Individual Tree Crown Width Prediction Models For Larix Olgensis Plantation

Canopy is an important part of forest and a crucial place for photosynthesis and respiration of trees.Canopy width is a significant index to describe canopy and has great application value.Larix olgensis is one of the representative tree species in northeast China.In this study,Larix olgensis plantation in Dongzhelenghe forest farm,Langxiang forestry bureau,Heilongjiang Province that was taken as the research object.The survey data were used to simulate the growth of canopy width.Three modeling methods were adopted:1.With the stand factors as the covariable,the conventional crown width-diameter regression model was constructed;2.The influence of sample plots and age groups was analyzed,and the nonlinear mixed effect model was constructed based on the conventional crown width-diameter model;3.Based on the constructed nonlinear mixed effect model and considering the correlation between the canopy width and the total canopy width,the crown width additivity model was constructed.Taking a series of evaluation indexes into consideration,a canopy width model system was constructed for the larch of this area.The main findings are as follows:?1?Among the 10 basic crown width-diameter models,the fitting accuracy of the three-parameter logistic model is the best,which is the best basic model for constructing the traditional regression model.Its equation is CW=a/?1+bexp?-cd??.Among other tree factors and stand variables considered to have an impact on canopy width,canopy length rate,height to diameter ratio and tree height had the highest correlation with canopy width growth.So they were used to participate in the modeling.The fitting accuracy of the model after they were added to the modeling was significantly improved compared with the basic model,in the form of:CW=?a+a1*CR?/?1+?b+b1*HDR?*exp?-?c+c1*H?*D??.?2?Based on the improved model in?1?,a nested two-level nonlinear mixed effects model was constructed by considering two levels of block and plot as random effects.The results show that the mixing model converges and has the best accuracy when the random effects of the level of the block and block*are applied to the parameter b2.Compared with the improved model,it can be seen that the AIC,BIC and-2ll values of the mixed model are all smaller than the improved model,so the model precision is better than the basic model and the improved model,in the form of:CW₄5)6))=(b1+(b2+₂₄)+u₂₄5)))*CR)/?1+?b3+b4*HDR?*exp?-?b5+b6*H?*D??+?₄5)6))?3?Based on the improved model in?1?and considering the correlation between crown width,the additive model of crown width was constructed with nonlinear seemingly uncorrelated regression.The result shows that the fitting accuracy of it is similar with?1?,and is worse than?2?obviously.At present,there are few studies on the canopy width of Larix olgensis plantation,especially using nonlinear mixed model and additive model.However,these two models have great advantages over traditional models.This study is helpful to predict the tree growth,productivity and the evaluation of wood quality grade in the plantation of Larix olgensis,which can provide a basis for the scientific management of the species and is of great significance in forestry production.The research can also have significant meaning for the research of other species.It has generalized significance.