| Researches on Landscape pattern change(Land-use change) is the frontier and hot issue of landscape ecology. Dynamic model can predict the change of landscape pattern more accurately. Yunnan "Three Parallel Rivers" world natural heritage site dotted with rare alpine valley landscapes, rich biodiversity, has outstanding universal value. But in recent years due to the growing population, gradually expand the scale of infrastructure construction, the impact of global climate change and other factors, the ecological risks of "Three Parallel Rivers" district increasing. Using a dynamic model to simulate the regional landscape pattern can identify the "hot spots" to provide an early warning mechanism for local government.The study takes Shangri-la County, the hinterland of Yunnan "Three Parallel Rivers" world natural heritage as research area, uses CLUE-S model coupling Markov chains model in non-space module, predicts demand for seven types of land use over the next15years; in space module, selects23driving factors and determines the probability allocation by using Binary Logistic regression. Finally the study combines the non-space and space modules, and obtain the landscape pattern allocation of the year2024.The main conclusion:(1) Using Markov chains model in non-space module leads to a more objective land-use demand prediction.(2) By comparing the regression precision ROC values, this study determines600m as the best simulation space scale.(3) In terms of the current development trend, in the year2024, forest area will reduce, forest in Gezan country, along Jinsha River, east of Balagezong will face more ecorisks; shrub area will increase and converted from forest; grass area will reduce; plough area will increase, spreading in valley dam districts; settlements areas will increase significantly in Jiantang and Jinsha River valley dam districts.(4) Dynamic simulation can provide scientific advice for various planning and space allocation. |
PDF Full Text Request