Removing atmospheric carbon (C) and storing it in the terrestrial biosphere is one of the options, which have been proposed to compensate greenhouse gas (GHG) emissions. Agricultural lands are believed to be a major potential sink and could absorb large quantities of C if trees are reintroduced to these systems and judiciously managed together with crops and/or animals. Thus, the importance of agroforestry as a land-use system is receiving wider recognition not only in terms of agricultural sustainability but also in issues related to climate change. The objective of this paper was to analyse C storage data in some tropical agroforestry systems and to discuss the role they can play in reducing the concentration of CO2 in the atmosphere. The C sequestration potential of agroforestry systems is estimated between 12 and 228 Mg ha-1 with a median value of 95 Mg ha-1. Therefore, based on the earth's area that is suitable for the practice (585-1215x10^6 ha), 1.1-2.2 Pg C could be stored in the terrestrial ecosystems over the next 50 years. Long rotation systems such as agroforests, homegardens and boundary plantings can sequester sizeable quantities of C in plant biomass and in long-lasting wood products. Soil C sequestration constitutes another realistic option achievable in many agroforestry systems. In conclusion, the potential of agroforestry for CO2 mitigation is well recognised. However, there are a number of shortcomings that need to be emphasised. These include the uncertainties related to future shifts in global climate, land-use and land cover, the poor performance of trees and crops on substandard soils and dry environments, pests and diseases such as nematodes. In addition, more efforts are needed to improve methods for estimating C stocks and trace gas balances such as nitrous oxide (N2O) and methane (CH4) to determine net benefits of agroforestry on the atmosphere.
