Simulation Study On Tree Diameter Growth Of Natural Coniferous And Broadleaved Mixed Forest In Muling Region

Compared with pure forests,natural coniferous and broadleaved mixed forests have higher biodiversity,more complex community structures,richer habitat characteristics,and more stable ecosystems,making them the main stand types for studying forest communities.The growth pattern of trees in natural coniferous and broadleaved mixed forests is an important basis for forest management decisions,and is affected by tree size and competition.In order to effectively quantify the impact of competition on the growth of natural coniferous and broadleaved mixed forests,this study took the natural coniferous and broadleaved mixed forests of Muling Forestry Bureau as the research object,based on three consecutive observation data from 44 fixed plots,and used the method of variance analysis to explore the differences in the growth of individual DBH of the main components of natural coniferous and broadleaved mixed forests under different competition levels,A basic model of single tree diameter growth independent of distance was established using the least square method.Due to significant differences in the growth of different tree species,a dummy variable single tree diameter growth model was constructed using tree species as dummy variables.Based on the dummy variable model of tree species,a single level linear mixed effect model was constructed by introducing random effects from sample plots.At the same time,independent inspection data were used to test the basic model of single tree diameter growth,the dummy variable model of tree species,and the mixed model.The main results are:(1)Construction of a basic model for single tree diameter growth:Using tree size and competition indicators as independent variables,and square growth at breast height as dependent variables,using the form of a logarithmic linear model,using a stepwise regression method to screen variables and their variants that have significant impact on the model,and using a variance expansion factor(VIF)to determine the multiple collinearity between the independent variables.Based on the criteria of VIF<10 and significant parameters(P<0.01),a basic model for the diameter growth of a single tree in a natural coniferous broadleaved mixed forest was constructed,with a coefficient of determination(R~2)of 0.6074 and a root mean square error(RMSE)of 0.7707 cm;(2)Construction of tree species dummy variable single tree diameter growth model:Based on the single tree diameter growth model,virtual variables are introduced,and tree species are used as dummy variables to construct a single tree diameter growth model that considers tree species dummy variables.The dummy variable model reflects the difference in DBH growth among tree species while ensuring the prediction accuracy.The prediction accuracy of the dummy variable model is significantly improved compared to the basic model,with the determination coefficient(R~2)increased from 0.6074 to 0.6274,Increased by 3.29%,and the root mean square error(RMSE)decreased from 0.7707cm to 0.7512cm,a decrease of2.53%;(3)Construction of mixed effect model:Based on the tree species dummy variable model,the random effect of the sample plot is introduced.When the diagonal matrix and its variance covariance matrix are used,the model has the best fitting effect,with an R~2 of 0.6633 and an RMSE of 0.7141cm.The prediction accuracy of the mixed effect model is significantly improved compared to the tree species dummy variable model,with the R~2 of the model increasing by 9.20%compared to the basic model and 5.72%compared to the dummy variable model.RMSE is reduced by 7.34%compared to the base model and 4.94%compared to the dummy variable model.The determination coefficient(R~2)and root mean square error(RMSE)are used as the evaluation indicators of the model;Average absolute error(MAE)and average relative error(MAE%)indicators were used to evaluate the predictive ability of the model.(4)Quantifying the impact of competition on the diameter growth of natural coniferous and broadleaved mixed forests:Using the hierarchical segmentation method,the contribution rates of different independent variables to the DBH square growth model were obtained;Tree size has the highest explanation rate for the variation in DBH growth,followed by competition.The indicator BAL representing competition is divided into five competition levels.After analyzing the differences in the growth of each tree species,it is found that the DBH growth of each tree species significantly decreases with the increase in competition intensity.Under the same competitive intensity,the growth of different tree species is different.When BAL<9,the growth of other tree species is the largest,while Acer mono is the smallest.When 9<BAL<36,the growth of spruce is the largest,while that of Acer mono is the smallest.When 36<BAL<45,the growth of Acer mono is the largest,while that of Birch is the smallest.