For “A Review on Spatial Scale Problems and Geostatistics Solutions”

This paper points out that nearly all environmental processes are scale-dependent in general because the spatial data are captured based on observations that is dependent on sampling framework with the particular scales of measurement – filtered version of reality. Moreover, the author reviews recent literature revealing scale problems in geography and holds a few discussions on the geostatistical techniques for re-scaling data and data models by introducing scales of measurement, scales of variation in spatial data and the modelling of spatial variation. Some approaches to changing the scale of measurement are suggested in the end. Adopting a conceptual framework that fits scales of the spatial variation and the scales of the spatial measurement and learning more details about the structure of the property do matter a lot when dealing with a scale-related geographic issue.

I appreciate this paper a lot for it helping me think more about scale issues in my thesis research. One of my research questions is to find if regular patterns do exist at large scale peatlands over the landscape by exploring the large-scale pattern of peatlands in one of the typical peatland landscapes–Hudson Bay Lowland. “Scale” referring as resolution and extent plays an important role when raising up my research project. The emergence of regular spatial pattern from the scale of several meters (hummocks and hollows), to tens of meters (pools, ridges and lawns) has been confirmed and the regular pattern together forming the stable individual peatland ecosystem (bogs, swamps and fens). However, there has been a lack of studies at the larger scale — hundreds of kilometers of massive peatland complexes. We inferred that the characteristics of regular patterns revealing the negative feedbacks are cross-scale transferrable which gives rise to the hypothesis that regular pattern still exist at large scale making itself a self-regulated system that is adaptive to the climate change. When it come to the implement of data collection and processing on remote sensing data, “which scale is most suitable to detect the heterogeneity between grids with a limited cost of image request” is the first step. How large the area I want to deal with(extent) and how much detail(resolution) I want in distinguish heterogeneity? If interpolation used for creating more high-resolution images, how much information is lost or mislead?

“SCALE issues” matters a lot, helps a lot, bothers a lot…

Comments are closed.