碩士 / 國立臺灣大學 / 森林學研究所 / 89 / Owing to the merits of the widely involving, securely recording, conveniently aligning, and time saving, it is more efficient for us to apply the photogrammetric techniques to the research of landuse than the traditional approaches do. In this thesis, Li-Yu-Tan reservoir watershed is our experimental area, and the used aerial photographs were taken on 1977,1985, 1993, and 2000, respectively. These photographs were first interpreted, sketched, and digitized as the landuse maps, and using the concepts of GIS, landscape ecology, the gray prediction of the gray theory, and Markov model to predict the variance of the future and monitor the change and model of the global landuse. In the analysis of landscape, it was found the largest block and area of the forests (the natural forests or established plantations) was smaller and smaller year by year, and the composition of landscape was more shattered than ever. In the characteristic of global space, the scale was complex more and more. The space arrangement is distributed in equilibrium, and there is the inclination of aggravation. It was also found the variance of the diversity index is very great in the structural index analysis of the landscape change, and the changes of these four periods are very conspicuous. Note that the Shannon index is more proper than the other ones in our experiments. In our observation, the extraneousness of the watershed continues varying, and the human disturbing has no any decrease in degree. In the analysis of the prediction model, our predictions are all appraised as "good" or above except for the landslide that does not consist with the structure and water prediction. This method is therefore proper for predicting the change of the land use, and the performance is high. By Markov model, we also predict the forestland will decrease severely (below half of the global area). It also revealed the prediction of the small time slot is more secure than that of the large one.
Identifer | oai:union.ndltd.org:TW/089NTU00360010 |
Date | January 2001 |
Creators | Kuo - Chen Teng, 鄧國禎 |
Contributors | Yeong - Kuan Chen, 陳永寬 |
Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
Language | zh-TW |
Detected Language | English |
Type | 學位論文 ; thesis |
Format | 94 |
Page generated in 0.0124 seconds