A Review on the Accuracy Analysis of Spatial Scaling Data
2012, 31 (12):
Scale is like a lens through which geographers observe and study the world. Scale dependence is one of the key studies and a challenge in geography, ecology, hydrology and meteorology in recent decades. One of the core scale issues is spatial scaling. Different scaling methods produce different resulting data with different degree of information loss. Thus, quantitative and qualitative assessment of scaling accuracy for spatial data is critical for correctly understanding and using scales. In the scaling research the basic theoretical frameworks include level theory, fractal theory, regional variable theory and first law of geography. In eco-geographical fields, scaling methods mainly are: spatial allometry, the dynamic model-based scaling method, wavelet analysis, autocorrelation analysis, fractal methods, semi-variogram method, and so forth. For the last ten years, remote sensing, GIS, and large-scale computer simulation technology have become more and more important in scaling research. As an important part of scaling process, accuracy analysis qualitatively or quantitatively evaluates the resulting data, based on the rules such as spatial composition, area precision, spatial pattern, patch characteristics, etc. Certain evaluation models and methods also involve type area statistics, landscape metrics, geostatistics, wavelets analysis, etc., with several indexes. In addition, comparing the simulation results of scaling data is also an important method in accuracy analysis. Although there are several methods of scaling accuracy analysis, few researchers have tried to develop a systematic methodology to evaluate resulting data. Therefore, it is an inevitable trend in scaling research to improve or even develop a new methodology for scaling accuracy analysis.
Related Articles |