Comparação de dois modelos empíricos para a classificação do grau de suscetibilidade a abatimento em áreas cársticas carbonáticas na Bacia do Rio Vieira, município de Montes Claros – MG

• Ano de defesa: 2023
• Autor(a) principal: Natália Dias Lopes
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais
• Palavras-chave em Português: Modelagem de dados – Aspectos ambientais; Rochas carbonáticas; Carste – Vieira, Rio, Bacia (MG)
• Link de acesso: https://hdl.handle.net/1843/58465

Slump processes are natural events that are part of the relief formation, giving rise to features such as sinkholes and uvalas. However, when they develop in occupied areas, whether due to urbanization, agriculture, livestock or other economic activities, it can cause damage to life and property. In Brazil, several municipalities have their territory on carbonate rocks and, in some of these, damage caused by abatement processes has already occurred. In this context, it is seen as important the knowledge and management of areas susceptible to subsidence linked to the complex karst environment. Trying to understand the natural aspects that corroborate the occurrence of slumps, this work carried out a bibliographical review in order to present the histories of the main events, the carbonate karst areas of Brazil, the types of inventories related to risk and susceptibility assessment maps, in addition to pointing out the main themes and natural characteristics addressed in models that use a multicriteria analysis technique to generate products that help manage the use and occupation of the karst territory in the face of this process. Knowledge of the various natural aspects that contribute to the classification of susceptibility to abatement for the study area was gathered and analyzed using two types of empirical models that use multicriteria analysis, the knowledge-based model and the data-based model. The first model uses the knowledge of specialists and applies Monte Carlo and Fuzzy Logic simulation techniques, resulting in a map that points out the different classes of susceptibility and the reliability of this information pixel by pixel. The second model was based on data, being elaborated through the application of machine learning. As a result, two susceptibility maps are presented that qualify the potential for natural abatement processes in the Rio Vieira basin in the municipality of Montes Claros. Despite the two models using the same databases, there was little similarity between them, mainly regarding the places with medium susceptibility class, with the highest agreement found in the low classes.