The impact of terrain forcing on cloud formation and precipitation over low mountain ranges is investigated by numerical experiments with the COnsortium for Small‐scale MOdeling (COSMO) model. The investigation comprises six case studies divided into strong and weak large‐scale synoptic forcing. To understand how the terrain affects the occurrence and intensity of precipitation, sensitivity runs with flattened mountains and incrementally smoothed terrain were performed at 500‐m horizontal grid spacing. On days with weak forcing (i.e. air‐mass convection), results show that low‐level wind convergence is crucial for the initiation of deep convection. Its strength varies between the simulations, depending e.g. on the formation of boundary‐layer rolls, and determines whether there is an increase or decrease of total precipitation in the simulations with flattened individual mountains compared with the reference run. For cases with strong synoptic forcing (i.e. passage of frontal zones), the large‐scale advection of precipitation interacts with local effects, as advected cells are not intensified by orographic uplift in the absence of mountains. As a consequence, the runs with flattened mountains show higher moisture contents over flat terrain, which can lead to more precipitation downstream. The model runs with smoothed external parameters (i.e. terrain height, land use, roughness length) show small changes on days with strong forcing and slightly larger effects under weak forcing. However, changing the resolution of the external parameters has only a relatively small effect on precipitation in high‐resolution simulations. The results from this study demonstrate the complexity of multiple processes, like lifting or flow deviation by mountains and wind changes due to thermal instabilities, on different spatial scales for the initiation of deep convection over complex terrain.