Simulating radio-off fractions in rotating radio transients

Aims. We aim to simulate the proportions of non-detectable emission, measured as radio-off fractions (f(off)), in rotating radio transients (RRATs). We also investigate the properties related to the underlying mechanism for such sporadic emission. Methods. From observations of intermittent pulsars, radio emission originates from two distinct emission states and it becomes non-detectable when the pulsar switches to an emission state characterized by magnetospheric plasma density of zero. We performed simulations of f(off) based on 10 000 samples, each with 10 000 rotations and using a model that tracks changes in the plasma density in a pulsar magnetosphere with multiple emission states. We assumed that (i) RRATs are radio pulsars, (ii) radio pulse intensity is correlated with the emitting plasma density as stated in the conventional models, and (iii) a pulse emission corresponds to a change in the plasma density under favorable conditions. Results. A best-fit distribution for f(off) is obtained when emission from RRATs is defaulted to radio-off. The resulting wait time distribution can be fitted by two functions of an exponential and a Gaussian, which is consistent with the observations. We demonstrate that the switch rate is low and that the burst rate is dependent on rotation period. In addition, the switch rate is related to the obliquity angle, which implies that the mechanism varies over time. Our results suggest that switching to radio-on is a random process, which implies that the burst rate is different for different RRATs. We show that RRAT emission and pulse nulling may share similar origins, but with different default emission. We discuss how the emission may change from that of RRAT to pulse nulling (or vice versa) as a pulsar evolves.