Figure 2 shows an auroral hiss event recorded starting at 0548 UT on March 16, 1997, lasting approximately 3 minutes, and extending from the lower bound of the instrument (30 kHz) into the AM-broadcast band above 500 kHz. Dark horizontal lines on this spectrogram represent fixed-frequency man-made transmissions. In contrast to the fixed-frequency signals, the auroral hiss is clearly elliptically polarized because alternate receiver sweeps differ in amplitude. Careful inspection reveals that the observed dark-light pattern corresponds to RP. This result is consistent with previous auroral hiss polarization measurements using several techniques [Tanaka, Y. et al.,1976]. Also, auroral hiss is believed to propagate in the whistler mode in the ionosphere, which implies RP, consistent with the observation. This observation of auroral hiss provides a natural calibration of the polarization detector.
Figure 3a (top panel) is a spectrogram showing three types of auroral radio emissions recorded 0445-0459 UT on April 4, 1997. Auroral hiss occurs below ~500 kHz beginning near 0452 UT lasting for ~5 minutes and again during the last ~30 seconds of the record. MF-burst is the broadband (~2 MHz) emission above ~1.5 MHz and correlated with the auroral hiss. Auroral roar is the relatively narrowband ( kHz) emission at ~3 MHz beginning near 0448 UT and ending near 0453 UT. Horizontal dark bands are fixed-frequency transmissions; the band from 550-1600 kHz is the AM broadcast band. The sweep identification marker is at 1-1.25 MHz. Clearly, alternate sweeps are light or dark, implying that they are elliptically polarized. Careful inspection shows that the MF-burst and auroral roar are LP and auroral hiss is RP.
Figure 3b (bottom panel) shows the polarization of the signals as a grayscale. In this display, white and black pixels correspond to LCP and RCP waves respectively. Elliptically polarized waves are represented as gray pixels between the two extremes, with linear polarization being at the middle of the gray band (half way between white and black). As expected, the auroral hiss shows up as dark pixels implying right-hand polarization. In contrast, both MF-burst and auroral roar are left-hand polarized. There are two sweeps at the onset of an auroral substorm near 0453 UT during which the polarization measurement momentarily implies that the auroral roar is RP, but at this time the auroral roar amplitude is probably highly time variable, as is known from fine structure measurements [e.g.,LaBelle, J. et al.,1997; Shepherd, S. G. et al.,1996], and under such conditions the polarization measurement cannot be trusted.
|Type of Event||Right||Left||Unknown|
Auroral emissions were seen on 38 of the 78 days of observation between March 15 and June 1, 1997. For purposes of accumulating statistics, an event is defined as an auroral emission that is detected for longer than 30 seconds and is separated by at least 10 minutes from other events. Table 1 shows the sense of polarization of all events which occurred during the observation period, as determined from spectrograms similar to those shown in Figure 3b. Table 1 shows that auroral hiss is right elliptically polarized and MF-bursts and 2 auroral roar are both left elliptically polarized. While the statistics establish that these emissions are almost entirely left-hand polarized, more statistics are needed to exclude the possibility that a few percent of these emissions are right-hand or linearly polarized. The few events in the unknown column of the table are due to interference from strong atmospherics. An atmospheric produces an output to the receiver which resembles a broadband signal that is strongly polarized in either direction, depending on the sign of the induced phase shift during the event. The impulsive nature of these signals obscure the polarization of coincident waves and illustrate the effect of a signal which varies in amplitude faster than the sweep period of the receiver.