Academia.eduAcademia.edu

Figure 3 – uploaded by Sanjay Giri

See full PDFdownloadDownload figure
Figure 3. Stage—discharge relationships showing hysteresis effect for longer-term (left) and annual observation data (right) at Bahadurabad (labeled i. in Figure 1).

Figure 3 Stage—discharge relationships showing hysteresis effect for longer-term (left) and annual observation data (right) at Bahadurabad (labeled i. in Figure 1).

Related Figures (17)

Figure 1. Study area (satellite image of January 2019, red lines show existing revetments). i. through vi. show key locations along the river system. vii. and viii. are the two major bifurcation points within the study area. ix. through xiv. are the key rivers within the study area. xv. through xvii. are the three major bridge crossings.
Figure 1. Study area (satellite image of January 2019, red lines show existing revetments). i. through vi. show key locations along the river system. vii. and viii. are the two major bifurcation points within the study area. ix. through xiv. are the key rivers within the study area. xv. through xvii. are the three major bridge crossings.
Figure 2. Satellite imagery from 2019 of the study area. Red lines on the map show ADCP mea- surement locations. Comparison between modeling and measured discharges (Q) as well as water levels (WL) at different branches and anabranches for the 2019 model are shown in the tables for each measurement location. Figure 2. Satellite imagery from 2019 of the study area. Red lines on the map show ADCP mea- 2.1.2. Numerical Hydraulic Modeling
Figure 2. Satellite imagery from 2019 of the study area. Red lines on the map show ADCP mea- surement locations. Comparison between modeling and measured discharges (Q) as well as water levels (WL) at different branches and anabranches for the 2019 model are shown in the tables for each measurement location. Figure 2. Satellite imagery from 2019 of the study area. Red lines on the map show ADCP mea- 2.1.2. Numerical Hydraulic Modeling
Table 1. Boundary conditions for verification run and simulation scenarios. 2.2. Results
Table 1. Boundary conditions for verification run and simulation scenarios. 2.2. Results
Figure 4. Water levels at Aricha (labeled vi. in Figure 1) in relation to the water levels at Bahadurabad (located on the Upper Jamuna, labeled i. in Figure 1), Sirajganj (located on the Upper Jamuna, labeled ii. in Figure 1) and the Hardinge Bridge (located on the Ganges, labeled xvi. in Figure 1). (VY) VVGALCL IN VEO AL SCALINIIG Since Aricha is the downstream boundary of the current study reach (also used in numerical models), it is important to analyze water level variation, particularly the interaction with the Ganges as it is located near the confluence of the Ganges (Padma) and the Lower Jamuna. To understand how the flow of the Ganges-Padma interacts with the Lower Jamuna, we tried to analyze the water level changes at Aricha (located on the Lower Jamuna near the confluence) in relation to the water levels at Bahadurabad (~160 km upstream of Aricha), Sirajganj (~70 km upstream of Aricha) and the Hardinge Bridge (~90 km from Aricha on the Ganges), as shown in Figure 1. The results are depicted in Figure 4.
Figure 4. Water levels at Aricha (labeled vi. in Figure 1) in relation to the water levels at Bahadurabad (located on the Upper Jamuna, labeled i. in Figure 1), Sirajganj (located on the Upper Jamuna, labeled ii. in Figure 1) and the Hardinge Bridge (located on the Ganges, labeled xvi. in Figure 1). (VY) VVGALCL IN VEO AL SCALINIIG Since Aricha is the downstream boundary of the current study reach (also used in numerical models), it is important to analyze water level variation, particularly the interaction with the Ganges as it is located near the confluence of the Ganges (Padma) and the Lower Jamuna. To understand how the flow of the Ganges-Padma interacts with the Lower Jamuna, we tried to analyze the water level changes at Aricha (located on the Lower Jamuna near the confluence) in relation to the water levels at Bahadurabad (~160 km upstream of Aricha), Sirajganj (~70 km upstream of Aricha) and the Hardinge Bridge (~90 km from Aricha on the Ganges), as shown in Figure 1. The results are depicted in Figure 4.
Figure 5. Discharge (Q) distribution at the major bifurcation, labeled vii. in Figure 1 (left), and the Salimabad channel, labeled viii. in Figure 1, located at the left branch (right) (ADCP measurement locations as well as comparison with the model result are shown in Figure 2). (a) Observed discharge distribution We assessed how (Figure 5). The results s the discharge distribution at the bifurca how that in 2017 the discharge entering the bifurcation (left plot) remained relatively consistent over most of conditions (at around 60%). The data show a 6% increase in discharge in the left branch during a high flow condition. This can be attributed to the fact that ess affected by bars and deep channels during high floods. The lef discharges through the left branch in 2012 (~79%) for discharge bankfull. This can be attributed to the presence of a single deep c branch and a milder branch angle in 2012 (despite the branch width being narrower). The result shows that in 2016 the discharge towards the left branch was significantly lower (65%) than in 2012 under similar upstream conditions. This can be attributed to the formation tions varied over time left branch of the major he upstream discharge the branch is wider and plot also shows higher conditions lower than hannel towards the left of two deep channels in 2016 conveying more flow towards the right branch. The results show that the left-branch discharge increased with increased upstream flow, particularly near bankfull conditions. This can be attributed to the lessened effect of the deep channel during bankfull flow conditions. Furthermore, the branch expanded at Chauhali due to bank erosion during 20 15-2016.
Figure 5. Discharge (Q) distribution at the major bifurcation, labeled vii. in Figure 1 (left), and the Salimabad channel, labeled viii. in Figure 1, located at the left branch (right) (ADCP measurement locations as well as comparison with the model result are shown in Figure 2). (a) Observed discharge distribution We assessed how (Figure 5). The results s the discharge distribution at the bifurca how that in 2017 the discharge entering the bifurcation (left plot) remained relatively consistent over most of conditions (at around 60%). The data show a 6% increase in discharge in the left branch during a high flow condition. This can be attributed to the fact that ess affected by bars and deep channels during high floods. The lef discharges through the left branch in 2012 (~79%) for discharge bankfull. This can be attributed to the presence of a single deep c branch and a milder branch angle in 2012 (despite the branch width being narrower). The result shows that in 2016 the discharge towards the left branch was significantly lower (65%) than in 2012 under similar upstream conditions. This can be attributed to the formation tions varied over time left branch of the major he upstream discharge the branch is wider and plot also shows higher conditions lower than hannel towards the left of two deep channels in 2016 conveying more flow towards the right branch. The results show that the left-branch discharge increased with increased upstream flow, particularly near bankfull conditions. This can be attributed to the lessened effect of the deep channel during bankfull flow conditions. Furthermore, the branch expanded at Chauhali due to bank erosion during 20 15-2016.
Figure 6. Discharge distribution for 2017 and 2019 models at the left branch (left) and the Salimabad channel (right) (ADCI measurement locations are shown in Figure 2). Similarly, we compared the computed and observed discharges for both bifurcations in 2017 and 2019 (Figure 6). While the model reasonably predicted flow division at the major bifurcation, the Salimabad division had some larger discrepancies, particularly at higher flows. The discrepancies can be attributed primarily to a lack of high-resolution data within the channel. Overall, the model performance was found to be satisfactory.
Figure 6. Discharge distribution for 2017 and 2019 models at the left branch (left) and the Salimabad channel (right) (ADCI measurement locations are shown in Figure 2). Similarly, we compared the computed and observed discharges for both bifurcations in 2017 and 2019 (Figure 6). While the model reasonably predicted flow division at the major bifurcation, the Salimabad division had some larger discrepancies, particularly at higher flows. The discrepancies can be attributed primarily to a lack of high-resolution data within the channel. Overall, the model performance was found to be satisfactory.
Figure 7. Simulated discharge (Q) distribution at the bifurcation (left) and Salimabad (right) for different discharge conditions. left) show that the discharge distribution noticeably alters for low flow (i.e., 3737 m” /s). A large portion of discharge (>90%) was found to be flowing towards the left branch in 2011, 2017 and 2018 under low flow conditions. However, it significantly dropped in 2016 and 2019. This can be attributed to the deep-channel formation and migration towards the downstream branches, e.g., a single deep channel towards the left branch in 2011 caused larger flow towards this branch. In 2016, the deep channels started to split into two, form- ing a deposition front at the bifurcation, leading to less discharge into the left branch. The deep channels became more prominent in 2017 and 2018, migrating towards the left branch and leading to more discharge in this left branch (this also caused bank erosion at the left side of the upstream channel). Deposition occurred in 2019 along with the development of a prominent deep channel towards the right branch (followed by larger erosion along the right bank of the upstream channel). This deep channel appeared to convey larger flows towards the right branch, although the resolution of the bathymetry data in 2016 and 2017 was rather low, which may cause some inaccuracies in the model prediction.
Figure 7. Simulated discharge (Q) distribution at the bifurcation (left) and Salimabad (right) for different discharge conditions. left) show that the discharge distribution noticeably alters for low flow (i.e., 3737 m” /s). A large portion of discharge (>90%) was found to be flowing towards the left branch in 2011, 2017 and 2018 under low flow conditions. However, it significantly dropped in 2016 and 2019. This can be attributed to the deep-channel formation and migration towards the downstream branches, e.g., a single deep channel towards the left branch in 2011 caused larger flow towards this branch. In 2016, the deep channels started to split into two, form- ing a deposition front at the bifurcation, leading to less discharge into the left branch. The deep channels became more prominent in 2017 and 2018, migrating towards the left branch and leading to more discharge in this left branch (this also caused bank erosion at the left side of the upstream channel). Deposition occurred in 2019 along with the development of a prominent deep channel towards the right branch (followed by larger erosion along the right bank of the upstream channel). This deep channel appeared to convey larger flows towards the right branch, although the resolution of the bathymetry data in 2016 and 2017 was rather low, which may cause some inaccuracies in the model prediction.
Figure 8. Bed topography measurements (in m +PWD) during 2011 and 2016-2019 showing changes in deep channels in upstream channel and branches (land boundary shows the planform of 2016).
Figure 8. Bed topography measurements (in m +PWD) during 2011 and 2016-2019 showing changes in deep channels in upstream channel and branches (land boundary shows the planform of 2016).
Figure 9. Bed topography measurements (in m PWD) during 2015-2019 showing changes in deep channels and bend scour at Chauhali bend and Salimabad channel (land boundary shows the planform of 2016).
Figure 9. Bed topography measurements (in m PWD) during 2015-2019 showing changes in deep channels and bend scour at Chauhali bend and Salimabad channel (land boundary shows the planform of 2016).
Figure 10. Water occurrence using JRC’s product (left) and NDWI average from raw images using MoMo (right) for 2019-2020. We made a comparison between the MoMo and Pekel’s water occurrence maps. The main difference is that Pekel computes binary water masks aggregated to the monthly time intervals, while the MoMo algorithm computes statistics of spectral water indexes without binarization. Therefore, the output is a continuous variable representing the probability of water/land presence. Our method can capture more local details (higher precision). A small example with a visual comparison of Pekel’s and our results for a small part of the Jamuna River is shown in Figure 10.
Figure 10. Water occurrence using JRC’s product (left) and NDWI average from raw images using MoMo (right) for 2019-2020. We made a comparison between the MoMo and Pekel’s water occurrence maps. The main difference is that Pekel computes binary water masks aggregated to the monthly time intervals, while the MoMo algorithm computes statistics of spectral water indexes without binarization. Therefore, the output is a continuous variable representing the probability of water/land presence. Our method can capture more local details (higher precision). A small example with a visual comparison of Pekel’s and our results for a small part of the Jamuna River is shown in Figure 10.
Figure 11. Reach-average estimation of bed levels (b,c) and scour level (d) for selected reaches outlined in the upper-lef plot (a) (the plotted bed topography data in (a) are from 2019).
Figure 11. Reach-average estimation of bed levels (b,c) and scour level (d) for selected reaches outlined in the upper-lef plot (a) (the plotted bed topography data in (a) are from 2019).
Figure 12. Sedimentation and erosion, quantified by using MoMo for different periods (green and light-blue colors indicate sedimentation and erosion, respectively).
Figure 12. Sedimentation and erosion, quantified by using MoMo for different periods (green and light-blue colors indicate sedimentation and erosion, respectively).
Figure 13. The satellite-derived differences for annual water occurrences computed using MoMo,
Figure 13. The satellite-derived differences for annual water occurrences computed using MoMo,
Figure 14. Annual, 5-year and decadal changes in eroded (negative) and accreted (positive) areas (right) for the Lower Jamuna reach (left) from 2000 to 2019 (e.g., 2019-2018 corresponds to subtraction in (2019-2020) — Lypwt (2018-2019)) for different threshold values of water occurrence frequency for the (upper layer) (0%, 25%, 50% and 75%, respectively, as shown in (upper-right plot)). areas reduce with increasing water occurrence (i.e., the upper layers). However, the pattern of variation in area changes is different in different years. For example, the changes in accreted areas during 2010-2011 for varying water occurrence is minimal. Accreted areas vary more drastically during 2011-2012 for varying water occurrence. This has to do with the dynamics of banks, chars and high-amplitude sandbars within the range of minimum water level and the floodplain. This does not reflect the morphological changes below minimum water level that are not captured by the images. (right) for the Lower Jamuna reach (left) from 2000 to 2019 (e.g., 2019-2018 corresponds to subtraction
Figure 14. Annual, 5-year and decadal changes in eroded (negative) and accreted (positive) areas (right) for the Lower Jamuna reach (left) from 2000 to 2019 (e.g., 2019-2018 corresponds to subtraction in (2019-2020) — Lypwt (2018-2019)) for different threshold values of water occurrence frequency for the (upper layer) (0%, 25%, 50% and 75%, respectively, as shown in (upper-right plot)). areas reduce with increasing water occurrence (i.e., the upper layers). However, the pattern of variation in area changes is different in different years. For example, the changes in accreted areas during 2010-2011 for varying water occurrence is minimal. Accreted areas vary more drastically during 2011-2012 for varying water occurrence. This has to do with the dynamics of banks, chars and high-amplitude sandbars within the range of minimum water level and the floodplain. This does not reflect the morphological changes below minimum water level that are not captured by the images. (right) for the Lower Jamuna reach (left) from 2000 to 2019 (e.g., 2019-2018 corresponds to subtraction
Figure 15. Annual changes in eroded (negative) and accreted (positive) areas within selected reach (left) from 2010 to 2019 (e.g., 2019-2018 corresponds to subtraction ier (2019-2020) — Layer (2018-2019)) for different threshold values of water occurrence frequency for the (upper layer) (0%, 25%, 50% and 75%, respectively, as shown in (upper left plot)). than erosion during 2014-2015, 2016-2017 and 2017-2018, whereas more erosion occurred in 2015-2016. As can be seen from the variation in the accreted and eroded areas under varying upper levels of water occurrence (indicated by 0% to 75% in Figure 15), the areas change differently in different years. For example, the variation in accreted area for varying water occurrence was more drastic during 2011-2012 in all reaches. This implies there was a larger change in the upper part (i.e., formation of high-amplitude sandbars) but less in the deeper part. For some other years the variation in accreted areas was more uniform for varying ranges of the upper and lowest water levels.
Figure 15. Annual changes in eroded (negative) and accreted (positive) areas within selected reach (left) from 2010 to 2019 (e.g., 2019-2018 corresponds to subtraction ier (2019-2020) — Layer (2018-2019)) for different threshold values of water occurrence frequency for the (upper layer) (0%, 25%, 50% and 75%, respectively, as shown in (upper left plot)). than erosion during 2014-2015, 2016-2017 and 2017-2018, whereas more erosion occurred in 2015-2016. As can be seen from the variation in the accreted and eroded areas under varying upper levels of water occurrence (indicated by 0% to 75% in Figure 15), the areas change differently in different years. For example, the variation in accreted area for varying water occurrence was more drastic during 2011-2012 in all reaches. This implies there was a larger change in the upper part (i.e., formation of high-amplitude sandbars) but less in the deeper part. For some other years the variation in accreted areas was more uniform for varying ranges of the upper and lowest water levels.
Figure 16. Sedimentation and erosion during 2018-2019, simulated by the morphological model (left plot) and estimated as the difference between measured bed topography (projected on model grid) of 2018 and 2019 (right plot).
Figure 16. Sedimentation and erosion during 2018-2019, simulated by the morphological model (left plot) and estimated as the difference between measured bed topography (projected on model grid) of 2018 and 2019 (right plot).
Figure 17. Simulated sedimentation and erosion during 2019-2020 (left plot) and 2020-2021 (right plot).
Figure 17. Simulated sedimentation and erosion during 2019-2020 (left plot) and 2020-2021 (right plot).
Related topics:
Environmental ScienceGeosciencesRiver Management
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved