ABSTRACT

Impermeable behaviour of manmade structures and urbanization make surface runoff and flooding more likely, especially in low-lying areas with poor drainage systems. This study aimed to simulate the performance of a sustainable stormwater drainage system designed in Kuergeng Town, Gambella, Ethiopia using EPA software (SWMM 5.1) and remote sensing data. A 30-meter Digital Elevation Model (DEM) and 30-year daily and monthly temperature and precipitation data from the National Aeronautics and Space Administration (NASA), land use/land cover data, and hydrological soil group data were among the important datasets. Using the Thiessen polygon methodology, the total area of 1,280.18 hectares was delineated and divided into six sub-catchments using ESRI software ArcGIS10.3. A 60-year return period obtained from plotting position methods and rainfall depth of 62.15 mm were considered for estimation of the peak runoff (30.31 m³/s) by the Soil Conservation Service Curve Number (SCS-CN) method. The peak runoff of 33.51 m³/s was obtained by simulating the system using SWMM v5.1, which resulted in 10.59% higher than the existing peak runoff. The study concluded that the drainage network should be redesigned considering a safety factor of 10.59% as a correction factor and establishing rain garden in the flood prone area in the town. The results also highlighted the importance of remote sensing data for sustainable stormwater management, especially in the areas with limited onsite data and economic vulnerability.

1 Introduction

Urbanization and climate change are typical challenges in stormwater management introduced in many regions around the world due to the impervious behaviour of man-made structures and unpredictable changes in climate scenarios (Erena and Worku 2018). The urbanized land has a heterogeneous feature most of which hinders the natural flow of stormwater thereby diverging some portion of rainfall intensity to form surface runoff (Hamilton et al. 2021). The precipitation that falls on the impervious surface is converted to runoff as the land is covered with natural or artificial material, but the portion that falls on the permeable surface is dispersed widely in numerous forms; approximately 10% of precipitation is accounted for runoff, while about 50% seeps into the soil to create or replenish groundwater, and the remaining 40% is losses through evaporation and plant uptake (Sundara Kumar et al. 2015). This hydrologic cycle is gradually altering as modernization transforms the natural surface into an artificial surface or feature.

Most stormwater drainage systems adopted in urban cities in Ethiopia are conventional drainage systems with a lower number of bio-retention ponds, which are designed with local standards, but not servicing as per the expectation of the residents or designers due to their incapability to handle the peak flow rates unaccounted for during the design phase and the siltation of the drainage section (Adisu and Hailemikael 2017; Dagne 2020). This issue prompts the stormwater drainage to dilapidate or even out of service within a limited time frame if not frequently maintained or redesigned (Zhou et al. 2021). Frequently maintaining drainage is a waste of money which may not be affordable in low-income regions. This study has hypothesized the fundamental cause of the malfunctioning of stormwater drainage systems as an underestimation of the drainage components.  

2 MATERIALS AND METHODS

2.1 Description of the study area

Gambella is the ninth region in Ethiopia categorized under a tropical rainforest climate (Seleshi and Zanke 2004). The rainfall pattern of the region is bimodal with high rain beginning from April to September, and low rainfall from November to December, with temperatures averaging at the mid-twenties during nighttime and 35–45°C during daytime (Berta et al. 2015). Gambella is comprised of three zones (Agnuak, Nuer, and Majangir) with five different nations. Lare district is in the northern part of Nuer Zone bounded by Jikow district (West), Itang special district (East), and Akobo district in the south (Figure 1). Lare district is crossed by the Baro River extending from east to west and the Nyandere stream, which is a tributary of the Jikow River from South Sudan (Figure 1). Lare district covers an average area of 685.17 km² in the lowland section of the region with an average monthly rainfall of 146.25 mm (Galgalo 2018). The terrain of Lare district is flat and its altitude ranges from 410 to 430 meters above sea level.

The soil formation of Lare district is black cotton (regur) under relatively high temperatures and low drainage capability. Lare Woreda includes a town known as Kuergeng, with a total population of 42,348. Based on gender; 21,892 are males and 20,456 are females. Based on residents, 11,764 live in the town and 30,584 are rural residents (CSA 2013). The geographical location of the study area is shown in Figure 1, below.

Figure 1 Geographical location of study area.

2.2 Climate data collection procedures

Three portals were considered as the primary sources of climate data for a detailed comparison of their magnitudes to utilize the data with relative accuracy. Data from the Climate Hazard Group Infra-Red Precipitation with Station (CHIRPS), Climatic Research Unit (CRU) and National Aeronautics and Space Administration (NASA) retrieved in July 2022 were considered for comparison. NASA sources of climate data accessed from https://power.larc.nasa.gov/data-access-viewer were considered as potential online sources of climate data, which accesses both monthly and daily rainfall data.

The study conducted by Rodrigues and Braga (2021) defined that NASA POWER is crucial for producing weather data sets when weather stations are unavailable or missing due to the goodness-of-fit with a Normalized Root Mean Square Error (NRMSE) of 11.2% and a mean Normalized Mean Bias Error (NMBE) of 0.1%. The performance of bias-corrected data was also acceptable, with an NRMSE of less than 12.1%. In this regard, NASA data was preferred for this study as an open-source climate data source, which has horizontal resolutions of 0.25^o (Thrasher et al. 2022).

2.3 Comparison of rainfall data and existing seasonal variation of the rainfall

Figure 2 is presented for the comparison of the rainfall variability on a seasonal basis using average annual rainfall data obtained from NASA data sources (1992 to 2021). The existing seasonal variation of the rainfall in the study area demonstrated the existence of rainfall from May to October, and a progressive decline from January to mid-March; this indicated the progressive decrease of the rainfall from autumn to summer, from summer to spring, and from spring to winter, respectively. As shown in Figure 2, the highest rainfall was recorded in summer and the lowest in winter, which shows the strong agreement of the existing scenario with the data presentation.

Figure 2 Seasonal variation of the rainfall.

2.4 Data preparation for intensity duration frequency (IDF) curve

The rainfall data obtained from NASA was aggregated to determine the maximum intensity of the shorter duration in a minute or hours. A short-time disaggregation method mentioned in Equation 1 was utilized due to its focus on the conversion of daily rainfall data into an hourly or sub-hourly basis (Koutsoyiannis 2003).

(1)

Where:

Pt	=	precipitation with time (h),
P24	=	maximum daily rainfall in 24 hours, and
T	=	time (h).

According to the Zimale and Beyene (2017), published by the Ethiopian Road Authority (ERA), the Gambella region is categorized under Group ‘B’ which uses the equation below for data preparation of existing IDF curve development.

(2)

Where:

Td	=	duration (minutes), and
Tr	=	return period (y).

2.5 Return periods and exceedance probability

The Recurrence Interval (return period) of the extreme event was determined by the formula adopted from Ganamala and Sundar Kumar (2017) as indicated by Equations 3 and 4.

(3)

(4)

Where:

Tr	=	return period (recurrence interval),
P	=	probability of exceedance, and
F	=	probability of non-exceedance.

Return period and probability distribution

The peak runoff estimated by the hydrologic model is lowest for return periods between 2 and 25 years, and progressively increases when the return period varies from 50 to 200 years (Adane et al. 2022). According to Mailhot and Duchesne (2010), design of the urban drainage system relies on an increase in the intensity and frequency of heavy rainfall events will most likely produce more evident trends. Design criteria should therefore be re-evaluated in the light of possible variations produced by climate change. An approach is suggested to update the design standards for urban drainage systems.

As shown in Table 1 below, the maximum return period was obtained using the Hazen method (60 years) with a probability of non-exceedance at 98.3%, which indicated the suitability of the return period for anticipation of the maximum rainfall and reduction of the system failure at extreme events. Table 1 shows the results of the ten (10) different plotting position methods for consideration of the maximum return period of all values obtained.

Table 1 Recurrence interval of the maximum rainfall.

S/N	Methods	T (y)	P (%)	F (%)	S/N	Methods	T (y)	P (%)	F (%)
1	Hazen	60	1.7	98.3	6	Weibull	44	2.3	97.7
2	Gringorten	54	1.9	98.1	7	Beard	44	2.3	97.7
3	Cunnane	51	2.0	98.0	8	Chegodayev	42	2.4	97.6
4	Blom	49	2.0	98.0	9	Adamowsky	41	2.5	97.5
5	Tukey	46	2.2	97.8	10	California	30	3.3	96.7

Development of IDF curve using existing IDF curve

The short-time disaggregation method presented in Equations 1 and 2 were considered as a conversion technique to develop new IDF curves and existing IDF curve of the study area, respectively. The IDF curve of seven commonly used return periods (2, 5, 10, 25, 50 and 100 years) with 24-hour storm durations (Figure 3a) corresponded to the existing IDF curve provided by the ERA (Figure 3b). This IDF curve indicated the decline in rainfall intensity as storm duration increased with longer return periods, as shown in Figure 3. According to Bibi and Tekesa (2023), the IDF curve produced a severe intensity and more inundation with a longer return period, which demonstrated the suitability of the IDF curve to the study. The huge difference encountered between these two IDF curves originated from the duration of the data and climate variabilities as the existing IDF curve was developed using old data with low resolution, which resulted in the sudden change of IDF curve relation (Kourtis and Tsihrintzis 2022).

Figure 3 (a) Developed, and (b) existing IDF in the study area.

Analysis of probability distribution methods

A goodness-of-fit test was essential with statistical tools used to assess how well the observed data matched with one of the probability distribution methods, particularly in hydrology for rainfall modeling and frequency analysis. Ghosh et al. (2016) assessed the suitability of monthly rainfall data to one of the probability distribution methods using statistical software (EasyFit v5.6); they ranked the results in ascending order of magnitude and selected the suitable methods based on the ranking order in the software. Likewise, this study considered frequently used probability distribution methods for the determination of the most suitable techniques using EasyFit v5.6. The results indicated the General Extreme Value (GEV) distribution as a good-fit to the data (Table 2).

Table 2 Fit test parameters, results, and summary.

Fitting Results			Kolmogorov-Smirnov		Anderson-Darling		Chi-square
S/N	Distribution methods	Parameters	Statistics	Rank	Statistics	Rank	Statistics	Rank
1	Gen. Extreme Value	K = 0.15164 σ = 8.0955  µ = 22.171	0.09807	1	0.24946	2	0.93331	1
2	Log-Pearson type-III	α = 16.54  β = 0.09702 ϒ = 1.6574	0.09834	2	0.23782	1	1.1478	3
3	Lognormal	σ = 0.38794 µ = 3.2621	0.11164	3	0.3433	3	1.5806	4
4	Normal	σ = 12.387 µ = 28.26	0.12615	4	0.97194	4	1.1415	2

GEV distribution techniques consist of Gumbel, Fréchet, and Weibull distribution techniques referred to as type I, II, and III respectively (Ailliot et al. 2011). Among the GEV family, Gumbel distribution methods were considered for analysis of extreme events due to their ability to focus on the maximum values of the extreme event (Patel 2020; Muça et al. 2019). Table 2 shows the statistical rank of the goodness-of-fit test and the results of each parameter.

Gumbel distribution methods

Based on Nwaigwe (2021), the Gumbel distribution method required these four steps for the estimation of the design flood:

1. Maximum of daily rainfall data obtained from NASA portal were assembled.
2. The statistical parameters like mean (x̅) and standard deviation (σ_x) were determined by Equations 5 and 6, respectively.

(5)

(6)

3. The flood frequency factor (K_t) was determined using Equation 7:

(7)

Where:

T

=

recurrence interval determined from plotting position method.

4. The design flood was determined using Equation 8:

(8)

2.6 Sources and analysis of surface information

The DEM (30 m x 30 m) obtained from the link https://earthexplorer.usgs.gov provided by USGS (retrieved in July 2022) was utilized for the analysis of surface formation of the catchment area.

Catchment area delineation

DEMs of the Lare District were divided into various basins using ArcGIS 10.3 (Figure 4A). Kuergeng Town was in the sub-basin extracted, as shown in Figure 4B. The total coverage delineated was 1280.08 ha, which has been divided into six sub-catchments (A-1 to A-6) with the ranges of 149.73–360.87 ha to maintain the ranges set under NRCS-CN methods (1–800 ha). The area has been classified using Thiessen polygon methods by setting certain points sensitive to flooding (Figure 4). Since the choice of drainage network alignment is influenced by the resting systems of the road (Owuama 2014). The road alignment in the town follows a tree pattern as seen in Figure 4C. The mixed uses indicated in Figure 4C represent a combination of buildings, water bodies, farmlands, and other human activities in the town.

Figure 4 Kuergeng Town delineation.

Stream network in Kuergeng Town

According to Godsey and Kirchner (2014), the patterns and processes of streamlining connections and disconnections in the catchment area manifest the information about surface runoff flow direction in the basins. The interconnection between streamlines in the town leads to the identification of the four possible locations of the rain garden (1, 2, 3 and 4), and six inlet and two outlet points (I, II) as shown in Figure 5 C. These inlets were chosen based on their potential to carry the maximum surface runoff of the catchment. According to the EBCS (2013), the discharged stormwater should be released into an area where it won't damage or cause a nuisance to the environment. Thus, the proposed outlet points were the best possible solution, as the location is a marshy area without any development activities established. Figure 5 (A and B) indicates the procedures for the accomplishment of delineations.

Figure 5 Stream network in Kuergeng Town.

Verification of stream network in the study area

The integration of spatial data is most important for detailed information about infrastructure and strategic guidance on flood management options (Dawson et al. 2020). Six points were considered for the comparison of the results with the inundated zones during a flash flooding incident in September 2022 and integrated with results obtained from the spatial analysis to verify the validity of the stream network and inundation status in the catchment area. As shown in Figure 6, the points considered coincide with the stream network in the catchment. In addition, the layout of the various infrastructures in the town were identified as contributing factors to the inundation due to the blockage of the natural flow path without the proper provision of stormwater flow path. These points require green infrastructure (rain gardens) and excess runoff disposal points to reduce stagnant rainwater and improve the environmental amenities. Figure 6 (A, B, C, D, E and F) shows the points susceptible to inundation and stagnant rainwater.

Figure 6 Susceptible points for flood inundation.

2.7 Topography and geology of the study area

The surface information regarded for this study was divided for appropriate analysis of the land features and allocation of appropriate options for urban environments (Salvan et al. 2016). The analysis of the catchment area’s surface characteristics was carried out as explained below.

Topography

The spatial information of Kuergeng Town was segregated into basins to indicate the presence of low, medium, and highly elevated points in the Town. As shown in Figure 7, more than half of the town has settled on higher terrain, and less than half on the lower terrain (flat). As per Hahmann and Usery (2015), the terrain of equal magnitude was interconnected by contour line. The contour line presented in Figure 7 shows both narrow and wide spaces, which define the low and high elevation profile in the town, respectively. The terrain presented from the contour and elevation categories revealed the topography of the town was inclined toward the southwest, which indicated the possible exit point of the surface runoff.  

Figure 7 Topographic map of Kuergeng Town.

Slope

The topography of Kuergeng Town naturally lacks a mountain, with more flat land features. The analysis of DEM data introduced a detailed generation of surface slope using ArcGIS 10.3 as a tool for analysis and presentation of the results in the form of a map (Figure 8). As per Seyedashraf et al. (2021), the overland flow regime is influenced by surface slope and the available time of concentration. Various studies classified the slope into flat (0–5%), gentle (5–10%), moderate (10–15%), steep (15–30%), and very steep (>30%). Based on these categories, surface slope is significant, with a steep slope of 25% (Jourgholami et al. 2021). As shown in Figure 8, more than 50% of the study area was categorized under steep slope, which would enhance the significant peak runoff by promoting fast conveyance and reduction of the infiltration rate. This slope has significant impact on surface runoff flow rate. Based on the results, the overland flow path defined was appropriate.

Figure 8 Surface slope of Kuergeng Town.

Land Use/Land Cover change

Natural Resources Conservation Service-Curve Number (NRCS-CN) largely requires soil mapping and land use data as input parameters to determine the curve number (CN) (Zhang 2019). As shown in Figure 9, Land Use/Land Cover (LU/LC) was analyzed using supervised image classification techniques through ArcGIS 10.3. Numerous LU/LC were identified and combined into five classes based on their compatibility on the surface runoff generations.

LU/LC, shown in Figure 9, were presented as forest (49.97%), shrub (24.89%), cultivated land (16.46%), build-up land (8.58%), and water body (0.1%). Concerning the capability to runoff generations; forest, shrub, and cultivated land can significantly reduce the runoff by enhancing soil infiltration and preserving some amount for plant growth (Armson et al. 2013). The building and water body in the catchment area were too small but had high potential for surface runoff. Thus, the result obtained demonstrated that 91.32% was covered by pervious surface, which allows for the possible reduction of surface runoff, and the remaining 8.68% indicated a high potential runoff generation.

Figure 9 Land uses land cover map of Kuergeng Town.

Geology

Hydrological soil data developed by USDA-NRCS with means of NASA remote sensor program to guide the predictions of moisture conditions of the soils were obtained from this link (https://daac.ornl.gov/SOILS/guides/Global_Hydrologic_Soil_Group.html) and retrieved in July 2022. Natural Resources Conservation Service (NRCS) has divided the soils into four groups (A, B, C, and D), and four dual groups (A/D, B/D, C/D and D/D) based on their runoff and water holding capacity. USDA-NRCS 2004 described the soil texture and runoff potential of the hydrological soil group as shown in Table 3.

Table 3 Presentation of hydrological soil group on soil texture and related pixel.

HSG (250m)	Soil texture and their descriptions	Runoff potential	Pixel value
A	Sand (>90% sand and <10% clay)	Low	1
B	Sandy loam, loamy sand (50–90% sand and 10-20% clay)	Moderately low	2
C	Clay loam, silty clay loam, sandy clay loam, silty loam, silt (<50% sand and 20–40% clay)	Moderately high	3
D	Clay, silty clay, sandy clay (<50% sand and >40% clay)	High, unless drain	4
A/D	Sand (>90% sand and <10% clay)	High, unless drain	11
B/D	Sandy loam, loamy sand (50–90% sand and 10-20% clay)	High, unless drain	12
C/D	Clay loam, silty clay loam, sandy clay loam, silty loam, silt (<50% sand and 20–40% clay)	High, unless drain	13
D/D	Clay, silty clay, sandy clay (40% clay)	High, unless drain	14

SOURCE: (Ross et al. 2018; USDA-NRCS 2004)

As shown in Figure 10 below, more than half of Kuergeng Town was covered by soil groups having dual characters, and the remaining areas fall into Groups C and D (Figure 10). According to Abraham et al. (2020), Soil group ‘C’ has soil texture ranging between 20 and 40% clay and less than 50% sand with high runoff and low infiltration rates measured in micrometres per second during wet conditions; Soil group ‘D’ has soil texture of more than 40% clay and less than 50% sand with a very-low infiltration rate measured in micrometres per second or less. Dual soil groups C/D and D/D should stand as ‘group D’ due to an undefined variation of physical state and runoff potential (Ross et al. 2018). Thus, the soil groups obtained in the study area were marked as soil under Groups ‘C’ and ‘D’.

Figure 10 Hydrological soil group of Kuergeng Town.

2.8 Estimation of peak flow

The parameters determined using the SCS-CN method generally include Curve Numbers (CN), initial abstraction, direct runoff and surface retention, and soil holding capacity of the catchment. The following equations adopted from Pathan and Joshi (2019) were used for determination of runoff depth, maximum retention, and initial abstraction in the study area.

(9)

(10)

(11)

Where:

Qr	=	direct runoff (mm),
P	=	storm rainfall (mm),
S	=	potential maximum retention of the soil texture (mm),
CN	=	weighted curve numbers under average condition, and
Ia	=	initial abstraction (mm).

The product of area occupied by land use/land cover class (A_i) and assigned curve numbers (CN_i) divided by total area of the watershed (A_t) of each hydrological soil group were calculated differently. The following equation was used for the determination of the weighted curve number (CN_w) under normal conditions.

(12)

The following equation, provided by NRCS, is applicable to the catchment area having an area of 1–800 ha with an average slope ranging from 0.5–64% was adopted due to the haphazard development of the study area.

(13)

Where:

Tc	=	time of concentration (hrs),
L	=	longest length of the catchment (m),
Y	=	average watershed slope (%), and
CNw	=	dimensionless weighted curve number.

SCS Peak runoff was calculated using Equation 14, adopted from Amiry and Mohammadi (2012). This equation was expected to generate reasonable peak discharge as applicable to the ungagged drainage basin.

(14)

Where:

Qpeak	=	peak discharge (runoff) (m3/s),
A	=	catchment area (km2),
Qr	=	runoff depth (mm), and
Tp	=	time to peak (hrs.) estimated from time of concentration (Tc) from Equation 15, as seen below:

(15)

Design of drainage network

The design methods of the drainage section were divided into two parts. The first part was the design of circular conduits, and the second part was the design of a triangular bio-swale located at the entrance and exit point of the flow in a free surface condition.

The following equations were considered for the estimation of the design parameters, which were essential for the design of the non-scouring drainage section. The design parameters include drainage cross-section area, flow depth, flow velocity, and grade of the conduit. These parameters were essentially used as input to both continuity and Manning’s equations with formulae presented in Equation 24 and 25, respectively.

A triangular-bio-swale was designed using the following commonly used equations stated as follows:

1. Depth of flow measured perpendicular to the swale bottom (Y):

(16)

2.  Top width (T):

(17)

3. Wetted area (A):

(18)

4. Wetted perimeter (P):

(19)

5. Hydraulic radius (R):

(20)

6.   Cross-sectional area (A_c) of triangular bio-swale:

(21)

7. Conveyance velocity of swale and conduit:

(22)

Where:

Qn	=	peak runoff (m3/s),
A	=	wetted area of bioswale (m2),
m	=	horizontal side slope of bioswale,
y	=	depth of the bioswale (m), and
b	=	horizontal distance to the center of the triangular bioswale (m).

8. Bottom slope of the conduits using Manning’s equation:

(23)

Where:

V	=	flow velocity (m/s),
R	=	hydraulic radius of the flow (m),
S	=	bottom slope of the swale, and
n	=	dimensionless Manning’s coefficient.

The circular conduit was proposed in premature to a bio-swale, and the cross-sectional area of both sections was designed after the derivation of the Manning’s and continuity equations. The diameter and depth of the conduit were estimated using Equations 24 and 25, respectively.

(24)

(25)

Where:

H	=	elevation difference of the conduit (m),
L	=	length of the conduits (m), and
S	=	bottom slope (%).

2.9 Configuration and simulation of peak flow

Peak runoff from the study area was simulated and analyzed using the Storm Water Management Model (SWMM) version 5.1, developed by the U.S. Environmental Protection Agency (EPA). The height, depth, and length of the drainage elements were considered as adjustable factors during the trial-and-error model calibration process. To increase the drainage network's anticipated peak flow, flow velocity, and slope accuracy, these parameters were iteratively changed (Niazi et al. 2017). The goal of the calibration was to make sure that the simulated outcomes closely matched the hydrologic behaviour that was anticipated given the site's characteristics and design specifications. The model calibration was completed, and the peak runoff simulation was carried out when acceptable parameter values were found. The outcome supported efficient stormwater planning and flood mitigation methods by offering crucial insights into the system's hydraulic performance.

3 RESULTS AND DISCUSSION

3.1 Variability of climate in the study area

The average temperature ranging from 41.72 to 46.33 °C and average rainfall ranging 624.88 to 1414.78 mm were considered for the analyses of the impact caused by urban improvement in the town. Urbanization and other anthropogenic activities on the land surface progressively increased the surface temperature and rainfall depth over time (Mirsanjari et al. 2021; Wu et al. 2021). As shown in Figure 11(a), temperature was increasingly fluctuating throughout the duration considered; the same holds true for Figure 11(b), where rainfall was fluctuating but extremely rising from 2015 to 2020. This indicated the improvement of the town in this specified period. Galgalo (2018) confirmed the significant vulnerability caused by erratic change of climate in the town is associated with the erratic change of the development sector in the given area. Regarding the climate scenario observed from the following two graphs, significant rainfall depth was expected due to ongoing developmental activities in the town.

Figure 11 Variation of temperature and rainfall depth in the study area.

3.2 Estimation of extreme event precipitation

To examine the variability of rainfall data using Gumbel distribution, the frequency variables are prematurely determined prior to the rainfall depth of the catchment (Arvind et al. 2017). Table 4 shows the result of the frequency variable, like mean of the rainfall data, standard deviation, coefficient of variation, and rainfall depth estimated using Equations 5, 6, 7 and 8, respectively, with a 60-year recurrence interval of the extreme event. The calculated depth based on average daily rainfall data was 62.15 mm and utilized for the estimation of peak discharge.

Table 4 Frequency variable of daily rainfall in the study area.

Recurrence interval (T)	Mean (X̅) (mm)	Standard deviation (σx) (mm)	Frequency factor (Kt) (mm)	Rainfall depth (Xt) (mm)
60 years	28.26	12.39	2.74	62.15

Surface runoff estimation using SCS-CN methods

The fundamental factor necessary for SCS-CN techniques, in addition to soil type and antecedent soil moisture condition, is the presence of LU/LC in the catchment area (Kumar et al. 2021). Table 5 shows LU/LC classes and their corresponding coverage in the town estimated using ArcGIS 10.3, which are essential for estimation of the peak runoff in each catchment. These areas were considered to minimize the numerous classes of the catchment, as this coverage complied with SCS-CN standard sizes.

Table 5 Land uses land cover class and their coverage in the town.

LULC class	A-1 (m2)	A-2 (m2)	A-3 (m2)	A-4 (m2)	A-5 (m2)	A-6 (m2)
Forest	683714.2	618246.2	1046977.0	1607681. 9	666078.5	809786.1
Shrub land	557914.51	519336.0	722184.8	1107994.1	506313.4	816891.1
Cultivated land	303556.29	156425.8	150897.8	531329.1	389417.6	190269.6
Buildup	108531.2	203314.1	113676.52	361740.93	376432.5	229986.8
Water body	18361.9	0.0	2428.9	0.0	0.0	1168.4
Total	1672078.0	1497322	2036165	3608746	1938242	2048102

Composite Curve Number Estimate

The Composite Curve Number (CN), ranging from 60 to 67, 67 to 76, and 76 to 100, indicated a low, moderate, and high runoff potential, respectively (Gezahegn and Suryabhagavan 2018). The town's composite CN value, ranging from 76 to 100 (excluding A-3), suggests a high runoff potential. Table 6 presents the direct runoff (Q_r), maximum retention (S), initial abstractions (I_a), and weighted curve number.

Table 6 SCS Parameters under normal conditions.

Sub-catchment area	Area (km2)	CNW	Ia (mm)	S mm)	Q (mm)
A-1	1.67	81	11.95	59.75	22.92
A-2	1.50	83	10.77	53.87	25.08
A-3	2.04	60	33.57	167.83	4.16
A-4	3.61	81	11.76	58.78	23.26
A-5	1.94	82	11.25	56.25	24.18
A-6	2.05	83	10.28	51.40	26.05

Peak runoff estimation using SCS-CN

As shown in Table 7, the peak runoff of the catchments was the lowest on the third sub-catchment (1.65%) and highest on the fourth sub catchment (26.39%), totaling 30.31 m³/s. Curve Number (CN) and Time of Concentration (Tc) likely influenced the minimum and maximum peak runoff observed in these sub-catchments. The result indicated that Tc values complied with the standard (>= 5 minutes) outlined in design manuals by Brown et al. (2013). Table 7 shows the result determined by the equations 13, 14, and 15.

Table 7 Peak runoff under normal condition.

Area	Runoff depth (mm)	CNW	Length (m)	Tc (hours)	Tp (hours)	Qpeak (m3/s)
A-1	22.74	81	5,296.10	7.55	4.53	4.86
A-2	25.08	83	5,312.57	7.17	4.30	5.05
A-3	4.16	60	6,873.14	16.43	9.86	0.50
A-4	23.26	81	7,752.12	10.11	6.07	8.00
A-5	24.18	82	6,664.66	8.77	5.26	5.15
A-6	26.05	83	5,904.16	7.63	4.58	6.75

3.3 Design of drainage network

Overly designed drainage could carry the flow adequately, but it is unaffordable; while underestimated drainage is inadequate to transmit the peak flow, but accomplished at a reasonable initial cost (Godyń 2022). Ramos and Besharat (2021) analyzed the urban flood risk and economic aspect of stormwater management options; the analysis has highlighted the scenario that reduces the peak runoff as an optimal option. The integrated circular conduit with a triangular bio-swale for sustainable performance of the system was designed one-by-one using the above-mentioned Equations 16 to 25.

Geometrical design of lateral conduits

Circular conduits should be effectives when it’s designed at a slope of at least 1% (Brown et al. 2013). The standard value of Manning’s coefficient for concrete varies slightly from 0.011 to 0.015, depending on the lining materials and flow conditions (Butler and Davies 2004). As shown in Table 8, the Manning’s coefficient, slope, velocity, and other hydraulic parameters designed complied with the standard. According to Federal Highway Administration (2024), a higher velocity from 0.9 to 6 m/s is acceptable for stormwater conduits. As indicated from Table 8, the velocity of the conduits varies from 2.28 m/s to 4.49 m/s, which complies with the recommended standard.

Table 8 Primary conduit design results.

Parameter	A-1	A-2	A-3	A-4	A-5	A-6
Peak runoff (m3/s)	4.856	5.052	0.497	8.004	5.152	6.745
Velocity (m/s)	3.646	3.943	2.276	4.132	3.962	4.491
Area (m2)	1.332	1.281	0.218	1.937	1.300	1.502
Diameter (m)	1.303	1.277	0.527	1.571	1.287	1.383
Hydraulic mean depth (m)	0.326	0.319	0.132	0.393	0.322	0.346
Slope (m/m)	0.010	0.012	0.013	0.010	0.012	0.014
Manning’s coefficient	0.013	0.013	0.013	0.015	0.015	0.013
Length (m)	100	100	90	100	100	100
Flow depth (m)	1.00	1.20	1.17	1.00	1.20	1.40

Design of the main conduits

As shown in Table 9, the slope and flow velocity of main drainage lines vary from 1 to 1.4% and from 2.06 to 4.95 m/s, respectively, with depth and length varying from 1 to 1.3 m and 100 m, respectively. Referring to Bloomberg and Strickland (2012) and USEPA (2021), the length and depth of the conduits should vary from 90 to 120 m, and 0.9 to 3 m, respectively. The velocity of the conduit should vary from 0.6 to 6 m/s to prevent sediment deposit and scouring effect of the pipe materials (Bloomberg and Strickland 2012). Therefore, the results of the hydraulic parameters complied with the standard.

Table 9 Velocity and slope of the main conduit.

Parameter	A-3 to A-2	A-2 to A-1	A-1 to A-6	A-6 to O-2	A-4 to O-1
Peak runoff (m3/s)	0.500	5.549	10.405	17.150	13.156
Velocity (m/s)	2.066	3.907	4.724	4.954	4.499
Area (m2)	0.242	1.420	2.203	3.462	2.924
Diameter (m)	0.555	1.345	1.675	2.100	1.930
Roughness coefficient	0.013	0.013	0.013	0.015	0.015
Flow mean depth (m)	0.139	0.336	0.419	0.525	0.482
Slope (m/m)	0.010	0.011	0.012	0.013	0.012
Conduit depth (m)	1.00	1.10	1.200	1.30	1.20

Geometric design of triangular bio-swale

The most suitable section of bio-swale is comprised of two design approaches: fully vegetated and open channel bio-swales (Ekka et al. 2021; Godyń 2022). These approaches are possible for sufficient conveyance and economical design consideration. In this study, the V-shaped bio-swale with a roughness coefficient of 0.03 and hydraulic retention time of 5 minutes lined by native plants was designed as a suitable geometrical section due to its ability to treat and convey surface runoff at reasonable land requirements. Table 10 shows the results of the parameters designed under V-shaped bio-swale.

Table 10 Designed parameters of V-shaped bio-swale.

Parameter	A-1	A-2	A3	A-4	A-5	A-6	O-1	O-2
Peak Q (m3/s)	4.86	5.05	0.50	8.00	5.15	6.75	13.16	17.15
Side slope	6.00	6.00	6.00	6.00	6.00	6.00	8.00	8.00
Flow depth (m)	0.76	0.75	0.35	0.90	0.73	0.80	1.00	1.10
Longitudinal slope	0.01	0.01	0.01	0.01	0.02	0.02	0.01	0.01
Area (m2)	3.38	3.84	0.74	4.86	3.20	3.84	7.26	8.64
Perimeter (m)	9.25	9.12	4.26	10.95	8.88	9.73	16.12	17.74
Hydraulic radius (m)	0.37	0.37	0.17	0.44	0.36	0.39	0.50	0.55
Top width (m)	9.12	9.00	4.20	10.80	8.76	9.60	16.00	17.60
Freeboard (m)	0.38	0.38	0.18	0.45	0.37	0.40	0.50	0.55
Velocity (m/s)	1.40	1.50	0.68	1.65	1.61	1.76	1.64	1.77
Volume (m3)	0.40	0.42	0.04	0.67	0.43	0.56	1.10	1.43
Total depth (m)	1.14	1.13	0.53	1.35	1.10	1.20	1.50	1.65

Figure 12 Geometrical section of the triangular bio-swale.

3.4 Simulation, calibration, and validation procedures of the models

This study considered dynamic wave routing and design standard as routing methods and model validation data, thereby considering the slope, velocity, and peak runoff as a dependent variable with length, depth, and elevation between the junction points as independent variable. The suitability measures of SWMM to conserve the flow in the system were indicated by continuity errors. In this regard, the continuity errors obtained from the study were 0.03%. This error represents the percent difference between initial storage + total inflow and final storage + total outflow for the entire drainage system, but if the percentage errors exceed 10%, then the result should be readjusted (Rossman and Simon 2022).

Hydraulic jump is the sudden change of flow from super-critical flow to sub-critical flow, typically accompanied by energy dissipation, and a turbulent and sharp rise of water depth, but this can be handled using dynamic wave routing. SWMM is commonly used for dynamic wave routing to simulate gradually varied flow and rapidly varied flow (Rossman and Simon 2022).

Prediction of velocity in the systems

The velocity shown in Figure 13 varied from 3.67 to 4.81 m/s, and from 1.25 to 1.81 m/s for conduits and bio-swales, respectively. Jurries (2003) indicated that velocity should be less than or equal to 0.46 m/s (1.5 ft/s) for the water quality design storm, and below 1.52 m/s (5 ft/s) for the peak flow design storm. This was generally presented as 0.91 to 1.83 m/s (3-6 ft/s) depending on vegetation type and erosion control fabric. The minimum and maximum velocity that doesn’t allow sediment deposition and scouring effects to the conduits vary from 0.9 to 5 m/s (Janitz 2003). However, the velocity obtained from both scenarios complied with the design standard of peak flow but relied on peak runoff conveyance for bio-swales.

Figure 13 Velocity variation in the drainage network.

As illustrated in Figure 13, the simulated velocity in the conduit was in increasing order from the third sub-catchment to the first sub-catchment (A-3 to A-1). This was because the conduits that carried flow from these sub-catchments and directed it to the second outlet point were conduit and carried flow with minor loss due to lining materials. In the bio-swale, the simulated velocity decreased slightly toward the first sub-catchment and increased toward the outlet. This demonstrated the possible reduction of flow rate due to lining materials that favour the infiltration of the soil (Charlesworth et al. 2016). Generally, the results obtained ranged from 3.67 to 4.81 m/s in the circular conduits, and from 1.25 to 1.81 m/s in the triangular bio-swale, but within the standard range considered for the design of both sections. This town has no significant ridges and valleys; therefore, installing the conduits at the roadside not only preserves pedestrian passage but also eliminates the need for fencing and the safety risk (Akter 2022).

Slope in the systems

The results presented in Figure 14 show the variation of slope from both the triangular bio-swale and circular conduit in the range between 1 and 2%. The slope of the bio-swale shown in the system varied from 1 to 2%, which maintains the water quality status of surface runoff flowing in the bio-swale (Jurries 2003). City of Dallas (2019) suggested the minimum slope of circular conduit required for self-cleansing velocity must not fall below 0.3%. By comparing the slope with the recommendation, the slope simulated in both sections complied with the recommended standard. Accurately simulated slope is essential for predicting flow behavior and ensuring system performance (Rossman and Simon 2022).

Figure 14 Status of the slope in the drainage system.

Flow performance in the system

The presence of vegetation in the triangular bio-swale reduced the flow rate of stormwater by allowing soil infiltration into the subsurface and trapping the sediment in the drainage (Deletic and Fletcher 2006). Bio-swales can significantly impact the peak flow rates when the morphometric parameters and geological formation of the catchment area favour the hydraulic conductivity of the media underlying the storm runoff flow path (Purvis 2018).

As illustrated in Figure 15, the predicted flow had decreased in the link near the outfall and increased in the bio-swale situated at the entrance link. The reason for the increase in flow rate at the entrance indicated the beginning of the flow before initial abstraction from the plant roots and continuous loss of surface due to environmental influence, as part of the system was designed as a free surface condition in each respective sub-catchment.

Figure 15 Status of the flow in the drainage system.

Rosa et al. (2015) conducted the calibration and validation of LIDs and traditional runoff control; the result indicated the simulated runoff volume was improved to within 5% and 12% of the actual value. As seen in Figure 15, the magnitude of the peak runoff predicted from both drainage sections (conduits and bio-swale) at peak timing of 16 minutes exceeds the estimated peak runoff (33.51 > 30.3 m³/s), which was 10.59% higher than estimated peak runoff. This variation was reasonable and expected due to more comprehensive and conservative adjustment of the input data during simulation process (Rossman and Simon 2022).

The graphical presentation of the predicted peak runoff was compared with the estimated peak runoff; the results exhibited a higher agreement with each other from their respective sub-catchments. This demonstrated that the peak runoff used during the design of the drainage section requires an adjustment safety factor of 10.59% for sufficient performance. The variation between the predicted and estimated peak runoff wouldn’t be severe, as a bio-swale has a higher capability to lessen the upsurge surface runoff and pollutant load of the surface runoff (Alamdari et al. 2017). Figure 15 shows the variation of simulated peak runoff and estimated peak runoff.

Profile plotting of the drainage node

Generally, 21 junction points including two outlets were established and simulated to examine the peak runoff status in the systems using the estimated values as an input to the models. Based on the results, some nodal points were flooding in the system, as presented in Figures 16 and 17. The nodes highlighted in each Figure 16 and 17 indicated the nodes linked with insufficient drainage conduits and channels; in this regard, J-16, J-18, J-19 and J-20 directed to the first outlet point reveals flooding during peak runoff (Figure 16). The flooding on the flow directed to the second outlet point was indicated in J-5, J-6, J-8, J-9, J-11, J-13, and J-14 (Figure 17). All junctions in the middle were flooded, and the remaining were sufficient to handle the peak runoff; this indicated that the conduit was insufficient to carry the peak runoff between each junction point, but sufficiently receiving and releasing the peak flow in the peripheral channels (bio-swale). Referring to Davies (2004), the presence of nodal flooding is induced by inadequacy of the link between each junction point. This can be enhanced by expanding the drainage section or interconnecting the drainage conduits so that flooding situations at the junction points should be optimized (Rangari et al. 2018). Thus, the peak runoff considered during the design of the existing drainage system required an adjustment peak factor of 10.59%.

Figure 16 Profile plot of the nodes directed to the first outfall.

Figure 17 Profile plotting of the nodes directed to second outfalls.

4 CONCLUSION AND RECOMMENDATIONS

4.1 Conclusion

The simulation of a sustainable stormwater management system using remote sensing data and SWMM (Storm Water Management Model) provided crucial information about the viability and potential suggestion of green infrastructure options in flood-prone areas. According to the study, the surface runoff and flooding were greatly exacerbated by the blockage of the flow path due to developmental activities and insufficient stormwater infrastructure.

The mapping of flood-prone areas using ArcGIS 10.3 and remote sensing data has significantly determined the best places for green infrastructure (like rain gardens) and the prediction of future shortcomings in the drainage section.

The sum of the peak discharge estimated from six delineated sub-catchments using SCS-CN techniques was 30.3 m³/s, and the peak discharge simulated in the system using SWMM v5.1 was 35.51 m³/s. In contrast, simulated peak flow was 10.59% higher than the estimated peak flow, which indicated an adjustment factor to consider for future drainage design in the town. The study concluded that remote sensing data in conjunction with hydrological modeling tools are a capable and economical method for urban storm water planning in data-scarce regions.

4.2 Recommendations

To better understand the implication of the results, future researchers and practitioners could address the following:

• This study utilized remote sensing data where the variation of remote and onsite data is imperative; however, to predict the performance of stormwater management systems, further study is needed to reveal the unfolded compatibility of these drainage sections (bio-swale and conduits) using on-site (station) data.
• Flow rate attenuates from triangular bio-swales to circular conduits after the simulation process executions; however, the amount attenuated between each section failed to be quantified. Perhaps future researchers should investigate the rate of change of flow from bioswale to conduits.
• In the state where nodal overflow becomes an issue, the flow from the node should be directed to the nearby bio-swales to maintain the flow circulation during the rainstorm. Regular inspection is equally important to overcome the blockages and ensure adequate grass in the bioswale without woody plant encroachment.
• The vegetation planted on the drainage section should not be mowed below the flow depth and must be drought-tolerant; a feeling of personal ownership is paramount for monitoring all components, which should be achieved through training before operation.
• The swale should be reinforced with rip-rap or turf reinforcement matting, which can sustain a higher velocity regarded as too high for grass cover in a particular swale design and the slope and cross-section.

References

1. Abraham, S., C. Huynh, and H. Vu. 2020. "Classification of soils into hydrologic groups using machine learning." Data 5 (1). https://doi.org/10.3390/data5010002
2. Adane, G.B., A.K. Kassa, A.T. Toni, and S.L. Tekle, S.L. 2022. "Spatial runoff estimation under different land uses and rainfall frequencies: case of flood-prone Dechatu River Catchment, Dire Dawa, Ethiopia." Arabian Journal of Geosciences 15 (11). https://doi.org/10.1007/s12517-022-10360-7
3. Adisu, M. and M. Hailemikael. 2017. An Approach to Drainage System Sustainability in Wolaita Soddo Town: A Case Study from Southern Ethiopia." International Journal of Waste Resources 7 (2). 
4. Ailliot, P., C. Thompson, and P. Thomson. 2011. "Mixed methods for fitting the GEV distribution." Water Resources Research 47 (5): 1–14. https://doi.org/10.1029/2010WR009417
5. Akter, A. 2022. "Stormwater Management." In: Rainwater Harvesting—Building a Water Smart City. Springer Water, 117–163. https://doi.org/10.1007/978-3-030-94643-2_4
6. Alamdari, N., D.J. Sample, P. Steinberg, A.C. Ross, and Z.M. Easton. 2017. "Assessing the effects of climate change on water quantity and quality in an urban watershed using a calibrated stormwater model." Water (Switzerland) 9 (7): 464. https://doi.org/10.3390/w9070464
7. Amiry, M.D. and A.A. Mohammadi. 2012. "Regional model for peak discharge estimation in ungauged drainage basin using GIUH, Snyder, SCS and triangular models." International Journal of Water Resources and Environmental Engineering 4 (4): 86–96. https://doi.org/10.5897/ijwree11.153
8. Armson, D., P. Stringer, and A.R. Ennos. 2013. "The effect of street trees and amenity grass on urban surface water runoff in Manchester, UK." Urban Forestry & Urban Greening 12 (3): 282–286. https://doi.org/10.1016/j.ufug.2013.04.001
9. Arvind, G., P. Ashok Kumar, S. Girish Karthi, and C.R. Suribabu. 2017. "Statistical Analysis of 30 Years Rainfall Data: A Case Study." IOP Conference Series: Earth and Environmental Science, 80, 012067. https://doi.org/10.1088/1755-1315/80/1/012067
10. Berta, A., T. Soromessa, and S. Belliethathan. 2015. "Carbon Sock of Gambella National Park: Implication for Climate Change Mitigation." Advances in Life Science and Technology 35, 41–56.
11. Bibi, T.S. and N.W. Tekesa. 2023. "Impacts of climate change on IDF curves for urban stormwater management systems design: the case of Dodola Town, Ethiopia." Environmental Monitoring and Assessment 195, 170. https://doi.org/10.1007/s10661-022-10781-7
12. Bloomberg, M.R. and C.H. Strickland Jr. 2012. "Guidelines for the Design and Construction of Stormwater Management Systems." New York City Department of Environmental Protection, July 2012. https://www.nyc.gov/assets/dep/downloads/pdf/water/stormwater/stormwater-design-construction-guidelines-2012-final.pdf
13. Brown, S.A., J.D. Schall, J.L. Morris, S. Stein, and J.C. Warner. 2013. Urban Drainage Design Manual Circular 22, Third Edition. National Highway Institute, United States Department of Transportation. https://rosap.ntl.bts.gov/view/dot/54350/dot_54350_DS1.pdf
14. Butler, D. and J.W. Davies. 2004. Urban Drainage,  Second Edition. Taylor & Francis e-Library, Spon Press, New York, NY.
15. Charlesworth, S., F. Warwick and C. Lashford. 2016. "Decision-Making and Sustainable Drainage: Design and Scale." Sustainability 8 (8): 782.  https://doi.org/10.3390/su8080782
16. City of Dallas. 2019. Drainage Design Manual. Dallas, TX.
17. CSA. 2013. Population Projection of Ethiopia for All Regions At Wereda Level from 2014 – 2017 (Issue August 2013).
18. Dagne, H.T. 2020. "Assessment of Stormwater Drainage System for Small Urban Watershed: Case of Shambu Town, Oromia Region, Ethiopia." Civil and Environmental Research 12 (9): 1–8. https://doi.org/10.7176/cer/12-9-01
19. Dawson, D.A., K. Vercruysse, and N. Wright. 2020. "A spatial framework to explore needs and opportunities for interoperable urban flood management." Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 378 (2168). https://doi.org/10.1098/rsta.2019.0205
20. Deletic, A. and T.D. Fletcher. 2006. "Performance of grass filters used for stormwater treatment – A field and modelling study." Journal of Hydrology 317 (3–4): 261–275. https://doi.org/10.1016/j.jhydrol.2005.05.021
21. EBCS. 2013. "EBCS-9 Ethiopian Building Code Standard: Plumbing Services of Buildings" (Issue August). Ministry of Urban Development and Construction, Addis Ababa, Ethiopia.
22. Ekka, S.A., H. Rujner, G. Leonhardt, G.T. Blecken, M. Viklander, and W.F. Hunt. 2021. "Next generation swale design for stormwater runoff treatment: A comprehensive approach." Journal of Environmental Management 279, 111756. 
23. Erena, S.H. and H. Worku. 2018. "Flood risk analysis : causes and landscape based mitigation strategies in Dire Dawa city , Ethiopia." Geoenvironmental Disasters 5, 16. https://doi.org/10.1186/s40677-018-0110-8
24. Federal Highway Administration. 2024. "Urban Drainage Design Fourth Edition. Publication No. FHWA-HIF-24-006. Hydraulic Engineering Circular No. 22." U.S. Department of Transportation, Washington, DC. https://www.fhwa.dot.gov/engineering/hydraulics/pubs/hif24006.pdf
25. Galgalo, D. 2018. "Impacts of Climate Variability and Households Adaptation Strategies in Lare District of Gambella Region, South Western Ethiopia." Journal of Earth Science & Climatic Change 09 (07). https://doi.org/10.4172/2157-7617.1000480
26. Ganamala, K. and P. Sundar Kumar. 2017. "A case study on flood frequency analysis." International Journal of Civil Engineering and Technology 8 (4): 1762–1767.
27. Gezahegn, M.A. and K.V. Suryabhagavan. 2018. "Developing Flood Hazard Forecasting and Early Warning System in Dire Dawa, Ethiopia." International Journal of Advanced Multidisciplinary Research 5 (9).
28. Ghosh, S, M. Kumar Roy, and S. Chowdhury Biswas. 2016. "Determination of the Best Fit Probability Distribution for Monthly Rainfall Data in Bangladesh." American Journal of Mathematics and Statistics 6 (4), 170–174. 
29. Godsey, S.E. and J.W. Kirchner. 2014. "Dynamic, discontinuous stream networks: Hydrologically driven variations in active drainage density, flowing channels and stream order." Hydrological Processes 28 (23): 5791–5803. https://doi.org/10.1002/hyp.10310
30. Godyń, I. 2022. "Economic Incentives in Stormwater Management: A Study of Practice Gaps in Poland." Water (Switzerland) 14 (23): 3817. https://doi.org/10.3390/w14233817
31. Hahmann, T. and E.L. Usery. 2015. "What is in a contour map? A region-based logical formalization of contour semantics." In: Fabrikant, S., Raubal, M., Bertolotto, …(eds) Spatial Information Theory. COSIT 2015. Lecture Notes in Computer Science, vol 9368. Springer, Cham. https://doi.org/10.1007/978-3-319-23374-1_18
32. Hamilton, B., N.C. Coops, and K. Lokman. 2021. "Time series monitoring of impervious surfaces and runoff impacts in Metro Vancouver." Science of the Total Environment 760, 143873. https://doi.org/10.1016/j.scitotenv.2020.143873
33. Jourgholami, M., S. Karami, F. Tavankar, A. Lo Monaco, and R. Picchio. 2021. "Effects of slope gradient on runoff and sediment yield on machine-induced compacted soil in temperate forests." Forests 12 (1): 1–19. https://doi.org/10.3390/f12010049
34. Jurries, D. 2003. "Biofilters (Bioswales, Vegetative Buffers, & Constructed Wetlands) for Storm Water Discharge Pollution Removal." State of Oregon Department of Environmental Quality. 
35. Kourtis, I.M. and V.A. Tsihrintzis. 2022. "Update of intensity-duration-frequency (IDF) curves under climate change: a review." Water Supply 22 (5): 4951–4974. https://doi.org/10.2166/ws.2022.152
36. Koutsoyiannis, D. 2003. "Rainfall disaggregation methods: Theory and applications." Proceedings from Workshop on Statistical and Mathematical Methods for Hydrological Analysis, May, 1–23. https://doi.org/10.13140/RG.2.1.2840.8564
37. Kumar, A., S. Kanga, A.K. Taloor, S.K. Singh, and B. Đurin. 2021. "Surface runoff estimation of Sind river basin using integrated SCS-CN and GIS techniques." HydroResearch 4, 61–74. https://doi.org/10.1016/j.hydres.2021.08.001
38. Mailhot, A., and S. Duchesne. 2010. "Design Criteria of Urban Drainage Infrastructures under Climate Change." Journal of Water Resources Planning and Management 136 (2): 201–208. https://doi.org/10.1061/(asce)wr.1943-5452.0000023
39. Mirsanjari, M.M., F. Mohammadyari, J.S. Visockiene, and A. Zarandian. 2021. "Relationship between land surface temperature and urbanization in Vilnius district." Environmental Monitoring and Assessment 193, 472. https://doi.org/10.1007/s10661-021-09209-5
40. Muça, O., A. Olloni, A. Musafaraj, and A. Kamberi. 2019. "Flood Frequency Analysis Using Gumbel’s Distribution: A Case Study of Komani Basin." Journal of Multidisciplinary Engineering Science and Technology 6 (8): 10532–10535.
41. Niazi, M., C. Nietch, M. Maghrebi, N. Jackson, B.R. Bennett, M. Tryby, and A. Massoudieh. 2017. "Storm Water Management Model: Performance Review and Gap Analysis." Journal of Sustainable Water in the Built Environment 3, 2. https://doi.org/10.1061/jswbay.0000817
42. Nwaigwe, C. 2021. "Log-normal and Log-Pearson Type III distributions." Proceedings of the 43^rd Annual National Statistical Conference 2019., Nigerian Statistical Association.
43. Owuama, C.O., E. Uja, and C.O. Kingsley. 2014. "Sustainable Drainage System for Road Networking." International Journal of Innovation, Management and Technology 5 (2). https://doi.org/10.7763/ijimt.2014.v5.491
44. Patel, M.B. 2020. "Flood Frequency Analysis Using Gumbel Distribution Method at Garudeshwar Weir, Narmada Basin." International Journal of Trend in Research and Development 7 (1): 2394–9333.
45. Pathan, H. and G.S. Joshi. 2019. "Estimation of Runoff Using SCS-CN Method and ArcGIS for Karjan Reservoir Basin. International Journal of Applied Engineering Research 14 (12): 2945–2951.
46. Purvis, R. 2018. "Bioswale Design Optimization for Enhanced Application and Pollutant Removal." Dissertation, Department of Biological and Agricultural Engineering,North Carolina State University.
47. Ramos, H.M. and M. Besharat. 2021. "Urban flood risk and economic viability analyses of a smart sustainable drainage system." Sustainability (Switzerland) 13 (24). https://doi.org/10.3390/su132413889
48. Rangari, V.A., S.S Prashanth, N.V. Umamahesh, and A.K. Patel. 2018. "Simulation of Urban Drainage System Using a Storm Water Management Model (SWMM)." Asian Journal of Engineering and Applied Technology 7 (1): 7–10. https://doi.org/10.51983/ajeat-2018.7.1.872
49. Rodrigues, G.C. and R.P. Braga. 2021. "Evaluation of NASA POWER Reanalysis Products To Estimate Daily Weather Variables in a Hot Summer Mediterranean Climate. Agronomy, 11 (6): 1–17. https://doi.org/10.3390/agronomy11061207
50. Rosa, D.J., J.C. Clausen, and M.E. Dietz. 2015. "Calibration and Verification of SWMM for Low Impact Development. Journal of the American Water Resources Association 51 (3): 746–757. https://doi.org/10.1111/jawr.12272
51. Ross, C.W., L. Prihodko, J. Anchang, S. Kumar, W. Ji, and N.P. Hanan. 2018. "HYSOGs250m, global gridded hydrologic soil groups for curve-number-based runoff modeling." Scientific Data 5, 180091. https://doi.org/10.1038/sdata.2018.91
52. Rossman, L. and M. Simon. 2022. Storm Water Management Model (SWMM) User’s Manual Version 5.2.1. U.S. Environmental Protection Agency (EPA), Center for Environmental Solutions and Emergency Response, Office of Research and Development, Cincinnati, OH.
53. Salvan, L., M. Abily, P. Gourbesville, and J. Schoorens. 2016. "Drainage System and Detailed Urban Topography: Towards Operational 1D-2D Modelling for Stormwater Management." Procedia Engineering 154, 890–897. https://doi.org/10.1016/j.proeng.2016.07.469
54. Seleshi, Y. and U. Zanke. 2004. "Recent changes in rainfall and rainy days in Ethiopia." International Journal of Climatology 24 (8): 973–983. https://doi.org/10.1002/joc.1052
55. Seyedashraf, O., A. Bottacin-Busolin, and J.J. Harou. 2021. "Many-Objective Optimization of Sustainable Drainage Systems in Urban Areas with Different Surface Slopes.” Water Resources Management 35, 2449–2464.
56. Sundara Kumar, P., T. Santhi, P. Manoj Srivatsav, S.V. Sreekanth Reddy, M. Anjaneya Prasad, and T.V. Praveen. 2015. "Storm Water Drainage Design (Case study Vijayawada)". International Journal of Earth Sciences and Engineering 8 (2), 507–511.
57. Thrasher, B., W. Wang, A. Michaelis, F. Melton, T. Lee, and R. Nemani. 2022. NASA Global Daily Downscaled Projections, CMIP6." Scientific Data 9, 262. https://doi.org/10.1038/s41597-022-01393-4
58. USDA-NRCS. 2004. National Engineering Handbook – Part 630 – Chapter 4, Storm Rainfall Depth. United States Department of Agriculture, Natural Resources Conservation Service.
59. USEPA. 2021. "Fiscal Year 2020 Annual Performance Report." Environmental Protection Agency, Washington, DC.
60. Wu, W., L. Ren, Y. Wei, and M. Guo. 2021. "Impacts of Urbanization on Extreme Regional Precipitation Events." Discrete Dynamics in Nature and Society, 2021, 2210184. https://doi.org/10.1155/2021/2210184
61. Zhang, W-Y. 2019. "Application of NRCS-CN method for estimation of watershed runoff and disaster risk." Geomatics, Natural Hazards and Risk 10 (1): 2220–2238. https://doi.org/10.1080/19475705.2019.1686431
62. Zhou, J., A.A. Bloorchian, S. Nassiri, and A. Osouli. 2021. "A Simplified Model for Predicting the Effectiveness of Bioswale’s Control on Stormwater Runoff from Roadways." Water (Switzerland) 13 (20): 2798. https://doi.org/10.3390/w13202798
63. Zimale, F.A. and T.D. Beyene. 2017. "IDF Equations for Similar Rainfall Regions in Ethiopia." Ethiopian Road Authority.