AN ASSESSMENT OF NON-WOOD FOREST PRODUCT RESOURCES FOR THE DEVELOPMENT OF SUSTAINABLE COMMERCIAL EXTRACTION

J.F.W. van Dijk

Abstract

The development of commercial NWFP extraction is often seen as an option to increase the income of forest dwellers and to conserve the overall forest resource. For the development and promotion of NWFP exploitation, the most promising resources are those that can be exploited in a sustainable way and which have a high yielding resource base. Within the framework of the Tropenbos Cameroon Programme (TCP) socio-economic and ecological surveys on NWFPs are being carried out in the Bipindi-Akom II region in the South Province of Cameroon. In this region the extraction of NWFP for markets is not particularly well developed; income from NWFPs is very important, but it is mainly generated by selling products such as bush meat and palm wine for which the trade is restricted to the village level only.

A survey of NWFP resources was recently carried out in order to gain an insight into the abundance and distribution of NWFP species and to obtain indications of the impact of exploitation on the available resources. From the surveys, it appears that even in the relatively small study area of 200 000 ha, the variation in abundance and distribution of NWFPs is high. Several frequently used species are restricted to specific localities within the area. In addition, an important number of the (potentially) commercial NWFP species find their maximum density in secondary forest types, e.g. the condiments njansang from Ricinodendron heudelotii and mbongo, the fruits of the rhizomatous herb Aframomum citratum. Many of the species providing products with a high commercial value, such as a number of oleaginous seed producing trees as Baillonella toxisperma, Panda oleosa and Poga oleosa, appear to be rare while other important NWFP species, like the well known bush mango, Irvingia gabonensis, have a moderate density.

An additional survey has concentrated on a high value NWFP species that is exploited commercially. Garcinia lucida, the bark and seeds of which are used for palm wine production as well as for medicinal purposes, is a small understorey tree that grows on steep slopes in high density stands over areas of 2 to 3 km² in high altitude forest. However, in areas where the bark is stripped for commercial purposes, the mortality is very high and these populations are becoming seriously threatened.

Key words: NWFPs, Cameroon, resource inventory, commercial extraction.

1. Introduction

At present it is generally recognized that NWFPs play an important role for subsistence and cash income for local people. The awareness is growing that effective conservation and management of NWFP resources should be included in natural forest management in order to meet the present and future needs of local populations. Moreover, the development of commercial extraction of NWFPs is often seen as a way to ensure forest conservation whilst improving rural peoples' living standards (Ros-Tonen, Dijkman & Lammerts van Bueren, 1995, de Beer and McDermott, 1989). Undoubtedly, some NWFPs can be extracted sustainably from forest ecosystems (Peters, this volume) and the attainment of this forms the basis of many conservation and development research programmes, including that of Tropenbos.

In general, spatial and quantitative data on the availability of NWFP resources and their use are lacking. Only recently have researchers begun to develop methods to assess the potential sustainability of NWFP resources. Within the Tropenbos Cameroon Programme (TCP) a broadly-oriented NWFP project is currently being undertaken. The project consists of a NWFP identification and utilization survey, a resource inventory, a harvesting impact study and socio-economic surveys. This paper will mainly review the results of the ecological inventory which can be best described as a reconnaissance survey.

During this survey, the abundance and distribution of NWFP species were assessed and the data were used to derive indications of the impact of different types of exploitation on the availability of a wide range of resources. This paper will focus on those plant NWFPs which are commonly traded in the south of Cameroon.

2. The setting

2.1. The study site

The TCP study site covers an area of 200 000 ha in the South Province of Cameroon and is located between 50 km and 100 km from the coast. The study site corresponds with past and present timber concession areas of a Dutch logging company and because of this the majority of the study area has been logged-over several times. However, this timber exploitation has been carried out selectively and is therefore undertaken at a low intensity; it is estimated that about 0.7 trees/ha are harvested with around 14% of the forest surface being affected (Van Leersum, pers. comm.).

The study area is geomorphologically diverse. In the western area, plains dominate whereas the eastern area is fairly mountainous. Because of this varied topography, the altitude ranges from 40m above sea level (asl) in the western area to more than 1 000m asl in the eastern area. Annual rainfall varies from 1 800 mm to 2 500 mm with two distinct rainy seasons and two drier seasons. The vegetation changes gradually from Low Altitude Evergreen Forest of Lophira alata to Mid Altitude Evergreen Forest rich in Caesalpiniaceae (Letouzey, 1985).

The population density is rather low, ranging from 5-10 inhabitants per km² with 90% of the population being of Bantu origin. These people are generally referred to as "villagers" or "farmers". They practice shifting cultivation, cocoa cash crop farming and hunting and gathering. The majority of the Bantu people belong to the Bulu tribe. Other Bantu groups in the area are the closely-related Fang, the Bassa and the Ngumba. The remaining 10% of the population consists of Bagyeli (Pygmy) people. They are mainly hunter-gatherers, although they have adopted farming and, to a lesser extent, cocoa growing.

In terms of logistics, road conditions and public transport access within the area are very poor, particularly as the logging company which used to maintain the road network has now departed from the majority of the area.

2.2. NWFP extraction in the study area

An enormous variety of NWFPs are commonly used or are known to be useful. About 200 animal species and 500 plant species were identified, the latter accounting for 1 200 different uses (Van Dijk, in press). Many of these species have a commercial value and are traded widely throughout the region. Appendix I presents an overview of commercial NWFP species of plant origin which occur in the study site.

Income from the sale of NWFPs is generally of importance but it is mainly generated by the sale of products such as bush meat and palm wine and its derivatives, for which the trade is restricted to the village level. The extraction and sale of NWFPs for markets outside the village range area is not particularly well developed. Table 1 shows the relative contribution from selling NWFPs and agricultural produce to household incomes. The figures are based on preliminary data only from an ongoing survey among eleven Bagyeli and nineteen Bantu families, and does not include adjustment for the costs on inputs such as wire for snares, bullets, etc. These data reflect the results over a period of four months at the end of 1997 which correspond to the peak period of cocoa harvest and sale. It appears that even in this period the income from the sale of bush meat is often equal to the cash obtained by selling cocoa. Bush meat represents 75% of the contribution of NWFPs to income.

Table 1. Household income from the sale of NWFPs and agricultural produce in the period September 1997-January 1998.

	% of income	# of families involved (N=30)
Bush meat/Chops Fish Honey Basketry Forest foods Palm wine/liquor Pharmaceutical inputs	31.2 0.4 0.7 1.3 3.3 6.0 1.0	23 4 9 5 15 12 2
Total NTFPs	43.9
Cash crops Food crops Fruits/nuts Animal husbandry	42.9 12.6 0.6 0.0	14 19 18 1
Total agriculture	56.1

Table 2 gives an impression of the actual degree of commercialization of a number of marketable NWFPs, which, in general, is not very high. Out of the 23 products which are said to be sold to retailers or assemblers, 18 were harvested during the four months' period. From these 18 products, only 10 were effectively sold. Bush mango, Irvingia gabonensis, is one of the most frequently harvested and by far the most frequently sold product. However, most of the harvest is allocated for household consumption and only 20% of the total quantity collected is sold. Surprisingly, less than half of the total number of families and none of the Bagyeli families are involved in selling. Njansang, a condiment derived from the oily seeds of Ricinodendron heudelotii, is another NWFP which is marketed in significant quantities in the South of Cameroon (Ndoye et al., 1997).

In the TCP study area, however, only two out of the thirty participating families appear to be involved in the trade of this product. In light of the data presented in Table 2, one might state that there exists a potential in the study area to develop the commercial extraction of NWFPs. A resource inventory can contribute to examine the opportunities for such a development.

Table 2. The degree of commercialization of marketable NWFP products among family households in the period September 1997-January 1998.

Species name	# of families involved in extraction	# of families involved in trade	% of harvest sold	Total revenue in CFA (4 months)
Coula edulis Cola acuminata/nitida Cola lepidota Dacryodes edulis Dacryodes macrophylla Elaeis guineensis (fruit) Garcinia lucida Irvingia gabonensis Ricinodendron heudelotii Strophanthus gratus	28 8 14 22 12 30 10 28 2 2	3 3 3 6 3 2 1 11 1 2	5 30 30 10 10 < 1 10 20 100 100	15 000 2 500 3 850 11 000 2 000 2 800 400 79 000 5 800 32 500

3. Data collection and analysis

The objectives of the ecological survey were to gain insight into the relative abundance and distribution of NWFP species and to obtain some indications of the impact of agriculture, logging and NWFP harvesting on the availability of these resources. The methodology was based on the sample design as developed by Hall and Bawa (1993). One-hectare plots in the form of 1 km long and 10 m wide transects were enumerated for all individuals >10 cm dbh. In each transect, 10 plots of 10 x 10 m were laid out at regular distances in which all individuals with a minimum height of 1 m and a maximum dbh of 10 cm were enumerated. The smallest individuals of less than 1m in height were counted in 2 x 2 m subplots.

Twenty NWFP species were selected for which the dbh was accurately measured using a tape. For other NWFP species encountered along the transects, the dbh was estimated using global size classes of 10 cm. Land characteristics in terms of topography and vegetation structure were recorded. Identification of each species was undertaken by local field assistants, using vernacular names. In order to determine the scientific names, a number of the transects were inventoried simultaneously by botanists and the local assistants. Moreover, herbarium vouchers collected during the field work were later examined in the TCP herbarium and the National Herbarium of Cameroon in Yaoundé.

The location of the transects was selected on the basis of an aerial photo-interpretation developed for an overall land inventory in which the planation level along with shifting cultivation were distinguished as the main differentiating factors for soils and vegetation (Touber, 1993). The selection included the inventory of fields, fallow lands and cocoa plantations. A total of 32 hectares were inventoried.

Two of the 20 selected NWFP species were initially not found during the inventory; Aframomum citratum and Garcinia lucida. For these species separate surveys were carried out in several alternative sites that were indicated by local harvesters.

The data analysis was undertaken by post-stratification in order to examine the factors influencing the variation in abundance and distribution of the NWFP species. This analysis resulted in the determination of the following habitat types:

· undisturbed lowland forest (<350m asl)

· mid-elevation or intermediate forest (350-500m asl)

· high altitude forest (>500m asl)

· swamp forest

· secondary forest

· fields and young fallow lands

· cocoa plantations

4. NWFP resources

4.1. Abundance and distribution of marketable NWFP species

With respect to the development of commercial extraction, the most suitable NWFP resources are those which can be extracted sustainably and which have a high potential yield per unit of land, defined by the number of individuals and the potential productivity. In Table 3, an overview is presented of the abundance and distribution of actual and potential commercial NWFP species which occur in the TCP study site.

In general, the average densities of marketable NWFP species may be considered low with the large majority of these species not achieving densities of more than 5 stems/ha with a dbh of more than 10 cm in any part of the area. Some species, for example the high-value oleaginous seed bearing trees Baillonella toxisperma, and Poga oleosa, can be classified as being rare, with less than 1 tree/ha greater than 10cm dbh. In addition, the fact that many of these species are medium to large trees which reach their reproductive stage only at larger size classes, for example Baillonella toxisperma which starts flowering at an average dbh of 70 cm (Debroux, 1996), enables one to conclude that the productivity per unit of area is low for many NWFP species.

The difference between the average and maximum densities of the listed species is an indication of the absence or presence of specific (distinct) habitat requirements. Most of the recorded NWFP species are more or less evenly distributed in the area. Nevertheless, there are some species which are far more abundant in specific forest or habitat types and in some cases they are completely absent in large parts of the area. The distribution patterns of these species indicate that the variation in availability of NWFP resources even in a small area such as the TCP study site is very important.

It is obvious that it is not only undisturbed forests that contain NWFP resources. An important number of species find their maximum densities in secondary vegetation types. A number of typical pioneer species provide important NWFPs such as the oil palm Elaeis guineensis, Ricinodendron heudelotii (njansang) and Aframomum citratum. The latter species, in particular occurs most commonly along logging roads and skid trails.

However, the fact that some species show the highest abundance in cocoa plantations (e.g. Dacryodes macrophylla, Hexalobus crispiflorus, Strophanthus gratus and Alstonia boonei) is undoubtedly the result of effective conservation, enrichment planting or other type of human intervention.

Figure 1 presents some examples of the abundance and distribution of NWFP species with regard to the various habitat types. The bush mango tree, Irvingia gabonensis, is an example of a species which is more or less evenly distributed in the area (Figure 1a). The tree occurs in every habitat type and the abundance varies from 0.4-3.0 and 2.0-3.6 stems/ha with a dbh > 10 cm in secondary and undisturbed forest types respectively (see Figure 1a).

Table 3. Abundance and distribution of NWFP species with a known market value

Figure 1. Abundance and distribution of four important NWFP species. Fhi: Forest high > 500 m asl; Fin: forest 350-500 m asl; Flo: forest < 350 m asl; Swa: swamp forest; Slo: logged-over forest; Sec: secondary forest; Cpl: cocoa plantation; Fi/Fy: fields and young fallow lands.

Ricinodendron heudelotii (Figure 1b) is also represented in the entire area. As stated above, this species is a typical pioneer which explains its higher abundance in all the secondary forest types.

Garcinia lucida (Figure 1c), is completely absent in the western area of the study i.e. the low altitude zone (<500 m asl) and the species occurs at altitudes of >500 m above sea level and uniquely on very steep slopes. The high average density of more than 70 adult stems/ha is undoubtedly due to the fact that the species grows in high density stands in restricted areas of an average of 2 to 3 km² within these rather specific ecological conditions. A comparable although inverse distribution pattern is shown by the fruit bearing tree Cola lepidota (Figure 1d). This species is also restricted to steep slopes. However, it occurs mainly within the lower altitude zone of <500 m asl (2.2 versus 15-19 stems/ha with a dbh >10 cm).

The four examples presented also demonstrate the importance of cocoa plantations as a source of NWFPs. The presence of Garcinia lucida in cocoa plantations results from enrichment planting by farmers. However, the majority of these plantings have not been successful due to prolonged insect attacks and damage.

4.2. The sustainability of NWFP harvesting: The case of Garcinia lucida

In general, it is not easy to evaluate the impact of NWFP harvesting on plant populations. The most appropriate way to get a first indication on the impact of harvesting is to analyze the size-class distribution of the population (Peters, 1996; Peters, this volume; Hall and Bawa, 1993), preferably with regard to differences in exploitation intensity. However, even with a rather exhaustive inventory as described above, the low densities and the high variation in abundance and distribution of the species puts serious restrictions on this type of analysis. The number of individuals is not sufficient to get reliable results.

However, the data collected on Garcinia lucida during the supplementary survey undertaken for this species (Guedje, 1996) do permit the plotting of a reliable size-class distribution. In total, 494 reproductive individuals (dbh >5 cm) were enumerated in four different sites over an area of 8.4 ha.

As stated above, this small understorey tree grows in restricted areas. The distance between the villages and these stands is rather important as the species is considered open-access resource. The extraction of the bark for commercial purposes is undertaken by stripping the bark from the entire bole, essentially ring-barking the stem and resulting in the immediate death of the individual. The mortality in the various sample locations ranged from 5 to 35% (with an average of 19%). Figure 2. presents the size class distribution after post stratification of the data in two sets by regrouping the transects in which the mortality was very high (more than 30%) and those in which the mortality was relatively low (less than 10%). This artificial grouping reflects the impacts of high and low exploitation.

The size-class distributions show clearly that it is preferable that the largest trees are exploited. In the dbh class of 5-10 cm hardly any tree is de-barked and, as expected, in the dbh class of >15 cm very few of the trees remain unaffected by bark exploitation. It can be confidently assumed that the elimination of the largest, and therefore most productive, trees has serious consequences for the regeneration capacity of the populations. However, at present, both exploitation regimes show an inverse J-shaped size-class distribution, indicative for an adequate recruitment. An ongoing monitoring programme on the population dynamics of Garcinia lucida will provide more information on the long term effects of the actual exploitation system. Moreover, future work will attempt to determine the sustainable harvesting levels by an experimental approach. A number of selected trees were subjected to various exploitation techniques and harvesting levels. Parameters on growth, health, reproduction and regeneration are being monitored.

Figure 2. The size class-distribution of Garcinia lucida including healthy trees (blank), de-barked trees (hatched) and dead trees (solid): a. lightly exploited sites; b. heavily exploited sites.

5. Conclusions and discussion

5.1. Methodological aspects

The adopted methodology of selecting sample locations on the basis of an aerial photo-interpretation is very appropriate for quickly gaining insight into the abundance and distribution of NWFP species in a heterogeneous area such as the TCP study site. Working mainly with vernacular names and "local taxonomists" provides generally sufficient detail, although some problems occurred. In some cases, several species are regrouped under the same vernacular name. This is for example the case for rattan species for which the generic term nlong is used. The same name is given to a particular rattan species, Eremospatha macrocarpa. The fact that this coincidence only became apparent in the course of the inventory, made it impossible to make a proper distinction whether nlong had been used to indicate this particular species or small-diameter canes as a whole. A second problem concerns those species which are hardly used but which could be of economic interest. They were not always recognized by the local assistants. This occurred with Gnetum spp. (providing the vegetable ocok or eru), which is very frequently consumed and heavily exploited elsewhere in Cameroon.

The attempt to combine the determination of variation in the abundance and distribution of NWFP species and obtaining indications on the impact of exploitation on the availability of resources was less successful. Due to the low densities of many important species the number of individuals is not sufficient for reliable size-class distribution analysis. Other methods should be applied to obtain indications of the impact of harvesting.

5.2. NWFP resources and their utilization: consequences for the development of extraction

In view of the results obtained from the various executed and/or ongoing ecological and socio-economic surveys undertaken in the study area of the Tropenbos Cameroon Programme, a number of preliminary conclusions can be drawn:

· The sale of NWFPs is almost as important to the income of households as the sale of agricultural products. However, the sale of NWFPs is mainly restricted to the village level and the income derived is dominated by the sale of bush meat (up to 75%);

· The products of an estimated 23 NWFP species are said to be of commercial value. Yet, despite the fact during a four month observation period many of these were seen to be exploited, only four of these species contributed effectively to the income of the households surveyed. These were Irvingia gabonensis, Strophantus gratus, Coula edulis and Dacryodes macrophylla;

· Even for a relatively small study area of 200 000 hectares, a large variation in abundance and distribution is observed,which implies that not every household has the same access to the same resources, making generalisations and extrapolations impossible;

· The densities of the majority of NWFP species with a commercial value may be regarded as low to moderate. Several species have less than one productive individual per hectare, which sets important constraints on the availability of resources and to the development of efficient extraction in natural forest;

· Those species which have a higher average density, and hence a higher potential yield, ^{for the whole study area are mostly restricted to specific habitat types. Except for those occurring in secondary forests, they are often absent in large parts of the area and therefore not available to every household;}

· Secondary forests and cocoa plantations are very important areas for NWFP extraction. The impact of active management of NWFP resources in, for example, cocoa plantations is reflected in the results of the ecological survey. Although the scale of active management is rather limited at present, it might indicate opportunities for further development of the availability of NWFP resources in such anthropogenic areas;

· The example of the impact of harvesting on Garcinia lucida populations demonstrates that high density stands of commercial NWFP trees considered as open-access resources are very vulnerable to over-harvesting. This phenomena is illustrated by the recorded high mortality rate. However, the fact that individuals in the smaller size distribution classes are reproductive, and that it is mainly the bigger trees that are de-barked, the impact might be less serious than expected. Further research to determine the sustainability of the exploitation of this species is continuing.

5.3. Which strategies to follow for a sustainable development of NWFP extraction?

The surveys demonstrated that there are quite a number of NWFP species with an economic value which can be extracted in a sustainable way. However, from the results obtained so far, it seems that the local people are not very interested in the extraction of NWFPs which have no subsistence value, like Ricinodendron heudelotii and Strophantus gratus and are only valuable on a commercial basis. Less than 10% of the families interviewed are involved in the exploitation of these species and harvesting and sale seems to be restricted to specialists.

In contrast, almost all families are involved in the harvesting of NWFPs with an important subsistence value, such as Coula edulis and Irvingia gabonensis with additional families generating income from the sale of these products. In other words, people seem to prefer to trade essentially the surplus of the quantity needed for auto-consumption. The still relatively low level of commercialization of these products is likely to be caused by restrictions on the harvesting capacity (for example time constraints, accessibility, property rights, subsistence needs) and on the marketing opportunities. The development of the extraction of these products seems to offer the most promising perspective for the improvement of income and sustainable resources use.

However, it is doubtful whether adequate resources can be found for a sustainable and efficient increase of existing extraction levels. In this respect, the compatibility of raising household incomes from NWFP extraction from natural stands whilst ensuring conservation of the forest as a whole might need to be reviewed. This is not only due to the risks of oft-practiced destructive harvesting practices, but also because of the low densities of many NWFP species in natural forest. Therefore, the intensification of the NWFP management on anthropogenic land types and the promotion of the domestication of appropriate NWFP species might be the most promising options for increasing household incomes whilst removing harvest pressure on the wild resource.

References

De Beer, J.H. & M.J. McDermott. 1989. The economic value of non-timber forest products in Southeast Asia. Amsterdam, IUCN.

Debroux L. 1996. La régéneration naturelle et la gestion de populations d'arbres: L'étude du Moabi (Baillonella toxisperma) dans la forêt du Dja, Cameroun. Synthèse à mi-parcours des travaux de doctorat, Gembloux, Faculté Universitaire des Sciences Agronomique de Gembloux.

Guedje, N. 1996. Evaluation de quelques aspects écologiques de Garcinia lucida, Aframomum citratum et A. melegueta dans la région de Bipindi-Akom II du Sud Cameroun. Yaoundé/Kribi, Rapport d'étudiant, Université de Yaoundé I/Programme Tropenbos Cameroun.

Hall, P. & K. Bawa, 1993. Methods to assess the impact of extraction of non-timber tropical forest products on plant populations. Economic Botany 47(3): 234-247.

Letouzey, R. 1985. Notice de la carte phytogeographique du Cameroun. Toulouse, France, Institut de la Carte Internationale de la Vegatation.

Ndoye, O., M. Ruiz Pérez & A. Eyebe, 1997. The markets of non-timber forest products in the humid forest zone of Cameroon. ODI-Network Paper 22c.

Peters, C.M., 1996. The ecology and management of non-timber forest resources. World Bank Technical Paper 322, The World Bank, Washington DC.

Ros-Tonen, M., W. Dijkman & E. Lammerts van Bueren, 1995. Commercial and sustainable extraction of non-timber forest products. Towards a policy and management oriented research strategy. Wageningen, Tropenbos Foundation.

Touber, L.,1993. Report on a mission in support of the Land Inventory Project in the Tropenbos Cameroon Programme. Wageningen, International Activities Report 32, SC-DLO.

Van Dijk, J.F.W. [in press]. Non-timber forest products resources in the Bipindi-Akom II area, South Cameroon. An economic and ecological assessment. Kribi, Tropenbos Cameroon Documents X.

APPENDIX 1: Commercialised and commerciable NWFP species occuring in the TCP study site