Batistella, Mateus 
  
Landscape indices were calculated using Fragstats (McGarigal and Marks, 1995). Index selection was based on their 
usefulness to address landscape fragmentation (Haines-Young and Chopping, 1996; Dale and Person, 1997). Some 
indices are particularly useful to test the hypothesis about landscape fragmentation processes in Machadinho d'Oeste 
and Vale do Anari. Among them, area indices and number of patches are of special interest. Edge, shape, and core area 
indices were calculated because of their relations with architecture design and landscape configuration. Mean nearest- 
neighbor index defines the average distance to the nearest patch of same patch type. Interspersion/juxtaposition indices 
may be useful to compare landscape configurations. 
5. RESULTS AND DISCUSSION 
Landscape fragmentation analysis in Machadinho d'Oeste and Vale do Anari was based on four LULC classes: forest, 
non-forest vegetation, bareland/cropland, and water (Figure 3). Table 1 shows selected landscape indices for the 
settlements in terms of LULC classes. Results for forest are particularly highlighted. The total area and percentage of 
landscape of forest is similar in both colonization projects. However, the number of patches and the mean patch size are 
significantly different. In Machadinho, the mean patch size of forest fragments is 106.59 ha (n=870), while in Anari the 
mean patch size is 76.94 ha (n=1165). The largest patch index does not depict these differences. Although there are 
large patches of forest in both settlements, their number and size differ among the landscapes. 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
Machadinho d’Oeste Vale do Anari 
Indices Forest |Non-forest| Bareland/| Water | Forest |Non-forest| Bareland/| Water 
Vegetation| Cropland Vegetation| Cropland 
Area (ha) 92729.52| 13311.54 | 18775.17 | 674.19 | 89638.2 | 15690.06 | 20063.88 | 98.28 
Percentage of landscape (%) | 73.89 10.61 14.96 0.54 71.43 12.5 15.99 0.08 
Largest patch index (%) 72.56 0.24 2.32 0.04 69.5 0.16 1.42 0.02 
Number of patches 870 3623 2132 111 1165 3943 2761 48 
Patch density (#/100ha) 0.69 2.89 1.7 0.09 0.93 3.14 2.2 0.04 
Mean patch size (ha) 106.59 3.67 8.81 6.07 76.94 3.98 7.27 2.05 
Total Edge (m) 3548130 | 3940050 | 3292800 |175020| 3905970 | 4472490 | 3692430 | 30030 
Edge density (m/ha) 28.27 31.4 26.24 1.39 31.13 35.64 29.42 0.24 
Shape index 26.05 28.81 24.25 2.24 28.57 32.57 27.07 1.22 
Total core area (ha) 77808.33| 1828.53 | 8343.09 | 53.64 |73264.41| 2233.17 | 8272.17 8.19 
Number of core areas 439 1732 1109 129 635 1999 1236 12 
Core area density (#/100ha) 0.35 1.38 0.88 0.1 0.51 1.59 0.98 0.01 
Mean core area (ha) 89.43 0.5 3.91 0.48 62.89 0.57 3 0.17 
Mean nearest neighbor 65.03 73.28 87.49 668.3 | 66.33 68.86 83.93 1940.78 
distance (m) 
Interspersion and 75.85 63.59 62.57 17.93 | 64.25 63.41 62.59 59.31 
juxtaposition index (%) 
  
Table 1 - Class indices for Machadinho d'Oeste and Vale do Anari in Rondónia, Brazilian Amazon. 
  
  
Certainly, settlements' configuration affects these results, based on the protection of many large patches of forest in 
Machadinho (forest reserves). Core area indices follow the same trend: total core area is similar but mean core area is 
89.43 (n=439) for Machadinho and 62.89 (n=635) for Anari. Higher patch density and core area density for forests also 
suggest a higher fragmentation in Anari and a better conservation of forest remnants in Machadinho. It is important to 
consider that forest fragmentation in both settlements, as indexed by area and core area metrics, may affect the 
propagation of disturbances across the landscapes. 
This discussion brings up the important role of landscape and patch type edges. The total edge and the edge density 
have higher values in Anari because edges are longer for all classes, except for water. Depending on the LULC class 
adjacent to the forest patches, different effects may be observed in terms of ecological processes. For example, if the 
forest patch borders bareland/cropland, this fragment may become more susceptible to disturbances in its structure and 
composition. Conversely, if the forest is adjacent to a fallow area (non-forest vegetation), secondary succession may be 
accelerated. Studies based on field data collection and further analysis are being carried out to address these 
relationships in both colonization areas. 
  
152 International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000.