>
This section looks
at: |
Resource assessment can examine:
· which resources are useful commercially;
· what the consequences of exploitation are on the resource base itself;
and can inform sensible and appropriate management of NWFP resources.
What is assessed?
Further reading:
Peters, 1994;
Peters, 1996a; Hall & Bawa, 1993
There are different types of assessment or study that can inform development and management of NWFP resources. Approaches can focus on either:
· the NWFP resource itself, including its abundance or potential for future supply, through resource inventory; or
· its use in the market, such as market or product surveys, biodiversity inventories (or species lists), and cultural studies.
An ideal development process might begin with species/product selection, and include market research, resource inventory, growth and yield forecasting, determination of sustainable harvest rates, management planning and monitoring. One possible strategy is described by the flow chart in Figure 1, which describes the main elements of resource assessment requirements. Whichever approach is taken, resource assessment has a key role in the management of NWFPs.
Figure 1: Flow chart of a basic strategy for managing NWFPs on a sustained-yield basis
Any NWFP development programme should involve assessment/study at all these stages. Not all stages require formal assessments, as information may already be available or can be collected through informal methods. The bold areas of the flow chart suggest where field assessment needs ideally to provide biometrically rigorous, statistically reliable data. In these areas quantitative methods should be used.
However, the simple adaptation of forestry techniques is hampered because of the variety of:
· objectives for assessments;
· life-forms of NWFPs; including ease of detection;
· distributions, often NWFPs are clumped rather than evenly spread across an area;
· seasonal productivity, which means that some NWFPs exist only in specific periods; and
· levels of time, money and skills available to do the assessment.
Who does the assessments?
The NWFP assessments have been carried out or commissioned by a range of stakeholders, including Forestry Departments, aid organizations and communities. Because there can be a wide range of reasons for doing assessments, the methods, knowledge and experience are scattered amongst people from very different professional disciplines. Sharing of experiences between disciplines is limited, and areas familiar to foresters may be unknown to wildlife specialists and vice versa.
This means that development of methodologies is uneven and patchy in terms of the ideal flow chart. In addition, some disciplines might miss products which are important to others - e.g. dealing with wildlife is typically not part of Forestry Department work, and quantitative assessment is rare in rural development approaches.
Interdisciplinary work is key to addressing some of these gaps. Collaboration will help to bring together needs and experiences and to develop and standardize both appropriate methodologies and terminology.
Who wants the information from resource assessments?
This is an important question, as the reason for doing the assessment influences how it is done.
The majority of published studies to date are at the local level. However, there are many national level NWFP inventories that have not been published. Table 3 indicates what the information from resource assessments is used for at different levels.
Although it is necessary to know the objectives of a specific study before judging whether biometrically rigourous methods are required, it seems certain that critical areas are how to sample, measure, monitor and analyse quantitative studies of NWFP resources. There is a particularly strong need for reliable methods for measuring the distribution and quantity of a resource at a range of scales, from local to international.
Local level |
· determining sustainable harvesting quotas · monitoring the state of the resource · demonstrating sustainability to persuade authorities to allow harvesting |
National level |
Strategic planning, including: · deciding whether to allow export quotas · considering promotion of resource-based industries |
International level |
Informing conservation of endangered species, e.g. CITES Note: This usually relies on national level data |
Other (usually international) |
Fora discussing: · criteria and indicators for sustainable forestry · certification · Convention on Biological Diversity |
Generally local level data are required in the preparation of detailed management plans for specific areas producing NWFPs. There is some debate about the level of biometric rigour needed for these data (see below). However, it seems that there is an urgent need for communities to be able to prepare `sustainable' management plans. Without a sufficiently robust biometric basis to the plans many regulatory authorities are reluctant to release land for community management. This is the case in the negotiation of community access rights in at least Bolivia, Brazil, Mexico, Cameroon, and Zimbabwe.
In other countries, such as Indonesia, biometric resource inventory can be important to establish indigenous land rights and secure adequate compensation for loss of access to NWFP collection grounds.
At the national or regional level data are required for the following purposes:
· Economic opportunities: Good data are needed in the planning for investment or the development of a sector. For example, in investigating the potential use of pine resin as raw material for chemical industries (turpentine, rosin), rattan for furniture, tannins as substitutes of imported polyphenols, glues for plywood production etc. Biometrically sound data also used to determine policy, for example, regarding financial incentives for import substitution or export promotion (e.g. import tariffs).
· Social criteria: Reliable data are needed to determine the potential role of NWFP in rural development programmes.
· Environmental criteria: Quantitative data should be used as the basis for conservation and sustainable exploitation of NWFPs.
Data requirements: details and constraints
Resource Status: The first consideration is deciding which species to collect information on. This requires some initial knowledge of utilizable species, their products and distribution. Basic information might also be required on what is being harvested, where it is coming from, how it is located, potential or actual yields, harvesting techniques and levels.
National Forest Inventories (NFI) (or agricultural census in case of domesticated products) may collect NWFP information. Inventory at this level requires high biometric rigour. At the management unit or operational level (MUL), lower biometric rigour may be acceptable depending on the size of the unit. Data collection can vary from the assessment of a few samples to a stock-survey type census. In addition to quantitative data, the following are often also required:
Social aspects. Information may be needed on:
· ownership and/or access to the resources/species (private, public ownership status and trends);
· level of dependence of livelihoods on the resource (who, where and how are resources harvested);
· impact of other sectors (agriculture, labour availability, farmers); and
· decision-making processes in the country (planning cycles).
Economic aspects. Information is needed on:
· how important is investment in NWFPs for the national economy, what are the trends;
· influence of (inter-) national markets (substitutes within and between NWFPs); and
· financial possibilities: joint ventures, World Bank loans and incentives, etc.
Institutional and policy aspects. Information may be needed on:
· (forest) legislation and rules (NWFP rights in `timber concessions'); and
· training/education needs.
Reporting to international/regional agreements. Statistics and other information on resource availability and use, e.g. distribution, quantity of resource base, production and trade data.
The increasing international interest in sustainability of forests over the last few years has created a need for some way of measuring whether a forest is being managed sustainably or not. Criteria and indicators (C&I) for sustainable forest management (SFM) have emerged as a tool to measureand monitor progress towards SFM. At the end of 2000, 149 countries were participating in one or more C&I processes at ecoregional levels (FAO, 2001).
Criteria define the essential elements against which the quality of forest management is judged. Each criterion is defined by quantitative or qualitative indicators, which can be measured and monitored to determine the impact of forest management over time.
Indicators are in essence a form of monitoring protocols, and thus for NWFPs there is a need to develop assessment methodologies as recommended in this publication. Good NWFP resource assessments are critical to determining current status and as a basis for determining trends over time. A key question concerns the level of biometric rigour needed in assessments of the indicators, as this influences overall sampling design.
Over the past decade there has been a strong drive for certification of forests and forest products, partly in response to non-governmental organizations' (NGOs) concern about poor forest management. Certification is the independent verification that certain minimum forest management standards have been achieved by the manager. A certificate covers a specific area of forest for a specific time period. It usually involves a connection to the market through chain of custody labelling of products from certified forests.
There are numerous approaches to certification of NWFPs:
· forest management unit certification - looks at a broad range of forest management issues, including environmental and social;
· environmental/organic certification - concentrates on the way the product is produced, certifying that no chemicals or unnatural additives are used; and
· people-orientated certification (fair trade) - ensures that local producers get a good deal out of the product.
The NWFP assessments are useful mainly in forest management certification, where assessing the impacts of management at the forest management unit level is important. Again, the question of biometric rigour can be critical, as certification inspectors need to know how to assess reliably and consistently harvest and monitoring data.
Monitoring of species endangered by overharvesting is critical to avoid further population decline.
Monitoring of species threatened by overharvesting is usually done through harvest and trade records. For example, international trade in CITES protected species is monitored through import and export statistics and TRAFFIC use elephant ivory seizures to indicate population levels.
This may be the easiest way of getting information on heavily traded species, but such assessments give no information at the field level, and may not be reliable reflections of the status of the actual populations.
Until better field level information is more widely available, it seems likely that broad international policy decisions will be made on the basis of market information of questionable reliability.
What does `biometrically sound' mean?
It is not just about collecting quantitative information - statistical principles must be met throughout the assessment. The main principles relate to:
· objectivity in sampling design;
· number of plots used; and
· independence of observations.
Precision and accuracy.
Good quality data allows precision and accuracy to be estimated.
Precision is high when errors are small.
Accuracy is high when the estimated average value is close to that of the whole population.
Ideally estimates from assessments should be both precise and accurate.
The main advantage of a biometric assessment is that the precision and accuracy of the results can be calculated. This means that it is possible to put some confidence in the results. Precision is how tightly clustered the sample estimates are while accuracy measures how close the estimates are to the true (or population) value.
Conventional statistics enables us to calculate the precision of the results (usually expressed as the sampling error - see box 1).
However, it is impossible to calculate their accuracy without knowing the true value and if we knew this we would not need to sample. The way around this is to try to minimize bias in the sampling design and to make the study as precise as possible. If the answer is precise and we are reasonably sure there is no bias then we expect that the result is also accurate. Figure 2 illustrates these concepts in relation to hitting a target with the centre point representing the true value.
Figure 2: Precision and accuracy of a biometric study
The implication of this is that in order to be considered biometrically rigorous an inventory needs to be:
· unbiased - usually achieved by using an objective sampling design; and
· precise - usually controlled through plot numbers.
Note that a small level of bias may be acceptable if results are precise and the level of bias is known.
Objectivity
Objectivity is about minimizing possible bias due to subjective choice of samples. In practice this means selecting samples using pre-determined and objective rules, such as taking random plots or ones falling at the intersections of a systematic grid. Random sampling, using random number tables to select sample plots at locations within a grid (or `sampling frame'), should ideally be used. Systematic sampling ensures a regular spread of plots and can be useful for mapping species distributions. With systematic sampling, care must be taken to ensure that plots do not line up with some regular feature in the landscape, as this will bring bias into the results.
It is not acceptable to:
· subjectively choose samples - e.g. deliberately choosing a location because it is judged to be typical of the area;
· opportunistically choose samples - e.g. to select a location because it is accessible. Occassionally this cannot be avoided, for example in flooded Amazonian forest only accesible along navigable channels. In these cases the level of possible bias should be estimated; and
· haphazardly select samples - e.g. by throwing a quadrat over your shoulder.
The number of plots is critical in ensuring that the results are precise. Precision is measured through the sampling error of an estimate - the smaller the sampling error, the more precise the estimate. Large numbers of plots reduce the sampling error. Often inventories are designed to deliver a specific sampling error (typically 10-20 percent) and thus it is important to know how many plots to use.
The actual number of samples required depends on:
· level of precision required;
· how variable the resource is - highly variable populations require more plots than homogenous ones to give the same sampling error; and
· cost of accessing and enumerating each sample/plot.
There are methods of deciding how many replicates are required, (see box 12) but these require some initial knowledge about the variability of the resource. This is rarely available. A very general guide is that more than 30 is good, and less than five is probably inadequate.
Sample plots should ideally not be close together, and certainly should not touch. This is to avoid the possibility of the presence of the presence of the species in one plot directly influencing whether it is present in another. For example, a large tree in one plot might influence the possibility of there being another tree or saplings in the adjacent plot. Touching plots also bring in dilemmas about how to deal with individuals on the touching edges.
Are current methods biometrically adequate?
Methods reviewed were assessed against the above criteria to judge their biometric strengths and weaknesses in different areas.
It was difficult to judge the biometric quality of many of the 97 reviewed NWFP studies because the protocols were not reported in enough detail (see the enclosed CD-ROM for details of the reviewed studies).
This is a concern because information from assessments is only useful to people not directly involved in the work if it is adequately reported, with a protocol that can be evaluated for different uses or replicated elsewhere.
Protocols should clearly report the following key elements:
Sampling design: without details of how plots are located, the reader can only assume that it may have been done subjectively and therefore be biometrically unreliable. Only 14 percent of studies reviewed gave adequate details.
Plot dimensions and numbers: Despite describing quantitative studies, 25 percent of the studies reviewed fail to say how many plots were used. Whilst in some cases this can be worked out from details of the systematic design used, this should not be necessary.
Enumeration techniques should give details of where and how each plant or animal was counted or measured, but such details are often poorly reported.
There is a range of sampling designs available, including: census, random, systematic, stratified and experimental designs, which are statistically sound and include adequate objectivity (see Table 4).
Design |
Number |
% of studies* |
Census |
5 |
6.0 |
Random |
18 |
21.7 |
Systematic |
24 |
28.9 |
Experimental designs |
3 |
3.6 |
Stratified |
21 |
25.3 |
Subjective |
18 |
21.7 |
Opportunistic |
11 |
13.2 |
* Percent of the 83 studies which reported sampling designs.
Note percentages do not add to 100 as many studied combined designs, i.e. stratified random, etc.
The main failings in sampling design were the use of:
· Subjective location of plots: this is not uncommon, despite frequent recommendations to avoid it. Subjective selection of samples or plots reduces the reliability of estimates about the population, because sampling errors cannot be calculated. Subjective sampling can be justified, but will always result in data that are difficult to generalize. Justifications include:
¬ locating rare, unusual or difficult to find individuals;
¬ minimizing plot numbers (and thus cost) by sampling `representative' areas;
¬ difficult terrain and access problems; and
¬ use of local knowledge.
Subjective site selection may be acceptable if actual samples within the site or area are selected objectively.
· Opportunistic sampling: i.e. when samples are selected simply because they are highly accessible or are the only known sites. Whilst in some cases this may seem valid (e.g. sampling birds along a trail) bias is always a possibility.
Assessments are often based on single plots. For example, ethnobotany studies often rely on 1-ha plots (originally this size was chosen using species-area curves but is now accepted as a standard), which are thought to be sufficient to capture most of the flora of a region. Whilst this method may be acceptable for describing a flora, it has limited value for management, especially as plots or samples are often subjectively chosen. Overall, 29 percent of assessments reviewed have less than 5 plots, 30 percent with 5 to 30 plots, with only 40 percent having more than 30 plots.
The distinction between plots and subplots is a common source of misunderstanding.
Plots should be independent of each other to avoid the risk of relationships between them. Plots which touch each other should never be treated as independent, but rather as subplots. However, many studies treat subplots and contiguous plots as independent plots - this is called `pseudo-replication'. Annex 2 provides a diagram to explain the difference between plots and subplots.
Table 5 notes the biometric performance of different types of assessment reviewed. The main criteria used to judge whether a study was biometric or not were:
· adequate reporting of protocols;
· use of objective sampling designs;
· use of more than one plot; and
· use of independent plots.
Study type |
Studies |
Protocols reported (%) |
Biometrically `good' (%) |
Comments/Main concerns |
Biodiversity |
3 |
66 |
0 |
Often subjective but justifiable? |
Demographic |
9 |
44 |
22 |
Often based on single study plots or stands |
Ethnobotany |
10 |
50 |
20 |
Including quantitative ethnobotany |
Experiments |
5 |
80 |
80 |
Insufficient replication of treatments |
Harvesting studies |
5 |
80 |
60 |
Insufficient replication of treatments |
Resource inventory |
42 |
69 |
57 |
Insufficient plots |
Mapping |
3 |
0 |
33 |
Biometric sampling not a major concern? |
Market studies |
2 |
50 |
0 |
Econometric not biometric criteria apply |
Methodology |
11 |
64 |
55 |
Problems with contiguous subplots |
Monitoring |
12 |
50 |
25 |
Different biometric criteria also apply |
Rapid assessment |
1 |
100 |
0 |
Rapidity and rigour not compatible? |
Remote sensing |
2 |
0 |
0 |
Do not report protocol for ground truthing |
Use of secondary data |
6 |
10 |
17 |
Do not report protocols for original data set |
Social surveys |
2 |
50 |
50 |
Sociometric not biometric criteria apply |
Yield studies |
13 |
46 |
8 |
Subjective selection of sample individuals |
All studies |
126 |
56 |
38 |
Only 38 percent of the 126 studies reviewed are biometrically adequate according to the four criteria, whilst some 43 percent of resource inventories and 90 percent of yield studies appeared to fail. However, 56 percent of the studies were not reported in sufficient detail to make a judgement on biometric quality.
The major problems with the studies are:
· inadequate reporting of protocols;
· use of subjective sampling schemes;
· use of few or single plots; and
· confusion between plots and subplots.
The fact that both resource inventories and yield studies commonly fail to report protocol and/or use poor design is of concern. Both are used to inform management, and need to be biometrically sound. If this review reflects the general picture, much of the information provided appears to lack credibility.
Although not all studies need to be biometrically rigorous, it is useful for users to understand why biometrics is important so that they can judge whether or not they need it.
There is an intense demand for information on NWFPs, but not all of it needs to be rigorous. It depends on the objectives, needs and expectations of the users of the information from assessments.
So, why use biometric methods?
Key debate: quantitative vs. qualitative data
For decades social scientists have argued about which of the approaches is better for recording social phenomena.
Combining advantages: More recently there is mainstream recognition that there should be integration of the best elements of the two.
When is biometrics relevant?
Biometric rigour is important because it provides reliable, good quality information. Such information is important in ensuring appropriate planning and management. It is critically important for:
Livelihoods - giving the right advice: Decisions based on resource assessments can influence the long-term survival of species and thus livelihoods (Cunningham, 1996b; Myers & Patil, 1995). Oversimplification of complex situations, risking giving poor recommendations, should be avoided. It is critical that community-based assessments provide useful and reliable information - advisers should see this as an ethical obligation.
Exploitation - avoiding overharvesting: Good quality information is important to ensure that decisions to not lead to decline of the target species, which may in turn put commercial ventures based on those species at risk. As yet, few NWFP enterprises base harvesting decisions on reliable data and overexploitation is not uncommon. In such cases, it is critical that robust monitoring systems are implemented to deal with any negative consequences and make corrective actions.
Valuation of tropical forest resources - allowing comparisons: The use of NWFP data by people not involved in the inventory requires some level of standardization of what is measured and data quality. It is difficult to compare results from assessments that are carried out differently. Table 6 shows common failings of biometric rigour and reporting protocols in NWFP assessments from the perspective of natural resource economists, and makes suggestions for how methods could be improved (Godoy et al., 1993).
Strategic overviews - planning and prioritization: Often the data used for national, regional or international statistics come from local assessments of NWFPs. Often called `meta-analysis' this synthesis of different studies is more than simple compilation of data, but rather involves further analysis for wider interpretation. Whilst it is a cost-effective way of generating large-scale data, it is only as reliable as the data it uses. It will only provide biometrically adequate results if the local assessments do.
Credibility - avoiding political bias: Ensuring that data are biometrically sound can add weight to recommendations based on that information. Where governments have to defend their reasons for setting quotas to those who lobby for higher (industry/trade) or lower (conservationists) levels, reliable data are important. Case study 1 provides a useful example of the role of reliable data in political debate concerning the national quota for Prunus africana bark in Cameroon.
Table 6: Summary of main failings of NWFP resource assessment for valuation studies
Information required |
Main failing |
Suggested methodology |
Data representative of forest |
Many studies only use one site and reasons for choice not given so not possible to use data for comparison or generalization |
Ideally a sample of study sites (to allow calculation of variance) or failing this presentation of reasons for site choice |
Population profiles suitable for generalization |
Informants in anthropological studies not randomized and sample sizes small |
Identification of main attributes of extractors (e.g. age, technology, income). Stratified random sampling of people in identified strata |
Data representative of seasonal pattern of NWFP use |
Few studies include more than 1 year's data |
Random selection of same number of weeks and days from each month through at least one year. Careful examination of climate and other variable, e.g. larger economy to understand representativeness of study period |
Quantification of product flows (quantities used by people) |
Some studies value the stock (inventory) which relates to neither present or sustainable flows |
Identify, count, weigh and measure products as they enter the village each day. Assess random sample of villages and households and either ask extractors or randomly observe and record their consumption |
Product weight |
Weights may not be measured |
If products too difficult to weigh in bulk, take seasonal subsamples for mean weights |
Product identification |
Irregular use of scientific names or use of local names hinders comparison between studies |
Collect specimens (vouchers, skulls, photographs) for definitive scientific identification |
Catchment area for product extraction |
Many studies do not record catchment area so not possible to determine yields per hectare |
Direct observation, participatory mapping, travel time assessment, aerial photographs, Global Positioning Systems (GPS), etc. |
Sufficient observations |
Insufficient if reliant on single researcher undertaking all observations |
Train and use extractors to collect information or keep personal diaries (be aware of possible biases) |
Value of product |
Some researchers use expenditure of labour or energy as a measure of value which is not consistent with modern valuation theory |
Use prices that exist for the commodity concerned or that prevail in related markets, e.g. use marketed good bartered for non-market product, use value of close substitute. Use contingent valuation (willingness to pay) methods |
Share of harvest going to the household and to the market |
Few studies have done this but it is important as household and market goods are priced differently |
Random sample of households asked to keep log books of daily income, expenses and amounts of NWFPs consumed or sold |
Shadow prices |
Important in providing an economic rationale for NWFPs that may not be financially profitable Required to estimate valuation from a national viewpoint |
Adjust for taxes and subsidies that cause price to deviate from opportunity cost of resource |
Environmental externalities |
No study has done this which means that conventional valuations underestimate economic benefits of NWFPs |
No suggestions made |
Marginal costs of extraction and processing |
No assessment of search times, cost of tools, etc., made for plant collection (has been done for animals in studies based on optimal-foraging theory) |
Interviews, direct observation (instantaneous sampling, focal subject sampling), extractors diaries/records, log movements out of and into village |
Wage rates |
Some researchers have used country's official wage rate but this should not be done uncritically |
Determine what people actually pay each other. Note that rural wages vary by season, age, sex and type of work |
Cost of capital |
Not often measured - use of market rate inappropriate |
Use social discount rate - may be calculated locally otherwise use 4-5% |
Sustainability |
Three views a) Indigenous people manage forest sustainability b) Indigenous people do not manage sustainability c) Sustainability is result of special conditions that must be identified in each case |
Indirect: comparison of distance, frequency and duration of collection forays, recall of yields over time, etc. Direct: comparisons of extraction and rates of reproduction/growth in the forest |
Use of plant and animal extraction in single valuation |
Not possible as botanists use returns per hectare while zoologists use returns per unit of labour |
Multi-disciplinary team comprising natural resource economist/economic anthropologist, botanist, zoologist; as well as indigenous people and local scholars |
International Conservation of Birds Programme; ONADEF: Office National de Développement des Forêts (parastatal with responsibility for forest inventory); MCP: Mount Cameroon project + Ministry of Environment and Forestry Taken from: Acworth et al.,1998; Cunningham & Mbenkum, 1993; Acworth, pers. comm. |