Utilize este identificador para referenciar este registo: http://hdl.handle.net/10400.5/1875
Título: Modelos estatísticos espaciais para ensaios de populações vegetais
Outros títulos: Spatial statistical models for plant populations trials
Autor: Gonçalves, Elsa Maria Félix
Orientador: St. Aubyn, António Gabriel da Silva
Palavras-chave: grapevine
mixed models
spatial model
experimental design
unreplicated trials
genetic variability
Data de Defesa: 2008
Resumo: Strategies for controlling spatial variation, for quantification of genetic variability and for an efficient genetic selection in large trials of vegetatively propagated perennial plants were studied. The mixed model was carefully studied, mainly concerning the variance parameters, the best linear unbiased predictors of genotypic effects and the heritability. Methodological studies revealed that when yield data from trials of vegetatively propagated plants were analysed (case study - grapevine clones), the decomposition of the spatial model-fit residuals in two components is possible: a spatially independent component that explain 40 to 60% of the total residual variance, and a spatially dependent component, characterized as an anisotropic power correlation function. With the aim of identifying the most suitable experimental designs for trials of plant populations, a simulation a study was carried out in order to assess the comparative efficiency of various experimental designs, fully replicated and unreplicated. The results obtained with fully replicated designs indicated a greater efficiency for row-column designs. The use of unreplicated trials as a way of preserving and quantifying the genetic variability of plant populations was feasible, mainly when collections are organized according to an alpha-alpha design containing over 250 genotypes and a check plot frequency higher than 33%.
Descrição: Doutoramento em Matemática e Estatística - Instituto Superior de Agronomia
URI: http://hdl.handle.net/10400.5/1875
Aparece nas colecções:BISA - Teses de Doutoramento / Doctoral Thesis

Ficheiros deste registo:
Ficheiro Descrição TamanhoFormato 
teseElsagoncalves.pdf1,39 MBAdobe PDFVer/Abrir    Acesso Restrito. Solicitar cópia ao autor!


FacebookTwitterDeliciousLinkedInDiggGoogle BookmarksMySpace
Formato BibTex MendeleyEndnote Degois 

Todos os registos no repositório estão protegidos por leis de copyright, com todos os direitos reservados.