User Tools

Site Tools


Building a social science knowledge-based model: Community Based Natural Resource Management

This exercise will acquaint you with a different type of knowledge base development activity. In the Cohead salmon example, the data used for analysis of fish habitat represented hard data, collected in the field. In this exercise, we will build a model based entirely on the subjective assessment of a human/environmental system. The intent of this exercise is to build a knowledge base that will assess the capacity of rural African communities to initiate natural resource management activities (i.e., community based natural resource management)1). While some encouraging examples of community involvement in sustainable natural resource management exist, shortcomings in development interventions can result in failure to enlist and integrate local people’s interests, needs, knowledge, skills, and cooperation in sustainably managing their resources.

Indigenous experience, external expert knowledge, evaluations, case studies and geo-referenced databases form the substantial base of existing knowledge about community based natural resource management (CBNRM) systems. Unfortunately, our understanding of CBNRM systems and their components as well as the effects of any of internal and external interventions (e.g., policy change, development of infrastructure, building of human capacity) is currently, and will always be, incomplete. Successful CBNRM depends on complex social, institutional, political, economic and biophysical arrangements. The challenge, therefore, is to critically analyze and evaluate experience and increase our understanding of both CBNRM systems and the effects of interventions so that CBNRM—as a development activity by communities, groups, or other associations that undertake efforts in Africa—can be improved.

Existing knowledge of CBNRM serves as the starting point for utilizing NetWeaver to facilitate the process of condensing, processing, filtering, organizing, categorizing, and analyzing disparate pieces of information and then presenting it as a new synthesis.

Getting Started

As in the Cohead salmon exercise, begin by opening a new project in NetWeaver. The outcome for this knowledge base is an assessment of a community’s capacity for initiating natural resource management activities. From this outcome, we will build down to a refined set of appropriate questions that will serve to characterize the assessment. Experts determined that four major components contribute to the assessment of CBNRM capacity: 1) Social Factors, 2) Economic Factors, 3) Political Factors, and 4) Biophysical Factors.

We are going to take this opportunity to demonstrate a productivity tool within NetWeaver that will expedite the construction of knowledge bases. From the Topics menu on the NetWeaver main menu select Batch Create Topics. A window will open called New Topics. Type this list of topics into this window:

  • CBNRM Initiation
  • Social Factors
  • Economic Factors
  • Political Factors
  • Biophysical Factors

Since we want these to be dependency networks, choose dependency networks from the drop down menu in this window’s toolbar (figure 1). Click the accept button to accept. NetWeaver will take the list, create a new dependency network for each line, and add them to the knowledge base (figure 2).

Figure 1:  The batch create list for CBNRM.  Names for dependency networks have been entered and the topic type changed from data links <nwbtn>topic type menu|common_87</nwbtn> to dependency networks <nwbtn>topic type menu|common_85</nwbtn>.

Figure 2:  The CBNRM knowledge base project window after adding the batch created dependency networks to the knowledge base (note that here the knowledge base has been named "CBNRM" already).

The First Network

We begin by building a dependency network that depicts the heuristics (i.e., rules) associated with assessing the capacity of a community to initiate CBNRM activities. We have already created this dependency network using batch create. We will now flesh out the network with appropriate topics. Open the dependency network window for CBNRM Initiation. It will only contain an OR node. Begin by hanging an AND node from the OR in the CBNRM Initiation dependency network, and from that node, hang the other four batch created networks in the following order: Social Factors, Economic Factors, Political Factors, and Biophysical Factors. The CBNRM Initiation dependency network should look like figure 3 when you are done.

The original model-building team did not immediately derive the knowledge base structure that we will build. They first developed an extensive list of potential factors that might contribute to CBNRM, organized that list into sub-lists containing related factors, and then developed several prototype knowledge bases that were repeatedly refined until they arrived at the present model. This iterative process of knowledge engineering will be part of most efforts at developing knowledge based models, as the process is often as much one of consensus building as it is of model construction.

See CBNRM Annex 1 for more detailed definitions and explanations of each of the factors and data links discussed in this section of the tutorial. Also, for additional information on the process of developing the CBNRM Initiation model, you can go to to review The Heron Group Reports Series reports numbers 105 and 106.

Figure 3: The CBNRM Initiation dependency network after hanging the other four batch created dependency networks for the key factors.

The Social Factors dependency network

We will now develop these four dependency networks in turn. This is most easily initiated by double clicking on the oval shape representing the dependency network as it appears in the CBNRM Initiation network. Double click on the Social Factors network object in the CBNRM Initiation network. You will see the dependency network window for Social Factors. It was determined that Social Factors are comprised of two contributing factors:

  • Cohesiveness
  • Extent of ability to manage

Hang an AND node to the OR node in the Social Factors dependency network, and to that AND node attach new dependency networks named Cohesiveness and Extent of Ability to Manage (figure 4). Alternatively, one can batch create these two dependency networks. Once they are a part of the knowledge base, they can hung from the AND node in Social Factors.

Figure 4: The Social Factors dependency network after adding the two networks for its contributing factors.

The Cohesiveness dependency network

Double click on the Cohesiveness dependency network object in the Social Factors dependency network editor window. This will bring up the dependency network window for Cohesiveness. Experts determined that Cohesiveness can be measured in two different ways: 1) A combination of Clear Leadership and Community Cohesiveness, or 2) Leadership Responsiveness. We cannot measure these parameters in the same way as gravel size or stream reach area in the Cohead Salmon example. However, someone trained in the ability to recognize and assess these community qualities could provide a qualitative assessment. We chose to permit an expert to rank these community attributes on a scale of 1 to 10. NetWeaver makes it easy to build a fuzzy argument to represent this ranking range.

We will build this qualitative model in the Cohesiveness dependency network window. First, hang an AND node from the topmost OR. Hang a new simple data link from the AND node. Name the new data link “Clear Leadership” and in the hint field write “To what degree is there a consensus in the community as to who is their leader? 1=low; 10=high”. Create a new fuzzy argument for this data link. Name the argument “0F-10T”. Build this argument as a simple fuzzy ramp from 0 (False) to 10 (True) (figure 5).

After you have finished creating the argument and attaching Clear Leadership to the AND node, repeat this procedure by creating a data link named “Community Cohesiveness”. Document this data link with the hint “How socially cohesive is the community? 1 = low; 10 = high”. You can create this fuzzy argument, or clone it from the argument in Clear Leadership. To do this, select the argument 0F-10T from the drop down menu on the argument list toolbar. This will clone the previously created argument and apply it to the this data link.

NOTE: You will recall that these community attributes are to be measured on a 1 to 10 scale, yet the fuzzy argument scales from 0 to 10. We did this to prevent any one measure of community capacity from unduly influencing the final community appraisal by being totally false. Some fuzzy arguments will be the inverse of the 0-10 arguments we have created here. In other words, 10 will be a poor score and 1 will be a good score. In these instances, the actual argument should scale from 1 to 11 for the same reason mentioned above.

Finally, create a new data link and hang it from the topmost OR node. Name this data link “Leadership Responsiveness”. Document this data link with the hint “To what degree is community leadership responsive to the needs of community members? 1=low; 10=high”. Create a fuzzy argument as you did for the previous two data links. The final dependency network should look like figure 6. The interpretation of this network is, simply, “Cohesiveness in a community is high when clear leadership exists and community cohesiveness, or when leadership is responsive to community needs”.

For the remainder of this tutorial example, you will develop the CBNRM model from a written narrative. This is to familiarize you with the process of making the transition from the elicitation of knowledge to the representation of knowledge. In other words, as in most knowledge base development efforts, the knowledge engineer (i.e., you) must conduct a dialogue with a domain expert as you attempt to identify key factors and relationships that are important in model development. We have simplified the process greatly here. Substantial modifications usually will be made after the initial development of the prototype model. In general, the final version of a knowledge base is smaller than the initial prototypes.

Read this narrative and develop the remainder of the CBNRM model. All your data link arguments will be like those you built for Cohesiveness (i.e., 0F-10T). In some cases the argument will be the inverse (i.e., 1T-11F). You have already built the part of the model that deals with cohesiveness. Also, review CBNRM Annex 1 if you have questions about definitions/explanations of different factors. Please note that this continues to be a work in progress, but use the information we provide to develop the model for this exercise.

Figure 5:  Fuzzy argument 0T-10F created in the data link Clear Leadership

Figure 6:  The Cohesiveness dependency network.

Narrative for CBNRM

The capacity of a community to conduct community based natural resource management is a function of four factors. Those factors are social, economic, political, and biophysical.

Social factors consist of cohesiveness and extent of ability to manage. Cohesiveness (which you have already built) in a community is high when there is clear leadership and community cohesiveness, or when there is leadership responsiveness to community needs.

Extent of ability to manage is more complex. Measuring a community’s ability to manage involves: 1) Breadth of participation (Is a significant share of community members prohibited from participating in activities associated with resource management?), 2) Labor mobilization (To what degree can labor be mobilized at the time it is needed for CBNRM activities?), 3) Extent of ability to negotiate (How able is the community to negotiate joint resource use and benefits with other communities and stakeholders?), and one or more of the following; 4a) Quality of the labor pool (To what extent is the community physically able to undertake a CBNRM program?), or 4b) Training (To what extent has the community benefited from training relevant to CBNRM?), or 4c) Community organizations (To what extent does the community have effective community based organizations?), or 4d) Level of innovation (To what extent do local communities display a capability and a willingness to innovate?).

Economic factors consist of 1) Perceived benefits/costs of CBNRM (To what extent do community members perceive that CBNRM will bring more to them than it costs?), 2) Distribution of benefits (To what extent can the benefits of CBNRM be distributed acceptably among the various stakeholders?), and either 3a) Infrastructure (To what degree does the infrastructure enhance the capture of the value of the resources?), or 3b) Financial resources (To what extent does access to financial resources constitute a factor in CBNRM initiation?). [Note: This is not good phrasing of a question as written since it implies no “direction”. It would be better written as “To what degree can a community access financial resources?”]

Political factors consist of two main areas, Legal factors and Institutional factors.

Legal factors consist of 1) Security of tenure (To what degree do people perceive tenure security in making natural resource management investment decisions?), and 2) Authority of communities (To what degree did the government grant to this and to other communities the authority to manage natural resources?), or 3) Legal framework (To what extent are community based resource management decisions within the legal framework of accepted bounds of national policy or tolerance?).

Institutional factors consists of 1) Linkage to national policy process (To what extent can the CBNRM effort be linked to various stages of the national policy process?), and 2) Vertical communication (To what extent do the community's decisions inform national authorities about the management of local resources and vice versa?), or 3) Decentralization (To what degree has authority devolved from the central government to lower levels of government?).

Biophysical factors can consist of natural disturbances including seismic activities, drought, flood, or fire. Since it is impossible to identify in advance the nature of the biophysical factors associated with a given community, we will restrict the assessment of biophysical factors as it (they) affect a community’s capacity to implement CBNRM to Resource manageability (What is the extent to which the resource(s) lends itself to management by the community?).


Details of this knowledge base, its rationale and the process of developing it are described in: Parker, Saunders, McFadden, and Miller. 2000. Knowledge Engineering Process Steps: NetWeaver Applied to Community-Based Natural Resource Management in Africa: Executive Summary. Submitted to USAID/AFR/SD/ENRM. Most of the substantive input on development of the CBNRM Initiation Model came from an array of individuals, most particularly Henri Josserand, Paul Bartel, and Mike McGahuey.