A Statistical Analysis of the Nulling Pulsar Population

Abstract

Approximately 8% of the $\sim$2800 known pulsars exhibit "nulling," a temporary broadband cessation of normal pulsar emission. Nulling behaviour can be coarsely quantified by the nulling fraction, which describes the percentage of time a given pulsar will be found in a null state. In this paper, we perform the most thorough statistical analysis thus far of the properties of 141 known nulling pulsars. We find weak, non-linear correlations between nulling fraction and pulse width, as well as nulling fraction and spin period which could be attributed to selection effects. We also further investigate a recently-hypothesized gap at 40% nulling fraction. While a local minimum does exist in the distribution, we cannot confirm a consistent and unique break in the distribution when we investigate with univariate and multivariate clustering methods, nor can we prove the existence of two statistically distinct populations about this minimum. Using the same methods, we find that nulling pulsars are a statistically different population from normal, radio, non-nulling pulsars, which has never been quantitatively verified. In addition, we summarize the findings of the prior nulling pulsar statistics literature, which are notoriously contradictory. This study, in context, furthers the idea that nulling fraction alone does not contain enough information to describe the behaviour of a nulling pulsar and that other parameters such as null lengths and null randomness, in addition to a better understanding of selection effects, are required to fully understand this phenomenon.