Close Window

Amazonia is one of the most important ecosystems of the planet, containing the largest extent of contiguous tropical rain forest on earth, over 5 million square kilometers. While most of the region remains forested, rapid development has led, over the past two decades, to the destruction of over 589,000 km(2) of forests in Brazil alone. Forest clearing can alter the transport of sediments, organic matter and associated nutrients to the rivers. In this article, we present the results of an integrated analysis of the landscape characteristics, including soil properties, river network, topography, and land use/cover of a tropical meso-scale river. This physical template was developed as a comprehensive tool, based on Remote Sensing and GIS, to support the understanding of the biogeochemistry of surface waters of the Ji-Parana river basin, State of Rondonia, Western Amazonia. Our primary objective was to demonstrate how this tool can help the understanding of complex environmental questions, such as the effects of land-use changes in the biogeochemistry of riverine systems. River sites and basin characteristics were calculated using the data sets compiled as layers in Arc-Info GIS. A land-use/cover map for 1999 was produced from a digital classification of Landsat 7-ETM+ images. To test the effects of the landscape characteristics on river water chemistry, we performed a multiple linear regression analysis. Average slope, river network density, effective cation exchange capacity (ECEC), and proportion of pasture were treated as independent variables. River water electrical conductivity (EC) and Na+, Ca2+, Mg2+, K+, Cl- and PO43- concentrations were the dependent variables. Spatially, higher values of all ions were associated with areas dominated by pasture, with the highest concentrations found in the central part of the basin, where pasture areas are at a maximum. As the river enters the lower reaches, forests dominate the landscape, and the concentrations drop. The percentage of the basin area covered by pasture was consistently the best predictor of EC (r(2) = 0.872), PO43- (r(2) = 0.794), Na+ (r(2) =0.754), Cl- (r(2) = 0.692) and K+ (r(2) = 0.626). For Ca2+, both ECEC (r(2) = 0.538) and pasture (r(2) = 0.502) explained most of the observed variability. The same pattern was found for Mg2+ (r(2) = 0.498 and 0.502, respectively). (C) 2003 Elsevier Inc. All rights reserved

Close Window