Previous Page Table of Contents Next Page

FAO's Global Fibre Supply Study: Context, method and modelling the future

G. Bull

Gary Bull is manager of the Global Fibre Supply Study, based at FAO headquarters in Rome.

An FAO effort aimed at developing policy-relevant information and analysis of industrial fibre sources.

In late 1995, on the recommendation of the FAO Advisory Committee on Paper and Wood Products, the FAO Forestry Department initiated the Global Fibre 1 Supply Study (GFSS). The study was intended to respond to questions raised by forest industry and the public such as: Where is the raw material going to come from to cover forest products needs and how much productive forest is needed to supply expected future fibre demand sustainably?

1 The meaning of fibre is to be understood as fibrous wood and non-wood raw material for primary industries producing sawntimber, wood-based panels, pulp and paper products. The majority of this fibre exists in standing form in forest undisturbed and disturbed by humans and plantation forests. Other kinds of fibre are in the form of recovered paper and non-wood fibres.

The study aims to contribute to worldwide forest policy development through the provision of reliable data, information and analysis of industrial fibre sources. In addition to the basic data, GFSS also includes a projection and analysis of future developments in fibre supply, based on an assessment of the major determining factors The intended audience for this information is broad and includes decision-makers in governments, financial institutions, industry and non-governmental organizations (NGOs).

The study complements other work by FAO such as the Asia-Pacific and African Forestry Sector Outlook Studies, and the upcoming Forest Resources Assessment 2000. FAO is also updating the statistics on forest plantations and these data are included in the GFSS database. The GFSS will also be linked with other regional and global studies on future prospects of forestry being undertaken by the FAO Forestry Department under its Global Forest Products Outlook Programme.

The GFSS is being monitored by representatives from the forest products industry and a special Steering Committee is providing guidance on the scope of the project and feedback on the information generated. The study was deliberately designed to avoid duplication of effort with other agencies working in the area of fibre supply. It incorporates data from and is consistent with analyses conducted by the International Institute of Applied Systems Analysis (for information on Russia), the UN Economic Commission for Europe and FAO Liaison Office with the UN, Geneva (for information on Europe) as well as the Canadian and United States Forest Services.

The potential utility of the study is at least twofold: it will help to sensitize industry, governments and NGOs to the critical policy issues that surround industrial raw material sources and sustainable uses of forest resources while highlighting the necessity for countries to improve their data collection and analysis in this field. Ultimately, the aim of the GFSS is to contribute to the sustainable utilization of forest resources worldwide and thus to sustainable forest management overall.

This article describes the steps undertaken in the GFSS process, including the establishment of key definitions and the development of a database and modelling structure. It also provides examples of how the model can be used to derive possible future scenarios.

Establishing definitions and classification of forests

In conducting a study of this nature it is critical to have a clear set of definitions for forest resource terms, and the definitions must have a number of characteristics if they are to be useful for ensuring consistency in statistical reports and for preparing outlooks:

· definitions must be consistent and compatible with terminology agreed on in international fore;

· definitions must be flexible enough to permit the utilization of data sets collected for other purposes;

· major terms must be able to accommodate the country-level data collected. Each country has its own unique way of classifying forest area and volumes and the standard definitions must be able to accommodate these differences;

· Terms must be readily understandable to a wide audience of users.

The critical definitions concern forest area, standing forest volume, growth rates and harvesting volumes. It is imperative for any long-term planning to establish standards which can be applied to a wide variety of circumstances.

Forest area is organized in the land classification as shown in Figure 1. Fibre is classified according to its source: natural forest (forest undisturbed by humans, i.e. never logged), forest disturbed by humans, industrial plantations, recovered fibre and non-wood fibre.

FIGURE 1 - Land classification for the Global Fibre Supply Study

There are significant complexities at each level of the study. For example, countries define forest area in many different ways, and some countries have boundary disputes which make statistics relating to total land area problematic. Moreover, many countries simply do not know how much of their forest is disturbed or undisturbed. Finally, the process of determining the area available for wood supply is very complex, since it depends on the interaction of dynamic economic, environmental and social factors.

A key statistic for fibre supply planning is the forest area available for wood supply, both disturbed and undisturbed.

Other key statistics that are essential for an assessment of the current situation and for predicting possible futures are the commercial species growing stock, the average annual increment of the forest with some indication of natural mortality and the average volumes harvested and removed for each cutting cycle. With these statistics it is possible to provide better information for the policy discussions related to sustainable forest management.

Establishing a database to collect all relevant forest inventory and fibre information

One of the early tasks was to construct an appropriate database for data compilation and information management. The major issues considered in the GFSS database design were:

Breadth of inquiry. The study focuses on industrial fibre supply and does not include fuelwood supply (including that for industrial use), as this would have required a separate methodology for assessment. Where possible, national industrial fibre uses were also considered. This is important because, in many countries, the available statistics cover only the logs or forest products slated for export, which does not reflect the amount of industrial roundwood actually removed from the forest.

Level of detail. The GFSS has collected or estimated data by forest type (as defined by each country) rather than just country-level averages, thus providing greater detail than previous studies and a better statistical basis for forecasting.

Other fibres. The database structure specifically allows for the inclusion of non-wood and recovered fibre materials since these sources may play an increasing role in future fibre supply.

Plantations. Industrial plantations will have an increasingly prominent role in future fibre supply. The database structure allows for the inclusion of data on plantation area and on growth (as per species grouping).

Resources available for wood supply. As mentioned, one important requirement of forest assessment for both the GFSS and the Forest Resource Assessment 2000 is to identify the area available for wood supply. This requires the adjustment of forest area to take into account such factors as protected area and economic accessibility in order to identify the area that may be used for industrial uses. Key elements affecting accessibility that were considered include: physical characteristics (e.g. steep area and swamps); transport distance and/or the lack of infrastructure. Another factor considered is logging bans. The volume of wood supply in areas available for harvesting will be determined by the forest type and forest conditions; low commercial volume or highly degraded forest usually result in a low volume.

Data gaps. For several countries there is simply no information on the volume of fibre removed by forest harvesting or on the growth rate of the forest. In order to present countries with complete estimates it was necessary to evaluate the volumes and growth under similar conditions in adjacent countries or by examining case studies within countries. This is an essential step in the preparation of preliminary estimates of future fibre supply and these estimates were improved with feedback from the countries in regional workshops held in Asia and the Pacific, Latin America and Africa.

Replicability of estimation processes. When producing data estimates it is very important to be able to replicate the results. The database is designed so that each data source is entered as a separate record and the current GFSS estimate is also a separate record. This allows any reviewer the opportunity to retrace the steps taken to develop preliminary estimates. This approach also permits the addition of new data.

Record keeping. It is critical that up-to-date computer tools are used to store forest statistics. The database developed makes statistics readily accessible in many forms to a wide variety of users using standard database software. This has substantial long-term benefits for forest resource analysis.

FIGURE 2 - Classification of forest volume, growth and harvest potential

Database structure

The GFSS database structure is summarized in Figure 3. The country list is drawn from the FAO World Agriculture Information Centre (WAICENT) database. The country list is linked to additional tables from either UN or FAO sources on population, the Forest Resource Assessment 1990 (FORIS) data files and the FAO Forest Products Yearbook for industrial roundwood production. The first data table on natural forest area includes both the GFSS estimates and the non-FAO source data on which the estimates were made. The second data table on natural forest area, drawn from the FAO Forest Resource Assessments 1980 and 1990 and the Organization's publication State of the World's Forests 1997, is included as both an information source and as a basis for comparison with the GFSS estimates. The industrial plantation data are drawn directly from the most recent information in the FAO plantation database. The alternative fibre tables contain both data and computed estimates on non-wood and recovered fibre. The data projection table is generated by a model built in the database and the conversion tables permit the reporting of standardized volumes. The reference list is linked to all fibre supply sources so that users can easily track the source for the supplied data.

FIGURE 3 - Structure of the Global Fibre Supply Study database

Before deciding on the final database structure, a pre-test was done using a selected number of countries to ensure that the structure would accommodate available statistics from data-rich countries without establishing an unrealistic data expectation for countries with poor information capacities.

Conducting data analysis and interpretation

Ideally, the GFSS forest resource estimate for each country is drawn directly from data produced by authoritative sources in that country. However, this frequently is not the case for a number of reasons:

· the country definitions of forest area and volume may differ from those adopted by the GFSS (thus requiring adjustment of country data);

· there may be conflicting sources of data from within a single country classes were established to indicate the reliability assigned to each source of information;

· as the data are frequently out of date some projections to the GFSS base year (1995) had to be made;

· there were often data gaps which had to be filled with estimates from neighbouring countries, with default values from the literature or estimates based on experience.

Thus, the analysis in GFSS is to be to viewed as preliminary or as a starting point for discussions with many developed and developing countries with a view to improving their data collection ability.

Figure 4 contains a sample of summary statistics in the database for an imaginary country (country XYZ) [Ed. note: In this article the presentation of actual data for a given country has been avoided to help focus attention on the process and not on country-level information].

FIGURE 4 - Sample of country summary statistics' used in the GFSS

Modelling possible (alternative) futures

The creation of possible future scenarios requires the manipulation of critical variables in a set of equations that forecast changes in fibre sources over time. The critical variables in the GFSS are:

· area available for wood supply;
· area change for natural and plantation forest;
· volume - both growing stock and commercial growing stock;
· growth - gross annual increment and mortality;
· fellings and removals - logging intensity and cutting cycle;
· recovered and non-wood fibre.

Modelling the future in today's forest policy context also requires the use of equations that have at least some ability to express the sustainability of supply and that have some linkage to the statistics or variables just described. The equations described in the Box were chosen for two reasons:

History. It is common with many governments and industry to start their discussion on sustainable fibre supply by using a yield regulation formula 2. This calculation frequently serves as the backbone of concession agreements with the timber and wood industries and is therefore critical in assessing the sustainability of timber supply.

2 This formula could be area- or volume-driven but, for our purposes, only the volume-driven approach is described. (See R. Catinot. 1997. The sustainable management of tropical rainforests. Paris, SCYTALE.)

Lack of alternatives. In the future it could be necessary for countries to have a yield regulation calculation, since it is frequently a very important starting point for discussions of sustainable forest management.

Establishing a starting point: static supply levels

One of the most difficult stages in developing a model for predicting possible futures is the establishment of a starting point or baseline - a rough indicator of supply levels based on current growing stock, forest increments, harvest intensity and forest losses. In the GFSS, the baseline was established by making a separate calculation for each major source of fibre (forest disturbed and undisturbed by humans, industrial plantations, non-wood and recovered fibre) and then combining them. The sections below describe how these calculations were made in order to give a complete fibre supply picture for each country.

Selected equations for supply calculations

Equation I:
Equation II: 1
Equation III:
Equation IV: 2
Equation V:
Gud = Commercial species growing stock in forest undisturbed by humans
Gd = Commercial species growing stock in forest disturbed by humans
Aud = Area available for wood supply in forest undisturbed by humans
Ad = Area available for wood supply in forest disturbed by humans
Hi = Harvest intensity
i = increment
c = Cutting cycle
r = Rotation period

Industrial plantations:
Non-wood fibres:
Recovered fibres:
i = mean annual area of plantation
Ap = net reported area of plantation
%NPC = percent non-wood pulping capacity
%WPR = percentage waste paper recovery
Tpc = total pulp production
Tpac = total paper production

1 These conversions were first suggested by Vanniere in FAO, 1975. Management possibilities of tropical high forest in Africa. Rome.
2 This equation approximates the Cotta Yield Determination which is included in FAO (in press). A handbook for the management of tropical forests. Rome.
3 This equation approximates the Von Mantel Method which is included in FAO (in press). A handbook for the management of tropical forests. Rome.

Forest disturbed and undisturbed by humans. There are a wide range of equations used to calculate area of forest undisturbed by humans and forest disturbed by humans. Based on an extensive review of the literature on yield regulation, particularly as applied to tropical forest conditions, the GFSS identified five equations as having the most promise (see Box).

Equations I and II rely on the commercial growing stock statistics in the forest undisturbed by humans and allow for the gradual transition over time from a undisturbed forest to a disturbed forest. Equation 11 allows for the application of a reduction factor to the increment statistics. Since in many countries the increment statistics are very difficult to obtain, the ability to reduce increment because of uncertainty, mortality, bark and other factors is very important. The primary disadvantage of the use of these formulae relates directly to the inadequacy of forest inventory data which only report volumes for large-diameter trees, usually more than 50 cm dbh. In calculating future potential supply this is simply not appropriate, since all trees, and certainly all those of more than 10 cm dbh, are of significance and should be reported. Unfortunately, the necessary conversion factors have not yet been developed, particularly to convert the inventories from a 50 cm class to a 20 cm class, for example.

Equations III and IV focus on logging intensity statistics for the gradual conversion of the undisturbed forest to disturbed forest. Once the conversion has occurred, the formulae use the growth in the forest as the driving variable in calculating supply.

Equation V is a barometer against which the results of the other four equations can be compared. It is useful to have a reasonable range of formulae for analysts to compare results and select the most appropriate future.

Normally it is most appropriate to calculate the amount of fibre to be removed on the basis of the standing commercial forest and the growth of the forest, i.e. using Equations I and II. However, since volumes are so difficult to determine from a supply point of view, the harvesting intensity applied to the area available for wood supply is a suitable surrogate for supply forecasting calculations. Therefore, Equation IV has been selected for demonstration purposes and is the formula applied throughout this article.

Industrial plantations, non-wood fibres and recovered fibres. In determining the static supply levels for industrial plantations, non-wood fibres and recovered fibres, the array of options are narrower than with forest undisturbed by humans and forest disturbed by humans. This is either because the data or information currently available are limited or the data can be handled in a more straightforward manner. The GFSS model identifies a single formula for the calculation of each of these components (see Fig. 5 A-D).

FIGURE 5 Formulae for calculating fibre supply for industrial plantations, non-wood fibre and recovered fibre - 5A - Static supply level

FIGURE 5 Formulae for calculating fibre supply for industrial plantations, non-wood fibre and recovered fibre - 5B - Future 1: Potential future fibre supply under current trends in area available for wood supply

FIGURE 5 Formulae for calculating fibre supply for industrial plantations, non-wood fibre and recovered fibre - 5C - Future 2: Increase in plantation fibre and rapid harvesting of natural forest

FIGURE 5 Formulae for calculating fibre supply for industrial plantations, non-wood fibre and recovered fibre - 5D - Future 3: The "greenish" future

Putting it all together

Using the sample data for country XYZ, Figure 5A presents the curves for natural forest (disturbed and undisturbed) under static supply conditions. As noted above, Equation IV is used for this demonstration.

The computed potential supply level from the forest undisturbed by humans is 36000000 m3. The forest disturbed by humans can continue to supply 38000000 m3. It is important to understand that this increases over time - as undisturbed forests are logged the forest area is added to the forest disturbed by humans. Over time, the basis for calculations also changes as forest growth becomes the driving variable instead of logging intensity. Hence, by the year 2043, all undisturbed forest in the area designated as being available for wood supply will have been harvested once. The potential annual volume harvested is reduced to approximately 40000000 m3 because the logging intensity is less in the forest disturbed by humans.

For industrial plantations the calculated supply is based on the forest growth, which for country XYZ is nearly 12000000 m3 of fibre, assuming 5 percent of the area is available for harvesting in any single year. The current afforestation rate of 250000 ha per year is being applied and this increases the potential available supply over time.

Under existing circumstances the recovered fibre supply is approximately 800000 m 3 and non-wood fibre capacity (capacity is an indicator of supply) is only 500000 m3. Country XYZ currently has only low levels of paper consumption and therefore a lack of available wastepaper. The levels of non-wood fibres are also very low, as is currently the case in most countries, the notable exceptions being India end Chine. [Ed. note: For a fuller discussion on non-wood sources of industrial fibre, see the article by Pande]

It is important to note that the total potential supply curve is a maximum supply line with few economic constraints. To get more realistic futures will require additional analysis whereby economic constraints are applied to "net down" the supply curve.

It is also important to bear in mind that the potential supply projections represent the supply as measured in the forest, not the supply as the volume of wood delivered to the mills for industrial roundwood production. The logging residue, which is the volume accounting for the difference between "bush" and "mill" supply can be greater than 30 percent in many countries. It is also important to recognize that logging damage to a forest once disturbed (see article by Pulkki) can also have profound impacts on the overall supply picture.

Selected major factors influencing fibre supply

Static supply

Future 1

Future 2

Future 3

Forests disturbed/undisturbed by humans

Land use (deforestation)





Sustainable management (as expressed by cutting cycle)





Legally protected area change





Industrial plantations

Land use (afforestation)





Development gains





Non-wood fibers

Non-wood fibre pulping capacity





Recovered fibres

Wastepaper recovery rate





So far in the examination of fibre supply no changes have been made to the calculations with regard to land use, harvesting efficiency, sustainable forest management or deforestation rate. For plantations, no changes have been made with respect to development gains or the officially reported afforestation rate. The fibre supply values specified for non-wood and recovered fibres are set using the latest available statistics with no foreseen increases. Therefore, under the static supply situation, the graph represents a complete picture of a static supply future and is based on a continuation of current policies and practices. To help create more realistic or alternative futures, some of these variables are integrated to enable the exploration of different policy options.

Using the model to predict alternative futures

Once an acceptable quantitative approach (the model) has been developed, alternative futures can be predicted for each country by varying critical variables. The Table lists the major factors identified in the GFSS future of fibre supply study. These factors are not an exhaustive set but they are part of a larger set of variables that have been raised in various studies within the last decade. They were selected based on their relative importance and the feasibility of obtaining information for each. In the future a wider range of factors could be added to the modelling framework.

The Table also indicates the variations introduced in order to produce three potential futures, which are displayed graphically in Figure 5 B. C and D using Equation IV.

The graphic representations shown in Figure 5 are, of course, only summary data. Much more detailed data are generated by the modelling exercise.

Figure 5B demonstrates the impact of a single factor, the deforestation rate of 0.96 percent per year, on the static supply line just described in the fibre supply forecasts.

Figure 5C explores the impact of decreasing the deforestation by 20 percent from the current level, giving less land in protected area status, and introducing shorter cutting cycles on the natural forest. For industrial plantations it was assumed that development gains increased by 50 percent and the afforestation rate increased by 20 percent. The non-wood fibre capacity decreased by 20 percent and the recovered fibres decreased by 10 percent.

Figure 5D explores what some might consider a "greener" future along with an aggressive plantation policy. It means the land use change is 5 percent higher than for Future 1, the areas placed under protected status are increased and the cutting cycles for forest harvesting are lengthened for the natural forest. Industrial plantation afforestation rates are also dramatically reduced - by 90 percent by the year 2010. The non-wood fibre capacity increased by 20 percent and the recovered fibre by 10 percent.

The model can be used to foresee futures based on dramatic departures from the static supply situation, but the most practical use is based on possible or achievable variations, determined on a country-by-country basis. The model provides a tool to explore the impacts of different forest policies, particularly those that readily articulate fibre supply aspects of sustainable forest management and present them to decision-makers for consideration.

Making the database and models widely accessible

In order to ensure transparency and accessibility, both the full database and a set of user-friendly models will be accessible online over the Internet. This allows users to access the baseline data for any country and to submit suggestions for data completion, updating, correction, etc. In addition, a user may run the model online, either to generate one of the three futures already modelled by the GFSS team, or to introduce new data in the critical variables to create alternative scenarios. The feedback mechanism built into the online presentation will also permit users to comment on the modelling approach itself and to suggest possible variations. Figure 6 shows the front page of the database and model.

FIGURE 6 - Capture of the GFSS front page


The GFSS project reflects a significant effort of the FAO Forestry Department and its partners to contribute to worldwide forest policy development through the provision of reliable data, information and analysis of industrial fibre sources. Thus far the process has been useful in developing an appropriate framework for data collection and presentation, for collecting the latest relevant forest area, volume, growth and removal statistics, for training country experts in data collection and analysis and for developing models to view alternative futures for forest resources.

A number of clear lessons have emerged from or have been reinforced by the process:

Obtaining good quality data requires extensive cooperation. Many countries have scattered subregional (frequently outdated) inventories as well as various field and research studies related to essential data needs. NGOs, industry, research institutes, universities, international agencies and governments all require a renewed commitment both to the production of better forest inventories for timber and resources and to a vastly improved process for sharing the information.

The production of up-to-date statistics is essential for efforts being made to implement sustainable forest management. In order to produce good statistics it is necessary to develop a suitable framework for the compilation and presentation of forest resource data and statistics. The GFSS contains just such a statistical framework which can be greatly improved over time as resources and time permit. The framework can also be modified to meet the needs of forest statistical reporting at the country level, thereby helping individual countries improve their reporting.

Statistics have to be maintained and continually improved. Once a statistical structure has been developed it needs improvement to include additional relevant statistics, input of new data as it is provided by countries and other sources and a constant analysis of data to ensure that the information is reflecting the situation in the country. The database also needs more input from other institutions which collect part of the data set used to conduct global studies.

Training has to be linked to data collection and analysis to make the statistical work sustainable. There is a strong need to establish and maintain an institutionalized network of contacts in all countries for coordination of statistical efforts and sharing of insights gained from local expert knowledge. With emerging communication tools there is an opportunity for far more effective interaction with country representatives at a comparatively low cost. Of course appropriate training is necessary for more effective communication.

The GFSS process represents the latest extensive global effort to collect data on forest resources with a direct link to attempts to describe sustainable industrial uses of those resources. Much more needs to be done to address the broader issues raised and described in the article by Duinker, Nilsson and Chipeta. Over time these issues will have to be merged into ongoing fibre data collection and analysis exercises as well as modelling studies regarding industrial fibre sources. In the meantime, the statistics and the modelling tools in the current GFSS project give policy-makers an interim approach to making their decisions on forest resource management and use.

In order to keep moving in the right direction, there are urgent applied research improvements that should and can be made to the work completed thus far and in which the FAO Forestry Department can work in partnership with countries and other institutions. They include:

Further statistical efforts. Additional statistics will have to be collected on various aspects, including the fibre volume of trees outside forests, the appropriate conversion factors in the harvesting and production process and on fuelwood supply. These supply factors require separate treatment and additional work.

More applied research. Current knowledge about forest growth and forest harvesting is very limited. Without significant additional research on growth and yield, the potential for lesser-known commercial species and forest harvesting systems, the sustainability of forest resources will remain difficult if not impossible to assess.

More examples of sustainable forest management. GFSS early research indicates that we can increase the security of supply and the amount of supply and, at the same time, incorporate other forest values into the decision-making framework with the application of sustainable forest management principles. However, many more practical examples or models need to be developed and promoted for demonstration purposes.

More forest inventories. Many countries still need to conduct country-level forest inventories in order to assist in planning for sustainable forest management.

Broader analysis to include critical social and economic factors. Decentralized decision-making in forestry is becoming more prevalent worldwide. In particular, there is more recognition of the need for participatory decision-making processes. It is anticipated that community forestry programmes will have an impact on future fibre supply, particularly as the need for local industrial roundwood consumption is formally recognized in the decision-making process.

Better understanding of the impact of fuelwood consumption and use on global fibre supply. Further study is required on how fuelwood consumption and use (a critical element in the lives of nearly 40 percent of the global population, as well as an increasingly important source of industrial energy) affects or can affect global fibre supply, and a methodology needs to be designed to integrate fuelwood considerations into fibre supply modelling exercises. Some early consideration of the fuelwood question is presented in the article by Horta Nogueira et al.

Industrial roundwood supply from trees outside forests is also an emerging issue in many countries. Understanding its contribution will have an important impact on future fibre supply forecasts in some regions.

Naturally these research improvements will have to feed into a much broader research agenda which will have to be coordinated with other forestry institutions to integrate issues such as biodiversity, economic development, land use change, forest conservation and amenities and climate change into the discussion of forest sustainability for both industrial uses and the many other forest functions. This is a significant challenge for forest analysts worldwide in attempting to increase the relevance of global forest policy debates.

Previous Page Top of Page Next Page