You discuss in an Editorial (Nature 462, 11; 2009) the promise of the emissions trading scheme REDD, whereby tropical countries will be rewarded for increased sequestration by forests. But it is important for negotiations to focus on the realities rather than on the ideal.

REDD countermeasures to tropical deforestation will affect food supplies and employment and will increase prices of forest products. They are likely to be contested by the powerful political forces that control logging, ranching, plantations and agricultural expansion in rainforests. It may therefore be premature to expect deforestation to be significantly reversed in the short term under REDD.

It may be easier to manage the politics and economics of emissions from degradation (that is, the thinning out rather than clearance of forest) in the world's dry forests and savanna woodlands. This type of degradation results primarily from the exploitation of forest by local communities as part of their livelihood. It has been tackled successfully in Nepal, India and Tanzania, for example, under programmes that promote community forest management.

Dry forests do not have the international status of the majestic Amazonian and Congo forests, and the dry-forest degradation option is relatively neglected in REDD debates. Although the carbon content of dry forests is considerably lower per hectare, more of their area is degraded because they are more densely populated. Carbon losses may be more easily adressed because the commercial value of dry forests is lower, and their use is not so contested.

REDD should give more attention to dry forests. It should strengthen local communities' rights to manage the forests and to hold tenure there. Accounting the carbon savings would depend on proper estimates of local dry-forest degradation rates (which are virtually unknown at present) and on devising ways to monitor the gradual increase in carbon stock resulting from community forest management.