Sustainable Cities and Society 67 (2021) 102715

Contents lists available at ScienceDirect

Sustainable Cities and Society

journal homepage: www.elsevier.com/locate/scs

Spatio-temporal patterns of urbanization in the Kolkata Urban

Agglomeration: A dynamic spatial territory-based approach
Saurav Chakraborty a, *, Indranil Maity b, Priyank Pravin Patel a, Hashem Dadashpoor c,
Suvamoy Pramanik d, Alexander Follmann e, Josef Novotný f, Utpal Roy b
a
Department Geography, Faculty of Natural and Mathematical Sciences, Presidency University, Kolkata, India
b
Department of Geography, University of Calcutta, Kolkata, India
c
Urban and Regional Planning Department, Faculty of Arts and Architecture, Tarbiat Modares University, Tehran, Iran
d
Center for Study of Regional Development, Jawaharlal Nehru University, New Delhi, India
e
University of Cologne, Institute of Geography, Cologne, Germany
f
Department of Social Geography and Regional Development, Charles University, Prague, Czechia

A R T I C L E I N F O A B S T R A C T

Keywords: This study analyzes the spatio-temporal patterns of urban growth for the Kolkata Urban Agglomeration (KUA) in
Urbanization eastern India using dynamic spatial territorial extents and compares the results with existing popular hypotheses
Diffusion-coalescence of urbanization patterns (diffusion-coalescence and three-growth-mode). Time series neighborhood-level built-
Urban growth
up densities were extracted and overlain to identify two broad urban spatial extents, inner and outer, which were
Spatial patterns
Urban agglomeration
further categorized into sub-territorial units. Landscape metrics and mode-wise urban growth statistics were
Peripheral spatial territory computed for the above spatial extents. Results reveal a marked increase in the urban land share, from 20 % in
2000 to nearly 50 % in 2018, when about 75 % of the built-up spaces comprised the high-density urban core.
Urban growth patterns in KUA show clear intra-urban variations between inner and outer spatial extents. The
inner city zone underwent coalescence with nearly 90 % of the built-up growth within inner city extents
occurring through infilling, whereas the outer extents showed complex growth patterns with outlying growth
dropping by almost 27 % points, with related increase in both infilling (11 % points) and edge-expansion (16 %
points). Diffusion and coalescence occurred simultaneously, but at varying intensities within the different sub-
territorial units. Examining such complex spatiality of urbanization is crucial to reduce its adverse environ­
mental effects and promote sustainable urban growth.

1. Introduction adversely impacts the natural environment through loss of fertile agri­
cultural lands (Follmann, Hartmann, & Dannenberg, 2018), forest/­
Rapid urbanization is the standout socio-ecological phenomenon in vegetation cover (Clement, Chi, & Ho, 2015) or water bodies (Mitra &
present-day Asia and Africa, with these regions slated to experience Banerji, 2018), and its impacts are felt well beyond city boundaries (van
about 90 % of the world’s urban population growth by the middle of this Vliet, 2019). These ultimately cause natural landscape fragmentation
century (United Nations, 2018). Indian cities are likely to house an (Dadashpoor, Azizi, & Moghadasi, 2019; Fahey & Casali, 2017) and the
additional 416 million people by 2050 (United Nations, 2018), mostly in decline/loss of valuable ecosystem services (Su, Xiao, Jiang, & Zhang,
its burgeoning and continually expanding metropolitan cities and large 2012). To ameliorate such impacts and foster sustainable urbanization,
urban agglomerations (UAs) (which are defined by the Census of India as enhanced understandings of the spatial patterns of urban growth are
a continuous urban spread, constituting of towns [both Statutory and therefore critical (Dadashpoor, Azizi, & Moghadasi, 2019; Liu, He, &
Census Towns (CTs)], villages and outgrowths)1 . Such swift city growth Wu, 2016; Zhu et al., 2019).

* Corresponding author at: Department of Geography, Faculty of Natural and Mathematical Sciences, Presidency University, 86/1 College Street, Kolkata, West
Bengal, 700073, India.
E-mail addresses: [email protected] (S. Chakraborty), [email protected] (I. Maity), [email protected] (P.P. Patel), h-
[email protected] (H. Dadashpoor), [email protected] (S. Pramanik), [email protected] (A. Follmann), [email protected]
(J. Novotný), [email protected] (U. Roy).
1
As per the 2011 Census, there are 475 UAs in India.

https://doi.org/10.1016/j.scs.2021.102715
Received 6 August 2020; Received in revised form 29 November 2020; Accepted 8 January 2021
Available online 14 January 2021
2210-6707/© 2021 Elsevier Ltd. All rights reserved.
S. Chakraborty et al. Sustainable Cities and Society 67 (2021) 102715

Spatio-temporal patterns of urban growth generally alternate be­ form changes (Dong, Jiao, Xu, Yang, & Liu, 2019; Jiao, 2015; Xu et al.,
tween diffusion and coalescence pathways/forms. It is usually assumed 2019). However, demarcating such dynamic spatial territories on the
that urbanization proceeds from the diffusion to the coalescence phase, basis of neighborhood level built-up land density distribution and
with the former typified by an increasing number of impervious/built- thereafter examining the diverse traces of spatio-temporal patterns of
up patches and the latter representing their merger into continuous urban growth therein, remains largely unknown.
built-up swathes (Dietzel, Herold, Hemphill, & Clarke, 2005). While The continuous evolution in Remote Sensing (RS)-based applications
many have deciphered such trends for different cities (Kantakumar, for urban areas (Zhu et al., 2019) and its integration with Geographic
Kumar, & Schneider, 2016; Liu et al., 2010, 2016; Xu et al., 2007; Yu & Information Systems (GIS) and landscape metrics has enabled the
Ng, 2007), exceptions have also been reported (He, Song, Liu, & Yin, formulation of customized techniques for measuring, monitoring and
2017; Jenerette & Potere, 2010; Wu, Jenerette, Buyantuyev, & Redman, mapping the spatio-temporal patterns of urbanization and its changes
2011) and the diffusion-coalescence hypothesis fallacies have been (Dadashpoor et al., 2019a; Hu, Qian, Pickett, & Zhou, 2020; Taubenböck
subsequently highlighted by the three growth modes proposition [i.e. et al., 2019; Tripathy & Kumar, 2019). Among these, Urban Sprawl
infill, edge-expansion and outlying urban growth (Li, Li, Zhu, Song, & Metrics (USM) have been widely used (Angel, Parent, & Civco, 2007;
Wu, 2013)]. While empirical studies exploring urban growth patterns Kantakumar et al., 2016; Sahana, Hong, & Sajjad, 2018; Sharma & Joshi,
and their changes at the city level have frequently verified the 2013), to demarcate the urban landscape into different spatial territories
diffusion-coalescence hypotheses (Fei & Zhao, 2019; He et al., 2017; He, based on urban land density functions (urbanness) and quantitatively
Zeng, Xie, Tan, & Wu, 2019; Li, Li, Zhu et al., 2013; Liu et al., 2016; Shi, defined qualitative urban extents (i.e. spatial territorial units delineated
Sun, Zhu, Li, & Mei, 2012), urban growth patterns can vary greatly based on neighborhood-level built-up densities- urban core, urban fringe
within a single urban entity, depending upon intra-urban variations in and scatter development zones). Besides this, multi-temporal analysis of
local conditions like the land availability (e.g. Jiao, 2015). However, urban landscapes using USM allows delineation of the changing char­
existing knowledge on intra-urban variations in urban growth patterns is acter/extents of these spatial entities/territories. Such spatial entities
still rare as only very few studies have examined in-depth how urban can also be grouped as inner extents (wherein these territories exist in
growth varies across the existent spatial structures and their both the most recent dataset as well as in previous time period datasets)
sub-divisions (e.g. Bosch, Jaligot, & Chenal, 2020; Dahal, Benner, & and outer extents (when such zones are evident only in the most recent
Lindquist, 2017; Li, Li, Zhu et al., 2013; Li, Li, & Wu, 2013; Peng, Wang, dataset).
Zhou, Zhao, & Yang, 2015). Studies on intra-urban variations can Based on the above theoretical underpinnings, this study utilizes a
empirically enhance the understanding of urban growth patterns and dynamic spatial territory-based approach to better examine the spatio-
conceptually add new dimensions to the existing urban growth hy­ temporal pattern of urbanization by:
potheses (e.g. Bosch et al., 2020). By focusing on intra-urban variations
in the Kolkata Urban Agglomeration (KUA), this study addresses this • Analyzing the spatio-temporal pattern of urban expansion in respect
important research gap. of three spatial urban typologies (urban core, urban fringe and
Urbanization is a complex and varied process (Liu et al., 2016; Yu & scatter development) and identifying the dynamic spatial territories
Ng, 2007; Wu et al., 2011; Zhu et al., 2019), and thus urban areas should (inner and outer extents),
be considered as multi-level systems. The distinct spatio-temporal pat­ • Computing the landscape metrics and modes of urban growth for
terns that prevail at each level (Bosch et al., 2020; Li, Li, Zhu et al., 2013) each of the spatial territories/extents demarcated above (the overall
are believed to be the primary reason behind the observed deviations unit and its components/sub-territories) in the various time periods
from the diffusion-coalescence hypothesis (Liu et al., 2016) and neces­ examined (2000–2010 and 2010–2018), and
sitates the inclusion of spatial extents to characterize the differential • Analyzing how urbanization trends have varied within each entity
patterns and processes of urbanization that are operational within them. and using this to compare our results with the diffusion-coalescence
Both qualitative criteria [i.e. administrative urban hierarchy (Li, Li, Zhu and three growth modes hypotheses.
et al., 2013) and urban-rural transect (Li, Li, Wu et al., 2013)] and
quantitative methods [i.e. bifractal city model- inner and outer zones We expect our results to offer additional evidence for inculcating
(Bosch et al., 2020)] have been used for such areal demarcations to multi-scalar urbanization studies within the ambit of diffusion-
analyze multi-level/scale dynamic growth patterns. Cities and Urban coalescence and three growth modes hypotheses, which has seldom
Agglomerations (UAs) usually have well-demarcated spatial structures, been done using such dynamically demarcated urban territories and
comprising of various types of spatial territories, commencing from the their sub-units. As the study area, we have chosen the KUA in eastern
city core, through an in-between space and terminating in rural locales India, for reasons outlined subsequently.
along the fringes of the territorial boundary limits.
The question of “location and spatial expansion patterns of such 2. Study area
territories” makes their demarcation quite difficult (Dadashpoor &
Ahani, 2019: 03). While the locations can be usually identified from Located in the eastern Indian state of West Bengal, the KUA is the
settlement boundaries, the blurring of rural-urban locales, due to the third-most populous UA in India (Census of India, 2011) and the 14th
unprecedented expansion of built-up areas beyond legally defined urban largest in the world (United Nations, 2016). It comprises of 4 Municipal
boundaries, makes this quite difficult (Jain, 2018; Simwanda & Mur­ Corporations (MC), 39 Municipalities (MU), 154 CTs (of which 75 were
ayama, 2018). As a result, many rural areas are now characterized by accorded this status only in the most recent Census of 2011), 367 vil­
built-up spaces similar to those found within denoted urban areas. This lages (of which 41 are expected to be declared as CTs in the next Census
growth of cities into its periphery triggers what Weng (2007:350) has of 2021) and 3 Outgrowths (Fig. 1.a). Its core is comprised by the Kol­
termed as the “relay-race effect”, wherein core urban areas expand while kata Municipal Corporation (KMC) and the Haora Municipal Corpora­
pushing rural-urban fringes further outward (Seto & Fragkias, 2005). tion (HMC), lying on opposite flanks of the River Hugli. Although, the
Thus, static urban extents or spatial territories rarely exist around such historical development of the city has resulted in its present dense
dynamic cities and therefore their spatial territorial boundaries are dy­ configuration of built-up spaces and congested neighborhoods (Chak­
namic due to expansion, relocation and shrinkage, irrespective of what ravorty, 2000; Haque & Patel, 2018; Haque, Rana, & Patel, 2020), it was
the demarcated administrative boundaries may be, whose the economic liberalizations of 1991 and the development of an infor­
re-demarcation often lags behind such rapid changes in impervious ex­ mation technology hub in the town of Bidhannagar (adjacent to Kolkata)
tents and space transformations. Urban land density and its distribution that has largely spurred on the recent urban growth herein. Given its
has been widely used to demarcate urban spatial structures and urban long urbanization history, large connurbation extents and marked

2
S. Chakraborty et al. Sustainable Cities and Society 67 (2021) 102715

Fig. 1. The study area, showing a. Entities within Kolkata Urban Agglomeration; b. Annual growth rate of population from 2001-2011; c: Population density in 2001;
d: Population density in 2011.
Source: Computed and compiled by the authors from Census of India, 2001 and 2011 maps and dataset.
Note: The Indian Census classifies urban areas into two broad categories- Statutory Towns (ST) (urban areas that are under urban governance, i.e. some urban
administrative body like Municipal Corporation and Municipality etc.) and Census Towns (CT) (urban areas that are governed under rural administration i.e. gram
panchayat). Potential CTs are those villages in 2011 that already meet the Census of India stated criteria to be designated as new CTs in the forthcoming Census of
2021. However, in this paper, such entities have been still considered as villages/rural areas. Outgrowths (OG) are a viable unit such as a village or part of a village
contiguous to a ST that possess most urban features in terms of infrastructure and amenities. The OG remains outside the statutory limits of a ST but falls within the
revenue limit of a village or villages contiguous to the town or city (Census of India 2011).

recent developments in and around the city limits that have endangered the KUA’s eastern periphery (Fig. 1.a), while fertile arable lands, surface
adjacent environmentally sensitive zones, the KUA was considered ideal water bodies and green belts continue to be converted into built-up
for the ensuing analysis. surfaces at unprecedented rates (Mondal et al., 2017; Sahana et al.,
As per the 2011 Census, the KUA has 14.72 million residents and an 2018), thereby impairing the critical balance between natural and
average population density of 7950 persons/sq.km. Nearly 83 % of its human-made landscapes and reducing the ecosystem health (Das, Das, &
total population lives in its legally urban areas (Statutory Towns- STs). Mandal, 2020). Comprehending the spatio-temporal patterns of KUA’s
However, during 2001–2011, the population growth rate in these areas expansion is therefore crucial to exploring its trajectories and the
was far lower (0.49 % per annum) than that of the rural areas (2.22 % areas/landscape elements most likely to be impacted, which is para­
per annum) located along the KUA’s periphery (Fig. 1.b). This high mount towards fostering sustainable urban growth.
population growth rate in the periphery, particularly within the CTs
therein, has further contributed to the increasing population density 3. Methodology
(Fig. 1.c and d) and the subsequent expansion of built-up areas. With
non-existent institutional capacities, lax/seldom enforced contracting/ An integrated remote sensing, geographic information system and
building regulations and proximity to a major urban centre, such legally landscape metrics based approach was adopted to prepare and compute
rural areas are hotspots for rapid and unplanned urban expansion, the required datasets and statistics (Fig. 2).
highlighting the unacknowledged/hidden urbanization occurring in
India. Such urbanization has adversely impacted the ecologically sen­
sitive East Kolkata Wetlands (a designated Ramsar site) situated along

3
S. Chakraborty et al. Sustainable Cities and Society 67 (2021) 102715

Fig. 2. Workflow for dynamic spatial territory based spatio-temporal patterns of urbanization.

3.1. Image classification and accuracy assessment into a different spatial entity at a later date (e.g. Kantakumar et al.,
2016).
Landsat Thematic Mapper (TM 4-5) images of 2000 and 2010 and USM can thus be used to identify different urban typologies and their
Landsat-8 OLI images of 2018 have been used in this study (Appendix exact spatial extents and changes over time, from multiple datasets.
Table A.1). Supervised image classification was employed using the Kantakumar et al. (2016) demonstrated that the urban core expands at
maximum likelihood method (Sahana et al., 2018), and eight broad the expense of the urban fringe and low-density scatter development.
LULC classes (built-up area, forest/vegetation, water bodies, river, These three entities were identified within the KUA for each of the
wetlands, open lands, agricultural fallow and cropland) were extracted studied timestamps (i.e. 2000, 2010 and 2018) and overlain on each
for each period. High-resolution Google Earth images were used for other. From the overlays, two broad spatial units were discerned; the
validation and the computed Kappa statistics, based on a confusion inner extent (which had particularly developed/become built-up during
matrix, helped confirm the achieved accuracy (Appendix Table A.2). the earlier study years of 2000 and 2010) and the outer extent (the
portion that had developed only in the current study year of 2018) (see
3.2. Urban Sprawl Metrics Supplementary Fig. F2). By comparing the respective boundaries and
their temporal changes for the spatial entities outlined above, we then
USM is a built-up density classification method (Angel et al., 2007; categorized the entire urban landscape into six zones- urban core 2000,
Sahana et al., 2018) that accords various urban typologies, and was urban core expansion 2000–2010, urban fringe of 2010, urban core
employed for demarcating the types of urban growth present and expansion 2010–2018, relocated urban fringe of 2018 and reducing
ascertain its spatial patterns/dynamics. All pixels designated as built-up scatter development in 2018 (Supplementary Fig. F3).
areas were accorded a certain urbanness class, based on the proportion The respective landscape metrics in each time period and the asso­
of built-up area present within a radius of 1000 m around them (based ciated growth mode type (i.e. infill, edge-expansion and outlying
on Angel et al., 2007). Pixels that obtained values more than 50 %, were growth) related statistics were subsequently computed for two hierar­
considered to constitute the Urban Core (UC), those that had between 30 chical spatial units, as follows:
%–50 % as part of the Urban Fringe (UF), and below this as Scatter
Development (SD) (following Angel et al., 2007. See supplementary A3 i Overall (i.e. for the entire agglomeration), the inner extent (which
and Fig. F1). included the urban core 2000, urban core expansion 2000–2010 and
urban fringe 2010) and the outer extent (which included the urban
3.3. Demarcating dynamic spatial territories core expansion 2010–2018, the relocated urban fringe 2018 and the
reducing scatter development 2018).
Angel et al. (2007: 02) opine that the above spatial territories/enti­ ii For the individual six sub-territories situated within the KUA and its
ties (i.e. USM classes) are “exclusive of each other”. However, continual inner and outer extents.
built-up growth within and around city cores and the unprecedented
transformations along a city’s peripheral regions make such territorial Additionally, the areal changes in the existing built-up typologies
boundaries dynamic. Thus, the definitive characteristics of a particular from one study year to another (and the possible alterations in their USM
territory/entity at a certain timestamp can alter due to changes in the class) and the transformation of formerly non-built-up areas into a
pattern and extent of its impervious cover and thereby metamorphose it particular urban typology (UC, UF or SD) were also calculated. For the

4
S. Chakraborty et al. Sustainable Cities and Society 67 (2021) 102715

latter, the existing non-built-up LULC classes were re-categorized using Table 2
the concept of Urbanized Open Space (UOS) and Non-Urbanized Open Descriptions of the enumerated landscape metrics.
Space (NUOS). The UOS was defined as “undeveloped/non-built-up land Metrics Description Category
surrounded by urbanness values of at least 50 %” (Angel et al., 2007),
Percentage of landscape Areal share of a particular land-use Composition
while “undeveloped/non-built-up land surrounded by urbanness values (PLAND) class
less than 50 %” were classified as NUOS (Sahana et al., 2018). The Patch density (PD) Number of patches per unit area Configuration
different LULC to built-up transformations were thereby classified into Euclidean mean nearest Mean patch shortest edge-to-edge Configuration
six categories based on the respective UOS and NUOS attributes neighbor distance distance to the nearest neighboring
(ENN) patch of the same class
(Table 1). Mean patch size (MPS) Total patch area divided by the Configuration
number of patches of the same time
3.4. Urban landscape pattern analysis Area-weighted mean Mean patch fractal dimension Configuration
fractal dimension weighted by relative patch area
(AWMFD)
Landscape metrics have been quite popular in assessing and under­
standing how landscape patterns alter in response to LULC changes Note: Built-up area belonging to different spatial territories were considered for
(whether natural or anthropogenic). Many studies have traced the the calculation of the respective landscape metrics. All metrics were calculated
impact of urbanization on the landscape pattern and its changes over at the landscape level using 8 cell neighborhood window in FRAGSTATS
software.
time (e.g. Dadashpoor et al., 2019a; Li, Peng, Yanxu, & Yi’na, 2017; Su,
Source: FRAGSTATS HELP v4.2.1 (McGarigal, 2015).
Jiang, Zhang, & Zhang, 2011) and deciphered that due to strong
multi-collinearity among the different landscape metrics, only some
selective landscape metrics are particularly useful in enumerating the the edge of a non-circular existing urban patch with landscape expan­
changes (Peng et al., 2010; Su et al., 2012). sion index value >50 %) (see Shi et al., 2012: 427 for details). In this
For assessing urban growth and its dynamics, many studies have used way, we were able to identify four different modes of urban growth:
a range of landscape patch metrics to highlight the composition (areal Infill Type-1, Infill Type-2, Edge-Expansion and Outlying Growth
share) and configuration (aggregation, isolation, edge, shape and frag­ (Fig. 3).
mentation) of urban landscape patterns and its changes (Bosch et al.,
2020; Dadashpoor et al., 2019a; Dietzel, Herold et al., 2005; Dietzel, 4. Results
Oguz, Hemphill, Clarke, & Gazulis, 2005; Li, Li, Zhu et al., 2013; Li, Li,
Wu et al., 2013). In this study, a total of five metrics were used for 4.1. Spatial patterns of urban typologies and their growth
quantifying the urbanization pattern in the KUA, allowing our results to
be compared against those elicited from similar studies (see Table 2). The entire built-up area of the KUA was divided into the UC, UF and
SD typologies and their respective areas were calculated and mapped in
3.5. Modes of urban growth each of the three studied years (Fig. 4). High-density UC was dominant
throughout the study period and increased from 56 % to 67 % in Phase-I
Xu et al. (2007) and Liu et al. (2010) have respectively proposed that (2000–2010). While UC’s growth rate was 11.81 % in Phase-I
new urban growth can be identified using common boundary analysis (2000–2010), it slowed down to 8.65 % in Phase-II (2010–2018; see
and the Landscape Expansion Index, but both these methods are limited Supplementary Table A4). Contrarily, the UF and SD typologies’
while distinguishing between particular infilling modes of urban growth respective coverage declined throughout. Interestingly, during Phase-II,
and engender confusions among the edge-expansion and outlying the UF growth rate increased marginally (by appropriating a chunk of
growth modes (Shi et al., 2012). The alternative method of Shi et al. the SD coverage), while the SD growth rate declined markedly (see
(2012) was thus used to overcome such issues and classify the newer Supplementary Table A4). The growth of these three built-up typologies
urban growth occurring within the KUA during the examined study during the entire study period occurred at the expense of different non-
periods. Besides, we have further categorized the infill growth type into built-up open spaces. In particular, the spatial pattern of such open
Infill Type-1 (for all types of infill growth that occurred within the spaces particularly changed due to the unprecedented spatial expansion
boundary of an already existing built-up tract with or without sharing its of the urban core (Fig. 5).
boundary with the existing built-up patches) and Infill Type-2 (growth at During Phase-I, nearly 52.17 sq.km and 18 sq.km of former UF and
SD areas were converted into UC, while this class simultaneously added
Table 1 another 133 sq.km of newly formed built-up area including nearly 69 sq.
Urban sprawl metrics-based classifications of the non built-up LULC classes. km from non-urbanized grey open space, 20 sq.km from urbanized grey
open space and about 4.8 sq.km from the urbanized blue open space
USM based Urbaness value and Non-built LULC USM typified
typology of open definition criteria class non built-up categories. During Phase-II, nearly 82 sq.km and 24 sq.km of the existing
space class UF and SD tracts, respectively, were converted into UC areas, while
Urbanized
about 165 sq.km of newly built-up zones were also added. About 56 sq.
Forest/ km of vegetated areas also got converted into UC, of which 16 sq.km
green open
vegetation
space were urbanized green open spaces and 40 sq.km non-urbanized green
non built-up pixels that
Urbanized Open Open land, open spaces, respectively (Fig. 4.d). Similarly, about 24 sq.km of ur­
are surrounded by > 50 Urbanized grey
Space (UOS) crop and fallow
% of built-up pixels open space banized grey open spaces and 52 sq.km of non-urbanized grey open
land
Water bodies, Urbanized blue spaces were consumed within the UC class by 2018 (Fig. 5.d).
river, wetlands open space Both UF and SD show dynamic territorial changes through the study
Forest/
Non-urbanized period. For example, in Phase II only 4.5 sq.km of existing UF remained
green open unchanged. The spatial growth of UF occurred from the transformation
vegetation
space
Non-Urbanized non built-up pixels that
Open land,
of former SD areas (34.95 sq.km), non-urbanized green open spaces
Open Space are surrounded by < 50 Non-urbanized (32.23 sq.km), non-urbanized grey open spaces (47.78 sq.km) and non-
crop and fallow
(NUOS) % of built-up pixels grey open space
land urbanized blue open spaces (1.9 sq.km). Similarly, 40.72 sq.km of the SD
Water bodies, Non-urbanized built-up typology remained in Phase II, while additions occurred at the
river, wetlands blue open space
expense of non-urbanized grey open spaces (37.11 sq.km), non-
Source: Based on Angel et al. (2007). urbanized green open spaces (27.1 sq.km) and non-urbanized blue

5
S. Chakraborty et al. Sustainable Cities and Society 67 (2021) 102715

Fig. 3. Graphical depiction of growth modes for newly grown urban patches (based on Shi et al., 2012).
Note: Misclassification of infill type 1 (b and c) into edge expansion and outlying may affect the statistics derived for these urban growth modes, which can impact the
results as well.

open spaces (2.09 sq.km). • The trends of the shape complexity of the built-up patches, judged
In summation, spatial expansion and further densification of the from the area-weighted mean fractal parameter (AWMFD), were
urban core territory result in a densely built urbanscape emerging as the similarly showing- a monotonic decline for the inner extent, while an
dominant built-up typology, while at the same time becoming more increasing trend for the outer extent and the total agglometation.
homogenous in character. The marked conversion of former open spaces
into high-desity built up areas is connected to multiple ecological bur­ The above landscape metrics were also calculated for each time-step
dens, including conversion of urbanized open lands. On the other hand, for the six sub-territories (three each for the inner and outer extents; see
the rise in moderate and low-density built-up areas (UF and SD) results Fig. 7). The PLAND showed a monotonic increasing trend for all sub-
in increasing landscape fragmention beyond the extents of the densely territories. However, during Phase-II its rate of increase slowed for the
built locales and is also connected to multiple environmental impacts (e. sub-zones within the inner extent while rising for those within the outer
g. conversion of wetlands, vegetation and crop lands). extent. The PLAND for the ’urban core 2000’ situated within the inner
extent reached almost 95 % in 2018. On the other hand, the ’urban core
expansion 2010− 2018’ sub-zone within the outer extent recorded the
4.2. Landscape patterns in the inner and outer agglomeration zones highest increase in PLAND during Phase-II (about 40 % points), followed
by that for the ’relocated urban fringe 2018′ and the ’reducing scatter
Fig. 6 displays the various landscape metrics calculated for the three development 2018′ .
different spatial entities- the agglomeration (entire KUA), the previously In the case of the inner sub-zones, all the metrics confirmed to a
demarcated inner and outer extents. The salient aspects noted were: similar trend, which was also reflected by the inner extent as a whole.
For example, the PD, ENN_MN and AWMFD showed a decreasing trend
• The built-up area proportion (PLAND) increased for all three spatial for all three sub-zones whereas the MPS increased monotonically in
entities in each time phase- the agglomeration showed an almost each. For the ’urban core 2000′ and the ’expanded urban core in 2010′ ,
uniform rising trend in both phases though this growth declined the increase in their respective MPS was reasonably higher during
slightly in Phase-II for the inner extent (wherein the PLAND was Phase-II (Fig. 7).
almost 90 % in 2018), possibly due to lack of suitable space for The outer extents’ sub-territories on the other hand, exhibited a
further increase. With space availability not an impediment in the completely different and mixed trend. For example, the PD followed a
outer extent, its PLAND rose most sharply by almost 20 % points in monotonic declining trend for the ’urban core expansion 2010− 2018’
Phase-II. zone, a unimodal pattern for the ’relocated urban fringe 2018′ zone and
• For Patch Density (PD), a unimodal trend was found for the an increased monotonic rate for the ’reducing scatter development
agglomeration and outer extent, with increase during Phase-I and 2018′ zone. For the MPS, a sharp increasing trend was noticed for the
decline in Phase-II. However, the PD declined in both phases for the ’urban core expansion 2010− 2018’ zone whereas it increased only
inner extent, albeit to a lesser extent in Phase-II. slightly for the other two sub-units. The ENN_MN and the AWMFD were
• The Mean Patch Size (MPS) increased continuously in each spatial the only two metrics for which a similar trend was recorded throughout
unit (possibly pointing towards more coalesence), with this being for all three sub-territories of the outer extent; similar to what had been
highest for the inner extent. observed for the outer extent as a whole, i.e. a decreasing mean
• The Mean Euclidean nearest neighbor distance (ENN_MN), denoting Euclidean urban patch distance and increasing patch shape complexities
the average distance between built-up patches, also declined for all (Fig. 7).
three spatial units. However, this was more apparent for the inner The indices presented above clearly indicate that within the inner
extent during Phase-II, while the other two had monotonic declining extents and its sub-territories, urban growth had almost reached satu­
trends throughout the study period. ration point, forming large continuous built-up patches and there is

6
S. Chakraborty et al. Sustainable Cities and Society 67 (2021) 102715

Fig. 4. The urban landscape of the KUA.

limited land available for further development. On the contrary, the edge-expansion and outlying growth), were shown to contribute vary­
outer extent and its sub-territories were the hot spot for urban growth, ingly to city expansion, as per their respective areal shares in the above
with the presence of dense, moderate and low density built-up patches spatial territories (Fig. 8).
making this entity a dynamic and complex part of the urban landscape. During Phase-I, both the agglomeration and inner extents displayed
Such dynamicity could be read in terms of the increasing patch size, nearly similar patterns in their relative dominance of all four growth
density and complexities of built-up tracts found in the outer extent modes. However, differences in their share of the infill type-1 and the
(AWMFD), which was quite opposite to that seen in the inner extent and outlying modes were particularly discernible. The outer extent, con­
its sub-territories (see Fig. 7). trastingly, was mainly dominated by edge-expansion and outlying
growth modes, with scant traces of the infilling types.
A very different picture arose during Phase-II. At the agglomeration
4.3. Modes of urban growth extent (i.e. for the entire KUA), the areal share for both infill type-1 and
edge expansion increased (about 5% points for each) at the expense of
The different modes of urban growth (infill type-1, infill type-2,

7
S. Chakraborty et al. Sustainable Cities and Society 67 (2021) 102715

Fig. 5. Spatial distribution of different non-built-up open spaces for three different study years (a, b and c); d. Transition of different LULC classes into various built-
up typologies, Phase-I (2000–2010) and Phase-II (2010–2018).

Fig. 6. Landscape metrics for the agglomeration, inner and outer extents of the KUA.

outlying growth (which dropped similarly- about 10 % points in total). inner extent, as space saturation occurred. Within the outer extents, a
During Phase-II, the inner extent was particularly dominated by the significant drop in the outlying growth mode occurred during Phase-II
infilling types (that together occupied 89 % of the new urban develop­ (about 27 % points), with a concomitant increase in the dominance of
ment occurring) and both the edge-expansion and outlying growth the edge-expansion category, which contributed 55 % of the total newly
modes declined and the latter became quite insignificant within the built-up land. Similarly, the dominance of infilling types rose in the

8
S. Chakraborty et al. Sustainable Cities and Society 67 (2021) 102715

Fig. 7. Landscape metrics for the six sub-territories situated within the inner and outer extents of the KUA.

9
S. Chakraborty et al. Sustainable Cities and Society 67 (2021) 102715

Fig. 8. Areal share (in %) under the different modes of urban growth (Phase - I: 2000 - 2010 and Phase - II: 2010–2018) as per the dynamic urban expansion
boundaries- a and b denote these for the respective broader spatial territories; c and d denote these for the sub-territories situated within the inner and outer extents.

outer extent during Phase-II. ’relocated fringe in 2018′ sub-territories witnessed a considerable new
Examinations of the relative dominance of growth modes according built-up addition via infilling. Edge-expansion was also recorded
to the respective sub-zones, offers further insight into occurring urban simultaneously in these two sub-zones and occurred therein at the
growth patterns. For example, within the inner extent, the ’urban core expense of outlying growth. This same trend was evident in the ’existent
2000′ was primarily dominated by infill types with a meager amount of fringe since 2010′ sub-territory within the inner extent. The ’shortened
outlying growth during Phase-I. The other two sub-territories therein scatter development in 2018′ sub-territory, had a significantly increased
mostly experienced edge-expansion and outlying growth. Contrarily, share of edge-expansion alongside a lower increase in infilling, all at the
infilling and edge-expansion were mostly observed for the ’urban core expense of outlying growth.
expansion 2010− 2018’ and ’relocated urban fringe 2018′ zones in the The results above documented that the inner extent, which was
outer extent. The ’reducing scatter development 2018′ territory mainly primarily dominated by densely built areas, grew mostly through edge-
experienced outlying growth while edge-expansion was dominant for expansion mode during Phase-I, but experienced more infilling in Phase-
the ’urban core expansion 2010− 2018’ zone in the outer extent. II. This is an expected growth scenaio, with the more developed urban
During Phase-II, the dominance of infilling types became more areas generally experiencing infilling growth. Such growth has mainly
evident in all three sub-zones within the inner extent. In the ’expanded occurred at the expense of existing open spaces, reducing its already
core in 2010′ sub-territory, the rise of the infill type-1 category occurred limited amount even further (see Fig. 5). In contrast, the sub-territory
at the expense of both edge-expansion and outlying growth. However, within the outer extent that was adjacent to the inner extent (i.e.
within the outer extent, both the ’expanded core in 2018′ and the urban core expansion 2010–2018) clearly offered a transition from the

10
S. Chakraborty et al. Sustainable Cities and Society 67 (2021) 102715

outlying to the infilling growth mode. Besides this, the existence of a present urban growth in the KUA is significantly higher in settlements
substantial proportion of both edge expansion and outlying growth (CTs and villages) under rural administration compared to those situ­
within the other two sub-territories in the outer extent reflected the ated within urban administrative units (MCs and MUs- see Fig. 1). This
complexity of the urban growth patterns/trajectories therein. highlights the existent informal urbanization, wherein the state pro­
motes deregulated growth in the peri-urban areas administered by rural
5. Discussions entities (2011, Roy, 2005; 2016). With available land scarce in the
saturated core city region, which primarily comprises of statutory towns
5.1. Changing urbanization pattern and dynamic spatial territory (Bhatta, 2009; Sengupta, 2006), this newer high-density urban core is
formation now expanding into its adjoining settlements in an unprecedented
manner (Supplementary Table A5) and transforming ecologically valu­
In this study, a neighborhood-level built-up density based approach able lands, i.e. forest/vegetation, crop lands and surface water bodies,
was employed to identify the three major spatial territories within the including the East Kolkata Wetlands, at an unprecedented rate (Sup­
KUA, i.e. urban core (high density built-up), urban fringe (moderately plementary Fig. F4). This will futher increase the ecological burden of
dense built-up) and scatter development (low density built-up). The city expansion and engender deterioration of the ecosystem health (Das
spatial extents of these different urban typologies in three timestamps et al., 2020).
were demarcated and overlapped to identify their temporal dynamics.
Inner and outer extents were suitably identified within the KUA together 5.2. The KUA’s growth trajectories and different hypotheses of
with their sub-territories (three within each). Results reveal that with urbanization
ongoing urbanization, high density built-up expectedly expands out­
ward, into previously moderate (UF) and low density (SD) zones. Across the world, the impacts of urban growth are not restricted to
Consequently, the previous UF zone is pushed outward due to the rapid within the city limits, being felt well beyond them (van Vliet, 2019).
UC expansion around the pre-existing city center and relocates itself Thus, deciphering the ambient spatio-temporal urban growth pattern
through occupation, conversion and partial replication of its previous and its trajectories is an important step in assessing how urbanization
form within the adjacent countryside. Low-density extents reduce affects the natural landscape in order to better inform any planning
considerably within/near the expanding city while fostering away from measures (Dahal et al., 2017; Dietzel, Herold et al., 2005; Liu et al.,
it, thus reshaping the urban landscape territorially. Thus, the urban- 2010). Such studies usually examine whether urban growth follows a
rural fringe is best considered as a dynamic frontier instead of as a generalized trend over time (i.e. diffusion at first, followed by coales­
static zone (e.g. Bosch et al., 2020; Dong et al., 2019), especially for cence; e.g. Fei & Zhao, 2019; He et al., 2017, 2019). This idea/concept is
fast-growing urban agglomerations. seemingly invalided when urbanization patterns are examined through
According to Seto and Fragkias (2005: 878), “the boundary of the a multi-scale perspective (e.g. Bosch et al., 2020; Li, Li, Zhu et al., 2013).
urban-rural fringe tends to shift through time as urban land expands Thus not considering such nuances may lead to inconclusive/erroneous
outward from the city centre”. While many have recognized such urban surmises about overall urban growth trajectories.
area restructuring/reorganization (Broitman & Koomen, 2015; Kant­ In this context, our results provide additional insights into how intra-
akumar et al., 2016; Sahana et al., 2018), discussions of the specific urban growth patterns vary within broader spatial extents (the demar­
pathways through which this occurs, the morphed nature of the spaces cated inner and outer zones) and in their sub-territories and differ from
themselves or how this impacts on urban growth are limited. Based on the theoretically expected trends. In case of the agglomeration, inner
the aforementioned UC-UF-SD alterations in the above two phases and outer extents demarcated within the KUA, landscape metrics
(2000–2010 and 2010–2018), six different spatial territories were confirm that urban growth has mostly occurred through coalescence.
categorized. While the baseline UC demarcated in 2000 occupied the The patch shape complexities (AWMFD) for the agglomeration and outer
same space in both 2010 and 2018, its boundaries had expanded into extents, which increased during the study period, also indicate a tran­
previously UF, SD and non-built areas. The UC could therefore re-classed sition from diffusion to coalescence (Dietzel, Herold et al., 2005; Dietzel,
as ’urban core 2000′ and ’urban core expansion 2000–2010 and Oguz et al., 2005). However, the computed sub-territorial multi-­
2010− 2018’. The UF growth in Phase-II (2010–2018) resulted from its temporal landscape metrics reveal the fuzzy nature of the overall growth
appropriation of former SD zones, thereby enabling its relocation and pattern. For example, all the three sub-territories within the inner extent
outward migration towards the peripheral areas of the KUA. The small conform to the overall pattern (coalescence) observed for the inner
portion of the existing UF of the 2010 dataset that was not merged with extent, while those within the outer extent deviate from this. The MPS
the newly developed UF of 2018 was classed as ’urban fringe 2010′ . The and ENN_MN values increased and decreased in all three outer extent
rapid UC growth and UF migration had led to marked shrinkage in the sub-territories respectively, indicating coalescence, with decreased PD
SD typology’s coverage in Phase-II, and this remnant zone was catego­ values for the ’urban core expansion 2010− 2018’ and ’relocated urban
rized as ’reducing scatter development in 2018′ . Such direct trans­ fringe 2018′ zones. In contrast, patch shape complexities increased for
formations of low-density scatter development territories into all three outer extent sub-territories implying transitions from diffusion
high-density urban areas within a small time span (2010–2018) to coalescence, with increasing PD values for the ’reducing scatter
strongly reflects the peri-urban dynamics of cities and urban agglom­ development 2018′ zone. Li, Li, Wu et al. (2013) opine that diffusion and
erations in the global south, particularly in Asia and Africa (Dong et al., coalescence should not be considered as two separate processes oper­
2019; Fei & Zhao, 2019; Follmann et al., 2018; Jiao, 2015; Xu et al., ating in mutually discrete time periods. Rather, they can occur simul­
2019). taneously and more importantly, ’diffusion rejuvenation’ may be
The economic liberalization of 1991 provided impetus for urbani­ discerned beyond the urban extents that have completed coalescing.
zation in the major metropolitan cities/UAs of India but in West Bengal, However, in contrast to Li, Li, Wu et al. (2013), our results reveal that
this was primarily manifested by paradigm shifts in the state’s economic coalescence may dominate in certain urban zones (within city inner
and industrial policies in the year 2000 to attract large-scale investment extents in particular) while within certain zones (e.g. sub-territories of
in the real estate sector (Shaw & Satish, 2007), leading to urban growth the outer extent), both diffusion and coalescence can occur simulta­
in the KUA’s south-eastern (Rajpur-Sonarpur and Baruipur) and neously, though at varying intensities. This is primarily begat by the
north-eastern (Bidhannagar, Barasat and Madhaymgram) tracts (Mon­ differing usability, accessibility and price/rent of lands within the con­
dal, Das, & Dolui, 2015). The establishment of large industrial hubs had fines of already built-up tracts against that offered by relatively
promoted such expansions in its south-western (Sankrail-Abada) and non-built spaces along the periphery. Thus, it may be erroneous to
north-western (Dankuni) zones (Vision 2025, KMA). Interestingly, the surmise that renewed diffusion can occur only after complete saturation

11
S. Chakraborty et al. Sustainable Cities and Society 67 (2021) 102715

(e.g. Li, Li, Wu et al., 2013). This simultaneous diffusion-coalescence Phase-II in our study area. This becomes detrimental from the envi­
process offers additional challenges to manage ecologically sensitive ronmental health perspective. Therefore, fast urbanizing nations in the
land cover (water bodies, open spaces) amid high-density urban growth global south, including India, need to devise alternative governance
on one hand and landscape fragmentation due to low density sprawl frameworks for such city peripheral areas as these show the most dy­
development on the other (Dadashpoor et al., 2019b; Li et al., 2017). namic urban growth, in order to redirect/regulate the occurring land­
The growth patterns of the overall agglomeration agree with the scape changes sustainably. Discerning these intra-urban variations in
theoretical assumptions of the three-growth mode proposition, i.e. the urban growth patterns is important, since the outer territory primarily
dominance of edge-expansion and the alternate growth/decline between consists of rural settlements (i.e. villages and Census Towns) that hardly
the infilling and outlying modes. However, deviations from the idealized have any ‘urban-administrative authority’ or framed policies to govern
conceptions were noted for the inner and outer extents and in the six such dynamicity. This allows infomal conversion of ecologically sensi­
sub-territories within them. For example, during Phase-I, the infilling tive land and water bodies into built-up tracts, resulting in severe
types were particularly insignificant in the outer extent whereas, during environmental impacts (Jain, 2018; Roy, 2005; 2016; Simwanda &
Phase-II, outlying growth was negligible within the inner extent. More Murayama, 2018). Although these peripheral zones should be of a high
interestingly, during Phase-II, the dominance of infill type-1 in the inner concern with respect to territorial planning, they actually typically
extent occurred at the expense of edge-expansion, while in the outer receive comparatively low attention. Planning these dynamic urban
extent, edge-expansion gained prominence at the expense of the peripheral areas can be seen as a key to make urban growth more sus­
outlying growth mode. Additionally, during the entire study period, the tainable in the near future.
infilling types in the three sub-territories within the inner extent grew at
the cost of both the edge-expansion and outlying growth modes, indi­
cating the dominance of the coalescing processes, while during Phase-II, 5.4. Methodological considerations
these infilling types were significant only in the two sub-territories
within the outer extent, i.e. the ’urban core expansion 2010− 2018’ Previous studies have examined the spatiality of urban areas using
and the ’relocated urban fringe 2018′ . This implies that infilling was both administrative (Li, Li, Zhu et al., 2013) and statistical criteria
insignificant only within the outermost sub-territory located in the outer (Bosch et al., 2020). However, such examinations are limited as they
extent of the KUA, where edge-expansion outweighed the outlying only consider different spatial territories present within certain pre­
growth mode. These observations i.e., the nonconcurrent presence and determined extents (as proposed by the bifractal city model) or hierar­
multi-directional change (e.g. increase dominance of infilling at the chical levels (administrative divisions), thereby possibly circumscribing
expense of edge expasnion and edge expasnion at the expense of analyses/surmises of how urban growth patterns differ across different
outlying) of three types of new urban development are thus deviations built-up typologies and their evolution.
from what is considered the norm as per the assumed pathways of urban The dynamic spatial territory extents employed in this study do not
growth postulated by the three growth modes hypothesis. Such findings depend on any specific location or any administrative hierarchies to
thus call for the generally accepted pathways of urban growth to be re- characterize urban growth spatiality (Supplementary Fig. F5). Instead,
examined, as more and more deviations from the norm become apparent the various spatial typologies demarcated through USM are used to
(e.g. Bosch et al., 2020; Dahal et al., 2017; Li, Li, Wu et al., 2013). The define such extents and consider the temporal changes in them
sub-divisions delineated within the broader spatial extents of the city accordingly. The rules proposed here for demarcating such units also
offer additional insights to such growth patterns and their enables differentiation of high, moderate and low-density urban devel­
pathways/trajectories. opment zones (Supplementary Figs. F6 and F7). Furthermore, the sub-
Therefore, since urbanization is an inherently complex process and territorial classifications enable comprehension of how urban growth
operates differently across various zones/territories (Dadashpoor & patterns vary even inside such broader spatial territories, i.e. within the
Ahani, 2019; Shi et al., 2012), any interpretations of its temporal growth inner and outer city extents. This is particularly true for the outer extent
pattern also need to be done at/for different spatial extent/entities to where admixtures of high, moderate and low-density built-up areas co-
test the validity of the diffusion-coalescence and three growth modes exist. Two distinctive paths of urban growth could also be traced be­
hypotheses. The differing characters of the sub-zones within the broader tween the existing/old high-density built-up areas and zones of newer
city extents can also skew any generalized/over-arching conclusions. urban growth. The former reflects infilling growth types causing coa­
lescence, whereas it is edge-expansion which precipitates merging of
5.3. Implications for fostering sustainable urbanization built-up patches in the latter. Most importantly, coalescence can
possibly be further typified into two categories: infilling-coalescence (a
An unprecedented conversion of valuable ecological land resources patch consumes its enclosed spaces) and expansion-coalescence (multiple
into impervious surfaces in KUA poses specific challenges towards patches together consume the countryside between/separating them).
mitigating its adverse impacts on the environment (Das et al., 2020; Furthermore, our results also confirm that diffusion does not commence
Sahana et al., 2018). Apart from simply highlighting the changes in the only after coalescence ceases, but that both can occur simultaneously,
natural land cover and its transition to built-up tracts, we show how particularly in the outer extents. This observation is particularly sig­
USM based re-categorization of the LULC can inform its transformations nificant for cities/UAs in developing nations wherein peri-urbanization
in a better way. However, the nature of such transformations differs drives land use/land cover changes and is difficult to plan for or regulate
from the urban core area to its periphery. In particular, the recent (Asabere et al., 2020; Shi et al., 2012).
spill-over of densely built areas into the urban fringe, particularly in the This study is limited by the non-consideration of secondary UCs (and
global south, makes their management further difficult (Follmann et al., their respective growth trajectories) situated within the KUA (Supple­
2018). The identification and inclusion of urbanized and non-urbanized mentary Fig. F8). This possibly partially skews the analyses undertaken
green and grey open spaces in urban planning can help local authorities and the enumerations obtained. Furthermore, the identified spatial
better manage such tracts. territories were not continuous, which could have subsequently influ­
The complexity of growth in an urban agglomeration’s outer extent, enced the calculated landscape metrics statistics. The inculcation of
as outlined in the study, provides particular challenges in managing this polycentric/nodal development patterns involving secondary UCs could
entity. Such peripheries are primarily governed under rural adminis­ further enhance similar studies. Discerning how these limitations may
tration and the absence/poor regulation of construction guidelines be overcome can provide avenues for future research in this domain, so
under this governance structure encourages accelerated urban growth that the spatio-temporal attributes of urban growth may be better
(Jain, 2018; Simwanda & Murayama, 2018), as was evident during comprehended.

12
S. Chakraborty et al. Sustainable Cities and Society 67 (2021) 102715

6. Conclusion References

This study has proposed an innovative dynamic spatial territory-based Angel, S., Parent, J., & Civco, D. (2007). Urban sprawl metrics: An analysis of global
urban expansion using GIS. May. In Proceedings of ASPRS 2007 Annual Conference
approach to test how far popular hypotheses of urbanization apply to the https://www.asprs.org/wp-content/uploads/2011/01/0003.pdf.
KUA’s spatial patterns of growth. Results show that the high-density Asabere, S. B., Acheampong, R. A., Ashiagbor, G., Beckers, S. C., Keck, M., Erasmi, S.,
urban core became the most significant spatial built-up typology in et al. (2020). Urbanization, land use transformation and spatio-environmental
impacts: Analyses of trends and implications in major metropolitan regions of
the study area in the examined time period. Such growth of the city core, Ghana. Land Use Policy, 96, Article 104707. https://doi.org/10.1016/j.
particularly its expansion towards the periphery, primarily happened at landusepol.2020.104707
the expense of both urbanized and non-urbanized open spaces. The Bhatta, B. (2009). Analysis of urban growth pattern using remote sensing and GIS: A case
study of Kolkata, India. International Journal of Remote Sensing, 30, 4733–4746.
spatial territories present in the periphery were also the most dynamic. https://doi.org/10.1080/01431160802651967
Landscape metrics and mode-wise urban growth statistics computed for Bosch, M., Jaligot, R., & Chenal, J. (2020). Spatiotemporal patterns of urbanization in
these entities further revealed that urbanization patterns varied mark­ three Swiss urban agglomerations: Insights from landscape metrics, growth modes
and fractal analysis. Landscape Ecology, 35, 879–891. https://doi.org/10.1007/
edly across the broad spatial extents demarcated. While on the scale of
s10980-020-00985-y
the agglomeration we found clear evidence for a shift from diffusion to Broitman, D., & Koomen, E. (2015). Residential density change: Densification and urban
coalescence, zooming into the inner and outer extents as well as sub- expansion. Computers, Environment and Urban Systems, 54, 32–46. https://doi.org/
territories within these, we have shown that urban growth patterns 10.1016/j.compenvurbsys.2015.05.006
Chakravorty, S. (2000). From colonial city to globalizing city? The far-from-complete
remarkedly differ from the theoretically expected trends - the inner city spatial transformation of Calcutta. Globalizing cities: A new spatial order? (pp. 56–77).
zone underwent coalescence, the outer extents displayed complex Clement, M. T., Chi, G., & Ho, H. C. (2015). Urbanization and land-use change: A human
growth patterns with diffusion and coalescence occurring simulta­ ecology of deforestation across the United States, 2001–2006. Sociological Inquiry, 85
(4), 628–653. https://doi.org/10.1111/soin.12097
neously, but at varying intensities within the different sub-territorial Dadashpoor, H., & Ahani, S. (2019). A conceptual typology of the spatial territories of the
units. Thus, we argue that these multi-scale analyses are important to peripheral areas of metropolises. Habitat International, 90, Article 102015. https://
better understand urban growth trajectories and develop well-adjusted doi.org/10.1016/j.habitatint.2019.102015
Dadashpoor, H., Azizi, P., & Moghadasi, M. (2019a). Land use change, urbanization, and
spatial policies to respond to these intra-urban variations. change in landscape pattern in a metropolitan area. Science of the Total Environment,
In line with previous research, our results also suggest that the nor­ 655, 707–719. https://doi.org/10.1016/j.scitotenv.2018.11.267
mally accepted theories/modes of urban growth may need to be re- Dadashpoor, H., Azizi, P., & Moghadasi, M. (2019b). Analyzing spatial patterns, driving
forces and predicting future growth scenarios for supporting sustainable urban
examined since its dynamicity cuts across spatial and administrative growth: Evidence from Tabriz metropolitan area, Iran. Sustainable Cities and Society,
hierarchies/units. Such deviations likely arise from unplanned/poorly 47, Article 101502. https://doi.org/10.1016/j.scs.2019.101502
regulated urban growth, the mushrooming of secondary city centers, Dahal, K. R., Benner, S., & Lindquist, E. (2017). Urban hypotheses and spatiotemporal
characterization of urban growth in the Treasure Valley of Idaho, USA. Applied
conscious policy decisions to lessen the congestion burden on the central
Geography, 79, 11–25.
city and the expected densification of older built-up spaces. Undertaking Das, M., Das, A., & Mandal, A. (2020). Research note: Ecosystem Health (EH) assessment
such research for a larger number of UAs (and examining variations of a rapidly urbanizing metropolitan city region of eastern India–A study on Kolkata
according to city size, within and across nations) would thus better Metropolitan Area. Landscape and Urban Planning, 204, Article 103938.
Dietzel, C., Herold, M., Hemphill, J. J., & Clarke, K. C. (2005). Spatio-temporal dynamics
inform urban growth hypotheses and help frame more effective policies in California’s central Valley: Empirical links to urban theory. International Journal of
to govern such rapidly morphing locales in order to promote sustainable Geographical Information Science, 19(2), 175–195.
urbanization. Future studies must take into account the extent, character Dietzel, C., Oguz, H., Hemphill, J. J., Clarke, K. C., & Gazulis, N. (2005). Diffusion and
coalescence of the Houston Metropolitan Area: Evidence supporting a new urban
of the dynamic spatial territories and intra-urban variations in growth theory. Environment and Planning B, Planning & Design, 32(2), 231–246.
patterns present within a city to properly identify its spatiality and Dong, T., Jiao, L., Xu, G., Yang, L., & Liu, J. (2019). Towards sustainability? Analyzing
growth trajectory to reduce its adverse impacts on environement. changing urban form patterns in the United States, Europe, and China. Science of The
Total Environment, 671, 632–643.
Fahey, R. T., & Casali, M. (2017). Distribution of forest ecosystems over two centuries in
Funding a highly urbanized landscape. Landscape and Urban Planning, 164, 13–24. https://
doi.org/10.1016/j.landurbplan.2017.03.008
Fei, W., & Zhao, S. (2019). Urban land expansion in China’s six megacities from 1978 to
This research did not receive any specific grant from funding 2015. Science of the Total Environment, 664, 60–71.
agencies in the public, commercial, or not-for-profit sectors. Follmann, A., Hartmann, G., & Dannenberg, P. (2018). Multi-temporal transect analysis
of peri-urban developments in Faridabad, India. Journal of Maps, 14(1), 17–25.
https://doi.org/10.1080/17445647.2018.1424656
Declaration of Competing Interest
Haque, I., & Patel, P. P. (2018). Growth of metro cities in India: Trends, patterns and
determinants. Urban Research & Practice, 11(4), 338–377. https://doi.org/10.1080/
No. 17535069.2017.1344727
Haque, I., Rana, M.d. J., & Patel, P. P. (2020). Location matters: Unravelling the spatial
dimensions of neighbourhood level housing quality in Kolkata, India. Habitat
Acknowledgements International, 99, Article 102157. https://doi.org/10.1016/j.habitatint.2020.102157
He, Q., Song, Y., Liu, Y., & Yin, C. (2017). Diffusion or coalescence? Urban growth
Saurav Chakraborty and Priyank Pravin Patel are thankful to Presi­ pattern and change in 363 Chinese cities from 1995 to 2015. Sustainable Cities and
Society, 35, 729–739. https://doi.org/10.1016/j.scs.2017.08.033
dency University, Kolkata for all supports. The authors are particularly He, Q., Zeng, C., Xie, P., Tan, S., & Wu, J. (2019). Comparison of urban growth patterns
thankful to the USGS for providing the satellite images freely. We are and changes between three urban agglomerations in China and three metropolises in
particularly thankful to Marti Bosch for his valuable suggestions to un­ the USA from 1995 to 2015. Sustainable Cities and Society, 50, Article 101649.
Hu, X., Qian, Y., Pickett, S. T. A., & Zhou, W. (2020). Urban mapping needs up-to-date
derstand the area-radius relationship regarding urban spatial organiza­ approaches to provide diverse perspectives of current urbanization: A novel attempt
tion. Without this we may not able to compare our method with other to map urban areas with nighttime light data. Landscape and Urban Planning, 195,
popular approaches like area-radius relation. Authors are very grateful Article 103709. https://doi.org/10.1016/j.landurbplan.2019.103709
Jain, M. (2018). Contemporary urbanization as unregulated growth in India: The story of
to the editor and two anonymous reviewers for their valuable comments census towns. Cities, 73(October), 117–127. https://doi.org/10.1016/j.
and suggestions to upgrade the manuscript. cities.2017.10.017
Jenerette, G. D., & Potere, D. (2010). Global analysis and simulation of land-use change
associated with urbanization. Landscape Ecology, 25, 657–670. https://doi.org/
Appendix A. Supplementary data
10.1007/s10980-010-9457-2
Jiao, L. (2015). Urban land density function: A new method to characterize urban
Supplementary material related to this article can be found, in the expansion. Landscape and Urban Planning, 139, 26–39.
online version, at doi:https://doi.org/10.1016/j.scs.2021.102715. Kantakumar, L. N., Kumar, S., & Schneider, K. (2016). Spatiotemporal urban expansion
in Pune metropolis, India using remote sensing. Habitat International, 51, 11–22.

13
S. Chakraborty et al. Sustainable Cities and Society 67 (2021) 102715

Li, H., Peng, J., Yanxu, L., & Yi’na, H. (2017). Urbanization impact on landscape patterns Sharma, R., & Joshi, P. K. (2013). Monitoring urban landscape dynamics over Delhi
in Beijing City, China: A spatial heterogeneity perspective. Ecological Indicators, 82, (India) using remote sensing (1998-2011) inputs. Journal of the Indian Society of
50–60. Remote Sensing, 41(3), 641–650. https://doi.org/10.1007/s12524-012-0248-x
Li, C., Li, J., & Wu, J. (2013). Quantifying the speed, growth modes, and landscape Shaw, A., & Satish, M. K. (2007). Metropolitan restructuring in post-liberalized India:
pattern changes of urbanization: A hierarchical patch dynamics approach. Landscape Separating the global and the local. Cities, 24, 148–163. https://doi.org/10.1016/j.
Ecology, 28, 1875–1888. https://doi.org/10.1007/s10980-013-9933-6 cities.2006.02.001
Li, J., Li, C., Zhu, F., Song, C., & Wu, J. (2013). Spatiotemporal pattern of urbanization in Shi, Y., Sun, X., Zhu, X., Li, Y., & Mei, L. (2012). Characterizing growth types and
Shanghai, China between 1989 and 2005. Landscape Ecology, 28, 1545–1565. analyzing growth density distribution in response to urban growth patterns in peri-
https://doi.org/10.1007/s10980-013-9901-1 urban areas of Lianyungang City. Landscape and Urban Planning, 105(4), 425–433.
Liu, X., Li, X., Chen, Y., Tan, Z., Li, S., & Ai, B. (2010). A new landscape index for Simwanda, M., & Murayama, Y. (2018). Spatiotemporal patterns of urban land use
quantifying urban expansion using multi-temporal remotely sensed data. Landscape change in the rapidly growing city of Lusaka, Zambia: Implications for sustainable
Ecology, 25(5), 671–682. urban development. Sustainable Cities and Society, 39, 262–274.
Liu, Z., He, C., & Wu, J. (2016). General spatiotemporal patterns of urbanization: An Su, S., Jiang, Z., Zhang, Q., & Zhang, Y. (2011). Transformation of agricultural
examination of 16 world cities. Sustainability, 8, 1–15. https://doi.org/10.3390/ landscapes under rapid urbanization: A threat to sustainability in Hang-Jia-Hu
su8010041 region, China. Applied Geography, 31(2), 439–449.
McGarigal, K. (2015). FRAGSTATS help. Amherst, MA, USA: University of Massachusetts. Su, S., Xiao, R., Jiang, Z., & Zhang, Y. (2012). Characterizing landscape pattern and
Mitra, D., & Banerji, S. (2018). Urbanisation and changing waterscapes: a case study of ecosystem service value changes for urbanization impacts at an eco-regional scale.
New Town, Kolkata, West Bengal, India. Applied Geography, 97, 109–118. Applied Geography, 34, 295–305.
Mondal, B., Das, D. N., & Dolui, G. (2015). Modeling spatial variation of explanatory Taubenböck, H., Weigand, M., Esch, T., Staab, J., Wurm, M., Mast, J., et al. (2019).
factors of urban expansion of Kolkata: A geographically weighted regression A new ranking of the world’s largest cities—Do administrative units obscure
approach. Modeling Earth Systems and Environment, 1, 1–13. https://doi.org/ morphological realities? Remote Sensing of Environment, 232, Article 111353.
10.1007/s40808-015-0026-1 Tripathy, P., & Kumar, A. (2019). Monitoring and modelling spatio-temporal urban
Mondal, B., Dolui, G., Pramanik, M., Maity, S., Biswas, S. S., & Pal, R. (2017). Urban growth of Delhi using Cellular Automata and geoinformatics. Cities, 90(January),
expansion and wetland shrinkage estimation using a GIS-based model in the East 52–63. https://doi.org/10.1016/j.cities.2019.01.021
Kolkata Wetland, India. Ecological Indicators, 83, 62–73. United Nations. (2018). 2018 revision of world urbanization prospects. New York: UNDESA.
Peng, J., Wang, Y., Zhang, Y., Wu, J., Li, W., & Li, Y. (2010). Evaluating the effectiveness United Nations. (2016). The world’s cities in 2016. Data Booklet (ST/ESA/SER. A/392) htt
of landscape metrics in quantifying spatial patterns. Ecological Indicators, 10(2), ps://www.un.org/en/development/desa/population/publications/pdf/urbanizatio
217–223. n/the_worlds_cities_in_2016_data_booklet.pdf.
Peng, W., Wang, G., Zhou, J., Zhao, J., & Yang, C. (2015). Studies on the temporal and van Vliet, J. (2019). Direct and indirect loss of natural area from urban expansion. Nature
spatial variations of urban expansion in Chengdu, western China, from 1978 to 2010. Sustainability. https://doi.org/10.1038/s41893-019-0340-0 (Table 1).
Sustainable Cities and Society, 17, 141–150. Weng, Y. C. (2007). Spatiotemporal changes in landscape pattern in response to
Roy, A. (2005). Urban informality: Toward an epistemology of planning. Journal of the urbanization. Landscape and Urban Planning, 81(4), 341–353.
American Planning Association American Planning Association, 71, 147–158. https:// Wu, J., Jenerette, G. D., Buyantuyev, A., & Redman, C. L. (2011). Quantifying
doi.org/10.1080/01944360508976689 spatiotemporal patterns of urbanization: The case of the two fastest growing
Roy, A. (2011). Re-forming the megacity: Calcutta and the rural-urban interface 93–109. metropolitan regions in the United States. Ecological Complexity, 8(1), 1–8.
https://doi.org/10.1007/978-4-431-99267-7_5 Xu, C., Liu, M., Zhang, C., An, S., Yu, W., & Chen, J. M. (2007). The spatiotemporal
Roy, A. (2016). What is urban about critical urban theory? Urban Geography, 37(6), dynamics of rapid urban growth in the Nanjing metropolitan region of China.
810–823. https://doi.org/10.1080/02723638.2015.1105485 Landscape Ecology, 22(6), 925–937.
Sahana, M., Hong, H., & Sajjad, H. (2018). Analyzing urban spatial patterns and trend of Xu, G., Dong, T., Cobbinah, P. B., Jiao, L., Sumari, N. S., Chai, B., et al. (2019). Urban
urban growth using urban sprawl matrix: A study on Kolkata urban agglomeration, expansion and form changes across African cities with a global outlook:
India. Science of the Total Environment, 628, 1557–1566. Spatiotemporal analysis of urban land densities. Journal of Cleaner Production, 224,
Sengupta, U. (2006). Government intervention and public-private partnerships in 802–810.
housing delivery in Kolkata. Habitat International, 30, 448–461. https://doi.org/ Yu, X. J., & Ng, C. N. (2007). Spatial and temporal dynamics of urban sprawl along two
10.1016/j.habitatint.2004.12.002 urban–rural transects: A case study of Guangzhou, China. Landscape and Urban
Seto, K. C., & Fragkias, M. (2005). Quantifying spatiotemporal patterns of urban land-use Planning, 79(1), 96–109.
change in four cities of China with time series landscape metrics. Landscape Ecology, Zhu, Z., Zhou, Y., Seto, K. C., Stokes, E. C., Deng, C., Pickett, S. T. A., et al. (2019).
20(7), 871–888. Understanding an urbanizing planet: Strategic directions for remote sensing. Remote
Sensing of Environment, 228(October 2018), 164–182. https://doi.org/10.1016/j.
rse.2019.04.020

14