Análise de séries temporais de focos de calor nos biomas brasileiros utilizando gráfico de recorrência

• Autor(a) principal: BARROS, Vaniele da Silva
• Data de Publicação: 2021
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da UFRPE
• Texto Completo: http://www.tede2.ufrpe.br:8080/tede2/handle/tede2/8768
• topic: Séries temporais; Bioma; Focos de calor; Gráfico de recorrência; CIENCIAS EXATAS E DA TERRA::PROBABILIDADE E ESTATISTICA

The scientific interest in studies that use recurrence analysis to approach the transitions between regular and chaotic behaviors, as well as, in the identification of structures of dynamic systems, has spread over the years. Among the main tools of this analysis, we highlight the Recurrence Graph method and the Recurrence Quantification Analysis, which are widely used in the analysis of time series supposedly coming from non-linear and even non-stationary dynamic systems. In particular, this work analyzed or evaluated the large and small scale patterns in the Recurrence Graphs of the series of hot pixels in the Amazon, Cerrado, Caatinga and Atlantic Forest biomes and to obtain the quantitative measures by the method of Recurrence Quantification Analysis. Daily series of hot pixels derived from data provided by National Institute of Space Research – INPE, of the biomes were analyzed for the period from July 4, 2002 to December 31, 2019. In Brazil, the annual average of number of hot pixels, between 2002 and 2019, is approximately 241,866 detections, being these most frequent events between the months of July to October. Considering the absolute values referring to the number of hot pixels in each biome, the highest concentration occurs in the Amazon biome, as it has the largest territorial extension, however, considering the number of hot pixels and the area of each biome, the Cerrado has the highest record per 𝑘𝑚2. The structures present in the Recurrence Graphs of the daily series of hot pixels of the biomes indicate low predictability, while for the series of anomalies, they indicate high predictability, in addition to presenting abrupt changes in the dynamics of the systems in both cases. The values of the various indices that serve as measures of process quantification confirm these results, were obtained through the application of the Recurrence Quantification Analysis method.