3D full-spectrum fitting: Algorithm comparison

Context. Full-spectrum fitting is the prevailing method for extracting stellar kinematic and population measurements from 1D galaxy spectra. 3D methods refer to an analysis of integral field spectroscopy (IFS) data, where the spatial and spectral dimensions are modelled simultaneously. While several 3D methods exist for modelling gas structures probed by a single emission, the more computationally demanding problem of 3D full-spectrum fitting for stellar recoveries has been investigated less frequently. Aims. We introduce and compare two algorithms for 3D full-spectrum fitting: the projected Nesterov Kaczmarz reconstruction method (PNKR), and a version of the Bayes-LOSVD software, which we modified to account for spatial correlations. The former is an iterative regularisation method that models the joint density over stellar populations and kinematics. The latter is a probabilistic method for kinematic recoveries. We aim to understand the strengths and weaknesses of the two algorithms and assess the effect of 3D methods for stellar inferences. Methods. We applied the two recovery algorithms to mock IFS data with a ground-truth model resembling a counter-rotating galaxy, over a signal-to-noise ratio (S/N) range from 20-200. We evaluated the quality of the recoveries compared to the known ground truth. Results. Accounting for spatial correlations in Bayes-LOSVD significantly improved the accuracy and precision of kinematic recoveries. 3D modelling with PNKR did not provide any significant improvement over 1D fits, but for S/N > 40, PNKR recovered the most accurate kinematics overall. Additionally, by modelling the joint distribution over kinematics and populations, PNKR successfully inferred trends between these quantities. For instance, it inferred local metallicity-velocity trends, but with a significant bias on the absolute metallicity. This successful demonstration of joint kinematic-population analyses includes cases when the counter-rotating components are not spectroscopically well-resolved. Conclusions. After demonstrating the advantages of (i) 3D modelling with Bayes-LOSVD and (ii) joint kinematic-population analyses with PNKR, we conclude that the two methodological advances will prove useful for detecting and characterising stellar structures from IFS data.