1 s2.0 S1110982317300054 Main
1 s2.0 S1110982317300054 Main
1 s2.0 S1110982317300054 Main
Research Paper
a r t i c l e i n f o a b s t r a c t
Article history: Land in the peri-urban interface is of vital importance as there is a lack of clarity in the planning and poli-
Received 9 August 2016 cies in this region whether they should be governed under the authority of rural administration. In gen-
Revised 23 November 2016 eral, these areas are frequently ignored as a specific area in the study of urbanization being neither pure
Accepted 5 January 2017
urban nor pure rural. In this regard, an attempt has been taken to detect the spatio-temporal dynamics of
Available online 19 January 2017
sprawl, the nature of land transformation taken place in the identified peripheral settlements with high
probability of urban development within the buffer zone of 2.5 km around the boundaries of the two
Keywords:
major urban bodies of Malda i.e. English Bazar municipality an Old Malda municipality. The secondary
Urban sprawl
Lateral spread
data of Census of India of 2011and Landsat TM imageries (1987, 2003, 2011 and 2015) have been
Encroachment employed to detect the evolutionary process with the view to capture the urban land-use growth by
Shannon’s Entropy Land Use–Land Cover classification. The result shows that the built-up area has increased around 30%
Built-up density over the course of 28 years span. Computations of Shannon’s Entropy, Urbanisation intensity index,
built-up density have helped to explore the degree of transition taking place in each spatial unit under
investigation. The spatial metrics with the gradient approach has revealed that beyond the buffer of
7 km from the city centre the increase of residential land uses is maximum which is characterised by a
fragmented pattern with nearest entropy value to log N. The suitability analysis in search of most
prospective future urban centre suggests the maximum distribution of urban units in the western front
of the ULBs.
Ó 2017 National Authority for Remote Sensing and Space Sciences. Production and hosting by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-
nd/4.0/).
1. Introduction formed the adjoining rural hinterland into semi urban or peri-
urban areas. Contemporary India’s urban policies and programmes
Before 1990s the growth of Indian cities was primarily driven (such as JNNURM and of more recent one like Smart City Project)
by large scale migration from rural to urban areas and natural have supported metropolitan based ‘polarized growth’. Lateral
growth of population. But the growth of urban population largely extension of the Indian cities has attracted attention of scholars
confined within the territorial limit of the city resulting in over- very recently (Shaw, 2003). The outward expansion of the cities
crowding of these urban areas and sprung of informal settle- beyond their territorial limits has resulted in mixed land uses
ments/slums. This pattern of urban growth changes significantly which are neither rural nor urban. These transition areas are pop-
after the economic liberalization, since then outward expansion ularly known as peri-urban. This terminology is derived from the
of the existing cities emerged as the major component of urban word ‘peripheral’. Effectively, these words also convey meanings
growth. As a consequence of the intense outward expansion of of being less important, incidental to main activities, outer edge,
the cities rapid changes in land use and occupation have trans- fringe to the main, spillover or over flown. Conceptually ‘Periurban’
has come to serve as a term to denote the intermediary zone
between the ‘rural’ and the ‘urban’, that is, a geographical space
Peer review under responsibility of National Authority for Remote Sensing and where the rural meets the urban (Narain et al., 2013).
Space Sciences. Recently, peri-urban areas of Indian metros (i.e. million plus
⇑ Corresponding author.
cities) extensively studied in the Indian literature although the
E-mail addresses: [email protected] (R. Shaw), [email protected]
peri-urban areas are not exclusive to metros only. The small and
(A. Das).
http://dx.doi.org/10.1016/j.ejrs.2017.01.002
1110-9823/Ó 2017 National Authority for Remote Sensing and Space Sciences. Production and hosting by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
160 R. Shaw, A. Das / The Egyptian Journal of Remote Sensing and Space Sciences 21 (2018) 159–172
medium sized towns are also playing a vital role in the process of affect their ecological integrity. Temporal analysis of spatial pat-
peri-urbanization in India but are not taken into account in the lit- tern of urbanization helps the municipal administration and town
erature. In some medium and small cities this procedure is rapidly planners to visualize and understand the growth of the ULBs so
expanding dependent on their location and specific characteristics that they can provide better resource planning to create a sustain-
(Dadras et al., 2015). The Peri urban growth of Indian cities results able urban area (Mosammam et al., 2017).
in sprawl by which the rural land uses are completely or partially
transformed into urban uses. The sprawl is occurring at an 2. Justification for the selection of the study area
unprecedented rate but in unregulated and unplanned manner in
response to intense peri-urbanization. From this perspective urban Located between latitudes 24°400 2000 N and 25°320 800 N, and lon-
sprawl can be visualized as dispersed growth of settlements in the gitudes 87°450 5000 E to 88°280 1000 E with an area of 3733.66 square
Peri urban areas. This rapid sprawling demanded immediate inter- kilometres (1441.6 sq mi) see Fig. 1 Malda is a district of West Ben-
ventions to mapping and monitoring by employing advanced gal with low level of urbanization (13.6%) against state’s average of
remote sensing techniques. 28.03%, which is hardly 2.93% of the entire land of the state
This paper brings out the extended pattern of urban sprawl of whereas only 2.43% of geographical area of Malda is belonged to
English Bazar Urban Agglomeration around 2.5 km from its outer that of urban areas. Subsequent urban growth in this region had
limit of official jurisdiction using temporal remote sensing data taken place more likely to say along the both sides of Mahananda
with zonal gradients and spatial metrics. The study area was river as a British commercial centre in the form of two municipal-
divided into eight zones and each zone is further divided into con- ities of the district i.e. English Bazar Municipality (EBM) and Old
centric circles of 0.5 km incrementing radii from the city centre Malda Municipality (OMM). But the current census data of 2011
extended up to 10.5 km to understand the patterns and extent of has brought a sudden twist in the record where Malda has been
urbanization at local level. Its land use analysis has revealed a captured as the fastest growing town of West Bengal among all
decline of water bodies from 23.61% (1987) to 6.69% (2015). During remaining districts in terms of growth rate of urban population
2015 the built up had constituted 32.51%, vegetation comprised of (124.81%). In this perspective, it is noteworthy to mention that
30.43%, whereas vacant land and agricultural land made up about the number of census towns in Malda has grown to 27 according
19.68% and 10.69% respectively. Increased Shannon’s entropy dur- to the last census, in 1991 where there was no census town in
ing 2015 highlights the tendency of sprawl that necessitated policy Malda.
interventions to provide basic amenities. Spatial pattern through Different backgrounds play vital roles in this unprecedented
metrics indicated a compact and simple structured growth at the growth. English Bazar municipality is the district headquarter of
centre of the urban body and a distributed complex shape in the Malda and the only service provider of 3447185 rural population.
buffer region. Further the metrics indicated that the municipalities It has such a significant and strategic position that it has been
are on the verge of becoming a single large urban patch that would emerging one of the major urban centre of North Bengal which acts
Fig. 2. English Bazar Urban Agglomeration in 2001. Fig. 3. English Bazar Urban Agglomeration in 2011.
demographic and economic data etc. was extracted from the cen- fication and computation of landscape metrics (Sudhira et al.,
sus records (2001 and 2011) and vector layers which were added 2004). Along with in order to detect and quantify the urban forms
to the database. The geographical area of villages was obtained such as built-up in terms of spatial phenomenon, the computation
from the census data of 1991. The convergence of GIS, remote sens- of Urbanisation Intensity Index (Suja et al., 2013) and the
ing and database management systems has helped in quantifying, Shannon’s entropy (Yeh and Li, 2004) have been carried out. In
monitoring, modelling and subsequently predicting this phe- order to predict the scenarios of future urban growth Suitability
nomenon. At the landscape level, GIS has aided in spatial analysis analysis method has been followed. These techniques are
to calculate the dispersion, compactness, built-up density, new discussed below.
development, directionality of open space encroachment.
Fig. 5. Comparison of Statutory Towns and Peripheral villages for Selected Service.
Table 1
Demographic and occupational profile of two ULBs of Malda and their peripheral villages.
Settlements Population Rate of population Population Density (per sq.km) Literacy rate (%) Non-farming workers to
growth in percent (2001–11) total main male workers (%)
EBM 205521 27.29 15511 86.46 98.16
OMM 84012 33.44 23467 75.6 99.53
Peripheral villages 110762 29.49 2049.8 54.47 74.64
Fig. 6. Land cover changes and Sprawl in 1987. Fig. 7. Land cover changes and Sprawl in 2003.
164 R. Shaw, A. Das / The Egyptian Journal of Remote Sensing and Space Sciences 21 (2018) 159–172
needs and welfare (Rawat and Kumar, 2015; Butt et al., 2015). Fol- are the areas of urban land use for years t + n and t respectively and
lowed by the intensive investigation of ground truths Land use n denotes the number of years. TLAi is the total area of the spatial
analysis has been carried out using the process of multistaged unit i.
unsupervised classification with 30 times iterations to detect the
pattern of sprawl and changes in the vast expanse of land cover 4.4. Zonal analysis
of the respective area.
The city boundary along with the buffer region was divided into
4.2. Built-up density eight zones at 45 degree interval considering the central pixels
(CBD) to detect the spatial directional variation of intensity of
The density of built-up cover per unit area of the village is nor- urbanization during 1987–2015. These directional zones in clock-
mally used in ecological investigations and now it is being wise order are WNW (West-North West), NNW (North-North
extended to enhance the understanding of the urban forms. Map West), NNE (North-North East), ENE (East-North East), ESE
density values are computed by dividing area of built up cover to (East-South East), SSE (South-South East), SSW (South-South
the total geographical area of a village. The computation of built- West), WSW (West-South West).
up density gives the distribution of built-up clusters in the study
area in varying degrees. High density of built-up refers to clustered
or more compact nature of the built-up and vice versa. 4.5. Division of zones in concentric circles
Fig. 8. Land cover changes and Sprawl in 2015. Fig. 9. Pattern of Sprawl in 1987–2015.
R. Shaw, A. Das / The Egyptian Journal of Remote Sensing and Space Sciences 21 (2018) 159–172 165
X
n
On the basis of the integration of the above said six indicators three
Hn ¼ Pi loge ðP i Þ zones of varying Urbanisation prospect have been identified. The
1
weightage scores have been provided based on the role of those
where Pi is the proportion of the variable in the ith zone determinants in the urban growth (see Table 7). In order to calcu-
and n is the total number of zones. Entropy value ranges from late the suitability score for each peripheral village these weightage
0 to Log (n) Where Log (n) = Maximum limit of Entropy. Here pattern of six parameters has to be compiled in raster calculator.
if the value closer to zero indicates very compact distribution
whereas the value closer to Log (n) indicates dispersed
5. Result & discussion
distribution. As the dispersed settlement indicates sprawl, larger
value of entropy reveals the occurrence of urban sprawl (Sudhira
In most of the developing world and industrializing countries,
et al., 2004).
the peri urban is often a zone of chaotic urbanization leading to
sprawl (Ravetz et al., 2013). Therefore, the peri-urban growth
4.7. Site suitability analysis around English Bazar Agglomeration can be discussed in the light
of three spatial measures mainly, Growth and density of the
This study has taken an attempt to uphold some growth centres built-up area, Urbanisation Intensity Index and nature of sprawls.
having high degree of urbanization through suitability analysis
method using GIS tools. In this perspective along with three major 5.1. Detection of sprawl & pattern of land cover/uses change
Urban criteria prescribed by Census of India, another three indica-
tors have been considered i.e. rising growth rate of population to Perhaps the most important definition is that of ‘urban sprawl’
indicate, high household density and maximum share of built-up – generally seen as a land use pattern with lower density, ineffi-
cover to total area. cient or wasted land-use, car dependency, and so on (Ravetz
166 R. Shaw, A. Das / The Egyptian Journal of Remote Sensing and Space Sciences 21 (2018) 159–172
Fig. 10. Areas under different land uses and land covers.
et al., 2013). The basic indicator to quantify urban footprints is the captured in 2003 with 15.34% increase among other study periods
proportion of built-up area and the reduction of other land use (see Figs. 9–11).
types (Suja et al., 2013). The results of the classification process The pattern of changes in land use and land cover is heteroge-
are shown in Figs. 6–8. neous in this study as the land cover categories other than the res-
The classified maps are displaying the incremental growth of idential land use have experienced asymmetric fluctuation in
built-up area (Figs. 6–8).Within the time span of 28 years (1987– growth. On the other hand, only the growth of vegetation and res-
2015) the build up area has increased 29.64% point i.e. total 6.64 idential land use is consistent and the most effected land use cat-
sq.km geographical land area. The share of residential land cover egory is water body that has constantly reduced with the rate of
to total geographical land under the study is 32.51% in the last 8.24 sq.km within interval of 16 years and 6.87 sq.km within
year. The maximum growth rate of build-up cover has been 12 years. In 2003 after the totalitarian flood (1998) the growth of
Fig. 11. Newly emerged residential areas in 2015. Fig. 12. Built-up Density in the study area.
R. Shaw, A. Das / The Egyptian Journal of Remote Sensing and Space Sciences 21 (2018) 159–172 167
Fig. 13. Distribution of built-up cover in different directions across buffers in 1987. Fig. 14. Distribution of built-up cover in different directions across buffers in 2015.
Table 3 Table 7
Urbanisation intensity index in each direction. Weightage Pattern of the selected parameters as per their role in making an urban
area.
Direction Built-up area (sq. Built-up area (sq. Urbanisation
km) in 2015 km) in 1987 Intensity Index (%) Parameters to be Integrated Integrating
Percentage
NNW 9.84 6.73 0.124
NNE 4.45 2.1 0.094 Rate of Population growth 25
ENE 1.46 0.92 0.021 Population Density 15
ESE 1.04 0.64 0.016 Population 25
SSE 2.14 2.19 0.002 Proportion of main male Non-farming workers to total 15
SSW 3.29 2.94 0.014 Proportion of built-up area to total 10
WSW 2.69 2.86 0.007 Household Density 10
WNW 4.11 3.98 0.005 Total 100
Source: Computed by the author by geo-spatial analysis of Landsat images. Source: Consideration of author.
Table 4
Conversions of 1987 land use types to residential cover in 2003.
Land use types Water Body Vegetation Residential Area Vacant Land Agricultural Land Total Residential
area in 2003
Area under Land uses in 1987 21.09 16.89 22.4 16.75 12.21
Proportion of land use types of 1987 transformed 4.06 (19.25%) 5.18 (30.67%) 10.1 (45.09%) 5.48 (32.72%) 1.06 (8.62%) 25.88
into residential cover in 2003
Area in sq.km., Figures in parentheses are percentage of land to total area under the concerned land use types in 1987.
Source: Computed by the author through geo-spatial analysis of Landsat images.
Table 5
Conversions of 2003 land use types to residential cover in 2015.
Land use types Water Body Vegetation Residential Area Vacant Land Agricultural Land Total Residential
area in 2015
Area under Land uses in 2003 12.85 20.5 25.88 16.49 13.62
Proportion of land use types of 2003 transformed 2.95 (22.96%) 4.82 (23.51%) 13.24 (51.16%) 5.93 (35.96%) 2.1 (15.42%) 29.04
as residential cover in 2015
Area in sq.km., Figures in parentheses are percentage of land to total area under the concerned landuse types in 2003.
Source: Computed by the author by geo-spatial analysis of Landsat images.
Table 6
Conversion of 1987 land use types to residential cover in 2015.
Land use types Water Body Vegetation Residential Area Vacant Land Agricultural Land Total Residential
area in 2015
Area under Land uses in 1987 21.09 16.89 22.4 16.75 12.21
Proportion of land use types of 1987 transformed 4.46 (21.15%) 5.75 (34.04%) 11.26 (50.27%) 5.97 (35.64%) 1.6 (13.10%) 29.04
as residential cover in 2015
Area in sq.km. Figures in parentheses are percentage of land to total area under the concerned land use types in 1987.
Source: Computed by the author by geo-spatial analysis of Landsat images.
R. Shaw, A. Das / The Egyptian Journal of Remote Sensing and Space Sciences 21 (2018) 159–172 169
The Entropy value for the all zones together is 1.85 in 1987 and
2.75 in 2015 while the log (n) for the whole study area is 3.04
(Fig. 15). Thus the results show that the distribution of built-up
in the region in 2015 is more dispersed than that of in 1987 tend-
ing to develop the sprawl-like features over that landscape and the
nature of sprawl is spatially sporadic. A centrifugal force contribut-
ing dispersed settlement was active significantly beyond 6.5 km
from the centre of the city in 2015.
Fig. 18. Urban transformation in the identified Peri urban areas. The invasion of natural land cover for living purpose promi-
nently indicates the ever-increasing urban development as well
as the resultant land loss. The land transformation contributes
greatly to the modification of the ecosystem also. Vacant land, for-
est or agricultural lands are reduced in expense of the expansion of
human habitation resulting into a dynamic influence of spatial pat-
tern of the locality.
The emergence of new settlements fundamentally belongs to
water body and agricultural land feeding pattern. From the geo
spatial analysis of the images of two periods it reveals that within
the time interval of 1987–2003 among other land cover categories
the conversion of the water bodies and vegetation to residential
land use was maximum. Table 4 clearly portrays that 32.72% of
vacant land existed in 1987 has been transformed to residential
cover in 2003. On the other side, like the previous case the vacant
land captured in 2003has retained the maximum share in consti-
tuting the residential area in 2015 through the process of transfor-
mation. Among the five land use – land cover categories vegetation
(captured in 2003) has largely modified by the expansion of built-
up cover in 2015. Thus the lateral expansion of urban area both
within and outside the administrative boundary is mainly respon-
sible for tremendous transformation of vacant land since 1987.
Plate 1. Godowns of oil & cement are commonly found in outskirts near
Like this, there is only 51.16% of pre-existing residential land cover
Narayanpur.
of 2003 remained same in 2015 (see Tables 5–7).
Hence the growth in built-up area is accounted for the reduc-
tion in agricultural land and water bodies. It can also be observed
5.3. Nature of sprawl that along the roads and waterways the growth of built up cover is
high especially in Old Malda municipality (OMM). This fact proves
Shannon’s Entropy was calculated based on the area under that it is not necessary that the reduction in any land use category
built-up impervious for all 21 buffer zones around city centre. It is meant for the increase in built up cover as the arrangement pat-
highlights that the land use is fragmented in all directions due to tern of the different land use categories is dynamic and mutually
emergence of new urbanpockes with time. transformed in different periods. However, there is a striking
R. Shaw, A. Das / The Egyptian Journal of Remote Sensing and Space Sciences 21 (2018) 159–172 171
Plate 3. Land is acquired for future extension of schools in vacant land of Jhangra.
feature of this study area that the old residential area has been
replaced by other land use categories while new places have been opment of an urban cluster (see Figs. 17 and 18). In this zone only
encroached for impervious built-up in the recent time. two mouzas i.e. Dilalpur and Arazi Dilalpur cannot fulfill the ideal
Land transformation map in Fig. 16 helps to understand the nat- condition to be treated as high potential urban units as Arazi Dilal-
ure and amount of encroachment of built up areas occurring in pur is an entirely uninhabited area and Dilalpur possesses the SSS
2015 through occupying the other land uses existing in 1987. score of 4.7 which is not far from becoming a high prospective area
Table 6 also depicts that out of 35.64% residential land cover found of that urban cluster. Among those villages Bagbari has already
in 2015 is developed by encroachment of vacant land existing in gained the Census given threefold criterion to be called as Census
1987. Besides, the wholesome figure of land use pattern change town in 2011.
within the time span of 28 years (1987–2015) indicates that the However, it is not necessary that all the contiguous villages of
major cause for the shrinkage in the volume of water bodies in this the municipalities have to satisfy the desired characters to be
area is the unprecedented urban spread. named as urban. For example, Itakhola in English Bazar and Maul-
pur in Old Malda experience negative rate of population growth as
per the latest census (16.35% and 10.82% respectively).
5.5. Prediction of future urban centres Jagadishpur (English Bazar) and Kaludewan (Old Malda) can also
be grouped under same category as there is comparatively low
The Site Suitability Analysis to recognise the prospective rural density of population recorded in these mouzas. On the other hand,
areas as future urban growth points, indicates the occurrence of most of the villages adjacent to the Old Malda municipality have
the maximum urban spread along the western and south- moderate prospect for urban growth in the future though some
western sides of English Bazar Municipality. The western part of peripheral villages are already recognised as census towns i.e.
English Bazar municipality bears the high possibility for the devel- Chhatianmor and Sahapur (see Plates 1–5).
172 R. Shaw, A. Das / The Egyptian Journal of Remote Sensing and Space Sciences 21 (2018) 159–172