OX, V2 ed ded AN 
Batistella, Mateus 
  
Rondönia is perhaps the state where processes of colonization in the region have attracted the greatest amount of 
attention, mainly because of LULC change associated with land appropriation. Development and colonization projects 
have marked the history of the state in the last 30 years (Fearnside, 1989). Two large projects took place during this 
time period: Polonoroeste, responsible for paving the main road crossing the state and providing access in the southeast- 
northwest direction; and Planafloro, which financed, among other things, the elaboration of a land zoning program. 
Negative and positive interpretations concerning the outcome of these projects were made but all agree that the state is 
facing a period of rapid change (Brown, 1998). 
Some studies have used specific metrics to understand landscape structure and fragmentation in the Amazon. There are 
metrics related to landscape composition, referring to features associated with the presence and amount of each patch 
type within the landscape but without being spatially explicit. Others are related to landscape configuration, referring to 
the physical distribution or spatial character of patches within the landscape (Baker and Cai, 1992; McGarigal and 
Marks, 1995). 
Dale et al. (1993) used a dynamic model consisting of three linked sub-models that simulate settlement patterns, land- 
use changes, and carbon release. The model incorporates both household- and regional-level analysis, showing that 
land-use changes in a fishbone colonization pattern in Central Rondónia are a function of variables such as lot size, 
land-use history, initial soil and vegetation conditions, land-use choices, and effects of illness on work. Dale ef al. 
(1994) used fractal dimension and contagion indices to quantify the effects of different land-management practices. 
Results indicated the decrease in complexity after 40 years of occupation for the typical- and worst-case scenarios. 
These scenarios were also used to address the effects of fragmentation in nine groups of animals in terms of gap- 
crossing ability and land area requirements. Species with large area requirements but with small gap-crossing ability are 
more affected (Dale ef al., 1994). Frohn et al. (1996) used satellite remote sensing analysis to evaluate the model 
proposed by Dale et al. (1993, 1994). Percent cleared, contagion, and fractal dimension were also used for comparisons 
of land cover patterns. The authors showed how changes in resolution affect landscape indices and how simulation of 
agricultural clearings at the lot level in Central Rondónia captures the majority of variance in the spatial and temporal 
patterns of deforestation. 
Examining further the quantification of habitat fragmentation, Dale and Pearson (1997) suggested a combination of 
spatial indices to depict the complicated process of LULC change in Rondónia. Indices such as area of habitat, 
frequency distribution of patch sizes, measures of patch shape, and length of edge between habitat types were of 
particular interest. They concluded that the percentage of area in a particular land cover type remains appropriate for 
land use change analysis. Furthermore, the ratio of the largest forest patch to the total forest area may be the most 
appropriate index for characterizing extreme conditions. 
Meanwhile, despite such important findings, few works have addressed the influence of different architectures of 
colonization on landscape fragmentation. Soares Filho (1998) developed a Markovian model to analyze landscape 
dynamics in three colonization patterns in Mato Grosso. The analysis was based on regional data simulating LULC 
change in a multi-temporal basis. Oliveira-Filho and Metzger (1999) looked for thresholds in landscape indices used to 
describe three different patterns of deforestation in the same state. Other authors have integrated distinct levels of 
analysis, from the household to the landscape to identify the human dimensions of LULC change (McCracken et al., 
1999; Moran and Brondizio, 1998; Moran et al., 1998). This paper is placed in this context, focusing on two distinct 
settlement designs in the state of Rondónia. 
3. ENVIRONMENTAL AND CULTURAL SETTING 
Within the Amazon Basin, particular areas offer a great potential to the development and testing of methods of spatial 
analysis related to forest fragmentation. One of these areas is located in the State of Rondónia (Figure 1), where 
deforestation rates have been the highest in the Amazon during the last twenty years (Alves, 1999). Following the 
national strategy of regional occupation and development, colonization projects initiated by the Brazilian government in 
the 1970's played a major role in this process (Pedlowski, 1999). 
In Rondónia, colonization projects were designed based on a fishbone scheme. Often, these projects were implemented 
with no consideration for environmental constraints and landscape characteristics of each region. Vale do Anari, 
established in the early 80's, is an example of this process. In 1982, the Brazilian Colonization and Land Reform 
Agency (INCRA) designed a new model of settlement, in which the road network followed the watershed topography 
(Figure 2). Within this project, called Machadinho d'Oeste, the lots almost always had access to a water stream. Studies 
of farming systems and socioeconomic characteristics at Machadinho d'Oeste have suggested this settlement to be a 
more adequate model of colonization design (Miranda et al., 1997). 
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000. 149