AVAILABILITY OF INFORMATION
FRA 2000 was launched with the ambition to cover a broad set of variables relevant to the forest sector at the international level. The intention was to broaden the global assessment approach from the previous strong focus on forest area statistics and address more complex forestry issues. The new subjects included qualitative aspects of forests such as biological diversity, biomass and availability for wood production as well as management parameters such as the status of forest management planning and protected areas.
FRA 2000 used the best available and most relevant country information on forest resources. Although some countries notably improved their inventories, and although the number of reports on forest resources increased in the 1990s, many countries still lack the basic data needed to accurately assess the state and changes of their forests. Most countries updated their forest cover estimates during the 1990s, often through remote sensing mapping, but in many cases the methodology was not directly compatible with that of previous surveys, making change estimates difficult. There is a scarcity of comparable multiple-date inventories and a need to improve both the accuracy and depth of information provided in forest inventories.
In the course of the study on forest area and area change there were found to be many documents and publications related to forest area, but some data were not representative or were derived from secondary sources. Information on forest area change could be derived with some precision, but data on qualitative changes such as forest degradation were generally missing, even in developed countries with relatively advanced forest inventory methodology.
Systematic field inventories that measure volume, biomass and productivity of the forests were carried out in many countries, but often within limited areas. As a result, national estimates for volume and biomass had to be extrapolated from local studies.
Although biological diversity has been widely studied, most studies focus on a specific ecosystem or species, and little quantitative and systematic information has been generated at the country level. Basic concepts for larger-scale assessment methodologies are still being developed. Simplified indicators such as the number of species under threat or spatial analyses of the degree of forest disturbance were proposed for FRA 2000 but no major advances were made in assessing this important aspect of forests.
Information on areas under forest management plans has generally not improved over the past decade, although the focus on certification has brought a higher level of information quality for the areas included in certification schemes. FRA 2000 collected estimates of forest areas that are covered by management plans and that are certified, but more work is needed for assessing the effectiveness of forest management for large areas. The increasing national commitments to implementing criteria and indicators for sustainable forest management offer hope that the next decade will see major strides in this important area. But a parallel important prerequisite to achieving sustainable forest management is an increase in efforts (including funding) to put more forest area under effective management.
Significant improvements have been made in obtaining estimates of protected forest areas. The data problem in this area is the apparent lack of consistency in countries' interpretation of the IUCN protected area management categories. Until the recent development of new methodology by the IUCN World Commission on Protected Areas (WCPA 2000) and the WWF/World Bank Alliance Rapid Assessment Methodology based on the WCPA framework, no generally applicable methodology for assessing management effectiveness in protected areas was available. It is still too early to see if the new methods will facilitate standardization of the assessment of management effectiveness in protected areas.
Several assessment parameters, such as forest fires, removals and non-wood forest products, would be relatively straightforward to assess if countries were willing and able to adopt a common approach to monitoring and reporting. FAO proposes to facilitate the development of definitions and reporting standards and to work with countries to implement them.
In summary, the availability of global and country information was not satisfactory for many subjects considered important for forest policy development. Furthermore, until baseline information is improved for important forest parameters, including diversity, degradation and productivity and their change over time, there is a danger that international policies and agreements such as the UN Framework Convention on Climate Change (UNFCCC) and the Convention on Biological Diversity (CBD) will be guided by general assumptions or extrapolations from partial, incomplete or invalid scientific studies.
ESTIMATING FOREST AREA AND AREA CHANGE
As in previous global assessments, the estimates of forest area and area change were the core of FRA 2000. Whether overemphasized or not, these parameters continue to be the most sought after from the global forest resources assessments. It is inevitable that the estimates will arouse controversy because of the political sensitivity of the subject material, especially as regards tropical deforestation (Matthews 2001; Stokstad 2001).
FRA 2000 provides transparent estimates that are traceable back to primary data and source documents. This approach will contribute to the establishment of a primary international data set and will help to counteract the recycling of inaccurate statistics in many reports. By indicating where information quality is low, FAO hopes to create an incentive to improve the baseline data. FRA 2000 reporting was done with the intention to establish a continuous improvement process where new information can be incorporated when it becomes available.
This approach does not make the results more reliable - only more accessible. Reliability can only stem from improvements in the quality and timeliness of national surveys and submissions to the global assessment in addition to the regional-level information generated directly by the assessment team, including remote sensing analysis. The collaboration with countries and national experts was very productive in FRA 2000 and the final findings are built on all available and relevant data on forest area and area change at the country level.
As concluded in Chapter 1, the precision of global-level estimates of forest area and area change is statistically valid, even though some source material may contain deviations from the true values. It is likely that an objective analysis would reveal some cases where forest area statistics reported by countries are too high, and others where they are too low. For example, one recent study in a moist tropical region suggested that regrowth of secondary forests is grossly underestimated in the forest area statistics reported by countries in the region. Therefore, it is important to continue to improve the science, methodology and consistency of forest area assessment and to try to ensure that the results of this work are processed as objectively as possible.
FRA 2000 conducted independent studies to support and complement information obtained from countries. Global mapping of forest cover, using 100 percent coverage by coarse-resolution remote sensing, provided an overview of the distribution of forests as well as a tool for global spatial analyses. However, the low spatial resolution of the remote sensing data meant that the results could not be used to enhance the findings on forest area or area change. By contrast, the FRA 2000 remote sensing survey of forest cover changes in tropical forests using higher-resolution images was very useful. The 10 percent sample, comparing the same areas that were sampled in 1990, provided statistically valid information on the area change dynamics at the regional and pan-tropical levels. The findings were not used to estimate forest area at the national level, but they provided an important validation of country data when aggregated at the regional level. This was especially useful for tropical Africa, where it was otherwise difficult to calibrate the area change estimate.
The documentation of the basis for estimating forest area and area change exceeds that of previous global forest assessments. A combination of well-developed inputs was used to draw the conclusions on the changes in forest area from 1990 to 2000. These include FRA 2000 national estimates on forests and plantations and the remote sensing survey of the tropics.
The result for net area change at the global level (-9.4 million hectares per year) indicates a lower net loss of forests compared with the results of earlier global assessments, mainly attributed to a greater expansion rate of new forests. While this lower rate of change is based on the most ambitious and accurate global assessment to date, the meaning and significance of the apparent trend is not yet known. While it may be a real trend, it could also be due to temporary conditions in the 1990s.
Confirming a shift in the rate of change is always a challenge, requiring long time series and careful analyses. In studying changes based on country reports, it is necessary to keep in mind that the average area-weighted reference year was 1994 for the world, and some years before that in many areas (Africa, for example). As a consequence the trends estimated for African countries relate mainly to the 1980s, with the change rates extrapolated to the 1990s.
In FRA 1990, the lack of updated information in many developing countries was addressed by developing a model that predicted the rate of change by country based mainly on population density and growth and climatic zone. This approach may be valid for a one-time estimate, but it has limited value in the study of trends over time, as the model will generate a time series based on its driving parameters rather than on real-world observations. Because of this potential bias and the desire to produce transparent estimates, the model approach used in FRA 1990 was not used in FRA 2000 to estimate area changes.
To support the country-level estimates, the FRA 2000 remote sensing survey was designed to study changes at the latest possible date, and to compare the 1980-1990 and 1990-2000 trends for the tropics. The statistical design made it possible to establish confidence intervals for the estimates. The area change estimate for the 1990s was lower than that for the 1980s, but the difference was not statistically significant. Thus, no reduced area loss could be confirmed for the tropics based solely on the remote sensing survey.
For the non-tropics as a whole, a considerable increase of forests in the 1990s was reported. Since the reference year was considerably later for these countries, it was valid to draw the conclusion that forests were expanding more in the non-tropics in the 1990s than in the previous decade.
Finally, it is useful to examine the possible reasons for this change in a larger social and land use context. The change of forest area depends to a large degree on the demand for land for other purposes. Two trends support the conclusion that the rate of deforestation may be decreasing. First, reduced demand for land by the agricultural sector and active afforestation programmes are leading to expansion of forests in temperate and boreal countries. Second, urbanization processes resulting from the development of national economies are significant in most parts of the world, which may reduce the demand for agricultural land in rural areas. In this case, agricultural land that is no longer needed often reverts to or is converted to forest, while land converted to urban uses may or may not be forest land.
In summary, FRA 2000 concluded that the global net change in forest area was lower in the 1990s than in the 1980s but that the rate of loss of natural forests remained at approximately the same level.
REMOTE SENSING
Remote sensing technologies are an area where there is promising potential to improve future assessments. Remote sensing technologies can provide images of physical or biological characteristics of the same land area at different points in time. FRA 2000 used remote sensing technology to create new global forest maps and to validate forest area change estimates. In theory, remote sensing is cheaper than traditional ground inventories when applied to large land areas (although the FRA 2000 remote sensing survey was paradoxically limited by financial constraints).
However, remote sensing can only address parameters that are well correlated with the information visible in images from above; it is thus excluded for many essential parameters and provides only limited precision and accuracy for basic biophysical variables such as wood volume and biomass. Furthermore, technology-intensive approaches generally preclude the participation of local people, thus limiting the ownership and local utilization of the information. Thus, a combination of remote sensing and ground-based assessment methodologies will continue to be needed for the foreseeable future.
STRENGTHS AND WEAKNESSES OF THE FRA 2000 APPROACH
FRA 2000 was carried out in response to global demand, represented notably by the recommendations of the FAO Committee on Forestry (COFO). The fourth session of the Intergovernmental Panel on Forests (IPF) endorsed the plan for the assessment elaborated during the Expert Consultation on FRA 2000 (Kotka III). The foundation for FRA 2000 was information provided by the countries of the world. It was recognized from the start that there would be considerable variation in the quality and completeness of this information, just as there is considerable variation in the institutional capabilities of nations in all other areas of endeavour. It was understood that major efforts would be required to validate and extrapolate from information provided by the countries. The problems related to inconsistent data quality and incomplete country information are the major weakness of FRA 2000. But the fact that FRA 2000 is built on country information is also its greatest strength. It is human nature that those who own and supply the information and are responsible for it are the most likely to be committed to use it and to improve it to influence policy decisions.
FRA 2000 also involved the active participation of international specialists and organizations in all phases of the work, starting with the expert consultation (Kotka III) that laid the framework for the assessment. For example, the global maps were produced in collaboration with the EROS Data Center in the United States. The protected forest estimates were made in collaboration with the UNEP World Conservation Monitoring Centre (UNEP-WCMC) in the United Kingdom. A transparent and forward-looking presentation of results was adopted, which encourages the continuous improvement of the baseline information. For the future, any individual, organization or country that develops more reliable or current information is encouraged to contribute it as soon as it is available so that it can be used to strengthen the next global assessment.
THE EXPANSION OF KNOWLEDGE IN FOREST ASSESSMENTS
FRA 2000 shared a major problem with most other forestry processes and programmes. Most of the process interaction was within the traditional forest sector. National forestry agencies provided data, and forestry experts in international organizations, universities and NGOs provided professional expertise. But most negative impacts on forests originate in other sectors. Furthermore, most countries that have succeeded in stabilizing their forest area are developed countries whose citizens do not need to exploit forest resources to try to escape from poverty. FRA 2000 would have benefited from greater cross-sectoral interaction. A major challenge for future global forest assessments will be to involve other sectors of society, to better integrate the assessment with other disciplines and to find new ways to use knowledge about forests to improve the lives of the world's citizens.
Interaction between the agricultural and forest sectors is central to how land is used and thus to the dynamics of forests. Economic development often leads to more capital-intensive agriculture and to a decrease in the area needed to produce agricultural outputs; in such cases (for example, throughout rural areas in the eastern United States) forests often expand on to former agricultural lands. Policies related to the development of infrastructure (e.g. roads and energy supply) often influence the use and size of forests. General development of economies may create employment opportunities in urban areas and less reliance on forest resources for basic needs such as fuelwood. Most of these topics fall outside the framework of a global forest resources assessment at the moment, but it would be important to incorporate cross-sectoral studies in the design of future global (and national) forest assessments.
At the local level, knowledge about forest resources is often relevant or essential to forest management. Such knowledge may also be relevant when aggregated to national or global levels. For example, the distribution of soil types and productivity (affecting carbon storage and fluxes), the location and dynamics of rare species (affecting biological diversity) and the impacts and benefits of uses of the land (affecting sustainable forest management) are important issues that are frequently studied and are discussed in this report. People who live and work in forests have huge amounts of knowledge about these subjects. This knowledge is usually reported in the form of local studies or case studies.
It is difficult to envision ways in which a global forest assessment can effectively address controversial issues such as illegal logging, but such areas should not be ruled out as impossibilities when plans are made for future assessments.
The present era is an exciting time of knowledge expansion, with new tools making it possible to share knowledge in unprecedented ways. One of the great challenges of forest assessments in the future will be to expand the participation of local people and of experts from other sectors and disciplines and to share and use knowledge in new ways.
BIBLIOGRAPHY
IUCN World Commission on Protected Areas (WCPA). 2000. Evaluating effectiveness: a framework for assessing management of protected areas, by M. Hockings with S. Stolton and N. Dudley. Best Practice Protected Area Guidelines Series No. 6. Cambridge, UK, WWF/IUCN Forest.
Matthews, E. 2001. Understanding FRA 2000. World
Resources Institute Forest Briefing No. 1. Washington, DC, WRI.
www.wri.org/pdf/fra2000.pdf .
Stokstad, E. 2001. UN report suggests slowed forest losses. Science, 291(5512): 2294.
