Oliveira, Ronaldo Pereira de 
  
  
However, there is no proper information on the book data reliability in terms of cartographic precision. Even more, 
there are minor islands inside the Guanabara Bay and other outer water bodies areas not include in the digital data set 
imported from SMAC. In the project framework, this difference was not considered of great importance, because the 
completeness of the continental cover. 
Land suitability results included 40.384,9 ha of suitable areas for reforestation, being 33,1% of the total district area and 
66,5% of the soil survey covering. From this result, the areas exclusively located at hilly slopes could be calculated as 
35.333,8 ha, 28,996 of the total area and 96,6% of the total suitable for reforestation. Suitable areas for irrigated 
horticulture have an extension of 16.679,5 ha, 13,7% of the total district. 
Primary criteria for land vulnerability, the sub-division of non-built areas into highlands and low lands resulted 
respectively in 36,850.0 ha and 23,885.3 ha. The areas considered of highest vulnerability for highlands (very high and 
extremely high vulnerability classes) are locate in the massifs of Tijuca, Pedra Branca, and Gericiné besides some 
isolated hills and mountains. Areas of lower vulnerability class for low lands are located at depositional slopes from 
Sepetiba, Pavuna, Jacarepaguá and Guanabara Bay. Highest vulnerability of low lands are presented close to 
Jacarepaguá, Tijuca, Camorim, and Marapendi lagoons; mainly close to dunes, mangroves, and sandbanks of 
Marambaia. Figure 3 gives an overall idea of land vulnerability class distribution for each one of the two main 
landscapes, as erosional environment of mountainous areas and depositional environment of piedmont areas. 
  
3% 4% 6% 
   
  
     
Low 
E] Intermediate Intermediate 
E] High E High 
E] Very High m Very High 
E] Extremelly High m Extremelly high 
Low Lands Highlands 
  
  
  
Figure 3. Percent distribution of vulnerability classes for depositional and erosional 
environments respectively low lands and highlands. 
Areas with a better land environmental quality for highlands are those which are still covered with native Atlantic 
Forest species, mainly located in the massifs of Tijuca, Pedra, Branca, and Gericin6 with 14,9% of the total district area. 
On the other hand, the worse environmental quality areas are close by mining activities, and mass sliding steps, which 
are in general spread all over the district at small sites representing 0,3%. At low lands, the best land environmental 
quality is the Marambaia sandbank, representing 0,4% of the total district area that is in a very good preservation state. 
In contrast, low lands have the worse environmental quality conditions at sand and sand stone mining that are mainly 
located in the west zone with total expansion of 0,3% of the total area of the municipal district. Mining activities also 
comprehends 1,7% of the total area, if considered exclusively at highlands. 
3.2 Integrated Geoinformation Model 
Although all the information and products that were generated in partnership with SMAC (Embrapa, 1999a; Embrapa, 
1999b), the use of GIS procedures was static and limited to exchange hand made steps by automated functions. 
Therefore, the ability to use spatial analysis as integrated part of a process driven study, providing means of scenario 
generation to improve environmental elements recognition, was left behind. The specific results from the proposed 
modelling and procedural evaluation activities were related to surveyor's integration on systems approach methodology. 
The involvement of few researchers has clarified to system analysts certain details from data structure modelling 
constructs, as polymorphism and typing. On the order hand, researchers have been aware of the existence methods and 
software technology that can help them to abstract terrain object classes as thematic categorisation of biophysical actors 
under a process driven analysis and design models (see Figure 4). 
Quantitative results form modelled processes were obtained by means of overlay procedures of previous generated land 
vulnerability and soil maps with the new slope map, which was calculated for each pixel from the generated DTM. 
Procedures were focus on land vulnerability information, because its evaluation has considered geomorphology as main 
criteria to define environmental limiting factors. Another procedure purposed as means of process optimisation was an 
image processing function, trying to refine previous executed supervised classification with detailed vegetation 
coverage and mining information from watershed level soil survey. Primary comparison of final area calculation results 
has shown that for highlands the DTM procedures can help with a more precise distinction between intermediate, high, 
and very high (M, A, and MA) land vulnerability classes, having up to 4096 from previous A class total area as 
  
  
  
  
  
  
1078 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000.