Documentation of Indigenous Pacific Agroforestry Systems: A Review of Methodologies

Recent interest in indigenous agroforestry has led to a need for documentation of these systems. However, previous work is very limited, and few methodologies are well-known or widely accepted. This paper outlines various methodologies (including sampling methods, data to be collected, and considerations in analysis) for documenting structure and productivity of indigenous agroforestry systems, using references to previous documentation studies carried out in other parts of the world. Interest in indigenous agricultural systems has grown enormously in the past decade or so, largely as the result of the shortcomings of the “green revolution.” The realization that many traditional systems are well integrated ecologically, economically, and socially at the local level has given a new impetus to research. Also, these systems offer valuable insights into adaptations to local environmental and cultural constraints. There is also a strong possibility that research into these agroforestry systems will lead to their improvement in terms of production and other development needs of the respective islands, and that information gained will be valuable in finding solutions to agricultural research problems in other areas. As a result, many scientists now see indigenous agriculture as dynamic systems which can serve as foundations for development efforts rather than as static obstacles to agricultural intensification. Unfortunately, due to largely being ignored in the past, research methods for studying indigenous agriculture have only just begun to be developed. The inherent difficulty in studying indigenous agriculture systems is the relative complexity of traditional mixed cropping systems compared to “western” agricultural practices. Indigenous agroforestry systems are the product of both natural and anthropogenic influences, so they are different than either natural ecosystems or modem agricultural ecosystems. Current research methods developed in various disciplines such as vegetation ecology, forestry, and agronomy need to be combined in their study. Collection of data from indigenous agroforestry systems is further complicated due to the lack of existing data on many indigenous crops and animals, the long-term nature of the perennial components, the subsistence nature of many agroforestry products, and the variation within and between farms and regions. Cultural practices and restrictions can also hinder research work in some areas. Finally, the lack of trained manpower and financial constraints must also be considered in any research project in developing countries. In the Pacific, most work, with a few exceptions (e.g., Handy 1940; Barrau 1958, 1961) has been in the form of general descriptions, with agriculture serving only as a component of the 1 An abbreviated version of this paper was presented at the Workshop on Research Methodologies and Applications for Pacific Island Agroforestry, July 16-20, 1990, Kolonia, Pohnpei, Federated States of Micronesia. 2 Researcher, College of Micronesia Land Grant Programs, Kolonia, Pohnpei, Federated States of Micronesia 96941. more general social or economic systems. Other work has focused on strictly botanical studies of natural or cultural vegetation (e.g., Fosberg 1959). It is not until quite recently that a concerted effort has been made to systematically and quantitatively analyze traditional Pacific island agriculture (e.g., Thaman 1975, Manner 1976, Raynor 1989). The goal of initial research should be to develop a general quantitative overview of the local indigenous agroforestry system. Among the data desired are floristic composition, vertical and horizontal structure, and phenology of agroforests, as well as information on production, seasonality, and yields of major products. Often, related information on farmer and farm family demographics, land use and tenure, and labor input and allocation is also desired. Methods for Characterizing Structure of Agroforestry Systems The initial focus of studies of indigenous agroforestry systems should be to characterize basic agroforest structure. Ecosystems have three basic structural components―vertical, horizontal, and temporal (Whittaker 1975). Vertical structure is the height and stratification of plants in the system, depending much on the floristic composition and light relations within and between species. Horizontal structure is the vegetation organization on the ground, affected by the environment, species interrelationships, and human management. Temporal or time relations include the phenology, age, and long-term development of the agroforest stand. Sampling is a key consideration in collecting indigenous agroforest structural data. It is usually impossible to measure the entire area where a system is practiced, so data must be recorded for samples of the agroforest, and then extrapolated to the larger area. Sample size also affects the precision of estimates obtained by sampling. A larger sample size gives greater precision, but often constraints of time and money limit the number of actual sampling sites. Selecting an unbiased sample is also important, especially if results are to be extrapolated to the general population. One way to generate a random, unbiased sample is the selection of sampling points (farms) on a map using a coordinate grid and random numbers. If areas to be studied are large, farms can be selected first using this method, then a systematic sub-sampling lay-out of smaller plots can be employed at each farm.