Assessing the accuracy and cost-efficacy of community-based monitoring for REDD+
- A simple methodology enabled community monitors with limited training to measure carbon stocks in complex tropical forests.
- Community monitoring of carbon stocks is accurate in relation to monitoring by professional foresters, and accuracy increases with repeated rounds of measurement.
- Community monitoring of carbon stocks is cost-effective in comparison to monitoring by professional foresters, and cost-efficiency increases over time.
Community involvement and participatory monitoring have been advocated as important sources of information for REDD+. Nonetheless, questions persist over the cost-effectiveness and accuracy of community-based monitoring in comparison to professional monitoring. Under the EU-funded I-REDD project a consortium of organisations led by Nordic Foundation for Development and Ecology tested the cost-effectiveness and accuracy of data collected by comunities for REDD+, in complex tropical forests in South-East Asia. The study compared the measurements of carbon biomass by community members and professional foresters.
The set of decisions reached at the UNFCCC Conference of the Parties in 2010 in Cancun, Mexico 2010, include the REDD+ safeguards and stress the importance of, “the full and effective participation of relevant stakeholders, in particular indigenous peoples and local communities” (Decision 12/Cp.17) within REDD+. Community-based monitoring for REDD+ is also mentioned in a number of national REDD+ strategies.
Despite this, the development of measuring, reporting and verification (MRV) and monitoring systems for REDD+ has so far largely focussed on remote sensing and ground based inventories by professional foresters. In contrast, more participatory forest monitoring systems have largely been overlooked. However, several studies have demonstrated that community collected data can be more cost-effective than monitoring by professional foresters (Danielsen et al. 2011; Skutsch et al. 2011). Community-based monitoring can enable real-time reporting of changes on a regular basis, which is important where local drivers of deforestation dominate but are difficult to identify through remote sensing. Community involvement can also help track the progress of reforestation and regeneration efforts.
The community monitors participating in the I-REDD project were identified by local community representatives as having a special interest or skills in forest monitoring. They were paid for their participation, receiving wages ranging from 5 to 16 USD per day across the nine sites in Year 1.
The project compared measurements by community monitors, who had all received primary education, with measurements by professional foresters, with academic degrees in natural sciences and an average of four years' work experience on forest assessments (Danielsen et al. 2013).
Measurements were made in 289 plots across nine sites in China (two forest types), Indonesia (one forest type), Laos (two forest types) and Vietnam (three forest types at Moi village and one forest type at Diem village).
Two rounds of measurements were undertaken, with a gap of 18 months, to assess changes in the accuracy and cost-effectiveness of community monitoring over time (Brofeldt et al. 2014).
The community monitors received one to two days’ training before undertaking monitoring in Year 1 and a further one day refresher course before monitoring in Year 2. In each team in Year 2 there were at least two monitors who participated in Year 1. The monitors were supervised during the process of locating plots and had a simple check-list of tasks in the form of a basic manual.
Above-ground biomass was measured with a simplified version of radial nested sampling, using 9 m and 15 m radii circles. The only equipment used was a GPS device, a measuring a tape, a rope for measuring plots, paint, field forms, and pencils. In all sites the number of trees in each plot (the indicator for plot demarcation accuracy) and the girth measurement (a proxy for Diameter at Breast Height) were recorded. In six sites in Year 1, and in all nine sites in Year 2, the trees were numbered to enable direct comparison of the girth measurements of each tree by community members and professionals.
The number of tree species in each plot was recorded.
Accuracy of data
Above-ground biomass measurements differed only slightly between community members and professionals in Year 1. While this difference was significant in one third of the sites, there was only a significant difference in the variance of biomass estimates in one of the nine sites (Danielsen et al. 2013). In Year 2, the difference in measurements between the community monitors and professionals declined, with only one of the nine sites showing a statistical difference (Brofeldt et al. 2014).
The number of trees recorded in each plot in both years was only significantly different in one site. In Year 1 there was greater agreement on the number of trees in each plot in the six sites where trees were numbered.
A significant difference in girth measurements was found in three of the six sites where trees were numbered in Year 1, and where a direct comparison between community and professional data was possible. Further analysis revealed that while the distribution of differences in girth measurements was slightly skewed, with community members consistently measuring slightly lower girths than professionals, there was a high level of agreement for most girth measurements. There was a substantial increase in accuracy in five of these six sites in Year 2, and no significant difference was found between community and professional measurements in the three sites that were numbered for the first time in Year 2.
This data demonstrates that community-collected data of carbon stocks, even with limited training, can be accurate in relation to data collected by professional foresters. It also indicates that accuracy increases over time as the community monitors’ skills improve. Where there are differences, these can be easily addressed with more training and practice.
The cost-effectiveness of community-based monitoring was assessed by calculating the costs incurred at each site for both communities and professional monitoring. Costs were calculated for the five sites in Indonesia, China, Laos, and Diem and Moi villages in Vietnam. The costs were classified as wages, travel, accommodation, and training. The cost of community monitoring was consistently higher than the average cost of monitoring by professionals in each of the five sites in Year 1. While wages, travel, and accommodation costs were lower for community members, these were outweighed by training costs.
However, in Year 2, community-based monitoring became more cost-effective than professional monitoring in four of the five sites. This was due to falling training costs, as skills increased, and the time required also decreased once paths were established, making it esier to find and move to plots. For both communities and professionals, monitoring was more cost-effective across larger areas.
Achievements and challenges
On the basis of data from 289 forest vegetation plots in four countries, this study demonstrated that community-based monitoring of carbon stocks can be accurate and cost-effective in relation to monitoring by professional foresters, even with very little training. Furthermore, accuracy and cost-effectiveness was shown to increase over time as skills increased and less training was required.
The use of a very simple methodology enabled community members to successfully monitor carbon stocks with very limited training. However, in the context of providing carbon stock data to national forest inventories, more complex protocols will likely need to be used by a coordinated team that can consistently replicate these protocols.