Previous Page Table of Contents Next Page

Chapter II


A fundamental precept in the formulation of the methodology of this study has been that qualitative and quantitative information on tropical forest resources is at the same time abundant, scattered and diverse. Each of these qualifications requires consideration. The first one may appear paradoxical: many persons indeed think that relevant information is scarce. However, if they had enough time to search for it, they would find, on the contrary, that it exists in abundance.

The main reason why these data are difficult to obtain relates to the second aspect mentioned, that is their dispersion. There are data not only in the national and international forestry organizations but also in a large number of institutes such as survey departments (and remote sensing centres), agricultural statistics services, colonization and land-use institutes, universities and research organizations in the concerned countries or in developed countries, consulting firms etc. In this respect many thematic mapping studies were carried out at regional, national and subnational levels, in the 70's thanks to the use of remote sensing techniques (Landsat satellite and side looking airborne radar imagery, small-scale aerial photographs). The following examples can be quoted as among the most important in tropical Africa: vegetation maps drawn by the FAO/UNEP Pilot Project on Tropical Forest Cover Monitoring in Benin, Cameroon and Togo, the vegetation map at 1/250 000 for the whole of Nigeria carried out from interpretation of the radar coverage (NIRAD project), the forest map of Mozambique at 1/1 000 000 scale drawn by the UNDP/FAO/Mozambique Forestry and Forest Industries Development Project using Landsat imagery, and the vegetation and landuse map at 1/500 000 scale of the UNDP/FAO/Sierra Leone Land-Use Project, using small-scale false colour aerial photography. Within the framework of this project a certain number of institutions were visited, discussions were held with experts and important correspondence was maintained with many others. It has not been possible, of course, to visit or contact the very large number of national institutions which could have been in a position to provide some useful information, in order to resolve contradictions in the vailable data, and correct erroneous interpretations found in the documents. It is important, however, to underline the fact that a large part of the project activities consisted of collecting as many as possible of the relevant data scattered around the world.

The third characteristic of this wealth of information, is its diversity which can be inspected from at least three different angles:

This study has consisted mainly in the selection, organization, compilation and interpretation of this abundant and diverse mass of information using a single framework of classifications and concepts for the 76 tropical countries studies (see section 2). However, in some countries, reliable base-line data on the areas of woody vegetation which could have been used for subsequent up-dating, did not exist at national level. In other countries the project was confronted with two or more sets of area information which could not be matched. In these cases it was decided to interpret available satellite imagery (of the years 1972 – 1976) to check and possibly correct area estimates obtained in a first phase (see section 3). For all countries it has been necessary to up-date the information at the end of 1980 on the basis of the trends observed in the last years and to project the situation at the end of 1985, on the basis of an estimation of the trends in the next 5 years (see section 4).


The value and usefulness of any forest resources assessment study depends, to a large part, on the concepts and classifications used. These must have several characteristics which are not necessarily compatible. In particular they must be:

In addition to the above conditions, the concepts and classifications of the study must be compatible with those already used in the former FAO World Forest Inventory reports for purposes of comparison and consistency. All these conditions can be fulfilled if one adopts forestry concepts currently in use and classifications which are not too detailed.

2.1 Concepts and classifications of natural woody vegetation

2.1.1 A large number of systems of tropical vegetation classification already exist, using various criteria (ecological, physiognomic, physiographic) at both national and regional levels. In this latter category the following classifications can be identified for tropical Africa:

Vegetation classification at national level are many. They have been used in this study for those countries where there was a corresponding map. As it has already been mentioned in chapter 1 a new generation of vegetation classifications and maps has emerged in the last ten years using classification based on the interpretation of satellite and radar imagery. Classification criteria and categories differ widely not only from one country to another but also within the same country.

2.1.2 When deciding on the forest vegetation classification to be used in this study, great care has been exercised to make it compatible with the Unesco one because of the useful aspects of this latter i.e.:

2.1.3 In addition to the distinctions mentioned above (tree/shrub and closed forest/mixed forest-grassland formation) there exists other essential classification criteria for woody vegetation formations both from the productive and environmental viewpoints, such as:

The simultaneous use of all these criteria provides a large number of classes. Some of these classes are not important. Others cannot be identified from the interpreted documents and images and their areas and characteristics cannot therefore be determined. The classification which has been finally adopted is limited to the most useful categories. The following chart presents this classification with the corresponding criteria. The various categories are described in detail below.

2.1.4 Classification of natural woody vegetation (N/n)

The only vegetation types which are considered are those for which woody elements cover more than 10% of the ground. Though it is often difficult if not impossible to estimate this percentage from the descriptions and this percentage is not always used in the classifications, it has been selected as the limit between the types in which the woody elements constitute actually a community and those where they are scattered or (in lines) in landscapes with a non-woody vegetation or without any other vegetation.

The word “woody” is used although the trees of some monocotyledons do not contain “wood” in the usual meaning of the word.

The adjective “natural” is used only in relation with plantations which can be considered as a purely artificial vegetation (see below section 2.2). This does not mean at all that there is no human or, more generally, biotic interference. On the contrary, a significant proportion varying with countries of “natural vegetation” corresponds indeed to degradation stages (after fires, clearings by shifting cultivation, overexploitation for wood, grazing) or reconstitution stages after degradation, or to forests disturbed by logging, with or without management.

1 These limits must be interpreted with flexibility, particularly the minimum tree height (and maximum shrub height) which may vary between 5 and 8 metres approximately.

A summarized definition of the various classes with their corresponding symbols, as used in the presentation of results, is given below (in the order they appear in the tables of area statistics):

NHCf1uv:undisturbed productive closed broadleaved forests not (intensively) managed;
NHCf1uc:logged-over productive closed broadleaved forests not (intensively) managed;
NHCf1u:productive closed broadleaved forests not (intensively) managed;
NHCf1m:(intensively) managed productive closed broadleaved forests;
NHCf1:productive closed broadleaved forests;
NHCf2i:closed broadleaved forests unproductive for physical reasons (stand and terrain characteristics);
NHCf2r:closed broadleaved forests unproductive for legal reasons;
NHCf2:unproductive closed broadleaved forests;
NHCf:closed broadleaved forests;
NHCa:forest fallow (of closed broadleaved forests).

Equivalent categories of coniferous, bamboo and closed forest have similar symbols in which NHC is replaced by NS, NHB and N. respectively.

NHc/NHO1:productive mixed broadleaved forest-grassland tree formations;
NHc/NHO2i:mixed broadleaved forest-grassland tree formations unproductive for physical reasons (stand and terrain characteristics);
NHc/NHO2r:mixed broadleaved forest-grassland tree formations unproductive for legal reasons;
NHc/NHO2:unproductive mixed broadleaved forest-grassland tree formations;
NHc/NHO:mixed broadleaved forest-grassland tree formations;
NHc/NHOa:forest fallow (of mixed broadleaved forest-grassland tree formations);
nH:(essentially) shrub formations (broadleaved).

2.2 Classification of plantations (P)

The term “plantation” corresponds to:1

Plantations in the sense used in this study do not include stands established by artificial regeneration and essentially similar to those they are replacing. These artificially regenerated forests are part of productive closed broadleaved (or coniferous) forests (intensively) managed (NHCf1m/NSf1m).

A distinction is made between industrial plantations (P..1) established totally or partly for production of wood for industry (sawlogs and veneer logs, pulpwood, pitprops mainly) and non-industrial plantations (or “other plantations”) (P..2) established mainly for one or severla of the following objectives:

Those tree plantations which are usually outside the competence of foresters are not accounted for. This is the case in particular of plantations of rubber trees, palm oil trees, coconut trees, and of the shade tree plantations for agriculture.

A distinction is made between plantations of broadleaved species, or hardwood plantations (PH.1/PH.2), and plantations of coniferous species, or softwood plantations (PS.1/PS.2).

Hardwood plantations are divided between plantations of fast-growing species (PHH1/PHH2) and plantations with other broadleaved species (PHL1/PHL2). Limit between these two groups of species corresponds approximately to a gross mean annual increment of 12–15 m3/ha/year. However, separation is made above all on the basis of the species. For instance Eucalyptus and Gmelina plantations are classified as fast-growing species (PHH1/PHH2), whereas teak plantations are included in the PHL1/PHL2 categories.

The categories of plantations are finally the following:

PHL1:industrial plantations of hardwood species other than fast-growing ones;
PHH1:industrial plantations of fast-growing hardwood species;
PH.1:industrial hardwood plantations;
PS.1:industrial softwood plantations;
P..1:industrial plantations.
PHL2:non-industrial plantations of hardwood species other than fast-growing ones;
PHH2:non-industrial plantations of fast-growing hardwood species;
PH.2:non-industrial hardwood plantations;
PS.2:non-industrial softwood plantations;
P..2:non-industrial plantations.

PHL-PHL1 + PHL2:plantations of hardwood species other than fast-growing ones;
PHH=PHH1 + PHH2:plantations of fast-growing hardwood species;
PH=PH.1 + PH.2 :hardwood plantations;
PS=PS.1 + PS.2 :softwood plantations;
P=P..1 + P..2 :all plantations.

1 The following categories were defined on the occasion of the World Symposium on Man-made Forests and their Industrial Importance (Canberra - Australia, 14–24 April 1967).

2.3 Concepts of volume

Three volume concepts (either mean volume per ha or total for a given forest category) are used throughout this study for closed broadleaved forests and coniferous forests (NHC-NS) and for productive mixed broadleaved forest-grassland tree formations (NHc/NHO1), which are:

VOB:gross volume over bark of free bole (from stump or buttresses to crown point or first main branch) of all living trees more than 10cm diameter at breast height (or above buttresses if they are higher);
VAC:(for forests not intensively managed): volume actually commercialised, that is volume under bark of logs actually extracted from the forest;
AAC:(for intensively managed forests): gross annual allowable cut, in general equated with current annual yield.


Lack of recent data at national level for Angola and Guinea prompted to decide the interpretation of Landsat imagery for these two countries. The main purpose of the work of satellite imagery interpretation was the checking and possible correction of estimates derived from a previous attempt to up-date maps and other available documents. Because of the global nature of this study, of the extent of the categories used and, in the case of Angola, of the impossibility of collecting detailed ground truth data, the work was limited to the visual interpretation of images, more precisely of the 1/1 000 000 scale positive transparencies of bands 5 and 7 and of the standard colour composition from bands 4, 5 and 7. The interpretation work benefited from the experience acquired by the FAO Forestry Department within the framework of the FAO/UNEP Pilot Project on Tropical Forest Cover Mointoring carried out in three countries of West Africa (Benin, Cameroun and Togo). The remote sensing expert in charge of all the interpretation work (J. Guellec) had participated in this pilot project.

The selected scenes are all images from Landsat 1 and 2, from 1972 to 1976 with cloud cover less than 10% above the territory of the countries concerned. The selection was made with the help of the Renote Sensing Unit of FAO using the microfiches of the print-out lists per country, and the microfilms of band 5 for checking the quality of each scene and the location and distribution of the clouds.

For both countries documents which could assist in the interpretation as “ground truth”, such as vegetation maps and forest inventory reports were used (a list of which is given at the end of the corresponding country briefs in the second part of this report), as well as the 1/1 000 000 scale aeronautical charts, particularly for the transfer of the international boundaries on to the images.

In the case of Angola the size of the country (more than one million km2) and the limited available time allowed the interpretation of half the images of acceptable quality only. Every second frame was selected on each orbit.

This global project does not aim at drawing forest maps but mainly at assessing the present situation and evolution of tropical forest resources qualitatively and quantitatively. Moreover the delineation of the various vegetation types from satellite images is not obligatory since area estimates can be obtained on a statistical basis through the identification of vegetation types at each dot of a grid. For this reason the latter device has been preferred for estimating the area of each interpretation class. A systematic dot grid on transparent stable material, with a 5mm by 5mm spacing in the directions parallel to the sides of the images, was applied on the transparencies observed on a mirror stereoscope. Band 5 or band 7 was usually visualised simultaneously with the colour composition on the stereoscope. The dot grid was limited to the effective part of the images, taking into account an average lateral overlap of 20% in the tropical regions1 and a forward lap along the orbit of 10%. Before the interpretation itself, some important features were indicated on one transparency of each scene, such as international boundaries, rivers, important roads in order to facilitate orientation.

This method has been adopted after having compared its results on an experimental basis with those of a more complete procedure including the delineation of vegetation classes and the subsequent use of a dot grid for area estimation. Differences for each class, such as closed forests, open forests, degraded forests, were not systematic and did not exceed 4% when these types were in the form of large areas. Differences were high and systematic for classes represented by scattered patches of small dimensions. In this latter case the smallest patches are not delineated and the total area of the corresponding class is underestimated, while the estimation by interpretation of dots is not biased from this aspect. Another advantage of the statistical method is to allow for reduction of the personal bias of the interpreter in the delineation of the class, more particularly in the transition zones where the drawing of the limits is often somewhat subjective.

The interpretation key is compatible with the general classification used in this project (see paragraph 2.1.3). The following distinctions have been introduced:

Other separations, for instance between mixed forest-grassland formation with trees and those with shrubs, were made on the basis of available phytogeographic maps and other documents.

Mangroves and large areas of swamp formations were also identified. Other distinctions, such as the separation between productive and unproductive forests, and undisturbed and logged-over forests were not feasible through the interpretation of Landsat images and the corresponding area estimates were obtained from the compilation of other documents.

In total 37 scenes corresponding to a total area of 79.2 million ha were used corresponding to approximately 31500 interpreted dots for the whole area of Guinea and approximately half that of Angola.

Use has also been made in this study of the results of remote sensing interpretation carried out during the last years for vegetation mapping at national level in the following countries: Benin, Cameroon, Guinea-Bissau, Mozambique, Nigeria, Senegal, Sierra Leone, Togo and Upper Volta.

1 The average overlap is approximately 14% on the equator and 24% on the tropics (23o27').


4.1 Country briefs

The first part of this report summarizes the results obtained for the whole of the 37 countries concerned of tropical Africa, while the second part contains the country briefs. 36 countries have been studied in detail. They are all those of continental tropical Africa south of the Sahara, plus Madagascar, less Mauritania, Djibouti and Zimbabwe. Lesotho, Swaziland and Republic of South Africa are not included as being outside the inter-tropical zone. Provisional results for Zimbabwe are inserted in the tables but there has been no detailed study for that country.

4.1.1 Text

The outline is the same for all country briefs. The present situation of forest resources and their trends are described in two separate sections, each of them with a part describing natural woody vegetation and another one on plantations.

Description of the composition and physiognomy of the various types of natural woody vegetation (paragraph 1.1.1) is followed by an estimation of areas of natural woody vegetation at the end of 1980 and by information on ownership, status, management and utilization of the forests (paragraph 1.1.2). The interpretation of available forest inventory results allows for an estimation of growing stock at national level at the end of 1980 (paragraph 1.1.3).

The comments on forest plantations contain an introduction (paragraph 1.2.1) dealing in particular with historical aspects followed by the estimation of forest plantation areas at the end of 1980, separately for industrial plantations and other plantations, each group by species categories and age classes (paragraph 1.2.2). Quantitative data on plantation characteristics, particularly on mean annual increments, are given in paragraph 1.2.3.

The 37 studied countries of tropical Africa 1


  1. Chad
  2. Gambia
  3. Mali
  4. Niger
  5. Senegal
  6. Upper Volta


  1. Benin
  2. Ghana
  3. Guinea
  4. Guinea Bissau
  5. Ivory Coast
  6. Liberia
  7. Nigeria
  8. Sierra Leone
  9. Togo


  1. Angola
  2. Cameroon
  3. Central African Republic
  4. Congo
  5. Equatorial Guinea
  6. Gabon
  7. Zaire


  1. Burundi
  2. Ethiopia
  3. Kenya
  4. Madagascar
  5. Malawi
  6. Mozambique
  7. Rwanda
  8. Somalia
  9. Sudan
  10. Tanzania
  11. Uganda
  12. Zambia
  13. Zimbabwe


  1. Botswana
  2. Namibia

1 Although area and volume estimates for Zimbabwe (35) are included in the summary tables of the regional synthesis, there is no corresponding “country brief”.

In section 2.1 on present trends of natural woody vegetation, an important distinction has been introduced between, on one hand, deforestation in the strict definition of the term (paragraph 2.1.1), i.e. alienation of forest areas to permanent or shifting agriculture or to other uses, and, on the other hand, degradation of woody vegetation (in particular of mixed forest-grassland formations) which results from other factors such as fire, overgrazing, overexploitation for fuelwood and charcoal, etc (paragraph 2.1.2). In most cases degradation does not show up so much as a decrease in the area of woody vegetation but rather as a gradual reduction of biomass, changes in specific composition and soil degradation. Unfortunately these changes are very seldom quantified, and never at national or sub-national levels. Taking into account trends in forest utilization (paragraph 2.1.3), area and growing stock estimates are then projected at the end of 1985 (paragraph 2.1.4).

In section 2.2, forest plantation programmes are mentioned as well as their probable rates of implementation in the period 1981–85. This allows for the projection of planted areas by species category up to the end of 1985.

A bibliography at the end of each country brief lists the main documents which have been used for assessing forest resources and their trends. These references are presented in chronological order, since the date of publication is particularly relevant to such a study.

4.1.2 Tables

In addition to some tables related to less important issues a number of basic tables have been set out to illustrate the text, as shown below:

4.2 Results at regional level

The presentation of results at regional level in the next chapter uses the same outline as that of the country briefs. In the tables each line corresponds to a country and the countries are grouped in the five following subregions:

Northern Savanna Region (6):Chad, Gambia, Mali, Niger, Senegal, Upper Volta;
West Africa (9):Benin, Ghana, Guinea, Guinea-Bissau, Ivory Coast, Liberia, Nigeria, Sierra Leone, Togo;
Central Africa (7):Angola, Cameroon, Central African Republic, Congo, Equatorial Guinea, Gabon, Zaire;
East Africa and Madagascar (13):Burundi, Ethiopia, Kenya, Madagascar, Malawi, Mozambique, Rwanda, Somalia, Sudan, Tanzania, Uganda, Zambia, Zimbabwe;
Tropical South Africa (2):Botswana, Namibia.

Previous Page Top of Page Next Page