Flow is the most abundant and experimentally best quantified collective phenomenon established for hot QCD matter in ultra-relativistic heavy ion collisions. In general, flow measurements in heavy ion collisions are correlation measurements. Correlations can be present in the fluctuating initial conditions, they can be generated in the fluid dynamic evolution and they can arise in the hadronization process. I outline strategies for disentangling these three sources of correlations. An essential element in these strategies is the observation that the fluid dynamic response to initial perturbations can be ordered as a power series in the amplitudes of initial fluctuations, and that a truncation of this series at low order is meaningful even for phenomenologically relevant fluctuations of large amplitude. Support for this statement will be given, and I shall discuss how this can be utilized to quantify the strength of hydrodynamically generated correlations. The seminar is mainly aimed at experts in the field (i.e. short introduction, technical details given), and the presentation will assume some familiarity with the experimental state of the art.

CERN (PH-TH)