THE BOSS Ly alpha FOREST SAMPLE FROM SDSS DATA RELEASE 9

We present the BOSS Lyman-alpha (Ly alpha) Forest Sample from SDSS Data Release 9, comprising 54,468 quasar spectra with z(qso) > 2.15 suitable for Ly alpha forest analysis. This data set probes the intergalactic medium with absorption redshifts 2.0 < z alpha < 5.7 over an area of 3275 deg(2), and encompasses an approximate comoving volume of 20 h(-3) Gpc(3). With each spectrum, we have included several products designed to aid in Ly alpha forest analysis: improved sky masks that flag pixels where data may be unreliable, corrections for known biases in the pipeline estimated noise, masks for the cores of damped Ly alpha systems and corrections for their wings, and estimates of the unabsorbed continua so that the observed flux can be converted to a fractional transmission. The continua are derived using a principal component fit to the quasar spectrum redward of rest-frame Ly alpha (lambda > 1216 angstrom), extrapolated into the forest region and normalized by a linear function to fit the expected evolution of the Ly alpha forest mean flux. The estimated continuum errors are less than or similar to 5% rms. We also discuss possible systematics arising from uncertain spectrophotometry and artifacts in the flux calibration; global corrections for the latter are provided. Our sample provides a convenient starting point for users to analyze clustering in BOSS Ly alpha forest data, and it provides a fiducial data set that can be used to compare results from different analyses of baryon acoustic oscillations in the Ly alpha forest. The full data set is available from the SDSS-III DR9 Web site.