Câmara et al. challenge our assertion that the unprecedented, planned expansion of highways and other transportation projects in Amazonia that was originally proposed under the “Avança Brasil” (Advance Brazil) program is likely to lead to a dramatic increase in forest loss and degradation and they argue that our earlier spatial models (1) were overly simplistic and “apocalyptic” in their projections. Three points about our models merit emphasis.

First, the projections of our models—that 28 to 42% of Brazilian Amazonia would be deforested by 2020 if all the Avança Brasil projects proceed immediately—are in fact very plausible and do not differ greatly from simple extrapolations using the current high rate of forest loss (2). Second, our models incorporated key components of regional heterogeneity in Amazonia, including spatial variability in forest vulnerability to fire, logging, and mining. Third, independently derived scenarios of future forest loss (3, 4), including a recent model that incorporates much of the region’s biophysical and economic heterogeneity (5), also indicate that new and planned highways are likely to play a central role in determining future patterns of Amazon deforestation.

If a new highway penetrates into a large forest tract and promotes spontaneous colonization by farmers, loggers, and ranchers, is the forest loss caused by the highway or the other drivers? Clearly, it is both—but the crucial point is that such transportation projects play a pivotal role in determining where forest destruction occurs. The truly alarming aspect of the Avança Brasil program is that it will criss-cross the Amazon with some 7500 km of paved highways and many other transportation projects that will penetrate deep into the heart of the basin. The net effect will be not only be increased deforestation but also fragmentation of forests on an unprecedented spatial scale (1). Rather than concentrating development in the vast expanses of land that have already been deforested, the projects that promote frontier expansion will do precisely the opposite.

Contrary to the claim by Câmara et al., the dramatic upsurge in Amazonian deforestation in 2002–03 includes many areas associated with highways and roads, including the notorious Santarém-Cuiabá Highway. Even the deforestation hot spot (São Félix do Xingu) emphasized by Câmara et al. is closely associated with privately financed roads (6). The point of our recent Letter is that Brazilian-government efforts to slow rampant Amazon deforestation are unlikely to succeed if the government proceeds with its most environmentally damaging transportation projects. We stand by this assessment.


   1. W. F. Laurance et al., Science 291, 438 (2001).

   2. If the deforestation rate in 2002–03 (2.35 million ha year-1) continues indefinitely, then 100 million hectares of forest will have disappeared by the year 2020. This is about 25% of the original forests in Brazilian Amazonia.

   3. G. Carvalho, A. C. Barros, P. Moutinho, D. C. Nepstad, Nature 409, 131 (2001).

   4. B. Soares-Filho et al., Global Change Biol. 10, 745 (2004).

   5. B. Soares-Filho et al., “A spatially explicit simulation model of deforestation for the Amazon Basin” (3rd Scientific Conference of the LBA Program, Brasília, Brazil, 2004).

   6. A. Alencar et al., Desmatamento na Amazônia: Indo Além da “Emergência Crônica” (Instituto de Pesquisa Ambiental da Amazônia, Belém, Brazil, 2004).