Investigating the high time-resolution statistics of pulsar radio signals using spectral self-noise

While observations of the stationary component of pulsar radio signals have in many ways formed the basis of our understanding of radio pulsars, the statistical deviations of these signals contain information that has become increasingly relevant. Using high time-frequency resolution data from the MeerKAT telescope, we study the self-noise of the autocorrelation function of six radio pulsars. The self-noise of the autocorrelation function is used to investigate the statistics of the observed radio signals on nanosecond time-scales and for five pulsars it is found to deviate from the expected form for a Gaussian process. Comparing the measured distribution of the intensity fluctuations of the on-pulse window to simulated models, we find that a mixture model comprising a Gaussian process and a Bernoulli-sampled Gaussian process is able to produce the excess self-noise while also producing the observed distribution of intensities. The parameters of the mixture model describing the signals are estimated for three of the pulsars in our sample group. Studies of the statistics presented in this work provide observational information for constraining the numerous theories of pulsar radio emission mechanisms. The mixture model suggested in this work would produce excess timing residuals for high signal-to-noise ratio observations when compared to that expected for a Gaussian process. Additionally, the measure of spectral self-noise provides a means of separating Gaussian and non-Gaussian processes that provides a potential basis for the development of alternative pulsar detection algorithms.

Automated summary

By studying the self-noise of the autocorrelation function of six radio pulsars, deviations from expected forms and the observed distribution of intensities were found. A mixture model, comprising a Gaussian process and a Bernoulli-sampled Gaussian process, was proposed to explain these deviations and produce the observed distribution.