Identificação de preditores para estimativa de produtividade da cana-de-açúcar, no leste de Pernambuco, utilizando análise de correlação canônica

• Ano de defesa: 2023
• Autor(a) principal: SILVA, Rodrigo Rogério da
• Orientador(a): MOURA, Geber Barbosa de Albuquerque
• Banca de defesa: NASCIMENTO, Cristina Rodrigues, GIONGO, Pedro Rogério
• 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 Engenharia Agrícola
• Departamento: Departamento de Engenharia Agrícola
• País: Brasil
• Palavras-chave em Português: Cana-de-açúcar; Produtividade; Correlação canônica; Variáveis climáticas; Temperatura; Ventos alísios; Pressão atmosférica
• Área do conhecimento CNPq: CIENCIAS AGRARIAS::ENGENHARIA AGRICOLA
• Link de acesso: http://www.tede2.ufrpe.br:8080/tede2/handle/tede2/9117

Brazil is the world's largest producer and exporter of sugar, accounting for approximately 39% of world exports in the 2019/20 harvest (FAO, 2020). Climatic factors have a great influence on sugarcane production, especially precipitation, which directly and significantly impacts agricultural yield values. The estimation of agricultural productivity of sugarcane through climate data in advance helps in agricultural planning and subsidizes decisions in the field and subsidizes public policies. Faced with the need to diversify methods for predictions, the objective of this dissertation was to find the best predictive variables through canonical correlation analysis in the trade winds, SST (Sea Surface Temperature), surface atmospheric pressure in the Equatorial Pacific Ocean and SST in the Tropical Atlantic (Dipole area), for the elaboration of sugarcane productivity forecast models, up to three months in advance. The study area comprised 58 municipalities in Pernambuco, located in the eastern region of the state, in a strip along the coast, in the Pernambuco forest region, and in transition zones between the forest and the wild. The hierarchical cluster analysis, represented by the dendrogram, produced three homogeneous groups of sugarcane productivity. It is assumed that the first canonical function approximates the results of multiple regression and the independent statistical variable represents the set of variables that best predicts the three dependent measures, mainly the set of dependent variables of the second productivity group (GruPro2). It was possible to notice that the best predictors are SST in the South Atlantic and North Atlantic, atmospheric pressure in Tahiti, the wind fields in the central Pacific and the SST fields in the El Niño 3 areas.