Numerical modeling shows great potential as a method for investigating and predicting the future development of ice shelves in a warming climate. The quality of ice shelf-ocean models is continuously improving but some limitations remain. For models using a terrain-following vertical coordinate, one such limitation is the enforcement of a minimum water-column thickness beneath ice shelves by adjustment of bottom topography. How this local distortion of bathymetry from reality affects modeled melt rates and cavity circulation is unknown so far. To quantify this effect, simulations with the Finite Element Sea ice–ice shelf–Ocean Model (FESOM) were executed on four different grids with minimum water-column thicknesses of 20 m, 50 m, 100 m and 200 m. While we use a global model grid, modifications of bathymetry are applied only to Filchner–Ronne Ice Shelf. We show that the choice of minimum water-column thickness does not affect the total basal melt rate of this cold-water ice shelf but does in fact impact the distribution of melt rates with significant differences between experiments in the magnitude of melting near the grounding lines.