Método de detecção de severidade de ataque e controle microbiano de Glycaspis brimblecombei (Hemiptera: Aphalaridae) em Eucalyptus urophylla

• Ano de defesa: 2024
• Autor(a) principal: Gabriella Silva de Gregori
• Orientador(a): Elisangela de Souza Loureiro
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Fundação Universidade Federal de Mato Grosso do Sul
• País: Brasil
• Palavras-chave em Português: Aprendizado de máquina; sensoriamento remoto; monitoramento de pragas; fungo entomopatogênico.
• Link de acesso: https://repositorio.ufms.br/handle/123456789/9758

This study addressed two important aspects of managing the red gum lerp psyllid (Glycaspis brimblecombei), a forest pest significantly affecting eucalyptus plantations. In the first experiment, the impact of different infestation levels of G. brimblecombei on the hyperspectral reflectance of Eucalyptus urophylla leaves was evaluated. Chlorophyll alterations caused by pest attacks affect the spectral behavior of the leaves. To classify infestation levels quickly and accurately, machine learning (ML) algorithms were tested, including artificial neural networks (RNA), decision trees (REPTree and J48), random forest (RF), support vector machines (MVS), and logistic regression (RL). The results showed clear differences in the hyperspectral behavior of the leaves depending on infestation levels, with the highest infestation level showing the highest reflectance values. Of the algorithms tested, logistic regression (RL) and support vector machines (MVS) achieved the best results in classifying infestation levels, with accuracy above 90%. Both algorithms had F-scores close to 0.90 and Kappa values above 0.8, ensuring high classification precision when the entire spectral range was used as input. In the second experiment, the efficacy of the entomopathogenic fungus Cordyceps (= Isaria) fumosorosea in controlling adult G. brimblecombei was evaluated under laboratory conditions. Leaves of Eucalyptus urophylla containing eggs and nymphs psyllids were treated with three solutions: control (sterilized distilled water), 0.5 L.ha⁻¹, and 1.0 L.ha⁻¹ of C. fumosorosea. The 0.5 and 1.0 L.ha⁻¹ doses caused 40 and 60% nymph mortality after 24 hours, reaching 100% mortality in 48 hours. Fungal extrusion was observed in all treated insect cadavers. Egg viability was also affected, with rates of 64.3% for the 0.5 L.ha⁻¹ dose and 95.2% for the 1.0 L.ha⁻¹ dose, respectively, after 120 hours. In the control group, insects remained alive for 10 days, with 100% egg viability. The fungus C. fumosorosea is effective against the eggs and nymphs of G. brimblecombei.