Assessment methods for non-timber forest products in off-reserve forests : case study of Goaso district, Ghana
von Francis Bih, 2007
This study was conducted with the objectives to develop assessment method for selected non-timber forest products (NTFPs) in off-reserve areas of the Goaso forest district, Ghana. In this frame the relative efficiency of adaptive cluster sampling method for inventorying selected NTFPs was investigated. The study also inventoried the selected species, explored their potential for income generation and sought how factors like accessibility influence their spatial distribution and abundance.
Literature on the significance of NTFPs to local people echoed the need for the inclusion of NTFPs in forest management. The different methods of assessing NTFPs were reviewed to establish that methods for timber inventory which were inefficient for NTFPs have been adapted for NTFP inventory and thus the need for methods which would be more efficient for NTFPs. Adaptive cluster sampling method which has been suggested to be efficient for clumped and rare species was developed and tested.
Community and market survey was conducted first to select important NTFPs obtainable in off-reserves of the study area. The importance here was based on the use of the NTFP for subsistence and/or income generation. Six communities in the study area were randomly selected and a random number of households in each community were interviewed. Two main markets in the study area were also visited to obtain an overview of NTFPs used for income generation. The survey sought to include local knowledge on ecology, harvesting techniques and history of abundance of the selected species in the inventory design. Through the results obtained the species were grouped into two, based on their distribution. Three species, Calamus deeratus, bamboo, and raphia palms were mainly found to be distributed in patches around streams and rivers and tree species or shrubs scattered across the mosaic off-reserve cover types. A pilot inventory was conducted and the results factored into the design of the main inventory.
Adaptive cluster sampling with a systematic base was designed to inventory the three species distributed around water bodies. A systematic base was used to allow comparison of the efficiency of systematic sampling and adaptive cluster sampling methods because systematic sampling is widely used in forest inventories. Arcview was used to randomly select 2% of streams and rivers in the study area. Using an initially chosen random point, 50m x 50m plots at 500m interval were systematically made along each selected stream or river. A total of 69 systematic plots were measured. Where any of the species of interest was found, 50m x 50m contiguous adaptive plots were made till no species of interest was observed in the edge units. Bamboo clumps were enumerated by counting the number of clumps, measuring the base area and counting the number of culms. A linear regression model was formulated to determine the validity of using the base area of a bamboo clump to predict the number of culms. The number of plants of Calamus deeratus and raphia palms were counted and grouped into mature and immature plants. Estimates were calculated using the Hansen-Hurwitz estimator. The relative efficiencies of adaptive cluster and systematic sampling methods, for inventorying each of the three species, were determined by comparing their sampling errors for each species.
To obtain an overview of the abundance and distribution of the other NTFP species in the study area, a systematic cluster sampling was used. Systematic cluster sampling method was used because of the relative cost efficiency of cluster sampling when inventorying larger areas. A 7km x 7km square grid was overlaid on a geo-referenced map of the study area. At each grid point four 50m x 50m plots were systematically made at the corners of a 100m x 100m square. A total of 128 plots were measured. The land cover type within which each plot occurred was identified and the diameter of all tree species above 10cm diameter at breast height (dbh) or at 30cm above the buttress were measured with a diameter tape. Each tree species was identified with its local and scientific name and grouped into commercial and non-commercial species. Where a cluster unit fell in a tree cover, 10m x 10m subplot was made at a corner of the 50m x 50m plot to access regeneration. Tree species of less than 10cm dbh to saplings of 1m height and above were recorded as regeneration.
The results showed that adaptive cluster sampling was about eight times more efficient as systematic sampling for all the three species inventoried. However, because the efficiency of adaptive cluster sampling depends on the density and clustering degree of the species which is usually not known before the start of the inventory, application of adaptive cluster sampling cannot be transferred easily from one area to another. The method should be applied with circumspection and pilot studies could be used to obtain the desired information about the species before main inventories are undertaken. The study also suggested testing of adaptive cluster sampling in the forest reserves because densities and degree of clustering of the species could be different within the reserves. Although statistically significant relation was found to exist between the base area of bamboo clump and the number of culms, base area of bamboo clumps was not found to be practically good predictive parameter of the number of culms in a bamboo clump. The validity of the linear regression model may have been weakened by stumps of harvested culms which formed part of the base area but were not counted.
When the inventory with systematic cluster sampling method was post stratified according to the cover types, single cluster units of different clusters occurred in different land cover types making the number of plots per cluster vary. The sample sizes of the different land cover types were thus different resulting in different precision levels in each cover type. A land cover map would be required for pre-stratification to make it more efficient.
The densities of the species were found to be low. Change in the land cover due conversion of the species habitats to agriculture may have accounted for this. Based on the inventory results, the capability of the off-reserves to provide sustained levels of the selected NTFP species was bleak as regeneration capability of the species could be hampered by agricultural conversion and the spread of Chlomolaena odorata.
Keywords: Ghana, NTFP, adaptive cluster sampling, Forest inventory