Bayesian estimation of spectral parameters of the 6.7-GHz methanol maser G339.884−1.259 from GRAO observations

Abstract

ABSTRACT Accurate decomposition of methanol maser spectra is essential for constraining the kinematics and physical conditions of high-mass star-forming regions, particularly in complex blended spectra where small differences in component structure can alter physical interpretation. Conventional profile fitting approaches often rely on fixed Gaussian decompositions that do not fully capture non-Gaussian spectral structure and provide limited statistical characterization of parameter uncertainties. To address this, we develop a Bayesian spectral decomposition framework that models the 6.7 GHz methanol maser emission using alternative Gaussian, Lorentzian, and Voigt profile families, with parameter posteriors inferred through Markov chain Monte Carlo sampling. This probabilistic framework enables simultaneous model comparison, uncertainty quantification, and statistically consistent estimation of spectral components. Application to the methanol maser G339.884–1.259 observed with the Ghana Radio Astronomy Observatory reveals a complex multicomponent spectrum composed of seven velocity-coherent features with tightly constrained parameters. Comparative analysis demonstrates that models incorporating both Doppler-like cores and Lorentzian wing structure provide the statistically preferred representation of the observed spectra, yielding the lowest information criteria (AIC $\approx 1.98 \times 10^{4}$; BIC $\approx 1.99 \times 10^{4}$), the smallest residual errors (RMSE $\approx 11.1$ Jy), and the highest goodness of fit ($R^{2} \approx 0.985$). In contrast, purely Gaussian or Lorentzian representations leave systematic residual structure near the line wings and strongest maser components. Elevated reduced $\chi ^{2}_{\nu }$ values across all tested models further indicate that unresolved spectral substructure, non-ideal noise properties, and intrinsic line-profile complexity remain important limitations in single-dish maser decomposition. These results demonstrate that Bayesian inference provides a robust and reproducible framework for analysing complex maser spectra while simultaneously quantifying parameter uncertainties and model limitations. The methodology is readily extendable to other molecular line studies and establishes a pathway towards integrating statistically rigorous spectral modelling with high-resolution interferometric observations to better constrain the dynamics and environments of massive star formation.