Previous PageTable Of ContentsNext Page


Charles M. Peters


Tropical trees and forests exhibit several ecological characteristics that make the sustainable exploitation of non-wood forest products (NWFPs) a more difficult proposition than it might first appear. The most important of these are the high diversity and low density of conspecific individuals, the strong reliance on animals for pollination and seed dispersal, the high mortality and low establishment rate during the seedling stage, and the sensitivity of population structure to changes in the level of natural regeneration. Ignoring these ecological constraints can lead to over-exploitation, resource degradation, and the gradual elimination of a species from the forest.

At a very basic level, designing systems for the sustainable exploitation of NWFPs requires two pieces of ecological information. We need to know the density and size-class structure of the plant populations that produce the NWFP (i.e. the resource stock), and we need to know how much of the desired resource these populations produce in a given period of time (i.e. the yield). The first data set is collected by means of a systematic forest inventory. The second data set requires periodic observations of the growth and productivity of a subsample of marked individuals of varying size. The relationship between resource stock and yield can be used to estimate a sustainable harvest level for many NWFPs. To ensure that this intensity of exploitation can be maintained over time, seedling and sapling numbers should be carefully monitored and harvest levels adjusted as necessary to provide a continual flow of new seedlings into the population. Data collection and monitoring activities are most effective when conducted by local communities that have been specifically trained for this purpose.

Key words: Ecological impacts, forest management, non-wood forest products, sustainability

1. Introduction

This paper presents the observations of a plant ecologist who has spent almost twenty years studying the ecology, use, and management of non-wood forest products (NWFPs) in the tropics. In fact, when I first started studying these plant resources they were called "minor" forest products and nobody really paid much attention to them. This situation has changed drastically in recent years. NWFPs are big business now, and numerous efforts are currently underway to promote the exploitation of these valuable and highly-publicised resources. Much attention has often focused on the economic side of things, e.g. developing markets for different products, implementing local processing and value-added strategies, and ensuring the equitable distribution of income generated. Social issues such as securing land tenure or usufruct rights for collector groups have also played a prominent role. It is somewhat surprising, however, that the ecological factors associated with the exploitation of NWFPs have only rarely been addressed. Maintaining a reliable income flow over time from a tract of forest requires that the forest resources upon which this flow is based be maintained as well. If these resources are depleted through over-exploitation, destructive harvesting, or poor management, no new market, cottage industry or land-tenure system will make very much difference. In the long term, ecology is arguably the bottom line for sustainability.

The purpose of this paper is threefold:

· to challenge the common assumption that the commercial harvesting of NWFPs has minimal impact on a tropical forest;

· to propose some simple data collection procedures for monitoring the ecological sustainability of forest exploitation, and

· to highlight several problems that might hinder the implementation of these monitoring activities.

Given the author's previous experience and interests, the discussion focuses primarily on techniques for the participatory management of NWFPs by local communities. Within the context of this paper, a sustainable system for exploiting NWFPs is defined as one in which fruits, nuts, latex, and other non-wood resources can be harvested indefinitely from a limited area of forest with negligible impact on the structure and dynamics of the plant populations being exploited.

2. Ecological impacts of forest use: The myth

Human cultures have developed a variety of different ways to use forest vegetation. Each form of land-use carries with it a particular suite of ecological costs. Perhaps the most intensive and costly way to use a forest is to cut it down, burn it, and plant something else (e.g. timber trees, agricultural crops, pasture grasses) on the site. The ecological impacts of forest conversion are immediate, highly visible, and, in most cases, highly severe. Current research in tropical forests suggests that the most important of these impacts include:

· the loss of biomass and species diversity

· the release of CO² and other greenhouse gases

· disruption of nutrient and hydrological cycles

· soil loss through erosion

· increased local temperatures and decreased local rainfall

To put some of these consequences in perspective, a one hectare tract of primary forest in the Brazilian Amazon may contain more than 200 tree species (>10cm dbh) and present an above ground living biomass of about 300 tons/hectare (Brown et al., 1995). Cutting and burning this forest would eliminate most of the biodiversity and release approximately 150 tons of carbon/hectare in the form of carbon dioxide and other greenhouse gases (Keller et al., 1991). The removal of the vegetation cover would increase water movement, soil erosion, and nutrient loss, decrease evapo-transpiration and total ecosystem productivity (Jordan, 1987), and potentially modify local climatic regimes because of the increased reflectance of solar radiation (Shukla et al., 1990). The site would be characterised by stumps, blackened tree trunks and, depending on the topography, a growing network of eroding gullies. It is obvious to the most casual observer that a major ecological disturbance has occurred here.

Another common use of forests is to selectively cut and remove the desirable timber trees. Although certainly less damaging than total forest conversion, selective logging is also known to produce a number of ecological repercussions. The most conspicuous of these are:

· Loss of some plant and animal species

· Damage to residual trees

· Soil loss through erosion

· Loss of nutrients through stem removal

· Change in forest structure and increase in light levels

A major problem with selective logging in tropical forests is that the crowns of many large canopy trees are lashed to those of their neighbours by a profusion of vines, lianas and climbers. When selected timber trees are felled, other canopy species are pulled down and the whole woody mass crashes through the lower canopy, snapping tree boles, breaking branches, and flattening a considerable proportion of the forest understory. Harvesting even a small number of stems can destroy up to 55% of the residual stand and seriously damage an additional 3% to 6% of the standing trees (Burgess, 1971; Johns, 1988). Associated impacts include soil compaction, decreased infiltration of water, increased rate of soil loss from erosion, disruption of local animal populations, increased susceptibility to fire (Uhl et al., 1988), and nutrient loss from the removal of sawlogs. Commercial tree felling produces a notable impact on a forest ecosystem, and the physical evidence of this disturbance is immediately apparent and persists in the form of logging roads, skid trails, and scattered stumps for many years.

A final form of forest use that has attracted a lot of attention recently involves the selective harvest of fruits, nuts, latex and other non-wood resources. Although relatively benign when compared with forest clearing and selective logging, this activity also produces a number of ecological impacts including:

· gradual reduction in the vigour of harvest plants

· decrease in rate of seedling establishment of harvest species

· potential disruption of local animal populations

· nutrient loss from harvested material

At first glance, these impacts seem insignificant. The harvest of non-wood forest products does not necessarily kill the plant, compact the soil, increase erosion, or cause a notable change in the structure and function of the forest. A forest exploited for non-wood resources, unlike a logged-over forest, maintains the appearance of being undisturbed. It is easy to overlook the subtle impacts of NWFP harvest and to assume a priori that this activity is something that can be done repeatedly, year after year, on a sustainable basis. This ubiquitous idea, or some variant of it, has appeared in books, scientific papers, conference proceedings, grant proposals, magazine articles, newspaper stories, on television and radio shows, in the annual reports of private companies, and even on the back of cereal boxes and ice cream cartons. Unfortunately, in the great majority of cases, this assumption is patently incorrect.

3. Some facts about tropical trees and forests

Tropical forests exhibit several ecological characteristics that make the sustainable exploitation of non-wood resources a more difficult proposition than it might first appear. One of the most fundamental and well-known features of these forests is their great species richness, or large number of plant species per unit area. To illustrate this point specifically for trees, floristic data collected from small tracts of tropical forests around the world are shown in Table 1. Although there is much variability from site to site, the results from these surveys show that tropical forests are extremely diverse and may contain from 100 to over 300 species of trees per hectare.

Table 1. Number of tree species (>10cm in diameter) recorded in small tracts of tropical forest.


Sample area (hectares)

Number of species >10cm in diameter


Cuyabeno, Ecuador



Valencia et al., 1994

Mishana, Peru



Gentry, 1988

Lambir, Sarawak



Ashton, 1984

Bajo Calima, Colombia



Faber-Lagendoen & Gentry, 1991

Sungei Menyala, Malaysia



Manokaran & Kochummen, 1987

Wanariset, East Kalimantan



Kartawinata et al., 1981

Gunung Mulu, Sarawak



Proctor et al., 1984

Campo, Cameroon



Sunderland et al., 1997

Rio Xingu, Brazil



Campbell et al., 1986

Barro Colorado, Panama



Lang & Knight, 1983

Oveng, Gabon



Reitsma, 1988

From a commercial standpoint, the high diversity of tropical forests is a mixed blessing. On the one hand, forests containing a large number of different species usually contain an equally diverse assortment of useful plant species, i.e. species richness and resource richness are usually correlated. The great interest in tropical forests as an undiscovered source of new foods, materials, and medicines is largely in response to the magnitude of the species pool in these ecosystems. Unfortunately, an additional correlate to high species diversity is that the individuals of a given species usually occur at very low densities. There is a limit to the total number of trees than can be packed into a hectare of tropical forest. If you have a large number of species, each species can only be represented by a few individuals.

This tendency of high species diversity coupled with low species density is illustrated in Figure 1 using inventory data collected from small tracts of forest in Brazil and Sarawak. As shown in the histogram, the great majority of the species at each site are represented by only one or two trees; less than ten percent of the species exhibited densities greater than four trees/hectare. Although there may be an abundance of resources in tropical forests, most of them are scattered throughout the forest at extremely low densities. Low density resources are difficult for collectors to locate, they require lengthy travel times, produce a low-yield per unit area, and they are extremely susceptible to over-exploitation. Clearly, none of these are desirable characteristics in a forest resource.

A second characteristic of tropical trees that represents a stumbling block to sustainability concerns the way that they move their pollen and disperse their seeds. The low density and scattered distribution of individuals in many tropical tree populations greatly complicates the process of pollination. Given that the distance between conspecific individuals may be greater than 100 meters in some cases, moving pollen from the flowers of one tree to another can be a difficult proposition. Many tropical trees have overcome this problem by co-evolving relationships with a variety of animals, ranging from tiny thrips and midges to bees and large bats, that act as long-distance pollen vectors. These relationships can be quite specific, with one type of insect being solely responsible for pollinating the flowers of a particular species, or even genus, of forest trees (e.g. Wiebes, 1979). The use of biotic vectors to transfer pollen is apparently the norm in tropical forests, and recent studies in Costa Rica (Bawa et al., 1985) suggest that over 96% of the local tree species are pollinated exclusively by animals.

Figure 1. Densities of different tree species within small tracts of tropical forest. Inventory data from Semengoh, Sarawak based on a 4.0 hectare sample plot (Ashton, 1984); Manaus based on a 1.0 hectare sample plot (from Prance).

Animals also play a very important role in dispersing the seeds produced by tropical trees. Studies conducted in Rio Palenque, Ecuador (Gentry, 1982), for example, have shown that 93% of the canopy trees produce fruit adapted for consumption by birds and mammals, while Croat (1978) estimates that 78% of the canopy trees and 87% of the subcanopy trees at Barro Colorado Island in Panama have animal dispersed fruits. These animals may either remove fruit and seeds directly from the tree (primary dispersers), or they may forage on fruits that have already fallen to the ground and split open (secondary dispersers).

The important lesson to be gained from these findings is that the production of fruits, seeds, and seedlings in tropical forests necessarily involves the collaboration of animals. Although it is very easy to overlook this fact, or to view forest animals solely as pests that damage or consume large quantities of fruit, sustainable resource use in tropical forests ultimately depends on the continual availability of pollinators and seed dispersers. In simple terms, no pollination means no fruits, no fruits and/or no dispersers mean no established seedlings, and no established seedlings means no next generation, no products, no profits and, importantly, no sustainability.

A final characteristic of many tropical tree species is that they have a very difficult time recruiting new seedlings into their populations. Even given abundant pollination, fruit set, and dispersal, there is still a very, very small probability that a seedling will become successfully established in the forest. The seed must avoid being eaten, it must encounter the appropriate light, soil moisture and nutrient conditions for germination, and it must be able to germinate and grow faster than the seeds of all other species that are competing to establish themselves on that microsite. The young seedling must then stay free of pathogens, be able to recuperate from the damage caused by herbivores, avoid falling branches and other hazards, and continue to photosynthesise and push its way upward into the forest canopy. Not surprisingly, mortality during the early stages of the life cycle of a tropical plant is extremely high.

A graphic example of the seedling mortality experienced by tropical trees is provided by the four survivorship curves shown in Figure 2. Brosimum alicastrum is a widely distributed canopy tree from the Neotropics (Peters, 1990a), Shorea curtisii and Shorea multiflora are dominant tree species in Southeast Asia (Turner, 1990), and Grias peruviana is an abundant lower canopy tree in western Amazonia (Peters, 1990b).

Figure 2. Seedling survivorship curves for Brosimum alicastrum, Shorea curtisii, Shorea multiflora, and Grias peruviana. Histograms show the percentage of seedlings surviving during the first year following germination. Data for B. alicastrum were collected in Veracruz, Mexico (Peters, 1990a), S. curtisii and S. multiflora were studied in Peninsula Malaysia (Turner, 1990), and G. peruviana data were collected in Peruvian Amazonia (Peters, 1990b).

As is illustrated in these histograms, seedling survival by these four species during the first twelve months following seedfall ranges from a high of 22% for S. curtisii to a low of 3% for B. alicastrum. Half-lives, or the time required to kill off 50 percent of the initial cohort, vary from two to five months. Taking into account seed predation and germination failure, less than 0.1% of the seeds produced by B. alicastrum become established seedlings (Peters, 1989). Only a very small fraction of these (approximately 1 in 1.5 million) will ever make it to the canopy and start producing fruit. Data such as these, which are by no means atypical for tropical trees, provide perhaps the most convincing demonstration of how difficult it is for a species to maintain itself in the forest, even in the absence of any type of resource harvest.

4. The reality of NWFP harvest

Given the low density of tropical forest species, their reliance on animals for reproduction, and the difficulty experienced in establishing their seedlings, the harvest of any type of plant tissue will necessarily have an effect on the species involved. The delicate ecological balance maintained in a tropical forest is easily disrupted by human intervention, and extractive activities that at first glance appear very benign can later have a severe impact on the structure and dynamics of forest tree populations. This impact may not be immediately visible to the untrained eye, but it is definitely occurring.

In general, the ecological impact of NWFP utilisation depends on the nature and intensity of harvesting and the particular species and type of resource under exploitation. Sporadic collection of a few fruits or the periodic harvesting of leaves for cordage may have little impact on the long-term stability of a tree population. Intensive, annual harvesting of a valuable market fruit or oil seed, on the other hand, can gradually eliminate a species from the forest. The felling of large adult trees can produce a similar ecological result in a much shorter time period.

Although the fact is seldom mentioned in much of the literature on the subject, a large number of non-wood forest resources are actually harvested destructively. Uncontrolled felling for fruit collection has virtually eliminated the valuable aguaje palm (Mauritia flexuosa) from many parts of Peruvian lowlands (Vazquez and Gentry, 1989). Destructive harvesting has also seriously reduced the local abundance of the ungurahui palm (Jessenia bataua), the babassu palm (Orbignya phalerata), and a wide variety of other important Amazonian fruit trees such as Parahancornia peruviana, Couma macrocarpa, and Genipa americana (Peters et al., 1989). Gharu trees (Aquilaria malaccensis) in Southeast Asia are routinely cut to harvest the resinous heartwood (Jessup and Peluso, 1986), and the collection of damar from Dipterocarpus trees in Peninsula Malaysia involves hacking a large box in the trunk of the harvest tree and then building a fire inside this cavity to stimulate the flow of oleo-resin (Gianno, 1990). Prunus africana trees in Cameroon are felled or completely stripped and girdled to harvest the bark tissue (Cunningham and Mbenkum, 1993). There are numerous other examples of forest species that are killed or fatally wounded by the harvest of non-wood products.

Even in the absence of destructive harvesting, the collection of commercial quantities of fruit and seeds can still have a significant ecological impact. In terms of simple demographics, if a tree population produces 1,000 seeds and 95% of the new seedlings produced from these seeds die during the first year, the population has still recruited 50 new individuals. If, on the other hand, commercial harvesting removes all but 100 of these seeds from the site prior to germination, the maximum number of seedlings that can be recruited into the population is reduced to only five. This ten-fold shortfall in recruitment can cause a notable change in the structure of the population.

In reality, this example is probably overly optimistic. First, it is assumed that all of the seeds left in the forest are positioned in precisely the right spot for germination and early growth. Second, there is always the possibility that the fruits and seeds left in the forest will experience a rate of mortality that is higher than 95%. Commercial collectors, in effect, are competitors with forest frugivores, and their activities reduce the total supply of food resources available. In response to the reduced abundance of fruits and seeds, frugivores might be forced to increase their foraging to obtain sufficient food. The net result would be an increase in the total percentage of seeds destroyed.

All of these factors interact in a synergistic fashion to inhibit the recruitment of new individuals into a plant population. Over time, this lack of recruitment will alter the size-class distribution of the population being harvested. If commercial collection continues uncontrolled, the harvest species can be gradually eliminated from the forest. This process of gradual population disintegration is illustrated in Figure 3 using demographic data for Grias peruviana and the stepwise results from computer simulations using a transition matrix model (Peters, 1990b). Size classes 1 to 3 are based on height measurements of seedlings, saplings, and juveniles, while classes 4 through 8 reflect a 5.0 cm diameter (dbh) interval for adults. For the purpose of the simulation, the intensity of harvest was set at 85% of the total annual fruit production. Note the change of scale at years 10, 50, and 80 to compensate for the gradual decrease in population size. The scatterplot in the lower half of the figure shows the total number of fruits harvested from the Grias population during each year of the simulation.

Figure 3. Simulated change in the population structure of Grias peruviana in response to excessive fruit collection. Results based on stepwise analyses using a transition matrix model and demographic data reported in Peters (1990b). Harvest intensity set at 85% of the total annual fruit production. Note change in scale in the latter three time periods to account for progressive decrease in population size.

As is shown at Year 0, the G. peruviana population initially displays an inverse J-shaped, or negative exponential, size-class distribution of a shade tolerant canopy tree with abundant reproduction. After ten years of fruit collection, however, the structure of the population has been notably changed. The infrequency of seedling establishment has caused a reduction in the smaller size classes; the greater number of stems in the intermediate size classes reflects the growth of saplings that were established prior to exploitation. By Year 50, the population has been even further degraded by the chronic lack of regeneration. Some of the intermediate size classes contain less than ten individuals, and the existing level of seedlings and saplings appears insufficient to re-stock the adult classes. The size-class histogram at Year 80 represents the culmination of a long process of over-exploitation. The population consists primarily of large, old adult trees, none of which are regenerating. In the absence of remedial action, it is only a matter of time before G. peruviana becomes locally extinct.

The important message to be gained from this simulation is that at no point during the process of over-exploitation is there any dramatic visual evidence (e.g. dead or dying trees) that something is going wrong. Fruit production and harvest levels don't even begin to drop below baseline until year 30, and commercial quantities of fruit continue to be available for several decades after this (see scatterplot in Figure 3). Even during the latter stages, the forest still contains a considerable number of adult G. peruviana trees that are producing fruit. Harvesting would undoubtedly continue unabated until these trees began to senesce, at which point collectors would be forced to move into a new area of forest in search of the Grias fruits.

The example shown in Figure 3 represents an extreme case of uncontrolled over-exploitation, and does not necessarily imply that every level of NWFP harvest leads directly to species extinction. The simulation is very useful, however, because it shows that even though the ecological impacts of this type of resource use are relatively subtle, very gradual, and essentially invisible, in the long run they can be as devastating as logging in causing the disruption of local populations and species extinction.

Finally, in addition to its impact on seedling establishment and population structure, the collection of non-wood forest products can also affect the genetic composition of the plant population being exploited (Peters, 1990c). A population of forest fruit trees, for example, will usually contain several individuals that produce large succulent fruits, a great number of individuals that produce fruits of intermediate size or quality, and a few individuals that produce fruits that, from a commercial standpoint, are inferior because of small size, bitter taste, or poor appearance. If this population is subjected to intensive fruit collection, the "inferior" trees will be the ones whose fruits and seeds are left in the forest to regenerate. Over time, the selective removal of only the best fruit types will result in a population dominated by trees of marginal economic value. This process, although more subtle and occurring over a longer period of time, is identical to the "high-grading" or "creaming" of the best tropical timbers that occurs in many logging operations.

5. Baseline data and monitoring to minimise ecological impact

Given the "boom and bust" cycles that have historically characterised the exploitation of non-wood forest products, it seems unlikely that the unfettered interaction of markets, commercial collectors, and tropical forest species will automatically produce a sustainable form of resource use. Achieving this objective will require more than blind faith in the productive capacity of tropical trees, an unwavering trust in a free market system, and the unquestioned assumption that local collector groups instinctually hold the goals of forest conservation above any desire for personal economic gain. Sustainable exploitation of NWFPs will require a concerted management effort by all of the parties involved. It will require careful selection of species, resource, and sites. It will require controlled harvesting and periodic monitoring of the regeneration and growth of the species being exploited. More than anything, however, it will require a greater appreciation of the fact that ecology and management are the cornerstones of sustainable resource use.

From an ecological standpoint, one of the most essential ingredients required to achieve a sustainable level of resource use is information. By this we mean information about the density and distribution of resources within the forest, information about the population structure and productivity of these resources, and information about the ecological impact of differing harvest levels. An overall strategy for collecting this information, and for applying it in such a way as to guarantee that the plant populations being exploited will maintain themselves in the forest over time, is presented as a flow chart in Figure 4. The overall concept and sequence of operations outlined is adapted from Peters (1994). The different procedures are sufficiently general that they can be applied to any class of NWFP procedures, at any scale, and in forests that have already been heavily exploited as well as in more pristine, undisturbed environments.

Figure 4. Flow chart of basic strategy for exploiting non-timber tropical forest plant resources on a sustained-yield basis. the complete process is composed of six steps: (1) Species Selection, (2) Forest Inventory, (3) Yield Studies, (4) Regeneration Surveys, (5) Harvest Adjustments, and (6) Serial Harvest Adjustments. See text for explanation of each management operation.

As is shown in Figure 4, the complete process is composed of six basic steps: (1) Species Selection, (2) Forest Inventory, (3) Yield Studies, (4) Regeneration Surveys, (5) Harvest Assessments, and (6) Harvest Adjustments. Taken together, these operations accomplish three fundamental management tasks. The species or resource to be exploited are first selected. Baseline data about the current density and productivity of these resources are then collected. Finally, the impact of harvesting is monitored and harvest levels are adjusted as necessary to minimise this impact.

The basic concept here is to provide a constant flow of diagnostic information about the ecological response of the species to varying degrees of exploitation. Sustainability is achieved through a continual process of reciprocal feedback, i.e. the demographic reaction of the target species must result in a corresponding adjustment in harvest levels. The exact nature of this "fine-tuning" process will depend on the site, the judgement of the resource manager, the precision of the diagnostic data collected, the effectiveness of harvest controls, and perhaps most importantly, the ecological behaviour of the plant population selected for management.

5.1. Species selection

The decision on which plant resources to harvest will be based largely on economic concerns. Those resources possessing the highest current market price and the greatest potential for future market expansion will usually be chosen first. Social factors can also come into play. Some forest resources may have a long history of extraction or traditional use in the region, and local people may have a strong cultural preference towards continuing to exploit these resources. Other resources (e.g. medicinal plants or other plants of ceremonial importance) may be subject to certain taboos that prohibit commercial exploitation.

In addition to economic and social factors, a third set of criteria that should also be considered is the overall potential of the resource to be managed on a sustained-yield basis. Although the fact is frequently overlooked, some species are inherently better able to withstand the continual perturbations caused by resource extraction than others. Important ecological factors to consider include the life cycle characteristics of the species (e.g. phenology of flowering and fruiting, pollination, and seed dispersal), the type of resource produced (e.g. fruits, stems, bark, etc.), the abundance of the species in the forest, and the size-class distribution of natural populations. The basic idea here is quite simple. Given a group of resources with similar economic profiles, why not select those that are the easiest to manage and have the highest potential for sustainable exploitation?

5.2. Forest inventory

Density and size-class structure data are the most fundamental pieces of information required for management. Just as foresters need to know how many cubic meters of timber occur in a particular forest, the management of non-timber resources also relies on estimates of the distribution and abundance of different species. These estimates can only be obtained through a quantitative forest inventory. Inventories also provide the baseline data necessary to monitor the impact of harvesting. Without some knowledge of initial density and size-class structure, the population could slowly go extinct with each successive harvest and never be noticed.

Forest inventories are time-consuming, somewhat costly, and extremely tedious to conduct. It is strongly recommended, therefore, that a professional forester or inventory specialist be involved in the planning of this fieldwork. In general, the inventory should be designed to provide the following types of information:

The inventory should provide a reasonably precise estimate of the total number of harvestable trees per hectare (i.e. the resource density) in different forest types. For fruit and oil seed species, this means the total number of adult trees. For latex-producing species, medicinal plants and other plants such as rattan, some juveniles may also need to be included.

The inventory should provide data on the current population structure or size-class distribution of adult trees. Collecting these data requires that the diameter (DBH) of all stems be measured. Height measurements can be substituted in the case of herbaceous plants, small understory palms or woody shrubs and in the case of rattan, stem length is the most important measurement.

The inventory should provide a preliminary assessment of the regeneration status of the species. Does the species appear to be maintaining itself in the forest? Are there a sufficient number of juveniles to replace the inevitable death of adult individuals? To begin answering these questions, smaller, non-productive individuals must also be counted and measured in the inventory.

5.3. Yield studies

Given an understanding of the density and size-class distribution of a forest species, the next question that needs to be addressed is "How much of the desired resource is produced by natural populations of the species?". Suppose 250 kilograms of fruit are harvested from the forest. Is this level of harvest sustainable? Well, that depends. How many fruits does the population produce? Is this only 10% of the total population seed production, or were 95% of all fruits removed? Clearly, it makes a difference. Just as foresters (theoretically) use growth data to avoid cutting timber faster than it is produced in the forest, the sustained-yield management of non-timber resources also requires information about the productive capacity of the species being exploited. This information is obtained through yield studies.

The basic objective here is to obtain a reasonable estimate of the total quantity of resource produced by a species in different habitats or forest types. In view of the fact that larger plants are invariably more productive than smaller plants, of particular interest is the relationship between plant size and productivity. Probably the easiest way to obtain these data is to train local collectors to weigh, count, or measure the quantity of resource produced by different sample trees during their normal harvest operations. These studies should be repeated every few years using the same group of sample plants to monitor the variation in yield over time.

5.4 Insights from inventory and yield data

For resources which involve the harvest of vegetative tissues (e.g. thatch from Raphia palms, rattan cane, bamboo), the baseline data from the forest inventory and yield studies can be used to provide a useful preliminary assessment of sustainability. The analysis is based on the general relationship between the current stock, or standing crop, of a resource and its annual production. In general, abundant species with a large stock produce the largest amount of growth in a year, while sparse, low density species exhibit an annual production rate that is much smaller. Ten rattan canes growing 50 centimetres/year will produce 5 meters of cane; one thousand rattan canes growing 50 centimetres/year will produce 500 meters of cane.

Given this relationship, if we want to exploit the same forest resource year after year, it is important that we harvest no more than its annual growth each time. If we harvest more than the growth, we diminish the current stock of the resource and, over time, the species can be eliminated from the forest. A graphic, albeit hypothetical, example of the effect of over-harvesting is shown in Figure 5. The solid bars in the histogram represent the current stock of rattan cane at the start of each year; the open bars shown at the top of each solid bar represent the total growth by the end of the year. The dotted bars show the amount of rattan cane that is harvested each year.

Figure 5. Hypothetical example of over-exploitation of rattan illustrating the relationship between resource stock and annual productivity. See text for explanation of stock, growth, and harvest parameters.

As illustrated in Figure 5, the current stock of rattan in the forest at the start of Year 1 is 1000 canes. This stock produces 500 new canes during the year. At the end of Year 1, 700 rattan canes are harvested, i.e. 200 more than were produced during the year, leaving a stock of 800 canes. These plants are left to grow for a year and they produce 400 new canes. The next year, 700 canes, i.e. 300 more than were produced, are harvested again. The remaining rattan plants produce 250 new canes yielding a total stock of only 750 canes. A final harvest of 700 canes at the end of Year 3 reduces the total stock to only 50. At this point, commercial rattan harvesting is no longer possible and the annual growth of the resource has been reduced to only 25 canes/year. If nothing is done to remedy the situation, rattan will probably disappear from the forest as it has in many parts of Southeast Asia (Dransfield and Manokaran, 1994).

Based on what we know about the density and yield of the rattan population in this example, a more sustainable level of offtake can be prescribed. The initial stock of rattan produces 500 new canes each year. By harvesting only this amount and leaving the basic stock untouched, rattan cane could be exploited for a long, long time on the site. The key is to only cut as much rattan as the basic stock produces in one year. Sounds simple, but only because there are inventory and yield data available to define the key parameters.

5.5. Regeneration surveys

Periodic monitoring activities are essential for defining and maintaining the sustainability of NWFP exploitation. For most species and resources, the effects of over-harvesting are most clearly visible in the seedling and small sapling stage. Harvesting may kill a large number of adult plants, may lower individual tree vigour to the point that flower and fruit production is affected, or may remove an excessive number of seeds from the forest. From a population standpoint, the net results of these activities are the same--all reduce the rate at which new seedlings are established in the population. This impact can be detected, and hopefully avoided, by periodically monitoring the density of seedlings and saplings in the populations being exploited. In essence, the seedling and sapling densities in each population are a demographic "yardstick" with which to measure the actual long-term impact of harvesting. To use a medical analogy, these data are the vital signs by which to assess the health or infirmity of the population.

5.6. Harvest assessments

Harvest assessments are an additional type of monitoring activity used to gauge the ecological impact of resource harvest. These are primarily visual appraisals of the behaviour and condition of adult trees that are conducted concurrently with harvest operations. In many cases, these quick assessments can detect a problem with reproduction or growth before it becomes serious enough to actually reduce the rate of seedling establishment. The sample plants selected and marked for the yield studies are perfect subjects for these observations. Examples of the type of information to be recorded during these assessments include: overall vigour of the plant, wounding caused by harvesting, trampling of seedlings by collectors, evidence of insect pests or fungal pathogens, and abundance of fallen flowers and immature fruits under the crown.

5.7. Harvest adjustments

The monitoring operations are used to appraise the sustainability of current harvest levels (see Figure 4). The seedling and sapling densities recorded in the original regeneration survey represent the threshold values by which sustainability is measured. As long as densities remain above this threshold value, and no major problems are detected in the harvest assessments, there is a high probability that the current level of exploitation can be sustained. If, however, seedling and sapling densities are found to drop below this value, immediate steps should be taken to reduce the intensity of harvest. The effectiveness of this harvest reduction will be verified during the next regeneration survey. Further reductions in harvest levels may be warranted if seedling and sapling densities fail to stabilise, or drop even lower during subsequent surveys.

In actual practice, achieving a sustainable yield in this manner will invariably involve a considerable number of harvest adjustments. There is frequently a time lag in a population's response to disturbance, and after several cycles of apparently stable results from the regeneration surveys, the population may exhibit a drastic fluctuation in seedling and sapling densities. The important thing is that these fluctuations do not go unnoticed. By gradually lowering, or even raising in some cases, the intensity of resource extraction, the level of seedling establishment should eventually approximate the threshold value established for the population.

6. Some hard questions about sustainability

In a perfect world, baseline data about the size-class structure and yield characteristic of different NWFPs would be collected, regeneration surveys would be conducted as a matter of routine, and harvest levels would be adjusted periodically as necessary to ensure the long-term sustainability of resource exploitation. The relatively sordid history of forest exploitation in the tropics, however, suggests that this has rarely, if ever, been the case. From a technical standpoint, there is absolutely no reason that non-wood forest resources cannot be managed on a sustained-yield basis. Why then, given all of the recent interest in the conservation, social, and financial benefits of non-timber forest products, has so little attention been focused on actually monitoring the sustainability of the resource base from which all of these benefits accrue? In closing, I would like to pose three questions, the answers to which will probably go a long way in explaining the total lack of sustainability which currently characterises the modern world of NWFPs.

6.1. Who is responsible for doing the monitoring?

It seems to me that this question has never been clearly defined. If local communities are to be given the responsibility of stewarding their own forests (an alternative that I decidedly favour), why haven't I witnessed a surge of collaborative programmes designed to train forest collectors to inventory, monitor, and manage their resource base under commercial levels of exploitation? There are literally hundreds of projects currently underway throughout the tropics that are focused on the development, marketing, and sustainable exploitation of non-timber forest products. Many of these involve the creation of a management plan. Most of these plans are being developed by expatriate development workers, university foresters and extension agents. I wonder how many of these plans are actually being developed with the enthusiastic participation of local community groups. I also wonder whether equal emphasis is being placed on the economic, social, and ecological aspects of the enterprise. Will the monitoring and management activities be continued after the outside technical assistance has been withdrawn?

6.2. Who is paying for it?

Forest inventories, yield studies, and the periodic survey of regeneration plots are expensive activities. Even given the local expertise to collect these data, where will the money come from to continue this fieldwork once the development project or research programme has finished? If we are really interested in maintaining the long-term sustainability of forest exploitation, these activities must be viewed as a fixed cost. Are any provisions being made to ensure that these costs will continually be covered from the profits generated by the sale of forest products?

6.3. How do you stop it if it's not sustainable?

Much of the current interest in NWFPs stems from the potential conservation benefits afforded by this type of land use. The forest can be used and conserved at the same time, ecosystem structure and function is preserved essentially intact, and local population experience a welcome improvement in their monetary situation and standard of living. At least that is how it is supposed to work. As an ecologist, however, I am always bothered by a disturbing variation on this scenario. Let's assume for the moment that everything works. New markets are created for a certain NWFP, all the baseline data has been collected and the monitoring systems are in place, a local cottage industry has been set up, and sales are increasing every year. The revenues from the enterprise make a significant contribution to the well-being of the community. During the third survey of the regeneration plots it becomes obvious that the current rates of harvest are not sustainable and that the resource is being progressively over-exploited. The management prescription is that harvest levels should be reduced by 20% which will cause an immediate and notable drop in the profits from the enterprise. Where does the incentive come from to follow the path towards sustainability?


Ashton, P.H. 1984. Biosystematics of tropical woody plants: A problem of rare species. pp. 497-518 in W.F. Grant (ed.), Plant Biosystematics. New York, Academic Press.

Bawa, K.S., S.H. Bullock, D.R. Perry, R.E. Coville & M.H. Grayum.1985. Reproductive biology of tropical lowland rain forest trees. II. Pollination systems. American Journal of Botany 72:346-356.

Brown, F.I., L.A. Martinelli, W.W. Thomas, M.Z. Moreira, C.A. Cid Ferreira & R.A.Victoria. 1995. Uncertainty in the biomass of Amazonian forests: An example from Rondonia, Brazil. Forest Ecology and Management 75: 175-189.

Burgess, P.F. 1971. Effect of logging on hill Dipterocarp forest. Malaysian Nature Journal 24:231-237.

Campbell, D.G., D.C. Daly, G.T. Prance & U.N. Maciel. 1986. Quantitative ecological inventory of terra firme and varzea tropical forests on the Rio Xingu, Brazilian Amazon. Brittonia 38:369-393.

Croat, T.B. 1978. Flora of Barro Colorado Island. La Jolla, Stanford University Press.

Cunningham, A.B. & F.T. Mbenkum. 1993. Sustainability of harvesting Prunus africana bark in Cameroon: A medicinal plant in international trade. People and Plants Working Papers No. 2. UNESCO.

Dransfield, J. & N. Manokaran. 1994. Rattans: Plant resources of South-East Asia No.6. PROSEA, Bogor.

Faber-Langendoen, D. & A.H. Gentry. 1991. The structure and diversity of rainforests at Bajo Calima, Chocó Region, western Colombia. Biotropica 23:2-11.

Gentry, A.H. 1982. Patterns of neotropical plant species diversity. Evolutionary Biology 15:1-84.

Gentry, A.H. 1988. Tree species richness of upper Amazonian forests. Proceedings of the National Academy of Sciences 85:156-159.

Gianno, R. 1990. Semelai culture and resin technology. Connecticut Academy of Arts and Sciences.

Jessup, T.C. & N.L. Peluso. 1986. Minor forest products as common property resources in East Kalimantan, Indonesia. Common Property Resource Management, National Academy Press.

Johns, A.D. 1988. Effect of "selective" timber extraction on rain forest structure and composition and some consequences for frugivores and folivores. Biotropica 20:31-37.

Jordan, C.F. 1987. Amazonian rain forests: Disturbance and recovery, New York, Springer-Verlag.

Kartawinata, K., R. Abdulhadi & T. Partomihardjo. 1981. Composition and structure of lowland Dipterocarp forest at Wanariset, East Kalimantan. Malayan Forester 44:397.

Keller, M., D.J. Jacob, S.C. Wofsy & R.C. Hariss. 1991. Effects of tropical deforestation on global and regional atmospheric chemistry. Climatic Change 19:139-158.

Lang, G.E. & D.H. Knight. 1983. Tree growth, mortality, recruitment, and canopy gap formation during a 10-year period in a tropical moist forest. Ecology 64:1075-1080.

Manokaran, N. & K.M. Kochummen. 1987. Recruitment, growth, and mortality of tree species in a lowland dipterocarp forest in Peninsula Malaysia. Journal of Tropical Ecology 3:315-330.

Peters, C.M. 1989. Reproduction, growth, and the population dynamics of Brosimum alicastrum Sw. in a moist tropical forest of central Veracruz, Mexico. Ph.D. dissertation, Yale University, New Haven.

Peters, C.M. 1990a. Plant demography and the management of tropical forest resources: A case study of Brosimum alicastrum in Mexico, pp. 265-272 In: A. Gomez-Pompa, T.C. Whitmore, and M. Hadley (eds.), Rain Forest Regeneration and Management, Cambridge University Press.

Peters, C.M. 1990b. Population ecology and management of forest fruit trees in Peruvian Amazonia, pp. 86-98. In A.B. Anderson (ed.), Alternatives to deforestation: Steps toward sustainable use of the Amazon rain forest, Columbia University Press.

Peters, C.M. 1990c. Plenty of fruit but no free lunch. Garden 14:8-13.

Peters, C.M. 1994. Sustainable harvest of non-timber plant resources in tropical moist forest: An ecological primer. Washington, Biodiversity Support Program.

Peters, C.M., M.J. Balick, F. Kahn & A.B. Anderson. 1989. Oligarchic forests of economic plants in Amazonia: Utilisation and conservation of an important tropical resource. Conservation Biology 3:341-349.

Procter, J., J.M. Anderson, P. Chai & H.W. Vallack. 1983. Ecological studies in four contrasting lowland rain forests in Gunung Mulu National Park, Sarawak. Journal of Ecology 71:237-260.

Reitsma, J.M. 1988. Végétation forestiére du Gabon. Tropenbos Tech. Series 1:1-142.

Shukla, J., C.A. Nobre & P. Sellers. 1990. Amazon deforestation and climate change. Science 247:1322-1325.

Sunderland, T.C.H., C.J. Ros, J.A. Comiskey & A. Njiamnshi. 1997. The vegetation of the Campo Faunal Reserve and the Ejagham Forest Reserve, Cameroon. International Cooperative Biodiversity Group / Smithsonian Institution - Man and the Biosphere Program.

Turner, I.M. 1990. The seedling survivorship and growth of three Shorea species in a Malaysian tropical rain forest. Journal of Tropical Ecology 6:469-478.

Uhl, C., J.B. Kauffman & D.L. Cummings. 1988. Fire in the Venezuelan Amazon. 2: Environmental conditions necessary for forest fires in the evergreen rainforest of Venezuela. Oikos 53:176-184.

Valencia, R., H. Balslev & G. Paz y Mino C. 1994. High tree alpha-diversity in Amazonian Ecuador. Biodiversity and Conservation 3:21-28.

Vazquez, R. & A.H. Gentry. 1989. Use and mis-use of forest-harvested fruits in the Iquitos area. Conservation Biology 3:350-361.

Wiebes, J.T. 1979. Coevolution of figs and their insect pollinators. Annual Review of Ecology and Systematics 10:1-12.

1 A similar version of this paper appears in: Current Issues in Non-timber Forest Products Research, M. Ruiz-Perez and J.E.M. Arnold (eds). 1996. Center for International Forestry Research, Bogor. Indonesia.

Previous PageTop Of PageTable Of ContentsNext Page