Individual Tree Growth Model Of Hunan Quercus Natural Forest Based On Mixed Effect

The growth model of natural forest has always been a difficult point in forestry research at home and abroad.The individual tree growth model is an important part of the tree growth and harvest model,and the density and site are important factors affecting the single tree growth.Therefore,this study uses the method of quantitative method I and the method of nonlinear mixed effect to explore and construct a single tree growth model of Quercus natural forest that reflects density,site differences and regional differences,objectively reflects its growth law,and provides theoretical support for stand estimation and forest quality improvement and provide a new idea for the site quality evaluation of Quercus natural forests.The main research contents and findings are as follows:(1)A growth model of average tree diameter at breast height of Quercus natural forests in Hunan based on mixed effects was constructed.Four theoretical growth equations were fitted,and the logistic model with the best fitting effect was finally selected as the optimal basic model(R2=0.339),the quantification method I was used to screen out altitude(ALT),slope(SLO),slope aspect(AS)and plant number density(ND)has a significant effect on the growth of average wood diameter at breast height.The optimal nonlinear mixed effect model with site factors and density is constructed as:Dj=(aj+al*ND)/(1+b*exp(-c*AGE))+λ,R2=0.930.(2)Established the average tree height growth model of Hunan Quercus natural forest based on mixed effects.The growth of average tree height is affected by many factors,among which altitude and region have the most significant impact on it.There are significant differences in average tree growth in different altitude conditions and different regions.From the fitting results,it can be seen that the Logistic model has the best fitting effect.The fitting accuracy is R2=0.669.The region and altitude are used as fixed effects and random effect factors to construct a nonlinear mixed effect model.It is finally determined that the model has the highest accuracy and the smallest error when the altitude is used as a random effect on the parameter a of the optimal basic model(R2=0.792),the R2 has increased by 18.35%.The model expression is Hj=aj/(1+b*exp(-c*AGE))+ε.(3)Established a quality evaluation model of natural Quercus forest site based on the growth equation of diameter at breast height of the thickest dominant wood.Through variance analysis,it was found that altitude,slope,slope position,and slope aspect are site factors that have a significant effect on the growth of the thickest dominant tree DBH;Richards formula is used to fit the dominant tree DBH growth model,which has a good simulation effect and model accuracy R2=0.732;the site factors that significantly affect the growth of diameter at breast height of the thickest dominant trees were combined and clustered to form site type groups,and the thickest dominant trees in Hunan Quercus natural forests with site type groups were constructed.The expression of the DBH-age is:Dj=aj*(1-exp(-c*AGE))^b+ε,the coefficient of determination R2 is 0.927,which improved by 26.7%.Significantly.Thus,a site quality evaluation equation based on site classification is derived with the diameter at breast height of the thickest dominant wood as the evaluation index.It’s(?).(4)Established a high growth model of the thickest dominant tree in Hunan Quercus natural forest based on mixed effects.The growth of the thickest dominant tree height is most significantly affected by the region,followed by slope and soil type.The theoretical growth equation is fitted by R language,and the power function model is finally determined as the basic model for the growth of the thickest dominant tree(R2=0.311).Taking slope and soil type as fixed effect factors,acting on the parameters,constructing a tree height model with site factors,and the determination coefficient R2 of the model is 0.355.Based on this,the region is used as a random effect factor to act on the model parameters,and a nonlinear mixed effect model including site and region is constructed.When the region is finally determined to act on the parameter a,the model has the highest accuracy,respectively R2=0.626,and the optimal nonlinear mixed-effects model expression is:Hj=aj*AGE(b+b1*PD+b2*TL)+ε.(5)Constructed a model of average dominant tree height-diameter at DBH of Hunan Quercus natural forest based on mixed effects.Fitting the correlation between average dominant tree height and altitude by theoretical growth equation,the final power function has the best fitting effect,the coefficient of determination(R2)is 0.405.Based on this,the selected altitude factor is used as a random effect on the parameter.When the altitude factor acts on the parameter b,the model determination coefficient is up to 0.797,which is improved The optimal model expression is:Hj=a*Dbj+ε.(6)It is clarified that the growth processes of different dominant tree species of Quercus in Hunan are basically similar or the same,which can be expressed by the same growth model.The tree species was added as a random effect to the optimal models of different target trees,and a mixed-effect model with tree species effects was constructed.The R2 of the final average tree DBH model is only increased by 1.6%;the R2 of the average tree height model is only increased by 1.0%;There was no significant change in R2 of the thickest dominant tree height model、the coarsest dominant tree DBH model and the average dominant tree height DBH correlation model.The results show that the same modeling unit can be used for the single-tree growth prediction model of different dominant tree species of Quercus in Hunan,which plays an important role and significance for scientific and rational management of Quercus natural forest in Hunan.