The rivers are the lifeline for the survival of mankind on earth. The river operates its normal function of erosion, transportation and deposition in a basin with respect to a base level. Rivers carry out their lives eroding valleys, transporting sediments and forming fluvial sculptures (Ashmore, 2015; M. Singh et al., 2020). The hydro-geomorphic functioning of a river has a potent capability of bringing changes in flow character, erosional-depositional architectures, sediment and ecological regimes (Das, 2019; Hupp et al., 2009; Sinha & Ghosh, 2012). Floodplains and alluvial rivers have been historically and still are one of the most attractive places on Earth for human life and agriculture activities (Allan, 2004; Boix-Fayos et al., 2007; Gordon & Meentemeyer, 2006). The constructions across the rivers bring potential threats to the overall river behavior (M. Biswas & Banerjee, 2018; Tarolli & Sofia, 2016). It increases the chance of sudden changes in the hydraulic geometry of cross sections, depositional character and most importantly the nature of human dependency on rivers (Ortega et al., 2014; Rudorff et al., 2014; Tipa, 2009). In anthropocene era, the man-made changes are inviting large scale devastations within the fluvial systems (Downs & Piégay, 2019; Wohl, 2020). River channel changes, such as bank erosion, down cutting and bank accretion, are natural processes for an alluvial river (Kummu et al., 2008).Bank erosion and channel migration can occur on different timescales (Ahmed & Fawzi, 2011; Bhattacharya et al., 2019; Dotterweich, 2013; Hooke, 2007; Simon & Collison, 2002). Riverbank erosion is not only influenced by climate change, amount of water discharge, type of soil, hydrological and physiological variation, but also anthropological activities, such as different construction along river, dam construction on river, land-use change, etc. (Golfieri et al., 2018; L. Li et al., 2007; S. Roy & Sahu, 2016). River bank erosion is a form of lateral channel expansion as a response to variations in fluid flow and sediment discharges. Lateral migration is therefore a process that can cause catastrophic local or regional changes (J. D. Das et al., 2007; J. D. Das & Saraf, 2007; Pati et al., 2008; Philip et al., 1989; Sinha et al., 2005; Thakur et al., 2012). River Bank shifting is often related to the event where one bank is forming and the opposite bank is eroding (Gazi et al., 2020). Observing the spatiotemporal morpho-dynamics of a river is very crucial to comprehend the river behavior (Leh et al., 2013).The river Ganga forms a marshy flood plain along its course after entering Malda district. As the river passes through different geomorphic characters in the deltaic region, the river repeatedly adjusts itself which leads to erosion and deposition (Nabi et al., 2016). This region is the main playfield of the river and evolved by the successive sedimentation over years and known as Diara. The region covers 1,99,493 hectares and is the result of river laden siltation in the moribund beds of the older Ganga or Mar Ganga (Debanshi & Mandal, 2014). The prolonged sedimentation and consolidation process results in to the formation of Bhutni Island in north-western part of Malda. The tract is supposed to turn into waterlogged flat marshy land considering the old Ganga-Kalindri piracy and mega flood events (Ghosh & Kar, 2018; Mehebub et al., 2015; Sinha & Sarkar, 2009). Severe flood flows of the river water grind down the bank during the monsoon period whereas in the winter, sandbanks become deposited on the both banks of the river.
The Farakka barrage construction commenced in late 1960s and was commissioned in 1975. Since 1975, the upstream river channel has bifurcated into four channels which rejoin together before reaching the barrage and has started meandering in the eastern part (Majumdar & Mandal, 2020a; Mazumder, 2004). In the post barrage period, a deltaic environment suddenly appeared in the upstream part with new adjustment to the artificial base level at Farakka (Islam & Guchhait, 2017; Khatun et al., 2018). It causes the water stagnancy and resultant hydraulic pressure against the river bank along the entire reach which causes repeated river bank failure through subsidence, slumping, toe erosion etc (Rudra, 2020; Thakur et al., 2012). The seepage mechanism allows the entry of rising flood water to the banks and later is released when water level recedes gradually. This process creates many voids in the bank walls and finally causes bank slumping. The erodible sandy composition of the left bank increases risk of bank failure. Thus, the river forms large meander bend at the south of Manikchak block in Malda district. It is a gradual process of the river channel to keep pace with the excess volume of water. The river exhibits a tendency to migrate to east. As a result, many ‘Charlands’ are emerged along the right bank of it (Chakraborty & Pal, 2020; Ullah et al., 2016). Geomorphologically, four channels have developed from west to east. After the construction of the barrage, sequential formation of three channels has been observed. Presently the middle channel is carrying the major flow but before 2006 the eastern most channels used to carry lion share of total river flow. The severity of bank erosion also changes with time here. After 2006, the eastern flow lost its previous flow energy and the main flow started through the central channel (Pal & Pani, 2019; Raj & Singh, 2020). It certainly reduces the high bank erosion at the entire Kaliachak-II and III Blocks (B. Das, 2011). So at present erosion is occurring in that part of left bank of Ganga where main flow through middle channel strikes the concave channel wall. The increased potential energy of the river due to rising water level to the upstream of Farakka barrage is the main cause of accelerated flood and subsequent river bank erosion through meander migration and stream avulsion as the flow has been partially obstructed by the barrage (Mandal, 2017; Thakur et al., 2012). So, bank erosion due to east ward expansion of the channel within the meander belt remains a problem for the inhabitants reside along the left bank. Many scholars from the field of geomorphology have already drawn the attention of international readers considering this prolonged history of river bank erosion in Malda from various angles viz. (i) using Bank Erosion Hazard Index (BEHI) & BANCS model (Majumdar & Mandal, 2020a; Ullah et al., 2016), (ii) using PAR model (R. Biswas & Anwaruzzaman, 2019), (iii) using planform and stream power indices (Majumdar & Mandal, 2020b) and (iv) erosion and LULC dynamicity models (Mandal, 2017; Mondal et al., 2016). In this paper, the authors tried to tie up all the existing research gaps especially the construction of a lucid and holistic model on identification of bank erosion potentials sites after 2020 considering–(i) field visits (ii) remote sensing data (iii) Geographical Information System (GIS) and (iv)regression model.
Remote sensing and GIS in combination can be used to capture empirical responses from historical and current morphological shifts in river channel problems, which would otherwise be challenging to achieve for time and land coverage purposes (Momin et al., 2020). Open-source remote sensing data that allows the rapid identification of morphological changes and how these affect river channels (Langat et al., 2019). Remote sensing and GIS techniques are widely used for the quantification of river bank erosion and change detection of riverbanks worldwide. Several studies have assessed channel change by means of geospatial techniques like time series analysis of different satellite images (Bhuiyan et al., 2017; Dabojani et al., 2014; Downward et al., 1994; Gurnell, 1997; Magliulo et al., 2016; Marston et al., 1995) in diverse categories of river systems. Remote sensing provides a wide scope of spatial modeling, simulation and forecasting of these complex dynamic phenomena with greater accuracy and accountability (Nong & Du, 2011). The spatial model assists the researchers and planners to assess the contemporary issue in hydrological, eco-hydrological modeling and urban planning enables them to predict the probable consequences of spatiotemporal changes of flood discharge, landuse-cover with response to runoff, gulley erosion etc (Arabameri et al., 2018; Mondal & Mandal, 2018) and urban growth (Hamdy et al., 2016; Hou et al., 2019; Mahmoud & Divigalpitiya, 2019; Maithani, 2010; Salem et al., 2019; Sarkar & Chouhan, 2020). Some popular works on soil erosion risk assessment has been done using empirical models like Universal Soil Loss Equation (Roger et al., 2020), the Unit Stream Power Based Erosion/Deposition model (USPED) (Mitas & Mitasova, 1998) and physically based models such as the Water Erosion Prediction Project (Flanagan and Nearing, 1995) (Rawls & Foster, 1987)and the European Soil Erosion Model (Marston et al., 1995; Merritt et al., 2003). While studying bank erosion in Indian and Bangladesh context, most of the research works gave more emphasize on-channel oscillation (Bag et al., 2019; Rudra, 2014; Sarif et al., 2021), bank line shifting (Aher et al., 2012; Gogoi & Goswami, 2014; Mallick & Mallick, 2016) and morpho-dynamics of rivers within the flood plain (Gazi et al., 2020; Nones, 2021; Process et al., 2020). This paper introduces a new angle of identifying potential sites of erosion along a river using combined application of RS and GIS technology.
In this study, the authors have adopted a new technique of bank erosion potential sites identification combining remote sensing and statistical approach. The main objective of this paper is to propose a new remote sensing indices and logistic regression based method for identifying potential sites of bank erosion along the left bank. This modeling has been done by considering nine parameters: i.e., NDVI, Modified NDWI, modified SAVI, modified NDMI, distance to river, distance to settlements, population density, soil dry density, soil bearing capacity and landuse-cover (LULC) categories. The present study also has the objectives of finding the most important causative or driving variable in causing high bank erosion in recent time (2020) using BLR model based on remote sensing data. The use of remotely sensed data in the modeling approach found to be very useful because- (i) the complex role of riparian vegetation adjustment with the soil, (ii) flood inundation (iii) LULC dynamics and (iv) soil characteristics in relation to the bank erosion can easily be included while framing the model. Most importantly, the combination of remote sensing and GIS approach in the first part of the modeling has made the spatial data extraction process much easier for the analyst.