Kyoto: Think Global, Act Local (K:TGAL) - community forest monitoring in 30 locations around the world

Guinea Bissau, India, Mali, Mexico, Nepal, Papua New Guinea, Senegal and Tanzania
2003-2010
Key Lessons 
  • The K:TGAL project was a collaborative effort between international researchers, local and regional NGOs, and local communities, with the ultimate goal of testing whether communities could monitor their carbon stocks independently.
  • The use of personal digital assistants (PDAs) and mobile phones, with GIS and GPS capabilities and in-built monitoring software, allowed community members to master monitoring techniques after 4-6 days of training. This included illiterate people, and those with little formal education and no experience of computers.
  • At sites where data reliability was assessed, biomass estimates generated using community collected data were found to be as reliable as those of professionals. However, local support was important to help pick up on and address errors that sometimes occurred.
  • Community monitoring could be carried out for 30-50% of the cost of professional surveys (although costs for more remote communities could be higher than the potential revenue from carbon credits, posing challenges for national REDD+ systems).
Overview 

The ‘Kyoto: Think Global, Act Local’ (K:TGAL) project trialled community-based forest monitoring in 30 sites in 7 countries between 2003 and 2009. It developed and tested a methodology and survey tools that could be used by communities to create and maintain forest inventories, and aimed to maximise the capacity of communities to control the monitoring process once they had received adequate training and support.


Community members were trained and assessed for their ability to monitor changes in forest carbon stocks. This was part of a broader aim: to investigate the potential for community based forest management to be included under international climate policy as an activity that reduces or prevents emissions from deforestation and forest degradation.


The project findings concluded that paying communities to monitor their forests would be more straightforward and equitable than paying them according to the quantities of carbon stocks they maintained or enhanced. The latter could lead to inequalities (for example due to some forest types being faster growing), as well as perverse incentives. The project found that if agreements focused on monitoring, they could also include a commitment to manage forest sustainably, without explicitly linking management results to payments.


The project was led by the Twente Centre for (Studies in) Technology and Sustainable Development (CSTM), at the University of Twente, Netherlands, in cooperation with  the International Institute for Geo-Information Science and Earth Observation (ITC) in the Netherlands.  It was  implemented with four regional partners: Environment and Development Action in the Third World (ENDA) in Senegal; the International Centre for Integrated Mountain Development (ICIMOD) in Nepal; the Geography Department at the University of Dar es Salaamin Tanzania, followed by the Department of Forest Management and Mensuration at Sokoine University; and, in Papua New Guinea, Silvestrum, a Dutch-based organisation. 

International Forest Agenda/s 
Monitoring Theme/s 
Carbon biomass
Indicators 
  • Tree species
  • Diameter at breast height (DBH)
  • Tree height
  • Tree status (living/dead)
  • Geo-reference of trees and other features
  • GPS location and geo-reference of sampling plots
  • Delimitation of forest ecotype strata (zones)
Governance
Indicators 
  • Community forestry management approaches, rules and practices
  • Boundaries of community and forest areas intended for a carbon payments project
  • Community’s land claims
  • Community infrastructure and land-use plans
  • Conflicts (spatial information about land use, boundaries etc.)
Deforestation drivers
Land use change
Indicators 

Forest degradation drivers:

  • Location of activities contributing to forest degradation (e.g. illegal logging, grazing, marginal agriculture, illegal settlements, urban encroachment)
  • Location of areas potentially affected by hazards (e.g. fire, soil erosion, soil and ecosystem degradation, flooding, strong winds, etc.)

Policy context


When the K:TGAL project started in 2003, the Clean Development Mechanism (CDM) was the only climate change mitigation option related to tropical forests under the Kyoto Protocol. However, credits under the CDM were restricted to afforestation and reforestation activities on land that had not been forested since at least as early as 1990, to avoid inadvertently incentivising the removal of existing forest in order to gain CDM finance to reforest it.


This meant that the CDM did not address emissions from tropical forests; only the voluntary market was buying credits from avoided deforestation projects. K:TGAL aimed to respond to this gap in international climate policy.


From 2005, with the emergence of RED (soon to become REDD+) under the United Nations Framework Convention on Climate Change (UNFCCC), the idea of rewarding activities that maintained or enhanced carbon stocks in standing forests began to take hold. The preparatory text on REDD+ also acknowledged the need to fully engage indigenous and local communities in monitoring forest carbon stocks.


K:TGAL proceeded in parallel with these developments, working to support the aims of REDD+, as well as local people’s rights, by exploring how community based forest monitoring and management might become part of such a mechanism.


During the latter stages of the project, a strong effort was made to lobby for community forest management and community forest monitoring to be better integrated into draft policy texts under the UNFCCC, with some success.


Funding


The project was funded by the Netherlands Development Cooperation branch of the Dutch Ministry of Foreign Affairs. Funding was extended three times to allow the project to run from 2003-2009.


Community participation


The project aimed to promote the highest possible level of involvement by local people. It worked on the basis of growing evidence that communities can manage and improve the state of forests when provided with the means (and sometimes incentives) to do so. It also found that, for effective REDD+ implementation, community based monitoring was essential, because changes in carbon stocks due to better management would be too small to be accurately detected by remote sensing from satellites.


The project could be categorised according to the monitoring typology proposed by Danielsen et al (2009) as collaborative monitoring with external data interpretation, with the aim of achieving collaborative monitoring with local data interpretation.


K:TGAL research teams initiated projects across the four regions by engaging and collaborating with local or regional non-governmental organisations (NGOs) with connections to local communities that were already carrying out community based forest management. The NGOs then acted as support organisations during the project.


The community members who would be trained to carry out forest inventory work were identified by the NGO in consultation with community leaders. The age, gender and other characteristics of those who were selected varied according to the local context.


Training was given to provide the community members with the technical skills to monitor their forests for carbon, and with an understanding of the REDD+ policy context including the theoretical possibility of financial rewards, while trying to avoid raising false hope of guaranteed income.


The villagers had no experience of computers and limited previous formal education, having attended primary school at most, often for less than seven years. Therefore, in addition to training, outside experts were needed for the technical set-up and for ongoing support. For example, an external technician was required to carry out the initial programming of the mobile equipment, and training and supervision were needed from people with experience in statistical sampling, in laying out permanent sample plots, and in using computer technology. Professionals were also needed to prepare the allometric equations to convert community measurements into biomass estimates.


All stages of the monitoring were then carried out by the community members, with the assistance of the external trainers. The methodology, training, and field guide were designed to ensure that the communities would eventually be able to take over all of the activities, including the technical tasks.


Methodology  


The methodology developed for the project for creating forest inventories was based on the Good Practice Guidelines for Land Use, Land-Use Changes and Forestry (IPCC, 2003) and other relevant literature.


Two levels of information gathering were used. The first was intended to establish the initial forest management scenario - ‘Year 0’ - and consisted of information at the community scale. This included information such as the boundaries of community land; land-use plans; drivers of forest degradation; and different forestry management practices.


Once identified, the boundaries and locations of these features and activities were mapped by the community with the trainer team. This was first done by sketching the features onto paper without using a base map, in order to understand the spatial concepts used by the community. Then, if boundaries were not already marked on formal maps, this was done using PDAs or smartphones with GIS and GPS, on a base map or geo-referenced image, e.g. using Google Earth.


The second level of information consisted of more detailed information to create the forest biomass inventories for each year. This information was gathered at plot level (e.g. the location of the sampling plots) and tree level (e.g. tree height, diameter at breast height, or DBH, height and status – living or dead). Tree species and dimensions were surveyed using hypsometers and tree tapes and/or callipers, with measurements being recorded onto the PDAs or smartphones and/or on paper.


The number of plots needed was calculated using the standard error found in the pilot measurements. Permanent plots were then set out on the basemap, distributed systematically within the different strata, on a transect framework with a random start point. The sampling plots were then marked in the forest, with the centres of the plots being staked out using a compass and measuring tape, and marked using paint or plaques. Identification codes for each plot were recorded using the PDA or smartphone. The species, DBH and height of trees in the plots were recorded. Tree and sapling stems were measured, as were bushes, shrubs, herbs and litter.


Allometric equations were used to calculate carbon stock from tree volume. Local allometric equations were used where they were available; otherwise default equations were used.


A database was created for maintaining the information, and monitoring plans were prepared for periodic monitoring of sample points. In many of the project locations, control sites were also identified, to assess the rate of change of carbon stock under unmanaged conditions.


Digital technology


Most research sites used handheld personal digital assistant (PDA) devices with GPS and GIS capabilities. The use of mobile GIS was useful in allowing maps and images to be viewed, and their attributes to be changed, in the field. Its greatest advantage was the ease with which data could be extracted, and therefore the potential for this data to be easily analysed and reported for carbon crediting. The GIS and GPS capabilities were valuable for monitoring carbon as well as mapping forest boundaries and stratification.


Data was stored using a MS Access database, and entered using Arcpad It was found that most community members who had basic numeracy and literacy skills, could use the handheld devices and GPS functions after brief training.


(Two projects in Mexico, which were established later, separately from K:TGAL, trialled newer smartphone devices. These had the added advantages of being simpler to use than PDAs, having greater storage capacity and web accessibility, and having built-in GPS, camera and video functions, reducing the amount of equipment that needed to be bought, carried and maintained.)


Accuracy of data


The reliability of community collected data was assessed for two sites in India and Tanzania. The results indicated that biomass estimates generated using community collected data were as reliable as those of professionals (i.e. the results were not significantly different).


Occasionally, there were problems with data reliability. For example, in one site in Nepal, it was found that community monitors had not completely understood how to use the GPS equipment, and this led to errors when they worked without supervision, with measurements being taken from different plots every year instead of being taken repeatedly from the sample plots.


This kind of error was rare, but highlighted the need for support from a local agency, to identify and address problems.


location forest type number of plots statistic above-ground biomass estimates from community measurements (tC/ha) Above-ground biomass estimates from professional measurements (tc/ha)  P
Dhaili, India Stratum 1: even-aged oak forest 14

Mean


SE

453.3


36.7

426.4


36.6

0.61

Dhaili, India

Stratum 2: dense oak forest 15

Mean


SE

283.4


40

279.9


40.5

0.95

Dhaili, India

Stratum 3: degraded oak 7

Mean


SE

41.7


4.6

38.1


3.7

0.55
Tanzania, Kitulangalo Savanna woodland (miombo) 89

Mean


SE

43.2


1.9

42.2


4.4

0.83

Table adapted from Margaret Skutch (ed.) 2011. Community Forest Monitoring for the Carbon Market: Opportunities Under REDD+. Earthscan: London.


Achievements and challenges


The community members involved in the project easily picked up the skills to carry out the measuring and mapping, and the accuracy of their results compared favourably to that of professional monitors.


The project found that the amount of carbon found to be saved under community management across the 30 sites ranged from an average of 5.4 tonnes per hectare per year of CO2 in dry savannah forests, to 7.2 tonnes per year in the temperate Himalayan sites, and to 21.24 tonnes in rainforests.


Estimates were made of the cost of monitoring for some of the sites. For two sites, these costs were compared with the cost of carrying out independent, professional inventories. It was found that community surveys (including all associated costs such as training) could be carried out for 30-50% of the cost of professional surveys.


However, in the context of a national REDD+ payment system, it was suggested that costs could be problematic for communities that are more expensive for external partners to access (for example if the support agency requires flights in order to reach remote areas). Variations in these costs, combined with variations in tree growth rates, would have implications for cost effectiveness of monitoring and managing different areas. This would present challenges for achieving emission reductions efficiently while also ensuring equity among communities.


Another potential challenge was found in the fact that the digital equipment used in the project could be prohibitively expensive for communities to buy and maintain. This could be solved through arrangements to borrow the equipment instead, for example from a local NGO.


A field guide was produced, which was found to be useful in the different countries, recognising that some deviation may be necessary depending on the local context. It provided information for communities, trainers and policy makers, and was an essential tool for enabling the communities to become independent in the long term. The project also developed a dedicated Tropical Forest Inventory Data Analysis (TROFIDA) package, for ease of data analysis. This was made publicly available for use by other practitioners.


Finally, it was found that soil carbon was important to consider, particularly in dry and temperate forests with high levels of carbon in the soil and litter layers, but that it would be difficult to measure and to incorporate in a crediting system, partly because measurement would require additional laboratory processes to be performed externally.