Cover Image
close this bookAgroforestry in the West African Sahel (BOSTID, 1984)
close this folderAppendixes
View the documentAppendix A: Selection and Use of Tree Species
View the documentAppendix B: Methodology for Diagnosis and Design of Agroforestry and Management Systems

Appendix B: Methodology for Diagnosis and Design of Agroforestry and Management Systems

John Raintree

If agroforestry is to live up to current expectations concerning its capacity to solve problems, significant improvements will have to be made in the methodology for relating research and development efforts to the actual needs and potential of tropical land use systems. This paper highlights the principal features of the core logic of an evolving diagnostic and design methodology that is intended to serve as a reliable tool for arriving at effective and adoptable agroforestry solutions to local land use problems the world over.


The ultimate practical aim is to improve agroforestry land management systems and technologies with specific capabilities to solve land management problems. Unfortunately, technologies implicitly designed for conditions that prevail at research stations, "high access" farms (Rolling 1980), and forest management units are often completely unsuitable when extended to the majority of land users in the same agroecological zone. The problem is not that the biophysical features of the zone have not been taken into account--on the contrary, they are usually well understood--but that discipline-focused researchers often fail to perceive that the existing land use system has its own internal organization and its own unique set of operational constraints and potentials.

The problem with an ad hoc approach to designing or prescribing technologies is that the technologists are rarely equipped to address the full set of design criteria. Rather than designing technology on the basis of only a partial set of criteria and then treating the nonadoption of the resulting technology as an "extension problem," it will almost always be more useful to place the onus of responsibility squarely on those developing the technology, recognizing that in the first instance, there is a design problem and no substitute for good design. This objective requires coordinated contributions from an interdisciplinary team of professionals as well as from the intended users of the eventual technology product.

A problem-oriented diagnostic approach to agroforestry design is the most direct and logical route to effective and adoptable agroforestry technologies and land management systems. A long, drawn-out survey process is neither necessary nor useful. The aim is to develop a practical, effective, and quickly accomplishable diagnosis and design method that can prove its usefulness by the results it obtains in a wide range of environments.


The success of the methodology will be judged not by the number or the elegance of resulting agroforestry technologies but by the impact it has had on the landscape, that is, how effective it has been in transforming landscapes into more productive and sustainable land use systems. A successful methodology must somehow guide the user toward agroforestry technologies that embody three essential attributes: productivity, sustainability, and adoptability.

The first two criteria are virtually axiomatic. Agroforestry is an approach that seeks to improve the productivity and sustainability of land use systems and has significant potential for achieving both objectives simultaneously. Productivity and sustainability are the most effective criteria by which to measure problems of existing land use systems and evaluate potential agroforestry alternatives. No matter how efficiently or elegantly a technology may solve a problem, however, it will have little impact unless it is adopted by a significant percentage of the intended users. In agroforestry diagnosis and design, there are many factors beyond technological irrelevance that may limit the adoptability of an otherwise promising technology.

Most of the possible adoption constraints have to do with the level of available resources and management skills in a given system, or with the incompatibility of the potential technology with existing practices or certain cultural factors associated with the general technological tradition of the area. It may be difficult, or even impossible, to diagnose all of the potential constraints on adoption before undertaking farm trials of the proposed technologies, but the process can be guided initially by the common sense assumption that the ability to solve a problem begins with the ability to define it (Steppler 1981, Steppler and Raintree 1981).

There are two practical implications for a strategy that focuses attention on the solution of perceived problems in existing land use systems. In the diagnostic phase, it becomes even more essential to involve the land users in the process inasmuch as only they can shed light on their perceived problems. Hence, it is important to emphasize analysis of perceived management problems and strategies at the household or unit management level.

In the design phase, not all of the problems that constrain the productivity and, particularly, the sustainability of a household land management system are clearly perceived by the manager; and even when the problem is perceived, its solution may not rank high in the farmer's priorities, and technologies designed to solve the problem may fail to awaken any adoption interest. Although often viewed as an "extension" or "education" problem, again it may be more productive to regard it as a design problem.

The multifunctional nature of many potential agroforestry technologies may enable the designer in such cases to find some attractive way to link the not necessarily wanted conservation function to some desirable production function of a well-chosen multipurpose technology.

For example, in Kenya, farmers with little or no present interest in erosion control (a severe problem in dry hill areas) nevertheless appear very interested in hedgerow planting of fast-growing leguminous trees to satisfy household fuel wood needs. By planting dense hedgerows of coppicing fuel wood trees on the contour with row spacings selected for effective erosion control, both problems can be solved with a single, adoptable design. Other farmers in Kenya, on the other hand, have expressed a definite and immediate interest in hedgerows for erosion control, but there is no currently perceived problem with fuel wood supply. Where trend analysis indicates a potential fuel wood problem, these farmers, with potential fuel wood production systems already in place, could then begin to manage the hedgerows for fuel wood. These two examples of cleverly designed multipurpose agroforestry systems illustrate the kinds of design considerations that follow from a diagnostic approach.

In making the analysis, it is helpful to distinguish between constraints and potentials of existing land use systems and those that pertain to the appropriateness of potential agroforestry technologies. These two levels of evaluation (dealing with constraints and potentials of different types) are part of a sequence of analyses outlined below:

Diagnostic Phase

1. Characterize the essential features of structure and function in the existing land use system and identify the output subsystems.

2. Evaluate the performance of the subsystems (that is, identify problems).

3. Determine what constraints limit the performance of the subsystems.

4. Identify general potentials for performance-improving (constraint-removing) interventions of an agroforestry nature (candidate technologies).

Design Phase

5. Determine constraints that condition the appropriateness of candidate agroforestry technologies (components and practices).

6. Identify remaining potentials for specific agroforestry technologies (existing or to be developed).

The following section discusses details of the logic of agroforestry diagnosis and design and considers what is needed at each of the above steps.

Identification of Output Subsystems

In analyzing land use systems, initial attention must be directed to the evaluation of the resource base. Subsequent priority should be given to the definition of land management units (or their equivalents) as the primary decision-making units and reference systems. Defining land use subsystems in terms of their output seems most appropriate because it is (a) the least restrictive modeling possibility, (b) the most compatible with various techniques of input-output analysis, and (c) the most consistent with the way in which land users manage their land--that is, to produce desired outputs. A "major output subsystem," then, may be defined as the set of activities, resources, and other land use factors that are involved in the generation of an output intended to satisfy one of the basic production objectives of the household,

In deciding specifically what output categories to consider as "basic," it is important, for a widely applicable methodology, to satisfy two general requirements: (1) general applicability, and (2) adequate representation of the idiosyncrasies of local land use systems. To satisfy both requirements and to facilitate ready linkage with categories of agroforestry technologies, it is fruitful to follow a "basic needs" approach. The output categories considered basic to the economic well-being of households everywhere are:

1. Food

2. Energy

3. Shelter--all forms of shelter (housing for people, livestock, and personal belongings; shade, windbreaks, etc.) and enclosure (fences, kraals, boundary markers, etc.)

4. Raw materials for home industry--all raw materials for household or village manufacture of everything from clothes and kitchen implements to medicinal preparations--that is, all locally manufactured consumer items, whether for home consumption or sale

5. Cash income

6. Community integration--all forms of "social" production and consumption (feasting, gift-giving, brideprice, taxes, education, etc.).

This approach assumes that (1) the needs identified in the list are basic and universal; (2) local systems will display great variety with respect to the preferred forms in which these needs are satisfied (food and fuel preferences, shelter types, etc.), but that these will all be variations on the same universal themes; and (3) local and regional land use systems are organized to produce goods aimed at satisfying these basic needs (whatever else they might also do). The way in which they do this will, of course, vary from system to system. In commercial land use systems, cash crop production for purchase of the basic commodities will be the predominant household strategy. In more subsistence-oriented economies the household land use system will be organized to satisfy the basic needs more directly.

The use of the term "basic needs" does not imply any restriction on the level of economic development. The needs that have been highlighted are basic in type, not necessarily in level of satisfaction.

Problem Identification

Once the basic needs subsystems have been identified, problems in the productivity and sustainability of the basic production subsystems can be identified by conducting intensive interviews with farmers. The following example from Kathama, Machakos District, Kenya, illustrates the application of the methodology to a semiarid zone, mixed farming system in the midlands of East Africa.

Problems in Household Basic Needs Supply Subsystems

1. Food. Seasonal staple food shortages are normal, and deficits must be made up by purchases; drought-related crop failure requiring famine relief occurs on the average of once every five years; low milk and meat production results from dry season feed shortage for livestock.

2. Energy. Insufficient fuelwood produced from personally owned land requires purchase for household and cottage industry uses; large trees for brick kilos are not available.

3. Shelter. Lack of construction-quality timber and poles requires purchase of expensive supplies; lack of large trees for brick making; lack of fencing and shade trees; problems with wind desiccation of crops.

4. Raw materials. Must purchase expensive fuelwood supplies for butchery and brick making.

5. Cash.Low net household income due in part to cash drain for staple foods, fuelwood, and construction wood; earning and savings potential of livestock enterprise limited by dry season feed gap.

6. Community integration. Difficulty in meeting expectations for cash contributions to numerous "harambee" community self-help projects; difficulty in meeting educational expenses.

Analysis of Land Use Constraints

Once the problem subsystems and the general nature of the supply problem have been identified, analysis of the land use system traces out the causes of the supply problem. In the Kathama example, the causal factors are:

Crop land

1. Low fertility and declining yields
2. Lack of manure
3. Soil or wind erosion and water loss due to poor infiltration and heavy runoff of rainwater
4. Waterlogging on low spots
5. Labor bottleneck at ploughing and weeding time
6. Pests

Grazing land

1. Small grazing area
2. Insufficient dry season feed production
3. Overgrazing and soil erosion
4. Uneven distribution of water supplies

In the first approach, the analyst has intensive discussions with the farmer to probe the cause of the problem and also observes the farm. Additional objective measures are also being developed to supplement interview and observation data with more quantitative measurements of land use problems. This approach provides (1) a spot diagnosis, and (2) sufficient information to establish a structural model of the problem's causes. With respect to the latter, a causal network diagraming technique (Figure B-1) is a useful tool in analyzing interrelationships among land use problems and identifying the critical constraints that limit the productivity and/or the sustainability of the system.

Identification of Potential Agroforestry Interventions

The resulting model or models of problem etiology, such as the partial model of cropping system constraints shown in Figure B-1, then serve as the basis for identifying points in the system where interventions could remove, reduce, or bypass specific constraints. The analyst simply studies the causal diagram(s) and, for each node in the causal network, asks "Is there anything trees can do to solve or mitigate this problem?" Ideally, this exercise should be an interdisciplinary brainstorming session about possible land use alternatives.

Nonagroforestry alternatives should also be considered. In certain situations, for example, traditional approaches to land use other than agroforestry may be more appropriate than agroforestry systems. Where these are clearly superior to agroforestry alternatives, they should be recommended. Agroforestry is not the solution to every land use problem, and there is simply too much real agroforestry work to be done in the world to squander resources trying to force agroforestry technologies into land use systems where they have no clear and significant role to play.

FIGURE B-1 Partial causal network model of cropping system problems in Kathama, Machakos District, Kenya (semiarid zone, mixed farming system). (International Council for Research in Agroforestry)

At a minimum, the diagnosis and design team identify agroforestry technologies that can solve land management problems by addressing specific end-use or service potentials in the system. Design indications are drawn from the Kathama example previously cited.

Specific Problem-Solving Agroforestry Alternatives

1. Alley cropping/mulch farming with leguminous and other suitable trees to control erosion, increase rainwater infiltration, reduce runoff, conserve soil moisture, improve soil fertility and structure, reduce the traction requirements for tillage (or the tillage requirement in general, by minimum tillage management), lessen the labor requirement for weeding, and possibly provide some measure of pest control through use of insect-repelling mulch species such as neem (see Figure B-2 for the logic of this intervention in the form of a causal network diagram).

FIGURE B-2 Alley cropping as a potential solution to cropping system problems in Kathama,Machakos District, Kenya. (International Council for Research in Agroforestry)

2. Elimination or reduction of the dry season feed gap by planting multipurpose fodder trees in grazing areas and as hedgerows in and around crop fields with concomitant erosion control and windbreak benefits and fuelwood and mulch coproduction possibilities; the improved feed situation should potentially allow dry season plowing and planting.

3. Hedgerows and living fences of high-yielding fuelwood species and fruit-producing thorn bushes for better livestock control; appropriate plantings can also function as a safeguard against famine in bad years and as a source of supplementary livestock feed in average years.

4. Multistory fruit tree plantings with undersown grass/legume pasture.

5. Cut-and-carry fodder trees for increased pen feeding of livestock to improve dry season nutrition and increase the amount of collectible manure.

Identification of Constraints on
Potential Agroforestry Interventions

The next step is to evaluate which of the agroforestry technologies identified in the previous step are promising in the context of a detailed analysis of site constraints. Gathering detailed data on site and land management characteristics can now be limited to those necessary to evaluate particular technological possibilities. This is done by eliminating those components rendered inappropriate by topography, soil, or other factors. For example, in the Kathama ease, the presence of large termite populations renders inappropriate any mulch species that provide a good habitat for these pests; and it encourages the use of mulch species that have the ability to repel or discourage termite infestation (for example, Azadirachta indica, Adhatoda vasica, Derris indica)

Next, the process identifies those practices that are unlikely to be adopted by virtue of their incompatibility with the local farming system because of resource requirements, labor bottlenecks, management incompatibilities, or conflicting government laws and regulations or the manner in which they are enforced. For example, in Kathama the establishment technique initially used to plant out the first round of alley-cropping farm trials was found to be incompatible with the local practice of plow weeding, which tended to bury the young tree seedlings under a heavy layer of soil. As in this case, it may not always be possible to identify all of the potential constraints prior to actual farm trial of the candidate technology, but such identification should be the aim of pretrial screening.

It may be possible to modify the local farming practice somewhat to accommodate the new technology (for example, a modified plow-weeding practice seems to be acceptable to the farmers in Kathama), or it may be necessary to look further for a suitable agroforestry alternative. A basic understanding of constraints on potential agroforestry interventions is of considerable importance in the planning process.

Finally, following this elimination process, we arrive at a set of feasible agroforestry alternatives that may be compared with each other, with existing land management practices, and with nonagroforestry alternatives to determine which, if any, should be incorporated into site-specific, problem-solving agroforestry designs.

Farm Trials and Field Station Follow-up

The "rapid appraisal" diagnostic and design procedures outlined above are merely the beginning of the technology research and development (R&D) cycle. For project development they should be followed, depending on the state of readiness of the technology in question, by immediate farm trials of "best bet" agroforestry technologies and/or by on-station R&D to develop "notional" or "preliminary" technologies for later incorporation into on-farm trials. These activities entail their own methodological needs. The International Council for Research in Agroforestry (ICRAF) intends to collect, develop, and disseminate information and methodologies for the full range of biophysical and socioeconomic research questions related to the development of agroforestry's potential as a solution to global land use problems (International Council for Research in Agroforestry 1982b).


International Council for Research in Agroforestry. 1982a. Concepts and procedures for diagnosis of existing land management systems and design of agroforestry technology. International Council for Research in Agroforestry, Nairobi, Kenya.

1982b. The ICRAF Programm of Work 1982-84. International Council for Research in Agroforestry, Nairobi, Kenya.

Rolling, N. 1980. Alternative approaches in extension. In Progress in Rural Extension and Community Development, G. E. Jones and M. Rolls, eds. John Wiley & Sons, Ltd,, Chicester, England.

Steppler, R. A. 1981. A strategy for the International Council for Research in Agroforestry. International Council for Research in Agroforestry, Nairobi, Kenya.

Steppler, R. A., and J. Raintree. 1981. The ICRAP research strategy in relation tO plant science research in agroforestry. In Plant Research and Agroforestry, P. A. Buxley, ed, International Council for Research in Agroforestry, Nairobi, Kenya.