Identificação e análise das unidades de paisagem do Parque Estadual da Serra do Rola-Moça e Estação Ecológica de Fechos, Quadrilátero Ferrífero - Minas Gerais

• Ano de defesa: 2016
• Autor(a) principal: Luiza Hoehne Mattos de Oliveira
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Parque Estadual da Serra do Rola-Moça; Estação Ecológica de Fechos; Unidades de Paisagem; Mineração de Dados; Modelagem Ambiental; Modelagem de dados Aspectos ambientais; Estação Ecológica de Fechos (MG); Áreas de conservação de recursos naturais; Paisagens; Parque Estadual da Serra do Rola-Moça (MG)
• Link de acesso: http://hdl.handle.net/1843/IGCM-AL6P6H

The Iron Quadrangle (QF), territory located in the central-southeastern portion of Minas Gerais, Brazil, is characterized historically by the extraction of mineral resources. This region stands out for presenting a unique biodiversity, scenic beauty and a high degree of endemism, which sets as a priority area for conservation. Within the limits of QF, it may be found a large number of protected areas, categorized as Conservation Units (UC) of Integral Protection and Sustainable Use. Among these, it was used as geographic scope of this work the Serra do Rola-Moça State Park (PESRM) and the Fechos Ecological Station (EEF), Integral Protection Conservation Units located in the metropolitan region of Belo Horizonte, capital of Minas Gerais. The PESRM and EEF are characterized by a particular biodiversity, supported by a peculiar geodiversity, by a high degree of flora and fauna endemism and house six major watersheds. This work aimed to establish a subdivision of the area in Landscape Units with conceptual basis of systemic analysis of the landscape and methodology in computational algorithms to identify patterns. Therefore, the following elements were considered: geology (CPRM), lithology (CPRM), hypsometry (SRTM), relief (IBGE), slope (SRTM), hydrology density (SRTM), vegetation cover and land use, pedology (CPRM) and zoning. The integration of these data was processed through the combination of their respective layers in ArcGIS 10.1.2 software, generating 2260 combinations of information. These were subjected to a data mining process with classification based on the decision tree, in the Weka Explorer software, using the J48 algorithm. The classification presented total accuracy rate of 94.69% and almost perfect performance, according to the Kappa index. The combinations generated subdivisions whose boundaries were refined due to the contiguous areas, allowing the identification of four landscape units in PESRM and EEF, which have been named and characterized. It is understood that the subdivision proposal in UPs details the subdivision Relief Units adopted in PESRM Management Plan and can support more adequately the policy planning and configuration of strategies for the conservation and preservation of those protected areas.