Simultaneous Multi-angle Observations of StrongLangmuir Turbulence at HAARP

Naomi Watanabe1 • Mark Golkowski1 •

James P. Sheerin2 • Brenton J. Watkins3

Received: 5 September 2014 /Accepted: 27 May 2015� Springer Science+Business Media Dordrecht 2015

Abstract We report results from a recent series of experiments employing the HF

transmitter of the High Frequency Active Auroral Research Program (HAARP) to generate

and study strong Langmuir turbulence (SLT) in the interaction region of overdense

ionospheric plasma. The Modular UHF Ionospheric Radar (MUIR) located at the HAARP

facility is used as the primary diagnostic. Short pulse, low duty cycle experiments are used

to avoid generation of artificial field-aligned irregularities and isolate ponderomotive

plasma turbulence effects. The HF pump frequency is close to the 3rd gyro-harmonic

frequency and the HF pointing angle and MUIR look angle are between the HF Spitze

angle and Magnetic Zenith angle. Plasma line spectra measured simultaneously in different

spots of the interaction region display differences dependent on the aspect angle of the HF

pump beam in the boresight direction and the pointing angle of the MUIR diagnostic radar.

Outshifted Plasma Lines, cascade, collapse, coexistence, spectra are observed in agreement

with existing theory and simulation results of Strong Langmuir Turbulence in ionospheric

interaction experiments. It is found that SLT at HAARP is most readily observed at a HF

pointing angle of 11� and UHF observation angle of 15�, which is consistent with the

magnetic zenith effect as documented in previous works and optimal orientation of the

refracted HF electric field vector.

Keywords Ionospheric heating � HAARP � Langmuir turbulence � Multi angle heating

& Naomi Watanabenaomi.watanabe@ucdenver.edu

1 Department of Electrical Engineering, University of Colorado Denver, Denver, CO, USA

2 Department of Physics and Astronomy, Eastern Michigan University, Ypsilanti, MI, USA

3 Geophysical Institute, University of Alaska Fairbanks, Fairbanks, AK, USA

123

Earth Moon PlanetsDOI 10.1007/s11038-015-9471-0

1 Introduction

Transmission of high power HF (2–10 MHz) waves into the Earth’s ionosphere is known

to increase the electron temperature, and generate a large number of nonlinear phenomena

including plasma turbulence (Bernhardt et al. 1988, Isham et al. 2005, and Milikh et al.

2010). When the frequency of the input HF wave, called the pump wave, is lower than the

local plasma frequency of the ionosphere, x\xp, the wave number k becomes imaginary.

Hence the HF wave decays in the ionosphere (Walker 1979) and couples to high-frequency

electron plasma waves, which are called Langmuir waves (LW), and low frequency ion

waves, which are called ion-acoustic waves (IAW) (Bernhardt et al. 1989; Stubbe et al.

1992; Djuth et al. 1994, and Kosch et al. 2004). This mechanism is called the Parametric

Decay Instability (PDI) (DuBois et al. 2001). A flow diagram of the multi-stage PDI is

shown in Fig. 1. The group velocity of the HF pump wave tends to zero upon approach to

the reflection height and the wave electric field becomes parallel to the geomagnetic field.

Consequently, the HF pump waves generate parametric instabilities just below the re-

flection height. The frequency matching condition of the PDI can be expressed as

x0 � x1 � x2 ð1Þ

where x1 is angular frequency of the IAW and x2 is angular frequency of the LW. The

IAW propagates in both directions along the magnetic field (Forme et al. 1993) as is

indicated by the observation of spectra in which one or both ion lines are enhanced

(Rietveld et al. 1991). Electrons oscillate fast in LWs and the ambipolar electrostatic force

makes both electrons and ions moves in phase with IAWs (Bryers et al. 2013). The

Langmuir waves, produced by PDI, can decay into another LW and another IAW if their

amplitudes exceed a threshold (Rietveld et al. 1993). This second decay process is called

Langmuir Decay, and the excitation of the instability is called Langmuir Decay Instability

(LDI) (DuBois et al. 2001). The upward LW decays into upgoing LW and down going

IAW (Forme 1999). The LW accelerates electrons, and which can produce intense optical

emissions (Ashrafi et al. 2007). The decay of waves continues until the threshold for the

decay is no longer exceeded (Kohl et al. 1993). The wave-matching conditions for LDI is

shown in Eqs. (2) and (3).

Fig. 1 Schematic of parametric decay and 1st and 2nd Langmuir decay. HF pump wave decays mother LWand IAW due to the parametric decay. The mother LW decays 1st daughter LW and another IAW due to theLangmuir decay, and the 1st daughter Langmuir wave decays 2nd LW and another IAW. The decays arecontinues

N. Watanabe et al.

123

x0L ¼ x1

L þ x1ia ð2Þ

k0L ¼ k1L þ k1ia ð3Þ

xL0 and kL0 are angular frequency and wave vector of the ‘‘mother’’ LW due to PDI, x1L

and x1ia are the 1st ‘‘daughter’’ LW and another IAW due to 1st LDI, and k1L and k1ia are

wave vector of the 1st daughter LW and another IAW due to 1st LDI (Forme 1993). The

LDI generates up going and down going LW and IAW (Wahlund et al. 2003). Langmuir

turbulence occurs typically between the matching height and the reflection height (Isham

et al. 2012, Isham et al. 1999). It has been observed that an HF pump frequency that is

slightly below the critical frequency and close to the third gyro-frequency generated can

effectively generate Langmuir Turbulence (Walker 1979).

The angle of the HF pump wave strongly affects the above mentioned processes re-

sulting from heating of the ionosphere (Rietveld et al. 2003). When the HF pump wave

angle is between the HF Spitze angle and the magnetic zenith angle, a strong enhancement

of the parametric decay has been observed (Rietveld et al. 2003). The Spitze angle is

shown in Eq. (4) (Ashrafi et al. 2007) and describes the range of angles for which an

O-mode pump wave can propagate to the highest reflection altitude (Isham et al. 2005). On

the other hand, an O-mode pump wave incident with an angle larger than the Spitze angle

reflects before reaching the highest altitude.

Spitze angle ¼ arcsinY

1þ Yð Þ

� �� �12

sin hB

!ð4Þ

where Y = xb/x0, xb is the electron-cyclotron frequency x0 is frequency of the HF pump

wave, and hB is magnetic zenith angle. Electron temperature strongly increases with the HF

pump wave at or close to the magnetic zenith angle (Rietveld et al. 2003). The features of

electron enhancement are different for different duty cycles of the HF pump wave.

Thermal and density perturbations are moderated with a low duty cycle, since the bulk

heating effect becomes small (Cheung et al. 1997).

If sufficient energy is coupled into the plasma, Langmuir solitons can be formed, which

are localized density depressions and electric field maxima. This process is referred to as

Strong Langmuir Turbulence (SLT). Observationally, incoherent radar observations of

spectrum features known as cascade, collapse, coexistence of cascade and collapse, and

outshifted plasma line (OPL) predict the existence of SLT. Cascade refers to the spectral

feature of the PDI and LDI (Depierreux et al. 2002) lines. The cascade is observed at or

below the PDI matching height (DuBois et al. 2001). The first cascade line shows the

‘‘mother’’ LW due to PDI, and the other cascade lines show ‘‘daughter’’ LWs due to LDI

(Fejer 1979). The cascade lines in a spectral sideband may be only odd multiples of the LW

frequencies. Cascade lines of even multiples LWs are not visible in the spectrum band-

width of the receiver channel since the even multiples of LWs have negative wavenumbers

(Kohl et al. 1993). The relationship between the frequency of cascade lines and the HF

pump frequency is governed by fn - fn?1 = 2(fHF - f1), where fn is frequency of the n-th

cascade line, fn?1 is frequency of the (n ? 1)-th cascade line, fHF is HF pump frequency,

and f1 is frequency of the 1st cascade line. Since the difference between HF pump fre-

quency and the frequency of the 1st cascade line, fHF - f1 is equal to the frequency of the

1st IAW, the frequency difference between the two cascade lines is the same as twice the

1st IAW (Kohl and Rietveld 1996). The process of cascading continues as long as sub-

sequent daughter waves exceed the threshold. The electron and ion temperature affect the

Simultaneous Multi-angle Observations of Strong…

123

spectral width of the cascade, and the total spectral power is proportional to the electron

density (Forme and Fontaine 1999).

Collapse is the spectral feature corresponding to the collapse of the Langmuir solitons

which occurs near the reflection height of the HF pump wave (DuBois et al. 1995). The

collapsing solitons are in effect cavitons meaning density cavities, which have the size of

approximately tens of meters (Wong et al. 1987) and are created by the ponderomotive

pressure (intensity) of electromagnetic waves (Fejer 1983). Strong Langmuir Turbulence

(SLT) is considered to be driven by collapsing cavitons (Sheerin et al. 1982). The density

cavity is deepened by the ponderomotive pressure, and which intensifies the localized field.

Nonlinear steepening of the trapped field leads to collapse (DuBois et al. 2001). Caviton

collapse dissipates most of the electrostatic energy through interaction with electrons

(DuBois et al. 1991). Moreover, DuBois et al. (1991) describe that the process of collapse

is apparent for approximately a microsecond after switch-off of the pump wave. Caviton

collapse has been observed during HF heating (Duncan and Sheerin 1985; Djuth et al.

1994) with cavitons formed near the reflection height of the HF pump wave (Sheerin et al.

1982; Duncan and Sheerin 1985). The coexistence of the cascade and collapse appears

around the PDI matching height (Hanssen et al. 1992; DuBois et al. 1991, 1993).

The outshifted plasma lines (OPL) in the spectrum refers to LW’s which appear at

frequency shifts above the HF pump wave frequency shift (Isham et al. 1996). OPL

signifies the presence of LWs with large frequencies that are generated by the accelerated

electrons which in turn is related to collapse. Langmuir turbulence is observed at the UHF

matching height where the wave vector of the HF pump wave is twice as much as the wave

vector of the UHF wave (Rietveld et al. 2000) which is called the Bragg scatter condition.

In this paper we describe results of an experiment wherein we investigate the generation of

Strong Langmuir Turbulence (SLT) as function of aspect angle.

2 Experiment

The experiment to investigate excitation of Strong Langmuir Turbulence (SLT) and plasma

instabilities as a function of aspect angle was carried out at the High Frequency Active

Auroral research program (HAARP) facility in Gakona Alaska (latitude 62.39� N and

longitude 145.15� W), over the course of several days from July 19 to July 26 in 2011. The

HF pump wave of the 3.6 MW (ERP 88.8–91 dBW) ionospheric heater was transmitted

into the F2 region of the ionosphere with O-mode polarization and frequency chosen just to

be below the f0F2-critical frequency based on ionosonde observations. The HF pointing

angles were 7�, 11�, and 14� geographically southward with respect to vertical as shown in

Fig. 2. The magnetic declination at the HAARP location is approximately 15�. The HF

transmission format consisted of a 60 ms pulse and a 12 s IPP giving a duty cycle of

0.5 %. The format was repeated for a 5 min sequence (4 min pulsing and 1 min off). The

very low duty cycle is used to avoid the creation of artificial field aligned irregularities that

occurs under prolonged bulk heating of the plasma, thus observations are dominated by

more rapid SLT processes. The Modular UHF Ionospheric Radar (MUIR) located at

HAARP facility was used for observations of parametric decay of the HF pump wave. The

frequency of the transmission of MUIR is 446 MHz. MUIR sends a radar pulse and

receives incoherent backscatter to detect plasma and ion lines which correspond to LW and

IAW, respectively. The pointing angle of the UHF MUIR signal was also varied at 6�, 12�,and 15�. Since the UHF IPP is only 3 ms, and the MUIR radar can change pointing pulse-

N. Watanabe et al.

123

to-pulse, the MUIR radar is able to diagnose each 60 ms HF pulse almost simultaneously at

the three different UHF angles. The time delay between the UHF transmission and re-

ceiving the backscatter echo provides the reflection height. The frequency shift of the

plasma lines detected by MUIR is close to the HF pump frequency (Sheerin et al. 2003).

The backscatter spectrum shows cascade lines that are observed LWs. The spectrum shows

only upshifted plasma lines that are cascade lines of odd multiples of LWs as discussed

above (Kohl et al. 1993). A schematic of the HF transmission and MUIR diagnostic is

shown in Fig. 2.

3 Observations

The strongest observed signatures of SLT were achieved during the transmission on 25

July 2011 from 03:05:00 UT to 03:44:48 UT. The HF frequency was 4.2 MHz. From

ionograms taken at the time, the critical frequency f0F2 was observed to be

5.125–5.262 MHz at an altitude of *250 km. The altitude where the pump frequency

matched the plasma frequency was observed to be 180–200 km. The HF pointing angles

were 7�, 11, and 14�, UHF angles were 6�, 12�, and 15�. The magnetic zenith angle was

14.5�. Figure 3 shows the spectra of the MUIR UHF echoes for some of the observations.

Figure 4 shows a detailed look at the PSD of MUIR observations at 03:28:00 UT when the

HF pointing angle is 11� and UHF angle is 15�. The features of cascade, collapse, coex-

istence of cascade and collapse, and OPL are identified in the spectrum.

Table 1 summarizes the observations of cascade, collapse, coexistence, and OPL for the

different combinations of HF and UHF angles. Most of the strong echoes were observed

when the UHF angle was 12� and 15� with less sensitivity to the HF pointing angle.

Although the UHF observations involved near simultaneous observation at all three UHF

angles, the HF angles were changed every 5 min beginning with 7� from 03:05:00 UT to

03:8:48 UT, 11� from 03:10:00 UT to 03:13:48 UT, 14� from 03:15:00 UT to 03:18:48 UT,

etc. All spectral features were most strongly observed during 03:15:00 UT to 03:28:48 UT.

Figure 5 shows ionograms taken at 03:10 UT, 03:25 UT and 03:40 UT. The ionogram

at -03:25 UT (the time of strongest SLT observations) is distinguished by a slightly lower

f0F2 critical frequency and slightly lower absorption.

Fig. 2 Schematic of HFtransmission directed southwardwith HF pointing angle of7�–11�–14� and MUIR (ModularUHF Ionospheric Radar) pointingangle of 6�–12�–15� fromvertical in Gakona, Alaska. Ascan be seen, the HF beam isapproximately three times widerthan the UHF beam

Simultaneous Multi-angle Observations of Strong…

123

As mentioned above, the strongest signatures of SLT were observed on 25 July 2011

from 03:05:00 UT through 03:44:48 UT. Here we briefly discusses results on other days

and times of the experiment. Earlier on July 25, 2011 from 01:15:00 UT through 02:29:48

UT the HF pump frequency was 4.8 MHz and a slightly different set of angles for HF and

UHF were used: the HF beam was vertical (0� HF) and UHF angles of 3� and 22� wereused. Most of the observations did not reveal detections for a 0� HF pointing angle and 22�and 3� UHF angles. The experimental time on July 24 2011 was from 01:45:00 UT through

02:34:48 UT. The HF frequency was 4.2 MHz, below the critical frequency f0F2 and

around the third electron gyro-harmonic frequency. Weak coexistence and OPL were

observed when the HF angle was 7� and 14� but only at UHF angle 15�. No detections

were observed at HF of 0� for any UHF angle. On July 24 2011, from 03:10:00 UT through

03:59:48 UT, the HF frequency was 4.2 MHz. The HF pointing angle was 0� while UHF

was pointing at 0�, 6�, and 15�. There were no signals detected for these parameters. From

01:35:00 UT through 02:24:48 UT on July 26, there were no detections with HF pump

frequency of 4.8 MHz with 7�, 11�, and 14� of HF pointing angle and 15� UHF angle.

Fig. 3 Power spectrum density (PSD) of the signal with HF pointing angle of 11�and a UHF of 15�, b UHFof 12�, and c UHF of 6� from 03:27:58 UT to 03:28:58 UT. Detections at 03:28:00 UT, 03:28:12 UT,03:28:24 UT, 03:28:36 UT, and 03:28:48 UT. Very strong cascades were observed at UHF of 15�. Strongcascades were observed at UHF of 12�; however, intensity and width of spectrum of cascade were strongerat UHF of 15� than 12�. Most of cascades were missing at UHF of 6�

N. Watanabe et al.

123

Since the most successful observations of SLT phenomena were observed during 03:05-

03:48 UT on July 25, 2011, we further analyze the strongest detections made during this

time interval. Figure 6 shows the spectrum intensity averaged over 20 pulses that were

transmitted from 03:25:00 UT to 03:28:48 UT for HF pointing angle of 11� and UHF

angles of 12� and 15�. Echoes from UHF angle of 6� are not shown since the they were

very weak. The overall spectrum for UHF 15� is seen to be 5-10 dB higher than for UHF

12�. The frequency and intensity of 1st to 5th cascades of the spectra in Fig. 5 are tabulatedon Table 2. Frequency difference between cascade lines are shown on Table 3.

The 1st cascade line shows the 1st LW due to the PDI and is the strongest feature in all

of the observed spectra with remaining features being proportionally lower. In Fig. 7 the

level of the 1st cascade line is shown for all combinations of HF and UHF angles, which

we use as a proxy of SLT observation. As Fig. 7 demonstrates, SLT is most favorably

observed when the HF pointing angle is 11� and UHF angle is 15�. Furthermore, it shows

that SLT is readily observed at all HF pointing angles, 7�, 11�, and 14� when the UHF

angle is 15�.

Fig. 4 a Power spectral density (PSD) of echo of the instabilities that MUIR detected. HF of 11� and UHFof 15� at 03:28:00 UT on July 25, 2011. b 2-D Intensity of the PSD of shown in Fig. 3a showing spectralfeatures associated with SLT

Table 1 The observation of cascade, collapse, coex, and OPL at HF at 7�, 11�, and 14� and UHF 6�, 12�,15� from 03:05:00UT to 03:43:48 UT on July 25, 2011

HFangle(�)

UHF angle (�)

6 12 15

7 Weak cascade and collapse, nodetection of coex, weakOPL or no detection of OPL

Cascade and collapse, weak coex,mostly weak OPL, isolated cases ofstrong OPL

11 Weak cascade and collapse,weak or no coex, OPLsometimes

Strong cascade andcollapse, strong coex,strong OPL

Very strong cascade and collapse,strong coex, strong OPL

14 Weak cascade and collapse,weak or no coex, OPLsometimes

Weak cascade andcollapse weak coex,weak OPL

Very strong cascade and collapse,strong coex, strong OPL

Simultaneous Multi-angle Observations of Strong…

123

4 Discussion

As Table 3 shows, the frequency difference between 1st and 2nd cascade lines to 4th and

5th cascade lines are almost the same as twice as the frequency difference between HF

pump frequency and 1st cascade line for both 11� of HF and 15� of UHF and 11� of HF and

12� of UHF. This confirms that frequency difference between n-th and (n ? 1)-th cascade

lines is equal to twice as the frequency difference between HF pump frequency and 1st

cascade line. The weaker dependence of SLT detections on HF pointing angle as opposed

to UHF angle is likely due to the fact that the HF beam is much broader than the UHF

beam. The HAARP beamwidth (FWHP) during the experiments, was 12.4� for North and

South pattern, and 9.9� for East and West pattern. Nevertheless, it is remarkable that we

detected SLT more strongly when the HF pointing angle was 11� than 14�. In fact the

strongest SLT signatures were observed for HF 11� and UHF 15�. Rietveld et al. (2003)

explain that when the electric field of the reflected HF pump wave is parallel to the UHF

Fig. 5 Ionosonde data for 25 July 2011. The ionogram at a 03:10:00 UT, b 03:20:00 UT, c 03:25:00 UT. Thepurple line shows the altitudewhere plasma frequency is equal to transmitted HF frequency and the orange lineshows the height at the critical frequency f0F2. The absorption level was slightly lower at the middle sequence(b) and this was the time when the strongest echoes and spectral features were observed. (Color figure online)

N. Watanabe et al.

123

beam at the UHF matching height where the Bragg scatter condition is satisfied, the radar

detects the strongest echo. Figure 8 is a schematic diagram that demonstrates refraction

and reflection of the HF pump wave and the geometry of its associated electric field vector.

The HF pointing angles are 7�, 11�, and 14�. As Fig. 8 shows for a 11� HF pointing angle,

the electric field of the reflected wave is almost parallel to the geo-magnetic field at the

UHF matching height. In our experiment, the UHF angle for which there were the strongest

detections was 15�. The magnetic zenith angle at HAARP on Jul 25 2011 was 14.5�. Thusthe magnetic zenith and UHF angle of 15� are almost the same. The preferable detection of

SLT for HF 11� may be due to the optimal electric field orientation of the reflected pump

wave begin parallel to the geo-magnetic field line at the UHF matching height. This

4.05 4.1 4.15 4.2 4.25 4.3 4.35-40

-35

-30

-25

-20

-15

-10

-5

Frequency [MHz]

Inte

nsity

[dB]

Average Intensity of 20 pulses 03:25:00 UT to 03:28:48 UT

HF11&UHF15HF11&UHF12

Fig. 6 Average intensity of 20 pulses of the HF transmission. The experimental time is from 3:25:00 UT to3:28:48 UT. The HF pointing angle is 11� and UHF angles are 15� and 12�. The plot of the signal with HF11� and UHF 6� is not shown because the signals were very weak

Table 2 Frequency and intensity of the 1st to 5th cascades from 03:25:00 UT to 03:28:48 UT for HFpointing angle of 11� and UHF angles of 12� and 15

Cascade’s number 11� of HF and 15� of UHF 11� of HF and 12� of UHF

Frequency(MHz)

Intensity(dB)

Frequency(MHz)

Intensity(dB)

1st 4.196 7.409 4.195 18.77

2nd 4.187 8.902 4.187 20.36

3rd 4.179 10.74 4.18 22.89

4th 4.171 16.04 4.171 27.66

5th 4.163 19.97 4.162 28.8

The frequency and intensity values are based on Fig. 5

Simultaneous Multi-angle Observations of Strong…

123

suggests an additional nuance to the well known magnetic zenith effect documented in

previous works.

5 Summary

The simultaneous detection of cascade, collapse, and OPL has been theoretically predicted

to occur during SLT. Our observation of these spectral features indicate that SLT was

excited by HF transmissions at HAARP when the HF pump frequency was 4.2 MHz, close

to the third gyro-harmonic frequency and below the critical frequency. The HF trans-

mission was kept at a low duty cycle (0.5 %), avoiding artificial field-aligned irregularities

in the ionosphere, permitting the interpretation of the spectra in the context of SLT pro-

cesses. The high time resolution of the UHF radar, 3.3 ms, allowed observation of SLT

simultaneously at different angles. There were high contrast detections of cascade, col-

lapse, coexistence, and OPL at UHF of 15� with HF pointing angles of 7�, 11�, and 14�.

Table 3 Frequency difference between n-th and (n ? 1)-th cascade lines for 11� of HF and 15� of UHF and11� of HF and 12� of UHF

Cascade’s number 11� of HF and 15� of UHF 11� of HF and 12� of UHFFrequency difference (kHz) Frequency difference (kHz)

HF pump—1st 4 5

1st–2nd 9 8

2nd–3rd 8 7

3rd–4th 8 9

4th–5th 8 9

HF pump of the table is the HF pump frequency 4.2 MHz. Each frequency difference is calculated using thefrequencies of Table 2

Fig. 7 Intensity level detection of signalswith 7�, 11�, and 14� ofHF pointing angle and 6�, 12�, and 15� ofUHFangle. The color shows intensity level with dB. Lighter color indicates stronger intensity. (Color figure online)

N. Watanabe et al.

123

This confirms that there were strong spectra of SLT with the HF pointing angle between

the Spitze angle and the magnetic zenith angles (Rietveld et al. 2003), and it is also

consistent with the magnetic zenith effect, a strong echo when the UHF radar angle is close

to the magnetic zenith angle as described by Rietveld et al. (2003). There were detections

of strong echoes when the HF pointing angle was 11� and 14� as compared to 7� for allcascade, collapse, coexistence, and OPL. Optimal direction of the refracted HF electric

field vector explains the strong observations for HF 11� and UHF 15�. OPLs were detectedmore readily than in these experimental runs than in previous experiments at HAARP

(Adham et al. 2010) perhaps due to more favorable ionospheric conditions. These obser-

vations confirm the existing theory and simulation results of SLT for RF ionospheric

interaction experiments.

Acknowledgments The authors are grateful to HAARP facility operators and Ed Kennedy for organizingthe HAARP Student Research Campaign in 2011.

References

N. Adham, J. P. Sheerin, M. R. Wood, R. G. Roe, J. M. Gerres, B. J. Watkins, W. A. Bristow, P.A. Bernhardt, C. A. Selcher AGU, 91 (52), Fall Meet. Suppl., Abs SA33A-1753 (2010)

M. Ashrafi, M.J. Kosch, K. Kaila, B. Isham, J. Geophys. Res. 112(A5), A05314 (2007)P.A. Bernhardt, L.M. Duncan, C.A. Tepley, R.A. Behnke, J.P. Sheerin, Adv. Space Res. 8(1), 271–277

(1988)P.A. Bernhardt, C.A. Tepley, L.M. Duncan, J. Geophys. Res. 94, 9071–9092 (1989)C.J. Bryers, M.J. Kosch, A. Senior, M.T. Rietveld, T.K. Yeoman, J. Geophys. Res. 118, 7472–7481 (2013).

doi:10.1002/2013JA019429P. Cheung, E. Mjølhus, D. DuBois, J. Pau, H. Zwi, A. Wong, Phys. Rev. Lett. 79(7), 1273–1276 (1997)S. Depierreux, C. Labaune, J. Fuchs, D. Pesme, V.T. Tikhonchuk, H.A. Baldis, Phys. Res. Lett. 89(4),

045001-1–045001-4 (2002)F.T. Djuth, P. Stubbe, M.P. Sulzer, H. Kohl, M.T. Rietveld, J.H. Elder, J. Geophys. Res. 99(A1), 333–339

(1994)D.F. DuBois, H.A. Rose, D. Russell, Phys. Rev. Lett. 66(15), 1970–1973 (1991)

Fig. 8 Schematic diagram of the HF pump wave with 7�, 11�, and 14� of the HF pointing angle. The HFpump wave was transmitted with 4.2 MHz at HAARP in Gakona, Alaska (latitude 62� N). The reflectionheight and UHF matching height are shown as dashed dotted lines. The green dashed dotted line shows thegeo-magnetic field. B0, and the blue line shows the electric field, E, of reflection of the HF pump wave. Thisfigure is an adaptation of the diagram of HF ray paths from Rietveld et al. (2003). (Color figure online)

Simultaneous Multi-angle Observations of Strong…

123

D.F. DuBois, A. Hansen, H.A. Rose, D. Russell, Phys. Fluids 5(7), 2616–2622 (1993)D.F. DuBois, D. Russell, H.A. Rose, Phys. Plasmas 2(76), 10.1063/1.871119 (1995)D.F. DuBois, D.A. Russell, P.Y. Cheung, M.P. Sulzer, Phys. Plasmas 8(3), 791–801 (2001)L.M. Duncan, J.P. Sheerin, J. Geophys. Res. 90(A9), 8371–8376 (1985)J.A. Fejer, Rev. Geophys. 17(1), 135 (1979)J.A. Fejer, J. Geophys. Res. 88(A1), 0148–0227 (1983)F.R.E. Forme, Geophys. Res. Lett. 20(21), 2347–2350 (1993)F.R.E. Forme, Ann. Geophys. 17(1172), 1181 (1999)F.R.E. Forme, D. Fontaine, Ann. Geophysicae 17, 182–189 (1999)A. Hanssen, E. Mjolhus, D.F. DuBois, H.A. Rose, J. Geophys. Res. 97, 12073–12091 (1992)B. Isham, C. La Hoz, H. Kohl, T. Hagfors, T.B. Leyser, M.T. Rietveld, J. Atmos. Terr. Phys. 58, 369–384

(1996)B. Isham, M.T. Rietveld, T. Hagfors, C. La Hoz, E. Mishin, W. Kofman, T.B. Leyser, A.P. van Eyken, Adv.

Space Res. 24(8), 1003–1006 (1999)B. Isham, T. Hagfors, B. Khudukon, R.Yu. Yurik, E.D. Tereshchenko, M.T. Rietveld, V. Belyey, M. Grill,

C. La Hoz, A. Brekke, C. Heinselman, Ann. Geophys. 23, 55–74 (2005)B. Isham, M. T. Rietveld, P. Guio., F. R. E. Forme, T. Grydeland, E. Mjølhu, Phys, Res. Lett., 108 (2012)H. Kohl, H. Kopka, P. Stubbe, M.T. Rietveld, J Atmos. Terr. Phys. 55(4/5), 601–613 (1993)H. Kohl, M.T. Rietveld, J. Goephys. Res. 101(A32), 5391–5395 (1996)M. J. Kosch, M. T. Rietveld, A. Senior, I. W. McCrea, A. J. Kavanagh, B. Isham, F. Honary, Geophys. Res.

Lett., 31 (2004)G.M. Milikh, A.G. Demekhov, K. Papadopoulos, A. Vartanyan, J.D. Huba, G. Joyce, Geophys. Res. Lett.

37, L07803 (2010). doi:10.1029/2010GL042684M. T. Rietveld, P. N. Collis, J. -P. St.-Maurice, J. Geophys. Res., 96(A11) 19, 291–19, 305 (1991)M.T. Rietveld, H. Kohl, H. Kopka, P. Stubbe, J. Atmos. Terr. Physis. 55, 577–599 (1993)M.T. Rietveld, B. Isham, H. Kohl, C. La Hoz, T. Hagfors, J. Geophys. Res. 105(A4), 7429 (2000)M. T. Rietveld, J. Geophys. Res., 108 (A4) (2003)M.T. Rietveld, M.J. Kosch, N.F. Blagoveshchenskaya, V.A. Kornienko, T.B. Leyser, T.K. Yeoman, J.

Geophys. Res. 108(A4), 1141 (2003)J.P. Sheerin, J.C. Weatherall, D.R. Nicholson, G.L. Payne, M.V. Goldman, P.J. Hansen, Solitons and

ionospheric modification. J. Atmos. Terr. Phys. 44(12), 1043–1048 (1982)J. P. Sheerin, J. M. Parzych, J. P. Mills, W. A. Bristow, K. M. Groves, American Geophysical Union, Fall

Meeting #SA12B-1092 (2003)P. Stubbe, H. Kohl, M.T. Rietveld, J. Geophys. Res. 97(A5), 6285–6297 (1992)J.E. Wahlund, A. Yilmaz, M. Backrud, D. Sundkvist, A. Vaivads, D. Winningham, M. Andre, A. Balogh, J.

Bonnell, S. Buchert, T. Carozzi, N. Cornilleau, M. Dunlop, A.I. Eriksson, A. Fazakerley, G.Gustafsson, M. Parrot, P. Robert, A. Tjulin, Geophys. Res. 30(11), 1563 (2003). doi:10.1029/2002GL016335

J.C.G. Walker, Rev. Geophys. 17(4), 534 (1979)A.Y. Wong, T. Tanikawa, A. Kuthi, Phys. Rev. Lett. 58(13), 1375–1378 (1987)

N. Watanabe et al.

123