Study On Artificial Neural Network Models Of Whole Stand Growth In Natural Dahurian Larch Forests

This study takes natural Dahurian Larch (Larix gemelinii Rupr.) forests in area of the Daxing'an Mountains as the object, in the systemic environment of MATLAB by making use of remeasured data from permanent sample plots and BP Artificial Neural Network (ANN), it constructs the stand growth dynamic predicting ANN model of natural Dahurian Larch forests in area of the Daxing'an Mountains.Site Class Index (SCI) ANN model, Stand Density Index (SDI) ANN model and Whole stand growth model are first constructed in this study. It also utilizes the permanent plot drawn in Daxing'an Mountains from forest areas in 1990,1995 and 2000 provide proper training and inspection for the settled model.Meanwhile, according to the result, compared with the corresponding model which is based on the growth equation, the fitting, precision, the average error, the average absolute error and the average absolute relative error are all superior to those of the model which is set based on the growth equation of existent trees.This study also intuitionisticly analyzes the fitting effect from the qualitative perspective that is to make use of the powerful drawing function of MATLAB, based on the graph of Site Class Index (SCI) ANN model and Stand Density Index (SDI) ANN model, drawing the graph of Site Class Index (SCI) ANN model and Stand Density Index (SDI) ANN model of Dahurian Larch Natural forest. If knowing the ages, the average height, the average diameter and the tree number per hectare, it will be quite easy to figure out the Index of Site Class and Stand Density conveniently from the corresponding graph. At the same time, it plots the three-dimensional comparison picture and regressing quantitative analysis of three-dimensional emulational curve and model predicting theory and the real checking data in the process of the stand growth which is based on the whole stand growth ANN model, the result shows that the fitting effect the settled model is good, and it has strong ability generalization.In order to improve the predicting accuracy of the model, based on the difference method, the Stand Density dynamic predicting model and Stand Volume dynamic predicting model are set up. The Stand growth ANN system of natural Dahurian Larch forests in area of the Daxing'an Mountains based on the settled models are constructed accordingly. This model systemically and objectively reflects the growing dynamics of natural Dahurian Larch forests in area of the Daxing'an Mountains, it can simulate the growing process of the stand thus providing the detailed growth information for the intensive management and productive policies of the natural mixed forests in this area. Therefore, this paper has a certain theoretical and practical value.