What is KinMS?Written on August 8th, 2022 by Tim Davis
The KinMS set of packages can be used to forward model 3D datacubes from interferometers and/or Integral Field Units (IFUs); or, with some postprocessing, 1D/2D data products and long-slit spectra.
KinMS standards for “KINematic Molecular Simulation”, as modelling the kinematics of molecular gas data was the original usecase (e.g. investigating the kinematics of gas in early-type galaxies, Davis et al., 2013a; and determining supermassive black-hole masses from interferometric observations, Davis et al., 2013b), but KinMS is useful for a wide variety of use cases beyond this. Some examples from the 40+ peer reviewed papers that use KinMS include:
- Creating mock datacubes from hydrodynamic simulations: Davis et al. 2019
- Modelling optical ionised-gas line profiles from MaNGA IFU/long-slit spectra in Red Geysers: Roy et al. 21
- Modelling the kinematics of edge on disc galaxies to search for non-circular motions: Hogarth et al. 2022
Why choose KinMS?
Various kinematic modelling packages are available, and widely used in the astronomical community. For instance TiRiFiC, 3D-Barolo and GalPaK3D. Each has its own strengths and weaknesses, and may be better/worse for your specific problem. Here I outline some of the advantages of KinMS, so you can determine if it is the package for you:
- KinMS allows you to go beyond tilted rings. It can be used as a classic tilted ring fitting code (but if that is your main use case and you don’t need any of KinMS’s extra functionality I recommend either TiRiFiC or 3D-Barolo), but shines when used to fit physically motivated functional forms for e.g. rotation curves and/or gas surface brightness profiles. For instance, it includes models for various potentials, allowing kinematic disc-bulge decompositions, the inclusion of compact objects, using MGE decompositions of stellar light, etc.
- KinMS is flexible. It can be used to fit non-axisymmetric structures such as spirals, bars and merging galaxies via its
inCloudsmechanism, and allows the user significant freedom in the types of circular/non-circular motions modelled.
- KinMS is Bayesian. By default the
KinMS_fitterroutines use a Markov Chain Monte Carlo (MCMC) approach, allowing the user to set arbitrary priors, explore degeneracies and obtain full ND posterior probability distribution functions.
- KinMS is implemented in pure python. This makes it simple to understand and use (to e.g. build a pipeline for fitting many objects), easy to modify for your needs (e.g. if you need to move beyond simple disc fitting), and effortless to install across different environments.