Study On Self-thinning Law And Growth Model Of Oak Natural Forest In Hunan Province

Oak species are widely distributed in China and are dominated by natural origin.Oak is a strategic tree species for forest management in China and great importance should be given to improve the quality of oak forest resources continuously and accurately.However,the original forest has decreased sharply and the secondary forest has increased due to various factors for a long time.It is urgent to study the key technologies of oak natural forest quality improvement and increment.In this paper,265 fixed sample plots of national forest resources continuous inventory oak natural forest in Hunan Province were selected,which are less disturbed and stressed.And 34 fixed plots in Huaihua,Yiyang,Yueyang and Shaoyang were retested.Meanwhile,26 typical temporal sample plots of oak natural forest were set up and investigated.The key problems related to the growth of oak natural forest,such as stand age,site quality and self-thinning law,were studied and analyzed within the guidance of system science,forestry,ecology and biostatistics.A series of growth models of oak natural forest were constructed to enrich the growth and harvest prediction theory of natural forest and provide technical guidance for the sustainable management of oak natural forest.The main research contents and results are as follows:(1)Estimation of stand age of oak natural forest.Two methods for estimating the age of natural forest based on multi-stage continuous monitoring were proposed,namely interval assignment method and initial value location method.The stand age estimated by initial value location method had no significant difference form the actual age,and the initial value location method performed stronger operability and applicability in forestry.The initial value positioning method can be expressed as follows:The initial age of the sample trees in the first stage is set according to the tree species.The age and DBH growth curve cluster with only lateral displacement is generated by using the Richards equation.According to the actual age when the tree height grows to 1.3 m,the Richards growth curve of the tree species can be accurately located.The age of each tree species can be estimated by curve and the age of natural forest stand can be obtained by weighted average of sectional area.(2)Site quality evaluation of oak natural forest.The growth of oak natural forest was significantly affected by the altitude,slope,slope direction and soil thickness by using the quantitative theory I method.The weight of dominant site factors was calculated by combining analytic hierarchy process and entropy weighting method,the results showed that the ranking of weight were altitude(0.348)>soil thickness(0.247)>slope(0.225)>aspect(0.180).According to the comprehensive score,the site quality grade was divided into five relative grades including excellent[0.8,1.0],good[0.6,0.8),medium[0.4,0.6),poor[0.2,0.4),inferior[0.0,0.2).The oak natural forest with good site quality(excellent,good and medium)accounts for 61.51%,and the oak natural forest with poor site quality(poor and poor)accounts for 38.49%.The average tree height and volume per unit area of stands with good site were significantly higher than those with poor site(P<0.05),indicating that the suitable site of oak natural forest in Hunan was wide,but there were significant differences in quality and efficiency.(3)Study on self-thinning model and self-thinning law of oak natural forest.Based on the self-thinning theory,a variable density self-thinning model(VDM)reflecting the self-thinning law of oak natural forest with the average DBH of stand and the average height of dominant tree species was constructed and deduced,that was,lnN=k+alnD+βlnH.The accuracy of the VDM and Reineke model was compared by using the least square method,quantile regression method and stochastic frontier analysis.The results showed that the fitting accuracy of VDM constructed by stochastic frontier analysis was significantly higher than that of Reineke model(P<0.05),and the VDM had more biological significance.The self-thinning rate is affected by the average diameter,while the average height of dominant tree species mainly affects the maximum density of stand.Meanwhile,the self-thinning rate is affected by stand types.Due to the greater absolute value of diameter coefficient,the death rate of oak pure forest is faster than oak conifer mixed forest and oak broad-leaved mixed forest.In addition,once the self-thinning occurred,the stand diameter distribution may change from inverted "J" type to left leaning single peak mountain curve or normal distribution curve in a relatively fast time.(4)Study on individual basal area increment model of oak natural forest.The individual basal area increment model was constructed by using stepwise regression method.Also,the effects of site quality grade,competition intensity and tree species composition on oak growth were evaluated by virtual variable method and mixed effect method.The results showed that the basal area growth of oak was affected by diameter,total basal area of stand,altitude and annual average precipitation,as well as the competition intensity and tree species composition.The mixed effect model performed best in the fitting of basal area increment model.(5)Study on diameter distribution model of oak natural forest.Six common probability distribution functions were selected to fit the diameter distribution,it was found that Weibull distribution function performed best.Three parameter prediction models including stepwise regression model,dummy variable model with stand type and self-thinning state,and the artificial neural network model(ANN)were established and compared.The results found that the ANN model with diameter class(C),average diameter(D)and stand type(T)as input variables had the highest prediction accuracy.Meanwhile,a stand basal area model based on the law of stand diameter distribution was established,and no significant difference was found between the basal area model and the dummy variable model(P<0.05).(6)Study on stand growth model of oak natural forest.The traditional oak compatible growth model with stand age and initial basal area as the independent variables was constructed.The site quality grade,self-thinning state and stand type were introduced into the traditional model as dummy variables,and the prediction accuracy of the two models was compared by t-test.The results showed that the compatibility of self-thinning state and site quality grade dummy variables can significantly improve the prediction accuracy of cross-sectional area and volume model.The R2 of the beginning stock volume equation,the ending sectional area equation and the ending stock volume equation increased by 0.026,0.031 and 0.069 respectively,and the RRMSE decreased by 2.98%,2.87%and 2.53%respectively.