Huanglongbing vector insect counting (HLB) by GAMLSS

Abstract

Citriculture is one of the most important agricultural activities globally, with Brazil being one of the leading world producers. Thus, such activity is essential for the country's economy and the producers who depend on it. In this sense, the fight against Huanglongbing, one of the most devastating citrus diseases caused by vector insects, is essential to guarantee the quality of the fruit and avoid economic losses. The present work analyzed the counting of insect vectors in a commercial orange orchard in an observational study carried out in the municipality of Paranavaí, state of Paraná, Brazil, using the methodologies of generalized linear mixed models (GLMM) and generalized additive models for location, scale, and form (GAMLSS), with Negative Binomial probability distribution. Data were obtained by counting insects trapped in sticky traps at twelve fixed points in the orchard at three different heights and collected over seven fortnights. The results indicated that the GAMLSS model presented better results by including the linear predictor for modeling the scale parameter associated with the study factors based on the AIC criterion and diagnostic analysis tools.