Abstract
Urban–rural differences seem particularly pronounced in India, especially when based on the official figures provided by the Census of India, which are heavily dependent on the administrative status of settlements. India, one of the world’s most dynamic and populous countries, still possesses an official urbanisation rate lagging well behind other developing economies. To investigate the extent of Indian urbanisation, this article develops a multi-step methodology using indicators specifically conceived for identifying urban structures in India. In this article, an emphasis is given to the conception and to the spatial analysis of two indicators: metropolitan ranking and meta-agglomerations. A method combining these indicators then allows identifying urban macro-structures acting as a larger organising framework in the regional space. Our results show a multitude of different functional areas that have developed specific urban morphologies over time. Some are particularly marked by high values of urban macrocephaly, small settlements taking the shape of nebulae, urban sprawl, etc.
Introduction
India is a large geographic entity possessing an impressive demographic growth (from 2001 to 2011, there was an increase in inhabitants of 181 million). In this respect, megacities in India suffer from rapid and ongoing urbanisation processes often resulting in overcrowding, infrastructure saturation, an increase in the number of people living in slums, etc. (UN-Habitat, 2001). Apart from the well-known Indian metropolitan cities (overrepresented in urban studies), this country also possesses an impressive network of small and mid-sized settlements (605,413 villages and towns in 2011) which are expected to accommodate an increasing share of the demographic growth. The Indian Republic is, above all and in all its aspects, a geographic space characterised by strong urban disparities. The Indian settlement structure is not a human-made single-purpose system. It is the result of several centuries of evolution in which India’s society has moved from a basically rural civilisation to a (partially) globalised and multi-layered socioeconomic compound. This country was politically and administratively unified only in 1947, incorporating urban and urban/rural systems that had previously grown separately. These disparities are the outcome of an old and complex history (Ramachandran, 1989), and we recall that the older a civilisation is, the more human settlements are articulated at all scales (villages, small and mid-sized cities, big cities) pervasively over geographic space (Moriconi-Ebrard, 1993). The ongoing urban transition happening within a country expected to become the most populous worldwide by 2022 (United Nations, 2015) raises the question of the location of the future urban dwellers.
In 2011, the official urbanisation rate of India only stood at 31.16%. The main difference between the previous censuses and the 2011 round is an absolute population growth more important in urban than in rural areas (with a gap of one million). Given that urban areas have been the dominant form of human habitat in the world since 2008 (United Nations, 2012), one might wonder: how is it possible that one of the world’s most dynamic and populous countries possesses an urbanisation rate lagging well behind other developing economies? 1 A major problem for urban studies is that official urban population figures depend on administrative census agencies (Moriconi-Ebrard, 1993). Each country produces and uses its own definitions and criteria to define what should be considered as urban, a process which may lead to underestimation or hyper-projection of the urbanisation phenomenon.
From this perspective, the aim of this article is to identify urban structures in India with no regards to the urban–rural official classification. The assumption is that despite the very low official urbanisation rate, it should be possible to identify and extract different urban structures acting as larger frameworks in regional space (urban macro-structures). The development of a new methodology maximising the use of urban morphological criteria and linking the official population count (the census) to the identified structures should shed new light on the current levels of urbanisation in India. The text of the article is organised as follows.
The first section addresses the complexity of urban transition and the identification of urban space in India. ‘Data’ presents the demographic and spatial sources (Census of India and e-Geopolis) used in this research. The third section presents the application of two methods allowing the identification of rank-1 metropolitan areas and the extraction of meta-agglomerations. ‘Identifying Indian urban macro-structures’ links the two methods in order to identify urban structures operating at a macro scale. The fifth section presents the external and internal features of the largest identified urban macro-structures. A final section concludes the article.
Problem setting
From a spatial and economic point of view, metropolises have become more and more connected to their broader hinterlands. Settlements surrounding the urban cores are playing an increasing role in structuring this wider space, thus creating integrated metropolitan regions. Mention can be made of America’s Northeast Megalopolis, a macro-structure located in the north-eastern seaboard of the United States and described as a chain of adjacent important urban centres (Gottmann, 1961). In such spaces, agriculture is operating in conjunction with more heavily urbanised spaces. Ascher (1995) coined the term ‘metapolis’, to put the accent on macro-structures that incorporate any component (urban or rural) into the daily functioning of metropolitan areas. Large cities are indeed recomposing themselves within a long-term cycle that gradually brings new centralities. The constitution of ‘metapolises’ should then be considered as an irreversible process since communication and exchanges are gradually intensifying.
This article extends this line of research by identifying macro-structures acting as a larger organising framework of urbanisation in India. The study of the Indian settlement structure is a complex challenge since this geographical area is composed of a wide range of different ethnic groups, languages, human and natural landscapes, etc. It appears that peculiarity and diversity are what give consistency to India as a single geographic entity. It is true that regions in India experienced urban transformations following various patterns that, according to Raman et al. (2015), defy a singular explanation. The urban transition cannot be grasped through a dual model opposing urban to rural, or metropolitan areas to small towns (Denis and Marius-Gnanou, 2011). Some authors point out, for example, dynamics leading to a progressive readjustment of the urban system with an overall increase in the number of cities of more than 100,000 inhabitants (Querci and Oliveau, 2015) or the increasing share of people living in settlements between 10,000 and 100,000 inhabitants (Raman et al., 2015). Others emphasise the degree of coherence of the urban structure within specific subspaces (Fusco and Perez, 2015; Oliveau, 2005, etc.), the importance of regional specificities as a key factor in growth distribution (Swerts et al., 2014) or the variation in growth according to urban centre size (Kundu, 2011). One trend is certain – deep and rapid changes are currently occurring within the urban hierarchy due to several co-existing realities such as the overall development of rural areas and the massive waves of migration towards urban areas. Yet, as we shall discuss, it is impossible to obtain a comprehensive picture of the urban processes occurring in India at a macro scale of analysis using only the census results.
In India, a settlement is categorised as ‘urban’ (i.e. a town) only if it satisfies the following three criteria: a minimum population of 5000 inhabitants, a minimum proportion of 75% of males employed outside agriculture and a population density of at least 400 inhabitants per square kilometre within the administrative area. The relevance of the urban classification operated by the Indian censuses is heavily criticised (Bhagat, 2005) since, based on such criteria, the urban population of India may be underestimated. For example, specialised agriculture can generate urban patterns (e.g. forestry; Ramachandran, 1989), villages may resist becoming officially urban in order to preserve state financial aids (Denis and Zérah, 2014), clusters of villages may generate urban agglomerations (Denis and Marius-Gnanou, 2011) and villages recently reached by the spillover of larger urban areas are still considered as rural. Accordingly, the number of official villages with more than 10,000 inhabitants surpassed the number of official towns in 2001 (Marius-Gnanou and Moriconi-Ebrard, 2007). Moreover, Sivaramakrishnan et al. (2007) point out that state governments are also granting the town status following political decisions, 2 regardless of the fulfilment of any criteria. Last but not least, urbanisation is officially unable to cross administrative borders (including the municipal ones), a resolution resulting in a constant reshaping of boundaries at every census. When urban sprawl crosses a boundary without reaching and being accounted for within another official settlement, these populations are counted as new units called ‘Out Growths’ (OG). There is a specific status, ‘Urban Agglomeration’, that aggregates both the towns that have physically merged and their related ‘OG parts’. However, the ‘Urban Agglomeration’ status is registered as a double counting within the census and does not integrate the villages which are physically merged with the agglomerations. There were 475 of these agglomerations made from at least two settlements detected by the government in 2011. Given all the aforementioned, it should be possible to find areas strongly urbanised in non-official towns and vice versa.
Data
The administrative borders used to perform the census surveys rarely match the physical borders of urban areas. As a result, if special precautions are not taken, comparing cities using raw data from official censuses can lead to comparing random urban space fragments. To avoid this statistical bias, this research relies on two sources of information: the Census of India and the e-Geopolis database (Moriconi-Ebrard, 1994). Several authors agree that the Indian population count is of an outstanding quality (e.g. Miranda, 1982; Oliveau and Guilmoto, 2005). For this reason, and from a strictly demographic point of view, this research makes use of the Indian Census. The e-Geopolis database, for its part, contains the digitised built-up areas of the world. In this database, every continuous built-up area surpassing 10,000 inhabitants is a statistical object considered as an ‘urban’ geographic entity, regardless of the official status of the locality, and whatever the criteria of national censuses are. The two e-Geopolis criteria, built using the recommendation of the United Nations (Department of Economic and Social Affairs, ESA) for the 1980 census round, are:
– A continuous urban area must have at least 10,000 inhabitants. – The distance between two built-up areas should not exceed 200 metres.
The built-up areas are digitised using polygons in a GIS software database and captured from maps, aerial photographs and high definition satellite images. Demographic data are collected from worldwide national censuses (in this case the Census of India) and then linked and aggregated to the polygons. In order to respect the characteristics of the Indian urbanisation (composed of a great nebula of settlements), the e-Geopolis population criteria have been dropped down to a minimum threshold of 5000 inhabitants for this research (instead of the usual 10,000 inhabitants). India is one of the last developing economies of the world where no settlement-based official GIS gazetteer exists. Consequently, there is no way to study the evolution of the urban population from a spatial point of view using only the Census of India. The geo-data which were used in e-Geopolis have thus been provided by a private Indian company, MapMyIndia. Only the settlements counting more than 2000 inhabitants were provided and this threshold was based on the 2001 census. Thus, an e-Geopolis urban area in India is composed of official settlements hosting at least 2000 inhabitants in 2001, as long as the overall built-up area hosts at least 5000 inhabitants. Finally, the overall database gathers 18,366 built-up areas digitised as polygons themselves containing 29,209 official settlements (official villages and towns).
Figure 1 shows the distribution of e-Geopolis Urban Areas (UA) at the national level within a district, and the distribution of official settlements within the main UA of this district. According to e-Geopolis, the entire continuous built-up area displayed here is composed of 17 different settlements accounting for 965,360 inhabitants in 2011, while the census office acknowledges only 874,408 inhabitants. For the metropolitan areas, the gap is often more significant (Chennai shows, for example, a gap of more than 2 million inhabitants).
The aggregative process of official settlements within an e-Geopolis Urban Area: The example of Aligarh.
Overall, the e-Geopolis urbanisation rate stands at 45.31%, 3 thus showing a difference of +14.15% with the government official rate. E-Geopolis also detects 2991 UA of at least two settlements (as compared with 475 agglomerations for the government; see previous section).
Methodology
This section is divided into two parts. First, the metropolitan ranking of the main urban areas is calculated using a scoring system. Second, rural spaces and urban corridors complementing the urban areas are extracted through a GIS procedure. Once combined, the results of these two methods allow identifying urban macro-structures (see below).
Metropolitan ranking of main urban areas
Metropolitan areas (large cities highly connected to the world economy) are the entry points for new global trends and the export hubs for locally adapted products and services, within cycles becoming shorter over time. As discussed previously, it is probable that the most intense growth will be hosted in small and mid-sized cities in the near future. Yet, the weight of the megacities within the Indian settlement structure cannot be ignored as today’s largest metropolitan areas will remain at the top of the urban hierarchy even if adjustments are made over time. Megacities act as cores, strengthening the structure of the urban systems on a broader scale and giving growth opportunities to small and mid-sized cities connected to them. Hence, defining and locating the main megacities becomes a prerequisite to finding their related urban macro-structures. This brings the following question: how can the metropolitan importance of the Indian urban areas be quantified? In addition to the urban–rural distinction, the Indian government is also classifying its settlements by using a population threshold (Class I: 100,000 and above; Class II: 50,000 to 99,999; etc.). However, when studying the influence of a city, several additional factors beyond population count should be considered. This is especially true for the metropolitan areas, which usually comprise a broad range of activities. To evaluate the weight of India’s megacities, the urban areas of at least 200,000 inhabitants in 2011 were selected (190 UAs). Afterwards, a multi-criteria scoring system was implemented. The aim was to obtain a ranking among the biggest urban areas using indicators that best describe the functional potential of urban areas. Subsequently, points were awarded to the 190 UAs based on the indicators in Table 1.
Allocation of marks for the 190 UAs of more than 200,000 inhabitants.
Notes: aGDP figures are usually only available at the state level in India, hence the acronym GSDP for Gross State Domestic Product. However, some of the former Planning Commissions released in a sporadic way the GSDP of their respective states at the district level. From this perspective, the most complete database we could obtain access to at the district level was for 2004–2005.
This method is inspired by previous works conducted on the functional classification of European cities (Brunet, 1989; Rozenblat and Cicille, 2003). Through an extensive battery of indicators, these authors covered important aspects characterising the international reach of European metropolitan areas. In this research, only a classification of India’s megacities is needed (in order to subsequently identify their related urban macro-structures). From this perspective, the multi-criteria scoring method can be considered as an adaptation made of a reduced number of indicators chosen through in-depth literature review. The following aspects are covered: overall demographic weight (population); economic modernity (workers within service and industry); potential growth of urban areas (construction GDP); and the weight of the tourism industry (Trade, Hotels & Restaurants GDP). Airport flows and the number of well-established universities are indicators related to the national and international influence of a megacity. As for specialised economic activities, bonus points were awarded to the UAs possessing an important commercial port as well as to those possessing at least 100 acres of Special Economic Zones (SEZ). The maximum possible score for an urban area is 26. A final discretisation of this score returns a ranking of five classes.
Figure 2 shows the five metropolitan levels arranged in descending order in terms of importance (from one to five). The maximum possible score for an urban area is 26 and the highest observed value is attributed to New Delhi with 24 points. The Indian administrative capital is followed by six senior centres that have 23 or 22 points (Kolkata, Mumbai, Hyderabad, Chennai, Bangalore and the Kerala urban agglomeration). These first seven major urban areas have been identified as rank-1 metropolitan areas. There is a difference of more than two points with Ahmadabad, the first rank-2 urban area.
The metropolitan areas in India ordered by rank.
The spatial distribution of these rank-1 metropolitan areas shows some patterns established during the time of the British Raj: i) Delhi, which had served as a capital of various kingdoms and empires, was structurally reinforced when it became the capital of the British Raj in 1911; ii) the three major port cities served as import/export platforms for manufactured goods and raw materials – Mumbai, Chennai and Kolkata.
The dynamic cities of southern India that were historically capitals of their own kingdoms before the British Raj also stand out: Bangalore in Karnataka and Hyderabad in Andhra Pradesh. The southern state of Kerala is added to this familiar pattern. The urbanisation processes along the coast have led to a continuous urban area from Kozhikode to Thiruvananthapuram (Figure 4). It is important to note that the differentiation between urban and rural is difficult here since this space possesses some similarities with a Desakota 4 (Denis and Marius-Gnanou, 2011; Pauchet and Oliveau, 2008). Unfortunately, the e-Geopolis database does not allow distinguishing of sub-spaces of higher or lesser density within a single urban area. Despite this difference in nature from the other rank-1 metropolitan areas, the Kerala UA qualifies as rank 1 within our multi-criteria analysis and will be considered as such within this research.
Meta-agglomerations extraction
The growth of the industrial sector, particularly intense after the reforms of the 1980s, contributed both to the expansion and reinforcement of urban corridors. The spaces not included within an urban area are neither homogeneous nor static. Proximity to a built-up area could be an important factor of dynamism, with spaces benefiting from positive externalities related to this particular environment. In order to identify the rural spaces and the urban corridors complementing the built-up areas and thereby forming larger settlement structures, a new concept named ‘meta-agglomeration’ (M-A) has been developed through an operational method aiming at extracting urban ‘envelopes’ encompassing urban areas.
To extract these specific patterns, a morphological closing was carried out in the GIS environment: a 5-km dilation of the urban areas followed by a 5-km erosion results in meta-agglomerations linking urban built-up areas closer than 10 km. Subsequently, a double filter was applied in order to remove automatically:
- The smallest and isolated M-As with a surface smaller than 78.5 square kilometres, which is the lower limit of the surface artificially generated by the 5-km dilation algorithm. This filtering also removes the small-sized and isolated urban spaces (40–70 km2).
- M-As with too low a level of urban compactness and thus possessing no real urban cores (they are artificially generated by the closing algorithms from numerous very small but not too far away built-up areas). They have been detected by removing the meta-agglomerated spaces which have not been detected by the 2.5-km algorithm (morphological closing filtered by a 19.3 km2 threshold) from the 5-km algorithm results.
In practice, several buffers and filters have been tested (2.5 km, 5 km, 7.5 km, 10 km). As shown in Figure 3, the selected distance of 5 km connects without gaps the major urban corridors, while this is not the case for the 2.5-km algorithm. 5 These macro-structures can take the shape of large patches or huge corridors, which deeply structure the Indian urban system. They contain most of the Indian urban areas and almost all the major cities. Figure 3 also shows the well-known corridor from Coimbatore to Salem in southern India, going through Erode and Tiruppur. This corridor connects many urban areas into a single set and stretches over 140 kilometres.
Meta-agglomerations extraction using a morphological closing.
Identifying Indian urban macro-structures
Since urbanisation and growth often reinforce each other, beyond a certain threshold the urban sprawl may extend across the administrative borders even on a regional scale. But most of the time, the administrative perimeters keep their own logic in terms of governance. This leads to a pattern in which strategies adopted by households and firms are influenced both by administrative borders (Districts and States) and by the physical realities of urbanisation. In order to define physical structures operating at a macro scale, the spatial extent of the meta-agglomerated spaces was characterised by the number of districts encroached upon. In other words, for each M-A, the number of concerned districts crossed into has been extracted to identify urban macro-structures. Sometimes, a district is concerned with more than one M-A. In those cases, its related macro-structure is the one with the largest M-A footprint encroaching upon its administrative area. Macro-structures are thus defined as the set of contiguous administrative districts encroached upon by a single M-A. It is thus possible to extract the administrative context supporting the Indian settlement structures.
Overall, 261 M-As and 50 macro-structures have been identified; 37.3% of the observations (i.e. 239 districts) are not affected by any M-A and therefore are not part of any macro-structure. The size of the macro-structures varies from one to 54 districts. Figure 4 shows the macro-structures related to the seven rank-1 metropolitan cities identified in the previous section. 6
The macro-structures related to the seven identified rank-1 metropolitan areas.
These seven macro-structures are home to almost 380 million inhabitants and cover 134 districts (Table 2). They can be divided into two categories: those possessing a huge hinterland (Kolkata–Bihar, the Kerala–Tamil Nadu corridors, Delhi), and the rest (Mumbai, Chennai, Hyderabad and Bangalore). Within the latter, important metropolitan areas seem disconnected from the main urban corridors. This is the case, for example, of Bangalore, which possesses a small macro-structure (4562 km2) containing only 20 urban areas.
Size of the macro-structures related to rank-1 metropolitan areas.
Notes: aAggregated population of the districts constituting the macro-structures.
On the other hand, the larger macro-structures possess a lot of small and mid-sized urban areas. They are roughly comparable to Europe’s ‘Blue Banana’ (Brunet, 1989) and to America’s Northeast Megalopolis (Gottmann, 1961). High population densities, typical of the Indian settlement structure, characterise these macro-structures, as they host a total population equal to that of the United Kingdom, Russia, Germany and France combined, on a surface nearly equivalent to Germany’s. The largest one is the Kolkata–Bihar macro-structure composed of 54 districts. This macro-structure follows the Ganges River from the state of Bihar to the Bay of Bengal. The megacity of Kolkata does not occupy a central place within this macro-structure since this city is located at the end of its south-eastern section. It is followed by Delhi and the Kerala–Tamil Nadu corridors macro-structures, stretching both to over 34 districts. The Kerala–Tamil Nadu corridors are mainly located in Kerala, and stretch north to south along the coast. This macro-structure also includes one main corridor and two smaller ones connecting inland Kerala and Tamil Nadu. The main one goes through Coimbatore to Salem (Figure 3). The smaller ones connect Kottayam to Dindugu and Kollam to Madurai. The Delhi macro-structure is related to the administrative capital of New Delhi. It possesses the same number of districts as the Kerala–Tamil Nadu corridors, but this is made easier by the fact that the core of the administrative capital is divided into nine small districts. The hyper-centre of New Delhi is located at the centre of the Delhi macro-structure and serves as a major crossroads distributing flows through seven main axes.
Characteristics of the main macro-structures
It is interesting to compare the three largest macro-structures in terms of shape, intensity and compactness of urbanisation in order to understand the urban–rural relationship at a regional scale. India’s three largest macro-structures seem to have diverging or even opposed features. The extent of these macro-structures (Table 2), coupled with their descriptive characteristics (Table 3), hint at three different kinds of urban structures.
Characteristics of the three biggest macro-structures.
Notes: aShare of urban areas’ surface within the macro-structures. bUrban areas’ ratio for 100k inhabitants. cShare of M-A surface within the macro-structures. dShare of urban areas’ surface within the M-As. e2011 population density within urban areas (inhabitants/km2). fShare of secondary and tertiary workers within the active population in 2011 (Census of India, 2011).
The Kerala–Tamil Nadu corridors macro-structure is characterised by an important urban area footprint and a relatively low urban density (by Indian standards). At first glance, these features confirm the concept of a Desakota, i.e. a place where the distinction between urban and rural space is questionable. Yet, this pattern is only partially true and needs a more attentive characterisation. Indeed, there is a duality with two different settlement structures that are linked together within this macro-structure. The first one is a major urban structure covering nearly the whole of Kerala and characterised by very low urban density and extremely high compactness (most of the M-A surface is covered by the urban area footprint). Small and mid-sized cities play the important role of ‘cement’, linking the main urban areas and thus ensuring a continuity within a functional framework. Along with bigger cities, they are part of a large-scale structure, operating as a single entity (from Kozhikode to Thiruvananthapuram). There is no single factor that can explain the appearance of a Desakota in Kerala. However, from a historical perspective, Firoz et al. (2014) point out that Kerala was ruled by more than 100 principalities before the British colonisation and that the confluence of exogenous factors prevented the establishment of a unified kingdom. Also worthy of mention is the fact that several European countries established trading stations in Kerala, and that the economic model of this period implied little contact between these locations. From this perspective, an urban structure already well established by the past associated with missing ‘cement’ could be a plausible assumption. As a matter of fact, the Desakota term is only half-relevant in this case, since a large and dominant city cannot be found and the urban density is much weaker than in a traditional Indonesian Desakota (McGee, 1991). The second settlement structure part of the Kerala–Tamil Nadu corridors macro-structure is found in a handful of urban corridors of different sizes connecting Kerala with Tamil Nadu. These corridors are mostly made up of dynamic mid-sized settlements located along the main roads. The meta-agglomeration compactness is much lower here (Share UA within M-A) than on the coastal part of Kerala, lowering the total compactness value for the Kerala–Tamil Nadu corridors macro-structure to 38%. For the whole macro-structure, the rate of secondary and tertiary workers within the active population reaches 72%. Despite a great number of small and mid-sized settlements and the inclusion of rural and natural areas in the meta-agglomerated space, agriculture is far from being the leading sector. Several decades ago, Chattopadhyay (1988) pointed out the special relationship of this area with the tertiary sector and the merging process of the urbanised areas. This provides further evidence of a functional area where the distinction between urban and rural patterns becomes problematic, especially along the coastline.
The Bihar–Kolkata macro-structure is characterised by a proliferation of small towns and big villages which are growing rapidly (the 2001–2011 decadal growth of Bihar and the West Bengal states reaches 20%). The strong population growth of the previous decades led to a progressive inclusion of these settlements into an oversized meta-agglomeration (Figure 4). The urban fragmentation ratio (Table 3) is nearly the same as in the Kerala–Tamil Nadu corridors macro-structure (respectively 1.91 and 1.95 per 100,000 inhabitants), but the population and the number of urban areas are much higher. Once again, two different models can be found within this macro-structure. The first model is located in the south of the macro-structure and is mostly characterised by a gigantic and very dense built-up area: Kolkata and its more recent urban sprawl. The second model follows the Ganges River and is mainly located within the Bihar state and in the northern part of West Bengal. The settlements within this model are much older since some of them have a history stretching back to the Vedic age (1700 BCE to 500 BCE), a period that saw the growth of a significant number of towns in this area (Prakash, 2005). Despite the spatial rearrangements that took place during the rise and fall of countless dynasties, some still-existing settlements played a historic role in structuring the urban structure on a macro scale. For example, Pātaliputra, the capital of the Mauryan Empire (300 BCE), was located on the same site as Patna, the current state capital of Bihar (Figure 4). We thus have a very old settlement structure characterised by a large number of small and mid-sized settlements, one close to the other, that were already in place long before the British Raj. These settlements are, de facto, forming built-up areas composed mostly of clusters of villages and small towns. Moreover, urban area coverage is less significant here (6.08%) than in the Kerala–Tamil Nadu corridors model (14.32%), but is nonetheless combined with a much more pronounced urban density (9714 vs 3569 inhabitants/km2), thus highlighting the clustering dynamic of this area. The aforementioned also suggests that there are ongoing rural-to-urban migrations feeding these clusters of dense, traditional (low rate of secondary and tertiary workers), small and mid-sized cities and big villages that do not present the sprawling characteristics observed in Kerala (and eventually around Kolkata). The persistence of agricultural activities within this macro-structure has led to an underestimation of the urbanisation phenomenon by the government (Census of India, 2011; 81% of rural population in Bihar), despite strong physical evidence of ongoing urbanisation dynamics. The literature suggests that the pressure on agricultural land in Bihar is concomitant to rural–urban migrations (Gupta, 1995). This leads us to conclude that clusters of small and mid-sized settlements have a role just as important concerning urbanisation processes as the official urban towns detected by the census.
The third biggest macro-structure is located around the Indian capital territory, New Delhi. It contains fewer built-up areas and a smaller M-A footprint than the aforementioned macro-structures. Yet, a lot of small and mid-sized settlements have already been incorporated into Delhi’s built-up area. The meta-agglomeration is structured by the presence of the gigantic Delhi agglomeration, accompanied by settlements gravitating around this centre. This is a metropolitan area in the conventional sense, with a megacity located at the core of an urban structure. The weaker urban fragmentation (ratio of urban areas per 100,000 inhabitants) and the very high urban density also show the prominence of a metropolitan core within the meta-agglomeration. This macro-structure can be described through a spatial model derived from central place theory since the weight of the biggest UA is even overwhelming the weight of the second biggest one (macrocephalic urban system; Berry, 1964). The small and mid-sized cities seem to be totally dependent on their proximity to the capital city. However, these settlements and the core feed each other, thus forming an interdependent system. By comparison, such an intense relationship between a megacity and a huge hinterland does not apply to Bangalore, Chennai, Mumbai or Hyderabad, where the megacities structure relatively smaller meta-agglomerations and macro-structures.
Conclusions
The methodology developed in this research turned out to be a good alternative to explore the urbanisation processes operating at a macro scale in India, despite the lack of official spatial information. In this research, urban spatial structures have been grasped through the use of both census data and morphological criteria. Coherent spatial structures have successfully been identified, from the smallest urban spatial objects (official settlements) to the urban macro-structures acting as larger organising frameworks in the regional space. Villages, small and mid-sized cities, urban areas, metropolitan areas, meta-agglomerations and finally macro-structures are the catalysts of India’s multi-faceted socioeconomic life. Indeed, our results show that urbanisation in India can take many different forms and is distributed in a heterogeneous way. From this perspective, macro-structures related to rank-1 metropolitan areas have proved to be a good entry point to identify sub-spaces that have a strong urban consistency while focusing only on the uppermost part of the urban hierarchy. There are different ways of ‘being urban’ for a given area in India (encroachment in UA footprint, M-A footprint, etc.), as well as different ways of functioning in an urban environment (share of secondary and tertiary workers within the working population, UA ratio for 100,000 inhabitants, etc.). At a macro scale of analysis, different functional areas that have developed complex and specific urban morphologies over time have been identified.
The macro-structures studied in this article will certainly be a major source of change in the near future, since they are less static than the deep rural space (migration, urban growth, intensive land use, etc.). Hence, the small and mid-sized cities that are part of a macro-structure will probably evolve differently from the ones located in the hinterland, far from the influence of metropolitan centres. As a comparison, the deep rural areas not directly under the influence of a large or medium urban space have been decreasing in countries of old industrialisation (Ascher, 1995). In any case, the role of small and mid-sized cities should not be underestimated. Generally, they are relegated to the rank of suburban areas within megacities or deep rural areas, when in fact the role and importance of these settlements differ according to the identified structure to which they belong.
Taken together, many small and mid-sized settlements with lower spacing between them can form a very impressive settlement structure, such as within the Kerala–Tamil Nadu corridors’ macro-structure. Within this spatial structure, urban sprawl is important and associated with a rather weak density. Despite the fact that this structure is not connected to any megacity, it remains nonetheless very dynamic, as witnessed by its economic modernity. The Bihar–Kolkata macro-structure is retaining some of the traditional features of Indian society, with an important rural heritage, a strong density and urban fragmentation concentrated on a relatively reduced urban surface. The Delhi macro-structure seems to operate in a more conventional sense, with settlements gravitating around the metropolitan core. Conversely, some megacities are part of a small urban structure (Hyderabad, Mumbai and Chennai) and even seem sometimes to be disconnected from their environment (Bangalore). In this article, careful consideration has been given to the macro-structures containing the most influential components of the urban system: the rank-1 metropolitan areas linking India to the world economy. Large morphological structures not containing rank-1 metropolitan areas have not been analysed, and deserve further investigation.
The method presented in this article allowed the identification of urban macro-structures but the results remain nonetheless silent on intra-urban density distributions. Knowing the role of intra-urban densities within metropolitan areas, this research is presently developing in several directions including more detailed analysis of urbanisation phenomena and urban forms. The reproducibility of these analytical methods will also enable the extraction of consistent urban phenotypes and macro-structures for other large emerging countries. A further direction of research is the implications of these findings on Indian urbanisation and on more general issues of India’s regionalisation in the age of globalisation. A complete clustering of Indian districts has, for example, been performed using a multi-dimensional dataset including the urban indicators presented in this article (Fusco and Perez, 2015).
Footnotes
Acknowledgements
Within the Renault group, the authors would especially like to thank Jean Grebert who greatly assisted the research. The authors would also like to thank the reviewers for providing many constructive comments and suggestions.
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 research was partially supported by a CIFRE contract between ESPACE and Renault (Industrial Agreements for Training Through Research funding).
