Initial Selection

Following is a short description of how the MHONGOOSE sample was put together. As one of the main goals of MHONGOOSE is to trace accretion and star formation processes over a large range in galaxy properties, it is important to ensure that within the limited observing time available a representative range of HI masses, star formation rates and rotation velocities is sampled. Furthermore, strongly interacting galaxies are preferably avoided.

A primary criterion for the sample selection is that galaxies should have been detected in HI before. We therefore made the Parkes HIPASS catalogue our natural starting point. To ensure the availability of a significant set of multi-wavelength data, we limited our selection to HIPASS galaxies that were also observed as part of the  SINGG and SUNGG surveys (PI. Meurer).  These surveys targeted a large fraction of the HIPASS-detected nearby galaxies and collected Hα, optical, infrared and ultraviolet data, with galaxies selected uniformly in bins of log(MHI) (i.e. equal number of galaxies in each bin of log(MHI).   

We use the SINGG/SUNGG survey to further define the MHONGOOSE sample. The main SINGG/SUNGG selection criteria are as follows:

  • HIPASS peak flux > 50 mJy (3.8σ in HIPASS)
  • galactic latitude |b| > 30o
  • projected distance from the center of the LMC > 10o
  • Galactic standard of rest velocity > 200 km s-1

The resulting SINGG/SUNGG list was divided in 0.2 dex bins in log(MHI). In each bin the closest 30-40 galaxies were chosen, giving a flat distribution in the range 8.5 < log(MHI) < 10.5, with a small number of galaxies down to log(MHI) ~ 7 and up to log(MHI) ~ 11. Most galaxies selected this way have z < 2000 km s-1 with a median redshift of 1300 km s-1 or 18.5 Mpc distance. The total number of galaxies selected this way is 468 (see Fig. 1, taken from Meurer et al 2006).

Requiring that a complete set of Hα, R-band, GALEX UV data is available reduces this number to 151 HI detections. Looking at the distribution of the number of galaxies versus distance (see Fig 2), it is clear that 30 Mpc is a natural distance to impose a further cut on the sample. The only unique galaxies removed by this are a small number of log(MHI)>10.5 galaxies. The type of lower-mass galaxies removed is also present at smaller distances. We limit this sample further by selecting only those galaxies with declination δ < 10o and we exclude the survey area of the Fornax Cluster Survey MeerKAT project (PI Paolo Serra). This reduces the list to 88 galaxies. 

As with SINGG/SUNGG we want a representative number of galaxies distributed as uniformly as possible over log(MHI). With the comparatively smaller number of galaxies we use bin sizes of 0.5 dex. The 88 galaxies are distributed over log(MHI) as shown in Table 1. For reference, some details on these proto-sample galaxies can be found here and here.

To get to the final sample of 30 galaxies we thus need to select 5 galaxies per bin. This selection cannot be purely random, there are certain well-defined inclination ranges where different science questions can be optimally addressed. 

These ranges are purely face-on, purely edge-on and an intermediate inclination around 50-60 degrees. Face-on allows the best characterization of the morphology of the ISM, as well as determination of vertical motions. Edge-on allows an unambiguous characterization of the vertical structure of the ISM. The intermediate inclination range allows (sometimes with a certain amount of modeling) determination of both, and can be used to tie results from edge-on and face-on  classes together. The selection of the final 30 galaxies is described here.

© Erwin de Blok 2017