Paper detail

Dissecting galaxy formation models with Sensitivity Analysis -- A new approach to constrain the Milky Way formation history

[Abridged] We present an application of a statistical tool known as Sensitivity Analysis to characterize the relationship between input parameters and observational predictions of semi-analytic models of galaxy formation coupled to cosmological $N$-body simulations. We show how a sensitivity analysis can be performed on our chemo-dynamical model, ChemTreeN, to characterize and quantify its relationship between model input parameters and predicted observable properties. The result of this analysis provides the user with information about which parameters are most important and most likely to affect the prediction of a given observable. It can also be used to simplify models by identifying input parameters that have no effect on the outputs of interest. Conversely, it allows us to identify what model parameters can be most efficiently constrained by the given observational data set. We have applied this technique to real observational data sets associated with the Milky Way, such as the luminosity function of satellite galaxies. A statistical comparison of model outputs and real observables is used to obtain a "best-fitting" parameter set. We consider different Milky Way-like dark matter halos to account for the dependence of the best-fitting parameters selection process on the underlying merger history of the models. For all formation histories considered, running ChemTreeN with best-fitting parameters produced luminosity functions that tightly fit their observed counterpart. However, only one of the resulting stellar halo models was able to reproduce the observed stellar halo mass within 40 kpc of the Galactic center. On the basis of this analysis it is possible to disregard certain models, and their corresponding merger histories, as good representations of the underlying merger history of the Milky Way.