PROGRESS IN GEOGRAPHY ›› 2013, Vol. 32 ›› Issue (6): 932-939.doi: 10.11820/dlkxjz.2013.06.010

• Application of RS and GIS Model • Previous Articles     Next Articles

Comparison of land-use structure fractal dimension based on vector and raster data: A case study of Suzhou City

ZHANG Jing1, PU Lijie1,2, ZHU Ming1, XU Yan1, LI Peng1   

  1. 1. School of Geographic and Oceanographic Sciences, Nanjing University, Nanjing 210023, China;
    2. Key Laboratory of Land Use, Ministry of Land and Resources, Beijing 100029, China
  • Received:2012-12-01 Revised:2013-04-01 Online:2013-06-25 Published:2013-06-25

Abstract: Land use system is the product of natural and human activities, and typically as a complex nonlinear dynamical system, its structure is irregular, unstable, complex and non-linear. Fractal theory, as a new technique, has been proved to be practical for analyzing irregular and nonlinear objects. The fractal dimension, one of the most important indices in fractal theory, is often calculated from raster data, but most land-use data are stored as vector data. Conversion of vector data to grid images to calculate fractal dimension may result in inaccurate values. Accuracy of the calculation on raster data is closely related to the grain size of the grid images. Taking a case study of the 1:100000 land use data of Suzhou City in 2008, this paper first analyzed the fractal characteristics of the study area by calculating the fractal dimension, investigated the scale effects of land use fractal dimension by changing the grain size of raster data, and then established a quantitative relationship between fractal dimension and the grain size, and lastly used the math model to calculate fractal dimensions from vector data based on the raster data. The results showed that land use structure of Suzhou City followed the general rules of fractal theory, which proved that this method was suitable for the analysis of the characteristics of land use system in such a rapidly urbanizing area. Furthermore, the overall land use degree of Suzhou City was high; human activities have different effects on the different land types. For example, under the influence of human activities the structures of arable land and construction land were relatively simple, but unused land and forest-grassland are quite complex. The morphology of water was less complex than other land use types, indicating that water was more affected by human activities such as water conservation facilities and irrigation ditches. The effect of the grain size on the fractal dimension in this area showed that the fractal dimension increased with expanding grain size, and the result of statistical analysis suggested that the relationship between fractal dimension and the grain size fit with the quadratic-polynomial-model which provided a bridge between the vector data and raster date for the calculation of the fractal dimension. If the vector data were viewed as raster data of 0 m grain size, vector fractal dimension can be calculated from raster data according to the quadratic-polynomial-model. The difference between the calculated results and the fractal dimension values directly using the vector data was minimal. Thus, fractal dimension of vector land use data (the grain size is 0 m×0 m) could be deduced by this relationship within the margin of error.

Key words: data type, fractal dimension, grain size, land use, Suzhou City