Study On The Variation Regularity Of Larch Bark Thickness

In this study, we selected the the different ages and different density of 25 larch plantation plots data of Daqing yang and Yong cui forest farm in Dailing forestry bureau, Heilongjiang Province, this study described the variation of larch bark thickness. This study used the data of analytic trees, we changed the data of DBH, tree height, relative height, crown length, rate of crown length through logarithmic,exponent,power function, simple linear function transformation, respectively, and use SAS program analysis to fit the model parameters, and draw scatter plots and residual plots, finally,selected the best model for each variable expression. Free combination the variables of DBH, tree height, relative height, crown length, rate of crown length, we use the SAS statistical software carried on multiple linear regression analysis through REG procedure and use the variance inflation factor (VIF) to determine multicollinearity among the independent variables, establishing the multivariate linear model of bark thickness at different height. Comprehensive analysis the domestic and foreign reference,we selected three types of 10 trunk diameter peeled prediction model:Grosenbaugh ratio equation, regression model and cut of the model, a comprehensive comparison of each model.The results showed that:logarithmic function could best reflects the relationship between the diameter and the diameter of the thickness of the bark, the relationship between bark thickness at DBH and DBH model was:bt=-0.1903+0.30721n (dbhob), Logarithmic function could best reflects the relationship between the diameter and thickness of the bark of tree height model was:bt=-0.0504+0.25011n (h). A linear function could best reflects the relationship between the larch bark thickness and diameter of the relative height model was:bt = 0.8154-2.0164rh, Logarithmic function could best reflects the relationship between the diameter of the larch bark thickness and crown length model was bt=-0.2521+0.42101n (cl). Using variable selection method of multiple linear statistical analysis constructed larch bark thickness optimal equation was:bt= 0.1237+0.0413dob-0.0179h+0.0968rh+0.0191cl, parameter estimation significance test (P<0.0001), model checking (F=423.57, P<0.0001) and multicollinearity test (VIF<10) have shown that the model achieved better fitting effect.With the increasing of the diameter, the Larch bark thickness increased which shown a positive correlation between the diameter and Larch bark thickness, indicating that the bark thickness from the bottom to the top of the tree become thinner. The different places of the same tree where bark thickness was different. Large-diameter where bark was more thickness. Grosenbaugh ratio formula has a lot of flexibility, no fitting parameter does not require a model. Overall, the model segments was significant, regression model has a smaller prediction error. especially Cao proposed the variables model which contained skin diameter, tree height and relative tree height, diameter and peeled skin. Due to the taper equation model does not contain skin variables, and therefore have a greater prediction error. Different types of models in the process of forest management have a certain adaptability.