Mudanças climáticas e ferrugem asiática da soja no Estado de Minas Gerais

• Ano de defesa: 2007
• Autor(a) principal: Lelis, Vicente de Paula
• Orientador(a): Costa, Luiz Cláudio
• Banca de defesa: Jesus Júnior, Waldir Cintra de , Justino, Flávio Barbosa
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Viçosa
• Programa de Pós-Graduação: Doutorado em Meteorologia Agrícola
• Departamento: Agrometeorologia; Climatologia; Micrometeorologia
• País: BR
• Palavras-chave em Português: Ferrugem da soja; Favorabilidade; Mudanças climáticas
• Palavras-chave em Inglês: Soybean rust; Favorableness: Climatic changes
• Área do conhecimento CNPq: CNPQ::CIENCIAS AGRARIAS::AGRONOMIA
• Link de acesso: http://locus.ufv.br/handle/123456789/1526

In 2001, the soybean Asian rust caused by the fungus Phakopsora pachyrhizi (Syd & P. Syd) appeared in the South American continent, as causing many damages to the crops in Brazil, mainly in the states of Mato Grosso, Goiás, Minas Gerais and Mato Grosso do Sul at the seasons 2004/2005 and 2005/2006. The temperature, air relative humidity and the presence of the liquid water on the leaves rather favor the development of this fungus. Two models were tested to identify the favorableness to the development of the disease in Minas Gerais State. The number of hours with air relative humidity higher or equal to 90% were considered in the first model, whereas the dew point depression lower than 2 oC was considered in the second one. The temperature range from 18 oC to 25 oC was considered as ideal for development of the fungus Phakopsora pachyrhizi. The research with data of the hourly temperature, minimum temperature and hourly air relative humidity obtained on 2005 and 2006 in 14 meteorological stations of the data collection platform pertaining to CPTEC-INPE. As the models showed similar results, the first model was chosen and again tested, by using data from two field experiments that were accomplished in 2006: one in Viçosa county located in Zona da Mata region; and another one in Uberlândia county located in Triângulo Mineiro. In this stage of the research, the measures of the severity in soybean crop were obtained. In analyzing the relationship between favorableness and the data (temperature and air relative humidity), the neural artificial networks (RNAs) were used. This procedure made possible the generation of neural nets for the separation and simultaneous obtainment of the information about the favorableness as a function of either temperatures and daily average relative humidity as input data into network. The networks with better performance contained 10 and 15 neurons, and the used learning type was the back propagation. After identification of the networks providing better response to relationship, the air relative humidity was calculated from the daily data of the average and minimum temperatures proceeding from the general circulation - HadCM3 model. Using the data, the favorableness projections were performed for the years 2020, 2050 and 2080, from which the results showed a tendency to increased favorableness to the development of the rust in almost all regions in the State.