Análise por regressão não-linear da incidência do cancro cítrico e determinação do tamanho da amostra na região noroeste do Estado do Paraná

• Autor(a) principal: Yamamoto, Akemi Yamagata
• Data de Publicação: 2009
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Repositório Institucional da Universidade Estadual de Maringá (RI-UEM)
• Texto Completo: http://repositorio.uem.br:8080/jspui/handle/1/1097
• topic: Cancro cítrico; Incidência; Modelos de regressão não-linear; Cluster; Citricultura; Doenças; Curva de crescimento; Identidade de modelos; Paraná (Estado); Brasil.; Citrus canker; Growth curve; Models of nonlinear regression. Identity of models; Paraná (State); Brasil.; Ciências Agrárias; Agronomia

Citrus canker was introduced for the first time in the State of Paraná in the county of Lupionópolis in 1957 and causes large damage to the orange-producing countries, being one of the most important diseases of citrus. The disease is caused by the bacterium Xanthomonas citri subsp. citri. It is easily to spread and prevention is the best way to combat it, so the sooner the disease is detected, the lower the damage. As there are few epidemiological studies on citrus canker, it was investigated the temporal aspect of the progress of the disease in two counties in the Northwest of Paraná State. Six plots of sweet orange (Citrus sinensis Osbeck), varieties "Pera', "Valência", 'Natal" and "Folha Murcha" were selected, and all plants of the plots were visually assessed for the presence or absence of symptoms of citrus canker. The data of the disease incidence were adjusted to the models of growth sigmoidal, based on the measurements of curvature intrinsic and parametric of Bates and Watts, Box bias, and asymmetry of Hougaard. Among the nonlinear regression models studied, logistic and Gompertz, it was not possible to find a single model with desirable level of reliability for all sets of data analyzed. Due to the nature of the data, the plots were divided into quadrats. The beta-binomial distribution and cluster sampling were used to determine the optimum sample size of each evaluation in each plot. For the data of each plot, there was no significant difference of growth curves in parameters, being possible to be represented by a single model.