Análise de fenômenos ambientais utilizando o grafo de visibilidade horizontal

• Ano de defesa: 2020
• Autor(a) principal: SILVA, Joelma Mayara da
• Orientador(a): STOSIC, Borko
• Banca de defesa: LALIC, Milan
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal Rural de Pernambuco
• Programa de Pós-Graduação: Programa de Pós-Graduação em Biometria e Estatística Aplicada
• Departamento: Departamento de Estatística e Informática
• País: Brasil
• Palavras-chave em Português: Grafo de visibilidade horizontal; Séries temporais; Queimada; Hidrologia; Bioma
• Área do conhecimento CNPq: CIENCIAS EXATAS E DA TERRA::PROBABILIDADE E ESTATISTICA
• Link de acesso: http://www.tede2.ufrpe.br:8080/tede2/handle/tede2/8761

Understanding the dynamic behavior of complex systems time series is a fundamental problem in several areas and is treated by different techniques. Recently, methods based on complex networks have been shown promising in temporal series analysis. One of the most used method is the Visibility graph that has two forms, Natural visibility graph and horizontal visibility graph that differentiate according to the visibility criterion. In this work we use horizontal visibility graph (HVG) to transform time series of heat spots (that are good proxy for vegetation fires) detected in Brazilian biomes and São Francisco River flow, to assess whether topological measures of HVG networks detect changes caused by human and natural factors. Vegetation fires can be beneficial or degrading for the ecosystem. Among the six Brazilian biomes, the most sensitive to fires are the tropical forests of the Atlantic Rain Forest and Amazon biomes, due to their high biodiversity and possibility of occurrence of changes in the ecosystem.When evaluating the hot pixels series, it was found that although they have different environmental characteristics, the dynamics of biomass burning in different biomes is very similar. Particularly for the Amazon, there is a big difference between the fire dynamics during the year, where in the first semester, during the rainy season, the hot pixel series are more chaotic, that is, less predictable. The temporal evolution of the clustering coefficient of hot pixels series detected in the Amazon highlighted years with the occurrence of drought, a phenomenon that makes the biome more susceptible to fires. Another complex environmental system is the hydrological cycle, whose main element is the hydrographic basin. In Brazil, basins and sub-basins are grouped in 12 hydrographic regions, among which one of the most important the São Francisco River basin, whose water is mainly used for irrigation, but also contributes to the country's energy sector through hydroelectric plants, among which are Sobradinho (in the state of Bahia) and Xingó (in Alagoas), that started operating in 1979 and 1994, respectively. When analyzing the flow at Juazeiro station that was affected by Sobradinho dam, it was observed that the characteristics of the HVG network change, after the construction. Similar result was obtained for Pão de Açúcar station in which case the flow regime was changed twice, first after the construction of Sobradinho dam, and later after the construction of Xingó dam. However, the HGV based temporal analysis revealed that the downstream flow was regularized by Sobradinho dam.