Towards a typology of urban centers and subcenters – evidence from German city regions

Abstract

This study aims to characterize urban spatial structure with respect to its multidimensionality. Using an encompassing data set of socioeconomic variables, as well as variables pertaining to the built environment, accessibility and land use, we suggest a quantitatively based urban (sub-) center typology. The fine-grained spatial scale of 1 km 2 grid cells permits a higher spatial resolution than that used in most previous studies. Our suggested typology is two-layered: a "macro layer" based on cluster analyses sheds light on urban spatial configurations. A corresponding threshold-based "micro layer" identifies distinctive types of centers and subcenters at the local level. The application of this multi-scale and multivariate typology to four German city regions indicates both the core cities' morphological dominance and the formation of subcenters with distinct profiles of regional economic importance, land use patterns and urban form. However, a substantial degree of spatial dispersion is observed because much activity is located in non-central locations.

Does the monocentric model work in a polycentric urban system? An examination of German metropolitan regions

“…This is a puzzling difference which could be the result of differences in the units of analysis: the Schmidt et al (2018) study examined administrative regions, while in this study we were interested in metropolitan regions. These differing results confirm the scalar complexity and multidimensionality of the current urban spatial structure (see also Krehl and Siedentop, 2019), and confirm Agarwal et al (2012: 440): 'polycentricity is fully consistent with monocentricity applied locally'.…”

Proposing new measures of employment deconcentration and spatial dispersion across metropolitan areas in the US

“…Another challenge is defining the base units of analysis that are then potentially aggregated into subcentres: some research has used units such as census tracts (Arribas‐Bel et al, 2015; Gordon, Richardson, & Wong, 1986; Huang & Yost‐Bremm, 2018) or block groups (Hajrasouliha & Hamidi, 2017), although some scholars argue that tracts are too large and therefore can be heterogeneous in that one part of a tract may belong in an employment subcentre whereas another part is simply residential. Therefore more recent studies have used smaller units such as grid cells (e.g., 1 km × 1 km cells) as the base units (Kane, Hipp, & Kim, 2018; Krehl & Siedentop, 2019). Certainly, smaller units are preferred in order to increase the plausibility of the assumption of homogeneity of business patterns within units, but smaller units can result in data and computational challenges.…”

“…One such approach uses spatial statistics, which enables researchers to identify subcentres in a less restrictive manner (McMillen, 2001; Redfearn, 2007). For example, one strategy uses local indicators of spatial association (LISA) statistics to define subcentres with a focus on the extent to which the number of jobs in clusters of cells/tracts are much higher than surrounding areas or spatially‐varying thresholds determined from the data of the region (Krehl & Siedentop, 2019; Riguelle, Thomas, & Verhetsel, 2007; Salvati et al, 2016). Another strategy uses non‐parametric regression to capture local peaks in the distribution of jobs across a metropolitan area (McMillen, 2001; Redfearn, 2007).…”

Polycentric regions: Proposals for a new typology and terminology

“…They result in polycentric spatial structures, both in historically PURs like the Rhine-Ruhr and the Randstad, or in (what were in the past) typically monocentric metropolitan regions like London, Paris and Tokyo. From this perspective, the transition to a polycentric system of settlement can be regarded as a universal phenomenon in post-industrial countries (Hall, 1997; Krehl and Siedentop, 2018).…”