Quantitative Genetic Analysis and Selection for Tree Height of a Coast Redwood (Sequoia sempervirens (D. Don) Endl.) Breeding Population

Coast Redwood (Sequoia sempervirens (D. Don) Endl.) is an ancient and iconic conifer tree species endemic to a narrow strip of coastline from southern Oregon to central California. The species is environmentally important, providing a diversity of ecosystem services, and the redwood region has been classified as a "globally significant" ecoregion by the World Wildlife Fund (Ricketts et al. 1999). The majority of the remaining 120,000 acres of ancient, old-growth forests are in protected land, while approximately 1.2 million acres are privately owned, generally younger forests managed for timber production (Noss 2000). Coast Redwood is the highest valued timber in California (CA BOE 2014). Stakeholders have established tree breeding and improvement programs to identify superior parents, families, and clones for timber yield and quality traits. The objectives of this research were to estimate genetic parameters for tree height across multiple variety trials from a coast redwood breeding population, to rank trees based on genetic merit, and to assess experimental design for optimizing efficiency and gain from selection for coast redwood tree improvement programs. Modified forms of Henderson's mixed effect "animal model" (Henderson 1985) were applied using restricted maximum likelihood (REML) and Bayesian Markov chain Monte Carlo (MCMC) methods for estimating population genetic parameters and genetic prediction for individual trees and tree families. Different variance structures were tested to account for environmental heterogeneity in tree test sites. Narrow-sense heritability for coast redwood height was estimated to be 0.09 to 0.26 and repeatability to be 0.23 to 0.50 for growth years 2, 4, and 9, and broad-sense heritability for height at 9 years to be 0.17 to 0.36. Progeny were ranked by average family breeding value, parents by general and specific combining ability, and clones by genotypic value predictions. Genotype by environment (GxE) interaction was found to be an important factor in estimating genetic parameters. GxE estimates can be applied to selection of genotypes that are superior for specific environments or those that are broadly adapted to a wide range of environmental conditions. GxE estimation was limited due to the practice of partial replication of variety trials across locations. In the future, it is recommended to completely replicate the experiments in multiple locations. In addition, sites with more entries and few blocks with multiple replicates per block consistently resulted in genetic estimates with higher accuracy than those with fewer. The findings from this research are pertinent to both the conservation and responsible management of the species. Scientific based decision making and an understanding of the in forest systems are necessary for maintaining society's need for natural resources and the preservation of the environment.