Optimal subtraction of OH airglow emission: A tool for infrared fiber spectroscopy

We have developed an optimal subtraction algorithm for removing the residuals of the sky background from multi-fiber spectra. In particular, this algorithm is suited to recover the flux of unresolved lines blended with atmospheric line features and works well in case of spectra dominated by the sky component. Our method is based on PSF matching and can compensate both imperfect wavelength calibration residuals and fiber-to-fiber PSF variations. Using simulated fiber spectra of an emission line object, we show how the algorithm improves the accuracy in line fluxes evaluation and retrieves diagnostic lines normally hidden, making them usable for scientific analysis.