Diversification in urban functions as a measure of metropolitan complexity

Abstract

Newly emerging relationships between form and function reveal the increasingly complex nature of metropolitan regions. The present study investigates spatial diversification in settlement forms and socioeconomic functions in metropolitan Attica (the administrative region including Athens, the capital of Greece), with the aim of implementing a holistic framework assessing urban complexity in contemporary cities. Taken as key components of urban complexity, morphological and functional diversity have been analysed using multi-domain indicators that describe settlement characteristics (land-use, soil sealing, building use, vertical profile of buildings, building age, construction materials) and socioeconomic functions (economic base, working classes, education levels, population age structure, composition of non-native population by citizenship, distribution of personal incomes), thus providing a comprehensive description of local-scale diversification in urban structures. A correlation analysis was used to verify the spatial coherency between individual dimensions of urban diversification. Analysis of global Moran’s spatial autocorrelation index reveals specific gradients of urban diversification that discriminate morphological attributes from socioeconomic functions. Municipalities were profiled on the basis of Pielou’s evenness indexes for each urban dimension: a factor analysis indicates latent patterns characterizing areas with high and low diversification in metropolitan functions. Urban and rural municipalities were, respectively, characterized as the most and least diversified in the study area, with peri-urban municipalities ranking in-between, evidencing a diversification gradient correlated with the distance from downtown Athens. A multidimensional analysis of the most relevant dimensions of metropolitan complexity has proved to be a promising tool for monitoring urban gradients, polycentric development and (latent) socioeconomic transformations in contemporary cities.

Keywords

Social mix economic structure spatial inequalities evenness Mediterranean city

Introduction

With metropolitan growth progressively more oriented towards residential decentralization, spatial rebalancing and functional polycentrism, an intense debate on the emerging relationships between urban form and function has consolidated in recent years (Burger and Meijers, 2012; Hirt, 2012; Kroll and Kabisch, 2012; Markusen and Schrock, 2006). While cities constitute an unsurpassed engine of growth for neighbouring regions, urban development follows distinctive paths reflecting place-specific transformations (Meijers and Burger, 2010). Diversity in form and function is intimately related with metropolitan development, where diverse and internally diversified cities develop and consolidate (Talen, 2006).

Spatial features of urban expansion, such as infilling, edge-expansion or leapfrogging, have been extensively studied in the literature to identify specific drivers (topological, socioeconomic or related to place-specific factors such as accessibility) and consequences in terms of compactness or sprawl of the resulting urban forms (Jiao et al., 2018; Mustafa et al., 2018; Sun and Zhao, 2018). This spatial pattern may have different declinations in specific contexts: urban districts with diverging ranks in the urban hierarchy may assume a comparable structure and, conversely, areas with similar size and population density may have divergent structures, hosting a different composition of economic activities and social functions (O’Donoghue, 1999). Socioeconomic functions are generally oriented along the urban gradient, as inner cities and the surrounding urban centres concentrate most of the activities, promoting a higher diversification of metropolitan structures (O’Donoghue and Townshend, 2005; Parr, 2004; Zambon et al., 2018). Moreover, functional changes in metropolitan hierarchies occur through complex trajectories of growth, since urban expansion no longer occurs through the physical addition of peripheral development units at the most basic scale, stimulating a progressively more polycentric and decentralized growth (Duranton and Puga, 2000; Ottaviano and Peri, 2006; Zambon et al., 2017).

In this regard, settlement expansion around sub-centres has altered metropolitan structures traditionally oriented along urban gradients, influencing the spatial dimension of urban diversification (Parr, 2008; Pili et al., 2017; Turok and Bailey, 2004). Exurban development has been demonstrated to create a sort of ‘metropolitan continuum’, with medium-density settlements embedded into a fragmented landscape, with spatially varying implications for processes of change at the base of urban complexity (Di Feliciantonio and Salvati, 2015; Hahs and McDonnell, 2006; Salvati, 2013, 2014; Salvati and Carlucci, 2016; Serra et al., 2014; Torrens, 2008; Vasanen, 2012). With changes in the spatial organization of metropolitan regions driven by dispersed urbanization, the analysis of metropolitan continuums based on diversification indexes is a challenging task in regional science, planning and geography (Dahly and Adair, 2007; Ellerman, 2005; Munafò et al., 2013; Salvati et al., 2016a). By identifying socioeconomic transformations along urban gradients, approaches based on a multidimensional analysis of diversification in key metropolitan functions have been demonstrated to be suitable for distinguishing inner cores and suburban areas, providing policy-oriented classifications of urban and rural districts (Colantoni et al., 2015; Comer and Greene, 2015; Cuadrado-Ciuraneta et al., 2017; Duvernoy et al., 2018).

Contributing to the inherent complexity of contemporary cities, functional diversification and urban entropy are taken as relevant traits of metropolitan growth in Europe (Batty and Longley, 1994; Cabral et al., 2013; Kazemzadeh-Zow et al., 2017; Page et al., 2001). On average, European cities reveal a greater mix and integration of socioeconomic functions than Northern American towns (e.g. Carlucci et al., 2017). In particular, European Mediterranean cities are examples of urban diversification and spatial entropy in their constituent elements, sharing common characteristics while reflecting distinctive history, socioeconomic development and settlement models (Beriatos and Gospodini, 2004; Di Feliciantonio and Salvati, 2015; Grekousis et al., 2013; Zitti et al., 2015). Among the main forces shaping urban transformations in Mediterranean Europe, economic polarization and social segregation became particularly relevant in metropolitan expansion and regional disparities (De Rosa, 2016; Malheiros, 2002; Maloutas, 2007; Rontos et al., 2016). Additionally, Mediterranean cities have often expanded in a partly unregulated manner resulting in disordered urban mosaics that reflect a fragmented economic base (Colantoni et al., 2016; Maloutas, 1993; Stanley, 2012). Long-term development processes have frequently reflected heterogeneous and self-organized metropolitan systems characterized by economic disparities, social isolation, settlement scattering and a fractal morphology (Carlucci et al., 2017). In this way, Mediterranean cities have represented key examples of socioeconomic transformations leading to morphologically dispersed and functionally discontinuous cities (Pili et al., 2017).

By contributing to a better comprehension of urban complexity, the present work implements an original framework to the analysis of complex metropolitan systems based on the use of multi-domain indicators and spatial statistics. The proposed approach sheds light on the spatial pattern of diversification in urban functions and morphological attributes, illustrating economic, social and environmental aspects at both local and regional scale in the Athens’ metropolitan region (Greece). Athens is considered a representative case of Mediterranean cities moving from a mostly informal expansion to planned development towards urban sustainability and a more balanced spatial structure. The complex (and, for some aspects, still unclear) relationship between socioeconomic functions and urban morphology in Athens, as in many other Mediterranean cities (Cuadrado-Ciuraneta et al., 2017; Di Feliciantonio and Salvati, 2015; Duvernoy et al., 2018; Morelli et al., 2014; Souliotis, 2013), justifies a multivariate analysis of urban diversification in such territorial context, under the hypothesis that metropolitan complexity reflects the increasing diversification in economic, environmental and social dimensions along urban gradients.

Methodology

Study area

The Athens Metropolitan Region (AMR) has mostly coincided with the boundaries of the administrative region of Attica, Greece. Up to 2011, the AMR was divided into 4 prefectures (Athens, Piraeus, East Attica and West Attica) and administered by 114 municipalities extending 3000 square kilometres, with a population density > 1200 inhabitants per square kilometre (2011). With the so-called Kallikratis law, a new local administrative regime was adopted in Greece since 2012, reducing the number of municipalities in the study area to nearly 60 (Rontos et al., 2016). Since statistical data referring to 2010 or 2011 were used as the main information source, the spatial structure adopted in this study refers to the administrative division (the so-called Kapodistrias spatial asset of local governance) in law at that time. The AMR includes dense urban and hyper-compact districts such as downtown Athens and Piraeus, suburban municipalities bordering the strictly urban area and rural municipalities with moderately low accessibility and a marginal role in the regional economic system (Pili et al., 2017). The urban core is traditionally associated with the ‘Greater Athens area’, a district with compact urban fabric extending 430 square kilometres and administered by 58 municipalities (‘Kapodistrias’ structure), hosting most of the population (85% in 2011) and the economic activities of the AMR (Colantoni et al., 2016).

Logical framework

Taken as basic components of metropolitan complexity, 12 dimensions of urban diversification were considered including morphological aspects (land-use, soil sealing, building use, vertical profile of buildings, building age, construction materials) and socioeconomic functions (economic base, working classes, education levels, population age structure, composition of non-native population by citizenship, distribution of personal incomes). To allow a coherent analysis of local-scale patterns of urban diversification, municipalities were adopted as the elementary domain of analysis in this study. Easily derived from statistical data sources at the municipal scale, such dimensions are intended to provide a comprehensive description of metropolitan complexity in the Mediterranean region and, possibly, in other socioeconomic contexts worldwide.

Morphological attributes

Six variables organized in different classes were computed at the spatial scale of municipalities: (1) land-use (hereafter ‘land’), (2) soil sealing profile (‘soil’), (3) building use (‘use’), (4) vertical profile of buildings (‘vert’), (5) building age (‘buil’) and (6) construction materials (‘mat’). Landscape composition (percentage class area of total municipal area) was derived from a high-resolution land-use map (1: 10,000 scale) referring to 2012 and disseminated through the European Urban Atlas initiative on behalf of the Global Monitoring for Environment and Security framework adopting a nomenclature system composed of 20 classes (built-up areas: code 1, cropland: code 2 and forests: code 3).

A soil sealing profile for each municipality was derived from a 100 metres grid land imperviousness map referring to 2012 and disseminated by the Land Copernicus initiative in collaboration with the European Environment Agency (2011). This considers all pavement structures (roads, sidewalks, driveways and parking lots) that are covered by impenetrable materials such as asphalt, concrete, brick, stone and rooftops with a continuous degree of land imperviousness ranging from 0 to 100%. Classification accuracy of built-up and non-built-up areas was higher than 85% per hectare, with omission and commission errors kept below 15% (European Environment Agency, 2011). A soil sealing profile for each municipality of the study area was determined by computing the percentage area of the total municipal area of 22 classes with different land imperviousness intensity (0, 1–5, 6–10, …, 91–95, 96–99, 100%). For both land-use and soil sealing variables, calculations were implemented using the ‘Tabulate areas’ tool provided with the ArcGIS software (ESRI Inc., Redwoods, USA) after the overlap between the land imperviousness map and a municipal boundary shapefile (Colantoni et al., 2015).

The other four variables assessing settlement characteristics were derived from the national census of buildings carried out by the Greek Statistical Authority (ELSTAT), aggregating elementary data (2011) at municipal scale (Zitti et al., 2015). Building types were classified using 18 categories distinguishing residential from industrial, commercial and service use, and the percentage share of each building class in the total municipal building stock was calculated accordingly. Building age was determined using 10 classes assessing the construction time period and computing the percentage share of buildings by age class of the total building stock. The vertical profile of buildings was assessed considering six classes that distinguish buildings according to their number of floors and calculating the percentage share of buildings by height class from the total building stock. Finally, buildings were classified according to the dominant construction material using six material types and calculating the percentage share of buildings by construction material from the total building stock.

Socioeconomic functions

Six variables organized in different classes were computed at municipal scale: (1) economic base (hereafter ‘econ’), (2) working class composition (‘work’), (3) educational level of the active population (‘educ’), (4) population age structure (‘pop’), (5) composition of non-native resident population by citizenship (‘fore’) and (6) distribution of personal incomes (‘inc’). The economic base was assessed considering the number of enterprises in the national business register (referring to 2010) based on a classification of activities composed of 15 sectors and compatible with the NACE-Rev2 nomenclature, calculating the percentage share of enterprises by sector from the total registered enterprises at municipal scale. Working class composition was studied, which refers to a classification of workers composed of 24 categories that distinguish different professional positions according to the national census of population and households held in 2011 in Greece. Based on these data, the percentage share of workers by professional position from the total workers was calculated for each municipality of the study area (Di Feliciantonio and Salvati, 2015). The education of the active population was assessed by referring to a nomenclature with 13 education levels adopted by the Greek population census (2011); the percentage share of the active population by education level from the total active population was calculated for each municipality. The population structure was investigated using census data that aggregate resident population into seven age classes and by computing the percentage share of the population by age class from the total resident population. The composition of the non-native population by citizenship was analysed by aggregating resident population into seven classes (including the most frequent nationalities in the study area and a residual class) based on census data (2011) and computing the percentage share of each class from the total resident population. Finally, the distribution of personal income across the population was investigated, considering elementary data from individual tax declarations (provided by the Hellenic Ministry of Finance and referring to 2014), aggregating resident inhabitants into 10 income classes.

Indexes of urban diversification

Twelve indicators of urban diversification, one for each variable described in ‘Logical framework’ section, were calculated using Pielou’s evenness index (J). This index has been extensively used to identify diversity-based functions, providing a comprehensive reading of pertinent aspects of urban growth and change (see references in Zambon et al., 2017). Based on a Shannon diversity function (the so-called H′ Shannon index grounded on information theory), Pielou’s J index is an entropy index standardized to the level of diversification in a given spatial domain (Salvati et al., 2016a), quantifying local-scale diversity and heterogeneity for each variable selected in this study. Ranging from 0 (complete homogeneity) to 1 (the highest heterogeneity according to the level of local-scale diversification), Pielou’s J index was calculated as follows

J = H' / H_{max}

where H′ is the Shannon diversity index calculated as

H' = \sum_{i = 1} n p_{i} \times ln (p_{i})

where p_i is the proportion of observations falling in the ith class from the total number of observations for each urban dimension. H_max is the logarithm of the number of classes with at least one observation.

Statistical analysis

A multi-step approach was implemented in this study with the following objectives: (i) to assess the spatial structure of each dimension of urban diversification, (ii) to identify pair-wise relationships between individual dimensions of urban diversification, (iii) to characterize the latent interplay between morphological and functional dimensions, (iv) to cluster municipalities on the basis of their diversification profile and finally (v) to relate the identified diversification profiles to contextual attributes associated with the metropolitan gradient. These objectives were addressed using an exploratory approach that integrates spatial statistics (global Moran’s I spatial autocorrelation index) with descriptive analysis/mapping (objective i), parametric and non-parametric correlations (objective ii), a principal component analysis (PCA) outlining latent relationships among variables (objective iii), hierarchical clustering (objective iv) and a final correlation analysis (objective v) to capture the latent relationship between local-scale diversification in urban functions and background variables, characterizing the urban gradients.

Spatial autocorrelation analysis

Under the assumption that Moran’s indexes of spatial autocorrelation are a powerful tool in urban studies (see, among others, Riguelle et al., 2007; Salvati et al., 2016b; Thomas et al., 2012; Tsai, 2005), a global Moran’s spatial autocorrelation index (z-score) computed on eight bandwidths (from 5 to 40 kilometres) was run separately for each dimension of urban diversification using municipalities as the elementary domain of analysis, contributing to a comparative investigation of the spatial structure of the study area.

Inferential analysis

A pair-wise correlation analysis using both parametric (Pearson linear product moment) and non-parametric (Spearman rank) coefficients was used to verify the spatial coherency between individual dimensions of urban diversification, testing for significance at p < 0.05 after Bonferroni’s correlation for multiple correlations. A comparison of parametric and non-parametric techniques based on the absolute difference between correlation coefficients (i) may indicate linear (or more complex) forms of the relationship between the studied variables and (ii) allows a coherent analysis of variables displaying a relevant deviation from normality. Coherency in sign and intensity of both parametric and non-parametric coefficients reflects a linear correlation between dimensions of urban diversification. Divergence in sign and intensity of correlation coefficients indicates more complex relationships among urban dimensions.

Multivariate exploratory analysis

A PCA was run on a data matrix composed of 12 columns (Pielou’s J index for each urban dimension) and 114 rows (municipalities) with the aim to extract a few (independent) components of urban diversification and to identify the related spatial patterns. A hierarchical clustering using Euclidean distances and Ward’s agglomeration rule was carried out separately on urban dimensions and municipalities with the aim of visualizing similarities in the spatial distribution of individual components of urban diversification and homogeneous groups of municipalities based on spatially consistent patterns of urban diversification. A non-parametric (Spearman) correlation analysis was finally run to identify pair-wise relationships between individual dimensions of urban diversification and contextual indicators (density, elevation, proximity to the sea coast, distance from four urban centres – Athens, Piraeus, Maroussi, Markopoulo Messoghias – and municipal surface area), testing for significance at p < 0.05 after Bonferroni’s correction for multiple comparisons.

Results

Assessing dimensions of urban diversification

The statistical distribution of individual Pielou’s J evenness indexes is presented in Table S1 (Supplemental material). On average, the highest evenness indexes (>0.7) were found for economic structure, working class structure, educational levels, vertical profile of buildings, building construction age, population age structure and personal income distribution. Municipalities belonging to the Greater Athens area emerged for a high diversification in specific morphological attributes (land-use, soil sealing profile, vertical profile of buildings and, to a lesser extent, building use) and socioeconomic dimensions (composition of non-native population by citizenship, distribution of personal income). A reverse diversification pattern was observed for specific morphological dimensions (building age, construction materials) and socioeconomic functions (educational levels). Heterogeneous spatial patterns – with no specific divides between urban and rural districts – were observed for diversification in economic structure, working class composition and population age structure, evidencing few sub-centres with medium-high socioeconomic diversification.

Spatial analysis

Using eight bandwidths ranging from 5 to 40 kilometres, global Moran’s coefficients of spatial autocorrelation showed inherent differences for the dimensions considered in this study (Table 1). Spatially correlated structures (significance tested at p < 0.05) were observed at all bandwidths for 6 out of 12 dimensions (5 morphological attributes and 1 socioeconomic function). Two dimensions (building age and distribution of personal incomes) had a spatially correlated structure for seven out of eight bandwidths. Two dimensions (diversification in working class composition and in educational levels) were spatially correlated at bandwidths > 15 kilometres. Finally, two dimensions (diversification in economic base and in population age structure) have no significant autocorrelation coefficients, indicating spatial heterogeneity with no specific patterns either at local or regional scale.

Table 1.

Global Moran’s spatial autocorrelation index (z-score) by bandwidth (km); bold indicates a significant correlation at p < 0.05 for local-scale spatial patterns (smaller bandwidths) or regional-scale spatial patterns (greater bandwidths).

Variable	5	10	15	20	25	30	35	40
Land-use	8.3	13.6	16.9	17.8	17.6	17.0	16.2	15.4
Soil sealing	11.1	18.1	22.0	23.2	23.0	22.0	21.0	20.1
Building use	5.8	7.7	8.6	9.0	8.8	8.5	8.4	8.0
Vertical profile	10.2	17.3	21.2	21.8	21.4	20.5	19.5	18.7
Building age	1.7	2.1	2.8	2.7	2.8	2.9	3.2	3.2
Construction materials	5.5	8.3	10.3	10.6	10.0	9.8	9.6	9.2
Economic base	0.4	0.3	1.0	1.2	1.4	1.5	1.6	1.6
Working classes	0.6	1.0	1.8	2.0	2.2	2.2	2.1	2.1
Educational levels	0.7	1.0	1.9	2.1	2.4	2.4	2.3	2.3
Population age structure	0.5	0.8	1.6	1.8	1.8	1.9	1.9	1.9
Composition of non-native residents	6.0	8.8	11.2	11.7	11.9	11.6	11.2	10.9
Distribution of personal incomes	1.2	2.2	3.4	3.6	3.7	3.7	3.6	3.6

Correlation analysis

Results of a pair-wise correlation analysis based on parametric (Pearson) and non-parametric (Spearman) coefficients are presented in Table 2. Significant pair-wise coefficients mainly reflect linear relationships among urban dimensions (similar correlation coefficients for both Pearson and Spearman analysis); more specifically, diversification in land-use and soil sealing profiles have similar correlation profiles, being negatively associated with diversification in building materials (and in population age structure, only for soil sealing profile) and positively associated with diversification in (i) vertical use of buildings, (ii) distribution of personal incomes and (iii) composition of non-native population by citizenship. Diversification of local-scale economic base was correlated positively with working class composition and building age, possibly indicating a higher economic diversification in consolidated urban settlements than in more recently developed districts. Diversity in working class composition increased linearly with diversity in personal incomes. Diversification in educational levels was correlated positively with diversification in building age and in personal income distribution within the population residing in each municipality. Diversification in the vertical profile of buildings was negatively correlated with diversification in building materials and positively correlated with diversification in personal income distribution and in the composition of the non-native population by citizenship. Finally, diversification in building materials was negatively correlated with diversification in the composition of personal income and non-native population by citizenship.

Table 2.

Correlation matrix among urban dimensions in the study area; Pearson linear product moment coefficients and Spearman non-parametric rank coefficients were used to assess, respectively, linear and non-linear pair-wise correlation between variables; bold indicates significant coefficients at p < 0.05 after Bonferroni’s correlation for multiple correlations.

Variable	Land	Soil	Econ	Work	Educ	Use	Vert	Buil	Mat	Pop	Fore
Pearson
Soil	0.91
Econ	−0.27	−0.19
Work	0.06	0.07	0.43
Educ	0.11	0.15	0.21	0.29
Use	0.01	0.15	0.07	−0.01	0.19
Vert	0.61	0.69	−0.04	0.28	0.13	0.25
Buil	−0.04	−0.01	0.38	0.10	0.33	0.08	−0.03
Mat	−0.49	−0.58	0.03	−0.25	−0.07	0.13	−0.68	0.16
Pop	−0.29	−0.32	0.09	0.24	−0.31	−0.15	−0.14	−0.15	0.21
Fore	0.50	0.53	0.06	0.26	0.12	−0.08	0.62	0.05	−0.65	−0.06
Inc	0.41	0.46	0.16	0.44	0.38	0.02	0.53	0.14	−0.46	−0.02	0.50
Spearman
Soil	0.88
Econ	−0.17	−0.11
Work	0.03	0.08	0.53
Educ	0.10	0.13	0.25	0.40
Use	0.00	0.16	0.04	0.05	0.17
Vert	0.54	0.63	−0.01	0.27	0.08	0.25
Buil	−0.03	−0.02	0.44	0.18	0.35	0.05	−0.13
Mat	−0.46	−0.56	−0.06	−0.26	−0.08	0.07	−0.68	0.19
Pop	−0.25	−0.31	0.06	0.11	−0.33	−0.18	−0.15	−0.15	0.29
Fore	0.46	0.49	0.06	0.34	0.12	−0.07	0.63	−0.01	−0.66	−0.07
Inc	0.47	0.52	0.29	0.43	0.37	0.05	0.55	0.20	−0.56	−0.13	0.57

For some significant pair-wise correlations, the absolute difference between the Pearson and Spearman coefficients was particularly high, evidencing a non-linear relationship between (i) educational levels and working class composition and (ii) distribution of personal income and diversification in construction materials. A non-linear correlation was also observed between diversification in the composition of the non-native population by citizenship and working class composition (a significant Spearman coefficient together with a non-significant Pearson coefficient).

Summarizing spatial patterns of urban diversification

PCA extracted four relevant components (eigenvalue > 1), explaining 72.9% of total matrix variance. A PCA biplot illustrates the joint distribution of urban dimensions and municipalities in the AMRs based on component loadings and scores on the selected components (Figure S1 in Supplemental material). Component 1 (34.3%) identifies a traditional urban gradient opposing spatial patterns of diversification in building materials (negative loadings) to diversification in land-use, soil sealing profile, the vertical profile of buildings, the composition of the non-native population by citizenship and the distribution of personal income (positive loadings). The highest scores on component 1 indicate urban municipalities belonging to the Greater Athens area, while the lowest scores characterized the remaining municipalities of the study area (Figure 1). Component 2 (17.2%) illustrates gradient distinguishing municipalities with high and low diversification in economic base, building age, educational levels and composition of working classes. Apart from a few exceptions, municipalities north, east and south of Athens received medium-high scores, while municipalities west of Athens had medium-low scores. Taken together, component 2 represents a gradient distinguishing wealth districts from economically disadvantaged areas, resembling the traditional east-to-west social differentiation in Athens. Component 3 (13.0%) identifies a gradient discriminating the spatial pattern of three socio-demographic dimensions (positive loadings: educational levels; negative loadings: working class composition and population age structure). Western Attica and specific districts of eastern Attica (experiencing dispersed urban expansion in the last decades) received the highest scores on this component. Component 4 (8.4%) was mainly associated with diversification in the vertical profile of buildings, with the highest scores associated with municipalities belonging to the Greater Athens area.

Figure 1.

Spatial distribution of principal component scores by axis. (a) Component 1, (b) component 2, (c) component 3 and (d) component 4.

Cluster analysis

A hierarchical clustering of municipalities based on Pielou’s J index of diversification in 12 dimensions identified two groups of municipalities: urban districts and peri-urban/rural areas (Figure 2). Both clusters are, in turn, separated into sub-clusters. The two sub-clusters referring to the ‘urban’ group distinguish municipalities with compact and dense settlements (left cluster), mostly situated in the western part of the Greater Athens area (including Athens and Piraeus) from municipalities with less dense and relatively more sparse settlements (right cluster) especially located in the eastern part of the Greater Athens area. The partition into two rural sub-clusters is also rather clear, distinguishing a left group of low-density, rural municipalities exclusively located in eastern and northern Attica and a right group including municipalities with medium-low population densities mainly (but not exclusively) from eastern Attica. Cluster analysis (Figure S2 in the Supplemental material) also grouped morphological attributes (e.g. land-use, soil sealing profile, vertical profile of buildings, settlement characteristics) and socioeconomic functions (e.g. working class composition, population age structure, distribution of personal incomes) on the basis of similarities in their spatial distribution over the study area.

Figure 2.

Hierarchical clustering of municipalities in the study area based on evenness indexes.

Local-scale diversification in urban dimensions and metropolitan gradients

Significant pair-wise Spearman correlation coefficients between each dimension of urban diversification and selected territorial attributes are reported in Table S2 in the Supplemental material. Eight dimensions (all morphological attributes and two socioeconomic functions) were significantly correlated with variables oriented along the urban gradient (population density or distance from the most relevant urban nodes in the Athens region) or indirectly associated with them (municipal area). Diversification in working class composition decreased with the distance from Markopoulo, a central place in the Messoghia district. Diversification in the local-scale economic base, education levels and population age structure were not correlated with any contextual variable, indicating a spatial structure more influenced by place-specific factors than regional gradients.

Discussion

Considering the ties between settlement morphology and socioeconomic attributes, metropolitan complexity depends on a strong diversification in urban functions (Van Oort et al., 2015). In this regard, our study introduces a multivariate concept of metropolitan complexity based on a multi-domain analysis of indicators assessing morphological and socioeconomic dimensions of urban diversity. To verify the spatial coherency of diversified urban functions, a multi-criteria spatial analysis was adopted, controlling for place-specific background variables. Indicator systems have often been proposed as more reliable and stable tools compared to individual variables (Salvati et al., 2016a). Since relevant and updated variables characterizing metropolitan continuums or discriminating urban from rural districts are still limited at both continental and country scale (e.g. Hahs and McDonnell, 2006; Hoyler et al., 2008; Youn et al., 2016), our study proposes spatial diversification in urban functions as a new indicator of metropolitan complexity. This approach may contribute to multi-domain information systems classifying urban, suburban and rural typologies of settlements in various regional contexts (Comer and Greene, 2015; Venerandi et al., 2017; Zambon et al., 2017).

The strength of the paper relies on the use of a relatively big data set of quantitative indicators made available at the municipal level together with a fine-grained spatial analysis of multi-faceted patterns of diversification in relevant urban dimensions, as reflected in the morphology of a Southern European metropolitan region (Rontos et al., 2016). However, this data-driven research approach could also be viewed as a weakness of the study, since it aimed to explore and detect relevant dimensions of metropolitan complexity and not to derive causal relations grounded on economic theory. In this regard, we think that the pros more than counterbalance the cons. Looking at the spatial diversification in settlement forms and socioeconomic functions through the lens of econometric modelling of the causal relationship between key socioeconomic drivers can produce partial (and sometimes misleading) results when applied to complex urban systems characterized by marked morphological and functional variety reflected in the feedback among landscape, urban form, social structures, economic activities and planning (Salvati et al., 2016a). On the contrary, exploratory data analysis can highlight, better than other quantitative methods, the relationship between morphological and functional variables, so strongly dependent on the interplay between socioeconomic and territorial indicators (Salvati and Serra, 2016; Stanley, 2012; Talen, 2006). Moreover, the approach proposed here is flexible enough to take into account place-specific components that characterize urban dimensions in different territorial contexts (Zambon et al., 2017), thus allowing its meaningful application to other metropolitan realities in (and possibly outside) the European region.

The results of our study have identified urban and rural municipalities in Athens, respectively, as the most and least diversified, with suburban areas ranking in between (‘Assessing dimensions of urban diversification’ section). The empirical results also demonstrate that local-scale morphological diversity is correlated with socioeconomic diversity in a non-linear manner (‘Correlation analysis’ section). A correlation analysis with external variables oriented along the urban gradient (‘Local-scale diversification in urban dimensions and metropolitan gradients’ section) indicates a different metropolitan structure characterizing each dimension of urban complexity. Territorial disparities result from the interaction among mixed social classes and self-governing micro-entities living in fragmented and heterogeneous ‘island’ settlements, outlining class segregation and economic inequalities at the spatial scale of neighbourhoods (Maloutas, 1993). In this regard, patterns of urban diversification in Athens reflect a divided spatial model, making this city a thought-provoking case – with functional and morphological traits that exemplify the recent expansion of Mediterranean cities (Di Feliciantonio and Salvati, 2015; Maloutas, 2007; Souliotis, 2013).

Under the assumption that urban diversification is intimately connected with the relationship between structure and functions in cities, the present study sheds light on the present (and, possibly, future) challenges concerning the interplay between form and function (Carlucci et al., 2017), as a contribution to a more comprehensive understanding of spatial patterns of diversification in metropolitan regions of Southern Europe (Cuadrado-Ciuraneta et al., 2017; Duvernoy et al., 2018; Grekousis et al., 2013). For instance, economic polarization and social segregation should be considered and managed together, distinguishing local-scale heterogeneity from more general patterns characterizing regional-scale configurations (Pili et al., 2017; Rontos et al., 2016; Souliotis, 2013). In this regard, employment diversification has traditionally benefited from local economies (Ejermo, 2005; Ellerman, 2005; O’Donoghue, 1999). Employment has shifted from agriculture and industry towards a service- and information-based economy, defining new spatial patterns of specialization and diversification associated with inter-sectoral employment changes (O’Donoghue and Townshend, 2005). Measuring the intrinsic diversity of economic activities in cities sheds more light on local development patterns and suggests optimal (or suboptimal) structures that promote economic expansion and more integrated urban systems (Youn et al., 2016).

A better understanding of socioeconomic transformations contributing to urban diversification requires (i) a more integrated analysis of major challenges including social disparities, demographic transformations, economic resilience and environmental change (Desrochers, 2001; Duranton and Puga, 2000; Wood and Dovey, 2015; Youn et al., 2016) and (ii) suitable planning measures recognizing that local-scale spatial structures may impact economic growth, social equity and sustainable urban development (Beriatos and Gospodini, 2004; Colantoni et al., 2016; Pili et al., 2017). New lifestyles, gentrification, economic polarization and industrial proliferation are some of the main themes that regional policies and spatial planning should tackle (Smets and Salman, 2008), considering that housing quality, residential satisfaction, neighbourhood-based social interactions and residential attitudes towards social mix are negatively influenced by uncertain policy goals at the pertinent spatial level (Kleinhans, 2004).

Although diversification patterns are increasingly heterogeneous and difficult to dealt with integrally (e.g. Hirt, 2016; Frenken K, Van Oort F and Verburg T, 2007), links among urban form and function can be managed, adapting land zoning targets to newly emerging metropolitan functions (Hirt, 2012; Talen, 2006; Vandermotten et al., 2008). While national policies targeting urban development have sometimes encouraged social mix and housing diversification in neighbourhoods under renewal and/or regeneration (O’Donoghue, 1999), a contemporary theory of difference should reinforce traditional interpretations of urban diversity and their connection with different political and economic contexts (Desrochers, 2001; Duranton and Puga, 2000; Ellerman, 2005; Tochterman, 2012). Planning strategies adapting to an increased diversification of urban functions and measures promoting local development and metropolitan identity are particularly suitable for managing urban complexity.

Conclusions

Diversification in urban functions may effectively reflect a new metropolitan geography better than more classical indicators of urban growth. In this regard, evenness indexes allowed discrimination of metropolitan contexts with different settlement forms and socioeconomic dynamics, contributing to the classification of urban, suburban and rural areas. A multivariate spatial analysis of a large set of diversification indexes proved to be a reliable tool to investigate local-scale urban complexity, identifying the latent relationship between the constituent dimensions, and may integrate decision support systems for diachronic analysis of urban growth. Future research should be aimed at verifying the informative potential of the holistic approach proposed here to other European metropolitan contexts experiencing heterogeneous urbanization patterns, consolidating the ‘core’ set of relevant indicators with ancillary, spatially explicit information. The increasing impact of recent urbanization processes on metropolitan regions definitely requires a diachronic evaluation of the intimate relationship between land-use heterogeneity and socioeconomic diversification, considering together results from spatial statistics, mapping, inferential techniques and multivariate models. Future steps could benefit from a comparative approach of case studies, distinguishing cities (and the related metropolitan areas) according to basic settlement and social characteristics (e.g. mono-centric versus polycentric patterns of growth, compact versus dispersed morphologies, disadvantaged versus affluent urban districts). Innovative approaches that investigate linear and non-linear relationships among urban dimensions should be refined with a coherent analysis providing planners, policy-makers, economic actors and other stakeholders with advanced knowledge to manage the increasing complexity of contemporary cities.

Supplemental Material

Supplemental material for Diversification in urban functions as a measure of metropolitan complexity

Supplemental Material for Diversification in urban functions as a measure of metropolitan complexity by Margherita Carlucci, Ilaria Zambon and Luca Salvati in Environment and Planning B: Urban Analytics and City Science

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Ministry of Education, Youth and Sports of Czech Republic within the National Sustainability Program I (NPU I), Grant Number LO1415.

Supplemental material

Supplemental material for this article is available online.

Margherita Carlucci is a Full Professor of Economic Statistics at La Sapienza University of Rome, Italy, Department of Social and Economic Sciences.

Ilaria Zambon is a PhD candidate in Rural planning at La Tuscia University of Viterbo, Italy.

Luca Salvati is a researcher at the Italian Council for Agricultural Research and Economics (CREA), Arezzo.

References

Batty

Longley

(1994) Fractal Cities: A Geometry of Form and Function. London: Academic Press.

Beriatos

Gospodini

(2004

) Glocalising urban landscapes: Athens and the 2004 Olympics.

Cities 21: 187–202.

Burger

Meijers

(2012) Form follows function? Linking morphological and functional polycentricity. Urban Studies 49(5): 1127–1149.

Cabral

Augusto

Tewolde

, et al. (2013) Entropy in urban systems. Entropy 15(12): 5223–5236.

Carlucci

Grigoriadis

Rontos

, et al. (2017) Revisiting a hegemonic concept: Long-term ‘Mediterranean urbanization’ in between city re-polarization and metropolitan decline. Applied Spatial Analysis and Policy 10(3): 347–362.

Colantoni

Grigoriadis

Sateriano

, et al. (2016) Cities as selective land predators? A lesson on urban growth, deregulated planning and sprawl containment. Science of the Total Environment 545–546: 329–339.

Colantoni

Mavrakis

Sorgi

, et al. (2015) Towards a ‘polycentric’ landscape? Reconnecting fragments into an integrated network of coastal forests in Rome. Rendiconti Lincei 26: 615–624.

Comer

Greene

(2015) The development and application of a land-use diversity index for Oklahoma City, OK. Applied Geography 60: 46–57.

Cuadrado-Ciuraneta

Durà-Guimerà

Salvati

(2017) Not only tourism: Unravelling suburbanization, second-home expansion and “rural” sprawl in Catalonia, Spain 2017. Urban Geography 38(1): 66–89.

10.

Dahly

Adair

(2007) Quantifying the urban environment: A scale measure of urbanicity outperforms the urban–rural dichotomy. Social Science & Medicine 64(7): 1407–1419.

11.

De Rosa

Salvati

(2016) Beyond a ‘side street story’? Naples from spontaneous centrality to entropic polycentricism, towards a ‘crisis city’. Cities 51: 74–83.

12.

Desrochers

(2001) Local diversity, human creativity and technological innovation. Growth and Change 32: 369–394.

13.

Di Feliciantonio

Salvati

(2015) ‘Southern’ alternatives of urban diffusion: Investigating settlement characteristics and socio-economic patterns in three Mediterranean regions. Tijdschrift voor Economische en Sociale Geografie 106(4): 453–470.

14.

Duranton

Puga

(2000) Diversity and specialisation in cities: Why, where and when does it matter? Urban Studies 37: 533–555.

15.

Duvernoy

Zambon

Sateriano

, et al. (2018) Pictures from the other side of the fringe: Urban growth and peri-urban agriculture in a post-industrial city (Toulouse, France). Journal of Rural Studies 57: 25–35.

16.

Ejermo

(2005) Technological diversity and Jacobs’ externality hypothesis revisited. Growth and Change 36(2): 167–195.

17.

Ellerman

(2005) How do we grow?: Jane Jacobs on diversification and specialization. Challenge 48(3): 50–83.

18.

European Environment Agency (2011) Mapping Guide for a European Urban Atlas. Version 1.1. Copenhagen: EEA.

19.

Frenken

Van Oort

Verburg

(2007) Related variety, unrelated variety and regional economic growth. Regional Studies 41: 685–697.

20.

Grekousis

Manetos

Photis

(2013) Modeling urban evolution using neural networks, fuzzy logic and GIS: The case of the Athens metropolitan area. Cities 30: 193–203.

21.

Hahs

McDonnell

(2006) Selecting independent measures to quantify Melbourne’s urban–rural gradient. Landscape and Urban Planning 78(4): 435–448.

22.

Hirt

(2012) Mixed use by default: how the Europeans (don’t) zone. Journal of Planning Literature 27(4): 375–393.

23.

Hirt

(2016) Rooting out mixed use: Revisiting the original rationales. Land Use Policy 50: 134–147.

24.

Hoyler

Kloosterman

Sokol

(2008) Polycentric puzzles: Emerging megacity regions seen through the lens of advanced producer services. Regional Studies 42: 1055–1064.

25.

Jiao

Liu

, et al. (2018) Proximity expansion index: An improved approach to characterize evolution process of urban expansion. Computer, Environment and Urban Systems 70: 102–112.

26.

Kazemzadeh-Zow

Zanganeh Shahraki

Salvati

, et al. (2017) A spatial zoning approach to calibrate and validate urban growth models. International Journal of Geographical Information Science 31(4): 763–782.

27.

Kleinhans

(2004) Social implications of housing diversification in urban renewal: A review of recent literature. Journal of Housing and the Built Environment 19(4): 367–390.

28.

Kroll

Kabisch

(2012) The relation of diverging urban growth processes and demographic change along an urban–rural gradient. Population, Space and Place 18(3): 260–276.

29.

Malheiros

(2002) Ethni-cities: Residential patterns in the northern European and Mediterranean metropolises – Implications for policy design. International Journal of Population Geography 8(2): 107–134.

30.

Maloutas

(1993) Social segregation in Athens. Antipode 25(3): 223–239.

31.

Maloutas

(2007) Socioeconomic classification models and contextual difference: The European Socioeconomic Classes (ESeC) from a South European Angle. South European Society and Politics 12(4): 443–460.

32.

Markusen

Schrock

(2006) The distinctive city: Divergent patterns in growth, hierarchy and specialisation. Urban Studies 43(8): 1301–1323.

33.

Meijers

Burger

(2010) Spatial structure and productivity in US metropolitan areas, Environment and Planning A 42: 1383–1402.

34.

Morelli

Rontos

Salvati

(2014) Between suburbanisation and re-urbanisation: Revisiting the urban life cycle in a Mediterranean compact city. Urban Research and Practice 7(1): 74–88.

35.

Munafò

Salvati

Zitti

(2013) Estimating soil sealing rate at national level – Italy as a case study. Ecological Indicators 26: 137–140.

36.

Mustafa

Van Rompaey

Cools

, et al. (2018) Addressing the determinants of built-up expansion and densification processes at the regional scale. Urban Studies 55(15): 3279–3298.

37.

O’Donoghue

(1999) The relationship between diversification and growth: Some evidence from the British urban system 1978 to 1991. International Journal of Urban and Regional Research 23(3): 549–566.

38.

O’Donoghue

Townshend

(2005) Diversification, specialization, convergence and divergence of sectoral employment structures in the British urban system, 1991–2001. Regional Studies 39: 585–601.

39.

Ottaviano

Peri

(2006) The economic value of cultural diversity: Evidence from US cities. Journal of Economic Geography 6(1): 9–44.

40.

Page

Parisel

Pumain

, et al. (2001) Knowledge-based simulation of settlement systems. Computers, Environment and Urban Systems 25(2): 167–193.

41.

Parr

(2004) The polycentric urban region: a closer inspection. Regional Studies 38: 231–240.

42.

Parr

(2008) Cities and regions: Problems and potentials. Environment and Planning A 40: 3009–3026.

43.

Pili

Grigoriadis

Carlucci

, et al. (2017) Towards sustainable growth? A multi-criteria assessment of (changing) urban forms. Ecological Indicators 76: 71–80.

44.

Riguelle

Thomas

Verhetsel

(2007) Measuring urban polycentrism: A European case study and its implications. Journal of Economic Geography 7(2): 193–215

45.

Rontos

Grigoriadis

Sateriano

, et al. (2016) Lost in protest, found in segregation: Divided cities in the light of the 2015 “Οχι” referendum in Greece. City, Culture and Society 7(3): 139–148.

46.

Salvati

(2013) Monitoring high-quality soil consumption driven by urban pressure in a growing city (Rome, Italy). Cities 31: 349–356.

47.

Salvati

(2014) Agro-forest landscape and the ‘fringe’ city: A multivariate assessment of land-use changes in a sprawling region and implications for planning. Science of the Total Environment 490: 715–723.

48.

Salvati

Sateriano

Saradakou

, et al. (2016a) ‘Land-use mixité’: Evaluating urban hierarchy and the urban-to-rural gradient with an evenness-based approach. Ecological Indicators 70: 35–42.

49.

Salvati

Serra

(2016) Estimating rapidity of change in complex urban systems: A multidimensional, local‐scale approach. Geographical Analysis 48(2): 132–156.

50.

Salvati

Venanzoni

Serra

, et al. (2016b) Scattered or polycentric? Untangling urban growth in three southern European metropolitan regions through exploratory spatial data analysis. Annals of Regional Science 57(1): 1–29.

51.

Serra

Vera

Tulla

, et al. (2014) Beyond urban–rural dichotomy: Exploring socioeconomic and land-use processes of change in Spain (1991–2011). Applied Geography 55: 71–81.

52.

Smets

Salman

(2008) Countering urban segregation: Theoretical and policy innovations from around the globe. Urban Studies 45(7): 1307–1332.

53.

Souliotis

(2013) Cultural economy, sovereign debt crisis and the importance of local contexts: The case of Athens. Cities 33: 61–68.

54.

Stanley

(2012) An historical perspective on the viability of urban diversity: Lessons from socio-spatial identity construction in nineteenth-century Algiers and Cape Town. Journal of Urbanism 5(1): 67–86.

55.

Sun

Zhao

(2018) Spatiotemporal dynamics of urban expansion in 13 cities across the Jing-Jin-Ji urban agglomeration from 1978 to 2015. Ecological Indicators 87: 302–313.

56.

Talen

(2006) Design that enables diversity: The complications of a planning ideal. Journal of Planning Literature 20(3): 233–249.

57.

Thomas

Frankhauser

Badariotti

(2012) Comparing the fractality of European urban neighbourhoods: Do national contexts matter? Journal of Geographical Systems 14(2): 189–208.

58.

Tochterman

(2012) Theorizing neoliberal urban development: A genealogy from Richard Florida to Jane Jacobs. Radical History Review 112: 65–87.

59.

Torrens

(2008) A toolkit for measuring sprawl. Applied Spatial Analysis and Policy 1: 5–36.

60.

Tsai

Y-H

(2005) Quantifying urban form: Compactness versus ‘sprawl’. Urban Studies 42(1): 141–161.

61.

Turok

Bailey

(2004) The theory of polynuclear urban regions and its application to central Scotland. European Planning Studies 12: 371–389.

62.

Vandermotten

Halbert

Roelandts

, et al. (2008) European planning and the polycentric consensus: Wishful thinking? Regional Studies 42: 1205–1217.

63.

Van Oort

de Geus

Dogaru

(2015) Related variety and regional economic growth in a cross-section of European urban regions. European Planning Studies 23(6): 1110–1127.

64.

Vasanen

(2012) Functional polycentricity: Examining metropolitan spatial structure through the connectivity of urban sub-centres. Urban Studies 49(16): 3627–3644.

65.

Venerandi

Zanella

Romice

, et al. (2017) Form and urban change – An urban morphometric study of five gentrified neighbourhoods in London. Environment and Planning B: Urban Analytics and City Science 44(6): 1056–1076.

66.

Wood

Dovey

(2015) Creative multiplicities: Urban morphologies of creative clustering. Journal of Urban Design 20(1): 52–74.

67.

Youn

Bettencourt

Lobo

, et al. (2016) Scaling and universality in urban economic diversification. Journal of the Royal Society Interface 13(114): 20150937.

68.

Zambon

Benedetti

Ferrara

, et al. (2018) Soil matters? A multivariate analysis of socioeconomic constraints to urban expansion in Mediterranean Europe. Ecological Economics 146: 173–183.

69.

Zambon

Serra

Grigoriadis

, et al. (2017) Emerging urban centrality: An entropy-based indicator of polycentric development and economic growth. Land Use Policy 68: 365–371.

70.

Zitti

Ferrara

Perini

, et al. (2015) Long-term urban growth and land use efficiency in Southern Europe: Implications for sustainable land management. Sustainability 7(3): 3359–3385.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.12 MB