Introduction

Flash floods are being considered as world’s devastating natural calamities (Costache 2019; Obeidat et al. 2021; Alam et al. 2021) amid natural calamities in view of property and human loss (Hong et al. 2018), chiefly owing to climatic and land-use changes, yet very little information is available at regional level (Kelman et al. 2018), as these phenomena befall at such spatiotemporal scales that conventional rain and river discharge gauging stations stand inadequate to assess them properly (Creutin and Borga 2003; Alam et al. 2021). This in turn results in poor understanding of the driving forces, accompanying a very unsure warning and hazard management procedures (Borga et al. 2014; Liu et al. 2021). Consequently, it is of paramount importance to exactly identify the combination and interplay of these driving forces responsible for triggering such extreme events. Watersheds with good environmental settings are crucial in delivering suitable habitats for wildlife, uncontaminated water to aquatic ecosystems and for drinking (EPA 2014), adapting to land-use and climate changes and adding toward sustainable development (Sadeghi et al. 2019). But, unfortunately, several studies in recent times have observed either watershed degradation or leaning toward deterioration due to human interferences and climate change (Sciera et al. 2008; Ferreira et al. 2017; Bhat et al., 2021; Dinata et al. 2021). The higher sediment rates in reservoirs and streambeds result in water quality degradation and reduce the potential for irrigation and power generation while augmenting occurrence and severity of floods downstream (Wolde 2016; Pourghasemi et al. 2021). Sustainable management of water resources like hydrological units in the form of watersheds must be conserved through tackling environmental hazards like water erosion, land degradation and flash floods (Pourghasemi et al. 2021). The ecosystem services potential of watersheds are dependent on their morphometric drainage characteristics. Morphometry offers insight toward interpreting the geomorphic structure, hydrological setup, mass wasting and water erosion features of the catchment drainage areas (Hajam et al. 2013a; Meshram et al. 2019). Morphometric attributes provide insights on hydrological behavior of watersheds to high-intensity rains and will be valuable for better understanding of flash flood potential of drainage basins and also helps in evading the devastations caused by the hazard (Alam et al. 2021). Quantifying morphometric parameters of watersheds by applying remote sensing and geographic information system (GIS) has a substantial importance in characterizing the river basin watersheds, wherein it encompasses computation of landscape attributes associated with its linear, aerial and relief aspects. Assessed applications of drainage morphometry include management of drainage basin resources and their evaluation (Pourghasemi et al. 2021), land surface form characterization (Thomas et al. 2012), identification of watershed hazards (Nitheshnirmal, et al. 2019), vegetation growth prospective (Kadam et al. 2017), sediment yield (Pourghasemi et al. 2021) and prioritization of catchment areas for safeguarding of water and soil resources (Romshoo et al. 2012; Rahmati et al. 2019).

Himalayan region is inherently elusive as it is characterized by expanded mountain systems, tectonically unstable and susceptible to natural hazards (Singh et al 2021). Flood assessment and management to comprehend aquascape landscape interaction processes are mostly shoddy in the Himalayan regions due to shortage of long-term hydrological datasets (Tabbussum and Dar 2020). Himalayan mountainous watersheds, which are complex in nature, are highly vulnerable to natural hazards, especially erosion and flash floods, causing extensive damage to life and property in particular and ecosystems in general (Altaf et al. 2014; Lepcha et al. 2021). The factors responsible for flash floods are necessary to be connected with a multi-criteria decision-making approach in a GIS-based environment (Singh et al. 2021). Unstable slopes and degraded watersheds can trigger flash floods and instigate an extreme hazard to lives, subsistence and infrastructure, both in the hilly areas and in the plains, and the most susceptible being underprivileged people, individuals with disabilities, children, elderly persons and women. For example, extraordinary and unprecedented flash floods of September 2014 in Kashmir region were one of the disastrous events which took more than 80 lives, with infrastructure and business loss of worth 8 billion USD. Despite the heavy destruction and serious impact that these flash floods and erosion inflict on socio-economy (Mikos 2011; Diakakis et al. 2016), the management of these hazards has not received adequate attention to reduce the magnitude of the risk. In the Jhelum basin of the northwestern Himalayan region, morphometry and prioritization of only some watersheds have been appraised with respect to sediment yield, erosion rates and flood hazard assessment (Altaf et al. 2014; Meraj et al. 2015; Bhat et al. 2019). Nonetheless, in these approaches also, watershed priority ranking was achieved through estimation of compound parameter value by providing equal significance to all morphometric parameters. Thus, the partiality linked with weights of individual parameters was ignored and led to imprecise ranking scheme results. Every watershed unit has its own unique characteristics which therefore renders that equal importance cannot be provided to all parameters for identifying any potential area for management and risk valuation.

In this context, the present approach involves delineation of morphometric attributes for sustainable watershed management planning of the Jhelum River Basin. Quantitative morphometric attribute analysis of the JRB is important due to the dearth of long-term hydrological records despite well-distributed gauged stations. Having said about the importance of healthy watersheds and the potential threats to disturb this relation and balance, together with very few works of this orientation, it was felt to take this research question. In order to overcome the above-noted limitation on previous studies, the aim of this research was to assess: (1) stream watershed characteristics in terms of morphometric attributes and slope classes, and (2) spatial prioritization of the JRB watersheds based on erosion and flash flood susceptibility index (EFSI) determined through principal component analysis (PCA)-based attribute weightage scheme (PCA-AWS). This methodology, therefore, shall come handy (especially in data-scarce ungauged regions) for enhanced understanding about effective control measures related to erosion and flash floods of basin watersheds besides safeguarding sustainable use of resources through better planning, decision making and management. The research work therefore can support in achieving the sustainable development goals (SDGs) related to land degradation, fluvial disasters, human and economic loss.

Materials and methods

Study area

The Jhelum River Basin (JRB) of Kashmir Himalayan region is of late Miocene evolved basin with elongated NW–SE Graben-type or pull-apart sedimentary trough (Alam et al. 2017) with dimensions of 130 km length and 40 km width and consists of about 15,000 km2 area. The Valley is encompassed by Pir Panjal mountains from the southwestern side and the Himalayan Great mountain range toward the northeastern flank (Fig. 1). Eleven tributaries drain from each of the two surrounding mountain ranges and radially confluence the main trunk of the River Jhelum (Bhat et al. 2019). The river meanders through the central city (Srinagar) of Valley before it enters into Wular Lake in the northwestern direction. The river, after exiting from the lake, cuts across Pir Panjal through Baramulla-Uri gorge and then flows into Muzaffarabad. The Jhelum waters harbor a rich resource of fisheries besides serving drinking water and agriculture sustenance in its catchment and therefore has a tremendous socioeconomics linked to it (Rather et al. 2016).

Fig. 1
figure 1

Map of study area showing the Jhelum River Basin (JRB)

Full size image

Geologically, the basin is characterized by heterogeneous rock types including Agglomeratic slate, Panjal traps, Gneissose granite, Shale, Quartzite inclusions, Limestone, Karewa formations and River alluvium (Bhatt 1989). Land system uses of the whole Valley consist of cropland, barren area, water plant foliage, bare rocks, settlement area, forests, scrubland, meadows, horticulture, plantation (willow and poplar), snow and glacier (Murtaza and Romshoo 2014). The characteristic weather of this region is of sub-humid temperate nature with erratic meteorological conditions. The month of July remains the hottest with average lower and higher temperatures of 16 °C and 32 °C, respectively, and the coldest period falls between December and January having − 15 °C and 0 °C as mean minimum and maximum temperatures, respectively. The Jhelum River Basin is selected for this study in view of witnessing clear signs of climate change in terms of disproportion in precipitation forms. At times, the basin experiences intense rainfall and thus preludes erosion and flash floods as environmental threats and at times beholds moderate to severe water shortages, thereby yearning mitigation measures in the form of management and planning.

Overview of the methodology

The overview of the methodology is represented in Fig. 2. The methodology consists of major steps which include watershed drainage system analysis, watershed slope analysis and attribute weightage scheme.

Fig. 2
figure 2

Overview of stepwise methods employed during the study period

Full size image

Tools and techniques used for watershed drainage system analysis

The extraction of watershed drainage morphometry of the JRB (Fig. 3) has been accomplished using SRTM-DEM (shuttle radar topography mission—digital elevation model) having 30 m spatial resolution and ground truthing incorporated in a well-founded framework applying GIS procedures. In ArcGIS (vs.10.2) software, the hydrology tool under spatial analyst program was adopted for extraction of drainage channels and other morphometric attributes (Dar et al. 2021a). The initial step for the generation of drainage network is to fill the sinks in DEM to make it devoid of any depressions for robust and accurate estimation of direction and accumulation of flow grids. In order to estimate flow direction of a grid cell, the D8 algorithm was applied (Altaf et al. 2014). The grid cell with steepest descent among the eight adjacent cells was used to determine each grid cell direction. Thereafter, stream networks for each watershed were generated using the accruing number of cells from upstream flanks which drain to the flow accumulation grid cells. A threshold level of 100 was assigned for stream definition in flow accumulation layer in order to create streams for all watersheds in the basin. Further, boundaries of each stream watershed were delineated by resorting to pour point delineation function in ArcGIS, considering this point at a location wherein drained water from whole watershed flows into the mainstream (Altaf et al. 2014; Meraj et al. 2018). The JRB was characterized into 18 major stream watersheds, and they were labeled, in this study, by the names of respective streams flowing in them (Fig. 4). Stream ordering was achieved in the ArcGIS environment, using Strahler’s method. The stream watershed morphometric attributes were assessed as per the standard mathematical formulae presented in Table 1.

Fig. 3
figure 3

Stream orders (based on Strahler 1964) of 18 major streams of the JRB

Full size image
Fig. 4
figure 4

Delineation of 18 major stream watersheds of the JRB and labeled after the names of streams flowing in them

Full size image
Table 1 Methods employed for estimating morphometric attributes of stream watersheds of the Jhelum River Basin
Full size table

Stream watershed slope analysis

The present study implements NRCC (1998) slope classification to describe the relation of watershed slope classes with erosion and flash floods. The procedure used by Tarboton (1989) has been applied for generating slope classes with the help of spatial analyst tool programmed in ArcGIS (v.10.2). Stream watersheds of the JRB have been characterized into ten slope classes based on guidelines provided by NRCC (1998). The slope classes comprise level (L, 0°–0.3°), nearly level (NL, 0.3°–1.1°), very gentle slope (VGS, 1.1°–3.0°), gentle slope (GS, 3.0°–5.0°), moderate slope (MS, 5.0°–8.5°), strong slope (SS, 8.5°–16.5°), very strong slope (VSS, 16.5°–24°), extreme slope (ES, 24°–35°), steep slope (StS, 35°–45°) and very steep slope (VStS, 45°–90°) (Fig. 5). In order to measure the watershed slope impact on its erosion and flash flood potential, EFSS values were consigned taking into consideration the percentage area covered by each slope class (Table 3).

Fig. 5
figure 5

Categorization of stream watersheds into 10 slope classes; level (L), nearly level (NL), very gentle slope (VGS), gentle slope (GS), moderate slope (MS), strong slope (SS), very strong slope (VSS), extreme slope (ES), steep slope (StS) and very steep slope (VStS)

Full size image

Attribute weightage scheme (AWS) based on principal component analysis (PCA)

Toward the assessment of erosion and flash flood vulnerability of 18 major stream sub-basins of the JRB, 23 morphometric attributes and 10 slope classes of watersheds were analyzed through PCA. In order to assign weights on these parameters, PCA-based AWS comprises the below-mentioned steps:

Apportioning of erosion and flash flood susceptibility scores (EFSS)

The first step is the allocation of erosion and flash flood susceptibility scores to the attributes as per their influence on processes which yield erosion and flash floods at a watershed scale. Morphometric attributes like mean bifurcation ratio, basin area, basin perimeter, shape index, form factor, circulatory ratio, stream frequency, drainage density, drainage intensity, drainage texture, infiltration number, basin relief, relief ratio and ruggedness number possess a direct association with erosion and flash floods, and consequently their higher values depict maximum amount of erosion and flood vulnerability. However, parameters such as elongation ratio, basin length, overland flow length and constant channel maintenance bear a converse relation with erosion and floods, and thereby maximum values of these attributes reflect less susceptibility to erosion and floods. In order to quantify combined effect of morphometric attributes on erodibility factors in all the 18 tributary watersheds, directly related parameter having the maximum value received an erosion and flash flood susceptibility score (EFSS) of 18 and the next higher value of watershed, in decreasing order, retains EFSS of 17, and so on. However, watershed holding the least value of conversely associated parameter keeps EFSS of 18 and the stream watershed having subsequent minimum value in an increasing level was consigned EFSS of 17, and so on.

Similarly, in the case of slope attribute, the relationship with erosion and flash floods was quantified using percent area occupied under different slope categories (Table 4). Among 18 tributary watersheds, the watershed retaining the minimum percentage of a slope class and having an inverse relationship with erosion and flood susceptibility was assigned EFSS of 18 and further apportioning of EFSS continues in an ascending fashion till the highest value of watershed percent area is reached. But, the watershed that occupies the maximum percentage of slope class and is proportionally linked to erosion and flash floods was provided EFSS of 18, and it continues in descending order until watershed with the lowest percentage area of slope category is scored.

Estimation of attribute weights (AW)

The PCA-based attribute weightage scheme is an empirical and semiquantitative method applied to evaluate the morphometric attribute weights. These weights are assessed with a reason that drainage areas of similar basins possess integrity with respect to geomorphology, lithology, pedology and topography and thus express homogeneous response and behavior to numerous land surface processes. During the current study, this relation was accomplished by performing PCA on 23 morphometric attributes and 10 slope categories of major stream watersheds of the JRB. Scree plot displaying principal component versus eigenvalues was used to decide the number of components incorporated for further examination. Three components having eigenvalue > 1 and explaining more than 80 percent variance in the database (Fig. 6) were taken for ascertaining AWS weights (Table 4). PCA is the most consistent pattern recognition method used to procure details by changing primary, interlinked variables into less uncorrelated parameters referred to as principal components (PCs) (Dar et al. 2021b; Islam et al. 2021). The input variables of PCA are interrelated, while the assumed factors (PCs) are orthogonal and acquired through linear combination of the experimental variables (Hatvani et al. 2014). The coefficients of correlation achieved from primary variables and principal components furnished the factor weightings which define the weights of PCs in the primary parameters. The PC is computed here as follows:

$$z{}_{ij} = a{}_{i1}x{}_{1j} + \;a_{i2} x_{2j} + \cdots + \;a_{im} x_{mj}$$
(1)

wherein z symbolizes component score, a indicates component weighting, x designates estimated parameter value, i is the component number, j infers the sample number and m implies the whole number of parameters.

Fig. 6
figure 6

Scree plot showing principal components (PCs) with eigenvalues. Three PCs having eigenvalues > 1 and explaining more than 80% variance were used

Full size image

In PCA, loadings on all correlated variables more than a threshold level are known as substantial loadings. This study selects a threshold value of 0.5 as strong substantial loadings based on the literature (Field 2009). Summation and normalization of substantial loadings of three PCs provide us AWS weights (AWSW) as shown in Table 5. Each individual PC consists of a group of attributes which govern erosion and flash flood susceptibility in a homologous manner. Therefore, this technique apportions weights based on pooled influence of homologous groups on erosion and flood susceptibility instead of individual influence of the attributes. When AWSW is multiplied with erosion and flash flood susceptibility scores (EFSS) of attributes, we obtain weightage-based scores (AWSS) (‘Appendix’). Lastly, AWSS of each stream watershed are summed up and the final EFSI (erosion and flash flood susceptibility index) is achieved. The PCA-based AWS model equation is expressed as follows:

$${\text{AWSS}}_{i} = {\text{AWSW}} \times {\text{EFSS}}$$
(2)
$${\text{EFSI}} = \sum\limits_{i = 1}^{N} {{\text{AWSS}}_{i} }$$
(3)

where AWSSi = attribute weightage scheme score of an attribute of ith watershed; AWSW = attribute weightage scheme weights; EFSS = erosion and flash flood susceptibility score of an attribute of watershed; EFSI = erosion and flash flood susceptibility index; N = number of attributes of a particular watershed.

Results and discussion

Analysis of morphometric attributes of stream watersheds

The PCA-based weightage technique used in this research study qualitatively parametrizes the erosion and flash flood yielding processes. Morphometric attributes of watersheds as inputs in the weightage scheme model established linkage between shape and dimensions of geomorphic features and various land surface processes (Rashid et al. 2011). In this study, nineteen morphometric attributes exhibited direct association with erosion and flash flood susceptibility (EFS), while four parameters showed opposite relation with EFS (Table 2).

Table 2 Morphometric attributes of major stream watersheds of the Jhelum River Basin and their corresponding erosion and flash flood susceptibility scores (EFSS) mentioned in parentheses
Full size table

Morphometric attributes exhibiting direct relation with EFS

The preliminary level of understanding about watershed drainage characteristics is accomplished through delineation of stream orders and subsequently by estimating number and length of streams. Higher stream orders correspond to a greater dissection of watersheds which results in more stream discharge and water current velocity (Costa 1987). The highest stream order of 9 was retained by Lower Jhelum followed by Pohru (8), Sind (7), Lidder (7) and Vishav (7) and thus possessed great potential for erosion and flash floods during intense rainfall. Watersheds like Rembiara, Doodhganga, Sukhnag, Madhumati, Erin, Dachigam, Arapal, Kuthar and Bringi are having sixth-order streams in them and thereby have erosion and flood developing capacity as well. Romshi, Ferozpur, Ningli and Sandran are fifth-order streams and hold least possibility of erodibility and flooding. The highest number of streams is possessed by Pohru, Sind and Lidder which reflects high discharge, more erosive potential and rapid peak flow in case of rain storm events. Ningli, Sandran and Dachigam under similar conditions contribute less flow and erosion because of containing less number of streams (Table 2).

Mean bifurcation ratio (Rbm) is an essential morphometric attribute to influence the peak level of hydrograph related to runoff (Jain and Sinha 2003). Higher values of Rbm depict rapid discharge and risk of erosion and flash flooding if rainfall prolongs (Rakesh et al. 2000). Rbm values of this study vary from 3.47 to 5.75 (Table 2), revealing mature topography, high dissected terrain and higher potential of discharge (Eze and Efiong 2010). Romshi, Sandran, Ningli and Ferozpur hold high Rbm values and hence are more susceptible to flash floods and related to higher sediment loads because of disturbed pedon structure (Altaf et al. 2014). Watershed basin area (A) and perimeter (Pb) determine spatial expanse and volume of sediment conveyance and flooding (Altaf et al. 2013). Consequently, larger A and Pb of watershed mean more erosion and flash flood potential. Among all 18 major stream watersheds of JRB, Pohru consists of largest basin area (2075.83 km2) and Sindh owns the most extensive basin perimeter (309.51 km), while Sandran registers the smallest area (290.51 km2) and shortest perimeter that of Dachigam (103.25 km).

Shape index (Sw), form factor (Ff) and circulatory ratio (Rc) are highly related to one another and are practiced to express configuration of watersheds (Soni 2017). Shape of watersheds fundamentally influences sediment and water yield potential from drainage basin. Sw is the highest for Pohru, Sind, Lidder and Vishav, reflecting higher yield of sediment and flash floods in these watersheds. Sandran, Ningli, Kuthar and Dachigam records the lowest Sw values which demonstrate minimum chances for high erodibility and water yielding. Likewise, high Rc values signify the circular shape of watersheds having medium to highly elevated relief with impermeable surface triggering higher water flows in a short period, and vice versa is true for low Rc (Sreedevi et al. 2005). Maximum Rc values are observed for Pohru, Arapal and Lower Jhelum, suggesting their peak flows in shorter duration. Moderate to low Rc values of other sub-basins indicate their low runoff potential, prolonged flow discharge times and longer basin lag times (Bhat et al. 2019).

Stream frequency (Fs) is inversely linked with perviousness, vegetation and infiltration magnitude; however, it is related directly to watershed relief. Fs values range from 2.84 to 3.41 per km2 (Table 2) in this study. In general, stream watersheds of great Himalayan side like Lidder, Sindh, Kuthar, Arapal, Sandran, Bringi and Erin endure high Fs values, thus revealing impervious subsurface material, maximum runoff capacity and early peak discharges to Jhelum River during high intense rainfall. On the contrary, low Fs values were observed for Pir Panjal flanked watersheds which include Rembiara, Doodhganga, Sukhnag, Romshi, Ningli, Ferozpur and Vishav taking longer durations for producing peak discharges to the Jhelum River. Similar kind of observations is also witnessed by earlier study related to only upper sub-basins of the Jhelum River (Bhat et al. 2019). High and low values of drainage density (Dd) in stream watersheds are procured depending on the nature of underlying substance (impermeable/permeable), relief (high/low), plantation (sparse/good), runoff (maximum/minimum) and flash flood volume (large/small) (Pallard et al. 2009). In this research, elevated values of Dd were witnessed for Romshi, Vishav, Sukhnag, Ferozpur, Lidder, Sindh and Arapal watersheds due to either impervious surface, sparse vegetation or high relief and thus contributing maximum runoff and sediment loads in less time durations. Sandran, Dachigam, Ningli and Lower Jhelum have the lowest Dd values because of low relief and good vegetation, implying low erodibility and flooding. Rembiara, Kuthar, Bringi, Doodhganga, Pohru, Madhumati and Erin have moderate Dd, which infers contribution of average water discharge toward flooding. The soft terrain (unconsolidated formations) covered by vegetation yields fine texture, and the hard rock terrain (consolidated formations) retains rough texture (Sreedevi et al. 2009). Drainage texture (Dt) is categorized among four classes: coarse (< 4/km), intermediate (4–10/km), fine (10–15/km) and very fine (> 15/km) (Smith 1950). Our observed Dt varies from 3.06 to 29.92, suggesting textures ranging from coarse to very fine (Table 4). Very fine Dt class includes Pohru, Lower Jhelum, Lidder, Sindh, Arapal and Vishav having high water yielding potential and small response time compared to other watersheds. Infiltration number (If) explains about infiltration physiognomies of watersheds and is conversely associated with permeation capacity of stream watersheds (Romshoo et al. 2012). The greater If values noticed for Arapal, Sind, Lidder and Vishav specify their higher susceptibility to erosion and water flows. Pohru, Lower Jhelum, Romshi, Erin and Sukhnag receive moderate If values and thus identify medium type of erosive and flooding vulnerability in them. Lowest values of infiltration number were registered for Dachigam, Sandran, Bringi, Kuthar, Madhumati, Ferozpur and Ningli and therefore have the least possibility of high water flows.

Relief aspect attributes (H, Rh and Rn) of stream watersheds furnish information related to denudation characteristics, morphological features, watershed steepness and erosional intensity. The Rh value of stream watersheds did not show much difference, but nevertheless Lidder, Rembiara and Erin sub-basins revealed steep sloping terrains and therefore higher basin energy along slopes with strong erosional processes and faster runoff, resulting in high peak discharges (Altin and Altin 2011). Higher values of H and Rn registered for Sind, Lidder, Erin, Kuthar, Lower Jhelum, Vishav and Romshi watersheds expressed an intense availability of erosive potential energy to generate high runoff and sediments down the slope (Altaf et al. 2014; Hajam et al. 2013b). Sandran, Ningli and Pohru showed smaller H and Rn values causing low water flows due to lesser degree of terrain complexity, while Bringi, Arapal, Dachigam, Madhumati, Doodhganga, Sukhnag and Rembiara sub-basins have moderate relief attribute values and therefore exhibit medium erosion and runoff potential (Bhat et al. 2019).

Morphometric attributes revealing inverse relation with EFS

Basin length (Lb), elongation ratio (Re), length of overland flow (Lg) and constant of channel maintenance (C) were inversely proportional to EFS. Lb is considered to be an important parameter in transporting sediment loads from a watershed. Higher values of Lb indicate lower sediment yield potential as sediments are provided sufficient time and longer distances to settle down prior to flowing into the main river. Of all the JRB watersheds, Pohru, Sindh, Lidder and Vishav have the longest Lb values, while Sandran and Ningli stream sub-basins possess the shortest basin lengths (Table 2). Re helps us to recognize hydrological character of drainage basins, and its value ranges between 1 for circular watersheds to 0 for elongated basins. Large values of Re yield rapid peak flows in lesser times and are considered as highly dangerous (Masoud 2016). Lidder, Sindh and Pohru retained the lowest values of Re and hence depict their elongated character among all stream basins. Watershed slope and soil stability are indicated by the constant of channel maintenance (C). Smaller values of C imply unconsolidated soils, sparse vegetation and precipitous mountain terrains and subsequent higher vulnerability to erosion and flash floods (Altaf et al. 2014). Lidder, Sindh, Vishav and Rembiara watersheds delivered the lowest C values and thus remain prone to erosion and flash flood susceptibility when compared to other stream basins. Lg influences development of hydrological and geomorphic setup of stream watersheds. The highest values of Lg were observed for Sandran, Ningli, Dachigam, Bringi, Madhumati, Lower Jhelum and Pohru and therefore characterized these watersheds by longer flow tracks and gentler gradients with minimum overflow and high infiltration.

Watershed slope as influencing factor in weightage scheme model

Slope classes of stream basins considerably impact the erosion and flash flood vulnerability. Percentage of slope area under different slope categories related to major stream watersheds of the Jhelum River Basin parametrizes the influence of mountain slope processes on water flow discharge and erodibility. Of all watersheds, Lower Jhelum, Madhumati, Lidder, Sindh, Bringi and Erin have the highest percentage of land area under steep slope classes (SS, VSS, ES, StS and VStS), thus depict the highest rugged topography and, as a consequence, lead to erosion and flash flood susceptibility upon intense rainfall events. A moderate kind of denudation and flooding are experienced by Vishav, Pohru, Kuthar, Arapal and Sandran watersheds as shown by their percent area in different slope categories of steep nature. Stream basins like Rembiara, Romshi, Doodhganga, Sukhnag, Ferozpur and Ningli have minimum area under steep slopes and would exhibit least erosion and deluging potential among the watersheds. The results of slope analysis and their corresponding EFSS apportioned to stream watersheds are mentioned in Table 3.

Table 3 Slope area (%) under different slope categories related to major stream watersheds of the Jhelum River Basin and their corresponding erosion and flash flood susceptibility scores (EFSS) given in parentheses
Full size table

PCA-based attribute weightage scheme (AWS) outcomes

Integration of weighted influence of 23 morphometric attributes and 10 slope categories on erosion and flash flood susceptibility of major JRB stream watersheds was accomplished by performing PCA-based weightage scheme. The model grouped all input parameters into three principal components (PCs) (Table 4) and then determined combined weights of parameters present in each PC (Table 5). In this way, the whole weightage setup evades any intricacy linkages which are usually witnessed in process-based simulations that perform on almost near-perfect weightage basis. The final equation for PCA-based attribute weightage scheme is given as follows:

$${\text{EFSI}} = 0.57 \times {\text{EFSS}}\left( {{\text{PC}}_{1} } \right) + 0.15 \times {\text{EFSS}}\left( {{\text{PC}}_{2} } \right) + 0.28 \times {\text{EFSS}}\left( {{\text{PC}}_{3} } \right)$$
(4)

where EFSI = erosion and flash flood susceptibility index, EFSS = erosion and flash flood susceptibility score and PC = principal component.

Table 4 Principal component loadings of morphometric parameters
Full size table
Table 5 Estimation of PCA-based weights (AWSW)
Full size table

EFSI values obtained from the above equation for all stream watersheds have been categorized into three priority classes using equal interval method in GIS environment (Table 6). Thereafter, susceptibility map related to all stream watersheds was generated and is shown in Fig. 7. The high-priority group comprising Sind, Lidder, Lower Jhelum and Pohru watersheds exhibited the highest EFSI values. These stream basins, consisting of 44.66% area of the JRB, possess the most hazardous erosion potential and flash flood vulnerability in the possibility of any rainfall storm event. This is followed by moderate-priority class encompassing Bringi, Vishav, Arapal, Rembiara, Madhumati, Erin, Doodhganga and Sukhnag. These watersheds constitute 38.45% of total area of Jhelum basin. Finally, 16.88% of basin area comes under low-priority class which includes Romshi, Ferozpur, Kuthar, Sandran, Dachigam and Ningli watersheds.

Table 6 Erosion and flash flood susceptibility index (EFSI) values (in decreasing order) of stream watersheds along with priority classes and percentage area occupied by them in the Jhelum River Basin
Full size table
Fig. 7
figure 7

Prioritization of stream watersheds of the JRB into high-, moderate- and low-priority classes based on EFSI (erosion and flash flood susceptibility index)

Full size image

The results of prioritization hence indicated that more than 80% of the Jhelum basin stream watersheds fall under high- to moderate-priority classes due to higher EFSI values. This is one of the main reasons that flood plains of Jhelum basin witness frequent floods in the eventuality of heavy downpours. The recent example is September 2014 floods causing havoc and huge damages to life and property in various districts of the Kashmir valley, which are considered to be worst in the past 60 years. Besides being destructive in nature, these waters detach huge quantities of sediments from mountainous watersheds and loads them downward. This phenomenon not only has deleterious effects on various aquatic ecosystems but causes shrinkage in river dimensions and spill channels. Eventually, this reduction in storage of downstream rivers and waterbodies build grounds for embankment breaches and river bank overflows, echoing more trouble in the flood plains, particularly in south Kashmir and Srinagar City. In light of this scenario, the government-run administration should have far in the past initiated the construction of many reservoirs in south Kashmir area to divert the water of Jhelum, Lidder, Vishav and Rembiara, and other tributaries for slow release of water into the Jhelum downward. This could have saved the inundation of South Kashmir during 2014 disastrous flood, which gets submerged due to backflow water near Sangam wherein three major tributaries—Lidder, Vishav and Rembiara enter into the main Jhelum within a distance of few kilometers.

In this study, thus, it is imperative to develop management strategies about stream watersheds which are at high degree of risk for soil erosion and flash floods. Prevention and mitigation approaches should rely on both ‘gray’ infrastructure and nature-based solutions referred as ‘green’ infrastructure. Gray infrastructure involves engineering solutions like formation of reservoirs, stable dykes, check dams, strategic cultivation practices in high-slope areas and desiltation of downstream rivers and spill channels.

Nature-based solutions include robust approaches and practices to exploit or refurbish natural land system uses and steer floodplain features inside watersheds, by either storage or minimizing flow down waters, to provide more duration for peak flow to get flatten (SEPA 2013). Such approaches not only reduce erosion and flooding but also deliver an array of further advantages including protection and fostering of biodiversity, improving water quality, recreation and resilience to climate change (Iacob et al. 2014). Growing more forests, tree plantation and other vegetation types in these stream watersheds would diminish runoff through intensified permeation using their root systems that augment macroporosity of soil. In addition, vegetation cover intercepts precipitation briefly by retaining it on surfaces besides the herb cover and litter layers and therefore gets returned to atmosphere before it reaches the ground (Attarod et al. 2015). Moreover, degraded lands like barren areas in high-risk watersheds should be delineated and converted to vegetation lands to enhance infiltration and lessen erosion and high discharge flows. Several well-established soil bioengineering methods like bush-layering, live crib walls and erosion control blankets can also be used to complement conventional methods for slope and embankment stabilization and erosion control. Restoration of wetlands and lakes across the fluvial plains can immensely act as sponges to store floodwaters and release them gradually.

Relevance of the study to land and water management and sustainable development goals

High susceptibility of watersheds to erosion and flash floods, as determined by EFSI, increases the risk of diminishing the value of land availability and freshwater resources. Intensive agriculture and deforestation are major causes of land degradation involving soil erosion (Dar et al. 2020), leaving large areas vulnerable to the loss of fertile top soil (Mir et al. 2017). This, along with the loss of associated nutrients and chemicals to water bodies, is a serious threat to sustainable agricultural production, environmental protection and food security (Kopittke et al. 2019; Thompson et al. 2020). Adopting appropriate conservation practices such as establishing water retention basins, building terraces and intercropping can significantly reduce soil erosion in high-risk areas like Sind, Lidder, Lower Jhelum, Pohru, Bringi, Vishav, Arapal, Rembiara, Madhumati, Erin, Doodhganga and Sukhnag. Further, watersheds such as Sind, Lidder, Lower Jhelum, Pohru, Bringi, Vishav, Arapal, Rembiara, Madhumati, Erin, Doodhganga and Sukhnag can be taken for agroforestry—wherein there is a deliberate integration of woody perennial plants (trees and shrubs) with crops or livestock. The trees and shrubs in agroforestry systems can be selectively protected and regenerated, or planted and managed. These high-erosion-prone areas should adopt conservation agriculture practices like minimal soil disturbance by reducing tillage, maintenance of crop residues on lands and diversification of crop species grown in associations. Such methods improve water infiltration, decreased runoff and erosion, and less evaporation from the soil surface, along with increased soil biological activity and soil organic matter (Beddington et al. 2012). During rainfall events, fair amount of runoff gets generated from high-susceptible areas of the JRB and thereby yearns for water harvesting techniques. These include planting pits, raised earthen barriers covered with grasses, line of stones placed along contours and ridge tillages. Nature of such tasks not only reduces land degradation but also maintains water table high enough which can be utilized in dry spells through extraction (Kumar et al. 2011). In order to manage water resource efficiently, construction of irrigation channels in a terracing pattern should be built in the above watersheds. Due to rugged topography and high slopes in the region, huge volume of waters get accumulated in the eventuality of any heavy and long spell of downpours. This necessitates the building up of reservoirs especially in high and moderate zones of watershed areas which would be of immense significance vis-à-vis hydropower generation, agriculture production, horticulture development and reverse land degradation as well as help rehabilitate the watershed’s natural resources and biodiversity productivity.

Additionally, this research has a substantial relevance vis-à-vis the 2030 Agenda for Sustainable Development with its 17 SDGs (sustainable development goals) at its core, as adopted by the UN Sustainable Development Summit in 2015. The study helps in identifying the watershed areas that fall under high- to moderate-priority classes and therefore aims in reducing the exposure and vulnerability of humans, especially the poor, to climate-related extreme events such as flash floods (SDG 1), since 80% of the JRB watersheds are vulnerable to erosion and flash floods and thus require management approaches that will progressively improve land and soil quality and subsequently ensure sustainable food production (SDG 2). Promotion of controlling factors in high erosion and flash flood-prone areas will substantially increase water use efficiency across these areas and would ensure sustainable withdrawals and supply of freshwater to address water scarcity challenge (SDG 6). The management procedures when applied to highly susceptible zones as recognized by the present work would significantly reduce the number of deaths and affected people besides considerably decreasing the direct economic losses from fluvial floods (SDG 11). Nature-based solutions may further support the region to attain other SDGs, such as protecting, restoring and promoting sustainable use of terrestrial ecosystems, especially forests, halting and reversing land degradation, and mitigating biodiversity loss (SDG 15) and combating climate change and its impacts (SDG 13).

Conclusions

Erosion and flash flood susceptibility index (EFSI), a new index, was determined through PCA-based weightage scheme using morphometric attributes and slope classes of 18 major stream watersheds of JRB. More than 80% of the JRB watersheds fall under high to moderate categories, thus highlighting that watersheds are vulnerable to erosion and flash floods and thus require proper management approaches. Engineering infrastructure like formation of reservoirs and check dams for water storage is developed in high- and moderate-slope areas such as Baramulla and Uri districts in Lower Jhelum watershed, Bandipora in Madhumati and Erin watersheds, Pahalgam area occupying Lidder watershed, Sonamarg, Kangan and Ganderbal areas located in Sindh and Kokernag area situated in Bringi watershed. Moreover, such infrastructures should also be developed for areas falling in Vishav, Pohru, Kuthar, Arapal and Sandran watersheds. In addition, construction of stable dykes, contour farming practices and desiltation of downstream rivers and spill channels must be prioritized. Such information rather is required by watershed managers, city/town planners and government authorities to prevent and mitigate the deleterious effects of such hazards instigating processes on ecosystems, life and property.