Evaluating the effectiveness of the city master plan in regulating future urban spatial growth of Varanasi city, India

This paper evaluates the effectiveness of the Varanasi City Master Plan 2031 in regulating urban growth by analyzing Land Use and Land Cover (LULC) changes. By comparing the model's predictions for 2031 with the Varanasi Development Authority's Master Plan, the study identifies discrepancies in the direction and extent of urban expansion. Rapid urbanization, driven by industrialization, migration, and infrastructural development, has dramatically reshaped Varanasi's spatial patterns. Utilizing remote sensing data from Landsat images (1990, 2000, 2010, and 2021) and integrating machine learning techniques, including the Multi-layer Perceptron and Markov Chain Analysis (MLP-MCA), this study simulates and predicts future urban expansion. The model's predictions, with an accuracy above 80%, offer critical insights for policymakers to revisit urban planning strategies. The built-up area has grown from 45.10 km² in 1990 to a projected 262.05 km² by 2031, representing a 480.95% increase over four decades. Simultaneously, agricultural acreage has declined from 908.23 km² to 656 km², a reduction of 252.23 km², or 27.77%, highlighting the shift from rural to urban land use. Notably, in the southwest, the Masterplan consistently exceeds predicted built-up areas across most zones, except in Zone 4 (9–12 km), with over-allocations around the Mughalsarai area. Furthermore, Sectors A, B, C, and D anticipate higher built-up areas, particularly in zones 6–9 km and 9–12 km. This study underscores the need for sustainable development planning to mitigate the negative impacts of rapid urbanization, such as loss of green spaces, environmental degradation, and urban heat island effects. The combined approach of remote sensing and machine learning provides a robust and replicable methodology for other rapidly urbanizing cities, ensuring future expansion aligns with sustainable development goals.

1 Introduction

Across the globe, urbanization is increasing rapidly. In 2010, urban areas made up more than 0.5% of the Earth's surface (Angel et al., 2011). As per a report of the World Economic Forum, 56.2 per cent of the global population has been residing in urban settings since 2010. Asian countries have also experienced almost three-fold growth in urban population since 1950. The United Nations (2017) estimates that by 2050, approximately 68% of the world's population will inhabit urban areas. India shares 36 percent of the urban population in its demographic settings 2022 (https://data.worldbank.org/). With an annual urbanization rate of 1.1%—higher than the global average of 0.9%—India is projected to have nearly half of its population residing in urban areas by 2050 8 (UN DESA, 2014).

The phenomena of urban sprawl and alterations in land use and land cover (LULC) are significant global concerns (Prestele et al., 2016; Long et al., 2007; Srivastava et al., 2012; Nahid et al., 2025). Anthropogenic activities such as rapid industrialization, rural-urban migration, infrastructural development in and around urban agglomeration, and resettlement are robust drivers of changes in Land Use Land Cover (LULC) and reshaping landscapes. Strong centripetal forces in the urbanization process increase the risk of ecosystem degradation resulting from LULC changes, concretization, loss of green space, depletion of the water table, environmental pollution, increase in area and intensity of heat island, urban and peri urban waterlogging, flooding, and biodiversity loss (Zhao et al., 2006; Grimm et al., 2008; Seto et al., 2012; Kumar et al., 2015; Bahrawi et al., 2020; Nahid et al., 2023; Kumar et al., 2024) are potential threat to environment in several ways. Most developing countries are seeing agricultural land converted into modern housing developments, which is a major driver of urbanization (Khalifa, 2015; Phuc et al., 2014; Zanganeh Shahraki et al., 2011) and Indian cities are no exception. These are the factors that affect urban and peri urban land governance and planning. The rapid transformation of land use and land cover (LULC) creates significant challenges for policymakers and governors in new urban areas, particularly during economic shifts. (You, 2017; Mohan et al., 2011; Zhang et al., 2011; Husain et al., 2023; Singh et al., 2024).

Urban systems, since they consist of many different interconnected components, are among the most complex ecosystems. The pace and process of urbanization is often closely tied to advancements in technology, economic growth, and cultural interactions and shifts in national or regional planning goals, as well as changes in land use rules, can sometimes lead to noticeable transformations in local neighborhoods and landscapes (Li et al., 2016; Medeiros and van der Zwet, 2020; Ul Din and Mak, 2021). Remote sensing, particularly through the utilization of high spatial and temporal resolution data, serves as a crucial instrument for acquiring land cover information in rapidly urbanizing areas (Soni et al., 2022; Salem et al., 2020). The classification of land use and land cover (LULC) from satellite imagery using efficient 2 methodologies—characterized by speed and accuracy—represents a significant area of research within the domain of satellite remote sensing (Hu and Nacun, 2018; Zhao et al., 2024). Recent literature indicates a growing acceptance and interest in emerging machine learning techniques 11 within this field (Wang et al., 2022; Adugna et al., 2022; Nahid et al., 2025).

The traditional supervised classification methods necessitate the extraction of a substantial number of features, a process that is both time-intensive and requires considerable expertise (Chollet, 2017; Chen et al., 2018). The accuracy and reliability of classification outcomes are often compromised due to the limited availability of training samples. Consequently, to obtain sufficient and efficient samples in a timely manner during supervised classification, considerable emphasis must be placed on the selection and extraction of training samples. The parameter classifier must effectively capture all pertinent spectral heterogeneity and adhere to the standards of probability sampling design within and between LULC categories, particularly concerning the constraints imposed by training sample limitations (Boser et al., 1992; Mountrakis et al., 2011). In contrast, multiscale segmentation object-oriented classifiers, such as non-parametric supervised classifiers—including Maximum Likelihood Classification (MLC), Classification and Regression Trees (CART), Support Vector Machine (SVM), Back Propagation (BP), Artificial Neural Network (ANN), and Random Forest (RF)—have gained widespread application in LULC classification due to their reduced limitations and superior classification accuracy compared to parametric classifiers (Kotsiantis, 2007; Yang et al., 2012; Attarchi and Gloaguen, 2014; Gómez et al., 2016; Hu and Nacun, 2018). Nevertheless, the requirement for a substantial volume of training data in non-parametric supervised classification methods presents a significant challenge that remains unresolved (Rodriguez-Galiano et al., 2012).

Researchers have conducted numerous studies to project urban expansion scenarios and examine their implications in order to determine right directions of future growth (Te Linde et al., 2011; Shi et al., 2017; Song et al., 2017; Lu et al., 2019; Nath et al., 2020). Spatial models and advanced Geographic Information System (GIS) methodologies have been extensively utilized regarding spatial algorithms to explain land use and land cover patterns, distribution, and directions as well as to forecast the amount of LULC dynamics. For instance, the cellular landscape with agent-based simulation is combined with the multisystem agent model to optimize the decision-making process of LULC changes (Schreinemachers and Berger, 2006; Le et al., 2008; Ralha et al., 2013; Salem et al., 2020). Likewise, for the modeling of land use changes, generalized linear modeling was applied within the empirical GIS environment (Aspinall, 2004; Rutherford et al., 2008). Zang and Huang (2006) used another type which is aggregated multivariate regression to simulate urban dynamics, particularly to assess dominant drivers and causal forces. The Markov model has emerged as a prominent tool for simulating future LULC transitions (Pahlavani et al., 2017; Sangermano et al., 2012) effectively analyzing spatiotemporal land use changes through the application of Markov chain methodologies (Rimal et al., 2018; Al-sharif and Pradhan, 2014; Yang et al., 2012; Arsanjani et al., 2011; Guan et al., 2011; Tang et al., 2007; Myint and Wang, 2006). The Artificial Neural Network (ANN) Multi-layer Perceptron (MLP) algorithm is 10 among the most widely utilized ANN applications (Sangermano et al., 2012) and the MLP-based ANN-Markov hybrid model has demonstrated significant efficiency in accurately predicting LULC changes (Mishra and Rai, 2016; Ozturk, 2015) because this model is particularly capable at managing ‘missing cases' in historical LULC training data and accommodating extensive training datasets (Iizuka et al., 2017; Riccioli et al., 2016). Furthermore, in contrast to other spatial models such as Slope, Land use, Exclusion, Urban extent, Transportation, and Hillshade (SLEUTH), which necessitate coefficient values for modeling LULC transitions (Chaudhuri and Clarke, 2013) or Cellular Automata (CA), which require potential transition maps as basic prior knowledge (Pontius et al., 2008), the MLP-based ANN-Markov model offers a distinct advantage in scenarios where prior knowledge is lacking. Additionally, this model has been successfully employed in various global 12 contexts to simulate complex LULC changes, producing impressive results with high simulation accuracy (Simwanda et al., 2021; Nurwanda and Honjo, 2020; Rimal et al., 2020; Shade and Kremer, 2019; Islam et al., 2018; Iizuka et al., 2017). R Studio and QGIS have been used to classify LULC changes while the MLP-MCA hybrid approach has been used to simulate future scenarios of LULC changes of Varanasi city to efficiently minimize errors in image classification and pattern simulation.

The objective of the present study is to map and monitor LULC of the Varanasi city of Uttar Pradesh, India to quantify spatial pattern and sprawl of urban area over the last decades. Based on previous years' LULC results, the study further simulates future spatial patterns. A comparison between the simulated urban sprawl map and the projected urban expansion map of Master Plan 2031 of the Municipal Corporation of Varanasi City was conducted. This comparative analysis of the model simulated and Master Plan projected LULC will help city planners and policymakers of Varanasi city administration to better understand and design future urban LULC plans to accommodate sustainable future urban growth. Findings definitely will help in revisiting the understanding of spatial patterns and the direction of land rent policy from the city core to the periphery.

2 Methods

2.1 Study area

Varanasi, one of the world's oldest continuously inhabited cities, is located in the middle Ganges plain in North India. It lies in the eastern part of Uttar Pradesh along the Ganga River at 11 coordinates 25°15′55.27″ N latitude and 82°58′56.92″ E longitude (Figure 1). The city has a long history of habitation dating back to around 1000 BCE, though its modern urban development took shape mainly in the early 18th century (Singh, 2015). Varanasi is positioned at an elevation of 80.71 meters (264.8 ft) and experiences a humid subtropical climate (Köppen climate classification Cwa), characterized by significant diurnal temperature variations. The average annual precipitation in the city is recorded at 1,110 mm (44 in). According to provisional data from the 2011 census (Census of India, 2011), the urban agglomeration of Varanasi had a population of 1,435,113, while the combined population of the Varanasi municipal corporation and the Cantonment Board was 1,212,610 in the same year. Varanasi attained the designation of a ‘million-plus city' (as an Urban Agglomeration, VUA) in 1991. In the Master Plan 2011, approximately 0.95% (82.50 ha) of the total area of 8,645 ha was designated for historical, heritage, and archaeological purposes; this allocation has since decreased to 0.38% (92.40 ha) in the proposed total area of 24,646 ha in the Master Plan 2031. The Urban Agglomeration (UA) spans an area of 119.52 sq. km, which previously consisted of a single core (Central Business District, i.e., Chauk). The city has since expanded significantly and is now referred to as the Varanasi Development Region, encompassing 477.34 sq. km. Given the ongoing conversion of the Grand Trunk (GT) road (National Highway 2) into a superhighway, along with its established status as a 21 prominent religious, cultural, and spiritual tourism destination, Varanasi is poised for substantial spatial growth in the future.

Figure 1

Figure 1. Location of study area.

2.2 Data

The study utilizes open-source ortho-rectified remote sensing satellite images of Landsat-5 Thematic Mapper (Landsat TM), Landsat Enhanced Thematic Mapper Plus (Landsat ETM+), and Landsat Operational Land Imager (Landsat OLI) collected from the official website of US Geological Survey (USGS) (http://glovis.usgs.gov). Landsat images for the years 1990, 2000, 2010, and 2021 have been collected to monitor LULC change, and urban expansion of the city. The ASTER DEM of 30-m spatial resolution has been used to create slope and aspect maps as variables. Digital Elevation Model (DEM) data was downloaded from the USGS website and the shape files of the roads were downloaded from the Open Street Map (OSM) (https://www.openstreetmap.org/). Shapefiles of road networks and administrative boundaries are the other data sets used in the study. The details of satellite data used in the study are given in Table 1.

Table 1

Table 1. Details of datasets used in this study.

2.3 Methodology

The study mostly used quantitative tools and techniques to qualify objectives. This research focuses on assessing and evaluating LULC change analysis, LULC change potential modeling, LULC change prediction, as well as comparative analysis of simulated urban growth with the proposed master plan 2031 for Varanasi city of Uttar Pradesh, India has been conducted. Before analyzing LULC changes, Landsat images underwent pre-processing steps such as cloud and shadow detection, atmospheric correction, and the application of composite/fusion/metrics techniques. The analysis of remote sensing data and the implementation of Multi-Layer Perceptron-Multi-Criteria Analysis (MLP-MCA) prediction tools were carried out within a Geographic Information System (GIS) (Figure 2).

Figure 2

Figure 2. Methodological framework.

2.3.1 DESU and LULC

2.3.1.1 Classification of images for DESU maps using Random Forest

Dynamics in Earth Surface Utilization (DESU) is a process of quantifying changes and identifying the location of changes from one category to another accurately on the Earth's surface in a single map between two points of time. A supervised classification was used over multispectral bands of TM/ETM +/OLI images for change detection and classification of land use and land cover changes during the study periods (1990, 2000, 2010, and 2021). The classification of Landsat imagery was carried out with R-Studio software using R-script based on a Random Forest machine learning algorithm where each image was classified into nine classes (Figure 3). False Color Composite (FCC) with different band combinations was used to classify different LULC category. Although it is theoretically possible to generate a square number of subclasses derived from main classes (Settlement, Vegetation, Acreages and Waterbody), the present study has selectively established classes pertinent to the study area. For instance, the establishment of subclasses such as “water body to settlement” has been deemed infeasible under any circumstances. Similarly, the subclass “vegetation to settlement” is deemed inapplicable within the study area due to the gradual nature of the transition from vegetation to settlement, a process unlikely to occur within the temporal span of 10 years. Consequently, classes deemed inapplicable remain vacant within the matrix in Figure 3.

Figure 3

Figure 3. Legend and matrix for DESU.

2.3.1.2 Reclassification of DESU maps to LULC

As the DESU map constitutes nine classes, it was essential to the reclassification of maps to reduce from nine sub-classes to four main classes namely Built up areas (including residential, commercial, and industrial buildings, as well as roads, transportation, mixed urban, rural settlements, and bare soil), Vegetation (encompassing deciduous forests, mixed forest lands, scrub, and other types), Acreage (comprising cultivated lands, floodplains, low-lying areas, marshy land, rills and gullies, swamps), and Waterbodies (including rivers, permanent open water, lagoons, lakes, ponds, and reservoirs). This reclassification was essential for the accurate recording and prediction of LULC changes and urban sprawl (Table 2).

Table 2

Table 2. Reclassification classes of LULC from DESU.

2.3.2 Prediction

2.3.2.1 LULC change analysis

LULC maps for 1990, 2000, 2010, and 2021 were analyzed using IDRISI's Land Change Modeler (LCM) for change analysis, simulation, and future predictions. It quantitatively assesses land use changes through classified classes such as net changes, swaps, gains, losses, and total changes (Eastman, 2015), utilizing consecutive date images. IDRISI's LCM, incorporates user-specified drivers, maps changes, identifies transitions, and models, and predicts future scenarios (Eastman, 2012) using Markov chain matrices for change potential assessment. LULC change analysis is conducted between time 1 and time 2, identifying transitions between classes and quantified through cross-tabulation analysis for 2000–2010 (period 1), 2010–2021 (period 2), and 2000–2021 (period 3). This method quantifies spatial and quantitative changes temporally between LULC classes (Mozumder and Tripathi 2014), examining gains, losses, and contributions to net change for the built-up area as well as analysis of the spatial trend of change, presented graphically for each period.

2.3.2.2 Selection of LULC transitions

This study focused exclusively on major transitions among LULC classes, excluding minor ones, as they significantly influence the area's dynamics. To optimize MLP neural network performance, two key transitions are integrated into the transition sub-model (Eastman, 2006), i.e., vegetation to built-up and acreages to built-up.

2.3.2.3 Selection of variables and model development

Land use change drivers vary by study area, requiring individual standardization. Physical constraints like existing built-up areas, streams, and roads limit expansion. Boolean constraint maps, assigning 1 for suitable and 0 for unsuitable areas (Eastman, 2012), were created, considering built-up areas and major roads as constraints. Unlike constraints as hard rules, factors here influence the suitability degree of a particular alternative for concern activity (Eastman, 2006) and provide a degree of suitability for an area to change on a distance basis.

Urban growth is shaped by key factors: Physical elements like slope and elevation (Braimoh and Onishi, 2007; Reilly et al., 2009; Ye et al., 2013), socioeconomic factors including proximity to roads (Luo and Wei, 2009), socioeconomic centers, water (Cheng and Masser, 2003; Luo and Wei, 2009), and overall determinants like GDP and population. Neighborhood effects consider that Non-built-up cells surrounded by built-up land are more likely to convert to built-up. Topographic and proximity factors, grouped into sub-models, were assessed for explanatory power using Cramer's V analysis. Seven variables, including distance to major roads and built-up areas as dynamic variables changing over time and elevation, slope, and aspect as static variables remaining constant, were considered as factors in the study.

Given that development typically aligns with roads, “distance from major roads” was chosen as a key factor. LULC transitions were grouped into sub-models with their primary driving factors, explaining specific transitions that occurred in a desired period with variables unique to each sub-model. Derived information was instrumental in generating transition potential maps to visualize future LULC suitability. The variables used in this study are shown in Figure 4.

Figure 4

Figure 4. Variables used in the study: (A) Distance from major roads, (B) aspect, (C) digital elevation model, (D) slope, and (E) urban distance map.

2.3.2.4 Transition potential modeling

To accomplish the modeling process, accurate transition potential maps were developed, identifying significant LULC transitions. MLPNN integrated into LCM, a powerful algorithm for LULC prediction, solves non-linear separable problems (Ahmed and Ahmed, 2012). The MLP is a feed-forward artificial neural network model based on the 9 supervised Back Propagation (BP) algorithm, which maps a set of input data to a corresponding set of outputs. This model is structured as a directed graph consisting of multiple layers of interconnected nodes. The training sites provide information for calculation. Being automatic mode and requiring minimal parameters, it surpasses other algorithms. The study employed MLP to analyze LULC changes between 2000, 2010, and 2021 maps, focusing on major LULC transitions. The neural network created a random sample of cells, following selected transitions in modeling. To adjust weights, compute training errors, and enhance accuracy, a weighted neuron network was employed. Samples from 2000 to 2010 LULC maps totaled 7,076 cells for MLPNN with 10,000 iterations. The sample was divided evenly-−50% for training and 50% for validation—allowing the MLPNN model to achieve a 71.46% accuracy rate, which measures how well the model was calibrated. The generated transition potential maps were then used to forecast future LULC changes.

2.3.2.5 Prediction and validation of LULC change

A hybrid model combining MLP and MCA predicts LULC changes for a specified future date, allocating changes through the LCM module (Eastman, 2006). MCA estimates change amounts using 2000 and 2010 LULC maps, determining required transition weights for Markov chain probability matrices with the help of MLPNN. The prediction for 2021 is calculated using the probability function including sub-model transitions and Markov chain analysis, and takes 2010 as the reference year. The output from the Markov chain, which consists of a matrix detailing the expected change quantities for each transition at projected dates, informs the prediction of the 2021 LULC map through the transition probability matrix established from 2000 to 2010. Two distinct outcomes were produced for 2021: a hard prediction (scenario) and a soft prediction (vulnerability). Model validation was conducted by comparing the predicted 2021 LULC map with the actual observed map, employing kappa index statistics for accuracy assessment. The components of the kappa index are kappa for no information (K_no), kappa for grid cell level location (K_location), and kappa for stratum-level location (K_{locationStrata}) in addition to kappa standard (K_standard) which is the same as kappa (Pontius, 2000; Geri et al., 2011; Cohen, 1960). Following this, the MLP-MC model was utilized to predict LULC changes for the year 2031.

2.3.3 . Comparison with masterplan 2031

For this purpose, the master plan of Varanasi city was procured from the Varanasi Development Authority (VDA), and the scanned image of the master plan has been georeferenced by taking coordinate values from the satellite images. After the georeferencing, the city master plan was digitized considering Vishwanath Temple as the center of the city. Six concentric zones of three-kilometer circles radiating outward were demarcated. Comparison between the simulated and predicted urban sprawl map and the proposed VDA master plan of 2031 was accomplished based on each concentric circle.

3 Results

The ancient city of Varanasi, revered for its cultural and spiritual significance, faces complex challenges due to rapid urbanization and climate change. As the urban population increase exponentially, there is an urgent need to reevaluate current urban planning strategies, particularly in historic cities like Varanasi where traditional infrastructures are strained by modern demands. Sustainable urban planning can mitigate adverse effects, such as the urban heat island phenomenon and escalating air pollution, by integrating green spaces and enhancing public transportation systems. Moreover, adaptive planning approaches that incorporate local knowledge and participatory governance are vital in fostering resilient urban ecosystems.The results of the present study have been presented in four distinct sections: population growth and density analysis; decadal LULC change analysis; prediction and validation of urban spatial growth and comparison of master plan 2031 of Varanasi Development Authority (VDA) and predicted urban spatial growth map of 2031 to maintain flow in sectional analysis as findings are not mutually exclusive.

3.1 Population growth and density in the study area

Varanasi city harbored a modest population of 226,105 persons in 1901 but experienced a dramatic shift in the city's demographic landscape during the mid-20th century. Figure 5 depicts that the population curve started showing a significant upward trend after 1941. District Census Handbook data reveals consistent population growth throughout the latter half of the 20th century. The city recorded a total urban population of 134,995 persons in 1971, 175,422 persons in 1981, and 1,057,972 persons in 1991 making it a million-plus city. The 2001 census documented a growth rate of 19.2%, followed by a comparatively sluggish growth rate of 12.94% in 2011. Age-old religious, cultural, and heritage tourism, and urban infrastructure development amidst economic liberalization act as pull factors for rural-to-urban migration and contribute significantly to absolute population growth. Master Plan 2031 further projected a sharp increase in urban population from 2011 to 2031. The projected urban population is 1,731,292 in 2021 and 2,671,994 in 2031 indicating an 87.67 per cent increase from 2011. As per Slum Free City Plan of Action (SFCPoA)-, the city's population density surged significantly, reaching 14,675 individuals per square kilometer in 2001 to a high of 15,062 in 2011. The growth and density of the population are tightly associated with the spatial expansion of the city in the rural-urban fringe area. If not planned appropriately, this results in the haphazard development of the city landscape.

Figure 5

Figure 5. DESU classification of Varanasi city.

3.2 Decadal Dynamics in Earth Surface Utilization (DESU)

The DESU analysis presents a summary of the change in the spatial pattern of LULC during the period between two consecutive years between base years and target years for each decade (Figure 5) The dynamic change is obvious and prominent in the study area due to the anthropogenic activities, especially in the cases of Built up-Built up, Acreages-Built up, Vegetation-Vegetation and Acreages-Acreages.

Varanasi city's LULC scenario is very complex with the dominant category being mixed land use. Built up to Built up change is showing no change in its area during 1990–2000 and 2000–2010 while it registered a modest increase from 44.51 sq km to 56.90 sq. km in the decade of 2010–2021 which can be seen primarily due to push and pull factor-induced rural-urban migration, infrastructural development (like road and rail network connectivity, upscaling religious and heritage tourism etc.) and post-liberalization market expansion in the city. Analysis of the table depicts the persistent increase in built-up due to the conversion of land from acreage to built-up category (Table 3). A substantial conversion from vegetation to Acreages (7.69 sq km to 53.52 sq km) has been registered between the decades 2000–2010 to 2010–2021 showing an increased magnitude of demand-driven conversion. Gradually over three decades around 62 sq km of land under Acreages has been converted into Built-up. This shows how the city landscape has been changing and urbanization is increasing at the expense of land under Acreages during the period. The Sharpe increase in the urban population in Varanasi city after 1990 (Figure 6) demanded more land for built-up category to accommodate them. The city experienced a more than 3-fold decrease in the conversion of land from Acreages to Vegetation in the three decades whereas a maximum decrease was noticed during the 2000–2010 to 2010–2021 period. Vegetation cover indicative of green spaces has registered an increase in area during three decades.

Table 3

Table 3. Decadal dynamics of Earth Surface utilization from one category to another.

Figure 6

Figure 6. Decadal growth of urban population of Varanasi city. Source: compiled from Census 2011 and VDA Master Plan 2031.

3.3 Land Use Land Cover (LULC) change

A LULC map of Varanasi city for the years 1990, 2000, 2010, and 2021 has been prepared to evaluate changes in each category. A future scenario of LULC of the city for the year 2031 was also predicted using the MLP-MCA tool. A substantial change in built-up and acreages has been registered in the city (Figure 7).

Figure 7

Figure 7. LULC classification of Varanasi city.

Spatial or vertical growth in the Built-up area is an important parameter of urban growth and has shown a significant increase in the area during the last 30 years. In the year 1990, it was 45.10 sq km and increased to 157.49 sq km in 2,000 covering 15.47 per cent of the total area. It covered 231.63 sq km in 2021 accounting for 22.71 per cent of the total area of the city. Built-up covered 25.74 per cent of the city space (262.05 sq km) in the predicted LULC map for 2031. More than a five-time increase in built-up from 1990 to 2031 is a clear indication of escalated spatial expansion and increased urban density of Varanasi city in the past, present, and future. The increase in vegetation cover (45.56 sq km in 1990 to 106.02 sq km in 2021) during the period in the city is because of some protected green areas/zones like Banaras Hindu University (BHU), Electric Locomotive Ward (ELW) colony, Sarnath UNESCO heritage site, designated community parks and green spaces, etc. are good for maintaining green spaces in the city. As has been stated earlier, acreages have declined substantially from 908.23 sq km (89.2 percent of the total area) in 1990 to 658.15 sq km (64.65 per cent) in 2021. It further shows a slight decrease of 2.15 sq km area in the predicted images of the year 2031. The pressure of rural-urban migration can be seen in the acreages category in the past, present, and future LULC of Varanasi city. Throughout the entire period of study, waterbody has shown negligible change in its area. A slight decrease has been noticed between the 2021 and predicted 2031 images (Table 4).

Table 4

Table 4. LULC change in Varanasi city from 1990 to 2031.

3.4 Prediction and validation/accuracy assessment

3.4.1 Software based

This study is based on the integration of multi-layer perceptron-Markov chain analysis (MLP-MCA) and has been employed to generate transition potential maps for various transitions with an accuracy of above 80%. Initially, LULC patterns for the year 2021 were forecasted using LULC maps from 2000 and 2010 through the MLP-MCA methodology. Subsequently, the predicted LULC map for 2021 was compared with the actual LULC map of the same year to validate the model's predictive capabilities and to ensure a reliable forecast for future LULC scenarios. For validation, we used the kappa index statistic. This allowed us to compare the predicted 2021 LULC map with the observed LULC map of 2021, and acquired the statistics for quantity and location. The results indicate the K_no value is 0.7680, the K_location value is 0.8532, the K_{locationStrata} value is 0.8532, and the K_standard value is 0.7163 (overall kappa), respectively. It showed that the kappa index values are acceptable, and the performance of the MLP-MCA method to identify grid cell level location of future change is quite satisfactory (here, the K_location value is 0.8352, where the K_location value of 1 is perfect). Following the successful prediction and validation of both hard and soft LULC maps for 2021, future LULC change maps were generated. Transition potential maps and a transition probability matrix were created using the 2021 LULC map as the reference. The predicted LULC change maps for 2031 are illustrated in Figure 7, with corresponding area statistics for various LULC classes presented in Table 4.

3.4.2 Manual validation

Apart from the software's validation of prediction results, there is another way to validate the prediction results produced by the software. In this method, the accuracy of the result has been calculated manually. In this process, we find out the absolute differences of areas between the observed image (latest LULC classified image) and the predicted image of the current situation and then convert them into percentages of differences in reference to the observed image. After calculating the percentage of differences, the under root squire values have been calculated to neutralize the positive and negative sign of values and are subtracted from 100, which produced a percentage of accuracy (Table 5).

$\begin{array}{l}\text{A}=100-\sqrt{{\left[\frac{(\text{I }-\text{ PI})}{\text{OI}}\times 100\right]}^2}\end{array}$

Table 5

Table 5. Difference between observed and predicted image in LULC classification.

Where,

A = Accuracy

OI = Observed Image (latest LULC classified image)

PI = predicted image (predicted image for current situation).

3.5 Comparison between predicted spatial growth of the city and VDA master plans 2031

The spatial growth of the city from the core toward the periphery results due to the conversion of semi urban landscape of the rural-urban fringe into an urban one. Several factors govern the pace and pattern of urban expansion. These factors can be grouped into physical such as rivers, mountains, etc., and anthropogenic such as policy interventions. Haphazard urbanization over time results in inappropriate land use and land cover compromising achieving Sustainable Development Goals (SDGs). Town and Country Planning Department of the state prepares future master plans of big cities to accommodate growing populations sustainably. Persistent upon natural city growth and regulated city growth, it is pertinent to look into a comparative analysis between predicted natural spatial growth and spatial growth of the master plan. It helps in understanding to what extent the master plan of the city complies with or regulates the future spatial growth of the city. Varanasi Development Authority (VDA) master plan 2031 and predicted spatial growth map of the same year are being compared to see the degree and direction of deviation in city growth from the master plan.

The first masterplan for Varanasi city was conceptualized by Varanasi Surveying Branch a sub-division of the Town and County Planning Department of Uttar Pradesh in 1964 under the “U.P. (Regulation of Building Operations) Act, 1958” and was accepted as Masterplan 1991 in 1973 (Master plan 2031). The Varanasi Development Authority (VDA) was established in 1974, and in 1982, the VDA reassessed the previous urban plans. Under VDA's direction, the Town and County Planning Organization (TCPO) formulated a comprehensive Master Plan for Varanasi covering the period from 1991 to 2011, anticipating that the population of the Varanasi Agglomeration would double (cf. Singh, 2009c, p. 327). A notable transformation during the 1991–2011 planning period was the significant expansion of the city's area (+112%).

The first Concentric zone which is the core of the city has a radius of three km from its center, i.e., Vishwanath temple is dominated by a built-up area covering the largest portion of this zone followed by acreages, vegetation, and waterbody. Masterplan 2031 however, produces almost the same pattern except waterbody. The second concentric zone is also dominated by built-up area for LULC 2031 as well as masterplan and follows the same stencil as of first one except for the total area which contributes slightly much area to each class (Figure 8).

Figure 8

Figure 8. Comparison between predicted spatial growth of the city and VDA Master Plans 2031.

Figure 9

Figure 9. Unauthorized unplanned haphazard built-up area development all across Varanasi city; (1) Lamhi-Danganj, (2) Lohta, (3) Kandwa, (4) Ratanpur-Padao, and (5) Dulhipur-Malokhar.

Unlike the previous two Concentric zones which were dominated by built-up areas covering the largest portion in their respective zones, this zone is dominated by acreages occupying the largest portion in this zone. The LULC scheme of Masterplan 2031 is also aligned with the predicted LULC pattern. The fourth concentric zone is unique in terms of having the greatest differences in the area of built-up land between predicted LULC and master plan 2031. it also has a significant difference in acreages class. Analysis of Masterplan 2031 and predicted LULC suggests that each class is marking a substantial difference in areas however, the same patterns are being followed by both Masterplan as well as LULC prediction. As far as the fifth Concentric zone is concerned it is also dominated by acreages for LULC 2031 as well as masterplan and follows the same trends as the fourth concentric zone except for sharp differences of corresponding classes and total area. The uttermost concentric zone is the sixth one which ranges from 15 Km to 18 Km from its center. It lies on the periphery of the study area occupying the least built-up area among all concentric zones in LULC as well as in the master plan. Since it is listed on the periphery apparently, covers the largest area (311.10 Km²) and would be dominated by acreages (Figure 8). It is also like the fourth concentric zone subjected to a sharp difference in terms of area in their concern classes between masterplan and predicted LULC.

4 Discussion

Given the projections that estimate an additional billion city dwellers globally by 2030, Varanasi exemplifies the pressing need for proactive planning. Strategies like upgrading waste management systems, implementing green building standards, and promoting renewable energy usage are essential to curb the environmental impacts of urban growth. Integrating climate-resilient architecture with traditional design principles has proven effective in reducing energy consumption. However, balancing modernization with cultural preservation remains a challenge. Digital tools and Geographic Information Systems (GIS) offer innovative ways to plan and monitor urban growth, enhancing decision-making and resource allocation. Moreover, innovative policies such as mixed land-use zoning and transit-oriented development can enhance livability while minimizing carbon footprints. The integration of smart technologies, like real-time air quality monitoring and data-driven urban management, further supports sustainable urban transitions.

To deal with a range of issues in urban space such as noise pollution, air quality, traffic congestion, urban heat island, urban flood, and summer heat shock an efficient and pragmatic Urban Land use and land cover (LULC) Planning in meso-level cities in India is essential. The primary objectives of this research were to analyze the spatiotemporal characteristics of urban expansion and its prediction to compare the master plan proposed for 2031 in Varanasi city located in the middle Ganges plain of northern India. Landsat images from various years (1990, 2000, 2010, and 2021) were examined to track urban expansion and forecast future land use in this Indian city. A CA-Markov model was applied to simulate potential development patterns until 2031. The model's accuracy was tested against 2021 data, and it demonstrated reliable performance. Taking into account how the spatial patterns of future urban expansion would look like in the upcoming decade, the Markov model successfully predicted the directions and magnitude of urbanization in Varanasi. Moreover, these directions and magnitude were strongly persuaded by driving forces, such as regions from immigrants coming, accessibility that includes vicinity to major roads and urban cores, and topography, including slopes and elevation. The consequences of rapid urbanization presents considerable challenges for policymakers and planners, particularly when it comes to very fertile plain regions, which are densely populated where established infrastructure and basic utilities are always legging behind. The results and modeling process for predicting urban growth in Varanasi hold significant value in multiple ways. Beyond illustrating the spatial and temporal patterns of urban expansion, the model can also simulate interactions between urban development and the surrounding landscape. Moreover, it effectively generates various growth scenarios and assesses their potential impact on urban ecosystems. In recent years, Varanasi has witnessed rapid and substantial urban expansion. Projections for 2031 indicate that this trend is likely to persist, driven by the conversion of open land and vegetation into urban areas. The findings further reinforce that urban expansion will directly replace open land and indirectly affect vegetation.

The alteration of open spaces into built-up areas is largely driven by urban population growth (You, 2017; Mohan et al., 2011; Zhang et al., 2011; Khalifa, 2015; Phuc et al., 2014), particularly due to natural population increases, higher fertility rates, and rural-to-urban migration. The population of Varanasi city was 10,00,747 in 1991, and 11,70,897 in 2001 adding 170150 (17%) persons to that of 1991. Likewise, 3,69,103 (31.52%) people were added in 2011 which reached up to 15,40,000 (census handbook). Master Plan 2031 estimates that the population of Varanasi city will be nearly 18,82,991 in 2021 and 22,74,623 by 2031. If we include the nearby villages, which are expected to happen in VDA this figure goes up to 25,71,256 in 2021 and 31,44,646 by 2031 (Master plan 2031).

Thus, the increased future population is affiliated with the expanding trend of urban agglomerations. Simulation outcomes indicate a significant rise in the built-up and 1 residential areas; for instance, it is projected that by 2031, an additional 30.42 square kilometers will be incorporated into the existing built-up area of Varanasi city. This rapid growth in built-up regions can be attributed not only to the rising urban population but also to factors such as immigration from eastern Uttar Pradesh and western Bihar, particularly driven by spiritual and cultural practices. If we closely observe the graphs from what we come up with the acquisition rate of the built-up area from acreages is continuously increasing in the last three decades which could be a major concern for the sake of reduction of agricultural land since this area lies in a very fertile Gangetic plan. Not only this vegetation is also acquiring its share from acreages though its acquisition rate is getting slower than that of built up. The constant built-up and constant vegetation categories are increasing. If constant built-up and constant vegetation categories are increasing that implies that these are also the acquirer of the open spaces. This fact is depicted by decreasing bars of the graph acreages-acreages in the same figure for the same period where constant Acreages hold descending figures of area (761.56, 703.04, and 601.14) between 1990 to 2000, 2000 to 2010, and 2010 to 2021 respectively.

Indeed, Land use pattern of this pain-located city brings up the significant impact of human activities on ecological, hydrological, and environmental structures. The peripheries of Varanasi are predominantly experiencing the expansion of residential and housing developments, often at the expense of converted acreages and vegetation. Having a strong affiliation with the political corridor since 2014, the infrastructural development of the city has been geared up. Constructing roads like the Babatpur-Varanasi highway has been a major driving force of urban growth and vehicle transportation in the last decade. In essence, procurement of required developmental resources as it is a politically influential constituency and the terrain conditions, which surround Varanasi city are determining factors of the magnitude and direction of future urban growth. Furthermore, since the city is expanding all around, especially the northern periphery of the city is acquiring a substantial pace of urbanization. Consequently, the absence of effective planning policies may result in uncontrolled urban sprawl and destruction of vegetation cover, open spaces, and lowland illegal settlements along either side of the Ganges.

Some studies conducted in non-developed nations (Zanganeh Shahraki et al., 2011; Singh, n.d) have indicated a significant increase in built-up areas in the last two decades and will follow the same trend in upcoming years. The rapid depletion of space and natural resources in this city resulted from ongoing urbanization poses a serious challenge to sustainable development and complicates long-term strategic planning, including the goals set for 2031 and future master plans. This decline in green spaces and natural resources has several adverse effects, including environmental degradation, poor air quality, increased rates of chronic illness, and the intensification of urban heat island effects. Thus, urbanization needs to deal efficiently with sufficient basic infrastructure, and proper monitoring of current happenings mainly transportation and planning with other social phenomenon.

To the best of our knowledge, this study is one of the first to utilize QGIS and RStudio software using machine-learning techniques with Random Forest algorithms to assess the dynamics of earth surface utilization in Varanasi city. This study also substantiates the capability and effectiveness of integrating remote sensing data with GIS for assessing and predicting LULC changes. However, the Landsat images with moderate resolution negatively affect the accuracy of LULC classification and prediction outcomes. On the other hand, some finer resolution datasets and some coding scripts could be developed to find out new prediction methods and achieve better results. In addition, such kind of research has significant potential to contribute in sustainable development initiatives in upcoming urban centers.

5 Conclusion

For Varanasi to thrive amidst these compounded challenges, collaborative efforts among government agencies, local communities, and academic institutions are crucial. Policies need to be inclusive, ensuring that vulnerable populations are safeguarded against climate risks, such as flooding and extreme heat events. By embedding sustainability into the core of urban planning, Varanasi can transform its challenges into opportunities, becoming a model of resilience and innovation in heritage city management. Ultimately, the city's future hinges on balancing preservation with progress, crafting a path that honors its past while embracing sustainable development. Overall, this research on spatial analysis and LULC prediction in Varanasi stands out as one of the most comprehensive studies of its kind, offering new insights into simulating urban LULC dynamics in the Gangetic Plain. Consequently, the findings of this study hold significant implications for urban planning research in other cities within this region. The use of advanced GIS and image processing techniques to identify areas vulnerable to urbanization is particularly valuable for shaping national strategies for future urban development.

Moreover, the simulation results from this study not only help address challenges related to urban sustainability in plain areas but also serve as a spatial reference for monitoring future trends in LULC dynamics. The ability to quantitatively assess land change patterns in this research is made possible through the integration of remote sensing and GIS techniques. However, future studies may enhance these findings by utilizing higher-resolution imagery and incorporating supplementary datasets to better understand the underlying drivers of land changes, ultimately improving knowledge of intra-urban variations in LULC dynamics.

Data availability statement

Publicly available datasets were analyzed in this study. This data can be found here: https://earthexplorer.usgs.gov/.

Author contributions

AR: Writing – review & editing, Writing – original draft, Investigation, Software, Data curation, Conceptualization, Methodology. PK: Formal analysis, Methodology, Validation, Conceptualization, Supervision, Writing – original draft, Writing – review & editing, Investigation, Resources. BD: Formal analysis, Resources, Visualization, Supervision, Writing – review & editing, Conceptualization. BG: Methodology, Supervision, Writing – review & editing, Validation, Resources, Visualization, Conceptualization. SS: Investigation, Software, Resources, Data curation, Writing – review & editing. A: Methodology, Writing – review & editing, Investigation, Data curation, Software, Formal analysis, Resources. AR: Data curation, Methodology, Writing – review & editing, Software, Resources, Formal analysis.

Funding

The author(s) declare that financial support was received for the research and/or publication of this article. Partial APC support was provided by the the Institution of Eminence (IoE), University of Delhi.

Acknowledgments

We acknowledge the Institution of Eminence (IoE), University of Delhi, for logistics, administrative, and partial APC support for Open Access Publication.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declare that no Gen AI was used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

Adugna, T., Xu, W., and Fan, J. (2022). Comparison of random forest and support vector machine classifiers for regional land cover mapping using coarse resolution FY-3C images. Remote Sens. 14:574. doi: 10.3390/rs14030574

Crossref Full Text | Google Scholar

Ahmed, B., and Ahmed, R. (2012). Modeling urban land cover growth dynamics using multi-temporal satellite images: a case study of Dhaka, Bangladesh. ISPRS Int. J. Geo-Inform. 1, 3–31. doi: 10.3390/ijgi1010003

Crossref Full Text | Google Scholar

Al-sharif, A. A. A., and Pradhan, B. (2014). Monitoring and predicting land use change in Tripoli Metropolitan City using an integrated Markov chain and cellular automata models in GIS. Arab. J. Geosci. 7, 4291–4301. doi: 10.1007/s12517-013-1119-7

Crossref Full Text | Google Scholar

Angel, S., Parent, J., Civco, D. L., Blei, A., and Potere, D. (2011). The dimensions of global urban expansion: estimates and projections for all countries, 2000–2050. Prog. Plann. 75, 53–107. doi: 10.1016/j.progress.2011.04.001

Crossref Full Text | Google Scholar

Arsanjani, J. J., Kainz, W., and Mousivand, A. J. (2011). Tracking dynamic land-use change using spatially explicit Markov Chain based on cellular automata: the case of Tehran. Int. J. Image Data Fusion 2, 329–345. doi: 10.1080/19479832.2011.605397

Crossref Full Text | Google Scholar

Aspinall, R. (2004). Modeling land use change with generalized linear models—a multi-model analysis of change between 1860 and 2000 in Gallatin Valley, Montana. J. Environ. Manag. 72, 91–103. doi: 10.1016/j.jenvman.2004.02.009

Crossref Full Text | Google Scholar

Attarchi, S., and Gloaguen, R. (2014). Classifying complex mountainous forests with L-band SAR and landsat data integration: a comparison among different machine learning methods in the Hyrcanian Forest. Remote Sens. 6, 3624–3647. doi: 10.3390/rs6053624

Crossref Full Text | Google Scholar

Bahrawi, J., Ewea, H.·, Kamis, A., and Elhag, M. (2020). Potential flood risk due to urbanization expansion in arid environments, Saudi Arabia. Nat. Haz. 104, 795–809. doi: 10.1007/s11069-020-04190-7

Crossref Full Text | Google Scholar

Boser, B. E., Guyon, I. M., and Vapnik, V. N. (1992). “A training algorithm for optimal margin classifiers,” in Proceedings of the Fifth Annual Workshop on Computational Learning Theory (New York, NY: Association for Computing Machinery (ACM)), 144–152. doi: 10.1145/130385.130401

Crossref Full Text | Google Scholar

Braimoh, A. K., and Onishi, T. (2007). Spatial determinants of urban land use change in Lagos, Nigeria. Land Use Policy 24, 502–515. doi: 10.1016/j.landusepol.2006.09.001

Crossref Full Text | Google Scholar

Census of India (2011). Urban Agglomerations/Cities Having Population 1 Lakh and Above. Provisional Population Totals; Census of India. Available online at: https://web.archive.org/web/20111113152754/http://www.censusindia.gov.in/2011-provresults/paper2/data_files/India2/Table_3_PR_UA_Citiees_1Lakh_and_Above.pdf

Google Scholar

Chaudhuri, G., and Clarke, K. (2013). The SLEUTH land use change model: a review. The Int. J. Environ. Resour. Res. 1, 88–105. doi: 10.22069/ijerr.2013.1688 Available online at: https://ijerr.gau.ac.ir/article_1688_7eecf8813c80d3ca017280879c5a6360.pdf

Crossref Full Text | Google Scholar

Chen, Y., Zhou, Y., Ge, Y., An, R., and Chen, Y. (2018). Enhancing land cover mapping through integration of pixel-based and object-based classifications from remotely sensed imagery. Remote Sens. 10:77. doi: 10.3390/rs10010077

Crossref Full Text | Google Scholar

Cheng, J., and Masser, I. (2003). Urban growth pattern modeling: a case study of Wuhan city, PR China. Landsc. Urban Plan. 62, 199–217. doi: 10.1016/S0169-2046(02)00150-0

Crossref Full Text | Google Scholar

Chollet, F. (2017). Deep Learning with Python. Greenwich, CT: Manning Publications Co.

Google Scholar

Cohen, J. (1960). A coefficient of agreement for nominal scales. Educ. Psychol. Meas. 20, 37–46. doi: 10.1177/001316446002000104

Crossref Full Text | Google Scholar

Eastman, J. R. (2006). IDRISI Andes Tutorial. Worcester, MA: Clark Labs.

Google Scholar

Eastman, J. R. (2012). IDRISI Selva Tutorial. Worcester, MA: Clark University.

Google Scholar

Eastman, J. R. (2015). IDRISI TerrSet Tutorial. Worcester, MA: Clark University.

Google Scholar

Geri, F., Amici, V., and Rocchini, D. (2011). Spatially-based accuracy assessment of forestation prediction in a complex Mediterranean landscape. Appl. Geograp. 31, 881–890. doi: 10.1016/j.apgeog.2011.01.019

Crossref Full Text | Google Scholar

Gómez, C., White, J. C., and Wulder, M. A. (2016). Optical remotely sensed time series data for land cover classification: a review. ISPRS J. Photogrammetry Rem. Sens. 116, 55–72. doi: 10.1016/j.isprsjprs.2016.03.008

Crossref Full Text | Google Scholar

Grimm, N. B., Faeth, S. H., Golubiewski, N. E., Redman, C. L., Wu, J., Bai, X., et al. (2008). Global change and the ecology of cities. Science 319, 756–760. doi: 10.1126/science.1150195

PubMed Abstract | Crossref Full Text | Google Scholar

Guan, D. J., Li, H. F., Inohae, T., Su, W., Nagaie, T., and Hokao, K. (2011). Modeling urban land use change by the integration of cellular automaton and Markov model. Ecol. Modell. 222, 3761–3772. doi: 10.1016/j.ecolmodel.2011.09.009

Crossref Full Text | Google Scholar

Hu, Y., and Nacun, B. (2018). An analysis of land-use change and grassland degradation from a policy perspective in inner Mongolia, China, 1990–2015. Sustainability 10:4048. doi: 10.3390/su10114048

Crossref Full Text | Google Scholar

Husain, M. A., Kumar, P., and Gonencgil, B. (2023). Assessment of spatio-temporal land use/cover change and its effect on land surface temperature in Lahaul and Spiti, India. Land 12:1294. doi: 10.3390/land12071294

Crossref Full Text | Google Scholar

Iizuka, K., Johnson, B. A., Onishi, A., Magcale-Macandog, D. B., Endo, I., and Bragais, M. (2017). Modeling Future urban sprawl and landscape change in the Laguna de Bay Area, Philippines. Land 6:26. doi: 10.3390/land6020026

Crossref Full Text | Google Scholar

Islam, K., Rahman, M. F., and Jashimuddin, M. (2018). Modeling land use change using cellular automata and artificial neural network: the case of Chunati Wildlife Sanctuary, Bangladesh. Ecol. Indic. 88, 439–453. doi: 10.1016/j.ecolind.2018.01.047

Crossref Full Text | Google Scholar

Khalifa, M. A. (2015). Evolution of informal settlements upgrading strategies in Egypt: from negligence to participatory development. Ain Shams Eng. J. 6, 1151–1159. doi: 10.1016/j.asej.2015.04.008

Crossref Full Text | Google Scholar

Kotsiantis, S. (2007). Supervised machine learning: a review of classification techniques. Informatica 31, 249–268. Available online at: https://datajobs.com/data-science-repo/Supervised-Learning-%5BSB-Kotsiantis%5D.pdf

Google Scholar

Kumar, P., Ashwani, D., eka, D., Rai, A., Kumar Rai, V., and Kumar Rai, A. (2024). Impact of climatic factors on air pollution and human health in the Lucknow city of Uttar Pradesh, India. Ecocycles 10, 51–69. doi: 10.19040/ecocycles.v10i2.454

Crossref Full Text | Google Scholar

Kumar, R. P., Ranjan, R. K., Ramanathan, A. L., Singh, S. K., and Srivatsav, P. K. (2015). Geochemical modeling to evaluate the mangrove forest water. Arab. J. Geosci. 8, 4687–4702. doi: 10.1007/s12517-014-1539-z

Crossref Full Text | Google Scholar

Le, Q. B., Park, S. J., Vlek, P. L. G., and Cremers, A. B. (2008). Land-Use Dynamic Simulator (LUDAS): a multi-agent system model for simulating spatio-temporal dynamics of coupled human–landscape system. I. Structure and theoretical specification. Ecol. Inform. 3, 135–153. doi: 10.1016/j.ecoinf.2008.04.003

Crossref Full Text | Google Scholar

Li, B., Chen, C., and Hu, B. (2016). Governing urbanization and the New Urbanization Plan in China. Environ. Urban. 28, 515–534. doi: 10.1177/0956247816647345

Crossref Full Text | Google Scholar

Long, H., Tang, G., Li, X., and Heilig, G. K. (2007). Socio-economic driving forces of land-use change in Kunshan, the Yangtze River Delta economic area of China. J. Environ. Manag. 83, 351–364. doi: 10.1016/j.jenvman.2006.04.003

PubMed Abstract | Crossref Full Text | Google Scholar

Lu, Q., Joyce, J., Imen, S., and Chang, N.-B. (2019). Linking socioeconomic development, sea level rise, and climate change impacts on urban growth in New York City with a fuzzy cellular automata-based Markov chain model. Environ. Plann. B: Urban Anal. City Sci. 46, 551–572. doi: 10.1177/2399808317720797

Crossref Full Text | Google Scholar

Luo, J., and Wei, Y. H. D. (2009). Modeling spatial variations of urban growth patterns in Chinese cities: the case of Nanjing. Landsc. Urban Plan. 91, 51–64. doi: 10.1016/j.landurbplan.2008.11.010

Crossref Full Text | Google Scholar

Medeiros, E., and van der Zwet, A. (2020). Sustainable and integrated urban planning and governance in metropolitan and medium-sized cities. Sustainability 12:5976. doi: 10.3390/su12155976

Crossref Full Text | Google Scholar

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. Arab. J. Geosci. 9:249. doi: 10.1007/s12517-015-2138-3

Crossref Full Text | Google Scholar

Mohan, M., Pathan, S. K., Narendrareddy, K., Kandya, A., and Pandey, S. (2011). Dynamics of urbanization and its impact on land-use/land-cover: a case study of Megacity Delhi. J. Environ. Prot. 2, 1274–1283. doi: 10.4236/jep.2011.29147

Crossref Full Text | Google Scholar

Mountrakis, G., Im, J., and Ogole, C. (2011). Support vector machines in remote sensing: a review. ISPRS J. Photogrammetry Rem. Sens. 66, 247–259. doi: 10.1016/j.isprsjprs.2010.11.001

Crossref Full Text | Google Scholar

Myint, S. W., and Wang, L. (2006). Multicriteria decision approach for land use land cover change using Markov chain analysis and a cellular automata approach. Can. J. Remote Sens. 32, 390–404. doi: 10.5589/m06-032

Crossref Full Text | Google Scholar

Nahid, S., Kumar, R. P., Acharya, P., Kumar, K., and Purohit, S. (2025). Forecasting urban expansion in Delhi-NCR: integrating remote sensing, machine learning, and Markov chain simulation for sustainable urban planning. Geo J. 90:72. doi: 10.1007/s10708-025-11317-5

Crossref Full Text | Google Scholar

Nahid, S., Mahato, S., Anees, M. M., Pandey, D. N., and Joshi, P. K. (2023). “Air pollutants-induced environmental critical zones in capital city of India,” in Advancements in Urban Environmental Studies. GIScience and Geo-environmental Modelling, eds. A. Rahman, S. Sen Roy, S. Talukdar, and Shahfahad (Cham: Springer). doi: 10.1007/978-3-031-21587-2_17

Crossref Full Text | Google Scholar

Nath, B., Wang, Z., Ge, Y., Islam, K., Singh, R. P., and Niu, Z. (2020). Land use and land cover change modeling and future potential landscape risk assessment using markov-ca model and analytical hierarchy process. ISPRS Int. J. Geo-Inform. 9:134. doi: 10.3390/ijgi9020134

Crossref Full Text | Google Scholar

Nurwanda, A., and Honjo, T. (2020). The prediction of city expansion and land surface temperature in Bogor City, Indonesia. Sust. Cities Soc. 52:101772. doi: 10.1016/j.scs.2019.101772

Crossref Full Text | Google Scholar

Ozturk, D. (2015). Urban growth simulation of Atakum (Samsun, Turkey) using cellular automata-Markov chain and multi-layer perceptron-Markov chain models. Rem. Sens. 7, 5918–5950. doi: 10.3390/rs70505918

Crossref Full Text | Google Scholar

Pahlavani, P., Omran, H. A., and Bigdeli, B. (2017). A multiple land use change model based on artificial neural network, Markov chain, and multi objective land allocation. Earth Observ. Geomat. Eng. 1, 82–99. doi: 10.22059/eoge.2017.220342.1006 Available online at: https://eoge.ut.ac.ir/article_64289_1f447d1da4afb649b18be68de046c583.pdf

Crossref Full Text | Google Scholar

Phuc, N. Q., Westen, A. C. M., and Zoomers, A. (2014). Agricultural land for urban development: the process of land conversion in Central Vietnam. Habitat Int. 41, 1–7. doi: 10.1016/j.habitatint.2013.06.004

Crossref Full Text | Google Scholar

Pontius, R. G. (2000). Quantification error versus location error in comparison of categorical maps. Photogramm. Eng. Rem. Sens. 66, 1011–1016. Available online at: https://www.asprs.org/wpcontent/uploads/pers/2000journal/august/2000_aug_1011-1016.pdf

Google Scholar

Pontius, R. G., Boersma, W., Castella, J.-C., Clarke, K., de Nijs, T., Dietzel, C., et al. (2008). Comparing the input, output, and validation maps for several models of land change. Ann. Reg. Sci. 42, 11–37. doi: 10.1007/s00168-007-0138-2

Crossref Full Text | Google Scholar

Prestele, R., Alexander, P., Rounsevell, M. D. A., Arneth, A., Calvin, K., Doelman, J., et al. (2016). Hotspots of uncertainty in land-use and land-cover change projections: a global-scale model comparison. Glob. Change Biol. 22, 3967–3983. doi: 10.1111/gcb.13337

PubMed Abstract | Crossref Full Text | Google Scholar

Ralha, C. G., Abreu, C. G., Coelho, C. G. C., Zaghetto, A., Macchiavello, B., and Machado, R. B. (2013). A multi-agent model system for land-use change simulation. Environ. Modell. Softw. 42, 30–46. doi: 10.1016/j.envsoft.2012.12.003

Crossref Full Text | Google Scholar

Reilly, M. K., O'Mara, M. P., and Seto, K. C. (2009). From Bangalore to the Bay Area: comparing transportation and activity accessibility as drivers of urban growth. Landsc. Urban Plan. 92, 24–33. doi: 10.1016/j.landurbplan.2009.02.001

Crossref Full Text | Google Scholar

Riccioli, F., El Asmar, T., El Asmar, J.-P., Fagarazzi, C., and Casini, L. (2016). Artificial neural network for multifunctional areas. Environ. Monit. Assess. 188:67. doi: 10.1007/s10661-015-5072-7

PubMed Abstract | Crossref Full Text | Google Scholar

Rimal, B., Sloan, S., Keshtkar, H., Sharma, R., Rijal, S., and Shrestha, U. B. (2020). Patterns of historical and future urban expansion in Nepal. Remote Sens. 12:628. doi: 10.3390/rs12040628

Crossref Full Text | Google Scholar

Rimal, B., Zhang, L., Keshtkar, H., Haack, B., Rijal, S., and Zhang, P. (2018). Land use/land cover dynamics and modeling of urban land expansion by the integration of cellular automata and Markov Chain. ISPRS Int. J. Geo-Inform. 7:154. doi: 10.3390/ijgi7040154

Crossref Full Text | Google Scholar

Rodriguez-Galiano, V. F., Ghimire, B., Rogan, J., Chica-Olmo, M., and Rigol-Sanchez, J. P. (2012). An assessment of the effectiveness of a random forest classifier for land-cover classification. ISPRS J. Photogrammetry Rem. Sens. 67, 93–104. doi: 10.1016/j.isprsjprs.2011.11.002

Crossref Full Text | Google Scholar

Rutherford, G. N., Bebi, P., Edwards, P. J., and Zimmermann, N. E. (2008). Assessing land-use statistics to model land cover change in a mountainous landscape in the European Alps. Ecol. Modell. 212, 460–471. doi: 10.1016/j.ecolmodel.2007.10.050

Crossref Full Text | Google Scholar

Salem, M., Tsurusaki, N., and Divigalpitiya, P. (2020). Remote sensing-based detection of agricultural land losses around Greater Cairo since the Egyptian revolution of 2011. Land Use Policy 97:104744. doi: 10.1016/j.landusepol.2020.104744

Crossref Full Text | Google Scholar

Sangermano, F., Toledano, J., and Eastman, J. R. (2012). Land cover change in the Bolivian Amazon and its implications for REDD+ and endemic biodiversity. Landsc. Ecol. 27, 571–584. doi: 10.1007/s10980-012-9710-y

Crossref Full Text | Google Scholar

Schreinemachers, P., and Berger, T. (2006). Land use decisions in developing countries and their representation in multi-agent systems. J. Land Use Sci. 1, 29–44. doi: 10.1080/17474230600605202

Crossref Full Text | Google Scholar

Seto, K. C., Güneralp, B., and Hutyra, L. R. (2012). Global forecasts of urban expansion to 2030 and direct impacts on biodiversity and carbon pools. Proc. Nat. Acad. Sci. U. S. A. 109, 16083–16088. doi: 10.1073/pnas.1211658109

PubMed Abstract | Crossref Full Text | Google Scholar

Shade, C., and Kremer, P. (2019). Predicting land use changes in philadelphia following green infrastructure policies. Land 8:28. doi: 10.3390/land8020028

Crossref Full Text | Google Scholar

Shi, Y., Wu, J., and Shi, S. (2017). Study of the simulated expansion boundary of construction land in shanghai based on a SLEUTH Model. Sustainability 9:876. doi: 10.3390/su9060876

Crossref Full Text | Google Scholar

Simwanda, M., Murayama, Y., Phiri, D., Nyirenda, V. R., and Ranagalage, M. (2021). Simulating scenarios of future intra-urban land-use expansion based on the neural network–Markov model: a case study of Lusaka, Zambia. Rem. Sens. 13:942. doi: 10.3390/rs13050942

Crossref Full Text | Google Scholar

Singh, R. (2015). Banaras, the cultural capital of India: visioning cultural heritage and planning. SANDHI 1, 100–122. Available online at: https://www.svabhinava.org/HinduCivilization/RanaPBSingh/BanarasCulturalCapitalofIndia.pdf

Google Scholar

Singh, R. P. B. (2009c). Ecospirituality: Organic Vision and Global Message; in His Uprooting Geographic Thoughts in India: Toward Ethics, Ecology and Culture in 21st Century. Newcastle upon Tyne: Cambridge Scholars Publishing, 115–152.

Google Scholar

Singh, R. P. B. (n.d.). Banaras, the Cultural Capital of India: Visioning Cultural Heritage and Planning. Available online at: https://www.researchgate.net/publication/320003111

Google Scholar

Singh, S., Kumar, P., Parijat, R., Gonengcil, B., and Rai, A. (2024). Establishing the relationship between land use land cover, normalized difference vegetation index and land surface temperature: a case of Lower Son River Basin, India. Geogr. Sust. 5, 265–275. doi: 10.1016/j.geosus.2023.11.006

Crossref Full Text | Google Scholar

Song, J., Fu, X., Gu, Y., Deng, Y., and Peng, Z.-R. (2017). An examination of land use impacts of flooding induced by sea level rise. Nat. Haz. Earth Syst. Sci. 17, 315–334. doi: 10.5194/nhess-17-315-2017

Crossref Full Text | Google Scholar

Soni, P. K., Rajpal, N., Mehta, R., and Mishra, V. K. (2022). Urban land cover and land use classification using multispectral sentinal-2 imagery. Multimed. Tools Appl. 81, 36853–36867. doi: 10.1007/s11042-021-10991-0

Crossref Full Text | Google Scholar

Srivastava, P. K., Han, D., Rico-Ramirez, M. A., Bray, M., and Islam, T. (2012). Selection of classification techniques for land use/land cover change investigation. Adv. Space Res. 50, 1250–1265. doi: 10.1016/j.asr.2012.06.032

Crossref Full Text | Google Scholar

Tang, J., Wang, L., and Yao, Z. (2007). Spatio-temporal urban landscape change analysis using the Markov chain model and a modified genetic algorithm. Int. J. Rem. Sens. 28, 3255–3271. doi: 10.1080/01431160600962749

Crossref Full Text | Google Scholar

Te Linde, A. H., Bubeck, P., Dekkers, J. E. C., de Moel, H., and Aerts, J. C. J. H. (2011). Future flood risk estimates along the river Rhine. Nat. Haz. Earth Syst. Sci. 11, 459–473. doi: 10.5194/nhess-11-459-2011

Crossref Full Text | Google Scholar

Ul Din, S., and Mak, H. W. L. (2021). Retrieval of Land-use/land cover change (LUCC) maps and Urban expansion dynamics of hyderabad, pakistan via landsat datasets and support vector machine framework. Rem. Sens. 13:3337. doi: 10.3390/rs13163337

Crossref Full Text | Google Scholar

UN DESA (2014). Available online at: https://www.iom.int/sites/g/files/tmzbdl2616/files/2018-07/WMR-2015-Background-Paper-RBhagat.pdf

Google Scholar

United Nations (2017). Available online at: https://www.un.org/uk/desa/68-world-population-projectedlive-urban-areas-2050-saysun#:~:text=68%25%20of%20the%20world%20population,2050%2C%20says%20UN%20%7C%20United%20Nations

Google Scholar

Wang, J., Bretz, M., Dewan, M. A. A., and Delavar, M. A. (2022). Machine learning in modelling land-use and land cover-change (LULCC): current status, challenges and prospects. Sci. Tot. Environ. 822:153559. doi: 10.1016/J.SCITOTENV.2022.153559

PubMed Abstract | Crossref Full Text | Google Scholar

Yang, W., Dai, D., Triggs, B., and Xia, G.-S. (2012). SARBased terrain classification using weakly supervised hierarchical Markov aspect models. IEEE Trans. Image Process. 21, 4232–4243. doi: 10.1109/TIP.2012.2199127

Crossref Full Text | Google Scholar

Ye, Y., Zhang, H., Liu, K., and Wu, Q. (2013). Research on the influence of site factors on the expansion of construction land in the Pearl River Delta, China: by using GIS and remote sensing. Int. J. Appl. Earth Observ. Geoinform. 21, 366–373. doi: 10.1016/j.jag.2011.10.012

Crossref Full Text | Google Scholar

You, H. (2017). Agricultural landscape dynamics in response to economic transition: comparisons between different spatial planning zones in Ningbo region, China. Land Use Policy 61, 316–328. doi: 10.1016/j.landusepol.2016.11.025

Crossref Full Text | Google Scholar

Zang, S., and Huang, X. (2006). An aggregated multivariate regression land-use model and its application to land-use change processes in the Daqing region (northeast China). Ecol. Modell. 193, 503–516. doi: 10.1016/j.ecolmodel.2005.08.026

Crossref Full Text | Google Scholar

Zanganeh Shahraki, S., Sauri, D., Serra, P., Modugno, S., Seifolddini, F., and Pourahmad, A. (2011). Urban sprawl pattern and land-use change detection in Yazd, Iran. Habit. Int. 35, 521–528. doi: 10.1016/J.HABITATINT.2011.02.004

Crossref Full Text | Google Scholar

Zhang, H., Zhou, L.-G., Chen, M.-N., and Ma, W.-C. (2011). Land use dynamics of the fast-growing Shanghai Metropolis, China (1979–2008) and its implications for land use and urban planning policy. Sensors 11, 1794–1809. doi: 10.3390/s110201794

PubMed Abstract | Crossref Full Text | Google Scholar

Zhao, Q., Haseeb, M., Wang, X., Zheng, X., Tahir, Z., Ghafoor, S., et al. (2024). Evaluation of land use land cover changes in response to land surface temperature with satellite indices and remote sensing data. Rangeland Ecol. Manag. 96, 183–196. doi: 10.1016/j.rama.2024.07.003

Crossref Full Text | Google Scholar

Zhao, S., Da, L., Tang, Z., Fang, H., Song, K., and Fang, J. (2006). Ecological consequences of rapid urban expansion: Shanghai, China. Front. Ecol. Environ. 4, 341–346. doi: 10.1890/1540-9295(2006)004[0341:ECORUE]]2.0.CO

Crossref Full Text | Google Scholar