1 Introduction

Post India’s independence in 1947, peace and stability followed, this led to a surge in population due to improved healthcare, economic development, and migration. Several government measures were taken after the independence resulting in reducing mortality rates and stimulating economic growth. Factors like the Green Revolution and migration also played important roles in influencing positive population growth trends [1]. However, the population growth is not uniform throughout the country. Compared to the hilly areas, the plainlands experienced faster growth due to easy access to resources and fertile lands like of the Gangetic alluvial plains. From the time of British India, the Himalayan mountain region served as a popular tourist destination due to its colder climate and scenic beauty. Post-independence, these regions experienced a rapid expansion in the tourism industry, and they became a major economic contributor to the Himalayan states. The hill stations such as Shimla, Nainital, and Darjeeling became major tourist destinations as the basic amenities were already established there. The hotels and hospitality industry also expanded complementing the tourism industry. These areas experienced steady growth in the annual number of tourists except during the global economic recession in 2008 and the COVID-19 pandemic in 2020 (Fig. 1a). This led to unprecedented growth in population and urbanization (Fig. 1b) in the region. It is mostly due to unregulated hotels and homestays established in the areas that provide a financial resource for the local people.

Fig. 1
figure 1

a Steady-state growth of tourism in the districts of Shimla and Nainital, temporarily hindered by the global economic recession of 2008 and the COVID-19 pandemic of 2020. b population and built-up area growth in three hilly Districts of Shimla, Nainital, and Darjeeling. Readers are referred to the online version of the manuscript for better visualization of the color figures

Full size image

Rapid urbanization puts pressure on the local ecosystem, and its effect is radical in vulnerable hilly regions often built up on semi-stable slopes [2]. India’s rapid population and economic expansion highlight the growing need for improved urban development planning [3]. The hilly regions are renowned for their tourism, and a deficiency exists in sustainable urban development which can accommodate a large and growing population [3]. This is due to the lack of well-informed urban planning practices [4]. We have chosen three Himalayan districts, specifically Shimla in the state of Himachal Pradesh, Nainital in Uttarakhand state, and Darjeeling in the state of West Bengal. These districts have had hill stations since before India gained independence. These three districts experienced an exponential population growth (Fig. 1b) that deteriorated the infrastructural facilities and reduced open spaces and agricultural lands [5]. Based on the observation of Fig. 1b it can be assumed that an increase in population is directly related to the increase in built-up areas.

The last few decades saw an exponential population growth in the regions which has also resulted in a decline in the quality of urban environments [6]. As a result, conflicts are often observed between population growth and the type of land use [7, 8]. It appears that some cities have exceeded their capacity for urbanization, which has led to several problems such as air pollution, noise pollution, extensive traffic jams, an increase in waste, and degradation of the land [9,10,11]. This may be attributed to a lack of planning and/or lack of full implementation of government policies for land use [12]. With the ongoing trend of urbanization, rapid population increase, and shift in climate patterns, a significant increase in the recurrence of natural disasters is also observed. Notably, in hilly regions, the frequency of rainfall and related natural hazards have increased significantly [13,14,15] but have not deterred the growth in the tourism industry and population [16]. With the dual challenges of expanding population and rising frequency of natural hazards, demand for identifying new areas suitable and sustainable for urban expansion in hill stations has become the need of the hour in recent years. Several factors should be considered before developing built-up areas in hilly regions sustainably and suitably to avoid the ill effects of many frequently occurring natural hazards like floods, landslides, earthquakes, etc.

The concept of land-use suitability in urban development is used primarily to determine whether a particular area of land is suitable for a particular land-use type and to assess the potential for alternative land uses based on a wide range of environmental, social, and economic criteria [17]. There have been many studies conducted recently to assess the suitability of land use in particular areas, cities, or a country [18,19,20,21]. Land use suitability depends not only on one factor, rather several factors have an influence on the decision making. And the degree of influence of each factor varies depending on whether it promotes or hinders the respective land use. In order to assess land use suitability, the most common method used is GIS-MCDA (Geographical Information System collaborative with Multi-Criteria Decision Analysis) or MCDMA (Multi-Criteria Decision-Making Approach) [22]. The MCDA approach considers a wide range of factors for decision-making and assigns a weight to each factor. The MCDA typically selects two categories of opportunity criteria as well as one category of constraint criteria, namely natural environment-social economy criteria [23] potential constraint force criteria [24], and resistance-motivation criteria [25]. There are several other methods under the MCDMA [26, 27]. Each of them has its own calculation method. The Methods can be selected by the type of desired outcome. If the outcomes are required as a comparison of values, AHP, MULTIMOORA, MAUT, Weighted Sum Method, Weighted Product Method can be used [26, 27]. While AHP, TOPSIS, and VIKOR, COPRAS, STEP can be used to get the desired outcome of finding the best alternative [28]. The methods PROMETHEE and ELECTRE are based on pairwise comparison [28]. In addition to overlay mapping and artificial intelligence, it is one of three methods that have been used worldwide for analyzing land use suitability.

Multiple criteria can be combined using different methods, including Weight Linear Combination (WLC) [18], Weighted Potential Constraint Method (WPCM), Land Suitability Index (LSI) [29], and Logic Scoring of Preference (LSP) [30]. These methods are used to weigh multiple factors in combination using the Analytical Hierarchy Process (AHP) [19], Analytic Network Process (ANP) [17] and Ordered Weighted Average (OWA) [31]. As a result of combining AHP with MCDA, which utilizes the LSP method, we have obtained suitability assessments, which are usually based on specific requirements or preferences [32]. By using GIS-MCDA, we have generated a very simplified and generalized but accurate and justifiable result to assess future urban development.

In this study, we have chosen a set of factors, such as slope, elevation, proximity to roads, tectonic faults, waterbodies, and urban amenities, landslide density, that affect the land use suitability for urbanization. The steep and rugged topography of the tectonically active Himalayan region with a well-distributed river drainage network, heavy monsoon, and triggered landslides, shows extensive soil erosion at some places. Thus, it is also an important criterion for urban development. The Revised Universal Soil Loss Equation (RUSLE) model is utilized to compute the soil erosion factor for the studied Himalayan districts.

A discernible trend has emerged indicating a direct correlation between population growth [16] and the expansion of built-up areas. By extrapolating population trends, it becomes evident that over the next 20 to 30 years, there will be a considerable surge in population, leading to increased urban development (https://ghsl.jrc.ec.europa.eu/ghs_bu.php) (Fig. 1a and b). This underscores the pressing need for a sustainable development strategy.

2 Study area

Hill stations captivate with their natural beauty but face natural hazards like landslides, cloud bursts, heavy rain, and earthquakes. Mountains are highly deformed, and their isolation from the populated areas worsens the adverse effects of disasters [33]. Rivers in hilly areas erode the mountain faster than in plains due to their strong flow. The Himalayas being a popular tourist destination attracts many tourists which is mostly owed to its scenic beauty and colder climate. The fact that the region is dominated by highly deformed rocks which are still actively deforming [34], makes the region vulnerable to huge property and life loss due to natural disasters. Thus, urbanization must be done with utmost care keeping a balance between human necessity and environmental exploitation. Our study focuses on three popular Himalayan hill stations namely Nainital, Darjeeling, and Shimla (Fig. 2). These areas face an exponentially increasing trend of population and built-up area growth.

Fig. 2
figure 2

Location of the study areas on the Indian subcontinent. a–c show the digital elevation model of Shimla, Nainital, and Darjeeling, respectively. Readers are referred to the online version of the manuscript for better visualization of the color figures

Full size image

Nainital which is situated in the southern part of Uttarakhand, is bordered by Almora and Champawat to the north and Udham Singh Nagar to the south. It spans 3860 km2, with plains in the south and mountains in the north, crossed by the Kosi River. The area experiences an annual rainfall of 330 cm (Government of Uttarakhand: https://nainital.nic.in/geography/). Slope in the region varies between 0° and 70° with a mean slope of 16°. The area experiences a vast number of natural hazards with earthquakes and landslides being the major hazards due to high rainfall intensity and the presence of tectonic faults [35]. It lies at the start of the middle Himalayas, between the Main Boundary Thrust (MBT) and the Main Central Thrust (MCT) [36]. The area experiences a significant number of earthquakes as it is seismically active and lies in zone IV of the earthquake zoning map of India [35].

Darjeeling district is situated in the northern part of Indian state of West Bengal. It is bounded to the north by Sikkim, to the east by Bhutan and to the west by Nepal [37]. It covers an area of 2092.5 km2 [38] in the Eastern Himalayas. We focused on the western part of Darjeeling district, the Sadar subdivision, which is susceptible to soil erosion due to its high elevation. The area experiences an annual rainfall of 310 cm [38]. Slope in the region varies between 0° and 65° with a mean slope of 25°. Heavy rainfall in the monsoon and steep slopes in the region are the main reasons for frequent landslides in the area. The major tectonic feature in the studied part of the Darjeeling district is the presence of the Main Central Thrust (MCT), which separates the Higher Himalayas from the Lesser Himalayan regions [34]. Recently, this area has been experiencing 2–3 major landslides every month. For instance, in June 2023, torrential rain for 24 h triggered multiple landslides, devastating several households, roads, etc. (https://www.telegraphindia.com/west-bengal/incessant-rain-over-past-24-hours-in-darjeeling-hills-triggers-multiple-landslides/cid/1947301). The major river system in the region consists of the Lodhoma River, the Rangeet River flowing along the northern boundary, and Teesta flowing along the western boundary of the region [34].

Shimla, once India's summer capital, lies in Himachal Pradesh's southeast and spans 5131 km2. These districts, vital in tourism and sitting around 2100 m above sea level, were chosen. The area experiences an annual rainfall of 142 cm (Government of Himachal Pradesh: https://hpshimla.nic.in/history/). Slope in the region varies between 0° and 75° with a mean slope of 29°. During the monsoon season of 2023, heavy rainfall triggered several landslides that stranded the tourists and pilgrims under the rubble. It called for an 11-day-long rescue operation and took around 20 human lives while many others went missing (https://timesofindia.indiatimes.com/city/shimla/shimla-landslide-search-operation-concludes-with-recovery-of-3-bodies-death-toll-reaches-20/articleshow/103031176.cms).

The tourism surge worldwide drives the need for more hotels and restaurants, leading to increased built-up areas. We have chosen these three districts to gauge their suitability for future urban development, considering factors like geography, faults, slope, and more, aiming for sustainable growth. Even though most of the developed areas in these three districts are in low erosion zones, it is concerning that they are built so close to the fault zones. For Shimla around 89% of the total built-up area, for Nainital around 78% and for Darjeeling around 86% of the total built-up area (Fig. 3a) is situated at very low to moderate distance from the faults. Fortunately, the built-up areas in these three districts are largely located far from the landslide-prone zones. Around 73% of built-up areas in Shimla, 86% in Nainital, and approximately 77% in Darjeeling are situated within low to very low-density landslide areas (Fig. 3b). Built-up areas near rivers in the plainlands provide easy access to daily water supply and transport. However, in the hilly regions, rivers are often susceptible to flash floods due to cloud bursts and landslides on river channels. During the twenty-first century, the adverse effects of global warming and climate change have accelerated the deglaciation process. The sudden increase in drainage volume triggers landslides near the riverbanks by toe erosion. Thus, closer proximity to rivers poses a serious threat to human lives and infrastructural facilities. Within the studied Himalayan districts, in Shimla 64%, in Nainital 80%, and in Darjeeling 70% of the total built-up area is situated close to very close of the rivers (Fig. 3c). This shows most of the built-up area is susceptible to above mentioned natural hazards. The built-up area data is collected from Global Human Settlement Layer (https://ghsl.jrc.ec.europa.eu/) and the process of determining the proximity is mentioned in the methodology section.

Fig. 3
figure 3

Distribution of 2020 built-up areas with the classification schemes of a proximity to faults, b proximity to landslide dense zone, and c. proximity to rivers for studied areas. Readers are referred to the online version of the manuscript for better visualization of the color figures

Full size image

After careful consideration of all these aspects, we have selected these three Himalayan districts for the study. While there are other districts of the Himalayan Mountain Belt that experience even harsher climatic conditions, these three regions were selected due to their historical popularity as hill stations and the ongoing growth in population and urbanization. The continuous population growth and subsequent environmental exploitation to meet the socio-economic demands put pressure on natural resources. Thus, a sustainable urbanization planning with proper food and environmental safety is need of the hour for the developing nations [39]. Furthermore, the escalating pressure exerted by expanding populations and the concomitant competition stemming from variances in land utilization necessitate enhanced efficiency in land use and management practices. Urbanization should not hinder the ecological and geological factors and must preserve the environment for future generation [39]. The main objective of this study is to identify such regions in the studied regions of tectonically active Himalayas. Furthermore, the core hypothesis of this study serves not only to pinpoint lands suitable for urbanization use but also aims to support the sustainable management of these lands, considering the characteristics of soil erosion. Additionally, it also reviews the effects of individual factors on site characterization for urbanization.

3 Methodology

3.1 Work plan

Our primary aim of the study is to determine the suitability of urban expansion, achieved through AHP under the MCDMA family of processes. To create a sustainable map for urban suitability, we concentrated on nine key factors: slope, elevation, LULC (land use and land cover), distance from roads, faults, rivers, urban amenities, landslide density, and erosion. Within these criteria, we specifically consider soil erosion, assessed via the RUSLE model. The workflow for the land use suitability analysis for urbanization is given in Fig. 4.

Fig. 4
figure 4

Work plan strategy for deriving urban suitability map

Full size image

We utilized the AHP technique within the GIS software ArcGIS to implement a strategic approach. This strategy involves categorizing factors into four distinct groups, Topography and Geology factors, Socioeconomic factors, Ecological factors, and Prohibitive factors. Subsequently, we further classified these groups into relevant subfactors. Then we clubbed all these factors together and used to combine all the suitability factors, generating a suitability map (Fig. 4). The data essential for this study was derived from various sources which are mentioned below in Supplementary Table S1.

3.2 RUSLE model

The revised Universal Soil Loss Equation (RUSLE) model is used to find soil erosion in three study areas. It is an empirical model to carry out soil loss analysis in larger areas like at district levels or state levels. Globally, many researchers have used this model for carrying out soil loss analysis [40,41,42]. As it is a very simple model and can be integrated with GIS, it has gained acceptance globally [43]. This model generates the average annual soil loss (A), calculated as the product of several factors: the Rainfall erosivity factor (R), Soil erodibility factor (K), Crop cover factor (C), Slope steepness and slope length factor (LS), and Conservation practice factor (P) [44]. Where C, P, and LS are dimensionless quantities, and R has a dimension of MJ mm/ha/h/yr., and K has a dimension of t.ha.h /ha/MJ/mm that results in the dimension of A as t/ha/yr.

3.2.1 Rainfall erosivity factor (R)

The R factor provides a significant measure of rainfall's capacity to displace soil particles at a given location. This factor relies on the kinetic energy of rainfall, which is influenced by the intensity, duration, and volume of rainfall [44]. We utilized a global erosivity map created by the European Soil Data Centre (ESDAC) (https://esdac.jrc.ec.europa.eu/), as comprehensive temporal and high-resolution data for our entire study area was unavailable. We extracted values specific to our study area from this global map. Using the extracted rainfall data, the R factor was calculated following the formula proposed by [45].

$$R=79+0.363\times {x}_{a}$$
(1)

where R describes the rainfall erosivity factor and \({x}_{a}\) denotes yearly rainfall in mm.

3.2.2 Soil erodibility factor (K)

The K factor is a metric that indicates the vulnerability of different soil types to erosion. It is normalized based on the rate and volume of runoff [46]. Soil characteristics such as grain size, porosity, permeability, water retention capacity, and infiltration rates vary across different soil types, influencing their erosion resistance. Determination of the K factor relies on soil texture, organic content, structure, and permeability [47].

The detailed small-scale soil data is inadequate for our study area. Thus prompting the utilization of the Digital Soil Map of the World (DSMW), which was clipped to fit our study area. To assess the K factor, we employed the EPIC model equation based on soil composition—sand, silt, clay, and organic carbon percentages [48]. We followed the EPIC equation as utilized by [49].

$$K= 0.1317\times { F}_{si-cl} \times { F}_{cSand} \times { F}_{Orgc} \times { F}_{hisand}$$
(2)

This equation was then simplified for each content of the soil as

$${F}_{si-cl}={\left[\frac{SIL}{CLA+SIL}\right]}^{0\cdot 3}$$
(3)
$${F}_{cSand}=\left[0\cdot 2+0\cdot 3expexp\left(-0.0256SAN\left(1-\frac{SIL}{100}\right)\right)\right]$$
(4)
$${F}_{Orgc}=\left[1\cdot 0-\frac{0\cdot 25C}{c+expexp\left(3.72-2\cdot 95C\right)}\right]$$
(5)
$${F}_{hisand}=\left[1\cdot 0-\frac{0\cdot 70SN1}{SN1+expexp\left(-5\cdot 51+22.9SN1\right)}\right]$$
(6)

SAN is Sand weight content (%), SIL is Silt weight content, CLA is Clay weight, TOC is Organic carbon content, \(SN1=1-\left(\frac{SAN}{100}\right)\). Here \({F}_{cSand}\) is for soil, \({F}_{si-cl}\) is for soil with high clay to silt ratio, \({F}_{Orgc}\) is for soil with high organic content and \({F}_{hisand}\) is for soil with high sand content [50].

3.2.3 Slope steepness and slope length factor (LS)

Topography, particularly the slope's length (L) and steepness (S), is a pivotal factor influencing erosion rate calculation in hilly regions. This dimensionless quantity shows values ranging from zero to above, roughly derived as the ratio of soil loss in an area with specific L and S to a standard RUSLE unit with L (22.13 m) and S (9%) [44]. Using the DEM, an LS equation [51] was developed to generate a topographic factor map. To compute LS, we generated a flow accumulation map of the study area and employed an equation derived from the works of [52, 53].

$$LS=\left({\left(\frac{Flow \,Accumulation \times Resolution}{22\cdot 1}\right)}^{0\cdot 5}\times \left(0.065+0.045\times \%+0.0065 \times (slope \%\times slope \%\right))\right)$$
(7)

We utilized a 30 m DEM to compute the LS factor in our study areas. The cell grid size significantly impacts the calculation of LS, particularly influencing the S factor, given that slope values vary with cell size [54].

3.2.4 Crop cover factor (C)

C represents the crop management factor, essential in regulating the impact of soil and crop practices on soil loss prevention. Within the RUSLE model, the C factor assigns values based on crop patterns, types, and cropping management practices to enhance model accuracy [55]. It signifies the ratio of soil loss under specific cultivation methods compared to clean-tilled continuous fallow conditions. The significance of this factor lies in its susceptibility to be influenced by farmers at a local scale. The regression equation used for calculating C was as follows:

$$C=1.02-\left(1\cdot 21\times NDVI\right)$$
(8)

As NDVI values exhibit a relationship with the C factor [56,57,58], our calculation of the C factor via NDVI involved regression analysis—a method previously employed by several researchers [59,60,61,62].

For the assessment of NDVI values, we employed satellite data sourced from LANDSAT 8 to determine this factor, aligning with the work of [41]. We used Band 4 and Band 5 from Landsat 8 to find NDVI.

$$NDVI=\frac{NIR-RED}{NIR+RED}$$
(9)

Here, NIR (near-infrared) and RED (visible regions) represent the spectral reflectance used for calculation. The NDVI values range from − 1 to 1, where a value of 1 signifies dense vegetation, while lower values correspond to water or bare ground [53].

The C factor's value hinges on various factors like vegetation type, growth, and crop cover percentage within the area [63]. It spans a range from 0 (well-protected) to 1 (bare soil), contingent upon land cover types [64]. Higher C values indicate a higher potential for soil erosion due to poor vegetation cover and poor land management practices.

3.3 Conservation practice factor (P)

Denotes the rate of soil loss associated with agricultural practices prevalent in the region, showcasing the impact of practices crucial in mitigating runoff water and, consequently, reducing erosion [65]. Commonly, three types of cropland practices prevail, such as strip cropping, contour cropping, and terrace cropping [66]. The P factor spans values from 0 to 1, where 0 signifies primarily human-induced erosion and 1 represents predominantly non-anthropogenic erosion [49, 67]. Values for cropping patterns have been pre-assigned (Supplementary Table S2), and we employed these values in our study, following the works of [68, 69].

3.4 Site suitability analysis with analytical hierarchy process (AHP)

The analytical hierarchy process serves as a tool that simplifies complex problems into a more manageable form, facilitating improved decision-making. We employed the MCDMA for our work, focusing on nine criteria through AHP analysis. The functioning of AHP is primarily guided by three principles outlined by [70, 71]: problem identification, comparative judgment, and establishment of relative importance.

Within AHP, the problem is deconstructed into multiple criteria, facilitating comparisons between them through pairwise assessments. These pairwise comparisons are pivotal steps, wherein rankings are computed for each factor via the eigenvector method. Subsequently, the consistency of the matrix is evaluated using the consistency ratio [70]. The Saaty scale, presented in Supplementary Table S3, governs the pairwise comparison scale. The consistency of weights assigned to factors, crucial for determining relative importance, is calculated using a specific formula:

$$Consistency \,ratio \left(CR\right)=\frac{CI}{RI}$$
(10)

where CI is the consistency index and RI is the randomness index.

$$CI=\frac{{\lambda }_{max}-n}{n-1}$$
(11)

where n is the number of criteria being compared or the order of the matrix and \({\lambda }_{max}\) is the largest eigenvalue of the matrix.

Saaty [70] has already provided RI values, which rely on the matrix's order, indicating the number of factors considered in AHP (Supplementary Table S3). For consistency to be deemed satisfactory, the CR values should be below 0.1. If CR values exceed 0.1, indicating inconsistency, a corrective measure involves reassigning weights for each factor in the pairwise matrix [72]. Subsequently, following the AHP process, all maps are overlaid to generate a composite suitability map.

$$\text{Suitability}=\sum \left(Criteria\times Weight\right)$$
(12)

3.4.1 Factors employed in AHP analysis

Elevation and slope The elevation data were obtained using the ASTER (30 m) Digital Elevation Model (DEM). The slope value is calculated using the DEMs in the ArcGIS environment. High elevation and steep slopes both pose serious challenges for urbanization as these areas are often disturbed by landslides. Thus, the higher classes of these factors signify a low suitability rank.

Distance from road Assessing proximity to roads involved calculating distances using district-specific road shapefiles and the Euclidean distance tool in ArcGIS. Proximity to roads provides better connectivity for the urban areas thus, it has a positive connotation on the land use suitability for urbanization. Hence, we provided a high suitability rank for the areas that are closer to the roads.

Distance from river, fault, urban amenities, and landslide density Determining distances from various features, such as rivers, faults, urban amenities, and landslide points, we employed corresponding shapefiles and the Euclidean distance tool in ArcGIS. Torrential rainfall and deglaciation often heavily flood the rivers in the hilly regions which increases the risk of landslides and flooding near the riverbanks. Thus, the urbanization should occur at a distance from the rivers. Hence we assigned a very low suitability class with a suitability rating of 1 to the nearest regions of the rivers. Subsequently, we assigned a very high suitability class with a suitability ranking of 5 to the distant regions. Similarly, the regions closer to the fault zones are assigned a very low suitability class and vice-versa. Since the Himalayan regions are tectonically active, a ‘very close’ proximity to the faults enhances the risk of earthquake-related damages [37]. The weak rocks in the fault damage zones also increase the chances of slope failure [34, 37]. Thus, all these three factors have negative connotations with urbanization suitability. Meanwhile, if the settlements are near basic amenities, it favors urbanization. Thus, it has a positive connotation with urbanization suitability.

Erosion Soil erosion plays a pivotal factor in hilly regions due to steep slopes and high elevation, and it was quantified using the RUSLE model. This factor significantly influences urban development planning and enhances the comprehensiveness of the AHP derived suitability assessment. It has a negative impact on the urbanization suitability. Thus, the areas with high erosion rates are assigned with low suitability class rank.

Land use and land cover (LULC) We extracted the LULC maps for the studied region from the opensource global LULC time series of the world. The data are produced by Microsoft, Impact Observatory and ESRI using 10 m Sentinel-2 L2A imageries for the years 2017 to 2023. Deep learning AI and land classification model of Impact Observatory is used to generate data for each year [73]. The model utilizes a fully convolutional neural network with a UNet architecture and predicts LULC for nine classes using over 5 billion human-labelled pixels [73]. Karra et al. [73] used the 6-bands of Sentinel-2 L2A surface reflectance data in the deep learning model, and validated the result with human-labelled classifications which were excluded from the training data set. They employed a “strict consensus” method for the assessment of data accuracy and found an accuracy level of 91%. The detailed methodology and accuracy result is described in Karra et al. [73] and Impact Observatory documentations (https://www.impactobservatory.com/legal/lulc-methodology-accuracy.pdf).

Detailed land use and cover maps offer insights into the current landscape, aiding in identifying distribution patterns related to barren land, crop cover, and built-up areas within the study region. This data enables comparison between existing built-up areas and the suitability map derived from AHP analysis. The waterbodies, forest regions, and swamp areas are not suitable for urbanization. Thus, we assign low suitability rank to these classes of the LULC. Bare ground and already built-up areas are favorable for urbanization and we assigned a high suitability rank to these areas. We used the LULC classification map of 2020 for the present study.

Classification of factors Each of the nine factors in the AHP analysis was classified into five classes—very low, low, moderate, high, and very high. This consistent classification approach ensures a comparable and uniform analysis for each factor in assessing urban suitability for land use.

4 Results

We applied the RUSLE equation to assess soil erosion in three hilly region districts: Darjeeling, Nainital, and Shimla. This equation considers five factors, producing a comprehensive soil erosion map for the area. The RUSLE model, widely used in land use management and soil erosion quantification due to its versatility and data availability [74], generates spatial patterns of soil erosion when integrated with a geographical information system (GIS) [75]. This spatial insight helps identify erosion-prone areas [76] and facilitates the implementation of soil conservation programs. Using the soil erosion factors with the other factors in site suitability analysis we calculated the suitability factor for urbanization in the study areas. In this section, we first display the results of the soil erosion analysis and then the site suitability results using the AHP technique.

4.1 Soil erosion

The rainfall erosivity (R) factor value in Shimla, which is derived from ESDAC global R data (https://esdac.jrc.ec.europa.eu/) varies from 794.1 to 7644.6 (Supplementary Fig. S1). Highest values are observed along a narrow zone at the NW part whereas, the NE part shows lowest value. It varies between 2595.34 and 7391.97 (Supplementary Fig. S2) in the Nainital district and in the Darjeeling region it ranges from 1874 to 10,178 (Supplementary Fig. S3). Elevated R values signify increased vulnerability to erosion, attributed to the regions’ high precipitation rate. The southern region of Nainital displays notably high R values, particularly around the foothills. The eastern and western margin of Darjeeling shows highest and lowest R factor value respectively.

The P factor values depends on the cultivation pattern of an area. The preferred method of cultivation in Shimla and Darjeeling is contour cropping, thus we assigned values ranging from 0.55 to 1 for different slope conditions (Supplementary Table S4) [77]. Agricultural practices in Nainital involve strip cropping and P factor values are assigned from 0.27 to 0.50 (Supplementary Table S4) as recommended by Shin [77]. Elevated P factor values in hilly regions suggest relatively poorer support practices.

The C factor shows a wide variation in Shimla which ranges from 0.04 to 1. Nainital and Darjeeling shows lesser variation, and it ranges from 0.429 to 1 and 0.362 to 1 respectively.

The Soil Erodibility factor, known as the K factor, primarily relies on soil type and its permeability [53]. The K factor for Shimla yields two values for the two principal types of soils found here, with K values ranging from 0.020 to 0.022 (Supplementary Fig. S1). The north-eastern part of the district shows lowest K factor value. We identified four distinct soil types in Nainital which are characterized by varying concentrations of sand, silt, clay, and organic carbon. The K values in this region range from 0.014 to 0.022 (Supplementary Fig. S2). A higher K values imply greater vulnerability to soil erosion. From the Fig. S1 it can be seen that a part of foothills area is showing highest K factor value. This area is dominated by loose soil, thus prone to higher erosion. The southern part of the district exhibits the lowest K value, indicating stability, while the hilly areas display moderate K values. Darjeeling has four distinct soil types, resulting in four different values for the K factor, which range from 0.013 to 0.020 (Supplementary Fig. S3).

Meanwhile, the LS factor serves as a direct erosion indicator; higher LS values signify increased susceptibility. Given our focus on hilly terrains, it is expected to encounter higher LS values, calculated considering flow conditions. The LS factor value goes from 0 to 9.641 for Shimla whereas, for Nainital it ranges from 0.01 to 4.39, and for Darjeeling it is from 0 to 2.797. It is predominantly higher in hilly regions and lower in flat areas (Supplementary Fig. S1, S2, and S3).

Upon calculating all five factors, our primary objective was to quantify soil loss. Integrating these parameters revealed soil erosion rate of 507.992 t/ha/yr in Shimla, 127.40 t/ha/yr in Nainital, and 398.910 t/ha/yr in Darjeeling district (Supplementary Fig. S1, S2, and S3), indicating significant soil erosion, consistent with prior studies [78].

4.2 AHP (analytical hierarchy process)

Our study harnesses AHP through MCDMA to derive an urban suitability map, offering a holistic view of the suitability landscape of the area. This map serves as a guide for identifying zones suitable for further development, utilizing the built-up area data from 2020 to provide a comprehensive understanding of ongoing development trends. As mentioned earlier, we chose nine key factors/criteria for the suitability analysis and these criteria are classified into five categories—Very Low, Low, Moderate, High, and Very High enhancing the clarity of suitability assessment.

We categorized these nine factors into five suitability classes (Supplementary Table S5) and established weights for each criterion ensuring a Consistency Ratio (CR) below 0.1 (Table 1).

Table 1 Pairwise comparison matrix of nine causative factors for urban suitability analysis
Full size table

The urbanization suitability results obtained through the AHP method show very high to high suitability near the central and southern parts of Shimla (Fig. 5a). The western part of the district i.e. Shimla town is already highly populated While the northern part shows very low suitability for urbanization, there are some patches of highly suitable areas along the NW boundary of the district. Suitability trends in Nainital show high to very high suitability in the foothill region of the district (Fig. 5b). But unlike Shimla, the hilly region shows very low to moderate suitability. In Nainital, the most suitable areas for urbanization are situated in the alluvial deposits of the Kosi and Gaula Rivers. In the Darjeeling district, suitability trends are patchy, and no concentrated high suitable area is found as observed in the other two districts. The area shows high to very high suitability in the southern region on the eastern boundary (Fig. 5c).

Fig. 5
figure 5

Results of the AHP land-use suitability for urbanization for the areas of a Shimla, b Nainital, and c Darjeeling. Readers are referred to the online version of the manuscript for better visualization of the color figures

Full size image

Then the resulting urban suitability map was compared with the 2020 built-up area for the studied regions to get an idea about the present urbanization trend in terms of suitability (Fig. 6). For Shimla, 73% (Fig. 6a) of the built-up area falls within the high to very high site suitability classification while for Nainital and Darjeeling this value is 56% and 51% respectively (Fig. 6b and c). Interestingly, we observed that only 7% area of Shimla falls within very low to low category, while for Nainital and Darjeeling this value is 24% and 23% respectively.

Fig. 6
figure 6

Intersection between AHP results and built-up area of 2020 for the areas of a Shimla, b Nainital, and c Darjeeling. The number in front of each suitability class denotes percentage of 2020 built-up area falling into that class. Readers are referred to the online version of the manuscript for better visualization of the color figures

Full size image

Rohru tehsil region of Shimla district is built on highly vulnerable areas which are low to very low in terms of site suitability. The hilly regions of Nainital district are mostly low to moderately suitable for urbanization and in the case of the Darjeeling district, 49% of the already built-up lies in the low to moderate class, and there is no concentrated distribution of high to very high suitability, rather it is randomly distributed throughout the district.

Notably, Shimla emerges as the most suitable district for urban planning, while Darjeeling appears to have the least suitable areas. This comparative analysis offers insights into varying suitability levels, aiding potential considerations for urban planning initiatives in each district. The resulting urban suitability map (Fig. 5) highlights that plane areas are more suitable for urban planning in the Nainital district, while hilly and mountainous regions are less suitable for urbanization.

5 Discussions

The district of Shimla is situated within hilly region of the Himalayas and shows maximum elevation distribution among the three studied regions (Fig. 2a). The area shows very low rainfall erosivity or R factor at the highest elevation (~ 5000 m) as these parts of the district is dominated by high mountain peaks where a colder climate prevails. The high R factor at the foothills of the Nainital district is owed to the southwest monsoon and Westerly Disturbances (WD) [78]. The whole Western Himalayan region receives rainfall in two phase, June–September and November-March and the latter is attributed to the WD [79]. According to an experimental study performed by Medina et al. [80], the low-level moist air originating from the Arabian Sea acquires more moisture from the foothills of Western Himalayas and when it is capped by dry and warm continental air flow, it causes heavy downpour at the foothills of the Western Himalayas. During the winters, due to pressure variation caused by the colder Tibetan Plateau and surrounding warmer oceans (northeastern wind), the region experiences rainfall from December to February. Nainital region is situated at the foothills and due to the orographic rainfall, a higher R factor is observed. Interestingly, the R factor in Darjeeling is highest compared to the other studied regions. Darjeeling is situated in the Eastern Himalayas, and the region receives more rainfall compared to the Western Himalayan region [81]. The air originating from the low-pressure zone of Bay of Bengal sometimes travels north to northeasterly and causes heavy rainfall in the Eastern Himalayas [82]. Thus, we see a higher R factor in the Darjeeling region.

Shimla and Darjeeling show a higher P factor (~ 1) indicating non-anthropogenic activities related soil erosion. Darjeeling has its 38% of the total area occupied by the forest cover and 30–40% area by tea garden [83]. Distribution of high C factor also corroborate to the fact that soil erosion is mostly due to anthropogenic activities. We also see overall high C factor in Shimla because the region is situated in mountainous region and deforestation of the hillslopes are done for agricultural practices (Fig. S1). Thus, for the two regions, agricultural practices are the most influential factor in soil erosion. We see overall low P factor value in Nainital and these values are spatially distributed over the foothills region which is an indicative of erosion influenced by anthropogenic activities. This area also shows dense built-up area distribution.

Soil erosion is lowest in Nainital because of its geographical location. The northern part of the district is within the hilly region and the southern part lies on the Himalayan foothills. The regional rivers carry extensive amount of sediments, eroding from the uplifting mountains, and deposits at the foothills. Continuous sediment recharge at the low slope region is responsible for the low erosion rate.

Our land use suitability evaluation for urban development encompasses various factors categorized into four key categories: topography and geology, socio-economic, ecological, and prohibitive factors [8]. We have previously delineated these factors into respective subfactors. Employing GIS software, our approach ensures flexibility, suitability, and improved results compared to prior studies, offering enhanced insights into land use suitability. Upon thorough analysis of population growth patterns across Nainital, Shimla, and Darjeeling districts, we have observed a consistent rise in population post-independence, leading to exponential growth in built-up areas (Fig. 1b). Projections reveal a concerning burden on local ecology due to this expansion, raising hazards. To mitigate risks, transitioning to more suitable land use practices becomes imperative.

Our findings indicate that a considerable portion of built-up areas falls within the moderate to high suitability range. Among the districts, Shimla emerges as the most suitable based on present day built-up area, while Darjeeling and Nainital show similar range of urban suitability (Fig. 6). The most suitable areas in Shimla are around the central and southern parts. The urbanization suitability of Nainital district is concentrated in the southern part along the foothills of the Himalayan mountains. Whereas in Darjeeling, site suitability analysis results show a relatively lesser percentage of high to very high distribution of suitability. While in the other districts, it is concentrated in a region but here it is sporadically present. Here the most suitable areas lie in the southern to central part of the district. Upon considering the present built-up area distribution and site suitability for all the three districts, urbanization trend follows the crop lands and range lands. The most suitable places for urbanization are also around these land use classes. These places are deforested, at low to moderate elevation, and low slope, and mostly used for agricultural purposes. Thus, it brings out a long persisting debate over urbanization and agricultural activities and maintaining food security [84, 85]. A detailed analysis on decreasing agricultural land and increasing food demand is beyond the scope of this work as the balance between these two factors depends on the economic shift of the concerned regions [85]. If a tourism-based economy is preferred over agriculture dependent economy, food security for the locals must be ensured. Over import of food source will impart heavy economical burdens for the residents of these areas. But with growing economy and rising buying power people move over the food chain [85, 86].

The most suitable areas in Nainital are concentrated in the foothills. While huge sedimentation from the Himalayas, as well as flat terrain, high rainfall, and low elevation makes this area suitable for farming, well established road connectivity and greater distance from landslides and seismicity prone areas makes it suitable for urbanization. Thus, a trade off must be set to control the food sustainability. Urbanization can also be focused on the hilly side of the region. This is also beneficial for the tourism industry as the scenic beauty of the Himalayas apart from the ‘tals’ (lakes at the foothills) are the main tourist attraction of the region. We see a moderately high suitability along the north-central side of the district and the region is also farther from the fault lines (Fig. 7). In the studied part of the Darjeeling, the AHP derived land use suitability for agriculture is marginal [87, 88]. Thus, based on slope conditions, elevation, distance from hazardous zones, Darjeeling has a good chance for further urbanization.

Fig. 7
figure 7

Built-up area for a Shimla, b Nainital, and c Darjeeling regions with dominant faults in the areas. Readers are referred to the online version of the manuscript for better visualization of the colour figures

Full size image

Future land use planning must prioritize high or very high suitable areas to ensure long-term suitability. Our study emphasizes nine critical factors contributing to land use suitability assessment, resulting in a comprehensive map delineating optimal zones for urban development. We place a specific focus on high and very high suitable areas, aligning with sustainable growth objectives. The amalgamation of these nine factors underscores their influential role in shaping urban suitability decisions. Each factor significantly impacts built-up development. For example, the urban amenities factor directs development concentration near existing built-up areas, promoting sustainable development. Moreover, the suitability map derived from AHP reflects the intricate interplay of these factors.

Although in the Shimla district, the western part (Shimla town) shows dense built-up area with most of the urban amenities, the area is developed around dense fault zones and landslide points (Fig. 7a). Despite that, the site suitability results show high to moderate suitability for urbanization as all the other factors like slope, elevation, and road connectivity, etc., exhibit positive feedback for urbanization. Since the Britishers already built Shimla town with basic amenities around that region due to its relatively low elevation and slope, the AHP technique prioritized that region in terms of urbanization. The already built-up areas in Nainital are situated at a distance from the faults but at some places near the frontal thrust of the Himalayas, built-up areas are in a close range to the fault zone (Fig. 5b and 7b). This area shows high suitability because of the presence of an alluvial fan of the Kosi River which provides perfect LULC conditions, a relatively flatter slope, and low elevation. Most of the urbanization in Darjeeling area happened near the N-E to NE-SW trending fault zone. This fault, also known as the Main Central Thrust (MCT), separates the Higher Himalayas at the south from the Lesser Himalayas at the north. Nearly all the urban amenities are present along the eastern boundary of the region, away from the main township. The site suitability result for Darjeeling shows moderate to very high suitability near the fault zones as the influence of all the other factors dominates (for example plenty of bare ground) over the effect of fault zones (Fig. 7c).

Our analysis reveals a substantial portion of the area lies in landslide-dense zones, seemingly suitable for urban development due to the influence of other factors. However, excluding regions near these zones reveals a highly suitable map for built-up development, mitigating landslide risks. Similarly, avoiding fault zones further enhances suitability, reducing the potential for casualties due to natural hazards in hilly areas.

Categorizing factors provides a unique perspective on urban development, offering distinct outcomes compared to considering all factors simultaneously. Ultimately, the output from AHP directly correlates with the number of factors considered and their assigned weights. In essence, while focusing on specific categories offers unique insights, the categorization of factors presents a more refined yet incomplete perspective. A comprehensive approach, considering all factors, remains essential for robust land use suitability assessment for urban development. A source of uncertainties lies in the expert decision on the weightage of the factors involved in in the AHP technique [89]. Although, the values of pair-wise comparison matrix were set after careful examination of cause-effect relationship between the factors, it is subjective. Thus, the consistency of the weights is assessed and a consistency ratio of 0.09 is found which is reasonable [87, 90]. Another fact is to be noted that since this method uses several causative factors, uncertainty in each layer can add up to an erroneous result. The topographic information (slope and elevation) collected using the ASTER DEM (30 m) has a mean absolute error of 5.35 m which is slightly greater than SRTM (30 mn) DEM [91]. But SRTM has several missing data in the hilly regions, thus we chose the ASTER DEM. The error margin is quite acceptable as the study is performed at a district-level scale. In a city-level spatial scale the inaccuracy of 5 m would result in large scale uncertainties. Although, a high-resolution soil map and ground truth verification is needed for more accurate assessment of the soil erosion, our result aligns with the published data for the studied regions [92, 93]. The LULC map was prepared with high resolution (10 m) Sentinel-2 data with relatively high accuracy. The built-up area data, collected from the Global Human Settlement Layer [94], has highest accuracy (94.8%) in the south-east Asia [95]. Thus, the overall result of site suitability has lesser uncertainty from the built-up area layer. The effects of uncertainties become severe when the analysis is performed at city or municipal scale. In those scenarios, each factor must be verified with the data collected in the field. For the uncertainties involved in each layer for the present study, we argue that it has lesser effect on the final product as the study is concerned with district-level spatial scale.

This study aims to identify the potential of urbanization for the three selected Himalayan hill stations. These hill stations are existing from a time when sustainable urbanization was less of a priority [83]. During post-independence time, these hill stations grew around the already developed parts of the areas which assured economic sustainability for the locals. Now at the twenty-first century, with growing economy urbanization must be done in a scientific way so that it can sustain the tourism industry and local livelihood. This study highlights the potential regions where urbanization can sustain. Moreover, it also highlights the urbanization trend for the studied areas. The age-old problem of “agricultural land reduction for urbanization” in the developing countries comes out as one of the important considerations for urbanization.

6 Conclusions

We utilized GIS, Multi-Criteria Decision-Making Approach (MCDMA), and Analytical Hierarchy Process (AHP) to assess land use suitability for urban development across three districts. Our approach integrates ordered weight averaging and logic scoring preference methods, mapping nine key factors including Slope, Elevation, LULC, Distance from fault, Distance from the river, Distance from the road, Distance from Urban amenities, Soil erosion, and Distance from landslide density into five suitability classes.

Hilly terrains face notable soil erosion risks, a focal point in our study. By leveraging GIS tool like the ArcGIS, we aim to delineate erosion-prone areas and quantify erosion rates. Incorporating soil erosion data enhances suitability analysis accuracy, crucial for robust decision-making in development initiatives. Our AHP-based criteria assign weights to parameters influencing erosion. The resulting consistency ratio validates the weights' alignment with our assumptions, ensuring a reliable assessment. Qualitatively analyzing erosion trends, our study classifies areas into five sustainability categories. Despite certain factors taking precedence, we consider nine factors for enhanced accuracy. Slope, elevation, barren land, and rainfall significantly impact erosion rates, aligning with our understanding of higher rainfall contributing to increased erosion.

Comparing urban suitability results derived from GIS-MCDMA and AHP with the 2020 built-up index offers valuable insights. Shimla emerges as the most suitable district, while Darjeeling ranks lowest. While existing plans consider suitability, areas with moderate and high suitability coinciding with landslide and fault zones require attention. Although these zones register as suitable based on nine factors, their proximity to risk areas implies that caution must be taken. Examining suitability through these lenses highlights that it is not determined by a single factor, emphasizing the need for comprehensive assessments. Based on the site suitability results derived from this multi-criterion decision-making approach, several recommendations for urbanization in the studied regions can be provided.

Although the north-western part of Nainital shows moderate to high site suitability for urban development, this region should be avoided due to the presence of the Main Central Thrust nearby. The planar area in Nainital shows moderate to very high suitability for development, and this can be used for future development, although a lot of built-up has already been done in this area. Future development in Nainital can be focused in the south-eastern part as this region is devoid of faults, and site suitability analysis shows moderate to high suitability. Urbanization can also be focused on the hilly side of the region assuring proper building techniques are used. A balance between agricultural land acquisition and urbanization must be maintained for future food security.

The central part of Shimla has several amenities already; this region is free from faults as well and shows moderate to very high suitability from AHP derived results. Similarly, in the southern part of Shimla, site suitability yields moderate to high suitability and this region is free from faults. These two areas can be focused for future development. Shimla town shows moderate to low suitability due to the presence of several faults, although built-up is dominant in this region due to lower slope and British planning for city development. This area should be avoided for further development, keeping suitability in mind.

The southern part of Darjeeling yields moderate to very high suitability from AHP technique, this portion is also distant from the faults. Suitability in Darjeeling is sporadically present, but the southern part and central part can be focused for future development in this region. The chances for urban development in Darjeeling are high as this area has a lot of bare ground or range land which are unsuitable for agricultural use due to its low soil nutrition. Landslide preventive measures must be taken, and good drainage system must be established to ensure a protection from natural preventive factors.

Although a generalized recommendation is made in this study, concerned authorities such as policy makers and government organizations can use this study as a basis for further development of the areas. The techniques described in this study can be used for urban suitability analysis for other areas. Additionally, the insights from this study, particularly the factors influencing urbanization suitability in hilly areas, can be applied to promote sustainable urban development in other similar regions.