MS12A
Jean-Paul Lanly1
An accurate analysis of deforestation and forest degradation requires that clear distinctions be made between these two terms, between the factors and underlying causes of these processes and between direct (e.g. the different forms of agriculture) and indirect (e.g. the development of logging road infrastructure)factors. A very large number of deforestation studies, essentially in the tropics, were carried out in the last 40 years at all levels and were facilitated by remote sensing. However they seldom included an objective assessment of the respective share of the various direct deforestation factors. The situation is even less satisfactory regarding forest degradation due in particular to the imprecision and multiple, and often subjective, interpretations of the term and the gradation it implies. FAO forest assessments of 1980 to 2000 tried to determine the relative importance of direct deforestation factors at regional and global levels: horizontal expansion of the various forms of agriculture and animal husbandry taken together remains by far the most important direct factor, but the share of shifting cultivation as well as that of ranching and colonization programmes decreased during that period. Forest area changes in industrialized countries, generally positive, have been less the subject of study, contrary to what has been the case of forest degradation resulting from air pollution, fires, insects and diseases. In conclusion, stress is put on the need for every country to strengthen its capacity in all forest inventory disciplines in support of sound forest land use and management.
It is important at the outset to carefully distinguish between the underlying causes of deforestation and forest degradation, about which there might be divergent views, and the actual factors, which could be part of the area of objective observation. For instance, in dealing with deforestation in developing tropical countries, the expansion, by means of clearing, of the different forms of subsistence agriculture, of cash cropping, or of ranching are all evident factors. The underlying causes are the triggering mechanisms for these factors. In the case of the first factor, we might say that the farmers' poverty obliges them to clear more land in order to enable them to buy the inputs that would allow them to produce more on less land, and that their poverty is itself engendered by their difficulty to sell what they produce, as a result of the insufficient farm price supports, due to the low priority given to the farming sector. We can go back a long way in this manner to find the underlying causes, with the increasing risk however of uncertainty, subjectivity and ideological posing. We will therefore limit ourselves to studying the factors.
The factors are either direct -in the case of deforestation in developing tropical countries, the replacement of a forest area by a field of coffee trees is due to cash cropping which is the direct and visible factor -, or indirect, as for example the opening of a road (public or for forest logging) in a forest zone, which, in addition to being a direct factor of relatively small significance (eliminating a strip of forest along the road) attracts farmers in search of land who will clear the forest for cultivation. With regard to the area cleared by these farmers, there is a direct factor - farming - as well as an indirect one - the colonization or forest logging programme. There is a question to be asked, as well as a precautionary measure to be taken in this regard:
The confusion that reigns between the two notions of deforestation and forest degradation has all too often been kept alive, or unconsciously made. In order to avoid any ambiguity, we might like to recall a number of elements that should be kept in mind:
The ambiguities of the term degradation, and the difficulties of estimating it, are additional reasons for clearly differentiating between deforestation and degradation.
It has often been the practice, particularly with regard to the tropics, to record separately changes in the "natural" or "semi-natural" forest areas - whether positive (natural reforestation) or negative (deforestation), from those related to man made forests (plantations), given the generally very different nature of these two categories of stands. Moreover, it is important to differentiate clearly between:
Characterizing deforestation at a given time and place involves as a rule determining with some certainty what the more or less long-term future of the deforested area will be. Can natural or artificial reforestation develop, or on the contrary, is the clearing or clearfelling the occasion or the first stage in the conversion of a forest area into one involving a non forest use? There is no deforestation if the clearfelling is done on an area that is meant to be maintained as a forest (as in the case of "temporarily unstocked" forests); deforestation on the other hand does exist - and this is actually the point of view of forest management - when the forest in question is cleared in order to be cultivated or abandoned for a long time, and if its regeneration cannot take place before several decades have passed.
In addition to the time factor, the size of the elementary land unit whose cover is qualified - which can vary from less than one hectare to as much as several hundreds of hectares - must be taken into account, or else the resolution of the remote sensing imagery used for classifying and mapping and/or estimating the areas of the different categories. One is often obliged to utilize the so-called "mosaic" categories made of a mixture of patches of different land cover categories. The estimated values of forest areas and the changes they undergo over time differ in particular according to the remote sensing tools and methods that are used, a fact that makes it difficult to compare estimates from one country to another or from one period to another.
We must also identify the deforestation factor: it is locally known, and its relative extent in relation to other factors can theoretically be estimated by aggregation at higher levels. The difficulty arises from the fact that it is not systematically recorded locally, and is generally inferred globally at the level of a country (or of each of its large divisions) from the prevailing socio-economic and land ownership conditions.
These two identifications - of the phenomenon and its factor - and the estimation of the areas involved, which are difficult enough with regard to deforestation, are even more so with regard to degradation.
Continual progress has been made during the past fifty years with regard to identifying deforestation and estimating the amount of deforested areas, thanks in particular to progress in remote sensing technology. The early phase of unlimited fascination with satellite and radar remote sensing in the 1970s and 1980s has fortunately given way to a more lucid awareness of the limitations of these tools, although considerable damage had already been done during the intervening period, namely, in the adoption of classifications and results that proved to be unusable by managers (and could not be applied to adjacent areas) and the reduction, up to total neglect, in the acquisition of "groundtruth" data and in field inventories.
Identifying and estimating deforestation implies the knowledge, at the relevant level, of the land cover at two dates. There are still unfortunately too few continuous forest inventories at the national or lower levels and most of them are carried out in the industrialized countries. At the international level, the assessments (beginning with the global ones of FAO) are only of value insofar as they are based on a reliable and repeated "groundtruth", which requires capacity building in this field in developing countries.
The degradation estimates of wooded lands are only accurate as a general rule at the local level on limited areas. At higher levels such estimates, when they exist, are most imprecise. Whereas the different forms of remote sensing are very useful tools for estimating deforestation, they are far less so for assessing degradation which most often calls for observations on the ground.
The difficulty in estimating degradation also lies in the need to differentiate between its forms and their degrees of gravity. The deforestation variable is binary - 0 or 1 -, whereas the degradation variable cannot be, at the risk of oversimplification; there is no comparison for example between the impact of selective felling on a tropical high forest for timber harvesting, and that of permanent overexploitation (fuelwood, grazing) of Mediterranean forest formations that can result in true deforestation.
The last three global forest assessments carried out by FAO with 1980, 1990 and 2000 respectively as reference years attempted to determine for each tropical country the areas occupied by the different types of forest cover, as well as forestation and deforestation.
The 1990 inventory introduced a number of interesting approaches for assessing changes in forest area but did not attempt to determine the respective share of the different deforestation factors. The 1980 inventory did this essentially with regard to shifting cultivation, and the 2000 inventory did this for the main factors.
The FAO 1980 inventory Annual deforestation in the tropical and subtropical world was estimated around 1980 at 11.3 million hectares (Mha), equivalent to 0.6% of the total forest area estimated at the time (1 935 Mha, 1 200 Mha of which were in high forests). Shifting cultivation was found to be the direct factor for 45% of deforestation; this proportion was nearly the same for high forests and for the open formations in the dry tropics but varied considerably according to the main regions, as the following table indicates.
Table 1. Proportion of deforestation attributed to shifting cultivation (with rotation) around 1980(in percentages)
Types of forest Regions |
high forests |
open forests |
all forests |
Tropical Africa |
70 |
60 |
64 |
Tropical America |
35 |
20 |
32 |
Tropical Asia |
49 |
27 |
47 |
Furthermore, global estimates (limited only to high forests and based upon the 1980 FAO inventory data and other sources) have been made, resulting in the following approximate proportions by direct deforestation factor at the beginning of the 1980s:
subsistence farming 63% }
permanent agriculture 16% } 84.5%
cattle ranching 6.5% }
(over)exploitation for fuelwood 8% }
(over)exploitation for timber 5.5%
infrastructure (roads, cities, mines...) 1%
In the humid tropics the horizontal expansion of the different forms of agriculture (and animal husbandry) constitutes the most important direct overall factor, since it is responsible for nearly 85% of deforestation. These figures are compatible with that of 45% of deforestation in the humid tropics due to shifting cultivation estimated by the 1980 FAO inventory, inasmuch as subsistence agriculture includes not only shifting cultivation as such, i.e. with rotation, but all forms of shifting cultivation without rotation as well: subsistence farming opening the way to cattle ranching in tropical America, or moving to very poor soils not likely to be recolonized by forest vegetation for a considerably long time (steep slopes in the Andes, the Himalayas and other mountain areas, and "laterized" soils after cropping, such as West African "bowes" or "alang-alang" areas in Southeast Asia).
These estimates also show the relative importance of ranching as a direct deforestation factor as practised for the most part in tropical America, which constitutes a particularly unsustainable and wasteful form of land use.
Although very approximate, these estimates also convey the fact that overexploitation - either for fuelwood or timber - can result in long-lasting deforestation, with the tree cover being reduced at a lower percentage than the minimum one set in the forest definition. This is the case of fuelwood collection around large towns for the supply of energy for households and small industries, and that of timber logging in forests rich in commercial species (such as monospecific edaphic stands and Southeast Asian Dipterocarp forests).
The FAO 2000 inventory ("FRA 2000 ")
By interpreting a sample of high-resolution satellite images that were taken around the years 1988 and 1997 respectively and spread over the entire inter-tropical belt, FRA 2000 analysed the transfers between the large classes of land cover. The negative transfers (in the forestry sense) - i.e. involving one or more forms of deforestation - between these six classes correspond to the boxes appearing above the diagonal of the change matrix below.
Table 2. Deforestation transfers between land cover classes
2000 Classes |
closed forest |
open forest |
long fallow |
fragmented |
short fallow |
other land uses |
closed forest |
(1) |
(1) |
(3) |
(3) |
(4) | |
open forest |
- |
(1) |
(3) |
(3) |
(4) | |
long fallow |
- |
- |
(2) |
(2) |
(4) | |
fragmented forest |
- |
- |
- |
(3) |
(3) | |
short fallow |
- |
- |
- |
- |
(2) | |
other land uses |
- |
- |
- |
- |
- |
These transfers are grouped into four land use categories as follows:
(1) "expansion of shifting cultivation into undisturbed forests";
(2) "intensification of agriculture in shifting cultivation areas": areas "in which shifting cultivation had become Intense (where the fallow period decreased)" and areas "where a complete transition from shifting to permanent agriculture had occurred";
(3) "direct conversion of forests to small-scale permanent agriculture" (conversion of small forest areas of less than 25 ha to agriculture);
(4) "direct conversion of forest area to large-scale agriculture".
By adding together the deforestation transfer areas that make up each of these four categories, one arrives at the following distribution of deforested areas for the whole tropical world during the period from 1988 to 1997.
Table 3. Distribution of deforested areas by main tropical regions (1988-1997)
Categories |
(1) |
(2) |
(3) |
(4) |
(1) to (4) |
Tropical Africa |
4.2% |
9.6% |
72.3% |
13.9% |
100.0% |
Tropical America |
3.1% |
5.3% |
19.8% |
71.8% |
100.0% |
Tropical Asia |
12.1% |
30.9% |
17.4% |
39.6% |
100.0% |
Tropical world |
6.2% |
13.8% |
35.2% |
44.8% |
100.0% |
Shifting cultivation [classes (1) and (2), the latter corresponding in part to the conversion of shifting to permanent agriculture] was only responsible for a 20% maximum of deforestation in the period from 1990 to 2000, compared to the 45% it was found at the end of the 1970s.
During approximately the last 20 years of the past century, shifting cultivation has undoubtedly played a reduced role in tropical deforestation if only because the available forest areas are becoming rare. It would nevertheless be imprudent to come to a too hasty conclusion concerning the importance of this reduction for at least two reasons: the imprecise nature of the 45% proportion that had been estimated in the 1980 inventory, and the simplification introduced in the 2000 inventory by the "translating" of transfers between land cover classes into changes in land use [categories (1) and (4)].
There are two factors other than shifting cultivation that will have undoubtedly declined: ranching, as a result of the elimination of tax incentives for companies investing in this form of speculation in Brazil, and permanent agriculture linked to governmental colonization programmes that have been suspended or reduced in certain regions.
Following the 1980 inventory of the forestresources of tropical and subtropical countries (published in 1982), between 1985 and 1995 and then again for the 2000 inventory, FAO analysed the situation and the evolution of forests in non-tropical developing countries without however attempting to determine the part played in deforestation by the different factors.
The Mediterranean countries (North Africa and the Near East) have been the subject of two studies, in 1987-88 and 1993-94 respectively. These studies indicated that the overall deforestation rate was greater than that of the tropical world (more than 1% as against 0.6%), which is all the more serious inasmuch as it affects countries in which the forest cover is already very reduced (an average 1% forest cover as against 40% for the tropical countries in 1980). In these countries, deforestation corresponds for the most part to the final stages of degradation resulting from overgrazing, compounded, in areas where oil products are too expensive for the majority of the population, by overexploitation for wood energy.
In the other non-tropical developing countries - southern Africa, the southern tip of the Americas, central Asia and China - the dominant factors in the clearing of natural forests are still the different forms of agriculture and animal husbandry.
In preparing the FRA 2000 inventory, the UNECC/FAO requested the industrialized countries to provide data with regard to the area changes of forests and other wooded lands, by comparing their national inventories for two different dates and adjusting the data for the two dates to the definition adopted for FRA 2000. But the countries involved were not asked to indicate what portion of the changes were due to the different deforestation factors.
In the great majority of industrialized countries, forests are expanding on lands abandoned by farming and animal husbandry. These changes generally occur more or less progressively through an intermediate stage of suffrutescent or shrubby wood formations that are classed in the vast category of "other wooded lands". The deforestation factors essentially include the development of residential units in the outlying areas around cities, as well as diverse infrastructures (particularly transportation and commercial areas).
Very stringent regulations - made necessary in the past by the negative economic and environmental effects of excessive clearing - combined with the growing need for nature of a largely urban population, have reduced this encroachment particularly in the European countries (excluding the Russian Federation). These countries, with a few rare exceptions including Albania, have seen an increase in their forest area, with their annual growth rate reaching or even exceeding 1% for those in the Mediterranean area. Forest lands have also been increasing in the other countries, with Russia being the only country to report a large reduction of its forest area (- 1.1 Mha per year), together with a considerable increase in "other wooded lands" (+ 1.6 Mha per year).
History has shown that human beings have most often considered the forest as a space that must be cleared in order to develop activities other than forestry (particularly farming), and used, eventually beyond its capacity to regenerate itself, as a wood and forage resource. It was only in the face of serious shortages in timber and wood energy, or the degradation of forest lands caused by deforestation or overexploitation that voluntary national policies aiming at the regeneration of forest lands were implemented in a generally authoritative manner. Following a more or less rapid phase of decline, the forest area of a country levels off and then increases by colonization of the anciently wooded areas, an evolution that is amplified by economic development, urbanization and the abandoning of the countryside. This general pattern, with its variations and breaks, unfolds over different periods depending upon the regions of the world. The industrialized and emerging countries are presently in a phase of forest restoration, unlike the developing world, which is generally in a deforestation phase.
Forest managers are aware that their margin of manoeuvre is limited; the occupation and utilization of forest lands are by their very nature essentially "residual" and most dependent on what happens in the other sectors of human activity. They must of course defend the forests and their sustainable management wherever the forestry stakes are high in terms of the well-being of local populations and national communities, while accepting the fact that the forests will disappear where clearing can be justified within the framework of sustainable development.
In all countries, the decisions to be taken with regard to the occupation of forest lands and the utilization and management of forests require a good knowledge of the situation and the changes involved at all times and at the appropriate levels. This presupposes the development of human resources and the strengthening of the units responsible for forest inventory in each country, as well as at the international level. This activity should be perceived as a whole with all of its disciplines, including the most modern and appealing, such as remote sensing and geographic information systems, as well as the more classical and frequently neglected ones of forest mensuration and taxonomy.
FAO. 1994. Evaluation of forestry resources 1990 - Non-tropical developing countries, Mediterranean Region - FO: MISC/94/3. Rome. 48 pp.
FAO. 1995. Forest resources assessment 1990 - Non-tropical developing countries. Technical Report (FO: GPC/INT/474/FRA - FO: GCP/INT/475/NET). Rome. 152 pp.
FAO. 1996. Forest resources assessment 1990 - Survey of tropical cover and study of change processes. FAO Forestry Paper 130. Rome. 152 pp.
FAO. 2002. Evaluation of world forestry resources 2000. Main Report. FAO Study: Forests 140. Rome. 466 pp.
FAO. 2001. Global forest resources assessment 2000. Main Report. FAO Forestry Paper 140. Rome. 479 pp.
Lanly, J.P. 1982. Tropical forestry resources. FAO Study: Forests 30. Rome. 113 pp.
Marchand, H. 1986. North African forestry resources. Working document. Rome. 34 pp.
Marchand, H. 1987. Near Eastern forestry resources. Working Document. Rome. 31 pp.
Myers, N. 1983. Conversion rates in tropical moist forests. In: Ecosystems of the world Volume 14a "Tropical rainforest ecosystems - Structure and functions". Elsevier.
UNECE/FAO. 2000. Forest resources of Europe, CIS, North America, Australia, Japan and New Zealand. Main Report. ECE/TIM/SP/17. Geneva. 445 pp.
1 Ingénieur Général honoraire du Génie Rural, des Eaux et des Forêts, 42 rue Albert Thomas, 75010 Paris, France.
[email protected]