Amostragem de sauveiros (Hymenoptera Formicidae) em eucaliptais, pelos métodos geoestatístico e de função de densidade de probabilidade

• Ano de defesa: 2010
• Autor(a) principal: Lasmar, Olinto
• Orientador(a): Bonetti Filho, Ronald Zanetti
• Banca de defesa: Carvalho, Geraldo Andrade, Calegario, Natalino
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: UNIVERSIDADE FEDERAL DE LAVRAS
• Programa de Pós-Graduação: DEN - Programa de Pós-graduação
• País: BRASIL
• Palavras-chave em Português: Geoestatística; Plano de amostragem; Transectos; Atta spp.; Eucalipto; Sampling plan; Geostatistics; Transects; Atta spp. Eucalyptus
• Área do conhecimento CNPq: CNPQ_NÃO_INFORMADO
• Link de acesso: https://repositorio.ufla.br/handle/1/2749

This study aimed to develop sampling plans of leaf-cutting ants in eucalyptus plantations through geostatistical and probability density function methods. A population census of nests of leaf-cutting ants was carried out in plots cultivated with eucalyptus. GIS maps containing the exact location of each nest were also made. The spatial pattern was determined considering the number and area of loose soil from the nests of the ants. Every analysis and maps were made using the statistical program R and the package geoR. In the first plots, the distribution of nests was set according to the Gaussian model for number of nests and the exponential model for area of loose soil of nests, both through ordinary least squares method (OLS). The second plot was adjusted to the exponential model for both, number and area of nests, through the OLS method. These models allow us to create kriging maps of areas infested by leaf-cutting ants, where chemical intervention is necessary in order to reduce the impact on humans and environment. The sampling plan of probability density function was developed with separation of nests into 3 size classes and 21 distance classes according to the transect. The data obtained were analyzed and compared with census and transect strip method. The Generalized Pareto cumulative model was the best to estimate of the density of nests of ants in relation to the probability of viewing at different distances. The probability density function of class size is more accurate than the independent of the size, but both are better than the transect strip method. Transects of 35 and 40m wide are more accurate than 30m when using a sampling intensity above 3 transects per plot.