Modelagem da dinâmica do desmatamento de uma fronteira em expansão, Lábrea, Amazonas

Abstract

Brazil’s “arc of deforestation” continues to expand across the Amazon region and has already reached the southeastern part of the state of Amazonas. A new focus of the deforestation has already affected a part of Lábrea, which is the municipality (county) that was recently found to have one of the highest deforestation rates in Amazonas. Pressure from the Arc originates in the neighboring states of Acre and Rondônia, which already have intense deforestation processes underway for the expansion and consolidation of the agricultural and ranching frontier. In addition to the impacts caused to natural ecosystems in Lábrea, land-use changes have induced a series of social conflicts through the process of “grilagem” (fraudulent appropriation of large areas of public land) and from agricultural and ranching activities that have contributed to the expulsion of extractive workers such as rubber tappers and Brazil nut gatherers. Consequently, traditional families demanded the creation of extractive reserves to protect themselves. In addition, as a part of the Program for the Acceleration of Growth (PAC) the Ministry of Transportation plans to reconstruct the BR-319 (Porto Velho- Manaus) Highway (which has been abandoned since 1988) and to recuperate a part of the marginally passable Transamazon Higway (BR-230) that connects the BR-319 to Lábrea. To avoid the environmental consequences of these projects, in 2006 the government proposed a series of protected areas in the area of influence of the BR-319, four of which have been recently created in the municipality of Lábrea. This study had the objective of modeling the future spatial dynamics of deforestation in Lábrea and evaluating the usefulness of the recently created protected areas in containing deforestation. Chapter I analyzes the effectiveness of protected areas that have already been created in Acre, Rondônia and southern Amazonas. The analyses revealed the usefulness of protected areas in containing deforestation: 90% of the protected areas had deforestation rates inferior to those in a 10-km-wide strip surrounding the protected areas. In addition, weights of evidence that represent the probabilities used in the AGROECO model to simulate future deforestation inside the proposed areas have been determined for the protected areas in this region, as well for the areas surrounding the protected areas (10-km buffer). These weights of evidence have been determined specifically for the category of use of each protected area and for the category of use in accord with its administrative level (federal or state). Indigenous reserves were the most effective in repelling deforestation, with a weight of -2.57; Integral Protected Areas were less efficient with a weight of -1.23, and, finally, Sustainable Use Protected Areas were the most vulnerable with a weight of -0.15. When considering weights of evidence as a function of the internal Euclidian distance, weights declined when the internal distance was progressing from the edge of the protected area to the center. Protected areas had weights of evidence that varied with the internal Euclidian distance. These weights have been used in Chapter II, where eight scenarios have been produced for the Lábrea region up to 2040 using the AGROECO model. Of the two groups of scenarios, one didn’t consider the recent creation of protected areas (I), Business As Usual while the other considered these areas, (II) Governance. In each scenario group, four study cases have been considered including: 1- homogenously distributed weights of evidence over the protected areas, 2- gradually distributed weights of evidence according to the internal Euclidian distance, 3- homogenously distributed weights of evidence over the protected areas and weights of evidence for the 10-km buffer areas, 4- gradually distributed weights of evidence according to the internal Euclidian distance and weights of evidence for the 10-km buffer areas. Creation of the protected areas has been little efficient because the majority of the Governance group scenarios had slightly less deforested area in 2040 than did the corresponding scenarios in the Business As Usual group. The use of gradually distributed weights of evidence according to the internal Euclidian distance appears to be the most realistic approach because it reduces the influence of deforestation occurrence in areas far from the protected area limits. Considering the weights of evidence for the 10-km buffer area, the surroundings of the protected areas have concentrated deforestation inside their limits because their high weights of evidence make them more attractive to deforestation in the simulations. However, this is believed to be a consequence of the years for which satellite data were available for calibrating the model, when the areas outside of the reserve buffers had already been heavily deforested, leaving little left to clear. The most realistic scenarios for our dataset are therefore considered to be those that use gradually distributed weights of evidence but do not use separate weights for the buffer areas. In the scenario (GOV-2), deforestation was reduced in the study area by 5,1 % (2.596 km 2 ) as a result of the creation of the reserves, when compared to the corresponding baseline scenario (BAU-2).