Building Forestry Data Warehouse For Forest Management Enterprise And Their Application

Forest resource data is the foundation for making forest management plan, designing corresponding management measures for stands and so on. In China, the forest resource data mainly come from national forest inventory, forest management inventory and forest inventory targeting a certain scientific problem. These forest inventory data contribute a lot to sustainable forest management, however, these data are not in rationally managed and are lack of integration with each other which results in low efficiency. In order to manage the data properly, realize the multi-dimensional data analysis and discover the knowledge that is on the background of the data, we introduced data warehouse and data mining to the forestry factor with the hope of its contribution to the forest management.We selected Baisha county in Hainan Province and Tropical Research Center in Guangxi autonomous region as the study areas. The data used in this study included field plot data, forest management inventory data and analytic tree data. From the field plot data, the forest manager want to get the forest management parameters such diameter distribution, vertical structure and species composition. From the analytic tree data, the forest manager want to analyze the dynamic of growth in multi-dimensional way. For the forest management inventory data, the forest resource was evaluated by the aspects such as forest resource scale, biodiversity, forest production function and forest ecological function as well as soil type distribution. The forest data warehouses were built for the three types of data, according to the demand from the data. The process of building data warehouse comprised of conceptual model, logical model and physical model. With the data warehouse, multi-dimensional analysis and data mining could be done so as to further explore the data and find the knowledge for sustainable forest management.With the data warehouse for field plot data, the light characteristics of the previous tree species was determined in Baisha county, Hainan Island; the forest development types were made for the restoration of tropical degraded secondary forests; The transferred plantation and the untransformed planation was compared and the results showed that the transferred plantation was gradually developed into uneven-aged close-to-nature forest that have multiple species and high productivity. With the data warehouse for the analytic tree, the growth dynamics for dominate trees, compressed trees and normal trees were analyzed and the results indicated that the dominate trees have the obvious advantage in dynamic growth followed by normal trees and depressed trees. The results also showed that the growth pattern between diameter growth and height growth tended to besimilar which was quite different from the volume growth. The results from data mining indicated that the average and successive volume growth was first determined by the total diameter growth and then tree type. With the data warehouse for forest management inventory, the forest resource was evaluated from different aspects according to multiple dimensions. The forest dynamics regulation was demonstrated and responding suggestion was made for the tropical research center. The decision tree algorithm in data mining showed that the first forest factor that determined the forest volume was forest category, followed by the protection level and as the level increase, the forest volume tend to increase.This dissertation first introduced the artificial intelligence that included data warehouse, multi-dimensional analysis and data mining to the forest sector. We used Microsoft Excel as the client tool of the data warehouse which contributed to the wide application of this tool, however, as it is not the software specially developed for artificial intelligence, it demand for interfere from clients. Therefore the development of a software with both broad application as well as friendly interface need to be further studied. Furthermore, the data warehouse for forest management inventory data did not include geographical information and hence the spatial data warehouse with GIS support should also further explored.