An Indigenous Pacific Island Agroforestry System: Pohnpei Island

The indigenous agroforestry system on Pohnpei was studied using circular plots laid out in transect across 57 randomly-selected farms. Data were collected on species and cultivar presence, size, density, frequency, as well as soil type, slope, aspect, and other related information. Through farmer interviews, farm family demographic data was also recorded. Seasonality of major crops was observed. Analysis shows indigenous agroforestry on Pohnpei to be a complex, but extremely well ecologically and culturally adapted, production system. Indigenous agroforestry is a dominant feature of both the landscape and culture on Pohnpei, the result of more than 2,500 years of development and refinement (Haan 1984). During this time, numerous crop and technology introductions have been made through continued waves of migration, and more recently, through direct and indirect efforts of colonial administrations 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. Researcher, Land Grant Programs, College of Micronesia, Kolonia, Pohnpei, F.S.M. 96941; Professor, Department of Soil Science and Agronomy, University of Hawaii at Manoa, Honolulu, Hawaii 96822. (Barrau 1961). Currently, agroforestry both employs and pro­ vides sustenance to a large majority of the Pohnpei population. The island of Pohnpei is located at 6°54' N latitude and 158°14'E longitude in the Caroline Islands group, about 4983 km southwest of the Hawaiian islands (fig. 1). It is the highest (772 m) and second largest (355 km) in the group and one of the few high islands. The island is of volcanic origin and is about five million years old (Keating and others 1984). Rainfall is high and temporally well-distributed, with an average of 4820 mm and 300 rainy days per year (NOAA 1987). At higher interior elevations, rainfall is estimated to reach 7,500 mm (Laird 1982, van der Brug 1984). Temperatures average 27°C year-round and humidity is high (NOAA 1987). The island is surrounded by a barrier reef and lagoon, with extensive mangrove forest devel­ opment around most of the shoreline. Pohnpei Island is typically volcanic, with a majority of the land area characterized as steep and mountainous. Vegetation is mainly upland forest (55.5 percent), mostly in the interior. Coastal areas and lower slopes are characterized by agroforest (33.4 percent) and secondary vegetation (5.2 percent). Agroforestry has been expanding rapidly in the last two decades, replacing primary forest and secondary vegetation (MacLean and others 1986). Figure 1-Location of Pohnpei in the Caroline Islands, Micronesia 42 USDA Forest Service Gen. Tech. Rep. PSW-GTR-140. 1993. Soils in areas under agroforestry are characterized by Typic Acrorthoxes in the lowlands and Typic Dystropepts on mountain slopes, with a few small areas of Typic Humitropepts (Laird 1982). Soils in the upland mountainous areas are generally deep, well drained, and commonly very stony. Use of these areas is limited by steep slopes and stoniness. Nearly level or gently sloping soils are generally moderately deep and moderately well-drained. Low fertility and wetness are limitations. Bottom land soils are generally poorly drained and are limited by wet­ ness (Laird 1982). Methods Selection of Survey Sites The area of this study was the entire agroforestry area on the island of Pohnpei. MacLean and others (1986), using aerial photos and ground surveys, estimated indigenous agroforestry to cover about 33.4 percent of the total land area of Pohnpei, or 11,865 ha as of 1984. It was desired to sample about 1 percent of the agroforest, so based on the reported area of agroforestry and assuming 2.5 ha as the average land parcel size, it was determined that about 50 farms would be surveyed. A map of Pohnpei was overlaid with a grid of intersecting lines corresponding to every 0.5 km, then 100 random pairs of numbers were generated, corresponding to x,y coordinates of farm survey points. Points that fell in the lagoon, mangrove, or uninhabited jungle areas of the island were discarded and farms on or nearest the remaining 57 survey points were identified (see map, fig. 2). Field Survey Methods In designing field methods, it was necessary to take into account that persons other than family members are not gener­ ally allowed to enter onto someone's land on Pohnpei. To allevi­ ate this, all farmers were visited several weeks before the actual survey took place. The local extension agent explained the pur­ pose of the study, people to be involved, and what would be done. If the farmer was agreeable, a date for the survey was set. Surveyors were limited to two people, the senior author and the extension agent, and survey methods were designed so that the farmer could accompany us on the survey. Figure 2-Map of Pohnpei Island showing farm survey sites USDA Forest Service Gen. Tech. Rep. PSW-GTR-140. 1993. 43 Upon arriving at a farm, the head of household, usually with his/her family, was interviewed using the prepared interview protocol (see Appendix). Then a rough land map, showing prop­ erty boundaries, buildings, and vegetation types, was sketched. The survey route was then determined with the farmer before starting. A systematic plot lay-out was used, working along a compass line from corner to corner of the survey farm, passing through or near the center, with plots taken at 40 meter centers along the line. If the distance across the farm was too short to make 10 plots, a second compass line branching at a right angle from the first line near the farm center was set and the remaining plots laid out on 40 meter centers along this line. Circular plots of 8 meter radius (201 m) were used for ease of layout. Slope and aspect were recorded with a clinometer and compass, respectively, and then weeds (grasses, ferns, and rec­ ognized weed species) were recorded by visual estimate of percent cover. All other species were recorded by local name, cultivar (if any), number and heights. On breadfruit trees, d.b.h. was measured. For bananas, taro, and sakau (Piper methysticum), number of stems were counted, and for yams, number of vines were recorded. This was repeated for each plot (see farm survey form in Appendix.) Species Identification Through farmer interviews, observation, and literature review, important data on each crop species were collected, in­ cluding genus and species (Glassman 1952, Falanruw and oth­ ers, in press), Pohnpei name (Rehg and Lawrence 1979, Falanruw and others, in press), life cycle (annual or perennial), seasonality, products, period of introduction (Glassman 1952, Bascom 1965), vegetation type group (Glassman 1952, MacLean and others 1986, Falanruw and others 1987), and other data, such as num­ ber of cultivars. Frequencies (percent of farms on which species occurred) and overall individuals per hectare were calculated for each species.