2005MNRAS.364...69S

We present numerical hydrodynamical models of the collision between the galaxies IC 2163 and NGC 2207. These models extend the results of earlier work in which the galaxy discs were modelled one at a time. We confirm the general result that the collision is primarily planar, that is, at moderate inclination relative to the two discs, and prograde for IC 2163, but retrograde for NGC 2207. We list 34 specific morphological or kinematic features on a variety of scales, found with multiwaveband observations, which we use to constrain the models. The models are able to reproduce most of these features, with a relative orbit in which the companion (IC 2163) disc first side-swipes the primary (NGC 2207) disc on the west side, then moves around the edge of the primary disc to the north and to its current position on the east side. The models also provide evidence that the dark matter halo of NGC 2207 has only moderate extent. For IC 2163, the prolonged prograde disturbance in the model produces a tidal tail, and an oval or ocular waveform very much like the observed ones, including some fine structure. The retrograde disturbance in the model produces no strong waveforms within the primary galaxy. This suggests that the prominent spiral waves in NGC 2207 were present before the collision, and models with waves imposed in the initial conditions confirm that they would not be disrupted by the collision. With an initial central hole in the gas disc of the primary, and imposed spirals, the model also reproduces the broad ring seen in Hi observations. Model gas disc kinematics compare well to the observed (Hi) kinematics, providing further confirmation of its validity. An algorithm for feedback heating from young stars is included, and the feedback models suggest the occurrence of a moderate starburst in IC 2163 about 250 Myr ago.

We believe that this is now one of the best-modelled systems of colliding galaxies, though the model could still be improved by including full disc self-gravity. The confrontation between observations and models of so many individual features provides one of the strongest tests of collision theory. The success of the models affirms this theory, but the effort required to achieve this, and the sensitivity of models to initial conditions, suggests that it will be difficult to model specific structures on scales smaller than about a kiloparsec in any collisional system.