Several year's worth of observations from the plasma wave instruments on both Magnetopause-to-Aurora Global Exploration (IMAGE) and Polar spacecraft are used to study the seasonal and solar cycle variations in the spectrum of auroral kilometric radiation (AKR). Only AKR observations when the spacecraft were in the Northern Hemisphere emission cones were used. The results from the seasonal analysis show significant changes in the AKR spectrum as a function of dipole tilt. The average AKR spectral peak for positive dipole tilt is ~150 kHz but is ~300 kHz during times of negative dipole tilt. In addition, the average emission spectrum for positive tilt is up to two orders of magnitude weaker over the 200--500 kHz frequency range when compared with the average emission spectrum for negative tilt. Assuming the cyclotron maser mechanism for AKR, these results imply that the AKR source region (the auroral density cavity) moves to higher altitudes during the summer and to lower altitudes during the winter. Using data from the DE-1 plasma wave instrument, the magnetic local time of average AKR source region is also investigated with dipole tilt. From these observations it is found that for negative dipole tilt a broad AKR source region exists, ranging from ~18 to ~24 MLT, with peak emission coming from ~20 MLT. In comparison, under positive dipole tilt the source region narrows (~20 to ~24 MLT) with peak emission at ~22 MLT. Taking into account the above seasonal effect, a comparison of the average spectra from IMAGE and Polar plasma wave data also demonstrates a solar cycle effect. The average AKR spectrum at solar maximum has the same structure with dipole tilt as at solar minimum but is typically lower (by as much as 1--2 orders of magnitude). The observations presented support the concept that the expected increases in ionospheric densities (with positive dipole tilt for the Northern Hemisphere and solar EUV flux increases during solar maximum) play a significant role in magnetospheric-ionospheric coupling by: (1) shortening the altitude range of the auroral plasma cavity, (2) confining the cavity to a smaller range of MLT and closer to midnight, and (3) decreasing the overall intensity of AKR by lessening the density depth of the auroral density cavity. The results of this study should be taken into account in future studies of using AKR as a substorm index and other statistical emission cone studies at both low and high frequencies.