Agroforestry Chapter 6 (Part 1): Planning, Constraints and Diagnosis and Design (D&D)
Agroforestry Notes · Chapter 6 · Part 1
Chapter 6 (Part 1): Planning, Constraints and Diagnosis and Design
How agroforestry is planned, the constraints farmers face, and ICRAF’s Diagnosis and Design methodology: key features, land use system, five stages, fifteen steps and comparison with FSR/E and Land Evaluation.
1. Planning and constraints
Planning in agroforestry is the management of interactions between trees, crops and livestock on each plot so that the farmer’s objectives are met. The manager uses the positive interactions and removes or reduces the negative ones. It applies to farms of all sizes, and the farmer takes the final decision on adoption.
- Tree interference can lower crop yield below sole cropping
- Tree canopy competes for light; felling can damage crops
- Water competition, worst when trees lack a deep taproot
- Trees may host pests; some release allelochemicals
- Long gestation period of trees delays returns
- Farmers prefer field crops with quick returns
- Land and tree tenure; restrictions on felling and transit
- Poor markets and price information for tree products
- Shortage of quality planting material
- Lack of location-specific packages
- Few trained extension workers
2. Diagnosis and Design (D&D)
D&D was developed by ICRAF (development directed by J.B. Raintree) as a methodology for the diagnosis of land management problems and the design of agroforestry solutions. It applies the medical principle “diagnosis should precede treatment”. It is an adaptation of Farming Systems Research/Extension (FSR/E) and Land Evaluation to agroforestry, because, in Raintree’s words, “for most FSR/E practitioners, the trees within the farming system tend to be invisible”.
A procedure that can be adapted to the needs and resources of different users and situations.
A rapid appraisal at the planning stage, with in-depth follow-up during implementation.
An open-ended, iterative learning process; it continues until no further improvement is needed.
- Criteria of a good design: productivity, sustainability and adoptability.
- D&D is system-specific: diagnosis leads to specifications, which suggest candidate technologies.
- Users: researchers, extension officers, government field workers and NGOs.
- Its last key concept: “If at first you do not succeed, try and try and try again.”
3. Stages and steps of D&D
| Stage | Basic question | Mode of inquiry |
|---|---|---|
| Pre-diagnostic | Which land use system to focus on? How does it work? | Comparison of land use systems, regional reconnaissance |
| Diagnostic | How well does the system work? What are the problems and their causes? | Diagnostic interviews and direct field observation |
| Design and evaluation | How to improve the system? | Specification and evaluation of technologies |
| Planning | What is needed to develop and spread the improved system? | Research design, project planning |
| Implementation | How to adjust to new information? | On-station and on-farm trials, re-diagnosis |
- Steps 1–4 pre-diagnostic; 5–7 diagnostic; 8–10 design; 11–12 evaluation; 13–14 planning; 15 implementation (15 steps under six stages in the e-course list).
- Sites are prioritised by severity of problems, agroforestry potential and regional representativeness.
- Ex-ante evaluation judges the design before it is tried, on productivity, sustainability and adoptability.
4. D&D compared with other methods
| Point | FSR/E | Land Evaluation | D&D |
|---|---|---|---|
| Origin | Failures of top-down Green Revolution transfer | FAO framework for land suitability | ICRAF, for agroforestry |
| Focus | Annual crops and livestock | Physical land qualities | Whole land use system including trees |
| Strength | Farmer participation | Environmental aspects | Social aspects, design step |
| Scale | Farm | Region | Variable (farm to region) |
- Young compared D&D with Land Evaluation: Land Evaluation is stronger on environmental, D&D on social aspects.
- Agro-ecosystem analysis is a simpler method for rapid rural appraisal with the same philosophy.
- These methods identify problems and suggest solutions; they are not a substitute for action.
5. D&D in more depth (Nair, Chapter 19)
Nair calls D&D the most significant tool for designing agroforestry systems in the 1980s. Its sections in his book are adapted, with permission, from the work of J.B. Raintree, who also wrote ICRAF’s user-friendly D&D manual (1987).
| Scale | Unit | Special tools |
|---|---|---|
| Micro | Household or family farm | Basic needs approach (food, fuel, fodder, shelter, raw materials for crafts, cash) and a troubleshooting procedure (what causes this problem, and why?) |
| Meso | Village, community or watershed | Landscape planning, e.g. aerial photos to plan contour hedgerows, check dams and gully plugs (Rocheleau, 1985) |
| Macro | Region, province, country or ecozone | Links with FAO Land Evaluation surveys (Scherr, 1989) |
- D&D is scale-neutral: the same logic works at farm, watershed and regional level.
- Women must be included in diagnostic interviews, because problems such as fuelwood shortage (often a “woman’s problem”) may be missed by men, while cash is usually the man’s domain.
- Meso example: low-resource farmers in the upper watershed grow fuelwood to sell to commercial farmers in the valley.
- Most D&D applications so far have been for development projects rather than research.
| How D&D differs from FSR/E (Raintree, 1987) |
|---|
| Broader diagnostic scope, with specific attention to trees in the farming system |
| A more elaborate technology design step for complex landscape interventions |
| Can be applied at variable scales |
| Greater emphasis on the iterative nature of diagnosis and design |
A conceptually simpler method for rapid rural appraisal; same philosophy as D&D.
From rangeland management; puts more weight on design than on diagnosis.
6. Field experiments in agroforestry (Nair, Chapter 20)
Huxley (1990): agroforestry, which has been mostly descriptive, must become experimental. Development projects often assume the technologies are ready, but what we have are mostly hypotheses.
| Basis (Table 20.1) | Categories |
|---|---|
| Organizational level | Ecosystem, farm/plot, component, cellular/molecular |
| Stage of technology generation | Exploratory, component/system management, prototype |
| Subject | Biophysical; social, economic, political |
| Nature | Methodological, experimental |
| Application of results | Basic; applied (strategic, adaptive) |
| Question addressed | What, why, how |
| Place | On-station, on-farm |
Studies processes and mechanisms (DNA, N-fixation, photosynthesis) to advance knowledge.
Innovative use of basic results for medium- to long-term problems, e.g. dwarf cereals, or developing alley cropping technology.
Location-specific technology for immediate problems, e.g. using alley cropping for soil conservation at one site.
“What” research is observational and done by technicians; “why” and “how” research is basic or strategic and done by scientists.
Fisher’s three principles (Fisher, 1947)
| Principle | Purpose |
|---|---|
| Randomization | Treatments allotted to plots at random, to remove hidden site effects and researcher bias |
| Replication | Each treatment repeated on several plots, to average responses and estimate variability (error) |
| Local control (blocking) | Plots grouped into uniform blocks, to reduce variability within each block |
| Experiment | Plot size or layout (Nair) |
|---|---|
| Early MPT screening (survival, growth) | Small: 20–30 m² |
| Testing an agroforestry technology | Larger: 50–200 m² |
| Alley cropping, hedge variables only | A single hedgerow is enough as the net plot |
| Alley cropping, alley width or soil effects | At least one alley with a hedgerow on each side (Roger and Rao, 1990) |
| Plots on slopes | Long and linear on the contour; larger guard area on the uphill side |
| Design | Key point |
|---|---|
| Completely randomized (CRD) | Simplest design, but seldom used in the field |
| Randomized complete block (RCBD) | Simplest and most used field design; every block has all treatments |
| Incomplete block (lattice, confounded) | When there are too many treatments for one uniform block, e.g. MPT screening (Cochran and Cox, 1957) |
| Split-plot | Large-plot factor (e.g. hedge species) on main plots, small-plot factor (e.g. pruning) on subplots. Main-plot effects are less precise (fewer error degrees of freedom) |
| Strip-plot | Two factors that both need large plots; the interaction is measured more precisely than either factor |
| Systematic (Nelder fan, 1962) | Spacing changes step by step across the site; useful for espacement and alley-width trials (Huxley, 1985) |
- Four kinds of field trials: MPT screening (the most common; “what” type), component and system management, component interaction, and prototype evaluation.
- Soil studies: Type I = the same site monitored over time (preferred but scarce); Type II = nearby sites of known age sampled at one time (quicker) (Sanchez, 1987).
- Big programmes of the 1980s: ICRAF trials in Africa; Winrock’s F/FRED project in Asia (USAID); the All India Coordinated Research Project on agroforestry (ICAR), 31 sites; NRCAF Jhansi; CATIE, IITA and ICRISAT.
- MPT breeding: Leucaena by J.L. Brewbaker (Hawaii); NFTA for Gliricidia, Erythrina, Sesbania and Dalbergia.
7. On-farm research (Nair, Chapter 21)
The old model was a one-way “transfer of technology”: researchers develop, extension teaches, farmers adopt (Chambers et al., 1989). On-farm research (OFR) means doing research or testing technologies in farmers’ fields so that farmers can help evaluate them.
- Related terms: participatory research; FSR/E (which uses OFR as its main tool); “Farmers First” (Chambers et al., 1989); “Farmer-Augmented Designs for Participatory Research” (Pinney, 1991).
- Farmers judge differently: researchers express yield in kg/ha/yr, farmers may count kg of grain per kg of seed sown.
- Farmers with similar conditions are grouped into recommendation domains (Shaner et al., 1982).
Modified Stability Analysis (Hildebrand, 1984)
Adapted from the method plant breeders use for genotype × environment interaction. The environmental index of each farm is the average yield of all treatments on that farm. Each treatment’s yield is regressed on this index:
Malawi, 14 farms in 2 villages, 2 maize cultivars × 2 fertiliser levels: the local maize was better on poor farms and the improved composite was better on good farms. MSA needs only one block per farm, but the index is relative; each environment must also be described.
- Poor understanding of how and why farmers grow trees
- Little empirical information on AF systems
- Systems are complex and variable
- Few locally validated technologies
- No data for policy (areas, productivity, markets)
- First learn how, where and why farmers grow trees
- Use simple designs such as randomized blocks
- Block by gradients; often one farm = one block
- On-farm plots are larger than on-station plots
- Select farms at random; use large sampling units
8. Quick revision box
- D&D: ICRAF, J.B. Raintree; “diagnosis should precede treatment”.
- Adapted from FSR/E and Land Evaluation; trees are “invisible” in FSR/E.
- Key features: Flexibility, Speed, Repetition.
- Good design: productivity, sustainability, adoptability.
- Land use system = land resources + technology + objectives of land user.
- 5 stages, 15 steps; regional reconnaissance step 2; ex-ante evaluation step 11.
- Young: Land Evaluation = environmental; D&D = social.
- D&D scales: micro (basic needs + troubleshooting), meso (watershed; Rocheleau 1985), macro (Scherr 1989).
- Fisher (1947): randomization, replication, local control. RCBD most used; split-plot main plots less precise.
- Strategic = developing alley cropping; adaptive = applying it at one site. Nelder fan design for spacing.
- AICRP agroforestry: 31 sites. Type I vs Type II soil studies (Sanchez).
- MSA (Hildebrand 1984): Yi = a + b·e; e = mean yield of all treatments at a site.
9. Interactive MCQ Quiz
27 questions. Tap an option to see the answer and explanation, then go to the next question.
Comments
Post a Comment