|
Context. The neutral hydrogen gas (HI) in the interstellar medium (ISM) can be broadly categorised into three phases: the bistable a cold and warm neutral medium (CNM and WNM respectively) and the thermally unstable lukewarm neutral medium (LNM). The distribution of these phases can be recovered by decomposing the 21-cm spectra into Gaussian components. Aims. We used various Gaussian decomposition algorithms with the aim of ascertaining if different algorithms lead to different properties recovered about the ISM. Methods. We generated synthetic datasets to compare Gausspy and Gausspy+ across several different types of spectral profiles. The best algorithm was then used to decompose the whole-sky HI4PI results into Gaussian components. We finally compared our decomposition results with those obtained by a previous investigation. Results. We first found that Gausspy+ performed better in recovering components especially in more complex spectral profiles. However, the results we obtained in the latter half were significantly influenced by a poor fitting of Gaussian components to the spectra. Conclusions. The differences we observed in the synthetic tests were a promising lead that the results recovered by algorithms do indeed differ. However, the distributions we obtained for the HI4PI observations were not convincingly indicative of a relation between choice of algorithm and the derived properties of the ISM. Further study into such an interdependence is needed. |
|