From fields to objects: A review of geographic boundary analysis

Abstract

Geographic boundary analysis is a relatively new approach unfamiliar to many spatial analysts. It is best viewed as a technique for de®ning objects ± geographic boundaries ± on spatial ®elds, and for evaluating the statistical signi®cance of characteristics of those boundary objects. This is accomplished using null spatial models representative of the spatial processes expected in the absence of boundary-generating phenomena. Close ties to the object-®eld dialectic eminently suit boundary analysis to GIS data. The majority of existing spatial methods are ®eld-based in that they describe, estimate, or predict how attributes (variables de®ning the ®eld) vary through geographic space. Such methods are appropriate for ®eld representations but not object representations. As the object-®eld paradigm gains currency in geographic information science, appropriate techniques for the statistical analysis of objects are required. The methods reviewed in this paper are a promising foundation. Geographic boundary analysis is clearly a valuable addition to the spatial statistical toolbox. This paper presents the philosophy of, and motivations for geographic boundary analysis. It de®nes commonly used statistics for quantifying boundaries and their characteristics, as well as simulation procedures for evaluating their signi®cance. We review applications of these techniques, with the objective of making this promising approach accessible to the GIS-spatial analysis community. We also describe the implementation of these methods within geographic boundary analysis software: GEM.

Unifying the spatial population dynamics and molecular evolution of epidemic rabies virus

“…The spatial clustering and geographic boundaries around clusters of sequence types (Fig. 1C) were then determined by categorical wombling (25) by using BOUNDARYSEER (26).…”

Spatial modelling-based approach to phytogeographical regionalization using grassland vegetation data

“…Does the uncertainty associated to a boundary segment/ single location exceed some threshold? Randomization procedures [9,26] and testing differences of mean values between neighbouring regions [27][28][29], as seen in this paper, are two possible methods. With regard to the geostatistical background of the proposed methodology, kriging variance [63], calculated along with kriging estimates, may be a proper measure to determine boundary uncertainty arising due to prediction errors.…”

“…To draw a border, it is assumed that the steepest gradient in EIT between adjacent grid cells indicates the most radical changes in floristic composition, typical of border locations [47]. This method is also referred to as lattice wombling [9,26,48]. The relevance of the generated division was evaluated using adjusted ANOVA accounting for spatial autocorrelation [30].…”

“…These may be point versus grid data, regularly versus irregularly distributed, abundance versus presence-absence, etc. [26]. Edge detectors, lattice wombling, scale-space filters, and moment measures [4] are appropriate tools in the case of grid data.…”

“…Randomization procedures are frequently used to determine whether the identified boundaries differ significantly from boundaries produced by randomizing the data [9,26]. Emphasis may also be placed on differences between proposed regions, in terms of mean values or other statistics.…”