Dispersion measure variations and their effect on precision pulsar timing
X. P. You, G. Hobbs, W. A. Coles, R. N. Manchester, Roderick Edwards, M. Bailes, John Sarkissian, J. P. W. Verbiest, W. van Straten, A. W. Hotan, S. M. Ord, Fredrick Jenet, N. D. R. Bhat, A. Teoh
Abstract
We present an analysis of the variations seen in the dispersion measures (DMs) of 20-ms pulsars observed as part of the Parkes Pulsar Timing Array project. We carry out a statistically rigorous structure function analysis for each pulsar and show that the variations seen for most pulsars are consistent with those expected for an interstellar medium characterized by a Kolmogorov turbulence spectrum. The structure functions for PSRs J1045-4509 and J1909-3744 provide the first clear evidence for a large inner scale, possibly due to ion-neutral damping. We also show the effect of the solar wind on the DMs and show that the simple models presently implemented into pulsar timing packages cannot reliably correct for this effect. For the first time we clearly show how DM variations affect pulsar timing residuals and how they can be corrected in order to obtain the highest possible timing precision. Even with our presently limited data span, the residuals (and all parameters derived from the timing) for six of our pulsars have been significantly improved by correcting for the DM variations.