A survey has been performed of the power spectra of the electric field measured by stratospheric balloon-borne instruments during the 1985--1986 South Pole Balloon Campaign. The survey reviewed all 468 hours of available data and covered the frequency range from 0 to 4 Hz. Three episodes of narrow-band emissions were detected in this frequency range. These emissions occurred about 5.5% of the time, with intense emission occurring about 1% of the time. These narrow-band emissions had frequencies in the horizontal component between 2 and 2.5 Hz. The frequencies observed in the vertical component were different, with several bands present. The fundamental frequency in the vertical component was around 0.8 to 1.4 Hz depending on event. The spectra of the events in the vertical component showed multiple emission bands, not all harmonically spaced. The vertical component showed much more pronounced time variations of the frequencies of the emission bands than the horizontal component showed. The amplitude of the emission in the vertical component was usually but not always larger than that in the horizontal components. The ratio EV/EH was typically ~2--3. The amplitude of the horizontal component of the emissions was quite large in one of the three events, with individual bursts having amplitudes of 60 mV/m. The polarization was complicated. In the horizontal plane, there were two narrow bands of opposite-handed circularly polarized emissions. Since the vertical axis of the emission had frequency components not present in the horizontal plane, the polarization in the zonal plane was linear, with a quasi-Lissajou character.

An extensive list of possible explanations for these emissions was considered. Since similar emissions with a similar occurrence rate have been seen by other workers, instrumental explanations were ruled out. Local external artifacts were also ruled out. Distant sources associated with tropospheric storms were considered in detail. During the most intense of the 2-Hz emission events, the nearest bad weather was more than 1000 km away. It proved impossible to account for the amplitude and polarization of the horizontal component with any reasonable tropospheric source. Explaining the narrow-band tuning was also a problem for any weather source model. An ionospheric model based on the concept of the ionospheric Alfv¿n resonator is capable of accounting for the tuning, amplitude, and polarization of the horizontal components. However, this model has moderate difficulty in accounting for the observed vertical component of the emissions. In the end, no available model proved fully satisfactory. ¿ American Geophysical Union 1995