Extended J-band spectroscopy of red supergiants: parameter identifiability and metallicity robustness

Abstract

Red supergiants (RSGs) provide a near-infrared metallicity diagnostic complementary to optical hot-supergiant spectroscopy, but their low-resolution extragalactic use requires tests of parameter identifiability and modelling robustness. We quantify the information content of extended J-band RSG spectra at R=2000 over 1.155-1.340 um and assess recovery of effective temperature, surface gravity, microturbulence, and metallicity. Using MARCS/TURBOSPECTRUM spectra and an emulator, we perform a six-node Fisher analysis and controlled mock experiments, and analyse 21 SpeX spectra of 20 RSG candidates towards Per OB1. All four parameters produce measurable responses, but only 8-10% of the raw metallicity information and 4-7% of the gravity information are independent of the other parameter directions. Extending beyond 1.22 um provides median fractions of 60% of the raw and 66% of the nuisance-orthogonal metallicity information. Gravity dominates the weakest Fisher mode, while its prior choice shifts association means by at most 0.03 dex. Matched mocks recover metallicity with an rms error of 0.04 dex. Standard continuum treatment biases solar-metallicity mocks by about -0.23 dex, reduced to about -0.01 dex by our adopted treatment. Plausible CNO mismatch adds a zero-point sensitivity of 0.04-0.05 dex, and regularised covariance changes association means by at most 0.03 dex. The preferred analysis gives a mean [Z] = -0.10 dex with a star-to-star dispersion of 0.09 dex. Extended low-resolution J-band spectra therefore yield recoverable metallicity and temperature constraints once continuum placement and weak parameter directions are controlled, although the absolute scale remains conditional on the adopted atmosphere and line-formation framework.