"This has clear implications for equity by potentially reducing local rights to forests," Ted Webb of the National University of Singapore told environmentalresearchweb. "It may also have important long-term implications for conservation because large-scale, centralized management has limitations to effectiveness and efficiency."

Together with colleagues Jacob Phelps of the National University of Singapore and Arun Agrawal of the University of Michigan, US, Webb analysed the strengths of both centralized and decentralized forest- – and carbon- – protection systems.

"We found that nobody had really addressed the obvious tensions between the fact that massively funded REDD+ programmes focusing on forest conservation appeared to require heavy central government control, and that decentralization of forest governance represented the relinquishing of control by the governments to local levels," he said. "Emissions reductions can be achieved with protected areas but they are also needed outside protected areas, where decentralized management has been extensive in many countries."

Although centralized governance can protect forests and enhance regrowth, it's costly to enforce, limited to within-park boundaries and can cause resentment in excluded users, say the researchers. But a national approach appears essential for REDD+ schemes to enable reliable monitoring and reporting for carbon-accounting purposes and to ensure permanence.

To get the best of both worlds, Webb and colleagues propose the use of "hybrid" governance systems. They say that these are contingent on "incentives and/or mandates for governments to incorporate decentralized strategies and promote community participation, while avoiding heavy recentralization; and research on REDD+ governance – particularly how to integrate decentralized strategies within a system that requires some level of centralized oversight".

So far, six developed countries have committed to investing $4.5 bn in REDD+ schemes in developing countries by 2012. Some believe that investment in the schemes may reach $30 bn a year by 2020.

"By articulating the ways in which REDD+ could change government incentive structures regarding forest governance, we help position policymakers to create a mechanism that ensures decentralized structures are incorporated into REDD+," said Webb.

The researchers believe that the REDD+ pilot projects currently under development need robust research agendas. "We must better understand trade-offs and synergies between rural livelihood activities, alternative land uses and REDD+ goals," they write in Science. "Research should also focus on how carbon sequestration varies with differing levels of community engagement and autonomy, and across diverse tenure regimes." Studying the practicality and acceptability to international markets of low-cost community-based monitoring, reporting and verification strategies would also be useful, they say.

But the urgency of reducing emissions means that it's unlikely that pilot studies will have time to reach their conclusions before global-scale implementation of REDD+ takes place.