The NANOGrav Nine-Year Data Set: Noise Budget for Pulsar Arrival Times on Intraday Timescales

The use of pulsars as astrophysical clocks for gravitational wave experiments demands the highest possible timing precision. Pulse times of arrival (TOAs) are limited by stochastic processes that occur in the pulsar itself, along the line of sight through the interstellar medium, and in the measurement process. On timescales of seconds to hours, the TOA variance exceeds that from template-fitting errors due to additive noise. We assess contributions to the total variance from two additional effects: amplitude and phase jitter intrinsic to single pulses and changes in the interstellar impulse response from scattering. The three effects have different dependencies on time, frequency, and pulse signal-to-noise ratio. We use data on 37 pulsars from the North American Nanohertz Observatory for Gravitational Waves to assess the individual contributions to the overall intraday noise budget for each pulsar. We detect jitter in 22 pulsars and estimate the average value of RMS jitter in our pulsars to be $\sim 1\%$ of pulse phase. We examine how jitter evolves as a function of frequency and find evidence for evolution. Finally, we compare our measurements with previous noise parameter estimates and discuss methods to improve gravitational wave detection pipelines.