The observations of the planetary radio astronomy (PRA) experiment aboard Voyager 2 reveal the existence of smooth and bursty radio emission. Both recur in a rather regular pattern with a 16.1-hour period (the Neptunian spin period). We describe the phenomenology of the smooth component in terms of frequency, polarization, and occurrence in magnetic longitude and latitude. The existence of both right-handed and left-handed polarized emissions is consistent with two sources (one in each hemisphere) which radiate independently in the R-X mode. Because Voyager passed Neptune at less than 5000 km from the surface at high northern magnetic latitudes, the radio sources were occulted by the planet near the encounter. We have taken advantage of this occultation to locate the northern hemisphere sources by calculating the radio horizon [based on the offset tilted dipole (OTD2) model> for two spacecraft positions close to the encounter. We find that the northern source is located at high magnetic latitudes Δm>40¿. By using a geometrical beaming model which assumes emission in a hollow cone pattern we fit the observed PRA intensity profile. The best fit is obtained for a radio source at L=6, thus confirming Δm>40¿. The longitudinal extent of the source is at least 180¿, from -90¿ to +90¿ magnetic longitude.

To locate the southern sources we use the large excursions of Voyager 2 to high southern magnetic latitudes. The source is found to be located also at high magnetic latitudes but is possibly more limited in longitude than the northern source. We estimate the uncertainty in source location due to the very limited knowledge of the magnetic field near the Neptunian surface, and we show that the current OTD2 model is satisfactory for the source locations for the lowest observed frequencies; however, an angular uncertainty of about 20¿ remains for sources in the northern hemisphere. The observed pattern of the smooth emission is strongly frequency dependent, which is in agreement with the azimuthal asymmetries of the magnetic field as predicted by the OTD2 model. ¿American Geophysical Union 1991