Areas deforested in Brazil increased from 152 000 km_ in 1976 to 517 000 km_ in 1996. That figure is the equivalent of the surface area of France (1). Deforestation is a complex process and involves a host of changing and widely differing situations. The factors behind it are many and varied. They include rising demand for agricultural land, international trade needs for timber and political decisions regarding strategic planning and development.

Researchers of the IRD Space Unit conducted an investigation from January 2000 to December 2002, on deforestation in the Amazon Basin. Their overall procedure is based on the use of remote-sensing data obtained from satellites and aircraft. The approach adopted, termed “descending hierarchical”, often used by geographers, was first applied in French Guiana. It consists in analysing the Amazonian forest cover making the most of a full range of remote-sensing tools which provide means of taking into account the different scales involved. These scales are: the regional scale, studied using satellite images of NOAA-AVHRR and Spot 4 VEGETATION sensors which give coverage of terrains of surface area of 2000 km_; the sub-regional scale, with Landsat and Spot images which provide more detailed data on areas of 185 and 60 km_ respectively; local scale (a few tens of kilometres square observed using aerial photography). Together these tools enabled the researchers were therefore able to demonstrate the size diversity of cleared plots in the Amazon forest, from extensive areas of deforestation like the pioneer fronts of Brazil to the small 1-ha plots epitomized by brushwood plots in French Guiana.

In parallel, the team processed a mass of field measurements and socio-economic data, for French Guiana in particular, in order to calibrate the available aerial or satellite observations, interpret them and place them in the prevailing context. Such socio-economic data take particular account of the land-ownership situation and strategic planning and redevelopment problems. The untimbered areas in Guiana correspond to spontaneous land-take for agriculture often liable to change rapidly and shift geographically, whereas the pioneer fronts in Brazil mark rather government intentions to enhance agricultural potential, expressed as a plan which indeed prompts the clearances and permanent occupation for pastoral purposes.

Examination of images recorded at different dates has allowed an assessment of the relative proportions of spontaneously cleared areas and the pioneer fronts driving the processes of deforestation attributable to small-scale clearances advanced at a rate of 0.2% per year in the Saint-Georges de l’Oyapock region between 1958 and 1988. Superposing aerial photographs taken at 40-year intervals shows ephemeral deforestation in certain places, and that fallowing had led to forest recolonization in some previously-cleared plots. However, the expansion of pioneer fronts in Brazil advanced at a rate of 1.2% per year between 1988 and 1998 on the site studied in the State of Paca (3), this time without apparent reforestation processes, the clearances appearing to be permanent.

Surveillance of the Amazon forest environment is therefore made possible by combined comparative analysis of images obtained using several different data acquisition systems (Landsat, Spot, radar, aerial photographs) and recorded at different dates. Because it takes into account the different spatial scales of deforestation in Amazonia, this remote-sensing approach is especially suitable for monitoring changes in the forest situation and the impact of human activity on this environment. Clearance for agriculture, spontaneous marginal urbanization, ORPAIILAGE and illegal forestry can be watched. The approach consequently promises to be valuable as an instrument for land-use assessment in French Guiana and could be an additional aid to forest management in the Amazon for regional cooperation schemes.
 
Notes for editor

(1) After estimates made by the Brazilian National Institute of Space Studies for1998.

(2) As part of a doctoral thesis by Moïse Tsayem Demaze supported by the CNES (Centre National d’Etudes Spatiales, the French National Space Studies Centre) and the IRD and supervised by Jean-Marie Fotsing and Laurent Polidori.

(3) From calculations performed on Landsat™ images

Contact Moïse Tsayem Demaze et Laurent Polidori, LRT, unité ESPACE, IRD Cayenne, BP 165-97323 Cayenne Cedex – Tél. 05 94 29 92 81, Fax. 05 94 31 98 55, courriel : mtsayem@nomade.fr et polidori@cayenne.ird.fr
Jean Marie Fotsing, Pôle Géomatique, IRD Orléans, 5, rue du Carbone, 45072 Orléans Cedex 2 – tél. 33 (0)2 38 49 95 16, Fax 33 (0)2 38 49 95 47, courriel : Jean-Marie.Fotsing@orleans.ird.fr

Contacts IRD Communication : Marie Guillaume (rédactrice), tél. : 01 48 03 76 07, courriel : guillaum@paris.ird.fr ; Bénédicte Robert (relations presse), tél. : 01 48 03 75 19, courriel :presse@paris.ird.fr

 

Peer reviewed publication and references

Tsayem Demaze M. – Caractérisation et suivi de la déforestation en milieu tropical par télédétection : application aux défrichements agricoles en Guyane française et au Brésil. Thèse de Doctorat de l’Université d’Orléans, 242 pages, 2002.

Tsayem Demaze M., Fotsing J.M., Huynh F. – La déforestation dans la région de Saint-Georges de l’Oyapock (Guyane française). In Cahiers d’Outre Mer, 55 (218), mars-juin 2002, p.p. 197-222.

Tsayem Demaze M., Polidori L., Fotsing J.M. – Caractérisation multi-échelle et multi-capteur de la déforestation tropicale amazonienne. In Bulletin de la Société Française de Photogrammétrie et Télédétection, n° 161, pp. 74-84, 2001.