I just received the following workshop announcement:
11th Birmingham-Nottingham Extragalactic Workshop - 1st Announcement "Semi-analytic models - are we kidding ourselves?"
Refeshingly honest conference title aside, this is a terrific topic for a workshop. Semi-analytic galaxy formation models are extremely useful tools, which consist of (1) an underlying prescription for the growth of dark matter halos and (2) a set of knobs for grafting complicated baryonic physics onto those halos. The first step is well-understood analytically, and has been reliably calibrated with N-body simulations. The second step, however, contains a lot of crafty juju (how much gas winds up inside the dark matter halos? At what rate does that gas cool? How and when does that gas convert into stars? How does the formation of stars and the subsequent supernovae affect the surrounding gas? How do mergers between dark matter halos change the spatial distribution of the stars and the temperature structure of the gas?). The developers of these models are smart folks, and make reasonably well-informed assumptions about all of these baryonic processes, leading to results that are decent matches to the ensemble properties of the galaxy populations. (Note that I didn't say "predict results" -- these models are ex post facto in most usages.) However, just because the models can be tuned to produce a rough statistical match to observations in no way means that the input assumptions are correct or unique descriptions of what actually happens. Moreover, there are serious discrepancies between the models and the observed properties of very low mass galaxies -- when the models are tuned to match the properties of relatively massive galaxies, they predict that the low mass galaxies are red and gas poor, whereas the observations say they're blue and gas rich. It's great that the community is looking at these issues head on, given the usefulness the semi-analytic galaxy models.
