A Study On Landscape Pattern Of Shanghai Metropolitan Region

Problems such as increasing crowding, traffic congestion, and deterioration of environment have come to urban area along with process of urbanization. Now the worldwide urbanization is accelerating and bringing a great challenge to the sustainable development of the world. Quantifying landscape pattern and simulating land use change are essential to the assessment and administration of urban landscape. Using 1994 land use data set of Shanghai, this dissertation attempted to discover the characters of the landscape pattern and to develope an Artificial Neural Network (ANN) to simulate the land use changes of shanghai metropolitan area.The advances in the study of urban ecology, landscape ecology and ANN were introduced at the start. Then, a gradient analysis was conducted based on Geographical Information System (GIS) and a landscape statistical software. The result indicated the gradient of landscape metrics along the rural-urban transect at two levels: 1) Landscape fell into more pieces, shape of patches got more complex, patches became samller and varied less in their area. Meanwhile, a decline of aggregation and a climb-up in diversity were also shown. 2) Land use types were arranged as follows for theirpredominance: Agriculture-Industry/Land for construction/Others-Residence/Road/Green space. Analyzing the causes of land-use evolution in shanghai, an ANN model was developed with five generators of urbanization as inputs and four landscape metrics as outputs. Impacts of generators on landscape pattern of urban-rural fringe were assessed according to the ANN model.Appropriate percent cover of road, residence and suitable population density to keep the integrality and aggregation of landscape were discussed. Countermeasures to potential problems were provided, including: 1) reducing area ratio of artificial landscape type and encouraging environmentally compatible lands use, 2) constraining road density, residence density and population density to their appropriate ranges, 3) optimizing the ratio of land use types. Finally, author summarized the whole work of this dissertation and referred to the future works.