Study On Scale Effect Of Scaling With Land Use Spatial Data

Scale issue of spatial data is undoubtedly one of the most fundamental aspects of any research. Researching the landscape pattern of the different land use in the different spatial scale is helpful to know the mechanisms and processes of the land-use/land-cover change at various levels. Selecting Huairou district as study area, based on the land use survey data, RS and GIS, this paper tries to discuss the scale effect of land use spatial data in the scaling from four aspects. The following conclusions can be inferred:(1) The error of the different land use types from raster data with the rule of maximum area(RMA) increases when spatial grain increases, while the error with the rule of centric cell(RCC) randomly distributes when spatial grain increases. According to RMA, the area accuracy loss of all types of land use has more correlativity with the grid size.(2) The study of scale effect impacted by elevation shows that: there are scale dependences between land use types and elevation. The holistic distributing rule can be got using the RMA. while the complexity and the character of its spatial distribution can be got using the RCC. The landscape indices, calculated from the original vector data, can indicate the law of the change of the area accuracy loss.(3) Landscape indices can indicate the grain effect with the scale changing. The scale domain and the first scale inflexion of landscape indices are various in maps with different scales. The different responds of different landscape indices show the diversity of information derived from the original data using the different methods of vector-raster conversion.(4) The land use spatial data derived from the different remote sensing images have different scale effect in some way. In the general, the regional landscape patterns, represented by the images of different spatial resolution from different sensors, are consistent, however it has great diversity in the boundary, shape and quantity of the land-use types; the landscape indices can reflect the information of various remote sensing images, and indicate the grain effect of landscape character information with scaling up methods from the different remote sensing data and different spatial resolution.