Reducing emissions from deforestation and forest degradation (REDD+) requires developing countries to quantify greenhouse gas emissions and removals from forests in a manner that is robust, transparent, and as accurate as possible. Although shifting cultivation is a dominant practice in several developing countries, there is still very limited information available on how to monitor this land-use practice for REDD+ as little is known about the areas of shifting cultivation or the net carbon balance. In this study, we propose and test a methodology to monitor the effect of the shifting cultivation on above-ground carbon stocks. We combine multiyear remote sensing information, taken from a 12-year period, with an in-depth community forest carbon stock inventory in Palo Seco Forest Reserve, western Panama. Using remote sensing, we were able to separate four forest classes expressing different forest-use intensity and time-since-intervention, which demonstrate expected trends in above-ground carbon stocks. The addition of different interventions observed over time is shown to be a good predictor, with remote sensing variables explaining 64.2% of the variation in forest carbon stocks in cultivated landscapes. Multitemporal and multispectral medium-resolution satellite imagery is shown to be adequate for tracking land-use dynamics of the agriculture-fallow cycle. The results also indicate that, over time, shifting cultivation has a transitory effect on forest carbon stocks in the study area. This is due to the rapid recovery of forest carbon stocks, which results in limited net emissions. Finally, community participation yielded important additional benefits to measuring carbon stocks, including transparency and the valorization of local knowledge for biodiversity monitoring. Our study provides important inputs regarding shifting cultivation, which should be taken into consideration when national forest monitoring systems are created, given the context of REDD+ safeguards.
