Sampling plans and application of neural networks to forecast the seasonal dynamics of Bemisia tabaci in soybean crops

• Ano de defesa: 2023
• Autor(a) principal: Arcanjo, Lucas de Paulo
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Universidade Federal de Viçosa
• Palavras-chave em Português: Bemisia tabaci; Mosca-branca - Controle; Soja - Doenças e pragas; Amostragem; Redes neurais (Computação); Entomologia agrícola
• Link de acesso: https://locus.ufv.br//handle/123456789/32060; https://doi.org/10.47328/ufvbbt.2023.453

Soybean Glycine max (L) (Merr) is the most produced, consumed, and traded legume worldwide. After its establishment in Brazil, Bemisia tabaci became a notorious sucking pest on soybean. Robust sampling plans and seasonal dynamic studies of B. tabaci in tropical soybean areas are essential to technicians and farmers early detect pest populations and plan sprays to manage this pest on time. The aim of this study is to determine a sampling plan and seasonal dynamics of B. tabaci in soybean crops through artificial neural networks. These studies were carried out in soybean commercial fields. Whitefly density, climatic elements and soybean age were assessed to support the dataset. In the seasonal dynamic studies, artificial neural networks were developed and selected to study this pest dynamic. The sampling design in this study is composed of 49 samples. The sampling unit and technique are the apical part of the soybean canopy and beating a plastic tray against plant apex, respectively, throughout the plant stages. The artificial neural network structure selected to determine the seasonal dynamic of B. tabaci in soybean crops has five entries (soybean age, average temperature, rainfall, wind speed, and atmosphere pressure) and four neurons in the hidden shell. This model previews whitefly adults with high accuracy from seven days of lag; it is reliable for modelling the seasonal dynamics of the whitefly B. tabaci in soybean crops. In conclusion, this study provides technical tools to scout, early detect, and plan sprays against the whitefly population, avoiding pest outbreaks. Keywords: Sucking pests. AI. Conventional sampling plan. Forecast.