Spectroscopic follow-up on Milky Way dwarf galaxies

[sazabi4]

We discuss here the science cases of having an 8-m wide FOV (~1 deg in diameter) telescope with a multi-object spectrograph on following up the dwarf galaxies in Milky Way. And maybe also a comparison with a 30-m telescope (but a much smaller FOV, 6' in diameter?)

@segasai @jobovy @kadrlica also with Jeff Newman and Matt Walker

[sazabi4]

One possible experiment is to follow-up all (or at least most of) the dwarf galaxies that LSST will discover up to virial radius of MW. We expect about 200 (+/- 100?) more dwarfs that will be discovered by LSST. However, these dwarfs will be mostly faint (Mstar around 100Msun, Cet II and Cet III like) and therefore spectroscopy will be hard. With an 8-m telescope, we can probably reach g~22-22.5 with 1-2 nights of observing and get 5-10 members on each dwarf. It is hard to determine the velocity dispersion (and therefore dynamical mass) to high precision. But maybe with a large sample of dwarfs, we could study the stellar mass - dynamical mass relation statistically. And maybe can be used to measure the subhalo mass function.

This experiment does not require a very big FOV but will need a lot of telescope time. (300 nights?)

[sazabi4]

The other possible project is to look at the outskirt of the currently known dwarf galaxies. Most of these dwarfs are followed up within the half-light radius rh, or say have a complete sample with r < rh but not r > rh. The mass estimation from stars of r < rh alone might raise some bias. A larger FOV instrument allows member search out to 3rh or more where the missing rate at r > 3rh are below 10%.

Also, it is worthy of checking possible signature of stellar population distribution vs radius in the dwarfs if we can get a nearly complete sample of members in the dwarfs. For example, many observations shows that BHB stars are more spatially extended than RGB stars. One may use this population vs radius to study the DM density profile in UFDs. Though the sample size is small for each UFD, maybe we could stack all of the UFDs to study cusp/core in these lowest masses where baryonic feedback is minimal.

Another possible measurement is to search for tidal disruption signal by measuring the kinematics at the outskirt of UFD.

[sazabi4]

One concern raised up is that a lot of the dwarf galaxies followed up recently (with Mv > -4) showed a small velocity dispersion (sigma_v < 3 km/s and some sigma_v < 1.5 km). Therefore we probably should start thinking about the next generation spectrograph should be designed more stable and have a smaller systemic velocity uncertainty, and we might need to go to a higher spectral resolution to get smaller systemic uncertainty as well.

One exercise that should be done is the trade off between wavelength coverage/range, spectral resolution, systematic uncertainty for different stars (metallicity, temperature, etc) at a given exposure time and a given mirror size. The higher the resolution of the instrument, the smaller the wavelength coverage and possibly the smaller the systematic uncertainty but also lower S/N (so larger statistical uncertainty) at a fix exposure time. It would be interesting/valuable to figure out what is the best resolution for an 8-10m and 30m to resolve a dwarf galaxy at a velocity dispersion of, say, 1-2 km/s, or even lower.

[kadrlica]

Adding @meiyuw as an expert on some modeling projections.

[sazabi4]

I think I missed something we discussed today. Please feel free to add more...

[janewman-pitt-edu]

Here are predictions for the survey time for the Kavli report version of the dwarf galaxy project (and others), which corresponds to 4000 hours on a PFS-like instrument. Scale the below appropiately.