Deforestation and Human Land-Use Dynamics in the Brazilian Amazon:
A Case Study Using Remote Sensing and GIS

Carolina Balzas
Bachelor's of Science in Environmental Science
May 2001

Since the 1960s, deforestation in the Brazilian Amazon has set in motion a series of inter-related land-use patterns. Using remote sensing and Geographic Information Systems (GIS), I assessed land-cover conversions and related land-use dynamics in the old frontier town of Paragominas, an area in the eastern Amazon, in the state of Para, Brazil. I used five satellite images (Landsat TM 1984, 1988, 1991, 1996 and SPOT 1999) to create a composite, multi-temporal image. Applying an unsupervised classification, I derived a series of estimates relating to the total area (ha) of different land-cover types, the main land-cover conversions and rates of land-cover conversion. I captured four main land-cover conversion patterns. These patterns include transitions from forest to pasture, pasture to secondary growth, logged/burned forest to recovering forest and pasture/secondary growth to plantation.

Using GIS, I captured the transition from logged forest to pasture. My results indicate that from 1984 to 1999 there was a 20% decrease in the forest cover (from 254,000 hectares to 184,000 ha), a 6% increase in the amount of pasture (87,000 hectares to 107,000) and an 8% increase in the amount of secondary growth (10,000 hectares to 38,000 hectares). GIS captured logging patterns more accurately than the unsupervised classification. Specifically, GIS estimates indicate that from 1984-1999 there was a 33% increase in logging (as opposed to a 6% increase derived from the unsupervised classification). I captured two separate stages of secondary growth-one that is 6-10 years old (as of 1999), another that is 16-20 years old (as of 1999). The rate of forest-to-pasture conversion peaked between 1991 and 1996. From 1984 to 1999, the rate of secondary growth increased steadily. Because this study suggests that that Paragominas is shifting towards an intensification of the human-cleared landscape, the amount of secondary growth forest may decrease as these areas are readily being converted to annual row crop agriculture. All primary data that I provide is derived from my remote sensing and GIS analyses.