The Effect of High Column Density Systems on the Measurement of the Lyman α Forest Correlation Function

We present a study of the effect of High Column Density (HCD) systems on the Lyman \alpha forest correlation function on large scales. We study the effect both numerically, by inserting HCD systems on mock spectra for a specific model, and analytically, in the context of two-point correlations and linear theory. We show that the presence of HCDs substantially contributes to the noise of the correlation function measurement, and systematically alters the measured redshift-space correlation function of the Lyman \alpha forest, increasing the value of the density bias factor and decreasing the redshift distortion parameter $\beta_\alpha$ of the Lyman \alpha forest. We provide simple formulae for corrections on these derived parameters, as a function of the mean effective optical depth and bias factor of the host halos of the HCDs, and discuss the conditions under which these expressions should be valid. In practice, precise corrections to the measured parameters of the Lyman \alpha forest correlation for the HCD effects are more complex than the simple analytical approximations we present, owing to non-linear effects of the damped wings of the HCD systems and the presence of three-point terms. However, we conclude that an accurate correction for these HCD effects can be obtained numerically and calibrated with observations of the HCD-Lyman \alpha cross-correlation. We also discuss an analogous formalism to treat and correct for the contaminating effect of metal lines overlapping the Lyman \alpha forest spectra.