Outstanding challenges for galaxy formation models and simulations
Organisateur(s) : Andrea Cattaneo (GEPI)
Semianalytic models and computer simulations have played a major role in shaping our understanding of galaxy formation, which has advanced considerably in the last ten years. Before 2006, the number density of massive galaxies was overpredicted. Then we understood how shock heating coupled to black hole feedback shuts down star formation in massive haloes, and the main problem became low-mass galaxies (why is star formation so inefficient in massive haloes ?) Today we are beginning to understand how stellar feedback works, even though many of the physics remain unclear. The baryon mass that is blown out of galactic haloes is relatively small but gas does not make stars efficiently because most of the time is on a fountain around the galaxy. The Tully-Fisher relation was another major problem for galaxy formation models (particularly for those calibrated on the galaxy stellar mass function), but now we know where it came from (disc rotation speeds differ from halo virial velocities in a way that depends on the ratio of galaxy to halo mass). The old challenge to form massive galaxies at high redshift has been largely resolved by realizing the impact of the Eddington effect on the evolution of the galaxy stellar mass function when the errors contain a dependence on redshift. Submillimetre counts used to be an insurmountable difficulty for semianalytic models but they are much less now that the importance of confusion has been realized.
However, new observations keep coming up and confronting models with new challenges. For instance, it is now possible to measure sizes of discs with z>>3 and we have studies of the morphological evolution of galaxies since z 3. In this new context, it is important to have meeting to discuss the outstanding questions and the main challenges that will dominate the field for the next decade(s). A workshop in autumn 2015 will be particularly timely, as it will fall after the release of the new semianalytic model GalICS 2.0, which will be presented as an important working tool to the researchers in our collaboration.