Unique concurrent observations of whistler modehiss, chorus, and triggered emissionsPoorya Hosseini1 , Mark Gołkowski1 , and Drew L. Turner2

1Department of Electrical Engineering, University of Colorado Denver, Denver, Colorado, USA, 2The Aerospace Corporation,El Segundo, California, USA

Abstract We present a unique 2 h ground-based observation of concurrent magnetospheric hiss, chorus,VLF triggered emissions as well as ELF/VLF signals generated locally by the High Frequency Active AuroralResearch Program (HAARP) facility. Eccentricity of observed wave polarization is used as a criteria to identifymagnetospheric emissions and estimate their ionospheric exit points. The observations of hiss and chorus inthe unique background of coherent HAARP ELF/VLF waves and triggered emissions allow for more accuratecharacterization of hiss and chorus properties than in typical ground-based observations. Eccentricity andazimuth results suggest a moving ionospheric exit point associated with a single ducted path at L ~ 5. Theemissions exhibit dynamics in time suggesting an evolution of a magnetospheric source from hiss generationto chorus generation or a moving plasmapause location. We introduce a frequency band-limitedautocorrelation method to quantify the relative coherency of the emissions. A range of coherency wasobserved from high order of coherency in local HAARP transmissions and their echoes to lower coherency innatural chorus and hiss emissions.

1. IntroductionChorus and hiss emissions are among two of the most intense naturally occurring electromagnetic ELF/VLFband waves in the Earth’s near-space environment. Both types of emissions are whistler mode waves andare believed to be generated in the equatorial plane either outside (chorus) or inside (hiss) of the plasma-pause [Santolık and Gurnett, 2003; Golden et al., 2009; Bortnik et al., 2008, 2009]. The exact generation pro-cesses of these emissions are not well understood but are known to involve wave-particle interactions andare therefore coupled to the energy dynamics behind space weather [Bortnik and Thorne, 2007]. Chorusand hiss emissions are mostly analyzed and classified using frequency-time spectrograms based on shorttime Fourier transforms of digitized data. There is general consensus that hiss emissions have characteristicsof an incoherent band limited noise-like signal, while chorus emission are characterized by varying levels ofcoherent frequency rising or frequency falling tones. The relation between hiss and chorus emissions has alsobeen a topic of active research. Bortnik et al. [2008, 2009] put forth the theory that chorus emissions are thesource of hiss emissions, with their spectral properties losing coherence under the influence of propagationin the magnetosphere involvingmultiple magnetospheric reflections. Earlier work suggested that chorus andhiss were driven by cyclotron resonance processes either in the linear (hiss) or nonlinear (chorus) regimes[Koons, 1981]. Recent in situ observations by Summers et al. [2014] suggest that the spectral properties of hissand chorus are actually more similar than previously recognized in that hiss also exhibits coherent featuresbut on shorter time scales. Chorus generation mechanisms have also been studied in the context of VLF trig-gered emissions, a closely related process where magnetospheric wave growth is excited by a coherent inputwave [Helliwell, 1988; Omura et al., 1991; Omura et al., 2008; Streltsov et al., 2010, Gołkowski et al., 2011]. VLFtriggered emissions have been studied using the Siple Station transmitter in Antarctica [Helliwell, 1988]and more recently the High Frequency Active Auroral Research Program (HAARP) array in Alaska[Gołkowski et al., 2011, 2010, 2008]. Although the availability of spacecraft observations has improved rapidlyin recent years, providing higher-resolution data of the magnetospheric emissions or energetic particles inthe radiation belts [Summers et al., 2014], ground-based observation of these emissions is still an importanttool. Ground-based receivers let researchers investigate a specific L shell of the magnetosphere over a longduration with the capability of higher storage capacity essential for certain long-term and statistical studies ofwave properties [Golden et al., 2009].

However, work with ground observation involves the added complication of trans-ionospheric propagationof the magnetospheric emissions, since the waves observed on the ground propagate mostly along the

HOSSEINI ET AL. WHISTLER WAVES AND VLF TRIGGERED EMISSIONS 6271

PUBLICATIONSJournal of Geophysical Research: Space Physics

RESEARCH ARTICLE10.1002/2017JA024072

Key Points:• Unique simultaneous observation ofhiss, chorus, and HAARP-inducedtriggered emissions

• Eccentricity and autocorrelation arevaluable tools for analysis of hiss andchorus emissions

• Observations suggest evolution ofmagnetosphere whistler mode source

Correspondence to:P. Hosseini,poorya.hosseini@ucdenver.edu

Citation:Hosseini, P., M. Gołkowski, andD. L. Turner (2017), Unique concurrentobservations of whistler mode hiss,chorus, and triggered emissions,J. Geophys. Res. Space Physics, 122,6271–6282, doi:10.1002/2017JA024072.

Received 21 FEB 2017Accepted 7 JUN 2017Accepted article online 9 JUN 2017Published online 21 JUN 2017

©2017. American Geophysical Union.All Rights Reserved.

geomagnetic field lines in the magnetosphere and need to pass through the ionosphere to reach the surfaceof the Earth. In most cases, subsequent propagation in the Earth-ionosphere waveguide from the ionosphericexit point to the receiver further modifies the wave properties. In this work we present a unique observationof simultaneous hiss and chorus emissions in Alaska in conjunction with an active experiment at HAARPinvolving the generation of ELF/VLF waves and excitation of magnetospheric triggered emissions. Theobservations of hiss and chorus in the unique background of coherent HAARP ELF/VLF waves andtriggered emissions allow for more accurate characterization of hiss and chorus properties than in typicalground-based observations. The emissions exhibit dynamics in time suggesting an evolution of amagnetospheric source from hiss generation to chorus generation or a moving plasmapause location. Thespectral properties of the chorus and triggered emissions are seen to evolve in similar fashion, suggestingpossible changes in the hot plasma distribution.

2. Observations

Observations are made in Chistochina, Alaska, on 11 December 2008 using VLF receivers with orthogonal air-core loop antennas as described by Cohen et al. [2008]. During this time the HAARP HF transmitter located36 km from the Chistochina receiver was active and modulating the overhead electrojet currents to generateELF/VLF waves for magnetospheric wave injection. The HF transmissions were X-mode with HF frequencies of2.75 MHz and 3.25 MHz. The ELF/VLF waves generated by HAARP were single-frequency pulses andfrequency-time ramps in the range of 0.5–3 kHz. The location of the Chistochina receiver and HAARP trans-mitter is shown in Figure 1a. Figure 1b shows the magnetospheric context of the observations. The receiveris located at L ~ 5. Observed waves of magnetospheric origin are assumed to have propagated in field-aligned density enhancements (ducts) allowing for penetration of the lower ionospheric boundary[Gołkowski and Inan, 2008]. Chorus and hiss emissions originate from source regions at the equator. TheVLF triggered emissions observed in the vicinity of the HAARP facility are known as “two-hop echoes” havingtraversed the equatorial region twice and experienced wave-particle interactions leading toamplification therein.

A two-dimensional schematic of the Earth-ionosphere waveguide is depicted in Figure 2a, which shows howthe three different types of magnetospheric emissions (hiss, chorus, and two-hop echoes) can arrive at theantenna from potentially different ionospheric exit points at variable distances from the receiver.Moreover, signals can arrive directly from an ionospheric exit point to a receiver or after undergoing a seriesof reflections in the Earth-ionosphere waveguide. These reflections change the wave characteristics, whichwe exploit to estimate the distance to the exit points. Figure 2b shows the polarization ellipse that can becalculated from the two orthogonal antenna measurements and also provides an important diagnostic.

The 2 h observation of interest is shown in Figure 3a as a spectrogram with time on the horizontal axis, fre-quency on vertical axis, and color showing wave power spectral density. Wave energy is initially spread over a

Figure 1. (a) Chistochina receiver and HAARP transmitter locations in Alaska. (b) A representation of propagation paths and assumed source regions of whistlermode chorus, hiss, and triggered emission (HAARP two-hop echoes) waves in the magnetosphere.

Journal of Geophysical Research: Space Physics 10.1002/2017JA024072

HOSSEINI ET AL. WHISTLER WAVES AND VLF TRIGGERED EMISSIONS 6272

range of frequencies spanning 1.5–2.8 kHz with this active band narrowing over the 2 h observation period.The maximum amplitudes on the color scale correspond to 0.5 pT. To show the data features of specificemissions in more detail, six 30 s time slots are shown in Figures 3b–3g corresponding to the vertical linedemarcations in Figure 3a. Time slot (Figure 3b) shows broadband hiss, plus local HAARP-generated pulsesand ramps; we call this hiss emission “Hiss 1.” In time slot Figure 3c, hiss emissions continue but withnarrower bandwidth and are designated as “Hiss 2.” Time slot Figure 3d shows chorus emissions whichcontinue to time slot Figure 3e when HAARP-induced two-hop echoes appear. In time slot Figure 3f onlytwo-hop echoes induced by HAARP-generated pulses and ramps transmissions are observed (with no hissnor chorus).

This data set is unique in that three of the most important whistler mode magnetospheric emissions occurconcurrently. Besides the natural magnetospheric waves, chorus and hiss, and the triggered emissions, thereare also local HAARP signals present which provide an important fully coherent reference signal in thesame band.

Although data shown in Figure 3 are from one antenna oriented in themagnetic east-west direction (EW), ourground-based VLF receiver consists of two loop antennas which are oriented orthogonal to each other.Recording two components of the wave magnetic fields enables us to calculate the arrival azimuth angleand the eccentricity of the received emissions. In this context we follow the method used by Gołkowskiand Inan [2008] and Maxworth et al. [2015] where the EW and NS channels are combined as the imaginaryand real part, respectively, of a complex signal. Applying the fast Fourier transform, the composite complexsignal is put into the frequency domain as

F ωð Þ ¼ ∫þ∞�∞ NS tð Þ þ iEW tð Þð Þ e�iωtdt (1)

The obtained frequency response can be written as positive and negative frequency components as follows:

F ωð Þ ¼ rþeiθþ F �ωð Þ ¼ r�eiθ� (2)

The major and minor diameters of the polarization ellipse (Figure 2b) can be found as follows:

A ¼ rþ þ r� B ¼ rþ � r�j j (3)

In this case, the arrival azimuth angle and eccentricity of the ellipse can be written as follows:

φ ¼ θþ � θ�2

ecc ¼ffiffiffiffiffiffiffiffiffiffiffiffiffi1� B2

A2

s(4)

The arrival azimuth and eccentricity are important tools for categorizing the received whistler mode emis-sions. The dispersion relationship of a whistler mode wave propagating parallel to the geomagnetic fieldrequires right-hand circular polarization (RHCP). Although in general magnetospheric whistler mode

Figure 2. (a) Two-dimensional schematic of Earth-ionosphere waveguide showing potential ionospheric exit points and paths to the receiver. (b) General polariza-tion ellipse in reference to two orthogonal loop antennas.

Journal of Geophysical Research: Space Physics 10.1002/2017JA024072

HOSSEINI ET AL. WHISTLER WAVES AND VLF TRIGGERED EMISSIONS 6273

Figure 3. (a) Spectrogram of the 2 h observation (color showing wave power spectral density). To show the data features in more details, six 30 s time slot are shownin (b) broadband hiss, plus HAARP-generated pulses and ramps Hiss 1; (c) hiss emissions continue but with narrower bandwidth and are designated as Hiss 2,(d) chorus emissions which continue to time slot (e) HAARP-induced two-hop echoes, and (f) only two-hop echoes induced by HAARP-generated pulses and rampstransmissions. Black arrows in Figures 3e and 3f show the two-hop echoes corresponding to HAARP-generated ELF/VLF signals. (g) No significant emissions observedin timeslot.

Journal of Geophysical Research: Space Physics 10.1002/2017JA024072

HOSSEINI ET AL. WHISTLER WAVES AND VLF TRIGGERED EMISSIONS 6274

emissions can propagate obliquely and have elliptical polarizations, the requirement for waves reaching aground-based receiver to have propagated in the ducted mode means that their polarization must beclose to circular (ecc < <1) in the immediate vicinity of the ionospheric exit point. However, once RHCPwaves propagate inside the Earth-ionosphere waveguide, they quickly convert to linear polarization uponreflection from the ground and ionosphere [Maxworth et al., 2015]. These reflections change theeccentricity of the polarization and make the waves be more linearly polarized (ecc ≈ 1). Thus, the relativeeccentricity of the waves can be used to diagnose the exit point locations.

3. Analysis3.1. Analysis of Eccentricity and Arrival Azimuth

To be able to track the evolution of emissions in the 2 h data set, we divide the data into 30 min consecutivesubsections characterized by the dominant emissions: Hiss 1, Hiss 2, chorus, or two-hop echoes occurringduring this time as shown in Figure 3. The eccentricity as a function of frequency and time of the mentionedemissions is calculated and shown in Figure 4. The blue parts of the color scales show RHCP waves which isconsistent with what we expect for magnetospheric whistler mode emissions observed close to their iono-spheric exit point. By comparing the blue color density of subsequent panels, it can be seen that the eccen-tricity level of the emissions is becoming more circular from Hiss 1 (first half an hour in Figure 4a) to two-hopechoes (fourth half an hour in Figure 4d). An important fact in this context is that the two-hop echoes areknown to exit the ionosphere directly above the HAARP heated region [Gołkowski et al., 2009, 2011]. Sincethe heated region has a diameter of ~30 km, the two-hop echoes in effect come in directly above the receiverand it is not surprising that their eccentricity is the most circular.

The local HAARP-generated ELF/VLF signals are expected to have eccentricity close to linear (ecc = 1) sincethe HAARP ELF/VLF generation process is known to be similar to that of an electric dipole antenna [Cohenet al., 2008; Maxworth et al., 2015]. This consistency with the theory and past work can be seen inFigures 4a and 4c with local HAARP pulses showing up in red (ecc ≈ 1).

Figure 4. Eccentricity as a function of frequency and time for (a) Hiss 1, (b) Hiss 2, (c) chorus, and (d) two-hop HAARP echo (color showing 0< ecc< 1).

Journal of Geophysical Research: Space Physics 10.1002/2017JA024072

HOSSEINI ET AL. WHISTLER WAVES AND VLF TRIGGERED EMISSIONS 6275

In addition to knowing the polarization of the received waves, another important parameter is the arrival azi-muth angle. In Figure 5 spectrograms of the four main emissions are illustrated with color scale showing arri-val azimuth. Again, taking the local HAARP pulses as a reference, it is clear that all HAARP pulses and rampsneed to reach our receiver with the same azimuth angle (Figures 5a, 5c, and 5d). Another important point isthat the dominant color scale components are changing in each panel. The implications of this azimuth tran-sition are discussed below.

As mentioned above, one of the biggest challenges in interpreting ground-based observations of magneto-spheric emissions, is that the emissions pass through the ionosphere and also propagate in the Earth-ionosphere waveguide to reach the receiver. Moreover, the Earth-ionosphere waveguide typically hosts alarge number of signals in the ELF/VLF band from both magnetospheric and terrestrial origin. Propagatingthrough the Earth-ionosphere waveguide changes the features of the emissions, especially polarization,making the eccentricity more linear. To deal exclusively with only magnetospheric emissions in the receivedsignals (and exclude emissions that have propagated a considerable distance in the Earth-ionosphericwaveguide), eccentricity can be used as a filter. Emissions with low eccentricities (less than 0.9) can beuniquely identified as magnetospheric emissions from nearby exit points. Moreover, the value of the eccen-tricity can provide an estimate of the distance between the receiver and ionosphere exit point.

We applied this eccentricity filtering process to amplitude and arrival azimuth spectrograms. Figure 6 showseccentricity-filtered amplitude spectrograms for the four main emissions. The color scale is adjusted to beable to have the white background reflect nonmagnetospheric emissions. Comparing the original spectro-grams (Figures 3b–3d and 3f) and eccentricity-filtered spectrograms (Figures 6a–6d) for the four mainemissions (Hiss 1, Hiss 2, chorus, and two-hop echoes) shows how the data records have been cleanedfrom background signals (with eccentricity above 0.9) leaving only magnetospheric emissions with lowereccentricity than 0.9. After applying this filter we repeated the azimuth calculations which showeddecreased variance.

The filtered arrival azimuth spectrograms, which have not shown here to save space, were averaged to createa polar statistical graph of arrival azimuth distribution for each half hour as is shown in Figure 7. Each panel

Figure 5. Arrival azimuth angle as a function of frequency and time for (a) Hiss 1, (b) Hiss 2, (c) chorus, and (d) two-hop HAARP echo (color showing�90< φ < 90).

Journal of Geophysical Research: Space Physics 10.1002/2017JA024072

HOSSEINI ET AL. WHISTLER WAVES AND VLF TRIGGERED EMISSIONS 6276

shows the normalized azimuth occurrence rate of the dominant emission with a color scale. The plots alsoexhibit the inherent 180° ambiguity in azimuth determination using two orthogonal horizontal loops.

White arrows in Figure 7 show what is believed to be a rotation or transition of dominant ionospheric exitpoint from �110° (Hiss 1) to �50° (Hiss 2), to �70° and �110° (Chorus), and to �120° (two-hop echoes).During two-hop echo observations we can see that the azimuth points toward the HAARP facility locationas expected. We note that the average eccentricity for each half an hour is obtained as 0.85 (Hiss 1), 0.8(Hiss 2), 0.7 (chorus), and 0.6 (two-hop HAARP echo). The falling eccentricity results suggest that the exit pointis getting closer to the receiver. Since we do not see evidence of multiple exit points, this leads us to believethat all of the emissions are propagating in the same duct. In other words, the same duct is deliveringfirst broadband hiss then hiss in a narrower band and then chorus emissions into the Earth-ionosphere waveguide.

Hosseini et al. [2017] quantitatively investigated how the eccentricity changes versus exit point’s distance, bysimulating the Earth-ionosphere waveguide with finite difference time domain method. Figure 8 summarizesthe mentioned exit points’ transition with actual simulated exit point distances. It shows that exit points aregetting closer to the Chistochina’s location, and also HAARP location, which is consistent with the fact thattwo-hop HAARP echoes should have exit point above the HAARP location.

3.2. Analysis of Emission Coherency

Traditionally, hiss and chorus have been categorized qualitatively with hiss defined as an incoherent noise-like signal and chorus as coherent rising tones or falling tones. A quantitative coherency metric is seen asnecessary to facilitate accurate discussion of these emissions. We note that Tsurutani et al. [2015] used crosscorrelation between different spacecraft magnetic probes to identify the level of wave coherency. However,such a technique is not applicable to ground-based data since in contrast to themagnetosphere, themediumat the point of observation is no longer anisotropic. In this context we propose a coherency measurementmethod based on frequency band-limited (~ 400 Hz) autocorrelation over 1 s duration of data. We have

Figure 6. Eccentricity filtered amplitude spectrograms as a function of frequency and time for (a) Hiss 1, (b) Hiss 2, (c) chorus, and (d) two-hop HAARP echo (modifiedcolor showing filtered wave power spectral density).

Journal of Geophysical Research: Space Physics 10.1002/2017JA024072

HOSSEINI ET AL. WHISTLER WAVES AND VLF TRIGGERED EMISSIONS 6277

applied this technique to the four different magnetospheric emissions as well as the local ELF/VLF HAARPsignals, which serve as the reference of highly coherent signals and band-limited white Gaussian noise,which serves as the reference for a truly incoherent signal.

The autocorrelation results are shown in Figure 9a. Themarkers plotted are the successive peaks of the result-ing autocorrelation, which in general have an oscillatory nature as expected for band-limited signals. Thevalues of the first four peak values have been extracted in Figure 9b. We use the values of these peaks as ameasure of coherency with the highest values associated with the highest level of coherence. As illustratedin this figure, HAARP pulses and Hiss 1 exhibit the highest and lowest coherency levels, respectively. HAARPfrequency-time ramps, two-hop echoes, chorus, and Hiss 2 have coherency measures that fall in betweenthese two boundaries of coherency. The lowest levels of coherency occur for Gaussian white noise (WGN)where we distinguish between “filtered WGN”where the bandwidth is limited to the same 400 Hz bandwidthused to analyze the emissions and “WGN”where the bandwidth is 50 kHz corresponding to half the samplingrate of the receiver hardware. Using our metrics, the filtered WGN shows higher coherency than the WGN butboth are less coherent than the observed hiss emissions. Specific values of our coherency measure for multi-ple peaks are shown in Table 1, and all show the same trend. Namely, there is decreasing coherency fromHAARP transmissions and their echoes to less coherent (natural chorus and hiss emissions), and all are

Figure 7. Polar statistical graph of arrival azimuth distribution for each half hour. Normalized azimuth occurrence rate of the (a) Hiss 1, (b) Hiss 2, (c) chorus, and(d) two-hop echo emission is shown with color scale. White arrows are related to the maximum normalized arrival azimuth occurrence.

Journal of Geophysical Research: Space Physics 10.1002/2017JA024072

HOSSEINI ET AL. WHISTLER WAVES AND VLF TRIGGERED EMISSIONS 6278

distinguishable from white Gaussian noise. This suggests that this technique could be used to quantitativelycategorize hiss, chorus, and background noise in a wide range of spacecraft and ground observations.

4. Discussion

Using eccentricity as a criteria to uniquely identify magnetospheric emissions in the 2 h ground-based obser-vation enabled us to accurately describe the characteristics of two different types of naturally occurring emis-sions and compare their properties to man-made HAARP transmissions and induced two-hop echoes. The

Figure 8. Maximum normalized arrival azimuth occurrence of the Hiss 1, Hiss 2, chorus, and two-hop echo emission isshown with constant eccentricity of 0.85 (light blue), 0.80 (dark blue), 0.70 (green), and 0.60 (black) circles, respectively(which is based on simulated propagation results for various exit point-receiver distances).

Figure 9. (a) Frequency band-limited (~ 400 Hz) autocorrelation over 1 s durations of data. (b) The markers plotted are the successive peaks of the resulting auto-correlation, showing the first four peak values.

Journal of Geophysical Research: Space Physics 10.1002/2017JA024072

HOSSEINI ET AL. WHISTLER WAVES AND VLF TRIGGERED EMISSIONS 6279

eccentricity and azimuth results both suggest a moving ionospheric exit point associated with a singleducted path at L ~ 5. The exit point is seen to rotate slightly in azimuth and approach the receiver achievinga final position in the HAARP heated region. Although the ionospheric exit point is moving, we believe themagnetospheric portion of the duct to remain little changed. This is supported by the fact that the two-hop echo propagation time remains constant. The exit point can be tracked, and we do not see evidenceof multiple exit point corresponding to multiple ducts. The nature of the magnetospheric emissions emanat-ing from this duct is seen to evolve from broadband hiss, to narrowband hiss to chorus to two-hop echoes.The coherence of these emissions during this time, which we quantify using autocorrelation, is seen toincrease. Below we present two possible explanations of what the observations suggest in terms of processesin the magnetospheric source region.

The first scenario we consider is that all the magnetospheric emissions originate from the same magneto-spheric source which is evolving in time. Initially, the source region has a hot plasma distribution favorableto broadband hiss generation. This could mean a distribution with anisotropy sufficient for linear growthof whistler mode waves but not sufficient for nonlinear growth associated with chorus and triggered emis-sions. The narrowing of the hiss band can be seen as evidence of the distribution changing to be more favor-able to growth in a narrower frequency range. This evolution continues as the source transitions to generatechorus emissions. Finally, the distribution which was already generating chorus is stimulated by the HAARPELF/VLF signals generating observable two-hop echoes. The chorus emissions are seen to disappear as thetwo-hop echoes dominate the last half hour of observations. This might suggest that the distribution energyis being coupled away from chorus to the triggered emissions and has important implications for the conceptof radiation belt mitigation [Inan et al., 2003]. Raghuram et al. [1977] report a similar phenomena where trans-missions near 5 kHz from Siple Station triggered frequency risers and suppressed hiss in a 200 Hz band belowthe transmitted frequency. Such competition between the HAARP-triggering signals and the natural chorusproducing instability suggests the exciting possibility of anthropogenic control of natural turbulent processesin the magnetosphere. In this context, we investigated the sweep rates of chorus rising tones and two-hopHAARP echoes either from pulses or ramps. The results are depicted in Table 2, where it is shown that evenacross two different emission types during the last 1 h of the observation, the sweep rates show a consistentpattern of decreasing. This smooth transition from chorus rising tones to two-hop echoes may be further evi-dence of a common source that is evolving in time and able to yield both naturally occurring chorus and arti-ficially triggered whistler mode waves.

Table 1. The First Four Peak Values of the Successive Peaks Markers of the Resulting Autocorrelation, Which in GeneralHave an Oscillatory Nature as Expected for Band-Limited Signals

Emission Type

Autocorrelation Peak

Second Peak Third Peak Fourth Peak

HAARP pulse 0.96 0.84 0.86HAARP ramp 0.96 0.90 0.85Two hop HAARP echo 0.88 0.73 0.43Chorus 0.89 0.70 0.42Hiss 2 0.90 0.67 0.42Hiss 1 0.89 0.64 0.36WGN (400 Hz) 0.89 0.62 0.33WGN (50 KHz) 0.03 0.03 0.2

Table 2. Sweep Rates Measurement of Chorus Rising Tones and Two-Hop HAARP Echoes Either From Pulses or RampsDuring the Last 1 h of the Observation

Emission Type

Time (Minutes After 2:00 UT)

80 84 86 94 96 100 102 107 109

Chorus 1000 900 800Two-hop HAARP echo ramp 780 666 630 666 700Two-hop HAARP echo pulse 411 500 625 580 400 350

Journal of Geophysical Research: Space Physics 10.1002/2017JA024072

HOSSEINI ET AL. WHISTLER WAVES AND VLF TRIGGERED EMISSIONS 6280

A second scenario we consider is that the sources of the different magnetospheric emissions are not thesame, and our observations are affected by a change in the plasmapause location. Initially, the receiver islocated at a magnetic field footprint inside the plasmapause and observes primarily hiss emissions. As theplasmapause moves in close to the Earth, the receiver is at a footprint outside the plasmapause and observeschorus emissions. HAARP two-hop echoes are observable only when the ducted path moves directly over theheated region. The ability of HARRP to excite two-hop echoes primarily only in overhead ducts has alreadybeen documented [Gołkowski et al., 2009, 2011]. Since the two-hop echoes are excited by a known source,their time of arrival can be used to determine the cold plasma density along the propagation path. Thetwo-hop echoes are seen to have a propagation time of 8.4 s for two magnetospheric traversals at L = 5.Using whistler mode dispersion techniques [Sahzin et al., 1992; Gołkowski et al., 2011], we determine theequatorial electron density to be 300 el/cm�3 at L = 5. Such a value is typical of densities inside the plasma-pause and therefore does not lend additional credence to the scenario of a moving plasmapause location.

Simultaneous Time History of Events and Macroscale Interactions during Substorms (THEMIS) in situobservations give more evidence of the actual location of the plasmapause. Magnetic local time (MLT)locations of the THEMIS A, D, and E spacecraft during 02:00 to 04:00 UT (time of our ground observations)are depicted in Figure 10a. It can be seen that Themis E is about 2 h behind Themis A and D. On theother hand, local time of the ground observation at Alaska is 18:00 to 20:00 (02:00 to 04:00 UT).Therefore, to investigate the location of the plasmapause during the 2 h time period, the plasma densityfor Themis A and D for 02:00 to 04:00 UT and for Themis E for 04:00 to 06:00 UT (2 h later) is shown inFigure 10b. The electron density data show a plasmapause location of L = 6. Equatorial electron density of300 el/cm�3 at L = 5 (from whistler mode dispersion techniques) is depicted by the yellow star and isconsistent with plasma density captured by THEMIS. THEMIS wave data are also analyzed, but no chorusor hiss emissions were observed on any spacecraft.

For completeness we also comment on the geomagnetic conditions at the time of observation and addi-tional possible effects of HAARP ionospheric heating. The magnetospheric conditions on 11 December2008 exhibit a period of quiet (max Kp = 2�) after substorm activity 5 days prior on 6 December (maxKp = 4). Such quieting after a substorm is precisely the type of condition that Gołkowski et al. [2011] identifiedto be favorable for ducting at L = 5 and observation of HAARP-induced triggered emissions. Magnetometerdata from the HAARP facility show peak values below 15 nT consistent with the presence of eletrojet currentsbut not highly disturbed conditions. Ionosonde records taken at the HAARP facility show that during the timeof the observations, the density of the ionosphere was decreasing. The peak frequency of the ionosphere(foF2) was 2.25 MHz at the start of the observation period (2:00 UT) and decreased to 1.3 MHz at 3:30 UT.Vartanyan et al. [2012] present evidence of the HAARP heater being able to create artificial magnetosphericducts. However, such cases involve O-mode heating at an HF frequency close to foF2. This was not the case inour observations. On the day of the observations the HAARP heater and ELF/VLF recording both commencedat 2:00 UT with X-mode transmissions. As can be seen in Figure 2a, magnetospheric hiss emissions were

Figure 10. (a) MLT locations and (b) plasma density versus L shell of the THEMIS A, D, and E satellites.

Journal of Geophysical Research: Space Physics 10.1002/2017JA024072

HOSSEINI ET AL. WHISTLER WAVES AND VLF TRIGGERED EMISSIONS 6281

immediately present at the start of recording. For these reasons as well as the duct footprint dynamicsdescribed above, we do not believe that the HAARP heater was involved in the creation of themagnetospheric duct responsible for the observations.

5. Summary

We have analyzed a unique 2 h observation with concurrent hiss, chorus, VLF triggered emissions, and localHAARP-generated ELF/VLF signals. Using calculated eccentricity of polarization as a novel filter, we show thatthe observations show a common ionospheric exit point for different types of magnetospheric emissions thatcan be tracked in time. We introduced a frequency band-limited autocorrelation method to quantify thecoherency of emissions. A range of coherency was observed from high order of coherency in HAARP trans-missions and their echoes to lower coherency (in natural chorus and hiss emissions). The introduced coher-ency metric is also able to distinguish between hiss emissions and white noise. Two possible scenarios ofmagnetospheric processes were presented to explain the evolution of emissions from hiss to chorus totwo-hop echoes. Time evolution of a magnetosphere source region is most consistent with the observations,including THEMIS in situ electron density measurements.

ReferencesBortnik, J., and R. M. Thorne (2007), The dual role of ELF/VLF chorus waves in the acceleration and precipitation of radiation belt electrons,

J. Atmos. Sol. Terr. Phys., 69(3), 378–386.Bortnik, J., R. M. Thorne, and N. P. Meredith (2008), The unexpected origin of plasmaspheric hiss from discrete chorus emissions, Nature, 452,

62–66, doi:10.1038/nature06741.Bortnik, J., W. Li, R. M. Thorne, V. Angelopoulos, C. Cully, J. Bonnell, O. Le Contel, and A. Roux (2009), An observation linking the origin of

plasmaspheric hiss to discrete chorus emissions, Science, 324(5928), 775–778.Cohen, M. B., M. Gołkowski, and U. S. Inan (2008), Orientation of the HAARP ELF ionospheric dipole and the auroral electrojet, Geophys. Res.

Lett., 35, L02806, doi:10.1029/2007GL032424.Golden D. I., M. Spasojevic, and U. S. Inan (2009), Diurnal dependence of ELF/VLF hiss and its relation to chorus at L = 2.4, J. Geophys. Res., 114,

A05212, doi:10.1029/2008JA013946.Gołkowski, M., and U. S. Inan (2008), Multi-station observations of whistler-mode chorus, J. Geophys. Res., 113, A08210, doi:10.1029/

2007JA012977.Gołkowski, M., U. S. Inan, A. R. Gibby, and M. B. Cohen (2008), Magnetospheric amplification and emission triggering by ELF/VLF waves

injected by the 3.6 MW HAARP ionospheric heater, J. Geophys. Res., 113, A10201, doi:10.1029/2008JA013157.Gołkowski, M., U. S. Inan, and M. B. Cohen. (2009), Cross modulation of whistler mode and HF waves above the HAARP ionospheric heater,

Geophys. Res. Lett., 36, L15103, doi:10.1029/2009GL039669.Gołkowski, M., M. B. Cohen, D. L. Carpenter, and U. S. Inan (2011), On the occurrence of ground observations of ELF/VLF magnetospheric

amplification induced by the HAARP facility, J. Geophys. Res., 116, A04208, doi:10.1029/2010JA016261.Gołkowski, M., U. S. Inan, M. B. Cohen, and A. R. Gibby (2010), Amplitude and phase of nonlinear magnetospheric wave growth excited by the

HAARP HF heater, J. Geophys. Res., 115, A00F04, doi:10.1029/2009JA014610.Helliwell, R. A. (1988), VLF wave simulation experiments in the magnetosphere from Siple Station Antarctica, Rev. Geophys., 26, 551.Hosseini, P., H. Chorsi, M. Golkowski, S. D. Gedney, R. C. Moore (2017), A new approach to locate ionospheric exit points of magnetospheric

whistler mode emissions, USNC/URSI National Radio Science Meeting.Inan, U. S., T. F. Bell, J. Bortnik, and J. M. Albert (2003), Controlled precipitation of radiation belt electrons, J. Geophys. Res., 108(A5), 1186,

doi:10.1029/2002JA009580.Koons, H. C. (1981), The role of hiss in magnetospheric chorus emissions, J. Geophys. Res., 86, 6745–6754.Maxworth, A. S., M. Gołkowski, M. B. Cohen, R. C. Moore, H. T. Chorsi, S. D. Gedney, and R. Jacobs (2015), Multistation observations of the

azimuth, polarization, and frequency dependence of ELF/VLF waves generated by electrojet modulation, Radio Sci., 50, 1008–1026,doi:10.1002/2015RS005683.

Omura, Y., D. Nunn, H. Matsumoto, and M. J. Rycroft (1991), A review of observational, theoretical and numerical studies of VLF triggeredemissions, J. Atmos. Terr. Phys., 53, 351–368.

Omura, Y., Y. Katoh, and D. Summers (2008), Theory and simulation of the generation of whistler-mode chorus, J. Geophys. Res., 113, A04223,doi:10.1029/2007JA012622.

Raghuram, R., T. F. Bell, R. A. Helliwell, and J. P. Katsufrakis (1977), A quiet band produced by VLF transmitter signals in the magnetosphere,Geophys. Res. Lett., 4, 199–202.

Santolık, O., and D. A. Gurnett (2003), Transverse dimensions of chorus in the source region, Geophys. Res. Lett., 30(2), 1031, doi:10.1029/2002GL016178.

Sazhin, S. S., M. Hayakawa, and K. Bullough (1992), Whistler diagnostics of magnetospheric parameters: A review, Ann. Geophys., 10, 293–308.Streltsov, A. V., M. Gołkowski, U. S. Inan, and K. D. Papadopoulos (2010), Propagation of whistler mode waves with a modulated frequency in

the magnetosphere, J. Geophys. Res., 115, A09209, doi:10.1029/2009JA015155.Summers, D., Y. Omura, S. Nakamura, and C. A. Kletzing (2014), Fine structure of plasmaspheric hiss, J. Geophys. Res. Space Physics, 119,

9134–9149, doi:10.1002/2014JA020437.Tsurutani, B. T., B. J. Falkowski, J. S. Pickett, O. Santolik, and G. S. Lakhina (2015), Plasmaspheric hiss properties: Observations from Polar,

J. Geophys. Res. Space Physics, 120, 414–431, doi:10.1002/2014JA020518.Vartanyan, A., G. M. Milikh, E. Mishin, M. Parrot, I. Galkin, B. Reinisch, J. Huba, G. Joyce, and K. Papadopoulos (2012), Artificial ducts caused by

HF heating of the ionosphere by HAARP, J. Geophys. Res., 117, A10307, doi:10.1029/2012JA017563.

Journal of Geophysical Research: Space Physics 10.1002/2017JA024072

HOSSEINI ET AL. WHISTLER WAVES AND VLF TRIGGERED EMISSIONS 6282

AcknowledgmentsWe would like to thank Stanford VLFGroup for access to the VLF data,which are available upon request fromMark Golkowski. We acknowledgeNASA contract NAS5-02099 andV. Angelopoulos for use of data fromthe THEMIS Mission. This work wassupported by the National ScienceFoundation under Award AGS-1254365and Award OPP-1542608 to theUniversity of Colorado Denver.