Geographically, local canals in Sylhet city, Bangladesh, mostly transport the rainwater to the Surma River as the city lacks typical drainage infrastructure for runoff control. Hence, proper hydraulic and hydrologic models are required to assess the current potential of these canals to withstand significant runoff and enhance the protection of flood problems during a severe storm. In this study, Malni Chara and Goali Chara sub-systems of Sylhet city's major drainage networks were calibrated and verified using the EPA Storm Water Management Model (SWMM), respectively, using the hydrological data from 2016 to 2019 and the meteorological data from 1975 to 2019. The models were suitable for measuring the runoff quantity since the simulated results matched the observed data well. For the Malni Chara sub-system, R2 of 0.94, Nash–Sutcliffe efficiency (NSE) of 0.92, d of 0.97, percent bias (PBIAS) (%) of 2.96%, and RMSE standard deviation ratio (RSR) of 0.05 have been found and for the Goali Chara sub-system, R2 of 0.96, NSE of 0.90, d of 0.93, PBIAS (%) of 1.54%, and RSR of 0.08 have been found.

  • The innovative SWMM-based flood model for Sylhet, Bangladesh, addresses data scarcity and serves globally.

  • Unique sub-system analysis improves drainage capacity accuracy.

  • The robust SWMM model calibration and verification ensure reliability.

  • Canal response to the 25-year design rainfall informs urban planning.

  • Transferable findings benefit data-scarce flood-prone areas worldwide.

Bangladesh has a low-lying landscape with more than 230 waterways, making it one of the most vulnerable disaster-prone countries in the world. As the lowest riparian in a huge transboundary river basin, Bangladesh faces an increasing threat of massive flood exposure. Though flood is a natural event and cannot be avoided, the damages can be reduced by responding appropriately and providing necessary drainage channels. However, urbanization and changing demographic features of urban cities have increased the exposure to urban flood hazards (Zhou 2014). Developing countries like it are facing the wrath of this kind of disaster, as the vulnerability level is very high across these regions due to poor socio-economic conditions and haphazard settlements (Khadka & Bhaukajee 2018). The nature of urban flooding is different from riverine flooding due to its hazardous characteristics, including environmental, social, and technical aspects. The flood peaks in urban areas increased from 1.8 to 8 times due to faster flow time (Rangari et al. 2016). So, the increasing trend of urban flooding is now drawing the attention of urban planners worldwide.

Due to the climatic and geological conditions, Sylhet city faces a large amount of rainfall each year (Hasnat et al. 2019). This heavy rainfall generates a large amount of runoff flow, which is evacuated through the channels (locally termed ‘charas’) linked to the Surma River (Munna et al. 2018). These natural channels act as the main drainage network system of the city. However, due to rapid urbanization, improper management, and siltation, these drainage canals within the catchment no longer could convey as much water as they formerly did within their active domain (Ghosh & Mistri 2015). Since the subsidiary natural channels are the main media to carry out the runoff to the river, it is necessary to know whether the existing cross-sections of these channels can carry out heavy runoff and mitigating the flood problem at the time of a heavy shower. This results in the inefficiency of drainage systems and canals and increases the risk of flooding and waterlogging (Ten Veldhuis 2010). In addition, the sewage is disposed of directly to these charas, which reduces the depth and deteriorates the water quality. Such inadequacy in the drainage systems spreads several types of diseases and increases the mortality rate. A defective stormwater management system contaminates the limited freshwater resources and pollutes the environment, whereas a sound management system can minimize the losses if the priorities and the choices are set up.

Watershed development and increased impervious surface development led to decreased infiltration capacity and increased runoff velocity. These circumstances effectively improve the efficiency of water transport to rivers. As rainfall and stormwater runoff from urban areas cause river pollution and watershed impairment, it is essential to integrate drainage system routing into these single and multiple storm event models (Tsihrintzis & Hamid 1998). Many hydrological models are used, such as Storage Treatment Overflow and Runoff Model (STORM), Technical Release Model (TR - 20/55), Hydrologic Engineering Centers River Analysis System (HEC-RAS), Storm Water Management Model (SWMM) and many more. However, studies on urban flood modeling in Bangladesh are limited. Rahman et al. (2014) assessed the drainage system of Chalna city, considering the city's response to hydrology analytically. They included a digital elevation map, satellite picture, and land use for comprehending the city's drainage network (Rahman et al. 2014). Khan (2015) developed a guideline for urban flooding and stormwater drainage for Mymensingh. Afrin et al. (2021) applied Hydrologic Engineering Centers River Analysis System (HEC-RAS) to find the peak runoff considering catchment delineation and the future land use of Dhaka. Kumar & Bhagavanulu (2007) studied an inundation map for a city on the Adyar River's bank. Hossain et al. (2016) performed a study that focused on designing a stormwater drainage system specifically for Sylhet Agricultural University, considering the frequent waterlogging caused by blockages in the existing drainage system, often resulting from sediment and solid waste accumulation (Hossain et al. 2016).

Previous studies have not addressed the development of a rainfall–runoff model specifically for the local canals (charas) within the Sylhet City Corporation (SCC) area. This gap is significant for performing capacity analysis of the major charas, which are crucial in managing most of the stormwater. Furthermore, there is a lack of understanding of the current conditions under various storm scenarios and existing land-use patterns. Notably, no prior research has simultaneously modeled two of the most critical sub-systems: the Malni Chara and Goali Chara sub-systems. These canal sub-systems are pivotal in the urban flooding crisis, often exceeding their capacity in many parts of the City Corporation area, thereby contributing significantly to the havoc caused by urban flooding.

Although many hydrologic and hydraulic models available can predict the effects of watershed changes and characterize stormwater runoff peaks and volumes (Yan et al. 2013), SWMM has been used in the study. SWMM has the key advantage of running the model using a dynamic routing system and is more user-friendly than the others. Though SWMM is developed to model urban drainage network settings, it can be well suited to model natural watersheds (Jang et al. 2007). In addition, the wide acceptance by the scientific society and engineers, user-friendly interface, and technological advancement make SWMM better suited for use. So, after properly calibrating and validating to provide realistic scenarios of runoff generation, hydrological models may be used to comprehend and assess these channel reactions to future land-use changes and climate changes (Du et al. 2012).

The specific objective of this research has been multi-fold. Firstly, rainfall–runoff models were developed for the significant canals (charas) of Sylhet city. Then, calibration and validation of the formulated hydrologic models were carried out respectively to assess the model performance. However, this study focuses only on the calibration and validation of the wide ranges of charas within different storm conditions and to find the sensitivity of different parameters in developing a rainfall–runoff model for the two significant sub-catchments within the area. Also, the study provides a comprehensive understanding of the stormwater management needs in this rapidly urbanizing region, offering significant insights for urban planners, stakeholders, and policymakers.

Modeling input data

The required input data were collected from different public institutions or websites. All the inputted datasets that have been used are summarized in Table 1.

Table 1

Datasets used for the study

Data typeSource of the dataLocation/WebsitesPeriod of data
Sentinel 2A image United States Geological Survey www.earthexplorer.usgs.gov 2017 
Rainfall data Bangladesh Meteorological Department Sylhet 1975–2019 
Chara discharge data BWDB Sylhet Water Development Board 
Soil map Soil Resource Development Institute www.srdi.gov.bd 2018 
Cross-section and elevation data of the canals (Charas) SCC Sylhet 2018 
Land-use data UDD http://www.udd.gov.bd/ 2018 
Data typeSource of the dataLocation/WebsitesPeriod of data
Sentinel 2A image United States Geological Survey www.earthexplorer.usgs.gov 2017 
Rainfall data Bangladesh Meteorological Department Sylhet 1975–2019 
Chara discharge data BWDB Sylhet Water Development Board 
Soil map Soil Resource Development Institute www.srdi.gov.bd 2018 
Cross-section and elevation data of the canals (Charas) SCC Sylhet 2018 
Land-use data UDD http://www.udd.gov.bd/ 2018 

Study area description

Sylhet and its neighboring areas are geologically located in the Sylhet Trough. Physio-graphically, it falls underneath the category of tertiary hills and comprises a tropical climate having an annual average highest temperature of 23 °C (August to October) and an average lowest temperature of 7 °C (January) and the mean annual rainfall is around 3,963 mm (Roy et al. 2014; Bari et al. 2015).

Drainage of sewers and stormwater to the Surma River relies solely on nine canals that cross the city from different parts of the city (Munna et al. 2018). These canals include Malni Chara, Kalibari Chara, Dhupa Chara, Gavier Khal, Goali Chara, Bhubi Chara, Jugni Chara, and Holdi Chara. Among these charas, the Goali Chara and the Malni Chara directly join with the Surma River. The Kalibari, Dhupa, and Gavier Khal join with the Malni Chara, whereas the Jugni Chara, Bhubi Chara, and Holdi Chara join with the Goali Chara and fall into the river Surma. In this study, the total drainage system of this local canal (charas) has been separated into two sub-systems named ‘Malni Chara sub-system’ (canals that are connected to the Malni Chara) and ‘Goali Chara sub-system’ (canals that are connected to Goali Chara). Furthermore, the Malni Chara sub-system was segmented into 11 sub-catchments, while the Goali Chara sub-system was broken down into 12 sub-catchments following the contour map. The sub-systems, including the sub-catchments are illustrated in Figure 1.
Figure 1

Study area map. (a) Malni Chara sub-system, (b) Goali Chara sub-system, and (c) Contour map of the area.

Figure 1

Study area map. (a) Malni Chara sub-system, (b) Goali Chara sub-system, and (c) Contour map of the area.

Close modal

Approach and modeling

In SWMM, the user must select the infiltration process as well as the flood routing method. For channel flow routing in SWMM, three different types of routing techniques can be used: steady flow routing, kinematic routing, and dynamic routing (Rossman et al. 2008). By modeling the pressured and backwater effects brought on by downstream flow constraints, the dynamic wave routing approach entirely solves the Saint-Venant equations (Rossman et al. 2008; Greenberg 2015). For this reason, we used the dynamic wave routing method (Greenberg 2015). Also, for the infiltration process, the Green-Ampt method was selected.

Model parameterization

EPA SWMM requires three major parameter categories for runoff quantity modeling including the physical catchment characteristics, rainfall, and infiltration data (Niyonkuru et al. 2018). The physical characteristics include sub-catchment area, percentage of impervious area, sub-catchment width, average slope, surface depression storage, and surface roughness (Paterne 2019a). This information has been derived by processing the topographic data using the ArcMAP and drainage data collected from the local authority.

However, the soil map for Sylhet city or the specific soil characteristics for Malni and Goali sub-systems were unavailable. Hence, the default value of Type C soil has been used.

Percentage of the impervious area has been calculated using the existing land-use data provided by the Urban Development Directorate (UDD). The land-use map was georeferenced and imperviousness (%) for every sub-catchment within the study area was calculated. The impervious area was divided by the total area to determine the percentage of imperviousness for each sub-catchment.

For conduit properties, ‘irregular’ shapes for the conduits were inputted using the cross-section editor. The cross-section map collected from the SCC provided the length, station distance, and elevations for the section concerned.

Manning's roughness value (n) was selected from the ASCE manual of practice for gravity sanitary sewer design and construction. As the charas are natural channels with an irregular section, the value 0.07 represents the channel roughness (ASCE 1982; Bizier 1982). Other input parameters are mentioned in Table 2.

Table 2

Input parameter for the sub-catchments

ParametersValueDescriptionSource
N-imperviousness 0.015 Manning's roughness coefficient for the impervious areas of the sub-catchments McCuen et al. (1996)  
N-perviousness 0.40 Manning's roughness coefficient for the pervious area of the sub-catchments 
D-store-imperviousness 1.533 mm Depth of depression storage on the impervious portion of the sub-catchments ASCE (1992)  
D-store-perviousness 5.08 mm Depth of depression storage on the pervious portion of the sub-catchments 
ParametersValueDescriptionSource
N-imperviousness 0.015 Manning's roughness coefficient for the impervious areas of the sub-catchments McCuen et al. (1996)  
N-perviousness 0.40 Manning's roughness coefficient for the pervious area of the sub-catchments 
D-store-imperviousness 1.533 mm Depth of depression storage on the impervious portion of the sub-catchments ASCE (1992)  
D-store-perviousness 5.08 mm Depth of depression storage on the pervious portion of the sub-catchments 

Model calibration

By contrasting the actual field-measured data with the simulated data, the SWMM model was calibrated (Zaghloul & Abu Kiefa 2001). In this study, the surface discharge data were collected from the Bangladesh Water Development Board (BWDB) at five locations (Figure 2) for individual charas to evaluate the performances of the models. The developed model has been calibrated against these discharge data by matching generated runoff patterns against actual rainfall events for a 3-month period from June to August 2016 (Figure 3), as the heaviest rainfall in the area generally occurs within this timeline. The most sensitive parameters for the rainfall–runoff model of the area for both sub-catchments have been identified through a sensitive analysis. A realistic range for each parameter was defined based on literature and local expert knowledge (Wu et al. 2017) (Table 3). After that, the parameters were varied according to the defined range, and the simulations were run. After that, the best fit between the simulated discharge and the actual discharge was achieved using the iterative adjustments of the parameters (Wu et al. 2017; Wu et al. 2021). Finally, the validation was checked for the runoff for the rainfall events of 2019 for the period of July to August.
Table 3

Ranges used for adjusted parameters for calibration

Parameters adjustedRangeReferences
Sub-catchment width 0–30 m Local experts 
Depression storage 1.27–5.08 mm ASCE (1992)  
Imperviousness (%) 35–50 Urban Development Directorate (2015)  
Roughness coefficient of banks 0.010–0.015 ASCE (1992)  
Saturated hydraulic conductivity (mm/hr) 0.06–0.57 Mockus (1964)  
Manning's ‘n’ for main channel 0.04–0.10 ASCE (1982)  
Parameters adjustedRangeReferences
Sub-catchment width 0–30 m Local experts 
Depression storage 1.27–5.08 mm ASCE (1992)  
Imperviousness (%) 35–50 Urban Development Directorate (2015)  
Roughness coefficient of banks 0.010–0.015 ASCE (1992)  
Saturated hydraulic conductivity (mm/hr) 0.06–0.57 Mockus (1964)  
Manning's ‘n’ for main channel 0.04–0.10 ASCE (1982)  
Figure 2

Runoff data collection for the sub-systems (a) Malni Chara sub-system and (b) Goali Chara sub-system.

Figure 2

Runoff data collection for the sub-systems (a) Malni Chara sub-system and (b) Goali Chara sub-system.

Close modal
Figure 3

Rainfal data for June, July and August of 2016.

Figure 3

Rainfal data for June, July and August of 2016.

Close modal

Parameter sensitivity analysis

Model outputs for the year 2016 are observed when the values of each chosen factors are adjusted in 10% additions within a scale of 100%, starting from −50% to +50% (Akdoğan & Güven 2016). By changing the value of one input parameter while keeping the other factors constant during the simulation, the sensitivity analysis was carried out (Niyonkuru et al. 2018). The model output variations resulting from changes in input parameters, the relative sensitivity of the outcomes to the various model parameters is identified, and the ratio of variations is calculated using Equation (1) (Akdoğan & Güven 2016):
formula
(1)
where I is the value of the input parameter, IBC is the value of the input parameter for the base-case scenario, O is the value of the output variable, and OBC is the value of output variable for the base-case scenario (Dubus et al. 2003).

Model efficiency criteria estimation

Nash–Sutcliffe efficiency

The Nash–Sutcliffe efficiency (NSE) (Equation (2)) is calculated as one minus the ratio of the error variance of the modeled time-series divided by the variance of the observed time-series:
formula
(2)
where Qsi is the model simulated output; Qoi is the observed hydrologic variable; is the mean of the data, which the NSE utilizes as a standard to measure the hydrologic model's performance; and n is the total number of observations (Muleta 2012). NSE values range from 0.5 to 1, where 1 shows a perfect model and 0.5 indicates a poor performance from the model.

Root mean square error

The evaluation criteria of root mean square error (RMSE) were used to compare the simulated model output with the observed data (Ghazavi et al. 2017). RMSE is calculated using Equation (3):
formula
(3)
where Qoi and Qsi, respectively, are the observed and simulated data and ‘n’ is the number of observations.

Index of agreement

Willmott (1981) proposed the index of agreement (d) (Equation (4)) to overcome the limitation of R2 described previously (Muleta 2012):
formula
(4)
where Qsi is the modeled output; Qoi is the observed parameter; and is the average of the total observations (Muleta 2012).

The standard range is defined between 0 and 1 for the index of agreement where a value of zero means no correlation at all and a value of 1 means that the dispersion of the prediction is equal to that of the observation (Paterne 2019a).

Percent bias

In 2007, Moriasi et al. recommended percent bias (PBIAS) (Equation (5)) as one of the measures that should be included in model performance reports (Moriasi et al. 2007). PBIAS describes whether the model simulations overestimate or underestimate the observations:
formula
(5)

RMSE standard deviation ratio

The RMSE standard deviation ratio (RSR) is calculated as the ratio of the RMSE and the standard deviation of measured data (Golmohammadi et al. 2014) (Equation (6)). The performance of the model simulation improves with decreasing RSR, RMSE, and NSE:
formula
(6)

Mean absolute error

Mean absolute error (MAE) is a measure of errors between paired observations expressing the same phenomenon (Nasirtafreshi 2022) (Equation (7)). It is thus an arithmetic average of the absolute errors, where is the prediction and is the true value. MAE is expected to be less sensitive to high flows and more sensitive to low flows than NSE and RMS and is expected to describe model performance more evenly (Muleta 2012):
formula
(7)

Sensitivity analysis

In hydrological modeling, understanding the high sensitivity to small environmental changes is crucial for effectively managing water resources. This is evident in Figure 4(a), which demonstrates the variation in runoff estimates in response to changes in different parameters. Notably, flow width (K), imperviousness, and precipitation significantly influence the model's output, highlighting their importance in calibration.
Figure 4

Influences of the parameters in the runoff estimation: (a) output variation vs multiplier and (b) ratio of variation vs multiplier.

Figure 4

Influences of the parameters in the runoff estimation: (a) output variation vs multiplier and (b) ratio of variation vs multiplier.

Close modal

Figure 4(b) illustrates the ratio of variations for different outputs which represents the sensitivity to changes in various parameters. The roughness of conduit has an average sensitivity of 0.55 for runoff while flow width (K) has an influence with 0.83. Imperviousness, conduit roughness, and conduit length have negative coefficients, which means that lowering this input parameter will lead to higher output values (Niyonkuru et al. 2018). From this figure it is clear that the imperviousness percentage, flow width (K), and roughness are the most vital parameters in the Goali Chara sub-catchments and Malni Chara sub-catchments.

Both figures collectively emphasize the substantial impact of impervious percentage and flow width on peak flow during simulations of tropical urban catchment runoff, both in terms of quantity and quality, as per SWMM modeling (Chow et al. 2012; Paterne 2019b). This is in line with Li et al. (2016), who found that conduit roughness and imperviousness percentage greatly affect runoff components, a conclusion also supported by our analysis. Table 4 shows the correlation and the sensitivity class for different parameters that have been used in this study.

Table 4

Correlation and classification of the sensitive parameters

ParameterSensitivity classCorrelation
Area High Direct 
Precipitation High Direct 
Flow width (KHigh Direct 
Imperviousness (%) Medium Direct 
Conduit roughness Medium Inverse 
Conduit length Low Inverse 
ParameterSensitivity classCorrelation
Area High Direct 
Precipitation High Direct 
Flow width (KHigh Direct 
Imperviousness (%) Medium Direct 
Conduit roughness Medium Inverse 
Conduit length Low Inverse 

Model calibration

The model was evaluated for proving its competencies with different performance markers using Equations (2)–(7). The results show that all the performance indicators are within the well-accepted ranges for all events denoting a good fit between the modeled and actual runoff. Figures 5 and 6 indicate the calibration curve for the Malni Chara and Goali Chara sub-systems. Moreover, Table 5 shows the statistical measurement for the model performance indicators, while Table 6 illustrates the standard ranges for the efficiency parameters.
Table 5

Statistical measurement of the model performance

Chara nameCalibration pointPerformance ratingR2RSRNSEMAEIndex of agreement, dPBIAS (%)RMSE
Malni Chara sub-system Very good 0.996 0.036 0.963 0.049 0.999 2.439 0.11 
Very good 0.982 0.048 0.951 0.088 0.996 2.923 0.14 
Very good 0.984 0.097 0.902 0.128 0.996 −9.033 0.15 
Very good 0.983 0.074 0.925 0.072 0.997 1.315 0.14 
Very good 0.995 0.071 0.928 0.069 0.999 −2.819 0.12 
Goali Chara sub-system Very good 0.938 0.049 0.950 0.135 0.988 −1.083 0.12 
Very good 0.997 0.021 0.978 0.046 0.999 1.480 0.13 
Very good 0.998 0.018 0.981 0.042 0.999 2.418 0.16 
Very good 0.998 0.025 0.974 0.040 0.999 −0.251 0.14 
Very good 0.994 0.074 0.925 0.059 0.999 4.069 0.13 
Chara nameCalibration pointPerformance ratingR2RSRNSEMAEIndex of agreement, dPBIAS (%)RMSE
Malni Chara sub-system Very good 0.996 0.036 0.963 0.049 0.999 2.439 0.11 
Very good 0.982 0.048 0.951 0.088 0.996 2.923 0.14 
Very good 0.984 0.097 0.902 0.128 0.996 −9.033 0.15 
Very good 0.983 0.074 0.925 0.072 0.997 1.315 0.14 
Very good 0.995 0.071 0.928 0.069 0.999 −2.819 0.12 
Goali Chara sub-system Very good 0.938 0.049 0.950 0.135 0.988 −1.083 0.12 
Very good 0.997 0.021 0.978 0.046 0.999 1.480 0.13 
Very good 0.998 0.018 0.981 0.042 0.999 2.418 0.16 
Very good 0.998 0.025 0.974 0.040 0.999 −0.251 0.14 
Very good 0.994 0.074 0.925 0.059 0.999 4.069 0.13 
Table 6

Standard range and performance rating for different parameters

PerformanceRanges
RatingRSRNSEPBIAS (%)RMSEMAEIndex of agreement
Very good 0 ∼ 0.5 0.75 ∼ 1 < ±10 0.1 ∼ 0.25 
Good 0.5 ∼ 0.6 0.65 ∼ 0.75 ± 10 ∼ ±15 0.25 ∼ 0.5 Not defined Not defined 
Satisfactory 0.6 ∼ 0.7 0.5 ∼ 0.65 ± 15 ∼ ±25 0.5 ∼ 0.75 Not defined Not defined 
Unsatisfactory > 0.7 < 0.5 > ±25 0.75 ∼ 1.0 
PerformanceRanges
RatingRSRNSEPBIAS (%)RMSEMAEIndex of agreement
Very good 0 ∼ 0.5 0.75 ∼ 1 < ±10 0.1 ∼ 0.25 
Good 0.5 ∼ 0.6 0.65 ∼ 0.75 ± 10 ∼ ±15 0.25 ∼ 0.5 Not defined Not defined 
Satisfactory 0.6 ∼ 0.7 0.5 ∼ 0.65 ± 15 ∼ ±25 0.5 ∼ 0.75 Not defined Not defined 
Unsatisfactory > 0.7 < 0.5 > ±25 0.75 ∼ 1.0 
Figure 5

Calibration curve for the Malni Chara sub-system.

Figure 5

Calibration curve for the Malni Chara sub-system.

Close modal
Figure 6

Calibration curve for the Goali Chara sub-system.

Figure 6

Calibration curve for the Goali Chara sub-system.

Close modal

Continuity error of the models

The continuity error is a critical metric used to evaluate the accuracy and reliability of the SWMM simulation. It represents the difference between the total inflow (including rainfall, runoff, external inflows, etc.) and total outflow (including outflow to downstream system, infiltration, evapotranspiration, etc.) from the system, plus any change in storage within the system over a given period. The continuity error can be calculated with the following formula in Equation (8):
formula
(8)

Continuity errors verify the accuracy of the SWMM's calculations. Two different kinds of continuity errors are calculated: one for flow routing and the other for runoff modeling (Greenberg 2015). These errors are calculated for the system by summing the final storage and total outflow and then subtracting it from the sum of the initial storage and total inflow (Rossman et al. 2008; Greenberg 2015). For maintaining the validity of the model and the system, the continuity error should not exceed 5% (Rossman et al. 2008). The runoff and routing continuity error found for the Malni Chara sub-system is −0.13 and 1.18%, whereas for the Goali Chara sub-system it is −0.15 and 1.49%.

Model validation

The model was validated using the input parameters produced during the calibration procedure. Figure 7 illustrates the observed and simulated runoff values for the two sub-systems. For the Malni Chara sub-system R2 of 0.94 which is close to 1; NSE of 0.92 which is between 0 and 1; d of 0.97 which is close to 1, PBIAS (%) of 2.96%, and RSR of 0.05. For the Goali Chara sub-system, R2 of 0.96 which is close to 1; NSE of 0.90 which is between 0 and 1; d of 0.93 which is close to 1, PBIAS (%) of 1.54%, and RSR of 0.08. The findings demonstrate that all performance metrics are within the permissible range for the events, demonstrating a solid match between the predicted and measured runoff. Research on using low-impact development techniques for canal restoration in urban catchments was carried out using the EPA SWMM, and both the R2 and NSE indices were employed for the model performance evaluation (Paterne 2019a). The R2 values varied from 0.89 to 0.99 for peak runoff, while the NSE values fluctuated from 0.79 to 0.99 (Paterne 2019a). Hence, the models were determined to be suitable for runoff quantity modeling for both Malni and Goali Chara sub-systems.
Figure 7

Observed vs simulated runoff values for the Malni Chara sub-system and the Goali Chara sub-system.

Figure 7

Observed vs simulated runoff values for the Malni Chara sub-system and the Goali Chara sub-system.

Close modal

Limitations

To enhance the accuracy of the runoff rainfall model, it is crucial to have access to high-resolution data with shorter intervals. However, the absence of such detailed data in the study area limited our ability to match the peak flow for the catchments precisely. Additionally, the lack of soil data for the canal area also constrained the model's robustness. Access to these data could have significantly improved the precision of runoff and groundwater recharge estimates. Additionally, for calibration, we utilized runoff data from 2016, and for validation, data from 2019 were used. The model's accuracy would have benefited from a broader range of recorded discharge data. We could only collect discharge data corresponding to rainfall for these two years due to data scarcity. Integrating more extensive recorded data would significantly enhance the model's accuracy and predictive capabilities.

This research focused on calibrating and evaluating the EPA SWMM for stormwater runoff modeling in Sylhet city's main drainage canals. A critical aspect of this study was the division of the total drainage system into two sub-systems and further segmentation into sub-catchments, facilitating precise identification and analysis. The EPA SWMM model parameters were meticulously developed, taking into account the unique characteristics of these sub-catchments, along with the area's topography and drainage network data. A significant part of this study was the parameter sensitivity analysis, which underscored the robustness of the chosen parameters. This robustness was further affirmed by the successful calibration and validation of the model within acceptable performance ranges. The indices used in the study uniformly indicated an excellent fit for the modeled data, reinforcing the model's reliability in simulating the runoff quantity of the city's drainage canals.

The practical implications of this study are significant. The calibrated model is a powerful tool for assessing the performance of drainage systems and crafting flood mitigation strategies. It offers a framework for other urban areas facing similar challenges, highlighting the potential of such models in urban planning and decision-making processes. By applying these findings, urban planners and policymakers can better evaluate the effectiveness of existing drainage systems, identify solutions to flood-related issues, and develop comprehensive stormwater management strategies for the watershed.

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

The authors have no relevant financial or non-financial interests to disclose.

Gulam Md Munna prepared the manuscript. Gulam Md Munna received cooperation from Md Mahmudul Hasan, Ahmed Hasan Nury, and Jahir Bin Alam in performing computation, developing the methodology, and manuscript preparation. Md Misbah Uddin, Mohammad Shahidur Rahman, and Shriful Islam assisted in analysis and draft preparation. Furthermore, the authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.

Afrin
S.
,
Islam
M. M.
&
Rahman
M. M.
2021
Adequacy assessment of an urban drainage system considering future land use and climate change scenario
.
Journal of Water and Climate Change
12
(
5
),
1944
1957
.
Akdoğan
Z.
&
Güven
B.
2016
Assessing the sensitivity of SWMM to variations in hydrological and hydraulic parameters: A case study for the city of Istanbul
.
Global Nest Journal
18
(
4
),
831
841
.
doi:10.30955/gnj.001717
.
ASCE
1982
Gravity Sanitary Sewer Design and Construction
.
American Society of Civil Engineers
,
Reston, VA, USA
.
ASCE
1992
Design and Construction of Urban Stormwater Management Systems
.
American Society of Civil Engineers and Water Environment Federation
,
Reston, VA, USA
.
Bari
S. H.
,
Rahman
M. T.
,
Hussain
M. M.
&
Ray
S.
2015
Forecasting monthly precipitation in Sylhet city using ARIMA model
.
Civil and Environmental Research
7
(
1
),
69
77
.
Bizier
P.
1982
Gravity Sanitary Sewer Design and Construction
.
American Society of Civil Engineers
,
New York
(Issues 60, WPCF Manual Practice No. FD-5. Water Environment Foundation (WEF)). doi:10.1061/9780784409008
.
Chow
M. F.
,
Yusop
Z.
&
Toriman
M. E.
2012
Modelling runoff quantity and quality in tropical urban catchments using storm water management model
.
International Journal of Environmental Science and Technology
9
(
4
),
737
748
.
doi:10.1007/s13762-012-0092-0
.
Du
J.
,
Qian
L.
,
Rui
H.
,
Zuo
T.
,
Zheng
D.
,
Xu
Y.
&
Xu
C. Y.
2012
Assessing the effects of urbanization on annual runoff and flood events using an integrated hydrological modeling system for Qinhuai River basin, China
.
Journal of Hydrology
464–465
,
127
139
.
doi:10.1016/j.jhydrol.2012.06.057
.
Dubus
I. G.
,
Brown
C. D.
&
Beulke
S.
2003
Sensitivity analyses for four pesticide leaching models
.
Pest Management Science: Formerly Pesticide Science
59
(
9
),
962
982
.
Ghazavi
R.
,
Rabori
A. M.
&
Reveshty
M. A.
2017
The effects of rainfall intensity-duration-frequency curves reformation on urban flood characteristics in semi-arid environment
.
Ecopersia
5
(
2
),
1799
1813
.
Golmohammadi
G.
,
Prasher
S.
,
Madani
A.
&
Rudra
R.
2014
Evaluating three hydrological distributed watershed models: MIKE-SHE, APEX, SWAT
.
Hydrology
1(
1
),
20
39
.
doi:10.3390/hydrology1010020
.
Greenberg
S. S. E.
2015
Urban Hydrological Modeling of the Malden River Using the Storm Water Management Model (SWMM)
.
Doctoral dissertation
.
Massachusetts Institute of Technology
,
Cambridge, MA, USA
.
Hasnat
G. N. T.
,
Kabir
M. A.
&
Hossain
M. A.
2019
Major environmental issues and Problems of South Asia, Particularly Bangladesh
. In:
Handbook of Environmental Materials Management
(Hussain, C., ed.). Springer, Cham
, pp.
109
148
.
doi:10.1007/978-3-319-73645-7_7
.
Hossain
M. A.
,
Ishaque
F.
,
Sarker
M. A. R.
,
Ritu
S. P.
&
Hussain
M. F.
2016
Sustainable storm water drainage system design for sylhet agricultural university
.
Journal of the Sylhet Agricultural University
3
(
2
),
271
280
.
Jang
S.
,
Cho
M.
,
Yoon
J.
,
Yoon
Y.
,
Kim
S.
,
Kim
G.
,
Kim
L.
&
Aksoy
H.
2007
Using SWMM as a tool for hydrologic impact assessment
.
Desalination
212
(
1–3
),
344
356
.
Khadka
J.
&
Bhaukajee
J.
2018
Rainfall-runoff simulation and modelling using HEC-HMS and HEC-RAS models : Case studies from Nepal and Sweden
.
TVVR 18/5009. Lund University, Lund, Sweden
, pp.
1
69
.
Khan
D. M. S. A.
2015
Guidelines for flood risk assessment and storm water drainage plan
.
Urban Development Directorate, Mymensingh, 300. Available at: https://msdp.gov.bd/reports/report_drainage.pdf
.
Kumar
S. V.
&
Bhagavanulu
D. V. S.
2007
Flood simulation and inundation mapping of Adyar river: A case study using GIS
.
Disaster and Development: Journal of the National Institute of Disaster Management
1
(
2
),
155
168
.
Li
C.
,
Liu
M.
,
Hu
Y.
,
Gong
J.
&
Xu
Y.
2016
Modeling the quality and quantity of runoff in a highly urbanized catchment using storm water management model
.
Polish Journal of Environmental Studies
25
(
4
),
1573
1581
.
McCuen
R.
,
Johnson
P.
&
Ragan
R.
1996
Highway Hydrology, Hydraulic Design Series No.-2, Pub No. FHWA-SA-96-067. September
.
US Department of Transportation
,
Washington, DC
.
Mockus
V.
1964
National Engineering Handbook
.
US Soil Conservation Service
,
Washington, DC, USA
, p.
4
.
Moriasi
D. N.
,
Arnold
J. G.
,
Van Liew
M. W.
,
Bingner
R. L.
,
Harmel
R. D.
&
Veith
T. L.
2007
Model evaluation guidelines for systematic quantification of accuracy in watershed simulations
.
Transactions of the ASABE
50
(
3
),
885
900
.
Munna
G. M.
,
Alam
J. B.
,
Ahmed
J.
,
Al Mamun
A.
&
Mamun
A. A.
2018
Evaluating discharge capacity of Major Chara's of Sylhet City Using GIS
.
Journal of Water Resource and Protection
10
(
02
),
167
181
.
doi:10.4236/jwarp.2018.102010
.
Niyonkuru
P.
,
Sang
J. K.
,
Nyadawa
M. O.
,
Munyaneza
O.
,
Nyawada
M. O.
&
Munyaneza
O.
2018
Calibration and validation of EPA SWMM for stormwater runoff modelling in Nyabugogo catchment, Rwanda
.
Journal of Sustainable Research in Engineering
4
(
4
),
152
159
.
Paterne
N.
2019a
Evaluating Drainage Systems Performance and Infiltration Enhancement Techniques as Flood Mitigation Measures in Nyabugogo Catchment, Rwanda
.
Paterne
N.
2019b
Evaluating Drainage Systems Performance and Infiltration Enhancement Techniques as Flood Mitigation Measures in Nyabugogo Catchment, Rwanda
.
Rahman
M.
,
Saha
R.
,
Haque
M. M.
&
Hossain
S.
2014
Storm Water Management for Urban Areas of Bangladesh by Analytical & Modelling Approach: A Case Study of Chalna Municipality. March
.
Rangari
V. A.
,
Umamahesh
N. V.
&
Patel
A. K.
2016
Development of Different Modeling Strategies for Urban Flooding: a Case Study of Hyderabad City. Proceedings of National Conference : Civil Engineering Conference Innovation for Sustainability (CEC 2016), 5(Cec), 10th
.
Rossman
L. A.
,
Dickinson
R.
,
Schade
T.
,
Chan
C.
,
Sullivan
D.
&
Burgess
E.
2008
Storm Water Management Model User's Manual Version 5.0
.
Roy
R.
,
Islam
A.
,
Miah
M. H. N.
,
Uddin
M. S.
&
Sikdar
A.
2014
Farmers’ opinion towards conservation and genetic erosion of citrus species at Jaintapur Upazila of Sylhet District in Bangladesh
.
Journal of the Sylhet Agricultural University
1
,
207
212
.
Ten Veldhuis
J. A. E.
2010
Quantitative Risk Analysis of Urban Flooding in Lowland Areas
.
Tsihrintzis
V. A.
&
Hamid
R.
1998
Runoff quality prediction from small urban catchments using SWMM
.
Hydrological Processes
12
(
2
),
311
329
.
Urban Development Directorate
2015
Hydrological Survey Report of the North-Eastern Zone of Bangladesh
.
Willmott
C. J.
1981
On the validation of models
.
Physical Geography
2
(
2
),
184
194
.
Wu
Q.
,
Liu
S.
,
Cai
Y.
,
Li
X.
&
Jiang
Y.
2017
Improvement of hydrological model calibration by selecting multiple parameter ranges
.
Hydrology and Earth System Sciences
21
(
1
),
393
407
.
doi:10.5194/HESS-21-393-2017
.
Wu
H.
,
Chen
B.
,
Ye
X.
,
Guo
H.
,
Meng
X.
&
Zhang
B.
2021
An improved calibration and uncertainty analysis approach using a multicriteria sequential algorithm for hydrological modeling
.
Scientific Reports
11
(
1
),
16954
.
doi:10.1038/s41598-021-96250-6
.
Yan
B.
,
Fang
N. F.
,
Zhang
P. C.
&
Shi
Z. H.
2013
Impacts of land use change on watershed streamflow and sediment yield: An assessment using hydrologic modelling and partial least squares regression
.
Journal of Hydrology
484
,
26
37
.
doi:10.1016/j.jhydrol.2013.01.008
.
Zaghloul
N. A.
&
Abu Kiefa
M. A.
2001
Neural network solution of inverse parameters used in the sensitivity-calibration analyses of the SWMM model simulations
.
Advances in Engineering Software
32
(
7
),
587
595
.
doi:10.1016/S0965-9978(00)00072-7
.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/).