During the last decade integrating information to obtain more impact-oriented outputs in development projects has been a new research challenge. Approaches have been focussed on the biophysical component; relevance of the social and economic components have been considered but not to the extent that they should. Attempts at a more holistic approach integrating community participation and local knowledge with geographic information system (GIS) tools to develop sustainable plans for natural resource management have been
implemented more recently. The goal of the research presented here was to develop methods for incorporating GIS-based tools, farmers' knowledge and formal models of representing their decision-making, along with scientific knowledge of cassava/ maize intercropping, into the planning and development of a participatory intercrop improvement programme for Colombia, many aspects of which are anticipated to have pa n-tropical application.
Using a GIS-tool, production systems were identified by modelling participatory land use and soil unit maps. For each prioritised production system, farmer types were identified and characterised using land tenure, farm size and other social and economic variables. Farmers' knowledge and decision-making criteria were acquired from small purposive samples of farmers, related to soil fertility conservation and enhancement, integrated pest management and genetic resources of cassava and maize. Generalisation was tested using a more
representative large, stratified, random sample. Scientific knowledge about cassava / maize intercropping was also generated in a large intercropping experiment in which root yield, dry matter, leaf area index and grain yield were measured. Results were integrated in terms of how farmers' knowledge was associated with the biophysical characteristics of production systems and varied amongst farmer types, and how this was reflected in farmers' decision-making
processes was explored.
Representing the soil and land use components using GIS showed the feasibility of identifying biophysical constraints for cassava / maize intercropping and the relative importance of this cropping practice on specific soil units. Participatory mapping was essential for production of an up to date land use map showing the location and extent of the various cropping practices and pasture. Clustering procedures, based on structural variables such as form and stability of land tenure and farm size , revealed different types of farmers within production systems. Farmers' articulated agro-ecological and socio-economic knowledge. Their decision-making could be represented as a logical reasoning process which followed four phases: detection of the problem, specifying the problem, analysis and choice amongst the possible alternatives, and finally the implementation of the decision by putting the chosen plan into action. Both farmers' knowledge and decision-making models varied according to the production system and, in some cases, farmer type, gender and age. Maize
varieties with high dry matter production and LAI and low grain yield were less suitable for intercropping. Cassava varieties with high vigour were able to compensate for the initial growth reduction caused by intercropping with maize. At high density of intercropping, cassava root diameter and length were affected simultaneously by more intense intra and inter-specific competition.
The integration of farmers knowledge to generate up to date land use maps, along with the physical constraints revealed during the GlS modelling, provided a clear definition of the research domain where socio-economic characteristics of farmers could be determined in detail. Farmers knowledge about cassava/ maize intercropping was based on the interactions amongst agro-ecological conditions and socio-economic determinants. The farmers' decision-making process was rational. and based on the knowledge that they held along with information about demand for products and prices, available from the marketing chain.
Controlled experimentation supported farmers' rationality about the use of local germplasm and severity of insect attack in relation to the phase of the moon in which crops were planted but not in respect of direct effects of maize pollen on cassava leaf function. There was some clear complementarity of local and scientific knowledge, supporting the potential value of integrating farmers' knowledge with that of researchers and extension agents when planning
research. Guidelines for the structure of a participatory intercrop improvement programme for cassava/ maize were outlined based upon the results of the research.