Análise Espaço-Temporal das variáveis meteorológicas associadas a culturas agrícolas em mesorregiões do estado do Paraná

• Ano de defesa: 2023
• Autor(a) principal: Gamero, Paulo
• Orientador(a): Opazo, Miguel Angel Uribe
• Banca de defesa: Opazo, Miguel Angel Uribe , Dalposso, Gustavo Henrique , Cima, Elizabeth Giron , Oliveira, Márcio Paulo de , Guedes, Luciana Pagliosa Carvalho
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual do Oeste do Paraná; Cascavel
• Programa de Pós-Graduação: Programa de Pós-Graduação em Engenharia Agrícola
• Departamento: Centro de Ciências Exatas e Tecnológicas
• País: Brasil
• Palavras-chave em Português: Cluster; Índice de Moran; Geoestatística; MULTISPATI-PCA
• Palavras-chave em Inglês: Cluster; Moran's Index; Geostatistics; MULTISPATI-PCA
• Área do conhecimento CNPq: Sistemas Biológicos e Agroindustriais
• Link de acesso: https://tede.unioeste.br/handle/tede/6848

Brazil is one of the world’s main agricultural producers, with an important contribution from the state of Paraná, particularly regarding soybean and maize crops. Maintaining this potential is crucial due to food demands, and one way to sustain and increase yield is through monitoring. The objective of this study is to evaluate the associations between productivity of first-harvest maize, second-harvest maize, and first-harvest soybean with meteorological variables such as rainfall, air temperature, dewpoint temperature, wind speed, and radiation. These variables are organized as accumulated or mean values per ten-day period for the historical series between 2010 and 2020 in municipalities located in several mesoregions of the state of Paraná. For spatial monitoring, the following methods can be used: Spatial Multivariate Analysis (MULTISPATI-PCA) to assess associations using cluster analysis; spatio-temporal geostatistical methods; and patterns and spatial correlations among areas. MULTISPATI-PCA reduces the dimensionality of the dataset into spatial principal components (SPC) composed of variables that show stronger associations with productivies. Furthermore, spatio-temporal geostatistics were used to fit models separable, sumMetric, metric, simpleSumMetric, and product-Sum for the three productivities in study, as well as the five meteorological variables, in order to determine the best model for the spatio-temporal representation of the variables. The best fitted model was based on the mean squared error (MSE). Additionally, to analyze the association between variables, the Global and Local Bivariate Moran's Index were used to indicate the degree of association among municipalities. Therefore, with the applied methods, it was possible to identify locations with associative clusters, as well as municipalities and variables with stronger associations with first-harvest maize, second-harvest maize, and firstharvest soybean. The best fits were obtained with the sumMetric and simpleSumMetric spatiotemporal models for the different variables in study. It can be concluded that the employed methods provided explanations regarding the degree of associations between the variables.