An Analysis of the Mosaic Structu Re of Regional Landscape Using GIS and Remote Sensing
徐建华,方创琳,岳文泽
DOI: https://doi.org/10.3321/j.issn:1000-0933.2003.02.021
2003-01-01
Acta Ecologica Sinica
Abstract:Based on the application of GIS, remote sensing and quantitative methods, this paper studies the landscape mosaic structure of the Xigu district in the city of Lanzhou. First, the region’s landscape of TM image was classified into ten types of land use patterns, they are: farmland, orchard, grassland, woodland, scattered woodland, bare land, water area, urban area, rural settlement, and industrial area far from settlement. Next, the classified information together with surveyed data was served as the inputs of a geographical information system. Then, a digital elevation model (DEM) and landscape mosaic map was processed in a GIS environment, which was followed by a subdivision of the region into four categories: valley area, northern mountain area, southern mountain area, and dais area. Finally, each of those areas was analyzed in terms of diversity, dominance, fragmentation, isolation and fractal dimension, with their results given below:;(1) Due to its flat terrain, convenient transportation, short distance to rivers, the valley area has been recognized as the center of human activities. This area is revealed with high values in the diversity and fragmentation indices as well as a low dominance figure as results of its flat terrain, low altitude and convenient irrigation. Though, few landscapes show strong dominance, their diversity indices are usually low due to mountainous steep slope and high altitude. On the other hand, farmland, grassland and woodland dominate the southern mountain area, and grassland dominates the northern mountain area. Due to its composition of more farmland, orchard and rural settlement landscape patches, the dais area’s diversity and dominance indices come in between those of the valley and the mountain areas.;(2) In general, the isolation indices for man-made and semi-made landscapes, such as rural settlement, farmland, orchard, are low in the valley area and the dais area, and high in the mountain area. The grassland shows a relatively large isolation index in the valley and dais areas, a smaller index in the south mountain area, and the smallest one in the north mountain area. On the other hand, the woodland has the smallest isolation index in the south mountain area and a relatively larger one in the valley area and dais area, but does not exist in the north mountain area. For the bare land, however, its smallest isolation index is revealed with the north mountain area, the second smallest is associated with the valley area, a relatively larger one with the dais area, and the largest one with the south mountain area. ;(3) Different fractal dimensions of each landscape show the patch forming itself in the same geomorphologic area. Fractal dimensions of same landscape in different geomorphologic areas are different; the differences show the influences of different landforms to the same landscape. The fractal dimensions of the man-made and semi-made landscapes, such as rural settlement, plant, farmland and orchard in mountain areas, are higher than those in valley and dais areas, but the fractal dimensions of woodland and grassland in valley and dais areas are lower than those in mountain area. ;(4) The main factors that control and affect the pattern of landscape are climate, landform, river, railway, roads, and human activities. The important countermeasure of landscape optimum management is to convert farmland in the southern mountain area with the slope greater than 25° to woodland and grassland; the farmland in the sunny side of the southern mountain area with the slope greater than 25° should be converted to grassland, but those in the shade side be converted to woodland. The farmland with the slope greater than 25° in the north mountain area should be converted to grassland.