0433-A1

Agent-based Simulation to Support Collaborative Forest Management and Decentralization Policy

Herry Purnomo^[1]

Abstract

International calls for sustainable forest management are being acted on in the era of decentralization policy. The importance of decentralization policy has been voiced in many countries, but how to implement it within ecosystems and social systems remains unclear. In many cases, this policy is implemented without considering the local reality of forest management and its stakeholders. This will not improve forest management. There is no single global arrangement of collaborative forest management and decentralization policy the can be implemented anywhere. Agent-based simulation is proposed as a tool to seek appropriate collaborative forest management. Agent-based simulation, a branch of artificial intelligence, integrates the complexity of the social and biophysical aspects of forest management. The feedback of this adaptive forest management, on a broader scale, will enhance the adaptability of decentralization policy.

Introduction

Increasing public concern over forest environment, pressure to downsize government in the economic crisis, and the recognition that local people should play an active role in forest management have all encouraged decentralization of forest management responsibilities in Thailand (Pragtong n.d.). Over the last three years, the Indonesian government has issues several important pieces of legislation aimed at transferring authorities to the provincial and district governments, and at allowing resource-rich regions to retain a larger share of the fiscal revenues generated within their jurisdictions. The most important of these have been Law 22 on Regional Governance and Law 25 on Fiscal Balancing, both of which were issued in May 1999. These laws have been supported by a variety of implementing regulations and sector-specific decentralization laws, including Law 41 of 1999, a revised version of Indonesia's Basic Forestry Law, which outlines of administrative authority in the forestry sector under regional autonomy (McCarthy 2001).

Collaborative Forest Management (CFM) is essentially a new management paradigm, which seeks to draw on the experience and knowledge of both professional foresters and local people - in a partnership arrangement that may also involve other stakeholders (Carter n.d.). Other terms similar to CFM are Joint Forest Management (JFM), shared forest management, co-management and participatory forest management. Co-management connotes a collaborative institutional arrangement among diverse stakeholders for managing or using a natural resource (Castro and Nielsen 2001). Bass et al. (1997) stated an element of make policies that work for forest and people is an appropriate decentralization, devolution and strengthening capacity.

A recent publication by the Food and Agriculture Organization asserts: "The promotion of collaborative management is based on the assumption that effective management is more likely to occur when local resource users have shared or exclusive rights to make decisions about and benefit from resource use" (Ingles et al. 1999). Stakeholders must aware the existence of different collaboration costs such as costs of specifying rights and obligation of each stakeholder; costs of enforcement of rights; and costs of collaboration monitoring. Ingles at al. (1999) stated that collaborative management of natural resources refers to: the arrangement for management that are negotiated by multiple stakeholders and are based on a set of rights and privileges (tenure) recognized by the government and widely accepted by resource users; The process for sharing power among stakeholders to make decisions and exercise control over resource use. Since simulation is a robust way to know the impact of any right arrangement scenario (Fahey and Randall 1998) then it was used to create possible collaborative forest management scenarios.

Indonesian production forests have been allocated legally to the timber logging companies. These companies have invested the areas. This situation creates a constraint for policy makers to make a very new arrangement of forest management. Policy makers need to consider the existing arrangement to make a smooth changing. Collaboration among logging companies and local communities is a possible scenarios. However, arrangement of collaboration including rights, responsibilities, returns and relations should be as fair as possible. Inappropriate arrangement of collaboration can make one stakeholder better off but the others are worse-off. This scenario cannot be implemented in the field naturally. Therefore, an appropriate collaboration scenario and decentralization policy is challenging to search.

This paper illustrates the possible use of agent-based simulation to seek appropriate collaborative forest management and decentralization policy. We use a case of East Kalimantan, Indonesia, where the simulation activities were carried out.

Methods

Multi-agent or agent-based simulation (MAS) is a promising way to examine natural resource and environmental management issues (Bousquet 1999). The hallmark of MAS is the recognition of "agents", which are entities with defined goals, actions, and domain knowledge, collectively known as their "behavior" (Stone & Veloso 1997). Some degree of agent autonomy is central to the notion of agent-based modelling (Weiss 1999). The research used CORMAS (Common Pool Resources and Multi-Agent System), a MAS platform specifically designed for renewable resource management systems (Bousquet et al. 1998). It provides a framework for developing simulation models of the interactions between individuals and groups that jointly exploit common resources. CORMAS facilitates the construction of a model by offering predefined elements, which the user can customize to a wide range of specific applications (CIRAD 2001). Key phases in the development of the model (Grant et al. 1997) are:

1. Forming a conceptual model: stating the model objectives, bounding the system of interest, categorizing its components, identifying relationships, and describing the expected patterns of model behavior.

2. Quantifying the model: identifying the functional forms of model equations, estimating the parameters, representing it in CORMAS and executing baseline simulations.

3. Evaluating the model: re-assess the logic underpinning the model, and comparing model predictions with expectations and with the real system.

4. Using the model: developing scenarios, testing hypotheses and communicating results.

This research used causal loop diagram (CLD) to illustrate the influence diagram of the possible use of simulation in the implementation of CFM and its links to decentralization policy. Making policy more adaptive is necessary. Ostrom (n.d.) stated the limits of fully decentralized systems due to: failure to organize in some localities, local tyrannies, stagnation, limited access to scientific organization, conflict among groups and inability to cope with large-scale problems. In contrast to the centralized and decentralized structure, a polycentric governing structure offers citizens to organize not one but many governing authorities (Ostrom 1993). Governance systems exist at multiple levels with some autonomy at each level. Polycentric systems are complex adaptive systems. Agent-based models or multi-agent system models are aimed to understand the properties of complex social systems through the analysis of simulations.

Results and Discussions

Simulation Case Study in East Kalimantan, Indonesia

A case study of using multi-agent system simulation had been carried out in a state own company, Inhutani II, located at 116^o28' E, 3^o14' N, District of Malinau, East Kalimantan Indonesia. Current regulations do not offer much flexibility for concessionaires to develop site-specific management, or to involve local communities in forest management. This research examined simulation techniques to explore scenarios of sustainable forest management addressing those limitations. Using multi-agent simulation to examine social and biophysical issues to be addressed developed several scenarios (Figure 1). A collaborative forest management involving both the concessionaire and the local community appeared to offer the most promising pathway toward sustainability. Table 1 shows the CFM benefits more that the status quo. Thus, multi-stakeholders manage forest better than a single actor manages the forest.

Primary forest remains higher in the collaborative scenario than the existing situation (without collaboration). This is because some areas of Inhutani II are allocated to communities for their use, provided that it is harvested using traditional or non-mechanized systems. Thus, community use of the forest will incur lower harvesting damage. The net revenue of Inhutani II decreases, but fees paid by the local communities could compensate for this loss. Harvesting techniques used by local communities appear critical to the sustainability of the forest. Typically, communities harvest only about 10 % of commercial trees, providing favorable conditions for regeneration and time for re-growth.

Table 1. Sustainable Forest Management (SFM) indicators of scenario of collaborative management compared with status quo

Figure 1. An overview of the model, showing three classes of actors (ovals) and their goals (rectangles), the pixel-based representation of the landscape, and a summary of the outputs to be monitored as indicators of sustainability.

Simulations have shown that collaboration between logging companies and local communities can lead to mutually satisfactory (win-win) outcomes. However, care is required to ensure that the specific arrangements regarding this collaboration are fair with respect to rights, returns and relations. Inappropriate arrangements can make some stakeholders better off and others worse-off. Finding a suitable arrangement to showcase such collaboration is challenging. Such an arrangement should draw on the comparative advantages of each stakeholder, drawing on the knowledge, techniques, experience and capital of the logging company and the local community to manage the forest sustainably. Local communities have a deep understanding of the forest as well as a spiritual relationship to forest than can be useful to protect the forest from illegal activities not covered in the collaboration scheme.

However, if the collaboration uses different scenario, Scenario A or Scenario C (Table 2) then the collaboration creates impacts that not satisfy the three main actors, which are local communities, timber, company and governments. The outcomes of Scenario A creates situation that worse-off for the timber company and governments. Scenario C creates worse-off situation for the timber company. The selected scenario (Scenario B) makes forest-standing stock better than the other scenarios. Therefore, collaboration could be worse-off for some stakeholders if appropriate scenarios are not used.

Table 2. Different scenarios of collaboration

This study had been presented and argued to Inhutani II, local governments and local communities. Currently the district of Malinau where the study took place, reallocating the land and forest. We do believe this study result could enhance the spirit of collaboration among those stakeholders. By having a tool of collaborative decision-making, decentralization policy created by the national government is meaningful for the local stakeholders. The simulation can make decentralization more adaptive.

Basic similarity perception of stakeholders on sustainability concepts could be a foundation towards collaborative management of forest where better outcomes of forest management are expected. Specifying the collaboration might differ from site to site. In other words, each forest management unit (FMU) can have different collaboration scheme or arrangement. Effective communication between stakeholders is an initial stage towards collaboration. Stakeholders have to compare benefits and costs of collaboration for every possible collaboration arrangement. Collaboration does not necessarily give better outcomes in terms of sustainability of forest as shown by the research result. The benefits and costs might include tangible and intangible benefits.

Adaptive Decentralization policy

Figure 2 shows the Influence diagram of any selected decentralization policy. There are many options the way to implement decentralization. What level of it, what rights are decentralized etc. Whatever selected scheme, it will affect to the CFM arrangement. A multi-stakeholder process on seeking an appropriate right arrangement is needed. A simulation can be used to know the impact of its different scenarios. Plans and collective or collaborative actions follow a selected scenario. The action outcomes give feedbacks to the decentralization option as well as the CFM.

To make forestry decentralization policy create a better result, it should link with CFM and its implementation. A simulation can facilitate the way CFM implemented in the field by generating multi-stakeholders understanding and commitment. A link between decentralization policies with outcomes of CFM is necessary. The policy makers can use the outcomes to seek an appropriate decentralization policy.

Facing complexity of ecosystem and social system including micro and macro politics in each district, for instance, decentralization should be adaptively implemented for each district. Predefined decentralization scheme will work difficulty. A democratic process of decentralization, through the involvement of all stakeholders, is necessary. This process will determine which parts of governing need to be decentralized, remain centralized, and people self-governed.

Figure 2. Influence diagram of a selected decentralization policy

The available literature provides little empirical evidence about whether decentralization is good for forests and people who depend on them. A possible advantages of decentralized natural resource management mentioned in the literatures is management decisions can incorporate local knowledge about the resource base (Brandon and Wells 1992; Carney 1995; Poffenberg 1990; Utting 1993 in Kaimowitz et al. 2002). Colfer et al. (1999) mentioned local knowledge as one of six dimensions of determining the relative importance of forest stakeholders. Therefore, incorporating local knowledge in the management of forest can increase the power of local communities in managing forest. In the simulation above, the local knowledge was traditional logging. Ignoring the role of traditional knowledge in the decentralization policy in Zimbabwe caused the failure of that policy (Lalonde 1993).

Figure 3. Plausible connection of decentralization policy that incorporates traditional knowledge

Empowering local communities by stating clearly that traditional knowledge of sustainable forest management has to be counted in managing forest will enhance the possibility of success of decentralization policy. Figure 3 shows the plausible influence of the decentralization policy that incorporates the use of traditional knowledge. The degradation of forest as well as the poor of local communities surrounding the forest enforces the formulation and implementation of decentralization policy. The negative loop indicates that there is a stable level of decentralization. The influence diagram shows that the forest degradation issue can be used to reduce forest degradation itself by first increasing public concern over forest and encouraging decentralization of forest management responsibilities. Using local knowledge such as traditional logging is a necessary condition to empower local communities. Finally, an appropriate CFM that is searched through simulation, enhances sustainability level.

Conclusion

Simulation can be used as a tool to seek an appropriate collaborative forest management. It had been done in East Kalimantan, where the collaboration scenario leads to better forest management. A key entry point of collaboration is using traditional knowledge that more environmentally friendly, such as traditional logging. Different appropriate collaboration of forest management for each different area will enhance the adaptive forest management. The decentralization policy that currently implemented has to adapt this variability. Adaptive decentralization policy is a necessary condition for reducing forest degradation.

References

Axelrod, R., 1997. The Complexity of Cooperation: Agent-based models of competition and collaboration. New Jersey: Princeton Univ. Press.

Bass, S., J. Mayers, J. Ahmed, C. Filer, A. Khare, N.A. Kotey, C. Nhira, V. Watson, 1997. Polices affecting forests and people: ten elements that work. Commonwealth Forestry Rev 76(3): p186-190

Bousquet, F., I. Bakam. H. Proton, and C. Le Page, 1998. CORMAS: Common-pool resources and multi-Agent systems. International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems. Lecture Notes in Artificial Intelligence 1416, 826-838, Berlin: Springer-Verlag.

Bousquet, F., O. Barreteau, C. Le Page, C. Mullon and J. Weber, 1999. An environmental modelling approach: The use of multi-agent simulations in F. Blasco and A. Weill (eds). Advances in Environmental and Ecological Modelling, http://CORMAS.cirad.fr/pdf/gowith.pdf. Accessed 15 January 2001.

Carter, J., n.d. Recent experience in collaborative forest management approaches: a review of key issues. Switzerland: Intercooperation.

Castro, A.P. and E. Nielsen, 2001. Indigenous people and co-management: implications for conflict management. J Environ Sci & Pol 4:229-239.

CIRAD, 2001. Natural resources and multi-agent simulations. Centre de Coopération Internationale en Recherche Agronomique pour le Développement. http://cormas.cirad.fr/en/outil/outil.htm, Accessed 25 February 2002.

Fahey, L. and R.M. Randall, 1998. What is scenario learning? in Fahey, L. and R.M. Randall (eds). Learning from the future: competitive foresight scenarios. New York: John Wiley & Sons, Inc. p 3-21.

Grant, J.W., E.K. Pedersen, S.L. Marin, 1997. Ecology and natural resource management: system analysis and simulation. Reading: Addison-Wesley.

Ingles, A.W., A. Musch, H. Qwist-Hoffmann. 1999. The participatory for supporting collaborative management of natural resources: an overview. Food and Agriculture Organization, UN. Rome.

Kaimowitz, D., C. Vallejos and P. Pacheco and R. Lopez, 1998. Municipal governments and forest management in lowland Bolivia. J Env. & Dev 7(1): 45-59.

Lalonde, A. 1993. African indigenous knowledge and its relevance to sustainable development. in Inglis, J.T. (ed) Traditional Ecological Knowledge: Concepts and Cases. International Program on Traditional Ecological Knowledge and International Development Research Centre. Canada.

McCarthy, J.M., 2001. Decentralization, local communities and forest management in Barito Selatan District, Central Kalimantan. Case Studies on Decentralization and Forests in Indonesia. Bogor: CIFOR Publ.

[ODA] Overseas Development Agency of UK, n.d. sharing forest management: Key Factors, Best Practice & ways forward.

Ostrom, E., L. Schroeder and S. Wynne, 1993. Institutional incentives and sustainable development: infrastructure policies in perspective. Boulder: Wesview Press.

Ostrom, E., 1999. Self-governance and forest resources. CIFOR Occasional paper no. 20. Bogor: CIFOR Publ.

Ostrom, E., n.d. Policentric Intututions: Blending local and global knowledge. Indiana University. http://www-2.ids.ac.uk/gdnet/fulltxt/ostrom[1].ppt. Accessed 25 April 2002.

Pragtong, K., n.d. Recent decentralization plans on the royal forest department and its implications forest management in Thailand. http://www.fao.org/. Accessed 25 April 2002.

Sarin, M., 2001. Disempowerment in the name of 'participatory' forestry? - Village joint forest management project in Uttarakhand. Forest Tree and People no 44.

Stone, P. and M. Veloso, 1997. Multiagent systems: a survey from a machine learning perspective. Carnegie Mellon University, http://www-2.cs.cmu.edu/afs/cs/usr/pstone/public/papers/97MAS-survey/revised-survey.html, Accessed 20 October 2001.

Sukwong, S., 2000. Linking local lessons to policy development. Paper presented at the 4^th International Workshop on Model Forests for Filed Level Applications of SFM; Japan, 23-27 October 2000.

Weiss, G., (ed), 1999. Multiagent systems: a modern approach to distributed artificial intelligence. Cambridge: MIT Pr.