Comparison of kilometric continuum latitudinal radiation patterns with linear mode conversion theory

<jats:p>Using 80 observations of kilometric continuum (KC) radiation patterns detected by the Radio Plasma Imager on the IMAGE spacecraft, the latitudinal variation of these radiation patterns is compared with the predictions of linear mode conversion theory (LMCT). Six of these comparisons are presented in this paper. Because the location and shape of the plasmapause at the source cannot be simultaneously determined with all of the KC observations, we estimated the minimum frequency‐dependent beaming angle possible based on a statistical upper limit plasma density versus <jats:italic>L</jats:italic>‐shell profile. This density versus <jats:italic>L</jats:italic>‐shell profile is equated to the 90th percentile plasma density versus <jats:italic>L</jats:italic>‐shell curve derived from over 40,000 in situ measurements of the upper hybrid resonance frequency made over the first 3 years of the IMAGE mission. For all cases, starting at frequencies approximately >150 kHz, the model beam center was found to lie just within or outside the radiation pattern. Unrealistic peak densities are required at the equatorial source <jats:italic>L</jats:italic>‐shell locations which are necessary to force the beam centers to lie deep within the observed radiation patterns. We present this as strong evidence that LMCT cannot be used to explain KC. In order to explain the latitudinal location of the high‐frequency end of these KC radiation patterns a theory is required in which these frequencies are beamed at equatorial angles that are much smaller than the predictions of LMCT. Hashimoto et al. (2005), analyzing one KC radiation pattern detected by CRRES, reached the same conclusion. We present one case in which the source location was inferred by equating the harmonic spacing in the KC radiation pattern with the model electron cyclotron frequency at the source. The LMCT model beams, for this source location, were found to lie outside of the radiation pattern at all frequencies, again consistent with Hashimoto et al. (2005). In addition a survey of KC observations by IMAGE in MLT is presented in which the probability of occurrence peaks in both the dawn and evening sectors.</jats:p>