At the first session of the Conference of FAO in the autumn of 1945, the need for up-to-date information on the forest resources of the world was fully recognized and it was recommended that an inventory should be undertaken as soon as possible. In May 1946 the Forestry and Forest Products Division was founded and work was immediately initiated on FAOs first worldwide assessment of forests (FAO 1948). After reviewing the results of the assessment in 1947, the sixth session of the FAO Conference in 1951 recommended that the Organization maintain a permanent capability to provide information on the state of forest resources worldwide on a continuing basis (FAO 1951). Since that time, various other regional and global surveys have been conducted every five to ten years. Each has taken a somewhat different form.
Statistics released by FAO on world forest cover from 1948 through 1963 were largely collected through questionnaires sent to the countries. The assessments since 1980 have taken a more solid technical form, being based on the analysis of country references supported by expert judgements, remote sensing and statistical modelling. FRA 2000 is the most comprehensive in terms of the number of references used and information analysed on forest cover, forest state, forest services and non-wood forest products (NWFP). FRA 2000 is also notable for applying for the first time a single technical definition of forest at the global level, based on 10 percent crown cover density.
Statistics from the different assessments are difficult to use for comparative purposes, owing to changes in baseline information, methods and definitions. However, better correlations can be achieved for time series in many countries for certain assessments, especially with information generated since 1980, when reporting parameters stabilized. Consistent definitions were applied for developing countries for subsequent assessments.
FAOS GLOBAL AND REGIONAL ASSESSMENTS 1946-1997
Forest resources of the world (1948)
In 1946, the year following the founding of FAO, a first global survey was conducted by the Organization; it was published in 1948 as Forest resources of the world (FAO 1948). Initially, a questionnaire was sent to all countries, of which 101 responded, representing about 66 percent of the worlds forests. Parameters included in the survey were forest area (total and productive), types of forest by accessibility, growth and fellings.
One of the noteworthy conclusions of the first world forest inventory report was that:
All these investigations made valuable additions to our knowledge, but all suffered from certain fundamental difficulties. Most important of these were the lack of reliable forest inventory information which existed and still exists in many countries, and the lack of commonly accepted definitions of some of the more important forestry terms. Hence, to the weakness of some of the quantitative estimates there was added doubt as to the real meanings of some of the qualitative descriptions (FAO 1948).This statement remains largely true today, over 50 years later. While technical and scientific advances have greatly increased the potential to improve the information base in countries, many still lack the training, institutional and financial resources to conduct periodic assessments.
World forest inventories (1953,1958 and 1963)
World forest inventories including all countries were carried out on three occasions during the 1950s and 1960s. Lanly (1983) describes these various inventories:
...126 countries and territories replied to the 1953 questionnaire representing about 73 percent of the world forest area. The picture was completed by information from the replies to the 1947 questionnaire for 10 other countries (representing 3 percent of the total world forested area) and official statistics for the remaining 57 countries, representing 24 percent of the world forest area. The results were published by FAO in 1955 under the title World forest resources -results of the inventory undertaken in 1953 by the Forestry Division of FAO.The main parameters assessed during the 1963 World Forest Inventory were forest area (total, productive, and protected), ownership, management status, composition (softwoods and hardwoods), growing stock and removals (FAO 1966).
The 1958 inventory of the FAO World Forest Inventory published in 1960 (World forest inventory 1958 - the third in the quinquennial series compiled by the Forestry and Forest Products Division of FAO) utilized the replies of the 143 countries or territories, representing the 88 percent of the world forest area, complemented by the replies to the 1953 questionnaire for 13 countries (2 percent) and to the 1947 questionnaire for 5 countries (3 percent). Necessary changes and precisions introduced in the definition of some concepts, more precise definitions of forests and changes in such concepts as forest-in-use and accessible forests affected comparability with the previous inventory. However, changes in area and other forest characteristics during the 1953-58 period were, for several countries, either reported directly from them or could be derived by comparison of the replies to both questionnaires (changes in area of permanent forests, in management status in forests-in-use, increase in accessible areas and in forest-in-use, afforested area between 1953 and 1957, etc.).
The World Forest Inventory 1963 published by FAO in 1965 witnessed a slightly lower rate of response of (105 compared to 130), at least partly accounted for by temporary strains on administration in countries gaining their independence as was reported in the document. Again comparability with the former enquiries was limited, and as pointed out by the authors of the report, large differences for some countries (between the results of the 1958 and 1963 enquires) resulted more from better knowledge about the forests, or stricter application of definitions, than from effective changes in the forest resources.
Major findings (1963)
Regional forest resources assessments (1970s)
During the 1970s FAO carried out no global surveys. Instead a series of regional assessments were made with the intention that each would be more regionally appropriate and specific. Beginning in the late 1960s FAO sent out questionnaires to all industrialized countries. The results were published in 1976 as Forest resources of the European Region (FAO 1976b). Questionnaires were also sent to Latin America and Asia and the results were published in Forest resources in Asia and the Far East Region (FAO 1976c) and Appraisal of forest resources of the Latin American Region (FAO 1976a). A similar questionnaire was sent to African countries by the Department of Forest Survey of the Swedish Royal College of Forestry and published in Forest resources of Africa - an approach to international forest resources appraisal, Part I: Country descriptions (Persson 1975) and Part II: Regional analyses (Persson 1977).
According to Lanly (1983), the regional assessments of the developing areas had the following main features in common:
Although FAO did not compile the regional findings into a global synthesis, a global survey was done outside FAO and published in World forest resources - review of the worlds forest resources in the early 1970s (Persson 1974). Finally, another FAO study, Attempt at an assessment of the worlds tropical moist forests (Sommer 1976), provided a summary on findings on the forest situation in all tropical moist forests.
FRA 1980 covered 97 percent of the land area of developing countries or 76 tropical countries: 36 in Africa, 16 in Asia and 23 in Latin America and the Carribean. FRA 1980 was distinguished by many features. Its breadth was the greatest to date, and in many cases remains unmatched by the present assessment. It is also notable as the first assessment to use a technical definition of forests, in which measurable parameters were indicated - notably 10 percent canopy cover density, minimum tree height of 7 m, and 10 ha as the minimum area for defining a forest. Previous assessments had relatively broad definitions which could be interpreted quite differently by different countries. The consistent definition provided parameters useful in adjusting country information to a common standard. An adjustment in time was also made using expert opinion to project the information to common reference years of 1976, 1980, 1981 and 1985.
FRA 1980 relied extensively on existing documentation from countries to formulate its estimates on forest cover (state and change), plantation resources and timber volume. Existing information from multiple sources in the countries was gathered and analysed. Dialogues with national and international experts on information utility and reliability helped to firm up the estimates for the countries. The assessment noted that information was abundant but hard to locate and synthesize in the coherent manner needed for a consistent global survey.
Extended narratives, explanatory text and qualitative information complemented the statistical data set. During the tenure of FRA 1980, FAO was conducting extensive work on forest inventories in tropical countries. Roughly one project existed for every two to three countries, and FAO experts employed in the projects provided valuable input to the 1980 assessment results.
In major forested areas where existing information was lacking, the assessment conducted manual interpretations of satellite imagery (1:1 000 000 scale). This was done for six Latin American countries, two African countries, two Asian countries and portions of two other Asian countries. The interpretations covered about 70 to 99 percent of these countries, with 55 satellite images used.
The final documentation for FRA 1980 included three volumes of country briefs (one for each developing country region) (FAO 1981a, FAO 1981b, FAO 1981c), three regional summaries and a condensed main report, published as an FAO Forestry Paper (FAO 1982). While the findings were not global, FRA 1980 was used again in 1988 to make an interim global assessment.
Interim assessment 1988
An interim report on the state of forest resources in the developing countries (FAO 1988) provided information on 129 developing countries (53 more than FRA 1980) as well as the industrialized countries. The report provided information on the state of the forests at the year 1980 and the changes over the period 1981-1985. Definitions varied between the industrialized and the developing countries, specifically in regard to canopy closure thresholds for forests, which were set at 20 percent for industrialized countries and 10 percent for developing countries. Information for the industrialized countries was collected by UNECE/FAO in Geneva, which drew on the report The forest resources of the ECE region (Europe, the USSR, North America) (UNECE/FAO 1985). Parameters also varied for the two groups of countries, so that a global synthesis of core elements was needed, in order to achieve a uniform global data set.
The elements of the global synthesis included forest, operable forest, inoperable forest, other wooded lands, broad-leaved forest and coniferous forest.
FRA 1990 (FAO 1995) covered all developing and industrialized countries and was distinguished by two innovations: the development and use of a computerized deforestation model which was applied to the developing country data for projecting the forest area statistics to a common reference year; and an independent pan-tropical remote sensing survey of forest change based on high-resolution remote sensing data.
FRA 1990 sought to improve estimates by eliminating the bias of expert opinions in the assessments through a statistical model to predict forest cover loss (and thereby deforestation rates). The model was based on forest cover change derived from the few comparable multi-date assessments available. Deforestation rates were then regressed against independent variables to determine the rate of forest loss relative to changes in population densities within specific ecological zones. Forest cover change rates were obtained by applying the model to existing baseline statistics available for the countries.
The advantages of the 1990 method were the near-uniformity achieved by applying the model equally to almost all developing countries and the ability to streamline production of statistics using computer routines. The disadvantages of the 1990 method were the low number of variables used in the deforestation algorithm and the low number of observations used to construct the model, introducing a relatively high random error (low precision) in country estimates.
Because of the many uncertainties involved in working with existing national data, FRA 1990 implemented a remote sensing survey to provide a quality-controlled set of statistics on forest resources and to complement the survey based on country information. The use of statistical sampling combined with a uniform data source (satellite imagery) and common data collection methods made this approach an important tool for providing a set of statistics to compare with the country data.
The survey relied on statistical sampling (10 percent) of the worlds tropical forests through 117 sample units distributed throughout the tropics to produce estimates of the state and changes of tropical forest at the regional, ecological and pan-tropical levels (but not at the national level). Each of the sample units consisted of three multi-date Landsat satellite images which provided the raw material for producing statistics on forest and other land cover changes from 1980 to 1990 and later to 2000.
FAO used an interdependent manual interpretation of satellite scenes at a scale of 1:250 000, conducted by local professionals where possible, and internationally experienced professionals in other areas. Multi-date image interpretations were manually registered to one another. Ground information was incorporated into about 50 percent of the interpretations. In some areas, ground truthing was not necessary owing to the high and consistent amount of forest. In other locations, especially where the composition of the landscape was highly differentiated, ground truthing was found to be highly valuable.
The principal output of the remote sensing survey was the change matrix, which illustrated and quantified how the forest and landscape change over time. The forest and land cover classification scheme of the remote sensing survey was linked closely to the FRA classes established for global reporting by countries.
Different definitions of forests for developing and industrialized countries again limited the utility of the final global synthesis, as did the absence of change information on forests in industrialized countries. Only changes in the area of forest combined with other wooded lands were assessed. (The definition of forest was set at 20 percent crown cover density for industrialized countries and 10 percent for developing countries.)
The assessment covered the parameters of volume, biomass, annual harvesting (tropics) and plantations. Brief summaries were also made on conservation, forest management and biological diversity. The country briefs prominent in FRA 1980 were unfortunately discontinued.
Interim 1995 assessment
An interim 1995 assessment was published in State of the Worlds Forests 1997 (FAO 1997). This report published new statistics on forest cover state and change for all countries with a reference year of 1995, and a change interval from 1991-1995. The definition of forest varied between the industrialized and the developing countries; canopy closure thresholds were set at 20 percent for industrialized countries and 10 percent for developing countries.
The baseline information set for the assessment with only a minimum of updates was drawn from the FRA 1990 data set and had an average reference year of only 1983. Although FAO contacted all developing countries and requested their latest inventory reports, updated information was only submitted and used for Brazil, Bolivia, Cambodia, Côte dIvoire, Guinea-Bissau, Mexico, Papua New Guinea, the Philippines and Sierra Leone.
The FRA 1990 deforestation model was used for adjusting developing country statistics to a standard reference year (1991 and 1995). No adjustments to standard reference years were made for the industrialized country statistics. Consequently, the industrialized and developing country data were not harmonized in terms of their definitions or reference year.
FAO. 1948. Forest resources of the world. Washington, DC.
FAO. 1951. Report of the sixth session of the Conference of FAO. Rome.
FAO. 1960. World forest inventory 1958 - the third in the sequential series compiled by the Forestry and Forest Division of FAO. Rome.
FAO. 1966. World forest inventory 1963. Rome.
FAO. 1976a. Appraisal of the forest resources in the Latin American Region. Document presented at the 12th session of the Latin American Forestry Commission, Havana, Cuba, February 1976. FO:LACF/76.
FAO. 1976b. Forest resources in the European Region. Rome.
FAO. 1976c. Forest resources in the Asia and Far-East Region. Rome.
FAO. 1981a. Tropical Forest Resources Assessment project (in the framework of GEMS) - Forest resources of tropical Asia. Rome.
FAO. 1981b. Los recursos forestales de la America tropical. Rome.
FAO. 1981c. Forest resources of tropical Africa. Rome.
FAO. 1982. Tropical forest resource, by J.P. Lanly. FAO Forestry Paper No. 30. Rome.
FAO. 1988. An interim report on the state of forest resources in the developing countries. Rome.
FAO. 1995. Forest resources assessment 1990 - global synthesis. FAO Forestry Paper No. 124. Rome.
FAO. 1997. State of the Worlds Forests 1997. Rome.
Lanly, J.P. 1983. Assessment of forest resources of the tropics. Commonwealth Forestry Review, 44 (6): 287-318.
Lanly, J.P. 1988. An interim report on the state of the forest resources in the developing countries. FO: Misc/88/7. Rome, FAO.
Persson, R. 1974. World forest resources - review of the worlds forest resources in the early 1970s. Department of Forest Survey, Reports and Dissertations No. 17. Stockholm, Royal College of Forestry.
Persson, R. 1975. Forest resources of Africa - an approach to international forest resource appraisals, Part I: Country appraisals. Department of Forest Survey, Reports and Dissertations No. 18. Stockholm, Royal College of Forestry.
Persson, R. 1977. Forest resources of Africa - an approach to international forest resource appraisals, Part II: Regional analyses. Department of Forest Survey, Reports and Dissertations No. 22. Stockholm, Sweden, Royal College of Forestry.
Sommer, A. 1976. Attempt at an assessment of the worlds tropical moist forests. Unasylva, 112-113: 5-27.
Space, J. 1997. Strategic plan, Global Forest Resources Assessment 2000. Rome, FAO. (unpublished).
UNECE/FAO. 1985. The forest resources of the ECE
region (Europe, the USSR, North America). Geneva.