| Forest site quality evaluation is the basis of forestry production,and scientific and reasonable evaluation is the first prerequisite for adapting to local conditions and avoiding blind afforestation.It is of great significance for improving forest ecological efficiency and accurately improving forest quality,and it is also the core demand for ecological restoration construction,reasonable forest cultivation and sustainable development.In this study,Cunninghamia lanceolata forest of Caijiaqiao State-owned forest farm in Jingde County,Xuancheng City,Anhui Province was taken as the research object.Based on the second-class survey data of forest resources,DEM data,standard deviation adjustment method,coefficient of variation adjustment method and proportion method,the status index table of Cunninghamia lanceolata in the study area was compiled.Based on the quantitative theory I and random forest regression algorithm,the evaluation model of Cunninghamia lanceolata status index and site factor suitable for unforested land was constructed.The influence of each site factor on the site quality was analyzed,and the site quality of the forest farm was evaluated,which provided theoretical basis for forest farm operation and decision making.The main research contents and results are as follows:(1)The average height growth of dominant trees of Cunninghamia lanceolata stands was fitted with the height growth equation of candidate 7 trees.The results showed that the logarithmic hyperbolic equation had the best fitting effect(R~2=0.871,Q=0.907),and the growth equation was lg(H)=1.1599-2.9933/A.Standard deviation adjustment method,coefficient of variation adjustment method and proportion method were used to compile and evaluate the status index table.The results show that the status index table prepared by standard deviation adjustment method has the best effect,and its test accuracy is 100%.(2)Two methods of quantitative theory I and Random Forest(RF)were used to construct the evaluation model of Cunninghamia lanceolata status index with elevation,slope direction,slope position,slope and soil thickness as independent variables.The results showed that the evaluation model based on random forest algorithm was better,with the determination coefficient R~2of 0.7549 and prediction accuracy of 83.938%.Classification is better.(3)The importance analysis of site factors in the study area was carried out based on the Random Forest algorithm model.The results showed that slope,soil thickness and altitude were the main factors affecting site quality in the study area,and their important values were 195.77,168.07 and 58.10,respectively.The relationship between the status index and site factors was analyzed according to the partial dependence graph.The results showed that the site quality of forest land was better when the slope was less than 25°,the soil thickness was more than 80 cm,the altitude was more than 400 m,the shady slope and downhill slope were more suitable for the growth of Cunninghamia lanceolata.(4)Based on the established status index evaluation model,the site quality of Cunninghamia lanceolata in the study area was evaluated.The results showed that in the small class with Cunninghamia lanceolata forest,the site quality grade was medium or above 88.142%,and the site quality grade was poor 11.859%.In the whole stand,Cunninghamia lanceolata forest accounted for 99.147%of the woodland with good site quality,and only accounted for 8.260%of the woodland with poor site quality.Based on the location index distribution and the current stand status,it is found that the current tree species selection in the study area is reasonable,and some suggestions on future forest management strategies are put forward accordingly. |
PDF Full Text Request