| Abstract | A procedure for estimating the internal reliability of the systemic velocity and the velocity width of galaxies inferred from observations of H I profiles is presented. A Monte Carlo approach is used to analyze the mean widths and variances of a single observation, regarding the signal as a peak in the profile, with noise being the rms variation in the flux. Attention is given to the S/N ratio, with gaussian noise added at an equal magnitude to the rms noise in the original profile. A Monte Carlo simulation then represents the effects of noise on the velocity measurements. A bias towards overestimation of the widths of narrow profiles was found, together with the correlation between predicted errors and S/N ratios. It is recommended that observers set S/N criteria for all observations, and that rms noise present in H I observations be published. |
