Abstract. The availability of dissolved silica (Si) in the ocean provides a major control on the growth of siliceous phytoplankton. Diatoms in particular account for a large proportion of oceanic primary production. The original source of the silica is rock weathering, followed by transport of dissolved and biogenic silica to the coastal zone. This model study aims at assessing the sensitivity of the global marine silicon cycle to variations in the river input of silica on timescales ranging from several centuries to millennia. We compare the performance of a box model for the marine silicon cycle to that of a global biogeochemical ocean general circulation model (HAMOCC2 and 5). Results indicate that the average global ocean response to changes in river input of silica is comparable in the models on time scales up to 150 kyrs. While the trends in export production and opal burial are the same, the box model shows a delayed response to the imposed perturbations compared to the general circulation model. Results of both models confirm the important role of the continental margins as a sink for silica at the global scale. Our work also demonstrates that the effects of changes in riverine dissolved silica on ocean biogeochemistry depend on the availability of the other nutrients such as nitrogen, phosphorus and iron. The model results suggest that the effects of reduced silica inputs due to river damming are particularly pronounced in the Gulf of Bengal, Gulf of Mexico and the Amazon plume where they negatively affect opal production. While general circulation models are indispensable when assessing the spatial variation in opal export production and biogenic Si burial in the ocean, this study demonstrates that box models provide a good alternative when studying the average global ocean response to perturbations of the oceanic silica cycle (especially on longer time scales).