Assessment of dredging impact on hydrodynamics of Surma River using hydrodynamic model: HEC-RAS approach

A flood is a type of natural catastrophe that happens when water spills into typically dry ground. Heavy rainfall, sediments, quick snowmelt, storm surges, or the collapse of a water containment device like a dam or levee are just a few of the variables that can result in flooding. Communities can be devastated by floods, which can also result in fatalities and damage to homes, businesses, infrastructure, and agriculture. Bangladesh is a nation that experiences frequent flooding, particularly from June to September when it rains the most (Brouwer et al. 2007).

Bangladesh's rivers have played a crucial role in the country's infrastructure ever since its inception, yet the country is currently facing challenges due to both natural and man-made causes (Uddin & Jeong 2021). Most of the Bangladesh's rivers come from India, which is upstream, and flow through Bangladesh, which is downstream. There are four main river networks in Bangladesh (Islam et al. 2015). They are the Brahmaputra-Jamuna river system, the Ganges-Padma river system, the Surma-Meghna river system, and Chottogram region river system. Surma is one of the principal rivers of the Meghna basin. It is a tributary of the Barak River, which originates on the Naga-Manipur watershed's southern slopes. In the Indian state of Assam's Cachhar district, the Barak splits into two branches. Surma, in actuality, is the northern branch of Barak, which flows west and then southwest to Sylhet (Hossain & Paul 2019). This river drains one of the world's wettest regions (Stefanidis & Stathis 2013). Every year, the huge amount of sediments come from upstream and is deposited downstream of the Surma River (Rahman et al. 2018), which increases the risk of flood and decreases navigability because of the deposition of the sediments. A number of districts in the Sylhet division, notably Sunamganj, Sylhet, and Moulvibazar, saw catastrophic flooding when the Surma River rose to its greatest level (Quader et al. 2023). Given that crops and fields were inundated with water, the flooding was especially disastrous for the many people who depend on agriculture for their livelihoods. To increase the channel size and to remove the sediments, Bangladesh Water Development Board (BWDB) has done dredging on the Surma River.

Dredging is an excavation process that is frequently done underwater, in shallow waters, or in freshwater regions with the intention of collecting sediments from the bottom and removing them to a different location (Vivian & Murray 2009). Ohimain (2012) defines dredging as the removal of sediments from a riverbed with the intention of regulating and modifying the watercourse's function to meet human needs. River dredging can affect the hydrology and geomorphology of a river in a number of ways, the specifics of which depend on the river's structure, sediment quality, dredging technique, the degree to which the floodplain is connected, and the nature of the surrounding environment. Dredging causes a change in hydrodynamic parameters such as river discharge, flow velocity, water level, and channel bed erosion deposition (Rahman & Yunus 2016). Although dredging has the potential to reduce water levels, this effect is highly context specific (Ismail & Samuel 2011). When water levels are lowered, the risk of flooding in a specific area is immediately mitigated. The other purpose of dredging is to clean out a river channel so that it can better carry water (Jain & Singh 2003).

Some of the relevant literature discussions are described in the following paragraph. Gibson et al. (2017b) discussed different sediment features of HEC-RAS. To undertake hydrodynamic modeling in the Brazilian Taquaraçu River, Prata et al. (2011) employed bathymetric data. They proposed numerous intervention approaches to alleviate the food. Gibson et al. (2017a) created an HEC-RAS model to demonstrate the potential for using sustainable dredging or sand mining as a method that can be used as part of a flood mitigation plan. This method will significantly raise flood levels, but it will do so during the normal flow, which would not cause drastic changes that would harm the area's existing flora and fauna. Gibson et al. (2017a) provide an overview of the creation of the Puyallup and White River Sediment models, explain the evaluation and selection of the calibration parameters, and show how the use of multiple calibration evaluation metrics (such as volume change and bed gradation) prevented any possible errors. Nelson (2020) created an HEC-RAS model to forecast upcoming dredging operations. The final model created as a consequence of this study may be helpful in estimating future channel maintenance requirements over this stretch of river. The main objective of the article by Haghiabi & Zaredehdasht (2012) was to locate significant erodible spots and regions with potential for silt aggregation along the Karun River.

Dredging can potentially reduce overbank flooding and improve the hydraulic efficiency of river channels, but its effects on flow hydrodynamics and the overall flooding regime need to be better understood. The aim of the study is to analyze the hydrodynamic changes, such as the change of water level, flow area, velocity, and discharge of the Surma River before and after dredging. A HEC-RAS 1D model has been developed for the modeling of the Surma River. After model development, impact assessment of dredging has been observed by the change of hydrodynamic parameters. Flood frequency analyses have been done for 5, 10, 25, 50, and 100 years return periods, and a comparison of water level, velocity, and discharge has been done before and after dredging.

From Figure 1, it is seen that in these locations, the stations of the BWDB have been found as in Table 2.

• The overarching objective of the study is ‘to investigate the impact of dredging on hydrodynamic characteristics as in flow, velocity, the water level of Surma River using state-of-the-art technology and justify the efficiency of proposed dredging locations’.

• To analyze the navigability of the river.

The science of hydrodynamics focuses on the motion of liquids, particularly water (Dey 2014). It is a field of science that examines fluid dynamics, particularly the motion of incompressible fluids.

The set of equations (Khalfallah & Saidi 2018) that describe the motion of fluids, which are derived from Newton's laws of motion, provide the basis of computational hydrodynamic models. These equations describe the action of force applied to the fluid or the changes in flow that ensue. Newton's second law, or the property of conservation of momentum, states that acceleration is dependent on applied force and inversely proportional to mass.

This is the equation of momentum, where A is the cross-sectional area, Q is discharge, S is the frictional slope, z is the water depth, x is the distance along with the flow, S_f is the friction slope, S_c is the bed slope, ϕ is the fraction to determine channel versus floodplain discharge, and t is time.

The following diagram depicts the steps of the methodology:

To support hydrodynamic modeling, secondary data as well as primary data have been collected from different sources. In connection with this study, the collection of data was from BWDB from 1990 to 2020. These data included cross section, water level, and discharge. Table 3 shows the list of datasets collected with their sources:

There are several types of distribution models (Rahman et al. 2013) that can be included in the Generalized Extreme Value, Gumbel or Extreme Value type 1 (EV1), Log-Normal, and the Log Pearson type III distributions. Gumbel distribution is effective for smaller sample sizes than 50 record data (Cunnane 1989). Of the four methods, Gumbel Distribution (EV1) gives gene expression programming, R² test value of 0.999, which is better than other frequency analyses (Onen & Bagatur 2017). Historical discharge data of Sylhet Station were used for frequency analysis. These 30 (1990–2020) years of monthly discharge data will be taken into consideration for frequency analysis and used to determine the different return period discharge information (i.e., 5, 10, 25, 50, and 100 years).

To perform the 1D hydrological modeling, HEC-RAS 5.0.7 is downloaded from its official website (https://www. hec.usace.army.mil/software/hec-ras/). The terrain data of the study area are downloaded from the earth data provided by USGS (https://earthexplorer.usgs.gov), followed by downloading the projection file of the DEM (Digital Elevation Model) from the spatial reference website (https://spatialreference.org). With the help of the preobserved data from the RAS mapper tool, this software has geospatial capabilities for digitizing bank lines, river centerlines, cross-sections, and channel flow lines (river geometry data). The Landsat 8–9, collection-2 level-2 imagery is used in this study, and the study area is extracted from the DEM. This digitization process had been earlier performed in ArcGIS with HEC-GeoRAS extension. This modeling application allows performing both 1D and 2D hydraulic modeling.

By using these files as input files in the RAS mapper, a new terrain is created by clicking the ‘create new terrain’ option available under the ‘Project’ menu in the RAS mapper window. This feature links the DEM to the Google Satellite Image, which then precisely incorporates the DEM into the study area. To build the proper steady-state model of a river in HEC-RAS, the model requires other crucial data. The following tools are used in the processing of the geometry data for the study area: centerline, banks, flow-track lines, and cross-sections of streams. These tools give key geometric data such as the river reach, left and right over the bank, flow path river station, cross sections, and main channel bank stations. The river stationing is then drawn downstream to upstream locations. The river system can also be modified by adding new points or shifting existing ones within the river's reach. After drawing the river system, the cross section data for the stationing points are added to a system editor.

The conditions or phenomena at the model's boundaries are referred to as boundary conditions (Bocquet & Barrat 2007). The boundaries of the one-dimensional model are the downstream water level and the upstream discharge. The boundary conditions for the model were calibrated using the observed data of water level for the year 2011. For calibration and validation, the boundaries of the upstream discharge and downstream water levels are employed to simulate unsteady flow.

Calibration is checked by two processes, the first one is by statistical analysis and the last one is by graphical representation of simulated and observed values. Statistical analysis is done by Nash–Sutcliffe efficiency (NSE) and R². By changing Manning's roughness coefficient ‘n,’ the water level has been calibrated using data from the 2011 flood year. The daily hydrograph was used to simulate the model for a whole year, from May to August. Subsequently, various values have been employed to demonstrate their suitability for simulating flow in the Surma River. Then comparison has been done between the observed and simulated stage hydrograph of the Sylhet Sadar (RMS29) gauging station shown on Figure 2. Finally, Manning's ‘n’ value of 0.029 has been rectified for the main channel. NSE and the coefficient of determination R² for unstable calibration have been found to be 0.9799 and 0.9706, respectively, indicating that the simulated value is closer to the observed value (Legates & McCabe 1999).

Model validation involves testing a model with a dataset representing ‘observed’ field data. This dataset represents an independent source different from the data used to calibrate the model. The model is validated using data from September to December of the 2011 flood year. Then comparison has been done between the observed and simulated stage hydrograph of the Sylhet Sadar gauging station shown on Figure 3. NSE and the coefficient of determination R² for unstable calibration have been found to be 0.9953 and 0.99, respectively, indicating that the simulated value is closer to the observed value.

The results of 5, 10, 25, 50, and 100 years return period flood frequency analysis based on maximum flow recorded upstream from the year 1990 to 2020 using the Gumbel Distribution (EV1) method are presented in Table 4.

Dredging hydrodynamics impact on Surma River is checked by the stations, which are located in the study area, and there are six stations whose impact has been checked. The stations are given in Table 5.

After calibration and validation, discharge was checked before and after dredging, by the unsteady flow analysis in the years 2011 and 2020. The comparison is given in Figures 34–38.

From Figures 4–38, it is clearly observed that, for the steady flow analysis, after dredging, the water level for 5, 10, 25, 50, and 100 years has decreased in a drastic manner. Impact checking is also done on the stations RMS14 and RMS21, which are not located in the dredging area, and it is seen that the water level of that station also decreases, which decreases the risk of flooding. On the dredging stations if the dredging has not been done, then in that area there was a high risk of flooding, but due to dredging, the water level goes down, which is seen in comparison from Figures 34–38. Dredging reduces the risk of flood, and the last station (RMS15) water level goes to a negative value for 5 and 10 years. Due to the deposition of sediments, velocity decreases, but after dredging, velocity increases, which is seen in Figures 39–43. The velocity of the first station (RMS 20) increases highly compared to other stations. Due to widening, the flow area for all stations for all years decreases, without gauging station (RMS18), and here, some amount of flow area increases after 5-year return period.

Many researchers have done 1D modeling of the Surma River. But the impact assessment of dredging has not been done earlier. The originality of these models is that all data taken from BWDB after model build-up calibration and validation have been done to check the model. This model is very innovative by which exact dredging location can be found, and predication of flood can also be determined. Bangladesh is a developing country with a low economy. This model has several benefits.

Design of flood control structures: Floodplain modeling can be used to determine the suitability of building flood control structures for preventative objectives (i.e., embankments, detention ponds).

Nonstructural approach to risk reduction strategy: This study can serve as the foundation for the design of nonstructural flood protection systems. Based on the results of this study, policies like floodplain zoning and river boundary lines can be planned. This can also aid in the design of evacuation and relief routes and the storage of equipment and supplies for emergency flood relief.

Sunamganj, Sylhet is a flood-prone area, and due to the deposition of sediments, the risk of flood increases. From the result, it is clearly understood that due to the dredging water level decreases, which decreases the risk of flood. So it is very important to do dredging at that location, but dredging has some environmental impact also, which are discussed as follows:

Sedimentation: Dredging may result in sedimentation, which can damage aquatic habitats by degrading the water's quality and obstructing light. Limiting their food supplies and interfering with their breeding cycles can have an impact on fish and other aquatic species (Erftemeijer et al. 2012).

Habitat destruction: Dredging can also ruin habitats for many types of plants and animals that live in and around the water. Critical ecosystems including seagrass beds, coral reefs, and wetlands, which act as spawning and feeding grounds for fish and other wildlife, can be damaged or removed by dredging (Okoyen et al. 2020).

Water quality: Pollutants that are released into the water as a result of dredging may harm aquatic life and alter its quality. For instance, dredging may cause the discharge of heavy metals, toxic compounds, and other toxins into the water, which may result in a decline in water quality and have adverse impacts on human health (Zhang et al. 2010).

Ecosystem disruption: Dredging has the potential to completely alter ecosystems, affecting food chains and the cycling of nutrients, which may result in species extinction or population decreases. In coastal environments, the removal of sediments may also modify the ratio of freshwater to saltwater, which may have additional effects on aquatic life (Thrush et al. 1998).

Noise pollution: Dredging activities can cause a lot of noise pollution, which is bad for animals like marine mammals who need sound for communication and navigation (Wenger et al. 2017).

Overall, dredging can have a large, long-lasting impact on the environment, which calls for thorough assessment and management to reduce harm to the environment.

This study presents a systematic way to assess the dredging impact of the Surma River using hydrodynamic models and Geographic Information System (GIS). The effectiveness of each dredged bathymetry was assessed by comparing the predredging and postdredging values of the corresponding simulated parameters. The change in discharge with time is also demonstrated in this study by using the unsteady flow analysis. This method's main tools are a one-dimensional numerical model called HEC-RAS and ArcView GIS for spatial data processing.

• Manning's n is used as a calibration parameter in the HEC-RAS module to analyze the river Surma's hydrodynamics.

• There is an increasing trend of discharge at all dredging-impacted stations. Flood frequency analysis is used to evaluate various return periods' predicted maximum flow for flood extent.

• Among the four distributions, the Gumbel distribution (EV1) is the best distribution for flood frequency analysis in this station. It is observed that flood frequency analysis by Gumbel distribution (EV1) discharges are 2,034.747, 2,233.649, 2,484.963, 2,671.402, and 2,856.465 m³/sec for 5-, 10-, 25-, 50-, and 100-year return periods, respectively.

• Water level and flow area are decreasing, and velocity is increasing, which means the flood risk is decreasing and the waterway is becoming more navigable.

All relevant data are available from an online repository at: http://www.hydrology.bwdb.gov.bd/.

The authors declare there is no conflict.