Channel bar development, braiding and bankline migration of the Brahmaputra-Jamuna river, Bangladesh through RS and GIS techniques

ABSTRACT Erosion of banks and channel migrations are usually common phenomenon and recurring geo-hazardous problem in the Brahmaputra-Jamuna river system of Bangladesh. The present study intensively analysed multi-temporal Landsat satellite imageries (Landsat MSS, Landsat TM, Landsat ETM+, Landsat OLI TIRS), historical documents (Hunter maps), and literature information employing Remote Sensing and Geographic Information System techniques, and estimation of braiding and sinuosity indices especially emphasizing channel bar development, braiding, and bank-line migration of the river. The computed braiding indices of the river were 1.5, 10.3, 13.4, 11.0, 10.6 and 14.6 for the year of 1853–1857, 1977, 1989, 2000, 2014 and 2018, respectively (overall deciphering increasing trend). Similar to the braiding indices, the sinuosity indices (2.58, 7.02, 9.92, 8.31, 7.42 and 11.25 for the same selected years) of the river have also been changed over time. During 1853–1857 to 2018, 766.7 and 1624.2 km2 of lands extensively eroded on the east and west bank, respectively at a rate of 4.8 and 10.1 km2 yr−1. However, from 1977 to 2018, the erosion rates were 14.4 and 12.5 km2 yr−1, respectively. High discharges, sediment transport associated with random bars formation related braiding expansion, and weak bank materials are presumably the governing factors for causing such simultaneous severe bank erosion and remarkable bank-line migration phenomenon. Although, from 1853–1857 to 1977 the erosion of the river followed the river flow direction. It is expected that the outcomes of this research can contribute further research on morphodynamic characteristics of the braided river for better understanding the river characteristics, the river hydromorphology, water and sediment transport, its bank erosion, and formulating protective measurement plan for their sustainable current flow path maintenance.


Introduction
The Brahmaputra River (BR, herein after) is a classic example of an active braided trans-boundary river, and one of the largest rivers in the world.It travels through China (1625 km) and India (918 km) before crossing through Bangladesh (337 km), with a total drainage area of approximately 573,394 km 2 (Coleman, 1969;Sarma 2005).It is located in a tectonically active region (Brammer, 2012;Coleman, 1969;Fergusson, 1863;Hirst, 1915;Morgan & McIntire, 1959;Rashid et al., 2015aRashid et al., , 2015bRashid et al., , 2015d;;Rashid et al., 2015cRashid et al., , 2018b) ) and originated from the Chema Yundung glacier of Tibetan Plateau in China (Sarma 2005).It is a sandy braided fluvial system with vast terraced flood plains of different ages in the central and lower reaches.It is geomorphologically complex, traditionally unstable and highly susceptible to frequent and rapid bankline shifting, channel migration and avulsion (Best et al., 2007;Bristow, 2009;Paszkowski et al., 2021;Rashid et al., 2021).The BR basin is one of the unique basins of the world because of its location and size, density of population, and catastrophic deposition of sediments.It merges locally with the Teesta River (Figure 1) on its right (west) bank in the extreme NW region of Bengal Basin (BB) (Bangladesh) and splits into two branches such as N-S (present Jamuna River, JR, runs N-S, length, 205 km) and NW-SE (Old BR, earlier mega flow path) directed routes (Bandyopadhyay et al., 2021;Rashid et al., 2021).The Old BR runs along the eastern margin of the uplifted Madhupur Terraces (MT) before joining the Meghna River in the southeast.The N-S directed route (JR) is mainly a sandy braided New BR (Goodbred & Kuehl, 2000;Morgan & McIntire, 1959;Sarker & Thorne, 2006).It (finally enters Bangladesh at Chilmari, Kurigram) confluences at Goalunda Ghat, Rajbari with the other giant river Ganga (locally regarded as Padma River) on its way to the Bay of Bengal (BoB) (length: ∼250 km).It is surrounded by Pleistocene sediments exposed in uplifted Madhupur (east) and Barind (west) Terraces and flowed through the central part of Recent New BR and the Teesta floodplains (composed of loosely compacted sand, silt and clay).The terraces are mainly composed of silt, silty sand, clayey silts and sticky silty clays, where reddishbrown layers of oxidized mud and fine-grained sand matrix with an elevation varying from 10 to 48 m above mean sea level.The uplifted areas appear to be associated with neotectonism, which acted to change the course of many tributary channels during the Pleistocene period.The Old BR was the main flow of the BR until the late eighteenth century, before the main channel of the BR shifted to the current Jamuna channel, associated with uplift of the Madhupur (Goodbred & Kuehl, 2000;Morgan & McIntire, 1959).In the past, it is worth to be noted that the JR was a very narrow meander channel before connected with the New BR mega flow through avulsion i.e. the current JR flow (Bandyopadhyay et al., 2021;Rashid et al., 2021;Rennell's, 1783;Sarker, Thorne, Aktar, & Ferdous, 2014).Then the JR has frequently been altered its thalweg position, modified its morphology, river dynamics, hydrogeological conditions, etc. and simultaneously facilitated to cause severe erosion and bankline migration.
Several studies have been heavily investigated, discussed and reviewed the avulsion of the JR into its current course and recommended several causes.They concluded that westward course migration of the JR caused in response to neotectonics (heighten of the Pleistocene Madhupur Tract) (Burger et al., 1988;Coleman, 1969;Morgan & McIntire, 1959;Winkley et al., 1994); disastrous floods (La Touche, 1910); regular fluvio-morphological processes (Best et al., 2007; Bristow, 1999;ISPAN, 1993;Sarker, 1996;Thorne et al., 1993).It has been migrated ∼10 km westward at a rate of ∼70 m yr −1 over the past 150 yrs (Coleman, 1969).It has devoured about 20-50 km 2 yr −1 of the mainland and caused homelessness of nearly 20 thousand families (Bristow, 2009).In addition, it also badly affects crops, loses huge productive in valuable agricultural landmasses, homestead land, forest, etc.; properties and infrastructures like bridges, culverts, buildings, etc. (Bandyopadhyay et al., 2021;Best, 2019;Best et al., 2007;Rahman, 2013;Rashid et al., 2021).The question regarding the westward movement of the JR appeared from the outer advancement of both banks of the river since the 1970 s.In the period of 1973-1992, the rate of eastward movement of the left (east) bank was 79 m yr −1 while the rate was 68 m yr −1 for westward movement of the right (west) bank (Best et al., 2007).Brice braiding index (BI) for JR was 4-6 (FAP24, 1996a).Sarker et al. (2014) stated that the average BI of the JR was between 2.2 and 2.3 from the late-1960 s to the early-1980 s, but after that BI has increased quickly, reaching a value approaching to 2.8 in 1990.The BI then alleviated at this higher value until the mid-1990 s.Then, the BI decreased to about 2.5 until 2005.However, Sarma (2005) and Sarma and Acharjee (2018) reported the average BI of the upper reach of the (BR) the same river in India (Assam valley) were 6. 1, 8.3, 8.7, 6.67, 6.58 and 7.70 for the years of 1912-1928, 1963-1975, 1996, 2000, 2007 and 2009, correspondingly which are notably much higher than the value reported by Sarker et al. (2014) at the lower reach of the JR in Bangladesh.Li, Lu, Gao, You, and Hu (2020) noted that the BI of the braided rivers is spatially identical.Similar findings were also reported in various low-altitude braided rivers in different regions of the globe (An et al., 2013;Kelly, 2006;Sambrook, Ashworth, Best, Woodward, & Simpson, 2005;Sapozhnikov & Foufoula-Georgiou, 1996).Hence, it essentially needs to reinvestigate this anomalous scenario and attract author's attention to resolve this issue for a braided river for exploring the hidden fact concerning severe bank erosion scenario, migration, avulsion, etc. phenomena as well as evaluate the current state of the river for considering future effective defensive measures.The river has supported many civilizations throughout the history and continues to play a vital role for transportation, supplying precious water for drinking, irrigation and industry to the people of the river region.In addition, due to the continuous increasing necessitate for natural scenery sightseeing, river marshland security, mineral resource improvement and travel trade buildings, plainimetric and morphodynamic characteristics of the river have attracted much universal attention particularly to the scientists, e.g.fluvial geomorphologists, sedimentologists, river engineers, researchers and academics (Ashmore et al., 2011;Bakker et al., 2019;Connor-Streich et al., 2018;Nicholas, 2013;Peirce et al., 2018;Sun et al., 2015b).So, the current research is inevitable to assess the current state of the river for its sustainable management, better understanding the existing appropriate advanced knowledge behind related to bank erosion and associated other aggressive events commonly caused by the river.As the neighbouring people of the river area are helpless due to undesired intense erosion or avulsion but no adequate protective strategic measures of the river have yet been adopted to reduce the people's countless sufferings.
Many scholars around the globe conducted research on morphological characterization of big braided rivers through Geographic Information System (GIS) and Remote sensing (RS) techniques (Coleman, 1969;Gilfellon et al., 2003;Goswami, 1985;Islam et al., 2017Islam et al., , 2019;;Khan & Islam, 2003;Langat et al., 2019;Lewin & Ashworth, 2014;Li et al., 2020;Mclelland et al., 2009;Negm et al., 2016;Nones, 2020;Richardson & Thorne, 1998).Morphological characteristics of any river including bar growth, the interconnection among the number of bars, bars area, river width, BI, sinuosity index (P T ), etc. are very important variables for characterizing frequent, gradual, and rapid bankline migration and erosion (Ashmore, 1991;Bertoldi & Tubino, 2007;Chalov & Alexeevsky, 2015;Egozi & Ashmore, 2008;Rashid, 2020;Sarma 2005).Therefore, the present research is an attempt to delineate the morphological characteristics and their relation to bankline shifting and avulsion of the JR by using RS and GIS techniques.The prime objectives of the current work are to (1) delineate channel bar formation of the river; (2) estimate the braiding and sinuosity indices; (3) characterize the erosional and depositional distribution pattern of the river; (4) investigate the bank line migration pattern and (5) evaluate the probable causes of river bank erosion, course migration/channel shifting.It is profoundly expected that the obtained information of this article may be useful to future protective plans for sustainable development of the surroundings for setting up an appropriate routine erosion maintenance approach and the progress of suitable supervision plan for defending riverine marshlands affected by overstocking, local engineering structure, and increased runoff deviation aggravated by climate change.

Study area, hydrology, and sediment supply
Bangladesh is a flat, humid subtropical, and mainly agricultural-based densely populated country (population: ∼210 million; density: ∼1265 per km 2 ) (Anon, 2021), occupying the major part of the tectonically active BB (Morgan & McIntire, 1959;Rashid et al., 2015aRashid et al., , 2015bRashid et al., , 2015dRashid et al., , 2018b) ) and comprises of dominantly alluvial-fluvial plains (∼80%) with riverine environments, some hilly (∼8%) and coastal (∼12%) areas (Figure 1).Geologically, the basin is located in a site of intense tectonic activity (Coleman, 1969;Morgan & McIntire, 1959).The average annual rainfall varies from 140 to 390 cm and about 80% of the total rainfall occurs during the summer monsoon period.Each of the Ganges-Brahmaputra-Meghna fluvial systems in the BB is strongly influenced by the wet summer monsoon, when up to 80% of the annual flow and more than 90% of the yearly sediments are discharged into the BoB (Goodbred & Kuehl, 2000).The BR receives enormous discharges and bed loads (mean annual discharge is ∼ 678,000 cusecs and ∼ 1.84 billion tons of sediment) from Chema-Yung-Dung glacier mass and the rising Himalayan Orogenic eroded belt (EGIS, 1997;Goswami, 1985;Milliman & Robert, 1983), and is characterized by high-velocity turbulent flow (Coleman, 1969).The average discharge of the JR is about 20,200 m 3 s −1 at Bahadurabad (Jamalpur) (see Figure 1 for location) with a minimum of dry period of about 2860 m 3 .s−1 (EGIS, 1997).However, it reached to about 100,000 m 3 .s−1 in the catastrophic flood of 1988 (EGIS, 1997).Bank-full discharge is hard to estimate since the overbank level and channel bank edge are ill-defined, but estimates vary from 45,000 to 60,000 m 3 .s−1 (FAP24, 1996d; Thorne et al., 1993).The average water slope of the river is 0.000076 over the first 130 km and more downstream, it is 0.000065 (FAP24, 1996b).The depth-averaged velocities can attain more than 3.5 m s −1 (FAP24, 1996c).Water level showed minimum values during dry season from November to April (Figure 2).Due to Himalayan snow melt in May, the hydrograph start to rise and it reaches its peak position between July and September during monsoon rainfall (Figure 2).The yearly hydrograph shows an annual variation in water level of about 6 m (FAP24, 1996a).

Data sources
Multi-temporal Landsat satellite images (Table 1) representing 1977, 1989, 2000, 2014, and 2018 were accumulated from the website 'http://glovis.usgs.gov'selecting the arid period (winter season/post-monsoon) for easily identifying exposed bars and utilized for this study having 30 m resolution for all, with the exception for Landsat MSS image of 1977 (60 m).The braiding intensity varies due to fluctuation of water levels (Li et al., 2020).Therefore, dry periods images are suitable for analysis such as ground-water boundaries, channel bar statistics, BI, and P T (Gan et al., 2013;Rashid, 2020;Rozo et al., 2014), because bars are exposed and easily identifiable during the post-monsoon period due to low water level (Figure 2).However, there is a variation in the temporal range between the considered periods (1977-2018) because cloudless images are not available at regular intervals.To contest the spatial resolution of Landsat TM/ETM+/OLI images, the Landsat MSS image of 1977 was comparable to 30 m employing the nearest neighbour method by Erdas imagine 2010.WGS 84 UTM Zone 45N projection system was considered for georeferencing of images.Moreover, with the recent satellite images, the historical documents (maps) of 1853-1857 (Hunter, 1876) were also employed.Previous scholars were also used these documents (Bandyopadhyay et al., 2021;Best et al., 2007;Rashid et al., 2021;Sarker et al., 2014).The coordinates system of the 1853-1857 maps (mapped on Everest spheroid) was transformed to the WGS-84 coordinates by ArcMap 10 to be compatible with the Landsat images (WGS 84 projection system) following Srivastava and Ramalingam (2013).

Delineating river course, bankline, and channel bars
The visual image analysis as well as digital image processing techniques (Agarwal & Garg, 2000;Islam et al., 2011;Alam & Islam, 2017;Lazaridou, 2012;Rashid, 2020;Rashid et al., 2018aRashid et al., , 2021;;Yanli, 2002) were applied to characterize the river route, its erosional and depositional outline, channel bar expansion, and frequent bank-line changing.Identifying the land-water boundary using Landsat image is a difficult job.To trounce this condition, an arrangement of various Landsat bands (Alam & Islam, 2017;Yang et al., 1999) and a soil-vegetation limit approach (Alam & Islam, 2017;Gurnell, 1997) were applied to recognize the landwater boundary.Earlier studies (Alam & Islam, 2017;Winterbottom, 2000;Yang et al., 1999) imply that these approaches are highly useful to differentiate river planform, resulting from varying water levels inside a channel.The channel bars and bank-line were delineated considering their dissimilar tone, texture, association with the river, its form, and its relationship with the river (Alam & Islam, 2017).The point mode technique was employed to digitize the images using ArcMap 10.
2.4.Calculation of number of bars, bars area, braiding and sinuosity indices, erosional and depositional rates, river width, and mutual relationship and association analysis Measurement of erosion and deposition pattern of both banks, polygons of two particular yrs were taken (superimposing) utilizing the union technique of analysis tool of Arc-Map 10.The braiding intensity (Table 2) was calculated following Germanoski and Schumm (1993) 1 for location), averaged over a 59-year period  and level is articulated relative to a standard low-water datum derived from long-term records (Data Source -Anon, 2020;Best et al., 2007;Sarker & Thorne, 2006).the extent of the entire bars, L r (Figure 3(a)) indicates the extent of the reach calculated along the centre procession of the river and N b is the entire amount of bars per reach length.The channel length ('sinuosity') index was estimated (Table 2) following Hong and Davies (1979) recommended Sinuosity Index, P r = L L /L r where L L (Table 2; Figure 3) is the whole length of channels.Statistical analysis was performed among river area, N b , bars area, average widths, BI, and P T using SPSS edition 20.Pearson association matrix was chosen for the mutual association and involvement analysis to categorize the scale of involvement of diverse constraints.The river widths from both banks of the channel at different times of dissimilar cross sections were estimated vertically to the opposite bank lines at the matching permanent geographical points employing the measure tool in Arc-Map 10 (Rashid, 2020;Rashid et al., 2021).Eight cross sections were constructed in different fractions of the river path and analysed with reasonable discussions (Figure 7).

Change of river area and channel bar formation of the JR
In braided rivers, there is a mixture of distinct channel bars, namely central bars, side bars, mid-channel bars, and islands of dissimilar dimensions that direct to the break and scatter in general planimetric morphology, so the twigs, chutes, and lobes are arbitrarily split and rejoined (Ashmore, 1991;Bertoldi & Tubino, 2007;Chalov & Alexeevsky, 2015;Li et al., 2020;Nicholas, 2013;Peirce et al., 2018;Rashid, 2020).Based on these inconsistent patterns, strong intra-and inter-annual erosional and depositional progressions were taking place during the flood and non-flood events which had increased the inconsistency of braided morphological uniqueness (Bertoldi et al., 2009;Leopold & Wolman, 1957;Li et al., 2017;Rashid, 2020;Reineck & Singh, 1980;Williams & Rust, 1969).These bars are randomly found in the JR (Figure 3).Some of them are small and others are large bars.The area of the individual bar ranged from < 1-125 km 2 .The large bars are elongated and pointed downstream.On the other hand, the small bars or newly developed bars are semi-circular and also pointed downstream.
During 1853-1857, the total number of bars of the river was only 31 (Table 3; Figure S2a).About 120 yrs later in 1977, the number of bars significantly increased (total number of bars = 299), where both the mean area of bars and the associated standard deviation were reduced.The number of bars of the river was dramatically increased within a short period of time (from 1977 to 1989), where both the mean area of bars and standard deviation were further declined.Reasonably it can be assumed that the declining bar areas and associated standard deviation were occurred owing to the unexpected episode major flooding event during 1988 (Dewan et al., 2003) in the river course.Major floods significantly affect river hydro-geomorphology, sediment load, and hydrodynamics because of discharging a huge volume of water and sediments with relatively higher velocity (Ashmore, 1988;Egozi & Ashmore, 2009;Rashid, 2020;Schuurman et al., 2016;Williams et al., 2016).As a result, existing bars face further erosion due to the high velocity of the water and simultaneously many new bars were also developed by the excessive sediment loads.Conversely, during normal flooding season (wet season/monsoon), the water velocity of the river is relatively low and the existing bars experience less erosion.Therefore, the sizes of the existing bars are generally increased instead of diminution.In many cases, adjacent bars are connected through siltation in connected channels.Therefore, the numbers of bars were increased after experiencing a major flood event, however the number of large bars were reduced.On the other hand, the number of bars decreased at normal flooding season but the number of large bars expanded.Consequently, in 2000 (during the normal flood time), the number of bars decreased but the number of large bars expanded.It is observed that various small bars were joined together and transformed into large bars and concurrently many small bars were extinct (Figure 4) by severe erosion.An alike trend was also observed in 2014.An opposite scenario was observed in 2018 while the number of bars, bar's means area and associated standard deviation were dramatically shortened up.It is possibly caused owing to another major flood event in 2017 (Anon, 2017) in the affected region.Therefore, the aforementioned observations suggest that the major flood events have been created an additional effect on the new bar formation, severe bank erosion, channel shifting, avulsion, and morphological and configuration changes of channels and bars of the river.
The analysis invokes that the overall area of the river and the bars area have progressively been increased at a rate of about 6.1 and 10.3 km 2 yr −1 , respectively over the past 16 decades (Figure S1a; Figure S2b).Though, during the period of 1853-1857 to 1977 the increasing rates were 3.0 and 9.9 km 2 yr −1 , respectively (Figure S1b, S1e).Then the areas were rapidly increased at a rate of about 14.7 to16.0 km 2 yr −1 (Figure S1c, S1f).Although, the overall area of the river was expanded through increasing channel bars area, forming many new bars, accumulating numerous small bars, expanding of individual bars configuration, whereas the water area was reduced at a rate of about 4.2 km 2 yr −1 over the past 16 decades (Figure S1d).

Erosion and deposition of the JR
Bank erosion, channel migration, and subsequent deposition are very common and familiar characteristics of the JR (Figure 5; Bristow, 1987;Sarker et al., 2014;Hassan et al., 2017).Within the period of 1853-1857 to 1977 about 175 km 2 in the east bank and 1113 km 2 in the west bank of the corresponding river's floodplains were eroded with a rate of about 1.5 and 9.3 km 2 yr −1 , respectively (Table 4; Figure S4).However, from 1977 to 1989, approximately 250.8 (east) and 207.5 km 2 (west) of landmasses were lost at a rate of about 20.9 and 17.3 km 2 .yr−1 , respectively.The analysis invokes that the erosion rates were significantly increased on both the river banks, about fifteen folds at the east bank and nearly two folds at the western bank during 1977 to 1989 relative to the period of 1853-1857 to 1977 (Table 4).From 1989 to 2000, the calculated area of the lost landmasses were of east-and west-banks about 176.9 and 159.6 km 2 at a rate of about 16.1 and 14.5 km 2 yr −1 , respectively (Table 4; Figure S4).The rates were decreased on both banks of the river.From 2000 to 2014, the rates were then become nearly half of the just immediate earlier period on both banks.About 126.4 and 133.3 km 2 of landmasses were eroded on the eastand west-bank, respectively at a rate of about 9.0 and 9.5 km 2 yr −1 (Table 4; Figure S4).Accordingly, from 2014 to 2018, the erosion rate was considerably decreased on the west bank of the river, while on the east bank the erosion rate was similar to the period (2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014).About 37.6 and 10.8 km 2 of landmasses were lost on the east-and west-bank, correspondingly at a rate of about 9.4 and 2.7 km 2 yr −1 (Table 4; Figure S4) indicating the frequently variations in migration trend.The overall observation implied that the erosion followed the similar trend as those of BI and P T (Tables 2-4).However, before the 1970 s the erosion (migration) of the river was followed the river flow direction (Figure 6(a)) and it was more prominent on the west bank.Based on flow direction, the river was partitioned into four sections A, B, C, and D (Figure 6).In section A (Figure 6(a)) (Kurigram to Fhulchari), the river was flowed from the northeast to the southwest and consequently eroded on the west bank.In section B which commences near Fhulchari and end at Sariakandi, the river was passed through almost north to south and concurrently eroded on both banks.And then, in section C (near Sariakandi to Kazipur), the river changed it's flow direction and discharged northwest to southeast occurring intense erosion simultaneously only on the east bank.In section D (near Kazipur to near Bera), the river was again flowed from northeast to south-west and eroded on the west bank.Finally, over the past 16 decades, around 766.7 (east) and 1624.2 (west) km 2 of landmasses were eroded by the JR, with a rate of 4.8 and 10.1 km 2 .yr−1 , respectively, (Table 4; Figure S4).The erosion rate of the river was nearly two times higher at the west bank compared to the east bank.However, noted that over the past 41 yrs , some 591.7 (east) and 511.2 km 2 (west) land area were washed away with a rate of 14.4 and 12.5 km 2 yr −1 , respectively (Table 4; Figure S4).It is also noticed that the erosion rates were significantly increased on both banks of the river after 1977.Noticeably, in the east bank erosion was more prominent and remarkably enhanced during the considered time scale.Though, the erosion rate has been considerably declined on both banks of the river since 2000 (Table 4).The observation also reports that the rate of erosion was varied over time suggesting the response of the widening process of the channel. Oer the past 16 decades (1853-1857 to 2018) about 997.6 and 210.1 km 2 lands have also been accreted on the east-and west-banks, with a rate of 6.2 and 1.3 km 2 yr −1 (Table 4), respectively.

Width of the river
Eight cross-sections (aa', bb', cc', dd', ee', ff', gg', and hh') were drawn, interpreted, and analysed in various reach of the JR course (Figure 7).In general, the investigation suggests that the average width of the river followed a rising trend with a rate of approximately 25 m yr −1 over the study period (Figure S5).Though, from 1853-1857 to 1977, it is observed that the average width of the river has considerably enlarged with a rate of around 4 m yr −1 (Figure S6a).And it was then significantly and rapidly increased with a high rate of about 95 m yr −1 (Figure S6b).Presently, the average width of the river is about 11.9 km (Table 5) which is very consistent with the other findings (Sarker et al., 2014).

Correlation coefficient analysis
The research (Table 6) discloses that the BI is considerably associated with the number of bars (r = 0.941, α ≤ 0.01).Related to BI, P T is also significantly correlated with the number of bars (r = 0.955, α ≤ 0.01).The BI and P T indices are also appreciably and positively linked (r = 0.982, α ≤ 0.01) suggesting that both indices have a strong linear relationship.Therefore, more bars generate more braiding and more sinuosity that finally causes further bank erosion.As, erosion phenomenon is an intrinsic feature of a braided river (Ashmore et al., 2011;Hundey & Ashmore, 2009;Rashid, 2020;Reineck & Singh, 1980;Sapozhnikov & Foufoula-Georgiou, 1996).The number of bars, BI, P T , and erosion rates of the river also support this statement (Tables 3  and 4).The study also indicates that the stream width and bars area of the river is significantly and positively correlated (r = 0.843, α ≤ 0.05) (Table 6) which reflects that the stream width is greatly controlled by the bars area.The bars area and the number of bars are also significantly linked (r = 0.610, 4. Number of bars in different times at different cross sections (See Figure 7 for cross section locations).α ≤ 0.05).The analysis also reveals that the river-and bars-area are remarkably correlated (r = 0.923, α ≤ 0.01) (Table 6).It is concluded that the river area is greatly controlled by the bars area.

Discussion
Severe river bank erosion is essentially concomitant with the undesired natural geo-hazards, and considered as the primary factor of frequently and rapidly occurred bank-line  migration of the JR banks.After changed the earlier course (Old BR) of the BR and followed the course via the JR, which has gradually been changed from narrow to wide becoming braided type (Figure 3).The obtained results reveal that the average width of the investigated JR was progressively enhanced over the past 16 decades (Table 5).The study further represents that the overall area and bars area of the river was also increased over the same period, whereas the total water surface area was reduced.
The computed BI index of the JR were ranged from 1.5 to 14.6 (mean 10.24 and σ = 4.59) and P T index were varied  Sarma (2005), Sarma andAcharjee (2018) andFAP24 (1996a).In general, the BI is spatially identical in a certain period and a certain segment in various low-altitude braided rivers in different region of the globe (An et al., 2013;Kelly, 2006;Li et al., 2020;Sambrook et al., 2005;Sapozhnikov & Foufoula-Georgiou, 1996).The BI and P T indices values of the JR were also varied over time.The rates of erosion of the JR were not identical over the time spans on both the banks as well as in diverse parts of both banks of the investigated JR.However, erosion followed the same trend as BI and P T , and consequently (bankline) migration.Though, before the 1970 s river bank erosion (migration) followed the same direction of river flow, and westward bank erosion was more prominent.It is found that the computed erosion rates in the studied segments were 1.5 and 9.3 km 2 yr −1 in the left (east) and right (west) banks, respectively from 1853-1857 to 1977 (Table 4; Figure S4).The erosion rate of the river was nearly two times higher at the west bank compared to the east bank.During 1977-2018, these then were 14.4 (east) and 12.5 (west) km 2 yr −1 , respectively (Table 4; Figure S4).The analysis identifies that the erosion rates were significantly increased on both banks of the river after 1977 i.e. 14.4    (east) km 2 yr −1 and 12.5 (west) km 2 yr −1 .This observation is consistent with Sarker et al. (2014) study.Authors documented that from 1973 to 2010, these calculated results were 13.4 and 10.1 km 2 yr −1 in the east and west bank of the JR, respectively.Sarma (2005) reported that from 1912-1928 to 1965-1975 and 1963-1975 to 1996, these were 11.9 (east) and 12.4 (west) km 2 yr −1 , and 11.2 (east) and 13.9 (west) km 2 yr −1 , respectively, in the BR in Assam India.Sarker et al. (2014) reported that in 1830, the JR was a meandering in nature until 1914.The morphology of the river dramatically changed in the mid-1900s and transformed to a braided form (Figure 3; Table 2).Sarma and Acharjee (2018) reported that due to the significant impact of the 1950 great earthquake caused in India (Assam), leading to the release of excessive quantities of sediments (a huge influx of clastic debris) from the Himalayan highlands, and concurrently accumulated in the depository basin i.e. the Jamuna Valley between the late-1960 s and the late-1990 s.Prior to the 1787 Assam flood (Brammer, 2012), the Old BR was the main channel of BR; since then the river has shifted its course westward along the Jhenai and Konai rivers to form the broad, braided Jamuna channel (New BR) (Bandyopadhyay et al., 2021;Rashid et al., 2021).It is postulated that giant seismo-event directly and greatly impacted on the regional geology, geomorphology, hydrogeology, sedimentation, river morphology, denudational pattern, morphodynamics, flow pattern, and path of the river.In addition, local lithological, geotechnical, hydro-geomorphological characteristics, neo-tectonics (Brammer, 2012;Coleman, 1969;Fergusson, 1863;Hirst, 1915, Morgan & McIntire, 1959;Rashid et al., 2015aRashid et al., , 2015bRashid et al., , 2015cRashid et al., , 2015dRashid et al., , 2018b)), engineering constructions, climatological and meteorological phenomena (Yu et al., 2010;Zhu et al., 2008), etc. also triggered the prevailing situation i.e. erosion and avulsion processes.Many new bars of diverse scales have been formed inside the JR (Table 3), and the river then progressively transformed to braided type.These processes probably influence on BI and P T values and represents larger values (Figure 3; Table 2).Li et al. (2020) noted that enormous sediment loads have a considerable effect on morphological transformations and unsteady vigorous conditions of the braided channel.These intricate connections among the morphology of flow, silt and braided rivers in various parts of the globe were also reported in provisions of qualitative study (Ashmore et al., 2011;Hundey & Ashmore, 2009;Sapozhnikov & Foufoula-Georgiou, 1996), flume tests (Bertoldi & Tubino, 2007;Egozi & Ashmore, 2009;Javernick et al., 2018;Peirce et al., 2018), and morphodynamic modelling (Schuurman et al., 2016;Sun et al., 2015a;Williams et al., 2016).Bank erosion is an important factor for braiding (Leopold & Wolman, 1957;Rashid, 2020).Braiding is familiar if high discharge available as well as weak banks prevailed (Rashid, 2020;Reineck & Singh, 1980;Williams & Rust, 1969).It is worthwhile that the bank's sediments of the JR are non-cohesive (loosely compacted silty sand) which have low stability and high erodibility (Coleman, 1969;Thorne et al., 1993).Therefore, high discharge (EGIS, 1997), abundant sediment supply (Goswami, 1985;Milliman & Robert, 1983), frequent large scale flooding, and low threshold of bank collapse (Coleman, 1969;Thorne et al., 1993) ultimately triggered the development of more new bars, more braiding, and consequently, erosion rates remarkably increased on the both banks of JR after the 1970s (Tables 2-4).Recent studies have also shown that the intensity of braiding is positively associated to discharge, stream supremacy, and bedload (Ashmore et al., 2011;Bertoldi et al., 2009;Li et al., 2020;Peirce et al., 2018).Hence, sediments in the JR are not only deposited in millions of tons and consequently it changes to more braided but are also highly susceptible to erosion.
However, after the 2000s the rates of erosion and BI values were remarkably declined on both banks owing to undertaking some protection measures (i.e.engineering intervention) against erosion and declining of 1950 earthquake related events (Sarker et al., 2014).Though, before the 1970 s the erosion (migration) of the river occurred along flow direction of the river (Figure 6(a)), prominently on the west bank.During that time the river was also nearly meandering type (Figure 3; Table 2).Therefore ongoing erosion and deposition processes on both banks of the studied river are mainly influenced by the morphology of the river (meandering/braiding), characteristics of bank matrices, sediment input, discharges, and flow direction.Goodbred et al. (2003) apprehend the effects of climate change (Yu et al., 2010;Zhu et al., 2008) and further potential seismic events as it lies in a tectonically active region (Brammer, 2012;Coleman, 1969;Fergusson, 1863;Hirst, 1915, Morgan & McIntire, 1959;Rashid et al., 2015aRashid et al., , 2015bRashid et al., , 2015cRashid et al., , 2015dRashid et al., , 2015aRashid et al., , 2018b)), may accelerate excessive sediment supply to the river.Consequently, further modifications would have occurred off the river for its adjustment.Finally, it could enhance more suffering to the nearby communities and create socio-economic problems (Rashid et al., 2021).Hence, for future planning and sustainable development of the area, careful systematic research and close regular monitoring the characteristics of the rivers are immensely important.The output of the research would be helpful to better understand the characteristics of the studied gigantic river.The obtained information may also be considered for river bank erosion and sustainable protection strategies, river engineering, facilitate plan makers in order to take required appraises for decreasing the erosional harshness and further future research and a basis for proper sustainable river basin management so as to reduce the people's countless sufferings and to protect invaluable public wealth, properties and other resources.

Conclusions
The current research emphasizes channel bar enlargement, braiding, and bank-line relocation of one of the largest and complex braided rivers in the globe.The BI of the JR has been changed over time.The P T of the river has also been changed with time and followed the same trend as BI.The erosion rates were not continuously the same on both the banks as well as at diverse parts along both banks of the channel.The erosion follows the same trend as those of BI and P T .However, before the 1970s the erosion of the river followed flow direction of the river when the river was less braided or meandering.High discharge, high sediment transport associated with numerous new bars formation and erodible river banks are the major controlling factors for the expansion of braiding and consequently massive riverbank erosion.Therefore, ongoing erosion and deposition on both banks of the river are mainly controlled by the morphology of the river (i.e.meandering and/or braiding), characteristics of bank materials, sediment input, water discharge, and river flow direction.The current study would be provided a set of valuable information about morphological characterization of the river as well as bank line shifting that would be helpful to undertake future protective plans against causing intense erosion and bankline shifting and sustainable flood protection strategies of the great rivers as well as sustainable river basin management.Outcomes of this research will be helpful to any upcoming attempt for setting up an erosion supervision plan for the JR of Bangladesh.Finally, further analysis based on high-resolution morphological records and hydrological information at finer chronological scales is suggested for better understanding the morphodynamics and hydrodynamics procedures for dynamic transforms of braided rivers in the JR at changeable spatial scales in the perspective of climate change.

Figure 1 .
Figure 1.(a) Regional map showing Bangladesh, Himalayas and surroundings with location of study area; (b) representing the major physiographic regions of the Bengal Basin, drainage system, and the study area.
introduced equation i.e.Braiding Index, BI = 2 L b /L r + N b /L r where, L b (Table 2; Figure 3(a)) denotes the figure of double

Figure 2 .
Figure2.The stage-hydrograph of the JR at Bahadurabad (see Figure1for location), averaged over a 59-year period and level is articulated relative to a standard low-water datum derived from long-term records (Data Source-Anon, 2020;Best et al., 2007;Sarker & Thorne, 2006).
from 2.58 to 11.25 (mean 7.75 and σ = 2.99) during 1853-1857 to 2018 period (showing overall increasing trend).The obtained values are nearly close to FAP24 (1996a) reported values (4-6) in Bangladesh, and Sarma (2005), and Sarma and Acharjee (2018) values (6.1-8.7) in India (Assam) during 1912-2009.While Sarker et al. (2014) noted values were 2.2-2.8 from the late-1960 s to 2005 for the same segment which are far below than the presently estimated results.The obtained values from present study are concomitant with the published results of

Figure 7 .
Figure 7. Banklines and different cross-sections of the Jamuna River in different times.

Table 1 .
Landsat Imagery used in this study.

Table 2 .
Braiding and Sinuosity index, and their matrix of the Jamuna River in different times.

Table 3 .
Channel and bar statistics as well as braiding and sinuosity indices of the Jamuna River from 1853-1857 to 2018.

Table 4 .
Statistics of erosion and deposition of both banks of the Jamuna River from 1853-1857 to 2018.

Table 5 .
River width (km) of the Jamuna River at different cross sections in different times.

Table 6 .
Mutual correlation matrix of different parameters of the Jamuna River from 1853-1857 to 2018.*Correlation is significant at the 0.01 level (2-tailed); * Correlation is significant at the 0.05 level (2-tailed). *