The aims of this thesis were; firstly, to obtain economic values for radiata pine traits to produce appearance and structural lumber, and secondly to analyze the selection of efficient logs and profitable trees to substantiate the development of breeding objectives for solid wood quality. The thesis included three approaches to value wood attributes: hedonic models, partial regressions and stochastic frontiers. Hedonic models generated economic values for pruned and unpruned log traits to produce appearance grades. Values for small end diameter were 0.33, 0.19 and 0.10 US $/mm, and for form 2.6, 1.4 and 0.63 US $ for the first, second and third log respectively. The value of mean internode length was 0.19 US $/cm. Branch size traits were non-significant to explain the log conversion return (p>0.05). The economic value of log traits to produce structural lumber with stiffness of 8, 10 and 12 GPa was estimated with a partial regression. The values were 1.1, 29.7, 0.3 and -0.4 NZ $/m3 for small end diameter (cm), stiffness (GPa), basic density (kg/m3) and largest branch (mm) respectively. Small end diameter and stiffness explained 73% of the variation of log conversion return. The economic values for structural attributes were also derived from a Cobb Douglas stochastic frontier, resulting in 2.1 NZ $/cm for small end diameter and 15.8 NZ $/GPa for stiffness. The change of values between approaches can be attributed to differences of model formulation. The stochastic frontier used aggregate volume of lumber with stiffness of 8 GPa or higher. The partial regression used the economic value of every lumber product derived from the logs, making it more sensitive to changes in wood quality. Data envelopment analysis (DEA) used structural traits and their economic values to assess the technical and economic efficiency of logs to produce lumber with stiffness of 8, 10 and 12 GPa. The most efficient logs had 1:4 ratios between stiffness and small end diameter, whereas logs that did not generate structural lumber had ratios closer to 1:8. Trait economic values from the partial regression analysis were used as attribute prices to estimate cost efficiency. Efficiency measures were significantly correlated with stiffness and log conversion return; however, they were non-significantly correlated with small end diameter and log prices. The technical efficiency of logs to produce structural lumber was also determined using a Cobb Douglas stochastic frontier which determined that the most efficient logs were characterized by a 1:5 ratio between stiffness and small end diameter. Selection of trees for deployment was analyzed with a portfolio model, where risk was represented as the mean absolute deviation of tree returns due to the variability of volume, stiffness and resin defects. Under high variability (risk), the model selected structural trees with large stiffness and high return. These results suggest an opportunity for narrowing genetic variability (via clonal or family forestry) to make the returns from radiata pine structural grades lumber less risky. As variability decreased the portfolio model opted for trees that produced appearance and structural lumber. These trees had a stabilizing effect on their returns, as there were phenotypic tradeoffs between stiffness and volume under optimistic and pessimistic growing scenarios. These results showed the benefits of product diversification at the tree level.

Submitted by Jaqueline Silva (jtas29@gmail.com) on 2016-11-10T17:37:30Z No. of bitstreams: 2 Tese - Flavia Martins de Souza - 2016.pdf: 2069042 bytes, checksum: fe83d06b4777198e60836ea748ef5c2d (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Jaqueline Silva (jtas29@gmail.com) on 2016-11-10T17:37:57Z (GMT) No. of bitstreams: 2 Tese - Flavia Martins de Souza - 2016.pdf: 2069042 bytes, checksum: fe83d06b4777198e60836ea748ef5c2d (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2016-11-10T17:37:57Z (GMT). No. of bitstreams: 2 Tese - Flavia Martins de Souza - 2016.pdf: 2069042 bytes, checksum: fe83d06b4777198e60836ea748ef5c2d (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2016-10-05 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / The use of selection indices as tools for animal genetic evaluation may be more efficient than other selection methods to represent the animal merit. The aim of this study was to propose indices applied to different breeding objectives for production systems of Nellore cattle raised in the Cerrado biome. This study consisted in two phases. First, traits of economic relevance in beef cattle were evaluated. Subsequently, the development of indices and different breeding objectives were considered. The production system evaluated was an operation of purebred Nellore cattle, located in the Northwest of Goiás state, Brazil, where the inputs about management, indices of productivity, income and expenses were obtained. A bio-economic model was used to calculate the economic values of the following traits: age at first calving (IPP), calving interval (IDP), stayability (STAY), earlier calving probability (3P), accumulated productivity (PAC), daily weight gain prior (GMDPre) and post weaning (GMDPos), weight at standard ages of 120 (P120), 210 (P210), 365 (P365) and 450 (P450) days, feed efficiency (EA), carcass dressing (RC), and longissimus muscle area (AOL). Economic values were obtained for a change in one unit of each trait, maintaining the remaining unchanged. Then the economic values were standardized by its standard deviation of each trait. To generate selection indices the genetic (co)variances components of the traits were estimated by Bayesian implementation via Gibbs sampling using bi-trait animal models. The covariance and variance means obtained were used to generate a new matrix (12 x 12) containing all traits. The "matrix bending" methodology was applied to obtain a positive-defined matrix. The selection index equation used was I = b1DEP1 + ... + bnDEPn, where DEP is the expected progeny differences and "b" is the index coefficient that maximize the correlation among the index and the breeding objective. The "b" coefficients were calculated as b = G11-1G12a, where G11 is the genetic (co)variance matrix of the criteria in the index, G12 is the genetic covariance matrix between the selection criteria in the index and the traits in the breeding objective, and "a" is the vector of corrected economic values. Three indices were constructed with different proposes: indices I represent an overall index, which objective is to select harmonics animals. The indices II and III were defined with the purpose to maximize the weaned calves and finished beef cattle production, respectively. Economic values of the traits varied between R$ 0.38 and R$ 68.29 per animal/year. The traits that more strongly affected the economic system were GMDPre (20.55%), IPP (15.70%), AOL (12.13%), GMDPos (11.13%) and P450 (8.98%). The greater economic gains were obtained with index I (R$ 129.12). Indices II and III represented the lowest gains for the system. Generally, the indices are very sensitive to market conditions. However, they may provide more total gains (genetic and economic) as they comprise a set of economically relevant traits to the production system. In addition, the indices may be applied to different purposes in order to attend specific market requirements. / A utilização de índices econômicos de seleção como ferramenta para avaliação genética animal é mais eficiente em representar o valor total de um animal. Por isso, objetivou-se propor índices aplicados a diferentes objetivos de seleção para sistemas de produção de Nelore no bioma Cerrado. O presente estudo consistiu em duas fases, em que na 1ª avaliou-se a relevância econômica de características em bovinos de corte e na 2ª fase desenvolveu-se índices econômicos de seleção com diferentes objetivos. O sistema de produção avaliado foi baseado em uma propriedade de criação de bovinos Nelore PO, localizada no noroeste do estado de Goiás-Brasil, na qual foram obtidas informações sobre manejo, índices zootécnicos, gestão, receitas e despesas. Utilizou-se modelo bioeconômico para o cálculo dos valores econômicos das características idade ao primeiro parto (IPP), intervalo de partos (IDP), stayability (STAY), probabilidade de parto precoce (3P), produtividade acumulada (PAC), ganho médio diário pré (GMDPre) e pós desmama (GMDPos), pesos padronizados aos 120 (P120), 210 (P210), 365 (P365) e 450 (P450) dias de idade, eficiência alimentar (EA), rendimento de carcaça (RC) e área de olho de lombo (AOL). Os valores econômicos foram calculados por meio do melhoramento em 1 unidade de cada característica, mantendo as demais constantes, e em seguida foram padronizados pelo respectivo desvio-padrão de cada característica. Para criação dos índices de seleção, estimou-se os componentes de (co)variâncias de características usualmente selecionadas em Nelore (P120, P210, P365, P450, Perímetros Escrotal aos 365 (PE365) e aos 450 dias (PE450) de idade, AOL, Acabamento (ACAB), IPP, IDP, PAC e STAY), por meio da metodologia Bayesiana via Gibbs sampling, em modelo animal bi-característico. As covariâncias e as médias das variâncias obtidas das análises bicaracterísticas foram utilizadas para formação de uma nova matriz (12 x 12) que envolveu todas as características analisadas. Para assegurar que esta matriz fosse positiva definida, foi aplicada a metodologia Matrix Bending. O índice econômico de seleção foi desenvolvido por meio da equação: I = b1DEP1 + ... + bnDEPn, em que DEP é a diferença esperada na progênie do animal e "b" é o coeficiente regressor que maximiza a correlação entre o índice e o objetivo de seleção. O "b" foi calculado por meio da seguinte equação: b= G11-1G12a, em que o G11 é a matriz de (co)variâncias genéticas entre as características do índice de seleção, o G12 é a matriz de covariâncias genéticas entre as características do índice e do objetivo de seleção e "a" é o vetor de valores genético-econômicos. Três índices foram construídos com diferentes propósitos: o I referiu-se a um índice geral, cujo objetivo foi selecionar animais harmômicos. Os índices II e III tiveram como finalidade a maximização da produção de bezerros desmamados e de animais terminados, respectivamente. Os valores econômicos das características avaliadas variaram de R$ 0,38 a R$ 68,29 /animal/ano. As características que mais impactaram economicamente o sistema foram GMDPre (20,55%), IPP (15,70%), AOL (12,13%), GMDPos (11,13%) e P450 (8,98%). O índice econômico que proporcionou maior ganho econômico para o sistema avaliado foi o I (R$ 129,12). Enquanto os índices II e III representaram menores ganhos para o sistema. Em geral, os índices são sensíveis às condições mercadológicas. No entanto, podem proporcionar maiores ganhos totais (genéticos e econômicos), por envolver um conjunto de características economicamente relevantes para o sistema de produção. Além disso, os índices podem ser aplicados a diferentes propósitos, de forma a atender às necessidades mercadológicas.