Propagation Characteristics of EMIC Waves in the High Latitude Ionospheric Waveguide Hyomin Kim1 (hyomin.kim@vt.edu), Marc Lessard2 (marc.lessard@unh.edu), Mark Engebretson3, Matthew Young2, Lasse Clausen1, Howard Singer4 and Brian Fraser5 1Center for Space Science and Engineering Research, Virginia Tech, Blacksburg, VA. USA. 2Space Science Center, University of New Hampshire, Durham, NH. USA. 3Physics Department, Augsburg College, Minneapolis, MN. USA. 4Space Weather Prediction Center, NOAA, Boulder, CO. USA. 5Center for Space Physics, University of New Castle, Callaghan, New South Wales, Australia Abstract Example Event: 3/5/2007 – Ground Observation Example Event: 10/13/2007 – Ground-GOES Conjunction EMIC wave data were acquired using a ground array of search-coil magnetometers predominantly in the pre-noon to post-noon sector of Antarctica. With extensive coverage from geomagnetic latitudes of 62 to 87 degrees ILAT (a distance within the ionosphere of 2920 km) and good alignment along the magnetic meridian, search-coil magnetometers operating aboard Automated Geophysical Observatories (AGOs) and manned ground stations show clear ducting of these waves with unexpectedly low attenuation in the waveguide. Halley Station, located at the lowest latitude, typically observes the highest wave power and well-defined band-limited signatures with the same wave event, showing less wave power, found at the other four stations at higher latitudes. This is a clear indication of the ducting effect (within a region centered around the electron density maximum near the F2 ionization peak). Since the events were observed over a very wide range in latitude, the observation of propagation within the ionosphere provides fundamental information regarding ducting efficiencies. This study presents the observations of over 100 Pc1 events showing spectral power attenuation in the waveguide. In addition, changes of the polarization properties such as ellipticity and polarization ellipse major axis direction during the propagation are also discussed. Finally, ionosphere effects (i.e., conductivity) are presented. (a) (b) (c) Introduction Figure 6 Spectrograms of the N-components (eastward) of (a) GOES -10 and (b) GOES-12. EMIC waves in the frequency range 0.1-0.35 Hz are measured from GOES-10 and GOES-12, simultaneously. (c) Stacked spectrograms in the Y (east-west) components of the search-coil data showing the ULF Pc 1-2 waves recorded at the Antarctic stations, HBA, P2, SPA, and P1. Data from P5 were unavailable. Wave activity with poleward attenuation over the network is apparent. Figure 4 Polarization ellipticity (ε) of the ULF Pc 1 waves observed by the Antarctic search-coil array. Each panel is represented in a color scale with -1 being LH circular polarization and +1 being RH circular polarization. Linear polarization (LP) is defined as having |ε| < 0.2. The polarization sense change appears during the poleward propagation. For example, the LHP mode event at HBA changes to an event in either LP or RHP mode at the higher latitudes while the RHP mode event at HBA changes to the LP or LHP mode events. Figure 5 Polarization angle (θ) of the ULF Pc 1 waves observed by the Antarctic search-coil array. The angle ranges between -90° and +90°. The sign represents the direction of angle with respect to the magnetic meridian in north-south direction (X component) with positive angle being counterclockwise and negative angle being clockwise. The panels on the left and right present positive (θ ≥ 0) and negative (θ < 0) angles, respectively. The temporal distribution of the positive angles at HBA shows a similar pattern to that of the ellipticities at HBA, which might indicate that the wave events in each polarization mode are propagated from each different localized injection region. The polarization ellipse observed on the ground becomes increasingly LP with the major axis pointing toward (or away from) the wave injection region as Pc 1 emissions propagate away from the injection region in the ionospheric waveguide (e.g., Greifinger [1972], Summers and Fraser [1972], Fujita and Tamao [1988]). Figure 3 Power spectra of the ULF Pc 1 events observed by the Antarctic search-coil array and the wave power attenuation (in dB) over the distance from HBA (in km) at four selected frequencies (0.55, 0.62, 0.69, and 0.76 Hz) during the two time periods, (a) 0220-0240 UT and (b) 0240-0300 UT on Mar. 5, 2007. The graphs for each time period display the results from both X and Y components. The attenuation factors appear to be ~16 to 24 dB/1000 km in the range of HBA-P2; ~2 to 13 dB/1000 km in the range of P2-SPA-P1-P5, depending on the wave frequency. Figure 2 Stacked 0-1 Hz Fourier spectrograms in the Y (east-west) components of the search-coil data showing the ULF Pc 1 waves recorded at the Antarctic stations, HBA, P2, SPA, P1 and P5 from 0200 to 0600 UT on Mar. 5, 2007. Wave activity with poleward attenuation over the network is apparent. Figure 1 Propagation of EMIC waves from the equatorial region in the magnetosphere to the ionospheric waveguide. ULF waves in the Pc 1-2 range (0.1 – 5.0 Hz) are typically associated with electromagnetic ion cyclotron (EMIC) waves, which are excited by the cyclotron instability of hot, anisotropic (T┴>T||) distributions of medium energy plasma sheet and ring current ions in the magnetospheric equator. Ionospheric waveguide is centered around the electron density maximum near the F2 region. LHP Alfvénic ULF pulsations in the Pc 1-2 range propagate along the field lines and couple to RHP compressional (fast) isotropic waves in the ionospheric waveguide. The walls of the waveguide are not perfect reflectors: wave power attenuation occurs during propagation. Meridional propagation is most efficient. Statistical Results – Ground Observation Statistical Results – Ground-GOES Conjunction Summary 138 ULF Pc 1-2 wave events detected by Antarctic search-coil magnetometer array in 2007 show poleward propagation in the ionospheric waveguide with very low attenuation (~11 dB/1000 km). The efficient wave propagation (compared to other previous studies) is perhaps due to the ground array that is positioned along the magnetic meridian (-62º to -87º MLAT, spanning ~3000 km), ideal for measuring such wave ducting events since ducting can occur most efficiently along the magnetic meridian. The attenuation and the frequency cutoff of the wave propagating in the ionospheric waveguide appears to be dependent on the ionospheric conductivity. In addition, the wave power attenuation increases with increasing wave frequency. Both LHP and RHP modes are dominant over LP mode at lower latitudes. The occurrences of both LHP and RHP modes decrease during propagation whereas the LP mode occurrences increase. 63 EMIC wave events have been identified from GOES-10 and/or GOES-12 and 65% of the events are observed in conjunction with the ground network. 93% of the conjunction events show propagation in the ionospheric waveguide in the Pc 1-2 range. Kp index and ionospheric conductivity might contribute to EMIC wave propagation from the magnetosphere to the ionosphere. 1. MLT Occurrences 2. Spectral Power Attenuation and Ionospheric Conductivity A total of 63 EMIC wave events have been identified from GOES-10 and/or GOES-12. 41 out of 63 events are observed in conjunction with the Antarctic ground magnetometer array. 93% of the conjunction events show that the ULF Pc 1-2 waves propagate horizontally in the ionospheric waveguide. 2.1. Spectral Power Attenuation A seasonal effect in attenuation is found in a different manner between near the injection region (HBA-P2) and the surrounding region (P2-SPA-P1-P5); higher attenuation during dark times in the injection region; higher attenuation during sunlit times in the surrounding region. This might indicate different wave propagation mechanisms are present in the two regions. That is, in the injection region, the wave attenuation in the ionospheric waveguide is governed by the Hall current induced by the incident Alfvén wave [Fujita and Tamao, 1988] as increased Hall current due to increased conductivity (equivalently, lower electric field intensity) under sunlit conditions results in lower attenuation. On the other hand, in the surrounding region, collisional processes in the ionospheric plasma control wave propagation in the waveguide as suggested by model predictions [e.g., Greifinger and Greifinger, 1968; Lysak, 2004] in which higher conductivity leads to higher attenuation due to either higher collision plasma frequency or lower induced magnetic field. 1. MLT Distribution of EMIC Waves Figure 11 MLT distribution of EMIC waves observed only from the GOES satellites (“GOES only”) and observed simultaneously from the satellites and on the ground (“Conjunction”). Data Set The search-coil magnetometer array in Antarctica covers an unprecedented latitudinal extent (-62º to -87º MLAT, ~2920 km) along a magnetic meridian. Antarctica is the only place where such extensive spatial latitudinal configuration is possible providing good meridional alignment, distinguishable ionospheric sunlit condition, and minimal geoelectric inhomogeneity. Range HBA-P2 P2-SPA-P1-P5 HBA-P2-SPA-P1-P5 dB/1000 km σ Sunlit 7.8 11.2 11.1 4.9 9.9 6.6 Mixed 11.4 9.3 11.9 6.2 12.3 4.6 Dark 17.2 10.1 8.4 5.1 13.8 7.1 Total 9.2 10.8 5.3 5.9 Figure 7 MLT distribution of ULF Pc 1-2 ducting events in the year 2007, indicating the Pc 1-2 ducting occurred at all local times, but was more common during daytime hours and showed peaks near 0700 MLT and 1400 MLT. 2.3. Ionospheric Conductivity vs Wave Frequency 2. EMIC Wave Occurrences and Kp index Figure 5 Kp indices during the observations of EMIC waves measured only from the GOES satellites (“GOES only”) and observed simultaneously from the satellites and on the ground (“Conjunction”), suggesting that the waves are less likely to propagate into the ionosphere with higher Kp. Note that most of the events occurred during quiet times (Kp < 4); only 4 out of the entire events occurred under moderate condition (Kp = 4) and no event occurred when Kp > 4. 3. Polarization Characteristics 2.2. Attenuation vs Wave Frequency Antarctic Search-Coil Magnetometer Array Station Geographic CGM MLT MN L* Distance from HBA (km) Lat. Long. Halley Bay (HBA) -75.5 333.4 -61.6 29.0 2:43 4.4 AGO P2 -85.7 313.6 -69.8 19.3 3:29 8.4 1170 South Pole (SPA) -90.0 ---- -74.0 18.4 3:35 13.2 1610 AGO P1 -83.9 129.6 -80.1 16.9 3:44 34.1 2252 AGO P5 -77.2 123.5 -86.7 29.5 2:52 2920 Figure 8 Spectral power attenuation versus Pc 1-2 wave frequency under three ionospheric sunlight conditions, indicating wave power attenuation appears to increase with increasing frequency. Attenuation in any band generally increases with frequency, being smallest at the cutoff frequency since signals are attenuated due to collisional process (signals with higher frequencies undergo more collisions) [Greifinger, 1968]. Attenuation decreases at the lower frequency because the ducted wave with a lower frequency has a longer vertical wavelength [Fujita, 1968]. Figure 9 Frequency distribution of the ULF Pc 1-2 waves under three ionospheric sunlight conditions (note that the occurrences in each sunlight condition are normalized). This result indicates ionospheric conductivity might be one of the controlling factors for wave propagation in the ionosphere, showing that the frequency of the waves decreases with the increasing ionospheric conductivity. At the lower frequency cutoff, the wavelength is longer than the physical extent of the ionosphere in altitude so that the ionospheric waveguide has little effect on the wave since most of the wave (in wavelength) is not confined in the waveguide. At higher frequencies, on the contrary, the wavelength becomes small enough to be confined in the waveguide and thus the wave propagation becomes more sensitive to the physical size of the waveguide, which is affected by the ionospheric sunlight condition. *Geomagnetic coordinates, dipole L-values, and MLT MN in UT are obtained from NASA GSFC Modelweb Website, http://modelweb.gsfc.nasa.gov/models/cgm/cgm.html, for epoch 2007, assuming an altitude of 100 km. 3. EMIC Wave Occurrences and Ionospheric Conditions Figure 10 Ellipticity occurrence percentiles over distance of the Antarctic array, showing that both LHP and RHP modes are dominant at lower latitudes. Both LHP and RHP modes decrease and the LP mode increases at the higher latitudes. Note that the events under dark ionospheric condition are not included due to insufficient number of events. Mode conversion from the incident LHP Alfvén waves to the ducted compressional mode waves occurs within the extent of the injection region and in the surrounding region, signals found on the ground should be identified as ducted waves in LP mode. Figure 6 EMIC occurrences under two ionospheric conditions (“Sunlit” and “Mixed”. See Data Set Section for description). The occurrences of the EMIC waves detected only from the GOES satellites (“GOES only”) are much less than the satellite-ground conjunction events under mixed conditions, which might suggest that ionospheric conductivity plays a role in wave injection from the magnetosphere to the ionosphere. Note that no event has been identified under dark conditions. The magnetic ground footprints of the GOES-10 and GOES-12 satellite orbits are well-aligned with the geomagnetic latitudinal and longitudinal coverage of the ground network. The satellites provide magnetic field data with the sampling rate of 512 ms, appropriate for the observations of EMIC waves. The data acquired in the year 2007 have been analyzed to identify EMIC waves from the satellites and ULF Pc 1-2 waves on the ground. The dates in the data set are categorized into three conditions based on the sunlight condition in the ionosphere – “Sunlit” if its start time was after sunrise and its end time was before sunset for all of the stations; “Dark” either if its start time and end time were earlier than sunrise or if both start and end times were later than sunset for all of the stations; “Mixed” if all events did not fall into one of these three cases. Acknowledgement This research was supported by NSF grants ANT-0838910, ANT-0839938, and ANT-0636874 to the University of New Hampshire and by NSF grants ANT-0838917 and ANT-0840133 to Augsburg College. The authors would like to thank Mike Rose and the engineering and logistics staff at the British Antarctic Survey for their deployment and operation of the Halley search coil instrument and the PENGUIn (Polar Experiment Network for Geospace Upper-atmosphere Investigations) team members for deploying/servicing the Antarctic search coil magnetometer array under NSF grants ANT-0840158 and ANT-0638587. Lasse Clausen acknowledges funding from the National Science Foundation under grant number ATM-0924919.