Paper abstract To meet decarbonisation targets, nations around the globe have made ambitious commitments to expand forested land. Operationalising these commitments requires choosing a planting strategy: how many trees should be planted, of which species, and where? Given those choices must be made now but have long term consequences, such decisions are plagued by uncertainty. For example, species that are well suited to present conditions may perform poorly under future climates, yet those future climates are themselves highly uncertain. Using the exemplar of the UK, a nation committed to achieving net zero emissions by mid-century, we quantify key uncertainties pertaining to co-evolving climate and economic conditions and examine how modern methods of decision-making under uncertainty can advise on planting choices. Our analysis reveals that the best planting strategy assuming a ‘high-emissions’ future is radically different to that for a future that remains on a ‘near-historic’ path. Planting for the former while experiencing the latter results in substantial net costs to UK society. Assimilating uncertainty into decision-making identifies planting strategies that diversify risk and significantly reduce the probability of high-cost outcomes. Importantly, our research reveals that the scope for mitigating risk through choice of planting strategy is relatively limited. Despite this persistent risk, we find that tree planting remains a highly cost-effective carbon removal solution when compared to alternative technologies, even when those alternatives are assumed to be riskless.