The Search for Solar Gravity-Mode Oscillations: An Analysis Using Ulysses Magnetic Field Data

In 1995 Thomson, Maclennon, and Lanzerotti (TML) reported on work where they carried out a time-series analysis of energetic particle fluxes measured by Ulysses and Voyager 2 and concluded that solar g-mode oscillations had been detected. The approach is based on finding significant peaks in spectra using a statistical F-test. Using three sets of 2048 hourly averages of Ulysses magnetic field magnitude data, and the same multitaper spectral estimation techniques, we obtain, on average, nine coincidences with the lines listed in the TML paper. We could not reject the hypothesis that the F-test peaks we obtained are uniformly distributed, and further statistical computations show that a sequence of uniformly distributed lines generated on the frequency grid would have, on average, nine coincidences with the lines of TML. Further, we find that a time series generated from a model with a smooth spectrum of the same form as derived from the Ulysses magnetic field magnitude data and having no true spectral lines above 2 μHz, when subjected to the multitaper F-tests, gives rise to essentially the same number of "identified" lines and coincident frequencies as found with our Ulysses data. We conclude that our average nine coincidences with the lines found by TML can arise by mechanisms wholly unconnected with the existence of real physical spectral lines and hence find no firm evidence that g-modes can be detected in our sample of magnetic field data.