The Data Analysis Pipeline for the SDSS-IV MaNGA IFU Galaxy Survey: Emission-Line Modeling

SDSS-IV MaNGA (Mapping Nearby Galaxies at Apache Point Observatory) is the largest integral-field spectroscopy survey to date, aiming to observe a statistically representative sample of 10,000 low-redshift galaxies. In this paper we study the reliability of the emission-line fluxes and kinematic properties derived by the MaNGA Data Analysis Pipeline (DAP). We describe the algorithmic choices made in the DAP with regards to measuring emission-line properties, and the effect of our adopted strategy of simultaneously fitting the continuum and line emission. The effect of random errors are quantified by studying various fit-quality metrics, idealized recovery simulations and repeat observations. This analysis demonstrates that the emission lines are well-fit in the vast majority of the MaNGA dataset and the derived fluxes and errors are statistically robust. The systematic uncertainty on emission-line properties introduced by the choice of continuum templates is also discussed. In particular, we test the effect of using different stellar libraries and simple stellar-population models on the derived emission-line fluxes and the effect of introducing different tying prescriptions for the emission-line kinematics. We show that these effects can generate large ($>$ 0.2 dex) discrepancies at low signal-to-noise and for lines with low equivalent width (EW); however, the combined effect is noticeable even for H$\alpha$ EW $>$ 6 \AA. We provide suggestions for optimal use of the data provided by SDSS data release 15 and propose refinements on the DAP for future MaNGA data releases.