Introduction

Landslides covering spatial and temporal dimensions can cause excessive loss to human lives, infrastructure, and the economy (Aleotti and Chowdhury 1999; Glade and Crozier 2005; Nandi and Shakoor 2008; Del Soldato et al. 2019; Cieslik et al. 2019). These mass wasting processes are dependent on natural causes like geology (Wakatsuki et al. 2005; Zheng et al. 2021), geomorphology (Lee and Min 2001; Lan et al. 2004; Anbazhagan and Ramesh 2014), and hydrology (Sidle and Ochiai 2006), as well as triggering factors such as rainfall (Guzzetti et al. 1999; Saito et al. 2018), earthquake (Malamud et al. 2004), volcanic activity (Di Traglia et al. 2018), and anthropogenic activities (Jones et al. 2021). Vulnerability is a dynamic, inquisitive concept that describes the levels or extent of susceptibility to damage or harm, inadequacy, marginality, and frailty of social, physical, economic, environmental, and cultural systems (Adger 2006). Turner et al. (2003) described it as the probability that the elements at risk suffer damage because of sensitivity, exposure, and lack of resilience. Thus, the highly dynamic, complex, multi-dimensional, and hazard-dependent nature of vulnerability has always made its quantification a significant challenge for the scientific communities (Fuchs et al. 2007). Still, it is an intrinsic criterion for risk estimation, assessment, and reduction. The physical vulnerability is quantified on a scale of 0 (no damage) to 1 (total damage) (Quan Luna et al. 2011; Guillard-Gonçalves et al. 2016; Papathoma-köhle et al. 2017) as it gives an idea about the degree of loss to elements at risk like buildings and roads due to hazard of some magnitude in an area. In the context of landslide hazard, the magnitude of damage to the infrastructures and buildings is a function of characteristics and location of the buildings, type of landslides and their frequency of occurrence, their intensity (quantified as landslide velocity), and magnitude (Dai et al. 2002; Thennavan et al. 2016). Different challenges exist in physical vulnerability assessment due to types of landslide movements, limited availability of the damage data, and assignment of scores as per expert’s opinion to calculate the resistance of the buildings, which restricts and limits the performance of the vulnerability model (Uzielli et al. 2015; Guillard-Gonçalves et al. 2016). Despite the challenges, the quantitative vulnerability assessment approaches are gaining momentum and attention amongst the researchers. It can calculate and assess the losses and damage occurred due to the landslide hazard though with limitations (Peng et al. 2015).

The Himalayan terrain is highly prone to landslides constituting about 70% of the world’s tragic and fatal landslides (Dikshit et al. 2020). Also, active geotectonics concomitant with slope failures have resulted in a massive increase in the mass wasting activities in the region (Paul and Mahajan 1999; Sah et al. 2018). The region encounters abundant rainfall during the monsoons and is one of the key triggering factors for landslide initiation (Dikshit et al. 2020). Nainital township located in the lower Himalaya of the Kumaon region of Uttarakhand has been affected by major calamitous slope failures in the past in areas like Sher-ka Danda, Kailakhan, and Balia Nala and are still prevalent in the region as studied and documented by Hukku et al. (1974), Bhandari (1987), Valdiya (1988), Sharma (2006), Pande and Pande (2008), Kotlia et al. (2009), Gupta et al. (2016), Kumar et al.(2017), Sah et al. (2018), Yhokha et al. (2018), Kuri et al. (2018), Vaid et al. (2019), Khanduri et al. (2019), and Deyet al. (2020).

Balia Nala is one of the active and significantly high landslide hazard zones of Nainital town (Kuri et al. 2018; Vaid et al. 2019). The landslides on the right side of Balia Nala were first recorded back on August 17, 1898, which was triggered after 8 days of continuous rainfall measuring 102 cm (Middlemiss 1898). In September 2014, the Rais Hotel colony located on the right side of Balia Nala was affected by a cataclysmic landslide following the intermittent rainfall. Also, before the landslides, cracks were noticed in the buildings and the ground, indicating continuous slow movement of the slopes (Gupta et al. 2016). The fury of Balia Nala landslide due to the excessive rainfall in September 2018 was to the extent that it wiped away 3–4 houses, destroyed the roads in the nearby area, and in light of these events, 17 households were given notices to vacate and shift to a new location (Khanduri et al. 2019). Considering these devastating landslide events, the region’s risk to life and infrastructure has escalated. Moreover, the haphazard constructions in the area, specifically over the landslides’ crown, have further exacerbated the physical vulnerability to the landslides.

Hence, the objective of the present study is to characterize the Balia Nala landslide and assess the site-specific physical vulnerability of the buildings lying on its slope. The significant portion of this research focuses on quantifying the site-specific physical vulnerability of 63 households in the Balia Nala area. This quantification would help the concerned authorities (like State Disaster Management Authority, Uttarakhand and Irrigation Department, Nainital) to undertake Disaster Risk Reduction (DRR) activities focused on reducing the intensity of landslide by implementing sustainable measures in the study area. Also, undertake efficient remedial measures for the highly vulnerable households and prioritize their relocation. The study provides a path to revise/frame the guidelines on building constructions and design specifically concerning landslides in the study area. The study also highlights the need to constantly monitor the growing illegal construction to reduce disaster risk and promote sustainable development. The Geographic Information System (GIS) is used to map and calculate the areal extent of the Balia Nala landslide. The site-specific physical vulnerability of the buildings was estimated using an indicator-based approach adopted by Singh et al. (2019a, b).

Previous studies

Papathoma-köhle et al. (2017) reviewed and discussed three varied approaches, viz., damage or vulnerability matrix, vulnerability curves, and an indicator-based approach for assessing the physical vulnerability of buildings specifically for debris flows. The damage or vulnerability matrix approach was first proposed by Leone et al. (1995) and resulted in the vulnerability value from 0 to 1 using damage states based on the process intensity. A new methodology for damage probability matrix by Erener and Düzgün (2011) also incorporated remote sensing and GIS for vulnerability assessment. Damage matrices on the basis of past events (Bründl et al. 2009) were also proposed. Jakob et al. (2012) developed a damage probability matrix by calculating the intensity index of the landslide and correlated it with the damage of the buildings using the HAZUS software. A matrix for Reinforced Concrete (RC) and non-Reinforced Concrete (non-RC) buildings based on impact pressure was proposed by Kang and Kim (2016). To study the effects of diverse variables on the bearing capacity of footings on soil slopes, Satvati et al. (2020) used a laboratory approach using three-dimensional geosynthetics called expandable braided sleeve and geogrid. Damage/vulnerability matrices clarify the relationship between the process and the consequence; hence is easily understandable by the common people. However, its subjective nature of assessment poses one of the significant limitations (Papathoma-köhle et al. 2017).

The vulnerability curve is a quantitative method to assess physical vulnerability and is mainly focused on quantifying damage loss of buildings into a monetary value. Fuchs et al. (2007) were the first to underscore the importance of this method and derived the vulnerability curves using the intensity of debris flow landslide and damage ratio of the buildings in the Alps, Austria. Vulnerability curves were also integrated into landslide vulnerability studies by Mavrouli et al. (2014) and Uzielli et al. (2015) for prestressed-concrete buildings. Finally, Chen et al. (2020) proposed a methodology to assess the physical vulnerability of the masonry buildings exposed to the effects of slow-moving landslides by using this method.

An indicator-based approach for assessing physical vulnerability has recently gained importance. Its process includes selecting the physical vulnerability indicators, identifying the variables, assigning weights followed by the final assessment. Papathoma-Köhle et al. (2007) were the first to attempt this approach to assess physical vulnerability for the hilly region hazards by assigning weights to the indicators based on expert’s judgement. Li et al. (2010) used an indicator-based approach with susceptibility of elements at risk and landslide intensity as inputs for assessing the physical vulnerability of the infrastructure. Kappes et al. (2012) further improved this method without considering the landslide’s intensity. Physical vulnerability and the probable losses of the buildings affected by the shallow landslides were assessed by Silva and Pereira (2014) at a municipality (regional level) using this approach by combining the ‘resistance’ and ‘intensity of process’ parameters. Du (2012) and Du et al. (2016) also used an indicator-based approach to assess the vulnerability of buildings due to landslides. Following the methodology of Du (2012), vulnerability assessment was conducted by Guo et al. (2020) using the same approach. Thennavan et al. (2016) carried out a vulnerability assessment of buildings in the high hazard landslide-prone area in the Nilgiris district using thematic layers of indicators related to buildings and calculated the vulnerability scores using empirical formulas scripting in a GIS environment. Ettinger et al. (2016) also used an indicator-based approach and further analysed the relationship between the vulnerability indicators and damage data using logistic regression. A study by Singh et al. (2017) reviewed several studies on qualitative and quantitative vulnerability approaches. It proposed a new methodology that considered structural details and typology of the buildings as indicators to assess the physical vulnerability. A study by Singh et al. (2018) adopted a modified semi-quantitative approach in which landslide intensity was quantified using velocity, volume, and building’s resistance of around 71 buildings in the Joshimath block of Uttarakhand, India. Physical vulnerability assessment of the buildings exposed to the rockfall hazard was conducted by Singh et al. (2019a, b) using empirical formula applied in Li et al. (2010). An indicator-based approach using 15 different indicators was used to assess the vulnerability to buildings at a site-specific scale in the ward of Chamoli district, India (Singh et al. 2019a, b). Bera et al. (2020) adopted indicator-based approach and used the Fuzzy-AHP approach to calculate the resistance scores of the buildings by considering the geomorphological conditions and applied it to 332 buildings in the high and active landslide hazard zone in the Kalimpong area.

The present study incorporates the methodology adopted by Singh et al. (2019a, b) for assessing the site-specific vulnerability of the buildings located in Balia Nala, Nainital, Uttarakhand. The indicator-based approach is preferred in this study over the damage/vulnerability matrix methodology due to its highly subjective nature of assessments. Moreover, one of the major limitations of vulnerability curves is that it is not applicable for the regions with different building types, and for obtaining a result, there is a need for a greater number of affected buildings (Papathoma-Köhle et al. 2017). Not many damaged/affected buildings were there in our study area due to the landslide. Hence, indicator-based approach was finalized, and the methodology adopted by Singh et al. (2019a, b) has been applied due to the similarity in the study area and landslide type (debris flow).

Study area

Nainital town, a famous hill station, is one of the flag bearers of the tourism industry of Uttarakhand. Ayarpatta, Sher-ka-Danda, Naina Peak, and Deopatha are the prominent hills surrounding the Naini Lake (Khanduri et al. 2019). About half of the township area is covered with debris accumulated due to the frequent mass wasting activities in the region (Valdiya 1988). The study area, Balia Nala, is located at 29°22′26.13″N latitude and 79°27′55.53″E longitude at an elevation of 1872 m in the Hari Nagar ward of Nainital town with the ward’s population of 2825 (Census of India, 2011) (Fig. 1). Mass wasting activities are common and frequent along both sides of the Nala, which flows in the S-E direction and is a natural outlet of Nainital Lake (Kuri et al. 2018). The continuous and rampant toe erosion of the Balia Nala landslide and burgeoning haphazard construction activities have added to the causes of slope instability in the area (Khanduri et al. 2019; Gupta et al. 2016; Rautela et al. 2014). Field photographs of the exposed and affected buildings to the Balia Nala landslide are shown in Fig. 2.

Fig. 1
figure 1

Study area—Balia Nala, Nainital, India, area with the landslide demarcated (source: Google Earth satellite imagery)

Full size image
Fig. 2
figure 2

Field photographs show Balia Nala landslide and its impact: a view of landslide, b and c exposed and affected households

Full size image

The township experiences salubrious summers and extremely cold winters. The temperature during the summers varies in the range of 10 to 28 °C and can even touch below 0 °C. In the monsoons (June to September), the area experiences moderate to heavy rainfall, which is one of the most important triggering factors for landslide reactivation in the area. As per the Irrigation Department of Nainital, the year 2010 recorded the highest average monsoon rainfall of about 1027 mm between the years 2005–2019. An increase in landslide activities was witnessed in the region following this heavy rainfall.

Geological setting

Nainital town majorly constitutes rocks like slate, shale, and limestone, which belong to the various formations viz. Krol, Blaini, and Tal of the Lesser Himalaya. The Krol group of rocks dominate in the Nainital township and mainly constitute grey and blue dolomitic limestone, argillaceous limestone, calcareous slate, siltstone, greywacke etc. The eastern and north-eastern parts of the Nainital Lake comprises mainly of the Blaini formation in which conglomerates associated with dolomite limestone, quartzitic, and purple slate dominate. In contrast, the Tal formation is majorly found near the south-western side of the lake and represented by fine-grained sandstone and siltstone, pyritic shale, and slate (Gupta et al. 2016). The area witnesses many faults, thrusts, and folds which transect the region of which the Nainital Lake fault trending in the NW–SE direction cuts the Main Boundary Thrust (MBT) and extends to the Balia Nala and is responsible for the weak geological setup (Valdiya 1988). It has resulted in weathering, further increasing the fragility of the dominantly found rocks in the area like slate, shale, and limestone leading to landslides.

Methodology and data collection

Quantitative vulnerability assessment for the hazards is recommended and preferred over qualitative assessments. It can evaluate or measure the losses, eliminating the communication gap between technical or political groups for better decision-making (Uzielli et al. 2008). The methodology developed by Singh et al. (2019a, b) is applied additionally with the aid of GIS to a different site, i.e. Balia Nala, a debris flow type landslide located in the foothills of the Kumaon outer Himalaya. The following steps have been followed to quantitatively assess the physical vulnerability of the buildings in the close periphery around the crown of the Balia Nala landslide (Fig. 3).

Fig. 3
figure 3

Methodology flowchart

Full size image

Spatial mapping of households and selection of building vulnerability indicators

A thorough field survey was carried out with the concerned authorities of the Irrigation Department, Nainital. The details of the hazard occurrences were gathered from literature surveys, newspaper articles, and most importantly, after the interaction with the local communities. Google Earth imageries of the years 2010, 2015, and 2019 were digitized to understand the areal extent and geometry of the Balia Nala landslide. Also, around 595 households visible in the Google Earth imagery of the year 2019 were mapped, which was followed by the creation of two buffer zones of 0–200 m and 200–400 m over the vulnerable slope section of the landslide and the digitized households were superimposed over the buffer zone. A total of 63 samples of households were randomly selected within these two buffer zones, and this pre-field map was taken as the footprint map for field data collection. Further, the building vulnerability indicators such as roof design and material, foundation material, type of construction (fully/partially engineered, non-engineered, and marginally engineered), and the height of the building, on which the resistance of a building depends, were taken into consideration. For the present study, most of the indicators are selected from the study by Singh et al. (2019a, b) due to the similarity of the terrain conditions of our study region in combination with the Indian Standard codes, viz., Indian Standard Code 14,804 on ‘Siting, Design and Selection of Materials for Residential Buildings in Hilly Areas’; National Building Code of India, 2005 (Group 1, ANNEX G, Clause 12.24) and National Building Code of India, 2016 (Group 1, ANNEX F, Clause 17) on ‘Special Requirements for development Planning in hilly areas’, literature surveys and based on the availability of data. Thus, 15 building indicators were selected, and a structured questionnaire was prepared based on those indicators as clearly tabulated in Table 1. The ground truth information was then collected with the help of the questionnaire by interviewing the inhabitants of the 63 selected sample households and the information gathered was compiled together to assess the vulnerability of the buildings.

Table 1 Questionnaire for physical vulnerability assessment (source: Papathoma-Köhle et al. 2007; IS 14458 (Part 2) 1997; BIS 2016; IS 14804 2000; Singh et al. 2019a, b; Joshi et al. 2019)
Full size table

Physical vulnerability assessment of the buildings

Physical vulnerability is considered in this study as a function of landslide intensity and building resistance. After the indicators were selected, raw scores were assigned based on interviews with locals, Indian Codal guidelines for hilly terrain (IS 14458 (Part 2) 1997; BIS 2016; IS 14804 2000), and a thorough literature review (Papathoma-Köhle et al. 2007; Singh et al. 2019a, b; Joshi et al. 2019). The raw score of 1 represents the ‘most vulnerable’ category, and the increasing raw scores signify the decrease in building vulnerability. These raw scores were further standardized in the range 0 to 1 by applying the formula proposed by Voogd (1983) as shown in Eq. (1).

$$Standardized\;Score=\frac{Raw\;Score}{Maximum\;Raw\;Score}$$
(1)

These standardized resistance scores of the 15 indicators were used as inputs in the resistance formula (Eq. 2) applied by Li et al. (2010) for calculating the resistance of the buildings, which plays a crucial role in vulnerability assessment and can be defined as the capacity of the structure to withstand the hazard of a given intensity. Further, the estimated resistance values were classified into three categories for a landslide-prone region in hilly terrains with its sliding material as debris by Singh et al. (2019a, b) and is presented in Table 2.

$$R={({\prod }_{i=1}^{{n}_{s}}{\xi }_{i})}^{\frac{1}{n}_{s}}$$
(2)

where R indicates the resistance capability of the building, ξi is the ith standardized resistance score of the building for a specific indicator, and ns indicates the total number of selected indicators.

Table 2 Resistance classes (source: Singh et al. 2019a, b)
Full size table

Another important factor considered for assessing the vulnerability and the magnitude of loss is the landslide’s intensity, which depends on the landslide’s moving mechanism and its type (Solari et al. 2020). Hungr (1997) defined it as the landslide’s level of destructiveness, which can be measured using a variety of parameters such as velocity, volume, deposition height, debris depth, deformation, pressure, momentum, kinetic energy (per unit area), mass thickness, run-out distance, peak discharge, and the area affected by the landslides (Ojeda-Moncayo et al. 2004; Singh et al. 2018). In the present study, two parameters, viz., the volume of the landslide (V) and its velocity (U) used by Cardinali et al. (2002), are considered for quantifying the landslide intensity (Is) as shown in Eq. 3.

$${I}_{s}=f(U,V)$$
(3)

Volume is an essential geometrical parameter that controls the run-out distance, the affected area, and the potential damages incurred on elements at risk (Jaboyedoff et al. 2020). It can be estimated using the average depth of landslide and surface area, but empirical formulae defining the relationship between landslide volume and area are generally adopted (Guzzetti et al. 2009). Thus, the area of landslide using Google Earth imagery shown in Fig. 1 was calculated in ArcMap software, and the volume was evaluated using the empirical formula proposed by Guzzetti et al. (2009) for debris flow as shown in Eq. 4.

$$V=0.074*{A}_{L}^{1.450}$$
(4)

Velocity is another crucial parameter that governs the flow momentum and the kinetic energy based on which the intensity of landslide can be determined. The methods developed for determining the velocity range from statistical-empirical relationships to three-dimensional run-out models like RAMMS, DAN3D, MassMov2D, and NGI (Mcdougall 2014). In the study, the already established empirical relationship by Rickenmann (1999) as shown in Eq. 5 has been used to calculate the debris flow velocity as our study also focuses on debris flow landslides.

Once the volume and velocity are evaluated, the intensity is estimated using the intensity matrix proposed by Singh et al. (2018).

$$U=2.1{Q}_{p}^{0.33}{S}^{0.33}$$
(5)

where ‘U’ is the landslide velocity, ‘S’ represents the local slope, and ‘Qp’ represents the peak discharge of boulder material and is given by Qp = 0.1V0.83 where V represents the volume.

Once the building resistance and the landslide intensity are quantified and calculated, the formula proposed by Li et al. (2010) as shown in Eq. 6 is further used for the calculation of the physical vulnerability (PV) of the sample buildings.

$$PV=f\left(I_s,R\right)=\left\{\begin{array}{lc}2\frac{I_s^2}{R^2}&\frac{I_s}R\leq0.5\\1-\frac{{2\left(R-I_s\right)}^2}{R^2}&0.5<\frac{I_s}R\leq1.0\\1.0&\frac{I_s}R>1\end{array}\right.$$
(6)

The proximity of these households to the landslide is then taken into consideration by using the 20-m contour interval derived using ALOS PALSAR digital elevation model (DEM) and the percentage decrease in the physical vulnerability of the respective households due to proximity from the crown of the Balia Nala landslide (Table 3), by using the empirical relationship proposed by Singh et al. (2019a, b) as shown in Eq. (7).

$$SSV=PV-A$$
(7)

where ‘SSV’ represents the site-specific vulnerability, ‘PV’ represents the physical vulnerability, and ‘A’ represents the percentage reduction in physical vulnerability with distance. The obtained site-specific physical vulnerability was further classified into three classes (class I, class II, and class III) as per the classification proposed by Singh et al. (2019a, b) and is shown in Table 4.

Table 3 Percentage reduction in physical vulnerability of buildings due to proximity from Balia Nala landslide (source: Singh et al. 2019a, b)
Full size table
Table 4 Site-specific vulnerability classes (source: Singh et al. 2019a, b)
Full size table

Results

Characteristics of Balia Nala landslide

Balia Nala landslide is a debris flow type landslide with two seepages present on the face of the slide. It is retrogressive and widening in nature with its progression towards the northern direction and has a maximum length of about 447 m and a maximum width of 282 m. The slide’s crown and toe are located at an elevation of EL ± 1820 m and EL ± 1560 m respectively and face towards the north-eastern direction with a minimum slope of 3°, a maximum slope of 68°, and an average slope of 32°. The landslide has been prominent in the region since the twentieth century but has shown major reactivation activity after 2010 (Gupta et al. 2016). Mapping of the landslide region using the Google Earth images of the years 2010, 2015, and 2019 showed an increase of landslide area from 70,698 to 72,708 m2 (a rise of 2010 m2) within 5 years (2015 to 2019) (Fig. 4). Thus, triggering factors like rainfall, Balia Nala fault, which passes through the toe of the slide, proximity to drainages, presence of seepages, higher natural slope conditions, haphazard and unplanned construction in the landslide area have resulted in the increasing instability of the Balia Nala slope.

Fig. 4
figure 4

Progression and areal extent of Balia Nala landslide from 2010 to 2019

Full size image

Estimation of resistance of each building

Resistance plays a vital role in vulnerability assessment and is the capacity of the structure to withstand the hazard of a given intensity. Based on the extensive field survey of the 63 buildings with the help of the structured questionnaire, it was found that the majority of the surveyed buildings in the region were constructed between the years (R1) 1901–2000 and are low rise buildings with two floors (R2) and have pitched roofs in the tin (R3) with no surrounding wall (R5). The recent constructions are made up of reinforced concrete (40%), whereas the older constructions are made up of stone masonry and unreinforced brick masonry (R4), which contribute to around 54%. The distance between 2 buildings (R7) and the distance between windows and doors (R6) is less than 1 m in 84% and 59% of the surveyed buildings, which indicates the haphazard construction in the study area. Openings on the side facing the slope (R8) could be found in 57% of the buildings and 92% having no forest or trees cover (R9) in its vicinity, which infers burgeoning urbanization in the region. In 18% of the cases, cracks were noticed in the old buildings and the ground settlement (R10), and only 35% of the surveyed buildings were retrofitted timely (R12); 73% of buildings showed no warnings of landslides (R13). Regular horizontal configuration (R14) contributes to the lesser vulnerability of the buildings; from the survey, 75% of the buildings showed this type of configuration. Of the surveyed households, 54% were engineered and belonged to the category of either newly constructed buildings or the buildings constructed during the British Raj (R11). Also, a maximum percentage of buildings were found at a distance of 51–100 m from the landslide slope (R15). The percentage of surveyed households of vulnerability indicators is given in Table 5. The raw scores assigned to the different physical indicators and their respective standardized resistance scores calculated using Eq. 1 is shown in Table 6. Likewise, Table 7 represents the values of resistance factors and the overall resistance (R) of the surveyed buildings calculated using Eq. 2. The following result concluded from the above calculations that out of the 63 buildings, about 89% of them belonged to the category of ‘moderate resistance class’; 11% fell under the category of ‘low resistance’; and none of the surveyed buildings fell in the ‘high resistance’ category. The buildings falling under the various resistance classes are shown in Fig. 5.

Table 5 Percentage of surveyed households of vulnerability indicators
Full size table
Table 6 Raw scores and resistance factor
Full size table
Table 7 Resistance (R) of each building in the study area
Full size table
Fig. 5
figure 5

Buildings falling under different resistance classes

Full size image

Estimation of the intensity of landslide

Landslide intensity is the other important parameter taken into consideration for vulnerability quantification which is further taken as a function of landslide’s volume and velocity. The estimated area of the Google Earth imagery dated 31/05/2019 is calculated to be 72,708 m2 with the aid of GIS; this value is further used in Eq. 4 for volume estimation of the Balia Nala landslide, which is obtained as 828,945 m3 and comes under the ‘high’ category (Sarkar et al. 2015). The velocities at the minimum, maximum, and average slopes were computed using Eq. 5 after classifying the slope of the landslide using ALOS PALSAR DEM (12.5 m) and are shown in Fig. 6. The values of landslide velocity estimated for minimum (3°), maximum (68°), and average (32°) slopes are 15.5 m/s, 43.4 m/s, and 33.8 m/s, respectively. Hence, the velocity of landslide lies in the range of 15.5 to 43.4 m/s, which belongs to the ‘extremely rapid’ category as proposed by Cruden and Varnes (1996). Landslide intensity is further categorized as ‘very high’ on a scale of 0 to 1; its value is 1 by using the landslide intensity matrix proposed by Sarkar et al. (2015).

Fig. 6
figure 6

Slope map of Balia Nala landslide

Full size image

Physical vulnerability of individual building and its quantification at site-specific scale

Physical vulnerability is calculated by putting together the values of resistance of individual buildings and landslide intensity value of 1 in Eq. 6, which results in a uniform physical vulnerability of value ‘1’ for all the surveyed buildings despite varying resistances of the buildings as the estimated landslide intensity is very high. To have a better understanding of the vulnerability quantification, the site-specific vulnerability of the individual buildings is calculated using Eq. 7, and the results are tabulated in Table 8 with a categorization of the buildings into three classes; class I, class II, and class III as proposed by Singh et al. (2019a, b). The following observations were made; out of the 63 surveyed buildings, 44% of them fall under class III (0.9 < SSV ≤ 1; very severe to complete), which infers that in the case of any severe future landslide, the buildings would get damaged with no scope of retrofitting. About 56% of them fall under class II (0.5 < SSV ≤ 0.9; moderate to severe), which signifies that the functionality of the buildings would be compromised and hence very high costs would be required for retrofitting when struck by a severe landslide, whereas none of the buildings surveyed falls under class I (SSV ≤ 0.5), which in the future would be subjected to low damage (Fig. 7). It shows that the buildings are highly vulnerable based on the proximity to the Balia Nala landslide. Few field photographs of the surveyed households are shown in Fig. 8.

Table 8 Calculated site-specific vulnerability of buildings
Full size table
Fig. 7
figure 7

Buildings falling under different SSV classes

Full size image
Fig. 8
figure 8

a Field photograph of building ID: H24 (R: 0.48, SSV class: III); b field photograph of building ID: H37 (R: 0.63, SSV class: II); c, d field photograph of building ID: H43 (R: 0.64, SSV class: II)

Full size image

Discussion

This study aims to assess the physical vulnerability of the 63 sample households using an indicator-based approach and GIS aid. The evaluation of physical vulnerability is an outcome of the complete process like selecting 15 building indicators, field survey, calculation and standardization of raw scores, calculation of resistance and intensity of the households. The final resistance calculated as a function of standardized resistance score shows that the majority of the buildings (89%) provide ‘moderate resistance’ (0.5 < R ≤ 0.9) and 11% are ‘low resistant’ (R ≤ 0.5). The least resistance evaluated was 0.44 (building ID H12) and the highest was 0.88 (building ID H57). As per the field observation of the study area, the majority of the households are illegally built, and they belong to people with a poor socio-economic background. Most of the households constructed in the encroached area were made of wood, tin, and bricks and were non-engineered. Owing to this fact, after considering all the vulnerability indicators, none of the households provides high resistance capability (0.9 < R ≤ 1) and falls in the ‘moderate’ or ‘low’ resistance classes.

The other most crucial parameter is the calculation of landslide intensity, which is calculated as a function of velocity and volume in this study. The velocity of landslide obtained ranges between 15.5 and 33.8 m/s (falls under ‘extremely rapid’ category), and the volume amounts to 8,28,945 m3 (falls under ‘high’ category). Using a landslide intensity matrix, the landslide intensity obtained is ‘1’ and falls under the ‘very high’ category. Damaged roads, demolished and abandoned houses, and continuous subsiding of grounds in the study area are all indicators of this result. It is a very alarming issue and urgently requires the implementation of suitable and sustainable remedial measures in this hill slope. The resistance and landslide intensity obtained are further used to evaluate physical vulnerability. This study got a constant physical vulnerability value of ‘1’ for all the surveyed households, owing to very high landslide intensity. The value is indicative of the fact that in a probable landslide scenario, the buildings would get totally damaged, which practically is not realistic, as proximity to the landslide also plays a major role. Hence, the site-specific physical vulnerability was calculated, and it was found that 56% of the surveyed households fall under class II (0.5 < SSV ≤ 0.9; moderate to severe damage) and 44% fall under class III (0.9 < SSV ≤ 1; very severe to complete damage). While in the first case, functionality of the buildings would get compromised and would demand retrofitting at a high cost, whereas in the other, complete damage would occur, offering no scope for retrofitting. This outcome is an issue of great concern as none of the buildings falls under class I (SSV ≤ 0.5; low damage). Interestingly, it is noticed that some of the moderately resistant buildings (building IDs H4–7, H9, H11, H13–17, H19–23, H25–28) fall under SSV class III, after considering the ‘proximity’ to the landslide zone. Overall, the results show that the buildings in the study area are highly vulnerable. Any triggering factor like a severe earthquake or heavy rainfall scenario would pose a risk to the lives and property. Hence, the concerned authorities need to pre-plan their relocation and take all the possible remedial measures.

The haphazard and illegal constructions in the study area also need serious attention. Specific guidelines on building design and construction in earthquake-prone regions are available in India; however, this is not the case with respect to the landslide-prone region. This research opens a new avenue to frame or revise proper guidelines on building constructions and design, specifically concerning landslides in the hilly terrain. It also highlights the need to take disaster risk reduction measures to reduce the landslide intensity and increase the resistance of buildings. Also, it recommends constantly monitoring the growing illegal constructions in the area with the objective to reduce disaster risk and promote sustainable development.

Conclusion

Balia Nala landslide has become an issue of great concern these days amongst the residents and researchers due to its frequent reactivation and burgeoning haphazard construction over its slope. These mandates characterize the Balia Nala landslide and quantify the physical vulnerability of the buildings. Thus, the indicator-based approach used by Singh et al. (2019a, b) with the aid of GIS has been adopted in the study area to assess the site-specific physical vulnerability of the buildings. The vulnerability of the 63 household samples was computed using 15 different indicators. Many other empirical relationships were employed to estimate the intensity of the landslide, resistance of the buildings, physical vulnerability (PV), and site-specific vulnerability (SSV). The landslide intensity obtained in the study is very high (1). About 89% of buildings fall under the category of moderate resistance (0.5 < R ≤ 0.9), whereas 11% fall under the low resistance (R ≤ 0.5) category. The physical vulnerability of all the buildings obtained is the same, i.e. 1. Since this is not realistic, proximity to the landslide is taken as a criterion. About 56% of the buildings fall in class II (0.5 < SSV ≤ 0.9), which is moderate to severe destruction and 44% of the buildings fall under class III (0.9 < SSV ≤ 1) that is prone to extreme destruction. Hence, this quantification of physical vulnerability of the buildings and the characterization of the Balia Nala landslide will entail for improved understanding of the landslide and can help in suggesting necessary remedial measures for the buildings to support mitigation and decision-making and has the potential to customize the guidelines for building constructions for landslide in the region. The study can benefit concerned authorities like State Disaster Management Authority (SDMA), Uttarakhand; Irrigation Department, Nainital; and town planners for taking necessary and efficient Disaster Risk Reduction (DRR) measures which can contribute in reducing the landslide intensity, increasing the resistance of the buildings, and prioritizing relocation of the vulnerable households in case of any major landslide disaster in the future.