| One of the main aspects of Digital Forestry is digital interpretations of static and dynamic forestry characteristics, analysis and decision-making. To achieve this goal, not only forestry survey data are required, but also professional models which can achieve dynamic updating and support management decisions are needed. So modeling of forest growth is the core and one of the basic technologies of forestry information. In this paper, Chinese pine which was the main coniferous forest species in the Beijing mountain area were selected as studied material, with the help of GIS, Geostatistic method, neural networks, genetic algorithms and dynamic planning, entire forest growth model and management model of Chinese pine were proposed, with a wide range of application, high precision and practicability. This model aimed at providing necessary and accurate theoretical foundation for afforestation planning, management, dynamic update, and four-dimensional expression of Beijing forest resources, moreover, providing basic information for building digital forestry, precision forestry, and other modern forestry technology systems, which promote the capital's sustainable forest management and sustainable development of forestry.The paper studied the growth simulation of Chinese Pine in Beijing mountain and optimization of management, the study focused on the limitations of current entire forest growth model and how to improve the accuracy of system-model. The studies included: (1) establishment of multiple site index model and the distribution of site index: facing the problem that for the implanted land, we can not access to the site index by using the entire forest growth model to forecast, multiple site index model was established. Take Fangshan district of Beijing for example, based on the method of interpolation in GIS, distribution of site index was built, so as to cimbine the entire forest growth model and specific location information, which raised the practicability and expansion of the entire growth system; (2) establishment of basal area growth model of stand: As the basal area growth model of stand is the core of entire forest growth model, the precision of model will directly affect the other predicted precision of stand-factors, therefore, much attention has been paid to the selection and establishment of basal area growth model of stand, to improve the predicted precision of entire forest growth model; (3) establishment of form height model of stand : stand volume reflects the state and level of scientific management of forests, which is an important index to evaluate contribution of ecological forest, form height model of stand play an important role in volume measurement and prediction. In order to obtain high-precision form height model, the method studied of improving the stand volume precision was carried out; (4) establishment of optimal density control model: in this part, the study concerned the density effect modeling, optimal density control modeling and visualization applications of density control model, these studies provided theoretic basis for optimal density control in the process of stand growth.By the studies, following conclusions can be drawn: (1)in the application of genetic algorithms in fitting multi-shape site index curve, the comparison analysis of parameter optimization between 5 selection function and 2 mutation function can help us draw the conclusion that none of the selection function resulted notable difference, Gauss function in the mutation function was distinctly superior to Uniform function; (2)both the neural network fitting model and its predicted precision were higher than stepwise regression model, so neural network fitting model can be used to estimate multi-site index; (3) by GIS and geostatistics theory, we can establish the site index distribution map of a certain tree species; (4) based on the improved Schumacher equation, basal area growth model by reparameterization methods was of more flexibility and precision than the model by Richards models and Korf-A equation. This provided a good reference for the building of basal area growth model of entire forest, so as to raise the predicted precision of entire forest model; (5) under conditions of the known breast diameter and tree height, we can get the different sectional diameters of stand by using neural network model. With help of these data, the standing tree volume estimated by average cross section sectional measurement method was of higher precision than by the of duality volume model; (6) density effect model established by stand density index was superior to the one by number of trees.The main innovation of the paper was the following five aspects: (1) the improved entire forest growth model system: By GIS technology and geostatistics theory, we established the Distribution map of site index for Chinese Pine in Beijing's Fangshan mountain area, which combined the spatial information and entire forest growth model, and improved the practicality of entire forest growth model ,this model system not only was applicable to the stand growth simulation, but also can be applied to the forest planning simulation.So that the compatibility of the entire forest growth simulation model has been expanded to the forest land and non-planed land. In the paper, the entire forest growth model system included not only the multiple site index model or the distribution map of site index, but also itself with much compatibility ; (2) Multiple site index modeling based on optimization neural network method of genetic algorithm ; (3) Stand basal area growth model based on improved Schumacher equation can paved way for entire forest basal-area growth model, which can improve the predicted precision for the entire forest model; (4) Standing tree volume measurement by optimization neural network methods of genetic algorithm and average cross section sectional measurement was of higher precision than binary volume equation. It served as a new research methods and ideas to improve the precision of stand volume measurement.
|
PDF Full Text Request