Application of Experimental Models and Spatial Analysis in the Breeding of Excellent Families of Casuarina equisetifolia

Using the 10 years old short branch Casuarina equisetifolia gene resource library from 109 families and 23 species sources of 10 countries introduced by the Australian Forest Seed Center as materials,ASReml software was used to mix linear models and BLUP methods to compare the effects of different experimental models and spatial analysis on the genetic effects of trees.Based on the breeding values estimated by the optimal model,excellent families were selected.The results indicate that AR1ȵRepFam of 7 models has the smallest AIC value and is the optimal model.Compared with non spatial analysis,spatial analysis showed 0.42% reduction in tree height additive variance,16.52% reduction in residual variance,and 18.87% reduction in individual plant heritability;the additive variance of breast diameter increased by 174.00%,the residual variance decreased by 9.29%,and the heritability per plant decreased by 20.00%;the additive variance of volume increased by 559.26%,the residual variance decreased by 42.56%,and the heritability per plant decreased by 25.00%;the additive variance of crown width increased by 8.10%,the residual variance decreased by 56.76%,and the heritability per plant decreased by 27.16%.17 fast-growing excellent families were selected based on the positive order of volume breeding values and a selection rate of 15%.The genetic gain percentages obtained for tree height,diameter at breast height,volume,and crown width were 12.57%,29.39%,65.64% and 30.00%,respectively.These excellent families of Casuarina equisetifolia can be used for promotion and planting.