Spatiotemporal analysis and prediction of urban evolution

patterns using Artificial Neural Network tool

Manuscript Creation- 1st August 2022 Number of Words- 4627 No. Figures- 7
No. of tables -4

Abstract – The precise quantification of Land Use Land Cover (LULC) plays a vital role in
preserving sustainability, which is being affected by growing urbanization. The study proposes
the comprehensive Geographical Information System (GIS) approach in integration with
Artificial Neural Network (ANN) to analyse the past development patterns of the city for
predicting future land transformations. In present study, land transformations from the year
1990 to 2020 have been analysed by the classified maps for Jaipur city, India, as a case study,
which reveals that the built-up land was increased by 48%. Subsequently, the simulated land
transformation for the year 2030 using the Multi-Layer Perceptron (MLP) and Cellular
Automata (CA) anticipates that the built-up land would be increased by 12.7% by cutting down
the barren land and vegetation by 9.4% and 3.3%, respectively. The simulation offers strong
evidence that most of the medium-built-up land density municipality wards transform into
high-density built-up land density wards during the next decade, which is visualized through
the exclusively developed ward-by-ward built-up land density maps. The utilization of the
simulated map in a proposed way helps to prepare the comprehensive micro-level urban
development plan by incorporating natural resource conservation and land use planning.

Keywords- Urbanization, Spatiotemporal analysis, Land Use Land Cover, Artificial Neural
Network.

1. Introduction –

Urbanization is nothing but the process of transforming the land from one class to another to
meet human requirements (Aspinall and Staiano, 2017). According to the United Nations
report on world urbanization, 50% of India's population currently lives in cities or urban areas
(Aboelnour et al., 2018). From the year 1991 to 2017, urbanization increased by 23% in 47
major cities in India due to the migration of people from rural areas (Dinda et al., 2019; Mishra
and Rai, 2016). Rapid urbanization causes decreases in sustainability due to significant land
transformation and deforestation over time (Busho, Wendimagegn, and Muleta 2021; Dutta
and Das, 2019). So to quantify unrestrained development and ecological characteristics of the
city, spatiotemporal analysis plays a vital role (Kafy et al., 2021; Gao et al. 2020; Osumanu
and Akomgbangre 2020). Nowadays, Geographical Information systems (GIS) and Remote
Sensing (RS) technologies have proven an effective technology for conducting various
spatiotemporal and temporal analyses of land and other natural resources (Aboelnour et al.,
2018; Wang and Maduako, 2018). In RS, satellite image datasets such as Landsat, Indian
Remote Sensing (IRS), and IKONOS are consumed to provide important data that can be
further utilized to perform various spatial analyses (Fu et al., 2019; Majumdar and Chatterjee,
2021). Implementing RS techniques makes it possible to quantify the LULC by utilizing multi-
temporal and multispectral satellite imageries over larger areas with precise and consistent data
sets (Chang et al., 2019; Fu et al., 2019; Pandey and Joshi, 2015). As a result of the
spatiotemporal analysis, the classified Land Use Land Cover (LULC) helps to build sustainable
land use control policies by considering past development patterns and current circumstances
(Al Faisal, Al Kafy and Roy, 2018).

In LULC maps, land use refers to the land that humans utilize, whereas land cover refers
to the available natural terrestrial surfaces (Lei, Flacke and Schwarz, 2021; Nong et al., 2018).
In urban areas, LULC changes are affected by various dynamic natural, geological,
topographic, and socio-economic factors that must be analysed to achieve sustainable
development goals (Sahana, Ahmed and Sajjad, no date). In past urban studies, artificial neural
networks (ANN) have been utilized to perform various spatial simulations and LULC analyses
(Dadashpour Moghaddam, Ahmadzadeh and Valizadeh, 2022; Das and Angadi, 2021). Along
with ANN, various tools such as logistic regression, decision support, and cellular automata
model were implemented through various software packages such as DINAMICA, Modules
for Land Use Change Evaluation (MOLUSCE), by incorporating multiple social and economic
parameters (Gao et al. 2020; Xia et al. 2019; Falah, Karimi and Harandi, 2020). The developed
LULC change model helps to detect the changes that occurred or will have the potential to
occur along with the time (Fu et al. 2019; Ghosh, 2019).

To perform large-scale spatial simulation and projection with high precision, the
integration of multiple tools and methodologies is required (Ramachandra et al. 2018; A Faisal
and Khan 2017). So, integrating the multi-layer perceptron (MLP) and Markov chain analysis
enriches the LULC modelling and restores the dimensional LULC changes in GIS (Aithal,
Vinay and Ramachandra, 2018). The created models allow the development of possible future
scenarios by incorporating factors such as population, topology, development, and distance
from the road, which are responsible for the land transformation (Al Faisal, Al Kafy and Roy,
2018). Most of the researchers have attempted to simply forecast future LULC patterns based
on past growth patterns, which limits the applicability of the predicted map. Also, existing
models focused on utilizing high-resolution data, which entailed more financial resources and
time to acquire the raw data. To facilitate the affluent LULC change analysis, the proposed
methodology utilizes the freely available Landsat satellite image data, Digital Elevation Model
(DEM), and Open Street Map (OSM) data acquired from the United States Geological Survey
(USGS) earth explorer. The proposed research practices the Arc GIS (10.7) software to create
LULC maps by unsupervised classification and Q GIS (2.18.16) software to simulate and
forecast future LULC maps, which are further utilized for creating ward-wise built-up land area
density maps.

2. Study Area

Jaipur is in the State of Rajasthan and was founded in the year 1727, which is currently the
capital of Rajasthan state and is the focus of the economic and demographic life of the State,
recognized as the pink city and a famous tourist destination. The Nahargarh hills, part of the
Aravalli hills ranges, surround the city to the north and Jhalana to the east. Jaipur (India) is
located at latitude 26° 55' north and 75° 49' east longitude. Its municipal boundaries run from
26° 46' north latitude to 27° 01' north latitude and from 75° 37' east longitude to 76° 57' east
longitude, as shown in Figure 1. Jaipur city is governed by the JMC and covers an area of 467
km2.

Jaipur is located in India's semi-arid zone and is distinguished by high temperatures,

low rainfall, and mild winters. Jaipur has experienced rapid physical and demographic growth,
with an expected population of 64.95 lakhs by 2025, by 5.3 % of annual growth rate (Dadhich
and Hanaoka, 2011). The city is conveniently located, with the capital of India, Delhi, at 258
kilometres, and Agra at 232 kilometres, forming part of the well-known Golden Triangle.
Currently, the city experiences multiple problems occurring due to urbanization. The most
significant problems among them are the inadequate quantity and quality of water, rapidly
growing demand, and the development of private peri-urban areas, including insufficient
governance structures. JMC was recently divided into two Municipal Corporations, “Greater
Jaipur Municipal Corporation” and “Jaipur Heritage Municipal Corporation,” with 150 and 100
wards. The field visit confirmed that maximum new urban settlements were taking place on
the outskirts of the Jaipur Municipal corporation (JMC) boundary area, within a 5 KM radius,
which extends the study area. The developed built-up land density maps for the study area
could assist the local government authority in prioritizing city areas for future development
projects and for carrying out land transformations.

Figure 1. Study Area Map

3. Materials and Data Methodology

The methodology used in this study deals with the transformation of LULC, along with area
change statistics utilized to forecast future probable land changes for Jaipur city over an
extended study area (5Km buffer). The study's framework is divided into five stages, beginning
with data acquisition, and ending with validation and analysis of the simulated map, as
represented in Figure 2.

Figure 2. Work Methodology

3.1 Data Acquisition

The current study entails the processing of Landsat satellite images for the extended study area.
The satellite images have been acquired from the USGS earth explorer over the last three
decades to perform the LULC analysis (i.e., from 1990 to 2020). For conducting change
analysis, Landsat 5 (Path/Row 147/41) for the date 18 March 1990, Landsat 7 (Path/Row
147/41) for the years 2000 and 2010, and Landsat 8 (Path/Row 147/41) for the year 2015 and
2020 were acquired and analysed. The detailed properties of the satellite images are shown in
Table 1.

Table 1. Satellite Image Properties

Satellite/Sensor Spatial Spectral Resolution

Path/Row Resolution(m)
Date Acquired
30 0.45-0.52(Visible Blue)
30 0.52-0.60(Visible green)
Landsat 5 (TM) 30 0.63-0.69(Visible Red)
147/41 30 0.76-0.90(NIR)
05 March 1990 30 1.65-1.75 (SWIR)
120 10.40-12.50 (Thermal)
30 2.08-2.35 (SWIR-2)
30 0.45-0.52(Visible Blue)
30 0.52-0.60(Visible green)
Landsat 7 (ETM+) 30 0.63-0.69(Visible Red)
147/41 30 0.77-0.90 (NIR)
12 February (2000,2010,2015) 30 1.55-1.75(SWIR)
60 10.40-12.50 (Thermal)
30 2.08-2.35(Mid Infrared)
30 0.45–0.51 (blue)
30 0.53–0.59 (green)
Landsat 8 30 0.64–0.67 (red)
147/41 30 0.85–0.88 (NIR)
30 December 2020 30 1.57–1.65 (SWIR-1)
30 2.11–2.29 (SWIR-2)
100 10.60–11.19 (TIRS-1)

Along with the satellite data, the ancillary data includes the Road shapefile and Digital
Elevation Model (DEM) data. The DEM data was obtained from the USGS earth explorer at
30m grid spacing from the Shuttle Radar Topographic Mission (SRTM), which includes a
terrain roughness index (TRI), topographic wetness index (YWI), slope length, and ground
steepness. The required road shapefile was downloaded from Open Street Maps (OSM) for
calculating the distance from road maps using the Euclidian distance method in Arc GIS
software.

3.2 Preparation for Classification

The study uses the Landsat series satellite image data, i.e., digital sensed data, to create LULC
maps for the entire city of Jaipur. Landsat data is widely available, and it is the most commonly
used satellite for RS data (Zheng et al., 2015). The various sensors from the Landsat satellite
reflect and record the electromagnetic spectrum, including the gamma and x-rays, and all
digitally recorded ground data is transmitted to ground stations (Hua et al., 2014). For
conducting unsupervised classification, all images were geographically referenced in the
Universal Transverse Mercator (UTM) projection and the World Geodetic System (WGS)
1984. The band composite (data management) tool was initially used to develop false-colour
composite (FCC) images for all the images, and each processed composite image was clipped
with a confined study area. Furthermore, a convolution function is used to smoothen raster
images by smoothing 5x5 pixels, reducing local variation and noise. Figure 3 provides a clear
visualization of the process required for classification (Pre-Processing of Image).

Figure 3. Preparation for Classification (Pre-Processing)

3.3 Unsupervised Classification

In unsupervised classification, the computer determines the pixel category based on the pixel
characteristics of the class. As shown in figure 4., the arc GIS 10.6 software was used to
perform unsupervised classification on the processed images to identify various features of
LULC. The unsupervised classification works effectively with coarse-resolution raster data
like Landsat because satellite image clusters consist of spectrally similar cells grouped into
classes with similar spectral properties (Pal and Ziaul, 2017). In unsupervised classification,
the image was initially classified into 15 different classes based on each pixel value, including
Cropland, Built-up Land, Mixed Forest, Shrubland, Barren Land, Fallow Land, Wasteland,
Water Bodies, Plantations, Aquaculture, Mangrove Forest, Salt Pan, Grassland, and Evergreen
Broadleaf Forest. The classified classes were reclassified to recognize the Jaipur region's
growth pattern, and all pixels were divided into three new categories: built-up land, barren land,
and vegetation. The reclassification images present the entire area's built-up land area,
vegetation, and arid land.

Figure 4. Unsupervised Classification

The classified maps show that the population of the JMC region has expanded
dramatically in the last year, whereas the population of the walled city remains constant as it is
an already densely populated area. The urban growth pattern for the extended study area of
Jaipur for the years 1990, 2000, 2010, and 2020 is analysed and presented in figure 5.
 Figure 5. Classified Image for Year 1990,2000,2010,2020

3.4 LULC Change Analysis

The LULC maps aid in comprehending a large amount of data that can be used to quantify,
analyse, and monitor the land changes that occurred over time (Sheikhi, Kanniah and Ho,
2015). As a result, updated LULC information was used to understand past-present trends and
to simulate future land transformations. In this study, the previous two decades (from 1990 to
2015) were analysed to perform the change analysis, allowing a better understanding of Jaipur
City's past growth pattern. The LULC analysis was carried out using the MOLUSCE plugin
from Q GIS 2.18.16. The MOLUSCE plugin contains a set of algorithms, including ANN,
Learning Rate (LR), Weight of Evidence (WoE), and Minimum Classification Error (MCE),
that were used to analyse and simulate Land utilization transformations, as well as the
validation of Kappa Statistics.

3.5 Area Change Map

The concept of area change is dependent on changes in areas that occurred between 1990 and
2015. The area change statistics for each predefined class were calculated (Built-up land,
vegetation, and Barren land) by multiplying the pixel count by its image resolution. The
obtained area change transition matrix was used to quantify changes that occurred in LULC.
The statistics of area change are represented in table 2.

Table 2. Area Change Statistics

LULC 1990 2000 2010 2020 Overall

Features Percentage

Land 483.5 372.69 388.24 311.47 -18.51

Built Up 85.32 218.27 260.87 532.97 48.23

Vegetation 359.72 337.59 279.04 83.68 -29.77

In addition to the area change statistics, a transition map and transition matrix were
created in the MOLUSCE plugin to identify class transitions from one class to another. The
quantitative transformation of the areas for the period was determined using this analysis. The
transition coefficients for the class are shown in Table 3.

Table 3. Transition Matrix

Class 1 2 3

1 0.543503 0.375035 0.081462

2 0.117784 0.824075 0.058142

3 0.416704 0.417093 0.166204

3.6 Transition potential Modelling

Initially, a transition model was built, followed by transition maps required to create the
area change map and used to examine the explanatory power of variables. The created maps
were also used to train the model through an ANN equipped with an MLP, with a learning rate
of 0.1 and 10 hidden layer numbers. MLP is the most practical method for classifying and
forecasting LULC changes (Islam and Ahmed, 2011). MLP is a feed-forward artificial neural
network model that converts input data sets into a set of suitable outputs. The MLP algorithm
comprises multiple layers of nodes that are fully connected with the successor nodes. Because
of appropriateness and minimum parameter requirements, MLP was used to analyse the LULC
changes from the map between 1990 and 2015. Along with MLP, a CA machine learning
algorithm is utilized to simulate the future LULC maps by incorporating transition potential
maps.

3.7 Validation

Validation of the simulated map is an important stage in all geospatial modelling as it ensures
the reliability and acceptance of the simulated map. In this study, simulated and satellite-driven
maps for the year 2020 were compared using kappa coefficients (overall kappa and percent of
correctness). The kappa values range from 0 to 1, representing the accuracy of the simulated
map in comparison to the reference map. The simulated model is strongly validated by the
kappa values obtained from the Q GIS software, with percent correctness of 85. The obtained
kappa values indicated a high validation agreement for the simulated map.

3.8 Prediction

The predicted future map aids in the long-term planning of new developments. The MLP and
CA models were used to forecast the future map for the year 2030. Figure 6. illustrates the
projected map for the year 2030. According to the change analysis conducted between 2020
and 2030, the net increase in a built-up land area is 12.68%, decreasing barren land and
vegetation by 9.4% and 3.2%, respectively. The predicted analysis shows remarkable growth
in a built-up land area and a significant declination in vegetation and barren land on the
outskirts of the main city area.
 Figure 6. Simulated Map for the Year 2030

4. Results

From the simulated map, most of the low-density areas on the west side of Jaipur city have
been transformed into high-density areas by converting barren land and forest land. The
municipality ward by-ward built-up land density maps were prepared for the years 2020 and
2030 to prioritize the area within the city region. Figure 7 shows the transformation and
modification of land areas in different wards over time in the Jaipur Heritage and Jaipur Greater
administrative region. As the Jaipur Heritage region is already densely populated, no drastic
increase in the built-up land area will be observed. Still, many wards in the Jaipur Greater
region have seen a valid increase in built-up land area in recent years.
 (a) (b)

Figure 7. (a) Built-Up land Area Density Map for the year 2020 and (b) Built-Up land Area Density Map for the
year 2030

In urban areas, built-up land areas are proportional to the urban compactness ratio of
the area. A built-up land area per meter square for an urban area is known as the urban
compactness ratio (UCR). The UCR is directly affected by sustainable development through
the energy consumption and utilization of resources which can be calculated using equation 1.

𝐵𝐷 = 𝐵𝐴/𝑈𝐴 equation (1)

Where,

BD = Built-up land area density

UA = Urban Area

BA = Built-up land area extracted from satellite

The transition study, conducted from 2020 to 2030, aids in analysing urban expansion and
density patterns. The built-up land density in 1990 was 9.1%, and by 2020 it had risen to 57.42
%. The analysis also indicates that the compactness ratio of Jaipur city would be 70% in the
year 2030. The performed analysis also helps to understand and control the unexpected
developments in the city area. To overcome the problems caused by rapid urbanization, it is
necessary to understand the city's growth pattern and stimulate future growth, which will be
useful for planning various situations.

5. Discussion

The analysis shows that the pattern of LULC change for Jaipur city is developing on the western
side by transforming mostly the Vegetation and barren land. The predicted map for the year
2030 is utilized to conduct the municipality ward-wise study for Jaipur city, India, where the
built-up land area was analysed to estimate the urban compacted ratio for the entire study
period. The calculated urban compactness ratio (UCR) depicts the distribution of ward-wise
built-up land density throughout the city, strongly associated with the population. Table. 4
shows the prospective built-up land density transformation of the municipality wards in Jaipur
Heritage and Jaipur Greater Corporation region. The study also validates a combination of RS
and Artificial Neural Network (ANN) technology applied to perform the LULC analysis and
to simulate future scenarios. In the current scenario, various factors such as industrialization,
available facilities, and IT hubs are responsible for the rapid urbanization, which indirectly
reduces the sustainability of the society by ignoring proper management and efficient
utilization. Considering the multiple factors needs the additional ancillary data as ground
survey data and aerial surveys, which will be further incorporated in a future study.

6. Conclusion

To facilitate the LULC analysis, the proposed research implemented Remote Sensing (RS)
technique to extract valuable spatial information from satellite images. In the study, the area
change analysis conducted from the year 1990 to 2020 shows that the urban expansion
increased in the western and southern direction of Jaipur city, India, by 48%, while available
barren land and vegetation decreased by 18% and 30%, respectively. Furthermore, the Multi-
Layer Perceptron (MLP) and Cellular Automata (CA) models were applied to predict the future
expansion and create the map for the year 2030. The created map reveals that the barren and
vegetation land has been transformed into an urban area with approximately 651 km2 out of
928 km2. Furthermore, a ward-by-ward study is conducted for the Jaipur Heritage and Jaipur
Greater Corporation to execute the predicted map, showing that most medium-density wards
will be converted to high-density wards by 2030, with an urban compactness ratio of 70%. The
presented land-use transition illustrates that the urban (built-up) land areas are rapidly
expanding by transforming the surrounding agricultural and barren land. The proposed
methodology and novel built-up land density maps for the city area may play a significant role

Year 2020
Density
Heritage (Ward No)
1,96,8,2,3,95,99,11,7,4,76,
Low Density 13,10,97,38,12,83,22,6,9,1
8,39,28.
47,35,78,17,100,48,41,40,2
Medium Density
649,55
50,86,14,34,88,53,36,
52,94,90,27,98,54,31,92,
37,77,46,82,20,32,93,42,
57,89,24,19,67,51,15,62,
High Density 43,29,91,65,44,23,33,16,
21,25,30,45,56,58,59,60,
61,63,64,66,68,69.70,71,
72,73,74,75,79,80,81,84,
85,87.
Table 4. Ward
development.
Transformation for
the year 2020-2030
Year 2030
Greater (Ward No) Heritage (Ward No) Greater (Ward No)
107,111,138,1,137
96,1,2,8,95,11,7,76,99,
56,112,131,35,97,120, 107,111,138,1.
3,97,4,13,10,3,83.
150,115,2,22,116.
125,53,114,96,109,118,119,1
36,90,95,63,104,99,87,91,98, 137,56,2,131,150,22,
18,6,5,12,22,9,28,78.
121,123,5,103,124,25,41,117 35,112,120.
,113,51,102,122,142,4,84,43.
125,97,53,115,116,136,109,5,95,9
86,76,94,82,106,71,66,44,9,8 6,114,104,87,90,103,4,123,102,99,
5,8,148,101,133,23,67,3,88,3 106,118,63,91,98,119,8,142,121,4
48,100,47,17,39,41,26,50,4,
1,29,108,65,48,128,47,147,1 3,86,3,41,76,25,117,84,133,113,44
35,34,40,14,53,88,87,32,36,
05,93,12,100,130,20,83,110, ,31,85,122,108,105,71,129,29,66,4
46,54,55,52,94,98,31,90,37,
80,75,59,21,49,74,6,37,89,64 8,51,128,94,124,82,148,101,69,67,
92,82,93,29,27,42,62,24,20,
,127,24,62,54,10,30,11,26,13 80,88,75,23,65,6,93,147,20,30,59,
19,89,77,15,67,51,64,33,16,
,46,68,144,78,3,7,27,149,126 26,21,72,83,110,60,127,62,47,24,4
44,43,91,21,23,25,30,45,56,
,55,60,70,61,17,45,72,50,52, 9,50,61,100,7,27,130,54,89,126,13
57,58,59,60,61,63,65,66,68,
38,40,32,58,143,77,39,81,36, ,37,10,68,70,7,38,46,144,73,55,78,
69,70,71,72,73,74,75,79,80,
140,139,92,42,28,12,134,16, 32,17,149,134,52,81,77,45,11,143,
81,84,85,86.
14,9,15,18,19,33,34,57,79,13 36,139,58,40,92,9,39,12,15,42,14,
2,135,141,145,146. 16,18,19,28,33,34,57,64,79,132,13
5,140,141,145,146.
in carrying out area planning and infrastructure development plans for sustainable
Declaration of Competing Interest

On behalf of all authors, the corresponding author states that there is no conflict of interest.

Funding

The authors are grateful to the Department of Science & Technology, New Delhi, for providing
financial assistance under the project vide grant no. DST/TMD/EWO/WTI/2K19/UWS-04(C1)
titled Structured Dialogues for Sustainable Urban Water Management (SDSUWM).
Data availability

Some or all data used are proprietary or is confidential in nature and thus may only be provided
with restrictions, stated clearly in the article.

Reference

Aboelnour, M. et al. (2018) ‘Application of Remote Sensing Techniques and Geographic

Information Systems to Analyze Land Surface Temperature in Response to Land Use/Land
Cover Change in Greater Cairo Region, Egypt’, Journal of Geographic Information System,
10(1), pp. 57–88. doi: 10.4236/JGIS.2018.101003.

Aithal, B. H., Vinay, S. and Ramachandra, T. V. (2018) ‘Simulating urban growth by two state
modelling and connected network’, Modeling Earth Systems and Environment, 4(4), pp. 1297–
1308. doi: 10.1007/S40808-018-0506-1/FIGURES/7.

Aspinall, R. and Staiano, M. (2017) ‘A Conceptual Model for Land System Dynamics as a
Coupled Human–Environment System’, Land 2017, Vol. 6, Page 81, 6(4), p. 81. doi:
10.3390/LAND6040081.

Busho, S. W., Wendimagegn, G. T. and Muleta, A. T. (2021) ‘Quantifying spatial patterns of

urbanization: growth types, rates, and changes in Addis Ababa City from 1990 to 2020’, Spatial
Information Research 2021 29:5, 29(5), pp. 699–713. doi: 10.1007/S41324-021-00388-4.

Chang, H. K. et al. (2019) ‘The Use of a Multiple Risk Level Model to Tackle the Duration of
Risk for Construction Activity’, KSCE Journal of Civil Engineering 2019 23:6, 23(6), pp.
2397–2408. doi: 10.1007/S12205-019-1757-8.

Dadashpour Moghaddam, M., Ahmadzadeh, H. and Valizadeh, R. (2022) ‘A GIS-Based

Assessment of Urban Tourism Potential with a Branding Approach Utilizing Hybrid
Modeling’, Spatial Information Research, 30(3), pp. 399–416. doi: 10.1007/S41324-022-
00439-4/TABLES/4.
Dadhich, P. N. and Hanaoka, S. (2011) ‘Spatio-temporal Urban Growth Modeling of Jaipur,
India’, http://dx.doi.org/10.1080/10630732.2011.615567, 18(3), pp. 45–65. doi:
10.1080/10630732.2011.615567.

Das, S. and Angadi, D. P. (2021) ‘Assessment of urban sprawl using landscape metrics and
Shannon’s entropy model approach in town level of Barrackpore sub-divisional region, India’,
Modeling Earth Systems and Environment, 7(2), pp. 1071–1095. doi: 10.1007/S40808-020-
00990-9/FIGURES/5.

Dinda, S. et al. (2019) ‘Integration of GIS and statistical approach in mapping of urban sprawl
and predicting future growth in Midnapore town, India’, Modeling Earth Systems and
Environment, 5(1), pp. 331–352. doi: 10.1007/S40808-018-0536-8/FIGURES/12.

Dutta, I. and Das, A. (2019) ‘Modeling dynamics of peri-urban interface based on principal
component analysis (PCA) and cluster analysis (CA): a study of English Bazar Urban
Agglomeration, West Bengal’, Modeling Earth Systems and Environment, 5(2), pp. 613–626.
doi: 10.1007/S40808-018-0554-6/FIGURES/9.

Al Faisal, A., Al Kafy, A. and Roy, S. (2018) ‘Integration of Remote Sensing and GIS
Techniques for Flood Monitoring and Damage Assessment: A Case Study of Naogaon District,
Bangladesh’, Journal of Remote Sensing & GIS, 07(02). doi: 10.4172/2469-4134.1000236.

Faisal, A. and Khan, A. H. (2017) ‘Application of GIS and Remote Sensing in Disaster
Management: A Critical Review of Flood Management’, Proceedings, International
Conference on Disaster Risk Mitigation, (January), pp. 2–5.

Falah, N., Karimi, A. and Harandi, A. T. (2020) ‘Urban growth modeling using cellular
automata model and AHP (case study: Qazvin city)’, Modeling Earth Systems and
Environment, 6(1), pp. 235–248. doi: 10.1007/S40808-019-00674-Z/FIGURES/8.

Fu, Y. et al. (2019) ‘Characterizing the spatial pattern of annual urban growth by using time
series Landsat imagery’, Science of The Total Environment, 666, pp. 274–284. doi:
10.1016/J.SCITOTENV.2019.02.178.

Gao, C. et al. (2020) ‘Modeling urban growth using spatially heterogeneous cellular automata
models: Comparison of spatial lag, spatial error and GWR’, Computers, Environment and
Urban Systems, 81, p. 101459. doi: 10.1016/J.COMPENVURBSYS.2020.101459.

Ghosh, S. (2019) ‘A city growth and land-use/land-cover change: a case study of Bhopal,
India’, Modeling Earth Systems and Environment, 5(4), pp. 1569–1578. doi: 10.1007/S40808-
019-00605-Y/TABLES/6.

Hua, L. et al. (2014) ‘Simulating Urban Growth Using the SLEUTH Model in a Coastal Peri-
Urban District in China’, Sustainability 2014, Vol. 6, Pages 3899-3914, 6(6), pp. 3899–3914.
doi: 10.3390/SU6063899.

Islam, M. S. and Ahmed, R. (2011) ‘Land Use Change Prediction In Dhaka City Using Gis
Aided Markov Chain Modeling’, Journal of Life and Earth Science, 6, pp. 81–89. doi:
10.3329/JLES.V6I0.9726.

Kafy, A. A. et al. (2021) ‘Cellular Automata approach in dynamic modelling of land cover
changes using RapidEye images in Dhaka, Bangladesh’, Environmental Challenges, 4, p.
100084. doi: 10.1016/J.ENVC.2021.100084.

Lei, Y., Flacke, J. and Schwarz, N. (2021) ‘Does Urban planning affect urban growth pattern?
A case study of Shenzhen, China’, Land Use Policy, 101(November 2019), p. 105100. doi:
10.1016/j.landusepol.2020.105100.

Majumdar, S. and Chatterjee, U. (2021) ‘Modelling urban growth using Urban growth
deterministic model in Kolkata Metropolitan Area: a geo-statistical approach’, Modeling Earth
Systems and Environment, 7(4), pp. 2241–2249. doi: 10.1007/S40808-020-00985-
6/FIGURES/5.

Mishra, V. N. and Rai, P. K. (2016) ‘A remote sensing aided multi-layer perceptron-Markov

chain analysis for land use and land cover change prediction in Patna district (Bihar), India’,
Arabian Journal of Geosciences, 9(4). doi: 10.1007/s12517-015-2138-3.

Nong, D. H. et al. (2018) ‘Quantifying urban growth patterns in Hanoi using landscape
expansion modes and time series spatial metrics’, PLOS ONE, 13(5), p. e0196940. doi:
10.1371/JOURNAL.PONE.0196940.

Osumanu, I. K. and Akomgbangre, J. N. (2020) ‘A growing city: patterns and ramifications of

urban change in Wa, Ghana’, Spatial Information Research 2020 28:5, 28(5), pp. 523–536.
doi: 10.1007/S41324-020-00313-1.

Pal, S. and Ziaul, S. (2017) ‘Detection of land use and land cover change and land surface
temperature in English Bazar urban centre’, The Egyptian Journal of Remote Sensing and
Space Science, 20(1), pp. 125–145. doi: 10.1016/J.EJRS.2016.11.003.

Pandey, B. and Joshi, P. K. (2015) ‘Numerical modelling spatial patterns of urban growth in
Chandigarh and surrounding region (India) using multi-agent systems’, Modeling Earth
Systems and Environment, 1(3), pp. 1–14. doi: 10.1007/S40808-015-0005-6/FIGURES/10.

Ramachandra, T. V. et al. (2018) ‘Modeling urban dynamics along two major industrial
corridors in India’, Spatial Information Research 2018 27:1, 27(1), pp. 37–48. doi:
10.1007/S41324-018-0217-8.

Sahana, M., Ahmed, R. and Sajjad, H. (no date) ‘Analyzing land surface temperature
distribution in response to land use/land cover change using split window algorithm and
spectral radiance model in Sundarban Biosphere Reserve, India’. doi: 10.1007/s40808-016-
0135-5.

Sheikhi, A., Kanniah, K. D. and Ho, C. H. (2015) ‘Effect of land cover and green space on land
surface temperature of a fast growing economic region in Malaysia’,
https://doi.org/10.1117/12.2194796, 9644, pp. 215–222. doi: 10.1117/12.2194796.

Wang, J. and Maduako, I. N. (2018) ‘Spatio-temporal urban growth dynamics of Lagos

Metropolitan Region of Nigeria based on Hybrid methods for LULC modeling and prediction’,
https://doi.org/10.1080/22797254.2017.1419831, 51(1), pp. 251–265. doi:
10.1080/22797254.2017.1419831.

Xia, C. et al. (2019) ‘Bidirectional urban flows in rapidly urbanizing metropolitan areas and
their macro and micro impacts on urban growth: A case study of the Yangtze River middle
reaches megalopolis, China’, Land Use Policy, 82, pp. 158–168. doi:
10.1016/J.LANDUSEPOL.2018.12.007.

Zheng, H. W. et al. (2015) ‘Simulating land use change in urban renewal areas: A case study
in Hong Kong’, Habitat International, 46, pp. 23–34. doi:
10.1016/J.HABITATINT.2014.10.008.