| The dynamic changes of forest have two distinct characters, long temporal span and large spatial span. Such two characters lead forestry research confront more difficult. In Virtual Forest Environment the complicated distribution of Forest data is analyzed by applying computer simulation technique, mathematics model and forest knowledge.The assumption of classical statistics depends on independent random distribution of sample data. However, the spatial data exist relatively one another. This relativity violates the basic assumptions of classical statistics. The relevance of spatial data must be considered in spatial data analysis first, in stead of directly using the classical statistical methods. spatial data has random property,providing the accordance with identical character of data distribution studied, the random numbers produced by computer can be used simulate the distribution of data of studied forest area.In this paper, Monte Carlo simulation method was introduced and a large number of computer-generated random numbers were used to simulate forest growth states and distribution for statistical test, and the problem solution obtained and test value were used to verify the validity. Through the analysis of spatial relevance, based on the covariance function and correlation function, the statistical characters of data generated by computer were consistent with the simulated forest areas. Therefore, the characteristic functions of tree such as height and diameter could be obtained. And furthermore, different forest spots were modeled according to these data. As sample space the spot set can once again be used as elements to implement Monte Carlo simulation, to simulate the distribution of holistic forest area. The work is to explore how to use of computer to simulate the forest distribution of spatial data of forest areas in order to reduce the field measurements, and provide a theoretical basis for replacing the burdensome data collection and analysis. |
PDF Full Text Request