Study On Estimation Of Site Index Of Tree Species In Reserve Forest Based On Spatial Statistical Analysis

In the traditional site quality evaluation,site index is used as evaluation index,which usually express as dominant height in the reference age.Actually,it is affected by site condition such as climate,topography and so on,so it is needed to establish site index model based on site factors.In this paper,Pinus koraiensisis taken as research object,which is atree species of reserves forest in Liaoning Province.Based on forest resources survey data of Liaoning Province in 2009,multi-source site data,such as climate,terrain,soil,etc.and using spatial statistical analysis methods and techniques to extract site factors,the spatial regression model of site index ofPinus koraiensisbased on site factors isestablished.Combined with forest resources survey data inXiuyan Manchu Autonomous County,Spatial analysis and valuation of sub-compartment-levelsite index of tree species in Pinus koraiensisreserve forest.Based on the spatial statistical analysis method,the site index can be more accurate to estimate site productivity and to provide a possible method for evaluating site quality.The main research contents are as follows:1.Based on the DEM data to extract the basic terrain factor and complex terrain factor;The space weather data interpolation software professional ANUSPLIN,combined with the interpolation of climate data to DEM data,obtained several climatic factors;Soil factors were extracted using HWSD data;Difference between site factors to use the past field collection of data,the extraction method of site factors and reduce the workload of manual field,saving cost,is more accurate and effective to evaluate the quality of forest site laid the foundation.2.Site index model of tree species inPinus koraiensisreserve forest in Liaoning province based on site factor is established,respectively is based on the multiple linear regression(MLR)model and the geographical weighted regression(GWR)model.The model variables are altitude,annual mean solar radiation,the annual average temperature,annual average relative humidity,soil clay content,soil stone percentage,soil sediment amount,and pH of soil.Based on the analysis and comparison of two models,GWR model is used to explain the extent of the site index of Pinus koraiensis is 30.7% better than the MLR model,so the GWR model is betterfitting on site index of Pinus koraiensis.3.The spatial variation characteristics of site index of tree species inPinus koraiensisreserve forest in Liaoning provinceare most suitable for using the exponential model,and the model fitting degree is high.Spatial distribution of site index is strongly related to spatial autocorrelation in the range of 0~26300m.Geographically weighted regression estimation and spatial Kriging estimation about site index ofPinus koraiensisinXiuyan Manchu Autonomous County is Carriedout.The results show that the prediction accuracy of the GWR model is less error than Kriginginterpolation and more suitable to analyze and predictthe spatial variability ofsite indexof Pinus koraiensis.Finally,the site quality of tree species in the Pinus koraiensis reserve forest in Xiuyan Manchu Autonomous County are analyzed,which can provide a reference for the base construction of tree species in the Pinus koraiensis reserve forest.