Participatory monitoring by Project Fauna
- Indigenous monitors collected a huge amount of high-quality data - the 335 monitors in Guyana and the partner activities in Brazil walked a total of 43,122 km of transects (a distance greater than the circumference of the world). The project generated nearly 50,000 wildlife sightings.
- Community data collection is generally highly reliable. The project analysed the factors affecting data quality, finding that this can be compromised when the motivation for participation is primarily financial, or when collectors do not have regular contact with broader governance structures.
- Data sharing, through community atlases tailored to each community, promoted understanding of the benefits of data collection and greater ownership of outcomes.
- Project participation provides various skills, including communication skills and scientific literacy which some community members valued to help them gain employment in the future.
Project Fauna was a collaborative project between indigenous communities in the Rupununi area of Southern Guyana and researchers at Stanford University and the State University of New York-ESF. Indigenous people of Raposa/Serra do Sol region of Roraima, Brazil, also took part in the project, though this case study focuses on the work in Guyana.
The study area included roughly 40 communities, ranging in population from 60 to 1200 people and predominantly located in seasonally-flooded lowland savannah within 5km of the edge of the forest. Of these, 23 communities took part in the monitoring project. The study area also incorporated the Iwokrama Forest Reserve, and most of the territory occupied by the Cariban-speaking Makushi and the Arawakan-speaking Wapishana people. Fishing, hunting and small scale agriculture are key livelihood activities.
The initial aim was to explore the relationships among indigenous culture, economic integration, hunting and wildlife dynamics. The collaborative and interdisciplinary research this was later expanded to include tree, soil and litter carbon stock assessments. Over three years, 335 indigenous data collectors were trained in demographic and socio-economic surveys, as well as various field ecology methods and technologies (e.g. setting transects, recording wildlife sightings, GPS, compass, clinometer, DBH and tape measurements). They collected scientific data on 300 vertebrate species, vegetation types, fruit production and hunting intensities.
The project provided insights into the potential for large-scale community-based monitoring programmes, with results showing that data collection by local people can be an efficient way to gather high quality data on a range of variables at this large scale. This can benefit both local people and the wider scientific and policy communities.
Returning primary data to the participants in the accessible format of 'Community Atlases' was a key part of the project. These provided the communities with information on the use and distribution of resources across their territories. This created an incentive to collect high-quality data, as the data were perceived to be valuable in informing hunting practices, territorial management plans and making the case for land title extensions. The long term employment and skills that could be used in future job applications were also motivating factors for communities and individuals who took part.
In addition, the data could potentially be used to inform carbon or biodiversity analyses as part of REDD+ or the CBD. Shared benefits of data collection are extremely important in scaling up monitoring, reporting and verification (MRV) schemes for REDD+, where data forgery is a risk due to financial incentives to overestimate carbon stocks.
Guyana is actively pursuing a REDD+ agenda as part of its Low Carbon Development Strategy. Monitoring, reporting and verification (MRV) of remotely sensed data is essential to ensure the accuracy of data used to underpin REDD+ payments. This is described in more detail in the policy context of another case study from Guyana, here.
Having trained community monitors on the ground ensures that REDD+ is applied in a way that is appropriate to local needs and that knowledge is shared by the community. However, a drawback to carbon-stock data is that the quality and accuracy can be compromised due to incentives to overestimate amounts, as well as differences in how stocks are measured. This project showed that even in a situation where incentives existed for falsification, data quality and reliability were high.
Community participation was a central goal of the project, with 335 indigenous community members trained to collect data. A key principle was for the project to generate data that the communities could then use in whatever way they felt appropriate. Difficulties arose in initial stages due to confusion over the goals of the project, with most having no experience with projects for the sake of knowledge and without the objective of creating a specific change. Once the aims were clarified, the community response was positive.
Although outside researchers collated and analysed the data, results were distributed in Community Atlases, with one for each community, showing their unique data. These atlases provide regional and community-specific information, often in map format, such as the importance of wildlife in their diet, numbers of animals per species annually consumed, carbon stocks per hectare, the kill sites of most hunted species, the location of important spiritual sites, and results of the socioeconomic surveys.
Journal publications were all distributed to the Amerindian organisations and leaders in the region as well as the Guyanan government. In addition, each community with titled land was informed about the size of its forest area and total carbon stock through the distribution of official ‘community carbon leaflets’. These were accompanied by presentations to explain in layman terms about CO2, climate change, REDD+ and the Opt-In Mechanism.
Twenty-three villages were involved in the study. Biodiversity and socioeconomic data were recorded in all villages, including through interviews. Carbon stock data was recorded in 17. Five sites were selected as controls: these were areas that were not used for hunting, logging or collecting, and were situated 15-40 kilometers away from villages.
Eight transects were established at each village study site. Four of these transects were established within 6km of the village, and four were set up 6-12km from the village centre. This created near and far 'buffer zones', to test for changes in variables (such as wildlife sightings) with distance from the village. The starting point and bearing for each transect was selected randomly, within the near or far buffer zone. The trained community monitors used GPS, compass and tape measures to cut a 4km transect straight through the forest. (See the Community Atlases for maps of the study area, including the buffer zones and transects, detailed results, and photographs.)
The information from the community monitors was combined with remote sensing data on forest cover and collated into GIS layers.
Twice each month, two or four community monitors walked the eight transects that had been established in buffer zones around the village centre or the central point of control plots. They recorded the date, time, location and species of signs or sightings on data sheets. In Guyana the project recorded information for about 300 vertebrate species, with about 48,099 wildlife sightings along 21,729 km of transect walked, and 84,028 records of animal signs along 21,393 km walked.
Plots were placed across a stratified random sample of land-cover types (high forest - flooded and upland, low forest - flooded and upland, savanna - flooded and upland, Ite swamp and Muri upland shrub). Carbon biomass was sampled in line with common protocols (such as GEM-RAINFOR). The diameter of all trees over 10cm were measured at breast height (DBH) each month. Height measurements with clinometer were found too variable and discarded. Equations that relate DBH to overall tree size were then used to calculate overall biomass, and halved to give the carbon value.
Hunting and spiritual sites
Once every week, each household was asked to mark sites of successful kills on a map and to record the species. These records were digitised and the kill sites of the five most hunted species were represented on maps that were given to the community as part of the community atlases.
The project identified a large number of spiritual sites, which are areas that hunters avoid or where they use extra caution. This data was gathered using one-time surveys with people that the communities identified as principal hunters. Knowledgeable community members (such as elders, the toshao - community leader - and shaman) were also interviewed about places considered to be dangerous or sacred. The proximity of kill sites to spiritual sites was measured to determine whether hunters avoid these sites when hunting.
Data was collected using surveys, which covered language proficiency (in Makushi, Wai Wai, Wapishana, English and Portuguese); number of males and females; main sources of protein; religious affiliation; cattle ownership; income and migration. Most of these interviews were conducted at the household level. The hunting and socio-economic data were collected from 9523 people in 24 villages in Guyana.
Achievements and challenges
The project showed that residents of forest communities can collect a huge body of data on a range of variables, working across large scales and across borders (although this case study focuses on the activity in Guyana, the project also operated in neighbouring Roraima, Brazil).
The project created a substantial dataset for the scientific community. For example, analysis of the data on hunting and spiritual sites suggest that hunters do deliberately avoid these sites; the researchers proposed that resource management should take account of social and cultural factors as well as environmental ones, as these spiritual sites could have particular conservation value, as source sites for wildlife populations.
In addition, the collection of data by trained community members provided local benefits. These include employment and skills, not only in terms of data collection, but also communication skills as the monitors reported regularly to their communities and village councils. Five of the indigenous assistants received training in managing scientific datsets. Communities were particularly interested in the data for the knowledge it could bring on hunting locations as well as making the case for land title extensions.
The project demonstrates the benefits of project co-management: the community monitors formed a two-way communication channel that facilitated a better understanding and trust in the project. They provided information to project managers on how the project was perceived throughout its duration, enabling the early signalling of problems that could then be addressed. The project managers report that the community monitors can create bridges between local and scientific perspectives on wildlife management.
The project took place in areas where at least five languages were spoken, suggesting that the model could be scaled up across other remote areas in the Amazon with potential language barriers. Using local monitors overcomes these barriers, providing better data as interviewees can express themselves accurately in their own language.
However, challenges were encountered in ensuring data quality, with some cases of forgery or general unreliability. The main factors that influenced the quality of data was individuals’ and communities’ motivation to participate in a project; whether there were effective village-level governance mechanisms; and whether the community was affiliated with local umbrella organisations. When motivations were mainly financial and local governance was weak, data were more likely to be low-quality. Identifying these factors helps to inform future projects; it was recommended that participants’ motivation should be identified early on, and that there should be regular contact between data collectors and other project members.
An additional issue that was encountered, but quickly addressed, was the fact that data collection was a full time commitment, leaving no time to complete household, farming or community commitments. In subsistence-oriented economies where households are responsible for growing or catching their food and completing repairs on their homes, significant time needs to be given over to these tasks. Once this was realised, substitutes were trained for times when one or both main data collectors had other commitments.
Information for this case study came from multiple sources, including:
Butt N, Epps K, Overman H, Iwamura T, & Fragoso J (2015) Assessing carbon stocks using indigenous peoples’ field measurements in Amazonian Guyana. For. Ecol. & Management 338: 191-99. dx.doi.org/10.1016/j.foreco.2014.11.014