Title
Urbanization between compactness and dispersion: designing a spatial model for measuring 2D binary settlement landscape configurations
Abstract
Measuring spatial patterns is a crucial task in spatial sciences. Multiple indicators have been developed to measure patterns in a quantitative manner. However, most comparative studies rely on relative comparisons, limiting their explanatory power to specific case studies. Motivated by advancements in earth observation providing unprecedented resolutions of settlement patterns, this paper suggests a measurement technique for spatial patterns to overcome the limits of relative comparisons. We design a model spanning a feature space based on two metrics - largest patch index and number of patches. The feature space is defined as 'dispersion index' and covers the entire spectrum of possible two-dimensional binary (settlement) patterns. The model configuration allows for an unambiguous ranking of each possible pattern with respect to spatial dispersion. As spatial resolutions of input data as well as selected areas of interest influence measurement results, we test dependencies within the model. Beyond, common other spatial metrics are selected for testing whether they allow unambiguous rankings. For scenarios, we apply the model to artificially generated patterns representing all possible configurations as well as to real-world settlement classifications differing in growth dynamics and patterns.
Year
DOI
Venue
2019
10.1080/17538947.2018.1474957
INTERNATIONAL JOURNAL OF DIGITAL EARTH
Keywords
Field
DocType
Land cover,spatial patterns,landscape metrics,dispersion index,sensitivity analysis,urban settlements,remote sensing,comparative spatial analysis
Dispersion (optics),Data mining,Feature vector,Ranking,Earth observation,Index of dispersion,Land cover,Geography,Spatial ecology,Binary number
Journal
Volume
Issue
ISSN
12.0
6.0
1753-8947
Citations 
PageRank 
References 
0
0.34
5
Authors
5
Name
Order
Citations
PageRank
Hannes Taubenböck115028.27
Michael Wurm212019.76
christian geis3203.09
S. Dech417128.32
Stefan Siedentop5122.51