Assessment of citrus canker severity in sweet orange genotypes by mixed models

Abstract

The citrus canker disease affects leaves and fruits of orange trees and causes great economic losses, therefore, the use of statistical methodologies in the analysis of citrus datasets is essential. In a lot of studies about the disease behavior, several measurements may be carried out in the same experimental unit, characterizing them as repeated measures, making it necessary to use statistical methods that take this fact into account. In this sense, linear mixed effects models have become an important analysis tool. The objective, in this work, was to study and apply the theory of linear mixed models to a dataset, coming from an experiment implemented in the Northwest region of the Paraná State, whose interest was to evaluate the severity of citrus canker in sweet orange leaves for fourteen different genotypes. Using available packages in the R statistical environment, it was observed that the mixed linear model provided a good fit of the model to the data. Even, it was concluded that some of the experimental varieties are more susceptible to the development of citrus canker than others.