouldliketothanktheoperatorsofHAARP,MikeMcCarrick,HelioZwiandDaveSeafolk-Koppwhohaveaccommodatedourlastminutechangesintransmissionformatsandhaveworkedtirelesslyonallthecampaigns.Last,butdenitelynottheleast,IwouldliketothankEdKennedyandLeeSynderfortheircontributioninrunningthecampaignsandprovidingmesupportandguidanceinthechatroom.Thisworkwouldnothavebeenpossiblewithouttheloveandsupportofsomeofmydearfriends,Sendhil,Heera,Bharani,JenniferJackson,ErinPatrick,LaureenRicks,Anitra,Gayatri,Pareena,Atchar,Suresh.Iofcoursecan'tlistallyournames,butyouknowwhoyouare.IamthankfultoAuntyPramila,AuntyRadha,UncleSubarna,UncleTirtha,AshaandSheelaforopeningtheirhometome,forthedelicioushomecookedmealsandforlovingandcaringmeliketheirown.Ialwayslookedforwardtospendingweekendsandholidayswiththem.Myparents,RenuandSusheelalthoughacrosstheAtlantic,havebeenmybiggeststrengththroughtheyears.TheirphonecallsandSkypecallswouldbrightenmyday.Ithankthemfromthebottomofmyheartfortheirlove,supportandcommitment.IthankmybrotherNikhilandhisfamilyforcheckingonme,advisingandsupportingme.Myin-laws,DonBKandDonyaZoila,CarolandJohnyfortheirloveandsupport.Lastly,mydearhusbandAlexforhislove,support,guidance,motivationandpatiencethroughtheyears.Iamlookingforwardtoourexcitingjourneytogether.ThisworkissupportedbyUSAirForcegrantFA9453-12-1-00246,DARPAcontractHR0011-09-C-0099,DARPAgrantHR0011-10-1-0061,ONRgrant#N000141010909,andNSFgrantsAGS-0940248andANT-0944639totheUniversityofFlorida. 5

PAGE 6

TABLEOFCONTENTS page ACKNOWLEDGMENTS .................................. 4 LISTOFTABLES ...................................... 8 LISTOFFIGURES ..................................... 9 LISTOFABBREVIATIONSANDSYMBOLS ...................... 12 CHAPTER ABSTRACT ......................................... 13 1INTRODUCTION ................................... 15 1.1TheIonosphere ................................. 15 1.2TheD-RegionIonosphere ........................... 17 1.3PreviousEffortstoCharacterizetheD-RegionIonosphere ......... 18 1.3.1RocketSoundingTechniques ..................... 18 1.3.2IncoherentScatterRadar(ISR) .................... 19 1.3.3Ionosonde ................................ 20 1.3.4LightningasaSourceforRemoteSensing .............. 20 1.3.5HFCross-Modulation .......................... 21 1.4Motivation .................................... 22 1.5Approach .................................... 23 1.6ScienticContributions ............................. 28 2DUAL-BEAMELF/VLFWAVEGENERATION:MODELIMPLEMENTATION .. 29 2.1NumericalAnalysis ............................... 29 2.1.1Multi-BeamHFHeatingModel ..................... 29 2.1.2TemperatureModulation ........................ 33 2.1.3ConductivityModulation ........................ 35 2.2Radiation .................................... 37 2.3ExtensiontoOtherAMWaveforms ...................... 45 2.4ExtensiontoX-andO-ModePolarizationoftheCWBeam ......... 49 3ELF/VLFWAVEGENERATIONUSINGSIMULTANEOUSCWANDMODULATEDHFHEATINGOFTHEIONOSPHERE ............... 52 3.1DescriptionoftheExperiment ......................... 52 3.2ExperimentalObservations .......................... 54 3.2.1ELF/VLFMagnitude .......................... 56 3.2.2ELF/VLFHarmonicRatio ....................... 59 3.2.3ELF/VLFPower-LawExponent .................... 61 3.3ModelPredictions ............................... 63 6

PAGE 7

3.3.1ELF/VLFMagnitude .......................... 63 3.3.2ELF/VLFHarmonicRatio ....................... 66 3.3.3ELF/VLFPower-LawExponent .................... 68 3.4Discussion ................................... 70 4ELF/VLFWAVEGENERATIONASAFUNCTIONOFPOWER,FREQUENCY,MODULATIONWAVEFORM,ANDRECEIVERLOCATION ........... 72 4.1DescriptionoftheExperiment ......................... 72 4.2DescriptionoftheDataSet .......................... 74 4.3Analysis ..................................... 78 4.3.1CWHFPower .............................. 78 4.3.2CWHFFrequency ........................... 81 4.3.3Beam1(Modulated)HFFrequency .................. 83 4.3.4ModulationWaveform ......................... 84 4.3.5ReceiverLocation ........................... 86 4.3.6Polarization ............................... 89 4.3.7ModulationFrequency ......................... 91 4.4Discussion ................................... 92 5TIME-OF-ARRIVAL(TOA)MEASUREMENTSASAMEANSTODETECTD-REGIONSTRUCTURE ...................... 94 5.1CoupledNatureofNeandTeProles .................... 95 5.2LinearTeProlesandExponentialNeProles ................ 96 5.2.1MatchingObservationson25July2011 ............... 100 5.3Dual-BeamELF/VLFTOAObservations ................... 102 5.3.1FittingaPiecewise-ExponentialNeProleUsingTOA ........ 108 5.3.1.1LowerSection ....................... 108 5.3.1.2AltitudeofIntersection(h0int) ................. 109 5.3.1.3TheMagnitudeofNe ..................... 110 5.3.1.4UpperSection ....................... 111 5.4BestFitPiecewise-ExponentialProle .................... 112 6SUMMARYANDFUTUREWORK ......................... 114 6.1WeightedLeastSquareImplementation ................... 115 6.2ExtensiontoHigherHarmonics ........................ 116 6.3HFFrequenciesWithinaCollisionFrequency ................ 116 6.4EvaluationofIonosphericCurrentDrive(ICD) ................ 116 6.5HFCross-ModulationUnderDual-BeamHeatingConditions ........ 117 REFERENCES ....................................... 118 BIOGRAPHICALSKETCH ................................ 125 7

PAGE 8

LISTOFTABLES Table page 1-1ReceiverLatitude,Longitude,andDistance(km)fromHAARP. ......... 25 2-1HFbeamparametersformodelinput. ....................... 31 4-11225HzSNRateachsiteforeachday. ...................... 78 5-1ExperimentalTOAobservations. .......................... 109 5-2Piecewise-ExponentialNe. ............................. 109 8

PAGE 9

LISTOFFIGURES Figure page 1-1ConstituentparametersintheD-region. ...................... 17 1-2Electronconcentrationsderivedfromvariousrocketsoundingtechniques. ... 19 1-3TheHAARPHFheatingfacility. ........................... 24 1-4Amapofthegroundbasedreceiverlocations. .................. 26 1-5ReceiversystemdeployedtomeasureradialandazimuthalB-elds.(PhotocourtesyofDivyaAgrawal.) ............................. 27 2-1D-regionambientelectrondensityandtemperatureproles. ........... 30 2-23-Dbeampatternsand2-DslicesoftheHFpowerpatternstransmittedbyHAARP. ........................................ 32 2-3AmplitudesofthreeharmonicsforHall,Pedersen,andParallelconductivities. 36 2-43-DrepresentationofHF-heatedregionsandradiation. ............. 38 2-5DifferenceinconductivitymodulationwithelectrojetE-eld. ........... 39 2-6Absolutemagneticeldstrengthforelectrojeteldof5mV/mand100mV/m. 41 2-7Numericalpredictions:1215HzHallcurrentB-eldamplitudes. ......... 42 2-8Numericalpredictions:1215HzPedersencurrentB-eldamplitudes. ...... 43 2-9Numericalpredictions:2430HzHallcurrentB-eldamplitudes. ........ 45 2-10Numericalpredictions:2430HzPedersencurrentB-eldamplitudes. ..... 46 2-11SteadystateTeasafunctionoftimeforvariousaltitudes. ............ 47 2-12MaxandminTeasafunctionofaltitudeforvedifferentmodulationwaveforms. 48 2-13MaximumandminimumTeachievedasafunctionofCWERP. ......... 48 3-1Acartoondiagramofthedual-beamHFheatingexperiment. .......... 53 3-2ThetransmissionscheduleforthemodulatedHFbeam. ............. 54 3-3MagnitudeofELF/VLFsignalsobservedatHAARP. ............... 55 3-4ELF/VLFmeasurements:Statisticaldistributions. ................. 57 3-5Magnitudeofthe1stharmonicobservedduringpeakpowertransmissions. .. 58 3-6Ratioofthesecondtotherstharmonicmagnitude. ............... 61 9

PAGE 10

3-7Thepower-lawexponentat1215Hzand2430Hzforeachpower-stepseries. 62 3-8Numericalpredictions:TotalB-eldmagnitudeatthereceiver. .......... 64 3-9Numericalpredictions:changeinB-eldmagnitude(CW-OFF/CW-ON). ... 65 3-10ELF/VLFharmonicratio. ............................... 67 3-11Numericalpredictions.(A)nat1215;(B)nat2430Hz;(C)ThechangeinnfromCW-OFFtoCW-ONconditions. ........................ 69 4-1Cartoondiagramofthedual-beamHFheatingexperiment. ........... 73 4-21225HzsignalmagnitudeobservedatParadise(PD). .............. 75 4-390-secondspectrogramsofELF/VLFobservationsatParadise. ......... 76 4-41225Hzsignalmagnitudesobservedon20July2011atPD. .......... 77 4-5Observationsandmodelpredictionsfor1225Hzsignalmagnitude. ....... 79 4-6NormalizedELFmagnitudeasafunctionofCWERP. .............. 82 4-7CW-OFF:ELF/VLFmagnitudeasafunctionofmodulationwaveform. ..... 85 4-8ELF/VLFmagnitudeasafunctionofCWERPfortheveAMwaveforms. ... 86 4-9NormalizedELF/VLFmagnitudeasafunctionofELF/VLFreceiverlocation. .. 88 4-10NormalizedELF/VLFmagnitudeversusCWHFbeampolarization. ....... 90 4-11NormalizedELF/VLFmagnitudefordifferentmodulationfrequencies. ..... 91 5-1ELFmagnitudedependenceonNeandTeproles. ................ 96 5-2ExponentialNeproleswithvaryingandh0. ................... 97 5-3LinearTeproles. .................................. 98 5-4ELFmagnitudeasafunctionof. ......................... 99 5-5ELFmagnitudeasafunctionofh0. ......................... 99 5-6ExponentialNewith0.05,0.1km)]TJ /F9 7.97 Tf 6.59 0 Td[(1andh090km. ............... 101 5-7NormalizedELFmagnitudeforof0.05and0.10km)]TJ /F9 7.97 Tf 6.58 0 Td[(1andh090km. ..... 102 5-8DominantsourceregionforHallconductivitywithCW-OFFandCW-peakpower. 103 5-9AbsoluteTOAandLOSamplitudeobservationswithCWERP. ......... 104 5-10RelativeTOAandnormalizedLOSamplitudeobservationswithCWERP. ... 105 10

PAGE 11

5-11AnalysisforexponentialNeasafunctionoffrom)]TJ /F1 11.955 Tf 9.3 0 Td[(0.1to0.8. ......... 106 5-12AnalysisforexponentialNeasafunctionoffrom0.2to0.4. .......... 107 5-13Varyingbetaofthelowerpiece. ........................... 110 5-14Varyingthealtitudeofintersectionofthetwopieces. ............... 110 5-15VaryingthemagnitudeofNe. ............................ 111 5-16Varyingbetaofupperpiece. ............................. 112 5-17Observationsalongwithbesttpiecewise-exponentialprole. .......... 113 11

PAGE 12

LISTOFABBREVIATIONSANDSYMBOLS ABBREVIATIONS:CWContinuousWaveELFExtremelyLowFrequency(3-3000Hz)ERPEffectiveRadiatedPowerHAARPHigh-frequencyActiveAuroralResearchProgramHFHighFrequency(3-30MHz)OAOasisReceiverlocation(3kmfromHAARP)PDParadiseReceiverlocation(98kmfromHAARP)SCSinonaCreekReceiverlocation(33kmfromHAARP)TOATime-of-ArrivalVLFVeryLowFrequency(3-30kHz)SYMBOLS:fcCenterFrequency(MHz)NeElectronDensityProle(elec/cm)]TJ /F9 7.97 Tf 6.59 0 Td[(3)TeElectronTemperatureProle(kelvin) 12

PAGE 13

AbstractofDissertationPresentedtotheGraduateSchooloftheUniversityofFloridainPartialFulllmentoftheRequirementsfortheDegreeofDoctorofPhilosophyAPPROXIMATINGAMBIENTD-REGIONELECTRONDENSITIESUSINGDUAL-BEAMHFHEATINGEXPERIMENTSATTHEHIGH-FREQUENCYACTIVEAURORALRESEARCHPROGRAM(HAARP)ByDivyaAgrawalAugust2013Chair:RobertMooreMajor:ElectricalandComputerEngineeringDual-beamELF/VLFwavegenerationexperimentsperformedattheHigh-frequencyActiveAuroralResearchProgram(HAARP)HFtransmitterinGakona,Alaskaarecriticallycomparedwiththepredictionsofanewlydevelopedionospherichighfrequency(HF)heatingmodelthataccountsforthesimultaneouspropagationandabsorptionofmultipleHFbeams.Thedual-beamHFheatingexperimentspresentedhereinconsistoftwoHFbeamstransmittingsimultaneously:oneamplitudemodulated(AM)HFbeammodulatestheconductivityofthelowerionosphereintheextremelylowfrequency(ELF,30Hzto3kHz)and/orverylowfrequency(VLF,3kHzto30kHz)bandwhileasecondHFbeambroadcastsacontinuouswaveform(CW)signal,modifyingtheefciencyofELF/VLFconductivitymodulationandtherebytheefciencyofELF/VLFwavegeneration.Ground-basedexperimentalobservationsareusedtogetherwiththepredictionsofthetheoreticalmodeltoidentifythepropertyofthereceivedELF/VLFwavethatismostsensitivetotheeffectsofmulti-beamHFheating,andthatpropertyisdeterminedtobetheELF/VLFsignalmagnitude.ThedependenceofthegeneratedELF/VLFwavemagnitudeonseveralHFtransmissionparameters(HFpower,HFfrequency,andmodulationwaveform)isthenexperimentallymeasuredandanalyzedwithinthecontextofthemulti-beamHFheatingmodel.Forallcasesstudied,thereceivedELF/VLFwavemagnitudeasafunctionoftransmissionparameterisanalyzed 13

PAGE 14

toidentifythedependenceontheambientD-regionelectrondensity(Ne)and/orelectrontemperature(Te),inturnidentifyingtheHFtransmissionparametersthatprovidesignicantindependentinformationregardingtheambientconditionsoftheD-regionionosphere.AtheoreticalanalysisisperformedtodeterminetheconditionsunderwhichtheeffectsofNeandTecanbedecoupled,andtheresultsofthisanalysisareappliedtoidentifyanelectrondensityprolethatcanreproducetheunusuallyhighlevelofELF/VLFmagnitudesuppressionobservedon25July2011.Finally,dual-beamELF/VLFtimeofarrival(TOA)observationsareanalyzedandusedtodemonstratethatthedual-beamexperimentissensitivetostructurewithintheD-regionionosphere. 14

PAGE 15

CHAPTER1INTRODUCTIONThisPh.D.dissertationexperimentallyandtheoreticallyinvestigatesanovelmethodtoestimatetheambientconditionsoftheD-regionionosphere.Thetechniquerequiresahighpower,highfrequency(HF,3-30MHz)transmittersuchasthatavailableattheHigh-frequencyActiveAuroralResearchProgram(HAARP)observatoryinGakona,Alaska.Adual-beamHFheatingtransmissionformatisusedtoquantifythesuppressionand/orenhancementoftheextremelylowfrequency(ELF,30Hzto3kHz)/verylowfrequency(VLF,3-30kHz)signalmagnitudegeneratedusinganamplitudemodulated(AM)beambytheadditionofacontinuouswave(CW)beamatvaryingHFpowerlevels,HFfrequencies,andHFpolarizations.Inordertounderstandtheimpactofthepresentedmaterial,thischapterprovidesbackgroundinformationabouttheionosphereingeneraltogetherwithahistoricalcontextregardingthecharacterizationofelectricalparameterswithintheweaklyionizedandhighlycollisionalD-regionionosphere.Wepresentabriefoutlineoftheapproachtotheproblemandsummarizethescienticcontributionsofthiswork. 1.1TheIonosphere ShortlybeforemiddayIplacedthesingleearphonetomyearandstartedlistening.Thereceiveronthetablebeforemewasverycrude-afewcoilsandcondensersandacoherer-novalves,noampliers,notevenacrystal.ButIwasatlastonthepointofputtingthecorrectnessofallmybeliefstotest.Theanswercameat12:30whenIheard,faintlybutdistinctly,pip-pip-pip.IhandedthephonetoKemp:Canyouhearanything?Iasked.Yes,hesaid.TheletterS.Hecouldhearit.-GuglielmoMarconi,December12,1901.Marconi'swordsfollowingthetransmissionoftherstwirelesstrans-AtlanticmessagebetweenPoldhu,CornwallonEngland'sSouthWestcoastandSaintJohns,NewfoundlandCanada,about3500kilometersapart,conveystheexcitementofthisdiscovery.ThisexperimentalobservationledtoKennellyandHeavisidemakingindependentsuggestionsin1902thatthereexistedanionizedlayerintheEarth's 15

PAGE 16

upperatmospherethatcouldreectradiowavesaroundthecurvedsurfaceoftheEarth.AstaunchscienticdebateensuedregardingtheexistenceofthesocalledKennelly-Heavisidelayer.Theconceptofa`radioreectinglayer';didnotreceivegeneralscienticacceptanceuntiltheearly1920's.Itisgenerallyacceptedthat Ap-pletonandBarnett [ 1925 ]and Appleton [ 1932 ]experimentallyprovedtheexistenceoftheatmosphericionizedlayeralongwithanestimateoftheheightofthereectinglayer.Inretrospect,itwasactually deForest [ 1912 ]whomadetherstapproximatemeasurementsoftheheightofthe`radioreectinglayer'anditisunfortunatethathiscontributionisoftenoverlookedintheworksofradiocommunication. Villard [ 1976 ]providesadetailedessayonthecontributionsofLeedeForrestandhiscolleagueLeonardF.Fuller.Nevertheless,withthescienticallyacceptedproofprovidedby Ap-pletonandBarnett [ 1925 ]and Appleton [ 1932 ],thetermIonospherereplacedthetermKennelly-Heavisidelayer.Theionosphereisanexampleofanaturallyoccurring,weakly-ionizedplasmaandconsistspredominantlyofneutralparticles,butalsooffreeelectronsandpositiveionsthatareassumedtobeapproximatelyequalinnumber.Basedonthelevelofionization,theionosphereisdividedintolayers,withtheD-regionextendingfrom50100km,theE-regionfrom100150km,andtheF-regionstartingat150km. Tellegen [ 1933 ]rstintroducedtheconceptofarticialionosphericmodicationbyhighpowerradiowaves.Highpowerradiowaveheatingoftheionosphereprovidesameanstoperformcontrolledionosphericmodicationexperiments.Since1933,plasmaphysicsandthescienticunderstandingoftheionospherehasdramaticallyimproved.Scientistsnowinvestigatetheionosphereforavarietyofreasons.Thereareeffectsrelatedtolightning(transientluminousevents,lightning-inducedelectronprecipitation),effectsdrivenbysolaractivity(X-rayares),effectsrelatedtointergalacticradiation(rays),effectsofradiationbeltdynamics(vialightning-inducedelectronprecipitation),andeffectsofatmosphericgravitywaves,tonameafew. 16

PAGE 17

Figure1-1. Concentrationsofpositiveions(N+2,O+2,NO+),negativeions(O)]TJ /F9 7.97 Tf -0.88 -7.97 Td[(2,)andelectrons(ne)versusaltitudeforaveryquietsun,adaptedfrom NicoletandAikin [ 1960 ]. 1.2TheD-RegionIonosphereTheD-regionionosphere(50100km)istheweaklyionizedandhighlycollisionalplasmalayeroftheionosphere,anditservesasaninterfacebetweentheneutralatmosphereandthemorehighlyionizedatmosphericlayers.IonizationinthisregionisproducedbysolarXrays,bothduringsolararesandundernormalsolarconditions.TheprimaryionizingradiationsfromthesunarethediscreteLyman-lineat1216Aintheultraviolet,andhardX-raysofwavelength<10A[ Davies 1990 ,p.33]. NicoletandAikin [ 1960 ]provideadetaileddiscussionontheprimaryconstituentsoftheD-region:molecularnitrogenandoxygen,nitricoxide,andpositiveandnegativeions.Disturbanceswithinthislayergreatlyaffecttheabsorptionofhighfrequencywavesandthereectionoflowfrequencyradiosignals.Figure 1-1 showsthevariousconstituents 17

PAGE 18

oftheD-regionionosphereasafunctionofaltitude.ThegeneralperceptionoftheD-regionconstituentshasnotchangeddramaticallysince1960.TheD-regionismainlycharacterizedbyelectrondensity,neutraldensity,andelectrontemperatureproles.ThecharacteristicparametersoftheD-regionvarywithday,season,solaractivity,zenithangle,andgeographiclocations,amongothers.RadiowavesareabsorbedastheypropagatethroughtheD-regionduetothehighelectron-neutralcollisionfrequency.Infact,theD-regionistheprimarysourceofradiowaveattenuationforwavespropagatingfromthegroundintospaceandviceversa.CharacterizingtheelectricalpropertiesoftheD-regionionosphereisthusanimportantgoalintheefforttounderstandthedynamiccouplingofradiowaveenergybetweenthegroundandspace.ThefollowingsectionprovidesanoverviewofpreviouseffortstocharacterizetheD-regionoftheionosphere. 1.3PreviousEffortstoCharacterizetheD-RegionIonosphereTheD-regionionosphereisparticularlydifculttocharacterizebecauseitsaltitudeistypicallytoohighforballoonmeasurementsand/ortoolowforradarandsatellitemeasurements.EffortstoprobetheD-regionhavebeenperformedusingrockets,radars,ionosondes,andtopsidesounders.ThelistcontinueswithdifferentialabsorptionanddifferentialphasemeasurementsofpartialHFreectionmeasurements,VLF,lowfrequency(LF),andmediumfrequency(MF)soundingtechniques.SomeofthemajoreffortsinthedirectionofcharacterizingtheambientD-regionelectrondensityarediscussedbelow. 1.3.1RocketSoundingTechniquesThesalientmethodstoprobetheD-regionionosphereusingrocketsoundingincludetheLFradiowavepropagationtechnique,theVLFDopplertechnique,andtheLangmuirprobetechniquethatmeasuresbothpositiveandnegativeorelectroncurrents[ Sechrist 1974 ,andreferencestherein].Thesetechniquesarebasedon 18

PAGE 19

Figure1-2. Electronconcentrationsderivedfromatwo-frequencyrocketdifferentialabsorption()andFaradayrotation(X,3385kHz;+,2225kHz)experiment.Thesolidlineistheuncalibrateddcprobecurrent;adaptedfrom Sechrist [ 1974 ]. measurementsofFaradayrotationanddifferentialabsorption.Forexample,Figure 1-2 showstheelectronconcentrationsderivedfromthedifferentialabsorption,Faradayrotation,anddcproberockettechniques.AnumberofotherscientistshavesuccessfullyperformedrocketexperimentstocharacterizetheD-regionelectrondensity[e.g., Sed-donetal. 1954 ; PrakashandPandey 1984 ; Ulwicketal. 1988 ].Althoughrocket-bornemeasurementsofD-regionelectrondensityproleexhibitgoodaccuracyandheightresolution,oneofthemajordrawbacksofusingrocketsisthattherocketexhaustmodiesthechemicalcompositionoftheionosphere[ ZinnandSutherland 1980 ; Pick-ettetal. 1985 ; Bernhartetal. 2005 ].RocketsoundingtechniquesthuscannotbeusedforthepurposeofprovidingareliablediagnosticaboveHAARPforuseduringongoingexperiments. 1.3.2IncoherentScatterRadar(ISR)ThetechniquetoprobetheionosphereusingISRinvolvestransmittingahigh-powerpulsedradiowaveupwardandrecordingtheweakreturnthatisscatteredfromthethermal,chemical,orturbulentcomponentsoftheionosphericplasmaatground-basedreceivers.TheRadarEquationisappliedtondtherelationshipbetweenthereceivedpowerandthetransmittedpower,inturnestimatingthescatteringcrosssectionand 19

PAGE 20

thustheelectrondensityorelectronconcentration(Ne)[e.g., Mathews 1984 1986 ; ChauandWoodman 2005 ].TheshapeofthereceivedpowerspectrumyieldsfurtherinformationabouttheionosphereintermsofNe[ Sechrist 1974 ]. Sechrist [ 1974 ]arguedthatISRmeasurementsarequiteaccurateforelectrondensitiesdowntotheledgeorthesteepgradientinelectrondensitiesbetween80and90km.Five-minuteaveragesproducedastatisticalerrorashighas50%forcaseswhentheelectrondensitieswereoftheorderof1000cm)]TJ /F9 7.97 Tf 6.59 0 Td[(3,whichistypicalatthehighestD-regionaltitudes.ISRmeasurementsthusprovideimportantboundsforNeatthehighestD-regionaltitudes,butareclearlynotsuitableforperforminghightime-resolutionquantication. 1.3.3IonosondeIonosphericdiagnosticsprovidedbytheionosondeconsistofshortpulseradioechoesoverabroadrangeofHFfrequencies.Itisalsoknownas`swept-frequencypulsesounding'[ Hargreaves 1992 ].Thesepulsesarereectedatanaltitudewheretheplasmafrequencyequalsthewavefrequency,andtheirechosarereceivedatagroundbasedreceiver.AnionogramisagraphofreectionheightversusHFfrequency,andunderoptimalconditionsitrepresentsadirectmeasureoftheelectrondensityasafunctionofaltitude.Thescaleheightisanotherimportantionosphericcharacteristicthatdescribestheshapeoftheionosphericelectrondensityprole.Ionosondemeasurementshavebeenusedfortheaccuratedeterminationofionosphericeffectivescaleheights[ Tulasietal. 2009 ].Althoughionosondemeasurementsprovideaccurateelectrondensitymeasurementintheupperregionsoftheionosphere(E-andF-regions),thismethodtypicallyproducesnoechofromwithintheD-region,duetothelowelectrondensitiesatthataltituderange[ RishbethandGarriott 1969 ]. 1.3.4LightningasaSourceforRemoteSensingItiswellknownthatthemajorportionoftheradiationofelectromagneticenergyfromlightningdischargesisintheVLFrange[ RakovandUman 2003 ]. Cummeretal. [ 1998 ]showedthatobservationsofELF/VLFradioatmosphericsorsfericsmade 20

PAGE 21

bygroundbasedreceiverscanbeutilizedtoderivebest-texponentialD-regionelectrondensityprolesalongthepropagationpath. Chengetal. [ 2006 ]usedasimilarmethodologyofradioatmosphericstoextractthenighttimeD-regionelectrondensityprolealongwithamodelforVLFpropagationintheEarth-ionospherewaveguide.Morerecently, Jacobsonetal. [ 2008 ], LayandShao [ 2011 ]performedVLF-LFionosphericsoundingoftheD-regionionosphereusinglightningradioemissions.Althoughtheseexponentialprolesreectthelong-pathaveragedD-regionelectrondensities,itremainstobeseenwhethertheycanprovidemoredetailedinformationaboutthelocalD-regionconditionsnearHAARP.ItwouldbeunrealistictoexpectthatenoughlightningwouldoccurataspecictimeandbeproperlylocatedtoprovideausablelocalizeddiagnosticatHAARPduringexperiments. 1.3.5HFCross-ModulationCross-ModulationprobingexperimentshavebeenwidelyusedtodeterminetheextentofionosphericconductivitymodulationintheD-regionionosphere.Themethodusesasequenceofpreciselytimedshort,high-powerdisturbingpulsesandshort,low-powerprobingpulses,suchthattheprobepulsereectsfromtheF-regionandasitpropagatesdownwards,undergoescross-modulationthroughthemodiedD-region,whichinturnleadstoadirectcalculationofthevirtualaltitudeofinteraction.Anumberofscientistshaveextensivelyexploredtheheater-modiedcharacteristicsoftheD-regionionosphereusingmeasurementsofcross-modulationexperiments[e.g., Weisbrodetal. 1964 ; Fejer 1970 ; Senioretal. 2010 ].Recently, LangstonandMoore [ 2013 ]demonstratedusingHFcross-modulationduringHFheatingexperimentsatHAARPthatitwaspossibletoquantizetheD-regionabsorptionproducedbyHFheatingbothduringtheinitialstagesofheatingandundersteady-stateconditions.Although,thisgivesanestimateoftheextendofionosphericconductivitymodulationintheD-region,thismethodologydoesnothoweverprovidedetailsontheD-regionstructure. 21

PAGE 22

1.4MotivationTheeffortofthisworktocharacterizethepropertiesoftheD-regionplasmaaremotivatedbythefactthattheD-regionistheprimarysourceofradiowaveattenuationforwavespropagatingfromthegroundintospaceandviceversa.CharacterizingtheelectricalpropertiesoftheD-regionionosphereisthusanimportantgoalintheefforttounderstandthedynamiccouplingofradiowaveenergybetweenthegroundandspace.HFheatingprovidesameanstoperformcontrolledionosphericmodicationexperiments.KnowledgeoftheambientD-regionandthusthevariableabsorptiontakingplaceintheD-regionisessentialforunderstandingthemechanismofELF/VLFgenerationandtoimprovetheefciencyofwavegeneration.ELF/VLFwavescanpropagateforlargedistancesoftheorderoftheEarth'sradiusandndtheirapplicationinglobalcommunicationandnavigation.Further,sinceELF/VLFwavespenetratedeepbelowthesurfaceoftheearth,theyareusefulforsubmarinecommunicationandinimagingundergroundstructures,e.g.,incavedetection.ExtensiveworkhasalsobeenperformedusingdirecttransmissionofVLFwavesfromthegroundtointeractwithradiationbeltelectronsinthemagnetosphere.AsidefromtheaforementioneddirectapplicationsofELF/VLFwavegeneration,knowledgeoftheambientD-regionaboveHAARPisessentialinunderstandingthephysicsthatoccursathigheraltitudes.Articially-producedsmall-scaleandlarge-scalestructuresintheplasmadensityintheE-andF-regionsoftheionosphereknownas`IonosphericIrregularities'presentrandomtemporaluctuationsinbothamplitudeandphasewhenreceivedatanantenna,knownas`IonosphericScintillation'[e.g., Djuthetal. 2006 ; Hysell 2008 ; Fallenetal. 2011 ].Scintillationmaycauseproblemsthatdegradethequalityofthesatellitenavigationsystems[e.g., Milikhetal. 2008 ; Kouetal. 2010 ].Furthermore,scientistsareinterestedingeneratingarticialemissionsspectrallysimilartothenaturalaurorabyhighpowerHF-heatingofthebottom-side 22

PAGE 23

F-region[e.g., Kendalletal. 2010 ; Holmesetal. 2011 ]andinunderstandingthemechanismofionoutowsrelatedtoarticialductformationbyHFheatinginboththebottom-andtop-sideF-regions[ Milikhetal. 2010 ].ManyothertypesofhigheraltitudeexperimentsareperformedatHAARP,andinallcases,itisnowimperativetohaveanunderstandingoftheHFpowerthatreachesthesealtitudes.TheD-regiondiagnosticproposedinthisworkdirectlyaddressestheseissues,andwillbeusedtoquantifytheHFpowerthatreachesE-andF-regionaltitudes.HavingprovidedadiscussionofthepreviouseffortstocharacterizetheD-regionionosphere,thedrawbacksassociatedwiththem,andthemotivationtoperformHFheatingexperimentstocharacterizetheD-regionionosphere,thefollowingsectionbrieydescribestheapproachusedinthisworktowardstheD-regioncharacterizationproblem. 1.5ApproachInordertocharacterizetheambientD-regionionosphere,thisworkutilizesELF/VLFwavegenerationusingdual-beamHFheatingexperiments.Ground-basedexperimentalobservationsofELF/VLFwavesareinterpretedinthecontextofatheoreticalmodel.TheELF/VLFwavemagnitudeisidentiedasthepropertyofthereceivedELF/VLFwavethatismostsensitivetotheeffectsofdual-beamHFheating.ThedependenceofthegeneratedELF/VLFwavemagnitudeonseveralHFtransmissionparameters(HFpower,HFfrequency,andmodulationwaveform)isthenexperimentallymeasuredandanalyzedinordertoidentifythedependenceontheambientD-regionelectrondensity(Ne)and/orelectrontemperature(Te).Itisshownthatthepower,frequency,andpolarizationoftheCWHFbeam(inadditiontothetypeofamplitudemodulationwaveformused)provideindependentinformationabouttheambientD-regionionosphere.Furthermore,conditionsunderwhichtheeffectsofNeandTecanbedecoupledaredeterminedtheoreticallybyevaluatingmodelpredictionsfornumerousexponentialelectrondensityproleswithdifferingdensitiesandslopesand 23

PAGE 24

Figure1-3. TheHAARPHFheatingfacility.(PhotocourtesyofDr.LeeSynder.) fornumerouslineartemperatureproles.Lastly,itisdemonstratedthatdual-beamELF/VLFtime-of-arrival(TOA)experimentsprovideadditionalkeyinformationabouttheambientNeprole.Usingthisinformation,wedemonstratethattheambientD-regionelectrondensitycanbeapproximatedusingapiecewise-exponentialNeprolethatisconsistentwithallobservationstodate.TheexperimentalworkiscarriedoutatHAARP,locatedinGakona,Alaska.ThegeographiccoordinatesoftheHFantennaarrayareapproximately62.39N,145.15W.Thefacilityisdesignedtotransmitanarrowbeamofhighpowerradiosignalsinthe2.8to10MHzfrequencyrange.IthasbeenshownthatforHFfrequencies>10MHz,theabsorptionofHFradiowavesgetsprogressivelysmaller.Thereare180towers 24

PAGE 25

ina1215rectangulargridwitheachtowersupportingtwopairsofcrosseddipoleantennas.Eachactivedipoleelementhasa10kWtransmitter,foratotalof3600kWavailablefortransmission.Figure 1-3 showsapictureoftheHFantennaarray.MoretechnicalinformationabouttheHAARPfacilitycanbefoundat http://www.haarp.alaska.edu/haarp/index.html .TheHAARParraycanbeusedtoindependentlybroadcastdifferentHFfrequenciessimultaneouslyusingasplit-arrayconguration.Theexperimentspresentedinthisworkutilizethesplit-arraycongurationtoperformdual-beamHFheatingexperimentsandtoassesstheveracityofamultiple-beamionosphericheatingmodel.Thedual-beamexperimentsperformedconsistoftwoHFbeamssimultaneouslyheatingthelowerionosphere:whileonebeammodulatestheconductivityofthelowerionosphereatELF/VLFfrequencies(AM),asecondHFbeamcontinuallyheatsthesamepartofthelowerionosphere(CW),modifyingtheefciencyofELF/VLFconductivitymodulationandtherebytheefciencyofELF/VLFwavegeneration.TheELF/VLFwavesgeneratedareobservedatvariousground-basedELF/VLFreceiversdeployedbytheUniversityofFlorida.Figure 1-4 showsthelocationsofHAARPandthereceiversonthemapofAlaska.Table 1-1 liststhereceiveracronyms,latitudeandlongitudeofeachreceiversiteandHAARP,andtheirapproximatedistance(inkm)fromHAARP. Table1-1. ReceiverLatitude,Longitude,andDistance(km)fromHAARP. ReceiverLatitudeLongitudeDist.fromName(N)(W)HAARP(km) HAARP62.39145.20Oasis(OA)62.35145.13SinonaCreek(SC)62.58144.633Paradise(PD)62.52143.298 Figure 1-5 showstheground-basedELF/VLFreceiversystem.Eachreceiversystemconsistsoftwoorthogonalmagneticloopantennasorientedtodetecttheradialandazimuthalcomponentsofthemagneticeldatgroundlevel,apreamplier,alinereceiver,andadigitizingcomputer.AccuratetimingisprovidedbyaGPSclock. 25

PAGE 26

Figure1-4. AmapofAlaska,showingthereceiverlocationsrelativetothelocationofHAARP.ReceiveracronymsandlatitudesandlongitudesaregiveninTable 1-1 Thereceiverissensitivetomagneticeldswithfrequenciesbetween300Hzand45kHz.TheELF/VLFreceiverhasbeenrigorouslytestedtodeterminewhethertheobservedELF/VLFsignalscouldbearticiallycreatedbynon-lineardemodulationofthehigh-powerHFwavearrivingatthereceiver.Ifthiswerethecase,onewouldexpecttoobservenonlineareffectsonotherELFandVLFsignalsrecordedinthedataatthetimeoftransmission,andtheseeffectsarenotobserved.Forinstance,modulationsidebandsarenotobservedonVLFtransmittersignals(inthe20-25kHzrange),andnaturalVLFsignalsdonotexhibitevidenceofreceiversaturationorothernonlinearities,despitethefactthatthesesignalsaretypicallymanytimesstrongerthantheELF/VLFsignalsgeneratedbymodulatedheatingofauroralelectrojetcurrents.Additionally,directmeasurementsofcommon-modeanddifferential-modesignalcouplingalsosuggestthattheobservedELF/VLFsignalsaregeneratedbymodulatedheatingoftheauroralelectrojetcurrents,ratherthanbynonlineardemodulationoftheHFwaveinthereceiverelectronics.Injectedcommon-modesignalsat1.6MHzwerereducedby40dBcomparedtosignalsat1kHz,andcommon-modesignalsathigherfrequencieswere 26

PAGE 27

Figure1-5. ReceiversystemdeployedtomeasureradialandazimuthalB-elds.(PhotocourtesyofDivyaAgrawal.) toosmalltobemeasured.Injecteddifferential-modesignalsmeasuredat1MHzwerereducedby40dBfromthe1kHzvalue.Higherfrequencydifferential-modesignalswerealsotoosmalltomeasureaccurately.Exceptwherenoted,theamplitudesandphases 27

PAGE 28

oftheELF/VLFtonesatthemodulationfrequenciesandtheirharmonicsaredeterminedinpost-processingusingdiscreteFouriertransforms.Havingprovidedthebackground,historicalreviewofpastD-regionprobingexperiments,motivation,andapproachtocharacterizetheD-regionoftheionosphere,theremainderofthisworkisorganizedasfollows:Chapter 2 ,describestheimplementationofmulti-beamHFheating;Chapter 3 identiestheELF/VLFmagnitudeastheELF/VLFwavepropertymostsensitivetoadditionalCWheating;Chapter 4 criticallyinvestigatesthedependanceofELF/VLFwavesonCWheatingasafunctionofHF-power,HFfrequency,modulationwaveform,andreceiverlocation;Chapter 5 demonstratesthatthedual-beamexperimentissensitivetostructurewithintheD-regionionosphere;andChapter 6 summarizesthepresentedmaterialandsuggestsfurthereffortsinthisarea. 1.6ScienticContributionsThefollowingscienticcontributionsaredemonstratedinthiswork: 1. Amulti-beamHFionosphericheatingmodelhasbeentestedandvalidatedusingobservationsatHAARP.Thefunctionalityofthemulti-beamHFheatingmodelhasbeensuccessfullyextendedtoaccountforvedifferentAMwaveforms,namelysquare,sinusoidal,square-root-sine(sqrt-sine),triangle,andsaw-tooth.ThemodelhasalsobeenextendedtoaccountforbothX-andO-modepolarizationoftheCWbeam. 2. IthasbeenexperimentallyestablishedthatthemagnitudeofELF/VLFwavegenerationistheparametermostsensitivetoadditionalCWheating. 3. ThetransmissionparametersthatprovideindependentinformationabouttheambientD-regionionosphereduringdual-beamheatingexperimentshavebeenexperimentallyidentied.Thesetransmissionparametersare:1)theCWpowerlevels,2)thefrequencyoftheHFCWbeam,3)themodulationwaveform,and4)thepolarizationoftheHFCWbeam. 4. InthecontextoftheHFheatingmodel,ithasbeendemonstratedthatobservationsperformedduringthedual-beamheatingexperimentaresensitivetostructurewithintheD-regionionosphere.Piecewise-exponentialelectrondensityapproximationsappeartoadequatelymatchallavailableobservations. 28

PAGE 29

CHAPTER2DUAL-BEAMELF/VLFWAVEGENERATION:MODELIMPLEMENTATIONThischapterprovidesadescriptionofthemulti-beamHFheatingmodelthatisutilizedtoprovidethemodelpredictionsinChapters 3 4 ,and 5 ofthisdissertation.Theextensionofthismulti-beamHFheatingmodeltoaccountforvedifferentAMwaveforms,X-,andO-modepolarizationoftheCWbeamisalsodescribed. 2.1NumericalAnalysisTheELF/VLFwavegenerationmodelpresentedhereinisimplementedusingtwodistinctcalculations:1)amultiple-HF-beamionosphericheatingmodelisusedtocalculatethefulltime-evolutionoftheionosphericconductivitymodulationasafunctionofspace,and2)asimpleradiationmodelisemployedtocalculatetheelectromagneticeldsatthereceiver.Thissectionprovidesadiscussionofeachofthesecalculations. 2.1.1Multi-BeamHFHeatingModelThemultiple-HFbeamionosphericheatingmodelisbasedonthesingle-beamHFheatingmodelprovidedby Moore [ 2007 ],whichhasbeenusedtosuccessfullymodelground-basedELF/VLFobservationsinanumberofworks[e.g., Payneetal. 2007 ; LehtinenandInan 2008 ; FujimaruandMoore 2011a ].Givenasetofionosphericproles,includingelectrondensityandelectrontemperatureheightproles,andgiventheparametersoftheHFheatingbeam,suchastheHFfrequency,HFpolarization,HFbeampattern,modulationfrequency,andHFpower,themodelpredictsthetime-variationofelectrontemperatureasafunctionofaltitudewithinthehighlycollisionalD-regionionosphere.Themodelaccountsfortheself-absorptionoftheHFwave[e.g., Tomko 1981 ]aswellasfornonlinearelectronenergylosses[e.g., Rodriguez 1994 ].Itneglectsanumberofionosphericprocessesthatareimportantathigheraltitudes(butthatarepresumablylessimportantintheD-region),suchaselectrondensitychangesthatmayresultfromlong-termHFheating.Theresultingvariationinelectrontemperatureisusedtocalculatethefulltime-evolutionofthe 29

PAGE 30

so-calledHall,Pedersen,andParallelconductivities,fromwhichtheamplitudesandphasesofconductivitymodulationatthemodulationfrequencyanditsharmonicsareextracted.TheambientelectrondensityandtemperatureprolesusedthroughoutthisworkareasshowninFigure 2-1 .TheseproleshavebeenextensivelyusedinpreviousELF/VLFwavegenerationanalyses[e.g., Mooreetal. 2007 ; AgrawalandMoore 2012 ; MooreandAgrawal 2011 ].Theelectrondensities(panelAofFigure 2-1 )representtenuous(I)todense(III)ionosphericconditions,withelectrondensitychangesbyfactorsof10at80kmaltitude.TheelectrontemperatureprolesshowninpanelBofFigure 2-1 arerepresentativeofayear-longsurveyofelectrontemperatureprolesprovidedbytheMSISE-90AtmosphereModelhostedbyNASAat http://ccmc.gsfc.nasa.gov/modelweb/ [ Labitzkeetal. 1985 ; Hedin 1991 ]. Figure2-1. D-region(A)AmbientElectronDensityproles(B)AmbientElectronTemperatureprolesasafunctionofaltitude. Inmultibeam-HFheatingexperiments,the1215HAARPantennaarray(Section 1.5 )isutilizedinasplit-arrayconguration.Thisdissertationfocusesondual-beamHFheating.Table 2-1 liststheHFheatingbeamparametersusedasinputstothedual-beam(615array)HF-heatingmodelthroughoutthiswork. 30

PAGE 31

Table2-1. Listof615HFsub-arrayparametersusedasinputtothemodel.Acronymsusedare:BW:Beamwidth,NS:North-South,EW:East-West,HPBW:HalfPowerBeamWidth,ERP:EffectiveRadiatedPower HF-BeamHFFrequency(MHz)Parameters3.254.55.86.9 BWNS()37.9027.7021.5018.00BWEW()13.909.907.606.40HPBW()23.1219.6712.8110.75MaxERP(dBW)78.984.286.487.9 Themodelisray-based,meaningthatalargenumberofraysareusedtocalculatethespatialextentofconductivitymodulation.Withalargeenoughnumberofruns,anyHFradiationpatternmaybemodeledincludingside-lobes,forinstance.Thetop-leftandtop-rightpanelsofFigure 2-2 areillustrativeexamplesofthe3-Dbeampatternsandthecorresponding2-DsliceoftheHFpowerpatternsfortheindividualsplitNorth615arrayat3.25MHz.Themiddlepanelshowsthe3-Dbeampatternsandthecorresponding2-DsliceoftheHFpowerpatternsfromtheSouth615arrayoperatingat5.8MHz.Inmultibeam-HFheating,thefrequencycombinationsarechosensuchthattheHFfrequencyoftheCWbeamisalwayssmallerthanthatofthemodulatedbeam,sothatthemodulatedregioniscompletelybathedwithintheCWheatedregion.Thebottompanelshowsthe3-Ddual-beampatternforthesplitnorth-southarray.Typically,thetotalnumberofmodelevaluationscanbereducedbycastingthesystemascylindricallysymmetric,althoughitisnotnecessarytodoso,ashasbeendemonstrated[ Payneetal. 2007 ].Inacylindricallysymmetricsystem,theEarth'smagneticeldisorientedperpendiculartotheEarth'ssurfaceatHAARP(15zenithangleinreality),andthisisagoodapproximationforD-regionohmicheating.TheHAARPHFheatingarrayisnotcylindricallysymmetric,however,particularlywhen6x15sub-arraysareutilized,asisthecaseinthiswork.Inordertoapproximatethesystemascylindricallysymmetric,thewidthsoftheHFbeamintheNorth-SouthandEast-Westdirectionsareusedtodeneasolidangle,andaneffectivecylindricallysymmetricbeam 31

PAGE 32

Figure2-2. 3-Dbeampatternandcorresponding2-DslicesoftheHFpowerpatternsfromHAARPforverticaltransmissionsatHAARPfor(toppanel)3.5MHz,(middlepanel)5.8MHzand(bottompanel)3.25-5.8MHzsplitarray(dual-beamtransmission).Notethatthe3-Dbeampatternforthedual-beamtransmissionlooksdistortedbecausethe3.2MHzbeam(toppanel)islowergain(solowertotalpower)thanthe5.8MHzbeam(middlepanel)andispartiallyhidden. 32

PAGE 33

widthischosensuchthatitproducesthesamesolidangle.Thischoicehastheeffectofproducingapproximatelythesametotalvolumeofmodulatedcurrents. 2.1.2TemperatureModulationThefollowingsectiondescribesthemodicationsmadetothesingle-beamHFheatingmodel[ Mooreetal. 2007 ]tocreateanewmultiple-beamheatingmodel;andthevalidityofthenewassumptionsareevaluated.WhiletheimplementationaccountsforonlytwoHFbeams,theassumptionsbuiltintothissystemareidenticaltothoseneededforasystemconsistingofmorethantwoHFbeams.Thisanalysis,therefore,maybeeasilyexpandedtoaccommodateanylargernumberofHFbeams.ThetwoprimarynonlinearitiesinvolvedinthemodulationoftheD-regionelectrontemperature,HFself-absorptionandnonlinearelectronenergylossrates,arecombinedwithinthewell-knownelectronenergybalanceequationgiveninEquation 2 [e.g., HuxleyandRatcliffe 1949 ; Maslin 1974 ; StubbeandKopka 1977 ; Tomko 1980 ; Rietveldetal. 1986 ; Rodriguez 1994 ].ForasingleHFbeam,theenergybalanceequationmaybestated: 3 2NeBdTe dt=2k(Te)S)]TJ /F2 11.955 Tf 11.95 0 Td[(L(Te,T0)(2)whereNeisthealtitude-dependentelectrondensity,BisBoltzmann'sconstant,Teisthetime-varyinglocalelectrontemperature,kistheHFfree-spacewavenumber,(Te)isthetemperature-dependentrateofabsorptionintheplasma(theimaginarypartoftherefractiveindex,n),Sisthetime-varyingpowerdensityoftheHFwave,andListhesumtotalofallelectronenergylossrates,whichdependingeneralonboththeambientelectrontemperatureT0andthetime-varyingelectrontemperatureTe.Energylossesduetoelasticcollisions[ Banks 1966 ]withrotational[ MentzoniandRow 1963 ; Dalgarnoetal. 1968 ],andvibrational[ StubbeandVarnum 1972 ; PrasadandFurman 1973 ]excitationofmolecularnitrogenandoxygenaretakenintoaccountinthemodel.Thisequationneglectsanytimevariationintheelectrondensity,andalsoneglectsheat 33

PAGE 34

conductionaswellasconvection,asistypicalforELF/VLFwavegenerationmodels.Convectioninparticularisusuallyneglectedconsideringthe1-millisecondtimescaleforELF/VLFwavegeneration.WhenaccountingfortwoHFbeams,theelectronenergybalanceequationrequiresanadditionalterm: 3 2NeBdTe dt=2k11(Te)S1+2k22(Te)S2)]TJ /F2 11.955 Tf 11.95 0 Td[(L(Te,T0)(2)wherethesubscripts,1and2,identifyquantitiesthatdependontheHFbeam.ThisadditionaltermrepresentstheenergyabsorbedbythelocalmediumfromasecondHFwave.Similarly,ifthenumberofHFbeamsisM,theelectronenergybalanceequationmaybewritten: 3 2NeBdTe dt=MXm2kmm(Te)Sm)]TJ /F2 11.955 Tf 11.95 0 Td[(L(Te,T0)(2)wheretheenergylocallyabsorbedbytheplasmafromeachoftheMwavesiscontainedwithinthesummationterm.TheHFheatingmodelsimultaneouslyandself-consistentlyaccountsforwaveabsorptionasitcalculatesthetrajectoryoftheHFraypaths(i.e.,asitperformsraytracing).WhenaccountingformultipleHFraypaths,itbecomesclearthatthefrequency-dependentrefractionandgroupvelocityofthewaveswithintheionospherewillcauseHFraysatdifferentfrequenciestobecomebothspatiallyandtemporallyseparated.ItisthusthecasethatanytwoHFwavesatdifferentfrequenciesthataretransmittedatthesametimeandwiththesameinitialtrajectorywillingeneralseparateinbothspaceandtimeasafunctionofpropagationdistance.ItisassumedthattheseeffectsarenegligibleforHFpropagationbelow100kmaltitude.Inordertoevaluatethisassumption,thetemporalandspatialseparationofHFbeamsat3.25and4.5MHzatanaltitudeof100kmforinitialHFrayanglesvaryingfrom0zenithangleisfoundforthetwelvepossiblecombinationsofelectrondensityandelectrontemperatureprolesshowninFigure 2-1 .Amongallofthevarious 34

PAGE 35

combinationsofelectrondensityandelectrontemperatureproles,themaximumlateralspatialseparationat100kmaltitudeiscalculatedtobe72meters,andthemaximumtemporalseparationiscalculatedtobe0.8microseconds.ForthepurposesofevaluatingthegenerationofELF/VLFconductivitymodulationwithintheD-regionionosphere,theseseparationvaluesarenotlikelytobesignicant.Forthisreason,themultiple-beamHFionosphericheatingmodelcalculatesthetrajectoryandtimingofeachraypathindependently,butassumestheraystobeco-locatedforthepurposesofevaluatingionosphericheatingandHFwaveabsorption.Itisnotablethatthedual-beamHFheatingmodeldoesnotautomaticallyaccountforlong-termchangesinelectrondensity.Thesechangesareexpectedtooccurontimescalesmuchlarger(byafactor>1000)thantheapproximatelymillisecondtimescalesofELF/VLFwavesofimportancetothiswork(1-3kHz).Forreference,thetheoreticalworkpresentedby MilikhandPapadopoulos [ 2007 ]predictsanelectrondensitychangebyafactorof2underlong-termHFheatingconditions. 2.1.3ConductivityModulationTheelectrontemperaturemodulationisrelatedtotheconductivitymodulationthroughanonlinearrelationship.TheconductivitytensorismostcommonlygivenbyEquation 2 [ Tomko 1981 ; Bittencourt 1986 ; Moore 2007 ]. =266664P)]TJ /F7 11.955 Tf 9.3 0 Td[(H0HP000jj377775(2)whereP,H,andjjdenotethePedersen,Hall,andParallelconductivities,respectively. P=4q2e 3meZ10avve3 av2+!ce2@fe,0 @vedve(2) H=4q2e 3meZ10!cev3e av2+!ce2@fe,0 @vedve(2) 35

PAGE 36

jj=4q2e 3meZ10v3e av@fe,0 @vedve(2)whereqeistheelectroncharge,meistheelectronmass,!istheangularfrequency,eistheeffectiveelectron-neutralcollisionfrequency,!ceisthecyclotronfrequencyandfe,0istheMaxwellianelectronvelocitydistributionfunction. Tomko [ 1981 ]and Bittencourt [ 1986 ](pg.172)provideadetaileddiscussiononthe`Maxwell-Boltzmann'or`Maxwellianvelocitydistributionfunction'whichisgivenby: fe,0=Neme 2BTe3=2exp)]TJ /F2 11.955 Tf 9.3 0 Td[(mev2e 2BTe(2)whereNeisthenumberdensityofelectrons,veistheelectronvelocity,BisBoltzmann'sconstant,andTeistheelectrontemperature. Figure2-3. AmplitudesofFirst,SecondandThirdharmonicsforHall,PedersenandParallelconductivities.Themodelpredictionsareforfc3.25MHz,Modulatedfc,5.8MHz,atmaxCWERP,100%peakpowerofmodulatedbeamusingIII-Dprole. 36

PAGE 37

Figure 2-3 representsillustrativeexamplesoftheamplitudeoftherst,secondandthirdharmonicsofthethreecomponentsoftheconductivitymodulationtensor.Hallconductivityshowsaltitudesofmaximumconductivitymodulationbetween80and85km,whichrepresentsthedominantsourcealtitude.ThePedersenconductivityshowstworegionsofincreasedconductivitymodulationwiththedominantsourceregionat85km.WhiletheParallelconductivityhasthestrongestamplitudeathigheraltitudescomparedtoHallandPedersen,thisquantityconvenientlycancelsforground-basedobservationsaswewillseeinSecion 2.2 2.2RadiationInordertocalculatethemagnitudeoftheelectromagneticwaveatthereceiver,theHall,Pedersen,andParallelcurrentsintheionospherearecalculated.Theamplitudesandphasesoftheconductivitymodulationcalculatedusingthedual-beamHFheatingmodelareinterpolatedontoaregularrectangulargridwith1-kmspacing(Figure 2-4 )andmultipliedbytheelectriceldoftheauroralelectrojet.ThespatialdistributionoftheelectrojeteldisassumedtobeconstantthroughouttheD-regionandorientedparalleltotheground,consistentwithpasttheoreticalworkandexperimentalobservations[e.g., BanksandDoupnik 1975 ; StubbeandKopka 1977 ; Papadopoulosetal. 2003 ; Payne 2007 ].ThemodelinherentlyaccountsforthegroupdelaysoftheHFwavestoeachoftheionosphericgridpointsandalsoaccountsforthetemporalreactionoftheplasmatothehigh-powerHFsignals.BothoftheseeffectsmodifythephaseoftheELFsourcecurrentsasafunctionofspace.Becausetheexperimentalobservationsperformedareallrelativeobservations,theactualmagnitudeoftheelectrojeteldstrengthdoesnotmatterinourcase.ComplicationsarisewhenoneconsidersthatD-regionheatingbytheelectrojetcurrentsthemselvesmayaffecttheamplitudeoftheradiatedELFelds,however.Itisagoodassumptionthattheelectrojeteldstrengthcanvarybetween5and100mV/m[e.g., BanksandDoupnik 1975 ; Stubbeetal. 1981 ; Papadopoulosetal. 37

PAGE 38

Figure2-4. 3-Dmodulatedheatedregion(orangecone:Modulatedheatedregion,gray:CWheatedregion)andtherectangularcuboid(blue).Thegreenvoxels,representthevolumeelementontowhichtheamplitudesandphasesofconductivitymodulationareinterpolated. 2003 ; Payne 2007 ].Figure 2-5 indicatesthattheconductivitymodulationproducedbymodulatedHFheatingtogetherwitha5mV/melectrojeteldstrengthislessthan0.2-dBdifferentthanthatcalculatedwitha100mV/melectrojeteld.ItisthusagoodapproximationtoassumethattheELFsourcecurrentsvarylinearlywiththeelectrojeteldstrength.AlthoughthismodelassumesthattheelectronenergydistributionremainsMaxwellianthroughouttheheatingprocess,theminimalimpactofelectrojetheatingonthecalculatedconductivitymodulationindicatesthatthisassumptionislikelytobevalidevenforfullykineticmodels. 38

PAGE 39

Figure2-5. ThedBdifferenceinconductivitymodulationasafunctionofaltitudefora5-mV/manda100-mV/melectrojetelectriceld. WiththedistributionofionosphericELFsourcecurrentsinhand,themagneticeldiscalculatedforagivenground-basedreceiverlocationassumingthattheELFsourcetakestheformofspatially-distributedsetofdipolesoveragroundplane[e.g., Payne 2007 ].Forreceiverlocationswithin75kmofHAARP,thisisagoodapproximation: Payne [ 2007 ]demonstratedthatwithin75kmofHAARPthemagnitudeofthemagneticeldcalculatedusingadistributeddipolemodelcloselymatchesthatcalculatedusingamorecompletemodelthataccountsforEarth-ionospherewaveguideeffectsandforthesecondaryionosphericcurrentsgeneratedduringthemodulatedheatingprocess.Inordertomitigatetheerrorsassociatedwithneglectingtheseeffects,theobservationsarenormalizedtothemagneticeldobservedduringCW-OFFperiods.ThisnormalizationcancelstheeffectsoftheEarth-ionospherewaveguideandthevaryingelectrojeteldstrengthtorstorder[ BarrandStubbe 1993 ],allowingforthedirectcomparisonofobservationsandmodelingresults.Itshouldbenoted,however,thattheeffectsoftheEarth-ionospherewaveguideareimportant.Forinstance,pronouncedwaveguideresonancesatmultiplesof2kHzhavebeenobservedinELF/VLFamplitudedata[e.g., Stubbeetal. 1982 ; BarrandStubbe 1984 ; Rietveld 39

PAGE 40

etal. 1989 ],andmultipleionosphericreectionshavebeendirectlyobservedduringELF/VLFpulsed-heatingexperiments[e.g., Papadopoulosetal. 2005 ; FujimaruandMoore 2011a ]. FujimaruandMoore [ 2011a ],showedviaTOA(TimeofArrival)thatthelineofsightpathcontributioncanbeseparatedfromtherstionosphericallyreectedcomponentusingchirped-frequencymodulation.ItisveryclearlythecasethattheEarth-ionospherewaveguideaffectstheamplitudeandphaseoftheELF/VLFsignalreceivedontheground.Inthisdissertation,weusethenormalizedELF/VLFB-elds.Becausewepresentonlynormalizedeldvaluesinthisdissertation,itisworthnotingthatthismodelisfullycapableofpredictingabsoluteeldstrengths.FortheprolesshowninFigure 2-1 ,andforreasonable(5mV/m)valuesoftheelectrojeteldstrength,themagneticeldmagnitudespredictedatareceiverlocated33kmfromHAARPvaryfrom40fTto20pT(Figure 2-6 ),whichareveryreasonableeldvaluescomparedtopastandpresentexperimentalobservations[e.g., Stubbeetal. 1982 ; Rietveldetal. 1986 1989 ; Villasenoretal. 1996 ; Cohenetal. 2010 ].Therangeofpossibleeldvaluescanbeeasilyextendedusingagreatervarietyofionosphericproles.Asmentionedabove,however,absoluteeldmeasurementsdependonEarth-ionospherewaveguideeffects,onthegenerationofsecondaryionosphericcurrents,andonthestrengthoftheelectrojetcurrents,inadditiontotheconductivitymodulationproducedbyHFheating.InordertoisolatetheeffectsofHFheating,thisdissertationexclusivelypresentsnormalizedeldobservations.Forthegivenassumptions,themagnitudeofthepredictedELF/VLFB-eldonthegroundisgivenbyp B2H+B2P,whereBHandBParetheamplitudesoftheB-eldsgeneratedbytheHallandPedersencurrents,respectively.TheassumptionthattheEarth'smagneticeldisperpendiculartothegroundmeansthattheparallelconductivityisalsoperpendiculartotheground.Asaresult,theradialandazimuthalcomponentsofthethedirect-pathandground-reectedB-eldsgeneratedbytheparallelconductivitycancel,andtheparallelB-elddoesnotplayaroleinthepredictedELF/VLFmagnitude. 40

PAGE 41

Figure2-6. AbsolutemagneticeldstrengthatSinonaCreek(SC)forelectrojeteldof5mV/mand100mV/m. Thequantityisconvenientlyindependentoftheorientationsofthereceiverantennas.ItisusefultoinspecttheeffectofCWheatingonthegenerationoftheHallandPedersenB-eldsindependently,however.EvenforrealisticB-elddirections(15off-zenithatHAARP),theparallelconductivitydoesnotcontributetothemagnitudeundertheassumptionthatthehighconductivityinthedirectionofthemagneticeldcancelsoutthemodulatedcurrentsinthatdirection[ Cohenetal. 2012 ].Figures 2-7 and 2-8 showtherstharmonicamplitudesoftheB-eldsgeneratedbytheHallandPedersencurrentsasafunctionofspacewithinthelowerionospherewitha1-kmgridspacing.Thecolorsrepresenttheamplitude(indB)oftheELF/VLFwaveobservedatthereceiverandgeneratedbyadipoleattheplottedlocation.Thelefthandpanelsshowtheeldamplitudesgeneratedundermodulatedsingle-beamheatingconditions(CW-OFF;sinusoidalAMat4.5MHz,X-mode),andtherighthandpanelsshowtheeldamplitudesgeneratedunderdual-beamheatingconditions(CW-ONat3.25MHz,X-mode;sinusoidalAMat4.5MHz,X-mode).Resultsareshownasafunctionofelectrondensity,fromProleI(top)toProleIII(bottom),forasingleelectrontemperatureprole(ProleB).Figuresdepictingthedependenceontheelectrontemperatureprole(notshown)demonstrateessentiallythesamespatialdistribution 41

PAGE 42

Figure2-7. Numericalpredictions.Firstharmonic(1215Hz)HallcurrentB-eldamplitudesasafunctionofsourcelocation.CW-ONandCW-OFFperiodsareshownasafunctionofelectrondensityprole. ofeldsshownhere,althoughtheabsolutemagnitudesaredifferent.WenotethatthespatialdistributionoftheB-eldsshownisessentiallycylindricallysymmetric.Inthiscase,thesymmetryresultsfromthefactthatthereceiverislocatedverycloseto(1.5kmfrom)theorigin,inadditiontothefactthattheconditionofcylindricalsymmetrywasenforcedinthecalculationoftheconductivitymodulation.Forinstance,plotscalculatedforareceiverdistantfromHAARPwouldshowhigheramplitudesinthedirectionofthereceiver. 42

PAGE 43

Figure2-8. Numericalpredictions.Firstharmonic(1215Hz)PedersencurrentB-eldamplitudesasafunctionofsourcelocation.CW-ONandCW-OFFperiodsareshownasafunctionofelectrondensityprole. Figures 2-7 and 2-8 demonstrateapronounceddependenceontheelectrondensityprole.UnderbothCW-OFFandCW-ONconditions,thespatialdistributionofbothHallandPedersencurrentsbecomesmorecompactinaltitudeastheelectrondensityvariesfromProleItoProleIII.Additionally,thealtitudeofthepeakamplitudedecreasessignicantly(by5kmperprole)andthepeakamplitudeitselfincreasessharply(by15dBperprole)astheelectrondensityvariesfromProleItoProleIII.Adetailed 43

PAGE 44

analysisontheaffectsofelectrondensityandelectrontemperatureprolesisprovidedinChapter 4 .ThedependenceonCW-heatingisalsoveryclearlydepictedinFigures 2-7 and 2-8 .Inallcases,theadditionofahigh-powerCWheatingbeampushesthewavegeneratingcurrentsupwardsandoutwards.Theaveragealtitudeofwavegenerationincreases,andthevolumeofradiatingcurrentsdecreasesatloweraltitudeswhileatthesametimeincreasesathigheraltitudes.AdetailedanalysisofthisaffectisdiscussedinChapter 5 .TheB-eldamplitudesaredramaticallyreducedatloweraltitudes,butinsomecasestheyincreaseslightlyathigheraltitudes.Inallcases,thepeakamplitudedecreasesunderCW-ONconditions.Figures 2-7 and 2-8 demonstratethatunderavarietyofelectrondensityandelectrontemperature(notshown)conditions,additionalCWheatingatpeakpowerwilltendtoreducethemagnitudeoftheELF/VLFB-eldreceivedontheground.Figures 2-9 and 2-10 showsthespatialdistributionofthesecondharmoniccomponentsoftheHallandPedersencurrentsforcomparisonwiththerstharmoniccomponents.ItisnotablethatduringCW-OFFperiods,thespatialdistributionofthesecondharmonicoftheHallB-eldisverysimilartothatoftherstharmonic,althoughthesecondharmonicofthePedersenB-eldissomewhatlowerinaltitudethantherstharmonic.DuringCW-ONperiods,however,thesecondharmonicoftheHallB-eldispushedhigherinaltitudethantherstharmonic,whereasthespatialdistributionofthesecondharmonicofthePedersenB-eldisverysimilartothatoftherstharmonic.ItisthusthecasethatduringbothCW-ONandCW-OFFperiods,thereisatleastonemajorcomponentoftheELF/VLFmagnitudesthathasaverysimilarspatialdistributionforboththerstandsecondharmonic,indicatingthattheELF/VLFsecondtorstharmonicrationaturallyminimizessecondordereffectsaswellasrstordereffects.AmoredetailedanalysisontheELF/VLFharmonicratioisdiscussedinChapter 3 44

PAGE 45

Figure2-9. Numericalpredictions.Secondharmonic(2430Hz)HallcurrentB-eldamplitudesasafunctionofsourcelocation.CW-ONandCW-OFFperiodsareshownasafunctionofelectrondensityprole. Havingprovidedadescriptionofthemodelimplementationformulti-beamHFheatingoftheionosphere,thenextsectiondescribestheextensionofthemulti-beamHFheatingmodeltoaccountfordifferentmodulationwaveforms. 2.3ExtensiontoOtherAMWaveformsInChapter 4 ofthisdissertation,theionosphereismodulatedusingasplit-arraycongurationandvedifferentmodulationwaveforms,(square,sinusoid,sqrt-sine,triangleandsaw-tooth)areemployed.Thephysicsandmodelimplementationleading 45

PAGE 46

Figure2-10. Numericalpredictions.Secondharmonic(2430Hz)PedersencurrentB-eldamplitudesasafunctionofsourcelocation.CW-ONandCW-OFFperiodsareshownasafunctionofelectrondensityprole. tothegenerationofELF/VLFwavesremainsthesame,asexplainedintheprevioussections,howevertheaffectofusingdifferentmodulationwaveformsismanifestedinthecalculationofS,thetime-averagedPoyntinguxofthemodulatedHFwave,makingsuretoaccountforthecontinuousordiscontinuousnatureofthemodulationwaveformproperly.Figure 2-11 showsmodelpredictionsforTemodulationatfourdifferentaltitudes,fortwosteadystateperiodsoftheELF/VLFwavegeneration.Themodelpredictions 46

PAGE 47

Figure2-11. ElectronTemperaturemodulation(at1225Hz)asafunctionofaltitudeforvedifferentmodulationwaveforms. shownareforamodulatingfrequencyof1225Hz,usingambientconditionsIII-D,forCWHFfrequency3.25MHz(78.9dBW)andmodulatedHFfrequency5.8MHz(86.4dBWpeakpowerand100%modulationdepth).Themaximumandminimumtemperaturesattainedatsteadystateasafunctionofaltitudeisnotaffectedbythemodulationwaveformused.MoredetailsonELF/VLFmagnitudegeneratedbydifferentmodulationwaveformscanbefoundinSection 4.3.4 .Theprimaryphysicaleffectofdual-beamheatingoftheionosphereatdifferentCWpowerlevelsisdescribedinFigure 2-12 .Thegureillustratesthemaximum(redtraces)andminimum(bluetraces)electrontemperaturesachievedasafunctionofaltitude 47

PAGE 48

Figure2-12. MaxandminTeasafunctionofaltitudeforvedifferentmodulationwaveforms. Figure2-13. (A)MaximumandminimumTeachievedduringtheheatingcycleforprolecombinationIII-DasafunctionofCWERPforsquarewaveAM.(B)ThecorrespondingamplitudeofHallconductivitymodulationasafunctionofCWERP. 48

PAGE 49

duringsinusoidalsteadystateforeachofthevemodulationwaveforms.Thedashedtraces,correspondtothesimulationswithCWat100%power,andthesolidtracescorrespondtosimulationswithCWat0%power.AstheCWpowerincreasesfrom0tofullpower,theminimumelectrontemperature(transitionfromsolidbluetodashedblue)increasestoagreaterextentthanthemaximumelectrontemperature(transitionfromsolidredtodashedred),resultinginanoverallreductioninelectrontemperaturemodulation.PanelAofFigure 2-13 demonstratedthiseffectasafunctionofvariousCWERPlevels.IncreasedCWheatingresultsinthesignicantreductionoftheHallconductivitymodulationatloweraltitudes(below85km)asisdemonstratedbypanelBofFigure 2-13 .ThereductioninconductivitymodulationresultsintheoverallreductionofmagneticeldstrengthobservedatthereceiverasafunctionofCWERP.DetailsofmodulationaffectsasafunctionofCWERP,arediscussedingreatdetailinChapter 3 andChapter 4 ofthisdissertation. 2.4ExtensiontoX-andO-ModePolarizationoftheCWBeamItiswellknownthatmodulatedX-modeheatingproduceshigheramplitudeELF/VLFwavesthatdoesmodulatedO-modeheating[ Stubbeetal. 1981 1982 ; Ferraroetal. 1984 ; Jamesetal. 1984 ; Villasenoretal. 1996 ].InordertoincreasetheSNRoftheobservations,wechoosetouseanX-modepolarizedHFbeamtomodulatetheionosphericconductivityforallexperimentsinthisdissertation.HoweverinSection 4.3.6 ,wetheoreticallyinvestigatetheeffectsofusingbothX-andO-modepolarizationontheCWbeam.TheprimarydifferencesinevaluatingELF/VLFwavegenerationusingX-andO-modepolarizationoftheCWbeamisinthecalculationoftherefractiveindex(n)andtherefractionoftheHFraythatresults[ Budden 1985 ,p.110].Therefractiveindex,nofaweaklyionized,magnetoplasmaisgivenbythewellknownAppleton-HartreeEquation: 49

PAGE 50

n2=1)]TJ /F2 11.955 Tf 124.98 8.08 Td[(X 1)]TJ /F2 11.955 Tf 11.95 0 Td[(jZ)]TJ /F6 7.97 Tf 19.49 4.71 Td[(Y2sin2 2(1)]TJ /F6 7.97 Tf 6.59 0 Td[(X)]TJ /F6 7.97 Tf 6.59 0 Td[(jZ)q Y4sin4 4(1)]TJ /F6 7.97 Tf 6.59 0 Td[(X)]TJ /F6 7.97 Tf 6.59 0 Td[(jZ)2+Y2cos2(2)wheretheterm,`+'termrepresentsO-modepolarization,`)]TJ /F1 11.955 Tf 9.29 0 Td[('representsX-modepolarization,istheanglebetweenthewavenormalandtheEarth'smagneticeld,and: X=!2pe !2Y=!ce !Z=e !(2)ThetermsX,YandZrepresentpropertiesoftheplasma.X,whichisafunctionoftheelectron-plasmafrequency!pe,canbethoughtofasthenaturalfrequencyatwhichtheelectronsareoscillatingintheplasmaandisrelatedtotheelectrondensity,Ne,thechargeofanelectron,qe,thepermittivity,0andthemassoftheelectron,me,throughEquation 2 : !pe=s Neqe2 0me(2)Ydependsontheelectrongyrofrequency,!ce,whichisthefrequencyatwhichtheelectronsgyrateaboutanexternallyappliedmagneticeld,suchastheEarth'smagneticeldinourcase. !ce=jqejB0 me(2)Lastly,Zrelatestheeffectiveelectron-neutralcollisionfrequency,eandthefrequencyofthepropagatingHFwave.Itisthistemperature-dependentquantity,ethatprimarilyundergoesmodulationintheHFheatingprocess.Therefractiveindex,ncanalsobeexpressedasn=)]TJ /F2 11.955 Tf 13.08 0 Td[(j,wheretherealpart,drivesthewavepropagationandtheimaginarypart,,determinestherateofabsorption.Thedeviationoftheraydirectionfromthek-vectordirectioniscalculatedas: 50

PAGE 51

tan=1 n@n @(2)and1 n@n @canbeobtainedbytakingthepartialdifferentialofEquation 2 withrespecttoasfollows: 1 n@n @=1 2n2"X 1)]TJ /F2 11.955 Tf 11.95 0 Td[(jZ)]TJ /F6 7.97 Tf 19.49 4.71 Td[(Y2sin2 2(1)]TJ /F6 7.97 Tf 6.59 0 Td[(X)]TJ /F6 7.97 Tf 6.58 0 Td[(jZ)q Y4sin4 4(1)]TJ /F6 7.97 Tf 6.59 0 Td[(X)]TJ /F6 7.97 Tf 6.59 0 Td[(jZ)2+Y2cos22#")]TJ /F2 11.955 Tf 9.3 0 Td[(Y2 (1)]TJ /F2 11.955 Tf 11.96 0 Td[(X)]TJ /F2 11.955 Tf 11.96 0 Td[(jZ)sincos1 21 q Y4sin4 4(1)]TJ /F6 7.97 Tf 6.58 0 Td[(X)]TJ /F6 7.97 Tf 6.58 0 Td[(jZ)2+Y2cos2"Y4 (1)]TJ /F2 11.955 Tf 11.96 0 Td[(X)]TJ /F2 11.955 Tf 11.96 0 Td[(jZ)2sin3cos)]TJ /F5 11.955 Tf 11.95 0 Td[(2Y2sincos##(2)wheretheterm,`+'termrepresentsO-modepolarization,`)]TJ /F1 11.955 Tf 9.29 0 Td[('representsX-modepolarization,istheanglebetweenthewavenormalandtheEarth'smagneticeld.ByproperlyaccountingforthepolarizationoftheCWbeamintherefractiveindexandthebendingoftheHFray,weevaluatethemodelpredictionsforbothX-andO-modeCWpolarizedbeams. 51

PAGE 52

CHAPTER3ELF/VLFWAVEGENERATIONUSINGSIMULTANEOUSCWANDMODULATEDHFHEATINGOFTHEIONOSPHEREInthischapter,themultibeam-HFheatingmodeldescribedinChapter 2 isutilizedtogetherwithgroundbasedELF/VLFobservationstoidentifythepropertyofthereceivedwavethatismostsensitivetotheeffectsofthemultibeam-HFheating.ThreepropertiesoftheELF/VLFwavesareassessed:theELF/VLFsignalmagnitude,theELF/VLFharmonicratio,andtheELF/VLFpower-lawexponent.Ground-basedexperimentalobservationsindicatethatsimultaneousheatingoftheionospherebyaCWHFwaveandamodulatedHFwavegeneratessignicantlylowerELF/VLFmagnitudesthanduringperiodswithoutCWheating,consistentwithmodelpredictions.TheratioofELF/VLFharmonicmagnitudesisalsoshowntobeasensitiveindicatorofionosphericmodication,althoughitissomewhatlesssensitivethantheELF/VLFmagnitudeandrequireshighSNR.Lastly,thepeakpowerlevelofthemodulatedHFbeamisvariedinordertoassessthepowerdependenceofELF/VLFwavegenerationunderbothsingle-anddual-beamheatingconditions.ExperimentalandtheoreticalresultsindicatethataccurateevaluationoftheELF/VLFpower-lawindexrequireshighsignal-to-noiseratio;itisthusalesssensitiveindicatorofionosphericmodicationthaneitherELF/VLFmagnitudeortheELF/VLFharmonicratio.TheworkpresentedinthischapterhasbeenpublishedintheJournalofGeophysi-calResearch[ MooreandAgrawal 2011 ]. 3.1DescriptionoftheExperimentDuringathirty-minuteperiodbetween0830and0900UTon2August2007,the1215HAARPHFtransmitterarraywasdividedintotwo615sub-arrays,eachwithapeakpowerof1800kW.Onesub-arraywasusedtogenerateELF/VLF-wavesintheionospherebytransmittingasinusoidalAMbeamat4.5MHz(X-modepolarization).Themodulationfrequencyalternatedbetween1215Hzand2430HzandthepeakHFpowerwassteppedin15distinctlog-basedsteps(from)]TJ /F1 11.955 Tf 9.29 0 Td[(12.5dBto0dBwith 52

PAGE 53

Figure3-1. Acartoondiagramofthedual-beamHFheatingexperiment.The3.25MHzCWbeamisbroaderthanthe4.5MHzmodulatedbeam. 1secondateachpowerlevel).Simultaneously,thesecondbeamoftheHAARPHFtransmittercontinuallyheatedthesamepatchofionosphereatpeakpowerat3.25MHz(CW,X-mode)foraperiodof8minutes.AlowerHFfrequencywasselectedfortheCWbeamsothattheCWbeampatternwouldbebroaderthanthatofthemodulated4.5MHzHFbeam.The8-minuteCWtransmissionblockwasfollowedbya7-minuteperiodwithoutCWheating(i.e.,therstbeamcontinuedtomodulateat4.5MHzwhilethesecondbeamwasOFF).AcartoondepictionoftheHFbeamcongurationcanbeseeninFigure 3-1 ,andadiagramofthemodulationfrequencyandHFpowerformatcanbeseeninFigure 3-2 .Thegainsofthetwosub-arraysdependonthefrequenciestransmitted.Forthepurposesofmodeling,wehaveapproximatedthepeakeffectiveradiatedpower(ERP)levels(using6x15sub-arrays)tobe78.9dBWat3.25MHzand84.2dBWat4.5MHz.The15-minuteexperimentwasrepeatedtwiceduringthe30-minutewindow,andtheKPindexwas2atthistime.ELF/VLFwaveobservationswereperformedataground-basedreceiverlocatedattheHAARPobservatory,approximately1.5kmfromtheHFtransmitter.Theradialandazimuthalcomponentsofthemagneticeldweremonitoredcontinually.In 53

PAGE 54

Figure3-2. Thetransmissionscheduleforthe4.5MHzmodulatedHFbeam.(A)Themodulationfrequency(sinusoidalAM)asafunctionoftime.(B)Thepeakpoweremployedasafunctionoftime.100%modulationdepthwasusedforallcases.Thetwopanelssharethesametimeaxis.This30-secondschedulerepeatedcontinuallyfor30minutes. post-processing,thenarrowbandELF/VLFamplitudesandphasesatthemodulationfrequenciesandtheirharmonicsweredeterminedusing1-second-longdiscreteFouriertransforms. 3.2ExperimentalObservationsFigure 3-3 showsthemagnitudeoftheELF/VLFsignalobservedatHAARPat1215Hzand2430Hzfortheentire30-minutedurationoftheexperiment,withtheCW-ONandCW-OFFperiodsindicatedusingagrayandwhitebackground,respectively.Good(>10dB)signal-to-noiseratios(SNR)wereobservedduringtherst15minutesoftheexperiment.Duringthesecond15-minuteperiod,theSNRdecreasedbyapproximately5dB.PanelBofFigure 3-3 showsseveralpower-stepseriesduringaCW-OFFperiodwithamagniedtimescale.From0seconds 54

PAGE 55

Figure3-3. ThemagnitudeofELF/VLFsignalsobservedattheground-basedreceiveratHAARP.(A)Alldataforthe30-minutedurationoftheexperiment,withCW-ONandCW-OFFperiodsindicatedwithgrayandwhitebackgrounds,respectively.(B)Severalexamplesofpower-stepseriesobservedduringaCW-OFFblock. 55

PAGE 56

andfrom30seconds,therstandsecondharmonics(at1215Hzand2430Hz,respectively)clearlyincreasewithtransmittedpower,whichincreaseslogarithmicallyover15seconds.From15secondsandfrom45seconds,asimilartrendisobserved,exceptthatthe2430Hzsignalistherstharmonic(i.e.,themodulationfrequencywas2430Hz).Whenthesecond-to-rstharmonicratioiscalculatedlaterinthissection,wewilldividethesecondharmonicmagnitudeat2430Hz(generatedbythe1215Hztransmission)bytherstharmonicmagnitudeat2430Hz(generatedbythe2430Hztransmission),effectivelycanceling(torstorder)thefrequency-dependenteffectsoftheEarth-ionospherewaveguide.Theslopesofthesetraces(effectivelyonlog-logscale)quantifythedifferentialincreaseofELF/VLFmagnitudewithpeakmodulatedHFpower.WewillnowdiscusstheobservedELF/VLFmagnitudesindetail. 3.2.1ELF/VLFMagnitudeTheleftpanelsofFigure 3-4 showtheprobabilitydensityfunctions(PDFs)fortheSNR,thereceivedELF/VLFmagnitude,andtheratioofELF/VLFmagnitudesrecordedduringCW-OFFperiodstothoserecordedduringCW-ONperiodsforobservationsat1215Hz.Therightpanelsshowthesametracesforobservationsat2430Hz.ThesePDFsarecalculatedusingallavailabledatapoints(includingallpowersteps)inordertoprovidestatisticalsignicance.TocalculatetheSNR,noiselevelsaredeterminedbyextractingtheELF/VLFmagnitudesat1170Hz(forcomparison1215Hz)andat2390Hz(forcomparisonwith2430Hz)asafunctionoftime.At1215Hz,theSNRisonaverage5.6dBhigherduringCW-OFFperiodsthanduringCW-ONperiods.At2430Hz,theSNRis4.4dBhigherduringtheCW-OFFperiodsthanduringtheCW-ONperiods.TheresultsaresimilarforELF/VLFmagnitudes:at1215Hz,theELF/VLFmagnitudeisonaverage4.9dBhigherduringCW-OFFperiods,andat2430Hz,itis4.1dBhigher.ThebottompanelsshowtheCW-OFFtoCW-ONELF/VLFmagnituderatio,calculatedbydividingtheobservedmagnitudesseparatedby8minutesintime.Forreference, 56

PAGE 57

Figure3-4. Statisticaldistributions(probabilitydensityfunctions).leftcolumn:1215Hz.rightcolumn:2430Hz.top:SNR.middle:magnitude.bottom:CW-ONtoCW-OFFratiocalculatedasdescribedinthetext. thenoisehasbeenprocessedinthesamemanner.Asexpected,theaverageresultsaresimilar:a4.9dBratioisobservedat1215Hz,anda4.5dBratioisobservedat2430Hz.ThefactthatthesedistributionsareverysimilartoeachothersupportsthestatementthattheELF/VLFmagnitudeissignicantlyreducedbyadditionalCWheating.InordertoassesstheeffectsofCWheatingonthemagnitudeofthereceivedELF/VLFsignalasafunctionoftime,weselectthemagnitudeoftherstharmonicatthepeakpowertransmission(i.e,the15thpowerstep).ThisselectionsuppliesobservationswiththehighestSNR.Thesemagnitudesareavailableonceevery30seconds,andtheyareshowninFigure 3-5 forboth1215Hzand2430Hz.ThemagnitudesexhibitanaturalvariationontheorderofseveraldBoverthe30-minuteexperiment.Thisvariationislikelydominatedbythevaryingstrengthoftheauroralelectrojetcurrents,butalsomaybeduetovariationsinelectrondensityandelectron 57

PAGE 58

Figure3-5. Themagnitudeofthe1stharmonicobservedduringonlythepeakpowertransmissionsthroughoutthe30-minuteexperiment.Foreachtrace,thereisonesampleevery30seconds. temperatureintheD-regionionosphere.Ontheonehand,thevariationsinionosphericparametersmaybeproduceddirectlybyHFheating;ontheotherhandtheymayalsooccurnaturally,produced,forinstance,byenergeticelectronprecipitationorothernaturalphenomena.TheobservedchangeinELF/VLFmagnitudebetweenCW-ONandCW-OFFperiods,however,canbedirectlyattributedtoHFheating.At1215Hzand2430Hz,themagnitudesoftheELF/VLFsignalsincreaseby8.6and8.1dB,respectively,whentheCWbeamisturnedoff.WhentheCWbeamisturnedonagainsevenminuteslater,the1215Hzand2430Hzmagnitudesdecreaseby7.4and9.1dB,respectively.ObservationsduringthesecondhalfoftheexperimentsufferfromlowSNR,althoughthedataarenotinconsistentwithobservationsperformedduringtherst15minutesoftheexperiment:theELF/VLFeldmagnitudesarestillhigherduringtheCW-OFFperiodthanduringtheCW-ONperiod.Thelarge(7dB)changesinELF/VLFmagnitudebetweenCW-ONandCW-OFFperiodsindicatethattheELF/VLFmagnitudemaybeusedasaverysensitiveindicatorofionosphericmodicationandthatmoredetailedexperimentsmaybeperformed. 58

PAGE 59

ByalternatingbetweenCW-ONandCW-OFFperiodsoncepersecond(orfaster),thechangeinELF/VLFmagnitudemaybetrackedasafunctionoftime,yieldinginsightintothevariationofionosphericparameterswithmuchhighertimeresolutionthanavailableinthepresentedexperiment.Additionally,thepowerandfrequencyoftheCWsignalmaybevaried,resultingindifferentchangesinELF/VLFmagnitudebetweenCW-ONandCW-OFFperiods.Basedonthelarge(7dB)changesinELF/VLFmagnitudepresentedinthischapter,itislikelythatthesesuggestedexperimentswouldyieldmeasurabledistinctchangesELF/VLFmagnitude,andwedirectlyassessthispossibilityinthemodelingsectionofthechapterandinChapter 4 .Wenowmoveontodiscussanothersensitiveexperimentalmethodtodetectchangesinionosphericproperties. 3.2.2ELF/VLFHarmonicRatioInthiswork,theELF/VLFharmonicratioistheratioofthesecondharmonicmagnitudetotherstharmonicmagnitude.Forthepresentedexperiment,ELF/VLFwavesweregeneratedusingsinusoidalamplitudemodulation,andthepowerenvelopeoftheHFtransmissionthusconsistsofarstandsecondharmonic.Ifsquare-waveamplitudemodulationhadbeenused,forinstance,anequivalentmeasurewouldbethethirdharmonictorstharmonicratio.Inanycase,theratioofthesetwomagnitudesgeneratedatthesametimeessentiallycancelsstrengthoftheauroralelectrojetcurrents(torstorder).PropagationwithintheEarth-ionospherewaveguideisstronglyfrequency-dependent,however.InordertocanceltheeffectsoftheEarth-ionospherewaveguide(torstorder),werequirethatthetwomagnitudesbemeasuredatthesamefrequency.Itisimpossibletodiscernrstandsecondharmonicsgeneratedatthesamefrequencyatthesametime,however.Asareasonableapproximation,wegeneratethesecondharmonicat2430Hzusinga1215Hztone,andtherstharmonicashorttimelaterusinga2430Hztone. BarrandStubbe [ 1993 ]usedthismethodforevaluating 59

PAGE 60

harmonicratioswithgreatsuccesstocancel(torst-order)thefrequency-dependenteffectsoftheEarth-ionospherewaveguide.BecausetheSNRofthesecondharmonicisnotparticularlyhighthroughouteachofthe15powersteps,weusethesecondharmonicmagnitudeduringonlythepeakpowerstep(6dBSNR)toanalyzethesecond-to-rstharmonicratio.Figure 3-6 showsthevariationoftheharmonicratiooverthecourseofthe30-minuteexperiment.DuringtherstCW-ONperiod,theharmonicratioisessentiallyconstantat)]TJ /F5 11.955 Tf 9.3 0 Td[(14.050.4dB.Weattributethetwosharp,buttemporary,deviationsfromthisleveltolightning-generatedsfericscouplingintothebandratherthantochangesinthepropertiesoftheionosphere.WhentheCWbeamturnsoff,however,theharmonicratioimmediatelydecreasesby3.75dBto)]TJ /F5 11.955 Tf 9.29 0 Td[(17.80.7dB.DuringtheCW-OFFperiod,theratioagainremainsrelativelyconstant,withtwosharpdeviationsthatlikelyresultfromlightning.Duringthesecond15-minuteperiodoftheexperiment,observationssufferfromlowSNR.Despitethisfact,somecomparisonscanbemade.UponturningtheCWbeamONforthesecondtime,theharmonicratioimmediatelyincreasesby4.5dB.DuringthesecondCW-ONperiod,theharmonicratiouctuatesrapidlybetween)]TJ /F5 11.955 Tf 9.29 0 Td[(12and)]TJ /F5 11.955 Tf 9.3 0 Td[(15dBduetolowSNR,althoughwenotethatthe)]TJ /F5 11.955 Tf 9.29 0 Td[(12to)]TJ /F5 11.955 Tf 9.3 0 Td[(15dBrangeincludesthe)]TJ /F5 11.955 Tf 9.3 0 Td[(14dBlevelobservedduringtherstCW-ONperiod.Approximately1minuteintothesecondCW-OFFperiod,theSNRincreasessomewhat,andtheharmonicratioremainscloseto)]TJ /F5 11.955 Tf 9.29 0 Td[(18dBfortheremainderoftheperiod,similartotherstCW-OFFperiod.TheassumptionthattheharmonicratioisasensitiveindicatorofionosphericchangeundergoodSNRconditionswillbeevaluatednumericallyinthemodelingsectionofthischapter.Becausetheharmonicratioisevaluatedonlyonceevery30seconds,theimmediacyofthe)]TJ /F5 11.955 Tf 9.29 0 Td[(3.75dBchangeandthe+4.5dBchangeatCW-ON/CW-OFFboundariescanonlybestatedwith30-secondresolution.Thismayeasilybeimprovedduringfutureexperiments,however,byomittingthepower-steppingfeatureofthepresentedexperiment. 60

PAGE 61

Figure3-6. Ratioofthesecondharmonicmagnitudetotherstharmonicmagnitude,calculatedasdiscussedinthetext,observedduringonlythepeakpowertransmissionsthroughoutthe30-minuteexperiment.Thereisonesampleevery30seconds. WewillnowdiscusstheobserveddependenceofELF/VLFmagnitudeonHFpower. 3.2.3ELF/VLFPower-LawExponentIntheearly1990's, Papadopoulosetal. [ 1990 ]and BarrandStubbe [ 1991 ]suggestedthattheELF/VLFmagnitudedependsonthepeakinputHFpowerasapowerlawwithindexn:AELF/PnHF.Inthiscontext,wewillrefertotheindexnastheELF/VLFPower-LawExponent(EPLE),whichshouldinprincipledependontheambientpropertiesoftheD-regionionosphere.Foreachpower-stepseriesperformedinourexperiment,theEPLEiscalculatedusingaweightedleast-squarettotheobservedELF/VLFmagnitudeasafunctionofHFpower:n=(PTHFWTWPHF))]TJ /F9 7.97 Tf 6.58 0 Td[(1PTHFWTWAELF,withtheweightsofthematrixWdeterminedbytheSNRofthedatapoints.Figure 3-7 showstheEPLEcalculatedforboth1215Hzand2430Hzoverthecourseoftheexperimentwith30-secondresolution.Duringtherst15-minuteperiod,theEPLEmeasuredat1215Hzis0.630.15duringtheCW-ONperiodand0.680.11duringtheCW-OFFperiod.NosignicanttrendsareobservedduringeithertheCW-ONorCW-OFFperiod,althoughtheymaybeobscuredbythenoise.TheEPLEexhibitsa 61

PAGE 62

Figure3-7. Thepower-lawexponentat1215Hzand2430Hzforeachpower-stepseries,calculatedasdiscussedinthetext.Thetwotracesareseparatedforclarity.Foreachtrace,thereisonesampleevery30seconds. verysubtleincreasecoincidentallywith(within30secondsof)thechangefromCW-ONtoCW-OFF.At2430Hz,theEPLEismeasuredtobe0.690.16duringtheCW-ONperiodand0.780.07duringtheCW-OFFperiod.Again,nosignicanttrendsareobservedduringeitherperiod,althoughsmalltrendsmaybeobscuredbythenoiseofthemeasurement.Inthiscase,theEPLEappearstoincreasecoincidentally(within30seconds)withthechangefromCW-ONtoCW-OFF.Duringthesecond15-minuteperiod,theSNRistoolowforareliablecalculationoftheEPLE.Thiseffectisevidentinthemarkedincreaseinmeasurementvariabilityduringthesecond15-minuteperiod. 62

PAGE 63

TheappearanceofslightincreasesinEPLEatboth1215Hzand2430HzduringCW-OFFperiodsmaybemisleading,however.Basedonthisdataset,turningofftheCWbeamincreasestheEPLEby0.050.26at1215Hzandby0.090.23at2430Hz.ThelargeuncertaintiesintheEPLEmeasurementsindicatethatitisnotaswell-suitedforevaluatingchangesinionosphericpropertiesastheELF/VLFmagnitudeortheELF/VLFharmonicratio.Furthermore,itwouldbedifculttoproperlyevaluatetheEPLEwithhightimeresolution.Theseconclusionswillbeevaluatedinthetheoreticalmodelingsectionofthischapter. 3.3ModelPredictionsInthefollowingsections,eachpropertyoftheELF/VLFwaveiscomparedwithmodelpredictionsofthedual-beamHFheatingexperiment,asdescribedingreatdetailinChapter 2 .TheambientelectrondensityandelectrontemperatureprolesusedforthepurposeofmodelingarethesameprolesasinFigure 2-1 3.3.1ELF/VLFMagnitudeFigure 3-8 showsthetotalB-eldmagnitudereceivedonthegroundat1215and2430Hzasafunctionofelectrondensityandelectrontemperatureprole.Atbothmodulationfrequencies,themagnitudeoftheB-eldatthereceiverincreaseswithincreasing(80km)electrondensityatarateofabout10dBperprole.Becausetheelectrondensityincreasesbyafactorof10betweeneachprole,thex-axisoftheseplotshaveanessentiallylogarithmicscale.Thus,afactorof2changeinelectrondensityat80km,assuggestedby MilikhandPapadopoulos [ 2007 ],wouldproduce5dBincreaseinELF/VLFmagnitudeonthegroundbytheseestimates.Thisvalueisonlyslightlysmallerthanthe7dBincreaseinmagnitudepredictedby MilikhandPapadopoulos [ 2007 ].Theeffectofambientelectrontemperatureislesspronouncedontheseplots,butstillimportant,astheymayproducea2dBchangeinELF/VLFmagnitude.Inmostcases,theELF/VLFB-eldmagnitudeincreaseswithdecreasingambientelectrontemperature.Theoneexceptionisthecombinationofelectrondensity 63

PAGE 64

Figure3-8. Numericalpredictions.TotalB-eldmagnitudeatthereceiverasafunctionofelectrondensityproleandelectrontemperatureprole.(A):1215Hz.(B)2430Hz. ProleIwithelectrontemperatureProleD.Together,theseplotsindicatethattheELF/VLFB-eldmagnitudeobservedonthegroundcouldbemoreeffectivelyenhancedbytheintroductionofachemicalprocessthatbothincreasestheelectrondensityandsimultaneouslydecreasestheelectrontemperatureintheD-region.Thetwoeffectsaretypicallycompetingeffects,however,asanincreaseinelectrondensityalsoproducesanincreaseinelectron-neutralcollisionfrequency.Figure 3-9 showsthechangeinELF/VLFmagnitudereceivedonthegroundduringCW-ONandCW-OFFperiods.PositivedBvaluesonthisplotindicatethattheELF/VLFmagnitudeishigherduringCW-OFFperiodsthanduringCW-ONperiods.Wenote 64

PAGE 65

Figure3-9. Numericalpredictions.ThechangeintotalB-eldmagnitudeatthereceiverfromCW-OFFtoCW-ONconditions. thatthismodelpredictsthattheB-eldonthegroundisalwayshigherduringCW-OFFperiodsthanduringCW-ONperiods,consistentwithobservationsatthesepowerlevels.Theobserved7-9dBchangesinELF/VLFmagnitude(showninFigure 3-5 )areslightly(2dB)higherthanthepredictedvaluesshowninFigure 3-9 .Nevertheless,thepredictedvaluesarereasonablyclosetotheobservedvalues.ConsideringthattheadditionalCWheatingtendstoincreasethealtitudeofthedominantELF/VLFsourcecurrents,itmaybethecasethatEarth-ionospherewaveguideeffects,whichdependuponboththealtitudeandfrequencyofthesourceandwhicharenotaccountedforinourwavepropagationmodel,maybeimportantinaccuratelyaccountingforthechangeinELF/VLFmagnitudeontheground.Consideringthechangespredictedforbothmodulationfrequencies,themodelpredictsthatthechangeinELF/VLFmagnitudeonthegroundisabout1dBlowerat2430Hzthanat1215Hz.ThiswasthecaseobservedduringtherstCW-ON/CW-OFFtransition,whichisveryencouraging,butitwasnotthecaseduringthesecondCW-ON/CW-OFFtransition,whenthechangein2430Hzmagnitudewasobservedtochangeby1dBmorethanat1215Hz.Whetherornotthisisthetypicalobservational 65

PAGE 66

casewillnotberesolvedinthischapter,butmayeasilyberesolvedbyadditionalexperimentalstudies.Thefrequency-dependenteffectsoftheEarth-ionospherewaveguidemaycontributetothediscrepancy,whichmayalsobeaffectedbytheassumptionthattheconductivitymodulationiscylindricallysymmetric.Despitetheseshortcomings,themodelcapturesinageneralsensetheeffectsofsimultaneousCWandmodulatedHFheating,inthatitconsistentlypredictslowerELF/VLFmagnitudesonthegroundduringCW-ONperiods,andinthatthepredictedchangesinmagnitudearewithin2dBofobservations.Theobservedlarge7-9dBchangesinELF/VLFmagnitudeindicatethattheELF/VLFmagnitudemaybesensitivetothefrequencyandpoweroftheCWbeamandweevaluatethispossibilityingreatdetailinChapter 4 .HavingdiscussedthepredictedELF/VLFmagnitudesingreatdepth,wenowproceedtoconsiderthetheoreticalpredictionsfortheELF/VLFharmonicratio. 3.3.2ELF/VLFHarmonicRatioAsdescribedearlierinthischapter,theELF/VLFharmonicratiocancelsfrequencydependentpropagationeffectstorstorder.Thesecondordereffectdependsonthespatialdistributionofthesourcecurrentsthatgeneratetherstandsecondharmonics.PanelAofFigure 3-10 showstheELF/VLFharmonicratioasafunctionofelectrondensityproleandelectrontemperatureproleduringCW-OFFperiods,andthebottompanelshowsthedBchangeinharmonicratiobetweenCW-ONandCW-OFFperiods.ThenegativechangesshowninthebottompanelindicatethattheharmonicratioismodeledtobehigherduringCW-ONperiodsthanduringCW-OFFperiods,consistentwithobservations.ThecombinationofelectrondensityProleIIIandelectrontemperatureProleDmatchesourobservationsveryclosely,bothintermsoftheCW-OFFharmonicratio()]TJ /F5 11.955 Tf 9.3 0 Td[(17.8dBmeasured,)]TJ /F5 11.955 Tf 9.3 0 Td[(18.3dBmodeled)andintermsofthechangeinharmonicratiobetweenCW-ONandCW-OFFperiods()]TJ /F5 11.955 Tf 9.3 0 Td[(3.75dBmeasured,)]TJ /F5 11.955 Tf 9.3 0 Td[(3.2dBmodeled).Weattributetheclosenessofthismatchtotheeffectivecancellationofbothrst-andsecond-orderpropagationeffects,whichwerenot 66

PAGE 67

Figure3-10. (A)ELF/VLFharmonicratio(asdescribedinthetext)asafunctionofelectrondensityproleandelectrontemperatureprole.(B)ThechangeinELF/VLFharmonicratiofromCW-OFFtoCW-ONconditions. convenientlycanceledbyothermeasurementtechniques.Interestingly,modelresults(notshown)indicatethattheELF/VLFharmonicratioisarelativelystablevalueintermsofobservationlocation,varyingbylessthan1dBwithin100kmoftheHAARPtransmitter.Theharmonicratioisverysensitivetoboththeelectrondensityproleandtheelectrontemperatureproleused,varyingbyseveraldBinbothcases.ThechangeinharmonicamplitudeatCW-ON/CW-OFFboundariesisalsoeasilydetectableandquicktoevaluate(onlytwoELF/VLFfrequenciesneeded).ItthusappearsthattheELF/VLFharmonicratioisideallysuitedtoevaluateHAARP-inducedelectrondensitychanges 67

PAGE 68

atthealtitudeofwavegeneration.PanelBofFigure 3-10 indicatesthat(forProleD)afactorof2increaseinelectrondensitywouldresultinanadditional0.25dBchangeintheharmonicratiobetweenCW-ONandCW-OFFperiods.Intheobservationsfromthecurrentexperimentthisisnotobservableabovethenoise.Wecanlimitthemaximumchangeinelectrondensitytoafactorof5,basedonthe0.7dBuncertaintyofthemeasurement,however.TherelativelyweakELF/VLFwavegenerationduringtheexperiment,however,indicatesthatfutureexperimentsmayhavemoresuccessapplyingthistechniquetofurtherlimitthepossibleheater-inducedchangeintheelectrondensitywithtime.HerewepointoutthatelectrondensityProlesIIandIII,togetherwithelectrontemperatureProleDhavemostcloselymatchedthechangeinELF/VLFmagnitudesatCW-ON/CW-OFFboundariesandtheyalsocloselyreproducechangesinELF/VLFharmonicratiosthatverycloselymatchtheobservationspresentedearlierinthischapter.Althoughwehavenotpresentedanexhaustivesetofelectrondensityandtemperatureproles,itisreasonabletoconcludethatelectrontemperatureProleDandsomecombinationofelectrondensityProlesIIandIIIarereasonableestimatesofthephysicalpropertiesoftheD-regionionosphereduringthepresentedexperiment. 3.3.3ELF/VLFPower-LawExponentTheEPLEvaluesasafunctionofelectrondensityandelectrontemperatureareshowninthepanelsAandBofFigure 3-11 .TwoimportantresultsareimmediatelyevidentfromFigure 3-11 .ForthehighHFpowerlevelsforwhichtheseresultswerecalculated,theEPLEdecreasessignicantlyastheelectrondensityvariesfromProleItoProleIII(i.e.,astheelectrondensityincreases).ThisresultstandsinstarkcontrasttothesimulationresultsfortheELF/VLFmagnitude,whichincreasessharplybetweenProlesIandIII.Together,thesesimulationresultssupporttheconclusionthattheEPLEdoesnotrepresenttheoverallefciencyofELF/VLFwavegeneration.AsecondresultthatisclearlydepictedinFigure 3-11 isthetemperingeffectofadditionalCWheating 68

PAGE 69

Figure3-11. Numericalpredictions.(A,B):nat1215and2430Hz.(C):ThechangeinnfromCW-OFFtoCW-ONconditions. 69

PAGE 70

ontheEPLE.AlthoughtheEPLEvariessignicantlyasafunctionofelectrondensityproleunderCW-OFFconditions,itisrelativelyconstantunderCW-ONconditions.ThiseffectyieldsthegeneralresultthattheEPLEmaybeeitherhigherorlowerduringCW-ONorCW-OFFperiods,dependingontheelectrondensityproleandtheelectrontemperatureproleemployed.FurtherconclusionsmaybedrawnbycloselyinspectingthedependenceofEPLEonelectrondensityandelectrontemperature.Afactoroftwoincreaseinelectrondensityat80kmaltitudewoulddecreasenbyapproximately0.05.07. MilikhandPapadopoulos [ 2007 ]predictthischangemayoccurwithatimescaleontheorderof1minute(althoughat85kmaltitude).ThelevelofnoiseinourexperimentalobservationsofEPLE,however,ismuchtoolargetodetectthissmallmodication.TheELF/VLFharmonicratioclearlyconstitutesamuchbettermeasurementforprovidinglimitsforpossiblechangesinelectrondensity.TheelectrontemperatureisalsoanimportantfactorindeterminingtheEPLE.Similartothedependenceonelectrondensity,ntypicallydecreasesastheelectrontemperaturedecreases,despitetheindicationthatthetotalELF/VLFmagnitudetendstoincreasewithdecreasingelectrontemperature.ThelargeuncertaintyinourexperimentallyobservedEPLEsmakesourobservationsconsistentwithalmostallofourmodelingruns.WhiletheELF/VLFmagnitudeandtheELF/VLFharmonicratioappeartobesensitivetoionosphericchangesevenunderlowSNRconditions,theEPLEderivedfromlowSNRobservationsisclearlynotagoodindicatorforionosphericmodication. 3.4DiscussionWehavepresentedexperimentalevidencealongwithmodelpredictionsindicatingthatthemagnitudesofELF/VLFwavesobservedonthegroundaresignicantlyreducedwhengeneratedtogetherwithabroaderCWheatingbeam.WedemonstratedthattheELF/VLFharmonicratioisalsosensitivetothepresenceoftheCWheatingbeam,althoughitissomewhatlesssensitivethantheELF/VLFmagnitudeandrequireshigh 70

PAGE 71

SNR.Lastly,theELF/VLFpower-lawexponentwasshowntobetoosensitivetoSNRtoprovideaccurateexperimentalobservationsrelatingtoionosphericmodication.Basedontheseobservations,weconcludethattheELF/VLFsignalmagnitudeasafunctionofHFtransmissionparameteristhepropertyofthereceivedELF/VLFeldthatismostsensitivetotheambientionosphericconditions. 71

PAGE 72

CHAPTER4ELF/VLFWAVEGENERATIONASAFUNCTIONOFPOWER,FREQUENCY,MODULATIONWAVEFORM,ANDRECEIVERLOCATIONBasedontheresultsofChapter 3 ,thischapterfocusesontheELF/VLFmagnitudegeneratedasafunctionofthecontrollabletransmissionparametersatHAARP:HFpower,HFfrequency,modulationwaveform,andHFpolarization.Theexperimentalobservationsarecomparedtothepredictionsofthedual-beamionosphericHFheatingmodel(Chapter 2 )toidentifythetransmissionparametersthatprovideindependentinformationabouttheambientD-regionionosphere.Thischapterprogressesinthefollowingmanner:Section 4.1 describestheexperiment;Section 4.2 presentsageneraldescriptionofexperimentalobservationsperformedduringthedual-beamHFheatingexperiment;andSection 4.3 directlycomparestheobservationsandmodelpredictions.Section 4.4 providesadiscussionandasummaryofthepresentedmaterial.TheworkpresentedinthischapterhasbeenpublishedintheJournalofGeophysi-calResearch[ AgrawalandMoore 2012 ]. 4.1DescriptionoftheExperimentDuringthreehalf-hourperiodson20,21,and25July2011,HAARPonceagainusedadual-beamheatingcongurationforwhichtheHFarraywassplitintotwo615(1800kW)sub-arrayscapableofsimultaneouslytransmittingtwoindependentHFbeamsatdifferentHFfrequencies.Therstsub-arraybroadcastanamplitudemodulatedHFsignalinordertogenerateELFwaves;wewillrefertothemodulatedbeamasBeam1.Atthesametime,thesecondsub-arraybroadcastaCWwaveatadifferentHFfrequencyandvariedthepowerofthetransmission;wewillrefertotheCWbeamasBeam2.ThecenterfrequencyforBeam1alternatedbetween5.8and6.9MHz(X-mode),andthepeakpowerwasheldconstantat100%.Themodulationwasdrivenat1225Hzusingvedifferentmodulationwaveforms:square,sinusoid,square-root-sinusoid(sqrt-sine),triangle,andsaw-tooth.Thecenterfrequencyfor 72

PAGE 73

Figure4-1. AcartoondiagramoftheDual-BeamHFheatingexperiment,showingthemodulatedHFbeam(constantpeakpower)andthepower-steppedCWbeam.TheCWbeamisbroaderthanthemodulatedHFbeam. Beam2alternatedbetween3.25and4.5MHz(X-mode),resultinginfourdifferentHFfrequencycombinationsbetweenBeam1andBeam2.ThecenterfrequencyofBeam2wasselectedtobelowerthanthatofBeam1inordertobathetheentiremodulatedionosphericregionwithCWpower,asdepictedinthecartoondiagramofFigure 4-1 .AlthoughthecartoondiagramisverysimilartothatshowninChapter 3 ,thetransmissiondetailsaredifferent.Inthiscase,thepeakpowerofBeam2increased(in1-dBsteps)from)]TJ /F1 11.955 Tf 9.3 0 Td[(8dBto0dB(fullpower),resultingintenCWpowerlevels(includingCW-OFF).EachCWpowerlevelwasheldconstantforaone-secondduration,andtheCWtransmissionalternatedbetweenCW-OFFandCW-ONeveryothersecondtoprovideameanstoevaluatechangesintheelectrojeteldstrength[ BarrandStubbe 1993 ].Foreachfrequencycombination,the18-secondtransmissionformatwasrepeatedvetimes:onceforeachmodulationwaveform.Figure 1-4 mapsthelocationsoftheELF/VLFreceiversitesrelativetoHAARP.ELF/VLFreceiverswerelocatedatOasis(OA,62.35N,145.1W,3kmfromHAARP),SinonaCreek(SC,62.58N,144.6W,33kmfromHAARP),andParadise(PD,62.52N,143.2W,98kmfromHAARP).Inpost-processing,theamplitudesandphasesofthe 73

PAGE 74

receivedELF/VLFtoneswereonceagaindeterminedusing1-second-longdiscreteFouriertransforms. 4.2DescriptionoftheDataSetFigure 4-2 showstherstharmonicELFsignalmagnitude(at1225Hz)receivedatParadiseduringthethreethirty-minutedurationtransmissionblockson20,21,and25July2011.Signalswithhigh(>10dB)signal-to-noiseratio(SNR)wereobservedthroughoutthehalf-hourtransmissionperiodson20Julyand25July.Ouranalysiswillfocusonobservationsperformedduringthesetwodays,whicharehighlightedwithagraybackgroundinFigure 4-2 .Onbothdays,theHAARPuxgatemagnetometerregisteredmagneticelductuationsofover100nTduringthetransmissionperiods,andthekpindexwas3+.Thelevelofabsorption,asmeasuredbythe30-MHzHAARPriometer,wasmuchhigheron20July(0.2dB)thanon25July(<0.1dB).Additionally,ionosphericelectrondensityproleestimationsperformedbytheHAARPdigisondeatthetimesoftransmissionindicatethattheionosphericprolesweredramaticallydifferentonthetwodays,evenintheD-region:theelectrondensityat100kmwas3.6104/cm3on20July,whereasitwaslessthan1.2104/cm3on25July.Thecomparisonofobservationsperformedonthesetwodayswillthusbeusedtoexperimentallyinvestigatedual-beamELFwavegenerationasafunctionofambientionosphericconditions.Figure 4-3 shows90-secondspectrogramsofthemagneticeldrecordingsperformedatParadiseon20July2011.ThetopandbottompanelscorrespondstotheNorth-South(NS)andEast-West(EW)channelsofthereceiver,respectively.Duringthis90-secondperiod,HAARPbroadcastve18-secondformats,oneforeachofthevedifferentmodulationwaveformsemployed.Inorder,theseare:square,sinusoid,sqrt-sine,triangle,andsaw-toothwaveforms.Inallcases,therstharmonicat1225Hzisclearlyvisibleinthespectrogramsforeachchannel.Higher-orderharmoniccontentisobservedtodependonthemodulationwaveform.Whileobservationsofthe 74

PAGE 75

Figure4-2. 1225HzsignalmagnitudeobservedatParadise(PD)on20,21,and25July2011.Thehighlightedgraybackgroundsidentifyperiodsofhigh(>10dB)SNR. higher-orderharmonicsareimportant,thisanalysiswillfocussolelyontherstharmoniccomponent.PanelBofFigure 4-4 showsthemagnitudeofthe1225HztonesobservedatParadiseduringthe30minutetransmissionperiodon20July2011.Thealternatinggrayandwhitebackgroundsrepresenttherepetitionofthe8-minutetransmissionformat,whichincludesfourdistinctHFfrequencycombinations(betweenBeam1andBeam2).Withineachsection,fourdistinctgroupsofdatapointsrangingfrom105to122dBareclearlydiscernible,andthesegroupscorrespondtoobservationsasafunctionofHFfrequencycombination.Thedatapointsrangingfrom80to92dBareperformedduringtransmitterofftimes,andthelargestoftheseamplitudesisusedtoestimatethenoiseoor.PanelAofFigure 4-4 providesanexpanded-timeviewofthe90-secondtransmissionperiodforBeam1at5.8MHzandBeam2at4.5MHz.Duringthis90-secondperiod,Beam1continuouslymodulatedtheconductivityofthelowerionosphere,changingthemodulationwaveformevery18seconds.Atthesametime,Beam2broadcastaCWwaveeveryothersecond(alternatingbetweenonandoff)increasingthepowerofthe 75

PAGE 76

Figure4-3. 90-secondspectrogramsofELFobservationsatParadise.StrongELFwavesaregeneratedat1225Hz,aswellasathigher-orderharmonics. transmissionin1-dBstepsoverthecourseof18seconds.TheobservationsshowninFigure 4-4 demonstratethattherstharmonicmagnitudeduringCW-OFFperiodsisclearlystableovereach18-secondperiod,varyingbylessthan1dB.TheCW-OFFsignalstabilityindicatesthattheionosphereandthestrengthoftheauroralelectrojetwerestableoverthetransmissionsequence.Therstharmonicmagnitudealsoclearlydependsonthemodulationwaveformemployed,ascanbeseenbytheseveral-dBchangesinmagnitudewhenthemodulationwaveformchanges.ThesechangesinmagnitudeareapproximatelyconsistentwithFourieranalysisofthepowerenvelopeofthetransmittedmodulationwaveform,asdescribedby BarrandStubbe [ 1993 ],andtheslightdeviationsfromFourieranalysiswillbediscussedinSection 4.3 .DuringtheCW-ONperiods,the1225HzsignalmagnitudeisobservedtodecreasewithincreasingCWpowerforallmodulationwaveforms,consistentwiththeobservationspresentedby MooreandAgrawal [ 2011 ]. 76

PAGE 77

Figure4-4. (A)1225Hzsignalmagnitudesovera90-secondtransmissionperiod.(B)1225Hzsignalmagnitudeforthe30minutetransmissionof20July2011. ThedatashowninFigures 4-2 4-4 aregenerallyrepresentativeofourobservationsatallreceiversitesandforalltransmissionperiods,althoughtheSNRvarieswithbothtimeandsitelocation.Table 4-1 summarizestheSNRobservedateachreceiversiteforalldaysofobservationsduringthetransmissionblocks.N/AentriesindicatethattheELFreceiveratOasishadnotyetbeendeployedtothesite.TocalculatetheSNRlevelsshowninTable 4-1 ,thenoiseoorisapproximatedduringperiodswhentheHFtransmitterisoff.Inthefollowingsection,weprovidedetailedanalysisoftheseobservationsasafunctionofCWpower,HFfrequencycombination,modulationwaveform,andreceiver 77

PAGE 78

Table4-1. 1225HzSNRateachsiteforeachday. DateTimeSNR(dB)(July2011)(UT)PDOASC 20053013-30N/A10-202107300N/A02507305-305-200-20 location,andwecomparetheobservationswiththeresultsofthedual-beamionosphericHFheatingmodel(Chapter 2 ). 4.3AnalysisInthissection,wecompareobservationswiththepredictionsofthedual-beamHFheatingmodel(Chapter 2 )asafunctionofCWHFpower,CWHFfrequency,modulatedHFfrequency,modulationwaveform,andreceiverlocation.Experimentalobservationsareusedtodemonstratethatthechangeinconductivitymodulation(Figure 2-13 )asafunctionofCWpowerisameasurablequantitythatissensitivetotheambientconditionsoftheD-regionionosphere.AtheoreticalanalysisconsideringthecasesofCWHFpolarizationandthemodulationfrequencyofthemodulatedHFbeamisalsopresented. 4.3.1CWHFPowerThefourpanelsofFigure 4-5 showtheobserved(blacktraces)andpredicted(colortraces)ELFwavemagnitudesasafunctionoftheeffectiveradiatedpower(ERP)oftheCWbeam,withthefourpanelscorrespondingtothefourBeam1/Beam2HFfrequencycombinationsemployed.Wepointoutthatthescalesofthefourpanelsaredifferentinordertoclearlydepicteachofthetraceswithineachpanel.Thepredictionsforalltwelvecombinationsofelectrondensityandelectrontemperatureprolesareshownforeachcase,andeachtracehasbeennormalizedsothat0dBcorrespondstothe1225HzmagnitudeobservedduringperiodswiththeCWbeamOFF.Asmentionedpreviously,thisnormalizationcancelstheeffectsoftheEarth-ionospherewaveguidetorstorderandprovidesameanstoaccountforthevariationinelectrojeteldstrengthoverthecourseoftheexperiment[ BarrandStubbe 1993 ]. 78

PAGE 79

Figure4-5. 1225HzsignalmagnitudeobservedatParadise(solidblack)togetherwithdual-beamHFheatingmodelpredictions(color)forsquare-waveamplitudemodulation.ThefourpanelspresentresultsforfourdifferentBeam1/Beam2HFfrequencycombinationsandfortwelvedifferentNe/Teprolecombinations. 79

PAGE 80

Firstconsideringonlytheexperimentalobservations,allfourtracesexhibitsimilarvariationswithCWERP:thenormalizedELFmagnitudedecreasesasafunctionofincreasingCWERP,andtherateofdecreaseincreaseswithincreasingCWERP.Althoughwehaveonlyshowntherstrepetitionof20July2011foreachfrequencycombinationonthisgure,allotheriterationsoftheexperimentexhibitthesesamecharacteristicfeatures.ForallpowerlevelswiththeCWbeamON,thenormalizedmagnitudesat1225HzarelessthanthoseobservedduringperiodswiththeCWbeamOFF,consistentwiththeobservationsreportedby MooreandAgrawal [ 2011 ].ComparingtheleftandrightpanelsofFigure 4-5 (i.e.,forconstantCWHFfrequency),subtledifferencesexistbetweentheobservationsasafunctionofCWpower,andwewillconsiderthesedifferencesindetailinsubsequentsubsections.Comparingthetopandbottompanels(i.e.,forconstantHFfrequencyofthemodulatedwave),itisevidentthatforagivenERPvalue,the3.25MHzCWsignalsuppressestheELFmagnitudetoagreaterextent(2dB)thandoesthe4.5MHzCWwave.WewillconsiderthisdependenceingreaterdetailinSection 4.3.2 .Nowconsideringthemodelpredictionstogetherwiththeexperimentalobservations,Figure 4-5 clearlydemonstratesthatalltraces(bothexperimentalandtheoretical)exhibitasimilardependenceonCWERP:thenormalizedELFmagnitudedecreasesasafunctionofincreasingCWERP,andtherateofdecreaseincreaseswithincreasingCWERP.Ingeneral,thepredictedELFmagnitudesshowgoodagreementwithobservations,althoughspecicdetailsofthetracesclearlydependuponthespecicambientelectrondensityandelectrontemperatureproleemployed.Nevertheless,comparingtheleftandrightpanelsofFigure 4-5 (i.e.,forconstantCWHFfrequency),thedifferencesbetweenmodelpredictionsasafunctionofBeam1frequencyareverysimilartothoseexhibitedbytheobservations.Furthermore,comparingthetopandbottompanels(i.e.,forconstantHFfrequencyofthemodulatedwave),themodelpredictsthatthe3.25MHzCWsignalwillsuppresstheELFmagnitudetoagreater 80

PAGE 81

extent(2dB)thanthe4.5MHzCWwaveforaconstantERPlevel,strikinglysimilartoobservations.HavingcomparedexperimentalobservationswiththeoreticalpredictionsasafunctionofCWpower,wecontinueouranalysisbycomparingresultsasafunctionoftheHFfrequencyoftheCWbeam. 4.3.2CWHFFrequencyPanelAofFigure 4-6 presentsexperimentalobservationsofthe1225Hzsignalmagnitudeperformedon20July2011(solidtraces)andon25July(dashedtraces)atParadise(PD)asafunctionofCWERP.ThepowerstepsforthetwodifferentCWfrequenciesemployedspantwodistinctERPranges.InordertodeterminethedependenceonCWfrequency,wecomparethe3.25/5.8(red)traceswiththe4.5/5.8(green)tracesandthe3.25/6.9(blue)traceswiththe4.5/6.9(purple)traces.Theobservationsperformedon20July2011exhibitnearlyidenticaldependenciesonCWERP(afterdiscountingforthedifferentCWfrequency-dependentgains):theinitialsuppressionoffsetandthespreadofsuppressionasafunctionofCWpower,asdenedinthegure,areessentiallythesame.Observationsperformedon25July2011,underdifferentambientionosphericconditions,however,clearlyindicatethatthesuppressionoffsetis1dBgreaterfor3.25MHzthanfor4.5MHzandthatthesuppressionspreadasafunctionofCWERPis1dBgreaterfor3.25MHzthanfor4.5MHz.Inthiscase,thesuppressionoffsetandthespreadarebothdifferentasafunctionofHFfrequencycombination.TheseexperimentalobservationsindicatethatthereceivedELFmagnitudeasafunctionofCWfrequencyissensitivelydependentupontheambientionosphericconditions.PanelBofFigure 4-6 presentsmodelpredictionsforfourdifferentionosphericprolecombinations(I-A,II-C,II-D,andIII-A).ThelevelofELFmagnitudesuppressiondiffersasafunctionofCWfrequency,andthisdifferencechangesasafunctionofionosphericprolecombination,dependentuponboththeambientelectrondensity 81

PAGE 82

Figure4-6. NormalizedELFmagnitudeasafunctionofCWERP,highlightingtheeffectsofBeam1(modulated)andBeam2(CW)HFfrequencyfor(A)ObservationsatParadiseon20and25July2011.(B)ModelPredictions. andelectrontemperatureproleemployed.ModelingresultsexhibitbothsmallandlargedifferencesinsuppressionoffsetandspreadasafunctionCWfrequency.Weexpectthatadifferentsetofionosphericproleswillreproducethelarge(3-dB)initialsuppressionoffsetobservedon25July2011,althoughwehavemadenoefforttodosohere.Mostimportantly,bothobservationsandmodelpredictionsindicatethatthelevelofELFmagnitudesuppressionbyadditionalCWheatingsensitivelydependsonthefrequencyoftheCWsignalandontheionosphericconditions.HavingcomparedexperimentalobservationswiththeoreticalpredictionsasafunctionofCWHFfrequency,wecontinueouranalysisbycomparingresultsasafunctionofthemodulatedHFfrequency. 82

PAGE 83

4.3.3Beam1(Modulated)HFFrequencyWecontinuetorefertoFigure 4-6 toinvestigatethedependenceonthemodulated(Beam1)HFfrequency.Inthiscase,wecomparethe3.25/5.8(red)traceswiththe3.25/6.9(blue)tracesandthe4.5/5.8(green)traceswiththe4.5/6.9(purple)traces.OnagivendayandforagivenCWHFfrequency,thenormalizedELFmagnitudefor5.8MHzisextremelysimilartothatfor6.9MHz.ForbothCWfrequencies(3.25and4.5MHz),thedifferencesinELFmagnitudesuppressionarenearlynegligibleatlowCWpowerlevels.ForaCWfrequencyof3.25MHz,athigher(>74dBWERP)CWpowerlevels,thesignalsgeneratedusing6.9MHzareslightlymoresuppressedthanthosegeneratedusing5.8MHz,withthedifferenceinsuppressionincreasingwithincreasingCWpoweruptoamaximumdifferenceof0.5.0dB(dependingontheday,andtherebyambientionosphericconditions).ForaCWfrequencyof4.5MHz,similardifferencesareobservedathigherCWpowerlevels(>80dBWERP),althoughthedifferenceinthelevelofsuppressionislessthanforthe3.25MHzCWsignal,maximizingat0.25dB.ThemodelpredictionsforthefourdifferentambientionosphericconditionsshowninpanelBofFigure 4-6 exhibitthesamegeneraltrendsexhibitedbytheexperimentalobservations:thedifferenceinthelevelofsuppressionbetweenBeam1frequenciesincreaseswithincreasingCWpower.Fortheambientionosphericcombinationsconsidered,thepredictionsindicatethatELFsignalsgeneratedusing5.8MHzmaybemoreorlesssuppressedthanthosegeneratedusing6.9MHz.Additionally,3.25MHzmaycreatelargerorsmallerdifferencesthan4.5MHz,dependingontheionosphericprolecombination.AtthehighestCWpowerlevels,thedifferenceinsuppressionmaybeashighas1.0dB,alsodependingontheambientionosphericprolecombinationemployed.Basedonbothexperimentalobservationsandtheoreticalpredictions,weconcludethatwhilethedependenceonBeam1(modulated)HFfrequencyismeasurableattheseCWpowerlevels,signicant(1-dB)differencesaredetectable 83

PAGE 84

onlyathigherCWpowerlevels,whentheELFSNRislower(andtheerrorbarsarehigher).Asaresult,thedifferenceinCWsuppressionasafunctionofBeam1frequencyisadifcultmeasurementtoperforminpractice.HavingcomparedexperimentalobservationswiththeoreticalpredictionsasafunctionofBeam1(modulated)HFfrequency,wecontinueouranalysisbycomparingresultsasafunctionofamplitudemodulation(AM)waveform. 4.3.4ModulationWaveformTheleftpanelofFigure 4-7 showsthenormalizedaverageELFmagnitudeexperimentallyobservedasafunctionofAMwaveformon20Julyand25July2011.TheseparticularmeasurementswereperformedduringCW-OFFperiods,andtheyarenormalizedbytherstFourierharmoniccomponentoftherespectiveidealsignalwaveforms,takingtheaveragesquare-wavesignalmagnitudeasa0-dBreference.Accountingfortheerrorbars,showningray,thedifferenceinnormalizedsignalmagnitudemeasuredbetweenthetwodaysvariesbetween0.10and0.80dBasafunctionofmodulationwaveform,withthelargestdifferenceoccurringforthesaw-toothwaveform.Wenotethatthe20July2011observationsarelargerthanthe25July2011observationsforallmodulationwaveforms.Becausethetransmissionformatdidnotchangebetweenthetwodays,thesedifferencesareattributabletothedifferentambientionosphericconditionsonthetwodays.ThemodelpredictionsforthenormalizedELFmagnitudeasafunctionofelectrondensity(forelectrontemperatureProleA)areshownintherightpanelofFigure 4-7 .Themodelpredictionsexhibitvariationsasafunctionofionosphericprolecombination,withthelargest(0.80-dB)variationsoccurringforthetrianglewaveform.Comparingobservationswiththepredictionsofthetheoreticalmodel,allexperimentallymeasuredvaluesarewithin0.5-dBofthetheoreticalresults,withtheworstcorrespondenceoccurringforthesqrt-sinemodulationwaveform.ConsideringthataCWbeamisnotrequiredtoperformthismeasurement,theSNRdoesnotsignicantlysufferasaresult(asopposedtotheBeam1frequency 84

PAGE 85

Figure4-7. CW-OFF:ELFmagnitudeasafunctionofmodulationwaveform,normalizedbythemagnitudegeneratedforsquareAM. case).ItthusappearstobethecasethatcarefulobservationsoftherelativemagnitudesgeneratedusingdifferentAMwaveformsmayyieldindependentinformationregardingtheambientionosphericconditions.NowconsideringtheeffectsofadditionalCWheating,panelAofFigure 4-8 showsthenormalizedELFmagnitudeat1225HzasafunctionofCWpowerforaconstantBeam1HFfrequencyof5.8MHz.Theobservationsarepresentedforboth20July(solidtraces)and25July(dashedtraces).Onbothdays,thenormalizedELFmagnitudeforallveAMwaveformsproducenearlyidenticalresultsatallCWERPlevels,withthesingularexceptionofthesquare-wavemodulationcasefor5.8/4.5MHzon20July.TheresultsforaBeam1HFfrequencyof6.9MHz(notshown)aresimilarinallrespects.PanelBofFigure 4-8 showsthemodelpredictionsfortheveAMwaveformsasafunctionofCWERP,usingtwodifferentsetsofambientionosphericconditions(I-AandII-D).ThepredictionsforbothsetsofambientprolesexhibitextremelysimilarvariationsofELFmagnitudeforallvemodulationwaveformsasafunctionofCWERP,withthelargestoffsets(only0.25-dB)occurringforthetriangleandsaw-toothwaveforms.Basedontheseresults,weconcludethatadditionalCWpowerdoesnotprovideadditionalindependentinformationabouttheambientionosphericconditionsasafunctionofthemodulationwaveform.ConsideringtheeffectsofCWheatingon 85

PAGE 86

Figure4-8. ELFmagnitudeasafunctionofCWERPforveAMwaveformsforCWfc3.25MHzandmodulatedfc5.8MHzfor(A)experimentalobservationsand(B)modelpredictions.OnlyverysubtlevariationsinELFmagnitudeareobservedasafunctionofAMwaveform. thehigher-orderharmoniccontentproducedasafunctionofmodulationwaveformisbeyondthescopeofthischapter.Inthefollowingsection,wecompareobservationsandmodelpredictionsfordifferentELFreceiverlocationsasafunctionofCWpower. 4.3.5ReceiverLocationPanelAofFigure 4-9 showsthenormalizedELFsignalmagnitudeobservedatatSinonaCreek(SC,33kmfromHAARP)andatParadise(PD,98kmfromHAARP) 86

PAGE 87

on20July2011asafunctionofCWERPforthefourcombinationsofHFfrequenciesemployed.Notethattwosetsoftracesinthetoptwopanelshavebeenfalselyoffsetby2dBandtwosetsofmodelpredictions(bottompanel)havebeenoffsetby1dBforaestheticpurposesinordertofacilitatethecomparisonasafunctionofreceiverlocation.ThenormalizedELFmagnitudesobservedatthetwolocationsareessentiallythesame(withintheerrorbars)forallCWpowerlevelsandallHFfrequencycombinations.TheerrorbarsfortheSinonaCreekmeasurementsarehigh(>2dBinsomecases),however.ThedeploymentofanELFreceiveratOasis(OA,3kmfromHAARP)enabledhighSNRmeasurementsattworeceiversiteson25July2011.PanelBofFigure 4-9 showsthenormalizedELFobservationsatParadiseandOasison25July.FortheBeam2CWfrequencyof4.5MHz,theobservationsareverysimilaratthetwositesandtheerrorbarsoverlapatalmostallCWpowerlevels.ThesameisnottruefortheBeam2CWfrequencyof3.25MHz.Atlow(<76dB)CWpowerlevels,theELFmagnitudegeneratedby5.8MHzobservedatParadiseislowerthanthatobservedatOasisby1dB.ThedifferencedecreaseswithincreasingCWERP.Asimilarvariationisobservedforthe6.9MHzsignalbelow73dBERP,althoughthemaximumdifferenceislessthan0.25dB.PanelCofFigure 4-9 showsthepredictednormalizedELFmagnitudeasafunctionofCWERPatallthreegroundbasedreceivers.Themodelpredictionsareingeneralagreementwiththeobservations,exceptatlowpowerlevels,andshowthatthenormalizedELFmagnitudesareexpectedtohavecomparablelevelsasafunctionofreceiverlocation.WeattributethedeviationobservedatParadiseforlowCWERPlevelstotheeffectsoftheEarth-ionospherewaveguideandpossiblytothegenerationofsecondaryionosphericcurrents[ Payneetal. 2007 ],neitherofwhichareaccountedforbyourpropagationmodel.Earth-ionospherewaveguideeffectsareexpectedtobeimportantatreceiverlocationsgreaterthan75kmfromHAARP[ Payne 2007 ],andParadiseis98kmdistant. 87

PAGE 88

Figure4-9. ThenormalizedELFmagnitudeasafunctionofELFreceiverlocationfor(A)Observationson20July2011,(B)Observationson25July2011and(C)ModelPredictions.Redandgreentraceshavebeenoffsetverticallytofacilitatevisualcomparisonbetweenthetraces. 88

PAGE 89

ItisinterestingtonotethatathigherCWERPlevelsandforhigherCWHFfrequencies,theobservationspresentedinFigure 4-9 areverymuchinlinewithmodelpredictions.WehypothesizethatthealtitudeoftheELFsourceregionplaysaroleindeterminingtherelativeimportanceofEarth-ionospherewaveguideeffectsatthereceiversite.Forinstance, MooreandAgrawal [ 2011 ]showed(theoretically)thatadditionalCWheatingincreasesthealtitudeoftheeffectiveELFsourceregion.Furthermore,theELFsourceproducedbymodulatedheatingathigherHFfrequenciesisexpectedtooccuratsomewhathigheraltitudesthanforlowerHFfrequencies[ StubbeandKopka 1977 ].Lastly,higheraltitudesourcesareexpectedtoexcitetheEarth-ionospherewaveguidelesseffectivelythanloweraltitudesources[e.g., Tri-pathietal. 1982 ].Together,theseconsiderationsmayexplainthenormalizedELFmagnitudesobservedatParadiseforlowandhighCWERPlevels.Basedontheobservationsandtheoreticalmodeling,weconcludethatadditionalmeasurementsatreceiverlocationswithin75kmofHAARPwillnotcontributeasignicantamountofadditionalinformationabouttheambientD-regionionosphere.Nevertheless,additionalobservationsatlocationsgreaterthan100kmfromHAARPcouldpossiblyprovidemoreinformationregardingothereffects,suchasthoserelatedtotheEarth-ionospherewaveguideortosecondaryionosphericcurrents.Havingcompletedourcomparisonofexperimentalobservationswithmodelpredictions,wenowpresenttheoreticalpredictionsforthenormalizedELFmagnitudeasafunctionofthepolarizationoftheCWHFbeamandthemodulationfrequencyofthemodulatedbeam. 4.3.6PolarizationInthissection,wetheoreticallyinvestigatetheeffectsthatareproducedbychangingthepolarizationoftheCWbeam,asopposedtothatofthemodulatedHFbeam.Figure 4-10 showsthemodelpredictionsforCWheatingusingbothX-andO-modepolarizationsfortheCWbeam.Resultsfortwoionosphericprolecombinations 89

PAGE 90

Figure4-10. TheoreticalpredictionsfornormalizedELFmagnitudeasafunctionofCWHFbeampolarization(X-vs.O-mode). (I-AandII-D)arepresented.Inbothcases,O-modeCWheatingsuppressesthenormalizedELFmagnitudetoalesserextentthanX-modeheating.ForProleI-A,O-modeheatingappearstohaveanalmostnegligibleeffectontheELFmagnitude,whereasforProleII-D,O-modeheatingproduces2dBofsuppressionatthehighestCWERPlevel.Basedonthesemodelpredictions,wesuggestthatthedifferenceinthelevelofELFmagnitudesuppressionproducedbyX-modeandO-modeCWheating1)ismeasurable,and2)producesindependentinformationregardingtheambientionosphericconditions.Forinstance,forProleI-A,thedifferenceinthesuppressionproducedbyX-andO-modeCWheatingincreaseswithCWERPfrom0.25dBatthelowestCWERPlevelto3.5dBatthehighestCWERPlevel.ForProleII-D,thedifferenceincreasesfrom0.75dBto5.5dB.Additionally,thesuppressionoffsetsandspreadsaresignicantlydifferent(>3dB),especiallyathighCWpowerlevels.MeasurementscomparingX-andO-modeCWsuppressioncouldsignicantlycontributetoananalysisofambientionosphericconditions. 90

PAGE 91

Figure4-11. TheoreticalpredictionsshowingthenormalizedELFmagnitudefordifferentmodulationfrequencies. 4.3.7ModulationFrequencyFigure 4-11 showsthepredictedELFmagnitudeasafunctionofCWERPfortwodifferentmodulationfrequenciesandfortwodifferentambientionosphericprolecombinations.Forbothmodulationfrequencies,increasingtheCWpowerincreasesthelevelofELFmagnitudesuppression,andforbothionosphericproles,thesignalgeneratedusingthehighermodulationfrequencyissuppressedtoalesserextent.Thedifferenceinsuppression(asafunctionofmodulationfrequency)increasesfrom0.25dBatthelowestCWpowerlevelto0.75.0dBatthehighestCWpowerlevelforbothionosphericproles.WhilethesedifferencesasafunctionofCWpowerlevelarenotaslargeasthedifferencesproducedbyX-andO-modeheating,theyaredetectable.Nevertheless,thesystemresponsewouldneedtobecalibratedtoaverytighttoleranceatthedifferentmodulationfrequenciestoprovidereliableobservationsasafunctionofmodulationfrequency.WethusconcludethatwhilemeasurementscomparingCWsuppressionasafunctionofmodulationfrequencyandCWpowercanprovideadditionalinformationregardingtheambientionosphericconditions,suchmeasurementswouldbedifculttoperformreliablyinpractice. 91

PAGE 92

4.4DiscussionThischapterhascomparedexperimentalobservationsperformedduringpowersteppeddual-beamELFwavegenerationexperimentsatHAARPtothepredictionsofadual-beamionosphericHFheatingmodel.ComparisonswereperformedasafunctionofHFpower,HFfrequency,modulationwaveform,andreceiverlocation.Modelpredictionsagreewellwithobservations,demonstratingthatthemodelincorporatestheessentialphysicsinvolvedinmultibeamHFheatingofthelowerionosphere.WeevaluatedthesensitivityofthereceivedELFwavemagnitudetothesecontrollableparametersandinterpretedthedependenceonambientionosphericconditions.Asaresult,wehaveidentiedthetypesoftransmissionsthatmayprovideasignicantamountofinformationregardingtheambientconditionsoftheD-regionionosphere.Whileaninverseproceduretoderivetheambientelectrondensityandtemperatureprolesfromthesemeasurementsremainstobepresented,theobservationsandmodelingpresentedhereinstronglysuggestthatdual-beamELFwavegenerationexperimentscanplayanimportantroleinapossiblefutureD-regiondiagnostic.Weconcludebyenumeratingourexperimentalandtheoreticalresults: 1. ForhighCWpowerlevels,theintroductionofadditionalCWheatingreducestheamplitudeofthereceivedELFwave.TherateofELFmagnitudesuppressionincreaseswithincreasingCWpower. 2. ThelevelofELFmagnitudesuppressiondependsontheCWfrequencyemployed,andthelevelofsuppressionasafunctionofCWfrequencyissensitivelydependentontheambientionosphericconditions. 3. ThelevelofELFmagnitudesuppressionalsodependsonthefrequencyofthemodulatedHFbeam,althoughtoalesserextentthantheCWfrequency. 4. TheELFsignalmagnitudeasafunctionofmodulationwaveform(withoutCWheating)alsodependsontheambientionosphericconditions,whereasthesuppressionsuppliedbyadditionalCWheatingisextremelysimilarasafunctionofmodulationwaveform. 5. ELFreceiverslocatedatsignicantlydifferentdistancesfromHAARP(3km)registersimilarnormalizedELFmagnitudesathighCWpowerlevels.Differences 92

PAGE 93

existatlowerCWpowerlevels,andweattributethesedifferencestotheeffectsoftheEarth-ionospherewaveguideatlargerdistances. 6. AtheoreticalanalysispredictsthatO-modeCWheatingmayprovideadditionalindependentinformationtoobservationsperformedusingX-modeCWheatingandthattheseobservationsaresensitivelydependentontheambientconditionsoftheD-region. 7. Lastly,wepredictthattheeffectofCWheatingwilldecreasewithincreasingmodulationfrequency. 93

PAGE 94

CHAPTER5TIME-OF-ARRIVAL(TOA)MEASUREMENTSASAMEANSTODETECTD-REGIONSTRUCTUREChapter 4 identiedthetypesofHFtransmissionsthatcanbeusedtoprovidesignicantinformationregardingtheambientcharacteristicsoftheD-regionionosphere.OneofthemajordifcultiesinprovidingaD-regiondiagnosticusingtheseparametersissimplytoseparatetheeffectsofelectrondensityandelectrontemperature.Inthischapter,wedemonstratethatNeandTeeffectscannotbefullydecoupledingeneral.Nevertheless,wedeterminethatlinearTevariationswithaltitudeproduceaminimalimpactonthemodelingofELFwavesgeneratedusingdual-beamheating,allowingforabest-telectrondensityproletobecalculatedinastraight-forwardmanner.Thisclassofelectrondensitysolutionsnecessarilyassumesalinearelectrontemperaturevariationwithaltitude.Whilewedonotpresumethattheidentiedelectrondensityandelectrontemperatureproleuniquelymatchobservations,weassertthatthecombinationofthetwomaybeusedtoprovideinsightintothestructureofthelowerionosphere.TheidentiedmethodisappliedtoproduceanelectrondensityprolethataccuratelymodelstheunusuallylargelevelofELFmagnitudesuppressionobservedon25July2011.Theseobservationsarefairlyaccuratelymodeled,andtheidentiedelectrondensityproleproducesadominantELFsourcealtitudelocatedatasharpridgeintheelectrondensityprole.Thisobservation,inturn,leadstotheinvestigationofELF/VLFtime-of-arrival(TOA)measurementsasameanstoprovideadditionalinformationabouttheambientD-regionelectrondensityprole.WedemonstratethatTOAobservationsperformedduringdual-beamheatingexperimentsareconsistentwithastructuredelectrondensityprole,andweprovideapiecewiseexponentialNeprolethatcloselymatchestheTOAobservations.ThesignalprocessingrelatedtotheTOAobservationspresentedhereinisthefocusofaUniversityofFloridaPh.D.thesisbyShujiFujimaru.Whilethischaptermakes 94

PAGE 95

useoftheTOAobservations,theprimaryfocusofthischapterisonthemodelingofobservations,specicallyunderdual-beamHFheatingconditions. 5.1CoupledNatureofNeandTeProlesContinuingwiththetypeofnumericalanalysispresentedinChapter 4 ,wepresentnumericalmodelingresultsindicatingthatthesuppressionofELFmagnitudebyadditionalCWheatingdependssensitivelyonboththeelectrondensityandelectrontemperatureproles.Figure 5-1 presentssummarychartsofELFmagnitudesuppressionasafunctionofCWERP,withthetracesorganizedbyelectrondensityprole.ItisclearthatforelectrondensityProles1and2,theelectrontemperatureproleemployedsignicantlyimpactstheresultingELFsignalmagnitude.ForelectrondensityProle3,however,theresultsareessentiallyindependentofelectrontemperature,indicatingthatforlargerD-regionelectrondensities,theeffectsofelectrontemperatureareminimized.TheresultssummarizedinFigure 5-1 demonstratethattheelectrontemperatureprolecanproducechangesinthemagnitudeoffsetofupto0.25dBandchangesinthemagnitudespreadofupto2.25dB.Becausesimilarresultsmaybeproducedbyslightlymodifyingeithertheelectrondensityorelectrontemperatureprole,itisclearthattheelectrondensityprolecannotbefullydecoupledfromtheelectrontemperatureprole.Atthesametime,modelingresultsindicatethattheelectrondensityproledominatesboththemagnitudeoffsetandthemagnitudespread,withtheeffectofelectrontemperaturebeingsecondary.Withtheseresultsinmind,weundertakethetaskofidentifyingaclassofelectrontemperatureprolethatminimallyimpactsthemodelingresults.Fromthisperspective,anelectrondensityprolemaybeidentiedandusedtointerpretthestructureoftheD-regionionosphere. 95

PAGE 96

Figure5-1. ModelpredictionsforthenormalizedELFmagnitudeasafunctionofelectrondensityproleandelectrontemperatureprole. 5.2LinearTeProlesandExponentialNeProlesInordertoassessthesensitivityofELF/VLFwavegenerationtotheelectrondensityandtemperatureprolesandtodeterminetheconditions(ifany)underwhichtheeffectsofNeandTemaybedecoupled,weevaluatethemultibeam-HFheatingmodelusingavarietyofexponentialNeprolesandlinearelectrontemperatureproleswithdifferingmagnitudesandslopes.Theexponentialelectrondensityprolesareparameterizedbyh0(km)and(km)]TJ /F9 7.97 Tf 6.58 0 Td[(1),asgivenby WaitandSpies [ 1964 ]: Ne(h)=1.43107exp()]TJ /F5 11.955 Tf 9.3 0 Td[(0.15h0)exp[()]TJ /F5 11.955 Tf 11.96 0 Td[(0.15)(h)]TJ /F2 11.955 Tf 11.96 0 Td[(h0)]cm)]TJ /F9 7.97 Tf 6.59 0 Td[(3(5)The WaitandSpies [ 1964 ]electrondensityprolehasbeensuccessfullyusedinpreviousD-regionworks[e.g., CummerandInan 1997 ; Cummeretal. 1998 ; McRaeandThomson 2000 ].Theparameterh0controlsthemagnitudeoftheelectrondensity 96

PAGE 97

Figure5-2. 16uniqueexponentialelectrondensityproles(plottedonlogscale)obtainedbyvaryingparametersandh0. proleandcontrolstheslopeoftheproleonlog-scale.Inthiswork,wevariedfrom0.38to0.8andh0from78to90kminordertocoveracompleterangeofreasonableelectrondensitiesconsistentwithanighttimeD-regionionosphere[e.g., Mooreetal. 2007 ; MooreandAgrawal 2011 ; FujimaruandMoore 2011a ].Figure 5-2 showsthe16distinctelectrondensityprolesemployed.Inordertoprovidemoderatelyrealisticproles,theseproleshavebeenlimitedathigheraltitudes(>80km)bythemaximumelectrondensitycalculatedusingtheIRIcalculator( http://omniweb.gsfc.nasa.gov/vitmo/iri_vitmo.html )for20August2011at7.5hoursUT(denotedbythesolidblacktrace).Neproleswiththesamecolortracerepresentproleswiththesamevaluewhileproleswiththesamelinestylerepresentthesameh0values.TheTeprolesemployedvarylinearlywithaltitude,withtheparametersselectedtocovertherangedelineatedbyprolesA,B,C,andDshowninFigure 5-3 .ProlesA,B,C,andDarerepresentativeoftheelectrontemperatureprolesmodeledoverthecourseofayearusingtheMSISE-90AtmosphereModel[ MooreandAgrawal 2011 ; AgrawalandMoore 2012 ].TheeightprolesusedinthisworkareshowninFigure 5-3 .TheseproleshavebeentruncatedbyProleAatloweraltitudestobemorerealistic. 97

PAGE 98

Figure5-3. 8uniquelinearelectrontemperatureprolesobtainedtocovertheentirerangeofreasonableTeproles. ThenormalizedELFmagnitudepredictionsforall128combinationsofexponentialNeandlinearTeprolesareshowninFigures 5-4 and 5-5 .Figure 5-4 highlightsthedependenceon,whileFigure 5-5 highlightsthedependenceonh0.Ineachofthegures,thetoppanelshowstheELFmagnitudewiththeCWbeamOFF,whilethemiddleandbottompanelsshowtheELFsuppressionoffsetandmagnitudespread,respectively,asafunctionofandh0.Inordertosimplifythedisplays,wehaveusederrorbarstorepresenttheextremesofthepredictionsforthe8linearTeproles.WenotethatthemodelpredictionscoverawiderangeofdetectableELFmagnitudes(1fTto1pT,toppanel),suppressionoffsets()]TJ /F1 11.955 Tf 18.59 0 Td[(1to+1dB,middlepanel),andmagnitudespreads()]TJ /F1 11.955 Tf 18.6 0 Td[(8to0dB,bottompanel).FocusingonFigure 5-4 ,thedifferentcolortracesrepresentthedifferenth0valuesusedtogeneratetheexponentialNeproles.PanelBofFigure 5-4 demonstratesthatforagivenh0,thesuppressionoffsethasacomplicateddependenceon:forh0=82,86,theoffsetincreasesandthendecreaseswith,whileforh0=78,90,theoffsetincreasesmonotonicallywith.PanelCofFigure 5-4 demonstratesthatforagivenh0,themagnitudespreadgenerallyincreaseswithincreasing,withonlyoneexceptionath0=86. 98

PAGE 99

Figure5-4. (A)ELFmagnitudewithCWbeamturnedoff.(B)NormalizedELFinitialsuppressionoffsetand(C)NormalizedELFsuppressionspread,asafunctionof. Figure5-5. (A)ELFmagnitudewithCWbeamturnedoff.(B)NormalizedELFinitialsuppressionoffsetand(C)NormalizedELFsuppressionspread,asafunctionofh0. 99

PAGE 100

TurningourattentiontoFigure 5-5 ,thedifferentcolortracesrepresentthedifferentvaluesusedtogeneratetheexponentialNeproles.PanelBofFigure 5-5 demonstratesthatforagiven,thesuppressionoffsettendstoincreasewithh0,withtheonlyexceptionbeingfor=0.38km)]TJ /F9 7.97 Tf 6.58 0 Td[(1.PanelCofFigure 5-5 demonstratesthatforagiven,themagnitudespreadhasacomplicateddependenceonh0:for=0.38km)]TJ /F9 7.97 Tf 6.58 0 Td[(1,themagnitudespreaddecreasesmonotonicallywithh0,for=0.52,0.66km)]TJ /F9 7.97 Tf 6.59 0 Td[(1,themagnitudespreadincreasesthendecreaseswithincreasingh0,andfor=0.80km)]TJ /F9 7.97 Tf 6.59 0 Td[(1,themagnitudespreadincreasesmonotonicallywithh0.ForbothFigures 5-4 and 5-5 ,thedisplayederrorbarsindicatethattheelectrontemperatureprolehasaminimalimpactonthecalculations.Theelectrontemperatureaffectsthesuppressionoffsetbyatmost0.75dB(h0=90km,=0.66km)]TJ /F9 7.97 Tf 6.59 0 Td[(1)andthemagnitudespreadbyatmost1.25dB(h0=90km,=0.66km)]TJ /F9 7.97 Tf 6.59 0 Td[(1).Theguresdemonstratethatthetypicalvariationwithelectrontemperatureismuchlessthanthesetwoquotedvalues.Basedonthesemodelingresults,andconsideringthatthetypicalSNRobservedduringdual-beamheatingexperimentsis0.5dB,weconcludethatelectrontemperatureprolesthatvarylinearlywithaltitudedonotsignicantlyimpactELFwavegenerationcalculationsforcomparisonwithexperimentalobservations.Inthenextsection,wediscusshowwecanachieveadditionalELFsuppressionoffsetsandmagnitudespreadsinordertomatchtheobservationsperformedon25July2011. 5.2.1MatchingObservationson25July2011Therangeofandh0showninFigures 5-4 and 5-5 stilldonotaccountfortheadditionalinitialELFsuppressionoffsetandmagnitudespreadobservedon25July2011atOA,asdiscussedinSections 4.3.2 and 4.3.5 .BasedonthemodelpredictionsshowninFigures 5-4 and 5-5 ,however,weconcludethatdecreasingandh0togethermaycloselymatchobservations.Figure 5-6 showstheelectrondensityprolesusingsof0.1and0.05andanh0of90km(bluedashedandbluedash-dottracerespectively). 100

PAGE 101

Figure5-6. ExponentialNeath0of90kmwithof0.05and0.10km)]TJ /F9 7.97 Tf 6.58 0 Td[(1(bluedashedandbluedash-botrespectively).Reddashedandreddash-dottracesrepresenttheprolestruncatedbyrealisticproles,I,IIandIII(leftpanelofFigure 2-1 ). TheseproleshavedecreasingslopesthroughouttheD-regionionosphere.Inordertomaketheseprolesmorerealistic,weinterpretthedecreasingslopeasindicatingthatledgesexistatthelowerandupperportionsoftheproles.WethusboundtheprolesusinganaverageofProlesIIandIIIattheloweraltitudesandProleIatthehigheraltitudes.Theseprolesarerepresentedbythereddashedandreddash-dottracesinFigure 5-6 respectively.ThenormalizedELFmagnitudepredictionsasafunctionofCWERPfortheseprolesareshowninFigure 5-7 .TheexperimentalobservationsofnormalizedELFmagnitudeobservedon25July2011atOAareoverlaidonthisgure.ThetwoNeprolesexhibitsignicantadditionalsuppression,bothintermsoftheELFsuppressionoffsetandintermsofthemagnitudespread.Themodelingpredictionsmatchtheobservationsreasonablywell,andcertainlyprovideabettermatchthantheprolecombinationsusedinChapter 4 (modelpredictionsfromprolecombinationII-D,alsooverlaidonthisgure).Thisanalysisindicatesthatthedual-beamELFwavegenerationexperimentmaybesensitivetoD-regionstructure,asexhibitedbythe 101

PAGE 102

Figure5-7. NormalizedELFmagnitudeat1225HzasafunctionofCWERPfor25July2011atOA,observationsforNeprolewithh090kmandof0.1and0.05km)]TJ /F9 7.97 Tf 6.59 0 Td[(1,(dashedanddottedrespectively).RedandgreentracesrepresentCWfc3.25and4.5MHzrespectivelywithmodulatedCWfc5.8MHz. electrondensityledgesrequiredtomaketheprolerealistic.Withoutadditionalinformation,itisimpossibletoverifywhetherthisstructureindeedexistedwithintheD-regionionosphereon25July2011.Whileperformingthistheoreticalexercise,however,wenoticedthatthedominantELFsourceregionwaslocatedjustabovethelowerridgeinelectrondensityforallCWpowersteps.Figure 5-8 showstheHallconductivitymodulationgeneratedintheionosphereusingaof0.1km)]TJ /F9 7.97 Tf 6.58 0 Td[(1andanh0of90kmforCWOFFandCWON(fullpower)transmissions.ThewhitetraceoverlaidoneachpanelrepresentstheNeprole,whilethewhitedotoneachpanelrepresentsthealtitudeofthedominantsourceregion.Inbothcases,thedominantsourcelocationisessentiallyatthelowerNeridgeatanaltitudeof73km.ThistheoreticalresultisaquantitythatcanbemeasuredexperimentallyusingELF/VLFtime-of-arrival(TOA)analysis[ FujimaruandMoore 2011a ],andthenextsectionfocusesontheTOAanalysisofexperimentalobservationsinthecontextofthedual-beamHFheatingexperiment. 5.3Dual-BeamELF/VLFTOAObservations FujimaruandMoore [ 2011a ]describeanELF/VLFtime-of-arrival(TOA)measurementmethodthatrequiresmodulatingtheionosphereusingafrequency-timerampformat, 102

PAGE 103

Figure5-8. ModulatedsourceregionforHallwithCW-OFF(A)andHallwithCWatpeakpower(B)withaltitude.OverlaidoneachpanelisthetruncatedNeprolewith0.1km)]TJ /F9 7.97 Tf 6.59 0 Td[(1andh090km.Thewhitedotoneachpanelrepresentsthealtitudeoftherespectivedominantsourceregions. ratherthantones.TheanalysisofTOAobservationsleadstotheabilitytoapproximatethelocationofthedominantELF/VLFsourcealtitude.Duringaveminuteperiodon3May2012between07:35:30and07:40:00UT,HAARPusedadual-beamheatingcongurationinwhichtheHFarraywassplitintotwo615(1800kW)sub-arrays.Therstsub-array(Beam1)modulatedtheionosphereusingafull-power4.5MHz(X-mode)beamthatwassquarewaveamplitudemodulatedwithalinear1)]TJ /F1 11.955 Tf 9.3 0 Td[(6kHzfrequency-timeramp(1kHz/secslope).Atthesametime,thesecondsub-array(Beam2)broadcastaCWwaveat3.25MHz.ThepeakpowerofBeam2decreased(in3-dBsteps)from0dB(fullpower)to)]TJ /F1 11.955 Tf 9.29 0 Td[(15dB.TheorientationofbothbeamswasheldverticalwithrespecttoHAARP.EachCWpowerlevelwasheldconstantforave-secondduration,andtheCWtransmissionalternatedbetweenCW-OFFandCW-ONeveryvesecondsinordertoprovideameanstoevaluatechangesintheelectrojeteldstrength[ BarrandStubbe 1993 ].Dataforvecompleteexperimentswasobtainedduringthistime.TheHAARPuxgatemagnetometershowedlargedisturbancesandregisteredmagnetic 103

PAGE 104

Figure5-9. (A)AbsoluteTOAand(B)lineofsight(LOS)magnitudeobservationsasafunctionofCWERPwithCWbeamturnedON(solid)andCW-OFF(dashed).Redandbluetracesrepresenttwotrialsoftheexperiment. elductuationsofbetween75nTand100nTduringthetransmissionperiod.Thekpindexwas2)]TJ /F1 11.955 Tf 12.62 0 Td[(atthistime.Usingtheanalysismethoddescribedby[ FujimaruandMoore 2011b ,Chapter2],theELF/VLFsignalsgeneratedbythemodulationoftheionospherewithfrequency-timerampsareextracted.PanelsAandBofFigure 5-9 showthetimes-of-arrivalandtheELF/VLFamplitudesdetectedduringthedual-beamTOAexperimentasafunctionofCWpowerrespectively.Inbothpanels,theblueandredtracescorrespondtotwotrialsoftheexperiment,withsolidanddashedtracesrepresentingobservationswiththeCWbeamturnedONandOFFrespectively.TheleftpaneldemonstratesthattheTOAincreaseswithCWpowerbetween530-secondsand555-seconds.Intermsofadominantsourcealtitude,thispropagationtimedelaycorrespondstoadominantELF/VLFsourcealtitudeof82kmwithCWOFF.PanelBofFigure 5-9 demonstratesthattheamplitudedecreaseswithincreasingCWERP,exhibitingamaximumsuppressionof7dB,asexpected.TheELF/VLFTOAsandamplitudesforCWOFF(dashedtraces)arefairlyconstantduringtheentiredurationoftheexperiment,indicatingafairlystableionosphereandelectrojeteldstrength. 104

PAGE 105

Figure5-10. RelativeTOA(A)andnormalizedLOSamplitude(B)observationswithCWERP.Theredandbluetracesrepresentthenormalizedobservationsfromtwotrialsoftheexperiment. Figure 5-10 showstherelativechangeinTOA(referredtoasTOAspread)andinELF/VLFamplitudeasafunctionofCWERPfortwotrialsofthedual-beamTOAexperiment.Intheleftpanel,bothtrialsdemonstratethatthenormalizedTOAincreaseswithincreasingCWERP(by16-seconds),consistentwiththepredictionsby MooreandAgrawal [ 2011 ]andindicatingthatthealtitudeofdominantsourceregionincreasesby2kmasafunctionofincreasingCWERP.TherightpanelshowsthatthenormalizedELF/VLFamplitudedecreasesby6.5.5dBasafunctionofincreasingCWERPwithreasonableerrorbarsof0.4dBatthelowestCWERPleveland1dBatthehighestCWERPlevel.Wecomparetheseobservationswiththepredictionsofthedual-beamHFheatingmodelusingexponentialelectrondensityproles.Themodelisevaluatedatthemidpointfrequencyofthefrequency-timeramp,i.e.,at3500Hz,usingexponentialNeproleswithvaryingfrom)]TJ /F1 11.955 Tf 12.62 0 Td[(0.1to0.8km)]TJ /F9 7.97 Tf 6.59 0 Td[(1andwithh0varyingfrom78to86km.Figure 5-11 showsthemodelpredictionsfortheCW-OFFTOA,theTOAspread(TOA 105

PAGE 106

Figure5-11. AnalysisforexponentialNeasafunctionoffrom)]TJ /F1 11.955 Tf 9.3 0 Td[(0.1to0.8,forh078,82and86km. CW-ON100%minusCW-OFFTOA),theinitialELF/VLFmagnitudeoffset,andtheELF/VLFmagnitudespreadasafunctionof.Thedifferentcoloredtracesoneachpanelrepresentthethreeh0semployed.ThetoppanelofFigure 5-11 showsthattheCW-OFFTOAincreaseswithforallh0.Ontheotherhand,theTOAspreaddecreasesuntilacertain,experiencesasharpincreaseat0.2,andthengraduallydecreaseswithincreasing.Thisisconsistentforallh0.TheoppositeistruefortheELF/VLFmagnitudeoffsetandmagnitudespread,however.Bothquantities(showninthebottom2panelsofFigure 5-11 )decreaseuntilacertainandthengraduallyincrease.Forbetween0.2and0.4,thevaluesofCW-OFFTOA,TOAspread,ELF/VLFmagnitudeoffset,andELF/VLFmagnitudespreadliewithinareasonablerangeofthepresentedELF/VLFTOAobservations.Figure 5-12 showsazoomedviewofFigure 5-11 forfourh0svaryingfrom78to90km.Alsoshownintherightpanelarethe16exponentialNeprolesemployed.Thetop2panelsofFigure 5-12 demonstratethattheCW-OFFTOAandtheTOAspreadexhibitcompetingdependenciesonforallh0s.Asincreases,theCW-OFFTOAincreasesfrom520-secondsto560-seconds,whereastheTOAspread 106

PAGE 107

Figure5-12. AnalysisforexponentialNeasafunctionoffrom0.2to0.4,forh078,82and86and90km. decreaseswithincreasingfrom20-secondsto10-seconds.TheELF/VLFmagnitudeoffsetandmagnitudespreadexhibitthesamedependenceon,however:bothquantitiesincreasewithincreasing.Examiningtheseresultsasafunctionofh0,wenoticethatforagiven,theCW-OFFTOA(toppanel)increasesash0increasesfrom78to90km.Similarly,inthesecondpanel,weobservethattheTOAspreaddecreaseswithincreasingh0.Lastly,inthebottompanel,itisdemonstratedthattheELF/VLFmagnitudespreadbecomesincreasinglynegativewithincreasingh0.Althoughthemodelingresultsforproleswithbetween0.2and0.4andforh0between78and90kmprovidevaluesforallfourquantitiesthatarewithinareasonablerangeofobservations,theredoesnotexistasingleproleamongthemthataccuratelymatchesalloftheobservationswithinerrorbars.Additionally,duetothetradeoffbetweenCW-OFFTOAandmagnitudespread,thereisnocleardirectioninwhich 107

PAGE 108

tomodifytheexponentialproleeitherintermsoforintermsofh0.InordertoovercomethecompetingaffectsofCW-OFFTOA,TOAspread,andmagnitudespread,weconsidertheconstructionofapiecewise-exponentialNeproleinthenextsection. 5.3.1FittingaPiecewise-ExponentialNeProleUsingTOABasedonthemodelingresultsprovidedthusfar,weproposethefollowingmethodforidentifyingapiecewiseexponentialelectrondensityprolethatmatchesTOAobservations.1)UsetheTOAdominantaltitudeastheinitialguessforthealtitudeofintersectionbetweenthetwoproles,2)usetheprolethatbestmatchestheTOAspreadandmagnitudespreadastheinitialguessforthelowerportionoftheprole,and3)usetheprolethatbestmatchestheCW-OFFTOAastheinitialguessfortheupperportionoftheprole.Basedonthisdescription,weidentifyaninitialintersectionaltitudeof82kmbasedontheTOAobservations,andweuseFigure 5-11 toidentifythecombinationof=0.2forthelowerportionoftheproleand=0.33fortheuppersectionoftheprole.Oncethisinitialguessprolehasbeenidentied,weiterativelymodifytheprolebyadjusting1)theofthelowersection,2)thealtitudeofintersection,3)themaximumelectrondensity,and4)theoftheuppersectioninordertobestmatchtheobservations.Tables 5-1 and 5-2 providereferencesfortheobservedandmodeledparameters.Thecolumnsofthe'Piecewise-Exponential'sectionprovidesasummaryoftheiterationsinorder.ThenaliterationresultsinaCW-OFFTOAof532-seconds,aTOAspreadof16-seconds,amagnitudeoffsetof)]TJ /F5 11.955 Tf 9.3 0 Td[(0.30dB,andamagnitudespreadof)]TJ /F5 11.955 Tf 9.3 0 Td[(6.9dB,allofwhicharewithintheerrorbarsoftheobservations.Inthefollowingsubsections,westepthrougheachiterationanddescribehowthechoicesweremade. 5.3.1.1LowerSectionKeepingtheandh0oftheuppersection(uandh0u)xedat0.33km)]TJ /F9 7.97 Tf 6.58 0 Td[(1and82kmrespectively,andkeepingtheheightofintersectionh0intxed,wevariedthe(l)ofthelowersection.TherightpanelofFigure 5-13 showsthelowersectionprolesthatwere 108

PAGE 109

Table5-1. ExperimentalTOAobservations. ParametersTrial1Trial2 Abs.TOA(sec)540530Rel.TOA(sec)1616Norm.offset(dB))]TJ /F1 11.955 Tf 9.3 0 Td[(0.500.4)]TJ /F10 11.955 Tf 9.3 0 Td[(0.400.4Norm.spread(dB))]TJ /F1 11.955 Tf 9.3 0 Td[(8.00.5)]TJ /F10 11.955 Tf 9.3 0 Td[(7.61 Table5-2. Piecewise-ExponentialNe. ParametersExponentialPiecewise-Exponential l(km)]TJ /F9 7.97 Tf 6.59 0 Td[(1)0.200.330.200.220.220.220.22u(km)]TJ /F9 7.97 Tf 6.59 0 Td[(1)0.330.330.330.330.33hint(km)8282868686log10(Ne)000-0.20-0.22Abs.TOA(sec)524551522534529532534Rel.TOA(sec)21133124201619Norm.offset(dB))]TJ /F1 11.955 Tf 9.3 0 Td[(0.30)]TJ /F1 11.955 Tf 9.3 0 Td[(0.14N/AN/AN/A)]TJ /F10 11.955 Tf 9.3 0 Td[(0.30)]TJ /F1 11.955 Tf 9.3 0 Td[(0.29Norm.spread(dB))]TJ /F1 11.955 Tf 9.3 0 Td[(6.80)]TJ /F1 11.955 Tf 9.3 0 Td[(5.1)]TJ /F1 11.955 Tf 9.3 0 Td[(6.7)]TJ /F1 11.955 Tf 9.3 0 Td[(6.2)]TJ /F1 11.955 Tf 9.3 0 Td[(6.4)]TJ /F10 11.955 Tf 9.3 0 Td[(6.9)]TJ /F1 11.955 Tf 9.3 0 Td[(6.4 evaluated.TheleftpanelofFigure 5-13 showsthevariationofthethreequantitiesasafunctionofl.SimilarcompetingeffectsareobservedforCW-OFFTOA,TOAspread,andmagnitudespreadaswasdescribedinassociationwiththeexponentialelectrondensityproles.Nevertheless,weobserveanimprovementintheCW-OFFTOA(topleftpanel).Basedonthisanalysis,wechoosealof0.22atthisstage. 5.3.1.2AltitudeofIntersection(h0int)Keepingtheandh0oftheuppersection(uandh0u)xedat0.33km)]TJ /F9 7.97 Tf 6.58 0 Td[(1and82kmrespectively,andkeepingthelxedat0.22(asderivedinSection 5.3.1.1 ),wevariedthealtitudeofintersectionofthetwoprolesections.TherightpanelofFigure 5-14 showstheprolesthatwereevaluated.TheleftpanelofFigure 5-14 showsthevariationofthethreequantitiesasafunctionofh0int.ThesamecompetingdependenciesareonceagainobservedfortheCW-OFFTOAandthemagnitudespreadaswasdescribedabove,buttheTOAspreadincreasesuptoacertainh0intafterwhichitdecreaseswithincreasingh0int(middlepanel).Weseeasignicantimprovementinmagnitudespread 109

PAGE 110

Figure5-13. Betaoflowersection. Figure5-14. Altitudeofintersectionofthetwopieces. (bottomleftpanel),however,thatisworththetradeoffinCW-OFFTOA.Asaresult,weselectanh0intof86kmatthisstage. 5.3.1.3TheMagnitudeofNeKeepingthesofbothsections(uandl)xedat0.33km)]TJ /F9 7.97 Tf 6.59 0 Td[(1and0.22km)]TJ /F9 7.97 Tf 6.59 0 Td[(1,respectively,andkeepingthealtitudeofintersectionofthetwopiecesat86km,we 110

PAGE 111

Figure5-15. MagnitudeofNe. modifythemagnitudeoftheentireNeprole.TherightpanelofFigure 5-15 showstheNeprolesevaluated,andtheleftpanelsofthegureshowsthevariationofthethreequantitiesasafunctionofNemagnitudeonlogscale.Itisclearfromthisgurethattwoprolematchingalloftheobservationswithinerrorbarshavebeenevaluated:log10(Ne)is0.20and0.22elec/cm)]TJ /F9 7.97 Tf 6.59 0 Td[(3. 5.3.1.4UpperSectionInthecaseathand,itisunnecessarytocontinuetheprocessfartherbecausewehaveidentiedanelectrondensityprolethatcanreproduceobservationswiththeerrorbars.Nevertheless,weevaluatethedependenceoftheparametersontheuppersectionforreference.Figure 5-16 demonstratesthattheCW-OFFTOAcaneitherincreaseordecreasewith,dependingontheslopeofthelowersection.TheTOAspreadtendstoincreasewithincreasinguppersection,althoughitchangesbylessthat10-seconds.Similarly,themagnitudeoffsetandmagnitudespreadvaryonlysubtly,andthedirectionofvariationwithuppersectiondependsontheofthelowersection. 111

PAGE 112

Figure5-16. Betaofuppersection. 5.4BestFitPiecewise-ExponentialProleFigure 5-17 comparesthemodelpredictionsatallCWpowerlevelsforthetwoprolesthattexperimentalobservationsoftheCW-OFFTOA,theTOAspread,themagnitudeoffset,andthemagnitudespread.Inallpanelsthesolidgreenanddashedgreenlinesrepresentthemodelpredictions;theblueandredtracesrepresenttheobservations.ComparingthevariationofCW-OFFTOA,TOAspread,andELF/VLFmagnitudeasafunctionofCWERP,bothmodelpredictionsgenerallycapturetheCWpowerdependenceofeachoftheevaluatedquantities.Basedontheseresults,wearenotabletochoosebetweenthetwomodelproles,butwenotethattheyareonly5%differentinelectrondensity.ThischapterhaspresentednewexperimentalobservationsusingTOAanalysis.Thetwobest-telectrondensityproles(calculatedassumingalinearelectrontemperatureprole)areonlydifferentby5%,suggesting5%accuracy.Furthermore,theexperimentalTOAobservationscouldnotbeaccuratelypredictedusingonly 112

PAGE 113

Figure5-17. Observationsalongwithbesttpiecewise-exponentialprole. exponentialelectrondensityproles;insteadapiecewiseexponentialelectrondensityprolewasrequiredtoreproducetheobservations.Thisresultindicatesthatthedual-beamexperimentissensitivetostructurewithintheD-regionionosphere,denedasadeviationfromlinearelectrontemperatureproleand/orexponentialelectrondensityprole,andthatthedual-beamTOAexperimentiswellsuitedtodetectthistypeofstructure. 113

PAGE 114

CHAPTER6SUMMARYANDFUTUREWORKInthisdissertation,weidentifyanovelmethodtocharacterizetheambientpropertiesoftheD-regionionosphereusingground-basedobservationsofELF/VLFwavesgeneratedduringdual-beamHFheatingexperimentsatHAARP.Chapter 1 providedbackgroundmaterialandashorthistoryofD-regiondiagnostictechniques.Chapter 2 describedthemulti-beamHFheatingmodel,includingtheupgradesperformedinordertomodelvariousmodulationwaveformsandX-andO-modepolarizationsoftheCWHFbeam.Chapter 3 outlinedthedual-beamheatingexperiment,discussinginparticularanextensiveexperimentconductedatHAARPin2007.ExperimentalobservationsandmodelpredictionsofELF/VLFwavegenerationwerecomparedasafunctionofmagnitude,harmonicratio,andpower-lawexponent,andtherstharmonicoftheELF/VLFmagnitudewasidentiedasthewavepropertythatismostsensitivetotheambientD-regionconditions.Chapter 4 presentedarigorousanalysisthatcomparedexperimentalobservationsoftherstharmonicELF/VLFmagnitudewithmodelpredictionsasafunctionofseveraltransmissionparameters(HFpower,HFfrequency,andmodulationwaveform).ThesensitivityofthereceivedELF/VLFwavemagnitudetothesecontrollableparameterswasevaluatedandtheirdependenceonambientionosphericconditionswasinterpreted.Lastly,Chapter 5 evaluatedtheconditionsunderwhichtheeffectsofNeandTecouldbedecoupled.Dual-beamTOAexperimentswereidentiedasanexcellentmeanstoprovidekeyadditionalinformationaboutthestructureoftheD-regionionosphere.Theconstructionofabest-tpiecewise-exponentialambientelectrondensityprolewasdescribedingreatdetail.Thefollowingscienticcontributionsweredemonstratedinthiswork: 1. Amulti-beamHFionosphericheatingmodelwastestedandvalidatedusingobservationsatHAARP.Thefunctionalityofthemulti-beamHFheatingmodelwassuccessfullyextendedtoaccountforvedifferentAMwaveforms,namelysquare, 114

PAGE 115

sinusoidal,square-root-sine(sqrt-sine),triangle,andsaw-tooth.ThemodelwasalsoextendedtoaccountforbothX-andO-modepolarizationsoftheCWbeam. 2. ItwasexperimentallyestablishedthatthemagnitudeofELF/VLFwavegenerationistheparametermostsensitivetoadditionalCWheating. 3. ThetransmissionparametersthatprovideindependentinformationabouttheambientD-regionionosphereduringdual-beamheatingexperimentswereexperimentallyidentied.Thesetransmissionparametersare:1)theCWpowerlevel,2)thefrequencyoftheHFCWbeam,3)themodulationwaveform,and4)thepolarizationoftheHFCWbeam. 4. InthecontextoftheHFheatingmodel,itwasdemonstratedthatobservationsperformedduringthedual-beamheatingexperimentaresensitivetostructurewithintheD-regionionosphere.Piecewise-exponentialelectrondensityapproximationsappeartoadequatelymatchallavailableobservations.Afewitemsthatarenotcoveredinthispresentdissertation,butwouldberatherinterestingareasforfutureresearcharediscussedbelow. 6.1WeightedLeastSquareImplementationApossiblemethodofestimatingtheNeandalsoTeproleswouldbetoimplementaweightedleastsquarealgorithm,bytreatingthemagnitudeoftheBeld,M(p B2H+B2P)asalinearfunctionofelectrondensityprole,Ne(h)andTe. M(n,h)=g(Ne(h),n)+f(Te(h),n)(6)wherenistheCWERPlevel.ByvaryingthecontrollableparametersofNe,,h0intandlog10(Ne)ateveryiterationandfurtherparameterizingTe,possiblyasapiecewise-linearproleandimplementingacomplexoptimizationalgorithm[ Boyd 2004 ],wecanndthebesttpiecewise-exponentialNeandpiecewise-linear(oranyreasonablyparameterizedfunction)TetobestmatchobservationsandgetbetterinsightintothestructureoftheD-region. 115

PAGE 116

6.2ExtensiontoHigherHarmonicsInthisworkpresented,wehaveonlycomparedobservationsandmodelpredictionsofthemagnitudeoftherstharmonicsofthereceivedELF/VLFwavesasafunctionofCWERP.WesawinFigure 4-3 thathigherharmonicsweregeneratedforallmodulationwaveforms,howeverwithasignicantlylowerSNRcomparedtotherstharmonic.Bychoosingthemodulationfrequency,suchthatthereishighSNRgenerationofsecondandpossiblythirdharmonics,dependingonthemodulationwaveformchosenandbyextendingthemodeltoproperlyevaluatethemagnitudeofthehigherharmonics,itwillbeinterestingtoevaluatethemagnitudevariationofthehigherharmonicsasafunctionofCWERP.ThisanalysiscouldprovideadditionalindependentinformationtodetectstructureintheD-region. 6.3HFFrequenciesWithinaCollisionFrequencyOneoftheassumptionsofthedual-beamHFheatingmodelusedinthiswork,isthatthefrequencyofthetwoHFbeamsaresuchthattheyaremorethanacollisionfrequencyfromoneanother,suchthatthetwoHFbeamsdonotinteractwithoneanother.Althoughchallenging,itwillbenoteworthytodevelopthedual-beamHFmodelinordertoevaluatethenon-linearitiesthatresultfromtheinterharmonicmixingbetweenCWandmodulatedsignalwhenthetwoHFfrequenciesarelessacollisionfrequencyapart.MorerecentlytherehavebeenseveralexperimentsconductedatHAARPtoexperimentallyevaluatetheseeffects,suchasSTF-CWstepsandSTF-TOA.ItwillbeexcitingtoevaluatethemagnitudeoftheELF/VLFwavesandconsequentlythesuppressionasafunctionofCWERPforthisnon-linearinteraction. 6.4EvaluationofIonosphericCurrentDrive(ICD)Recently,therehavebeeneffortstogenerateULF/ELFwavesusingmodulatedF-regionHFheatingwithoutthepresenceofanelectrojetcurrent[e.g., Papadopoulosetal. 2011a b ; Eliassonetal. 2012 ].ThismethodusestheheatertomodulatetheF-regionatlowfrequencies(<50Hz)togenerateMagnetoSonic(MS)waves.By 116

PAGE 117

incorporatingchangesinelectrondensitywithtimeinourexistingdual-beamHFmodel,wecouldgetmoreinsightintothegenerationofULF/ELFwavesbyICD.ThisisadirectapplicationthatcouldprovidesomeimportantinformationaboutNeasitdependsonthemodulationofNe,usinglowerfrequenciesandaccountingforNechangeswithtime. 6.5HFCross-ModulationUnderDual-BeamHeatingConditionsInSection 1.3.5 ,wediscussedHFCross-ModulationprobingexperimentsthathavebeenusedtodeterminetheextentofionosphericconductivitymodulationintheD-regionionosphere.Recently, LangstonandMoore [ 2013 ]demonstratedusingHFcross-modulationduringHFheatingexperimentsatHAARPthatitwaspossibletoquantizetheD-regionabsorptionproducedbyHFheatingbothduringtheinitialstagesofheatingandundersteady-stateconditions.ItwouldbeinterestingtocalculateHFcross-modulationunderdual-beamheatingconditions,andinturnquantifytheD-regionabsorptionasafunctionofCWERP.Byevaluatingtheabsorptionproducedbythepiecewise-exponentialNeproles,wecouldpossiblyfurtherconrmthestructureintheD-region. 117

PAGE 118

REFERENCES Agrawal,D.,andR.C.Moore(2012),Dual-beamELFwavegenerationasafunctionofpower,frequency,modulationwaveform,andreceiverlocation,J.Geophys.Res.,117(A12305),doi:10.1029/2012JA018,061. Appleton,E.V.(1932),Wirelessstudiesoftheionosphere,J.Inst.Elec.Engrs,71,642. Appleton,E.V.,andM.A.F.Barnett(1925),Localreectionofwirelesswavesfromtheupperatmosphere,Nature(London),115,333334. Banks,P.(1966),Collisionfrequenciesandenergytransfer:electrons,Planet.SpaceSci.,14,1085. Banks,P.M.,andJ.R.Doupnik(1975),Areviewofauroralzoneelectrodynamicsdeducedfromincoherentscatterradarobservations,J.Atmos.Terr.Phys.,37,951. Barr,R.,andP.Stubbe(1984),ELFandVLFradiationfromthepolarelectrojetantenna,RadioSci.,19,1111. Barr,R.,andP.Stubbe(1991),ELFradiationfromtheTromsoSuperHeaterfacility,Geophys.Res.Lett.,18(6),1035. Barr,R.,andP.Stubbe(1993),ELFharmonicradiationfromtheTromsoheatingfacility,Geophys.Res.Lett.,20,2243. Bernhart,P.A.,P.J.Erickson,F.D.Lind,J.C.Foster,andB.W.Reinisch(2005),ArticialdisturbancesoftheionosphereovertheMillstoneHillIncoherentScatterRadarfromdedicatedburnsofthespaceshuttleorbitalmaneuversubsystemengines,J.Geophys.Res.,110(A5),A05,311. Bittencourt,J.A.(1986),FundamentalsofPlasmaPhysics. Boyd,S.(2004),ConvexOptimization,CambridgeUniversityPress. Budden,K.G.(1985),ThepropagationofRadioWaves:TheTheoryofRadioWaves:TheTheoryofRadioWavesofLowPowerintheIonosphereandMagnetosphere,CambridgeUniversityPress,CambridgeandNewYork. Chau,J.L.,andR.F.Woodman(2005),DandEregionincoherentscatterradardensitymeasurementsoverJicamarca,J.Geophys.Res.,110(A12),doi:10.1029/2005JA011438. Cheng,Z.,S.A.Cummer,D.N.Baker,andS.G.Kanekal(2006),NighttimeDregionelectrondensityprolesandvariabilitiesinferredfrombroadbandmeasurementsusingVLFradioemissionsfromlightning,J.Geophys.Res.,111(A5),doi:10.1029/2005JA011308. 118

PAGE 119

Cohen,M.B.,U.S.Inan,M.Golkowski,andM.J.McCarrick(2010),ELF/VLFwavegenerationviaionosphericHFheating:Experimentalcomparisonofamplitudemodulation,beampainting,andgeometricmodulation,J.Geophys.Res.,115(A02302),doi:10.1029/2009JA014,410. Cohen,M.B.,M.Golkowski,N.G.Lehtinen,U.S.Inan,andM.J.McCarrick(2012),HFbeamparametersinELF/VLFwavegenerationviamodulatedheatingoftheionosphere,J.Geophys.Res.,117(A05327). Cummer,S.A.,andU.S.Inan(1997),Measurementofchargetransferinsprite-producinglightningusingELFradioatmospherics,Geophys.Res.Lett.,24,1731. Cummer,S.A.,U.S.Inan,andT.F.Bell(1998),IonosphericDregionremotesensingusingVLFradioatmospherics,RadioSci.,33(6),1781,doi:10.1029/98RS02381. Dalgarno,A.,M.B.McElroy,M.H.Rees,andJ.C.G.Walker(1968),Theeffectofoxygencoolingonionosphericelectrontemperatures,Planet.SpaceSci.,16,1371. Davies,K.(1990),IonosphericRadio,580pp,PeterPeregrinusLtd.,London. deForest,L.(1912),Absorption(?)ofundampedwaves,Electrician,69,369. Djuth,F.T.,B.W.Reinisch,D.F.Kitrosser,J.H.Elder,A.L.Snyder,andG.S.Sales(2006),ImagingHF-inducedlarge-scaleirregularitiesaboveHAARP,Geophys.Res.Lett.,33(4),doi:10.1029/2005GL024536. Eliasson,B.,C.-L.Chang,andK.Papadopoulos(2012),Generationofelfandulfelectromagneticwavesbymodulatedheatingoftheionosphericf2region,JournalofGeophysicalResearch:SpacePhysics,117(A10),n/an/a,doi:10.1029/2012JA017935. Fallen,C.T.,J.A.Secan,andB.J.Watkins(2011),In-situmeasurementsoftopsideionosphereelectrondensityenhancementsduringanHF-modicationexperiment,Geophys.Res.Lett.,38(8),doi:10.1029/2011GL046887. Fejer,J.A.(1970),Radiowaveprobingofthelowerionospherebycross-modulationtechniques,J.Atmos.Terr.Phys.,32(597). Ferraro,A.J.,H.S.Lee,R.Allshouse,K.Carroll,R.Lunnen,andT.Collins(1984),CharacteristicsofionosphericELFradiationgeneratedbyHFheating,J.Atmos.Terr.Phys.,46,855. Fujimaru,S.,andR.C.Moore(2011a),AnalysisofTime-of-ArrivalObservationsPerformedduringELF/VLFWaveGenerationExperimentsatHAARP,RadioSci.,46(RS0M03),doi:10.1029/2011RS004,695. 119

PAGE 120

Fujimaru,S.,andR.C.Moore(2011b),Time-of-ArrivalanalysisappliedtothespatiallydistributedELF/VLFsourceregionaboveHAARP,Master'sthesis,UniveristyofFlorida,Gainesville. Hargreaves,J.K.(1992),TheSolar-TerrestrialEnvironment,CambridgeUniversityPress. Hedin,A.E.(1991),ExtensionoftheMSISThermosphericModelintotheMiddleandLowerAtmosphere,J.Geophys.Res. Holmes,J.M.,T.R.Pedersen,andT.J.Mills(2011),RF-Inducedairglowobservedusingcompositemultispectralimaging,PlasmaScience,IEEETrans.,39(11),27142715. Huxley,L.G.H.,andJ.A.Ratcliffe(1949),Asurveyofionosphericcross-modulation,Proc.Inst.Elec.Eng.,96,433. Hysell,D.L.(2008),30MHzradarobservationsofarticialEregioneld-alignedplasmairregularities,AnnalesGeophysicae,26,117. Jacobson,A.R.,R.Holzworth,andX.-M.Shao(2008),Low-frequencyionosphericsoundingwithNarrowBipolarEventlightningradioemissions:energy-reectivityspectrum,AnnalesGeophysicae,26(7),1793,doi:10.5194/angeo-26-1793-2008. James,H.G.,R.L.Dowden,M.T.Rietveld,P.Stubbe,andH.Kopka(1984),SimultaneousobservationsofELFwavesfromanarticiallymodulatedauroralelectrojetinspaceandontheground,J.Geophys.Res.,89,1655. Kendall,E.,R.Marshall,R.T.Parris,A.Bhatt,A.Coster,T.Pedersen,P.Bernhardt,andC.Selcher(2010),Decameterstructureinheater-inducedairglowattheHighfrequencyActiveAuroralResearchProgramfacility,J.Geophys.Res.,115(A8),doi:10.1029/2009JA015043. Kou,Y.,X.Zhou,Y.Morton,andL.Zhang(2010),ProcessingGPSL2Csignalsunderionosphericscintillations,inPositionLocationandNavigationSymposium(PLANS),pp.771782,IEEE/ION. Labitzke,K.,J.J.Barnett,andB.E.(eds.)(1985),HandbookMAP16,SCOSTEP,UniversityofIllinois,Urbana. Langston,J.,andR.C.Moore(2013),HightimeresolutionobservationsofHFcross-modulationwithingtheD-regionionosphere,Geophys.Res.Lett.,p.doi:10.1002/grl.50391. Lay,E.H.,andX.M.Shao(2011),Multi-stationprobingofthunderstorm-generatedD-layeructuationsbyusingtime-domainlightningwaveforms,Geophys.Res.Lett.,38(23),doi:10.1029/2011GL049790. 120

PAGE 121

Lehtinen,N.G.,andU.S.Inan(2008),RadiationofELF/VLFwavesbyharmonicallyvaryingcurrentsintoastratiedionospherewithapplicationtoradiationbyamodulatedelectrojet,J.Geophys.Res.,113(A06301). Maslin,N.M.(1974),Theoryofenergyuxandpolarizationchangesofaradiowavewithtwomagnetoioniccomponentsundergoingselfdemodulationintheionosphere,RoyalSocietyofLondonProceedingsSeriesA,341,361. Mathews,J.D.(1984),TheincoherentscatterradarasatoolforstudyingtheionosphericD-region,JournalofAtmosphericandTerrestrialPhysics,46(11),975986,doi:10.1016/0021-9169(84)90004-7. Mathews,J.D.(1986),IncoherentScatterRadarProbingofthe60-100kmAtmosphereandIonosphere,IEEETrans.onGeoSci.andRemoteSensing,GE-24(5),765. McRae,W.M.,andN.R.Thomson(2000),VLFphaseandamplitude:daytimeionosphericparameters,JournalofAtmosphericandSolar-TerrestrialPhysics,62(7),609. Mentzoni,M.H.,andR.V.Row(1963),Rotationalexcitationandelectronrelaxationinnitrogen,Phys.Rev.,130,2312. Milikh,G.M.,andK.Papadopoulos(2007),EnhancedionosphericELF/VLFgenerationefciencybymultipletimescalemodulatedheating,Geophys.Res.Lett.,34. Milikh,G.M.,A.Gurevich,K.Zybin,andJ.Secan(2008),PerturbationsofGPSsignalsbytheionosphericirregularitiesgeneratedduetoHF-heatingattripleofelectrongyrofrequency,Geophys.Res.Lett.,35(L22102). Milikh,G.M.,E.Mishin,I.Galkin,A.Vartanyan,C.Roth,andB.W.Reinisch(2010),IonoutowsandarticialductsinthetopsideionosphereatHAARP,Geophys.Res.Lett.,37,L18,102,14. Moore,R.C.(2007),ELF/VLFwavegenerationbymodulatedHFheatingoftheauroralelectrojet,Ph.D.thesis,StanfordUniversity,Stanford,California. Moore,R.C.,andD.Agrawal(2011),ELF/VLFwavegenerationusingsimultaneousCWandmodulatedHFheatingoftheionosphere,J.Geophys.Res.,116(A04217),doi:10.1029/2010JA015,902. Moore,R.C.,U.S.Inan,T.F.Bell,andE.J.Kennedy(2007),ELFwavesgeneratedbymodulatedHFheatingoftheauroralelectrojetandobservedatagrounddistanceof4400km,J.Geophys.Res.,112(A05309). Nicolet,M.,andA.C.Aikin(1960),TheFormationoftheD-RegionoftheIonosphere,J.Geophys.Res.,65(5),1469. Papadopoulos,K.,C.L.Chang,P.Vitello,andA.Drobot(1990),OntheefciencyofionosphericELFgeneration,RadioSci.,25,1311. 121

PAGE 122

Papadopoulos,K.,T.Wallace,M.McCarrick,G.M.Milikh,andX.Yang(2003),OntheefciencyofELF/VLFgenerationusingHFheatingoftheauroralelectrojet,PlasmaPhys.Rep.,29,561. Papadopoulos,K.,T.Wallace,G.M.Milikh,W.Peter,andM.McCarrick(2005),ThemagneticresponseoftheionospheretopulsedHFheating,Geophys.Res.Lett.,32(L13101). Papadopoulos,K.,N.A.Gumerov,X.Shao,I.Doxas,andC.L.Chang(2011a),Hf-drivencurrentsinthepolarionosphere,GeophysicalResearchLetters,38(12),n/an/a,doi:10.1029/2011GL047368. Papadopoulos,K.,C.-L.Chang,J.Labenski,andT.Wallace(2011b),Firstdemonstrationofhf-drivenionosphericcurrents,GeophysicalResearchLetters,38(20),n/an/a,doi:10.1029/2011GL049263. Payne,J.A.(2007),Spatialstructureofverylowfrequencymodulatedionosphericcurrents,Ph.D.thesis,StanfordUniversity,Stanford,California. Payne,J.A.,U.S.Inan,F.R.Foust,T.W.Chevalier,andT.F.Bell(2007),HFmodulatedionosphericcurrents,Geophys.Res.Lett.,34(L23101). Pickett,J.S.,G.B.Murphy,W.S.Kurth,C.K.Goertz,andS.D.Shawhan(1985),EffectsofChemicalReleasesbytheSTS3OrbiterontheIonosphere,J.Geophys.Res.,90(A4),3487. Prakash,S.,andR.Pandey(1984),RocketbornestudiesofelectrondensityirregularitiesinequatorialDandEregions,J.EarthSystemScience,93(3),283. Prasad,S.S.,andD.R.Furman(1973),Electroncoolingbymolecularoxygen,J.Geophys.Res.,78,6701. Rakov,V.A.,andM.A.Uman(2003),LightningPhysicsandEffects,ISBN0521583276,PBISBN0521035414,CambridgeUniversityPress. Rietveld,M.T.,H.Kopka,andP.Stubbe(1986),D-regioncharacteristicsdeducedfrompulsedionosphericheatingunderauroralelectrojetconditions,J.Atmos.Terr.Phys.,48,311. Rietveld,M.T.,P.Stubbe,andH.Kopka(1989),OnthefrequencydependenceofELF/VLFwavesproducedbymodulatedionosphericheating,RadioSci.,24,270. Rishbeth,H.,andO.K.Garriott(1969),IntroductiontoIonosphericPhysics,NewYork:AcademicPress. Rodriguez,J.V.(1994),ModicationoftheEarth'sionospherebyvery-low-frequencytransmitters,Ph.D.thesis,StanfordUniversity,Stanford,California. 122

PAGE 123

Sechrist,C.F.J.(1974),ComparisonoftechniquesformeasurementofD-regionelectrondensities,RadioSci.,9(2),137. Seddon,J.C.,A.D.Pickar,andJ.E.Jackson(1954),ContinuousElectronDensityMeasurementsUpTo200km,J.Geophys.Res.,59(4),513. Senior,A.,M.T.Rietveld,M.J.Kosch,andW.Singer(2010),Diagnosingradioplasmaheatinginthepolarsummermesosphereusingcrossmodulation:Theoryandobservations,J.Geophys.Res.,115(doi:10.1029/2010JA015379),A09,318. Stubbe,P.,andH.Kopka(1977),ModulationofpolarelectrojetbypowerfulHFwaves,J.Geophys.Res.,82,2319. Stubbe,P.,andW.S.Varnum(1972),Electronenergytransferratesintheionosphere,Planet.SpaceSci.,20,1121. Stubbe,P.,H.Kopka,andR.L.Dowden(1981),GenerationofELFandVLFwavesbypolarelectrojetmodulation:Experimentalresults,J.Geophys.Res.,86,9073. Stubbe,P.,H.Kopka,M.T.Rietveld,andR.L.Dowden(1982),ELFandVLFwavegenerationbymodulatedHFheatingofthecurrentcarryinglowerionosphere,JournalofAtmosphericandTerrestrialPhysics,44(12),1123. Tellegen,B.D.H.(1933),Interactionsbetweenradiowaves,Nature(London),131,840. Tomko,A.A.(1980),Dregionabsorptioneffectsduringhigh-powerradiowaveheating,,RadioSci.,15,675. Tomko,A.A.(1981),Nonlinearphenomenaarisingfromradiowaveheatingofthelowerionosphere,TechnicalReportPSU-IRL-SCI-470,IonosphereResearchLaboratory,PennsylvaniaStateUniversity,UniversityPark,Pennsylvania. Tripathi,V.K.,C.L.Chang,andK.Papadopoulos(1982),ExcitationoftheEarth-ionospherewaveguidebyanELFsourceintheionosphere,RadioSci.,17,1321. Tulasi,S.R.,S.Y.Su,C.H.Liu,B.W.Reinisch,andL.A.Mckinnell(2009),Topsideionosphericeffectivescaleheights(HT)derivedwithROCSAT-1andground-basedionosondeobservationsatequatorialandmidlatitudestations,J.Geophys.Res.,114(A10),A10,309. Ulwick,J.C.,K.D.Baker,M.C.Kelley,B.B.Balsley,andW.L.Ecklund(1988),ComparisonofSimultaneousMSTRadarandElectronDensityProbeMeasurementsDuringSTATE,J.Geophys.Res.,93(D6),6989. Villard,O.G.J.(1976),Theionosphericsounderanditsplaceinthehistoryofradioscience,RadioSci.,11(11),847860. 123

PAGE 124

Villasenor,J.,A.Y.Wong,B.Song,J.Pau,M.McCarrick,andD.Sentman(1996),ComparisonofELF/VLFgenerationmodesintheionospherebytheHIPASheaterarray,RadioSci.,31,211. Wait,J.R.,andK.P.Spies(1964),CharacteristicsoftheEarth-ionospherewaveguideforVLFradiowaves,Tech.rep.,NationalBlur.ofStand.,Boulder,Colo. Weisbrod,S.,A.J.Ferraro,andH.S.Lee(1964),Investigationofphaseinteractionasameansofstudyingthelowerionosphere,J.Geophys.Res.,69(2337). Zinn,J.,andC.D.Sutherland(1980),Effectsofrocketexhaustproductsinthethermosphereandionsphere,TechnicalReportLA-8233-MS,LosAlamosScienticLab.,NM(USA). 124

PAGE 125

BIOGRAPHICALSKETCH DivyaAgrawalreceivedherBachelorofScienceinElectronicsandCommunicationengineeringin2005fromSriJayachamarajendraCollegeofEngineering,India.SheearnedaMasterofSciencedegreeandaPh.D.inElectricalandComputerEngineeringin2008and2013,respectively,fromtheUniversityofFlorida,Gainesville,Florida.Herinterestsincludesignalprocessing,antennadesign,andelectromagneticeldtheory. 125