Research On Method Of Site Quality Evaluation And Model-Assisted Construction System

Site quality evaluation plays an important role in forest management and decision-making,which is the prerequisite for scientific afforestation and subsequent management.In this paper,taking the Cunninghamia Lanceolata Plantation in Guangxi as the research object,the construction method of site quality evaluation model was studied,and three models were completed by using variable coefficient adjustment method,standard deviation adjustment method,quantification method and potential growth method based on BP network.Based on the three models,the model-assisted construction system of site quality evaluation is designed and developed.The specific research content and results are as follows:(1)Variable coefficient and standard deviation adjustment method are used to construct the site index model,which is parameterized and suitable for forested land.During the construction of the model,the Richards equation was determined as the best growth equation of Cunninghamia Lanceolata.And according to the change rate of variable coefficient,it was determined to be 20a as the base age of Cunninghamia Lanceolata.Finally,the two models were tested by stem analysis data.The result shows that the overall accuracy of the variation coefficient method is higher than the standard deviation method,RMSE was reduced by 10%,and R~2 is higher.(2)Considering the situation of site quality evaluation of suitable forestland,this paper uses the Quantitative method I theory to construct a site quality evaluation model,which can be used for suitable forestland.The input variables are seven environmental factors while the output variable is site index.Through factor contribution analysis,the environmental factors affecting the site quality of Cunninghamia Lanceolata in Guangxi according to the contribution degree are soil layer thickness,humus layer thickness,slope direction,slope,litter layer thickness,slope position,and altitude.Compared with previous studies,it is found that the thickness of the soil layer is the most important factor that affects the site quality of Cunninghamia Lanceolata in many areas,and should be paid more attention in the management of Cunninghamia Lanceolata forest.(3)To construct a site quality evaluation model for the potential growth of Cunninghamia Lanceolata in Guangxi,the BP neural network was used to model the growth of stand area,and the parameterized model of site index based on the variable coefficient method was introduced to the modeling process to simplify the modeling process.Finally,using the density index as an adjustable variable,it iteratively calculated the potential growth of the cross-sectional area of different site index at the base age and showed the optimal density.The result shows that the changing trend of the potential growth amount as same as the site index and testifies the accuracy of the model;after comparing with the actual growth amount,it is concluded that the production potential of the research area is between 36%and 52%.(4)The most basic principle of forest management is planting the tree in the fitting place.The site quality evaluation model has its certain spatiotemporal adaptability.The traditional site quality evaluation system is usually developed for a tree species in a certain area,so it has relatively large limitations.Our system abstracts the model building process into nodes,develops a visual interface for model building,and realizes the semi-automatic construction of the model.Users can utilize the model construction method stored in the system to build a model suitable for the local area.During the modeling process,users are allowed to adjust the modeling parameters to improve the accuracy of the model.