Introduction

Humans are facing severe natural disasters such as floods, landslides, earthquakes, cyclones, etc., since the beginning of civilization (Nasiri et al. 2016; Sarkar and Mondal 2020). Over time, the potency and magnitude of all-natural hazards and disasters have increased manifold across the globe, attributing to a host of factors (Erena and Worku 2018). Physical factors, coupled with human activities, have accelerated its pace posing large-scale devastation to the environment and ecosystem (Ali et al. 2019; Sarkar and Mondal 2020). Amongst the natural hazards, floods are the most common hydrometeorological and catastrophic event evoking immense risk to social, economic, and physical aspects (Indrayani et al. 2018; Panhalkar and Jarag 2017).

Criss-crossed by multiple rivers, India, aptly called the land of rivers, also experiences recurrent flood annually. Though high-intensity floods are regarded as a hazard but at the same time, these flood also helps in replenishing the river ecology and soil fertility, thus can be well-said that flood acts both as a boon and bane for the economy (Das et al. 2019; Gourav et al. 2020). According to Srivastava (2011), it is evident that the climatic and topographic features make various regions of India susceptible to several natural hazards. It is stated that 54% of land areas are highly prone to earthquakes, whereas about 40 million hectares of land areas are highly vulnerable to flood. Based on Andrewet al. (2018), India is ranked as the high-risk class for humanitarian crises and disasters with a risk management index of 5.7, while in terms of flood hazard index, the value was about 8.5, respectively. However, CWC (2018) highlighted that the total land area inundated by flood in the country is around 5.74 million hectares, and the entire population affected was recorded as 18.64 million.

Eventually, the Indian sub-continent also possesses a longer coastline of 7517 km, witnessing several hazards, becoming one of the most vulnerable and potent zones (Vignesh and Madha Suresh 2018). It is noteworthy that coastal cities being regarded as dynamic interface zones experience significant coastal threats owing to rapid urban, industrial, and population growth aggravating the flood situation (Dhiman et al. 2019; Di Risio et al. 2017; Nayak and Bhaskaran 2014; Roy and Blaschke 2015). In this context, Kanyakumari is also worth mentioning as the region situated adjacent to the coastal belt has been encountering flood havoc since millennia (Vignesh and Madha Suresh 2018).

The term “vulnerability” indicates the measurement of potential risk with a combination of socio-economic ability to tackle the worst situation resulting from the disastrous event. The target of flood vulnerability assessment is to determine the probability and intensity of the flood hazard occurring for the protracted time (Al Baky et al. 2019; Hoque et al. 2019; Mundhe 2019). According to the (Cafiero and Vakis 2006) the vulnerability assessment aimed to determine the probability over the extended periods of decades to assist the activities of risk management. Though natural calamities cannot be stopped yet, its risk and effects can be reduced if proper planning and strategies are considered. Therefore, a vulnerability assessment is a challenging approach to map the potential vulnerable areas and initiate constructive efforts for managing sustainably (Danumah et al. 2016).

Amongst all approaches and techniques used to examine flood hazard and its aftermath, a comprehensible method propounded by Saaty 1980 developed within the framework of MCDM is Analytical Hierarchy Process (AHP). It is a structured technique and measurement involving a complex set of multiple decision-making criteria. These criteria are arranged in a hierarchical order based on decision maker’s preference that helps in determining the relative importance of the parameters applied in a precise and systematic manner (Al-Harbi 2001; Boulomytis et al. 2019; Chen et al. 2011; Gupta 2017; Lawal et al. 2012; Nandy et al. 2012; Saaty 1980; Seejata et al. 2018; Velasquez and Hester 2013). The present paper attempts to assess the flood vulnerable areas of the Kanyakumari district by the integration of the MCDM based AHP model and the geospatial techniques. Located along the coastline of the Indian Peninsula, Kanyakumari district also encounters severe annual floods posing more significant threats.

Study area

The study area is confined to the southernmost district of Tamil Nadu, i.e., Kanyakumari (Fig. 1), comprising a total geographical area of 1684 km2. Its longitude and latitude extend from 77° 6′ 1″ to 77° 35′ 26″ East and from 8° 4′ 36″ to 8° 34′ 43″ North. The district is bounded by the Tirunelveli district on the North and East, while the Gulf of Mannar forms its South-eastern border. Moreover, the Indian Ocean and the Arabian Sea surrounds the district on South and South-west direction and Kerala state on West and North-west. The chief perennial river flowing through the central part of the district is Thamirabarani, along with Karaiyar and Paralayar as its tributaries. Being situated at the southern tip of peninsular India, the region receives rainfall from both South-west and North-east monsoon during summer and winter seasons. It also undergoes pleasant tropical climate all-round the year with summer temperature rising from 22 to 35 °C and winter temperature from 21 to 34 °C, respectively.

Fig. 1
figure 1

Location map of Kanyakumari district, Tamil Nadu

Full size image

Materials and methods

Materials

The evaluation and mapping of the FVZ of Kanyakumari district, various datasets comprising sequential steps of methods have been followed. The present research has been carried out to framework the AHP model coupled with RS and GIS in the geospatial environment using ArcGIS 10.3. The formulation of the model requires the ten-flood influencing physical variables viz. rainfall, slope, drainage density, LULC, DEM, soil, geology, geomorphology, surface runoff, and TWI (Ali et al. 2019). The datasets used in this research are described accordingly in Table 1.

Table 1 Datasets used in the study
Full size table

Methods

FVZ preparation in this study involves the series of steps: identifying and defining the complex problem, generate the AHP model-based hierarchical structure for the selected criteria, perform the pair-wise comparison matrix analysis for the chosen influencing variables (binary comparison), assignment of priorities values and the determination of relative weights for each variable, measurement of consistency value for the evaluations and decisions, synthesize the conclusions on priority variables to identify the FVZ (Dandapat and Panda 2017; Danumah et al. 2016; Roy and Blaschke 2015; Sulaiman et al. 2015). The considered flood influencing factors were compared with the application of MCDM to compute the relative weights (or priorities) among the factors to attain a significant outcome are discussed below in detail. MCDM-based AHP model was used to quantify the weightage of ten selected flood initiation factors, responsible for the determination of the potential of individual elements in inducing the flood hazard (Chakraborty and Mukhopadhyay 2019; Lawal et al. 2012). The above influencing factors were further reclassified into the sub-classes according their relative ranks for the computation of their relative weights (Danumah et al. 2016; de Brito and Evers 2016).

These relative ranks were allocated between 1 and 5, according to the Saaty 1980 preference descriptor (Table 3) for the construction of the thematic layers of each influencing factor. The maximum weightage value indicates very high vulnerability, followed by high, moderate, low, and very low, respectively (Table 5). All the obtained thematic layers have been combined and interpreted in the geospatial environment through the Weighted Linear Combination (WLC) approach using the geospatial platform to achieve the FVZ as the outcome. The flowchart of the proposed methodology is shown in Fig. 2.

Fig. 2
figure 2

Methodology flow chart

Full size image

Flood-inducing factors

Rainfall

It forms the most striking factor since the Kanyakumari district receives enough rainfall from both north-east and south-west monsoons. For the preparation of the rainfall distribution map, the rainfall data of all the rain gauge stations have been calculated through the Inverse Distance Weighted (IDW) interpolation tool sin GIS platform. The district blocks witnessing high rainfall ranging from 75 to 190 mm are Thovalai, Thuckalay, Agasteeswaram, Kurunthancode, Munchirai, and Melpuram, constituting as high flood vulnerable regions. On contradictorily, lower rainfall of 0–75 mm is experienced by southern and eastern parts of the Thiruvattar, northern part of the Thovalai, Killiyoor, and north part of the Rajakkamangalam intending to less flood vulnerability (Fig. 3a). The rainfall map was reclassified into four classes viz. 0–50 mm, 50–75 mm, 75–108 mm, and 108–190 mm.

Fig. 3
figure 3

a Rainfall map and b Slope map

Full size image

Slope

The flood occurs in the low elevated area due to the downward inundation of water from the higher region. The slope map (Fig. 3b) has been created from the ASTER DEM satellite image of the study area wherein the area ranges from 0° to 75°. The southern region initiates by the least slope category from 0 to 15°, increasing towards the northern side of the district, indicating high proficiency to flood hazard. On the contrary, the steepest slope category from 60 to 75° covers a small portion in the northeast, extending up to the middle south-eastern part of the district. The slope map was grouped into five classes as 0–15°, 15–30°, 30–45°, 45–60°, and 60–75°, respectively.

Drainage density (DD)

DD plays an influential role in the determination of flood vulnerability. DD map has been computed in the GIS environment from ASTER DEM (30 m) using the hydrological tools. The entire region is classified into three classes from low (0–0.016 km/km2), moderate (0.016–0.112 km/km2) and high (0.112–0.454 km/km2). Blocks with high drainage density are Thiruvattar, Thovalai, the eastern part of Thuckalay, and the northeastern part of Kurunthancode, which is very less vulnerable to flood. The lower density blocks such as Munchirai, Killiyoor, Rajakkamangalam, Agasteeswaram, and the southern part of the Melpuram are vulnerable to the flood. Adjacently, the regions such as Killiyoor, Kurunthancode, Rajakkamangalam, Agasteeswaram are highly vulnerable to flood because of the existence of coastal belt (Fig. 4a).

Fig. 4
figure 4

a Drainage density map and b LULC map

Full size image

LULC

LULC is the primary driver in changing the landscape of a specific area. LULC map has been extracted from Landsat-8 using supervised classification wherein, four classes, such as agricultural land, built-up, forest, and waterbodies, are identified. Agrarian land constitutes 867.6 km2, and forest covers about 545.09 km2 while built up and waterbodies comprise 206.09 km2 and 65.12 km2, respectively (Fig. 4b). The study highlighted that the agricultural land, built up, and water bodies are regarded as highly susceptible, while forest-covered regions are least vulnerable to flood.

DEM

DEM (Fig. 5a) posturizes the elevation of the district ranging from 0 to1816 m. The north and northeastern sides signify high altitude (1090–1816 m), whereas southern, southwestern parts belong to a low elevation (0–727 m) category. Thiruvattar, Melpuram, Thovalai blocks of 727–1090 m elevation are moderately vulnerable to flood while the low elevation coastal blocks, namely Rajakkamangalam, Agasteeswaram, Kurunthancode, Killiyoor, and Munchirai leads to highly flood-prone region.

Fig. 5
figure 5

a DEM map and b Soil texture map

Full size image

Soil texture

Soil texture has a sequential relationship with the rate of infiltration and surface runoff. The soil texture map and its classification has been generated based on the National Bureau of Soil Survey and Land Use Planning (NBSSLUP). The maximum traces compose of sandy clay covering 831.7 km2, i.e., 50% of the total area followed by clay and sandy clay loam representing 400 and 238 km2, indicating a high flood vulnerability along the district. While clay loam, loamy sand, and rock land cover 92.51 km2 described as less flood-prone area (Fig. 5b). The produced map has been classified into eight soil texture classes such as clay, clay loam, loamy sand, rock land, sand, sandy clay, sandy clay loam, and sandy loam.

Geology

Based on the historical records, it is evident that the Kanyakumari district was formed during the Achaean period covered with Archean metamorphic rocks. According to the Geological Survey of India (GSI), the Archean gneisses are the most predominant rocks separating two districts, namely Kanyakumari and Tirunelveli. This separation occurred due to the formation of lofty ridges of Western Ghats. Prolonged erosion led to a reduction of gneiss in height that appeared as subdued hills. Figure 6a depicts the geology map that has been prepared based on the GSI, where the plan was reclassified into seven categories. The middle parts of the district are covered with Peninsular gneisses and a considerable amount of Migmatite complex rocks. Subsequently, the south-eastern sides of the region comprise of fluvial sediment deposits, whereas the western, northwestern sides of the district are covered with pockets of Khandolite groups of rocks.

Fig. 6
figure 6

a Geology map and b Geomorphology map

Full size image

Geomorphology

Geomorphologically, the entire district has been classified vividly as structural hills, denudational hills, coastal plains, alluvial plains, peneplain, and aeolian plains. The middle and southern portion of the district comprises of aeolian, alluvial and coastal plains that are highly susceptible to flood. Contiguously, the northern, northeastern, and northwestern part are covered by structural and denudational hills that is resistant to flood (Fig. 6b).

Surface runoff

Surface Runoff is directly proportional to slope condition, soil saturation, rainfall, and LULC types (Mu et al. 2015). Preparation of surface runoff map involves parameters such as LULC, soil texture, rainfall, additionally flow direction, flow accumulation, a watershed delineation that were extracted from ASTER DEM. Moreover, the Curve Number (CN) for Antecedent Moisture Condition (AMC) was assigned using the LULC and the Hydrological Soil Group (HSG), which determines the soil condition of the study area. The union of all these parameters concludes the surface runoff model for the Kanyakumari district, which has been reclassified into five classes Fig. 7a. The northern parts record very low surface runoff ranging from 0 to 31 mm, but due to the availability of forest cover, the water gets infiltrated into the soil. Due to higher elevation, there is a chance of water getting discharged downwards to the lower region. Rajakkamangalam, Munchirai, Killiyoor, and Agasteeswaram blocks are experiencing very low surface runoff that tends to be highly vulnerable to flood hazards. However, the presence of dense forest cover around Thovalai, Thiruvattar, and Melpuram causes these blocks to be less sensitive to flood accompanied by very low surface runoff.

Fig. 7
figure 7

a Surface runoff map and b TWI map

Full size image

TWI

TWI was prepared using the ASTER DEM, which excerpts the slope and catchment area, providing an accurate TWI for the study area. Most of the blocks in northern parts such as Melpuram, Thiruvattar, and individual pockets of Thovalai, Munchirai are less flood-prone because of less TWI that occurs due to higher slope recorded in the district towards the northern side. The inundated water can easily penetrate the unsaturated soil, and the nominal amount of water flows into the low-lying areas of the district. Adjacently, the blocks in the southern side of the region, such as Rajakkamangalam and Kurunthancode, indicates a higher value of TWI due to higher saturation of the soil and lower elevation of the terrain. During seasonal rainfall, these areas water does not tend to penetrate the ground and hence creates a havoc situation. The areas considerably far from the coastal zone are moderately vulnerable to flood. The primary area covering 1005.5 km2 represents a high topographic wetness index gradually decreasing to 674 km2, while the lowest TWI is found in the minimal region of the district. The TWI map has been grouped into three classes depicted in (Fig. 7b).

Analytical hierarchy process (AHP)

AHP is the most reliable and flexible multi-criteria mathematical approach, was adopted in this current research to map the FVZ (Chen et al. 2011; Dandapat and Panda 2017; Zelenakova et al. 2015). Grouping the prepared thematic maps into five vulnerable categories, an AHP-based pair-wise comparison matrix of variables described above is constructed (Table 2). Noteworthy, AHP is applied to assign varied weights for comparing the ten individual factors, and according to their relative importance, these ten factors are rated from 1 to 9 (Table 3) on an absolute number scale (Saaty 1980). The values of the size range from less to more flood-influencing factors. The priority scales were synthesized by dividing each element of the matrix by its column total. Moreover, the priority vector is derived by calculating the row averages.

Table 2 Pair-wise comparison of 10 × 10 decision matrix
Full size table
Table 3 Pair-wise comparison scale for AHP preferences (Saaty 1980)
Full size table

Consistency check

To ratify the constructed pair-wise matrix and its allotted weightage (Saaty, 1980), evaluation through the Consistency Ratio (CR) was formulated where the acceptable CR must be below 0.1. In the present study, the consistency of the derived Eigen vector–matrix following the index below found is 0.093 concludes that the set of decisions considered is acceptable.

$$\mathrm{CR}= \frac{\mathrm{CI}}{\mathrm{RI}},$$
(1)

where CR is the consistency ratio, CI is the consistency index, and RI is the random index. RI refers to the consistency of the randomly evolved pair-wise matrix depicted in Table 4. The values provided in the table are subject to various parameters involved in AHP. Therefore, based on the ten variables, the derived RI obtained in the study is 1.49.

Table 4 Random Index (RI) value
Full size table

CI is computed using Eq. 2

$$\mathrm{CI}= \left(\frac{{\lambda }_{\mathrm{max}}- n}{n-1}\right),$$
(2)

where λmax stands for the maximum eigenvalue of the comparison matrix, n for number factors. Furthermore, the Flood Vulnerability Zone (FVZ) map has been generated using the Flood Vulnerability Index (FVI) values obtained from Eq. 3

$$\mathrm{FVI}= \sum_{\mathrm{i}=1}^{n}\mathrm{Wi}\times \mathrm{Ri},$$
(3)

where Wi is the weights for individual flood conditioning factors, and Ri is the rating class.

Finally, the obtained FVZ was reclassified into five groups viz. very high, high, moderate, low, and very low using natural breaks (Rocha et al. 2020).

Results and discussion

Current research emphasized on the FVZ preparation of Kanyakumari district using the MCDM AHP model. The 10 × 10 decision matrix was formed, where diagonal elements are equal to 1 (Kazakis et al. 2015) using the inducing variables with the derived CR = 0.093 (< 0.1 thresholds) as indicated in Table 2. The CR value obtained in the AHP analysis signifies that the weightage derived for the flood influencing factors such as rainfall (0.22), slope (0.124), drainage density (0.154), LULC (0.153), DEM (0.109), soil (0.068), Geology (0.052), Geomorphology (0.048), Surface Runoff (0.042) and TWI (0.03), respectively, are in acceptable range (Table 5) and can used in the FVZ preparation using the WLC.

Table 5 Weights of sub-classes using AHP comparison matrix
Full size table

Table 6 shows the resultant normalized pair-wise comparison, wherein each element of the 10 × 10 decision matrix was divided by the total sum of inter-comparison of priority rankings (Saaty 1980) of the considered parameters. Normalization indicates the scaling of the ten parameters from 0 to 1.

Table 6 Normalized pair-wise comparison
Full size table

FVI values of the flood vulnerable zone map range from 2.23 to 4.75 obtained using Eq. 3 and reclassified into five classes such as very low, low, medium, high, and very high to generate the FVZ map (Fig. 8) of the Kanyakumari district.

Fig. 8
figure 8

Flood vulnerable zone map

Full size image

Assessment of flood vulnerable zone map is represented in Table 7 where its respective areas along with percentages covering under distinctive FVZ categories have been derived through the AHP application. Considering the total area of the district, about 102.82 km2 (6.11%) falls under a very highly vulnerable zone while 813.48 km2 (48.31%) are recognized as highly sensitive regions. On the other hand, a total area of 305.94 km2 representing 18.17% is recorded as moderately vulnerable areas. Accordingly, low sensitive and very low vulnerable regions occupy a total of 337.78 km2 and 123.97 km2, respectively.

Table 7 Flood vulnerability status
Full size table

Additionally, the block-wise flood vulnerability zone is also shown in Table 8. The nine administrative blocks constituting the Kanyakumari district have been analyzed individually based on the differential vulnerable categories. The result highlights that 89.64 km2 (22.31%) of Kurunthancode block have been delineated as a very highly sensitive zone followed by Agasteeswaram (15.99%) and Munchirai (14.71%). Meanwhile, most of the blocks excluding Thiruvattaru, Thovalai and Melpuram are categorized as high vulnerability regions wherein chronologically, 97.64% area of Killiyoor, 87.03% area of Rajakkamangalam, 85.14% area of Munchirai, 74.79% area of Agasteeswaram and 73.24% area of Kurunthancode block are identified as highly flood vulnerable regions. This can be attributed due to adjacency to the coastal belt, torrential downpour, high population density, gentle slope coupled with haphazard urban growth, injudicious anthropogenic activities, sparse vegetation and forest cover, etc. On the contrarily, the moderately vulnerable blocks involve Melpuram representing 30.42%, followed by Thovalai (28.54%) and Thiruvattaru (21.27%). Similarly, the very low and low sensitive blocks are recorded as Thovalai and Thiruvattaru, respectively. These regions witness less exposure towards flood as it is dominated by dense vegetation and forest cover, presence of extensive plains as a hindrance for settlement, and other developmental activities.

Table 8 Block-wise area of flood vulnerability
Full size table

Conclusion

The present research elucidates flood vulnerability mapping as a prerequisite approach in identifying and predicting the degree of susceptibility of various regions towards the natural calamity. Being the most efficient and widely acceptable technique, MCDM-based AHP model approach was used to carrying out the flood vulnerability analysis of the Kanyakumari district. In this regard, according to the decision maker’s preference, the weights are allotted to the ten influencing variables that represent their respective priorities. Amongst all the variables, rainfall with (0.22) weightage is highly significant, while TWI is of least significance (0.03) that substantially contributes to occurrence of flood in the region. Similarly, the layers of the parameters have also been reclassified into five groups (Table 5) from 1 to 5, wherein one indicates very low, and five shows very high that represents their priorities in causing such hazard. Thus, the study reveals that a majority of Kanyakumari district has been delineated as very high to high vulnerability zones covering an area of 916.3 km2. However, the southern parts of Kurunthancode, coupled with the southwestern part of Rajakkamangalam, south and south-east part of Agasteeswaram, south and western part of Thuckalay, south part of Killiyoor and southwestern part of Munchirai are regarded as highly susceptible blocks to the flood havoc. The total area found under the obtained five classes in percent (Tables 7, 8 and Fig. 8) is represented as very high (6.11), high (48.31), moderately (18.17), low (20.05), and very low (7.36) vulnerable.

The most causative factor behind the presence of vast flood vulnerable zones is its adjacency to the coastlines triggered by coastal floods. Unplanned urbanization, aided by rampant population growth, is also another notable factor that needs to be addressed in the study area. Such type of modeling and mapping of selected sites enhances to recognize the flood-prone regions that could be assumed as the foremost step in preparing flood plans. Moreover, the executed analysis might provide a baseline for planners, civil engineers, Non-Governmental Organization (NGOs), government officials, policymakers as it would help to implement manifold strategies required for effective planning so that these natural hazards could be monitored and mitigated to a certain extent. Thus, multi-criteria-based models could further help in executing the unpredictable hazards in a sustainable manner.