Tropical forests are home to a large proportion of the global terrestrial biodiversity, help regulate the worlds’ climate and are essential for the livelihoods of millions of people across the tropics. However, it is estimated that the global forest cover has decreased with approximately 129 million ha between 1990 and 2015. Since COP 14 in Poznan in December 2008, reduced emissions from deforestation and forest degradation (REDD) have become an essential part of the global strategy to mitigate climate change and after almost ten years of negotiations, REDD+ (which also includes the enhancement of existing carbon stocks) is expected to come into effect as part of the Paris agreement. A central strategy of REDD+ is the inclusion of local and indigenous peoples, and their indigenous knowledge, in the monitoring of carbon stocks, emission activities and biodiversity and livelihood safeguards. This has made the potential of community based monitoring (CBM) to inform national forest monitoring systems an important and active field of study. Numerous studies have investigated data quality, costs and motivation in CBM programs but a number of central aspects relating to the use of CBM in REDD+ remain unexplored: i) Monitoring for REDD+ will require repeated measurement over time but no studies have assessed the development in community capacity to undertake monitoring of carbon stocks and emission activities using repeated measurements. ii) CBM of biodiversity will require local people to apply their local ecological knowledge (LEK) to identify tree species, but no studies have addressed how the distribution of LEK in local communities and the reliance on key informants, may affect the quality of local species identifications. iii) The monitoring of emission activities is a priority in REDD+ and the use of emerging tools like information and communication technology (ICT) has the potential to greatly strengthen the documentation of monitored resources and activities. However, no studies have investigated the capacity of local communities to use ICT to produce documentation of emission activities. iv) Incentives for local communities to engage in CBM for REDD+ are generally based on payment or on decentralization of forest management and promotion and enforcement of local forest rights. Increased decentralization of forest management and promotion of local forest rights have been found to increase local communities’ benefits from forests and local feelings of ownership, leading to strengthened biodiversity protection and local level monitoring of forest resources. However, most existing studies of CBM for REDD+ have focused on CBM projects where incentives have been based on payment. These are all important considerations for the future of REDD+, especially as the signing of the Paris agreement will require CBM to be implemented at large spatial and temporal scales as part of national forest monitoring systems. The research in this PhD thesis seeks to inform the future implementation of CBM as part of national REDD+ activities by presenting five key papers based on field research carried out in eight sites across Southeast Asia, as part of three separate projects. The papers address four overall objectives by providing empirical knowledge on: i) the quality of data produced by community based monitoring of carbon stocks, biodiversity safeguards and emission activities (papers 1-4); ii) the costs of involving local communities in monitoring of carbon stocks, biodiversity safeguards and emission activities (papers 1, 2 and 4); iii) the capacity of local communities to be highly involved in monitoring using ICT (paper 4) and iv) the motivations of communities to engage in CBM on a voluntary basis (paper 5). Papers 1 and 2 are based on two consecutive years of field research in China, Vietnam, Laos and Indonesia. Paper 1 investigates the quality and cost of CBM of above ground carbon stocks. Paper 2 explores the quality and cost of community- and scientist-led tree species identification, which can be used for both carbon stock- and biodiversity monitoring. Papers 1 and 2 provide evidence that data on carbon stocks and tree species rarely differ significantly from data collected by professional foresters and botanists. The results also show that the cost of the monitoring is consistently lower than the cost of the monitoring done by professionals. The results of paper 1 further show that the correspondence between carbon stock measurements done by community monitors and professional foresters increase, while the cost decreases, from the first to the second year of monitoring. Paper 3 is based on field research in two sites in Indonesia and Vietnam respectively. It examines how different levels of LEK in the two sites affect the quality of tree species identifications. The paper documents a notable difference in the quality of identifications in the two sites and that the use of key informants in both sites does not fully mitigate these site-level differences. The results also show that the age and level of active forest use of the informants influence the identification rate, consistency and level of detail of the provided identifications. Papers 4 and 5 are based on applied research in Cambodia within a community led forest monitoring project. Paper 4 investigates the use of ICT methods for CBM of forest resources and logging activities and the related implications for data quality, cost and the long-term sustainability of the monitoring program. The paper provides evidence that community monitors are able to provide large amounts of valid documentation of illegal logging activities and important forest resources. They are able to do this at a cost that is lower than the cost of professionally operated monitoring programs, although using ICT has made the monitoring program more expensive than CBM programs using simpler methods. The results further indicate that the complexity of ICT tools do not negatively impact community monitors ability to use them correctly if adequate training and supervision is undertaken. Finally, the results suggest that the inherent issues related to mobile network connectivity when using ICT in remote areas is a significantly stronger constraint to the production of valid data, than the capacity of local communities. Paper 5 focusses on the motivations of local communities to engage in voluntary forest monitoring within a setting featuring a high level of conflict. It presents evidence that community monitors are characterised by being active forest-users motivated to engage in monitoring activities by desires to protect the forest against illegal logging in order to preserve their own lands and forest uses. It further documents that the community monitors feel that their monitoring is successfully stopping the logging, and that this is a major contributing factor to their continued engagement in the monitoring program. The findings presented in this PhD show that communities can collect data on carbon stocks, biodiversity and emission activities of a quality that is strong enough to inform results based payment for REDD+ and document the implementation of REDD+ biodiversity safeguards. The cost of employing communities in this role is significantly lower than the cost of employing professional foresters and botanists and the costs of operating CBM programs will decrease over time, as the experience of community monitors build. This is a strong argument for future implementers of REDD+ to consider CBM a feasible way to meet the large monitoring demands that will be associated with national and international scale REDD+. The complexity of ICT does not necessarily negatively affect the capacity of community monitors to produce valid information using it, but CBM using ICT is more expensive than CBM using simpler methods. As ICT necessitates professional assistance for data management it is most feasible as a way to standardize data collection for national and international scale REDD+, where such costs can be reduced through standardization and cost sharing. Providing incentives for communities to engage in CBM for REDD+ by promoting and enforcing local forest rights also has the potential to decrease the overall cost of monitoring significantly but this will require that monitoring objectives are relevant to local concerns of forest protection and forest rights. Designing monitoring systems that can provide such incentives will be a major challenge for future REDD+ implementers. Finally, the findings show that the level of forest dependence also influences the distribution of the LEK required to make species identifications in CBM of biodiversity. Differences in the level of LEK between individuals and villages make it important to consider how participants in monitoring for REDD+ are selected – especially when addressing monitoring objectives such as biodiversity, that require accurate species identifications. This will become increasingly important for REDD+ implementers to consider, as social development and changing availability of natural resources continue to affect the level of LEK in local communities.