Forest Ecosystem Health Assessment On The Basis Of Artificial Neural Network

In order to push forward the sustainable development of forestry and improve the management of forest, forest health management was listed in the main contents of experiment and demonstration of the sustainable development of forestry by National Forestry Bureau.National Forestry Bureau, in cooperation with National Forestry Bureau of America, established the demonstration Project of China-America Forest Health Management.Considering the forest health of Beijing and the convenience for applying the results of forest health,Beijing was chosen for demonstration area of the management of forest health after Xinfeng County of Jiangxi Province,Foping County of Shanxi Province,Lijiang County of Yunnan Province,Majiang County of Guizhou Province.The paper belongs to one part of the study of Beijing forest health.On the theory or method of ecosystem ecology, vegetation ecology, biodiversity and artificial neural network and Geographic Information System, this paper mainly discusses the assessment of the Badaling forest ecosystem health using the scientific data obtained by investigating in wild forestry and experimenting in the laboratory. At the same time,some important parameters and indices calculated with these scientific data by mathematical software can be put into efficient in assessing the construction and function of Badaling forest ecosystem.There are four indices as follows:the first is the biomass index of every community which can illustrate the health index of ecosystem.The second is the biodiversity index which can illustrate the structure of forestry community. The third is the forest fire index which was used as resistance index. At last, we synthetically analyze all of the health indices of the main forest construction of 121 sub-compartments and apply the artificial neural network in assessing the ecosystem health grade, and make a dynamic analysis for ecosystem health.The Result from analysis above shows that:In Badaling forest, using BP neural network model in assessing the health of every sub-compartment, we can divide it into the very healthy,the healthy, the sub-healthy,the unhealthy and the others.The sub-healthy counts for the majority,58.02%.The very healthy counts for the minority,0.07%.The healthy counts for 22.22%,the unhealthy counts for 0.10%.It is useful and helpful for the assessment of ecosystem health in the efficient management of ecological forestry and the development of ecology to combine neural network model and Geographic Information System.