Large-scale biomass plantations (BPs) are a common factor in climate mitigation scenarios as they promise double benefits: extracting carbon from the atmosphere and providing a renewable energy source. However, their terrestrial carbon dioxide removal (tCDR) potentials depend on important factors such as land availability, efficiency of capturing biomass-derived carbon and the timing of operation. Land availability is restricted by the demands of future food production depending on yield increases and population growth, by requirements for nature conservation and, with respect to climate mitigation, avoiding unfavourable albedo changes. We integrate these factors in one spatially explicit biogeochemical simulation framework to explore the tCDR opportunity space on land available after these constraints are taken into account, starting either in 2020 or 2050, and lasting until 2100. We find that assumed future needs for nature protection and food production strongly limit tCDR potentials. BPs on abandoned crop and pasture areas (~1,300 Mha in scenarios of either 8.0 billion people and yield gap reductions of 25% until 2020 or 9.5 billion people and yield gap reductions of 50% until 2050) could, theoretically, sequester ~100 GtC in land carbon stocks and biomass harvest by 2100. However, this potential would be ~80% lower if only cropland was available or ~50% lower if albedo decreases were considered as a factor restricting land availability. Converting instead natural forest, shrubland or grassland into BPs could result in much larger tCDR potentials ̶ but at high environmental costs (e.g. biodiversity loss). The most promising avenue for effective tCDR seems to be improvement of efficient carbon utilization pathways, changes in dietary trends or the restoration of marginal lands for the implementation of tCDR.