Agent-based modeling of hunting and subsistence agriculture on indigenous lands: Understanding interactions between social and ecological systems
Graphical abstract
Introduction
Much of the Amazon Basin's forest cover and biodiversity is found in areas inhabited by indigenous people (Nepstad et al., 2006, Schwartzman et al., 2000). Modeling the resource and land use dynamics of these people will help us understand their contribution to Amazonian ecosystems, and may also assist indigenous peoples in their design of sustainable management and livelihood plans. Previous efforts to model the sustainability of indigenous people's resource use activities have typically focused on a few aspects of their livelihoods, and often represented the scope of a particular discipline. Some have primarily investigated the sustainability of hunting (Bodmer et al., 1997, Fragoso et al., 2000, Hill et al., 2003, Silvius et al., 2004, Sirén et al., 2004, Damania et al., 2005, Levi et al., 2009), while others have been more interested in land use decisions and patterns of land cover change (Wilkie and Finn, 1988, Brondizio et al., 1994, Deadman et al., 2004, Nepstad et al., 2006, Sirén, 2007). Sustainability, however, hinges on the feedbacks and balances between social and ecological systems, and should be investigated with a holistic framework (Ostrom, 2007). For example, habitat fragmentation can cause the sudden decline of animal abundance around villages, and lead to agricultural expansion to compensate for food loss due to unsuccessful hunting (Bennett, 2002, Damania et al., 2005). A few studies focus on clarifying the connection between hunting and deforestation (e.g, Wilkie et al., 1998), but are unable to incorporate feedbacks between these activities and natural resources due to the dearth of detailed datasets needed to develop and calibrate complex models.
This paper describes our modeling effort for the indigenous communities in the Rupununi region of Guyana, in the Guiana Shield region of the Amazon, where we collected an extensive dataset on social and ecological aspects of the lives, land, biodiversity, and environment of the Makushi and Wapishana people (Luzar et al., 2011). We apply a holistic framework to investigate human-environment interactions, which is under the rubric of social-ecological systems (SESs) (Ostrom, 2009) or coupled human-natural systems (Liu et al., 2007). Our objective is to develop a model to investigate the conditions under which indigenous communities relying on hunting and subsistence agriculture alter their impacts on an ecological system through land use change.
The Rupununi region provides a unique setting in which to investigate such research questions. Its ancient geological history (1.7 billion years) and mixed upland and lowland savannas, forests and wetlands have resulted in a highly diverse fauna and flora (Hammond, 2005). The region supports approximately 20,000 predominantly Makushi and Wapishana indigenous people (Hammond, 2005, Luzar et al., 2012). The indigenous communities remain largely isolated due to their remoteness from Guyanese population centers on coastal areas, to which they are poorly connected by an inadequately maintained dirt road and scattered airstrips. As is the case in other parts of Latin America (Geist and Lambin, 2002, Lambin et al., 2003), impact assessments for the region (Conservation International et al., 2009) find that government-led infrastructure establishment is expected to trigger deforestation in the Rupununi. Land outside of demarcated indigenous territory in this region is federal state land subject to long term leasing including by foreign industries, which makes public land use policies very important in driving land use change.
Our spatially-explicit household simulation model is meant to analyze the feedback between human activities and natural resource systems by using agent-based modeling (ABM) (Gilbert and Troitzsch, 2005). ABM is a bottom up approach to model decision making by individual “agents” to explain macro-scale behaviors (Parker et al., 2003, Brown et al., 2005, Filatova et al., 2013). It is widely used to simulate social behavior, including land use change based on household-level data (Deadman and Gimblett, 1994, An et al., 2001, Evans and Kelley, 2004, Schreinemachers and Berger, 2011). It has been applied in sectors for which a large amount of data is available, such as transportation (Bonabeau, 2002), insurance (LeBaron, 2006), mobile telecommunication (Fricke et al., 2001), health (Lambin et al., 2010, Linard et al., 2009) and computational social networks (Bonabeau, 2002).
Our model takes advantage of the accumulation of ABM literature on household decision making for subsistence agriculture. Deadman et al. (2004) describes land use change among colonist farmers in the Brazilian Amazon based on satellite images and interviews with local people using spatially explicit ABM. An et al. (2005) investigates the dynamism of demographic changes and their impacts on deforestation in giant panda habitats in China. Entwisle et al. (2008) applies ABM to study the interactions between demographic and land use changes in Thailand. Evans et al. (2011) investigates the agricultural transition from subsistence agriculture to rubber production in Laos. Walsh et al. (2013) combines an ABM (Entwisle et al., 2007) with ecological modeling to identify suitable areas for future agricultural expansion. ABM has also been applied to represent decision making by indigenous people regarding land use (Lim et al., 2002, Berman et al., 2004, Deadman et al., 2004, Cabrera et al., 2012). To our knowledge, only Berman et al. (2004) has modeled both hunting and other activities in indigenous communities. Their study in arctic Canada uses an extensive empirical social data set; however, it does not address changes in animal abundance, distribution and diversity.
Our simulation model presents a more holistic framework incorporating indigenous hunting and agricultural activities as well as changes in demography, land cover, and animal abundance, distribution and diversity. It thus contributes to the study of sustainability of indigenous communities and their environments by providing a tool to investigate complex interactions and feedbacks between human and natural systems. For example, a greater involvement in agriculture activities is explicitly modeled as the result of the decision making of each household managing their energy budgets. Efficient energy intake can be achieved through a mix of hunting and agricultural activities. Successful energy intake for each household results in a demographic increase in a village, which then affects animal abundance, distribution and diversity through wildlife meta-population dynamics, and vegetation succession.
The major constraint to the application of ABM to socio-ecological systems has been insufficient empirical information to parameterize real-world complexity (Robinson et al., 2007; Windrum et al., 2007, Filatova et al., 2013). In our study, we use a wide range of data on indigenous peoples and communities of the Rupununi (Luzar et al., 2011) to parameterize our simulation model, including interview-based surveys of demographic and socio-economic characteristics of all households within the study area, along with field data on animal kill locations, and the diversity, distribution and abundance of animal species (Luzar et al., 2011, Read et al., 2010, Luzar and Fragoso, 2013, Fragoso et al., 2010), and a time series of land cover change from satellite images. The limited connectivity to the outside world of villages means that the entire system can be modeled based on local factors. This enables us to circumvent a “boundary problem” commonly found in other studies where the extent of the modeling effort is arbitrarily defined (Meadows, 2008).
This paper describes the model based on a spatially-explicit ABM to understand interactions between indigenous people and their natural environment in Guyana's Amazon. We seek in to simulate the interplay between subsistence agriculture and wildlife hunting and their impacts on animal populations and landscape through a bottom-up modeling framework. While our model is implemented through rich datasets from fieldwork, remote sensing and literature review, extensive sensitivity analysis is also conducted to show how robust simulation results are to a few unknown parameters. We validated the model with the field dataset for two different villages to examine its relevance to the real-world settings.
Section snippets
Study area
The Rupununi region of Guyana (W58°6′23″, S3°17′58″; W59°37′22″, S2°6′55″) lies within the Guiana Shield of the northern Amazon region (Fig. 1). The area is difficult to access due to a mountainous terrain, un-bridged river network, and dirt roads that provide only seasonal access to most communities. The land is poorly suited for agriculture due to poor soils and a strongly seasonal flood regime (Luzar et al., 2011). To date, the Makushi and Wapishana people in the area maintain much of their
Model evaluation
We evaluate our simulation model in two ways: verification (internal consistency) and validation (external consistency). With verification, we confirm that the system logics, as described above, are correctly represented in the program. This includes a sensitivity analysis on the eight parameters for which we do not have data. We followed the protocol of Pattern Oriented Modeling (POM) (Grimm et al., 2005). POM provides a rigorous framework for model verification and validation for
Results
Our simulation results show some clear patterns in terms of human–nature interactions for the study region of the Amazon. With the establishment of a human community, animal abundance and biodiversity as well as carbon stocks decrease gradually but slowly. Initially, land conversion is limited to areas near the village (Fig. 4A). Over time, the distance between animal-kill locations increases and the average body mass of killed animals decreases. To compensate for the increasing difficulty of
Discussion
In this study, we have modeled interactions of social and ecological systems for indigenous peoples and their lands through household-level decision making on hunting and agricultural activities. This model incorporates demographic change, agricultural expansion, wildlife hunting, animal abundance, distribution and diversity, and forest succession based on empirical information (Table 2). Simulation of the establishment of a new village represents interactions between the human population size
Conclusion
Sustainability is best analyzed by examining feedbacks between social and ecological systems (Berkes and Folke, 1998, Ostrom, 2009, Schluter et al., 2012). In this paper, we describe a simulation model based on rich dataset for indigenous people and their lands in Guyana's Amazon region to examine the dynamics of socio-ecological systems. Our model applies a nutrition model as the decision making mechanism of a household agent in an indigenous territory. Sensitivity analysis shows that
Acknowledgments
The National Science Foundation (NSF; Grant BE/CNH 05 08094) provided funding, as did the Gordon and Betty Moore Foundation (GBMF). We thank the Guyana Environmental Protection Agency, and the Ministry of Amerindian Affairs for authorizing the field study; the Iwokrama International Centre for Rainforest Conservation and Development, the North Rupununi District Development Board, The Bina Hill Institute and the South Central Peoples Development Association (SCPDA) for acting as in country
References (98)
- et al.
Simulating demographic and socioeconomic processes on household level and implications for giant panda habitats
Ecol. Model.
(2001) - et al.
Plant-animal subsistence ratios and macronutrient energy estimations in worldwide hunter-gatherer diets
Am. J. Clin. Nutr.
(2000) - et al.
A role for goal-oriented autonomous agents in modeling people-environment interactions in forest recreation
Math. Comput. Model.
(1994) - et al.
Multi-scale analysis of a household level agent-based model of landcover change
J. Environ. Manag.
(2004) - et al.
Spatial agent-based models for socio-ecological systems: challenges and prospects
Environ. Model. Softw.
(2013) - et al.
The ODD protocol A review and first update
Ecol. Model.
(2010) - et al.
Mapping from ASTER stereo image data: DEM validation and accuracy assessment
ISPRS J. Photogram. Remote Sens.
(2003) - et al.
Infestation of dried cassava (Manihot esculenta Crantz) by Prostephanus truncatus (Horn( (Coleoptera: Bostrichidae)
J. Stored Prod. Res.
(1985) - et al.
Bringing home the biggest bacon: a cross-site analysis of the structure of hunter-kill profiles in neotropical forests
Biol. Conserv.
(2003) - et al.
A multi-agent simulation to assess the risk of malaria re-emergence in southern France
Ecol. Model.
(2009)
Nonlinearities in biodiversity incentive schemes: a study using an integrated agent-based and metacommunity model
Environ. Model. Softw.
An agent-based simulation model of human–environment interactions in agricultural systems
Environ. Model. Softw.
Early hominid hunting and scavenging: the role of meat as an energy source
J. Hum. Evol.
Design of an agent-based model to examine population–environment interactions in Nang Rong District, Thailand
Appl. Geogr.
A spatial model of land use and forest regeneration in the Ituri forest of northeastern Zaire
Ecol. Model.
Integrating landscape and metapopulation modeling approaches: viability of the sharp-tailed grouse in a dynamic landscape
Conserv. Biol.
Exploring complexity in a human-environment system: an agent-based spatial model for multidisciplinary and multiscale integration
Ann. Assoc. Am. Geogr.
Hunting and gathering in tropical rain forest: Is it possible?
Am. Anthropol.
The Impact of Ox-weeding on Labour Use, Labour Costs and Returns in the Teso Farming System
Home range, population density, and food resources of Agouti paca (Rodentia: Agoutidae) in Costa Rica: a study using alternative Methods1
Biotropica
Is there a link between wild meat and food security?
Conserv. Biol.
Linking social and ecological systems for resilience and sustainability
Adaptation and sustainability in a small arctic community: results of an agent-based simulation model
Arctic
The Actual and Potential Market for Cassava in Guyana, AAACP Paper Series
Hunting and the likelihood of extinction of Amazonian mammals
Conserv. Biol.
Agent-based modeling: methods and techniques for simulating human systems
Proc. Natl. Acad. Sci. U. S. A.
The Condition of Agricultural Growth. The Economics of Agrarian Change under Population Pressure
Land use change in the Amazon estuary: patterns of caboclo settlement and landscape management
Hum. Ecol.
Tapirs: Status Survey and Conservation Action Plan. IUCN, Gland and Cambridge
Spatial process and data models: toward integration of agent-based models and GIS
J. Geogr. Syst.
Exploring Demographic and Lot Effects in an ABM/LUCC of Agriculture in the Brazilian Amazon
Principles and Concepts of Agent-based Modelling for Developing Geospatial Simulations
South American Regional Infrastructure Development, Forests and REDD: Implications for Guyana
Digital change detection methods in ecosystem monitoring: a review
Int. J. Remote Sens.
The integration of agent-based modelling and geographical information for geospatial simulation
Agent-based Models Geogr. Syst.
For Logs, For Traditional Purposes and For Food: Identification of Multiple-Use Plant Species of Northern Amazonia and an Assessment of Factors Associated with their Distribution (Dissertations – ALL)
Paper 17
A bioeconomic analysis of bushmeat hunting
Proc. Royal Soc. B: Biol. Sci.
Subsistence hunting among the Waimiri Atroari Indians in central Amazonia, Brazil
Biodivers. Conserv.
Colonist household decisionmaking and land-use change in the Amazon Rainforest: an agent-based simulation
Environ. Plan. B
Cassava in Amazonia: lessons in utilization and safety from native peoples
Int. Workshop Cassava Saf.
Generating carbon finance through avoided deforestation and its potential to create climatic, conservation and human development benefits
Phil. Trans. Royal Soc. B: Biol. Sci.
An agent-based model of household dynamics and land use change
J. Land Use Sci.
An agent-based model of agricultural innovation, land-cover change and household inequality: the transition from swidden cultivation to rubber plantations in Laos PDR
J. Land Use Sci.
Updating the Minimum Dietary Energy Requirements, FAO Methodology for the Measurement of Food Deprivation
Sociometabolic transitions in subsistence communities: Boserup revisited in four comparative case studies
Hum. Ecol. Rev.
Wildlife management at the Rio das Mortes Xavante Reserve, MT, Brazil: Integrating indigenous culture and scientific method for conservation
Animals Overhunted or Animals Hiding?
Agent-based telematic services and telecom applications
Commun. ACM
Proximate causes and underlying driving forces of tropical deforestation
Bioscience
Cited by (65)
Use of logging roads by terrestrial mammals in a responsibly managed neotropical rainforest in Guyana
2023, Forest Ecology and ManagementEffectiveness of community-based monitoring projects of terrestrial game fauna in the tropics: a global review
2023, Perspectives in Ecology and ConservationBibliometric analysis of publications discussing the use of the artificial intelligence technique agent-based models in sustainable agriculture
2022, HeliyonCitation Excerpt :Group 3 (sustainable development, land usage, food safety, climate change, environment policy, environmental economy): The third group is closely related with group 1 as seen in Fig. 8. This group encompassed research articles that contribute to sustainable development, performing simulations using agent-based models to determine land usage [94, 98, 100, 101, 109, 110, 114, 115, 116], considering climate change [95, 101, 109, 110, 117], and environment policies and economies [100, 107, 110, 117, 118, 119, 120], to ensure food safety [101, 107, 110, 116, 120, 121, 122]. In summary, group 3 has a more environmental-focused approach than the other groups favoring finding the best land usage conditions and promoting sustainable development.
Understanding human-environment interrelationships under constrained land-use decisions with a spatially explicit agent-based model
2022, AnthropoceneCitation Excerpt :Exploring human-environment interactions in this study is a representative case for understanding the coupled nature-human complexity (Liu et al., 2007). Through the lens of labor migration and cropland abandonment, the model shows that the nonlinearity of social-ecological outcomes with time emerges due to feedback loops within the land system, recognizing the bidirectional influences between the social and ecological systems (An, 2012; Iwamura et al., 2014; Schulze et al., 2017). One interesting outcome is the different patterns and trajectories between the abandoned land and migrants through time (Fig. 3).
Environmental vulnerability assessment of Brazilian Amazon Indigenous Lands
2022, Environmental Science and Policy