For several years auroral radio emissions have been continuously monitored at various northern and southern hemisphere sites using a programmable stepped frequency receiver (PSFR). This receiver is usually programmed to sweep from 30 kHz to 5 MHz every 2 seconds in 10 kHz steps. Data are collected, digitized, and stored by a local computer and sent back to Dartmouth College monthly.
In March 1997, the PSFR at Churchill (N, E, invariant latitude) was modified to measure the polarization of received signals. The modified antenna system consists of two vertical 2.5 m loop antennas oriented 90 to each other in an approximately N-S/E-W position. Figure 1 shows a schematic of the two antennas and a simplified block diagram of the system. (In the figure, the two antennas are separated for clarity, but in fact they are physically located on the same vertical mast.) The polarization detector, receiver, and computer are located ~100 meters from the antenna to minimize noise pickup.
In the polarization detector, a 90 phase lag is introduced into the signal from the N-S loop. On alternate sweeps this signal is inverted, effectively shifting the phase of the N-S loop signal from to relative to the E-W loop. The input to the receiver is the shifted and switched N-S loop signal summed with the signal from the E-W loop. If the original signals induced in the antenna loops are equal in amplitude but differ in phase by exactly , as would be expected for vertically incident right- or left-circularly polarized waves, the input signal to the receiver alternates between zero and twice the induced signal strength. On the other hand, a linearly polarized signal induces in-phase signals in the antenna loops which result in a constant input signal to the receiver, there being no difference between shifting the N-S signal forward or backward in phase in this case. To assist in data analysis, a marker is recorded during part of the noninverted sweep to allow sweep identification, and a calibration signal that simulates a received right-cirularly polarized signal is periodically inserted into the antennas. In order to determine polarization using this technique, the measured signals must be relatively constant in amplitude and polarization during two consecutive meaurements (~2 s); signals whose amplitude varies faster than that may register a false or indeterminant polarization. The sense of polarization (right or left) is determined by noting the relative signal strengths of the two sweeps and comparing that to the marker.
The interpretation above assumes that the electronics are perfect. It is difficult to shift a single signal by 90 over a wide bandwidth, but it is easy to shift both the N-S and E-W loop signals such that the phase of the N-S loop signal lags the E-W loop signal by ~90 over a range of approximately two decades. The error in phase shift over the 0.05-5.0 MHz frequency range is less than ten degrees, implying that the maximum amplitude difference between consecutive sweeps for either right- or left-circularly polarized vertically incident signals is about 20 dB which is less than the observed differences of the real signals described below.
In this paper all wave polarizations are measured with respect to the local magnetic field. The electric field vector of a right-handed circularly polarized (RCP) wave rotates clockwise in time as viewed in the direction of the magnetic field. In this case, for a receiver in the Northern Hemisphere the electric field vector rotates clockwise as seen by an observer looking down on the antennas. This definition is standard in plasma physics [e.g., Chen, F. F.,1984] and is used in previous auoral radio emission polarization studies [Tanaka, Y. et al.,1976].