S. DembnerThe Forest Resources Assessment 1990 Project was launched in 1989 to provide current and objective information about the existing conditions of the world's tropical forest resources. This article, based on a report to the tenth session of the FAO Committee on Forestry, held at FAO headquarters in September 1990, provides a general overview of the Forest Resources Assessment 1990 Project and highlights from provisional data on deforestation in humid tropical forests.
Stephen A. Dembner is the Editor of Unasylva.
The status and condition of the world's forests are a source of increasing concern to the international community. The Forest Resources Assessment 1990 Project falls within FAO's overall mandate to "collect, analyse, interpret and disseminate information relating to nutrition, food and agriculture... including... forestry and primary forest products" (Article 1, FAO Constitution).
The first FAO survey of forest resources was carried out in 1946 on the basis of questionnaires sent to all countries. Similar surveys provided information on world forests with 1953, 1958 and 1963 as reference years.
The most recent worldwide forest resources assessment, carried out with 1980 as the reference date, took a completely new approach, dividing the overall effort into two parts, each with a distinct methodology: the FAO/ECE Survey, covering most developed countries; and the FAO/UNEP Tropical Forest Resources Assessment Project.
The work of the global assessment for 1990 also comprises two main components: a survey of forest resources of the developed countries, again coordinated by the FAO/ECE Agriculture and Timber Division in Geneva; and an assessment of the forest resources of the developing world, carried out by the Forest Resources Assessment 1990 Project at FAO, Rome, which will also be responsible for the global synthesis.
In line with its overall goal, mentioned above, the project has three immediate objectives:
· to provide current, objective and globally consistent information on the state of the world's tropical forests, including recent trends of deforestation and forest degradation, and the underlying causes and effects;· to provide a tested methodology and a forest resources classification standard that will permit ongoing integration and updating of global and national level information on forest resources;
· to coordinate international activities and strengthen the facilities and capabilities of national institutions to monitor forest resources on a continuing basis.
Project data sources
The tropical forest resources assessment is based on the analysis and interpretation of many types of data, both existing and new, classified broadly as follows:
· statistical data at the subnational level (province, state), including forest resources data, population data and socio-economic data;· map data covering vegetation, eco-floristic zones, country and subnational boundaries;
· satellite data, including coarse-resolution satellite imagery (global coverage), and multidate high-resolution satellite imagery (sample coverage).
Planned project outputs
Upon completion in mid-1992, the project is expected to have produced the following outputs:
· global, regional and country reports on the status of tropical forest resources;· models of tropical deforestation, biomass estimation and associated environmental/socio-economic causes and effects;
· a tested methodology for global continuous forest inventories;
· a global standard for classifying tropical forest conditions and land uses;
· a series of statistical and spatial data bases;
· an archive comprising multidate satellite data;
· strengthened global, regional and national institutions and support networks for implementing continuous assessment of forest resources.
There has been steadily growing concern that deforestation has accelerated over the past decade and a persistent demand for reliable information on the subject. Therefore, in September 1989 the FAO Committee on Forest Development in the Tropics (CFDT) suggested that FAO provide an interim report on the condition of the tropical forests in selected areas, well before the date foreseen for the completion of the Forest Resources Assessment 1990 Project (i.e. mid-1992). Accordingly, it was decided that action should be initiated to produce global estimates of the status of the tropical moist forest as soon as possible.
The estimates of the rates of deforestation provided below should be considered provisional, pending completion of the analysis of existing data and use of additional information generated by the planned remote-sensing studies of the project.
Database for the preliminary report
This preliminary data covers 62 countries, comprising nearly 80 percent of the geographic area of the countries, which are located predominantly in the moist tropics. Existing data were collected and analysed for 400 subnational-level units (provinces in most cases).
Most of the existing information on forest cover comes from remote sensing-based inventories and studies which included field checking. Data sources include reliable observations (on a single date as close to 1990 as possible) of the forest cover area for all 400 subnational units; reliable observations of the forest cover area, on two or more dates, for a subsample of 78 well-distributed units; forest cover area estimates for western and central Africa, derived from coarse resolution satellite data; and other existing information, such as area distribution by eco-floristic zone and multidate population statistics.
Estimation technique
At an early stage of the project, analysis indicated that a reliable assessment of forest cover area for 1990 could be made using existing forest cover data (i.e. one observation per unit). To estimate the rate of change, a two-step procedure was formulated. First, a function was developed correlating the rate of change of the forest cover for the 78 units for which multiple observations were available with the initial observations of forest cover, estimated population on corresponding dates and area distribution by eco-floristic zone for the units. Correlation was highly significant (R2 = 0.94). The function was tested by applying it to subnational units of five countries for which forest cover observations for two dates were available but were not included in development of the model. The forest cover estimated by the model for the second date differed from the observed forest cover by 1.6 percent, and the estimated rate of change differed from the observed rate by less than 10 percent over a ten-year period.
In the second step, the function was used to estimate the rate of deforestation for the 400 units which had one date forest cover data available, demographic data on two dates and eco-floristic zone data. The average national rate of deforestation was derived from the subnational results, the regional rate from the national rates, and the global rate from the regional rates. Keeping this summation process in view, the reliability of the regional estimate is higher than that of the national estimates, while the reliability of the global estimate is higher than that of the regional ones.
Results of statistical analysis
For the countries studied, provisional estimates of the status of forest cover area and rates of deforestation are given in Table 1.
TABLE 1. Provisional estimates of forest cover and deforestation for 62 countries In the tropical regions
|
Continent |
Number countries studied |
Total land area |
Forest area 1980 |
Forest area 1990 |
Annual deforest. 1981-90 |
Rate of change 1981-90 (%/year) |
|
(thousands of ha) |
||||||
|
Africa |
15 |
609800 |
289700 |
241800 |
4800 |
-1.7 |
|
Latin America |
32 |
1263600 |
825900 |
753000 |
7300 |
-0.9 |
|
Asia |
15 |
891100 |
334500 |
287500 |
4700 |
-1.4 |
|
Total |
62 |
2764500 |
1450100 |
1282300 |
16800 |
-1.2 |
Note: Countries include almost all of the moist forest zone, along with some dry areas. Figures are indicative, and should not be taken as regional averages.
Forests are defined here as ecological systems with a minimum of 10 percent crown cover of trees and/or bamboos, generally associated with wild flora, fauna and natural soil conditions, and not subject to agricultural practices. Deforestation refers to change of land use or depletion of crown cover to less than 10 percent. Changes within the forest class which negatively affect the stand or site and, in particular, lower the production capacity are termed forest degradation. Degradation is not reflected in the estimates.
It is important to note that the data in Table 1 covers countries of the moist tropical zone only. For this zone it gives indications that may be considered reliable within the limits of the validity of the input data and methods used. However, it cannot be assumed that parallel developments have taken place in the dry and mountainous tropical zones. Therefore, the rate of deforestation for the whole tropical zone, as calculated by the 1980 FAO/UNEP Tropical Forest Resources Assessment Project; cannot currently be updated. However, comparison with the 1980 assessment is possible for 52 countries covered by both assessments. The figures for the 52 countries in common are provided in Table 2.
TABLE 2. Comparison of final 1980 and provisional 1990 forest assessment results for 52 countries covered in both assessments
|
Estimate |
1980 FAO/UNEP assessment |
1990 FAO assessment |
|
(thousands of ha) |
||
|
Forest area |
||
|
by end 1980 |
1529000 |
1449300 |
|
by end 1990 |
|
1281500 |
|
Annual rate of deforestation |
||
|
during 1976-80 |
9200 |
- |
|
during 1981-90 |
- |
16800 |
While there is a reasonable convergence of the total forest area estimates for 1980 in both the 1980 and 1990 assessments, the estimates of deforestation rates, for the period 1976-80 by the former and for 1981-90 by the latter assessment, are significantly different. Since the definitions of forest and deforestation are basically the same in the two assessments, the differences must be attributed to one or more of three factors' an actual increase in the rate of deforestation between the 1976-80 and 1981-90 periods; an underestimation of the rate of deforestation in the 1980 assessment; and an overestimation of the rate of deforestation in the 1990 assessment.
It is not possible at this stage to assess the relative contribution of the various components. However, a number of general remarks may be made. First, comparison of estimated deforestation rates for the two decades 1971-80 and 1981-90 shows that deforestation has accelerated in the tropical moist region as a whole. It has slowed down in a group of densely populated Asian countries, not because adequate forest policies have been implemented, but mainly because what little forest is left is mostly located in inaccessible areas. Second, FAO now believes that the 1980 assessment underestimated the rate of deforestation for the period 1976-80 in some large Asian countries.
Finally, it should be noted that the provisional results of the 1990 assessment presented here are also subject to error, attributable to both the input data and the models used. From tests made there is reason to believe that the model errors are small. It should be stressed, however, that the input data used originated from inventories made in the countries with varying aims, methods, accuracy goals and definitions. Only the final results will be based on uniform remote-sensing observations specifically made for the project.
Analysis of map-based data
In addition to the estimations of forest cover area and its changes over time, the project has begun to quantify an aspect of forest distribution through the use of the Geographic Information System (GIS). Using a forest cover map derived from coarse-resolution satellite imagery, forest edges were measured to derive fragmentation indices, i.e. the amount of boundary area between forest and non-forest land compared with the area of total forest. A comparison of fragmentation for seven West African countries is shown in the figure. The higher the index value, the more complex is the edge between closed forests and other land conditions. The unevenness of distribution carries a number of socio-economic and environmental implications, and may also be an indicator of the likelihood of further deforestation.
Biomass degradation
The project is also involved in analysis of changes in natural forest area and forest biomass, based on inventories at national or subnational level for at least two periods. The results to date generally show that the decrease in forest biomass is considerably greater than the decrease in forest area. There are two possible explanations for this: deforestation is occurring disproportionately on forest land with higher biomass levels; or remaining forests are being degraded through removal of biomass. Biomass degradation seems to be related to increases in the fragmentation index; the more a forest is fragmented, the more it appears to be subject to degradation. This analysis points to the need for improved land-use planning in order to conserve forest resources, with attention given to location of human settlement and infrastructure development in the remaining forest areas.
In the course of its implementation, the project has made an in-depth study of the state of forest inventory in the developing world, compiled existing reliable data on forest resources (including area, volume/biomass, production, etc.), collected new data using remote-sensing and made a special study on the deforestation process. It has been observed that most forest inventories to date, particularly in tropical countries, have been separate and unrelated exercises carried out at varying intervals and therefore do not directly provide reliable information on forest change. Furthermore, deforestation processes are complex interactions of human and environmental factors which require time series data for complete understanding. There is a clear need to develop a system of continuous national and global monitoring of forest resources. The current critical state of the resources and the depth of concern for their conservation demand nothing less. Based on project activities to date, two important conclusions can be drawn.
First, there is a need to expand the scope of forest inventories in order to make them responsive to a greater variety of user needs and applications,: including questions of global climate change, bio-diversity, biomass, ecological zone influences, wildlife habitat, desertification in dry areas, carbon uptake, etc.
Second, national forestry action plans, including those being developed under the Tropical Forestry Action Plan, should give priority to the strengthening of the institutional infrastructure and technical capabilities needed for national continuous forest resources monitoring.
Unasylva will continue to keep its readers abreast of progress in the implementation of the Forest Resources Assessment 1990 Project, and will publish a summary of the final data when they become available in mid-1992.
